Productivity growth is slowing around the world. In 2014, according to the Conference Board’s Total Economy Database, growth of total factor productivity (TFP) hovered around zero for the third year in a row, down from 1% in the glorieuse decennia of 1996-2006 and 0.5% in 2007-2012.
The decline in TFP growth is not limited to the advanced countries. TFP growth has been falling in China. It is negative in Brazil and Mexico. It was barely positive in India in 2014. It has declined and is near zero in a number of relatively poor countries that are still far from middle-income status. TFP growth fell in sub-Saharan Africa from 1999-2006 to 2007-2012, and again in 2013-2014. It fell in Russia, Central Asia, Southeastern Europe, and Latin America as a whole.
This is one of the most disturbing and, no doubt, important phenomena affecting the world economy. The question is what lies behind it and whether it might be reversed.
While it may be possible to get some purchase by comparing and contrasting the extent of the TFP growth slowdown in different countries in recent years, this approach has limitations. It is not possible to distinguish the importance of global and country-specific factors, for example, because all countries experienced the same changes in the state of the global economy in, say, 2014. Limiting one’s attention to the recent spate of TFP slumps does not enable one to gauge the extent to which current trends are unprecedented or have been witnessed before.
We take advantage of the fact that there have been productivity slumps before. We build a comprehensive database of a large number of countries, focusing on episodes in which the rate of TFP growth has declined. We look at the distribution of per-capita incomes at which these slumps have occurred. We consider the distribution of TFP slumps over time. We then consider global and country-specific correlates of these TFP slumps.
We take data on TFP growth from the Penn World Tables 8.1 (Feenstra et al. 2015), covering the period ending in 2011. This is the second PWT release with estimates of TFP. We consider all countries aside from oil exporters with a per capita income of at least US$4,000 (2005 prices). We identify TFP slumps and recoveries using a methodology in the spirit of our earlier work (Eichengreen et al. 2012) on growth slowdowns. We consider five successive periods and isolate episodes where the growth rate of TFP was at least 1% lower on average in the second period than in the first.
Incidence of TFP slumps
77 countries have experienced TFP slumps according to this criterion. Figures 1 and 2 show the distribution of TFP slowdowns by year and per capita income.
Figure 1. TFP slumps by year (countries with incomes above $4,000)
In Figure 1, a cluster of such episodes is evident in the early 1970s at the time of the global productivity slowdown. We see another cluster in the late 1980s and early 1990s, just prior to the 1995-2005 acceleration in productivity growth popularly associated with the IT revolution, and a third cluster in the second half of the 1990s in the run-up to the Asian crisis. There is then a fourth cluster in the mid-2000s, at the point in time when the 1995-2005 IT-related productivity surge petered out and just prior to the Global Crisis. That many of these episodes are clustered at particular points in time is suggestive of a role for global factors in TFP-growth slowdowns.
Figure 2. TFP slumps by per capita income (countries with incomes above $4,000)
In Figure 2, we see three modes around $4,000, $11,000 and $33,000. It is tempting to interpret these in terms of the productivity problems of low-, middle-, and high-income countries.
TFP slumps versus growth slowdowns
Readers may worry that we are simply capturing the same slowdowns in GDP growth as in Eichengreen et al. (2012), since TFP (as opposed to capital and labour input) is the component of gross output whose rate of increase often falls most abruptly when the rate of growth of GDP slows.
In fact, although growth slowdowns and TFP slumps do sometimes coincide, there are also many instances where they do not. Recall, for example, the Asian crisis in the late 1990s, following which the rate of GDP growth slowed noticeably across the region. In some countries that slowing was associated with an abrupt downshift in the rate of growth of TFP. But in other instances, it was associated instead with a downshift in the investment rate.
We have a total of 1,052 country-year observations that satisfy our five-year – 1% criterion. Ask now how many of these years are also categorised as a growth slowdown, and to bias the results in the direction of overlap include also growth slowdowns in the preceding or successive year. The answer is 455, less than a half of our TFP-slowdown country-year observations.
Circumstances in which TFP slumps occur
Finally, we attempt to shed light on the circumstances in which TFP slumps occur. In a regression analysis, we find a negative association between the incidence of TFP slumps and educational attainment as measured by average years of schooling.
- Countries with better educated populations are evidently better able to avoid TFP slumps.
- Countries with stronger political systems as measured by their Polity2 scores are similarly less susceptible to TFP slumps.
- In contrast, countries with high investment shares of GDP are more susceptible to TFP slumps, consistent with the existence of a trade-off between extensive and intensive growth that places a priority on, respectively, capacity expansion and productivity growth.
In addition, we find an important role for global factors. Global financial stringency (as measured by the level of LIBOR), risk aversion (as measured by the TED spread), and world oil prices are all positively and significantly associated with the likelihood of TFP slumps.
The determinants of productivity growth and productivity slumps are notoriously elusive. It follows that the patterns described here are necessarily suggestive. There is no guarantee that countries investing heavily in education, avoiding excessive investment, or developing relatively strong political systems will necessarily avoid TFP slumps. The good news is that there are in the historical record not just TFP growth slowdowns but also TFP growth accelerations and even recoveries (accelerations following slumps). Whether such accelerations and recoveries are now on the horizon, only time will tell.
Conference Board (2015), “Global Productivity Growth Stuck in the Slow Lane with No Recovery in Sight,” Total Economy Database – Key Findings, New York: Conference Board.
Eichengreen, B, D Park and K Shin (2012), “When Fast-Growing Economies Slow Down: International Evidence and Implications for China,” Asian Economic Papers 11, pp.42-87.
Eichengreen, B, D Park and K Shin (2015), “The Global Productivity Slowdown: Common and Country-Specific Factors,” NBER Working Paper (forthcoming).
Feenstra, R, R Inklaar and M Timmer (2015), “The Next Generation of the Penn World Table,” American Economic Review (forthcoming).