firms and inequality
I agree with Adam Ozimek (aka Modeled Behavior) that the role of firms in rising income inequality is a Big Open Question. I've been lucky enough to see "Firming Up Inequality" by Song, Price, Guvenen, Bloom, and von Wachter presented twice in the past two years ... this research keeps getting more insightful and remains thought provoking.
Here's the abstract of the paper:
"Using a massive, new, matched employer-employee database that we construct for the United States, we show that the rise in earnings inequality between workers over the last three decades has primarily been a between-firm phenomenon. Over two-thirds of the increase in earnings inequality from 1981 to 2013 can be accounted for by the rising variance of earnings between firms and only one-third by the rising variance within firms. This rise in between-firm inequality was particularly strong in smaller and medium-sized firms (explaining 84% for firms with fewer than 10,000 employees). In contrast, in the very largest firms with 10,000+ employees, almost half of the increase in inequality took place within firms, driven by both declines in earnings for employees below the median and sharp rises for the top 50 or so best-paid employees. Finally, examining the mobility of employees across firms, we find that the increase in between-firm inequality has been driven by increased employee segregation—high- and low-paid employees are increasingly clustering in different firms."
One simple chart from the paper drives home their main result:
Variance Decomposition of Annual Earnings (for all firms)
Differences in individual income within a firm (the green line) have always been larger than differences in average income between firms (red line). BUT, the sizable increase in overall income inequality (the blue line) in recent decades has largely come from an increase in the differences between firms. Put simply, who you are has always been more important than where you work for explaining income inequality, but where you work is increasingly important. (It's more complicated than that but differences between firms have risen.)
In 60+ pages, the authors carefully go through a lot more empirical analysis ... thanks, in part, to a massive, high-quality data set ... to get at why this shift has occurred. Yet, it is the mark of a good paper that its results come through so clearly in the summary stats. Not surprisingly, it has already gotten plenty of attention in its earlier version (for example, WSJ-1, WSJ-2, and WSJ-3).
Still I think this research opens up a lot more questions ...
1. How much of the rise in the within-firm vs across-firm inequality is driven by changes over time in what "a firm" is?
Rising income inequality is not news ... but the emphasis on widening differences between firms, as in this paper, has been gaining traction. (See p 4-5 for related research.) For those more accustomed to thinking about inequality across individuals or households, I think it is important to reflect on what is a firm here. Technically, the unit of observation in their earnings data is still the individual, but the groupings by firm are key to the analysis. (And yes, groupings can matter, see also this FEDS Note on inequality using families versus tax units.) The "Firming Up Inequality" paper defines a firm by its employer identification number (EINs) -- an identifier used for corporate tax purposes. While EIN is a sensible definition of a firm, it is far from the only way to group workers and employers. Census uses a different system to classify firms but the firms counts are comparable (see appendix chart A.1-c). Still EINs may not conform to some popular notions of a firm, for example, as the authors point out NYSE listed firms have an average of 3.2 EINs each (pg 8). And much of the business dynamics literature focuses on the more dis-aggregate level of establishments. Something to keep in mind when comparing across other firm/business studies. (A recent example of the latter is this by Decker et al.)
More broadly, what a firm is goes well beyond how you measure it. This analysis takes the set of firms as given, yet firm structure is a choice and one that is likely to evolve with changes in tax codes, regulations, and business practices. Also their results by firm size leads to even more questions about how firm structure is chosen. Section 5.1 which discusses outsourcing is a great start to this inquiry but more can be done. While the authors worry in the conclusion that "increased segregation itself may obscure the underlying inequality" because workers have not seen themselves fall further behind their coworkers ... this may be a feature not a bug. Suppose skill-biased technical change or some other shift in the marginal product across workers meant that overall income inequality was going to rise, then wage segregation between firms might be the least harmful way in terms of worker well being to structure it. (See this new NBER working paper with experimental evidence on the harmful effects of within-firm pay disparities ... probably not news to HR managers.) I am not disputing the potential costs of increased wage segregation but this paper does not really tell us why it is happening.
2. How should we think about the increase in worker segregation and does this paper tell us anything about the decrease in labor market fluidity?
To start more simply, what does "worker segregation" here even mean? The authors argue (with the help of an econometric model) that the rise in between-firm inequality has been driven by a shift in the mix of workers at firms (those who would earn a lot anywhere are now more likely to work together at top firms than in the past) as opposed to a shift in the wage premium from working at the very best firms. The key to separating out these two hypotheses is to watch what happens to the income of workers who switch firms. Section 4 of the paper is a bit of a slog unless you like acronyms, can look past some typos, and aren't too worried about the rise of purple squirrels (workers who are a perfect match for a firm). Still the results seem plausible. I do wonder about the firm movers in their rolling regressions, particularly with the overall rate of job switching falling over the period. Molloy and coauthors have done some interesting work on the aggregate decline in geographic mobility and labor market fluidity. Their work doesn't identify a clear cause though some broad-based decline in the net benefits of moving or switching jobs appears to have occurred. I wonder how the "Firming Up Inequality" paper with its greater worker segregation and lower benefits of working at large firms may relate to reduced fluidity.
Several results in the paper differ for small and large firms, but the one that caught my eye was the declining premium for most workers from working at very large firms.
Change in the Distribution of Annual Earnings
At firms with more than 10,000 workers (which employ about 30 percent of all workers), the real income of the median worker has fallen since 1981,and even the gains at the 90th percentile are fairly meager (when compared to smaller firms, see Figure 5a). Some of this is an equalization between small and large firms. In the 1980s the within-firm inequality at large firms was slightly less than at small firms ... and those lower down in the income distribution got paid relatively more by working at a large firm. Over time that large-firm premium has eroded for most workers... well except if you happen to be a tippy top earner, then it has gotten bigger. But the high fliers (or the 0.1%) are only about 70,000 people so I don't expect them to affect the aggregate trends in mobility much. But maybe the reduced benefits for many from working for one of the really big (and less prevalent) firms might have lowered the benefits of moving/job switching?
3. Is rising worker segregation a sign of reduced competition, greater
economic rents, or is it telling us about a change in the nature of
Even with their careful econometrics, I still think it is very hard to disentangle the between-firm effects ... or least how to interpret it. The authors give the rising dispersion in the profit rates between firms, as an example of what might be a firm-specific premium might look like (as opposed to segregation, see footnote 13). But I am not sure that's how I would interpret this chart from Furman and Orszag's work.
Clearly, this is a different way to cut firms but it is not hard to imagine that a greater concentration of highly-skilled workers at top firms might also push up the returns on invested capital. (Those computers don't write their own code yet.) Worker segregation is probably not a fluke. More importantly, it is hard to see how any of this work so far resolves the question of why. Is the rise in between-firm differences a sign of economic rents/reduced competition or is it an unavoidable outcome given the changes in production? Maybe agglomeration effects require more skilled workers in a firm? Ok, so this topic quickly gets out of my depth, but I find worth pondering. The OECD's recent survey on the U.S. economy has good discussion (starting on p 83) on the role of market power, including a box on industry-level price mark ups. Not sure how well the bits there map to the "Firming Up Inequality" paper, but these are more questions to pursue.
In summary, firms and inequality is a worthy addition to a Big Open Question list.
PS Photo credits go to my 6-year old son. He left his under-sea "firm" next to my desk at home ... seems those workers have more to worry about than income inequality.
**Opinions here are mine and should not to be attributed to anyone with whom I work.**