Go to the Globe and Mail homepage

Jump to main navigationJump to main content

(Dragan Grkic/Getty Images/iStockphoto)
(Dragan Grkic/Getty Images/iStockphoto)

Cross and Lee

Why the Gatsby curve is a poor measure of income mobility Add to ...

Miles Corak and Mark Stabile in a recent op-ed defended the relevance of the ‘Great Gatsby’ curve, which asserts that more income inequality in one generation leads to less income mobility for the next generation. The U.S. has more inequality and less mobility compared with Denmark, which has less inequality and more mobility. Plotting this trade-off across several nations yields the so-called Gatsby curve. However, there are problems with interpreting the differences across nations and the measure of mobility used in their analysis, which limits its policy implications.

More Related to this Story

Whether the negative relationship observed between inequality and mobility truly reflects an underlying dynamic about incomes and class has been questioned by Greg Mankiw, former head of the Council of Economic Advisors to the U.S. President. He argues that comparing a large diverse country like the U.S. with smaller, more homogeneous countries like Denmark is misleading. The U.S. is a vastly more diverse country in terms of geography, ethnicity, and race, and therefore would be expected to produce a wider range of outcomes than Denmark.

A more ‘apples to apples’ comparison is not the U.S. and Denmark, but Rhode Island and Denmark. In terms of diversity, the U.S. and the EU would be a more valid comparison, but still would not be ‘oranges to oranges’ because the EU does not have the same integrated labour market as the U.S. Unemployed workers can move from Ohio to North Dakota, but not from Greece to Germany, if only because of the language barrier.

The question is whether more inequality by itself triggers a mechanism that reduces mobility for the next generation, and why this mechanism would exist across societies with different institutions and demographics. The Gatsby curve invites policy makers to think that there is a trade-off between inequality and mobility – if we chose to pursue less inequality, increased mobility will automatically follow.

Does greater income inequality in countries like the U.S. by itself short-circuit mobility between income strata relative to countries like Denmark? A necessary corollary would be that more inequality in Denmark by itself would lead to markedly lower mobility. That seems an unlikely outcome, because of all the institutional factors that encourage mobility in Denmark. This is why carefully examining the Gatsby curve reveals a narrow range of inequality co-exists with a wide range of mobility outcomes, outside of the Nordic countries. The importance of institutional factors is why reducing inequality in the U.S. would not change mobility significantly without changes to other institutions such as health care, education, the criminal justice system and family structure (over half of poor children in the U.S. live in single parent families).

Without taking these other factors into account, the Gatsby curve begins to look more like a convenient narrative stringing together unrelated facts, posing as a meaningful insight into income and class dynamics. Tempting trade-off menus have been presented before to policy makers that over-simplified an issue, notably the Phillips Curve that alleged a trade-off between inflation and unemployment. This misled policy makers for years before it was revealed to be entirely an illusion.

A further limitation of the Gatsby curve is that it measures inter-generational mobility only between fathers and sons. This probably is why Toronto Centre Liberal candidate Chrystia Freeland made her much-criticized remark about how your job prospects increasingly depend on the job your father had, although the irony of her saying this in the presence of Justin Trudeau was too rich for the media to overlook (academic research and politics can be a toxic mix). Researchers have to limit the income data to fathers and sons in order to abstract from the impact of the marked improvement of labour market outcomes for women in recent decades. Doing so, however, exaggerates the importance of the long-run deterioration of incomes for young men while ignoring the gains made by women.

Using data only for men puts limitations on the policy implications of this research which are not fully appreciated. While restricting data to men is technically appropriate to preserve the sanctity of the assumptions underlying the equations, it means sacrificing any relevance to policy, which is the main conclusion of their oped. Policy makers live in the real world. If labour market outcomes for sons are deteriorating, but daughters are doing better, is there really a problem of inter-generational mobility? Perhaps policy should focus more on the growing number of the ‘disconnected’—young men neither in school nor in the labour force, as Peter Edelman called them in his recent book on poverty in the U.S.

So there is reason to hope, only not because of the hoped-for efficacy of government programs to help kids in low income homes, but because the problem may not be as great as suggested from comparing results only for men across countries with inherently different institutions and diverse populations.

Philip Cross is a Research Fellow at the Macdonald Laurier Institute and the former Chief Economic Analyst at Statistics Canada. Ian Lee is assistant professor, Sprott School of Business, Carleton University

Follow us on Twitter: @GlobeDebate

In the know

Most popular video »


More from The Globe and Mail

Most Popular Stories