Philippon (2015): Has the US Finance Industry Become Less Efficient?

Summary

Thomas Philippon's "Has the U.S. Finance Industry Become Less Efficient? On the Theory and Measurement of Financial Intermediation" (American Economic Review, 2015) measures the unit cost of financial intermediation — what it costs the economy to move a dollar from savers to borrowers and keep it invested. He defines the user cost of finance as r + ψ (the return paid to savers plus the intermediation cost ψ), and measures ψ as finance income divided by the stock of assets intermediated.[1]

The central finding is a puzzle. Across ~140 years (1886–2012) the unit cost of intermediation "has an annual cost of 1.5% to 2% of intermediated assets" and is "remarkably constant" — even after Philippon quality-adjusts for the changing characteristics of borrowers (Figure 3, raw vs. quality-adjusted).[1] Finance's share of GDP rises and falls (high in the 1920s, low in the 1960s, high again after 1980), but that is explained mostly by the quantity of assets intermediated, not by a falling price of intermediation.[1]

Why It's a Puzzle — and Why It Matters Here

In every other information-intensive industry, computerization drove the cost of handling information sharply down. In finance it did not: "improvements in information technologies do not appear to have led to a significant decrease in the unit cost of intermediation," and, Philippon notes, "explaining this puzzle is an active area of research."[1] Something absorbed the efficiency gains that IT should have delivered.

For this wiki's rent gradient, that is the crux of the finance-rents question. A stable-or-rising unit cost despite a technological revolution is consistent with finance capturing its productivity gains as economic rent — extracted through information asymmetries, market power, regulatory privilege, and the partly zero-sum growth of trading — rather than competing them away to customers. It is the strongest single quantitative anchor for the claim that a large slice of FIRE-sector income is rent rather than the price of a competitively-supplied service.

Honest Limits

The finding is a puzzle, not a proof of rent. A constant measured unit cost is consistent with rent capture, but also with rising unmeasured quality of intermediation, with the growth of genuinely valuable-but-costly new activities, or with measurement problems in dividing "finance income" by "assets intermediated." Philippon himself frames it as an open question, and the magnitude and interpretation are debated in follow-up work (e.g., Bazot's European estimates and Philippon's own later papers). The page carries it as powerful, disciplining evidence that finance did not pass on its efficiency gains — which any account of finance as productive intermediation must explain — not as a settled measurement of the rent share.

See Also

• The FIRE Sector — the finance-insurance-real-estate rent question
• Rentier · Economic Rent
• The Rentier Economy (narrative)
• Corporate profits increasingly reflect economic rents — the parallel non-financial case
• Geoism — the rent-domain program and its gradient

Sources

1. Thomas Philippon (2015), "Has the U.S. Finance Industry Become Less Efficient? On the Theory and Measurement of Financial Intermediation," American Economic Review 105(4), 1408–1438 — used for the ~1.5–2% unit cost of intermediation, its ~140-year stability (raw and quality-adjusted), and the "no IT-driven cost decrease" puzzle (B-claims; verified against the paper this session). Author PDF · AEA