The industry anatomy of the transatlantic productivity growth slowdown: Similarities outweigh the differences
Since 2005, productivity growth in america and Europe has dipped below 1%. Using new industry-level from the united states and ten EU countries, this column implies that that the industrial composition of the slowdown was similar in Europe and the united states. Falling multifactor productivity growth explains both magnitude and composition of falling productivity growth on both sides of the Atlantic. Decelerating technical change, instead of slowing investment, was the principal driving force in the transatlantic slowdown.
Slowing labour productivity growth for just about any given income share of labour directly results in a slower growth rate of real wages and a nation’s quality lifestyle. Each percentage point where productivity growth declines results in the same decrease in the growth rate of potential output, reducing a nation’s capacity to finance national security, education, healthcare, and old-age pensions.
Tepid rates of productivity growth below 1% in both US and Western Europe since 2005 have raised the spectre of secular stagnation, blamed variously on lagging innovation, low investment, growing corporate concentration, reduced business dynamism, and rising income inequality. The productivity slowdown has taken centre stage in monetary policy deliberations through the downward pressure it has exerted on the true natural interest, giving central banks less room to drop rates to combat future recessions. For instance, estimates of the united states natural rate by the Federal Reserve Bank of NY (2019) predicated on work by Laubach and Williams (2003) show a decline from over 4% in the 1970s to under 1% in the 2010s.
A notable contrast over the Atlantic may be the speedup folks productivity growth from 1.25 percentage points each year in 1972-1995 to 2.17 points in 1995-2005 – commonly related to widespread adoption of PCs, the web, and other information and communications technology (ICT) – pitched against a slowdown over the same intervals from 2.31 to at least one 1.26 percentage points within an aggregate of 10 EUROPEAN nations (the EU10). 1 Explanations for why Americans seemingly benefited more from ICT adoption than Europeans often appeal to country-specific factors such as for example differences in incentives for innovation or labour market institutions.
Our recent research (Gordon and Sayed 2019) requires a different view, suggesting that the slowdowns in US and European growth have significantly more in keeping than previously recognised. We utilise new industry-level data from the united states and the EU10 to compare productivity growth across time intervals divided in 1972, 1995, and 2005. We emphasise the contributions of specific industries to growth in labour productivity and its own three growth-accounting components – investment (capital deepening), multifactor productivity (measuring innovation and technological progress), and labour quality. 2
Figure 1 displays the annual growth rates of output each hour since 1950 for the full total US economy in red and an aggregate of the EU10 in blue.3 Averages over selected intervals are listed in Table 1. US productivity growth averaged 2.54 percentage points each year from 1950-1972, fell to at least one 1.25 points from 1972-1995, rebounded to 2.17 between 1995 and 2005, and lastly fell to 0.87 points during 2005-2015. Rather than exceptional same up-down-up-down pattern as the united states, Table 1 demonstrates the EU10 experienced a continuing slowdown, with average annual productivity growth averaging 4.85 percentage points from 1950-1972, slowing to 2.31 during 1972-95, falling again to at least one 1.26 between 1995-2005, and finishing at 0.63 points during 2005-15.
Figure 1 US vs. EU10 total economy labour productivity growth 1950-2015, centred five- year moving averages
Source: EUKLEMS data aside from EU10 1950-72 for the full total economy, which is GDP each hour from the Conference Board Total Economy Database, weighted together for the ten EU nations using real GDP weights.
Table 1 Annual average growth rates of labour productivity by sector, selected intervals, 1950-2015
Source: All cells are computed from the merged KLEMS database aside from the EU10 in 1950-72, which originates from the Conference Board Total Economy Database.
The postwar history of the transatlantic productivity comparison could be told either with growth rates or comparative degrees of labour productivity, which may be measured as purchasing-power-parity output divided by hours. The EU10 level in 1950 was just half that of the united states, reflecting the economic dislocations of both world wars and the interwar period, when the united states degree of productivity leapt before Europe. By 1972, the amount of European productivity had reached 81% of the united states and soared further to 106% in 1995 before falling back again to 90% in 2005.
These level data claim that the burgeoning European productivity growth of 1950-1972 could be interpreted as catching-up to the pre-1950 accomplishments of the united states. During 1972-1995, Europe slowed to an interest rate comparable to the original 1950-72 postwar growth rate of the united states. Timmer et al. (2011) similarly argue that European productivity played ‘catch up’ with the united states from 1950-73 through technology and institutional imitation, with 1973-95 growth slowing as the EU ‘caught up’.
Table 1 implies that taking US labour productivity during 1950-1972 as isomorphic compared to that of the EU10 during 1972-1995 allows a fresh interpretation of postwar history. Between 1950-72 and 2005-15, US productivity growth slowed by 1.67 percentage points, while between 1972-95 and the same 2005-15 endpoint European growth slowed up by almost the same amount – 1.68 points. Similarly, an aggregate of commodities-producing industries (such as for example manufacturing, agriculture, and mining) slowed by 2.13 points in america and 2.15 points in Europe, while an aggregate of services-producing industries (such as for example retail trade, communications, and finance) slowed by 1.29 points in america and 1.05 points in Europe.
The strong correlation across industries in the extent of the slowdown is plotted in Figure 2. The vertical axis measures the change between 1972-95 and 2005-15 in the productivity growth rate of individual industries in the EU10, as the horizontal axis measures the slowdown for the same industries in america between 1950-72 and 2005-15. Observe that seven of the 16 industries are plotted in the southwest quadrant, indicating a decline for both US and the EU10. Five of the – utilities, agriculture, transportation, manufacturing, and education – lie on the grey 45-degree line, indicating a similar slowdown on both sides of the Atlantic. The correlation coefficient of the partnership in Figure 2 is 0.81 when the outlier mining industry is excluded.
Figure 2 Regression of EU10 1972-95 to 2005-15 productivity slowdown on US 1950-72 to 2005-15 productivity slowdown
When dividing the economy between commodities-producing industries and the ones producing services, commodities dominate changes across intervals in america. Commodities contributed to the entirety of the united states slowdown after 1972, falling from 3.39 percentage points each year to at least one 1.45, while services experienced no net change. The revival folks productivity growth after 1995 was also mostly because of a 2.15-point rise for commodities instead of a revival of only 0.64 points for services. The post-2005 US slowdown was shared between commodities and services, both slowing by about 2 percentage points. In Europe, growth in commodities and services slowed after 1995 and 2005, however the slowdown in annual commodities growth was almost doubly large (2.15 percentage points) as that in services (1.05 points).
As noted above, labour productivity growth could be split into three components – capital deepening, multifactor productivity growth, and the contribution of labour quality. Ignoring the latter as relatively unimportant, we are able to investigate how capital deepening contrasted with multifactor productivity growth as a way to obtain overall productivity growth. For the united states, the contribution of capital deepening was relatively steady before 2005, so the changes in productivity growth across time intervals largely mirrored changes in multifactor productivity growth, which explains 1.03 of the 1.29-point post-1972 slowdown and 0.71 of the 0.92-point post-1995 revival. On the other hand, multifactor productivity growth explains only 0.53 points of the 1.30-point post-2005 slowdown. Data for the EU10 reveal a roughly co-equal role of multifactor productivity growth and capital deepening in explaining both post-1995 and additional post-2005 productivity growth slowdowns.
This pattern of changes in multifactor productivity growth supports our broad theme that the industry anatomy of the productivity slowdown was similar over the Atlantic. Contrasting the slowdown ending in 2005-15 and beginning with 1950-72 in america, and 1972-95 in the EU10, we find that five of the six industries with the biggest slowdowns in multifactor productivity growth will be the same in both regions -agriculture, transportation, utilities, construction, and manufacturing. Four of the six manufacturing sub-industries with the biggest multifactor productivity slowdowns are also the same – chemicals, petroleum, food, and machinery not elsewhere classified. Combined with tight correlation between your magnitudes of slowdowns folks and European industries, this shows that falling multifactor productivity growth explains both magnitude and composition of falling productivity growth on both sides of the Atlantic. This supports our theme a diminished role of innovation, not slowing investment, may be the most important underlying reason behind the transatlantic productivity growth slowdown.
One interesting difference between your US and Europe may be the contribution of the electronics manufacturing industry to aggregate productivity growth. In america, this industry achieved productivity growth of 13.5 percentage points a year from 1972-1995, 17.6 points from 1995-2005, and fell to (still impressive) 7.07 points after 2005. European electronics manufacturing growth through the same three periods was much slower – only 5.47, 4.78, and 2.95, respectively. Beyond electronics, another outlier in the info is apparently mining in america, the only industry to visit a meaningful rise in productivity growth from -1.87 in 1995-2005 to 2.98 in 2005-2015, likely because of the development of fracking that didn’t occur in Europe (where growth in mining productivity was negative before and after 2005). Additionally, chemicals manufacturing is apparently an outlier because of its stronger productivity growth in Europe across all schedules.
Despite these differences, our findings shed a fresh light on the transatlantic productivity slowdown by highlighting some striking similarities between your US and Europe. The industrial composition of the slowdown was similar in Europe and the united states, and we indicate decelerating technical change as the principal driving force in the transatlantic slowdown. Fluctuations in commodities-producing industries explain more of the period-to-period slowdowns of every region than do services-producing industries. Further insight in to the resources of the productivity slowdown should be fought in the trenches of detailed studies of individual industries, and our study has helped to indicate those particular industries that are most looking for further insight and evaluation.
Laubach, T and J C Williams (2003), “Measuring the natural interest”, Overview of Economics and Statistics 85(4): 1063-70.
Timmer, M P, R Inklaar, M O’Mahony and B Van Ark (2011), “Productivity and economic growth in Europe: A comparative industry perspective”, International Productivity Monitor 21: 3-23.
 The ten nations are Austria, Belgium, Denmark, France, Germany, Italy, Luxembourg, holland, Spain, Sweden, and the united kingdom.
 Our data decompose total economic activity into 16 industries, 12 which are available in the market sector and four which are in the nonmarket sector. We likewise have an additional decomposition of the manufacturing sector into 11 manufacturing sub-industries.
 All references to productivity growth in this column make reference to the full total economy including agriculture, government, and households, not the additionally cited productivity data for the nonfarm private business sector. Total economy productivity growth is always slower than in the nonfarm private business sector, but by varying amounts.