S1E2: "The number game"

Discounting struggles in the 1950s and 1960s

This post is part of a series on the history of how economists model the future with the Ramsey formula, based on joint work with Pedro Garcia Duarte. Episode 1 here. Full Paper here

Before delving into the economists who pioneered the use of the Ramsey formula for discounting, let me describe discounting debates in postwar US – the epicenter of this development. This overview draws from the literature on the history of cost-benefit analysis, where the bulk of discounting debates took place. There were exceptions, of course. Theoretical models of exhaustible resources, for instance, saw significant activity in the early 20th century, with the importance of discount rates recognized (as evidenced by Hotelling's work). Otherwise, discounting gradually emerged as a separate concern as cost-benefit analysis spread from water resource management into other areas of public administration.

Early US government guidelines, such as the 1950 Green Book, offered minimal guidance on rate selection. Discount rates incorporate time and risk elements, they read, and should consider “competing demands that exist for limited supplies of savings available for capital investments.” Soon, picking a discount rate became what economist Warren Gram describes in 1963 as a “number game,” whose political implications were well understood (so much so that Wisconsin Senator William Proxmire, which the New York Times later described as “the longtime gadfly of the United States Senate who thrived on exposing frivolous federal spending,” organized a series of congressional hearing on discounting in 1967. He criticized it as part of governmental agencies’ “hocus pocus”). Pick 4% and that new dam, lake, highway, classroom hire, laser, building would likely be funded. Choose 12% and it would be rejected. It was a game played by not just by practioners, but by entire institutions.

More economists entered the fray. But they did not just play the number game. They hunted for underlying economic rationales for these numbers. Discounting debates expanded from paragraphs to book chapters, dedicated academic articles, and special issues. However, this didn't lead to consensus. After over a decade of work, Harold Somers opened a review article with the stark observation that "the search for a single discount rate in the evaluation of government projects has failed." What those US economists disagreed on, essentially, was whose discount rate should be used. Who should value tomorrow’s flows in cost-benefit analysis?

Investors, RAND economists such as Jack Hirschleifer argued. Complaining about “overinvestment” in water infrastructure, they advocated for a 10% rate, one that should reflect a “fundamental postulate of neutrality…between publicly and privately owned enterprises.” They emphasized the opportunity cost of private investments, as reflected in capital market rates.

But it was the late 1950s, a time when MIT economist Francis Bator coined the term “market failures” to encapsulate a growing body of work in taxation, spending public finance, externalities and public goods. 1n increasing number of scholars, including Harvard PhDs Otto Eckstein and John Krutilla working for the Harvard Water Project and Resources for the Future emphasized that such market imperfections prevented an efficient assessment of risks by lenders. They added that since public investment was funded by taxes -which further deviated from the marginal rate of return in the private sector- the relevant question was “how much consumption is divested by the public investment.” The discount rate should be that of consumers, not investors.

The emphasis on consumers sparked another debate: should economists model the discount rate as a consumer’s time preference, that is, draw on the intertemporal decision toolbox developed by Fisher or Samuelson? Or should it represent a collection of consumers? Or even society at large? Represented by a government, a planner? In the latter case, they sought a "social cost of public capital" (Eckstein and Krutilla) or "social discount rate." Attaching the term "social" to "discount rate" compelled scholars to consider what "society" represents and to what extent a collective body should be modeled differently from an individual. This didn't bring further consensus.

Welfare economists were the most divided. William Baumol acknowledged the social nature of the discount rate but argued it should primarily be interpreted as the social opportunity cost of capital paid by the government. He was particularly concerned with whether the discount rate should include a risk premium. Conversely, Amartya Sen emphasized that market and collective decisions were fundamentally different "in nature," advocating for a wider gap between market and social discount rates. Stephen Marglin, who had worked on the Harvard Water program before contributing to the OMB's cost-benefit guidelines revision, agreed that the social discount rate should reflect ethical preferences and the government’s "stewardship" of present and future generations. This perspective pointed toward a lower discount rate (2-3%).

By the mid-1960s, yet another Harvard applied microeconomist, Martin Feldstein, thus summarized the state of the debate:

Economists had thus moved beyond the institutional "number game" in an attempt to instill more economic rationale. However, rifts quickly appeared between those who viewed the discount rate as a market rate, those who aligned it with individual behavior theory, and those who saw it as reflecting some sort of social entity. Nor did they agree on whether it should reflect the effect of public investment on private investment or consumption, or how to account for risk.

Some of these debates found an echo in a distinct community, one concerned with building optimal growth theory models.

Next (S1E3). “They do it in completely arbitrary ways”: discounting in optimal growth theory