Estimating Frisch elasticity
Swiss Gruyere-factory worker takes home some extra cheese.
R.Babakin
It’s a central question in economics, but researchers have long struggled to pin down exactly how much people adjust the amount they work when their earnings change.
Standard theory says that if wages go up, people will work more. On the other hand, employees can’t always adjust their hours or pick up side jobs very easily.
In a paper in the American Economic Review, authors Isabel Z. Martínez, Emmanuel Saez, and Michael Siegenthaler broke new ground estimating how sensitive workers actually are to temporary income changes.
The authors studied an unusual feature of Switzerland’s push to modernize their tax system.
In the late 1990s and early 2000s, the Swiss government exempted income tax during special transition periods. These tax holidays were staggered across the subnational states, called cantons, over three different periods.
Figure 5 shows how employment and earnings changed from 1990 to 2010 for each tax-holiday period.
Figure 5 from Martínez et al. (2021)
Panel A indicates the share of wage earners in the population, and it shows very little change during tax-exempt periods. For example, the solid line with brown squares represents the share of workers in the three cantons with a tax holiday from 2001 to 2002. There was almost no increase in employment over that holiday, similar to other cantons in 2001–2002.
When it came to incomes, the picture was a little murkier. Panel B shows the earnings per person in thousands of Swiss francs. There is a slight uptick in earnings among the cantons that transitioned early (both blue lines with circles) and transitioned late (brown line with square) compared to other areas. But in the middle period, cantons that had a tax holiday (both the green lines with triangles) performed the same as cantons that paid taxes.
The authors used several methods for estimating the responsiveness of workers to these tax changes—what economists call the Frisch elasticity—but found no significant effect on employment and only a small effect on incomes.
The finding that the Frisch elasticity is lower than expected could have important implications for how economists understand fluctuations in the overall economy.