Keynes, Minsky, and the Economics of Uncertainty

JOHN MAYNARD KEYNES

CAMBRIDGE—Consideration of economic and financial conditions today calls to mind the economic theorists who most influenced my own thinking, as an academic and as a venture capitalist: John Maynard Keynes and Hyman Minsky. Their relationship, though virtual, is essential. It was Minsky who revealed the profound conflict between “the economics of Keynes” and the “Keynesian economics” that dominated the teaching and practice of macroeconomics for at least a generation, and which remains embedded in contemporary “New Keynesian” models of the economy.

Minsky completed his doctorate at Harvard under the supervision of Joseph Schumpeter, whose concept of “creative destruction” illuminated how technological innovation drives economic transformation, and emerged as a “heterodox economist”—a dissident from the prevailing Keynesian doctrine developed at MIT by the Nobel laureates Paul Samuelson and Robert Solow.

Based at Washington University in St. Louis for much of his career, Minsky developed a reading of Keynes’s work that contrasts fundamentally with that formulated by Samuelson and Solow. In their Neoclassical Synthesis, Keynesian macroeconomic policy would ensure that resources are fully employed. Then the traditional neoclassical microeconomics of efficient markets could be deployed, devoid of the uncertainty that pervades Keynes’s own work. The economist Joan Robinson called this “Bastard Keynesianism.”

I myself was sufficiently immersed in Keynes’s own thinking that, rather than teach the Neoclassical Synthesis, I embarked on a 35-year sabbatical as a venture capitalist. My one significant brush with academia in these years was when I met Minsky in the mid-1980s. The relationship deepened when he joined the Levy Economics Institute, conveniently close to New York City. Minsky sponsored a paper I presented to the Annual Meeting of the Association for Evolutionary Economics in December 1985.

That paper, “Doing Capitalism: Notes on the Practice of Venture Capitalism,” examined the differing profiles of the “financial agent” in the works of Fernand Braudel, Karl Marx, Schumpeter, and Keynes, drawing parallels between each and the role of the modern professional VC investor. It served as the seed that would grow into Doing Capitalism in the Innovation Economy, which I published almost 30 years later, in 2012. Still following where Keynes had led, I focused on the economics of innovation, where investment at the frontier of technology necessarily takes place under conditions of radical uncertainty and volatile financial conditions.

As L. Randall Wray shows in his excellent 2015 book, Why Minsky Matters, Minsky’s most important message is that economists’ fixation on defining equilibrium conditions evades a central, existential truth: stability is itself destabilizing. Minsky set out to explain what he identified as Keynes’s “investment theory of fluctuations in real demand and a financial theory of fluctuations in real investment.” He began by walking carefully through “the conventional wisdom, the standard interpretation of Keynes,” which had served to obscure what Keynes had achieved and derailed the revolution in economic theory he had launched.

Throughout his 1975 book, John Maynard Keynes, Minsky repeatedly cites Keynes’s invocation of uncertainty as the fundamental factor conditioning economic and financial decisions. “Keynes without uncertainty is something like Hamlet without the Prince,” he observed. As Keynes himself had emphasized in a 1937 commentary for the Quarterly Journal of Economics, uncertainty is effectively an ontological condition of the universe:

What Money Does

This argument moves from metaphysics to the sublunary plane of financial economics as soon as Keynes (and Minsky) turns to money. It is all very well to note the convenience that money offers as a medium of exchange, eliminating the simultaneous “coincidence of wants” that would otherwise be necessary to motivate trade. But why hold money as an asset, as something to hoard, when it yields no income? “Why,” Keynes asked, “should anyone outside a lunatic asylum wish to use money as a store of wealth?” The answer, of course, is that it provides insurance against all that cannot be known in advance.

Money is unique among assets for its extreme liquidity, which complements its lack of return. Keynes’s description of “liquidity preference” follows from his recognition that interest rates represent the “premium” that must be offered to induce investors not to hold cash. But liquidity preference cannot be a stable function, because it is not an attribute of an asset. Rather, it is an attribute of the market conditions in which it becomes subjectively desirable or even objectively necessary to convert a given asset into cash. As anyone who has worked in financial markets knows, liquidity has the perverse attribute of being less available the more you need it. Liquidity preference thus becomes one avenue through which uncertainty informs investment decisions.

The motivation to acquire and hold any asset other than money, of course, stems from the yield it will return to the owner. For the yield on capital assets, Minsky follows Keynes in using Alfred Marshall’s term “quasi-rent,” which incorporates both the uncertainty about the size of future yields as well as a forecast of the liquidity that might be available to the owner of the asset if needed. Minsky then addresses the demand price for capital assets, the physical embodiment of investment. Here, he criticizes Keynes for inviting confusion by viewing the relationship between investment and the variables that determine it (the investment function) largely in terms of interest rates rather than asset prices. For it is in the pricing of capital assets that uncertainty about their future yield becomes a critical factor.

The economist John Hicks walked through the door that Keynes left open when he proposed a radically and misleadingly simplified alternative to Keynes’s economics. Hicks argued in a 1937 Econometrica paper that investment can be treated as a simple function inversely related to the rate of interest and equal to savings in equilibrium. He reduced liquidity preference to a simple demand function for money, equivalent in equilibrium to a given supply of money set by the central bank. Known as “IS-LM,” this formulation banished expectations and uncertainty not only from center stage but even from the wings of the theatre. Unfortunately, it became the near-universal summary of Keynesian economics.

What Drives Investment?

For me, the core of Minsky’s analysis of Keynes’s economics comes in his explication of the process through which investment is determined. This process involves three types of assets: money, debts (or “contracts exchanging present money for future money”), and real capital assets (“characterized by expected yields that may vary for a number of reasons”). Minsky writes:

Four decades after presenting his IS-LM formulation, Hicks published “IS-LM: An Explanation,” setting forth the various reasons why he himself had “become dissatisfied” with what by then had long since become the standard formulation of Keynesian economics. The central factors encapsulated in IS-LM, he emphasized, are equilibrium relationships, which must by definition be maintained through time: “Expectations must be kept self-consistent. … That can only be possible if expectations … are right. Equilibrium over time thus implies consistency between expectations and realizations.”

The problem, of course, is that “there is no sense in liquidity, unless expectations are uncertain.” Thus, Hicks pulled a central brick from the foundations of the Neoclassical Synthesis, belatedly providing a critique that neither Keynes—first sidelined by a mild heart attack, and then thoroughly occupied with financing the war effort against the Axis Powers—nor his students had delivered previously.

For Keynes, the fragility of any transient equilibrium among the relationships that together determine the rate of investment is explicit in the central role of “long-term expectations,” which are always subject to rapid revisions as the unknown future morphs into the ever-evolving present. As Minsky subsequently wrote, “every reference by Keynes to an equilibrium is best interpreted as a reference to a transitory set of system variables toward which the economy is tending; but … as the economy moves … endogenously determined changes occur which affect the set of system variables toward which the economy tends.”

More simply put: “[T]he core of Keynes’s system consists of an analysis of capitalist finance in the context of uncertainty, and of how capitalist finance affects the valuation of items in the stock of capital assets and thus affects the pace of investment.” (Emphasis in the original.) Given the relatively passive nature of the relationship between consumption and disposable income (the consumption function), fluctuations in investment drive fluctuations in aggregate demand and in the economy as a whole. Here we have, in sum, the Investment Theory of Aggregate Demand and the Financial Theory of Investment.

From this perspective, the supposed link between the technical productivity of the “K” (the aggregate capital stock) specified in the neoclassical production function and the rate of investment is confounded by the variability of the prospective yield from the asset (taking into account uncertainty and liquidity preference) and the variability between the discount rate applied to the expected yields from owning specific assets and the money rate of interest. Moreover, both the demand and supply prices of capital are denominated in nominal money terms, as is the debt assumed by investors to finance their purchase.

The neoclassical Keynesian tradition has a long history of evading the fact that in actually existing labor markets, employers and employees can bargain only over money wages. Equally, in actually existing financial markets, debt liabilities are denominated in nominal terms. The real wages embedded in production functions and macro equations are realized only after the fact.

This confusion has sustained the fundamentally flawed belief that the root cause of persistent unemployment lies in money wages being too high, implying that they must be lowered to reduce unemployment. Both neoclassical and New Keynesian macroeconomics use “wage stickiness” to justify policy interventions to augment aggregate demand, thus avoiding the bitter struggle to force wages down. In theory, if wages and prices were fully flexible, then interventions to boost demand in times of recessions would not be needed.

By contrast, Minsky correctly interpreted Keynes to mean that, in a world of debt contracts set in nominal money terms, downwardly flexible money wages and prices will threaten to unleash the destructive debt-deflation cycle (as described by Irving Fisher) that the United States had experienced in 1930-33. To Keynes and to Minsky, wage stickiness is a feature of the system, not the bug that mainstream theory portrays.

The Financial-Instability Hypothesis

Minsky’s restatement of The General Theory is enriched by his linking the acquisition and ownership of assets to the liabilities incurred to finance their purchase. His economic units are like “little banks” that pay for assets by issuing liabilities. They sit at the intersection of cashflows from the assets they own and those they are obliged to make to their creditors.

Minsky was building on Keynes’s identification of “the entrepreneur’s or borrower’s risk … of his actually earning the prospective yield for which he hopes,” and the “lender’s risk” of voluntary or involuntary default by debtors. “The fundamental fact about both borrower’s and lender’s risks is that they reflect subjective valuations,” Minsky writes. He then quotes Keynes: “During a boom the popular estimation of both of these risks … is apt to become unusually and imprudently low.”

Here we can find the roots of the financial-instability hypothesis for which Minsky is best known. In a 1992 Levy Economics Institute working paper, he proposed two theorems. First, “the economy has financing regimes under which it is stable, and financing regimes in which it is unstable.” Second, “over periods of prolonged prosperity, the economy transits from financial relations that make for a stable system to financial relations that make for an unstable system.” Remember, stability is destabilizing: the credit system and the real economy progress (or degenerate) through successive stages of rising confidence and heightened risk-taking.

The initial, conservative stage is that of hedge finance: the operating cash flows of borrowers are sufficient to service outstanding debts and to repay them as they mature. The business operations of the borrower are themselves an effective hedge against the obligation to service and repay the debt. As expectations of borrowers and lenders are validated by experience, they jointly move into the phase of speculative finance. Operating cash flow is sufficient to make timely payment of interest, but the principal must be rolled over and refinanced to prevent default and hence is exposed to changing market conditions: this is the speculative element. Finally, the system moves into the stage of Ponzi finance, where debtors must borrow to pay the interest they owe to indulgent lenders in order to maintain the fiction of solvency.

Minsky operationalized the liability structure that these firms engender as “budget constraints.” Through a neat model, he shows how the tightness of budget constraints at the firm and aggregate levels turns on the extent to which investments can be funded from operating cash flow versus external finance. Where and when the budget constraint will bind, in turn, depends on the firm’s access to liquidity, whether from current cash holdings or the potential to liquidate assets.

Within mainstream neoclassical economics, the 2008 global financial crisis generated a wave of responses that often explicitly invoked Minsky. But even before the fact, anyone attuned to Minsky’s analysis of financial fragility could have noted a phenomenon that had emerged in the leveraged-buyout market by 2006, fully two years before Lehman Brothers collapsed. The popular rediscovery of Minsky was typically expressed in the meme, “The Minsky Moment.” But Minsky had defined a process, not a moment, and the stage that it had reached in 2006 could have been inferred from the fact that banks were funding their customers with “PIK-Toggle” debt instruments.

This meant that debtors could service their obligations by issuing more debt (PIK stands for “payment in kind”), and that the decision to do so was entirely within the borrower’s discretion (the “Toggle”). There could have been no clearer evidence that the system was in the Ponzi phase. Moreover, markets were so confident in the illusion they had created that one could purchase a three-year put on the S&P 500 (a bet that the index would fall) for only 2% per year.

My own experience as a VC had validated Minsky’s theorizing. As my brilliant and challenging mentor Fred Adler put it, “corporate happiness is positive cash flow.” The ability to pay your debts because your customers pay you more in cash than what it costs to develop and deliver what you are selling liberates the firm from dependence on the problematic, essentially unknowable future state of financial markets.

For a VC funding startups, the imminence of Minsky’s “survival constraint”—the point at which all sources of cash from operations, securities issuances, and asset sales have been exhausted—is always a real-world threat. And the episodic bubbles that magically relieve such threats represent the quintessence of uncertainty: you know that each will burst; you just don’t know when.

Big-State Capitalism

In The General Theory, Keynes analyzed the dynamics of “small-state capitalism.” In 1929, the US public sector amounted to only around 4% of national income, half of which was accounted for by the federal government (principally the postal and customs services, plus very modest armed services) and the other half by state and local governments (the largest component was schoolteacher salaries). The federal share then doubled between the 1929 Wall Street crash and Franklin D. Roosevelt’s inauguration in 1933, but that was because the decline in nominal national income had reached almost 50%, half of which was “real” and half from deflation. Meanwhile, state governments were required by their own constitutions to balance their budgets, cutting expenditures as tax revenues fell. The federal government’s share thus grew to 4% as the Hoover administration passively absorbed the debt-deflation juggernaut.

During the summer of 1964, I was a research assistant to Princeton’s Lester Chandler, a doyen of old-school money-and-banking finance (who would soon publish American Monetary Policy, 1928-1941, a neglected corrective to Milton Friedman and Anna Schwartz’s Monetary History of the United States). At the time, the White House Council of Economic Advisers was promulgating the concept of a full-employment budget deficit to justify tax cuts, and Chandler guided me through a back-of-the-envelope estimate of the magnitude of the federal deficit in 1933 that would have been roughly consistent with full employment. As I recall, the number we came up with was about three times the size of total federal expenditures in 1932, a year in which net private-sector investment was negative (because new investment spending was less than the depreciation of the capital stock). The conclusion was simple: to stabilize the economy, the government had to become much bigger.

In reality, by the 1980s, when Minsky addressed the question of whether “it”—a Great Depression—could happen again, the US public sector had grown large enough to offset investment fluctuations, thanks to the dual emergence of the welfare state (from Social Security through Medicare) and the Cold War warfare state. To explain the big state’s stabilizing role, he identified three complementary contributions that substantial fiscal deficits make to a financially stressed economy. First, they augment aggregate demand by increasing income and employment; second, they generate cash flows that protect firms from the threat of default due to their own reduced investment and increased precautionary savings by households; and third, they supply low-risk investment instruments for risk-averse investors.

Minsky continues:

Minsky’s vision extended beyond the stabilizing macroeconomic corrective with which history had endowed the US. He sought a much more comprehensive extension of the New Deal order, characterized by the public orchestration of large, capital- and debt-intensive public works and the direct provision of employment to idle workers. While massive defense budgets had become a stabilizing force, he worried that they made for a poor alternative to socially productive investments. Sadly, with the current US administration hell-bent on destroying the capacity of the US state to deliver on any of its socioeconomic responsibilities, Minsky’s policy agenda appears to be off the table.

Minsky’s Legacy

Minsky’s version of Keynes’s economics had a minimal impact on mainstream economics during his own lifetime and in the years that followed. But the 2008 global financial crisis set in motion a process that continues—however problematically and haphazardly—to reconstruct the discipline of economics, and Minsky and his work may yet play a role.

Economists have responded to the 2008 crisis in two ways. The first is the “empirical turn” toward analysis of distinct aspects of economic behavior. It is as if a broad range of scholars reflected: “We thought we knew how the world works. The global financial crisis suggests we were wrong. We should study how the world really works.”

A second response has been a shotgun re-marriage of economics and finance. Modern finance theory and analysis had found an academic home in the leading business schools, while mainstream macroeconomists’ models largely excluded the possibility that any developments in financial markets and institutions could affect real economic variables such as employment and consumption. Largely in response to the 2008 crisis, a new field of “financial macroeconomics” has emerged—more or less explicitly incorporating lessons from Minsky.

Following the crash, the macroeconomists who rediscovered Minsky’s insights included Richard Koo of the Nomura Research Institute and the Nobel laureate Paul Krugman, both of whom looked back to Japan’s “lost decades” after 1990 to identify the post-2008 Great Recession as a “balance-sheet recession.” Informed by the global financial crisis, in 2011 Patrick Bolton, Hui Chen, and Neng Wang published a paper analyzing budget constraints at the level of the firm. They proposed “a model of dynamic investment, financing, and risk management for financially constrained firms,” highlighting “the central importance of the endogenous marginal value of liquidity (cash and credit line) for corporate decisions.”

More recently, the macroeconomic implications of borrowing constraints as explored by Minsky have been examined by a growing number of economists, generally showing how such “frictions” serve to amplify economic contractions as in the Great Recession. But one paper by Nuno Coimbra and Hélène Rey develops an endogenous process through which financial fragility emerges—as Minsky proposed—during the boom phase of a macro-financial cycle. This paper is of particular significance because Rey has just been named to the strategic role of Economic Adviser and Head of the Monetary and Economic Department of the Bank for International Settlements, “the central banks’ central bank.” Thus, 30 years after his death, Minsky and his reading of Keynes are entering the mainstream of the discipline.

One highly relevant place to examine how the world really works right now is the extraordinary boom in AI-related investment to fund construction of the physical infrastructure needed to enable the training and deployment of large language models. These investments are motivated by problematic long-term expectations with respect to commercially useful and financially rewarding applications of generative AI.

The question from the world of financial economics is whether, in aggregate and with respect to specific players, the cash flows generated by these applications will be sufficient—and sufficiently timely—to validate the investments now committed. In turn, funding from the operational cash flows of the established, monopolistic tech giants has been giving way to rapidly growing issuance of debt securities. The entire phenomenon cries out for analysis in the spirit of Minsky’s Keynes.

It is good news that economists are increasingly exploring how a complex financial system interacts with an equally complex real economy, one that is structured through dynamically evolving networks of production and consumption. This is a project to which Minsky would have much to contribute, especially by bringing to bear the central element in Keynes’s economics: the inescapable uncertainty that attends all economic and financial commitments.

William H. Janeway is a distinguished affiliated professor in economics at the University of Cambridge and author of Doing Capitalism in the Innovation Economy (Cambridge University Press, 2018).