The Public Sphere 2016 Edition by Estúdio Chaleira

JOURNAL OF PUBLIC POLICY I 2016 EDITION

2016

2015-16 Editorial team Genevieve Joy, Editor-in-Chief Samira Noronha, Lead Editor Niranjana Narayanan, Managing Editor

Senior Editors Stephanie Campbell Madeleine Dâ&#x20AC;&#x2122;Arcy Tarini Fernando Julia Hug Ruobing Wang

Contributing Editors Daniel Almeida Dinker Bhardwaj Hana Chambers Zachary Clemence Salimah Kassamali Ashley Lau Mari Magnussen-Landsem Arjun Puri Silvana Rebaza Zachary Rosentzveig Gabriela Smarrelli Jaime Urrego Andrea Villarreal Diana Zambonino

Copyright ÂŠ 2016 by The Public Sphere Journal Design: Cristiane Viana - www.estudiochaleira.com.br Cover photo: Julia Hagen ISSN: 2056-533X

www.publicspherejournal.com

The Public Sphere is a journal of international policy studies produced by postgraduate students at the London School of Economics and Political Scienceâ&#x20AC;&#x2122;s Institute of Public Affairs (IPA). The Public Sphere Journal, now in its fourth year, publishes evidence-based research in economic policy, public management, international affairs, development studies and social policy. The journal offers contributors and readers an active space to rethink, critically analyse and address important policy problems. The Public Sphere welcomes submissions that contribute to vibrant debate from advanced students in relevant degree programs, academics and practitioners. This publication was made possible with the support of the LSE Annual Fund.

CONTENTS

Letter from the editors

Asian carp: The value of understanding the political, social, and economic effects of prospective policies to mitigate a Laurentian Great Lakes invasion Rachel Binder

The new Eurasian land bridge: Opportunities for China, Europe, and Central Asia Gabor Debreczeni

3 4

Tajikistan: Political Economy of drug addiction Harald Edinger

Getting out of poverty for the first time: Evidence from the United Kingdom David Escamilla-Guerrero

The political economy of food subsidies in Egypt: Reforms and strengthening of social protection Flavia Lorenzon

105

Data localization laws in a digital world: Data protection or data protectionism? Neha Mishra

135

Evaluating the strength of commitments made to swing constituencies: An analysis of the Conservativesâ&#x20AC;&#x2122; decision to block Heathrow Expansion Ben Stevenson

159

Determinants of change in the international fight against corruption: Do freedom of press and the executive composition matter? Vincent Zimmer and Britta SchĂźnemann

175

[Letter from the editors]

LETTER FROM THE EDITORS This has been a year of growth for The Public Sphere. The team was the biggest yet, and their motivation, drive and ideas have results that extend beyond this journal. Our fourth edition will be the first to live on our new, dedicated website that brings the sphere into the digital realm. The thoughtful, rigorous debate that happens within these pages will be reflected online on a continuous basis and will give life to the authors and the arguments that are driving policy. The Public Sphere was conceived at the London School of Economics and Political Science by students in the Master of Public Administration, a two-year program that is unique for its studentsâ&#x20AC;&#x2122; diversity of background and work experience and its practical approach. It is only natural and fitting that our product reflects who we are and what we value. This year, we received more submissions than ever before. The eight articles in this journal were the most applied, relevant and captivating topics. You, reader, will be able to learn about food subsidies in Egypt, engage with policy implications of data localisation laws, and finally understand voting dynamics around the Heathrow Airport expansion outside London. Econometric analysis will deepen your understanding about poverty in the United Kingdom and international corruption. And then we will push you to think about topics that do not enter the everyday sphere like drug addiction in Tajikistan, the New Eurasian Land Bridge, and invasive Asian carp in the Great Lakes of the United States. As editors, we facilitate. We select articles based on relevance, novelty and interest; we edit to help authors clarify their thoughts and drive their points home; and we publish in the hopes that people like you will learn something new or be forced to question old beliefs. This yearâ&#x20AC;&#x2122;s developments have allowed our form to

THE PUBLIC SPHERE | 2016 ISSUE

follow function, thanks to generous funding from the LSE Student Union Annual Fund. Our new branding, journal format and website are vessels to help The Public Sphere reach as many minds as possible. Turn the pages ready to learn. Ready to disagree. Ready to walk away thinking about our lifeblood – policy. Thank you for reading. We hope you enjoy. Genevieve, Samira and Niranjana

THE PUBLIC SPHERE 2016 ISSUE

Asian carp The value of understanding political, social, and economic effects of prospective policies to mitigate a Laurentian Great Lakes invasion Rachel Binder

MA in Public Policy and Administration, Class of 2015 Northwestern University

ÂŠjunrong

ABSTRACT Bighead (Hypophthalmichthys nobilis) and silver Asian (Hypophthalmichthys molitrix) carp are invasive species that threaten to populate the Laurentian Great Lakes, located on the United States-Canada border in northeastern North America. Elected officials, at the urging of Laurentian Great Lakes stakeholders, are debating eight possible policy alternatives to stop an Asian carp invasion. The alternatives have been justified scientifically; however, the political, social, and economic impacts of the alternatives are yet to be articulated. This study will discuss these impacts using a cause and effect analysis of each policy alternative. The results of this analysis show that the policies currently being considered, although viewed as scientifically effective, are not politically, socially, or economically feasible. There are doubts about scientific certainty, varying citizen opinions on alternatives and little to no detail about who will pay for policy implementation. Elected officials must gather credible data regarding political opportunities and obstacles, public opinion and economic externalities to recommend and implement the best policy alternative.

[ Asian carp ]

The Laurentian Great Lakes (the Great Lakes) are a system of five lakes bordering eight states within the United States and Canada, and the largest bodies of freshwater in the world. They provide drinking water for 40 million people, a home for 250 species of fish, a clear passage for 200 million tons of commercial goods annually and a US$4 billion sports fishery industry (GLERL, 2014). In the last decade, scientists have warned that the services citizens depend on the Great Lakes to provide could be jeopardised by a serious, imminent threat â&#x20AC;&#x201C; an invasion of bighead and silver Asian carp. Stakeholders of the Great Lakes, including politicians, citizens, environmental committees and trade organisations, agree that a US government policy to prevent an Asian carp invasion is an urgent necessity. The chosen policy will target the Chicago Area Waterway System (CAWS), the geological area where an Asian carp invasion will likely occur, and will include several structural and non-structural control methods. The US Army Corp of Engineers released a study in 2014 outlining the eight most realistic policy alternatives. Stakeholders disagree about which policy should be implemented. The 2014 study evaluated the alternatives on scientific merit alone and did not consider the political, social, and economic effects that each prospective policy would have on the Great Lakes and its users. The lack of comprehensive and well-organised data about each policy alternative is a key reason why a policy to prevent an Asian carp invasion of the Great Lakes has not been chosen. Conducting a cost/benefit analysis of each policy alternative that includes social, political and economic effects would allow stakeholders and elected decision-makers to better understand the strengths, weaknesses, opportunities and threats of each alternative, and perhaps expedite the choice and implementation of an official U.S. policy. This study provides a clearer description of prospective policy alternatives for effectively mitigating Asian

THE PUBLIC SPHERE | 2016 ISSUE

carp in the Great Lakes by incorporating social, political and economic aspects not previously considered.

BACKGROUND Bighead and silver Asian carp currently inhabit the Upper and Lower Mississippi River basins of the southern United States, which connects to the Great Lakes basin area of the northeastern United States. Local, state, and federal governments as well as environmental committees, the general public, and trade associations are concerned that Asian carp will establish populations in the waters of the Great Lakes in the near future. Asian carp are invasive because they are adaptable, reach maturity quickly and eat phytoplankton and zooplankton – behaviour that can potentially threaten the life and habitat of native fish and plants in the Great Lakes (Reeves, 2014). Government officials, the general public, environmental committees and trade organisations are currently considering various policy alternatives to mitigate the spread of Asian carp into the Great Lakes.

HISTORY The Great Lakes were formed between 14 000 and 4 000 years ago by glacial retreat (McKenzie, 2014). They are the world’s largest group of freshwater lakes, and account for 21 percent of global freshwater. Edward Mills defines an invasive species as “successfully reproducing organisms transported by humans into regions where they did not exist in historical times” (in McKenzie, 2014, p. 94). Humans have introduced invasive species to the Great Lakes for hundreds of years. Invasive species can be damaging to an eco-system because they have less natural predators, reproduce at rapid rates and ex-

[ Asian carp ]

perience little competition for natural resources. This causes a decrease in native plant and animal populations and may harm the ecological and economic activity of the region (McKenzie, 2014). The Great Lakes now face these risks because of Asian carp. In particular, silver carp can grow up to 18 kg and jump several times their body length in the air, which has caused panic and bodily harm for sport fisherman and recreational lake users. Bighead carp are massive in size. They can grow up to five feet in length and weigh up to 45 kg. Asian carp were introduced to the US in Arkansas and Louisiana, imported largely from China in the 1970s for pest-control in man-made fish ponds. Asian carp spread to the Mississippi River during flooding in the 1990s and have since travelled to the Illinois River, which is linked to the Great Lakes near Chicago (Rodriguez, 2011). Bighead and silver carp compete for food and space with native fish (Hinterthuer, 2012), and will threaten the multibillion-dollar fishing industry of the Great Lakes if populations are established. (Wittmann, Cooke, Rothlisberger, & Lodge, 2014)

GEOGRAPHIC AREA OF CONCERN The aquatic pathways between the Upper Mississippi and Great Lakes basins are the geographical area of concern for most policies. The CAWS â&#x20AC;&#x201C; a large area of channels and canals, nearly half of which are man-made â&#x20AC;&#x201C; is key. Five major aquatic pathways between the Mississippi River and Great Lakes basins lie within CAWS, making it the primary area of focus for prevention of Asian carp invasion into the Great Lakes (USACE, 2014). The CAWS is a significant water body serving a range of mitigation, provisioning and recreational purposes. Most of the water in CAWS comes from Lake Michigan. CAWS is used for managing storm water and preventing flood; supplying water to industries,

THE PUBLIC SPHERE â&#x20AC;&#x2020;| â&#x20AC;&#x2020;2016 ISSUE

power generation and emergency response systems; as well as for commercial navigation, and recreational boating and fishing (USACE, 2014). The uses and users of CAWS are taken into account while examining the efficacy and cost of policy alternatives in this geographical area. While it is possible that Asian carp can enter the Great Lakes through rivers leading to Lake Erie, one of the five Great Lakes, the Chicago area is the focus of this analysis.

STAKEHOLDERS US policy stakeholders agree that it is important to keep Asian carp out of the Great Lakes. It is likely that no solution will appease every stakeholder and it is desirable to avoid a solution that greatly benefits one party to the detriment of another (Rodriguez, 2011). The policy chosen by decision-makers for implementation should consider the perspectives and concerns of government officials, the general public, environmental committees, and trade associations. Government officials. Politicians want to choose the best policy for mitigating an Asian carp invasion because it is in the public interest to protect natural resources. Political figures have a role in the Asian carp debate because they are authorised to take official action. The U.S. Congress will approve and fund the chosen alternative, multiple governors and state legislatures will implement federal policy, and ground-level administrators and public officials will oversee infrastructure projects. Environmental committees. Formal committees conduct research and advocate for policies to prevent Asian carp from populating the Great Lakes. The Asian Carp Regional Coordinating Committee (ACRCC) and the Great Lakes Restoration Initiative (GLRI) are made up of 20 local, state, and federal agencies that work together to protect and restore the Great Lakes. The ACRCC is funded with US$104 million from GRLI (Hinterthuer,

[ Asian carp ]

2012). The Alliance for the Great Lakes (AGL) and National Resources Defense Council (NRDC) are action groups who conduct their own studies to determine how the City of Chicago can best redesign its water infrastructure to allow for hydrological separation in order to keep Asian carp out of the Great Lakes (Hinterthuer, 2012). Trade associations. The Chemical Industry Council of Illinois (CICI) is a trade association that aims to protect the economic interest of the chemical industries of Illinois. Unlock Our Jobs is a coalition of over 60 agriculture, business, communities, and citizens. Both associations are specifically fighting for a solution to keep Asian carp out of the Great Lakes without closing the Chicago locks.

GOVERNMENT ROLE Federal and state US governments play a powerful role in funding, regulating, and protecting policies and laws that affect the Great Lakes. Every viable solution to stop the invasion of Asian carp into the Great Lakes has political considerations, such as the cost to government and taxpayers, the economic impact on stakeholders, and the level of support among citizens. Major federal funding began in 2010, when President Obama formed ACRCC to stop the spread of Asian carp into the Great Lakes and funded it through the GLRI. Congress funded a comprehensive USACE study of feasible policy alternatives to prevent the spread of Asian carp into the Great Lakes and has the power to fund and endorse the policy alternative of their choice (Reeves, 2014). The federal government has used its legislative power to protect the Great Lakes from invasive species. In 2010, the Asian Carp Prevention and Control Act made it unlawful to import Asian carp, specifically the bighead carp, into the United States

THE PUBLIC SPHERE | 2016 ISSUE

by labelling it “an injurious species under the Lacey Act” (Rodriguez, 2011, p. 27). While this law does not address the Asian carp population already present in the United States, it is a crucial step toward invasive species management. State governments have tested their power to influence a solution to an Asian carp invasion using legal channels. The State of Michigan filed a lawsuit against the State of Illinois and the USACE in 2009, requesting a preliminary injunction to close the locks from CAWS into Lake Michigan. Attorneys general from the states of Minnesota, Ohio, Pennsylvania, and Wisconsin joined the lawsuit in 2011. The request for injunction was eventually denied by the US Supreme Court (Rodriguez, 2011).

POLICY CONTEXT AND ALTERNATIVES Policy alternatives Short and long-term policies are being considered for implementation to stop the spread of Asian carp into the Great Lakes. The 2014 USACE Great Lakes and Mississippi River Interbasin Study Report (GLMRISR) was the first comprehensive government research highly anticipated by stakeholders. It set out eight policy alternatives to mitigate the spread of Asian carp into the Great Lakes. This section presents a high level overview of the key themes that are important in policy alternatives before introducing the specific alternatives. Long-term policy alternatives. The baseline policy alternative (status quo) is to continue what is currently being done: USACE operation of two electric barriers approximately 40 kilometers downstream from Chicago and genetic surveillance (eDNA) of CAWS and the Great Lakes. Other methods of controlling the spread of aquatic invasive species currently used include struc-

[ Asian carp ]

tural approaches such as locks, electric barriers, treatment plants, screened sluice gates, and physical barriers. Locks are a two-gate system that open and close when boats pass through and function as a control point to physically prevent invasive species from entering the waterway. Electric barriers located in the bottom of a water channel create an electric field that repels or stuns fish. Treatment plants within the lock system replace the water that contains invasive species, using ultraviolet light, screening, and filtration. Sluice gates are used during flooding periods, allowing water to pass through while controlling fish travel, while physical barriers made from concrete and sheet pile completely separate one body of water from another (USACE, 2014). Non-structural controls are also part of the baseline approach, including the use of nets and chemicals to remove or control fish and plants in or from a water body, the control of boat access to waterways, the development of boat cleaning stations for vessels passing into different bodies of water and public awareness and education (USACE, 2014). Three of the eight policy alternatives presented by USACE include a combination of structural and non-structural techniques and are estimated to cost US$18 billion and take 25 years to implement (Reese, 2014; USACE, 2014). Concerns relating to time and money are part of the reason why policy alternatives presented by GLMRISR have not been chosen or implemented to date. Genetic surveillance measures, such as the collection of environmental DNA, are particularly effective at evaluating the presence of an invasive species prior to the establishment of a population (Jerde et al, 2013). Environmental DNA has been found in the CAWS above the USACE electric barrier. This alerts scientists to the possibility that Asian carp populations are already in the Great Lakes waters, although no fish have been positively detected to date. eDNA cannot provide information about how many Asian carp are in the water, whether they are dead or alive, or if

THE PUBLIC SPHERE â&#x20AC;&#x2020;| â&#x20AC;&#x2020;2016 ISSUE

they were transported by an outside source such as a ship. The long-term benefits of genetic surveillance of the Great Lakes region for Asian carp will become clearer with time and additional research (Jerde et al, 2013). Short-term policy alternatives. Human consumption of Asian carp as an invasive species management strategy is a potentially cost-effective, market-based solution. However, Varble and Secchi (2013) caution that humans in the Great Lakes region should only be allowed to consume Asian carp while a population problem exists and argue that Asian carp should not be farmed or otherwise cultivated for sustained profitable commercial activity. Varble and Secchi, after conducting a national survey in 2013, suggest that Americans are willing to sample or even purchase Asian carp, which shows promise for this strategy. Commercial harvesting is another short-term policy alternative that includes overfishing Asian carp in the Illinois River to control population numbers. Both human consumption and commercial harvesting activities as policy alternatives have not received the same scrutiny from stakeholders as long-term alternatives. This is because, despite their low cost of implementation and absence of significant negative externalities they do not significantly mitigate the spread of Asian carp into the Great Lakes.

PREVIOUS EVALUATION OF POLICY EFFECTIVENESS Each policy alternative proposed in the GLMRISR was scientifically evaluated on its ability to keep Asian carp out of the Great Lakes. According to Wittmann et al (2014), hydrological separation, with 99 per cent containment of Asian carp access, is the most effective. Underwater electronic and acoustic-bubble-strobe barriers are 92-95 per cent effective.

[ Asian carp ]

The use of empirical data to evaluate the efficacy of possible solutions is unfeasible due to financial, ethical and time constraints. Structured expert judgment (SEJ) is more appropriate, because it “aggregates expert knowledge to quantify the uncertainties associated with invasive species pathway management” (Wittman et al, 2014, p. 2152). SEJ also suggests that expert responses should not include the identities of the experts. This helps “quantify scientific uncertainties on important societal issues without the risk of political confrontation, and for policy makers to utilize this information to make risk-based decisions in a transparent and defensible way” (Wittmann et al, p. 2155). Empirical data and structured expert judgment together are accepted tools for comprehensive scientific evaluation of policy alternatives and solutions.

ALTERNATIVE VIEWPOINTS AND LIMITATIONS Not everyone agrees on how policy alternatives are evaluated. A few scientists and scholars doubt the published success rates of measures to stop the spread of Asian carp into the Great Lakes, because of the “expense or infeasibility of experimental tests at realistic field scales” (Wittmann, Cooke, Rothlisberger, & Lodge, 2014). Instead they believe that SEJ should carry more influence in adjudicating policy alternatives. However, expert judgment has its limitations and cannot determine the cost of prevention strategies, the costs of an Asian carp invasion of the Great Lakes, or a relationship between carp numbers and the likelihood they will establish in the Great Lakes (Wittmann et al, 2014). Questions have been raised about whether Asian carp in the Great Lakes would be as disastrous as speculated by scientists and the media. It is difficult to know the full effect of Asian carp on the Mississippi and Illinois Rivers because

THE PUBLIC SPHERE | 2016 ISSUE

changing water and temperature patterns makes studying rivers difficult (Hinterthuer, 2012). Media, in particular social media, has illuminated the risks of an Asian carp invasion of the Great Lakes and shaped stakeholders’ understanding of the threat and their opinions about the various policy alternatives (Colon, 2014). A YouTube search yields over 40 000 videos of silver carp rapidly flying out of waterways, leading the viewer to believe that these invasive species are likely to cause bodily harm to fishers. Scientists may either downplay the risk, or only be able to influence a small general public audience with their research, while media may over-exaggerate the threat to millions of viewers. The GLMRISR report, absent of a direct policy recommendation, sparked animated conversations at public town hall meetings along the Mississippi River and Great Lakes basin regions (Reeves, 2014). Some concerned citizens of the Great Lakes basin, frustrated by the abundance of research without decisive action from the US government, wonder if Asian carp will populate the Great Lakes before a policy is implemented and speculate about the health of the Great Lakes for future generations.

ANALYSIS Method This study provides a better understanding of prospective policy alternatives for effectively mitigating the spread of Asian carp in the Great Lakes by incorporating political, social and economic effects not previously considered. This study uses a cause and effect analysis of GLMRIS policy alternatives because it allows stakeholders to easily and thoroughly make sense of important policy alternatives being considered. This will help expedite an efficient policy planning and implementation process.

[ Asian carp ]

A Goeller Scorecard is used to compare each policy alternative. Goeller Scorecards are a commonly used tool for biodiversity analysis whereby a subjective numerical weight is assigned to various alternatives to allow comparison between alternatives. Assignment of the weight to different alternatives is a critical part of the process. In this analysis, the assignment of relative importance of different policy effects was informed by an extensive literature review. A subjective weight was placed on each policy effect (scientific, political, social, and economic), using a scale of 1 (not important) to 5 (very important). Scientific effects are rated 5 for importance, because policy makers primarily use scientific principles to adjudicate each alternative. The results of analysis are sensitive to this subjective weighing of effects and could be tailored in the future to accommodate new data and information not known at this time. Political effects are rated 3; while politics alone will not mitigate an Asian carp migration, it sets the timeline and climate for policy choice and implementation. Social effects are rated 2; while public opinion is not likely to influence policy choice, the will of the people as a basic tenet of democracy must be acknowledged and considered. Economic effects are rated 4, because the policy alternative chosen should be economically feasible for stakeholders; economic effects are only second in importance to scientific effects. Each policy alternative was then given four scores â&#x20AC;&#x201C; one for each policy effect on a scale of 1 (not feasible) to 5 (very feasible). These scores were then multiplied by the predetermined weight placed on each policy effect totalled (see Table 1).

Definitions Baseline conditions: the current actions to mitigate the spread of Asian carp to the Great Lakes â&#x20AC;&#x201C; scientific, political, and economic data is not provided because the baseline is al-

THE PUBLIC SPHERE â&#x20AC;&#x2020;| â&#x20AC;&#x2020;2016 ISSUE

ready implemented and supported on these merits. Going forward, each policy alternative is evaluated as a continuum of baseline conditions. Economic effects: aspects of a policy alternative that directly involve money and commerce such as total cost of policy implementation and the effect on operating and capital budgets, private industry, employment, infrastructure, and economic development are considered. People: includes assumptions based on voluntary, documented opinions of policy stakeholders, either through a public comment period in town hall meetings and interviews conducted, or information gathered by research professionals. Political effects: qualities of implementing and managing a policy alternative; the role of political actors, and secondary policies that must be developed to mitigate particular consequences of a policy alternative are included. Scientific effects: observations that are measured and judged using scientific principles; specifically, the effectiveness of proposed actions, water quality, potential ecological contamination, and risk of flooding are mentioned. Social effects: relate to any and all concerns and opinions of the general public. Examples include impressions of timelines for implementation and anticipated results, beliefs about the effectiveness of each policy alternative, other inherent (mitigating) costs of each policy alternative, and information about public opinions in each neighbouring state.

Limitations There were several limitations to the data collection process. Records of public opinion show that citizens have minimal knowledge about the specific details of policy alternatives. There is a lack of credible primary data involving political, social, and economic considerations relating to a potential Asian carp inva-

[ Asian carp ]

sion of the Great Lakes despite many editorial articles providing biased and often inaccurate information. Table 1 Goeller Scorecard with weighted criteria Policy alternative

Scientific effect Weight = 5

Political effect Weight = 3

Social effect Weight = 2

Economic effect Weight = 4

Totals

#1 Baseline

5 (1 x 5 = 5)

3 (1 x 3 = 3)

2 (1 x 2 = 2)

4 (1 x 4 = 4)

5 (1 x 5 = 5)

15 (5 x 3 = 15)

10 (5 x 2 = 10)

20 (5 x 4 = 20)

15 (3 x 5 = 15)

15 (3 x 3 = 9)

6 (3 x 2 = 6)

12 (3 x 4 = 12)

20 (4 x 5 = 20)

6 (2 x 3 = 6)

6 (3 x 2 = 6)

12 (3 x 4 = 12)

#5 - #8

25 (5 x 5 = 25)

6 (2 x 3 = 6)

4 (2 x 2 = 4)

8 (2 x 4 = 8)

The table below outlines the eight prospective policy alternatives being considered by USACE for mitigating an Asian carp invasion of the Great Lakes. Policy alternatives are itemised by row, and policy effects are listed in columns. Information in the table represents the key elements for consideration for each effect policy and the corresponding effect. Note that GLMRISR policy alternatives #5 - 8 are grouped and analysed together because of their core similarities (Jatkevicius, 2003).

THE PUBLIC SPHERE â&#x20AC;&#x2020;| â&#x20AC;&#x2020;2016 ISSUE

Table 2 Comparison of GLMRIS policy alternatives and known effects Policy alternatives

Scientific effect

Political effect

Baseline- sustained activity

N/A

Time until reduction of threat: N/A Effectiveness: Not proven to reduce the risk of Asian carp invasion of the Great Lakes Impact on water quality: None Ecological contamination: None Risk of flooding: None

Time until Implementation: Immediate Role of political actors: Very little. The federally-funded Great Lakes Restoration Initiative pays for commercial harvesting activity Secondary policies: None

Time until reduction of threat: 25 years. Effectiveness: 92% effective at keeping Asian carp out of the Great Lakes Impact on water quality: None Risk of flooding: Present, but quantified

Time until implementation: 25 years Role of political actors: Federal funding for existing electric barriers promotes political cooperation between affected states (little effect on state budgets/politics). Secondary policies: Requires modification and implementation of floodwater management plans in numerous Chicago areas. Support: Commercial haulers/ shippers

(GLMRIS Alt. #1) Nonstructural control technologies Includes the use of nets and chemicals, controlling boat access to waterways, commercial harvesting, human consumption, and public education about invasive species (GLMRIS alternative #2)

Structural control methodsNo physical separation Includes two electric barriers, three reservoirs, and non-structural control technologies (GLMRIS Alternative #3)

[ Asian carp ]

Social effect

Economic effect

Beliefs about effectiveness: 98% believe this alternative is not effective enough on its own.

N/A

Beliefs about effectiveness: 75% believe this alternative is not effective enough on its own Beliefs about other inherent (mitigating) Costs: Human consumption is not favoured because of negative media portrayal of Asian carp State that Show Majority Support: Inconclusive

Dollar cost of policy implementation (operating budget): US$68 million/year Capital budget(s): Unknown Private industry: Potential to make money from the private sale of Asian carp though details unknown Employment: None Infrastructure: None Economic development: None

Beliefs about effectiveness: 35% support policy alternatives that minimise the effect on navigation Beliefs about other inherent (mitigating) costs: Unknown State populations that show equal or majority support: Ohio, Minnesota, Illinois, Indiana, Louisiana, and Ontario (Canada)

Dollar cost of policy implementation (capital budget(s)): US$15.6 billion Operating Budgets: Annual costs not conclusive Private industry: Private boaters and fisherman have to navigate around electric barriers, may choose other areas. Commercial haulers must limit and adjust their loads. Employment: Unknown Infrastructure: US$6.4 billion, unknown who will pay for how much. Economic development: Jobs created through new infrastructure projects and management.

continue

THE PUBLIC SPHERE | 2016 ISSUE

Table 2. Comparison of GLMRIS policy alternatives and known effects (continuation) Policy alternatives

Scientific effect

Political effect

Structural control methodsTwo (2) physical barriers Also includes three electric barriers and non-structural control technologies

Time until reduction of threat: 25 years Effectiveness: 92-99% effective at keeping Asian carp out of the Great Lakes. Impact on water quality: Implied concern in CAWS without full description Risk of flooding: Potential flooding of Chicago land, streets

Time until implementation: 25 years Role of political actors: Michigan and several other states have petitioned U.S. Congress, unsuccessfully, to impose physical barriers in all major water arteries of CAWS and CSSC. States of Michigan and Illinois disagree about respective economic impacts of this policy. Secondary policies: Requires significant changes to floodwater management plans Support: Natural Resources Defense Council (NRDC), National Wildlife Federation (NWF), and the Alliance for the Great Lakes (AGL)

Time until reduction of threat: 25 years Effectiveness: 99% effective at keeping Asian carp out of the Great Lakes. Impact on water quality: Expected contamination of Lake Michigan and CAWS water Risk of flooding: Significant to the City of Chicago and suburban areas. Other consequences: Would exacerbate existing concerns about Chicago’s flood water systems

Time until implementation: 25 years Role of political actors: Michigan and several other states have unsuccessfully petitioned the U.S. Congress to impose physical barriers in all major water arteries of CAWS and CSSC States of Michigan and Illinois disagree about respective economic impacts of this policy. Secondary policies: Support: Natural Resources Defense Council (NRDC), National Wildlife Federation (NWF), and the Alliance for the Great Lakes (AGL)

(GLMRIS Alternative #4)

Structural control methodsthree (3) + physical barriers Also includes two electric barriers, and two reservoirs, and two screened sluice gates (GLMRIS # 8 only), and non-structural control measures (GLMRIS Alternative #5, 6, 7, 8)

Source: United States Army Corp of Engineers. (2014). Great Lakes and Mississippi River Interbasin Study Report. Chicago, IL., Colon, S. (2014). Occupying nature: Fishing for meaning in the Asian carp. Transforming Anthropology 22(1), 24-30., Reeves, A. (2014). Carpocalypse now: Can the US Army Corps contain the Great Lakes carp invasion? Alternatives Journal, 40(5), 16-22. United States Army Corp of Engineers. (2014). Great Lakes and Mississippi River Interbasin Study Report Public Comment Summary. Chicago, IL: Retrieved from http://glmris.anl.gov/ documents/docs/glmris_report_public_input.pdf/

[ Asian carp ]

Social effect

Economic effect

Beliefs about effectiveness: 45% support policies that involve physical separation of the Great Lakes and Mississippi River basins; believe this is the most effective method of mitigating a Great Lakes invasion of Asian carp Beliefs about other inherent (mitigating) costs: some people believe that mitigating costs are inflated in policy cost estimates State populations that show equal or majority support: Michigan, Wisconsin, Ohio, New York, and Minnesota

Dollar cost of policy implementation (capital budget(s)): US$7.2 billion Operating budgets: Annual costs not conclusive Economic impact: An estimated US$70 million loss to the State of Illinois. Private industry: US$18 billion loss to commercial shipping industry Employment: Unknown. Infrastructure: US$500 million expenses incurred by State of Illinois. Economic development: Possibility of new marine rail/lift systems in place of ships/ navigation vessels, jobs created through new infrastructure projects and management

Dollar cost of policy implementation (capital budget(s)): US$8.35 - 15.52 billion, depending on the alternative Operating budget: Annual costs not conclusive Economic impact: An estimated US$70 million loss to the State of Illinois. Employment: Unknown. Infrastructure: US$500 million expenses incurred by State of Illinois. Economic development: Possibility of new marine rail/lift systems in place of ships/ navigation vessels, jobs created through new infrastructure projects and management

THE PUBLIC SPHERE | 2016 ISSUE

RESULTS The results of the Goeller Scorecard and cause and effect analysis show that the most feasible policy alternative, non-structural control technologies (#2), is also the least likely to prevent an Asian carp invasion of the Great Lakes. Policy alternatives (#4 and #5) with the highest scores for scientific feasibility, or the greatest potential to prevent an Asian carp invasion, also have the lowest overall feasibility scores. The most effective alternative is the least feasible, while the most feasible alternative is the least effective. It is no surprise that a new policy has not been chosen or implemented to date. While the scientific effectiveness of each policy is acknowledged, it is not a perfect justification for choosing a policy to prevent an Asian carp invasion of the Great Lakes. For example, five of eight prospective policy alternatives would have an effect on flooding and water quality in the City of Chicago and nearby suburban areas. The management of current mitigation methods (baseline condition) is a stable and effective political process. If one of the eight policy alternatives is chosen, however, management and funding of mitigation under the alternative scenario may face political scrutiny. While the GLMRIS outlines eight policy alternatives, people usually adopt one of three opinions – support for hydrologic separation, support for electric barriers, or support for protection against Asian carp with no specific opinion about which method should be implemented. Public opinion varies based on geographic location. For example, stakeholders who work or live in communities that are directly affected by commercial hauling are more likely to oppose hydrologic separation. Stakeholders who work or live in communities that are directly affected by recreation and tourism are more likely to support hydrologic separation. The media presentation of the Asian carp as a looming threat to the Great Lakes has increased the sense

[ Asian carp ]

of urgency within the citizen base to implement a solution. The negative portrayal of the Asian carp however, may cause doubt about human consumption as a viable control method. It is unclear how each policy alternative would affect the economies of states, cities, towns, and citizens in the Great Lakes basin region. It is also unclear who would be responsible for funding the chosen alternative.

RECOMMENDATION Decision-makers can choose the policy that ranks the highest in scientific effectiveness even if this policy is not politically, socially, or economically feasible. This will fulfil the primary goal of mitigating an Asian carp invasion of the Great Lakes. However, the recommendation is to increase the political, social, and economic feasibility of a scientifically effective policy. Policy alternatives should be presented to government officials, the general public, environmental committees and trade organisations only when they include specific and concise information about scientific, political, social, and economic effects. A failure to acknowledge these effects is why policy planning and implementation are currently stalled. Comprehensive information behind each policy alternative will allow for an accurate cost/benefit analysis of each policy alternative, and will facilitate better policy decisions and a speedier policy implementation.

CONCLUSION Policy decision-makers must increase the political, social, and economic feasibility of the most scientifically effective policy in order to best mitigate the spread of Asian carp in the Great Lakes. Policy alternatives are best understood when they address

THE PUBLIC SPHERE â&#x20AC;&#x2020;| â&#x20AC;&#x2020;2016 ISSUE

the political, social, and economic concerns of every stakeholder. Failure to understand the multiple facets of a policy decision may lead to profound consequences for the Great Lakes and its users, beyond inheriting a sustainable and invasive Asian carp population. If decision-makers wait to choose an official government policy, the Asian carp threat to the Great Lakes might change. Problems tend to amplify or worsen with time if they are not properly addressed. The facts that support a particular policy choice now may be invalid or irrelevant later. The cost of implementing an Asian carp prevention policy in five years may be greater than it is now, and political considerations may change as new elected officials take office. Policy analysts have quantifiable data regarding the scientific validity of each policy alternative, but do not have primary data that positively identifies the political or social considerations. Stakeholders and decision-makers need a viable tool for measuring political and social opinion, such as a Great Lakes regional survey. At minimum, the potential economic externalities facing stakeholders needs to be factored into a proper cost/benefit analysis of each policy alternative.

REFERENCES Colon, S. (2014). Occupying nature: Fishing for meaning in the Asian carp. Transforming Anthropology 22(1), 24-30. http:// dx.doi.org/10.1111/traa.12019 Great Lakes Commission and Great Lakes and St. Lawrence Cities Initiative. (2012). Restoring the natural divide: Separating the Great Lakes and Mississippi River basins in the Chicago Area Waterway System. Ann Arbor, MI, Chicago, IL: Retrieved from http://projects.glc. org/caws//#reports Great Lakes Environmental Research Laboratory (GLERL). (2014). About our Great Lakes: Facts about the Great Lakes Basin. Ann Arbor, MI: Retrieved from http://www.glerl. noaa.gov/pr/ourlakes/facts.html Hinterthuer, A. (2012). The explosive spread of Asian Carp: Can the Great Lakes be protected? Does it matter? BioScience, 62(3), 220-224. http://dx.doi.org/10.1525/ bio.2012.62.3.3

[ Asian carp ]

Jatkevicius, J. (2003). Satisficing, public libraries, and internet filtering: A case study of public policy development. Library Administration & Management, 17(3), 135-140. Jerde, C.L., Chadderton, W.L., Mahon, A.R., Renshaw, M.A., Corush, J., Budny, M.L., Lodge, D.M. (2013). Detection of Asian Carp DNA as part of a Great Lakes basin-wide surveillance program. Canadian Journal of Fisheries & Aquatic Sciences, 70(4), 522-26. http:// dx.doi.org/10.1139/cjfas-2012-0478 McKenzie, S. (2014). Laurentian and African Great Lakes- different strategies in the fight against invasive species. Indiana International and Comparative Law Review, 24, 93-111. Reeves, A. (2014). Carpocalypse now: Can the US Army Corps contain the Great Lakes carp invasion? Alternatives Journal, 40(5), 16-22. Rodriguez, N. (2011). The Asian Carp threat to the Great Lakes: An analysis of alternatives for preventing Asian carp migration. Policy Perspectives, 18(1), 25-39. http://dx.doi. org/10.4079/pp.v18i0.9352 United States Army Corp of Engineers. (2014). Great Lakes and Mississippi River Interbasin Study Report. Chicago, IL: Retrieved from http://www.glmris.anl.gov/glmris-report/ United States Army Corp of Engineers. (2014). Great Lakes and Mississippi River Interbasin Study Report Public Comment Summary. Chicago, IL: Retrieved from http://glmris. anl.gov/documents/docs/glmris_report_public_input.pdf/ Varble, S., Secchi, S. (2013). Human consumption as an invasive species management strategy. A preliminary assessment of the marketing potential of invasive Asian carp in the US. Appetite, 65, 8-67. http://dx.doi.org/10.1016/j.appet.2013.01.022 Wittmann, M.E., Cooke, R.M., Rothlisberger, J.D., Lodge, D.M. (2014). Using structured expert judgment to assess invasive species prevention: Asian carp and the Mississippi - Great Lakes hydrologic connection, Environmental Science Technology, 48, 2150-2156. http:// dx.doi.org/10.1021/es4043098

The New Eurasian Land Bridge Opportunities for China, Europe, and Central Asia Gabor Debreczeni

MA in International Development and International Economics, Class of 2015 Johns Hopkins University School of Advanced International Studies

ÂŠJohn Cobb

ABSTRACT This article evaluates the present and considers the future of the intermediary role of Central Asia in overland trade routes from China to Europe. Focusing primarily on technical and pragmatic issues, it discusses the advantages and costs of potential freight modes and trade routes from China to Europe, finding that rail freight already operates successfully and at high efficiency via Central Asia, albeit at a small scale. Central Asian countries play a successful economic role in the overland trade, but could further benefit if they took part in, rather than just facilitated, China-Europe trade. China and Europe benefit from faster and cheaper trade with each other, and would further benefit from the inclusion of nations in between for either their import demand or their potential development as low-cost manufacturers. On the other hand, Russia’s policy regarding overland trade is driven by the opportunity of continued re-integration with Central Asian nations in the name of facilitated trade. The first section discusses the advantages and costs of potential freight modes and trade routes from China to Europe. The second section describes the Central Asian route’s emergence and the current state of its freight operations. The third section considers the long-term outlook of the route and the challenges of high-speed upgrades mooted by China. The fourth section analyzes Central Asian economies’ current roles in the trade occurring between China and Europe on the Central Asian trade route, as well as potential opportunities for their further engagement with China-Europe trade, while the last section discusses risks for the route’s future importance and growth. Finally, the report concludes by considering the relevant state actors’ potential future policies regarding overland trade between China and Europe.

[The New Eurasian Land Bridge ]

THE NEW SILK ROAD The Silk Road, the legendary trade route between China and Europe, has held an important place in the popular imagination of societies across the Eurasian landmass ever since trade began on the route over 2000 years ago. Like “El Camino Real” (Masters 2015) or the triangular trade of the Atlantic, the term “Silk Road” never referred to one fixed route, but instead to a trade pattern that shifted with alliances, climate conditions and the changing nature of banditry. Central Asia’s role has always been that of the “in between” – the secondary player between two great powers. The original inspiration for the trade route came from a mixed Greek-Central-Asian society (in present day Tajikistan) impressing Chinese leaders with the quality of their war horses (Mark, 2014). Kazakh President, Nursultan Nazarbayev, has recently maintained this Central Asian, in between identity by claiming that Kazakhstan looks both east and west and is a combination of Asian and European civilizations saying that he has “tried to realize the idea of Eurasianism in Astana, [the capital of Kazakhstan]” (Kazinform, 2009). The idea of an overland trade route from China to Europe re-emerged after the fall of the Soviet Union. Before and immediately after the fall of the Soviet Union, the route was crippled by extreme versions of the bottlenecks that still plague it today: red tape, low capacity, and complicated border procedures (Winterbottom, 2012). Practically speaking, any New Silk Road from China to Europe must go through Russia. Any route further south is extremely complex, either requiring a circuitous detour around the Caspian Sea or traversing the Himalayas, followed by crossing Afghanistan and Iran or Pakistan and Iran. This has caused much promise to be wished upon a new rail route via Central Asia, dubbed the New Eurasian Land Bridge. The Central Asian route’s recent popularity has origin in China’s spectacular growth. China began to re-emerge as an exporting

THE PUBLIC SPHERE | 2016 ISSUE

1. Generally, and in this article, Central Asia refers to Kazakhstan, Kyrgyzstan, Uzbekistan, Turkmenistan, and Tajikistan.

power in the 1980s (Silva-Ruete, 2006), and now exports US$2 trillion of products per year across all routes, with roughly US$500 billion of that to Europe and US$15 billion to the five Central Asian countries1 (Observatory of Economic Complexity, 2015). A disproportionate amount of the gains in incomes from China’s rise have taken place in the country’s east, the part of the country that contains the biggest cities and sea shores, allowing for agglomeration economies and easy logistics for export-based businesses that ship via ocean freight. As a result, labour costs have been increasing in eastern China, and China’s low-wage manufacturing sectors have begun to shift to the interior of the country, to cities like Chongqing and Chengdu (Bradsher, 2013). As a result of this shift, and considering that the Chinese government had been pushing for economic growth in these inland provinces, export growth from the region has recently annually grown at a rate of 25 per cent (Winterbottom, 2012). These manufacturers have poor access to ports (Bradsher, 2013) and to the eastern Russian route, rendering a competitive advantage to the New Eurasian Land Bridge route via Kazakhstan. From the inland city of Chengdu, use of the route to Europe via eastern Russia would add roughly 3,000 kilometres – three or four days’ travel time to the freight journey. This article considers the present and future of overland trade routes from China to Europe focusing on the intermediary role of Central Asia. The first section discusses the advantages and costs of potential freight modes and trade routes from China to Europe. The second section describes the Central Asian route’s emergence and the current state of its freight operations. The third section considers the long-term outlook of the route and the challenges of high-speed upgrades mooted by China. The fourth section analyzes Central Asian economies’ current roles in the trade occurring between China and Europe on the Central Asian trade route, as well as potential opportunities for their further engagement with China-Europe trade, while the last section discusses risks for the route’s future importance and growth.

[The New Eurasian Land Bridge ]

Finally, the report concludes by considering the relevant state actorsâ&#x20AC;&#x2122; potential future policies regarding overland trade between China and Europe.

A TRADE ROUTE THOUGH CENTRAL ASIA Shipping modes and railâ&#x20AC;&#x2122;s advantage In international freight shipping the cost of shipment is inversely proportional to time of delivery. Global producers are willing to pay a premium for faster transit due to their increasing reliance on just-in-time supply chains (The Chamber of Commerce of the United States, 2006) and because faster transit allows them to optimize their profits by responding quickly to changes in demand in markets located far from their production operations (Bradsher, 2013). The 13,000-kilometres Silk Road is long enough that rail, sea, and air shipments take their natural places in the freight hierarchy, for instance sea is slowest and cheapest, air is fastest and most expensive, and rail is in between on both metrics (Hutchinson, 2015). A 40-foot (12.2 metres) container travelling warehouse-to-warehouse from China to central Poland will take approximately three days to arrive by plane, two weeks by train and six weeks by boat, with an associated cost of US$40,000, US$10,000 and US$5,000, respectively (Mount, 2014). Because trucking from the interior of China to a port is expensive, the difference between train and sea shipping is less extreme - about 25 per cent for shipments that originate in western China (Bradsher, 2013)2. For time sensitive but heavy products - such as car parts, processed food, and high technology, the price-speed structure of rail freight is more appealing than either the cheaper or faster alternatives. Shipping by rail also provides additional value as a way to diversify methods of export for China in response to three areas

2. For some companies a further reason to use rail (or sea) instead of air freight is that the carbon emissions of air freight are 30 times higher (Bradsher, 2013).

THE PUBLIC SPHERE | 2016 ISSUE

of risk. Firstly, the prices of all four major methods of transportation – rail, road, sea, and air, are differentially sensitive to changes in fuel prices. Secondly, the speed of sea shipping reduces dramatically as operators reduce speeds to save fuel when costs are high. In addition, Chinese sea shipping is overwhelmingly reliant on safe passage through the Malacca Straits, the existential importance of that passage to China’s exporters is a geopolitical issue for the country – giving China a further incentive to find ways to diversify its export routes (Arduino, 2014).

The New Eurasian Land Bridge The first reintroduction of a land route between China and Europe occurred via China’s border with Russia bypassing Central Asia. The chief weakness of this route is that the freight must cross into Russia in far northeastern China, which is an indirect route to Europe and – more importantly convenient only to China’s eastern port cities that are already well serviced by port infrastructure. In addition, much of the demand for rail freight is in fact coming from interior Chinese cities. In response to these challenges, an alternate route constructed in 1990 came into regular use in 2011 (Winterbottom, 2012). This route originates in eastern China and cuts through its interior before passing through Kazakhstan, Russia, Belarus, Poland and Germany with some trains continuing on to France and Spain. The 8,000-mile route, from the city of Yiwu in Zhejiang province to Madrid, is the longest continuous train route ever. Many slightly different routes have been labeled under the umbrella term of the New Eurasian Land Bridge – some link Chongqing to Duisburg, while others link Beijing to Hamburg, however it is the middle section of the route, from Kazakhstan to Poland, which seems to define it (Hutchinson, 2015). In the past two years, this train route has evolved into a naturally and privately sustainable venture, rather than simply a

[The New Eurasian Land Bridge ]

symbolic, unprofitable, government-run project. Since 2014, the Central Asian route has been used by five trains each week, running from China to Germany and back (Mount 2014), operated by a joint venture between Deutsche Bahn and Russian Railways (Brautlecht 2014). Customers on this line, between Chongqing and Duisberg, have included Foxconn, Hewlett-Packard, and Acer (Brautlecht 2014). The Chengdu-Poland line, operated by a Polish company called Hatrans, also uses the Central Asian route, runs weekly, and takes 14 days (Michalik 2015). It regularly runs at full capacity and its clients include DHL, Philips, Ericsson, Seagate, Samsung, and FIAT. This train also makes delivery stops in Russia and Kazakhstan (Arduino 2014). Rail freight from China to Europe is growing. Container traffic on the trans-Siberian railway grew by 15 per cent in 2013 and 22 per cent in 2014, with a total of 865,600 20-foot (6.1 metres) containers of freight in the first six months of 2014. The average container spent 14 days in transit, confirming industry claims on the speed of China-Europe rail freight transit (International Transport Journal, 2014). While Kazakhstan might have been overly optimistic in predicting a tripling of freight traffic through Kazakhstan every year from 2013 to 2020 (Bradsher, 2013), further growth does seem likely as the early movers solve logistical problems on the route, making newer entrantsâ&#x20AC;&#x2122; issues less pronounced3. It should be noted that Spain is the destination for which the rail routeâ&#x20AC;&#x2122;s comparative advantage is weakest, since the freight journey is lengthened by one week as a result of the journey from Germany to Spain, and Spain is well-serviced by ports and thus sea shipping (Mount, 2014). It is possible that the extension of the route from Germany to Spain will be more efficient if more trains use it, since the Germany-Spain route should take only two to three days at international best practice average freight times (Bureau of Transportation Statistics, 2008)4. Also, the European Economic and Social Committee recently suggested that

3. One example of such a solved issue is that Hewlett-Packard helped Chinese customs solve a software design flaw whereby customs software only allowed countries adjacent to China to be chosen as destination countries (Bradsher, 2013). 4. The freight trains in Europe have much lower priority on the rails compared to passenger trains in the United States. This slows rail freight speeds in Europe and makes the current speeds of the trains traveling on the New Eurasian Land Route more impressive (Furtado, 2013).

THE PUBLIC SPHERE | 2016 ISSUE

Spain adopt the European standard for freight rail gauges, which would remove a source of major delay at the France-Spain border (Bouron, 2009).

The question of high-speed rail Sea shipping makes up 80 per cent of global freight volume (GeoBeats News, 2014), but 95 per cent of China-Europe trade. Most of the remaining 5 per cent consists of rail freight (Winterbottom, 2012). While rail is a comparatively minor player in freight, the large amount of trade between the two regions – 40 million 20-foot containers crossing the Suez Canal each year (Rastogi and Arvis, 2014) means that rail freight from China to Europe is still a sizeable market. Some minor inefficiencies were addressed on the New Eurasian Land Bridge route, including that most countries on the route accept uniform freight documentation and that goods imported into China via the route are only inspected by customs upon arrival at their destination (Winterbottom, 2012). However, some challenges remain, including the fact that due to gauge changes, the containers must be transferred between two sets of carriages at three separate points: on the China-Kazakhstan, Belarus-Poland, and France-Spain borders (Hutchinson, 2015). That said, the line is operating at near-peak efficiency. In the United States, which is a highly developed freight market, freight trains average travel speeds of roughly 22 miles per hour (Bureau of Transportation Statistics, 2008). At that speed, the Chongqing-Duisburg segment would be covered in 13 days, and current freight delivery times on the route are only two days longer. The Chengdu-Lodz route is even closer to best practice speeds – its record time of 12 days is only a few hours slower than what the United States rail freight average would imply (Arduino, 2014). The route’s inefficiencies are more pronounced with regard to capacity than speeds, with bottlenecks caused by limited ca-

[The New Eurasian Land Bridge ]

pacity at borders and gauge transfer stations, and seasonality caused by extreme cold winter temperatures (Arduino, 2014). At current capacity, a peak-utilized New Eurasian Land Route could transport 20 per cent of China-Europe cargo (Rastogi & Arvis, 2014; Fronda, 2013). European rail freight authorities are also working on peak capacity issues on the western end of the route to improve infrastructure in order to increase the maximum length of freight trains allowed into the EU by 150 per cent (Mount, 2014). China has considered increasing the rail optionâ&#x20AC;&#x2122;s appeal by building a US$150 billion high-speed rail link from China to Turkey, via Kazakhstan, Turkmenistan, and Iran (Mann, 2014). Without considering border, customs, loading, or unloading delays, the journey could take as little as three days, despite being much longer because the route has to travel south of the ideal path to avoid Russia and the Caspian Sea, a move that is necessary to avoid gauge changes on the Russian-built train systemâ&#x20AC;&#x2122;s borders with China and Poland (Fronda, 2013). It is unclear from the plans alone whether the operational costs of such a system would make it unfeasible, a complement to traditional standard speed freight, or cause it to completely replace traditional rail freight. The answer will partially depend on issues of capacity â&#x20AC;&#x201C; if border, loading, and unloading bottlenecks are not addressed, then it is possible that both the new high-speed route and the old traditional route could at some point, simultaneously run at peak capacity, and still be only a minor if important complement to sea shipping. Given that China is currently planning to complete a 30,000 kilometre network of high-speed rail within the country (Badkar, 2011) as part of a US$300 billion investment plan (Fronda, 2013), funding and engineering a high-speed Silk Road route is well within its capabilities. The most difficult challenges are, of course, political, as the line would have to cross two difficult countries in Iran and Turkmenistan, and that China might

THE PUBLIC SPHERE | 2016 ISSUE

have to come to a compensatory arrangement with Russia, who stands to lose much from the completion of a freight route that bypasses it entirely.

The role of Central Asia Kazakhstan has already begun to benefit from the existence of the New Eurasian Land Bridge. As of 2013, the Kazakhstan rail system employed about 160,000 workers, representing 2 per cent of the country’s labor force. Much of the system’s work is related to trade with China – as of 2013, a quarter of rail freight within Kazakhstan was coming from or going to China and that number is growing (Bradsher, 2013). In 2011, the rail system’s freight division earned approximately US$100 million in profit, with 76 per cent of that profit coming from integrated freight operations and the remainder from freight operations where the Kazakh group engaged only in freight forwarding or operating (Kaztemirtans, 2012). In addition, locomotives used on the route were built in Kazakhstan, and General Electric is moving further production facilities to the country (Bradsher, 2013). In contrast, the real gains would materialize if the region were to latch onto the freight network as an endpoint for either exports from, or imports into, the Central Asia region. Currently, the trains return from Europe to China with cars, car parts, and manufacturing equipment, but many containers are empty on the return voyage (Mount, 2014). As mentioned earlier, the Chengu-Lodz trains have begun to make stops in Kazakhstan and it is natural that these trains might make it easier for Kazakhstan to import European goods. As of 2012, Kazakhstan imported US$28 billion of goods from Europe. The makeup of imports can best be described as diversified manufactured goods (Observatory of Economic Complexity, 2015). Central Asian economies are overwhelmingly dependent on the export of natural resources. Kazakhstan also exported US$16

[The New Eurasian Land Bridge ]

billion of goods to China in 2012, mostly consisting of raw materials – crude petroleum, iron, copper, radioactive materials, and zinc (Observatory of Economic Complexity, 2015). Given the complementarities of this trade, it is difficult to argue that freight containers should be returning to China empty. Both national governments and international organizations have made it a priority for the Central Asian economies to diversify into manufacturing. An obvious issue with this strategy is that the countries are landlocked and have no access to a port. As a result, most of the limited manufacturing that these countries engage in – such as the car manufacturing industry in Uzbekistan has focused on the massive markets provided by the region’s neighbours – namely India, Pakistan, China, and Russia. The Eurasian Land Bridge, in theory adds the European Union to this list, which is a larger market in purchasing power than the other four countries combined. Moreover, relying solely only on its neighbours exposes Central Asian exporters to political power plays, if China decides to restrict Central Asian business, there are few alternative points of sale available. The EU does have the ability to act in unison but it is still a somewhat independent pool of 28 actors and Central Asia is less likely to find itself frozen out. This could provide further incentive for Central Asian entrepreneurs to consider the manufacturing export space going forward, though of course this is all predated on these companies being able to compete with Chinese and European companies’ willingness to pay for cargo space on the New Eurasian Land Bridge. In addition, further gains could be realized from the presence of the New Eurasian Land Bridge if Central Asian countries were to begin to engage in China-Europe supply chains, perhaps as a supplier of low-cost labour to align with western China’s relatively higher-skill and higher-cost labor supply. This would particularly be an opportunity for the poorer Central Asian countries – Tajikistan, Kyrgyzstan, and Uzbekistan. According to Rastogi and

THE PUBLIC SPHERE | 2016 ISSUE

Arvis (2014, p.95), obstacles to engaging in these global supply chains include that very little voice is given to the users of logistics such as retail companies or exporters of manufactured and time-sensitive goods who actually suffer from the supply-chain inefficiencies…the region is one of the most isolated from international logistics knowledge.

Moreover, such integration pre-supposes the integration of rail networks within Central Asia, which is a matter of constant discussion – most recently involving Kazakh National Railways’ attempt to connect China to Iran via Turkmenistan (Bradsher, 2013), but is far from a complete project. If such integration were to take place, and the gains from the presence of the New Eurasian Land Bridge began to reveal themselves, it could also serve as a major incentive for those Central Asian countries who have not joined the Eurasian Customs Union to do so.

5. An instructive demonstration of why this is such an existential risk for rail is Paul Theroux’s famous rail voyage in 1973 in the book The Great Railway Bazaar. At the time, he traveled through Europe, Turkey, Iran, Afghanistan, Pakistan, India, Myanmar, Thailand, Laos, Malaysia, Singapore, Vietnam, Japan, and the Soviet Union without issue. At least four of these countries are impassible by rail today and a few countries he could not have traveled to, including China, are wide open today.

The risks Air freight and sea shipping will always have one advantage over rail freight: the ability to avoid contentious countries5. Another concern is that the route could lose favour with Chinese exporters, perhaps if there were a decline in demand for “justin-time” manufacturing components. While it is possible that this would lower demand for the route, allowing Central Asian exporters to become more competitive, it is more likely that the loss of the scale that was provided by the massive China-Europe trade will render the operation of the route unprofitable. China is clearly, and reasonably, trying to diversify the outlets for its products, and the New Eurasian Land Route is not its only rail route option. Another project it has recently funded is to build a rail connection to a port in Pakistan. Understandably, Russia is also keen not to lose freight traffic on the easternmost sec-

[The New Eurasian Land Bridge ]

tion of the trans-Siberian railroad, which is a direct – if slower, competitor of the New Eurasian Land Bridge. To this end, in 2013 it announced US$43 billion in infrastructure spending for improving the Trans-Siberian route’s connections with China (Bradsher, 2013). Fuel costs also pose a risk to shipping by rail that is difficult to predict. Rail’s advantage over air freight depends on its cost – an advantage that can expand or erode in an unpredictable and tremendous way given the volatility of fuel prices and the fact that fuel expenses make up a disproportionate amount of total air freight costs. In turn, rail’s advantage over sea shipping is its speed, and this too is variable with fuel prices, as shipping operators slow their cruising speeds to save fuel in times is high fuel prices (Bradsher, 2013).

CONCLUSION The emergence of the New Eurasian Land Route is an equally important advance in international trade for China, Central Asia, and Europe. It allows China to diversify its export transit options via a method that is a clear complement to maritime and air freight options. European companies who are moderately time-sensitive due to quickly changing demand for its products or “just-in-time” supply chain issues but unwilling to pay for air freight, now have a new method of transit to use for their products. Central Asia has shown the ability to service many aspects of the new industry, by managing tracks and trains and even by manufacturing carriages. Moreover, greater consequences, both positive and negative, could be at hand for all parties from the further emergence of overland trade routes between China and Europe. Central Asian countries could gain by becoming a part of, rather than just the facilitators of trade between China and Europe. Because of Cen-

THE PUBLIC SPHERE | 2016 ISSUE

tral Asia’s low labor and energy costs – lower than even in China, they could position themselves as a low-cost manufacturer of goods bound for either China or Europe. Capitalizing on the new overland trade routes is crucial for these countries given that their lack of ports and small domestic consumer markets make them peripheral to the modern economic order. Both China and Europe have much to gain from integrating the Central Asian countries into the trade supply chain, especially if the proposed high-speed route is realized, which could pass through large importing markets in Pakistan and Iran. Lastly, realizing the opportunity afforded to Central Asian nations by the overland trade, Russia has been quick to cooperate on the route as long as the Central Asian countries agree to further re-integration with Russia via the Eurasian Customs Union. Given Russia’s recent actions along its western borders, this should hardly be a surprise and it should be expected to continue as policy.

REFERENCES Arduino, A. (2014). The New Silk Road. Europe China Research and Advice Network. Retrieved 4th May 2015 from http://eeas.europa.eu/china/docs/division_ecran/ecran_is114_paper_91_the_new_silk_ road_alessandro_arduino_en.pdf. Badkar, M. (2011). What Chinese High-Speed Rail Will Do For Tourism. Business Insider. Retrieved 4th May 2015, from: http://www.businessinsider.com/map-of-the-day-chinas-world-heritage-spots-connected-by-high-speed-rail-2011-5?IR=T. Bouron, P. (2009). Technical, Socio-economic and Supply/Demand study regarding the transport of the FERRMED Great Rail Network (Scandinavia-Rhine-Rhône-Western Mediterranean). FERRMED. [Online] 27th October. Available from: http://www.eesc. europa.eu/resources/docs/5technical-analysis-patrice-bouron_new.pdf. [Accessed: 3rd May 2015]. Bradsher, K. (2013). Hauling New Treasure Along the Silk Road. The New York Times. [Online] 20th July. Available from: http://www.nytimes.com/2013/07/21/business/global/ hauling-new-treasure-along-the-silk-road.html?_r=1 . [Accessed:3rd May 2015]. Bradsher, K. (2013). Kazakhstan’s Bet on Rail. The New York Times. [Online] 20th November. Available from: http://www.nytimes.com/2013/11/21/business/international/kazakhstans-bet-on-rail.html. [Accessed:3rd May 2015].

[The New Eurasian Land Bridge ]

Brautlecht, N. (2014). Germany Plans to Expand Chinese Rail Link as Xi Visits Duisburg. BloombergBusiness. [Online] 28th March. Available from: http://www.bloomberg. com/news/articles/2014-03-28/germany-plans-to-expand-chinese-rail-link-as-xivisits-duisburg. [Accessed: 3rd May 2015]. Bureau of Transportation Statistics. (2008). Rail Freight Average Speeds, Revenue Ton-Miles, and Terminal Dwell Times. Retrieved 3rd May 2015, from http://www.rita.dot.gov/bts/sites/rita.dot.gov.bts/files/publications/transportation_statistics_annual_report/2008/html/chapter_03/table_03_04_08.html. Fronda, A. (2013). High-Speed Rail: the future of freight. World Finance. Retrieved 4th May 2015, from: http://www.worldfinance.com/infrastructure-investment/project-management/high-speed-rail-the-future-of-freight. Furtado, F. (2013). U.S. and European Freight Railways: The Differences That Matter. Journal of the Transportation Research Forum. Retrieved 4th May 2015, from: http://www. trforum.org/journal/downloads/2013v52n2_04_FreightRailways.pdf. Geobeats News. (2014). World’s Longest Train Journey Completed. [Online] 11th December. Available from: https://www.youtube.com/watch?v=MNGjifWYmbg. [Accessed: 3rd May 2015]. Hutchinson, J. (2015). Homeward bound on the world’s longest rail route. The Daily Mail. [Online] 25th February. Available from: http://www.dailymail.co.uk/travel/travel_news/ article-2968433/The-world-s-longest-rail-route-Train-finally-arrives-China-fourmonth-16-000-mile-journey-Spain-back.html. [Accessed: 3rd May 2015]. International Transport Journal. (2014). Box numbers steadily riding on Eurasian land bridge. [Online] 19th September. Available from: http://www.transportjournal.com/en/ home/news/artikeldetail/box-numbers-steadily-rising-on-eurasian-land-bridge. html. [Accessed: 3rd May 2015]. Kazinform. (2009). Nazarbayev: I’ve tried to realize the idea of Eurasianism in Astana. [Online] 29th June. Available from: http://www.inform.kz/eng/article/2182517. [Accessed 4th May 2015]. Kaztermirtrans. (2012). Annual Report is Published by “Kaztemirtrans”. Press Service of “Kaztemirtrans”. [Online] 13th September. Available from: http://www.railways.kz/en/ node/3284. [Accessed 4th May 2015]. Mann, W. (2014). China considers high-speed ‘silk railroad’ to Europe. New Civil Engineer. [Online] 7th August. Available from: http://www.nce.co.uk/china-considers-high-speedsilk-railroad-to-europe/8667079.article. [Accessed 4th May 2015]. Mark, J. (2014). Silk Road. Ancient History Encyclopedia. Retrieved 4th May 2015, from: http://www.ancient.eu/Silk_Road/. Masters, N. (2013). How El Camino Real, California’s ‘Royal Road,’ Was Invented. [Online] 4th January. Available from: http://www.kcet.org/updaily/socal_focus/history/la-as-subject/el-camino-real.html. [Accessed: 2nd December 2015]. Michalik, K. (2015). Train China to Lodz. Hatrans Logistics. Retrieved 3rd May 2015, from http://www.hatrans.pl/en/rail.html. Mount, I. (2014). Spain to China by rail: A 21st Century Silk Road riddled with obstacles. Fortune. [Online] 24th December. Available from: http://fortune.com/2014/12/24/spain-chinarail-trade/. [Accessed:3rd May 2015]. Observatory of Economic Complexity. (2015). The Observatory of Economic Complexity. Retrieved 4th May 2015, from https://atlas.media.mit.edu/en/. Rastogi, C. & Arvis, J. (2014) The Eurasian Connection: Supply-Chain Efficiency along the Modern Silk Route through Central Asia. 24th June. Available from: https://openknowl-

THE PUBLIC SPHERE | 2016 ISSUE

edge.worldbank.org/bitstream/handle/10986/18779/888910PUB0Box300EPI199120June122014.pdf ?sequence=1. [Accessed: 14th November 2015]. Silva-Ruete, J. (2006). The Development of China’s Export Performance. International Monetary Fund. Retrieved 4th May 2015, from https://www.imf.org/external/np/speeches/2006/030706.htm. The Chamber of Commerce of the United States. (2006). Land Transport Options Between Europe and Asia: Commercial Feasibility Study. Retrieved 3rd May 2015, from http://www. osce.org/eea/41310?download=true. Winterbottom, V. (2012). From Camels to Horsepower. Beijing Review. 17th September. Available from: http://www.bjreview.com.cn/Cover_Stories_Series_2014/2012-09/17/ content_609212_2.htm. [Accessed: 3rd May 2015].Winterbottom, V. (2012). The Silk Railroad. China Today. Retrieved 3rd May 2015, from http://eds.b.ebscohost.com/ ehost/pdfviewer/pdfviewer?sid=78a8641b-f4af-4d4a-a99f-7f895bb7df87%40sessionmgr111&vid=5&hid=112.

Tajikistan Political economy of drug addiction

Harald Edinger

MA, Class of 2014 Johns Hopkins University School of Advanced International Studies

ÂŠ Tracing Tea

ABSTRACT Narcotics are the lifeblood of Tajikistan’s economy: drug trading makes up for at least 30 per cent of the country’s GDP. This article addresses the questions of why and how this dependency materializes, and what the effects are on the economy and population. The analysis is structured along the following pattern: i) Causes of addiction: for what reasons is the economy addicted, and why is there no reduction of trafficked volumes despite international efforts? ii) Extent of addiction: how dependent is Tajikistan on drug trafficking in economic terms, and are there sensible ways to measure it? iii) Consequences of addiction: what short- and long-run economic, sociological and political effects is Tajikistan facing if it stays “hooked”? The article concludes with a brief discussion of iv) Counter narcotics policies: what is the current approach to Tajikistan’s problems, and does it work? This article argues that Western donor governments should reconsider the way they are combatting drug trafficking in Tajikistan. A sound approach will recognize that the production of opiates is largely demand-driven, and should thus be aimed primarily at demand reduction. Opium and heroin bring death and destruction along the paths they travel, but they also inject an immense amount of wealth into the Tajik economy. Counter narcotics policies must, therefore, be evaluated with a view to the potentially destabilizing effects on the socioeconomic fabric, in particular to which segments of the population they are hurting.

[Tajikistan]

INTRODUCTION: THE TRADE OF AFGHAN OPIATES There are few production and cultivation facilities for opium in Tajikistan itself. It is the country’s southern neighbour Afghanistan, which regularly produces over 70 per cent of the world’s opium output, according to yearly estimates by the United Nations Office on Drugs and Crime (UNODC, 2014). Roughly 25 per cent of Afghan opiates – that is opium and opium derivatives such as heroine or morphine – are transported on the “northern route,” across the Afghan-Tajik border, and then on towards the target markets in the Russian Federation and Europe – markets that are currently worth a combined US$33 billion (UNODC, 2011). The country has become the most attractive transit and distribution point for narcotics, mainly due to geographical and ethnic reasons – among other reasons, which will be discussed later on. It shares an approximately 1,200km-long land border with Afghanistan, posing the ideal entry point for smugglers. In terms of ethnic and social ties, 25 per cent of the Afghan population is composed of ethnic Tajiks (Engvall, 2006, p.828). Tajikistan still has strong connections to Russia due to the large number of Tajik migrant workers in the country, which also maintains its largest military outpost there. A look at drug seizure1 data seems to confirm Tajikistan’s status as a major transit hub. Between 1996 and 2006, Tajikistan effected 48 per cent of total opium seizures and 73 per cent of heroin seizures across all Central Asian countries – the amount of heroin seized is roughly four times as much as has been interdicted by other Central Asian Countries (UNODC, 2008).The exact volume of the trade flow cannot be determined with certainty. Most of the data is based on seizure volumes, including the figures reported by the UNODC. The volume of seized drugs, however, is an inaccurate indicator of the actual scope of trafficking in the region. The UNODC estimates that, on average, seizures amount to less than

1. The drawbacks of using seizure data in order to estimate the total flow of opiates will be discussed later on. However, seizure volumes can help to make regional comparisons and draw rough estimates of opiate flows across the Central Asian countries.

THE PUBLIC SPHERE | 2016 ISSUE

4 per cent of estimated total opiate flows in heroin equivalence and that the throughput increases annually (12 per cent increase in 2006) due to copious supply from Afghanistan (UNODC, 2008). FIGURE 1 HEROIN SEIZURES BY PROVINCES IN TAJIKISTAN, 2006

Source: UNODC, Illicit Drug Trends in Central Asia, 2008

The UNODC further assumes that approximately 60 per cent of all Afghan opiates that cross through Tajikistan enter the country in Western Tajikistan, namely the Khamdoni, Nizhni Pyanj, Shurabad and Parhar districts. Once inside the country, most drugs pass through the capital Dushanbe, for repacking, distribution and onward trafficking (UNODC, 2008). The map above shows the volumes of heroin seized in the different Tajik provinces. While most of the opiates appear to be coming in through the southwestern province of Khatlon, it should be noted that the southeastern border is particularly porous. Traffickers use trails through the mountains and across “green borders” to avoid detection. Even easier than through the

[Tajikistan]

mountains is crossing the Pyanj River, “which forms [a large] part of Tajikistan’s border with Afghanistan, is thinly guarded, and difficult to patrol. It is easily crossed without inspection at a number of points” (U.S. Department of State, International Narcotics Control Strategy Report, 2003). It can be assumed that the ratio of drugs seized to the total amount trafficked is a lot lower in this region than in the more accessible west.

FACTORS CAUSING DEPENDENCY ON DRUG TRAFFICKING Regarding Tajikistan as a transit country, there arise several paradoxes that differentiate it from other countries also struggling with drug trafficking. Firstly, one would assume that higher volumes of narcotics traded increase the level of drug-related crime. Nevertheless, this does not seem to be not the case: Since the deposition of the Taliban in 2002, which ended their ban on opium production, the flow of opiates through Tajikistan has steadily increased. The level of drug-related crime, however, has fallen against the expectations of the United Nations (UN) and other experts (Lewis, 2010, p.40). In fact, drug-related crime and convictions in Tajikistan were observed to be the lowest among the Central Asian countries (UNODC, 2008). As Lewis (2010) observes, though, the fact that fewer criminal incidents and convictions are reported does not necessarily mean that there is less crime: “Even where low-level drug crimes are uncovered, major trafficking figures are almost never arrested or charged.” Another paradox concerns seizure figures. While Afghan output – and thus the likely share of total output that is transported through Central Asia – has increased overall since the mid-1990s, the proportionate increase in seizure numbers lags behind the production figures (UNODC, 2012).2 Even though Tajikistan boasts the highest seizure rates in Central Asia – below four per cent

2. A comparative look at presumed output and seizure figures will be presented in the next section.

THE PUBLIC SPHERE | 2016 ISSUE

3. For example, see De Danieli (2011) where he talks about “narco-terror”; Engvall (2006), on the “crime-terror nexus”; or Chouvy (2011), on “From the ‘War on Drugs’ to ‘The War on Terrorism’” (p. 161).

makes it the highest rate –, this does not compare to India or Pakistan, which interdict roughly 15 per cent of the opiates transported through their territory (UNODC, 2012). This trend persists despite a steady supply of aid from the international community to help Tajikistan counter the flow of narcotics, which seems to lead to the conclusion that, as Lewis (2010, p.40) argues, “the more that’s spent on the problem, the deeper the problem becomes.” Lastly, a lesson learned from violent drug trends in Central America and other regions is that drug trafficking causes political instability and may be a source of funding for anti-government insurgency. This also does not seem to be the case in Tajikistan. Lewis (2010, p.40) argues that none of the extremist groups listed by the UN was active in Central Asia. Contrasting evidence suggests, though, that money from the drug trade does in fact fund Islamic extremist and other groups in the region.3 Tajikistan has done a good job in reducing poverty. Between 1994 and 2004, the proportion of the population affected by poverty dropped from 80 per cent to 32 per cent. This should, however, not obfuscate the fact that Tajikistan still does very poorly in terms of non-monetary policy, which expresses the country’s limited access to education, heating and proper sanitation (World Bank, 2015). The 7.5 average per cent growth of the Tajik economy over the last decade, combined with the increase in Afghan opium production and the volumes of opiates trafficked along the northern route, lead to the supposition that a considerable proportion of the economic growth was fuelled by drug money. The question that arises when assuming that Tajikistan is an “addicted economy”, however, is why are the symptoms of addiction that were described above so different from other transit countries?

The economic aftermath of the civil war The political power structures that exist in Tajikistan today were formed mostly during the period of the civil war of

[Tajikistan]

the 1990s. As Paoli et al (2007, p.954) describe in their article on the “narco-state”, “[d]espite being formally included in government structures, some of these civil servants and politicians-cum-trafﬁckers virtually remain warlords, disposing of their more or less private armies and ﬁefdoms.” The International Crisis Group (2001) points out that the problem of drug money corrupting the political system in Tajikistan exceeds that seen in other Central Asian states because both government and the opposition forces participated in the drug trade during the civil war. As a result of the 1997 peace treaty, many of these traffickers took over government positions. The ones that did not have been marginalized by the larger trafficking networks or, as has been suspected, fallen victim to purges by President Emomali Rahmon, who is said to be controlling individual trafficking networks operating in the country.

Corruption on all levels Corruption is high in Tajikistan. According to the yearly survey conducted by Transparency International, it continuously ranks in the lowest group – rank 152 or lower out of 174 countries in the last three years. The fact that many high-ranking civil servants are known to be involved in the drug industry, but very few of them are brought to court, is another indicator of the prevailing high levels of corruption in the executive and judicial branches. There are also a number of accounts that establish a direct link between the president and the trafficking enterprise: [f]oreign diplomats say they have no proof that he [Tajikistan’s President Emomali Rahmon] is involved in the drug trade, but they point out that those managing the smuggling routes owe their jobs to him or to other high-level officials. They maintain local security in exchange for the state’s hands-off approach… Although there are regular busts, they usually involve low-level upstarts with no

THE PUBLIC SPHERE | 2016 ISSUE

connections. In this way, diplomats say, officials protect their own trafficking networks, allowing no footholds for smaller players. (The Economist, April 21, 2012).

The case of Ghaffor Mirzoyev is particularly interesting. Formerly the influential head of the Presidential Guard, he later became head of the Drug Control Agency (DCA). At first a close ally of President Rahmon, Mirzoyev fell out with the president in what seems to have been a struggle for control over the drug trade. Mirzoyev was jailed for life in 2006 on a host of charges, including tax evasion, corruption, coup plotting and murder (Lewis, 2010). Testimony of diplomats and other foreign officials in Tajikistan points to corruption that is widespread, includes the highest levels of government, and serves to protect and develop the drug trafficking industry. Quantitative indicators – opium production and seizure figures, for example – confirm the tendency too, as will be outlined in the subsequent part of this analysis.

Ineffective border protection

4. More details on the source and scope of financial assistance in the field of counter narcotics will be provided in part III.

Even if law enforcement agencies and border protection units were not hijacked by political motives and ridden by corruption, they would still lack the equipment and skills to effectively tackle the activities of criminal groups (De Danieli, 2011, p.132). The nonexistence of adequate technological equipment to interdict (more) drugs flowing north from Afghanistan is particularly sobering in light of the substantial assistance payments coming from the United States, the European Union, and other Western countries.4 It should be noted again, though, that the 1,200km-long Afghan-Tajik border, by nature, is hard to patrol. In addition to inadequate equipment and border protection personnel, there are also organizational and structural factors that hinder border protection. As De Danieli (2011, p.133) describes,

[Tajikistan]

collaboration and coordination between different law-enforcement agencies is very weak, as competences are not clearly defined, and specific interests of single agencies often prevail over national priorities. Lack of coordination between central headquarters and peripheral outposts is another major problem.

De Daniele (2011, p.133) describes how low-level corruption has an influence as well. As soldiers and police officers at the border earn very low salaries, they can easily be incentivized to let large shipments of opiates pass the border unhindered. It can also be argued that counter narcotics policies, in the way they are executed now, merely perpetuate the problem while pretending to address it: this is because the increase in international assistance is affecting small players more than the powerful drug mafias, which is illustrated by the fact that the majority of people arrested on charges of drug trafficking were simple couriers, or even drug users (De Danieli, 2011, p.138). In this sense, disrupting activities of small “start-up” drug networks and couriers might actually favour operations of the established (and allegedly government-tolerated) networks.

Stabilizing effects of the drug trade on the economy In reality, there seems to have been little incentive to do something about the problem – for the Tajik administration and the West alike. For the political establishment in Tajikistan, many members of which are presumed to be affiliated with the drug trade, pursuing an alternative path seems to carry way too many risks: taking serious action against powerful figures in the industry – and therefore political players – could spark serious violence and possibly reignite the civil war. Western initiatives are aimed at reducing the volume of opiates trafficked on the northern route. However, as this essay will argue, the influx of drugs into Western nations is largely driven by demand,

THE PUBLIC SPHERE | 2016 ISSUE

not supply. In other words, if the drugs do not pass through Tajikistan, they will find another way. Some foreign officials have even admitted that they had little incentive to challenge the authorities’ hold on drug smuggling and that the status quo was preferable to taking a large amount of cash out of the economy, which could plunge the country into chaos (The Economist, April 21, 2012).

DETERMINING THE EXTENT OF DEPENDENCY ON THE DRUG TRADE Data on drug seizures, primarily compiled by the UNODC and its subsidiaries, is currently the main source for estimating the volumes of opiate traded worldwide. Unfortunately, this kind of data is flawed in a number of ways. As mentioned earlier, the scope of opium seized is only a small fraction of the traded volumes. No reliable relation can be drawn between seizures and the total volume of drugs trafficked. Even if Tajikistan consistently reports the highest seizure rates in the region – which, at four per cent of estimated flows, is still not particularly high – this does not mean the government is doing all it can. Leaders of the most powerful drug trafficking groups are believed to occupy high-ranking government positions, and seizures and other initiatives to control drug trafficking may actually serve the purpose of taking out rival or new trafficking networks (Lewis, 2010, p.44, and Paoli et al, 2007, p.952). The unreliability of seizure data is reflected in the large variations of opium and heroin seizures that can be observed on a regional level, both comparing regions with each other, and particularly comparing regions over time. While Dushanbe and the Republican Subordinated Regions, for example, witnessed drops in recorded seizures in 2006, the Khatlon province reported a large increase of 203 per cent in seizures (UNODC, 2008). The charts on the next page contrast the development of seizure figures, presumed Afghan opiate production, and the devel-

[Tajikistan]

opment of Tajikistan’s GDP, along with the development of foreign direct investment (FDI) in per cent of GDP and household final consumption spending, in terms of GDP.5 The objective is to uncover potential correlations between the different indicators and draw conclusions on the interconnectedness of the opiate trade and Tajikistan’s economic development in recent years.

5. Both indices follow the definition by the World Bank.

FIGURE 2 REPORTED SEIZURES OF OPIUM AND HEROIN IN TAJIKISTAN (IN KG)

Source: UNODC data FIGURE 3 POTENTIAL PRODUCTION OF OVEN-DRY OPIUM IN AFGHANISTAN (IN METRIC TONS)

Source: UNODC data

THE PUBLIC SPHERE | 2016 ISSUE

FIGURE 4 TAJIKISTAN GDP (IN MN. US$), FDI (% OF GDP), HOUSEHOLD SPENDING (% OF GDP)

Source: UNODC data

Contrasting Afghan opium production with Tajikistan’s GDP, the evolution of both indicators is not quite similar enough to establish a direct correlation. However, it can be observed that there is no onset of fundamental economic growth until 2002, which coincides with the lifting of the Taliban’s ban on opium production in late 2001 and the resumption of Afghan opiate production on a large scale. The new millennium also marks what many believe was the ultimate consolidation of President Rahmon’s power and influence not only over Tajikistan’s political apparatus, but also, allegedly, the drug trading industry. To that effect, it may or may not be a coincidence that figures from Tajikistan’s Drug Control Agency show an 80 per cent drop in opiate seizures since 2001 – which is also observable in the decreasing trend in opiate seizure figures by the UNODC in Fig. 2 – while during the same period, production in Afghanistan increased overall (The Economist, April 21, 2012).

[Tajikistan]

Foreign direct investment (FDI), or lack thereof, has been a particularly big problem for Tajik economic development. The rate of net FDI has been decreasing since 2006 and even turned negative in 2010. Negative FDI can be interpreted as revenues – part of which may have be generated through illicit activity – that are invested abroad (e.g. in property) or transferred to offshore accounts. In the absence of other sources of economic growth – government spending had been almost flat in the years before 2010, Tajikistan has been running trade deficits since 2000, and there were also no significant improvements in indicators of productive capacity (National Bank of Tajikistan, Trading economics data, 2015) – it is difficult to explain the steady economic growth since the early 2000s based solely on legitimate economic activity. The recent decline in the rate of household consumption, while experiencing economic growth, is peculiar for a developing economy and another indicator that there is excess income in the economy that is, if not saved, spent on things other than quantifiable consumer goods, and possibly reinvested in the trafficking enterprise. It should also be mentioned that there are other informal economic activities ultimately affecting peoples’ available income. The UNDP (2007) estimated that the size of the shadow economy from tax avoidance constituted 32.98 per cent of Tajik GDP in 2005, the volume of home production of food products for own consumption was 14.74 per cent of GDP, and the incomes from wages paid inkind and barter exchanges were 13.21 per cent of GDP. On the one hand, this implies that not all unaccounted economic growth can be traced back to revenues from drug trafficking. On the other hand, the UNDP (2007) report asserts that the volume of underreported gross value added was to some extent overlapping with the volume of illegal activities because at least some part of incomes from illegal activities, such as drug trafficking was likely to be spent on goods and services produced in Tajikistan. Therefore, by capturing the size of underpaid value-added tax, the UNDP figures were capturing some share of

THE PUBLIC SPHERE | 2016 ISSUE

income from illegal activities. Adding up the volume of underreported gross value added that is connected to properly registered legal enterprises and the value generated by illegal enterprises such as drug trafficking networks – which could evidently not be included in the UNDP sample –, the actual size of the Tajik shadow economy is likely to be much greater than the 60.9 per cent reported by the UNDP (2007). The next level of analysis will examine seizure figures. These numbers, as has been mentioned, need to be taken with a grain of salt. As De Danieli (2011, p.137) describes, seizure figures were often “artificially inflated by the Tajik authorities in order to show the donors that they are ‘doing well’.” The data itself refutes the claim that there is a reliable connection between production and seizures: the large regional variations – across districts and over periods- fail to correspond to the development of Afghan opiates output. Sudden switches between specific transportation routes, too, are unlikely to be responsible for such intense variations. The Drug Control Agency of Tajikistan (DCA), set up in 1999 with assistance from the UNODC, is responsible for these figures. The UN and individual UNODC donor countries, the Government of Tajikistan, and a number of experts in the field have since described it as an effective institution to counter the flow of narcotics through Tajikistan. Matveeva (2005) claims that “[s]election processes of the staff are transparent, and the agency does not have a reputation for corruption.” There is doubt, however, whether the DCA deserves the praise it has been getting. De Danieli (2011, p.137), for example, argues that many of the objectives that were stated when the DCA project was launched had not been achieved. The agency had been set up with the precise purpose of coordinating the national law-enforcement bodies involved in counter-narcotic operations, and stimulating positive competition among them. Instead, the DCA had transformed into a donor-oriented agency

[Tajikistan]

that was isolated from other domestic law enforcement bodies. Engvall (2006, p.849) goes further than that, saying that “the appointment of [Ghaffor] Mirzoyev as head of the DCA shows that internationally funded drug control is used as a tool in the domestic politics of the country.” The life-long jail sentence for Mr. Mirzoyev would appear to back up that claim.

Alternatives to seizure data Economists have begun to address the problem of weak data. The most notable contribution to date is “The World Heroin Market - Can Supply Be Cut?” by Letizia Paoli, Peter Reuter and Victoria Greenfield (2009b). By recruiting collaborators in eight nations that are important to understanding the global production and trafficking of opium and its derivatives – among them Tajikistan –, they aimed at building their own stock of qualitative and quantitative data on the heroin trade. The researchers also attempted to draw conclusions from the market reaction to the 2000-2001 Afghan ban on opium production (Paoli et al 2009a). One result of their research – and a central working hypothesis for this paper – is the figure they put on dependency of the Tajik economy of the drug trafficking industry. The authors conclude that in the case of Tajikistan, it was “unlikely that opiate trafficking adds less than 30 per cent to the recorded gross domestic product (GDP)” (Paoli et al 2007, p.952). This number might seem high. However, other scholars and officials involved in counter narcotics policy-making go even further than that. Matthew Kahane, United Nations Development Programme (UNDP) head in Tajikistan, for example, estimates that “the drug trade accounts for 30 to 50 percent of the economy” (Paoli et al 2007, 952). The World Bank also estimates that between 30 and 50 per cent of the country’s economic activity is linked to narcotics trafficking (Engvall, 2006, p.844). In addition, as will be outlined below and in the next section of the paper,

THE PUBLIC SPHERE | 2016 ISSUE

the shadow economy in Tajikistan, for which drug trafficking undoubtedly forms a strong foundation, is believed to make up more than half of the country’s GDP (UNDP, 2007).

An addicted shadow economy Looking at three years of steady economic growth – and not a whole lot to show for it –, the question arises of where the money has gone. It seems that the legitimate economy has not received a large share of it. This is reflected in the very low rate of private investment, which has been stagnating around five per cent, according to the World Bank (2013), and poses “the biggest challenge in the coming years.” A 2007 UNDP study concludes that in Tajikistan, 60.9 per cent of GDP constitutes activities attributable to the shadow economy. The study focuses only on unreported value added from tax evasion and unaccounted household incomes, but it assumes to be implicitly accounting for income stemming from illegal activities (UNDP, 2007). Dependency on drug trafficking constitutes the main reason why counter-narcotics measures must be implemented with great caution. As Paoli et al (2007, p. 971) explain in their article, “cracking down on heroin trafficking would result in a substantial deterioration of living conditions for a considerable segment of their impoverished people.” A key question for future research on drug trafficking in Tajikistan will be to determine the fraction of the Tajik population that profits from the industry – so that counter narcotics policies can be evaluated with a view to their potential destabilizing effects on the social fabric. Besides the individuals directly profiting from the trade, one also has to consider those profiting from it indirectly, e.g. by working in legitimate companies funded by drug money, or in service industries demanded by the drug traffickers. There is reason to believe that “the share of the Tajik

[Tajikistan]

population supplementing their legitimate income with opiate revenues is not insignificantâ&#x20AC;? (Paoli et al, 2007, p.971).

Demand drives production One might assume the heroin market is supply-driven because of the high margins on production and cultivation of opiates, but data indicates it is rather a demand-driven industry. Herein seems to lie a major flaw in traditional counter-narcotics policies. As Reuter describes it, the producing nations are blamed for their failure to control production; accusations that are sharpened by the corruption that is ubiquitous around drug production and by the large rewards that accrue to some developing country players in the trade (2008, p.1).

While it is true that the rewards are large in absolute terms, other parts further up the value creation chain profit much more from the handling of narcotics. The total value of the global opiates market (opium and heroin), in retail terms, amounted to an estimated US$65 billion in 2008 (UNODC, 2010). The regional breakdown in terms of value-creation is displayed in Fig. 5. As becomes apparent, half of the opiate market value is accrued in Europe and Russia, with China and North American being other attractive markets. The producers (Afghanistan) and traffickers (Tajikistan) do not even appear in the visualization. The value of the Afghan opium economy (farm-gate) was estimated at approximately US$2.4 billion. The bulk of the value of drug trading is in trafficking and dealing premiums, created in the demanding countries in Europe, Russia, China, and North America. What is thus necessary is an increased acceptance on the side of rich countries that it is actually domestic demand â&#x20AC;&#x201C; created by users, as well as traffickers and dealers â&#x20AC;&#x201C; that drives the industry.

THE PUBLIC SPHERE | 2016 ISSUE

EFFECTS OF TAJIKISTAN’S DEPENDENCY ON DRUG TRAFFICKING The costs of drug abuse A lesson learned from previous country cases of drug trafficking is that addiction and disease usually follow along on the trading routes. It is difficult to put a number on the current level of drug abuse in Tajikistan. Official figures are probably far off because few drug addicts will be registered. The UNODC in 2008 estimated that about 20,000 people in Tajikistan regularly use opiates (UNODC, 2008). Many experts, though, suggest this estimate is entirely too low. Engvall (206, p.842) goes so far as to say: taking into consideration the weak health infrastructure and the subsequent lack of tools for monitoring drug addiction in combination with the field work conducted by the UNODC, as well as the fact that Tajikistan is the major entry point for Afghan produced opiates smuggled northwards, it does not seem too speculative to assume that the country has one of the highest prevalence rates of opiate abusers in the world.

Two trends are particularly worrisome: the numbers are rising, and there is a strong tendency towards heroin, which is problematic in light of its extremely addictive and destructive potential. The UNODC reports that heroin has already become “the overwhelming drug of choice in all regions of the country,” with 81 per cent of Tajikistan’s drug abusers using heroin (UNODC, 2008). The prevailing form of abusing heroin is through injection, which explains another negative development: an upsurge in the number of HIV and Hepatitis infections throughout the region. The UNODC reports that Tajikistan has the fastest growing HIV rate in the region (UNODC, 2008). That there is a strong corre-

[Tajikistan]

lation between drug trafficking and HIV becomes obvious if one considers that 91 per cent of the cases booked during the period 1997-2001 were related to intravenous drug use (UNODC, 2003). Looking at the long-term spread of disease, however, policymakers often fall victim to a dangerous misconception: that, as long as the spread of HIV is restricted to the drug user community, it remains controllable. However, as it has been illustrated in India, Pakistan, Russia, and other former Soviet Republics where HIV has taken root way beyond this group, other forms of transmission can lead to explosive growth in infection rates and even lead to an HIV epidemic (Engvall, 2006, p.843, and Paoli et al, 2007, p.965). Engvall (2006, p.843) emphasizes that this is especially a threat for countries that lack the resources necessary for effective preventative measures or treatment. The author claims that: [t]he average cost of life-sustaining treatment for one AIDS patient is around $10,000 annually, however, in 1999 Tajikistan had a health budget of $13 per capita, and one commentator has pointed out that for the majority of the people, a diagnosis of AIDS would therefore mean a death sentence.

Just like the market for heroin is global, so are the problems that follow along with injecting drug use. The dangerous potential of heroin â&#x20AC;&#x201C; both addiction and the spread of disease- affects the final markets in Europe and Russia as well. A rising death toll and high treatment costs due to heroin addiction and exploding numbers of HIV infection among drug addicts make it obvious that the West cannot neglect the problem.

Drug trade strengthens the shadow economy Under the assumption that drug trafficking continues to add to the flourishing shadow economy, continued structural weakness

THE PUBLIC SPHERE | 2016 ISSUE

of the state and an uninviting business environment will remain prevalent. As the political risk consulting firm Oxford Analytica (June 23, 2011) describes in one of their briefs on Central Asia: The prevalence of informal economies will continue to have a deleterious impact on state-society relations, lowering public expectations and reinforcing the unaccountability of state bureaucracy. Blurred lines between public and private interests will reinforce rent-seeking behaviour at every level of government, and widespread tax evasion will reinforce this vicious circle by depriving formal state institutions of much-needed funding.

Little research has been done on the crowding-out of the legitimate economy that is caused by the drug trading industry and in how far it prohibits legitimate growth. It seems likely, though, that as long as the economy is so “addicted” to opiates, there is little incentive to build or produce anything else. With regards to building the state institutions required for an appealing economic environment, scholars have identified corruption inherent in the current drug-based economy as another major problem: Corruption emanating from criminally motivated special interests prevalent in the society is a major obstacle for developing a proper state underpinned by regular economic activity, where the output of public goods is a natural part of the relationship between the government and the public. (Engvall, 2006, p.832)

Likewise, De Danieli (2011, p. 131) states: in some situations, for example in the context of a post-conflict transition, when the government lacks the resources to impose and maintain order, the existence of this criminal-institutional partnership can represent a positive factor for stability, but in a long-term

[Tajikistan]

perspective it will preclude the building up of legitimate and accountable institutions. (De Danieli, 2011, p.131)

Ineffective and costly counter narcotics policies The implementation of counter narcotics policies in Tajikistan, especially under the cover of security assistance, is increasingly costly for the international community. Russia has traditionally been a strong partner for the Tajik government in the field of security. Russian troops, for example, were solely responsible for the protection of Tajikistanâ&#x20AC;&#x2122;s borders up until 2005, when they handed over responsibility to Tajik forces. Over the last 10 to 15 years, Tajikistan has become an area of key strategic interest for Western governments and organizations â&#x20AC;&#x201C; the UN, the United States and the European Union, in particular. For example, in 2007, of the US$45.2 million of total assistance by the U.S. government to Tajikistan, US$20.3 million were destined for law enforcement and security assistance (in: De Danieli, 2011, p.134). The European Union is also directly involved in counter narcotics assistance: two major projects, Border Management in Central Asia (BOMCA) and the Central Asia Drug Addiction program (CADAP), were launched in 2003. A European Community strategy paper talks about an amount of EUR85 million that has been allocated for activities in the area of Justice and Home Affairs under the Regional Tacis program, in addition to the â&#x201A;Ź27 million from the Central Asia program (2002-2006) (EC, 2007-2013).

In addition, both the EU and the United States invest on the multilateral level, for example through the UNODC or the International Organization for Migrations (IOM).

THE PUBLIC SPHERE | 2016 ISSUE

UNODC, IOM and UNDP are the most important agencies active in the field of counter narcotics. About 55 per cent of the total UNODC budget for Central Asia goes to projects in Tajikistan, and in particular for the project E24, “Strengthening control along the Tajik Afghan border” – approximately US$8 million – and the project H03, “Tajikistan Drug Control Agency” (approximately US$11.5 million) (UNODC, 2013). Donor governments and the involved international organizations praise the successes that have been brought about by these programs. However, as De Danieli (2011, p.130) argues, donors were lacking instruments to monitor the effective implementation of the projects. It has been mentioned how cases of corruption have been reported within the DCA (Engvall 2006, p.849), how seizure figures are often artificially inflated to please donor governments (De Danieli 2011, p.137), and how border protection affects the small players more than the powerful established trafficking networks. Overall, it seems that counter narcotics dollars are currently not spent wisely: Foreign donors have spent hundreds of millions of dollars on border security and counter narcotics projects designed to cut trafficking and boost Tajikistan’s legal economy. Instead, these initiatives help to control the trafficking by helping officials take out rivals. Donors subsidize basic government functions, including the drugs agency, while watching [President] Rahmon construct flamboyant new palaces and the world’s tallest flagpole (The Economist, April 21, 2012).

A BRIEF DISCUSSION OF COUNTER NARCOTICS POLICIES Experts point out that the traditional approach to counter narcotics policies, which is aimed at border enforcement and

[Tajikistan]

the strengthening of interdiction capabilities, fails to take into account the wider political, economic and social dimensions of the drug trade. Scholars like De Danieli believe that this current “military” focus leads to two unintended outcomes. The first is the so-called “militarization of the state,” which allows the government to “invest all the resources at its disposal in strengthening power ministries, such as the Interior and Security services” (De Danieli, 2011, p.136). It seeks to “ensure law and order, suppress domestic opposition and project enough fear to discourage expression of popular discontent” (Mateeva, 2005, p.154). Secondly, the efforts put in place by the international community in order to improve the capacities of Tajik law-enforcement agencies were paradoxically strengthening the political-criminal nexus (De Danieli, 2011, p.140). Paoli et al (2007, p. 965) hold that: progress in border control and law enforcement achieved thanks to international assistance has, on the one hand, favoured the emergence of strategic partnerships between drug mafias and governmental actors around the exploitation of shadow revenues, while on the other it had the effect of driving small players out of business.

So what are the alternatives? It is fair to assume that the market for opiates – in particular heroin – is largely demand-driven. However, the fact that just two countries (Afghanistan and Burma) dominate current production is evidence that programs focused on those countries could have some potential to reduce global production, and thus, consumption as well, if only in the short to medium term. Traditionally, three types of programs have been used to reduce source country supply: eradication involving either aerial spraying or ground operations, alternative crop development, aimed at increasing farmers’ earnings from other crop products, and in-country enforcement against refin-

THE PUBLIC SPHERE | 2016 ISSUE

ers, more vigorous pursuit of cultivators in order to increase their risks and production costs (Reuter, 2008, p.17). However, as Paoli et al (2009a, p.7 notice in their most recent paper, “the cliché that so long as the demand for an addictive drug exists there will be a supply is thus reasonably accurate as a long rung condition.” Policies aimed at cutting supply can thus not constitute a long-term response to the problem. Interdiction, too, is problematic, as the example of Tajikistan illustrates perfectly. Scholars like Reuter have therefore introduced non-traditional drug control methods into the discussion, which should be given some thought: a de-facto legalization of production and trafficking would probably serve to reduce corruption around the trade (Reuter, 2008, p.24) and help move the drug trading economy out of the shadows. However, as Reuter (2008, p.25) argues, developed nations would likely not agree with such a policy, the legalizing state would probably not be able to tax the industry effectively, and it would create an ambivalence to the state in enforcing generally agreed upon norms. Buying up the entire crop output is another option for Western governments that is sometimes suggested. While such a policy may be successful in making heroin much harder to obtain, heroin prices would naturally soar, making future buy-ups much more costly for Western taxpayers (Reuter, 2008, p.26). Moreover, production of opium poppy would likely increase as a response to the price hikes, thus rendering the policy counterproductive. Policies aimed at demand reduction could potentially be the most successful, such as, offering treatment and raising awareness of the deadly consequences of addiction. As long as demand is not suppressed, some nations, like Tajikistan today, will have to bear the consequences of the flow of narcotics through its territory. Drug production constitutes a “global public bad, like toxic waste disposal” (Reuter, 2008, p.26). Policy decisions, at least in the medium-term should therefore try and channel production

[Tajikistan]

and trafficking in ways that are bearable for the affected nations. It is clear that a small nation such as Tajikistan may be substantially more corrupted by a large trading volume of narcotics than a comparatively larger country like Pakistan. In this context, Reuter (2008, p.27) brings up compensation mechanisms that keep weak nations from seeking out such illicit industries. However, he does acknowledge that “there is something disturbing about such policy realism and it is not clear that it is a politically stable option” (Reuter, 2008, p.29).

CONCLUSION Thirty to 50 per cent of economic activity depends on the illicit trade (Paoli et al, 2007, p.952). Curing an economy so dependent on, and intertwined with, drug trafficking is a cumbersome and problematic process. There are certain positive effects on the Tajik economy stemming from drug trafficking: providing a certain degree of stability, incomes for the state, and economic opportunities for otherwise unemployed and poverty-stricken individuals. However, the negative consequences for Tajikistan, particularly in the long run, far outweigh these temporary benefits. Scholars have claimed that Tajikistan is a failed and a weak state, a condition that would give rise to the drug trade. This may be imprecise: with regard to drug trafficking, Tajikistan may be complicit, rather than weak. Undoubtedly, the state has failed in certain ways, but it is also highly functional in others. As it is, the drug trade – with the help of token counter narcotics initiatives that apparently only serve to make Western donors happy – fuels and stabilizes a sophisticated network of politicians, border control forces and couriers. If the proclaimed goal is to turn Tajikistan into a stable, legitimate economy, with functioning institutions that operate

THE PUBLIC SPHERE | 2016 ISSUE

according to transparent procedures, the West should work on controlling domestic demand for opiates, and bring real reforms to the field of counter narcotics policymaking for Central Asia. Further research on the dependency of the Tajik economy will be required in order to make sure that the measures implemented do not reinforce trafficking networks – as seems to be the case now –, or threaten the livelihood of a large part of the Tajik population, which could ultimately send the country spinning into another civil war.

REFERENCES Chouvy, P.-A. (2010). Opium: Uncovering the Politics of the Poppy. Cambridge: Harvard University Press. De Danieli, F. (2011). Counter-narcotics policies in Tajikistan and their impact on state building. Central Asian Survey, 30 (1), 129-145. The Economist (2012), Drugs in Tajikistan: Addicted. 21 April 2012. Retrieved from: http:// www.economist.com/node/21553092 Engvall, J. (2006). The State under Siege: The Drug Trade and Organized Crime in Tajikistan. Europe-Asia Studies, 58 (6), 827-854. European Community (2013). Regional Strategy Paper for Assistance to Central Asia. retrieved from: http://eeas.europa.eu/central_asia/rsp/07_13_en.pdf International Crisis Group (2001). Central Asia: Drugs and Conflict. retrieved from: http:// www.crisisgroup.org/en/regions/asia/central-asia/025-central-asia-drugs-and-conflict.aspx Lewis, D. (2010). High Times on the Silk Road: The Central Asian Paradox. World Policy Journal. Spring 2010, 139-149. Matveeva, A., Tajikistan: evolution of the security sector and the War on Terror. (2005). In: Ebnöther, A. et al (eds.). Facing the terrorist challenge: Central Asia’s role in regional and international cooperation. Geneva: DCAF. Oxford Analytica (2011). Daily Briefs: Kyrgyz and Tajik shadow economies impede development. 23 June 2011.retrieved from: http://www.oxan.com/display.aspx?ItemID=DB168930 Paoli, L., Greenfield, V., and Reuter, P. (2009a). Modelling the World Heroin Market: Assessing the Consequences of Changes in Afghanistan Production, Project Summary for the Netherlands Ministry of Justice and Scientific Research Documentation Centre (WODC). Paoli, L., Greenfield, V., and Reuter, P. (2009b). Modelling the World Heroin Market: Can Supply be Cut? Don Mills: Oxford University Press Canada. Paoli, L., Rabkov, I., Greenfield, V., and Reuter, P. (2007). Tajikistan: The Rise of a Narco-State. Journal of Drug Issues, 37 (4), 951-979.

[Tajikistan]

Reuter, P. (2008). Can production and trafficking of illicit drugs be reduced or merely shifted?, Policy Research Paper, The World Bank Development Research Group. Rubin, A. (2012). Opium Cultivation Rose This Year in Afghanistan, U.N. Survey Shows, The New York Times. 20 November 2012. retrieved from: http://www.nytimes. com/2012/11/21/world/asia/afghan-opium-cultivation-rose-in-2012-un-says.html?_ r=0 Trading Economics (2015). Economic Indicators Tajikistan. retrieved from: http://www. tradingeconomics.com/tajikistan Transparency International (2015). Corruption by country: Tajikistan. retrieved from: http://www.transparency.org/country#TJK UNODC (2011). Estimating illicit financial flows resulting from drug trafficking and other transnational organized crimes. retrieved from: http://www.unodc.org/documents/ data-and-analysis/Studies/Illicit_financial_flows_2011_web.pdf UNODC (2008). Illicit Drug Trends in Central Asia. retrieved from: http://www.unodc.org/ documents/regional/central-asia/Illicit%20Drug%20Trends_Central%20Asia-final. pdf UNODC (2003). The Opium Economy in Afghanistan: An International Problem. retrieved from: http://www.unodc.org/pdf/publications/afg_opium_economy_2003.pdf UNODC (2010). Tajikistan: Drug Abuse and Drug Dependence Treatment Situation. retrieved from: http://www.unodc.org/docs/treatment/CoPro/Web_Tajikistan.pdf UNODC (2013). Tajikistan: Ongoing Projects. retrieved from: http://www.unodc.org/centralasia/en/ongoing-projects/drug-trafficking-and-border-control.html UNODC (2010). World Drug Report 2010. retrieved from: https://www.unodc.org/documents/wdr/WDR_2010/1.2_The_global_heroin_market.pdf UNODC (2012). World Drug Report 2012. retrieved from: http://www.unodc.org/documents/data-and-analysis/WDR2012/WDR_2012_web_small.pdf UNODC (2012). World Drug Report 2014. https://www.unodc.org/documents/wdr2014/ World_Drug_Report_2014_web.pdf U.S. State Department (2003). International Narcotics Control Strategy Report: Europe and Central Asia. retrieved from: http://www.state.gov/j/inl/rls/nrcrpt/2003/vol1/ html/29838.htm World Bank (2013). Country Overview Tajikistan. retrieved from: http://www.worldbank. org/en/country/tajikistan/overview World Bank (2015). Country Overview Tajikistan. retrieved from: http://www.worldbank. org/en/country/tajikistan/overview

Getting out of poverty for the first time Evidence from the United Kingdom David Escamilla-Guerrero

MSc in Development Economics Class of 2011, Universidad Carlos III MSc in Economics, Class of 2015 University of Essex PhD candidate in Economic History, London School of Economics and Political Science

ÂŠi4lcocl2

ABSTRACT For the period 2000 - 2007, married women living in London with some education had considerably higher chances of getting out of poverty for the first time. This reflects that, in the UK, transitions out of poverty might be determined by personal decisions when poverty is identified with a monetary-relative threshold. In consequence, policy makers must design economic incentives for the poor and influence the reduction of poverty through market mechanisms.

[Getting out of poverty for the first time]

INTRODUCTION Poverty is one of the most important social problems in the world. For this reason, knowing its magnitude, location and mechanisms that perpetuate it is essential to design, implement and evaluate policies aiming its reduction. In general, it is agreed that poverty is a condition that violates the dignity of the human beings, limits their freedoms, prevents the satisfaction of their basic needs and precludes their social inclusion. The objective of this paper is to identify the factors that determine the transition out of poverty for the first time in the United Kingdom (UK) and highlight the role of personal decisions when poverty is measured from a relative approach. To identify these drivers, first it is necessary to measure this phenomenon. The following section presents a general review of select poverty measures and their characteristics. The second part offers some results using the British Household Panel Survey (BHPS).1 In the third section, I develop an econometric tool to identify the factors that might influence poverty transitions. The fourth section concludes and offers some policy perspectives.

1. A prospective indefinite life panel survey without sample replacement in which people are reinterviewed annually.

MEASURING POVERTY Traditionally, poverty is considered a one-dimensional phenomenon. From this perspective, money, through income or consumption, is the sole criterion to identify the poor. As a result, this framework considers monetary resources to be the unique instruments to mitigate poverty. However, approaching poverty from a multi-factorial perspective has caught the attention of researchers and policy makers recently. A multidimensional analysis, in contrast with the one-dimensional perspective, incorporates non-monetary aspects of life. It considers some important deprivations related

THE PUBLIC SPHERE | 2016 ISSUE

to education, health, social security, housing and food (CONEVAL, 2010; Kakwani and Silber, 2008). Hence, the aspects of poverty that most affect the population can be identified with this approach. Regardless of how poverty is defined, measuring it involves two main processes: the identification of the poor, and the aggregation of their characteristics (Sen, 1976). Identifying the poor requires the comparison of the wellbeing of an individual against a fixed parameter that is considered to be a ‘reasonable minimum’, normally called ‘poverty line’ (Feres and Mancero, 2001). Regarding the aggregation process, there are some desirable axioms that any measure of poverty must satisfy.

Poverty indices and robustness axioms To be considered a robust measure, any increase in the income of a non-poor person must not affect the overall poverty index (Focal Axiom, Anand, 1977). This axiom guarantees that a measure of poverty is only sensitive to changes in the income distribution of the poor population. Also, a measure of poverty should increase when the income of a poor person decreases (Monotonicity Axiom, Sen, 1976). Another condition that must be satisfied is the Transfer Axiom (Sen, 1976). It implies that a transfer of income from any person to a less poor individual should increase the poverty index. This means that the measure should reflect how wealth is distributed among the poor. In addition, a measure of poverty must increase if the poverty level of a particular group of people increases. (Subgroup Monotonicity Axiom, Foster, Greer and Thorbecke, 1984). The above axioms arose from the constant improvements of various measures, some of which are the following. The headcount index measures the share of population below the poverty line – that is, the incidence of poverty. While this measure is easy to interpret, it only fulfills the focal axiom.

[Getting out of poverty for the first time]

A measure of poverty that satisfies the focal axiom and monotonicity axiom is the poverty gap index. This measure increases if a poor personâ&#x20AC;&#x2122;s income decreases. However, it does not satisfy the transfer axiom. Finally, the FGT index is an index that can be interpreted as a poverty gap which assigns a greater weight to the poor that are farther away from the poverty line. Unlike the previous index, it adds an â&#x20AC;&#x2DC;inequality aversionâ&#x20AC;&#x2122; factor, captured with the coefficient of variation (CV). This factor gives greater importance to the poorer than those who are closer to the poverty line, i.e. the higher the inequality aversion, the greater the weight given to the poor. In this way, the FGT index satisfies all the previous axioms (Foster, Greer and Thorbecke, 1984).

POVERTY IN THE UK Despite its intellectual attractiveness, a multidimensional measurement requires the definition of the poverty dimensions as well as the theoretical justification of their thresholds. It is necessary to define how these dimensions will interact and to specify their relative weights. Also, all the previous axioms must be proved for in each dimension and for the overall measure (Alkaire and Foster, 2011). The aim of this research is not to propose a measure of poverty, but to study the dynamics underlying this phenomenon. Thus, the following results are in regard to the one-dimensional approach. Nonetheless, future research can verify whether they hold when considering the multiple dimensions of poverty. I use the monthly per capita income as a proxy for well-being to avoid seasonality effects, normally experienced when using consumption data.2 This number is obtained by dividing the annual household income into even monthly tranches. The BHPS does not capture informal wages, but this is not highly relevant

2. The income is obtained from the annual household income after applying economies of scale.

THE PUBLIC SPHERE | 2016 ISSUE

3. Information available at: www. stats.oecd.org.

since the non-observed economy in the UK is estimated at 2.3 per cent of the gross domestic product (GDP) (OECD, 2014). Because we only need to identify the poor and the introduction of an inequality weight is not necessary for the purposes of this research, the poverty measure used is the headcount index. The threshold or poverty line selected is 60 per cent of the median income of each year. This is the most widely quoted measure of income poverty in the UK and the rest of the European Union (Belfield, 2014). A relative parameter is necessary since the study covers a high-income developed country. If we choose a threshold under the absolute approach, like the 2.5 daily U.S. dollars, developing this study would be meaningless due to the lack of poverty in the UK, by this measure. The period under analysis covers eight years, from 2000 to 2007. Since I am measuring poverty through income, it is necessary to choose a period assuring macroeconomic stability. This is important because the loss of employment and income are natural phenomena under economic contractions. Thus, an increase of poverty would be expected during these periods, especially if it is measured through monetary variables. Future research might analyse poverty transitions in periods with macroeconomic shocks and contrast the findings of this paper. The period chosen offers a positive gross domestic product (GDP) annual growth rate (3.0 per cent) as a well as a constant inflation rate (1.5 per cent).3 This ensures that any transition out of poverty could be associated with sociodemographic determinants and not to macroeconomic shocks. Before addressing the determinants of poverty transitions, it is important to know some of its characteristics. In 2000, the value of the poverty threshold was GBP 666 per month, while in 2007, it was GBP 905 pounds. This increase of 39 per cent reflects an improvement in the income per capita of the overall population. Nevertheless, the share of poor was constant during the period

[Getting out of poverty for the first time]

(on average 21 per cent). Its highest level was in 2004, when 22.3 per cent (13.4 million) lived below the poverty line (Table 1). Table 1 Poverty in the UK (millions) YEAR - WAVE 2000

2001

2002

2003

2004

2005

2006

2007

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

Not poor

46.7

47.0

46.8

47.2

46.6

48.0

48.3

48.1

79.2

79.5

78.9

79.3

77.7

79.4

79.5

78.5

Poor

12.2

12.1

12.6

12.4

13.4

12.4

12.5

13.2

20.8

20.5

21.1

20.7

22.3

20.6

20.5

21.5

Total 58.9 Source: Author.

59.1

59.4

59.6

60.0

60.4

60.8

61.3

Looking at the characteristics of the poor as proportions across regions and demographics, it is possible to observe that it is not a homogeneous phenomenon. For instance, looking at the regions in the UK, the West Midlands Conurbation and South Yorkshire had the highest proportions of poverty. In addition, the statistics confirm that females presented higher rates of poverty than males during the period (Table 2). Table 2 Proportions of poor population by sociodemographic characteristics YEAR - WAVE 2000 (1)

2003 (4)

2007 (8)

1. Inner London

17.6

18.9

13.4

2. Outer London

13.5

16.7

23.9

3. South East

14.7

14.3

15.2

4. South West

20.7

21.6

19.8

5. East Anglia

22.6

22.8

29.2

Region

continue

THE PUBLIC SPHERE â&#x20AC;&#x2020;| â&#x20AC;&#x2020;2016 ISSUE

Table 2 Proportions of poor population by sociodemographic characteristics (continuation) YEAR - WAVE 2000 (1)

2003 (4)

2007 (8)

6. East Midlands

23.5

7. West Midlands Conurbation

32.2

35.6

33.1

8. West Midlands

19.5

19.3

20.0

9. Greater Manchester

25.5

29.2

21.3

10. Merseyside

21.2

21.3

28.1

11. North west

16.9

16.1

23.1

12. South Yorkshire

28.8

26.5

28.9

13. West Yorkshire

21.8

25.5

28.9

14. Yorks & Humberside

22.3

23.2

28.0

15. Tyne & Wear

34.4

25.9

18.4

16. North

24.5

19.2

20.9

17. Wales

21.6

22.2

21.1

18. Scotland

26.4

24.8

20.1

19. UK

20.8

20.7

21.5

1. Male

16.2

17.3

18.1

2. Female

24.9

23.8

24.4

1. Married

15.0

15.7

16.2

2. Living as a couple

11.1

14.9

16.6

3. Widowed

53.0

49.1

45.1

4. Divorced

42.1

34.6

36.0

5. Separated

34.8

37.1

33.0

6. Never married

19.2

19.7

22.3

Sex

Marital status

Source: Author.

[Getting out of poverty for the first time]

The marital status variable shows that almost half of the widowed population was poor in 2007. Similarly, the divorced and separated individuals had high rates of poverty (36 and 33 per cent). In contrast, only 16.2 per cent of the persons living as a couple were below the poverty threshold. The previous results only describe some characteristics of the poor that might explain poverty dynamics. The following section addresses this matter using a balanced panel with eight waves (2000â&#x20AC;&#x201C;2007). This structure allows us to follow the same individuals over the period and identify which socioeconomic and demographic characteristics could explain their transition out of poverty.

GETTING OUT OF POVERTY FOR THE FIRST TIME In life, many events or decisions change our present and future economic well-being. Sometimes, culture and social environments influence the opportunities each person will have. But, due to discrimination or inequality, demographic characteristics (sex, age, etc.) or disabilities may determine the well-being of an individual. Besides decisions and sociodemographic characteristics, there are structural conditions that impact poverty dynamics. Iceland (1997) argues that restructures in the economic markets, skills mismatches, racial residential segregation and changes in welfare benefit levels could explain the transitions out of poverty. Using data from the United States, he finds that the previous factors substantially impact African Americans, while their effects are weak for the white population. Four types of multivariate models are commonly used when studying poverty dynamics: 1) longitudinal poverty pattern models, 2) transition probability models, 3) income variance component models and 4) structural models (Andriopoulou and Tsakloglou, 2011). To explore the determinants that influence

THE PUBLIC SPHERE | 2016 ISSUE

the transition out of poverty, I implement two logit models. A logit model is used when the dependent variable is binary and it estimates the probability of observing an event. The two models implemented identify the factors that might affect the likelihood of getting out of poverty for the first time, conditional to individual sociodemographic characteristics. The binary transition indicator (dependent variable) used is defined as getting out and not getting into poverty because the latter might capture an inequality effect. When using a relative measure of poverty, leaving everything else equal, a person could become poor due to an increase in the rich population. In contrast, the former captures the individuals that registered sufficient income to be considered non-poor despite the income distribution of each wave. Measuring the transition out of poverty (‘getting out’) allows the identification of the socioeconomic characteristics of those who increased their income enough to be above the poverty threshold, regardless of the inequality level. The regressors capture some factors that might modify the monthly income per capita. Income is not included in the independent variables because it is the unique variable considered in the poverty identification process. Hence, including it would produce a type of endogeneity known as reverse causality or twoway causation. These regressors estimate the probability of getting out of poverty conditional on the age, marital status, highest qualification, health condition, type of job contract, hours worked per week, saving behavior, amount saved and region of residence. This group of explanatory variables represents voluntary decisions and individual characteristics that could influence the probability of overcoming poverty. Following Lichter (1992), the regressors were lagged to avoid simultaneity bias (when one or more of the independent variables is jointly determined with the dependent variable). The population under 18 years old was excluded and only individuals

[Getting out of poverty for the first time]

interviewed personally were considered. Controlling by gender is relevant because in the UK, 60 per cent of the poor were women; hence, it is desirable to analyse separate cohorts to understand the different drivers behind the gender gap in poverty. Generic index models (i.e. for any data set-up) are specified following Wooldridge (2010): P[y = 1|x] = G(xβ)

(1)

where x is 1 × K, β is K × 1, and G(·) takes on values in the open . unit interval: 0 < G(z) < 1, for all z In logit models, G is a logistic function that arises from an underlying latent variable model: y* = xβ + e (2)

Thus, the specific binary choice model used in this research, which uses a panel setup, can be specified as: P[Mit = 1|xit] = P[xitβ + eit ≥ 0|xit] (3)

where Mit {0,1} identifies the transition out of poverty for the first time. xit represents the regressors vector and eit the unobserved characteristics, which have a standard logistic distribution (Wooldridge, 2010). i identifies the individual and t the wave or year. The regressors are: 1) age: age at the moment of the interview; 2) married: 1 if married or living in couple and 0 otherwise; 3) job: 1 if the employment is permanent and

THE PUBLIC SPHERE â&#x20AC;&#x2020;| â&#x20AC;&#x2020;2016 ISSUE

0 otherwise; 4) save: 1 if the person saves from current income and 0 otherwise; 5) health: a five multinomial variable that captures the number of visits to the doctor, where 1 is none and 5 is more than ten; 6) Q: a five multinomial variable that captures the highest qualification, where 1 is high school degree, 2 is university diploma, 3 is advanced education level, 4 is ordinary education level, 5 is none; 7) InLon: 1 if living in inner London and 0 otherwise; and 8) jbhrs: number of hours worked per week. The marginal effects estimations in Table 3 show that age does not impact the chances of getting out of poverty for men or women. For females, being married increases the probability of getting out of poverty for the first time by 11 per cent. This result may reflect the gains from marriage (economies of scale, complementarities of inputs, risk diversification, etc.) argued by Becker (1973). These gains are not observed for men, since the variable married is not significant in the second model. Moreover, these findings complement the results of Jarvis and Jenkins (1999). Using the first four waves of the BHPS, they argue that marital splits generate declines in income for separated wives, while for men this effect is much lower. Table 3 Getting out of poverty for the first time Marginal effects (dy/dx) WOMEN

MEN

(Model 1)

(Model 2)

Lag_age

-0.000

0.001

(0.001)

(0.002)

1.Lag_married

0.110***

0.038

(0.032)

(0.050)

-0.018

-0.023

(0.057)

(0.069)

Out of poverty

1.Lag_job

continue

[Getting out of poverty for the first time]

Table 3 Getting out of poverty for the first time (continuation) Marginal effects (dy/dx) Out of poverty 1.Lag_save

2.Lag_health

3.Lag_health

4.Lag_health

5.Lag_health 1.Lag_Q

2.Lag_Q

3.Lag_Q

4.Lag_Q

1.Lag_In_Lon

Lag_jbhrs

Observations

WOMEN

MEN

(Model 1)

(Model 2)

-0.007

0.028

(0.031)

(0.048)

0.001

-0.044

(0.039)

(0.046)

-0.028

0.060

(0.044)

(0.060)

0.041

0.012

(0.054)

(0.087)

-0.081

0.090

(0.062)

(0.113)

0.277***

0.367***

(0.077)

(0.073)

0.304***

0.302**

(0.071)

(0.121)

0.187***

0.284***

(0.059)

(0.067)

0.138***

0.182***

(0.043)

(0.063)

0.325***

0.061

(0.107)

(0.125)

0.009***

(0.001)

1,134

513

Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 Source: Author.

THE PUBLIC SPHERE â&#x20AC;&#x2020;| â&#x20AC;&#x2020;2016 ISSUE

The regressors job, save and health are not significant in either model. It seems to be that the job contract nature does not influence the probability for being above the poverty line. Surprisingly, the decision to save is not significant either. Savings ease off temporal shocks and could be invested in capital or education. Looking at the distribution of this decision, on average, 24 per cent of the poor saves from their current income. Regarding the non-poor population, 46 per cent saves, which shows that there is variation in savings among these populations, but it does not influence the exit of poverty for the first time. Some of the health proxies capture a negative impact on the probability of getting out of poverty but none of them are significant. Following Jenkins and Rigg (2003), employment rates and average income fall with disability, and they continue to fall the longer the spell lasts (the disability condition could be caused by illness or accident). In this sense, a higher number of visits to the doctor reflects more constraints for developing labor activities, reducing the chances for being above the poverty line. Regarding the education variables, all are significant and robust for men and women. This result suggests that education may be the best instrument to increase the chances of getting out of poverty for the first time. For instance, earning a university degree increases the probability for overcoming poverty by 27 per cent (women) and 36 per cent (men). Consistently, these effects are smaller in lower qualifications for which these results are on average 6.2 per cent higher for men, reflecting the presence of gender inequality (Figure 1). The regional factor is significant but only for women. This result confirms that living in a region with lower shares of poverty increases the probability of overcoming it. In addition, being a resident of a region with a dynamic economy, such as inner London, provides more available income sources and more diverse economic activities. In this case, women living in inner London have 32.5 per cent higher chance to exit from poverty.

[Getting out of poverty for the first time]

.5 .4 .3 .2 .1

Effects on Pr(First time getting out of poverty)

FIGURE 1 AVERAGE MARGINAL EFFECTS OF EDUCATION LEVEL

1.Lag_Q

2.Lag_Q

3.Lag_Q

4.Lag_Q

1.Lag_Q

.5 .4 .3 .2 .1 0

Effects on Pr(First time getting out of poverty)

2.Lag_Q

3.Lag_Q

4.Lag_Q

1.Lag_Q

2.Lag_Q

3.Lag_Q

4.Lag_Q

Education level - women

- Men

3.Lag_Q

Education level - wome

Education level - Men

Source: Author.

The decision to prioritise labor over leisure is an individual choice that does not always impact the income of a person. The re-

THE PUBLIC SPHERE | 2016 ISSUE

sults show that the number of hours worked are significant and positively impact the probability of being above the poverty threshold. The evaluation of this effect at a different level shows that it starts to disappear after a certain number of hours. For men, working 35 hours per week increases the probability of getting out of poverty by 1.0 per cent while working 45 generates a 0.8 per cent effect. For women, working eight hours more (from 8 to 16) only increases the probability of overcoming poverty by 0.1 per cent (Figure 2). Certainly, these marginal effects suggest that the number of hours a person works matters but the quality of the job and the salary received is more important. In the sample, on average, the poor work eight hours less than the population living above the poverty line. However, the previous results neglect heterogeneity across individuals. This can be corrected with a model allowing for unobserved individual effects that are currently in the error term, eit, which is composed of two parts: an effect that does not vary across time ci, and an idiosyncratic error uit. The first can be drawn out and (2) can thus be written as: Yit* = xit β + ci + uit. (4) Furthermore, it is possible to obtain consistent estimators by assuming that: 1. ci and xit are independent: Cov(xit,c_i )=0, t=1,2,…,T 2. ci has a normal distribution with mean zero and variance , and 3. the regressors’ vector is strictly exogenous. In econometric parlance, the previous implies that ci is an ‘individual random effect’ (Wooldridge, 2010). The random effects capture favourable characteristics (willingness to escape poverty, capacities or social networks) that are difficult to observe and

[Getting out of poverty for the first time]

which make it easier for some individuals to overcome poverty (i.e. unobserved heterogeneity).

.014 .012 .01 .006

.008

Effects on Pr(First time getting out of poverty)

.01 .005

Effects on Pr(First time getting out of poverty)

.015

FIGURE 2 AVERAGE MARGINAL EFFECTS OF HOURS WORKED

Hours worked per week (men)

.014 .012 .01 .008

Effects on Pr(First time getting out of poverty)

.006 35

r week (men)

Hours worked per week

37 39

Hours worked per week (women)

Source: Author.

THE PUBLIC SPHERE â&#x20AC;&#x2020;| â&#x20AC;&#x2020;2016 ISSUE

4. Further information available at: http://www.stata. com/manuals13/ xtxtlogit.pdf .

In STATA, the package xtlogit fits random effects in logistic models. In other words, it includes an individual-specific random intercept ci in the linear predictor to obtain a random-intercept logistic regression model (Rabe-Hesketh and Skrondal, 2012).4 Table 4 presents the final results, which corroborate the sign and significance of the previous results. Table 4 Getting out of poverty for the first time Random effects Out of poverty Lag_age 1.Lag_married 1.Lag_job 1.Lag_save 2.Lag_health 3.Lag_health 4.Lag_health 5.Lag_health 1.Lag_Q

WOMEN

MEN

(Model 1)

(Model 2)

0.001

0.009

(0.008)

(0.012)

0.536***

0.265

(0.187)

(0.273)

-0.012

-0.003

(0.302)

(0.398)

-0.060

0.145

(0.181)

(0.267)

-0.002

-0.329

(0.222)

(0.259)

-0.197

0.176

(0.247)

(0.348)

0.155

-0.106

(0.299)

(0.480)

-0.492

0.325

(0.364)

(0.715)

1.769***

1.991***

(0.424)

(0.494)

2.Lag_Q

1.571***

1.746**

(0.435)

(0.691)

3.Lag_Q

1.097***

1.471***

(0.331)

(0.378) Continue

100

[Getting out of poverty for the first time]

Table 4 Getting out of poverty for the first time (continuation) Random effects Out of poverty 4.Lag_Q 1.Lag_In_Lon Lag_jbhrs

Observations

WOMEN

MEN

(Model 1)

(Model 2)

0.789***

0.925***

(0.264)

(0.329)

1.745**

0.354

(0.827)

(0.744)

0.054***

0.053***

(0.009)

(0.012)

1,134

513

Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 Source: Author.

Future research addressing the effect of personal decisions on poverty dynamics must consider the time lapsed before exiting poverty (time effects) as an explanatory variable. For example, using the European Community Household Panel, Andriopoulou and Tsakloglou (2011) found a negative relationship between the time spent into/out poverty and the probability of exiting/ re-entering poverty.

CONCLUSION For the period 2000â&#x20AC;&#x201C;2007, it is clear that education was the main factor that increased the probability of getting out of poverty for the first time. For women, earning a diploma in any discipline implies increasing the chances of being above the poverty threshold by 12 per cent. Most of the diploma courses in the UK last one year and several can be studied on a part-time basis. Thus, any policy increasing the populationâ&#x20AC;&#x2122;s educational level could generate positive spillovers and ameliorate the overall social wellbeing. This result echoes previous findings arguing that

101

THE PUBLIC SPHERE | 2016 ISSUE

education reduces poverty persistence and entry rates (Cappellari and Jenkins, 2002). The geographical distribution of poverty is an important factor that must be considered as well. London has the lowest shares of poverty as compared to the rest of the UK, but overcoming poverty in London is easier than in any other region, at least for women. This fact is important to design sustainable policies over time, since it is desirable to promote egalitarian development and avoid the concentration of poverty in specific locations. Even though the persistence of illness constrains the integral development of an individual, a well-functioning health and social security system can offset these events. This seems to be the case in the UK because no evidence was found regarding the impact of illness on the probability of exiting poverty. Measuring poverty through a relative monetary approach precludes the use of income as a main driver of poverty transitions. If used, the presence of different types of endogeneity could generate biased estimates. The implementation of logistic regressions using lagged explanatory variables that affects the income level avoids endogeneity but at the same time recognises the income effect. Finally, by not including monetary independent variables, it is possible to confirm that poverty is a multidimensional phenomenon and the probability of overcoming it is influenced by personal decisions, market constraints and specific individual preconditions. Thus policy makers must design economic incentives for the poor and influence the reduction of poverty through market mechanisms.

REFERENCES Alkaire, S. and Foster, J. (2011). “Counting and multidimensional poverty measurements”, Journal of Public Economics, Vol. 95, Issues 7-8, pp. 476-487. Anand, S. (1977). “Aspects of Poverty in Malaysia”, Review of Income and Wealth, Vol.23, pp. 1-16.

102

[Getting out of poverty for the first time]

Andriopoulou, E. and Tsakloglou, P. (2011). “The Determinants of Poverty Transitions in Europe and the Role of Duration Dependence”, IZA Discussion Paper Series, No. 5692, pp. 1-38. Becker, G. (1973). “A Theory of Marriage: Part I”, The Journal of Political Economy, Vol. 81, No. 4, pp. 813-846. Belfield, C. et al. (2014). “Living Standards, Poverty and Inequality in the UK: 2014”, Institute for Fiscal Studies, pp. 1-123. Cappellari, L. and Jenkins, S. (2002). “Who stays poor? Who becomes poor? Evidence from British Household Panel Survey”, The Economic Journal, Vol. 112, No. 478, pp. C60-C67. CONEVAL. (2010). “Metodología para la medición multidimensional de la pobreza en México, CONEVAL, pp. 1-122. Feres, J. and Mancero, X. (2001). “Enfoques para la medición de la pobreza. Breve revisión de la literatura”, Estudios estadísticos y prospectivos, No. 4, CEPAL, pp. 1-47. Foster, J.; Greer, J. and Thorbecke, E. (1984). “A class of descomposable Poverty Measures”, Econometrica, Vol. 52, pp. 761-766. Iceland, J. (1997). “Urban Labor Markets and Individual Transitions Out of Poverty”, Demography, Vol. 34, No. 3, pp. 429-441. Jarvis, S. and Jenkins, S. (1999). “Marital splits and income changes: Evidence from the British Household Panel Survey”, Population Studies, Vol. 53, No. 2, pp. 237-254. Jenkins, S. and Rigg, J. (2003). “Disability and disadvantage: Selection, onset, and duration effects”, CASE Paper, No. 74, Center for Analysis of Social Exclusion, pp. 1-27. Kakwani, N. and Silber, J. (2008). “Quantitative Approaches to Multidimensional Poverty Measurement”, Palgrave Macmillan, xviii. Lichter, D. et al. (1992). “Race and the retreat from marriage: A shortage of marriageable men?, American Sociological Review, Vol. 57, No. 6, pp. 781-799. OECD (2014). “The Non-Observed Economy in the System of National Accounts”, Statistics Briefs, No. 18, pp. 1-12. Sen, A. (1976). “Poverty: An Ordinal Approach to Measurement”, Econometrica, Vol. 44, pp. 219-232. Sen, A. (1984). “Values, Resources and Development”, Harvard University Press, Cambridge. Wooldridge, J. (2010). “Economic Analysis of Cross Section and Panel Data”, The MIT Press, pp. 561-630.

103

The political economy of food subsidies in Egypt Reforms and strengthening of social protection Flavia Lorenzon

MSc in Political Economy of Late Development, Class of 2015 London School of Economics and Political Science

ÂŠNick Scheerbart

ABSTRACT Egyptâ&#x20AC;&#x2122;s widely used food subsidies have been controversial since they were first introduced in 1941 by President Gamal Nasser. Subsidies are implicitly nontargeted transfers that reduce the risk of food insecurity. At the same time, they impose a heavy fiscal burden on governments. Over time, the majority of Egyptians have come to see food subsidies as the most concrete social benefit received from the government. Therefore, despite the fact that the current food subsidies system inadequately targets the poorest part of the population, past attempts at reform have proven difficult to implement. This paper has two objectives: 1) to assess historical linkages between the two main food subsidies programmes and socio-political stability in Egypt, and 2) to present alternatives for improving Egyptâ&#x20AC;&#x2122;s food-based social safety nets. This study finds that better targeting in lieu of eliminating food subsidies is the optimal option to avoid social unrest. In addition, this paper concludes by analysing the alternative safety nets and their limitations given the political context.

[The political economy of food subsidies in Egypt]

INTRODUCTION On January 18 and 19, 1977, a country-wide protest unfolded in the streets of Egypt’s main cities following President Anwar Sadat’s decision to implement relatively small changes to the food subsidy system at the urging of international actors such as the World Bank and the International Monetary Fund (IMF). At that time, nine out of ten Egyptians lived in poverty, and the male unemployment rate reached 50 per cent of its population (Staff, 2011). The so-called ‘food riot’ was likely triggered by a growing dissatisfaction with poor standards of living, lack of opportunities and most importantly, the strong popular perception that food subsidies were the main service the government could offer. Protests like this one provided an important channel for bottom-up communication in the Egyptian political system. They were a signalling mechanism. The food riot lasted only two days since Sadat decided to re-establish former levels of food subsidies quickly after the riots started (Sachs, 2012). In countries in the Middle East and North Africa (MENA), generalised price subsidies have been part of the ‘social contract’ for many years, especially with respect to food and fuel. Food subsidies have been embedded in Egyptian politics since 1941 when President Gamal Nasser created the programme to cope with food scarcity during wartime (Alderman, Braum, and Sakr, 1982). In emergency or war contexts, subsidies are often a good way for the government to reach the vulnerable population without targeting. However, Egypt’s subsidies continued long after World War II and became embedded in the government’s social benefits system. Egypt’s food subsidy system has two components: the subsidised bread that is provided for the whole population and ration cards that are provided to eligible households. These cards allocate monthly quotas of basic foodstuffs according to house-

107

THE PUBLIC SPHERE | 2016 ISSUE

hold size (WFP, 2008). Drawing on registration data for the ration cards, the World Food Programme (WFP) found most of the beneficiaries to be literate and part of a non-vulnerable (i.e. food-secure) group. Food subsidies in Egypt are considered inefficient for two main reasons: a) they do not fully reach the most vulnerable and food insecure populations; and b) they represent a high cost for the government (World Bank 2005, Gutner 2002 & Korayem 2013). Egypt’s food subsidy system does not seem to have one purpose; or if it did at its inception, that purpose has evolved over time. From a historical point of view, it seems that food subsidies were first provided by Nasser as a way to offset food scarcity during war. Since then, it has also been asserted as a benefit for all Egyptians regardless of their income level; a non-targeted social transfer to the poorest; and a signalling mechanism for the government to show its commitment to its people while withholding political freedoms. Finally, regardless of the initial intent, the public’s perception of the food subsidy system is also a reason that it has existed so long. This paper will not attempt to separate a singular purpose of the subsidies; rather it will look at the effectiveness of these different objectives and the efficiency of the system given the current political context. Notwithstanding the complexity of the system, one irrefutable role of food subsidies in Egypt is to address the nutritional needs of the most vulnerable by providing an in-kind food transfer to the poorest households. However, the current targeting still leaves a great part of the poor in Egypt without access to benefits. In 2011 for example, only 15 per cent of the benefits from subsidies on bread reached the poorest quintile (Sdralevich et al., 2014). In addition, food subsidies are costly for the government and come at the expense of investments in, for example, food production that could posit a better macroeconomic solution for controlling food prices in the long-term. In 1996–1997, the

108

[The political economy of food subsidies in Egypt]

absolute costs of the programme stood at (Egyptian Pounds) LE 3.7 billion (approximately US$ 472 million): nearly the same as profits from the tourism sector during the same period (Ahmed et al., 2001). Despite several attempts, a successful reform of food subsidies in Egypt has not yet been mounted. A potential explanation is that the government fears that changes might result in more episodes like the 1977 food riots. It is possible that, despite minor changes over the years, the political functions of the food subsidies system in Egypt is still the main rationale for keeping it in place. Indeed, some scholars associate the rising prices in Egypt due to the global food crisis in 2007 with the chain of protests that generated the Arab Spring in 2011 (El-Bakry, 2008). By mid-2008, 88 per cent of the total budget designated to food subsidies in the country had been used to control price increases, while lines in front of government-owned bakeries were growing at a fast pace (El-Bakry, 2008). In the same period, however, when trying to suggest a change in the system for direct cash transfers instead of food subsidies, the government of Hosni Mubarak came under public attack for the suggestion. In light of the close link between food subsidies and social stability in Egypt, this article intends to provide a chronological view of the recent political economy history and to present the current options for reforming a system believed to be inefficient in poverty reduction. Whilst defending the reform of the food subsidies system for a more efficient impact, this article walks through the main elements present in the political economy of the system that may explain its long-lasting popularity. In addition, by assessing the primary alternatives for strengthening social protection in Egypt, the paper brings to light the recent discussions of how to address the countryâ&#x20AC;&#x2122;s social needs that were contributors to the mass protests during the 2011 Arab Spring.

109

THE PUBLIC SPHERE | 2016 ISSUE

THEORETICAL FRAMEWORK Food subsidies as social transfers Many governments use subsidies to meet their social protection goals. In many cases, politicians prefer in-kind transfers as they are easier to track and manage in comparison with cash transfers. Food subsidies are the biggest part of the Egyptian Social Protection System and thus the government’s most significant safety net, both in terms of cost and number of beneficiaries (Rohac, 2013). Their primary objective is to address poverty through improving food security. However, the impact of the food subsidies goes beyond purely food security matters and needs to be analysed through the following aspects: their stabilising political function within the framework of a social contract, their importance for political stability given the public’s perception of their right to receive subsidies; and their efficiency as a social safety net. The World Bank defines safety nets as: ‘non-contributory transfers designed to provide regular and predictable support to targeted poor and vulnerable people. These are also referred to as “social assistance” or “social transfers”’ (Gentilini et al., 2014). Often safety nets are part of a broader social protection scheme, incorporated under a strategy for poverty reduction, resilience building and food security enhancement. The theoretical argument underlying social transfers is that the provision of regular and guaranteed support to beneficiaries allow them to make thoughtful decisions about how to use the transfer in order to plan ahead, invest or save (Davis et al., 2009). Their ability to ensure minimal support against shocks, and permit beneficiaries to allocate their budgets to activities that increase human capital, supports widespread transfer programmes around the world. In the case of food subsidies (non-targeted, in-kind transfers), the links with human capital accumulation are not as straightfor-

110

[The political economy of food subsidies in Egypt]

ward as cash transfers (Arnold, Conway, and Greenslade, 2011). Even if, at times, the cash provided is not allocated to future investments, the immediate burden of vulnerable individuals is alleviated (Ames et al., 2001). The evidence of the impact of in-kind transfers on household spending and savings choices is inconclusive. Some suggest that they can in fact distort prices and contribute to overconsumption, thereby imposing welfare costs on the larger community (Sdralevich et al., 2014). There are substantial signals that food subsidies affect consumption as they have a direct influence on purchases that often exceed the amount desired (Alderman, 2002). The main purpose of the Egyptian governmentâ&#x20AC;&#x2122;s provision of food subsidies is unclear although the general idea of providing food subsidies is to ensure food security. However, this practice can have secondary, unforeseen outcomes such as: a) significantly influencing local diets; b) inducing the overconsumption of the subsided items; and finally c) not providing individuals with the freedom to choose what to eat, impacting individual liberties (Gutner, 2002). These negative consequences suggest that subsidies may not be the most effective form of safety net but citizens consider them entitlements, suggesting that their political appeal outweighs their practicality (Gutner, 2002). There are indications that the average expenditure on food does not stop at subsistence levels, partially a result of having more incentives and means to acquire certain food items (Alderman, 2002). This affirms the idea that food subsidies promote overconsumption. Given the high cost of maintaining food subsidies, this dependency is likely to become unsustainable for the government to support at some point. Two main problems have been identified in relation to the use of food subsidies as social transfers. First, the programmes in Egypt are the inefficient and not well targeted since they do not correspond to real levels of consumption needs. Second, the costs to the government of supplying the subsidised item to meet the demand it creates may be

111

THE PUBLIC SPHERE | 2016 ISSUE

higher than the benefits, given that national expenditures with food subsidies increased over the years, but poverty and food security levels did not (World Bank, 2004; World Bank, 2010). Despite their potential as a targeted safety net, the food subsidies were not well connected to an adequate strategy for poverty alleviation until recently. Furthermore, because of their inefficiency as a safety net for the poor, they also impose a heavy fiscal burden on public finances (Sdralevich et al., 2014). Time and time again, the government of Egypt has tried to reform this system to optimise costs and its impact. So far, these efforts have not been successful, as no government wants to change a system greatly entrenched within Egyptian society. The common idea of subsidies as entitlements has transformed the programme into the biggest source of social justice that could be provided by the government. This idea is further assessed in this article, taking into account the political nuances of reforms in the light of a current desire held by the government of Egypt to no longer be seen as the main provider of goods and subsidies to its citizens (World Bank, 2005).

HISTORY OF FOOD SUBSIDIES IN EGYPT Although the use of food subsidies as a social protection tool is still being examined, their role in maintaining political stability has always been recognised. In societies where civil liberties are limited, the power of food subsidies as a symbol of the social contract between the government and its citizens cannot be underestimated (Singerman, 1995). Egypt remains a society that has weak political and social freedoms despite some economic liberties (Weinberger, 2012). The complex relationship between social unrest and food security is reflected in some of the socalled ‘bread riots’ that ultimately affected the political economy of food provision (Patel and McMichael, 2009).

112

[The political economy of food subsidies in Egypt]

Under Nasser’s government in the late 1950s, the government’s provision of subsidised food worked as temporary emergency relief at first. In order to respond to rising food prices, however, the government started to import large quantities of wheat and provide it to government-owned shops in a turn away from strictly emergency-based provision. During this period, subsidised bread became the symbol of the government’s social contract with its people. This has generated an ingrained mindset in Egypt that the government’s duty is to ensure affordable food and since the Nasser era the state has explicitly pursued this mandate (Khouri-Dager, 1996). Modernisation and social investments were hallmarks of the Nasser government, which introduced social support within the context of tight government control on personal freedoms. As Singerman (1995) noted, The Egyptian Government’s policies of political exclusion have gone hand in hand with their public commitment to provide the basic needs of the population. Because it limits and controls mass participation, the government maintains its legitimacy by providing goods and services to the population.

More specifically, rationing food to cope with war times began in 1941 and underwent several reform attempts throughout the following years. In the face of rapid urbanisation and population growth, the need to control food prices in the 1960s was reflected in the expansion of food subsidies (Gutner, 2002). During the same period, Egypt became the recipient of a steady stream of surplus U.S. wheat that boosted the availability and consumption of the subsidised crop in Egypt given that the country was not sufficient in domestic production of wheat (Sharp, 2010). Although it is fair to assume that any public policy can be reflective of the government’s stance at the time when it was first established, Egypt’s food subsidies programme has not fulfilled this primary purpose – whatever it may be (Pearce, 1982). If the

113

THE PUBLIC SPHERE | 2016 ISSUE

FIGURE 1 WHEAT DOMESTIC CONSUMPTION AND PRODUCTION IN EGYPT, 1960-2014

Wheat consumption

Sources: FAOSTAT and Agricultural Production Statistics by Country, no date.

Wheat production

114

thousand tonnes

[The political economy of food subsidies in Egypt]

idea for the programme was to function as a temporary relief mechanism to cope with food shocks during times of war or scarcity, the government could have maintained non-targeted and broadly accessible subsidies only during times of crisis and price increases. On the other hand, if it intended to boost agricultural production to keep a supply of subsidies to domestic consumers, it could have invested in local cereal (mostly wheat) production. If the original purpose of the grain subsidies in Egypt was to improve food security, then it has also failed because it does a poor job of targeting the most vulnerable. This lead us to believe that, rather than having a strong food security rationale, the system has become a powerful symbol of the broader social contract between the government and the population (Singerman, 1995). Figure 1 shows the constant trend of upward wheat production and consumption that indicates Egypt’s remaining dependence on imported wheat.

The bread riots After the death of Nasser in 1970, Sadat came to power. In contrast with his predecessor, Sadat’s policies had a stronger appeal to the right-wing elites who were inclined towards ‘Western’ aspirations for a political economy (Sachs, 2012). The 1970s and 1980s were marked by trends of economic liberalisation measures strongly advocated by international economic institutions of the Bretton Woods. The World Bank, the IMF, private financial institutions, and the oil-producing Arab states all played a role in influencing Egypt’s economic policy (Korany et al., 2008). At the same time, due to economic and oil crises, and unstable food prices during the first decade of Sadat’s presidency, food subsidies were increased (Table 1). The most notable episode involving food subsidies during Sadat’s rule was the bread riots (Patel and McMichael, 2009). The sequence of protests that flared up in Cairo and Alexandria in Jan-

115

THE PUBLIC SPHERE | 2016 ISSUE

899

1974 765

329

1973 730

1972

661

1970/71

Table 1 Total public expenditure, 1970–1980

Expenditure

Food subsidies

Central and local government expenditures subsidies

1,628

1977

450

2,012

1978

880

2,375

1979

472

1,754

1980*

1976 1,670

313

1975 1,298

322

(LEmillion)

491

*1980 figures are actual figures from period January 1, 1980 to June 30, 1980. Source: Alderman, Braum, and Sakr, 1982.

116

[The political economy of food subsidies in Egypt]

uary 1977 targeted the government, where demonstrators protested the food subsidy cuts proposed by Sadat. The social unrest was organised after the president complied with IMF negotiators who were advocating for a reduction in the subsidies system in the light of pressures for deep macroeconomic reforms (Gutner, 2002). Due to their connection to the IMF’s structural adjustment policies, these violent political demonstrations have also been called the ‘IMF riots’ (Walton and Seddom, 1994). Two days after the protests started, the government rescinded the cuts and decided to respond by further expanding the same programmes (Sachs, 2012). The 1977 riots were a protest not only of subsidy cuts, but also economic changes in general. The event signalled two core issues regarding the role of food subsidies in political economy. On one side, the riots highlighted the strategies policymakers should avoid when reforming food subsidy systems. On the other side, they made evident the fact that externally imposed need for economic reforms did not seem to take internal political dynamics into account (Korany et. al, 2008). The policy to reduce food subsidies as a part of economic structural adjustments did not carefully think through the role the subsidies system had (and still has) as a political tool in Egypt. The riots were an important reminder of the role food plays in uniting civil opposition in Egypt. According to political theory, those in power (‘the elites’) will only choose democracy over the status quo when income redistribution is less costly than a credible threat (Olson, 1993). If redistribution is too costly for the rich and insufficient to motivate the poor, this may be the cause for turning against democracy (Benhabib and Przeworski, 2005). Revolution is more likely if governments resist addressing the needs of the poorest. In 1977, just before the bread riots, the government decided between cuts affecting the military, industrialists, financial leaders, and cuts affecting the poor. The first three groups had enough access to political leaders to shift the burden of reform to the Egyptian masses (Hopkins, 1988).

117

THE PUBLIC SPHERE | 2016 ISSUE

Huntington and Fukuyama (1997) argue that political instability arises from tensions within political systems, particularly unfulfilled popular expectations, social change, and political mobilisation of dissatisfied groups. This idea is applicable to the 1977 food riots in Egypt. The fact that the government responded by increasing subsidies indicates that they were aware of this dynamic and decided to address popular expectations again as a way of meeting the less costly demands of the people in terms of political power. The act of increasing subsidies served as a ‘symbolic fulfilment of a social contract’ (Hopkins, 1988): that is, the government fulfilled public expectations by providing ‘material side payments’ (in-kind food in this case) and not addressing the basis of civil rights and democratic participation in social policies.

Changing the subsidy status quo By 1980–1981, total expenditures on food subsidies in Egypt marked almost 14 per cent of total government spending whereas levels in previous decades reached as little as 0.2 per cent (Sadowski, 1991). When Hosni Mubarak became president in 1981, another strategy attempting to gradually reduce the subsidy system was put in place. Mubarak acknowledged the importance of reforming the system, but initially avoided specific proposals for how to do so. Instead, he ended up improving the targeting and quality of provided food items but did not remove the overall provision of subsidies during the 1980s (Nasser and Gomaa, 1998). Learning from Sadat, Mubarak initially adopted low profile, gradual reform measures that improved targeting but did not remove the benefits (Korany et al., 2008). In this regard, the gradual reform strategy was more successful in avoiding political instability in comparison to former attempts. At the same time, elements of political restraint are constantly present in most political analyses of Mubarak’s government and

118

[The political economy of food subsidies in Egypt]

are said to have contributed to the general social stability as well. According to Sachs (2012): Subsidy reform in the early years of the Mubarak era was marked by a gradual implementation of covert changes and distribution controls (...) domestically speaking, the government engaged in selective policy retreat and repression to calm popular anger.

As some civil liberties were granted in tandem with the opening of the market economy, the government justified withdrawing other political freedoms. Egyptians were granted some democratic civil liberties while they lost most of the political freedoms of expression and association they had in previous decades. Given this context, continuing to provide food subsidies at the cost of social access was an old tactic that was maintained during this period. As expected, however, a few protests driven by food price peaks also happened in the 1980s. Once again, bread riots in 1984 and 1988 provoked responses that combined retreat and expansion of subsidies followed by repression when the government directly intervened to control food prices (Sachs, 2012). In the 1990s, the rate of change was accelerated as the government removed goods such as frozen meat, fish, tea, and rice from the subsidy system (Sachs, 2012). By 1995, only baladi bread, wheat flour, edible oil, and sugar were included â&#x20AC;&#x201C; down from 20 goods in the 1980s. These changes to the status quo should have brought about opposition from the public based on historical reactions, but the gradual pace of the changes along with the political repression were strong enough that there was little room for protest (Sachs, 2012).

Arab Spring Between 2008 and 2009, a global food crisis strongly affected food-importing countries such as Egypt, where prices

119

THE PUBLIC SPHERE | 2016 ISSUE

quickly increased. The cost of the minimum food basket in Egypt increased by 47 per cent between 2005 and 2008 (FAO et al., 2008). Consequences included increasing public discontent, which echoed the belief that neither the government nor political elites were able to address minimal food needs. Mubarak’s failures to direct political dialogue and his constantly repressive actions fuelled feelings of general anger and willingness for a change. The government’s practice of balancing repression with new social benefits started to fail, as the need for a more participatory political system was demanded. Once expectations of government assistance were damaged, the regime in Egypt was threatened. In addition to high food prices, political oppression and restrained rights, there was a sequence of socioeconomic discontent that can be linked to the 2011 Egyptian Revolution. Youth unemployment, for example, reached 38.8 per cent of the labour force between 15 to 24 years (World Bank metadata). The widespread anger that mobilised Egyptians to Tahrir Square in January 2011 fit into a broader set of concerns over corruption, the ineffectiveness of the government and lack of social policies. Following Mubarak’s resignation, expenditure levels of the food subsidies programme were maintained, likely due to political apprehension: the 2012 budget remained as high as the years before, LE 18.9 billion in comparison to LE 13.6 billion in 2011 and LE 14.1 billion in 2009/10 (Ministry of Finance in Egypt, 2011). Following years of increasing subsidies since the Arab Spring, the government of President Abdel Fattah el-Sisi projected and decreased its subsidy spending in 2014/15 (Figure 2). Despite the decrease, it is still too early to understand if the projections are linked to system reforms or if gradual cuts are taking the 2014 political stabilisation of Egypt into account. The Arab Spring signalled to the government that the basis of its social contract had limitations (Ghanem, 2014).

120

[The political economy of food subsidies in Egypt]

FIGURE 2 EGYPTIAN GOVERNMENT SPENDING ON FOOD AND ENERGY SUBSIDIES, 2009/10â&#x20AC;&#x201C;2014/15 Food subsidies (billions of Egyptian pounds) Energy subsidies (billions of Egyptian pounds) Food and energy subsidies as a % of government expenditures 180

30%

120

20%

10%

2009/10

10/11

11/12

12/13

13/14

2014/15 (projected)

Source: Egyptian Ministry of Finance, July 2015 via Your Source for Political Risk Analysis, 2015.

FOOD SUBSIDY REFORM Strengthening social protection In Egypt, social protection schemes are heavily dominated by the subsidy programmes that account for nearly the entirety of social spending in the country (Silva, Levin and Morgandi, 2013). In 2013, the Egyptian government spent roughly LE 4.31 billion on food subsidies alone (Oliver, 2015). Poverty levels remain high, even though food subsidies remain an important programme. According to the World Bank, the per centage of the population under the poverty line in Egypt increased from 18.4 in 1999 to 19.6 per cent in 2004 (World Bank, 2005). Budgetary pressure and political sensitivity have made policy changes difficult. International pressure from governments

121

THE PUBLIC SPHERE â&#x20AC;&#x2020;| â&#x20AC;&#x2020;2016 ISSUE

and economic institutions for structural economic reforms also play a significant role in how government approaches the reform of food subsidies. Although social safety nets are not, in theory, designed to substitute wide-ranging macroeconomic and social policies, they can play various indirect roles in relief and socio-economic development (Mkandawire, 2004). One argument addresses the consequences of removing subsidies in the shortterm, such as macroeconomic effects like high inflation (Sdralevich et al., 2014). The use of safety nets based on cash transfers has been suggested as an alternative to the current food subsidy programme by various scholars and international development agencies (World Bank, 2005; Gutner, 2002; Ghanem, 2014; Lofgren & El-Said, 2001). The rationale for cash transfers is straightforward: poverty is multidimensional and inconsistent income is a common reason for this. The current cash transfer programme in Egypt is still relatively small when compared to the food subsidies. It represented as little as 0.12 per cent of GDP in 2005, and it is still disconnected from other social benefits (World Bank, 2005). The recent announcement of a World Bank-designed US$ 400 million cash transfer programme in Egypt is bound to change perceptions of this safety net in the country (World Bank Press Release, 2015). However, substituting food subsidies for cash transfers in Egypt is an alternative that has been still under-researched. If implemented in the absence of complementary policies and strategies for exiting the programme, it could also represent a fiscal burden for the government in the long run. Developing a comprehensive approach to its social safety net would require a high level compromise by the government, and would entail an analysis of which economic and fiscal dimensions to include. Most importantly, the government will have to keep in mind how to balance the delicate political aspects that have been present throughout the existence of the system.

122

[The political economy of food subsidies in Egypt]

The inefficiency of food subsidies in Egypt In spite of its generally successful political function as a social stabiliser, the food subsidies system in Egypt has not reached the poorest households efficiently. The subsidised baladi bread and wheat flour, for example, are expected to be available to all citizens. This is primarily limited by the proximity to the governmentâ&#x20AC;&#x2122;s bakery and beneficiariesâ&#x20AC;&#x2122; willingness to wait in line for it (Ahmed et al., 2001). With respect to distribution, a study conducted by the World Bank (2005) disaggregated the Egyptians by income into five quintiles of expenditure levels. Further assessing national household surveys, the Bank demonstrated that the richest quintile received the largest portion of food subsidies (24 per cent) while the poorest quintile received the smallest portion (17 per cent). Even if the primary goal of the government is not to reach the poorest but rather the general population, it is still doing a comparably bad job at reaching the most vulnerable as an equal share overall. FIGURE 3 DISTRIBUTION OF FOOD SUBSIDY VALUE BY EXPENDITURE

Poorest quintile 1 Quintile 2

24%

17%

Quintile 3

18%

Quintile 4 Richest quintile 5

22% 19%

Source: World Bank, 2005.

Leakage, which is when a subsidised item does not reach the intended beneficiary, is historically high1. Greater availability

1. Baladi bread leakages cost LE 3.7 billion to the Government of Egypt in 2008 (World Bank, 2010).

123

THE PUBLIC SPHERE | 2016 ISSUE

of grain also induces overconsumption and waste. Egypt is the world’s biggest importer of wheat and one of its biggest consumers too. In 2013, Egyptians consumed between 180 to 200 kilograms of the grain (AEGIC, 2015). If leakages could be eliminated and the coverage of the programme narrowed, savings could be up to LE 27 billion. Additionally, using IFPRI’s 1997 regional poverty head-count (Datt and Jolliffe, 1999) and national data on distribution, Ahmed et al. (2001) identified that the regional supply of food subsidies was not sensitive to regional distributions of poverty. What the data suggests is that the system does an extremely poor job of targeting the needy. Furthermore, in the case of wheat, which is the foundation of the Egyptian diet, incentives to overconsume not only affect people’s diets but also ensure that the country remains dependent on the import. Despite the prevalent agreement among the main policymakers on the need to reform the system, the political sensitivity of the policy has challenged the government in finding the most feasible strategy to do so.

Reform options An IMF report on the reform of subsidies in MENA found that the main determinants of accomplishing subsidy reforms are ownership, communication and consultation (Sdralevich et al., 2014). Egypt, however, has not yet been successful in these three elements. Other countries in the region like Tunisia were effective at reforming their food subsidies in the 1990s after experiencing public riots in reaction to attempts to change the system. An important lesson that Tunisia can share with Egypt is the notion that, to reduce subsidies, the government should convince the wealthiest sectors and perhaps more importantly, the large middle class, that they do not need the benefit. To do so, Tunisia used a media campaign to sensitise the public to impending

124

[The political economy of food subsidies in Egypt]

changes and to explain the new compensation measures it would introduce to mitigate any negative impacts that these changes might have on the poor (Sdralevich et al., 2014). While this approach has not been used by Egypt, it is a successful tactic by a nearby country that could be considered. Before focusing on modes of communicating reform, discussions need to focus on whether the system should be substituted by a different safety net or if it should be improved through better targeting. In the latter case, basing a reform on better targeting of the food subsidy system to reach the poorest households is an option that does not add significant costs for the government; indeed, it may reduce them. For instance, increasing the use of ration food cards (or smart cards) would provide beneficiaries with the freedom to choose if they want to purchase food goods, and if so which ones and how much (Korayem, 2013). In any of the cases, modifying the current food subsidies system would have serious impacts on poor households if not accompanied by alternatives for strengthening social protection structures (Rohac, 2013). Seeking to understand the importance of the subsidised food goods to poor Egyptians, Korayem (2013) assessed the average expenditure by this group on subsidised food commodities. She calculated the average household expenditure on subsidised food commodities as a per centage of the total food expenditure and applied it to the different expenditure groups. By using data from the 2004/05 HIECS, Korayam found that subsidised food items were more significant in the overall budgets of the lower expenditure households when compared to all other quintiles. The importance of subsidised goods for the poorest Egyptians does not come as a surprise. In order to ensure that reforming the system is done progressively on the basis of informed political dialogue with society, the public must be aware of the economic benefits of such a reform, which does not seem to have been the case in Egypt. In addition, any adjustments to the food subsidy

125

THE PUBLIC SPHERE | 2016 ISSUE

system need to consider how to meet the needs of both rural and urban poor; targeting mechanisms to ensure pro-poor measures; and a gradual process of linking government’s social expenditures with human capital formation (Oliver, 2015). Maintaining in-kind benefits, on the other hand, would ensure that the minimal necessary commodities stay at low prices. Pro-poor oriented, in-kind benefits could potentially benefit the poorest households that have a bigger concern with subsistence rather than choosing what to eat. As Korayem (2013, p.20) concludes: For the subsidy beneficiary, choosing between two kinds of subsidy systems is equivalent to choosing between having freedom of choice of what to buy, but always being insecure regarding the quantities given potential price changes, or choosing to guarantee a certain amount of basic commodities at affordable prices but being deprived of the freedom to choose what.

Investing in the implementation of a cash transfer programmes as a way to reform food subsidies is an alternative frequently supported by the international community. Although an impact evaluation on the effects of cash transfers in Egypt should be conducted before any policy is cleared to replace the current safety net, such programmes should be considered. Cash transfers provide an interesting aspect of both protecting living standards of households, as well as promoting wealth creation through income’s increase. Depending on context, they may also help prevent households from suffering shocks and transform relationships within society and between citizens and the state. (Arnold, Conway & Greenslade, 2011). For a country like Egypt, the use of this safety net can have the potential to address the multifaceted factors of poverty that the other alternatives cannot do. Furthermore, if there is social resistance against the idea of providing cash transfers to the poor, adding conditions to the

126

[The political economy of food subsidies in Egypt]

scheme could encourage general acceptance. As scholars like Hickey believe: ‘Conditional transfers have an intuitive appeal to elites as they help make better citizens of the poor’ (Hickey, 2006). Still, conditioning the cash transfer would require further budgetary investments for the government as beneficiaries would need to be provided with educational and health services, for example. Analysing how the conditionalities would benefit citizens and as well the Egyptian government is not assessed on this article but suggests that future research should be dedicated to the issue. Nevertheless, it is important to remember that there is no one ‘right’ modality (Hoddinott et al., 2013). The relative effectiveness of different models is highly dependent on contextual factors that reflect the kind and geographic incidence of poverty in Egypt (2013). Whereas the discussion of potential reform options is important, the readiness of the government of Egypt to coordinate and implement it is an important concern. While this article believes in the importance of enhancing Egypt’s social protection network, it recognises the need for further analysis of reform within the present political context. The feasibility of the most adequate model needs to be appraised. For this reason, considering the reform under a real political process will be essential for its future success.

CONCLUSION Despite the myriad efficiency issues, the food subsidy programme in Egypt has shown at least one positive impact for food security. The popular baladi bread programme has especially been important in providing the minimum necessary amount of calories for the poorest Egyptians despite its nutritional and macroeconomic consequences in the long-term. As history has shown, changes to the food subsidy system involve a risky politi-

127

THE PUBLIC SPHERE | 2016 ISSUE

cal trade-off. However, subsidies and rationing systems are more appropriate for emergencies or ad hoc interventions. The macroeconomic effects, in the long run, can produce distortions in consumption patterns and import rates of the subsidised food, thus representing a fiscal burden for governments. It is therefore in Egypt’s best interest to reform the system while balancing the political costs. Finding the best political process to reform the system is just as important as the reform itself. Perhaps the most acceptable options combine a progressive reduction of the benefits to the wealthiest population while increasing compensations for the most vulnerable. Policy options including improved targeting are likely to be the preferred strategies in moving towards more pro-poor policies that can better affect and reach the poor and reduce impact on government expenditure. At the same time, the process has to strongly consider the political elements linked with the food subsidies programme. The national history of social unrest could indicate that attempts to reform subsidies systems without discussing Egypt’s social contract were failures. The use of conditionalities (attendance rates at schools, vaccinations, and other mechanisms that try to ensure that people use the money but with certain obligations) is an interesting possibility as they could exclude the wealthiest -- who do not need benefits – from the subsidies whilst showing them that whoever is receiving the benefits is ‘deserving’ of them. Taking into account the political importance of the food subsidies, conditioning better targeted programmes for food assistance and poverty alleviation can be a way for the government to make the idea of a transfer more acceptable to the top quintiles (Fiszbein & Schady, 2009). It is important for Egypt to understand what reform process would be most feasible for the introduction of better social protection schemes given the country’s socio-political and economic context. While pro-poor policies are less costly for the govern-

128

[The political economy of food subsidies in Egypt]

ment than the current food subsidies system, it is essential to identify the financial feasibility of new programmes in order to implement the reform effectively, doing what no Egyptian government has successfully managed to do. The present government has to define a path to food subsidy reform in parallel with its need for other related reforms in order to realistically improve development, thereby promoting economic growth. Better targeting of the poorest populations that are more vulnerable to food price shocks may be an interesting first step, especially in a programme like food subsidies that, in its current shape, ends up benefiting the elites without fully addressing the poor.

REFERENCES Acemoglu, D. and Robinson, J. A. (2006) ‘De Facto Political Power and Institutional Persistence’, American Economic Review, 96(2), pp. 325–330. doi: 10.1257/000282806777212549. Agriculture and Food: Key Tables from OECD, Producer support estimates (subsidies): As a percentage of gross farm receipts (no date) Available at: http://www.oecd-ilibrary. org/agriculture-and-food/producer-support-estimates-subsidies_20755104-table1 (Accessed: 10 August 2015). Agricultural Production Statistics by Country (no date) Available at: http://www.indexmundi.com/agriculture (Accessed: 15 August 2015) Ahmed, A.U., Bouis, H., Gutner, T., Lofgren, H., (2001) The Egyptian Food Subsidy System: Structure, Perfomance, and Options for Reform. Washington, D.C.: International Food Policy Research Institute (Research Report 119). Alderman, H. (2002) ‘Subsidies as a Social Net: effectiveness and challenges’, World Bank Institute, no. 0224(Social Safety net Primer series). Alderman, H. and Braun, J. von (1984) The effects of the Egyptian food ration and subsidy system on income distribution and consumption. Washington, D.C.: International Food Policy Research Institute. Alderman, H., Braum, J. von and Sakr, S. A. (1982) ‘Egypt’s Food Subsidy and Rationing System: A Description’. Research Report 34 edn. International Food Policy Research Institute. Ames, B., Brown, W. and Devarajan, S. (2001) ‘Macroeconomic Policy and Poverty Reduction’. Washington, DC: International Monetary Fund. Arnold, C., Conway, T. and Greenslade, M. (2011) ‘Cash Transfers Literature Review- . UK Department for International Development (DFID-UK). London, UK. Australian Export Grains Innovation Centre, AEGIC (2015) Global Grain Market Series Egypt. Available at: http://www.aegic.org.au/media/49851/aegic_egypt_factsheet_ web0415.pdf .

129

THE PUBLIC SPHERE | 2016 ISSUE

Barrientos, A. (2012) Social Transfers and Growth: What Do We Know? What Do We Need to Find Out?, World Development, 40(1), pp. 11–20. doi: 10.1016/j.worlddev.2011.05.012. Benhabib, J. and Przeworski, A. (2005) ‘The political economy of redistribution under democracy’, Economic Theory, 29(2), pp. 271–290. doi: 10.1007/s00199-005-0002-5. Blaydes, L. (2010) ‘Elections and Distributive Politics in Mubarak’s Egypt’. Cambridge University Press. Bourdieu, P. (1991) Language and Symbolic Power. United Kingdom: Harvard University Press. Cambanis, T. (2015) The Arab Spring was a revolution of the hungry - The Boston Globe. Available at: https://www.bostonglobe.com/ideas/2015/08/22/the-arab-spring-was-revolution-hungry/K15S1kGeO5Y6gsJwAYHejI/story.html (Accessed: 25 August 2015). Cassandra. (1995). ‘The Impending Crisis in Egypt’. Middle East Journal 49 (1): 9–27. Cassing, J., Nassar, S., Siam, G. and Moussa, H. (2007) ‘Distortions to Agricultural Incentives in Egypt’, World Bank’s Development Research Group. Clarke, D. L. (1997) ‘U.S. Security Assistance to Egypt and Israel: Politically Untouchable?’ Middle East Journal 51 (2). Available at: http://www2.ari.net/mei/clark.hl (Accessed: 13 August 2015). Clezadlo, A. (2012) Day of Honey: A Memoir of Food, Love, and War. New York: Free Press. Cohen, M. J. and Smale, M. (eds.) (2011) Global Food Price Shocks and Poor People: Themes and Case Studies. London: Taylor & Francis. Datt, G. and Jolliffe, D. (1999) Determinants of poverty in Egypt, 1997. Available at: http:// ebrary.ifpri.org/cdm/ref/collection/p15738coll2/id/125744 . Davies, M., Barrientos, A., Devereux, S., Hickey, S., Sabates-Wheeler, R., Guenther, B. and Macauslan, I. (2009) ‘DFID Social Transfers Evaluation Summary Report’, IDS Research Reports, 2009(60), pp. 1–36. doi: 10.1111/j.2040-0217.2009.00060_2.x. De Schutter, O. (2014) Final Report: The Transformative Potential of the Right to Food, report submitted to the UN Human Rights Council by the Special Rapporteur on the right to food, Olivier De Schutter, New York: United Nations Human Rights Council. Durojaye, E. (2015) ‘The ILO Social Protection Floors Recommendation 202, and poverty reduction’, Development in Practice, 25(2), pp. 270–276. doi: 10.1080/09614524.2015.1003531. El-Dahan, M. (2011) Egypt inches towards far-reaching food subsidy reform. Available at: http://www.reuters.com/article/2011/09/28/us-egypt-food-idUSTRE78R2SD20110928 (Accessed: 25 August 2015). Fabiosa, J. and Soliman, I. (2008) ‘Egypt’s Household Expenditure Pattern: Does It Alleviate a Food Crisis?’, Iowa State University. Center for Agricultural and Rural Development-CARD Working Paper Series, CARD Working Paper 08-WP 475. Farouk Ghoneim, A. (2014) ‘The Political Economy of Food Price Policy in Egypt’, in Food Price Policy in an Era of Market Instability. Oxford University Press (OUP), pp. 253–274 FAO, WFP, IFAD, ILO and NEPAD (2008) Inter-Agency Mission Assessment on Initiative on Soaring Food Price - November to December 2008. Available at: https://www.ids. ac.uk/files/dmfile/FSInteragencymissionreport.pdf FAOSTAT (no date) Available at: http://faostat3.fao.org/compare/E (Accessed: 31 August 2015). Food and Agriculture Organization (FAO) ‘The State of Food Insecurity in the World 2015’ (no date) Available at: http://www.fao.org/hunger/en/ (Accessed: 25 August 2015). Fukuyama, F. and Huntington, S. P. (1997) ‘Political Order in Changing Societies’, Foreign Affairs, 76(5), p. 215. doi: 10.2307/20048214.

130

[The political economy of food subsidies in Egypt]

Gentilini, Ugo; Honorati, Maddalena; Yemtsov, Ruslan. (2014) The state of social safety nets 2014. Washington, DC: World Bank Group Ghanem, H. (2014) ‘Egypt needs a new social protection system’,Brookings, 25 July. Available at: http://www.brookings.edu/blogs/up-front/posts/2014/07/25-egypt-social-protection-system-ghanem (Accessed: 19 August 2015). Graham, P. C. (1996) ‘Safety Nets, Politics, and the Poor: Transitions to Market Economies’, The Journal of Developing Areas, Vol. 30, No. 3 (Apr., 1996), pp. 400-402. Gutner, T. (2002) ‘The political economy of food subsidy reform: the case of Egypt’, Food Policy, 27(5-6), pp. 455–476. doi: 10.1016/s0306-9192(02)00049-0. Gutner, T., Gomaa, S., Nasser, S., (1999). ‘The Political Economy of Food Subsidy Reform in Egypt’. International Food Policy Research Institue, Washington, DC. Harsch, Ernest (2008) “Price Protests Expose State Faults: Rioting and Repression Reflect Problems of African Governance,” African Renewal, XX, 2, 15–17. Available at: http:// www.un.org/ecosocdev/geninfo/afrec/vol22no2/222-price-protests.html Hickey, S. (2006). The Politics of What Works in Reducing Chronic Poverty. A synthesis report for the Ministry of Foreign Affairs, The Netherlands. CPRC Working Paper 91, Chronic Poverty Research Centre. Hoddinott, J., Gilligan, D., Hidrobo, M., Margolies, A., Roy, S., Sandstrom, S., Schwab, B. and Upton, J. (2013) Enhancing WFP’s Capacity and experience to Design, Implement, Monitor and Evaluate Vouchers and Cash Transfer Programmes: Study Summary. Washington, D.C.: International Food Policy Research Institute (IFPRI). Ibrahim, S. E., and H. Lofgren. (1996). Successful Adjustment and Declining Governance? The Case of Egypt. World Bank Private Sector Development Department Essays on Governance, Leadership and Communication. Washington, D.C.: World Bank Kathleen Oliver, K. (2015) Targeting the Egyptian Food Subsidy System: Urban Versus Rural Reform. UChicago Undergraduate Business Journal. Khouri-Dager, N. (1996). The State, Urban Households, and Management of Daily Life: Food and Social Order in Cairo. In Development, Change, and Gender in Cairo, ed. D. Singerman and H. Hoodfar. Bloomington: Indiana University Press. Korany, B., Dessouki, A. H. E. and Ahmad, A. Y. (2008) The foreign policies of Arab states: the challenge of globalisation. 1st edn. Cairo: American University of Cairo Press. Korayem, K. (2013) ‘Food subsidy and the Social Assistance program in Egypt: Targeting and Efficiency Assessment’, Topics in Middle Eastern and African Economies, 15(1). Lachine, Nadime. (1977) Class Roots of the Egyptian Regime. MERIP Reports 56 (1977): 3-7. Löfgren, H. and El-Said, M. (2001) Food subsidies in Egypt: reform options, distribution and welfare, Food Policy, 26(1), pp. 65–83. doi: 10.1016/s0306-9192(00)00030-0. McDermott, A. (1988) Egypt from Nasser to Mubarak: A Flawed Revolution. New York: Croom Helm. Ministry of Finance in Egypt (no date) Available at: http://www.mof.gov.eg/ (Accessed: 31 August 2015). Mkandawire, T. (2007) ‘Transformative Social Policy and Innovation in Developing Countries’, The European Journal of Development Research, 19(1), pp. 13–29. doi: 10.1080/09578810601144236. Mkandawire, T. (ed.) (2004) Social Policy in a Development Context. Nature Publishing Group. Nasser, S., and S. Gomaa. (1998) Politics and Policies of Food Subsidies in Egypt. Cairo University, Public Administration Research and Consultation Center.

131

THE PUBLIC SPHERE | 2016 ISSUE

Olson, Mancur (1993) Dictatorship, Democracy, and Development. American Political Science Review, vol. 87, no. 3, pp. 567-576. Patel, R. and McMichael, P. (2009) A Political Economy of the Food Riot. Available at: http:// rajpatel.org/wp-content/uploads/2009/11/patel-mcmichael-2010Review321.pdf .review, xxxii, 1, 2009, 9–35 Pearce, J. (1982) ‘Government policy and food imports—The case of wheat in Egypt’, Agricultural Administration, 11(2), pp. 156–157. doi: 10.1016/0309-586x(82)90059-0. Pinstrup-Andersen, P. (2014) Food Price Policy in an Era of Market Instability: A Political Economy Analysis. WIDER Studies in Development Economics Podeh E., & Winckler O., (2004). Introduction. In: Podeh E., and Winckler O., Rethinking Nasserism: Revolution and historical memory in modern Egypt, USA: University Press of Florida Press Release (2015) ‘US$400 Million Program to Support 1.5 million Poor Egyptian Families through Enhanced Social Safety Nets’, The World Bank, 10 April. Available at: http://www.worldbank.org/en/news/press-release/2015/04/10/us400-million-program-to-support-15-million-poor-egyptian-families-through-enhanced-social-safety-nets (Accessed: 14 August 2015). Ramadan, R. and Thomas, A. (2011) ‘Evaluating the impact of reforming the food subsidy program in Egypt: A Mixed Demand approach’, Food Policy, 36(5), pp. 638–646. doi: 10.1016/j.foodpol.2011.06.006. Raymond F. Hopkins. (1988). ‘Political Calculations In Subsidizing Food’, in Food Subsidies In Developing Countries: Costs, Benefits, And Policy Options. John Hopkins University Press. Pp. 107-126. Regional Health Observatory Data Repository (no date) Available at: http://rho.emro.who. int/rhodata/node.country.country-EGY-stats?lang=en (Accessed: 27 August 2015). Rehab el-Bakry, (2008) Public outcry over talk of subsidy reform, Business Monthly, American Chamber of Commerce in Egypt, January 2008. Rodrik, D. and Fernandez, R. (1991) ‘Resistance to Reform: Status Quo Bias in the Presence of Individual-Specific Uncertainty’, The American Economic Review, 81(5), pp. 1146–1155. Rohac, D. (2013) ‘Solving Egypt’s Subsidy Problem’, Cato Institute. Sachs, R. (2012) On Bread and Circuses: Food Subsidy Reform and Popular Opposition in Egypt. Center for International Security and Cooperation Stanford University. Sadowski, Y.M. 1991. ‘Political vegetables? Businessman and bureaucrat in the development of Egyptian agriculture’. Washington, D.C.: Brookings Institution. Samir, M. (2015) Annual inflation drops to 8.3% in July: CAPMAS. Available at: http://www. dailynewsegypt.com/2015/08/10/annual-inflation-drops-to-8-3-in-july-capmas/ (Accessed: 25 August 2015). Sdralevich, C., Sab, R., Zouhar, Y. and Albertin, G. (2014) Subsidy Reform in the Middle East and North Africa: Recent Progress and Challenges Ahead. Washington, D.C.: International Monetary Fund. Sen, A. (1981) Poverty and Famines: An Essay on Entitlement and Deprivation, Oxford: Oxford University Press. Sharp, J. (2010) U.S. Foreign Assistance to the Middle East: Historical Background, Recent Trends, and the FY2011 Request. Available at: https://www.fas.org/sgp/crs/mideast/ RL32260.pdf . Siam, G. (2006) An assessment of the impact of increasing wheat self-sufficiency and promoting cash-transfer subsidies for consumers in Egypt: A multi-market model. Cairo: ESA Working Paper No. 06-03 Food and Agriculture Organisation.

132

[The political economy of food subsidies in Egypt]

Silva, J., Levin, V. and Morgandi, M. (2013) Inclusion and Resilience: The Way Forward for Social Safety Nets in the Middle East and North Africa. Washington, DC: World Bank Publications. Singerman, Diane. (1995) ‘Avenues of Participation: Family, Politics, and Networks in Urban Quarters of Cairo’. Princeton, NJ: Princeton UP. Staff, O. (2011) ‘Egyptians Riot in the Streets in 1977’. Available at: http://www.cbsnews. com/news/egyptians-riot-in-the-streets-in-1977/ (Accessed: 25 August 2015). The Economist 17 March 2012. Let them eat baklava. Available at: http://www.economist. com/node/21550328 (Accessed: 22 August 2015). The World Bank (2005) Egypt – Toward a More Effective Social Policy: subsidies and social safety net. Cairo, Egypt: Social and Economic Development Group. Available at:http://siteresources.worldbank.org/INTPSIA/Resources/490023-1171551075650/ Egypt_PSIA_121605.pdf (Accessed: 20 August 2015) The World Bank (2010) Egypt’s food subsidies: benefit incidence and leakages. Cairo, Egypt: Social and Economic Development Group. Thompson, Edward P. (1971). ‘The Moral Economy of the English Crowd in the Eight- The Moral Economy of the English Crowd in the Eighteenth Century,’ Past and Present, L, 76–136 Thomson, Anne. (1983) Egypt: Food Security and Food Aid. Food Policy 8.3. Tilly, Louise A. (1971). ‘The Food Riot as a Form of Political Conflict in France’, Journal of Interdisciplinary History, II, 23–57 Tuck, L. and Lindert, K. (1998) in From Universal Food Subsidies to a Self-Targeted Program: A Case Study in Tunisian Reform, Vol. 351. Washington, DC: World Bank Publications. Walton, J. and Seddon, D. (eds.) (1994) Free Markets & Food Riots. Wiley-Blackwell. Weinberger, D. (2012) Bread, Freedom, Social Justice: A Report From Egypt. Available at: http://www.huffingtonpost.com/david-weinberger/bread-freedom-socialjust_b_1114936.html (Accessed: 22 August 2015). Wille, D. (2015) Will Sisi succeed in taking on Egypt’s subsidies? | GRI. Available at: http:// globalriskinsights.com/2015/07/will-sisi-succeed-in-taking-on-egypt-subsidies/ (Accessed: 22 August 2015). World Bank Metadata (no date) Available at: http://data.worldbank.org/indicator/ SL.UEM.1524.ZS (Accessed: 26 August 2015). World Food Programme (2008) Vulnerability Analysis and Review of Food Subsidy in Egypt. Available at: http://documents.wfp.org/stellent/groups/public/documents/ ena/wfp198322.pdf XE Currency Converter - Live Rates (no date) Available at: http://www.xe.com/currencyconverter/ (Accessed: 30 August 2015). Your Source for Political Risk Analysis (2015) Available at: http:// http://globalriskinsights. com/wp-content/uploads/2015/07/Untitled4.png (Accessed: 05 August 2015).

133

Data localization laws in a digital world Data protection or data protectionism? Neha Mishra

LLM in Public International Law, Class of 2011, London School of Economics and Political Science Master in Public Policy, Class of 2015 National University of Singapore

ÂŠJaroslaw PuszczyĹ&#x201E;ski

ABSTRACT Data localization laws are emerging as a pernicious form of non-tariff barrier which significantly harms the growth of trade in a digitally powered world. An International Political Economy approach provides a more comprehensive analysis of the policy rationale behind such laws, as compared to a purely economic approach, which only focuses on economic losses resulting from protectionism. On a closer analysis, it is found that different countries may have different policy rationales for implementing data localization laws – while some promote their domestic ICT industry through forced localization measures, others have concerns regarding national security, privacy, and ensuring sovereign control in the highly privatized world of internet governance. It is not always possible to demarcate the “protectionist” rationale from that of rational “data protection”. To address data localization effectively and facilitate digital trade, it is not sufficient to negotiate for free flow of data in trade agreements without Governments and companies being open and transparent about the related issues of privacy, national security and consumer protection. Particularly, the role of US Government as well as leading US-based technology companies will be instrumental in this regard. At the same time, it may be necessary to develop policy initiatives both to encourage transparent and clear international standards on data security, as well as to enable higher levels of digital innovation in developing countries such that they can harness the benefits of evolving internet technologies.

[Data localization laws in a digital world]

INTRODUCTION The new wave of policies preventing free flow of data across borders is feared to be one of the most critical barriers to 21st century trade. Several experts both within the industry and outside consider such restrictions to be unsustainable policy practice as the free flow of data through the internet powers a majority of transactions in the world today (Ezell et al, 2013; McKinsey Global Institute, 2014; Donnan 2014; Chander and Le, 2014). Furthermore, critics are sceptical as to whether data localization initiatives generate any value addition to the domestic economy at all (Baeur et al, 2013; 2014; 2015). Nonetheless, several Governments continue to adopt a range of data localization laws for a variety of policy objectives, from safeguarding the data privacy of individual citizens and guarding their (data) sovereignty to promoting the growth of a domestic digital economy (Chander and Le, 2014; Castro and Mcquinn, 2015, US Chamber of Commerce and Hunton & Williams, 2014). It is particularly striking that data localization policies are proliferating across both liberal/democratic states such as Australia, Canada, and India, as well as illiberal regimes such as China, Vietnam and Iran. The latest to join was Russia, which implemented its data localization law in September 2015, with the stated policy objectives of national security and the protection of privacy of Russian citizens (Bowman, 2015; Kurochkin et al, 2015). Sceptics suspect that these policies are more likely to be instrumental in repressing any political dissent through online platforms and preventing the free flow of information from the outside world to Russia. The impact of Russiaâ&#x20AC;&#x2122;s laws is further expected to have an adverse affect on businesses and cross-border trade, with estimated economic losses amounting to 0.27 per cent of GDP â&#x20AC;&#x201C; particularly harmful in a time of severe economic recession (Lee-Makiyama, 10 June 2015). The purpose of this paper is to study data localization laws from an international political economy (IPE) perspective, its re-

137

THE PUBLIC SPHERE | 2016 ISSUE

percussions, and its policy implications for rules governing digital trade. This exercise is vital to understanding the environment in which data localization laws take force. It is also an important starting point to assess the extent to which trade policy is suited to address regulatory challenges of the digital world today. The majority of existing scholarship on this issue tends to focus on the liberalization of digital trade and the necessity of the cross-border flow of data, which, while compelling, is insufficient to unravel the complex regulatory dilemmas associated with digital data management. While some believe that forced data localization is a deliberate strategy to protect the domestic economy and undercut competition from American IT giants (Lee-Makiyama, 2013; Chander and Le, 2014; Aaronson and Maxim, 2013), others are more sympathetic towards concerns regarding data privacy, surveillance, and guarding the data sovereignty of countries (Rubin, 2015; Kong, 2010). This paper emphasizes that the policy rationale for implementing data localization laws in a country may often be contextual and an outcome of the interaction between “markets” and “politics” in which a given government operates (Gilpin, 2001). The first part of the paper provides an overview of data localization laws across the world, their desired policy objectives, and their impact on the innovation economy at large. The second part of the essay is a critical evaluation of the IPE of data localization policies. A clear distinction is drawn between the economics and the IPE of data localization policies to evaluate the larger policy environment in which laws for digital trade operate. The third part provides recommendations to engage with the issue of data localization more effectively and meaningfully, taking into account both the economic and political realities. The paper concludes by suggesting that trade rules are one of the constituent elements of the larger digital trade economy and may have limited impact in addressing data localization problems unless countries are willing to negotiate on interconnected issues of privacy, con-

138

[Data localization laws in a digital world]

sumer protection and data sovereignty. The latter necessitates enhanced standards of transparency, international cooperation and political compromise at a global level.

DATA LOCALIZATION LAWS AND THEIR IMPACT ON TRADE Implementation of data localization laws became vigorous in the aftermath of the Snowden affair in June 2013, which brought damning evidence to the international community of the extent to which the US National Security Agency had been surveilling online information of both American and foreign citizens and companies (Dhont and Woodcock, 2015; Donnan, 14 August 2014; Donohue, 2015; Hill, 2014). However, the move towards data localization predates the Snowden incident. For instance, as early as 2005, the Government of Kazakhstan passed a law requiring that all data sited in the .kz domain be located domestically, but later made an exception for technology companies such as Google (Castro and Mcquinn, 2015). Requirements for data localizations or restrictions on free flow of data have been made in recent years in countries such as Vietnam, Indonesia, Brunei, Iran, China, Brazil, India, Australia, Korea, Nigeria and, most recently, Russia (For details, see Chander and Le, 2014; Castro and Mcquinn, 2015, Dhont and Woodcock, 2015; Svatesson, 2010). It should be noted that while some of these countries impose a blanket ban on the transfer of all categories of personal data abroad, others, such as Australia and South Korea, impose specific restrictions on the transfer of data in sectors such as health and finance on grounds of protecting citizensâ&#x20AC;&#x2122; sensitive data. For some countries, such as Malaysia and the Philippines, strict consent requirements and regulatory approvals for overseas data transfer exist, which tend to slow down the process and often result in forced data localization. Some countries such as India

139

THE PUBLIC SPHERE | 2016 ISSUE

also require foreign companies to enter into local partnerships to provide various IT services (Hill, 2014). In many of these countries, it appears that data localization requirements are pushed as a protectionist measure to boost the domestic digital economy while – especially in the case of Russia, China, Vietnam and Iran – there are added concerns of state control and censorship of data (Atkinson, 2010; Hill 2014; Chander and Le, 2014; Bajoria, 5 June 2014). The EU is considered to be one of the strictest regimes in the world for data privacy. EU regulations prohibit the transfer of data belonging to EU residents to outside jurisdictions lacking “adequate” data privacy regimes. However, companies who meet the required standards under the regulation are deemed to qualify for “safe harbour” protections (Directive 95/46/EC, 24 October 1995). In 2012, the European Commission proposed a unified regime called the General Data Protection Regime (“GDPR”) which imposes much stricter standards on data privacy and protection in light of new technologies such as cloud computing and the rise in social networking. A study commissioned by the American Chamber of Commerce (Bauer et al, 2013) estimated that the implementation of GDPR would reduce the GDP of the EU between 0.8 to 1.3 percentage points. Furthermore, if the ‘right to be forgotten’ rule was incorporated in the GDPR, the losses to GDP would be far higher, in the range of 1.5 to 3.9 percentage points (See also Ezell et al, 2014). While some argue that the EU data privacy regime is grounded in the cultural context (Milberg et al, 2000), others highlight that the proposed regime is disproportionate and excessive, and potentially protectionist (Lee-Makiyama, 2013; Baeur et al, 2013). The EU maintains that their privacy regulations make a legitimate distinction between rational protection of data and data protectionism (Panel discussion on Trade Agreements and Data Flow [see statement of Ignatio Irrurarezaga, Head of Unit on Services, DG Trade], 30 July 2015). It is not easy to assess this distinction in practice.

140

[Data localization laws in a digital world]

For instance, in recent years, several technology giants such as Microsoft, Google, Apple and Amazon have built data centres in Dublin and Denmark (The Irish Times, 5 March 2015; Apple Press Info, 23 February 2015). It is perhaps also noteworthy to mention that, from the Snowden revelations to the dispute between Max Schrems and Facebook (Judgment in Case C-362/14, Maximillian Schrems v Data Protection Commissioner, â&#x20AC;&#x153;Schrems caseâ&#x20AC;?), it has been contended that several American companies which enjoy safe harbour protection do not have privacy protection tantamount to the EU legal standard (Crawford, 26 March 2015). The recent CJEU judgment in the Schrems case invalidated the US-EU Safe Harbour Argument and thereby raises problematic questions on current business models of digital trade, even though press reports had previously indicated the possibility of a political compromise on the EU-US Safe Harbour Agreement (Reuters, 2015). As expected, in several recent trade negotiations such as the Transatlantic Trade and Investment Partnership (TTIP), the EU is under pressure to advocate for higher standards of data privacy in order to win greater public trust within their domestic jurisdiction(s), while there is strong pushback from the US. One of the basic problems with complying with data localization laws for companies is the difficulty in determining which categories of data need to be locally stored and which can be moved abroad. A recent study indicates that distinguishing personal data from non-personal data for purposes of data localization is a complex issue (Bauer et al, 2015). Legal experts also have a problem in assessing the legal liability of foreign companies who do not have a business presence in a country but may be handling large quantities of data of citizens in the course of normal day-to-day business transactions in the global marketplace. Furthermore, when companies are forced to relocate their servers to jurisdictions with low levels of internet security, concerns arise regarding possible breaches of consumer trust,

141

THE PUBLIC SPHERE â&#x20AC;&#x2020;| â&#x20AC;&#x2020;2016 ISSUE

which may invite further legal liability. Finally, foreign companies are also worried to operate data centres in several authoritarian jurisdictions where state censorship and surveillance laws are over-encompassing and create significant liabilities for service providers. While these laws fail to satisfy their objective of providing adequate security and privacy to user data, they create several economic and technical drawbacks which can prejudice the resilience and the utility of the internet as a platform for communication and trade. Discussed below are the main drawbacks of data localization, which not only prejudice economic benefits from digital trade, but also hamper data security and interfere with the broader architecture of the internet. First, several studies have shown that the free flow of data is not only critical for information and communication technologies (ICT) services, but of paramount importance to trade in goods and services in general (For an elaborate discussion, see Van-der Marel, 2015). This not only affects the global economy, but also countries that implement such measures, thus counteracting any potential policies to boost the local market (Ezell et al, 2013; McKinsey Global Institute, 2014; Donnan, 2014; Baeur et al, 2013; 2014; 2015). For instance, a study by McKinsey (2014) showed that 75 per cent of the value addition from data flows goes to traditional manufacturing industries while another study by United Nations Conference on Trade and Development (UNCTAD) showed that 50 per cent of all traded services are enabled through the ICT industry (Castro and Mcquinn, 2015). A recent study by ECIPE (Bauer et al, 2014) has shown that restriction on cross-border data flows adversely impacts countries which adopt such laws â&#x20AC;&#x201C; for instance, in Indonesia, data localization laws could reduce GDP by 0.7 per cent and reduce investments by 2.3 per cent. Similar results were also recorded for South Korea and the EU (Baeur et al, 2013; 2014; 2015). As cross-border trade increasingly moves towards e-commerce and relies on the use of internet technologies such as cloud computing

142

[Data localization laws in a digital world]

and big data, data localization policies pose a major threat to the economy. Not surprisingly, several business associations (mostly consisting of global market leaders based in the US) (AmCham China, 2015; Information Technology Industry Council, 15 October 2014; Asia Internet Coalition, 2014; US Chamber of Commerce, 2015) as well as a few governments (most evidently, the US government) (Office of the United States Trade Representative [USTR], 2015), have presented a strong economic case for the free flow of data in various regional trade agreements such as the Trans Pacific Partnership (TPP), Trade in Services Agreement (TISA) and TTIP. Second, data localization laws do not necessarily provide a solution to problems of data breaches or boosting data security. On the contrary, by compelling forced localization of data, these laws are very likely to make data more vulnerable to both security attacks and natural disasters, as the data no longer undergoes sharding1(Heidt, 2015). Particularly, in countries with poor IT security systems, data localization defeats the purpose of data protection. Further, several foreign governments such as the US use sophisticated malware for data surveillance – hence, simply relocating data is of no use. Moreover, there is an increased risk of local surveillance through implementation of such laws (Chander and Le, 2014; 2015). Third, mandating localization of data centres is against the economic logic of the technology industry, which is primarily based on global economies of scale. Furthermore, localization raises costs, and reduces competitiveness and productivity for both local consumers and businesses (Ezell and Atkinson, 2010). It is also detrimental to several modern-day innovations in big data, cloud computing and “Internet of Things”, which bring several benefits to any local economy, including higher levels of efficiency and cost-effectiveness in businesses. More importantly, data centres are largely automated – hence, they do not generate significant levels of employment, though upfront investment and

1. In the process of sharding, data is partitioned and stored over multiple physical locations, rather than one location.

143

THE PUBLIC SPHERE | 2016 ISSUE

long-term energy costs are very high (The Wall Street Journal, 13 November 2013; ZD Net, 2 April 2013). Finally, data localization initiatives adversely impact the overall structure of the current form of internet governance, which is based on a world-wide network of exchange of information and data. Localized data centres may result in splintering the internet (“splinternet”), which is unlikely to be rewarding to businesses and consumers alike (US Chamber of Commerce and Hunton & Williams, 2014). Moreover, if country A does not trust country B to host its citizens’ data, the reverse is also likely to be true, ultimately leading to disruption of an interconnected network. From the discussion above, it is evident that the costs of data localization are too high and the achievement of desired policy outcomes mostly uncertain. Particularly, Governments have the option of adopting alternate standards such as enforcing strict end-to-end encryption standards based on a recognized international standard instead of imposing measures that restrict trade and innovation. However, governments increasingly adopt localization measures, both in the developing and developed world alike. This begs the fundamental question regarding policy motivations that result in implementation of data localization laws.

DATA LOCALIZATION FROM AN IPE POINT OF VIEW From an IPE perspective, a dominant view is the emergence of “innovation mercantilism” in digital trade today (Ezell et al, 2013; Atkinson, 2010). Prior to the widespread adoption of liberal values, mercantilism (or economic nationalism) prevailed from the 16th to 18th centuries. The prevailing policies in mercantilism include a favourable balance of trade, protection of domestic industries, boosting local employment, species accumulation, and manipula-

144

[Data localization laws in a digital world]

tion of exchange rates to keep exports competitive (The Economist, 23 August 2013; Douglas, 1991; Rodrik, 2013). However, scholars repeatedly indicate that certain aspects of mercantilist practice play an important role in doctrines of classical liberalism (Grampp, 1952). For instance, classical liberalists like Adam Smith argued that national security was an important exception to free trade. He also acknowledged that states cannot merely be driven by an economic logic of the market, but will also be driven by social interests such as protecting the people who emerge as losers in a free market (Walter, 1996). Even liberal governments regulate economic activity and provide protection or financial support to some groups within their country based on their political power or, more rarely, on grounds of social justice. With this background in mind, it is important to explore how the political economy has evolved as digital data has become the â&#x20AC;&#x153;new currencyâ&#x20AC;? in international trade. There is a perception that in a digital world, countries that have access to more data are in a better position to maximize their wealth and power. It is therefore not surprising that more and more developing countries make an effort to prevent the export of data across borders or even tax the flow of data, with the hope that it would drive greater innovation, generate more revenues for local enterprises, and drive investment into the domestic economy. To trade practitioners, this is a familiar case of infant industry protection through non-tariff measures. Further, countries (particularly those with larger populations and/or resources, such as China) are likely to consider data localization as a strategic tool to gain control over more data at home and abroad, thereby providing credible competition to some of the biggest American players who dominate the market today. Again, many trade practitioners would recognize this is as a use of a non-tariff barrier to gain greater market shares. However, this narrative is incomplete in terms of recognizing several other strategic interests (aside from gaining trade) that drive domestic policies in the digital world today.

145

THE PUBLIC SPHERE â&#x20AC;&#x2020;| â&#x20AC;&#x2020;2016 ISSUE

Services such as cloud computing, e-commerce, and big data processing now allow some of the biggest American internet companies to collect and control vast amounts of data. Many developed countries and fast-developing economies such as China recognize that overpowering American leadership in digital space is only possible if they develop indigenous data processing facilities that reach a larger chunk of the global market. The trade war between the US and China is particularly instrumental in understanding the policy context. Other developed countries such as Germany, France, Norway and Australia also aspire to become key leaders in the digital economy (Chander and Le, 2015). However, the power competition is not merely for economic gains, but equally for political ones. As both American and Chinese companies have access to greater amounts of data, their political power and control increases significantly and extends well beyond the realms of trade. This was evident when the US blocked the use of Huawei products on the suspicion that it would provide backdoor access to data located within the US to the Chinese government (BBC News, 2014). Similarly, China has successfully blocked the likes of Google and Facebook from operating within their jurisdiction â&#x20AC;&#x201C; arguably not only with the intention of generating more business opportunities for local players, but also based on strategic political interests. The developments in the EU are also equally interesting. With the increased suspicion of American surveillance, the EU, which enjoys a reputation for strict enforcement of data protection, now has new players from France, Germany and Norway who claim that their digital services are surveillance-free, aiming to appeal to both the EU and the global market (Chander and Le, 2015). At the same time, a large number of developing countries, including Brazil, India, Indonesia and several other fast developing countries in Southeast Asia, are aiming to expand their reach in global markets through data localization policies. Even some relatively less developed countries, such as Nigeria, are eager to

146

[Data localization laws in a digital world]

expand their digital economy and are therefore keen on building local data centres. Often, their policy impetus for data localization laws is stated to be privacy and national security, although it is arguably a policy tool to conceal protectionism. It is often assumed that local data centres will create greater employment, skills upgrading and an overall improvement in the economy; in other words, help developing countries to move up the global value chains. This narrative can be very politically appealing, particularly given that it evokes a sense of nationalism and self-sufficiency. Consider the example of Indonesia: the strong wave of economic nationalism in various spheres of economic activity resulted in massive support for the government, but has so far resulted in limited returns to the economy (The Economist, 7 May 2015; The Economist, 7 May 2015a). It particularly results in causing harm to smaller, local businesses, who are deprived of cost-effective, secure services provided by several foreign players, which consequently tends to increase their transaction costs significantly (ITI, JEITA and Digital Europe, 2014; eBay, 2015). In countries with monopoly telecom providers (usually stateowned), there is a high incentive for telecom lobby groups to block out foreign competition. Business estimates increasingly show that many of these companies are entering the cloud computing markets and ousting their foreign counterparts (Kehl et al, 2014). For instance, in Russia, the two most important gainers from the data localization law would be Rostec, a stateowned entity, as well as Rostelcom, the monopoly telecommunications provider in Russia (Data Centre Knowledge, 2015). Several of these developing countries, however, do not have appropriate infrastructure in place to set up secure and efficient data centres. This includes a secure technical infrastructure as well as an appropriate legal regime. For instance, internet fraud rates are highest in the world in Brazil, while in the past years Indonesia has shared a similar reputation for cybercrime (NPR, 2015). Even the Data Centre Risk Index, which ranks coun-

147

THE PUBLIC SPHERE â&#x20AC;&#x2020;| â&#x20AC;&#x2020;2016 ISSUE

tries in terms of technical, economic and political security of data centres, finds that several jurisdictions implementing data localization laws such as China, Brazil, Indonesia and Russia are ranked very low (Cushman and Wakefield, 2013). Further, supporting laws such as privacy laws, data protection laws, IP laws, as well as laws to protect against political persecution, are essential to maintain integrity and stability of data centres; this is often missing in many developing countries (UNCTAD, 2013; Cushman and Wakefield, 2013). There is no systematic study to date that shows that there are any foreseeable positive returns to the economy through data localization, either from the additional investments or the technology transfer that comes in when a foreign company relocates its servers to a country. In fact, companies have strong economic (and often strategic) reasons for locating their servers in other parts of the world â&#x20AC;&#x201C; hence, forced localization of data centres may not reap the expected economic returns. However, states are not just interested in increasing revenues for their own economy, but also in their position relative to other countries, commonly known as relative gains. States are often driven by relative gains, even at the cost of sacrificing absolute gains (Krasner, 1976). For instance, the Indonesian or Vietnamese government would be more keen to have the likes of Google and Microsoft open up data centres in their country rather than in neighbouring Thailand or Malaysia, even though it is very likely that their domestic markets are less equipped to host data centres and would benefit more by using foreign data centres. Although economically counter-intuitive, the idea of relative gains is an important driver of national policy as it finds surprising levels of political support (Drezner, 2012). The other important idea driving state policy is the protection of national autonomy. First, it is important to remember that the internet governance mechanism in the past has been heavily privately managed (by bodies such as ICANN) and has

148

[Data localization laws in a digital world]

often been predominantly US-centric. The idea of sovereign control over flow of information over the internet was not given much policy attention for several years. However, this resulted in a “complex cyber regime complex” with a multitude of actors (often private) and institutions. As Nye (2014) effectively puts it, this was a compromise between “a single coherent legal structure and complete fragmentation of normative structures” in cyberspace. Progressively, it is becoming clearer that a sovereignty-based model is irrelevant in a digitally connected world. Rather, a multi-stakeholder approach with a variety of institutional and private actors, and a variety of interests going well beyond trade (human rights, internet governance, surveillance and privacy), may be necessary. This increasingly appears as a threat to the domestic sovereignty that countries enjoy in trade negotiations in bodies such as the WTO. Moreover, US-based companies currently enjoy an almost hegemonic status in the digital world. As a result, there is significant apprehension that the US government exercises strong control over the data goldmines, and reap disproportionate economic and political benefits. These views have been particularly strengthened post-Snowden. Many governments are able to generate political buy-in for adopting stringent consent regimes for data transfer and high thresholds of data protection, as it is likely to weaken American hegemony channelled through technology giants based out of the US. Furthermore, the privately-regulated nature of the internet, along with non-transparent actions of the US Government, are likely to make countries wary of letting data on its citizens flow freely out of the country. In fact, trade forums such as the WTO barely touch upon privacy and transparency issues within the internet. Therefore, despite the strong economic logic for letting data move around freely, it appears legitimate for countries to be concerned about national security and the preservation of autonomy. These exceptions are also well built in the WTO Agreements such as under Article XIV and XIV: bis of GATS.

149

THE PUBLIC SPHERE â&#x20AC;&#x2020;| â&#x20AC;&#x2020;2016 ISSUE

One of the most effective ways of allowing for safe transmission of data online is to execute end-to-end encryption of data. This raises interesting questions of sovereignty and standard-setting. While technology companies should ideally develop trustworthy and secure standards to encrypt data, bodies such as the Federal Bureau of Investigation in the US have required companies to have a backdoor mechanism by which they could break the encryption in case of sensitive investigations (The Guardian, 8 July 2015). This raises important questions of transparency and trust in the standards that are used in digital services and goods. Particularly, since standard regulation of the internet tends to be based on private protocols and often set by American companies, other sovereign states are likely to consider it an affront to their sovereignty because they have no say in the adoption of such standards. Further, once data resides abroad, they do not exercise any jurisdictional control over such data. Therefore, countries may also develop indigenous standards for data encryption in order to avoid extra-territorial surveillance. For instance, Chinese digital products and services provide for end-to-end encryption, which is developed locally. For the American Government, this is not only viewed as a cause of suspicion, but also an affront to the position of domination of American companies in technology trade. The issue of law enforcement in the digital world also creates an incentive for data localization. Particularly, with the increase in the rates and scale of cybercrimes, governments cannot afford to ignore issues of law enforcement. Mulvenon and Denmark (2010) argued that the global cyber commons is comparable to the Wild West of the 1870s and 1880s, where rule of law barely prevailed and it was up to each individual to take care of themselves. Similarly, in the absence of any cohesive rules in the current digital space, internet users are more or less responsible for ensuring their security. Evidently, there is a discord between the economic logic behind allowing cross-border data flows â&#x20AC;&#x201C; discussed in the previ-

150

[Data localization laws in a digital world]

ous section â&#x20AC;&#x201C; and the political economy of data localization policies. In fact, the discussion dictates that existing dialogue on data localization touches upon clashing political philosophies and is not solely a case of straightforward data protectionism. The existing discourse on data localization makes compelling economic and technical arguments as to why data localization policies are harmful for trade, investment, cyber security and innovation. However, it fails to understand the strong political rationale driving such policies and the complex political economy of the digital world.

STRIKING THE RIGHT BALANCE The contentious issues with respect to data localization extend well beyond free trade versus protectionism into some delicate, complex and legitimate political concerns, such as technology transfer and IP rights, privacy, human rights, and national security, which is currently missing (and expectedly so) on most trade agendas. To take an example: in the course of the TTIP negotiations, it became clear that addressing issues of data protection and privacy along with the free flow of data was inevitable to reach any kind of consensus between two great powers; the US and the EU (Kong, 2010). Similar possibilities exist with respect to the TPP, even though the US holds a dominant negotiating position. In fact, time and again, the US Governmentâ&#x20AC;&#x2122;s reluctance to engage on issues of privacy and transparency, combined with its agenda on unhindered liberalization of data flows, has caused severe discontent amongst both developing and developed countries in negotiations of various Regional Trade Agreements (RTAs). Some scholars and business associations (often US-centric) have suggested taking stronger actions in trade forums such as the WTO or using the instrumentalities of USTR and mega-regionals such as the TPP to prevent restrictions on

151

THE PUBLIC SPHERE | 2016 ISSUE

cross-border data flow. For instance, the Trade Bill introduced in the Senate in 2013 (S. 1788, 113th Congress, 2013-14), inter alia, proposed negotiating for free flow of data and prevention of data protectionism in US trade negotiations. These measures are often viewed as a US strategy to maintain its position in the digital market and wipe out potential competitors. Notably, the WTO is also ill-equipped (with its current arrangements) to become a focal centre for negotiations on cross-border data flows or addressing disputes related to these matters. Issues related to data protection and privacy often involve a plurality of interests and competing points of view (often embedded in deeper political philosophy of the countries and other relevant actors (See Post, 2001)); a consensus-based forum such as WTO barely facilitates any possibility of a compromise. Countries such as India and Brazil feel unprepared to negotiate on issues related to e-commerce (Livemint, 30 July 2015), which is likely to include discussions around free flow of data. Chander (2009) recommends the adoption of a “glocalization” approach in digital space, where laws can be harmonized globally, paying specific attention to local interests. This paints an optimistic scenario of extreme international cooperation on issues of internet standards and interoperability, mutual respect for domestic cultures and legal cooperation through treaties such as the Mutual Legal Assistance Treaties to enable free flow of data. A realist may dismiss this as a utopian state of affairs. A more pragmatic approach to address concerns regarding data localization could be to incentivize governments across the world (particularly the developing world) to allow data to flow across borders freely. This could be done by creating opportunities for local businesses to harness the multiplier effect of innovations in big data and cloud technology (say for instance, the potential opportunity for digitally powered local companies and SMEs to participate in worldwide e-commerce markets [eBay, 2015]), while removing misunderstandings surrounding secu-

152

[Data localization laws in a digital world]

rity of cloud and big data technologies. Rather than using pressure tactics in trade forums such as RTA negotiations, the US government could show more engagement with issues related to sensitive political concerns. Countries such as the US, Australia and Singapore, which host global data centres, are likely to be viewed with suspicion, particularly when American companies are involved. There is a need to push for greater commitment and transparency from the US Government and American IT giants, in particular, to engage in a public dialogue to gain trust from their foreign counterparts. This can be coupled with a greater effort by technology companies to emphasize that the setting up of data centres alone has limited economic returns but investments in technology innovations (whether foreign or local) generates manifold increase in economic returns to traditional manufacturing sectors. It is also important to highlight to governments the potential dangers of having insecure domestic systems, particularly if they store sensitive data. Ensuring data security is not simply based on the geographical location of the data servers, but also critically depends on the security and robustness of IT systems. More often than not, larger companies are better situated to ensure these services than local companies. In particular, this is likely to raise issues of standard-setting. Through transparent processes and open deliberations, international organization such as the ITU can enable such discussions amongst countries.

CONCLUSION It is much a more comprehensive and viable policy approach to view trade negotiations for the free flow of data as a political economy issue rather than a strictly economic one. Although data localization laws have the capacity to fracture the digital trade ecosystem, political priorities cloud the policy space with

153

THE PUBLIC SPHERE | 2016 ISSUE

illusive promises of economic growth and data sovereignty, which need to be addressed head-on. Economic efficiency is not the sole benchmark in policy choices – concerns related to trust in digital goods and services cannot be ruled out. Powerful internet companies, along with their host governments, will need to be more transparent and ethical regarding the use of personal data. This will also increase the possibility of engaging in more policy innovations in the digital space, which is likely to undercut the wave of data localization. A good example in this regard is the adoption of data categorization standards in cloud computing, which allows for automatic cross-border transfer of most routine transaction data, and only imposes additional legal and security requirements for politically sensitive or highly secure data. In particular, as most of these policy innovations tend to start in the private realm, transparency is critical. Recognition and reconciliation of the multitude of interests and ideologies in digital trade is not easy – however, ignoring these complexities in trade agreements will most likely be counter-productive to maintaining the integrity of a digital trade ecosystem.

REFERENCES Aaronson, Susan and Maxim, Rob (2013). “Data Protection and Digital Trade in the Wake of NSA Revelations.” http://elliott.gwu.edu/sites/elliott.gwu.edu/files/downloads/ research/aaronson-Data%20Protection%20and%20Digital%20Trade%20in%20 the%20Wake%20of%20the%20NSA%20Revelations.pdf AmCham China (2015). “Protecting Data Flows in the US-China Bilateral Investment Treaty.” http://www.amchamchina.org/policy-advocacy/policy-spotlight/data-localization Apple Press Info (23 February 2015). “Apple to Invest €1.7 Billion in New European Data Centres.” https://www.apple.com/ie/pr/library/2015/02/23Apple-to-Invest-1-7-Billionin-New-European-Data-Centres.html Asia Internet Coalition (2014). “Coalition for cross-border data flows.” http://www.asiainternetcoalition.org/wp-content/uploads/2014/10/Data-Resource-Paper-July-3.pdf Atkinson, Robert (2010). “Time to end rampant mercantilism.” G8G20 Summit 2010. http://www.itif.org/files/G8G20_RAtkinson.pdf Bajoria, Jayshree (5 June 2014). ”India’s Snooping,” Wall Street Journal (India). http://blogs. wsj.com/indiarealtime/2014/06/05/indias-snooping-and-snowden/

154

[Data localization laws in a digital world]

Bauer, Matthias et al (2013). “The Economic Importance of Getting Data Protection Right: Protecting Privacy, Transmitting Data, Moving Commerce.” https://www.uschamber. com/sites/default/files/documents/files/020508_EconomicImportance_ Final_Revised_lr.pdf Bauer, Matthias et al (2014). “The Costs of Data Localization: Friendly Fire on Economic Recovery.” ECIPE Occasion Paper no. 3/2014, http://www.ecipe.org/app/uploads/2014/12/ OCC32014__1.pdf Bauer, Matthias et al. (2015). “Data localization in Russia: a self-imposed sanction.” ECIPE Policy Brief no. 6/2015. http://www.ecipe.org/app/uploads/2015/06/Policy-Brief-062015_Fixed.pdf BBC News (16 October 2014). “Huwaei boss says US ban not very important.”http://www. bbc.com/news/business-29620442# Bowman, Catherine (27 August 2015). “A Primer on Russia’s New Data Localization Law (blogpost).” Privacy Law Blog. http://privacylaw.proskauer.com/2015/08/articles/international/a-primer-on-russias-new-data-localization-law/ Castro, Daniel and Mcquinn, Alan (2015). “Cross-Border Data Flows Enable Growth in All Industries.” http://www2.itif.org/2015-cross-border-data-flows.pdf Chander, Anupam and Le, Uyen (2015). “Data Nationalism.” Emory Law Journal 64, no. 1: 677-739. http://law.emory.edu/elj/content/volume-64/issue-3/articles/data-nationalism.html Chander, Anupam (2009). “Trade 2.0.” UC Davis Legal Studies Research Paper Series Research Paper No. 173. http://www.law.uchicago.edu/files/files/465-chander-trade.pdf Chander, Anupam and Le, Uyen (2014). “Breaking the Web: the Global Internet v/s Data Localization.” UC Davis Legal Studies Research Paper Series Research Paper No. 378. http:// papers.ssrn.com/sol3/papers.cfm?abstract_id=2407858 Crawford, Gail (26 March 2015), “Snowden’s legacy: Safe harbour under fire at the CJEU (blogpost).” http://www.globalprivacyblog.com/privacy/snowdens-legacy-safe-harbor-under-fire-at-the-cjeu/ Cushman and Wakefield (2013). “Data Centre Risk Index.” http://www.cushmanwakefield. com/~/media/global-reports/data-centre-risk-index-2013.pdf Dani Rodrik (2013). “The new mercantilist challenge,” http://www.project- syndicate.org/ commentary/the-return-of-mercantilism-by-dani-rodrik Data Centre Knowledge (21 July 2015). “Firms rethink Russian data strategy.” http:// www.datacenterknowledge.com/archives/2015/07/21/russian-data-localization-law-spurs-data-center-strategy-changes/ Dhont, Jan and Woodcock, Katherine (2015). “Data localization requirements: Growing trends and impact for company compliance,” Corporate and Ethics Professional. http:// www.lorenz- law.com/wp-content/uploads/Data-localization-requirements-Growing-trends-and-impact-of- compnay-compliance1.pdf Donnan, Shawn (14 August 2014). “Digital trade: Data protectionism.” The Financial Times. http://www.ft.com/intl/cms/s/0/93acdbf4-0e9b-11e4-ae0e- 00144feabdc0.html#axzz3WD30VpD4 Donohue, Laura (2015). “High Technology, Consumer privacy and US National Security,” Georgetown Law Faculty Publications and Other Works. Paper 1457. http://scholarship. law.georgetown.edu/facpub/1457/ Drezner, Daniel (20 June 2012). “A most unusual foreign policy poll.” Foreign Policy. http:// foreignpolicy.com/2012/06/20/a-most-unusual-foreign-policy-poll/ eBay Inc. (2015). “Commerce 3.0: Enabling Small and Medium Enterprises.” http://www. ebaymainstreet.com/sites/default/files/asean_commerce_3_0_final_1.pdf

155

THE PUBLIC SPHERE | 2016 ISSUE

Ezell, Stephen (2012). “The Benefits of ITA Expansion for Developing Countries.” http:// www2.itif.org/2012-benefits-ita-developing-countries.pdf Ezell, Stephen (2014). “Safeguarding Digital Trade is vital to ensure thriving innovation economy,” Bridges no. 39. http://ostaustria.org/bridges-magazine/item/8177-safeguarding-digital-trade-is- vital-to-ensuring-a-thriving-global-innovation-economy?tmpl=component&print=1 Ezell, Stephen and Atkinson, Robert (2010). “The Good, the Bad, the Ugly, and the Self-Destructive of Innovation Policy .” http://www.itif.org/files/2010-good-bad-ugly.pdf Ezell, Stephen et al. (2013). “Localization barriers to global trade: threat to the global economy.”http://www.itif.org/publications/localization-barriers-trade-threat-global-innovation- economy Gilpin, Robert (2001). Global Political Economy. New Jersey: Princeton University Press. Grampp, William (1952). “The Liberal Elements in English mercantilism.” The Quarterly Journal of Economics 66, no. 4: 465-501. Heidt, Franz (2015). “The Harms of Forced Data Localization,” https://www.leviathansecurity.com/blog/the-harms-of-forced-data-localization/ Hill, Jonah (2014). “The Growth of Data Localization Post-Snowden.” Lawfare Research Paper Series 2, no. 3:1-40. Information Technology Industry Council (15 October 2014). “ITI Joins Broad International Coalition Urging TPP Negotiations Ensure Robust Cross-Border Data Flow Provisions.”http://www.itic.org/news-events/news-releases/iti-joins-broad-international-coalition-urging-tpp-negotiations-ensure-robust-cross-border-data-flow-provisions Irwin, Douglas (1991). “Mercantilism as a strategic trade policy.” The Journal of Political Economy 99, no. 6: 1296-1314. ITI, JEITA and Digital Europe (2014). “The Tokyo Resolution on Combatting Data Localization Requirements.” http://www.digitaleurope.org/DesktopModules/Bring2mind/ DMX/Download.aspx?Command=Core_Download&EntryId=841&PortalId=0&TabId=353 Kehl, Danielle et al (2014). “Surveillance Costs: The NSA’s Impact on the Economy, Internet Freedom and Cyber security.” http://newamerica.net/publications/policy/surveillance_costs_the_nsas_impact_on_the_economy _internet_freedom_cybersecurity Kong, Lingjie (2010). “Data Protection and Transborder Data Flow in the European and Global Context.” European Journal of International Law 21, no. 2: 441-56. Krasner, Stephen (1976). “State power and the structure of international trade”, World Politics, pp. 317-347 http://www.indiana.edu/~gradipe/docs/krasner.pdf Kurochkin, Dmirt et al (2015). “Russia’s New Server Localization Law: Implications for foreign companies.” World Data Protection Report 15, no. 2: 1-3. https://www.dechert.com/ files/Uploads/Documents/Bloomberg%20-%20Russia%20New%20Server%20Localization%20Law%20-%20Dechert%20LLP%20-%20February%202015.pdf Lee-Makiyama, Hosuk (2013). “European leaders show leave data flows open.” http:// www.euractiv.com/infosociety/european-leaders-leave-data-flow-analysis-530785 Lee-Makiyama, Hosuk (2015). “Data localization requirement in Russia.” http://www.ecipe.org/blog/data-localisation-russia/ Livemint (30 July 2015). “India, US on a collision course over e-commerce, IP norms.” http:// www.livemint.com/Politics/FNMRmLJElPC6zwA7K4NZQL/India-US-on-a-collisioncourse-over-ecommerce-IP-norms.html McKinsey Global Institute (2014). “Global flows in a digital age.” http://www.mckinsey. com/insights/globalization/global_flows_in_a_digital_age

156

[Data localization laws in a digital world]

Milberg, Sandra et al. (2000). “Information Privacy: Corporate Management and National Regulation.” Organizational Science 11, no. 1: 35-57. Mulvenon, James and Abraham Denmark (2010). “Contested Commons: The Future of American Power in a Multipolar World.” http://www.cnas.org/files/documents/publications/CNAS%20Contested%20Commons_1.pdf NPR (15 June 2015). “Brazil’s Cybercrime Free-For-All: Many Scams and Little Punishment.” http://www.npr.org/sections/parallels/2015/06/15/414622197/brazils-cybercrimefree-for-all-many-scams-and-little-punishment Nye, Joseph. (2014). “The Regime Complex for Global Cyber Activities.” https://www.cigionline.org/sites/default/files/gcig_paper_no1.pdf Office of the United States Trade Representative (2015). “Trans-Pacific Partnership: Summary of US Objectives.” https://ustr.gov/tpp/Summary-of-US-objectives Panel Discussions (30 July 2015). “Trade Agreement and Data Flows: Safeguarding the EU Data Protection Standards.” http://europe-liberte-securite-justice.org/2015/07/30/ trade-agreements-and-data-flows-safeguarding-the-eu-data-protection-standards/ Post, Robert (2001). “Three concepts of privacy.” Faculty Scholarship Series. Paper 185. http:// digitalcommons.law.yale.edu/fss_papers/185 Reuters (US Edition) (5 August 2015). “EU close to sealing deal with US on data sharing. ” http://www.reuters.com/article/2015/08/05/us-usa-eu-data-idUSKCN0QA1XB20150805 Robert, Atkinson (2014). “The Rise of Innovation Mercantilism.” International Economy 28, no. 2: 30-32 and 55-56. Rubin, Ryan (2015). “Companies should prepare for EU’s forthcoming data protection regulation.” http://www.euractiv.com/sections/infosociety/companies-should-prepare-eus-forthcoming- data-protection-regulation-312487 Svantesson, Dan (2010). “Privacy, the Internet and transborder data flows - An Australian perspective,” Masaryk University journal of law and technology 4, no.1: 1-20. The Economist (23 August 2013). “What was mercantilism?.” http://www.economist.com/ blogs/freeexchange/2013/08/economic-history The Economist (7 May 2015). “ Spicing up growth.” http://www.economist.com/news/finance-and-economics/21650586-bad-policy-much-bad-infrastructure-holding-indonesia-back-spicing-up The Economist (7 May 2015a). “Jokowi’s to-do list.” http://www.economist.com/news/ leaders/21650544-indonesias-president-should-ditch-his-economic-nationalismand-if-necessary-his-party-jokowis The Guardian (8 July 2015). “FBI chief wants ‘backdoor access’ to encrypted communications to fight Isis.” http://www.theguardian.com/technology/2015/jul/08/fbi-chiefbackdoor-access-encryption-isis The Irish Times, “Ireland’s Data Centre Boom to Continue,” March 5, 2015, http://www. irishtimes.com/business/technology/ireland-s-data-centre-boom-set-to-continue1.2126081 The Wall Street Journal (13 November 2013). “Brazil legislators bear down on Internet Bill.” http://www.wsj.com/articles/SB10001424052702304868404579194290325348688 UNCTAD (2013). “Information Economy Report.” http://unctad.org/en/PublicationsLibrary/ier2013_en.pdf US Chamber of Commerce (19 May 2015). “Safeguard Cross-border data flows.” https:// www.uschamber.com/issue-brief/safeguard-cross-border-data-flows

157

THE PUBLIC SPHERE | 2016 ISSUE

US Chamber of Commerce and Hunton & Williams (2014). “Business without borders.” https://www.huntonprivacyblog.com/files/2014/05/021384_BusinessWOBorders_final.pdf Van der Marel, Erik (2015). “Disentangling the flows of data: Inside or Outside of the Multinational Company?.” http://www.ecipe.org/publications/flows-data-inside-outside-multinational-company/?chapter=all Walter, Andrew (1996). “Adam Smith and the Liberal Tradition in International Relations.” Review of International Studies 22, no. 1: 5-28. ZD Net (2 April 2013), “Is it better to build or outsource a data center?” http://www.zdnet. com/article/is-it- better-to-own-or-outsource-your-data-center/

158

Evaluating the strength of commitments made to swing constituencies An analysis of the Conservativesâ&#x20AC;&#x2122; decision to block Heathrow Expansion

Ben Stevenson

Master of Public Administration Candidate, Class of 2016 London School of Economics and Political Science

ÂŠMaciej Bledowski

ABSTRACT This paper uses the Conservativesâ&#x20AC;&#x2122; commitment to block Heathrow expansion to shed light on the role of core and swing communities in distributive politics. While previous studies have focused on the allocation of benefits, the analysis presented here assesses a geographically specific tax, namely the placement of activities that are unpopular with nearby constituencies. By studying the political conditions leading up to the 2010 general election, the essay evidences that the Conservativesâ&#x20AC;&#x2122; behaviour exhibited a strong swing-voter bias. However, subsequent shifts within the political landscape bring the durability of their commitment into question, since the electoral incentives may have reverted to supporting the partyâ&#x20AC;&#x2122;s core, pro-business ideology.

[Evaluating the strength of commitments made to swing constituencies]

INTRODUCTION Whether parties target benefits to core or swing voters is a longstanding debate within distributive politics. The logic that politicians would convey selective, non-programmatic transfers to the most loyal supporters was enunciated by Cox & McCubbins (1986) but challenged by Dixit & Londregan (1996), who demonstrated the considerable electoral incentives of targeting swing voters. Many have tested the competing theories by examining the determinants of clientelistic or “pork-barrel” rewards (Stokes, 2005; Jablonski, 2014; Vaishnav & Sircar, 2012), however, few have analysed the allocation of activities that are unpopular with local voters. By framing the expansion of London’s Heathrow airport as a tax on local constituents, the following essay assesses whether the core-swing framework helps to illuminate the Conservative Party’s decision to block the construction of a third runway in the run-up to the UK’s 2010 general election. In accordance with Dixit & Londregan (1996), the analysis finds that the notion of swing-bias helps explain why the Conservative party turned against its core, pro-business ideology, but recent events highlight the fragility of their commitment. After reviewing the literature, the essay details the conditions under which the Conservatives would decide to U-turn on the issue of Heathrow expansion, and in doing so suggests an amendment to Dixit & Londregan’s argument; a static view of core and swing leads to an inflexible understanding of distributive politics, as shifting political constellations can precipitate a rapid change in strategy, and challenge the foundations of political promises.

REVIEW OF THE CORE-SWING DEBATE Core-swing models of distributive politics tend to assume a two-party system, splitting each group’s support into i) the core, whose views strongly reflect the party ideology, ii) its opposition,

161

THE PUBLIC SPHERE | 2016 ISSUE

who are firmly at the other end of the spectrum, and iii) the swing, whose preferences are not clearly aligned. In a bid to explain the stability of political coalitions in the United States, Cox & McCubbins (1986) argued that targeting the core was the more effective strategy. Their work foregrounded the importance of credible commitments, predicting that the swing’s lack of allegiance makes them a risky investment, which encourages politicians to prioritise the core, who are more responsive to their promises. Contrastingly, Dixit & Londregan (1996) predict a swing-voter bias. In their model, voters align with the party closest to their fixed, ideological preference, but will defect if the other party conveys a sufficiently large benefit. A group that is weakly attached to its political ideology can be “bought” relatively cheaply, especially if its average income level is low. Under such conditions, both parties target swing-voters, unless they are particularly efficient at targeting the core. Empirical research has advanced the debate by considering a variety of contextual factors that shape political strategies. Milesi-Ferretti et al. (2002) highlight that certain electoral structures lend themselves to specific forms of distributive politics, with single-member district systems having a greater swing-voter bias than those with proportional representation. Rather than focusing on the overlying electoral system, Jablonski’s study of aid flows in Kenya (2014) stresses the importance of the underlying identity politics; his findings suggested that ethnic ties, coupled with patronage networks, encouraged politicians to adopt a core-voter strategy. Stokes (2005) echoes the importance of the underlying political relationships, but argues against a static framework for assessing the distributive outcomes. Studying clientelistic transfers in Argentina, she asserts that patronage networks help improve efficiency, but uses a multi-period model to stress that it is a dynamic, iterated exchange, with politicians employing a variety of techniques to encourage voters to stick to their commitments through time, even when voting in secret ballots. Following Stokes, this essay stresses the importance of political dynamics by considering how the situa-

162

[Evaluating the strength of commitments made to swing constituencies]

tion changes from one electoral period to the next. However, rather than assessing whether voters stick to their end of the bargain, it considers whether politicians can renege on their own, allowing them to reallocate benefits from the swing to the core. The analysis is based on a regional “tax”, more specifically, the environmental and social costs of a new airport runway. Since it is unpopular with local constituents, a commitment to block its construction can be thought of as a targeted group transfer of “welfare,” which is realised after the election. It therefore falls under the heading of “pork barrel” politics, as opposed to clientelism, which is associated with a more direct quid pro quo for an individual’s vote (Hicken, 2011). Taking the form of geographically concentrated benefits, studies of “pork” often focus on swing-constituencies rather than swing-voters (Golden & Min, 2013). For example, Vaishnav & Sircar (2012), tested whether school building in India was higher in constituencies with narrower political margins. This study is in a similar vein, but differs in the sense that school buildings are difficult to rescind after construction, whereas commitments not to build something are more vulnerable to periodic shifts in electoral incentives.

HEATHROW EXPANSION AND THE 2010 GENERAL ELECTION Speaking at a West London rally in 2009, David Cameron, the Leader of the Conservative Party, stated: “No ifs, no buts, there will be no third runway.” The statement cemented his party’s commitment to block the expansion of Heathrow, the U.K.’s largest airport, if he became prime minister after the 2010 general election. At the time, the incumbents, Labour (centre left) were facing competition from two main challengers: the Conservatives (centre-right) and the Liberal Democrats (centrist liberals). The challenging parties were both in strong positions, with Liberal

163

THE PUBLIC SPHERE | 2016 ISSUE

Democrats experiencing a considerable surge in popularity over the course of the 2000s, particularly in London. The arguments in favour of expanding Heathrow are largely economic and are in tune with the Conservatives’ core, pro-business ideology. The Labour Leader, Gordon Brown, supported the development, and many expected David Cameron to follow suit. It was a great shock therefore when he sided with the Liberal Democrats, and came out in opposition. It was not only a surprise to the party’s core voters; various commercial and aviation groups had been lobbying intensively and were confident of Conservative support (Hayden, 2014). To rationalise the strategy, Hayden (2014) argued that it formed part of the political absorption of environmental values, which was part of the “detoxification” of the Conservative brand, and aimed at bringing environmentally conscious voters into the Conservative fold. This appears to fit with a Downsian prediction that parties gradually adopt the views of pivotal voters. However, by suggesting that the strategy was designed to capitalize on a countywide wave of environmentalism, Hayden’s analysis is effectively limited to the national scale. By doing so, it fails to consider the policy’s local, distributional element, and therefore provides only a partial understanding of the multifaceted electoral incentives. While capturing the national environmentalist vote may have contributed to the decision, the policy was instrumental in local-level politics. Despite the economic benefits from expansion, the vast majority of constituents in Heathrow’s vicinity opposed it on the grounds of increased road congestion and aircraft noise pollution. Local opposition crystallised in particular around the latter, indeed, a 2008 survey in Richmond, which is one of the worse affected constituencies, found 89 per cent of residents were opposed to expansion due to the noise (Richmond Gov, 2008). As shown in Table 1, many of the worst affected boroughs were marginal constituencies, which provided an opportunity for political parties to garner local support by committing to block the move.

164

[Evaluating the strength of commitments made to swing constituencies]

As shown in Figure 1, the airport sits at the edge of West London, in the constituency of Hayes and Harlington. To the West, are the counties of Surrey, Berkshire and Buckinghamshire; historically the Conservative heartland. However, in the 2005 general election, the urban constituencies (labelled in Figure 1), elected MPs from Labour and the Liberal Democrats. The 2005 results are shown in Table 1, and illustrate the narrow margins in Ealing Central, Brentford & Isleworth, and Richmond. In line with Dixit & Londregan’s prediction, these narrow margins would alter the Conservatives’ political calculus. Despite being unpopular with the national, core support base, the decision to match the Liberal Democrats’ offer of blocking expansion, would give them a chance to retake pivotal voters, and shift the balance of power in West London. FIGURE 1 MAP OF LONDON SHOWING HEATHROW AIRPORT AND SURROUNDING CONSTITUENCIES

Source: Wikicommons, 2015.

165

THE PUBLIC SPHERE | 2016 ISSUE

Conservative

Liberal Democrat

Labour

7.3

30.7

23.0

39.0

Brentford & Isleworth

2.2

4.8

31.2

30.6

33.4

Ealing, Central

18.6

5.6

28.3

19.2

46.9

Ealing, North

38.4

5.6

18.9

18.2

57.3

Ealing, Southall

19.3

6.4

28.8

16.7

48.1

Feltham & Heston

28.7

6.1

27.7

9.8

56.4

Hayes & Harlington

7.1

4.5

39.6

46.7

9.2

Richmond Park

19.2

4.6

32.4

51.6

11.4

Twicken-

Table 1 2005 election results (%): London constituencies near Heathrow (Shaded: win. Bold: victory margin <10%)

Others 8.3

ham

Margin of victory over Conservatives

Data Source: BBC (2010)

166

[Evaluating the strength of commitments made to swing constituencies]

These marginal seats were already under the flight path, but their generally affluent constituents feared they would be amongst the worst affected by the third runway (particularly through reduced property values), and would be too far from the development to experience benefits such as labour market improvements. A recent map published by the Heathrow group in 2014, suggests that the constituentsâ&#x20AC;&#x2122; fears about noise were well founded, as they predicted that the third runway would exacerbate noise pollution in Ealing, Brentford and Richmond (which are circled in Figure 2). Cameronâ&#x20AC;&#x2122;s commitment capitalised on these concerns, consolidated support under the Heathrow flight path, and arguably tipped the balance. As shown in Table 2, which displays the 2010 election results, all three constituencies swung to the Conservatives. FIGURE 2 MAP SHOWING NOISE IMPACTS OF A THIRD RUNWAY AT HEATHROW, WITH EALING, BRENTFORD AND RICHMOND, CIRCLED

Source: Heathrow; Taking Britain Further, 2014.

167

THE PUBLIC SPHERE | 2016 ISSUE

Conservative

Liberal Democrat

Labour

Data

5.5

37.2

23.7

33.6

Brentford & Isleworth

4.3

38.0

27.6

30.1

Ealing, Central

5.5

30.9

13.2

50.4

Ealing, North

3.8

29.8

14.9

51.5

Ealing, Southall

8.7

34.0

13.7

43.6

Feltham & Heston

7.1

29.4

8.7

54.8

Hayes & Harlington

2.5

49.7

42.8

5.0

Richmond Park

3.8

34.1

54.4

7.7

Twickenham

Table 2 2010 election results (%): London constituencies near Heathrow (Shaded: win. Bold: victory margin <10%)

Others Source: BBC (2010)

168

[Evaluating the strength of commitments made to swing constituencies]

It is beyond the scope of this study to estimate how much of the vote gains can be attributed to Cameron’s commitment to blocking Heathrow expansion. However, it seems that the experiences of Ealing Central, Brentford & Isleworth and Richmond, are quite distinctive when compared to the rest of London. Of the 74 constituencies in Greater London, seven swung to the Conservatives in the 2010 election, suggesting that political allegiances were quite stable in the majority of London boroughs. The circle in figure 2. surrounds the block of the West London constituencies, and points to one of the main areas of change between the two periods, especially within the external ring of London’s more suburban boroughs. This suggests that there was something distinctive about the area’s interaction with the Conservatives. This is by no means proof that the airport bargain was the determining factor in the electoral outcomes (as it could still be attributed to other borough-specific characteristics that resonated with the Conservative message, such as the level of affluence of the constituents) however it is certainly consistent with the argument that the commitment was instrumental in political outcomes. FIGURE 3 ELECTORAL RESULTS IN 2005 AND 2010. EALING CENTRAL, BRENTFORD & ISLEWORTH, & RICHMOND CIRCLED

Source: Parliamentary commons, 2015.

169

THE PUBLIC SPHERE | 2016 ISSUE

LAYING THE FOUNDATIONS FOR A U-TURN

1. Subsequent to writing the paper, the report has been published, and argues that a third runway at Heathrow would be the most effective way to increase airport capacity.

170

In the wake of the 2010 election, the Conservatives formed a coalition government with the Liberal Democrats. Their joint commitment to blocking airport expansion was formalised in the coalition’s programme. Many assumed that David Cameron’s commitment would persist, but there has been mounting concern that the Conservatives have reassessed their position and are preparing to row back on their commitment. The party rhetoric has shifted from “no ifs, no buts” to a more pragmatic, pro-business stance. In 2012, the government established the Airports commission, which will provide an independent review of the case for the third runway. The results are due to be released at the beginning of the 2015 parliament.1 Most probably, if the Conservatives are in power, and the commission finds in favour of expanding Heathrow, David Cameron will argue that it is in national interests that they renege on their 2010 manifesto commitment. Such a U-turn would result in considerable reputation loss in the West London constituencies, but the swing voters in those areas may be less numerous than they were in 2010. If the Conservatives thought that the runway issue was still a critical component of their electoral strategy, they would probably have reiterated their 2010 commitment, but their 2015 manifesto is non-committal, simply stating that they will “respond to the Airports Commission’s final report” (Conservative Manifesto, 2015, p.14). From a strategic perspective, muddying of the waters around the Conservative commitment can be interpreted as a strong signal that they would be willing to revert to their core, pro-business position. In 2010, the commitment to block expansion helped to deliver a pivotal group of swing voters. A wholehearted adoption of Dixit & Londregan’s thesis would suggest that failing to renew that commitment would overturn those political gains. However, despite the considerable damage to the Conservatives’ reputa-

[Evaluating the strength of commitments made to swing constituencies]

tion, the political calculus may now lean towards a reversal of the commitment. The implicit calculation is to weigh the negative implications of the reputation damage, which could result in the loss of seats in West London, against the benefits of attracting voters who support Heathrow expansion, spread more evenly throughout the country. Two key elements, relating to the shifting political context between 2010 and 2015, have meant that the negative electoral implications, in that block of West London boroughs, would be dampened to a considerable extent. First, the compromises forced on the Liberal Democrats in the coalition (such as raising university tuition fees) have undermined their liberal-minded core supporters, and reduced the pressure on the Conservatives in affluent areas like Richmond. Second, there has been a shift in the relative popularity of Labour and the Conservatives. A flagship policy of Ed Milliband’s Labour opposition has been a progressive “mansion tax” on houses worth over £2 million. A study by Knight Frank, a London-based estate agent, highlighted that West London would be the most affected part of the country by the tax, and indeed found that Richmond would be one of the six most affected constituencies (Telegraph, 2015). Such changes give the Conservatives a greater degree of certainty that the swing voters gained in 2010 would not defect, even if the party failed to uphold its end of the airport bargain. If, leading up to the 2015 election, the Conservatives are more insulated from political competition in those West London seats, then the commitment would be unlikely to alter political outcomes in those boroughs. According to Downsian theory, the Conservatives may be encouraged to revert to a median-voter strategy. While Hayden (2014) might argue that prioritizing the environment has significant power to determine political outcomes, the discussion in the lead up to the 2015 election leaned heavily towards economic stewardship. Concerns that a Labour government would destabilise a fragile economic recovery have

171

THE PUBLIC SPHERE | 2016 ISSUE

provided conditions that encourage Cameron to reinforce the Conservatives’ reputation as the party that sets economic stability as its first priority. Thus, from the Conservatives’ perspective, the apparent erosion of the strategic importance of voters that are firmly against Heathrow expansion would allow the party to revert to their core position.

CONCLUSION The issue of airport expansion offers a useful lens through which to assess Dixit & Londregan’s swing voter theory. The analysis fits with their prediction of swing voter bias, but suggests a subtle shift in emphasis when thinking about long-term political outcomes. While swing voter theory certainly helps explain why transfers are established, the assumption of fixed voter preferences hampers our ability to assess their durability. With existing commitments, the question becomes whether swing constituencies are still strategically important enough to warrant a tax on the core. This has implications for understanding the evolution of distributive politics, as parties will constantly reevaluate the situation between electoral periods, assessing whether intervening political shifts allow them to revert to their core preference. In terms of the policy implications, the case is quite distinctive in the sense that supporting Heathrow expansion has become an issue of national political significance, as a touchstone of a party’s willingness to trade economic growth for environmental sustainability. At the same time however, local externalities mean that elections in the near vicinity are highly sensitive to a party’s political stance. These conflicting priorities are shared by other nationally significant projects that come up against local resistance. In the U.K., issues that would fit into this category include: the construction of new housing, the allocation of energy infrastructure (on-shore wind farms, new-generation nuclear power

172

[Evaluating the strength of commitments made to swing constituencies]

plants), unconventional oil exploration (fracking) and decisions about routes for new transport links. In these cases, the durability of political commitments that supposedly steer such projects away from swing constituencies may be limited, as the strategic importance of these groups of voters may well change.

REFERENCES BBC(2010) Data from:news.bbc.co.uk/1/shared/election2010/results/constituency/c04. stm Conservative Party Manifesto (2015) Available at:http://www.theguardian.com/politics/2015/apr /14/conservative-party-manifesto-2015-the-full-pdf Cox, Gary & Mathew McCubbins (1986) “Electoral Politics as a Redistributive Game.” The Journal of Politics 48(2):370–389. Dixit, Avinash & Londregan, L. (1996) “The Determinants of Success of Special Interests in Redistributive Politics.” The Journal of Politics 58(4):1132–1155. Golden, M & B. Min (2013) “Distributive politics around the world” Annual Review of Political Science, 16 (2013), 73–99 Hayden, Anders (2014) “Stopping Heathrow Airport Expansion (For Now): Lessons from a Victory for the Politics of Sufficiency” Journal of Env Policy & Planning, 16:4, 539-558 Hicken, A. (2011) “Clientelism” Annual Reviews of Political Science 14, 289–310 Heathrow; Taking Britain Further (2014) at: http://www.heathrowairport.com/about-us/ company- news-and-information/airports-commission/our-reports Jablonski, Ryan (2014) “How Aid Targets Votes: The Impact of Electoral Incentives on Foreign Aid” World Politics, 66(2), pp. 293-330 Milesi-Ferretti, G & Perotti R.(2002) “Electoral systems and the composition of public spending” Q. J. Econ. 117:609–57 Parliamentary Commons (2015) Map available at: http://en.wikipedia.org/wiki/List_of_ Parliamentary _constituencies_in_London#Results) Richmond Gov (2008): Available at: http://www.richmond.gov.uk/heathrow_expansion_history Stokes, Susan. (2005) “Perverse accountability: a formal model of machine politics with evidence from Argentina” American Political Science Review 99(3):315–25 Telegraph (2015), article available at: http://www.telegraph.co.uk/news/politics/labour/ 11433155 /Most-2m-homes-hit-by-mansion-tax-are-in-Home-Counties-and-Londonsuburbs.html Vaishnav, Milan & Sircar, N. (2012) “Core or Swing? The Role of Electoral Context in Shaping Pork Barrel” Department of political Science Colombia University available at: https://nsircar.files.wordpress.com/2013/02/vaishnav_sircar_03-12-12.pdf Wikicommons (2015): map available at:Wikicommons:http://commons.wikimedia. org/ wiki/File: Greater_London_UK_constituency_map_(blank).svg

173

Determinants of change in the international fight against corruption Do freedom of press and the executive composition matter? Vincent Zimmer

MSc in Public Policy and Administration, Class of 2014 London School of Economics and Political Science Managing Director and co-founder of Kiron Open Higher Education

Britta Schünemann

BSc in Psychology Class of 2015 Georg-August University of Göttingen

©Picsfive

ABSTRACT This paper uses a cross-country panel to examine the determinants of change of corruption levels, paying particular attention to freedom of press and executive composition (parliamentary or presidential). The results are analysed in the theoretical perspective of collective action theory, which defines corruption as an equilibrium. The empirical evidence confirms previous results on the role of freedom of the press and parliamentary systems in determining the prevalence of corruption. In addition, the paper finds evidence of a significant interaction effect. If change of corruption is defined as the dependent variable, only the variables ‘freedom of press’ and ‘share of Protestants’ remain significant. This indicates that the focus of previous studies on causes of corruption might have to shift toward determinants of change. Referring to collective action theory, this paper argues that freedom of the press is able to shift the collective action equilibrium toward less corruption because it increases accountability and is exogenous to the equilibrium. Factors regarded as exogenous determinants of change might be even more important than their potential to foster radical change as suggested in previous studies.

[Determinants of change in the international fight against corruption]

INTRODUCTION Corruption, often defined as the misuse of public office for private gain (Kolstad and Wiig, 2009: 521), has devastating effects on economies and their citizensâ&#x20AC;&#x2122; quality of life (Graf Lambsdorff, 2005). Various studies provide evidence that corruption reduces gross domestic product (GDP) growth, foreign investment and the efficiency of government expenditures (Graf Lambsdorff, 2005). The global annual costs of corruption are estimated to reach around US$ 1 trillion from bribery alone (Kaufmann, 2005). Thus, an increasing number of studies have been analysing possible causes of corruption over the last two decades to raise awareness and inform anti-corruption policies (Judge et al., 2011). This paper contributes to this strand of literature by testing executive composition (presidential or parliamentary) and freedom of the press as independent variables for correlation with change of corruption as the dependent variable. Previous studies do not shed light on the question of whether both variables are working in concert or against each other (Gerring and Thacker, 2004). This paper departs from the dominating framework of principal agent (PA) theory in previous studies, which often only addressed single causes of corruption like Kolstad and Wiig (2009) who focus entirely on transparency and free press without taking cultural or economics factory into account. Accordingly, papers focusing on single causes neglected explanations for the systematic failure of previous anti-corruption policies (Mungiu-Pippidi, 2013). In contrast, collective action (CA) theory explains this almost systematic failure by characterising corruption as an equilibrium where the existence of corruption is driven by various factors and leads to a strategic situation where no agent can choose a non-corrupt strategy without harming his own interest. Therefore, the corrupt system is maintaining itself in an equi-

177

THE PUBLIC SPHERE â&#x20AC;&#x2020;| â&#x20AC;&#x2020;2016 ISSUE

librium even if anti-corruption policies are addressing one of the several determinants (Persson et al., 2013) The CA model justifies the focus on executive composition and a free press as effective solutions to CA problems (Lederman et al., 2005). They are highlighted as pathways to a less corrupt equilibrium by increasing the accountability in the system (Lederman et al., 2005). By addressing previous methodological shortcomings and by making first attempts to analyse determinants of change rather than causes of corruption, this paper is opening new avenues for future research. The investigation is guided by the research question: Are executive composition and freedom of press determinants of change of corruption levels?

THEORY AND HYPOTHESES As the theoretical background and previous empirical evidence indicates, corruption is a regional and sector-specific phenomenon, with great degrees of variability within as well as across societies (Peisakhin and Pinto, 2010: 265). Six hypotheses are derived from the literature under the assumption that causes of corruption have a similar impact on change of corruption as on corruption itself. Figure 1 illustrates the research design, including all hypotheses and using dotted arrows to indicate the tested correlations and direction of those correlations based on the described underlying mechanism and previous findings. In order to allow measurement and comparison across countries, the following section defines independent, dependent and control variables precisely, discusses them from the theoretical perspective of the CA model and outlines testable hypotheses.

178

[Determinants of change in the international fight against corruption]

FIGURE 1 RESEARCH HYPOTHESES Independent variable: corruption

H1 control variables

parliamentarism

H6(I)

interaction effect

H6(II)

free press

Independent variable: change of corruption

Source: Author.

Dependent variables: corruption and change of corruption Despite the fact that corruption is commonly defined as misuse of public office for private gain (Kolstad and Wiig, 2009: 523), the concept is ambiguous (Rothstein, 2014). For example, the chosen definition does not specify the standard against which abuse of power is measured. Bribes, for example, could be engrained within social practices and become a de facto standard in a specific country like for example in Papua New Guinea where local authorities provide services for fees that are legally not implemented but still not considered as corrupt by citizens (Walton, 2014). This variation illustrates the difficulties for cross-country investigations to pin down a definition of corruption (Rothstein, 2014; Svensson, 2005). However, for the purpose of cross-country comparisons this paper defines corruption as â&#x20AC;&#x2DC;perception of

179

THE PUBLIC SPHERE | 2016 ISSUE

corruption’ which allows it to compare between different cultural settings and especially analyze the change over time in a certain country (Treisman, 2000). This definition adheres to the CA model since the observation that other agents are corrupt is the central mechanism that stabilises the political equilibrium: as long as corrupt agents observe and ‘interpret life in terms of corruption’, this ‘deeply held system of beliefs’ does not give incentives to single actors to change their behaviour (Rothstein 2011: 234). It is described as a belief because the agents cannot control all other agents at all times. However, as long as they believe that all other agents follow a corrupt strategy they will act accordingly. This models corruption as a self-reinforcing equilibrium (Persson et al., 2013). The empirical evidence suggests that the perception of the past level of corruption determines the current level (Dong et al., 2012), which indicates that perception also shapes the change of corruption levels. However, previous analyses did not differentiate between determinants that stabilise the equilibrium and determinants that act as drivers of change towards a lower (or higher) equilibrium. Therefore, corruption and change in corruption levels are both analysed as dependent variables in order to answer independent hypotheses. This focus on the CA model within the rational choice approach (RCA) relies upon an instrumental account of rationality. This allows clear and testable predictions for the independent variables, which is essential for the purpose of an empirical cross-country investigation.

Independent variable: executive composition Executive composition is understood as the degree of separation between president and parliament as well as their relative power. Parliamentarism is a governmental system in which the executive is elected and therefore responsible to the elective

180

[Determinants of change in the international fight against corruption]

body (Gerring and Thacker, 2004). This constellation ‘creates a single sovereignty at the national level’ (Gerring and Thacker, 2004: 305). Therefore, parliamentarism centralises political power, which reduces the number of potential veto points1 (or points of access). In contrast, sovereignty in presidentialism is distributed between the two directly elected bodies, the legislature and the president (Gerring and Thacker, 2004). This decentralises the power. Traditional advocates of public choice theory are more in favor of presidential systems (Asthana, 2012). They argue that corruption is decreased (Asthana, 2012) by the greater autonomy of and checks and balances between sub-jurisdictions (Fisman and Gatti, 2002). In contrast, parliamentary systems are more likely to increase the internalisation of externalities by centralising political power in the hands of national party leaders and central government bureaucrats (Bannikova, 2013). Therefore, parliamentary systems enable the legislature to monitor the executive more closely because its members can be removed from the executive office by parliament (Bannikova, 2013). This solves collective action problems by bringing the ‘net social costs and benefits of policy into the purview of the relevant political authority (Gerring and Thacker, 2004: 324). More recent empirical evidence seems to suggest however that presidentalism is associated with higher levels of corruption (Treisman, 2007), or has no effect at all (Pellegrini and Gerlagh, 2008), while parliamentarism is associated with lower levels of corruption (Lederman et al., 2005). However, these previous studies lack some control variables such as GDP or uninterrupted establishment of democracy identified in other studies (Serra, 2006). Their results have to be retested and the following hypotheses are derived from those studies.

1. ‘Veto points’ are representing the power of individuals and groups to influence governmental actions, even if they cannot blockade such actions (Tsebelis, 2002).

H1: Corruption is lower in countries with a parliamentary system than in countries with a presidential system.

181

THE PUBLIC SPHERE | 2016 ISSUE

H2: Negative change of levels of corruption is more likely to occur in countries with a parliamentary system than in countries with a presidential system.

Independent variable: free press The idea that greater transparency of the government´s activities decreases corruption is assumed by several scholars (Lindstedt and Naurin, 2005). More generally, Kaufmann et al. (2002) argue that ‘voice-related variables, including freedom of speech and freedom of press, have a large effect on corruption and the quality of services’ (Kaufmann et al., 2002: 3). Freedom of press is considered to be the most important operator for transparency (Kolstad and Wiig, 2009). A lack of transparency makes corruption less risky and more attractive for agents because they do not have to fear consequences as the problems remain hidden from the public (Kolstad and Wiig, 2009). Therefore, it becomes harder to use incentives to make public officials act cleanly. Taking this evidence into account, the research question is further defined H3: The level of corruption is lower in countries with higher levels of freedom of press than in countries with lower levels. 2. In order to avoid an overfitting, not all possible control variables suggested by the literature were incorporated into the estimation model (Wermuth, 1992). Simultaneously, it was important to avoid problems of omitted variable bias (OVB) (incorrectly leaving out causal factors) (Kaufmann et al., 2002).

182

H4: Negative change of levels of corruption is more likely to occur in countries with higher levels of freedom of press.

Control variables The selection of control variables was guided by previous literature and the criteria of robustness used by Serra (2006) and Treisman (2007).2 Robustness is defined as follows: the estimated coefficient of an explanatory variable that remains statistically significant for the 1 per cent threshold under different sets of

[Determinants of change in the international fight against corruption]

control factors (Serra 2006). This paper focuses on the robust factors identified by Serra (2006) and Treisman (2007) as they represent the most complete set of variable in the literature so far and reduce the danger of OVB. These variables are economic development measured via GDP, democratic development, legal culture measured via a British colony dummy and share of Protestants. It is important to note that this paperâ&#x20AC;&#x2122;s chosen focus does not presume to â&#x20AC;&#x2DC;claim the irrelevance of all the other explanatory variablesâ&#x20AC;&#x2122; (Serra, 2006: 250). Therefore, OVB might still be a problem within this study even if empirical robustness is the best criteria available and the presented set of variables complete according to the status quo in the literature.

GDP The empirical evidence exposes a negative correlation between corruption and GDP per capita (Serra, 2006), but many scholars acknowledge that causality remains unclear as high levels of corruption, for example, decrease economic growth (Mauro, 1995). The literature suggests that a long exposure (at least 30 years) to uninterrupted democracy undermines the foundations of corruption (Serra, 2006). This is explained through the combination of electoral competition and institutional checks and balances on executive power, which increase the costs of corruption for corrupt agents. They often fear being exposed and punished as corrupt, according to experimental findings (Barr et al., 2009).

Former British colonies Treisman (2000) argues that the civil service of former British colonies has a strong focus on procedural aspects of law. This results in judges who follow procedures also against state officials, which may lead to a lower level of corruption as corrupt agents have to fear a higher risk of exposure (Treisman, 2000).

183

THE PUBLIC SPHERE | 2016 ISSUE

Socio-cultural factors Among socio-cultural factors, the presence of Protestants in the population is associated with lower corruption in comparison to more state-religions like Catholicism or Islam that are officially recognized and supported as the official religion of the nation by the government (Pellegrini and Gerlagh, 2008) because of deliberative and less hierarchical structures within the Protestant church (Armantier and Boly, 2011). The analysis with ‘change of corruption levels’ as a dependent variable may reveal to what extent causes of corruption also function as determinants of change. The evidence is used to test the following hypothesis: H5: Root factors of corruption are significantly correlated with factors determining changes in corruption levels.

Interaction effect In addition, the investigation has to check for an interaction effect between the independent variables. The central argument for such an effect is that both factors increase the degree of accountability in the system. Accountability is defined as the degree of competition in the political system, the existence of checks and balances mechanisms across different branches of government and the transparency of the system (Lederman et al., 2005). Furthermore, previous studies argue that the effect of free press is not unconditional. Lindstedt and Naurin (2005) provide empirical evidence that the effect is dependent on fair elections and democracy, which are highly correlated with executive composition (Gerring and Thacker, 2004). They argue that only the connected impact of these factors is effective in reducing corruption (Lindstedt and Naurin, 2010). In addition, empirical evidence provided by Lessmann and Markwardt

184

[Determinants of change in the international fight against corruption]

(2010) as well as Treisman (2007) suggests that freedom of press is an intermediating factor for separation of power to decrease corruption: separation of power does not have an effect on its own, only when interacted with freedom of the press. This adheres to the theoretical definition of executive composition. Thus, the following hypothesis is derived: H6: There is a positive interaction effect between freedom of press and executive composition which entails (i) corruption reduction and (ii) negative changes of corruption levels.

METHODOLOGY AND DATA This paper uses a cross-country regression analysis to investigate the determinants of changes in corruption using panel data. Previous cross-national studies have often been limited to 80 or fewer country cases (Treisman, 2007). Such partial coverage may cause problems of sample bias, leverage (influential cases that screw the whole analysis in a certain direction as for example the case of Hong Kong is a huge success story but also historically not compariable to other Asian countries), and misspecification (Bollen and Jackman, 1985). Furthermore, many studies have not included the full range of control variables identified in the literature (Serra, 2006). Consequently, this paper includes a wider country set of 130 countries. Moreover, panel data regressions are conducted to ensure that the determinants of corruption as suggested by Serra (2006) are indeed associated with corruption or change in corruption. The results are presented in Tables 7 and 8 (see Appendix B). Variables that were statistically significant at the 10 per cent threshold were included in the model. Following suggestions from the empirical literature, this paper uses the block-recursive linear regression method (Wermuth, 1992)3, which ensures that there

3. The method allows testing different types of non-decomposable independent hypotheses and does not face problems of overparameterisation (Wermuth, 1992). Each parameter is a regression coefficient the interpretation directly derives from the model graphs, which simplifies the interpretation of the results. The crosssectional method has also the advantage of avoiding a serial correlation of the residuals and is not dependent on the assumption that the relation between the variables is stable over time. These characteristics make the method superior to alternative approaches as covariance selection or path analysis (Wermuth, 1992).

185

THE PUBLIC SPHERE | 2016 ISSUE

are no correlation between residuals and does not rely on assumption that the relation between the variables is stable over time. This makes the methodology to a fixed-effect regression as ithas the advantage of avoiding various complicating aspects of the use of data drawn from various points in time, such as serial correlation of residuals. It also has the advantage that the data analysis does not rely on the assumption that the nature of the relationships between variables is stable over time. This assumption does not adhere to the analysis of corruption levels because the nature of the determinants is changing over time. For example, the variable executive composition differentiate only between presidential and parliamentary system but the in reality this variable has to be understood as a continuum consisting also of many semi-presidential and semi-parliamentary system. If the paper would assume fixed effects, it would impose time independent effects for each entity that are possibly correlated with the regressors. Additionally, this paper analyses the effect power of the two independent variables ‘executive composition’ and ‘freedom of press’ following Cohen (1992) and explores the interaction effects between the two variables (Brambor, 2005). This analysis spans the years 2003 to 2013. Corruption is measured via Transparency International’s (TI) annual index of perceived corruption (Transparency International, 2014). The index represents a ‘poll of polls’ from a number of individual surveys in each country (Treisman, 2000). It is converted to a scala from 0 to 10, so that higher values indicate a higher level of corruption. It has the biggest database among all available indices (Donchev and Ujhelyi, 2014). Following Golder (2005), the executive composition on governance is coded with a three-part coding scheme: 1, presidential; 2, semi-presidential; 3, parliamentary. In accordance with the literature (Kolstad and Wiig, 2009), an index of freedom of information produced by Freedom House is used as a proxy for

186

[Determinants of change in the international fight against corruption]

free press or transparency4. The further selection of control variables and the sources of data are derived from Serra (2006) and are briefly described in the following paragraph. Drawing on the World Bank, economic development is measured as the logarithm of GDP per capita5. The degree of democracy is measured through a dummy defining the degree of basic political rights as constructed by Freedom House. For the legal culture variable, a British colonial heritage dummy (‘British colony’) is used with data from Treisman (2000). The dummy is constructed based on historical evidence whether the country was ever a British colony or not. Data for ‘Protestants’ is obtained from Porta et al. (1997) that indicates the per centage of the population belonging to the Protestant religion in 1980 and their calculations are repeated with new data obtained from the CIA World Factbook – Religions (Central Intelligence Agency, 2010).

4. https:// freedomhouse. org/report-types/ freedom-world

5. http://data. worldbank.org/

RESULTS The descriptive statistics of all the variables are presented and described in detail in Table 6 in Appendix A. Independent analysis for the starting and end phase of the analysed time period indicates that overall the level of corruption seems to be almost constant across countries as the mean level of corruption only changed slightly from 4.46 to 4.17 (SD = 2.3). The descriptive statistic exposes no outliers, but countries with no experience as a democracy by 2003 are excluded from the analysis. In a non-democratic system, the freedom of speech is not guaranteed. It is therefore unclear if the observed data on, for example, the perception of corruption is reliable since the data is based on polls. Additionally, the analysis has filtered out countries with a semi-presidential system when analysing the executive composition, as the theoretical lens is focused on the

187

THE PUBLIC SPHERE | 2016 ISSUE

6. The analysis uses the generalised formula for panel data by Bhargava (1982) resulting in a value of 1.647 and a corresponding p-value of 0.141. 7. VIF provides an index based on a ordinary least squares (OLS) regression that measures how the variance of the estimated regression coefficient is increased because of collinearity. No high values (>7.5) are resulting from the analysis presented in Tables 1 and 2, which indicates that no problematic multicollinearity occurs.

188

difference between parliamentary and presidential systems. All restrictions are in alliance with previous studies (Gerring and Thacker, 2004) and reduce the number of countries from 130 down to 123 for the analysis of freedom of press and 112 for the analysis of the executive composition. Autocorrelation, which is a common problem of cross-sectional analyses, is checked via the Durbin Watson test6. Results show no strong autocorrelation and therefore are not considered a problem for the overall analysis. Multicollinearity was checked in the final models using the variance inflation factor (VIF) analysis to assess its robustness7. Results show no multicollinearity problems.

Relationship between free press and the executive composition on corruption Table 1 presents the results for the multivariate regression analysis, putting ‘perceived corruption in 2013’ as the response variable. All factors are added one by one in order to indicate the additional impact of each factor, creating six different models that are all significant at the one per cent threshold (P<F). When fitting the best multivariate regression models, being a former British colony is dismissed from the analysis, as it does not reach the significance threshold. In the multivariate model 5 for perceived corruption in 2013, after controlling for all the other variables, the coefficient for freedom of press means that an increase of freedom of press by 1 index point would is associated with a perceived corruption of 0.027 index points. This is equivalent to a decrease of perceived corruption by 0.27 per cent on average, ceteris paribus at a global level. The coefficient for the governmental system shows that a parliamentary system is associated lower perceived corruption by -0.317 index points than a presidential country. Therefore, a change from a presidential

[Determinants of change in the international fight against corruption]

to a parliamentary system is associated with a decrease in perceived corruption by 3.17 per cent on average ceteris paribus at a global level. Table 1 Multivariate models for corruption, 2013 (Model 1) Dependent variable: perception of corruption 1

Ln GDP

0.001*** 0.001*** 0.001*** 0.001*** 0.001*** 0.001***

-1.040

-0.664

-0.464

-0.426

-0.393

-0.438

Standard error

0.087

0.095

0.104

0.107

0.103

0.113

VIF

1,63

Free press

0.001*** 0.001*** 0.001*** 0.001*** 0.001***

-0.040

-0.035

-0.031

-0.027

Standard error

0.006

VIF

2,15

Democracy

0.001*** 0.001*** 0.001*** 0.013**

-0.019

-0.017

-0.016

-0.013

Standard error

0.005

VIF

1,78

Executive composition

0.018**

0.016**

0.03**

-.331

-0.317

-0.310 continue

189

THE PUBLIC SPHERE | 2016 ISSUE

Table 1 Multivariate models for corruption, 2013 (Model 1) (continuation) Dependent variable: perception of corruption 1

Standard error

0.138

0.133

0.142

VIF

1,2

Protestant

0.005*** 0.011**

-0.014

Standard error

0.005

VIF

1,43

British colony

0.691

-0.108

Standard error

0.270

VIF

1,07

Constant

14.376

13.482

12.056

11.986

11.826

11.926

Standard error

0.732

0.634

0.697

0.694

0.671

0.740

Adj. R²

0.603

0.717

0.744

0.756

0.773

0.767

Change R²

0.114

0.027

0.012

0.017

-0.006

149.726

124.862

95.750

76.814

67.790

54.364

P<F

0.000

Source: Author’s calculations. Note: For all tables, defined significance levels are: *P<0.10; **P<0.05; ***P<0.01.

Relationship between free press and executive composition on change of corruption Table 2 presents the results of the multivariate regression analysis, putting change of perceived corruption from 2003 to 2013 as the response variable. The six different models are all significant at the five per cent threshold (P<F).

190

[Determinants of change in the international fight against corruption]

In the multivariate Model 4, after controlling for all the other variables, the coefficient for freedom of press indicate that an increase of freedom of press by one index point is associated with a decrease in perceived change of corruption by 0.017, which equals a decrease of perceived change of corruption by 0.17 per cent on average, ceteris paribus at a global level. After controlling for all the other variables, the coefficient for the executive composition is not significant. Table 2 Multivariate models for change of corruption (Model 2) Dependent variable: Change of perception of corruption between 2003 and 2013 1

Free press

0.013**

0.063*

.156

0.032**

0.019**

0.027**

-0.014

-0.013

-0.010

-0.017

-0.019

-0.018

Standard error

0.006

0.007

0.008

VIF

1,85

Executive composition

0.780

0.837

0.576

0.397

0.651

-0.045

-0.033

-0.089

-0.142

-0.080

Standard error

0.161

0.159

0.167

0.176

VIF

1,34

Protestant

0.278

0.071*

0.068*

0.147

-0.007

-0.012

-0.010

Standard error

0.006

0.007

VIF

1,17

Democracy

0.019**

0.106

0.074*

0.013

0.010

0.011

Standard error

0.005

0.006

VIF

1,56 continue

191

THE PUBLIC SPHERE | 2016 ISSUE

Table 2 Multivariate models for change of corruption (Model 2) (continuation) Dependent variable: Change of perception of corruption between 2003 and 2013 1

Ln GDP

0.300

0.611

0.134

0.072

Standard error

0.129

0.140

VIF British colony

0.259

-0.380

Standard error

0.335

VIF Constant

0.338

0.356

0.447

0.343

-0.453

-0.073

Standard error

0.341

0.348

0.355

0.348

0.840

0.919

Adj. R²

0.051

0.042

0.044

0.089

0.090

0.087

-0.009

0.002

0.045

0.001

-0.002

Change R² F

6.299

3.159

2.507

3.393

2.934

2.548

P<F

0.014

0.047

0.064

0.012

0.017

0.025

Source: Author’s calculations.

Effect power and interaction effect In order to allow a more detailed interpretation and understanding of the regression results, the relation between the executive composition and freedom of press will be further analysed in terms of an interaction effect and the effect power of R². Thise paper will refer back to the effect power as used by Cohen (1992), which compares the difference between the average of two groups with equal numbers (n) and equal variance. Cohen’s effect power is ranges from 0 to 1. The results are presented in Table 3.

192

[Determinants of change in the international fight against corruption]

Table 3 Effect power R² difference to 0

Effect power of R²

0-<0.0196

No effect power

0.0196-<0.1300

Weak effect power

0.1300-<0.2600

Medium effect power

0.2600-<1.00

Strong effect power

F, E

F F

F, E

Source: Cohen (1992).

After adding the control variables and conducting a multivariate analysis, freedom of press only had a medium effect power on perception of corruption and a weak effect power on change of corruption. The executive composition had no effect power in both models. In order to test for an interaction effect between freedom of press and executive composition, the estimation for Model 5 in M1 and Model 4 in M2 is replicated after standardisation8 of freedom of press in order to reduce problems of multicollinearity and to simplify interpretation.

8. Standardised variables are centred around zero and have a standard deviation of one.

Table 4 Interaction effect U1

Ln GDP

0.001***

0.487

-0.468

0.093

Standard error

0.105

0.133

Free press

0.001***

0.018

-0.044

-0.033

Standard error

0.011

0.014

Democracy

0.001***

0.119

-0.013

0.010

Standard error

0.005

0.006

Executive composition

0.001***

0.284 continue

193

THE PUBLIC SPHERE | 2016 ISSUE

Table 4 Interaction effect (continuation) U1

-0.374

-0.183

Standard error

0.134

0.169

Protestant

0.001***

0.034

-0.017

-0.014

Standard error

0.005

0.007

Interaction effect

0.001***

0.071*

0.077*

0.208

1.000

-0.333

1.000

-1.984

Standard error

.088

0.183

.560

1.565

Constant

14.376

15.127

-2.220E-16

-0.083

Standard error

0.732

1.929

0.286

0.886

Adj. R²

0.568

0.779

0.022

0.096

130.068

58.454

3.188

2.729

P<F

0.000

0.077

0.017

Source: Author’s calculations.

The results for the interaction effect are presented in Table 4 and indicate that the interaction is significant at the ten per cent threshold if ‘perception of corruption in 2013’ is the dependent variable. The effect disappears for ‘change of perceived corruption’ if all control variables are added to the regression. In both cases, the interaction effect is positive, which indicates that the variables are more likely to act in concert together rather than being rival antecedents. Interestingly, 32.8 per cent of the variation for parliamentary system is explained by the interaction effect.

Linking evidence and hypotheses The investigation provided mixed empirical evidence in relation to the hypothesis: no clear conclusion seems to apply.

194

[Determinants of change in the international fight against corruption]

The results are illustrated in Figure 2. Confirmed hypotheses are marked with green and hypothesis not supported by evidence are marked with red. If the evidence leads to unclear conclusions, the hypothesis is in white. Hypotheses H1 and H3 are confirmed, as the analysis exposes that countries with free press and parliamentary government countries are associated with lower levels of corruption. This supports previous findings of Gerring and Thacker (2004) as well as Peisakhin and Pinto (2010). FIGURE 2 LINKING EVIDENCE AND HYPOTHESES Independent variable: corruption

H1 control variables

parliamentarism

H6(I)

interaction effect

H6(II)

free press

Independent variable: change of corruption

Source: Author.

If a free press exposes corrupt actions in a presidential system that tends toward corruption, it could motivate other agents to become corrupt as well according to the RCA. However, the exposed positive interaction effect reveals that freedom of press has a greater impact in reducing corruption in presidential systems. One reason for this might be the fact that in a parliamentary system the opposition functions as a con-

195

THE PUBLIC SPHERE â&#x20AC;&#x2020;| â&#x20AC;&#x2020;2016 ISSUE

trol institution for governmental actions, whereas the free press fulfills this role in a presidential system, for example in China. This explanation seems to be more plausible than the theoretical argument provided by RCAs. Freedom of press and the executive composition are moving in the same direction and are correlated together with a change in corruption. Yet an analysis of the effect power exposes that only freedom of press has a moderate effect power. In contrast, the executive composition stays significant, but does not effectively shape the perception of corruption. Therefore, freedom of press seems to be more relevant than executive composition in decreasing corruption. H2 is not confirmed as executive composition is not correlated with the change of corruption levels. In contrast, freedom of press has a significant impact on change of corruption at the five per cent level, which confirms H4. Despite the significant effect on corruption, the interaction effect between freedom of press and the executive composition disappears for change of corruption levels, once the control variables are added. Therefore, the second part of H6 is not confirmed. The analysis provided no clear results for H5 as the chosen causes, freedom of press and share of Protestants, had a significant impact on change of corruption but not the other variables.

Discussion and limitations While the discussed empirical evidence seems to offer clear evidence in favor of freedom of press as a determinant of change in corruption levels, several limitations regarding the theoretical perspective of CA as well as the empirical methodology have to be taken into account. Therefore all findings presented must be interpreted with caution. This section highlights possible explanation for the causal mechanism behind changes in corruption and discusses limitations of the proposed analysis.

196

[Determinants of change in the international fight against corruption]

Endogenous versus exogenous factors In the context of the CA model, it might be plausible that parliamentarism alone is not associated with change of corruption, as parliamentarism to solve the collective action problem will not exist if the elected officials are part of the corrupt equilibrium. For example, they may block reform initiatives to ensure their personal benefits or legalise their traditional system of patronage (Rothstein 2011). Corrupt systems might also be disrupted by an exogenous factor like changes in the degree of free press. Free press functions as a check and balance to the government if the parliamentary opposition is not able to do so. This indicates that this underlying mechanism to parliamentarism might not work if the corrupt political equilibrium also includes the free press or if free press lacks the credibility or motivation to fight against corruption (Besley and Prat, 2006). On the other hand, in a globalised world, international newspapers or the internet can also satisfy the role of free press if the national media is not effective at embracing this function and could help to decrease corruption (Lio et al., 2011). Therefore, a free press cannot be solely treated as national variable: it cannot be truly endogenous to a national corrupt equilibrium. This might explain why freedom of press is a better determinant of change in corruption levels as compared to other historical, cultural or institutional factor bound to the national level and which are likely to be an endogenous factor in the corrupt equilibrium (Kolstad and Wiig, 2009). Generalising these results indicates that the premises of the CA model might need to be adjusted (Pieth, 2014) as scholars have mostly been advocating for a radical change towards more accountability to overcome a corrupt political equilibrium (Rothstein, 2011). However, less attention was given to the question of whether this change can be triggered by endogenous factors or if it has to be an exogenous shock (Persson et al., 2013). The presented findings in this paper indicate that it depends on the

197

THE PUBLIC SPHERE â&#x20AC;&#x2020;| â&#x20AC;&#x2020;2016 ISSUE

initial spread of corruption within the system. Yet it seems likely that only exogenous factors like international press can facilitate a break away from the corrupt equilibrium (Peisakhin and Pinto, 2010). This argument would support institutional and policy reforms that would focus on determinants of change rather than on causes of corruption.

Challenging the collective action model assumptions Nevertheless, the theoretical perspective of the CA model is limited to its assumptions. It has, for example, a shortcoming when explaining the formation of this corrupt equilibrium. Therefore, this paper is limited to finding ways of changing the corrupt equilibrium rather than explaining why it is there in the first place (Treisman, 2007). This approach often underestimates cultural differences between countries (Collier, 2002). In Asia, for example, bribes are often perceived as presents (Barr and Serra, 2010) and there is a different Asian press model (Winfield et al., 2000). Hallin and Mancini (2011) provide a framework with a compilation of media systems beyond the Western world that differentiates between Western media systems as opposed to communist, authoritarian, revolutionary and developmental media systems around the world. The function of a free press differs among these systems, but it was not in the scope of this paper to explore these culture differences in detail. However, the purpose of an empirical cross-country investigation justified the adoption of the CA perspective, as it sheds light on the basic characteristics that indicate corruption as a persistent problem in many countries around the world (Persson et al., 2013).

Limitations From a methodological point of view, several limitations apply. First, there is a concern about measurement error. Corruption

198

[Determinants of change in the international fight against corruption]

is measured as an index of perception, but in a field experiment in Indonesia, Olken (2009) exposes that although villagers report corruption correctly, the most significant cases are successfully hidden by officials. This indicates potential large differences between real corruption and perception. Furthermore, the perception of corruption is strongly influenced by current discourses in the media (Markovits and Silverstein, 1988). In more advanced democracies, corruption tends to be overestimated in by public and press. By contrast, corruption might be underestimated in corrupt regimes by virtue of the absence of free public opinion and press. Moreover, perception of corruption is often shaped by the number of corruption events reported in the media rather than the relative level. This might lead to an overestimation of corruption in large countries (Donchev and Ujhelyi, 2014). An experiment from Falk and Fischbacher (2002) found that a high perception of corruption could cause future corruption. However, in their meta study, Judge et al. (2011) evaluated around 100 previous empirical papers and find a strong significant correlation between all corruption measures. Therefore, no available index seems to be superior among the available measurements and â&#x20AC;&#x2DC;perception of corruptionâ&#x20AC;&#x2122; provides the largest data base. Overall, the regression analyses in this paper cannot test causality but only highlight correlations. Future research should include additional quantitative and qualitative method to enhance the understanding of the causal mechanisms behind changes in corruption, in particular to address measurement errors highlighted previously as well as OVB related problems For example, is the analyses biased by the fact that only determinants with available data are considered. Therefore, future research might identify relevant additional variables not considered here via explorative and qualitative methods. Another reason to assume that OVB might be problematic is the fact that all variables are highly correlated and it is therefore, not possible to identify the specific impact of each determinant. The GDP for example is a

199

THE PUBLIC SPHERE â&#x20AC;&#x2020;| â&#x20AC;&#x2020;2016 ISSUE

rather broad indicator as it gives very limited information on the direction of causality regarding corruption. ItÂ´s also not including information on how the resources are distributed among the society which on the other side might help to explain corrupt behavior. Future research should use experimental methods to analyze the impact of single factors to ensure the used set of variables is complete and to provide evidence on the direction of causality.

CONCLUSION This paper analysed how executive compositions and freedom of press affected change of corruption levels between 2003 and 2013 and explores the interaction between both variables. The selection of these factors is justified by the CA model as they can change the incentives for all agents, increase accountability and be influenced more easily by policymakers as compared to cultural, historical or economic antecedents. Parliamentarism is associated with a centralisation of power, resulting in fewer veto points and allow a closer monitoring by the central government. A free press works as a check and balance mechanism for the government. The empirical results are mixed: both independent variables were found to be associated with lower levels of corruption. Additionally, the investigation finds a positive interaction effect, indicating that both variables work in the same direction. Interestingly, this interaction effect as well as the correlation with composition of the government disappears if change of corruption is analysed as the dependent variable as suggested by the CA model. The empirical model only indicates a significant correlation between freedom of press and change of corruption levels. Some theoretical limitations and methodological problems cannot be ruled out completely. However, the applied method provides more robust empirical evidence than previous studies as the criteria for robustness was checked for more variables and applied.

200

[Determinants of change in the international fight against corruption]

The results can be explained by the ability of a free press to overcome a corrupt equilibrium if it is not endogenous to and therefore, part of the corrupt system. In contrast with the executive composition, a free press is not limited to the national level and can affect corruption via the global media and the internet. This adheres to a new avenue of research (Persson et al., 2013), which explains the failure of previous anti-corruption reforms through the focus on causes of corruption. Corruption, analysed on the macro level, seems to be a rather technical or moral issue, but the large associated costs indicate that it is a major challenge to policymakers all over the world. Therefore, the fight against corruption should continually receive high priority and attention by scholars as well as policymakers. Future research should explore the underlying mechanism of free press in detail and should search for other endogenous determinants of change. This paper recommends that policymakers primarily support a free global press to achieve a less corrupt equilibrium, as it is less likely to be endogenous to the current equilibrium.

APPENDIX A The results of the descriptive analysis are presented in Table 6. On average, between 2003 and 2013 the highest score for freedom of press was 90 in Denmark and the lowest was 10 in Uzbekistan. The world average was 56.94. The mean for executive composition indicates that countries in the sample tend to be parliamentary rather than presidential. Regarding corruption, the country with the lowestindex in 2003 was Finland and in 2013 it was Denmark; whereas the country with the highest level of corruption was Bangladesh and Sudan in 2013. However, the change between 2003 and 2013 of 0.29 is rather small but the minimal value of -6.1 and the maximal value of 5.8 indicates that in some countries radical change has taken place in the level of corruption.

201

THE PUBLIC SPHERE | 2016 ISSUE

Table 6 Descriptive statistics Variables

Mean

Standard deviation

Standard error

Min

Max

Free press average, 2003 and 2013

130

43.06

22.51

1.98

10.00

94.00

Executive compensation

130

1.84

0.95

0.08

1.00

3.00

Corruption perception, 1997

130

4.46

1.34

0.12

1.34

6.46

Corruption perception, 2003

130

4.25

2.20

0.19

0.20

9.10

Corruption perception, 2013

130

4.17

2.30

0.20

1.30

9.70

Change of corruption between 2003 and 2013

130

0.29

1.09

0.10

-6.10

5.80

GDP per capita average between 2003 and 2013, US$

130

8.46

1.56

0.14

5.55

11.46

Democracy since 1900 until 2003

130

33.55

30.37

2.66

0.00

103.00

Political rights, average 2003 and 2013

130

2.86

1.86

0.16

1.00

7.00

Protestant population

130

12.80

23.26

2.04

0.00

98.00

British colony

130

0.28

0.45

0.04

0.00

1.00

Source: Author’s calculations using data obtained from the World Bank, CIA World Factbook and Transparency International 2014.

The GDP per capita ranges from Malawi, with US$ 5.55 thousand per capita, and Luxembourg, with US$ 11.86 thousand per capita. Sixty-one countries have a Protestant population, within which there is an enormous range between 1 and 98 per cent. Thirty-seven countries have a former history as a British colony, which is above the suggested n<30 line. Some countries such Switzerland or the United States have been a democracy for more

202

[Determinants of change in the international fight against corruption]

than 103 years, whereas other countries only became democratic recently, such as Algeria in 1995. However, in a majority of states the political rights of the citizens are respected as the mean is 2.86 over the 5.00 threshold, which some studies set for consider the country as truly democratic (Gerring and Thacker, 2004).

APPENDIX B The univariate analysis was conducted for every year between 2003 and 2013 to generate more measurement points. However, as no contradicting results occurred, Table 7 presents only the results for ‘setting perception of corruption in 2013’ as the response variable. Table 7 Univariate corruption 2013 (U1) Variables

Sign

R²

Adj. R²

Executive composition

-1.350

0.167

-0.634

-8.078

0.001***

0.402

0.396

112.00

British colony -0.410

0.446

-0.085

-0.918

0.361

0.007

-0.001

123.00

Democracy

-0.049

0.004

-0.714

-10.981

0.001***

0.510

0.505

123.00

Free press

-0.070

0.006

-0.755

-12.414

0.001***

0.571

0.567

123.00

GDP

-1.040

0.070

-0.779

-13.390

0.001***

0.607

0.604

123.00

Protestant

-0.049

0.070

-0.550

-7.084

0.001***

0.302

0.296

123.00

Source: Author’s calculations.

Apart from being a former British colony, the two independent variables free press and the executive composition as well as all included control variables reach the one per cent threshold. The coefficient for the executive composition would mean that a parliamentary system faces a -1.350 lower level of perceived corruption compared to a presidential system. It is therefore expected that parliamentary ruled countries have 13.5 per cent less perceived corrup-

203

THE PUBLIC SPHERE | 2016 ISSUE

tion in the observed country set. For a democracy, this would mean that for every additional year of being a democracy the perception of corruption dropped by -0.049 index points, which means that every additional year of democracy is expected to reduce corruption by 0.49 per cent on average, ceteris paribus on the global level. In addition, for each additional index point for freedom of press, the perception of corruption dropped 0.070 points, which means that it is expected to reduce perceived corruption on the global level by 0.7 per cent. An increase of GDP per capita of one index point is associated with an drop of corruption by -1.040, which means that is expected to reduce perceived corruption by 10,4 per cent on average, ceteris paribus on the global level. For the share of Protestants in the population, an increase of one per cent would indicate a drop of -0.049 of the perceived corruption, which means a reduction of 0.49 per cent on average, ceteris paribus on the global level. As a result, it can be stated that the robust variables are able to explain between 29.6 per cent and 60.4 per cent of the regression, whereas the variable Protestant explains the lowest and GDP the highest share of the variation in the data set. The results of the univariate analysis presented in Table 8 put the change of perception of corruption between 2003 and 2013 as the response variable. Table 8 Univariate model for change of corruption (U2) Variables

Sign

R²

Adj. R²

Executive composition

-0.209

0.125

-0.153

-1.668

0.098*

0.027

0.017

112.00

British colony

-0.462

0.264

-0.161

-1.746

0.084*

0.026

0.017

123.00

Democracy

0.002

0.004

0.049

0.532

0.595

0.002

-0.006

123.00

Free press

-0.010

0.005

-0.183

-1.999

0.048** 0.033

0.025

123.00

GDP

0.010

0.074

.013

0.142

0.887

0.000

-0.008

123.00

Protestant

-0.007

0.005

-0.131

-1.420

0.158

0.017

0.009

123.00

Source: Author’s calculations.

204

[Determinants of change in the international fight against corruption]

The executive composition and being a former British colony reaches the ten per cent threshold whereas free press is significant on the five per cent level. All other control variables fail in the univariate analysis to reach at least the ten per cent threshold. The coefficient ß indicates that the executive composition parliamentary systems foster a more negative change compared to presidential systems, which is estimated to be -0.209 index points. This indicates that parliamentary systems reduce corruption 2.09 per cent more than presidential systems on the global level. The coefficient for British colonies indicates that the change is -0.462 index points as compared to other countries, which means that former British colonies reduce corruption by 4.62 per cent more than other countries. One index point increase for freedom of press results in a -0.010 lower increase of corruption, which means that it is expected to reduce perceived corruption on the global level by 0.01 per cent. It is notable that the R² is much lower than for the univariate analysis on perceived corruption as the lowest level is reached with 0.017 by the executive composition and being a former British colony, which indicates that a model based on this factors would only make 1.7 per cent fewer mistakes in order to estimate the change of perceived corruption. However, the highest value is reached by freedom of press with 0.025, which indicates that even the most significant factor could only help to reduce errors by 2.5 per cent if used for an estimation of change of perceived corruption.

REFERENCES Armantier, Olivier and Amadou Boly. 2011. “A controlled field experiment on corruption.” European Economic Review 55 (8):1072–82. Asthana, Anand. 2012. “Decentralisation and Corruption Revisited: Evidence from a natural Experiment.” Public Administration and Development 32 (1):27–37. Bannikova, Marina. 2013. Corruption in representative democracy. Documents de treball del Departament d’Economia;2013-17. Universitat Rovira i Virgili. Departament d’Economia.

205

THE PUBLIC SPHERE | 2016 ISSUE

Barr, Abigail, Magnus Lindelow, and Pieter Serneels. 2009. “Corruption in public service delivery: An experimental analysis.” Journal of Economic Behavior & Organization 72 (1):225–39. Barr, Abigail and Danila Serra. 2010. “Corruption and culture: An experimental analysis.” Journal of Public Economics 94 (11):862–9. Besley, Timothy and Andrea Prat. 2006. “Handcuffs for the grabbing hand? Media capture and government accountability.” The American economic review 96 (3):720–36. Bhargava, Alok; Franzini, L.; Narendranathan, W. 1982. “Serial Correlation and the Fixed Effects Model”. Review of Economic Studies 49 (4): 533–549 Bollen, Kenneth and Robert Jackman. 1985. “Regression diagnostics an expository treatment of outliers and influential cases.” Sociological Methods & Research 13 (4):510–42. Brambor, Thomas. 2005. “Understanding Interaction Models: Improving Empirical Analyses.” Political Analysis 14 (1):63–82. Central Intelligence Agency. 2010. “The world factbook 2010.” Central Intelligence Agency, Washington, DC. Cohen, Jacob. 1992. “A power primer.” Psychological bulletin 112 (1):150–5. Collier, Michael. 2002. “Explaining corruption: An institutional choice approach.” Crime, Law and Social Change 38 (1):1–32. Donchev, Dilyan and Gergely Ujhelyi. 2014. “What Do Corruption Indices Measure?” Economics & Politics 26 (2):309–31. Dong, Bin, Uwe Dulleck, and Benno Torgler. 2012. “Conditional corruption.” Journal of Economic Psychology 33 (3):609–27. Falk, Armin and Urs Fischbacher. 2002. ““Crime” in the lab-detecting social interaction.” European Economic Review 46 (4):859–69. Fisman, Raymond and Roberta Gatti. 2002. “Decentralization and corruption: evidence across countries.” Journal of Public Economics 83 (3):325–45. Gerring, John and Strom Thacker. 2004. “Political institutions and corruption: The role of unitarism and parliamentarism.” British Journal of Political Science 34 (2):295–330. Golder, Matt. 2005. “Democratic electoral systems around the world, 1946‐2000.” Electoral Studies 24 (1):103–21. Graf Lambsdorff, Johann. 2005. “Consequences and causes of corruption: What do we know from a cross-section of countries?” Working Paper Passauer Diskussionspapiere: Volkswirtschaftliche Reihe 34 (4):1–35. Graf Lambsdorff, Johann and Björn Frank. 2011. “Corrupt reciprocity – Experimental evidence on a men’s game.” International Review of Law and Economics 31 (2):116–25. Green-Pedersen, Christoffer and Stefaan Walgrave. 2014. Agenda Setting, Policies, and Political Systems: A Comparative Approach: University of Chicago Press. Hallin, Daniel and Paolo Mancini. 2011. Comparing media systems beyond the Western world: Cambridge University Press. Judge, William, Brian McNatt, and Weichu Xu. 2011. “The antecedents and effects of national corruption: A meta-analysis.” Journal of World Business 46 (1):93–103. Kaufmann, Daniel. 2005. “Myths About Governance and Corruption.” Finance and Development 42 (3):2–41. Kaufmann, Daniel, Gil Mehrez, and Tugrul Gurgur. 2002. “Voice or public sector management? An empirical investigation of determinants of public sector performance based on a survey of public officials.” Washington, DC: World Bank.:1–45. Kolstad, Ivar and Arne Wiig. 2009. “Is Transparency the Key to Reducing Corruption in Resource-Rich Countries?” World Development 37 (3):521–32.

206

[Determinants of change in the international fight against corruption]

Lederman, Daniel, Norman Loayza, and Rodrigo Soares. 2005. “Accountability And Corruption: Political Institutions Matter.” Economics and Politics 17 (3):1–35. Lessmann, Christian and Gunther Markwardt. 2010. “One Size Fits All? Decentralization, Corruption, and the Monitoring of Bureaucrats.” World Development 38 (4):631–46. Lind, Georg. 2012. Effektstärken: Statistische, praktische und theoretische Bedeutsamkeit empirischer Studien: Bibliothek der Universität Konstanz. Lindstedt, Catharina, and Daniel Naurin, eds. 2005. Transparency and corruption: The conditional significance of a free press. Budapest. Lindstedt, Catharina and Daniel Naurin. 2010. “Transparency is not enough: making transparency effective in reducing corruption.” International Political Science Review 31 (3):301–22. Lio, Mon-Chi, Meng-Chun Liu, and Yi-Pey Ou. 2011. “Can the internet reduce corruption? A cross-country study based on dynamic panel data models.” Government Information Quarterly 28 (1):47–53. Markovits, Andrei and Mark Silverstein. 1988. The politics of scandal: Power and process in liberal democracies: Holmes & Meier New York. Mauro, Paolo. 1995. “Corruption and Growth.” The Quarterly Journal of Economics 110 (3):681–712. Mungiu-Pippidi, Alina. 2013. “Controlling corruption through collective action.” Journal of Democracy 24 (1):101–15. Olken, Benjamin. 2009. “Corruption perceptions vs. corruption reality.” Journal of Public Economics 93 (7-8):950–64. Peisakhin, Leonid and Paul Pinto. 2010. “Is transparency an effective anti-corruption strategy? Evidence from a field experiment in India.” Regulation & Governance 4 (3):261–80. Pellegrini, Lorenzo and Reyer Gerlagh. 2008. “Causes of corruption: a survey of cross-country analyses and extended results.” Economics of Governance 9 (3):245–63. Persson, Anna, Bo Rothstein, and Jan Teorell. 2013. “Why Anticorruption Reforms Fail— Systemic Corruption as a Collective Action Problem.” Governance 26 (3):449–71. Pieth, Mark. 2014. “Collective Action and Corruption.” In Preventing Corporate Corruption: Springer, 93–108. Porta, Rafael, Florencio Lopez-De-Silanes, Andrei Shleifer, and Robert Vishny. 1997. “Legal determinants of external finance.” The Journal of Finance 52 (3):1131–50. Rothstein, Bo. 2011. “Anti-corruption: the indirect ‘big bang’ approach.” Review of International Political Economy 18 (2):228–50. Rothstein, Bo. 2014. “What is the opposite of corruption?” Third World Quarterly 35 (5):737–52. Serra, Danila. 2006. “Empirical determinants of corruption: A sensitivity analysis.” Public Choice 126 (1-2):225–56. Svensson, Jakob. 2005. “Eight Questions about Corruption.” Journal of Economic Perspectives 19 (3):19–42. Transparency International. 2014. Corruption Perception Index. Berlin. http://www.transparency.org/ (Accessed August 30, 2014). Treisman, Daniel. 2000. “The causes of corruption. A cross-national study.” Journal of Public Economics 76 (3):399–457. Treisman, Daniel. 2007. “What Have We Learned About the Causes of Corruption from Ten Years of Cross-National Empirical Research?” Annual Review of Political Science 10 (1):211–44. Tsebelis, George. 2002. Veto players: How political institutions work: Princeton University Press.

207

THE PUBLIC SPHERE | 2016 ISSUE

Walton, Grant. 2014. “Defining Corruption where the State is Weak: The Case of Papua New Guinea.” Journal of Development Studies 50 (7):1–17. Wermuth, Nanny. 1992. “On Block-Recursive Linear Regression Equations.” Revista Brasileira de Probabilidade e Estatística 6 (2):1–56. Winfield, Betty, Takeya Mizuno, and Christopher Beaudoin. 2000. “Confucianism, collectivism and constitutions: Press systems in China and Japan.” Communication Law and Policy 5 (3):323–47. World Bank. World Development Indicators. http://data.worldbank.org/data-catalog/ world-development-indicators (Accessed August 30, 2014).

208