Summary
Puerto Rico-Michigan Link
Financial Dependence
On the Econometric Specification
Two-Step Approach
Discussion of “The Transmission of Quasi-Sovereign Default Risk: Evidence from Puerto Rico” by Chari, Leary and Phan Marı́a Pı́a Olivero, PhD GIC Conference - LeBow College of Business Drexel University
February 28th , 2020
The Transmission of Quasi-Sovereign Default Risk
Summary
Puerto Rico-Michigan Link
Financial Dependence
On the Econometric Specification
Two-Step Approach
Summary of Results
The Government Demand Channel: An increase in default risk disproportionately affects employment in industries that are more dependent on demand from the Puerto Rican government
The Transmission of Quasi-Sovereign Default Risk
Summary
Puerto Rico-Michigan Link
Financial Dependence
On the Econometric Specification
Two-Step Approach
Take-Aways
Novel way to isolate the macro effects of an exogenous shock, i.e. bankruptcy and non-bailout of Detroit → Puerto Rico
The Transmission of Quasi-Sovereign Default Risk
Summary
Puerto Rico-Michigan Link
Financial Dependence
On the Econometric Specification
Two-Step Approach
Take-Aways
Novel way to isolate the macro effects of an exogenous shock, i.e. bankruptcy and non-bailout of Detroit → Puerto Rico New government demand channel 6= well known financial dependence channel
The Transmission of Quasi-Sovereign Default Risk
Summary
Puerto Rico-Michigan Link
Financial Dependence
On the Econometric Specification
Two-Step Approach
Take-Aways
Novel way to isolate the macro effects of an exogenous shock, i.e. bankruptcy and non-bailout of Detroit → Puerto Rico New government demand channel 6= well known financial dependence channel Interesting case!! default risk can be disentangled from currency risk
The Transmission of Quasi-Sovereign Default Risk
Summary
Puerto Rico-Michigan Link
Financial Dependence
On the Econometric Specification
Two-Step Approach
The Instrument?
Michigan vs Puerto Rico GDP (in millions of current $) 600000.0
120000
500000.0
100000
400000.0
80000
300000.0
60000
200000.0
40000
100000.0
20000
0.0
1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018
0
GSP_MI
GDP_PR
The Transmission of Quasi-Sovereign Default Risk
Summary
Puerto Rico-Michigan Link
Financial Dependence
On the Econometric Specification
Two-Step Approach
Publicly-Traded vs Smaller Firms in Manufacturing 1.5
1
.5
0
-.5
-1 310
320 FINDEP_ASM
The Transmission of Quasi-Sovereign Default Risk
NAICS
330 FINDEP_Compustat
340
Summary
Puerto Rico-Michigan Link
Financial Dependence
On the Econometric Specification
Two-Step Approach
5 Most Financially-Dependent Industries
(1) (2) (3) (4) (5)
Compustat Chemicals Beverages and tobacco Computer and electronics Furniture and related products Miscellaneous
The Transmission of Quasi-Sovereign Default Risk
ASM Computer and electronics Paper Nonmetallic minerals Plastics and rubber Wood products
Summary
Puerto Rico-Michigan Link
Financial Dependence
On the Econometric Specification
Two-Step Approach
5 Least Financially-Dependent Industries
(1) (2) (3) (4) (5)
Compustat Apparel Printing and related support activities Leather and allied products Wood products Food
The Transmission of Quasi-Sovereign Default Risk
ASM Leather and allied products Furniture and related products Apparel Textiles Petroleum and coal
Summary
Puerto Rico-Michigan Link
Financial Dependence
On the Econometric Specification
Lags in the Econometric Specification
The transmission channel government default → government demand → employment in “government-intensiveâ€? industries Lags? In D’Erasmo, Moscoso-Boedo and Olivero (forthcoming) ∆lit = Îątk + β1 Ei2001 + β2 F Ci2001 + β3 Xit−3 + it
The Transmission of Quasi-Sovereign Default Risk
Two-Step Approach
Summary
Puerto Rico-Michigan Link Financial Dependence Onvariables the Econometric The controls at the bank level are intended to account for other that can Specification impact the
banks’ availability of loanable funds. As such, the vector Xit−3 includes banks’ liquidity and leverage, the level of loans to the private non-financial sector, and profitability as measured by banks’ net income.17 We include the most stringent fixed effect such that we are able keep the variables of interest in the regression.18 We measure liquidity using the ratio of liquid assets to total assets and bank leverage using the inverse of the equity to asset ratio. Our measure of bank size is standard and given by the value of total assets. Table 2 presents the results.
Lags in the Econometric Specification Bank-Level Effects of Sovereign Debt and Foreign Currency Exposure Dep. Variable Government Exposure Sov. Debt Exposure (Ei,2001 )
∆`it -0.845** (0.047)
-0.923** (0.030)
-0.985** (0.018)
-0.721* (0.095) -0.298 (0.118)
-0.747* (0.084) -0.495** (0.014)
-0.847** (0.045) -0.386* (0.051)
1.353*** (0.000) -0.622*** (0.000)
1.553*** (0.000) -0.755*** (0.000) 0.0501** (0.024)
1.239*** (0.000) -0.633*** (0.000)
1.440*** (0.000) -0.824*** (0.000) 0.0692*** (0.003)
yes 2003-2005 3,220 0.029
yes 2003-2005 3,220 0.030
1.393*** (0.000) -0.431*** (0.001) 0.0219 (0.314) 2.387*** (0.000) yes 2003-2005 3,220 0.077
yes 2003-2005 3,220 0.029
yes 2003-2005 3,220 0.032
1.306*** (0.000) -0.487*** (0.000) 0.0370 (0.109) 2.371*** (0.000) yes 2003-2005 3,220 0.078
FC Exposure (F Ci,2001 ) Bank Characteristics Liquidityi,t−3 Leveragei,t−3 (log) Real Assetsi,t−3 Net Incomei,t−3 Bank Type×Time FE Period N. of Observations R-squared
Note: “Sov. Debt Exposure” refers to the ratio of domestic government bonds to total assets in 2001. “FC Exposure” refers to the ratio of non-deposit foreign currency liabilities to total assets in 2001. p-values in parentheses. *p<0.10, **p<0.05, ***p<0.01.
The estimated coefficients indicate that exposure to sovereign debt prior to the time of default and devaluation has a negative and significant effect on the availability of credit postdefault.19 This prima facie evidence suggests that the supply channel of default is statistically 17 18
Appendix A-3 presents several robustness checks that involve including additional controls. Results do not change if we include only time fixed effects, only bank-type fixed effects, or time and
The Transmission of Quasi-Sovereign Risk bank-type fixedDefault effects separately.
Two-Step Approach
Summary
Puerto Rico-Michigan Link
Financial Dependence
On the Econometric Specification
Two-Step Approach
Diff-in-Diff vs Two-Step Approach á la Kashyap and Stein (2000) Exploit the panel nature of the data in two separate steps Relaxing the linearity assumption regarding the effects of default-risk on employment logEit ∂ 2 logEit ∂GOVi ∂P ostDetroitt
= α + βGOVi × P ostDetroitt + ......... = β̂
logEi
=
α + βGOVi + .........
β̂
=
δP ostDetroitt
∂ logEit ∂GOVi ∂P ostDetroitt
=
∂ β̂ = δ̂ ∂P ostDetroitt
2
The Transmission of Quasi-Sovereign Default Risk
∀t