Population Change and Lifecourse

Research Brief #4 - February 2012

An International Comparison of Lifetime Inequality:
How Continental Europe Resembles North America

Summary

Is earnings inequality in North America as high as previous research has suggested? And how does North America compare to Europe? Previous studies on this topic have found a higher level of earnings inequality in North America than in Continental Europe. However, these studies have focused largely on earnings in a single year. In their forthcoming study on earnings inequality, authors Audra Bowlus and Jean-Marc Robin develop a new methodology for investigating and comparing earnings inequality in North America and Europe.

The methodology developed by Bowlus and Robin constructs a measure of lifetime earnings in order to compare lifetime earnings inequality across countries. By focusing on lifetime earnings, the authors find that earnings inequality is smaller in North America than single year earnings comparisons would suggest. In fact, due to cross country differences in how earnings change over an individual’s lifetime, there may be no difference in lifetime earnings inequality between North America and Europe.

Key Findings

Background

Researchers, policy makers, and the media have long been interested in issues of economic equality and fairness. In its Global Risks 2011 report, the World Economic Forum identified economic disparity, including income inequality within countries, as one of the greatest economic risks facing the world in the next decade.

In an attempt to quantify inequality within a country, many studies have relied on earnings data from a single year. However, if single-year earnings inequality differs from the long-run pattern, then incorrect conclusions may be drawn about the degree of inequality that actually exists and the policies that might reduce it.

Prior Research

While many researchers aim to investigate lifetime earnings inequality, such an investigation typically requires large datasets in which the same sample of individuals is tracked over a period of many years.

Additionally, in order to make meaningful cross-country comparisons, data for each country should represent the same time period. This adds to the difficulty in finding the necessary data to perform such an analysis. As a result, many researchers have instead relied on an analysis of current earnings inequality.

Even those studies that focus on lifetime earnings fail to account for important influences on an individual’s earnings.

First, they treat all variation in earnings as relating to the individual. However, some earnings variation comes as countries move through the business cycle (recessions and economic expansions). This is particularly problematic when trying to compare inequality across countries, as recessions and expansions do not always occur at the same time or have the same effects in each country.

Second, most research on lifetime earnings inequality still does not include employment risk in the analysis.

Key Contributions and Objectives

Data Sources

The study focuses on earnings patterns within five countries in the late 1990s. Panel datasets, which include data on the same sample of individuals collected periodically over several years, are required. The datasets used are as follows.

Analysis and Results

The authors estimate a statistical model of earnings and employment risk. Using this statistical model, they simulate lifetime earnings paths for each individual. This simulation incorporates country-specific characteristics such as differences in average retirement age and unemployment insurance systems.

Earnings of high earners (90th percentile) are divided by earnings of low earners (10th percentile) as a measure of earnings inequality. This analysis is conducted using current-year earnings and simulated lifetime earnings in order to analyze differences between the prior approach to measuring inequality and the new methodology.

Base-year earnings comparisons confirm previous findings that current earnings inequality is higher in North America than in Continental Europe, with the U.K. falling in between.

In 1998, high earners in the U.S. received 4.88 times the earnings of low earners. France showed a much lower current earnings inequality in that year, when high earners received only 2.55 times the earnings of low earners.

[These results are for males only. The same analysis was performed for females. Estimates of current earnings inequality were higher for females than for males, but cross-country rankings were similar.]

The lifetime inequality picture looks much different. When earnings mobility and employment risk are taken into account, all five countries show a similar level of lifetime earnings inequality.

The U.S. is still the most unequal, with high earners receiving 2.76 times the income of low earners. However, this is only slightly higher than Canada, France, and Germany, where high earners receive approximately 2.65 times the income of low earners.

[These results are based on a lower bound of the lifetime earnings inequality estimate. The upper bound estimates still show some cross-country differences in lifetime earnings inequality, but these differences remain smaller than those suggested by current earnings inequality comparisons.]

The results indicate that earnings mobility lowers lifetime earnings inequality in all countries, as expected. However, the inclusion of employment risk decreases lifetime earnings inequality in the U.S., Canada, and the U.K., but increases it in France and Germany. When the two components are combined, the five countries look quite similar.

Policy Implications

References

About the study

“An International Comparison of Lifetime Inequality: How Continental Europe Resembles North America” is forthcoming in the Journal of the European Economic Association. This paper was written by Audra Bowlus, Ph.D., University of Western Ontario, and Jean-Marc Robin, Ph.D., Sciences Po and University College London.

For more information, please contact Audra Bowlus.

The analysis was carried out at the University of Western Ontario Research Data Centre. The Research Data Centre program is part of an initiative by Statistics Canada, the Social Sciences and Humanities Research Council, the Canadian Institutes of Health Research and university consortia to strengthen Canada’s social research capacity.

This research brief was prepared by Emilie McHugh Rivers.