Body
I Introduction
As the impact of the Great Recession gradually recedes on both sides of the Atlantic, political and economic attention will invariably shift toward prospects for renewed job creation. This paper will show how, in contrast to the outcome of earlier recessions, European labor markets have weathered the effects of the Great Recession noticeably better than that of the United States, how on both sides of the Atlantic many of the jobs lost during the Great Recession have likely disappeared permanently, and describe how the United States and Europe face different policy challenges in addressing this shock. In the United States, the key issue will be to ensure adequate access to retraining for many unemployed, who will have to look for jobs outside their previous sectors of employment. Meanwhile, it is crucial that Europe maintains its recent successful focus on "job rich growth" by continuing to support worker retraining and further liberalizing core parts of European labor markets.
When thinking about this issue in a trans-Atlantic context, a striking feature of the Great Recession is the pronounced "leveling effect" it has had on labor markets in the United States and the EU. Europe's economic downturn was considerably deeper, over 4 percent of GDP compared with approximately 2.5 percent of GDP experienced in the United States. But disproportionately it was American jobs that disappeared during this period, with U.S. headline unemployment roughly doubling to between 9-10 percent throughout 2010. For the first time in almost 30 years U.S. unemployment has for a sustained period of time been about the same level as in the EU, where unemployment rose by only 2-3 percentage points during the Great Recession. As described by the OECD (2010), various government supported short-time-work schemes, combined with reductions in overtime hours and other employer initiatives meant that most of the EU labor input reduction was achieved through a reduction in Europeans' hours worked, rather than through the destruction of European jobs. In contrast, U.S. employers shed workers aggressively during the Great Recession, causing the traditional Okun's law relationship between contractions in GDP and unemployment to break down1. As a result, the United States and EU has experienced widely diverging productivity trends since 2008, with strong output per hour gains in the United States and declines in most EU countries2.
Labor hoarding in many European countries has reduced the social costs of the Great Recession by avoiding large increases in unemployment.. However, at the same time European countries face a potential jobless recovery, as their lower hours have been associated with a reduction in hourly productivity. This would be the outcome across Europe, if hours worked per employee and hourly productivity were to rise back to pre-crisis levels – a substantial increase in GDP without new jobs being created. Meanwhile, judged principally on the large number of job cuts and countercyclical productivity growth during the recession, the United States looks well poised for substantial job creation during the recovery.
II Longer-Term Trends in Employment Rates in the United States and EU-15
However, analysis of longer-term labor market developments on both sides of the Atlantic suggests that the outlook for employment growth is less certain. Looking beyond the short-term jobs picture and instead on the longer-term employment trends adds to the understanding of today's employment projections. Figure 1 shows the comparable overall employment rates3 in the United States and EU-15 from 1992 to the present for the overall population, and figure 2 for men only.
The employment/population ratio is the most intuitive metric for the ability of an economy to employ its working age population. As such it is a better indicator of the state of the labor market than the headline unemployment number, which is heavily influenced both by entries to and exits from the overall measured labor force and by the number of unemployed4.
Several clear trends are visible in figures 1 and 2; first of all, figure 1 shows how after the recession of 2001 the United States never regained the employment rates of the 1990s. Instead, US job creation during the housing boom and economic expansion from 2001 to 2007 merely managed to stabilize employment rates before the dramatic, roughly five percentage point decline during the Great Recession. The latest available data puts the total U.S. employment rate for the working age population at just 66.6 percent, an extremely low level last seen in the U.S. economy in the early 1980s, well before the full entry of American women into the workforce. Or, to put it another way, when controlling for population growth, every U.S. job created since Ronald Reagan's first term was destroyed during the Great Recession.
Figure 1 also shows how the EU-15 generally managed to raise its employment rate from the early 1990s to the Great Recession, and from 2008 to 2010 saw just a two percentage point decline in employment. As a result, the long-standing U.S. advantage in working age employment rate vs. the EU-15, an advantage of at least 10 percentage points, has largely disappeared in recent years. The latest available EU-15 employment rate for Q3 2010 stood at 65.7 percent, while the U.S. rate was just about one percent higher. In addition figure 1 shows how in 2010 Germany and the UK enjoyed employment rates considerably above the U.S., with France slightly lower and Spain and Italy significantly below. Moreover, it is evident how among the large EU-15 members only Spain, which suffered an even bigger housing bust than in the United States, suffered a large decline in employment during the Great Recession, whereas German employment rose substantially.
Figure 2 illustrates how the longer-term overall trans-Atlantic employment trends are the same for men. The U.S. working age male employment rate dropped to just 70 percent in Q4 2010, the lowest level since WWIII. Meanwhile, EU-15 male employment rates have generally held up far better in the last two decades and are now comparable to those of the United States. Moreover, German and British male workers are now substantially more likely to be employed than their U.S. counterparts, with only EU-15 Mediterranean males less likely to work.
Figure 1 and 2 clearly illustrate that recent dramatic declines in U.S. employment are less of an unprecedented one-off situation than a striking acceleration of longer-term trends. The relatively smaller employment impact of the Great Recession in the EU-15 comes as a continuation of a generally improving employment situation over the last two decades. The traditional notion of the U.S. labor market as a superior "job creation machine" is far less true in the post-crisis era than in earlier periods.
As discussed above, reductions in the number of hours Europeans work accounted for a substantial part of the region's reduction in labor input during the crisis. While this will tend to bias trans-Atlantic employment rate comparisons in favor of Europe5 (as seen in figures 1 and 2), the same general trend is found also when looking at hours worked/capita data. "Hours worked/capita" is a broad metric that has the advantage of incorporating both changes in employment and hours worked per employee. As such, it takes account of the effects of both job losses, reductions in working time and increases in the frequency of part-time work. The most recent data from the Conference Board6 suggests that U.S. hours worked per capita dropped by 70 hours annually from 835 in 2008 to 765 hours in 2010, a level last seen in the early 1980s. Concurrently, hours worked/capita in the EU-15 dropped just 26 hours, from 735 hours in 2008 to 709hour in 2010. Thus, while aggregate labor input per capita remains higher in the U.S. economy than in the EU-15, this is less true after the Great Recession than before, and the trend toward a relatively smaller U.S. employment advantage persists.
III Sectoral Trends in Employment Gains and Losses in the United States and EU-15
The next comparative question to ask relevant to the prospects for trans-Atlantic job creation going forward is the sectoral distribution of job losses during the Great Recession and whether job gains and losses are structural or cyclical in character. For these purposes, it is informative to apply the labor market analysis methodology utilized by Groshen and Potter (2003) and Kirkegaard (2009), which breaks down the total economy by industrial sector and estimates which sectors have gained and lost jobs relative to the overall economy over the economic recession and recovery periods that account for the two components of the peak-to-peak business cycle. Sectors that have lost relatively more jobs than the overall economy during both the economic contraction and the subsequent economic expansion are deemed to be experiencing "structural job losses", while sectors gaining more jobs than the overall economy over the cycle are seeing "structural job gains". Sectors that see faster job growth during expansions and contract more during recessions are "pro-cyclical", while sectors with slower employment gains than the total economy during expansions and smaller losses in recessions are "counter-cyclical".
Figure 3 shows the trends for the latest peak-to-peak business cycle (i.e. the Great Recession and the subsequent expansion from June 2009 to January 2011) in the United States using the latest available Bureau of Labor Statistics employment data and using the Great Recession dates as identified by the National Bureau of Economic Research. Figure 4 shows the EU-15 trends with the latest available Eurostat employment data7. As the EU has no official institution identifying the beginning and end of EU recessions, and no monthly employment data is available, the EU-15 Great Recession dates are instead identified by the peak and trough of quarterly GDP growth rates (i.e. the Great Recession and the subsequent expansion from Q2 2009 to Q3 2010).
Figure 3 shows several things; first in the lower left corner, it is evident how the construction sector has lost far more jobs than the overall U.S. economy during both the recession (X-axis) and the expansion (Y-axis. This puts the sector into this methodology's "structural losses" category, indicating that many of these jobs will not come back to the U.S. economy. That is similarly true for the U.S. manufacturing sector (although less so during the recent expansion), again indicating that most of the manufacturing jobs lost in the current recession are unlikely to come back. The same however is true for several large services sectors in the U.S., too, as figure 3 shows how the Trade (retail/wholesale), Transportation, Utilities, Financial Activities and Information sectors also have seen "structural job losses" since the beginning of the Great Recession.
Figure 3 (unsurprisingly) also shows how the government is a counter-cyclical employment sector in the U.S., and that there have been relatively few pro-cyclical private sector employers since the Great Recession began, just Professional and Business Services and Mining and Logging.
Figure 3 shows how the continued structural job gains in the U.S. economy since 2008 has been concentrated in the two large services sectors of Leisure and Hospitality (by far the largest sub-category is restaurants8) and Education and Healthcare Services (Healthcare is by far the largest sub-category).
In many ways, Figure 4 for the EU-15 (allowing for the EU's different data classification system and for the fact that the education sector in Europe is almost wholly inside the government sector) shows trends remarkably similar to those in the United States. The EU-15 Construction and Manufacturing sectors have also seen structural job losses, as have the European Trade, Transportation and Financial and Insurance sectors. Meanwhile, European governmental sectors like Public Administration and Defense and Education have been more stable employers. As in the United States, the majority of EU-15 structural job gains have been in healthcare related sectors and in Accommodation and Food Service Activities.
Figures 3 and 4 demonstrate that in both the United States and EU-15, the Great Recession and the subsequent economic recovery accelerated the long-term decline of manufacturing employment9, saw large declines in the construction sector, as well as several large services sectors. The disproportionate employment losses among men in both the United States and EU-15 are verified by this sectoral pattern. As such, on both sides of the Atlantic, the ongoing employment shift away from manufacturing and into services continues, but notably into only a limited number of services sectors, concentrated in healthcare related and food preparation sectors. Broad employment growth across most services sectors has disappeared with the Great Recession. With relatively few pro-cyclical sectors in the economy it furthermore seems clear that the opportunities for workers to be "re-hired" back into their old sector is quite limited, as so few of the economic sectors which saw big job losses during the contraction returned to strong employment growth during the expansion. Most people who have lost jobs on both sides of the Atlantic during the Great Recession consequently seem likely to have to look for new employment outside their old sector of employment.
IV Implications for Trans-Atlantic Capitalism and Future Job Creation
With many trans-Atlantic job losses during the Great Recession likely to be of a structural nature and requiring the unemployed to retrain and seek skills for reemployment in a new sector, without broad access to skill upgrading and retraining, the prospects for many Americans for rapid reemployment are dim and subdued job creation will continue in the short to medium term. Expanding demand alone looks unlikely to raise employment levels sufficiently. With U.S. long-term unemployment rates at persistent historically high levels, the risk of harmful long-term declines in the re-employability of many workers in the U.S. labor market is already elevated. This will in many ways be a novel labor market policy challenge in the United States. Meanwhile in Europe, where the problem of long-term unemployment is long-standing, this continuing challenge will require a redoubling of existing EU-15 efforts to combat it.
In the face of this challenge, the United States currently spends relatively little on worker retraining. Significantly increasing public funding for such skill building will be of paramount importance. The goal must be that all American workers who want to retrain have that opportunity. One option might be to give the unemployed a "Worker Retraining Voucher" repayable to the government by the recipient only after success in finding a new job. This is the equivalent of extending "student loan" type programs, where students gets subsidized government loans payable after graduation, to today's unemployed.
At the same time in the EU-15, where job losses have been concentrated among young, low-skilled workers, often in temporary positions, crisis-induced efforts to loosen employment protection legislation in many countries (noticeably in the eurozone periphery) must be accelerated and expanded in other large economies, especially France and Italy. Excessive redundancy costs for protected prime age (usually male) "insider workers" must be slashed to facilitate broad-based job creation benefitting vulnerable "outsider groups", many still closed services sectors opened up and pan-European acceptance of certifications and qualifications accelerated.
Only by embracing these two distinct sets of policies can the trans-Atlantic economy hope to begin producing broad-based employment gains again following the Great Recession.
References used:
- Bureau of Labor Statistics Current Establishment Survey (CES) Database, available at http://www.bls.gov/ces/data.htm
- Daly, M. and Hobijn (2010) Okun's Law and the Unemployment Surprise of 2009. FRBSF Economic Letter March 8, 2010. Available at http://www.frbsf.org/publications/economics/letter/2010/el2010-07.html
- Eurostat Quarterly LFS database, available at http://epp.eurostat.ec.europa.eu/portal/page/portal/statistics/search_database.
- Kirkegaard, J.F. (2009) Structural and Cyclical Trends in Net Employment over US Business Cycles, 1949–2009: Implications for the Next Recovery and Beyond. Peterson Institute For International Economics, Working Paper 09-5. Peterson Institute, Washington, D.C.
- OECD (2010) OECD Employment Outlook 2010. OECD, Paris
- Summers, L. H. (2009) Rescuing and Rebuilding the US Economy: A Progress Report. Prepared remarks at an event held at the Peterson Institute for International Economics
July 17, 2009. Available at /commentary/speeches-papers/rescuing-and-rebuilding-us-economy-progress-report.
Notes
1. See Daly and Hobijn (2010) and Summers (2009)
2. Only in Austria, Ireland and Spain did – as in the United States – total hours worked decline more than GDP in 2008-09. See OECD (2010:43)
3. The age group 15-64 is the standard OECD definition of the "working age population". The most commonly used U.S. employment rate estimated by the Bureau of Labor Statistics is for the population group 16 years and above. This includes all older age groups where employment rates are lower than for the standard working age population. The data points presented in figures 1 and 2 for the United States is therefore higher than the standard referenced BLS data for the U.S. employment ratio. In Q4 2010, the seasonally unadjusted total employment rate for 16-64y reported in figure 1 was 66.6 percent, whereas the total population employment rate for the 16y and above age group was a lower 58.4 percent.
4. Unemployed people that stop looking for work, consequently drop of out the labor force and hence reduces measured unemployment levels is a particular concern.
5. This type of employment data is not available as "full-time-equivalent" (FTE) data.
6. Data from "The Conference Board Total Economy Database, January 2011, http://www.conference-board.org/data/economydatabase".
7. Sectoral classification distinctions exist between NAICS and NACE based data. Correspondingly, direct comparisons of individual data values should be avoided. NAICS classification descriptions available at http://www.census.gov/cgi-bin/sssd/naics/naicsrch?chart=2007; NACE classification descriptions available at http://epp.eurostat.ec.europa.eu/cache/ITY_OFFPUB/KS-RA-07-015/EN/KS-RA-07-015-EN.PDF.
8. See Kirkegaard (2009) for a more detailed breakdown of the NAICS meta-sectors.
9. Similar analysis for individual countries in the EU shows that this is true in Germany, too, despite its strong manufacturing goods export performance.