Cochrane (2013): Finance — Function Matters, Not Size

Summary

John Cochrane's "Finance: Function Matters, Not Size" (Journal of Economic Perspectives 27(2), 2013, pp. 29–50) is the direct companion piece — same issue, immediately following — to Greenwood & Scharfstein's "The Growth of Finance". Where Greenwood–Scharfstein document the scale (finance's GDP share tripling, 1950–2006) and flag that persistently high asset-management fees "generate economic rents," Cochrane's paper is the era's most-cited mainstream reply arguing that the scale facts do not establish rent extraction. It is the wiki's central citable steelman for the position that finance's growth reflects genuine intermediation services rather than rent capture, and the honest counterweight this domain's pages (the FIRE sector, finance-growth-is-land-credit) need to carry.

The Core Argument — Ask "Function," Not "Size"

Cochrane opens by mocking the rhetorical move of treating an industry's size as self-evidently wasteful — "the US economy spends $170 billion a year on advertising"; "wholesale and retail trade... cost 14.6 percent of GDP, while all manufacturing is only 11.5 percent" — before stating his method: economists should ask whether a market or government failure explains an apparently puzzling industry size, not assume waste from size alone. On finance specifically: "I don't claim to estimate the socially optimal 'size of finance' at, say, 8.267 percent of GDP. It's just the wrong question."[1] He proposes instead to "use size as an organizing principle for studying function and dysfunction" — auditing each channel of finance's growth for whether it reflects genuine value or an identifiable distortion, rather than treating the aggregate total as the object of concern.[1]

Reading Greenwood–Scharfstein's Facts as Demand, Not Rent

Cochrane's baseline account of why finance grew directly engages the paper it follows: "Greenwood and Scharfstein nicely review the key facts... Their most basic story is: quantity increased a lot, but prices didn't fall." His reinterpretation: "This description suggests a simple economic interpretation: The demand for financial services shifted out. People with scarce skills supplying such services made a lot of money."[1] On this reading, rising house and equity prices increased the base against which proportional fees (mortgage origination, asset-management fees) are charged — a demand-side, not a rent-extraction, story. He also notes finance's size is not one-directional: "Demand that shifts out can shift back again... [it] evaporated with the decline in housing and asset values in the 2008 recession," with employment in finance falling back to roughly its 2000 level by the time he wrote — evidence, he argues, against a simple rising-rent-share narrative.[1]

The Active-Management-Fee Puzzle: The Berk–Green Model

The paper's most developed case study is mutual-fund and hedge-fund fees, the component of "the size of finance" easiest to measure directly. Cochrane reports that fee rates fell over the period finance's revenue rose — "the average actively managed equity mutual fund fee fell from 2.19 percent in 1980 to 1 percent in 2007" (citing French 2008), and average bond-fund fees fell from 2.04 to 0.75 percent (citing Greenwood & Scharfstein's own data) — even as total fee revenue rose, because assets under management rose faster than fees fell.[1] That is a direct, if partial, tension with Philippon's finding that the economy-wide unit cost of intermediation stayed flat: Cochrane's component-level fee schedules did fall, but the composition of assets shifted toward costlier, higher-fee vehicles (hedge funds, private equity — "managers charge 1.5–2.5 percent of assets each year, and also 15–25 percent of profits"), which can offset a falling price with a rising mix — precisely the kind of composition effect Philippon's own quality-adjustment tries to net out.[1]

To explain why sophisticated investors keep paying high fees for active management despite decades of evidence that average fund alpha is roughly zero, Cochrane leans on Berk & Green's (2004) supply-and-demand model: a manager with scarce skill attracts assets until inflows drive the manager's edge to its scale limit, at which point "the manager still generates $1 million alpha, but now he collects $1 million in fees. His investors get exactly zero alpha, the competitive rate of return. Everyone is acting rationally."[1] Cochrane's point is that this is a competitive equilibrium, not a market failure — fees rise to fully capture the scarce manager's value-added, leaving investors with only the ordinary market return. The wiki notes, without adopting Cochrane's framing, that this is structurally close to a description of rent flowing to a scarce factor (managerial skill) rather than being competed away to customers — the same quasi-rent-to-scarce-talent pattern EDITORIAL's rent gradient flags as the Schumpeterian, harder-to-tax frontier case. Cochrane and a rent-extraction reading can agree on the mechanism and disagree only about whether "rent to scarce skill in a competitive market for that skill" counts as the kind of rent geoist policy should target.

What Cochrane Concedes

The paper is not an unqualified defense of finance's size, and the wiki records the concessions as part of the steelman:

• Regulation, subsidy, and capture, not size, are "the more fertile fishing ground." Cochrane's own framing: "The run-prone nature of the US financial system, together with its massive regulation, subsidies, government guarantees, and regulatory capture, looks to be a more fertile fishing ground for trying to understand market and government failures than does mere size."[1]
• Some financial engineering was regulatory arbitrage, not value creation. Discussing the mortgage-backed-securities boom: some innovation — "creating off-balance-sheet, special-purpose vehicles and tailoring securities in order to game credit ratings" — "counts as a regulatory failure needing reform, rather than a market failure needing additional regulation," but he does not dispute it was wasteful in itself.[1]
• High-frequency trading is the one activity he treats as a genuinely open question, leaning skeptical. "It's especially hard to see why high-frequency trading is needed. Price discovery every millisecond doesn't seem necessary to guide corporate investment or individual risk sharing and hedging."[1] Reviewing the 2010 "Flash Crash" and 2012 stock-price "sawtooth" episodes, he concludes "these palpable inefficiencies suggest a market with very little 'liquidity provision,' not the opposite" — a rare point where Cochrane's own evidence reading cuts against the "size reflects genuine service" thesis he otherwise defends.[1]
• An older literature already found finance modestly oversized. He cites Lucas (2000), who "concluded that finance was about 1 percent of GDP too big" using the welfare cost of cash-management activity — a small but real concession that some resources devoted to finance are excess, on a measure Cochrane treats as credible.[1]

Why It's the Steelman, and What It Doesn't Refute

For this wiki's rent gradient, Cochrane is valuable precisely because he does not deny the scale facts — he shares them with Greenwood–Scharfstein and does not dispute Philippon's later unit-cost puzzle — but supplies the alternative causal story (demand shifted out; fees paid for scarce skill and real risk-bearing) that any rent-extraction reading of finance's growth has to rule out, not merely note evidence consistent with. He does not address land or mortgage collateral directly — his mortgage discussion treats loan-origination fee growth as an ordinary demand response and separates state-dependent refinancing fees from GDP, but never engages the land-vs-structure decomposition (Knoll–Schularick–Steger) that grounds the land-credit reading. His paper is about whether finance's size indicates waste, not about whether its collateral is land — the two literatures answer different questions and neither settles the other.

Honest Limits

• This is an analytical response essay, not new data. Cochrane presents no new dataset; he reinterprets Greenwood–Scharfstein's and others' numbers and leans on existing models (Berk–Green). Its evidentiary weight is argumentative, not empirical, in the way Philippon's or Jordà–Schularick–Taylor's is.
• It predates and does not engage Philippon (2015). The two papers are close in time (Cochrane 2013, Philippon 2015) and Cochrane cannot address Philippon's 140-year quality-adjusted unit-cost series directly; the wiki notes the tension (component fee rates fell; the aggregate unit cost measured by Philippon did not) without resolving it — this is an open empirical question in the literature itself.
• Institutional standpoint. Cochrane identifies himself, in the paper's own author note, as affiliated with the University of Chicago Booth School, the Hoover Institution, and as an "Adjunct Scholar" of the Cato Institute — market-liberal institutions with a general prior against regulation-based size critiques. This does not make the argument wrong, but the wiki flags it for the same source-diversity transparency it applies to advocacy sources on the pro-rent side (e.g., Hudson's institutional position is likewise noted on his pages).
• The Berk–Green concession noted above cuts both ways and is presented as the wiki's own observation, not Cochrane's — the model is consistent with either "fees reflect real value" (his reading) or "fees are rent captured by scarce skill in a market too concentrated for competition to erode it" (a reading he does not adopt but does not foreclose either).

See Also

• The FIRE Sector — where this paper is the wiki's strongest citable counter-view to the rent-extraction reading
• The growth of modern banking is largely mortgage credit against land — the problem page this paper's challenged_by entry disciplines
• Greenwood & Scharfstein: The Growth of Finance — the paper Cochrane replies to, in the same JEP issue
• Philippon (2015): the finance-efficiency puzzle — the pro-rent quantitative anchor this page is in honest tension with
• Borio: the financial cycle — the complementary steelman on the credit-cycle side (a different argument: booms are general monetary phenomena, not specifically land)
• Economic Rent · Rentier
• Geoism — the rent-domain program and its gradient

Sources

1. John H. Cochrane, "Finance: Function Matters, Not Size," Journal of Economic Perspectives 27(2), 2013, pp. 29–50 — used for the "function, not size" method; the "8.267 percent of GDP... wrong question" framing; the reinterpretation of Greenwood–Scharfstein's "quantity up, prices flat" as demand-shift; the mutual-fund and hedge-fund fee-rate figures; the Berk–Green (2004) active-management fee model and its "everyone is acting rationally" conclusion; the "more fertile fishing ground" concession on regulation/subsidy/capture; the regulatory-arbitrage characterization of some mortgage-backed-security innovation; the high-frequency-trading skepticism and Flash Crash discussion; and the Lucas (2000) "1 percent of GDP too big" citation (B- and D-claims; verified against the full PDF this session). Author PDF · AEA (open access)