Is product market concentration rising or falling?
Core argument: The share of industries classified as "highly concentrated" at the product level fell from 44.4% in 1994 to 36.6% in 2019, a decline observed across goods and services both nationally and locally across 29 state groups.
Since 1994, product market concentration has been decreasing, with the number of industries classified as "highly concentrated" dropping from 44.4% in 1994 to 36.6% in 2019. This trend is observed across a broad range of goods and services, both nationally and locally in 29 state groups. The decrease in product-level concentration contrasts with rising concentration at broader sector levels, suggesting firms are expanding into adjacent product markets within the same sector. Despite high concentration levels, there is no evidence of worsening market power over time. The median Herfindahl-Hirschman Index (HHI) for product markets is 2180, with 42.2% of industries having an HHI above 2500, indicating potential market power higher than previously thought. However, this finding contradicts popular perceptions driven by large firms and high-profile mergers.



















Steve Comment: had a few questions about your “Profit Puzzle” paper. I’m sitting here looking at Figure 14 and find the results really surprising. Could the divergence btw private and public firm profits (given your using return to capital) largely be a function of the lower capital the intensity of private service firms? I’m shocked at the public firm series. I would have thought that would have had an upward slope given US firms’ international profits. Are taxes skewing this (Apple booking stuff in Ireland, etc)?
James Traina Comment: Thank you for reading! Capital intensity and tax differences are good hypotheses here. For the former, could you expand on what you have in mind? e.g. Are you thinking about physical vs financial capital differences? For the latter, we show in the “Solving the Puzzle” section that public vs aggregate tax rate differences are there, but they’re small and actually pointing the other way — they’re higher for public firms. That also relates to the rise of S-corps, which folks have attributed to tax advantages. The international dimension is much harder because we don’t have good data on it. Basically, there’s still a mismatch when we make our comparisons because “domestic” in Compustat means US incorporation, while “domestic” in the IMAs means US operation. It’s hard to say which direction this would bias our results. One thing that I find helpful to think about, but we didn’t fit into the paper: You can find the same kinds of results in *all* the standard profits / capital measures, e.g. ROA, ROIC, etc. So any explanation would have to work for all these measures jointly.
Steve Comment: Yes I have in mind firms of engineers, architects, or lawyers that have little physical or financial capital, but a lot of human capital. Could those firms be driving the high ROI of private firms relative to public? I’m genuinely curious about this, because it feels like a failure of economic efficiency to have private firms yielding so much more than public firms.
James Traina Comment: Ah yes, that’s possible! You’d need an accounting mismeasurement, though, where it doesn’t show up in labor income. You might be interested in this paper: https://bfi.uchicago.edu/insight/research-summary/the-rise-of-pass-throughs-and-the-decline-of-the-labor-share/ Public firms’ returns on the book value of assets are down ~ 50% from 1980 and private firms’ returns have doubled. @EconTraina @ASollaci @CarterDavisFin (135)