Schumpeterian Growth Theory and the Dynamics of Income Inequality
GAICIL Insights
More college graduates should have shrunk the college premium.1 They did the opposite, and that puzzle sits at the center of Philippe Aghion's 2002 Econometrica lecture on growth theory and the dynamics of income inequality. The lecture argues that growth theory in which growth is driven by a sequence of quality-improving innovations can shed light on two puzzles in rising wage inequality.
A hard question for anyone alarmed by early AI wage data: ask how much of the dispersion is people in motion. The lecture's framework suggests that freezing workers in place to reduce measured inequality could trade a temporary gap for a permanent one, or for unemployment, as the French case hints.
Puzzle one is between educational groups. Wage inequality between them rose substantially in the United States and the United Kingdom over the two decades before the lecture, despite a sharp increase in the supply of educated labor. A larger supply would ordinarily cut the premium. Its persistence and growth suggest technological change raised the demand for skills faster than supply could follow. Demand outran the classroom.
Puzzle two is within educational groups. This accounts for a large fraction of the observed increase in wage inequality, yet it mainly affected the temporary component of income. Workers with similar education saw increasingly different wages, but the differences were largely transitory, reflecting mobility across jobs and sectors during technological change. Two graduates with the same degree could be in very different places for a while.
Innovation-driven growth theory explains both. A new technological paradigm raises demand for workers able to use it and increases turnover as workers move across jobs and sectors. These forces act on between-group and within-group inequality together.
The research agenda is where the lecture gets bold. During the transition to a new paradigm, growth-enhancing education or training policies can raise the mobility, and therefore the within-group inequality, of workers across jobs and sectors, while still producing a more egalitarian distribution of permanent incomes. That raises the question of how to handle the insurance concerns created by greater variability in lifetime incomes. Short-run churn can buy long-run equality.
Countries diverge. Germany and France did not experience income widening of the same magnitude as the United States and the United Kingdom. In France, high legal bounds on the cost of unskilled labor may explain why skill-biased technical change showed up as higher unemployment in place of a higher college premium. Institutions decide whether technological change lands as a wage gap or a jobs gap.
Now the AI angle. Artificial intelligence is a new technological paradigm of exactly this kind. It may raise demand for workers able to use it, potentially lifting the skill premium, and increase mobility as workers move across jobs and sectors. Both kinds of inequality may move.
The temporary versus permanent distinction is the key to reading AI's early data. Within-group inequality driven by transition mobility may be transitory. Education and training that support mobility can help workers reach new AI-created opportunities, leading to more equal permanent incomes even if short-run variability rises. A snapshot of wage dispersion in the middle of the transition may badly misjudge where it ends.
Insurance is the hinge. Greater variability in lifetime incomes during transitions creates risk for workers. Social insurance and transition support can address it while letting mobility proceed. Designing those policies is a central challenge of an AI transition. Workers will move more readily if a bad year does not become a ruined decade.
The cross-country lesson cuts both ways. Labor market institutions determine whether change appears as higher wage inequality or higher unemployment. As AI spreads, each country's institutions will shape its effects on workers, and combining flexibility with support may avoid both rising inequality and persistent unemployment. Rigidity moves the shock to a different address.
Introduction
In this lecture, it is argued that Schumpeterian Growth Theory, in which growth is driven by a sequence of quality-improving innovations, can shed light on two important puzzles raised by the recent evolution of wage inequality in developed economies. The first puzzle concerns wage inequality between educational groups, which has substantially risen in the US and the UK during the past two decades following a sharp increase in the supply of educated labor. The second puzzle concerns wage inequality within educational groups, which accounts for a large fraction of the observed increase in wage inequality, although in contrast to between-group wage inequality it has mainly affected the temporary component of income.
Takeaway
In this lecture we have argued that Schumpeterian Growth Theory can provide plausible explanations to outstanding puzzles raised by the observed increases in between-group and within-group wage inequality in developed countries over the past thirty years. We now want to conclude brie‡y by suggesting three potential avenues for further research. First, our discussion of between- versus within-group wage inequality, has potentially important implications for labor market policy during the transition to a new technological paradigm. For example, our analysis suggests that growth-enhancing policies (in education or training) can be designed which increase the mobility -and therefore within-group inequality- of workers across jobs and sectors during the transition , but still lead to a more egalitarian distribution of permanent incomes. Now, how can we deal with the insurance concerns raised by the increased variability of life-time individual incomes? Second, whilst trying to explain the sharp increases in wage inequality in countries like the US and the UK, we have left aside the question as to why countries like Germany or France have not experienced income widening of the same magnitude. In France, the existence of a high legal bounds on the cost of unskilled labor, may explain why skill-biased technical change has resulted in an increase in unemployment instead of an increase in the college premium. In Germany, where unemployment has also sharply risen during the last decade (although largely as a consequence of the German reunification), we believe that the persistence of strong unions in manufacturing, may have affected the direction and speed of technical change (Acemoglu-AghionViolante (2000)).1
Finally, in this lecture we have concentrated our attention on the dynamics of wage incomes. But Schumpeterian growth theory can shed light on other aspects in the evolution of income inequality. For example, in AghionHowitt (1998; Ch. 3) we derive a simple relation between the aggregate growth rate and the distribution of relative productivity levels, and therefore the distribution of sectoral incomes.
Notes
- Philippe Aghion, Schumpeterian Growth Theory and the Dynamics of Income Inequality, 70 Econometrica 855 (2002), https://doi.org/10.1111/1468-0262.00312. Takeaway source: Concluding Comments, reproduced from a free author or repository copy. ↩
Cite as: Philippe Aghion, Schumpeterian Growth Theory and the Dynamics of Income Inequality, 70 Econometrica 855 (2002), https://doi.org/10.1111/1468-0262.00312
Header image: Qatar, Doha, Skyscrapers, Qatar.jpg by Vyacheslav Argenberg, CC BY 4.0, via Wikimedia Commons.
