74
GLOBAL PRODUCTIVITY
CHAPTER 1
ANNEX 1A Cyclical and technology-driven labor productivity developments This annex describes the SVAR used to separate supply (technology) and demand-side influences on labor productivity growth. The methodology used to identify supply-side “technology” drivers of labor productivity uses a Spectral identification. “Technology” shocks are identified as those that explain the majority of productivity fluctuations at frequencies longer than 10 years—this approach disregards fluctuations at higher (shorter) frequencies and is robust to contamination in economies where productivity is affected by many other factors such as demand shocks. This approach identifies longlasting innovations to labor productivity, assuming that these highly persistent changes are likely to consist of structural supply-side factors. The methodology is further explored in chapter 6.
Estimation Each SVAR is estimated using annual data and consists of the natural log-difference of labor productivity, the log-level of employment, the share of investment and separately consumption in GDP, the consumer price inflation rate, and, where available, the shortterm policy interest rate. Table 1A.1 provides summary statistics on the data length available in each income group. TABLE 1A.1 Median sample periods Labor productivity
TFP
Spectral
AEs
1962-2018
1951-2018
1973-2018
EMDEs
1972-2018
1971-2018
1981-2018
LICs
1981-2018
1981-2018
1981-2018
Source: World Bank. Note: AEs = advanced economies; EMDEs = emerging market and developing economies; LICs = low-income countries; TFP = total factor productivity.
Shock decomposition The decomposition for each region or income grouping is based on individual estimations which are aggregated using GDP weights on GDP at 2010 U.S. dollar prices and exchange rates. Historical decompositions of labor productivity growth, Yt , can be written as a function of the structural shocks identified through the Spectral identification ϵ t , initial condition X0 (which accounts for the lack of data before the start of the sample), and the constant, C: t −1
t −1
Yt = F i∈t −i + At X 0 + F i C. i =0
i =0
In the decompositions shown in figure 1.5, the identified technology shock, initial condition, and constants are included in the “technology” category, given that they
