Debunking Economics

Misunderstanding the Crisis

Misunderstanding the Great Depression and the Great Recession

Steve Keen

Bernanke’s Essays on the Great Depression (Bernanke 2000) is near the top of my stack of books that indicate how poorly neoclassical economists understand capitalism. Most of the other are books of pure theory, like Debreu’s Theory of Value (Debreu 1959), or textbooks like Varian’s Microeconomic Analysis (Varian 1992). Bernanke’s distinguished itself by being empirical: he was, he claimed, searching the data to locate the causes of the Great Depression, since:

To understand the Great Depression is the Holy Grail of macroeconomics. Not only did the Depression give birth to macroeconomics as a distinct field of study, but also—to an extent that is not always fully appreciated—the experience of the 1930s continues to influence macroeconomists’ beliefs, policy recommendations, and research agendas. And, practicalities aside, finding an explanation for the worldwide economic collapse of the 1930s remains a fascinating intellectual challenge. (Bernanke 2000, p. 5)

However, what Bernanke was actually doing was searching for an explanation that was consistent with neoclassical theory. Statements to this effect abound throughout the Essays, and they highlight the profound difficulty he faced—since according to neoclassical theory, events like the Great Depression should not occur. This disconnection between reality and neoclassical theory had at least the following manifestations that Bernanke admitted to in his Essays:

• Monetary variables affect inflation, but are not supposed to affect real variables—money is supposed to be “neutral”:

Of course, the conclusion that monetary shocks were an important source of the Depression raises a central question in macroeconomics, which is why nominal shocks should have real effects. (p. 7)

the gold standard theory leaves unsolved the corresponding “aggregate supply puzzle,” namely, why were the observed worldwide declines in nominal aggregate demand associated with such deep and persistent contractions in real output and employment? Or, in the language of contemporary macroeconomics, how can we explain what appears to be a massive and very long-lived instance of monetary nonneutrality? (p. 277)

• A prolonged macro downturn is inconsistent with rational micro behavior:

my theory … does have the virtues that, first, it seems capable of explaining the unusual length and depth of the depression; and, second, it can do this without assuming markedly irrational behavior by private economic agents. Since the reconciliation of the obvious inefficiency of the depression with the postulate of rational private behavior remains a leading unsolved puzzle of macroeconomics, these two virtues alone provide motivation for serious consideration of this theory. (p. 42; emphasis added)

• Rational behavior by agents should lead to all prices—including money wages—adjusting rapidly to a monetary shock, so that its impact should be transient:

slow nominal-wage adjustment (in the face of massive unemployment) is especially difficult to reconcile with the postulate of economic rationality. We cannot claim to understand the Depression until we can provide a rationale for this paradoxical behavior of wages. (p. 7)

• Rapid adjustment of prices should bring the economy back to equilibrium:

the failure of nominal wages (and, similarly, prices) to adjust seems inconsistent with the postulate of economic rationality (p. 32; emphases added)

Bernanke began well when he stated that the causes of the Great Depression had to lie in a collapse in aggregate demand—though even here he manifested a neoclassical bias of expecting capitalism to rapidly return to equilibrium after any disturbance:

Because the Depression was characterized by sharp declines in both output and prices, the premise of this essay is that declines in aggregate demand were the dominant factor in the onset of the Depression.

This starting point leads naturally to two questions: First, what caused the worldwide collapse in aggregate demand in the late 1920s and early 1930s (the “aggregate demand puzzle”)? Second, why did the Depression last so long? In particular, why didn’t the “normal” stabilizing mechanisms of the economy, such as the adjustment of wages and prices to changes in demand, limit the real economic impact of the fall in aggregate demand (the “aggregate supply puzzle”). (Bernanke 2000, p. ix)

However, from this point on, his neoclassical priors excluded both salient data and rival intellectual perspectives on the data. His treatment of Hyman Minsky’s “Financial Instability Hypothesis”—which is outlined in Chapter 13—is particularly reprehensible. In the entire volume, there is a single, utterly dismissive reference to Minsky:

Hyman Minsky (1977) and Charles Kindleberger (1978) have in several places argued for the inherent instability of the financial system but in doing so have had to depart from the assumption of rational economic behavior.. [A footnote adds] I do not deny the possible importance of irrationality in economic life; however it seems that the best research strategy is to push the rationality postulate as far as it will go. (Bernanke 2000, p. 43)

As we shall see, this is a parody of Minsky’s Hypothesis. He devoted slightly more space to Irving Fisher and his debt-deflation theory, but what he presented was likewise a parody of Fisher’s views, rather than a serious consideration of them:

The idea of debt-deflation goes back to Irving Fisher (1933). Fisher envisioned a dynamic process in which falling asset and commodity prices created pressure on nominal debtors, forcing them into distress sales of assets, which in turn led to further price declines and financial difficulties. His diagnosis led him to urge President Roosevelt to subordinate exchange-rate considerations to the need for reflation, advice that (ultimately) FDR followed.

Fisher’s idea was less influential in academic circles, though, because of the counterargument that debt-deflation represented no more than a redistribution from one group (debtors) to another (creditors). Absent implausibly large differences in marginal spending propensities among the groups, it was suggested, pure redistributions should have no significant macro-economic effects… (Bernanke 2000, p. 24)[1]

There are many grounds on which this is a misrepresentation of Fisher,[2] but the key fallacy is the proposition that debt has no macroeconomic effects. From Bernanke’s neoclassical perspective, debt merely involves the transfer of spending power from the saver to the borrower, while deflation merely increases the amount transferred, in debt servicing and repayment, from the borrower back to the saver. Therefore, unless borrowers and savers have very different propensities to consume, this transfer should have no impact on aggregate demand.

The contrast with the theoretical case that Marx, Schumpeter, Keynes and Minsky made about debt and aggregate demand (see pages 228-230) could not be more stark—and in the next chapter I’ll make the empirical case that a collapse in debt-financed demand was the cause of both the Great Depression and the Great Recession. Bernanke’s neoclassical goggles rendered him incapable of comprehending the best explanations of the Great Depression, and led him to ignore the one data set that overwhelmingly explained the fall in aggregate demand and the collapse in employment.

The three reasons he ultimately provided for the Great Depression were (a) that it was caused by the then Federal Reserve’s mismanagement of the money supply between 1928 and 1931 (b) that the slow adjustment of money wages to the fall in aggregate demand is what made it last so long; and (c) that the Gold Standard transmitted the collapse internationally. His conclusion on the first point was emphatic:

there is now overwhelming evidence that the main factor depressing aggregate demand was a worldwide contraction in world money supplies. This monetary collapse was itself the result of a poorly managed and technically flawed international monetary system (the gold standard, as reconstituted after World War I). (Bernanke 2000, p. ix)

He was also emphatic about his “smoking gun”: the Great Depression was triggered by the Federal Reserve’s reduction of the US base money supply between June 1928 and June 1931:

The monetary data for the United States are quite remarkable, and tend to underscore the stinging critique of the Fed’s policy choices by Friedman and Schwartz (1963)… the United States is the only country in which the discretionary component of policy was arguably significantly destabilizing… the ratio of monetary base to international reserves … fell consistently in the United States from … 1928:II … through the second quarter of 1931. As a result, U.S. nominal money growth was precisely zero between 1928:IV and 1929:IV, despite both gold inflows and an increase in the money multiplier.

The year 1930 was even worse in this respect: between 1929:IV and 1930:IV, nominal money in the United States fell by almost 6 [percent], even as the U.S. gold stock increased by 8 [percent] over the same period. The proximate cause of this decline in M₁ was continued contraction in the ratio of base to reserves, which reinforced rather than offset declines in the money multiplier. This tightening seems clearly inconsistent with the gold standard’s “rules of the game,” and locates much of the blame for the early (pre-1931) slowdown in world monetary aggregates with the Federal Reserve. (p. 153)

There are four problems with Bernanke’s argument, in addition to the fundamental one of ignoring the role of debt in macroeconomics. Firstly, as far as smoking guns go, this is a pop-gun, not a Colt 45. Secondly, it has fired at other times since WWII (once in nominal terms, and many times when adjusted for inflation) without causing anything remotely like the Great Depression. Thirdly, a close look at the data shows that the correlations between changes in the rate of growth of the money supply[3] and unemployment conflict with Bernanke’s argument that mismanagement of the monetary base was the causa causans of the Great Depression. Fourthly, the only other time that it has led to a Great-Depression-like event was when Bernanke himself was Chairman of the Federal Reserve.

Between March 1928 and May 1929, Base Money fell at an average rate of just over 1 percent per annum in nominal terms, and a maximum rate of minus 1.8 percent.[4] It fell at the same rate between 1948 and 1950, and coincided with a garden-variety recession, rather than a prolonged slump: unemployment peaked at 7.9 percent and rapidly returned to boom levels of under 3 percent. So the pop-gun has fired twice in nominal terms, and only once did it “cause” a Great Depression.

It could also be argued, from a Neoclassical perspective, that the Fed’s reduction in Base Money in the lead-up to the Great Depression was merely a response to the rate of inflation, which had turned negative in mid-1924. Neoclassical theory emphasizes money’s role as a means to facilitate transactions, and a falling price level implies a need for less money. On this point Milton Friedman, whom Bernanke cited as a critic of the Federal Reserve for letting Base Money fall by 1 percent per annum, argued elsewhere that social welfare would be maximized if the money supply actually fell by 10 percent per year.[5]

When the inflation-adjusted rate of change of Base Money is considered, there were numerous other periods where Base Money fell as fast as in 1928-29, without leading to a Depression-scale event. The average inflation-adjusted rate of growth of M₀ in mid-1928 to mid-1929 was minus 0.5 percent, and even in 1930 M₀ fell by a maximum of 2.2 percent per annum in real terms. There were six occasions in the post-WWII period where the real rate of decline of M₀ was greater than this without causing a Depression-like event[6] (though there were recessions on all but one occasion). Why did the pop-gun fire then, but emit no smoke?

The reason is, of course, that the pop-gun wasn’t really the guilty culprit in the crime of the Great Depression, and Friedman and Bernanke’s focus upon it merely diverted attention from the real culprit in this investigation: the economy itself. Capitalism was on trial because of the Great Depression, and the verdict could well have been attempted suicide—which is the last verdict that neoclassical economists could stomach, because they are wedded to the belief that capitalism is inherently stable. They cannot bring themselves to consider the alternative perspective that capitalism is inherently unstable, and that the financial sector causes its most severe breakdowns.

To neoclassicals like Friedman and Bernanke, it was better to blame one of the nurses for incompetence, than to admit that capitalism is a manic-depressive social system that periodically attempts to take its own life. It was better to blame the Fed for not administering its M₀ medicine properly, than to admit that the financial system’s proclivity to create too much debt causes capitalism’s periodic breakdowns.

It is therefore a delicious if socially painful irony that the only other time that the pop-gun fired and a Depression-like event did follow was when the Chairman of the Federal Reserve was one Ben S. Bernanke.

Bernanke began as Chairman on February 1, 2006, and between October 2007 and July 2008, the change in M₀ was an inflation-adjusted minus 3 percent—one percent lower than its steepest rate of decline in 1930-33. The rate of change of M₀ had trended down in nominal terms ever since 2002, when the Greenspan Fed had embarked on some Quantitative Easing to stimulate the economy during the recession of 2001. Then, M₀ growth had turned from minus 2 percent nominal (and minus 6 percent real) at the end of 2000 to plus 11 percent nominal (and 8 percent real) by July 2001. From there it fell steadily to 1 percent nominal–and minus 3 percent real—by the start of 2008.

Whatever way you look at it, this makes a mockery of the conclusion to Bernanke’s fawning speech at Milton Friedman’s 90^th birthday party in November 2002:

Let me end my talk by abusing slightly my status as an official representative of the Federal Reserve. I would like to say to Milton and Anna: Regarding the Great Depression. You’re right, we did it. We’re very sorry. But thanks to you, we won’t do it again. (Bernanke 2002)

Either Bernanke forgot what he learnt from Friedman and his own research once in office—since Friedman and Bernanke’s criticism of the 1920’s Fed was that it let the growth rate of M₀ drop too low before the crisis—or the advice itself was irrelevant. The latter is of course the case. As I argue in the next chapter, the key to preventing Depressions is to prevent an explosion in the ratio of private debt to GDP, so that debt-financed demand cannot reach a level from which its collapse will trigger a Depression. Far from explaining what caused the Great Depression, Friedman and Bernanke’s simplistic perspective diverted attention from the real culprit—the expansion of private debt by the banking sector—and ignored the enormous growth of debt that occurred while the Central Bank was under the thrall of neoclassical economics.

The relative irrelevance of changes in base money as a cause of changes in unemployment, let alone a cause of serious economic breakdown, can be gauged by looking at the correlation between the growth of M₀ and the rate of unemployment over the period from 1920 till 1940—across both the boom of the Roaring Twenties and the collapse of the Great Depression (see Figure 92). If too slow a rate of growth of M₀ can trigger a Depression, as Bernanke asserts, then surely there should be a negative correlation between the change in M₀ and the rate of unemployment: unemployment should fall when the rate of change of M₀ is high, and rise when it is low.

The correlation is has the right sign for the period from 1920 till 1930 (minus 0.22 in for changes in nominal M0 and minus 0.19 after inflation) but the wrong one for the period from 1930 till 1940 (plus 0.28 for nominal M0 and 0.54 after inflation), and it is positive for the entire period 1920-1940 (plus 0.44 for nominal change to M0, and 0.61 for the inflation-adjusted rate of change). Therefore unemployment increased when the rate of growth of M₀ increased, and fell when it fell. Lagging the data on the basis that changes in M₀ should precede changes in unemployment doesn’t help either—the correlation remains positive.

Figure 92: Change in M0 and Unemployment 1920-1940

On the other hand, the correlation of changes in M₁ to unemployment is negative as expected over both the whole period (minus 0.47 for nominal change and minus 0.21 for inflation adjusted change) and the sub-periods of the Roaring Twenties (minus .31 for nominal M₁ and 0.79 for inflation-adjusted) and the Great Depression (minus 0.62 for nominal and 0.31 for real). So any causal link relates more to private-bank-driven changes in M₁ than to Central-Bank-driven changes in M₀.

There are only two interpretations of this, neither of which support the case that Bernanke made against the 1920s Fed.

The first is that, far from changes in M₀ driving unemployment, the unemployment rate drives changes in M₀. The Fed largely ignored the level of unemployment when it was low (during the 1920s), but went into panic policy mode when it exploded during the Great Depression. It therefore increased the level of M₀ when unemployment rose, and decreased it when unemployment seemed to be falling. The causation between changes in M₀ and unemployment is therefore the reverse of the one Bernanke sought to prove.

The second is that other factors are far more important in determining the rate of unemployment—and by extension, causing Great Depressions as well—than the Fed’s quantitative monetary policy. Two hints that the private financial system was the culprit are given by the negative relationship between changes in M₁ and unemployment, and by the fact that the relationship of M₀ to M₁ shifted dramatically when the Great Depression hit.

Before the Great Depression, there was a positive relationship between changes in M₀ and changes in M₁, and changes in M₀ appeared to lead changes in M₁ by about 1-2 months. This is the direction of causation expected by the conventional model of money creation—the “Money Multiplier”—which argues that commercial banks need reserves in order to be able to lend, though the magnitude is lower than might be expected

After the Great Depression, this relationship broke down completely, and changes in M₁ appeared to lead changes in M₀ by up to 15 months. This contradicts the conventional theory—a point I elaborate upon shortly.

So Bernanke’s analysis of what caused the Great Depression is erroneous, and to make matters worse, he didn’t even follow his own advice prior to the Great Recession when he Chairman of the Federal Reserve. But he certainly took his own analysis seriously after the Great Recession began—increasing M₀ as never before in an attempt to turn deflation into inflation.

After the Great Recession: Bernanke to the rescue?

Bernanke foreshadowed that he might do this in a speech for which he gained the nickname “Helicopter Ben” in 2002. With the unfortunate title of “Deflation: Making Sure “It” Doesn’t Happen Here”, it proved to be remarkably unprescient in terms of the economic future, since the US did slip into deflation. But the speech accurately signaled what he did do, once what he had hoped to avoid actually occurred:

Like gold, U.S. dollars have value only to the extent that they are strictly limited in supply. But the U.S. government has a technology, called a printing press (or, today, its electronic equivalent), that allows it to produce as many U.S. dollars as it wishes at essentially no cost. By increasing the number of U.S. dollars in circulation … the U.S. government can also reduce the value of a dollar in terms of goods and services, which is equivalent to raising the prices in dollars of those goods and services. We conclude that, under a paper-money system, a determined government can always generate higher spending and hence positive inflation…

Normally, money is injected into the economy through asset purchases by the Federal Reserve. To stimulate aggregate spending when short-term interest rates have reached zero, the Fed must expand the scale of its asset purchases or, possibly, expand the menu of assets that it buys. Alternatively, the Fed could find other ways of injecting money into the system—for example, by making low-interest-rate loans to banks or cooperating with the fiscal authorities. Each method of adding money to the economy has advantages and drawbacks, both technical and economic. One important concern in practice is that calibrating the economic effects of nonstandard means of injecting money may be difficult, given our relative lack of experience with such policies. Thus, as I have stressed already, prevention of deflation remains preferable to having to cure it. If we do fall into deflation, however, we can take comfort that the logic of the printing press example must assert itself, and sufficient injections of money will ultimately always reverse a deflation. (Bernanke 2002)

In late 2008, Bernanke turned on the printing presses as never before. In late 2008, he doubled base money in a mere 5 months, when the previous doubling had taken 13 years.

In inflation-adjusted terms, he expanded M₀ at a rate of over 100 percent a year, when its average annual rate of growth for the preceding 5 decades was 2.3 percent. By the time Bernanke finally took his foot off the M₀ accelerator one and a half years later, base money had jumped from $850 billion to $2.15 trillion (see Figure 98).

Figure 98: The volume of base money in Bernanke's "Quantitative Easing" in historical perspective

There is little doubt that this massive, unprecedented injection of base money did help reverse the deflation that commenced very suddenly in 2008, when inflation fell from plus 5.6 percent in mid-2008 to minus 2.1 percent a year later—the sharpest fall in inflation in post-WWII history. But I expect Bernanke was underwhelmed by the magnitude of the change: inflation rose from minus 2.1 percent to a peak of 2.7 percent, and it rapidly fell back to a rate of just 1 percent. That is very little inflationary bang for a large amount of bucks.

According to the conventional model of money creation—known as the “Money Multiplier”—this large an injection of government money into the reserve accounts of private banks should resulted in a far larger sum of bank-created money being added to the economy—as much as $10 trillion. This amplification of Bernanke’s $1.3 trillion injection should have rapidly revived the economy—according to neoclassical theory. This is precisely what President Obama, speaking no doubt on the advice of his economists, predicted when he explained the strategy they had advised him to follow, twelve weeks after he took office:

And although there are a lot of Americans who understandably think that government money would be better spent going directly to families and businesses instead of banks – “where’s our bailout?,” they ask – the truth is that a dollar of capital in a bank can actually result in eight or ten dollars of loans to families and businesses, a multiplier effect that can ultimately lead to a faster pace of economic growth. (Obama 2009, p. 3; emphasis added.)

Only that isn’t what happened. The dramatic increase in bank reserves spurred only a tiny increase in money in circulation: the 110 percent growth rate of M₀ resulted in only a 20 percent rate of growth of M₁.

The difference in growth rates was so great so that there is now less money in check accounts and currency in circulation than there is money in the reserve accounts of the commercial banks.

The “eight or ten dollars of loans to families and businesses” from each extra “dollar of capital in a bank” simply didn’t happen. What went wrong?

The Mythical Money Multiplier

Few concepts are more deserving than the “Money Multiplier” of Henry Menchen’s aphorism that “Explanations exist; they have existed for all time; there is always a well-known solution to every human problem—neat, plausible, and wrong”.[7]

In this model, money is created in a two stage process. Firstly, the government creates “fiat” money, say by printing dollar bills and giving them to an individual. The individual then deposits the dollar bills in her bank account. Secondly, the bank keeps a fraction of the deposit as a reserve, and lends out the rest to a borrower. That borrower then deposits this loaned money in another bank account, and the process repeats.

Let’s say that the amount created by the government is $100, the fraction the banks keep as a reserve (known as the “Reserve Requirement” and set by the government or Central Bank) is 10 percent, and it takes banks a week to go from getting a new deposit to making a loan. The process starts with the $100 created by the government. One week later, the first bank has created another $90 by lending 90 percent of that money to a borrower. A week later, a second bank creates another $81—by keeping $9 of the new deposit in reserve and lending out the other $81. The process keeps on going so that, after many weeks, there will be $1,000 created, consisting of the initial printing of $100 by the government, and $900 in credit money created by the banking system—which is matched by $900 in additional debt. There will be $900 of credit-money in circulation, facilitating trade, while another $100 of cash will be held by the banks in reserve (see Table 17).

Table 17: The alleged Money Multiplier process

Week	Loans	Deposits	Cash kept by bank	Sum of Loans	Sum of Cash Reserves
0	$0	$100	$10	$0	$10
1	$90	$90	$9	$90	$19
2	$81	$81	$8	$171	$27
3	$73	$73	$7	$244	$34
4	$66	$66	$7	$310	$41
5	$59	$59	$6	$369	$47
6	$53	$53	$5	$422	$52
7	$48	$48	$5	$470	$57
8	$43	$43	$4	$513	$61
9	$39	$39	$4	$551	$65
10	$35	$35	$3	$586	$69
Total after 10 weeks	$686.19	$586.19	$68.62	$586.19	$68.62
Final totals	$1,000	$900	$100	$900	$100

In this simple illustration, all the notes remain in the banks’ vaults, while all commerce is undertaken by people electronically transferring the sums in their deposit accounts. Of course, we all keep some notes in our pockets as well for small transactions, so there’s less credit created than the example implies, but the model can be modified to take account of this.

This process is also known as “Fractional Reserve Banking”, and it’s the process that Obama, on the advice of his economists, relied upon to rapidly bring the Great Recession to an end. Its failure to work was superficially due to some issues that Bernanke was well aware of,[8] but the fundamental reason why it failed is that, as a model of how money is actually created, it is “neat, plausible, and wrong”.

The fallacies in the model were first identified by practical experience, and then empirical research.

In the late 1970s, when Friedman’s Monetarism dominated economic debate and the Federal Reserve Board under Volcker attempted to control inflation by controlling the rate of growth of the money supply, the actual rate normally exceeded the maximum target that the Board set (Lindsey, Orphanides et al. 2005, p. 213). Falling below the target range could be explained by the model, but consistently exceeding it was hard to reconcile with the model itself.

Empirical research initiated by Basil Moore (Moore 1979; Moore 1983; Moore 1988; Moore 1997; Moore 2001) and later independently corroborated by numerous researchers including Kydland and Prescott (Kydland and Prescott 1990) confirmed a simple operational observation about how banks actually operate made in the very early days of the Monetarist controversy, by the then Senior Vice-President of the New York Federal Reserve, Alan Holmes.

The “Money Multiplier” model assumes that banks need excess reserves before they can make loans. The model process is that first deposits are made, creating excess reserves, and then these excess reserves allow loans to be made, which create more deposits. Each new loan reduces the level of excess reserves, and the process stops when this excess has fallen to zero.

But in reality, Holmes pointed out, banks create loans first, which simultaneously creates deposits. If the level of loans and deposits then means that banks have insufficient reserves, then they get them afterwards—and they have a two week period in which to do so.[9] In contrast to the Money Multiplier fantasy of bank managers who are unable to lend until they receive more deposits, the real world practicality of banking was that the time delay between deposits and reserves meant that the direction of causation flows, not from reserves to loans, but from loans to reserves.

Banks, which have the reserves needed to back the loans they have previously made, extend new loans which create new deposits simultaneously. If this then generates a need for new reserves, and the Federal Reserve refuses to supply them, then it would force banks to recall old or newly issued loans, and cause a “credit crunch”.

The Federal Reserve is therefore under great pressure to provide those reserves. It has some discretion about how to provide them, but unless it was willing to cause serious financial ructions to commerce on an almost weekly basis, it has no discretion about whether those reserves should be provided.

Holmes summed up the Monetarist objective of controlling inflation by controlling the growth of Base Money— and by inference the Money Multiplier model itself—as suffering from “a naive assumption”:

that the banking system only expands loans after the [Federal Reserve] System (or market factors) have put reserves in the banking system. In the real world, banks extend credit, creating deposits in the process, and look for the reserves later. The question then becomes one of whether and how the Federal Reserve will accommodate the demand for reserves. In the very short run, the Federal Reserve has little or no choice about accommodating that demand; over time, its influence can obviously be felt. (Holmes 1969, p. 73; emphasis added)

With causation actually running from bank lending and the deposits it creates to reserve creation, the changes in credit money should therefore precede changes in fiat money. This is the opposite of what is implied by the “Money Multiplier” model (since in it government money—Base Money or M₀—has to be created before credit money—M1, M2 and M3—can be created), and it is precisely what Kydland and Prescott found in their empirical analysis of the timing of economic variables:

There is no evidence that either the monetary base or M₁ leads the cycle, although some economists still believe this monetary myth. Both the monetary base and M₁ series are generally procyclical and, if anything, the monetary base lags the cycle slightly… The difference in the behavior of M₁ and M₂ suggests that the difference of these aggregates (M₂ minus M₁) should be considered… The difference of M₂-M₁ leads the cycle by even more than M₂, with the lead being about three quarters… (Kydland and Prescott 1990, p. 4)

Well before Kydland and Prescott reached this statistical conclusion, the Post Keynesian economist Basil Moore pointed out the implication of the actual money creation process for macroeconomic theory. When macroeconomic models actually considered the role of money, they treated the money supply as an exogenous variable under the direct control of the government—this is an essential feature of Hicks’s IS-LM model, for instance. But since credit money is created before and causes changes in government money, the money supply must instead be endogenous. The “Money Multiplier” model of money creation was therefore a fallacy:

This traditional view of the bank money creation process relies on the bank reserves-multiplier relation. The Fed is posited to be able to affect the quantity of bank deposits, and thereby the money stock, by determining the nominal amount of the reserve base or by changing the reserve multiplier…

There is now mounting evidence that the traditional characterization of the money supply process, which views changes in an exogenously controlled reserve aggregate as “causing” changes in some money stock aggregate, is fundamentally mistaken. Although there is a reasonably stable relationship between the high-powered base and the money stock, and between the money stock and aggregate money income, the causal relationship implied is exactly the reverse of the traditional view. (Moore 1983, p. 538)

It is possible to interpret this reverse causation as representing “a lack of moral fiber” by Central Bankers—accommodating banks’ loan-creation rather than regulating it in the interests of the economy—but Moore pointed out that the provision of reserves by Central Banks to match loan-creation by banks merely mirrored the standard behavior of banks with respect to their business clients. Businesses need credit in order to be able to meet their costs of production prior to receiving sales receipts, and this is the fundamental beneficial role of banks in a capitalist economy:

In modern economies production costs are normally incurred and paid prior to the receipt of sales proceeds. Such costs represent a working capital investment by the firm, for which it must necessarily obtain finance. Whenever wage or raw materials price increases raise current production costs, unchanged production flows will require additional working capital finance. In the absence of instantaneous replacement cost pricing, firms must finance their increased working capital needs by increasing their borrowings from their banks or by running down their liquid assets. (Moore 1983, p. 545)

Banks therefore accommodate the need that businesses have for credit via additional lending—and if they did not, ordinary commerce would be subject to Lehman Brothers-style credit crunches on a daily basis. The Federal Reserve then accommodates the need for reserves that the additional lending implies—otherwise the Fed would cause a credit crunch:

Once deposits have been created by an act of lending, the central bank must somehow ensure that the required reserves are available at the settlement date. Otherwise the banks, no matter how hard they scramble for funds, could not in the aggregate meet their reserve requirements. (Moore 1983, p. 544)

Consequently, attempts to use the “money multiplier” as a control mechanism—either to restrict credit growth as during the Monetarist period of the late 1970s, or to cause a boom in lending during the Great Recession—are bound to fail. It is not a control mechanism at all, but a simple measure of the ratio between the private banking system’s creation of credit money and the government’s creation of fiat money. This can vary dramatically over time: growing when the private banks are expanding credit rapidly and the government tries—largely vainly—to restrain the growth in money; collapsing when private banks and borrowers retreat from debt in a financial crisis, and the government tries—again, largely vainly—to drive the rate of growth of money up.

This is something that Bernanke should have known from his own research on the Great Depression. Then, the “Money Multiplier” rose from under 6 in the early 1920s to over 9 in 1930, only to plunge to below 4.5 by 1940 (see Figure 102).

Figure 102: The empirical "Money Multiplier" 1920-1940

Perhaps he did remember this lesson of history, since his increase in Base Money was far greater than that of his predecessors. He may well have put such a massive influx of money into the system simply because he feared that little or no additional credit money would be forthcoming as a result. Better then to flood the economy with fiat money and hope that that alone would cause the desired boost to aggregate demand.

We will have to await his memoirs to know, but even if so, he (and Obama’s other neoclassical economic advisors) made the wrong choice by putting this injection of fiat money into the reserve accounts of the banks, rather than giving it to the public—as Obama considered in his “where’s our bailout?” counterpoint in his April 2009 speech.

The money drove up the unused reserves of the banking sector as never before (from $20 billion before the crisis to over $1 trillion after it) and the “Money Multipliers”—which in reality, is no more than the ratios of the three measures of the broad money supply M₃, M₂ and M₁ to Base Money—collapsed as never before. The M₃ ratio fell from over 16 to under 8, and has continued to fall to below 7 since then; the M₂ ratio—the one most comparable to the M₁ ratio back in the 1920s-1940s—fell from 9 to below 4, while most embarrassingly of all, the M₁ ratio fell below 1, hit as low as 0.78, and is still below 0.9 two years after Bernanke’s fiat money injection.

Some “multiplier effect”. Obama was sold a pup by his neoclassical advisors. The huge injection of fiat money would have been far more effective had it been given to the public, who at least would have spent it into circulation.

Don’t mention the Data

As this book details, neoclassical economics is awash with examples of its internal contradictions being ignored by its believers, so in one sense their practice of pretending that the Money Multiplier determines the amount of money in the economy is just another example of neoclassical economists believing in something that doesn’t exist. However the Money Multiplier is different in at least two ways. Firstly, many neoclassical economists know that it doesn’t exist, and secondly, its non-existence is empirically obvious. So rather than ignoring the problem because they are unaware of it, or of its ramifications—as with the Sonnenschein-Mantel-Debreu conditions—they ignore it simply because it is inconvenient to acknowledge it.

Admitting that the money multiplier doesn’t exist is inconvenient because, if so, then the supply of money is not exogenous—set by the government—but endogenous—determined by the workings of a market economy. This in turn means that this endogenous process affects real economic variables like the level of investment, the level of employment and the level of output, when it has always been a tenet of neoclassical theory that “money doesn’t matter”. So acknowledging the empirically bleedingly obvious fact that the money multiplier is a myth also means letting go of another favorite neoclassical myth, that the dynamics of money can safely be ignored in economic analysis. Consequently, clear evidence that the money multiplier is a myth has been ignored even by the neoclassical economists who know otherwise.

One of the clearest instances of this is the difference between the very emphatic conclusion that Kydland and Prescott reached about the importance of credit, and their subsequent theoretical work. In their conclusion to their empirical paper, they made a clear case for the need to develop a theory of endogenous credit:

The fact that the transaction component of real cash balances (M₁) moves contemporaneously with the cycle while the much larger nontransaction component (M₂) leads the cycle suggests that credit arrangements could play a significant role in future business cycle theory. Introducing money and credit into growth theory in a way that accounts for the cyclical behavior of monetary as well as real aggregates is an important open problem in economics. (Kydland and Prescott 1990, p. 15; emphasis added)

However they have done nothing since to develop such a theory. Instead, they have continued to champion the “Real Business Cycle Theory” that they developed prior to this empirical research, and Carpenter and Demiralp note that Kydland continues “to refer to the very narrow money multiplier and accord it a principle role in the transmission of monetary policy” (Carpenter and Demiralp 2010, p. 2, commenting on Freeman and Kydland 2000).

This charade of continuing to believe in a concept whose non-existence was an empirically fact could be maintained for as long as the money multiplier didn’t have any real world significance. Unfortunately, the “bailout the banks” strategy that Obama was advised to follow by Bernanke depended crucially on the money multiplier working to turn the huge increase in reserves into an even larger increase in private sector lending. It was an abject failure: excess reserves increased by a factor of fifty, but private sector lending fell, as did credit-money.

A recent paper by Federal Reserve Associate Director Seth Carpenter entitled “Money, Reserves, and the Transmission of Monetary Policy: Does the Money Multiplier Exist?” (Carpenter and Demiralp 2010) finally acknowledges this:

Since 2008, the Federal Reserve has supplied an enormous quantity of reserve balances relative to historical levels as a result of a set of nontraditional policy actions. These actions were taken to stabilize short-term funding markets and to provide additional monetary policy stimulus at a time when the federal funds rate was at its effective lower bound.

The question arises whether or not this unprecedented rise in reserve balances ought to lead to a sharp rise in money and lending. The results in this paper suggest that the quantity of reserve balances itself is not likely to trigger a rapid increase in lending… the narrow, textbook money multiplier does not appear to be a useful means of assessing the implications of monetary policy for future money growth or bank lending. (Carpenter and Demiralp 2010, p. 29; emphasis added)

This acknowledgement of reality is good to see, but—compared both to the data and the empirically-oriented work of the rival “Post Keynesian” school of thought—it is 30 years and one economic crisis too late. It also post-dates the effective abolition of the Reserve Requirement—an essential component of the “Money Multiplier” model—by about two decades.

Since 1991, the publicly-reported Reserve Requirement has been effectively applicable only to household bank accounts, which are a tiny fraction of the aggregate deposits of the banking system (see Table 12 in O’Brien 2007, p. 52). As Carpenter and Demiralp note, today reserve requirements “are assessed on only about one-tenth of M₂”:

Casual empirical evidence points away from a standard money multiplier and away from a story in which monetary policy has a direct effect on broader monetary aggregates. The explanation lies in the institutional structure in the United States, especially after 1990.

First, there is no direct link between reserves and money—as defined as M₂. Following a change in required reserves ratios in early 1990s, reserve requirements are assessed on only about one-tenth of M₂.

Second, there is no direct link between money—defined as M₂—and bank lending. Banks have access to non-deposit funding (and such liabilities would also not be reservable), so the narrow bank lending channel breaks down in theory. Notably, large time deposits, a liability that banks are able to manage more directly to fund loans, are not reservable and not included in M₂. Banks’ ability to issue managed liabilities increased substantially in the period after 1990, following the developments and increased liquidity in the markets for bank liabilities.

Furthermore, the removal of interest rate ceilings through Regulation Q significantly improved the ability of banks to generate non-reservable liabilities by offering competitive rates on large time deposits. Additionally, money market mutual funds account for about one-fifth of M₂, but are not on bank balance sheets, and thus they cannot be used to fund lending. These facts imply that the tight link suggested by the multiplier between reserves and money and bank lending does not exist. (Carpenter and Demiralp , pp. 4-5).

The effective freedom of banks to decide how much they money will keep in reserve—and thus not use as a source of income—versus the amount they will lend, effectively leaves the private banks free to create as much credit as they wish. This is a freedom they have exploited with gusto, as I detail in the next chapter.

After the Great Recession II: Neoclassical Responses

One would hope that the complete failure of neoclassical models to anticipate the Great Recession might lead to some soul-searching by neoclassical economists: was there not something fundamentally wrong in their modeling that they could be blindsided by such a huge event?

Unfortunately, they are so wedded to their vision of the economy that even an event like the Great Recession can’t shake them. Their near-universal reaction has been that it was simply an extreme event—like a sequence of a dozen coin-tosses that all resulted in “Heads”, which is a feasible though very rare outcome.[10] Though such a thing is possible, when it will happen can’t be predicted.

In saying this, they of course ignored the public warnings from myself and others, as documented by Bezemer (Bezemer 2009, 2010), despite the fact that those warnings were made, not merely in non-mainstream academic publications, but in the media as well. Here I can’t resist quoting the Governor of my own country’s Central Bank, Glenn Stevens:

I do not know anyone who predicted this course of events. This should give us cause to reflect on how hard a job it is to make genuinely useful forecasts. What we have seen is truly a ‘tail’ outcome – the kind of outcome that the routine forecasting process never predicts. But it has occurred, it has implications, and so we must on it. (Stevens 2008)

That speech, made in Sydney in December 2008, ignored not only the well-known warnings in the USA by Peter Schiff and Nouriel Roubini, but my own in Australia since December 2005. These had included appearances on the leading current affairs programs 60 Minutes (60 Minutes 2008) and The 7.30 Report (The 7.30 Report 2007).

Central Bankers like Stevens and Bernanke had to live in a cocoon not to know of such warnings, and neoclassical economics provides the silk of this cocoon, because they refuse to consider any analysis of economics that does not make neoclassical assumptions. Since those who predicted the crisis did so—as they had to—using non-neoclassical tools, to Bernanke and his brethren around the world, those warnings did not exist.

Unfortunately, the Great Recession does exist, and neoclassical economists have been forced to consider it. Their responses have taken two forms: tweaking the “exogenous shocks” to their models until the models generate results that look like the Great Recession; and adding additional tweaks to the core neoclassical model that at least to some degree incorporate the effects of debt. Both approaches completely miss the real causes of this crisis.

It’s Just a Jolt to the Left…

As of February 2011, there were two neoclassical papers that attempted to comprehend the Great Recession using New Keynesian models which, of course, had completely failed to anticipate it (McKibbin and Stoeckel 2009; Ireland 2011). Since the underlying theory generates tranquil equilibrium growth rather than crises, the authors instead looked for a plausible set of exogenous shocks that, if simulated in their models, generate something that resembled the Great Recession. These shocks remain unspecified however, beyond stating that they emanate from “households”, or “technology”. Neither even considered modifying their models to include the role of private debt.[11]

Ireland started promisingly, with the thought that perhaps the underlying theory itself should be challenged:

Indeed, the Great Recession’s extreme severity makes it tempting to argue that new theories are required to fully explain it. (Ireland 2011, p. 31)

However the apostate road was quickly abandoned, with the assertion that “it would be premature to abandon existing models just yet”. One ground given for persevering with neoclassical models displayed the standard neoclassical ignorance of dynamic modeling, by asserting that:

Attempts to explain movements in one set of endogenous variables, like GDP and employment, by direct appeal to movements in another, like asset market valuations or interest rates, sometimes make for decent journalism but rarely produce satisfactory economic insights. (p. 32)

Having dismissed the need for a change of approach, he went in search of “shocks” that might explain why the economy so suddenly and for so long diverged from its … equilibrium, with the objective of showing that the Great Recession was really no different to “the two previous downturns in 1990-91 and 2001”:

this paper asks whether, in terms of its macroeconomics, the Great Recession of 2007-09 really stands apart from what came before… (p. 32)

Using his small scale “New Keynesian” model, Ireland concluded that unspecified “adverse shocks” to the household’s consumption preferences and the firm’s technology caused all 3 recessions:

the Great Recession began in late 2007 and early 2008 with a series of adverse preference and technology shocks in roughly the same mix and of roughly the same magnitude as those that hit the United States at the onset of the previous two recessions…

What made this recession different however, was that the shocks went on for longer, and got bigger over time:

The string of adverse preference and technology shocks continued, however, throughout 2008 and into 2009. Moreover, these shocks grew larger in magnitude, adding substantially not just to the length but also to the severity of the great recession… (p. 48)

Ireland stated his positive conclusions for the New Keynesian approach halfway through the paper, claiming that his results:

speak to the continued relevance of the New Keynesian model, perhaps not as providing the very last word on but certainly for offering up useful insights into, both macroeconomic analysis and monetary policy evaluation. (Ireland 2011, p. 33)

This is laughable, given both the author’s methodology, and manifest ignorance of the fallacies in neoclassical thought—as evidenced by the manner in which he measured the gap between output during the recessions and the ideal level of output. He envisages a “benevolent social planner”, who can derive a “social welfare function” that reconciles all social conflict over the distribution of income, reproducing—I am sure, without knowing the source—Samuelson’s bizarre vision of capitalism as one big happy family:

it is helpful to define a welfare-theoretic measure of the output gap, based on a comparison between the level of output that prevails in equilibrium and the level of output chosen by a benevolent social planner who can overcome the frictions associated with monetary trade and sluggish nominal price adjustment. Such a planner chooses the efficient level of output and the efficient amounts of labor to allocate to … production … to maximize a social welfare function reflecting the same preference orderings over consumption and leisure embedded into the representative household’s utility function (p. 38; emphases added)

McKibbin and Stoekel use a larger scale with six household-firm agents—one for each of six economic sectors (energy, mining, agriculture, manufacturing durables, manufacturing non-durables, and services)—and 15 countries as well. As a New Keynesian model it allows for various “imperfections”, and tellingly they remark that without “short-run nominal wage rigidity” and a stylized but trivial role for money (“Money is introduced into the model through a restriction that households require money to purchase goods”), the model would simply predict that full-employment equilibrium would apply at all times:

The model also allows for short-run nominal wage rigidity (by different degrees in different countries) and therefore allows for significant periods of unemployment depending on the labor-market institutions in each country. This assumption, when taken together with the explicit role for money, is what gives the model its ‘macroeconomic’ characteristics. (Here again the model’s assumptions differ from the standard market-clearing assumption in most CGE models.)…

Although it is assumed that market forces eventually drive the world economy to neoclassical steady-state growth equilibrium, unemployment does emerge for long periods owing to wage stickiness, to an extent that differs between countries owing to differences in labor-market institutions (McKibbin and Stoeckel 2009, p. 584; emphases added)

As with Ireland, they manipulate the shocks applied to their model until its short run deviations from the steady state mimic what occurred during the Great Recession, and as with Ireland, one shock is not enough—three have to be used:

1. the bursting of the housing bubble, causing a reallocation of capital and a loss of household wealth and drop in consumption;

2. a sharp rise in the equity risk premium (the risk premium of equities over bonds), causing the cost of capital to rise, private investment to fall, and demand for durable goods to collapse;

3. a reappraisal of risk by households, causing them to discount their future labor income and increase savings and decrease consumption. (p. 587)

Not even this was enough to replicate the data: they also needed to assume that two of these “shocks”—the risk tolerances of business and households—changed their magnitudes over the course of the crisis. A previous paper had found that a “a temporary shock to risk premia, as seems to have happened in hindsight, does not generate the large observed real effects”, so they instead considered an extreme shock, followed by an attenuation of it later:

The question is then, what would happen if business and households initially assumed the worst—that is, a long lasting permanent rise in risk premia—but unexpectedly revised their views on risk to that of a temporary scenario 1 year later whereby things are expected to return to ‘normal’? (p. 582)

The procedure adopted in both these papers amplifies Solow’s acerbic observation that “New Keynesian” models fit the data better than “New Classical” ones do, simply because the modelers add “imperfections … chosen by intelligent economists to make the models work better…” (Solow 2001, p. 26). Now, to cope with the Great Recession—whose characteristics cannot be fitted even by the base New Keynesian model—the modeler also adds shocks that make the imperfections fit the data better, and even manipulates the shocks themselves until the model’s output finally appears to match reality.

This is not science, but evasion. Adding tweaks to a deficient model—now including adding variable shocks—to avoid confronting the reality that the model itself is has failed, is the behavior of a “degenerative scientific research program”, to use Lakatos’s phrase.

Krugman’s paper should have been better than these, in that at least he admits that one key component of reality that has been omitted in neoclassical economics—the role of private debt—needs to be incorporated to explain the Great Recession.

“Like a Dog Walking on its Hind Legs”: Krugman’s Minsky Model

While Krugman’s “Debt, Deleveraging, and the Liquidity Trap: A Fisher-Minsky-Koo approach” (Krugman and Eggertsson 2010) deserves some praise as the first neoclassical attempt to model Minsky after decades of ignoring him, the paper itself embodies everything that is bad in neoclassical economics.

This reflect poorly, not so much Krugman—who has done the best he can with the neoclassical toolset to model what he thinks Minsky said—but on the toolset itself, which is so inappropriate for understanding the economy in which we actually live.

Attempts to increase the realism of the neoclassical model follow a mould that is as predictable as sunrise—but nowhere near as beautiful. The author takes the core model—which cannot generate the real world phenomenon under discussion—and then adds some twist to the basic assumptions which, hey presto, generate the phenomenon in some highly stylized way. The mathematics (or geometry) of the twist is explicated, policy conclusions (if any) are then drawn, and the paper ends.

The flaw with this game is the very starting point, and since Minsky put it best, I’ll use his words to explain it:

Can “It”—a Great Depression—happen again? And if “It” can happen, why didn’t “It” occur in the years since World War II? These are questions that naturally follow from both the historical record and the comparative success of the past thirty-five years. To answer these questions it is necessary to have an economic theory which makes great depressions one of the possible states in which our type of capitalist economy can find itself. (Minsky 1982, p. xii; emphasis added)

The flaw in the neoclassical game is that it never achieves Minsky’s final objective, because the “twist” that the author adds to the basic assumptions of the neoclassical model are never incorporated into its core. The basic theory therefore remains one in which the key phenomenon under investigation—in this case, the crucial one Minsky highlights of how Depressions come about—cannot happen. With the core theory unaltered, the performance is rather like that of a dog that learns how to walk on its hind legs on command, but which will revert to four legged locomotion when the performance is over.[12]

Krugman himself is unlikely to stop walking on two legs—he enjoys standing out in the crowd of neoclassical quadrupeds—but the pack will return to form once this crisis ultimately gives way to tranquility.

However, one way in which Krugman doesn’t stand out from the pack is how he treats rival schools of thought in economics: he ignores them.

The scholarship of ignorance and the ignorance of scholarship

Krugman’s paper cites 19 works,[13] three of which are non-neoclassical—Fisher’s classic 1933 “debt deflation” paper, Minsky’s last book Stabilizing an Unstable Economy (Minsky 1986), and Richard Koo’s The Holy Grail of Macroeconomics: Lessons from Japan’s Great Recession (Koo 2009). The other 16 include one empirical study (McKinsey Global Institute 2010) and 15 neoclassical papers written between 1989 (Bernanke and Gertler 1989) and 2010 (Woodford 2010)—5 of which are papers by Krugman or his co-author.

Was this the best he could have done? Hardly! For starters, the one Minsky reference he used was, in my opinion, Minsky’s worst book—and I’m speaking as someone in a position to know. Anyone wanting to get a handle on the Financial Instability Hypothesis from Minsky himself would be far better advised to read the essays in Can “It” Happen Again? (Minsky 1982), or his original book John Maynard Keynes (Minsky 1975)—which despite its title is not a biography, but the first full statement of his hypothesis.[14]

Krugman’s ignorance of Minsky prior to the crisis was par for the course amongst neoclassical authors, since they only read papers published in what they call the leading journals—such as the American Economic Review—which routinely reject non-neoclassical papers without even refereeing them.[15] Almost all academic papers on or by Minsky have been published in non-mainstream journals—the American Economic Review (AER), for example, has published a grand total of two papers on or by Minsky, one in 1957 (Minsky 1957) and the other in 1971 (Minsky 1971). If the AER and the other so-called leading journals were all you consulted as you walked up and down the library aisles, you wouldn’t even know that Minsky existed—and most neoclassicals didn’t know of him until after 2007.

Before the “Great Recession” too, you might have been justified in ignoring the other journals—such as the Journal of Post Keynesian Economics, the Journal of Economic Issues, the Review of Political Economy (let alone the Nebraska Journal of Economics and Business, where several of Hyman’s key papers were published) because these were “obviously” inferior journals, where papers not good enough to make it into the AER, the Economic Journal, Econometrica and so on were finally published.

But after the Great Recession, when the authors who foresaw the crisis came almost exclusively from the non-neoclassical world (Bezemer 2009; Bezemer 2010), and whose papers were published almost exclusively in the non-mainstream journals, neoclassical economists like Krugman should have eaten humble pie and consulted the journals they once ignored.

That might have been difficult once: which journals would you look in, if all you knew was that the good stuff—the models that actually predicted what happened—hadn’t been published in the journals you normally consulted? But today, with the Internet, that’s not a problem. Academic economists have as their bibliographic version of Google the online service Econlit, and there it’s impossible to do even a cursory search on Minsky and not find literally hundreds of papers on or by him. For example, a search on the keywords “Minsky” and “model” turned up 106 references (including three by yours truly–Keen 1995; Keen 1996; Keen 2001).

27 of these are available in linked full text (one of which is also by yours truly; Keen 1995), so that you can download them direct to your computer from within Econlit, while others can be located by searching through other online sources, without having to trundle off to a physical library to get them. To not have any references at all from this rich literature is simply poor scholarship. Were Krugman a student of mine, he’d have failed this part of his essay.

So in attempting to model a debt crisis in a capitalist economy, Krugman has used as his guide Fisher’s pivotal paper, Minsky’s worst book, and about 10 neoclassical references written by someone other than himself and his co-author. How did he fare?

Mishandling an “omitted variable”

One thing I can compliment Krugman for is honestly about the state of neoclassical macroeconomic modeling before the Great Recession. His paper opens with the observation that “If there is a single word that appears most frequently in discussions of the economic problems now afflicting both the United States and Europe, that word is surely “debt”” (Eggertsson and Krugman 2010, p. 1), and then admits that private debt played no role in neoclassical macroeconomic models before the crisis:

Given both the prominence of debt in popular discussion of our current economic difficulties and the long tradition of invoking debt as a key factor in major economic contractions, one might have expected debt to be at the heart of most mainstream macroeconomic models—especially the analysis of monetary and fiscal policy. Perhaps somewhat surprisingly, however, it is quite common to abstract altogether from this feature of the economy. Even economists trying to analyze the problems of monetary and fiscal policy at the zero lower bound—and yes, that includes the authors—have often adopted representative-agent models in which everyone is alike, and in which the shock that pushes the economy into a situation in which even a zero interest rate isn’t low enough takes the form of a shift in everyone’s preferences. (p. 2)

This, along with the unnecessary insistence on equilibrium modeling, is the key weakness in neoclassical economics: if you omit so crucial a variable as debt from your analysis of a market economy, there is precious little else you will get right. So Krugman has taken at least one step in the right direction.

However, from this mea culpa, it’s all downhill, because he made no fundamental shift from a neoclassical approach; all he did was modify his base “New Keynesian” model to incorporate debt as he perceived it. On this front, he fell into same trap that ensnared Bernanke, of being incapable of conceiving that aggregate debt can have a macroeconomic impact:

Ignoring the foreign component, or looking at the world as a whole, the overall level of debt makes no difference to aggregate net worth — one person’s liability is another person’s asset. (p. 3)

This one sentence established that Krugman failed to comprehend Minsky, who realized—as did Schumpeter and Marx before him—that growing debt in fact boosts aggregate demand:

If income is to grow, the financial markets… must generate an aggregate demand that, aside from brief intervals, is ever rising. For real aggregate demand to be increasing… it is necessary that current spending plans, summed over all sectors, be greater than current received income … It follows that over a period during which economic growth takes place, at least some sectors finance a part of their spending by emitting debt or selling assets. (Minsky 1982, p. 6)

Krugman also has no understanding of the endogeneity of credit money—that banks create an increase in spending power by simultaneously creating money and debt. Lacking any appreciation of how money is created in a credit-based economy, Krugman instead sees lending as simply a transfer of spending power from one agent to another: neither banks nor money exist in the model he built.

Instead, rather than modeling the economy as a single representative agent, he modeled it as consisting of two agents, one of whom was impatient while the other was patient. Debt was simply a transfer of spending power from the patient agent to the impatient one, and therefore the debt itself had no macroeconomic impact—it simply transferred spending power from the patient agent to the impatient one. The only way this could have a macroeconomic impact was if the “impatient” agent was somehow constrained in ways that the patient agent was not, and that’s exactly how Krugman concocted a macroeconomic story out of this neoclassical microeconomic fantasy:

In what follows, we begin by setting out a flexible-price endowment model in which “impatient” agents borrow from “patient” agents [where what is borrowed is not money, but “”risk-free bonds denominated in the consumption good” (p. 5)], but are subject to a debt limit.

To then generate a crisis, Krugman had to introduce an ad-hoc and unexplained change to this debt limit:

If this debt limit is, for some reason, suddenly reduced, the impatient agents are forced to cut spending; if the required deleveraging is large enough, the result can easily be to push the economy up against the zero lower bound. If debt takes the form of nominal obligations, Fisherian debt deflation magnifies the effect of the initial shock. (Eggertsson and Krugman 2010, p. 3; emphasis added)

He then generalized this with “a sticky-price model in which the deleveraging shock affects output instead of, or as well as, prices” (p. 3), brought in nominal prices without money by imagining “that there is a nominal government debt traded in zero supply… We need not explicitly introduce the money supply” (p. 9), modeled production—yes, the preceding analysis was of a no-production economy in which agents simply trade existing “endowments” of goods distributed like Manna from heaven—under imperfect competition (p. 11), added a Central Bank that sets the interest rate (in an economy without money) by following a Taylor Rule, and on it went.

The mathematics was complicated, and real brain power was exerted to develop the argument—just as, obviously, it takes real brain power for a poodle to learn how to walk on its hind legs. But it was the wrong mathematics: it compared two equilibria separated by time, whereas truly dynamic analysis considers change over time regardless of whether equilibrium applies or not. And it was wasted brain power, because the initial premise—that aggregate debt has no macroeconomic effects—was false.

Krugman at least acknowledged the former problem—that the dynamics are crude:

The major limitation of this analysis, as we see it, is its reliance on strategically crude dynamics. To simplify the analysis, we think of all the action as taking place within a single, aggregated short run, with debt paid down to sustainable levels and prices returned to full ex ante flexibility by the time the next period begins. (p. 23)

But even here, I doubt that he would consider genuine dynamic modeling without the clumsy neoclassical device of assuming that all economic processes involve movements from one equilibrium to another. Certainly this paper remained true to the perspective he gave in 1996 when speaking to the European Association for Evolutionary Political Economy:

I like to think that I am more open-minded about alternative approaches to economics than most, but I am basically a maximization-and-equilibrium kind of guy. Indeed, I am quite fanatical about defending the relevance of standard economic models in many situations…

He described himself as an “evolution groupie” to this audience, but then made the telling observation that:

Most economists who try to apply evolutionary concepts start from some deep dissatisfaction with economics as it is. I won’t say that I am entirely happy with the state of economics. But let us be honest: I have done very well within the world of conventional economics. I have pushed the envelope, but not broken it, and have received very widespread acceptance for my ideas. What this means is that I may have more sympathy for standard economics than most of you. My criticisms are those of someone who loves the field and has seen that affection repaid.

Krugman’s observations on methodology in this speech also highlight why he was incapable of truly comprehending Minsky—because he starts from the premise that neoclassical economics itself has proven to be false, that macroeconomics must be based on individual behavior:

Economics is about what individuals do: not classes, not “correlations of forces”, but individual actors. This is not to deny the relevance of higher levels of analysis, but they must be grounded in individual behavior. Methodological individualism is of the essence. (Krugman 1996; emphases added)

No it’s not: methodological individualism is one of the key flaws in neoclassical macroeconomics, as the SMD conditions establish. Economic processes have to be modeled at a higher level of aggregation, as Kirman argued (Kirman 1989, p. 138) and Minsky, in practice, did.

So while Krugman reached some policy conclusions with which I concur—such as arguing against government austerity programs during a debt-deflationary crisis—his analysis is proof for the prosecution that even “cutting edge” neoclassical economics, by continuing to ignore the role of aggregate debt in macroeconomic dynamics, is part of the problem of the Great Recession, not part of its solution.

Conclusion: Neat, Plausible, and Wrong

Menchen’s aphorism suits not merely the money multiplier, but the whole of neoclassical economics: “neat, plausible, and wrong”. If we are to avoid another Great Depression—more bleakly, if we are to get out of the one we are still in—then neoclassical economics has to be consigned to the dustbin of intellectual history. But that by itself is not enough: we need a replacement theory that does not make the many methodological mistakes that have made neoclassical economics such a singularly misleading and dangerous guide to the management of a capitalist economy.

The manner in which neoclassical economists have dealt with the crisis also makes a mockery of the basis on which neoclassical macroeconomics was based: its criticism of the preceding IS-LM “Keynesian” models that they were based on many “ad hoc” parameters—as Solow observed, “the main argument for this modeling strategy has been a more aesthetic one: its virtue is said to be that it is compatible with general equilibrium theory, and thus it is superior to ad hoc descriptive models that are not related to ‘deep’ structural parameters” (Solow 2007, p. 8). However, to cope with the Great Recession, neoclassical economists are now introducing ad-hoc changes to these “’deep’ structural parameters”—in order to explain why risk is suddenly re-evaluated and so on—and even introducing “ad hoc” shocks. Neoclassical attempts to reproduce the crisis therefore fail the Lucas Critique which gave birth to this approach in the first place.

A complete, ready-made replacement does not exist. But there are alternative ways of thinking about economics that provide a good foundation on which an empirically grounded, non-ideological theory of economics can be built. I now turn to these alternatives, starting with the perspective that enabled me to be one of the very few economists who saw the Great Recession coming.

---

[1] Bernanke went on to rephrase debt-deflation using several concepts from neoclassiccal microeconomics—including information asymmetry, the impairment of banks’ role as adjudicators of the quality of debtors, and so on. He also ultimately developed a cumbersome neoclassical explanation for nominal wage rigidity which gave debt a role, arguing that “nonindexation of financial contracts, and the associated debt-deflation, might in some way have been a source of the slow adjustment of wages and other prices” (Bernanke 2000, pp. 32-33). By “nonindexation”, he meant the fact that debts are not adjusted because of inflation. This is one of many instances of Bernanke criticizing real-world practices because they don’t conform to neoclassical theory. In fact, only one country ever put neoclassical theory on debts into practice was Iceland—with disastrous consequences when its credit bubble burst.

[2] For a start, Fisher’s process began with over-indebtedness, and falling asset prices were one of the consequences of this.

[3] There are numerous measures of the money supply, with varying definitions of each in different countries. The normal definitions start with currency; then the “Monetary Base” or M₀, which is currency plus the reserve accounts of private bank at the Central Bank; next is M₁, which is currency plus check accounts but does not include reserve accounts, then M₂ which includes M₁ plus savings accounts, small (under $100,000) time deposits and individual money-market deposit accounts, and finally M₃—which the US Federal Reserve no longer measures, but which is still tracked by Shadowstats—which includes M₂ plus large time deposits and all money market funds.

[4] It then grew at up to 2.2 percent per annum until the October 1929 (the month of the Stock Market Crash) and then turned sharply negative, falling at a rate of up to 6 percent per annum by October 1930. However here it is quite likely that the Fed was being swamped by events, rather than being in control, as even Bernanke concedes was the case by 1931: “As in the case of the United States, then, the story of the world monetary contraction can be summarized as “self-inflicted wounds” for the period through early 1931, and “forces beyond our control” for the two years that followed.” (Bernanke 2000, p. 156).

[5] “When prices are stable, one component of the cost [of holding money balances] is zero—namely, the annual cost—but the other component is not—namely, the cost of abstinence. This suggests that, perhaps, just as inflation produces a welfare loss, deflation may produce a welfare gain. Suppose therefore that we substitute a furnace for the helicopter. Let us introduce a government which imposes a tax on all individuals and burns up the proceeds, engaging in no other functions. Let the tax be altered continuously to yield an amount that will produce a steady decline in the quantity of money at the rate of, say, 10 per cent a year.” (Friedman 1969 p. 16; emphases added). Friedman went on to recommend a lower rate of deflation of 5 percent for expediency reasons (“The rough estimates of the preceding section indicate that that would require for the U.S. a decline in prices at the rate of at least 5 per cent per year, and perhaps decidedly more”, p. 46), but even this implied a rate of reduction of the money supply of 2 percent per annum—the same rate that he criticized the Fed for maintaining in the late 1920s.

[6] These were June 1946 till January 1949, June 1950 till December 1951, 1957-1958, June 1974 till June 1975, 1979-1982 and December 2000 till January 2001.

[7] See http://en.wikiquote.org/wiki/H._L._Mencken.

[8] The minimum fraction that banks can hold is mandated by law, but banks can hold more than this, weakening the multiplier; and the public can decide to hang onto its cash during a financial crisis, which further weakens it. Bernanke considered both these factors in his analysis of why the Great Depression was so prolonged: “In fractional-reserve banking systems, the quantity of inside money (M₁) is a multiple of the quantity of outside money (the monetary base)… the money multiplier depends on the public’s preferred ratio of currency to deposits and the ratio of bank reserves to deposits… sharp variations in the money multiplier … were typically associated with banking panics, or at least problems in the banking system, during the Depression era. For example, the money multiplier in the United States began to decline precipitously following the “first banking crisis” identified by Friedman and Schwartz, in December 1930, and fell more or less continuously until the final banking crisis in March 1933, when it stabilized. Therefore, below we interpret changes in national money stocks arising from changes in the money multiplier as being caused primarily by problems in the domestic banking system.” (Bernanke 2000, pp. 125-126)

[9] “the reserves required to be maintained by the banking .system are predetermined by the level of deposits existing two weeks earlier.” (Holmes 1969, p. 73)

[10] Such a sequence has a one in four-thousand chance of occurring.

[11] The word “debt” doesn’t even appear in the Ireland paper, and while McKibbin and Stoeckel’s model does incorporate borrowing, it plays no role in their analysis.

[12] Samuel Johnson’s aphorism, that something is “like a dog’s walking on his hind legs. It is not done well; but you are surprised to find it done at all”, is one of those phrases that was offensive in its origins—since Johnson used it to deride the idea of women preaching—but utterly apt in its usage today.

[13] An update in February 2011 made no changes to the paper apart from adding an additional 11 works, only one of which—a 1975 paper by James Tobin—could even remotely be described as non-neoclassical.

[14] I actually posted a comment to this effect on Krugman’s blog when he announced that he had decided to read Minsky and had purchased this book.

[15] A paper based on the model that I described in this chapter (Keen 2011) was rejected unrefereed by both the AER and the specialist AER: Macroeconomics, before being accepted by the Journal of Economic Behavior and Organization.