The triumph of the shiftability view in the 1935 Bank Act meant that, from then on, the Fed was prepared to act fully as lender of last resort, accepting as collateral any “sound” asset and not limiting itself to short-term self-liquidating paper. Two years later, in a communication by the Federal Open Market Committee (FOMC) issued in April 1937, the Fed went even further, committing itself to maintaining “orderly conditions in the money market” quite generally. What this meant was that, instead of waiting passively for banks to request loans, the Fed was prepared to intervene proactively by buying and selling securities in the open market. In 1935, shiftability was supposed to be provided by private dealers, operating with the knowledge that the Fed’s discount window would be available to them if they got into trouble. That wasn’t enough, so in 1937 the Fed took over responsibility for providing shiftability itself. In effect, the Fed committed itself to act as a security dealer.

In retrospect, this 1937 commitment represents a very significant step, since it brought to the forefront of policy consideration the whole question of capital asset pricing. After all, how is the Fed supposed to know when the market is “disorderly,” and how is it supposed to know what kind of intervention is needed? In the end, the symptom of disorder always comes down to deviation of asset prices from some idealized norm. If sellers of the seven-year bond find that they have to make significant price concessions in order to attract buyers, the Fed is supposed to notice and step in as buyer, while at the same time selling other securities for which buying interest is stronger, say the two-year bond. Such intervention requires the Fed to have in mind a more or less clear norm indicating what current asset prices should be, so that it can use deviation from that norm as an indication of which issues it ought to be buying and which issues it ought to be selling.

In 1937, the more or less obvious idealized norm was something economists call the expectations hypothesis of the term structure (EH). According to this theory, the return on a longterm bond should be just an average of expected short-term interest rates over the life of the bond. The logic is simple. Suppose you have money that you want to invest for the next two years. You can either buy a bond with two years to maturity or you can buy a short-term bill with only one year to maturity and then reinvest the proceeds in another one-year bill. Since investors can freely choose between these two strategies, they must be equally attractive in market equilibrium; both two-year bonds and one-year bills have to be held willingly by someone. One measure of the attractiveness of an investment strategy is simply its yield to maturity. Accordingly, the EH suggests that the expected yield to maturity for each of the two investment strategies should be the same.

Appealing as this theory is, it cannot be the whole story because in actual fact the term structure typically slopes upward. Short-term interest rates are typically lower than long-term interest rates, so a long-term investment typically yields more than a series of short-term investments. One way to understand this apparent anomaly is that investors are influenced by something other than expected yield to maturity. Maybe the long-term investment is more risky in some way, and the extra yield is compensation for bearing that risk? Certainly it is true that long-term investments can potentially fluctuate in value much more widely than shortterm investments, but this fluctuation matters only if for some reason you have to sell before maturity. This line of thinking thus suggests that the extra yield on the long-term investment is a kind of “liquidity premium” that compensates the long-term investor for the fact that he may have to take a loss if for some reason he needs to convert his investment into cash before maturity.

This way of thinking about the term structure was more or less state of the art in 1937, and the Fed’s 1937 commitment to maintain orderly conditions in the money market can therefore be understood as a commitment relative to this norm. In 1937, the Fed’s commitment meant that it would take the “liquidity premium” as given by the market, and ensure shiftability at that premium. Over time, interest rates would change, depending on market conditions. Short rates might rise or fall by more than long rates, making the term structure flatter or steeper, but that was not supposed to be any concern of the Fed. “Orderly conditions in the money market” meant smoothing price changes, including changes in the liquidity premium, not preventing them.

Why did the Fed think that such smoothing was an appropriate policy goal? Here, perhaps by analogy to previous experience with seasonal smoothing, the Fed seems to have understood itself as intervening to produce the outcome that the market was trying, unsuccessfully, to achieve. Behind this approach was the idea that, in a well-functioning market, private dealers themselves would be doing the smoothing, using their own balance sheets to absorb temporary imbalances of supply and demand. But even in a wellfunctioning market it might occasionally happen that the bulk of some particular security issue was locked away somewhere out of the reach of the dealers, so that the market price of that security came to reflect a scarcity premium. In such a case, the Fed could help the market by disgorging its own holding. If this problem could arise in normal times, then a fortiori during a depression, when private dealers were weakened and private investors reluctant to disgorge their only safe securities, the market could use the Fed’s help. The Fed’s April 1937 announcement was intended to signal that such help would be forthcoming.

Had events not intervened, this reassuring signal might have been the first step toward reconstruction of private capital markets on a new, more solid, foundation, based now explicitly on shiftability rather than self-liquidation. As the capitalization of the dealers was restored, the time would eventually have come when dealers themselves would have taken over the task of smoothing, leaving the Fed free to concentrate instead on general credit policy. In this respect, it is telling that, at the very same moment when the Fed was sending its reassuring signal about the shiftability of all government securities, it was also doubling reserve requirements in an effort to forestall a possible unhealthy credit expansion on the basis of the burgeoning excess reserves that had built up in the banking system. In 1937, the Fed was preparing for return to normalcy.

What kind of normalcy might have been built on these new foundations? As mentioned earlier, the Fed’s commitment to ensure shiftability brought to the forefront the whole question of capital asset pricing. The EH was a good starting point, especially as augmented by the idea of liquidity preference that generated risk premiums in the yields of longer-maturity assets. But matters would never have remained at such a primitive level. Once you start down the road of intervening to correct deviations from appropriate pricing, you also start down the road of constantly improving your theory of appropriate pricing, if only to protect yourself from profit-seeking counterparties on the other side of your trades.

Just so, the evolving role of the Fed inevitably would have led to further investigation of the foundations of the concept of “liquidity premium” in order to understand where it comes from, why it takes on particular values for particular instruments, and why those values change over time. In Britain, John Maynard Keynes and John Hicks were already showing a possible road forward with their “normal backwardation” theory of why longterm rates tend to be higher than short-term rates (see chap. 4). Events intervened, however, and these early contributions got put on the shelf. The hoped-for restoration of private capital markets, and the institutional evolution that would have followed from that restoration, were both diverted, first by renewed economic downturn in 1938, and then by World War II.

Monetary Policy and the Employment Act

During World War II, the Fed once again took on responsibility not only for maintaining orderly markets in government debt, but also for fixing prices of that debt. Private capital markets essentially disappeared and were replaced by an expanded market for government debt, for which the Fed served as market maker. As new credit went preferentially to government, businesses and households used any excess funds first to pay down their own debts, and then to accumulate government debt. After the war, at the insistence of the Treasury, the Fed continued to fix prices, with the result that government bonds served essentially as a kind of interest-bearing cash. So flush was the private sector with its accumulation of war debt that reconstruction of private capital markets could and did wait until these wartime accumulations had been drawn down.

Meanwhile, the Employment Act of 1946 crystallized the new political consensus in favor of using the power of the federal government “to promote maximum employment, production, and purchasing power.” The experience of wartime spending had produced a new appreciation of the power of fiscal measures to pull the economy out of depression, while the experience of wartime wage and price controls had produced a new appreciation of the power of direct measures to stem even very powerful inflationary pressures. Even more, the experience of low and stable interest rates during the war had produced a new appreciation of the power of finance to support important social goals. In all these dimensions, wartime lessons shaped postwar understanding of economic policy in ways that involved a significant downgrade of the role of monetary policy and the Fed. The Employment Act made no mention of a possible role for the monetary authorities in achieving the new goals; in practice the role of the Fed was limited to fixing the price of government debt for a full five years after the Employment Act was passed.

Nevertheless, inside the Fed, preparations were being made for a return to normalcy. The first step was to regain control over short-term interest rates. Pegged at only 3/8 percent during the war, when long bonds were supported at the more generous rate of 2 1/2 percent, the short-term bill had attracted very little private interest, with the consequence that the Fed wound up owning almost all of the outstanding bills. By raising short-term yields on assets it already owned, the Fed was finally able to attract private buyers and so to begin the process of rebuilding a proper money market. After a few years of this, in 1951 the Fed was also finally able to rid itself of responsibility for pegging bond prices. Under the new Fed chairman William McChesney Martin, this abdication of wartime responsibility was presented as a crucial step toward rebuilding a proper money market for use in postwar monetary policy intervention. The key document was the 1952 report of the FOMC’s Ad Hoc Subcommittee on the Government Securities Market, a committee chaired by Martin.

In retrospect, this report, and the “bills only” doctrine that emerged from it, can be understood as an adaptation of the FOMC’s 1937 commitment (to maintain “orderly conditions in the money market”) to the significantly changed conditions of 1952. In both cases, the central idea was to support rebuilding of the dealer infrastructure of private capital markets. In 1937, the Fed thought it could confine its involvement to the government securities market, and it relied on private dealers to bring orderly conditions to private capital markets as they recovered. By 1952, however, the government market was more or less the entire capital market and the private dealer system was essentially moribund. It seemed to follow that the task of rebuilding the infrastructure of private capital markets had to begin at a more primitive stage by first attracting private dealers to act as market makers in the government securities market.

“Bills only” was thus a way of signaling that the Fed was leaving the long end of the market to the private dealers. (It was of course also a way to avoid political pressure to keep long rates low and stable.) Thenceforth the Fed would maintain orderly conditions at the short end, and rely on arbitrage and the private dealers to bring orderly conditions to the long end. In 1937, the EH had served as the norm for the Fed’s own intervention across the term structure of Treasury debt. In 1952, the same EH was to serve as the norm for private profit-maximizing dealers.

In practice, even this simple idealized norm proved to be unrealizable in the rigid financial conditions inherited from New Deal reforms. Regulatory constraints on portfolio investment created demand for specific issues that was not readily diverted into other issues by mere price movements. Thus, although deviations from the EH norm offered profit opportunity for dealers, exploiting those opportunities was possible only if the dealers were prepared to hold the resulting position to maturity; the very existence of such deviations was a sign that the position would not be easy to liquidate. Assets that theory treated as close substitutes, and thus ideal candidates for arbitrage, were in fact not close substitutes at all in the portfolios of the ultimate wealth holders, and hence not for anyone else either.

Allan Sproul, the president of the New York Fed, had anticipated exactly this problem and thus had spoken out in opposition to “bills only.” He thought the Fed should retain its ability to operate in each of the multiple fragmented and disconnected credit markets, as well as in the long end of the Treasury market. Experience of the new policy proved him right, but only in the short run. Over the longer run, the vision of institutional arrangements laid out by Martin in the 1952 report proved remarkably prescient, as we shall see. It just took longer than he anticipated.

Meanwhile, the next step for the Fed after regaining control over interest rates was to revive its own role in managing them. We have seen how, during the 1920s, following the Strong rule, the Fed targeted borrowed reserves as a measure of money market tightness and slack, and used open market operations to increase or decrease borrowing depending on whether it felt that conditions warranted a bit more discipline or a bit more elasticity. In the 1950s, the Fed adopted a quite similar procedure, based on a quite similar theory. The only difference was that now it targeted so-called free reserves, defined as the difference between excess reserves and borrowed reserves.

This operational change can be understood, first of all, as a nod to the developing federal funds market, in which banks with excess reserves were increasingly able to lend their excess to banks with deficient reserves. Such interbank lending had the effect of decreasing both the sum of excess reserves and the sum of borrowed reserves by the same amount. One advantage of the free reserves concept, therefore, relative to the former focus on borrowed reserves alone, was that it was not affected by fluctuations in the volume of interbank lending in the federal funds market.

Another advantage was that the free reserves concept could be either positive (in times of slack) or negative (in times of tightness). By contrast, the borrowed reserve concept seemed to have a natural floor at zero since, once no bank was borrowing reserves from the Fed, there was no way for the Fed to “make discount rate effective.” (During the early years of the Depression, this apparent floor had led the Fed erroneously to conclude that monetary conditions were maximally expansive.) The new procedure recognized that, even when no bank was borrowing from it, the Fed could still exert additional downward pressure on shortterm interest rates by buying securities outright; the effects of such open market operations would show up as an expansion of free reserves.

As in the 1920s, the purpose of the Fed’s post–Fed-Treasury Accord operations was not merely to ensure orderly conditions in the money market, but also to contribute to economic stabilization more generally by constraining speculative excess on the upside and by supporting markets during liquidation on the downside. In 1955, addressing a joint meeting of the American Finance Association and the American Economic Association, New York Fed president Sproul explicitly endorsed the goals of the Employment Act, and urged the importance of the Fed for achieving those goals: “We must be alert to oppose both inflationary and deflationary pressures, either one of which can upset the precarious balance of a high-employment, high-production, high-income economy.” In the same address, Sproul also asked the assembled academics for their help: “It seems to me that this matter of open market techniques involves problems of economic significance beyond its immediate technical application, and that it deserves your study and your published findings.”

Listening to the Academics

Be careful what you ask for. Academic advice had not always been welcomed by the Fed, and for good reason. During the 1920s, the Yale economist Irving Fisher had spearheaded a campaign to pass legislation that would have required the Fed, as a matter of law, to stabilize the price level. Fisher’s analysis rested on his version of the quantity theory of money, developed in his 1911 book The Purchasing Power of Money. According to Fisher, both inflation and deflation could easily be avoided simply by manipulating the money supply. Since they could be avoided, they should be avoided, both as a matter of justice between creditors and debtors and as a matter of efficiency in economic behavior. When price levels shift around, he argued, people tend to make mistakes in their economic decisions because they fail to distinguish movements in individual prices (which signal changes in profit opportunities) from movements in the price level (which contain no such signal). These mistakes, which he called “money illusion,” he thought were an important cause of business fluctuations. Thus, he concluded, a policy of manipulating the money supply in order to stabilize the price level should also help to stabilize the economy more generally.

In the 1920s, the Fed had resisted Fisher’s attempt to tie its mission explicitly to price stabilization, but not because of any particular objection to price stabilization as a goal. Rather, it rejected the idea that such stabilization was as easily achieved as Fisher thought, and rejected furthermore the idea that manipulation of the money supply was the most efficient operational mechanism toward that end. The yawning gap between Irving Fisher of Yale and Benjamin Strong of the New York Fed was not so much about the mission of the Fed as it was about the gap between academic monetary theory and central banking practice. For Strong, the focus of attention was not so much on the quantity of money as it was on the price of credit, which is the rate of interest. Strong’s idea was that the pace of lending depends on the profitability of lending, which is the difference between the loan rate of interest and the money market rate of interest. Central bank intervention thus properly focused on the money rate of interest, using control over bank reserves as leverage, to influence the pace of lending.

In the 1920s, emergent central banking practice (Strong) easily trumped a priori academic monetary theory (Fisher), but only to be itself trumped by depression and war in the 1930s and 1940s. The result was a substantial downgrading of monetary policy, in both academic and policy circles. In the 1950s, when peace and then prosperity returned, so too did the old debate, albeit in somewhat different form and with different protagonists.

In academia, Milton Friedman, professor of economics at the University of Chicago, reprised Irving Fisher with his proposal to require the Fed to stabilize, not prices directly, but rather the growth of the money supply. Three percent, said Friedman, was about right to achieve long-run price stability, given historical patterns of long-run growth in the real economy. For our story the important point is that, in a nod to the weak empirical connection between money and prices in the short run, Friedman advocated abandoning the goal of active countercyclical stabilization. In this respect, Friedman’s monetarism diverged from Irving Fisher’s, and in a direction that took him even further away from central banking practice; the postwar gap between academic theory and central banking practice threatened to yawn even larger than the prewar gap.

Meanwhile, reprising Benjamin Strong at the Fed, Allan Sproul and William McChesney Martin quite definitely saw a role for countercyclical monetary policy (“leaning against the wind”). The difference from the 1920s was that, after the Employment Act, the commitment to economic stabilization was much more widely held, and a broader range of instruments was available for the task. monetary policy no longer had to do everything, so central bankers could pick and choose those dimensions of the stabilization project for which their instruments were best suited. By stabilizing money markets, the Fed could do its own bit toward the broader social goals of stabilizing the price level and economic activity more generally. Friedman’s monetarism was definitely not the kind of academic work that Sproul had meant to encourage.

Closer to what Sproul probably had in mind was the work of John G. Gurley and Edward S. Shaw, who reprised Harold Moulton in their 1960 book Money in a Theory of Finance. Moulton, it will be recalled, had sparked the reconceptualization of liquidity as shiftability, arguing that the shiftability concept was better suited to American conditions in which banks were much more involved with long-term capital finance. Taking their theme from Moulton, Gurley and Shaw expressed concern that bank regulations introduced during the New Deal were having the unintended consequence of suppressing capital accumulation. New financial intermediaries, especially pension funds and insurance companies, were taking over the role formerly played by banks, but inadequately so. The liquidity preference of households meant that they still tended to channel their savings preferentially to the relatively safe and liquid liabilities of banks.

In the Gurley-Shaw idealization, each financial intermediary issues a characteristic type of liability that attracts funds from a distinct segment of final savers, funds that the intermediary then uses to acquire a characteristic type of asset issued by a distinct segment of final borrowers. This image of the institutional organization of financial markets was intended to capture the rigidities that had been built into the American financial system by regulatory strictures (the very rigidities that Sproul had emphasized in his opposition to “bills only”), but Gurley and Shaw went further. regulatory constraints on banking meant that asset holders were not able to find in the marketplace sufficient assets to satisfy their liquidity preference, and the consequence was higher liquidity premiums than necessary, which stifled general capital accumulation and long-run growth. Furthermore, attempts by the central bank to manage money by managing bank credit risked stifling those particular forms of capital accumulation that were most reliant on bank credit. In both respects, longrun growth was being sacrificed to short-run cyclical stabilization; the Fed could do better.

At the Fed and in academia, the decade of the 1950s was devoted largely to monetary reconstruction, in terms of both institutional structures and intellectual frameworks. In both respects, by 1960 we were more or less back to 1930; money mattered again. Given the enormous discredit that had befallen both the Fed and monetary economics during the 1930s, it was a tremendous achievement. But getting money back on the agenda was only the beginning, since meanwhile the world had moved on. The problem after 1960 was to find a way to integrate the new appreciation of money within the larger institutional and intellectual framework of macroeconomics, which had undergone tremendous change in the decades since 1930.

Monetary Walrasianism

One way to understand the catastrophe of the Great Depression is not so much as a failure of monetary policy narrowly but rather as a failure of the decentralized market system more generally. Price determination in response to the fluctuation of supply and demand was apparently not sufficient to ensure satisfactory performance of the economy as a whole. The experience of instability and unemployment was unacceptable, and the consequence was an opening to explore alternatives, both institutional and intellectual. One possible alternative was to abandon the market economy entirely and replace it with a centralized command economy; the success of the U.S. war economy showed well enough that such a possibility could work. But maybe there was another way, a middle ground in which government managed rather than commanded. The commitment to exploring that middle ground is what the so-called Keynesian revolution was all about, at least in the United States.

We have seen already how the Fed’s 1937 commitment to maintain “orderly conditions in the money market” makes sense only by reference to an idealized norm, at that time the liquidityaugmented expectations hypothesis of the term structure. No less did the broader 1946 commitment “to promote maximum employment, production, and purchasing power” make sense only by reference to an idealized norm. By 1960, when money finally came back into the picture, economists and policymakers had already chosen the ideal norm against which to compare imperfect reality. That norm was the general equilibrium model first put forth by Léon Walras in 1874.

It seems to have been John Hicks, first in his 1937 article “Mr. Keynes and the ‘Classics’” and then in his 1939 book Value and Capital, who most clearly set forth the general equilibrium model of Walras as the idealized norm against which to calibrate real world deviations. Walras famously envisioned the economy as a set of simultaneous equations, each one setting the supply of a particular commodity equal to the demand for that commodity. The equilibrium of the system is the set of prices that satisfies all of the equations simultaneously. By 1960, Arrow and Debreu had developed the Walrasian general equilibrium idea into a fully rigorous mathematical formalism, but their version of the model had a problem; there was no place in it for money.

Those who wanted to bring money into the picture had therefore to find a different starting point, which they did in a 1938 article by Jacob Marschak titled “Money and the Theory of Assets.” Here is the origin of monetary Walrasianism, the operational form taken by the Keynesian revolution in terms of money. The idea was, by analogy to Walras, to treat the financial side of the economy also as the solution to a set of simultaneous equations, each one involving the supply and demand for a particular financial asset, money being only one of many such assets.

It took a while for the Marschak approach to gain acceptance, in part because of war, and because of monetary discredit too. But in 1952 Harry Markowitz reprised the Marschak approach in his seminal “Portfolio Selection,” and in 1958 James Tobin used the approach to develop a theory of money demand in “Liquidity Preference as Behavior towards Risk.” Once these works demonstrated the apparent viability of the new approach, the next question was how to bring the new approach to money into contact with the institutional specificity of the American economy as emphasized by Gurley and Shaw, on the one hand, and with the practical stabilization policy goals as emphasized by Sproul and Martin, on the other hand. That’s what James Tobin, among others, was doing during the decade of the 1960s, work summed up in his 1969 “A General Equilibrium approach to Money.”

In fact, Gurley and Shaw had already paved the way for Tobin’s monetary Walrasianism by presenting household liquidity preference as a matter not just of money demand but also more generally of portfolio choice. In the Gurley and Shaw framework, the institutional specificity of the American economy involved various distortions in the supply of assets, while the price of assets moved to ensure that all assets were held by someone. For Tobin, these distortions were the source of deviations from the ideal, but they were also the source of leverage for policymakers attempting to get closer to the ideal. Just so, Tobin would emphasize how the ability of the monetary authority to affect asset prices and hence real activity stemmed from the fact that “the interest rate on money is exogenously fixed by law or convention.” Other institutional rigidities, such as “prohibition of interest on demand deposits and a ceiling on time deposit interest,” provided additional sources of policy leverage.

In effect, Tobin used Gurley and Shaw to bridge the gap between academic theory and central bank practice, but once the bridge was built the traffic that flowed over it was largely from academic theory to practice, not the other way around. In Tobin’s hands, the Gurley-Shaw vision merely enriched the specification of the standard Hicks-Hansen IS/LM model, used by everyone for short-run comparative statics exercises. Tobin’s paper concludes, “There is no reason to think that the impact \[of monetary policies and other financial events\] will be captured in any single exogenous or intermediate variable, whether it is a monetary stock or a market interest rate.” This is a potshot both at academic monetarists who focus attention narrowly on the money stock and at central bank practitioners who focus attention narrowly on the money rate of interest. It is also an attempt to make common cause with the Sproul/Martin conception of the role of the Fed as stabilizer of the financial sector quite generally.

The Marschak-Tobin framework thus became the template for monetary practice as well as theory, once it got operationalized as the financial sector of the Fed’s large-scale econometric model of the United States. The idea was to build an empirically calibrated model of how all the various possible levers of government policy affect variables of economic interest — such as output, employment, and capital investment — with a view to informing the use of those levers. monetary policy was to be conceived broadly as encompassing any and all levers involving the financial sector of the economy. It was the high point of the age of management, when the new Keynesian economic science seemed to show how it was possible, merely by manipulating a few strategic policy levers, to achieve the ambitious goals of the 1946 Employment Act.

In all the excitement, no one seems to have noticed that the monetary Walrasianism of Tobin amounts to nothing less than an apotheosis of the shiftability view of the nature of liquidity. Notwithstanding institutional rigidities on supply, all assets in Tobin’s model are assumed to be salable at a price determined by the balance of supply and demand. In effect, market liquidity is assumed for all assets equally. The demand for money is not a demand for the ultimate liquid asset but only a demand for the ultimate riskless asset, as all assets are assumed to be liquid.

Recall that the Arrow-Debreu version of Walras could not be used as a norm for monetary policy because there was no place for money in it. By contrast, the Marschak-Tobin version of Walras had something in it that could be labeled money, but it had a different problem, namely, abstraction from liquidity as the defining feature of money. The Marschak-Tobin idealization posed no immediate problem in the rigid institutional setting of the immediate postwar period; after all, the overwhelming majority of financial assets were government securities, and all of them were fully and equally liquid because of the Fed’s backstop. But once rigidities were relaxed and private capital markets restored, the abstraction from liquidity would become increasingly problematic.

Tobin himself hints at what was to come. In his 1969 model, all policy levers depend on institutional rigidities, so what happens if those rigidities are ever relaxed, as they would be over the coming decades? Tobin was presciently clear. “There would be no room for discrepancies between market and natural rates of return on capital, between market valuation and reproduction cost. There would be no room for monetary policy to affect aggregate demand. The real economy would call the tune for the financial sector, with no feedback in the other direction.” Thus, according to Tobin’s own model, once rigidities are relaxed the ideal norm would be realized without any policy intervention. Crucially, that ideal norm was the full market-clearing Walrasian equilibrium, a model with no place for money.

For the present story, the important point is that the ideal norm that emerges when all rigidities are relaxed is a world in which liquidity is a free good. In that ideal world everything, both real commodities and financial assets, is perfectly shiftable. This is the norm against which reality was to be measured and toward which policy intervention was to be directed. As a measure of how unreal and unreachable that norm was, it is sufficient to note that, in this ideal world, the EH holds true, even without any liquiditypremium add-factors. Nevertheless, that was the ideal and so it became the Fed’s job to achieve it. Since the ideal was a world in which liquidity is a free good, it seemed to follow that the job of the Fed was to supply liquidity as a free good.

A Dissenting View

In historical retrospect, we can appreciate that the Marschak-Tobin framework was not the only possible road forward; the road not taken was a reconstructed money view. Unlike the Marschak-Tobin model, the logic underlying classic central banking practice did not depend on any rigidities or inefficiencies. Instead, classic central banking literature developed a theory of bank rate management that rested ultimately on the role of the central bank as lender of last resort in times of crisis. Understanding the origin of crisis as a matter of the inherent instability of credit, classic central banking practice sought to intervene before the crisis to prevent buildup of speculative excess. True, American banking practice was different, and as a consequence American central banking practice came to focus more on open market operations than on discount operations. Nevertheless, the classic vision of the central bank’s problem and mission could have been adapted simply by focusing on the Fed’s ability to use its control over funding liquidity to influence market liquidity.

In the American academy, the most prominent voice that retained contact with the older traditions of central banking thought was that of Hyman Minsky, professor of economics at Washington University in St. Louis. 17 A student of Joseph Schumpeter at Harvard, Minsky placed at the center of his own thought Schumpeter’s idea that the capital development of the nation was crucially dependent on the organization and operation of the financial system, and particularly the banking system. Minsky’s first published paper, on “Central Banking and Money Market Changes” (1957) represents the earliest attempt by any academic to grapple with the implications of new money market instruments, specifically in the federal funds and repo markets, for the operational effectiveness of the Fed.

The elasticity provided by these markets would, Minsky suggested, ultimately undermine attempts to use monetary policy for aggregate stabilization, but would nevertheless leave in place the older and more fundamental central bank function as lender of last resort. Even as the American academic discussion was getting organized around a debate between monetarists (Milton Friedman) and Keynesians (James Tobin), Minsky was working to carve out a third position that reformulated classic British central banking practice for modern American conditions. Minsky’s source for understanding British practice was not Hawtrey but rather Richard Sayers, whose magisterial history Bank of England Operations, 1890–1914 explained how the vanished prewar system had worked. Nevertheless, Minsky’s conclusion, which he termed the “Financial Instability Hypothesis,” has more than a little flavor of Hawtrey insofar as the emphasis is on the inherent instability of credit.

The credit that concerned Minsky was not Hawtrey’s merchant trade credit but rather business investment credit, bank lending to finance capital investment spending by American business. When businesses borrow, they commit themselves to a stream of future payments, and their ability to make those payments depends on realizing a net positive cash flow from their investment. The solvency of a business depends on the balance between the current valuation of payment commitments (liabilities) and expected cash flows (assets). The liquidity of a business, however, depends on the match between the time pattern of those payment commitments and realized cash flows. The basic problem in a capital-using economy is that illiquidity is a fact of life, given that long-lived capital assets are the source of so much realized cash flow. Such capital assets generate cash flow only over an extended period of time, which means that liquidity is always a problem for the economy as a whole, and hence for each agent within the economy as well.

For Minsky, the inherent instability of credit is all about the shifting balance between cash commitments and cash flows. “Hedge” finance structures, in which promised future cash commitments are always less than realized cash inflows, are inherently stable; a business financed in such a way can never run into liquidity problems and therefore can focus its attention on other matters. The problem is that, over time, Hedge structures tend to be replaced by speculative and then Ponzi finance structures, in which firms promise payments that they cannot necessarily meet from concurrent cash flow. In good times, these more fragile finance structures cause no trouble; when the promised debts come due, they are just rolled over to a future date. But the fragility is there nonetheless, since any dislocation in the refinance mechanism can cause disruption. When the dislocation comes, the size of the resulting disruption depends on the prevalence of speculative and Ponzi finance structures that are vulnerable to such a dislocation.

Why the tendency toward fragility? The reason ultimately is the liquidity preference of wealth holders. Unwilling to tie their money up for the life of the real capital asset they are financing, investors are willing to accept a lower promised yield in order to acquire a shorter-dated financial asset. In this way, they essentially bribe borrowers to accept some part of the liquidity risk inherent in long-term capital investment (everything after the maturity of the note), while they take the rest (everything before the maturity of the note). For the lender, having made the loan, the only way to get his money back faster is to sell that loan to someone else; this is market liquidity. For the borrower, having committed to the series of payments specified in the loan, the only way to put off those payments when they come due is by paying someone else to make them, that is, by rolling debts as they come due; this is funding liquidity.

In Minsky’s thought, the tendency toward fragility comes from the interaction between asset valuation and creditworthiness. Whereas Hawtrey emphasized the valuation of inventories and the feedback of that valuation on the availability of trade credit, Minsky emphasized the valuation of capital assets and the feedback of that valuation on the availability of capital credit. Creditfinanced spending by one person creates income for others, both present and prospective income, and the capitalization of that income raises asset prices and thus improves creditworthiness for another round of spending. The resulting instability is more substantial than anything Hawtrey ever imagined, simply because the price of capital assets has a lot more room to move than the price of inventories. That wide range of price movement makes it also more difficult to control instability with mere interest rate policy.

In such a world, the best thing to do, according to Minsky, was to use collateral policy at the central bank’s discount window to discourage speculative financing structures and to encourage hedge financing structures. If people know that only hedge financing structures will be eligible for discount when a crisis comes, that will tend to mute somewhat the push toward fragility on the way up, and hence also the scale of the collapse on the way down. Just as the Fed’s operations in government debt had come to support the use of Treasury bills as a secondary reserve, so too would Fed operations in private hedge finance debt structures create a liquid market for that debt. In a crisis, that debt would move onto the balance sheet of the Fed, and once normalcy returned it would move back out.

That’s how the world looked to Minsky, but not to either the monetarists or the Keynesians, whose debate dominated the academic airwaves. Their debate was all about managing aggregate fluctuations around a determinate equilibrium, and not at all about stemming the inherent instability of credit. Minsky’s pessimism about the possibility of active management put him outside the optimistic Keynesian camp, and his pessimism about the inherent stability of a private credit economy put him outside the optimistic monetarist camp. Minsky’s views were continuous with the great tradition of central banking thought, but that continuity proved to be a disadvantage in a postwar era looking to put the past behind it and build anew.

And so, instead of a Minskian reworking of central banking verities for modern circumstances, we got active money management along monetary Walrasian lines. In times of rising prices the Fed tightened monetary policy, causing money market interest rates to rise, thus creating incentive for private liquidity provision. As a consequence we got financial innovations such as the certificate of deposit, bank commercial paper, and Eurodollar borrowing. Higher rates also meant that refinance of maturing speculative positions was achieved only by pledging even greater future cash payments, which meant that even successful refinance tended to increase fragility. Thus, the natural thrust toward fragility was amplified, not dampened, by the operations of the financial authority. In a typical cycle, eventually refinance would become impossible for the most overextended units and crisis would erupt, forcing central bank intervention to mop up the mess ex post.