How much do workers' outside options and employer concentration affect wages?
Core argument: A $1 increase in the value of workers' outside options is associated with $0.24–$0.37 higher wages in imperfectly competitive labor markets.
In imperfectly competitive labor markets, the value of workers' outside options significantly impacts wages. A $1 increase in the value of outside options results in $0.24-$0.37 higher wages. Employer concentration, moving from the 75th to the 95th percentile, reduces wages by 5%, explaining 21% of wage variation within occupations across cities. Additionally, outside-occupation options account for 13% of wage variation. The interaction between employer concentration and outside options is crucial; the effect of concentration on wages is three times higher for occupations with low outward mobility. This highlights the importance of considering outside-occupation options in understanding wage dynamics and suggests that higher-skilled workers are attracted to cities with deeper labor markets due to better outside options.
Ed Comment:What companies/industries represent the 95th percentile of concentration? Are the skills in those industries truly unique to those companies? Why don’t worker just use their skills in other industries? Ie why doe industry concentration matter?
Steve Comment:Sadly they don’t give data on local area concentration by sector/industry in the paper So there concentration measure is by local labor market (so in an extreme example, if you were a male living in Aspen in 1900 you could either work at the mine or leave, few other options, and those wages would likely be set relative to what the mining company would pay) Here is the example she uses: “…Our instrument exploits the fact that (a) increases in local labor market concentration are often driven by specific, already-large, firms growing, (b) these large firms usually operate across many labor markets, (c) that different local labor markets are differentially exposed to the growth of these large national firms based on the initial share of employment with that firm, and (d) that the employment growth of these large firms nationally is likely to be orthogonal to productivity changes or other conditions in a specific local labor market. The intuition behind the instrument can be explained with a hypothetical example of baristas in Portland, OR, and in Hartford, CT. Imagine there are two large coffee chains in each city: Starbucks, and Dunkin’. In Portland, Starbucks is relatively more dominant than Dunkin’, and the reverse in Hartford, meaning that Starbucks’ local employment share is much larger in Portland than in Hartford. In years where Starbucks grows substantially faster than Dunkin’ nationwide, employer concentration of baristas will grow by more in Portland than in Hartford…” They are using a Herfindahl-Hirschman Index by metropolitan areas using the Burning Glass database of online job postings (this is at the bottom of page 2 top of 3) Workers with other skills who can find work outside of the concentrated industry do have higher wages, it just seems by there measures most workers don’t have the skill or desire to move to new occupations (Figure 11)
Ed Comment:Sure sems weird. I get how an isolated factory or coal mine in a rural area might enjoy lower wages than the national average for their skill level, but 1) that seems rare and 2) exponential not linear, 3) the increase demand it creates surely raises the wages in the town from what they would have been. Baristas in urban coffee shops seems unfathomable unless the workers are getting nonpecuniary benefits from working somewhere—job security or growth/development opportunities with faster growing competitors, for example. Regardless, how does she reconcile it to the other paper that says it’s not the case?
Steve Comment:I’m working on an answer to that question, the answer likely is the monopsony factor (so firm power) is a factor but just a very small factor versus the “worker power” factor that she wrote about with Summers, and the later (worker power, inability to extract higher rents) is more convincing an explanation versus local industry concentration overall
Ed Comment:The last paper says monopsony is a significant fact. The prior paper said it was not. Monopsony would pocoure in isolated towns where employers have very high share. Loss of labor power should occer everywhere.
Steve Comment: Her answer: “…Dear Steve, Thank you for your email! Your question makes a lot of sense. The way to reconcile them is that one paper focuses on levels and the other focuses on changes. So: I would say that the Monopsony and Outside Options paper suggests that there is relatively high monopsony power impacting workers in specialized occupations and/or in low-population density places (our findings suggest employer concentration affects probably about a quarter of workers to some degree). The Declining Worker Power Hypothesis paper does not argue that there is no monopsony power, but argues that there is no evidence that monopsony power has increased over time. So, while monopsony power may be important for some workers, it cannot explain things like the decline in the aggregate labor share.I hope this answers your question but happy to discuss more if not! All the best, Anna…”
Ed Comment:Makes sense. no rise. Aren’t rural workers choosing monopsony? I’m always suspicious that their finding mean what I said they do: that rural factories or coal mines enjoy lower wages than the national average for the skill level of their workforce. And this provides evidence of monopsony, even though the increase demand for employment the employer creates raises wages in the town from what they would have been otherwise, the latter of which is never mentioned by an intentionally misleading author.
Steve Comment:You can add another layer on that in the rural factory towns in the Midwest/upper Midwest rust belt are depopulating. So the individuals that remain are very much choosing monopsony.