Given the interest and policy relevance (as well as Miles Kimball’s immigration tweet day), I thought I’d write a post on the theory and empirics of the effects of immigration in the labor market.
A simple starting point for thinking about the labor market effects of immigration begins with the supply and demand for labor.
An increase in the number of immigrants increases aggregate labor supply. You can see this in the chart as a rightward shift in labor supply (from S to S’) since the number of people who are willing to work for a given wage increases for all wages. Based partially on this simple framework, some believe that immigration decreases wages. Prominent economists in the immigration literature such as George Borjas espouse this “labor demand is downward sloping” view of immigration and argue that increased immigration adds to the supply of workers and depresses the wages of natives (say from W0 to W’).
However, increased immigration does more than shift supply right, it also increases demand. David Card is among the most prominent economists who have emphasized the idea that “demand also shifts out.” This view of immigration holds that the effect of immigration on natives’ wages is ambiguous – it depends on the relative size of rightward shifts in supply and demand. Here are a few reasons why demand shifts out:
- Consumption: Immigrants consume goods and services and increase demand (and thus labor demand) for those products.
- Long-run K/L Ratio: Common production functions predict a constant capital to labor ratio, which predicts that higher L will eventually mean more investment and thus more labor demand (since the value of each worker is larger when more machines are around).
- Trade enables local production (and local labor demand) to exceed local consumption.
- Productivity: Immigrants could help increase productivity, which can do many great things for the economy. Increases in productivity will not only increase labor demand directly, but it will also spur more economic activity through the first channel mentioned above. In a recent paper that I posted, Peri et al roughly estimate that an influx of foreign STEM workers can explain a quarter of productivity growth in the 1990-2000 decades. So this productivity channel can be huge, especially for high skilled immigrants. In addition, having more domestic customers increases the size of product markets and may encourage new investments and innovation as potential profits and “addressable markets” expand in size.
So where does that leave us? Immigration causes outward shifts in both supply and demand – knowing which shift is bigger and hence whether wages go up or down is an empirical matter.
Taking this simple model to the data, where we would try to estimate how inflows of immigrants affect wage changes, requires thinking hard about a few issues:
- How do we define the labor market? Should we have separate figures for low and high skilled groups? How do these markets complement one another? Are low skilled natives and low skilled immigrants full substitutes? What about different geographies? These issues matter for interpreting how big the increase in supply is since adding 1 person to a 100 person market is much more important than adding 1 person to a million person market.
- What would have happened to wages otherwise? If immigrants come to places that have thriving labor markets, then observing little effect on natives wages may be misleading if natives wages would have increased 5% otherwise. We need good counterfactuals to have good estimates of the effects of immigration.
I don’t have time today to write a full review of this literature, but this nytimes magazine article has a nice account of some of the key papers and their estimates of the impacts of immigration on wages and employment. Finally, I should mention that Peri et al find that an influx of STEM workers of 1% of total employment raises the wages of college educated workers (both STEM and non-STEM) by 4-6%. An overly simplistic way to think about why STEM and college educated natives could be complements is that finance and legal work complement the high-tech industry. Finally, below find two sets of estimates (that reflect the Supply and Demand views respectively) from the Hamilton Project: