Why Some Countries Do Better
Pietro Garibaldi, Paolo Mauro
©2000 International Monetary Fund
[Job Creation: Why Some Countries Do Better]
The Economic Issues Series aims to make available to a broad readership of nonspecialists some of the economic research being produced on topical issues by the International Monetary Fund. The raw material of the series is drawn mainly from IMF Working Papers, technical papers produced by IMF staff members and visiting scholars, as well as from policy-related research papers.
The following paper draws on material originally contained in IMF Working Paper 99/109, "Deconstructing Job Creation," by Pietro Garibaldi and Paolo Mauro. Citations for the research referred to in this shortened version are provided in Working Paper 99/109, which readers may purchase ($7.00) from IMF Publication Services, or download from www.imf.org. Charles S. Gardner and Jeremy Clift prepared this version.
Why Some Countries Do Better
Over the past decade, the United States has been very successful at creating jobs. Some other industrial countries have clearly lagged behind. But what is the reason why some countries are more successful than others at creating employment? Are there common factors that explain job creation?
The challenge for many European countries is to create more jobs. The unemployment rate has been notoriously higher in Continental Europe (10 percent in the Euro area in 1999) than in the United States (4½ percent), but there have also been considerable differences within Continental Europe, where the unemployment rate recently ranged from 4½ percent in Portugal to 16 percent in Spain.
Many studies have attempted to explain why some countries have higher unemployment rates than others, but less attention has been devoted to countries' relative performance in job creation, or net employment growth. This paper presents the findings of a new study by IMF staff that has systematically analyzed job creation over the past two decades in the industrial countries, focusing particularly on differences within Europe.
Shifting the focus to job creation has four advantages:
A legitimate argument against this approach is that over very long periods of time, employment growth should simply equal the growth of the working-age population. However, in the medium term--20 years or so--government policies and institutional differences can have a big impact on the share of the working-age population that decides to participate in the labor market. Furthermore, the pace of job creation affects the growth of the working-age population itself, since countries adding jobs fastest will tend to attract more immigrants.
Over the past two decades, net growth of employment varied widely among advanced countries. Outside Europe, the United States, Canada, Australia, and New Zealand created far more jobs than most European countries. And within Europe, France, Italy, and some of the Nordic countries lagged in job creation, while the Netherlands and Ireland were high-fliers, especially in the 1990s.
The study by IMF staff applies standard statistical methods to a wide variety of data reflecting these countries' varied experiences with job creation. For each country, it studies age and gender groups, economic sectors, and type of contracts (part-time versus full-time, and temporary versus permanent), and how they relate to employment growth. These factors dominate the public policy debate about overall job market performance, even though most theories of the labor market focus on aggregate employment and unemployment.
This appears to be the first attempt at a detailed cross-country comparison of job creation over a medium-term period among 21 members of the Organization for Economic Cooperation and Development (OECD), but far more study is needed to confirm some of the policy implications of the work. Simple statistical correlations and accounting techniques can be highly misleading in identifying cause and effect, for example.
Caution is also warranted by the fact that this study deals throughout with numbers of jobs rather than the level or flexibility of wages. Some studies of pay argue that the stronger job performance of the United States compared with Europe may be due to greater U.S. wage flexibility, a factor not assessed here.
A look at job creation in all the advanced countries over the past 20 years shows that some countries did especially well in a limited number of sectors, such as the large U.S. gain in retail jobs, or that some countries with fast job growth over the period began with a favorable mix of jobs, such as relatively small shares in agriculture or manufacturing.
Further analysis, however, shows that these special factors account for only a small portion of the superior employment performance of the non-European countries. Instead, the data indicate that a policy package consisting of low dismissal costs and low taxation may have been far more important in producing rapid job creation, and accounted almost entirely for the difference between Europe and the high-performing non-European countries.
When the focus is narrowed down just to the European countries, the statistical evidence is much less conclusive in accounting for country differences. Notably, the Netherlands' success is largely explained by a remarkable surge in part-time employment. And while the Netherlands' experience is impressive and merits further study, throughout Europe it is also evident that the growth of part-time work has had a high cost in the loss of full-time jobs.
These issues are explored below in five sections:
The contrasts in job creation over the past two decades are remarkable among the 21 countries considered. The study's sample period--1980 through 1997--eliminates shorter term economic ups and downs from the comparisons. In this perspective, non-European countries including Australia, the United States, Canada, and New Zealand outstripped most Continental European countries, except the Netherlands and Switzerland. The non-European high-fliers raised their total number of jobs an average of 1½ percent a year, compared with less than ½ percent a year in Continental Europe (See Figure 1).
In absolute terms, these differences are huge: for a country the size of Italy, for instance, a 1 percentage point difference in employment growth amounts to a difference of some 200,000 jobs a year, or 3½ million jobs over the 17-year period.
Do such sharp differences reflect a need for countries to reform their labor markets, or does the explanation lie in other factors, such as increases in working-age population, total output, or investment in capital? The data offer some insights, starting with changes in working-age population.
If a country's job creation keeps pace with the growth of its working-age population, it is usually considered to be doing well. Indeed, countries with more rapid working-age population growth typically end up creating relatively more jobs. By this yardstick, the United States has truly experienced an employment miracle, creating many more jobs than needed to keep pace with population growth and bringing a dramatic decline in unemployment. Over the last 20 years, the country's ratio of employment to working-age population rose by more than 7 percentage points, despite sizable immigration.
Applying this same yardstick, the ranking of Australia, Canada, Ireland, and New Zealand is somewhat lower, while that of the United Kingdom and Belgium is somewhat higher. But the overall picture remains similar, as it does when growth in output is examined.
Output growth and job creation are intrinsically linked. If it were assumed (probably unrealistically) that labor and capital are used in production in the same proportions regardless of the level of output, job creation and output growth would simply be mirror images of each other. In reality, the proportions vary, and increases in productivity are reflected in differences between employment and output growth.
Over long periods, these differences provide clues to the sources of employment growth. For example, a country developing a new product or becoming more competitive internationally for any reason experiences greater demand for its output, and thus its employment rises to meet the new demand. Some of Ireland's rapid job growth, with briskly rising investment particularly in high technology, may thus make it more an output growth miracle than an employment miracle. The reverse, low output growth, may have hampered job growth performance in countries such as Greece and Sweden.
The study finds that a majority of the European economies substituted capital for labor to a greater extent than the non-European economies with dynamic job growth. Countries with flexible labor market institutions may thus create more jobs because they meet rising demand through hiring as well as through additional capital investment. Some analysts conclude that heightened demands by trade unions in Europe beginning in the late 1970s led to substantial substitution of capital for labor. Canada also increased its capital stock far more rapidly than the number of its employed workers, which suggests that some potential to create jobs was left unexploited.
Focusing only on the 1990s, the performance of some European countries, especially Ireland and the Netherlands, becomes more impressive. Ireland led all advanced countries with 3 percent average annual employment growth for 1990–97. More recently, since 1995, Spain's job creation has been this brisk, too, but it is too early to tell if this simply reflects a business upswing. By contrast, Switzerland's employment growth, rapid until 1990, has turned sluggish, consistent with a slowdown in output. For most other countries, however, the ranking based on 1990–97 is similar to that for 1980–97.
Recent studies argue that historical variations between countries in the mix of jobs by economic sector--particularly farming and manufacturing--play a large role in their different rates of job growth. One suggests that the United States enjoys higher job creation than France because of rapid U.S. gains in retail trade jobs. This is in line with the popular view that most U.S. job creation has been in low-skill, low-wage jobs. Indeed, IMF research confirms that retail job growth added considerably to total employment growth in all the fast job creators, amounting to an annual average of ½ a percentage point over 1983–94 not only in the United States, but also in Australia and Canada.
The study by IMF staff analyzed employment data for 11 economic sectors in 11 countries between 1982 and 1994. Although this is fewer countries than in Section I, it includes rapid job creators such as Australia, Canada, the United States, and the Netherlands as well as some of the slowest: Italy, France, and Sweden.
The analysis supports the possibility that the initial mix of jobs by economic sector is a determinant of overall job creation. In 1982, for instance, several slow job creators, including France and Italy, had large shares of employment in agriculture and industry, where most industrial economies have lost jobs.
But while sectoral factors are significant, for most countries they explain only a small portion of overall job creation, and do not much affect country rankings based on total employment growth. Nor do the results support the idea that retail trade accounts for most of the differences in job growth. If countries' job creation is computed under the extreme assumption that no jobs were created in the retail trade sector at all, the high-performing non-European countries remain by far the most rapid job creators, and the overall ranking is unchanged.
The effects of historical differences in the mix of jobs were tested using an accounting approach that estimates what each country's overall job creation would have been if its job composition in 1982 were the same as the average for all the countries in the sample. The results show that all slow job creators suffered from adverse historical conditions, just as the high-fliers benefited from favorable circumstances, such as relatively low employment in farming and manufacturing. Nevertheless, the country rankings remain broadly unchanged.
Exceptions showed up in a few countries, especially in southern Europe, which had lots of jobs in agriculture at the start. Taking the rough accounting results at face value, if Italy's sectoral mix of employment had been the same as the sample average in 1982, by 1994 it would have had at least 1,200,000 more jobs, and its ranking among countries would have been noticeably better.
The advantages some countries got--and the burden for others--can also be assessed through a similar accounting exercise, estimating what overall job creation would have been in each country if each of its sectors had grown at the same rate as the average for all the countries. This shows that differences in job composition by sector at the start explain only about one fifth of the total differences among countries.
A final illustration that differences in the functioning of individual country job markets may tell much more about their performance than their different historical patterns of employment by sector is that, for example, the United States created more jobs than Italy in all of the 11 sectors considered (See Figure 2). Even though Italy is the country where sectoral effects appear to be the starkest, the chart shows that these effects are unlikely to explain why Italy did worse than the United States.
Since sector differences only partly explain variations in employment growth, the key explanations must be sought elsewhere.
A promising area for investigation is the relationship between employment growth and labor market policies and institutions. The IMF study examined such factors as unemployment benefits, the strength of trade unions and their bargaining practices, and the level of taxation. It also looked at a group of practices that affect the costs of dismissal, also referred to as employment protection legislation. These include the amount of notice an employer must give an employee and how many months' wages must be paid to laid-off workers, as well as the complexity of the procedures required before layoff.
Earlier studies, both theoretical and empirical, have generally focused on unemployment trends, and they have reached a variety of conclusions about the role of such factors.
Higher unemployment benefits, for example, result in higher unemployment and lower job creation in most theoretical models of the labor market, and in practice are associated with higher unemployment. Trade union strength, measured by the proportion of workers covered by union contracts, leads to higher wages and higher unemployment--but less so when unions and firms coordinate their bargaining activities.
The role of taxation and the various kinds of employment protection legislation is less clear-cut, as witnessed by continuing sharp public debate about them. Theory suggests that the effects of changes in taxation on unemployment depend largely on the extent to which higher taxes are shifted to labor in the form of lower compensation, together with how responsive the supply of labor is to changes in pay. For example, one study shows that tax cuts increase employment only if they raise the ratio of net wages to unemployment compensation.
A separate analysis argues that, in the long run, any tax on labor is borne by the employees, as seen by the absence, in practice, of any connection between total taxation and labor costs. Other research, by contrast, argues from both theory and labor market experience that higher taxes raise unemployment and lower output. This study compiles evidence that European unions have had the power to shift part of the increases in the tax burden onto firms. Unemployment has also been seen to rise with total taxation, but not with higher payroll taxes.
Most theoretical studies predict that high dismissal costs will not affect unemployment. The logic here is that since employment protection increases the cost to a company or an industry of making changes in its labor force, both hiring and firing will be lower, but the net effect on average employment will be ambiguous. That view is supported by the small flows into and out of unemployment in European countries compared with those observed in North America. However, another view is that increases in dismissal costs may lead entrepreneurs, over a decade or two, to substitute capital for labor, which is consistent with the decline in the labor share of income in the major advanced countries.
A focus on growth in employment over the medium term confirms several empirical relationships identified by earlier studies of unemployment. The IMF research finds less job creation where unions are relatively powerful, as well as a close link between working-age population growth and job creation. By contrast, country differences in unemployment benefits seem to have negligible effect on their job creation.
More suggestive from a policy perspective, extensive employment protection appears to dampen job creation, and so does a higher level of overall taxation. Further tests to try to quantify these relationships confirmed a stronger association between employment protection legislation and job creation, but the relationship with total taxation is also fairly robust.
Specifically, measurements for this study suggest that, on average, if we drop a country down by five positions in the ranking of all countries in the strength of their employment protection legislation, its average job creation rises by 0.1–0.2 percentage point. For Italy, for example, this amounts to some 20,000–40,000 jobs a year, or some 400,000–800,000 jobs over 20 years. A cut in total taxation by 1 percentage point of output is associated with higher average job creation of some 0.05 percentage point.
How strong are these conclusions? The study's statistical results must be interpreted with caution, especially given the small number of observations and the high degree of association among some of the policy variables. Trying to identify cause and effect is particularly hazardous, yet the results fit better with some earlier findings than with others. For example, a finding that high dismissal costs are associated with weak job creation appears consistent with the idea that they also lead employers to substitute capital for labor.
The findings might, in principle, also be consistent with the traditional view that employment protection reduces both hiring in upswings and dismissals in downswings, with no net impact on employment. This interpretation would be more convincing if 1980–97 could be seen as a cyclical upswing, with employment increases smaller in countries with high dismissal costs, as in Continental Europe, and larger in countries such as the United States with low dismissal costs. Both Europe and the United States have, however, experienced cyclical ups and downs over the last two decades, so it is harder to accept that higher dismissal costs had no impact on employment.
The relationship found between the overall taxation and employment growth is consistent with the view that, as taxes rose in Europe, more and more of the additional burden was shifted onto employers, who reduced payrolls as a result. In this instance, however, researchers had expected higher payroll taxes to hold down employment more than the overall tax burden. Yet the IMF study found no statistical support for this expectation.
Higher taxes and dismissal costs may help explain differences in job creation between high-performing non-European countries like the United States and Australia and most of Continental Europe, but the wide variations within Europe remain unexplained. For insights into Europe, the study looked at differences among 11 European Union countries over 1983–97. It examined the composition of job creation by type of contract (part-time versus full-time, and temporary versus permanent), the broad economic sectors in which jobs were created, and the age and gender of the newly hired. It also looked at interactions among these factors, such as the extent to which full-time jobs were filled by young men. The most striking finding is that the best European performer, the Netherlands, has had about half of its job creation since the mid-1980s in part-time jobs taken by women aged 25–49, typically in the service sector.
The study found that the kinds of jobs created are mainly determined by developments in technology or labor supply. In virtually all European Union countries, employment growth was much faster for women than men, mirroring much higher increases in labor force participation by women. Declines in youth employment seem to be related partly to more years spent in schooling, as well as to labor market conditions such as high firing costs.
While all countries increased employment for workers aged 25–49, their performance was mixed for the 50–64 age group, partly reflecting the trend toward early retirement in a number of countries. Economic sector changes also affected employment by age and gender, notably generating rapid employment growth for women aged 25–49 as all countries' service sectors expanded.
Contracts and Job Creation
A key policy question is whether increases in the share of part-time jobs are associated with higher overall job creation. The Netherlands clearly stands out, with half its employment creation attributable to part-time contracts. Reforms undertaken by the Netherlands in the early 1980s seem to have spurred overall employment through a sharp rise in part-time work. At the same time, for the 11 European countries as a whole, the study found no clear evidence that a bigger share of part-time jobs led to higher total job creation, either for Europe as a whole or for any of the three broad economic sectors: agriculture, industry, and services.
In an effort to estimate the extent that increases in part-time jobs are associated with the loss of full-time jobs, the study by IMF staff compared overall employment growth in Europe with the increase in the share of part-time jobs.
What, the study asked, is the employment creation associated with a country's addition of 100 part-time jobs? It looked at three benchmarks. First, if overall employment also rose by 100 jobs, there was no crowding out of full-time jobs at all. Second, if overall employment rose by 50 jobs, there was no net gain or loss of total hours worked--given that the average weekly hours of part-time jobs is about half that of full-time jobs. Third, if overall employment remained unchanged, the addition of 100 part-time jobs completely crowded out full-time jobs.
The results of this approach suggest that increases in part-time employment have typically been associated with some crowding out of full-time jobs, yet also with some increase in the total number of jobs.
Looking at part-time job creation by sector in individual countries--for example, at France alone by industry, agriculture, and services--provides a richer set of data, letting us estimate the relationship between increases in part-time jobs in a given sector to total employment, and the contribution of that sector to overall job creation in the country. This highlights the extent to which technological considerations may determine how much part-time jobs substitute for full-time jobs.
In the services sector, the study found that increases in part-time employment were associated both with increases in the overall number of jobs and also with partial crowding out of full-time jobs. The results strongly suggested the possibility that there had been no net increase in the number of hours worked, but clearly, more investigation is warranted here.
Turning from part-time to temporary work, Spain stands out: its net job creation over the past two decades was entirely in temporary jobs. Here again, reforms in the early 1980s, allowing temporary employment against a background of extremely high dismissal costs, seem to have sparked the dramatic rise in temporary jobs, while overall employment grew very slowly. Spain's share of temporary employment now stands at one-third, by far the highest among advanced countries.
Given that Spain and Italy have the highest dismissal costs among advanced countries, it is tempting to surmise that countries with high dismissal costs have more temporary employment or that they have a rising share of temporary jobs. However, statistical evidence for this idea is not strong, especially when Spain is dropped from the sample.
The strikingly different experiences of the Netherlands and Spain suggest that part-time contracts may be a more promising avenue of job creation than temporary contracts--a conclusion strengthened by evidence that workers tend to be happier with part-time contracts than with temporary contracts.
Surveys in the European Union in 1997 found that about 58 percent of part-time workers did not want a full-time job instead, and in the Netherlands, 72 percent of part-time workers did not want a full-time job instead. By contrast, only 7 percent of temporary workers in Europe said they did not want a permanent job instead of temporary work, while in Spain, 87 percent of temporary workers were unable to find permanent jobs and the share who did not want a permanent job was negligible.
Some of the main findings from this research are: