Europeans work fewer hours than Americans. This column uses new survey data to disentangle the demographic dimensions and the drivers of this gap. In Eastern and Southern Europe, the gap is driven by lower employment rates, while in Western Europe and Scandinavia it is driven by fewer hours worked per person per week. Europe’s more generous holiday allowance alone accounts for between a third and a half of the gap.
By Alexander Bick, Bettina Brüggemann and Nicola Fuchs-Schündeln*
It is a well-documented fact that aggregate hours worked per person are lower in Europe than in the US. This large difference in hours has a bearing on many policy-relevant issues, such as the measurement of labour productivity (GDP divided by total hours worked) or welfare differences (how much leisure does the average person enjoy) across countries. But what causes the difference? This question has recently sparked an active literature, tracing lower aggregate hours in Europe back to, amongst other things, differences in labour income taxation (e.g. Prescott 2004, Rogerson 2006, Olovsson 2009, McDaniel 2011), institutions (Alesina et al. 2005), and social security systems (Erosa et al 2012, Wallenius 2014, Alonso-Ortiz 2014).
To understand the driving forces of differences in aggregate hours worked, it is very useful to document differences on the disaggregate level – that is, the importance of the different components of hours per person for the aggregate difference in hours, and the role played by differences in the demographic structure. Are fewer Europeans working, or do they work fewer hours per work week, or do they simply enjoy more holiday? Are all Europeans working less than Americans, or is this only true for specific groups? Raised early on in Rogerson (2006), until now these questions could only be answered to a limited extent based on existing data readily available to researchers from the OECD or the Total Economy Database (TED).
In a recent paper, we present new data compiled from national labour force surveys (LFS) that are well suited to provide answers to these questions (Bick et al. 2016a). LFS data allow us to study each of the different components of hours per person separately – namely, employment rates, weeks worked per year, and weekly hours worked per employed. Additionally, we can cut the data along a large number of demographic dimensions, most notably gender, age, education, and sector of employment.
Constructing internationally comparable estimates of hours worked per person, employment rates, and hours worked per employed from labour force surveys is not trivial. One particular challenge that we address in the construction of our dataset is that not all weeks of the year are sampled; in fact, the number of weeks covered by each survey is country-specific and varies considerably over time. This is a major problem for the consistent measurement of hours worked per employed because of issues of seasonality and under-sampling of vacation days. By using external data sources to adjust for vacation days, we are able to successfully deal with both of these issues.
Even in the aggregate, it matters whether one uses statistics based on labour force surveys or from the National Income and Product Accounts (NIPA). While LFS data indicate that Europeans worked 19% fewer hours than Americans in the years 2005-2007, NIPA data only imply a difference of 7%.1 This is mainly due to the fact that national statistical agencies responsible for calculating the NIPA measures rely on vastly different kinds of data sets and methods to obtain their hours estimates, which also vary by country. As a consequence, the OECD itself cautions that their data are not suitable for cross-country comparisons.
Given these large discrepancies, which data one uses plays a role for macroeconomic analyses. First, European labour productivity based on LFS hours (measured as GDP divided by total hours worked) is, at 86%, much closer to the US level than with the 78% based on the NIPA hours. Second, relying on the work by Prescott (2004), we show that through the lens of a neo-classical growth model average tax rate differences and differences in the consumption/output ratios can account for half of the difference in hours worked between Europe and the US if these are measured by LFS hours, as opposed to more than the full difference if they are measured by NIPA hours. Thus, the explanatory power of taxes for the Europe-US hours gap is smaller when relying on LFS data than when relying on NIPA data. Furthermore, in Bick et al. (2016b) we document that the aggregate data by the OECD and TED are subject to substantial revisions over time, which crucially alter the results for the two examples given above. The LFS data in turn are not subject to comparable revisions.
The crucial advantage of LFS data over NIPA data is the ability to construct disaggregate measures of hours worked and employment rates. Exploiting the rich information in the micro data, we are able to decompose the differences in hours worked per person between Europe and the US into their different components, and also analyse to what extent these differences are driven by differences in demographic and sectoral composition.
Figure 1 shows a strong negative correlation between two of these components of hours per person: weekly hours per employed during a regular work week, and the employment rate. Countries with high employment rates feature at the same time low weekly hours worked, and vice versa. While in Eastern Europe (blue circles) and Southern Europe (green triangles) weekly hours lie above the median and employment rates below, for Scandinavia (red squares) the opposite picture emerges. By contrast, the third component, the number of work weeks during the year, is uncorrelated with both hours worked per week and employment rates (not shown here). This tells us that while Europeans uniformly work fewer weeks than Americans, in Eastern and Southern Europe lower hours are additionally driven by lower employment rates, whereas in Western Europe and Scandinavia they are driven by fewer hours worked per work week.
Do these three components of hours worked contribute differently to the overall gap between European and US hours if we also take into account differences in demographic and sectoral composition? For instance, if older people worked on average fewer hours than younger people in all countries, and European countries had an on average older population, this could partly account for the lower hours in Europe. We investigate differences in the composition by gender, age, education, and sectors. Any of these factors can play a role in accounting for differences in hours across countries only if both the composition across countries is different and different groups exhibit different labour supply behaviour within a country. Since the age and gender compositions across countries are in fact very similar, these two factors turn out to not play a role.
By contrast, the educational compositions differ vastly across countries. In general, Europe has a higher share of low- or medium-educated individuals and a lower share of highly-educated individuals than the US, especially in Eastern and Southern Europe. While weekly hours worked are similar for the three educational groups, employment rates increase substantially by education in all countries. Therefore, we find that differences in the educational composition between the US and the European countries are very important for understanding the hours differences via their impact on employment rather than on weekly hours worked. Differences in the sectoral structure turn out to matter only minimally in the statistical decomposition because differences in weekly hours per worker between sectors are small.
Figure 2 summarises the main results – the greater European holiday weeks alone account for between one third and one half of the difference in hours between Europe and the US, and Europe’s larger share of low- and medium-educated people accounts for another one third to one half of the difference. However, because of the strong negative cross-country correlation between employment rates and weekly hours worked, this does not imply that these two components alone account for almost all of the differences. In Scandinavia and Western Europe, lower hours per person than in the US are driven by lower weekly hours, with employment rates in Scandinavia being substantially higher than in the US. In Eastern and Southern Europe, by contrast, lower employment rates explain large fractions of the Europe-US hours difference, whereas weekly hours tend to be higher than in the US.
Implications for policy
The potential implications of these findings for public policy are far-reaching.
- Why are the greater holiday weeks in Europe not offset by higher employment rates or weekly hours worked?
This indicates that the number of weeks worked is determined by other driving factors than those determining employment and weekly hours worked.
- Which policies shape the strong negative correlation between employment and weekly hours worked?
The availability and remuneration of part-time work, or the availability and cost of childcare could be such driving forces.
- Which policies shape the differences in educational compositions across countries?
One potential candidate has been proposed by Guvenen et al. (2014), who show that progressive taxation in Europe distorts the incentives to invest in human capital.
- Our results point also to the importance of understanding why employment rates increase strongly with education within a country, but hours worked per employed do not.
Policies that affect the incentives to work differentially by education could, for example, be welfare systems or different degrees of flexibility in the labour markets for low and high-income earners (for example, due to a minimum wage being binding only for low-income earners). Yet, it remains an open question whether these factors influence the extensive margin more than the intensive one. Thus, the new data that we construct and the results we present provide not only a lot of new avenues for research, but also for the public policy debate.
*About the authors:
Alexander Bick, Assistant Professor of Economics, W.P. Carey School of Business, Arizona State University
Bettina Brüggemann, Assistant Professor of Economics, McMaster University
Nicola Fuchs-Schündeln, Professor for Macroeconomics and Development, Goethe University Frankfurt and CEPR Research Fellow
Alesina, A, E Glaeser, and B Sacerdote (2005) “Work and leisure in the United States and Europe: Why so different?”, NBER Macroeconomics Annual, 20: 1-64.
Alonso-Ortiz, J (2014) “Social security and retirement across the OECD”, Journal of Economic Dynamics & Control, 47: 300-316.
Erosa, A, L Fuster and G Kambourov (2012) “Labor supply and government programs: A cross-country analysis”, Journal of Monetary Economics, 59: 84-107.
Bick, A, B Brüggemann and N Fuchs-Schündeln (2016a) “Hours worked in Europe and the US: New data, new answers”, Working Paper.
Bick, A, B Brüggemann and N Fuchs-Schündeln (2016b) “A note on the Europe-US gap in hours worked per person: Data revisions and different formulas”, Working Paper.
Guvenen, F, B Kuruscu and S Ozkan (2014) “Taxation of human capital and wage inequality: A cross-country analysis”, Review of Economic Studies, 81: 818-850.
McDaniel, C (2011) “Forces shaping hours worked in the OECD, 1960-2004”, American Economic Journal: Macroeconomics, 3: 27-52.
Olovsson, C (2009) “Why do Europeans work so little?”, International Economic Review, 50(1): 39-61.
Prescott, E C (2004) “Why do Americans work so much more than Europeans?”, Federal Reserve Bank of Minneapolis Quarterly Review, 38: 2-13.
Rogerson, R (2006) “Understanding differences in hours worked”, Review of Economic Dynamics 9: 365-409.
Wallenius, J (2014) “Social security and cross-country differences in hours: A general equilibrium analysis”, Journal of Economic Dynamics and Control, 37: 2466-2482.
 Time trends for hours are similar in LFS and NIPA data.
Please Donate Today
Did you enjoy this article? Then please consider donating today to ensure that Eurasia Review can continue to be able to provide similar content.