Here is a presentation that I’ll give today at the University of Denver at the annual J. Fagg Foster honors ceremony. Most of you will not know of Foster, but you should. While he did not publish much, he was the professor of a number of prominent institutionalists who attended DU in the early postwar period. I was lucky to have studied with his student, Marc Tool, and was introduced to Foster’s work at the very beginning of my studies of economics. My presentation below is based on two of Foster’s articles: J. Fagg Foster (1981) “Understandings and Misunderstandings of Keynesian Economics”, JEI, vol XV, No 4, p. 949-957.; and (1981) “The Reality of the Present and the Challenge of the Future”, JEI vol XV, No 4, p. 963-968. Both are from 1966, republished in a special issue of the Journal of Economic Issues, 1981. You should read them.
Is this the age of Keynes? That’s the question raised by Fagg Foster in 1966.
In the 1960s the answer seemed obvious. Keynes dominated economics—or, at least, macroeconomics—and Keynesianism dominated policy. And it worked! Or, so most thought.
Foster wasn’t sure. While he agreed that “[t]here probably has been no instance in history in which a pattern of ideas has had so much effect on the everyday life of everyone in so short a time”, he thought most of Keynes’s followers misunderstood his theory.
Further, Foster wasn’t convinced the theory provided a firm basis for policy.
Finally, he lamented that “among all post-Keynesian economists, the institutionalists seem to have been least affected by Keynes’s theory…. The institutionalists have not even contemplated the possibility of any generic relationships between the Keynesian theory and their own.”
A decade later, so-called Keynesian economics was in disarray, a casualty of the apparent failure of policy to fine-tune the economy. Stagflation at the end of the 1970s delivered the final blow, and fueled the rise of increasingly preposterous approaches such as Rational Expectations, Real Business Cycle theory, the Efficient Markets Hypothesis and hence on to DSGE with a single representative agent standing in for the whole economy.
In truth, even in the heyday of Keynesianism, policy was directed to stimulate the sentiments of business undertakers—precisely what Keynes recommended against—with supply-side tax cuts and a cornucopia of subsidies to the captains of industry.
While a parallel approach developed calling itself New Keynesian, the only thing new was the adoption of the craziest “new” orthodox ideas (witness rational expectations). And the only thing “Keynesian” was the presumption that sticky wages and prices prevent instantaneous market clearing—which was actually the old Neoclassical explanation of unemployment that Keynes had dispatched.
With friends like these, Keynes doesn’t need enemies.
In retrospect, Foster might have been a bit hard on the institutionalists. While it took them a while to embrace Keynes, I think it would be accurate to claim that most today accept Keynesian macro theory as their own.
Still, there has been some discomfort with Keynes’s theory—or just about any theory—on the argument that theory is too mechanical, that it abstracts too much from institutions.
It is true that in the hands of the “Keynesians”, the whole argument became a simple arithmetic problem (multiply the inverse of the marginal propensity to save by autonomous spending).
But Foster argued that Keynes’s own exposition was dynamic, general, and “open”:
“[g]enerality is attained by identifying income as the instantaneous concomitant of any particular combination of the whole range of possible relationships between the three independent variables. Since the theory is equally applicable to any, it is applicable to all patterns of relationships between the propensity to consume, the marginal efficiency of capital, and the rate of interest.”
This makes Keynes’s theory “open-ended in the sense in which that quality is associated with scientific theory as such…The theory is…subject to indefinite development.”
By contrast, neoclassical theory is “tautological”—closed and unscientific.
Let me give two examples. In General Equilibrium theory, the goal was to find a vector of equilibrium prices that would clear all markets given assumptions about behavior, tastes, and endowments.
By contrast, Keynes’s theory allows for equilibrium (of a different sort) at any level of output and employment, with the point of effective demand determined by the three independent variables: PC, MEK and interest rate (LP).
To put it another way, firms hire the amount of labor they think they need to produce the amount of output they think they can sell. In this way, expectations and hiring decisions of business undertakers determine the point of effective demand, which need not clear any market.
As another example, Keynes distinguishes between the Neoclassical “natural rate” of interest and his own “neutral rate”. In Neoclassical theory, market forces produce a unique rate of interest that equilibrates saving (a function of the rate of time preference) and investment (a function of the marginal productivity of capital).
This ensures that Say’s Law operates—which is the low brow Neoclassical analogue to the Arrow-Debreu general equilibrium.
Keynes argues that saving equals investment regardless of the rate of interest, hence, any interest rate is “natural” in the Neoclassical sense—and each different natural rate has its own corresponding level of effective demand. However, only one interest rate is consistent with full employment–the “neutral rate”.
His system is “open” in the sense that any interest rate and any level of effective demand is possible. The Neoclassical system is “closed” because the assumptions ensure there is a unique “price” (interest rate) that clears the loanable funds market.
As Foster argues about the GT,
“almost any theory of the rate of interest which does not involve the rate of money savings as a determinant of the rate of interest would be compatible with the Keynesian general theory…. In any event, the three independent variables are presented as institutionally determined….they can, for example, be affected by public policy.”
Somewhat remarkably, Foster recognizes Keynes’s Chapter 17 theory of asset prices. I say “remarkably” because few economists recognized what Keynes was trying to do.
Most look to Chapters 13 and 15, in which the interest rate is determined by “money supply and money demand”. That conventional approach lead to Hicks’s fundamentally flawed ISLM analysis.
Some of the followers of Keynes (Townshend, Minsky, Davidson, Kregel) insist we should look at Chapter 17 instead.
Further, as especially Kregel has pointed out, the S&D approach is subject to Sraffa’s critique of the whole Marshallian edifice. That is doubly ironic since Keynes had prodded Sraffa to produce the criticism—aimed at Hayek’s capital theory—and had together with Sraffa developed the commodity “own rate” approach of Chapter 17.
Foster recognized the generality of Ch17: “Even in the absence of money, judgments would have to be made about whether to add to inventory and what the components of the inventory should be. All commodities have their ‘own rates’.”
Keynes presented a liquidity preference theory of own rates, with a preference for liquidity institutionally determined as a pattern of “correlated human behavior”.
In my own work I provide an institutionalist alternative to the orthodox approach—which is based on a money supply fixed by the central bank and a downward sloping money demand that together determine “the” interest rate.
Post Keynesians turned this on its head, making the money supply “horizontal”. The central bank accommodates reserve demand, and banks accommodate loan demand. The money supply is “endogenous”, interest rates are “exogenous”.
While this is an improvement, it is not very satisfying. I won’t go into my critique of Horizontalism. Instead, I want to hold it up to Foster’s critique of “Keynesians”. Is this “general”? Is it “scientific”? Is it “institutional”?
Many institutionalists have argued that money is an institution; Dillard argued that it might be the most important institution in the capitalist economy. Yet, most economists, including institutionalists, identify things as money: wampum, shells, metal coins, paper notes, demand deposits.
How can a thing be an institution? As Foster argues, “the very word institution connotes patterns of correlated human behavior; it does not pertain to nonhuman phenomena.” Things are not human behavior!
What do the things that many people call money have in common? Most economists, including institutionalists, identify money as something we use in exchange.
In The Treatise, Keynes began with the money of account in which we denominate debts and credits, and, yes, prices. Following Knapp he argues that for the past 4000 years the money of account has been chosen by the state authorities.
Keynes’s generalization holds up well, as there are few exceptions to the rule that monies of account are chosen by authorities.
Units of measurement are necessarily social constructions. I can choose my own idiosyncratic measuring units for time, space, and value, but they must be socially sanctioned for wide adoption.
This is particularly true when valuing heterogeneous things that share no obvious physical characteristics. It is relatively easy to develop weight units—typically an important grain food is counted out to obtain the weight measure—or length units (the King’s foot will do).
But money value is a conceptual leap. We know the early money units came from the weight measures, but that leap is still difficult because the money values of a cow versus a basket of fruit cannot be obtained by their weight equivalent in barley grains.
The earliest money units were equivalent to a month’s grain ration, and authorities measured and established the money value of other things in posted price lists.
So, one commonality is that all monies are measured in a money of account. All those things economists declare to be money are denominated in the money of account. But the nature of money must amount to more than that if money is an institution.
We have the inch that we use to measure the length of a couch, but no one claims that a couch is an institution, even though we certainly agree the inch is a social convention.
Markets are institutions, and while Neoclassical economists usually assume one kind of market (the famous Auctioneer taking bids and offers), we know there are many types. Many economists identify money as that which is used to intermediate market exchange. But that seems to reduce money to a thing we agree to use to intermediate exchange in the market—rather than an institution in its own right.
What is the institutional nature of those money things? The most obvious shared characteristic of some of them is that they are evidence of debt: coins and central bank notes are government debts; bank notes or deposits are bank debts; and we can expand our definition of money things to include shares of money market mutual funds, etc.
If we go back through time, we find wooden tally sticks issued by European monarchs as evidence of debt (notches recorded money amounts).
According to Knapp we can even view the claim ticket issued by the coat check attendant as a debt that is redeemed by returning one’s coat.
Note that use of cowry shells or the huge stone “wheels” of the Yap islanders or the tobacco leaves of early America would not seem to fit—they seem to be “things”, not “records” of money debt. However, as Forstater has shown, the cowry shells used as money things actually were issued by authorities as a debt that could be used to pay taxes.
And the stone wheels were converted from ceremonial “primitive valuables” to “money-like” things by German colonizers who “seized” them by painting an “X” on them and requiring the Yap people work to earn their return.
The American tobacco example seems to come close to barter exchange of things, although with the money value of tobacco administered by the authorities. Hence, it was not really a barter of things, but rather an exchange of money values.
What we have, then, is a socially created and generally accepted money of account, and debts that are denominated in that money of account. Within a modern nation, most socially sanctioned money-denominated debts are written in the nation’s money of account.
Some kinds of money things “circulate”, used in exchange and other payments (ie paying down one’s own debts). The best examples are currency (debt of treasury and central bank) and demand deposits (debt of banks). Why do we accept these in payment?
It has long been believed that we accept currency because it is made of precious metal or redeemable for metal—we accept them for their “thing-ness”. In truth, coined precious metal almost always circulated well beyond the value of embodied metal; and redeemability of currency for gold at a fixed rate has been the exception not the rule.
Hence, most economists recognize that currency is today (and in the past) “fiat”.
One could concoct a story about why someone would accept another’s “fiat” debt denominated in the national money of account. I will accept your IOU if you promise to redeem it later for something I want—a commodity, a service, someone else’s IOU.
I might be able to pass off your IOU before redemption day if I can find someone willing to accept it. Some IOUs might be easier to pass along, and could become generally recognized media of exchange and payment.
That is the typical story. It usually takes the form of an infinite regress: I accept BillyBob’s IOU because I think BuffySue will take it; she accepts it because she believes she can dupe some other dope to take it.
This is, I suppose, an institutional approach: the institution is “trust” or “delusion” (depending on how well it works). But it is hard for me to believe that such an ephemeral social relation underlies “the most important institution of the capitalist economy”.
When this is applied to the sovereign’s currency, it is said that we accept it both because it is “legal tender” and because we think there are plenty of dopes who will take it. However, sovereign currencies are accepted without legal tender laws, and are refused with them—as Knapp put it, those laws are little more than “pious hope”.
The dupe a dope explanation relies on weak institutions—perhaps even weaker trust and more fragile delusion than in the case of IOUs of some private issuers. (Enter Bitcoin!)
The argument of Smith, Knapp, Innes, Keynes, Grierson, and Lerner is that currency will be accepted if there is an enforceable obligation to make payments to its issuer in that same currency. Hence, “taxes drive money” in the sense that the state can impose tax liabilities and issue the means of paying those liabilities in the form of its own liabilities.
Here there is an institution, or a set of institutions, that we can identify as “sovereignty”. As Keynes said, the sovereign has the power to declare what will be the unit of account—the Dollar, the Lira, the Pound, the Yen. The sovereign has the power to impose fees, fines, and taxes, and to name what it will accept in payment. When the fees, fines, and taxes are paid, the currency is “redeemed”—accepted by the sovereign.
While sovereigns sometimes “redeem” their currency for precious metal or foreign currency, that is not necessary. The agreement to “redeem” currency in payment of taxes, fees, tithes and fines is sufficient to “drive” the currency—to create a demand for it.
While it could be true I’m more willing to accept the state’s IOUs if I know I can dupe a dope, I will definitely accept it if I must pay taxes in the state’s currency.
This is the sense in which we say “taxes are sufficient to create a demand for the currency”.
There are other reasons to accept a currency—maybe Biff will accept it, maybe I can exchange it for gold or foreign currency, maybe I can hold it as a store of value. These uses derive from the obligations (such as taxes, fees, tithes, and fines).
Innes posed a fundamental “law” of credit: the issuer of an IOU must accept it in payment. This the principle of redeemability. If one expects she will need to make payments to some entity, she will want to obtain the IOUs of that entity. This goes part way to explaining why the IOUs of non-sovereign issuers can be widely accepted.
Acceptability can be increased by promising to convert on demand one’s IOUs to more widely accepted IOUs. This is called leveraging and while it sounds similar to the deposit multiplier there is no simple, fixed ratio of leverage.
Stephanie Bell/Kelton, Duncan Foley, and Minsky have all used the metaphor of a pyramid of liabilities, with those lower in the pyramid leveraging those higher in the pyramid, and with the sovereign’s liabilities at the apex.
The power of the sovereign brings me back to Foster’s other great piece, with the same title as this presentation. It contains the most profound economics statement:
“Whatever is technically feasible is financially possible. To the perpetual question ‘Where is the money coming from?’ the answer is now clear. It comes from the only two institutions we permit to create money funds: the treasury of the sovereign government and commercial banks. And the rate at which we permit either to create funds is pretty much a matter of public policy.”
He argues that there is no theoretical limit to the ability to create funding, so “the only question is should they be made available.” Finance is not a scarce resource. The state cannot run out of its own money (currency).
He goes through a list of technically feasible projects from a 1966 Govt report, including:
* guaranteed minimum income of $3000 for every family in the US
* provision of 14 years of free education to every qualified person
* providing jobs to “all of those unable to find employment elsewhere”.
The report “suggested institutional adjustments to implement its recommendations. In all of this, there is no finding of financial incapacity. The only question is, again, whether these recommendations should be adopted, since they are technically feasible.”
These are the policy conclusions that MMT reached in the beginning. We have provided close institutional analysis of monetary and fiscal operations to back-up Foster’s claim.
I’ve been surprised at the reaction of some institutionalists—who argue that ability to finance projects–such as the employer of last resort program–is limited by operational rules of behavior. Today’s institutions constrain financial feasibility.
But Foster argued, “the very word institution connotes patterns of correlated human behavior” and “all answers to all social problems take the form of institutional modifications”. Any social theory that assumes static institutions is “without significance”.
Those who argue against Foster’s claim that society can afford desirable and technologically affordable program because there exist constraining institutions are unscientific (like Neoclassical economics). The general, open, scientific statement is that government can financially afford full employment.
The institutional barriers can be changed, and any pronouncement on what is not possible that presumes static institutions is “without significance”. The answer to the problem of unemployment is institutional modification.
Let us turn to Foster’s view on economic growth. The mainstream view is that growth is promoted through thrift—more savings allow for more investment. While no one who has read Keynes’s GT should fall into the paradox of thrift trap, almost all do.
Foster adds, however, that because the propensity to consume is less than one, the “level of economic activity cannot remain constant; it either increases or decreases. The level of income and employment either contracts or expands.” I won’t go into that in detail, but it is the Domar problem that Vatter and Walker verified in a number of books.
In recent months, the mainstream has rediscovered secular stagnation, as both Krugman and Larry Summers warned that we cannot find a stable source of demand to keep the economy growing; permanently higher unemployment is the new normal.
Summers went on to proclaim that stagnation has been relieved over the past three decades only by speculative financial bubbles (dotcoms, stocks, commodities, houses). Bubbles are all we’ve got; Bubbles R Us.
Foster, however, argues “We have learned to use autonomous investment (mostly government purchases of public capital) to counteract any deficiency in ordinary induced investment, and directly in proportion to our astuteness in this regard we find our previous problems of prolonged recession less serious.”
Our policymakers over the past three decades have been considerably less “astute” than they were in the early postwar period. We suffer stagnation and crumbling public physical and social infrastructure and massive quantities of idle resources.
And, yet, it is said that Uncle Sam is “broke”; he cannot afford infrastructure; he cannot afford to put the unemployed to work. Spending more today would burden future generations with debt. We need to reduce spending today to increase saving for the future. The only hope for a rising tide of baby boomer retirees is to increase saving now.
We close with another of Foster’s brilliant statements: “Aggregate income cannot be transferred from one time period to another.” (ibid p. 967) He goes on to explain:
This means that one generation cannot implement or diminish the income of future generations by expending less or more than its own means…. The community at large cannot “save money”; it can save only by investing, and its savings are constituted by that investment. The instantaneity of the equation between aggregate sales and aggregate purchases forces the instantaneity of the equation between saving and investment. … [o]ne can see why we have come to use the federal government to incur debt without accumulating equivalent assets in order that the rest of us can accumulate unobligated liquidity—so that the rest of us can save money. (ibid pp. 967-8)
An entire course in macroeconomics is contained in that paragraph.
Foster destroys the loanable funds argument and the “intergenerational warriors” like Kotlikoff (who claims we cannot “afford” retiring baby boomers—and calculates a looming Federal Government shortfall of $200 trillion due to unfunded commitments).
Foster clears up the confusion caused by the mechanical investment multiplier (in which investment raises income through a series of steps so that saving only gradually rises to equality with investment after some indeterminate time period).
And he clearly distinguishes between the options available to individuals versus the adding-up constraints at the macro level—in other words, the fallacies of composition that befall all of mainstream macro theory.
Uncle Sam’s spending is our income. His deficits are our surpluses. His debt is our net financial wealth. He can financially afford to put us to work to mobilize our resources for progress, rather than for decay.
There is nothing inevitable about stagnation. There is nothing inevitable that makes growth environment-destroying. We need to adjust the institutions that have encouraged growth to destroy our environment.
We have some of the technological know-how already, and we can develop more of it. What we already know how to do is financially affordable; what we learn how to do in the future will also be affordable.
The reality of the present is that policy is failing us. The challenge of the future is to recognize that through application of Keynes’s “scientific theory”, “the quality of a much greater part of human life will be in man’s own hands.”