Viewed in this light, the European productivity performance was disappointing compared to the U.S. since the mid 1990s: Hourly labor productivity in the euro area grew at 2.4% per year on average between 1980 and 1995, but grew at just 1.3% on average between 1996 and 2005. In the U.S., by contrast, productivity growth was 1.3% in 1980–95, but 2.4% in 1996–2005.
Labor productivity is defined as follows:
Simply comparing the labor productivity headline numbers reported by different countries is misleading. The following are a series of necessary adjustments:
Labor input: Output per hour worked (as reported by the U.S. and Germany) is the ideal indicator for labor productivity comparisons as it is not influenced by any variables which influence the scale of the labor input rather than its efficiency. These include:
– differences in labor force participation rates across countries;
– differences in unemployment rates;
– differences in the number of hours worked per week (length of regular working week, the share of part-timers);
– differences in the number of days worked per year (number of public holidays, days of paid vacation)
Real Output: Nominal output in OECD countries is measured in the national accounts according to a harmonized system referred to as “93SNA”. Nonetheless, there are a few important differences between the U.S. and many European countries:
– Investment in Software: The U.S. enters a much larger share of software purchases by businesses as investment in the national accounts, boosting the GDP level by as much as 1 percent.
– Quality change and the impact of hedonic prices: Many observers argue that American statistics underestimate inflation in the ICT (information and communication technologies) sector by generously allowing for real quality improvements instead of treating price increases as inflation. The resulting smaller ICT deflator thus yields a larger investment share in real terms.
The OECD Compendium of Productivity Indicators provides consistent estimates of labor productivity levels and growth rates, measured as GDP per hour worked, for 29 OECD countries for the period 1970-2005. Even with adjusted numbers, U.S. labor productivity grew more than 1 percentage point per year faster than Euro area labor productivity over the period 2000-2005 (see Table A2 in the OECD report).
What is driving this divergence? The ECB provides a sector-by-sector breakdown of the gap between the euro area and U.S. labor productivity growth in 1990-1995 and in 1996-2002, dividing sectors into ICT producing, ICT using and all remaining non-ICT industries. Each sector is further broken down into manufacturing and services, respectively.
Diverging labor productivity trends since 1996 seem to reflect primarily developments in ICT-usingservices such as retail, wholesale and financial services. The data therefore point away from the focus of the “New Economy” discussion, which revolves more around the ICT-producing industry.
Kenneth Rogoff notes that “together with a few sister ’big box‘ stores Wal-Mart accounts for roughly 50% of America’s much vaunted productivity growth edge over Europe during the last decade! … The US productivity miracle and the emergence of Wal-Mart-style retailing are virtually synonymous.” Rogoff questions the long-term sustainability of the current productivity momentum in the U.S. as the gains from the replacement of smaller retail stores with Wal-Mart and other big-boxes will occur only once. The same might hold true for the ‘auxiliary activities to financial services’ once the U.S. first mover advantage in offshore outsourcing in this sector runs its course.
While the acceleration in US labor productivity growth was concentrated in a few sectors, the decline in the Euro area productivity growth occurred in all countries across the board, either in terms of labor or total factor productivity (see OECD report, table A2 and A6 on page 17 and 25, respectively.) Robert Gordon shows that after 1995 Europe made labor less expensive, raising hours per capita and reducing the growth rate of labor productivity. This led to a de facto substitution of labor for capital. The fact that many of the newly employed are part-timers or lower-skilled employees is captured by the negative correlation between employment and labor productivity growth in the non-ICT sector as computed by the ECB (see Figure 6 on page 21 of the ECB report).
Looking at long-term projections of labor productivity in the EU member countries the European Commission stresses above all the need to increase labor participation rates and to reduce unemployment. In the short run this might reduce output per hour worked – as shown by Robert Gordon – due to the economy’s capacity constraints. In the long run, however, the dynamics of higher employment growth will shift the productive capacity to a permanently higher level. In this context, the Macroeconomic Policy Institute stresses the importance of sound fiscal policy and a supportive monetary policy by the ECB to foster employment and hence productivity growth in the long run.