This archived content originally appeared at, the predecessor site to, and does not carry the sponsorship of the Cato Institute.

A 1920-21 Recovery Myth

The Forgotten Depression, Jim Grant's excellent book about the 1920-21 downturn and the recovery that followed, has generated a burst of critical commentary from persons anxious to reject the principal conclusion Grant draws from that episode. That conclusion, in brief, is that the U.S. was able to recover relatively quickly from at least one deep slump (and the '21 slump was deep, to judge not only from price statistics but also from available if sketchy unemployment statistics) despite authorities' refusal to resort to either fiscal or monetary stimulus. On the contrary, Grant observes, both fiscal and monetary policy were, according to today's Keynesian-influenced understanding, more contractionary than expansionary.

I've no desire to plunge into the general controversy concerning what other lessons one might safely draw from the 1920-21 episode, except to point out (as many of Grant's critics fail to do) that Grant himself resists drawing many other conclusions. He never claims, first of all, that Harding-administration-type policies might have been a dandy solution in 2008. Nor does he insist that post-2008-style expansionary fiscal and monetary policies would have made for a less satisfactory recovery had they been employed in '21. "We can't know what might have been," Grant writes (p. 2) "if Wilson and Harding had intervened as presidents in of the late 20th and early 21st centuries are wont to do." Grant merely settles for observing that "When, as 31st president, Hoover did intervene–notably, in an attempt to prevent a drop in wages–the results were unsatisfactory" (ibid.) The results of FDR's more aggressive interference with price and wage cuts, through the NRA and AAA, were, I would add, still more so.

If there's a foolish generalization lurking about here, or anywhere else in Grant's book (say, for instance, a "citation of the 1921 economic recovery as somehow refuting everything we’ve learned about macroeconomics since then," or an assertion to the effect that "If only we had let wages and prices crash in 2009, we would be in la la land right now,") I hope someone (Paul? Barkley?) will be so kind as to point it out to me. I also hope Barkley will explain to me why, in purporting to refute Grant's thesis, he compares what happened in 1920-21, not with what transpired in 1929-33 (which is the one episode concerning which Grant himself draws comparisons) but with what happened in various post-WWII recessions to which Grant himself never even refers.

My concern here, in any event, isn't with the general lessons that either should or shouldn't be drawn from the post-21 recovery, but with a particular myth concerning that recovery, namely, the myth that, contrary to what Grant and others have suggested, the Fed did in fact help out, and help out in a big way, by loosening of monetary policy.

Barkley Rosser has been particularly anxious to make hay with this claim, especially in the post (linked above) written in response to the recent Cato Book Forum over which I presided, featuring Grant's book. (For his part Krugman settles for a mere link to Barkley's post–this in a post implicitly accusing Grant, whose book Krugman almost certainly didn't bother to read, of laziness!) "In 1921," Barkley writes, "the Fed reversed course and lowered the discount rate back down to 4%. The economy then went into its rapid rebound. I note that in his remarks at Cato, at least Larry White did note this point as a caveat on all the proceedings. Bordo et al also note that both Irving Fisher and also Friedman and Schwartz pinpointed the role of the Fed in all this and declared it to have behaved very irresponsibly in the entire episode. But for Grant and Samuelson, the Fed barely even existed then."

The claim about Fed easing having ended the 21 slump has been repeated by many others, including The Economist, which in its review of Grant's book observes that "The Fed brought on the 1920-21 depression with high interest rates. Those rates drew in gold anew, which, along with deflation and political pressure, eventually caused the Fed to relent and lower rates. The slump and recovery were thus not the spontaneous product of the free market but of deliberate policy, much as in later recessions." Another proponent of this view is Daniel Kuehn, who has written two articles and several blog posts countering Austrian claims about the implications of the 1920-21 episode. In a comment responding to a laudatory David Glasner post concerning his work on the subject, for example, Kuehn claims that Fed "loosening…definitely played a prominent role in the recovery" from the 1921 slump.

What, then, are the facts of the matter? One fact, or set of them, to which Barkley and Co. refer, is that the Fed banks did indeed lower their discount rates, from 7%, where they'd stood since June of 1920, to 6.5% in May 1921, and then all the way to 4.5% in November 1921. (The further reduction to 4% to which Barkley refers did not occur until June 1922.) But, as Scott Sumner has been tirelessly observing for some years now, even under an interest-rate targeting regime, a low policy rate doesn't necessarily mean easy money. Instead, low rates can reflect slack demand for funds, and indeed tend to do just that in any slump. A Wicksellian would say that what matters isn't where rates stand absolutely, but where they are relative to their "natural" counterparts.

But treating the discount rate as an indicator of the stance of monetary policy with reference to the 1920-21 episode is even worse than treating it so in reference to more recent experience. In recent times, you see, the relevant policy rate has been, not the Fed's discount rate–the rate at which it extends discount-window loans–but the federal funds rate, to which, in the good old day's before the recent recession, it assigned a target value, to be achieved using open-market operations, by means of which the supply of federal funds (that is, overnight loans of bank reserves) would be either increased or reduced sufficiently to bring the funds rate to its target level. A decision to "lower interest rates" by the Fed thus tended to imply a decision to expand the monetary base by adding to the Fed's security holdings. Thus, although low rates didn't necessarily mean "easy" money, a decision to target lower rates did at least tend to mean more money.

Back in the 20s, on the other hand, a lowering of the Fed's policy rate–here, not the federal funds rate but the discount rate–might not even imply an increase in Fed lending or security purchases. In reducing its discount rate from 6.5% to 4.5%, for example, the Fed merely allowed banks possessing the requisite commercial paper to discount that paper with it at the newly reduced rates. Whether they would do so, however, depended on whether the rates in question were low, not merely compared to previous rates, but relative to market rates generally or, again, to "natural" rates. If not, the volume of discounting might not budge, and the lower rates would not imply any actual monetary expansion, except perhaps relative to the contraction that might have ensued had rates remained high.

So, did the Fed, by lowering its discount rate, actually give the U.S. economy a dose of monetary stimulus? It did not, as can be readily seen by referring to the chart below, reproduced from Nathan Lewis's New World Economics blog:


As you can see from the chart, although there was some increase in "bills discounted" in response to the Fed's lowering of its discount rate, the increase was slight compared to the massive decline in total Fed non-gold assets since 1920. What's more, it was more-or-less perfectly–and by implication quite intentionally–offset or "sterilized" by means of Fed sales of government securities. The Fed's contribution to recovery, in short, consisted, not of any actual monetary stimulus, but of a mere cessation of what had been a precipitous decline in its interest-earning asset holdings.

This isn't to say that monetary expansion played no part in the post-1921 recovery. In fact, it played a significant part. But the expansion that took place was due solely to gold inflows, which were themselves encouraged by relatively high interest rates as well as by falling prices–that is, by the normal working of the price mechanism rather than by activist Fed policy. (In the 30s as well, by the way, such recovery as took place was entirely the result not of Fed easing–or of fiscal stimulus–but of the dollar's devaluation and subsequent gold inflows from Europe.) That gold flows (as opposed to Fed easing) contributed to the post-1921 recovery is itself a fact that Jim Grant readily acknowledges; his book's 17th chapter is called "Gold Pours into America."

In fine, far from having overlooked the real cause of the recovery, as his critics claim, Grant seems to have gotten it just right, whereas they all seem to have been led astray by an interest-rate red-herring.

While preparing this post I was unaware of Bob Murphy's reply to Krugman's remarks, which is very much worth reading.

  • Lord Keynes

    "If there's a foolish generalization lurking about here, or anywhere else in Grant's book (say, for instance, a "citation of the 1921 economic recovery as somehow refuting everything we’ve learned about macroeconomics since then," or an assertion to the effect that "If only we had let wages and prices crash in 2009, we would be in la la land right now,") I hope someone (Paul? Barkley?) will be so kind as to point it out to me."

    George Selgin,

    The latter is exactly what many Austrians (especially internet Rothbardians) imply all the time.

    As for monetary policy, the nominal cuts in the discount rate were sufficient to cause significant changes in business expectations and confidence.

    Also, you seemed to have missed a proto-form of quantitative easing by the Federal Reserve in which there were open market operations in late 1921–1922 to aid recovery, and in which the Federal Reserve bought government bonds from November 1921 to June 1922 and tripled its holdings from $193 million in October 1921 to $603 million by May 1922 – a fact even noted by Rothbard, America's Great Depression, 2000, p. 133, who says that, to the Federal reserve officials, "Inflation seemed justified as a means of promoting recovery from the 1920–1921 slump, to increase production and relieve unemployment" (p. 134).

    • George Selgin

      "The latter is exactly what many Austrians (especially internet Rothbardians) imply all the time."

      So what? The topic is Jim Grant's book, and it is to that book that Barkley and Krugman (by indirection) refer. If every economist sympathetic to free markets is to be held responsible for the thoughtless ejaculations of internet Rothbardians, then of course the free market case would be hopeless.

      Your claim about the rate cuts changing expectations is of course difficult to either prove or disprove. However it is, I think, irrelevant, because gold inflows would have induced such cuts in any sort of banking system. The best that can be said for the Fed is that it did not try to thwart the tendency more aggressively.

      Regarding the third point, the diagram I supplied displays the necessary facts, magnitudes included; these suggest that the operation in question was of minor imortance.

  • Paul Marks

    I notice that "Lord Keynes" does not actually deny the central point that George Selgin makes – i.e. that the Federal Reserve did NOT practice an "easy money" policy in 1921, contrary to the claims of the Economist magazine and other Keynesians.

    In my own reply to the Economist magazine review of the book at the time, I failed to make this point clearly. I know it will come as a terrible shock to people on this site but I can be mite short tempered at times – and this can get in the way of my making a point as clearly as I want to.

    As for "la la land" and so on, as a chronic gloomy guts, I am never going to assume that any policy is going avoid some horror or other. Once a credit bubble economy has been created (such as the credit bubble economy of World War One) the end of it is never going to be pretty, and a credit bubble economy must (eventually) end.

    But, yes Lord Keynes, I do believe that it is vital that wages (a form of price) need to be allowed to adjust freely to a bust – indeed the failure to allow that after the 1929 bust (Herbert "The Forgotten Progressive" Hoover's frantic efforts to PREVENT prices and wages adjusting to the bust – thus preventing markets clear) was a very bad thing indeed.

    As for fiscal policy – the record is plain enough. In reaction to the 1921 bust Warren Harding massively cut government spending, in reaction to the 1929 bust Herbert Hoover increased government spending – in real terms as a proportion of the economy. I believe the fiscal response of 1921 was more sensible than the fiscal response after 1929.

  • McKinney

    Good job! As Julien Noizet has often pointed out, mainstream economists are extremely ignorant about banking. That ignorance helps them maintain their ideology while fooling others who are equally ignorant.

  • RickHull

    > If every economist sympathetic to free markets is to be held responsible for the thoughtless ejaculations of internet Rothbardians, then of course the free market case would be hopeless.

    Ouch. This is quite the denouncement of internet Rothbardians. Did you mean to be so specifically harsh?

    • George Selgin

      Rick, I'm glad you asked me to clarify. I meant not the academic Austrians whose work appears on the internet as well as in print, but the "pop" Austrians who aren't really economists at all. I assume that Lord Keynes mainly has them in mind, but admittedly there's a thin line separating these from some of the weaker academic Austrians.

  • ___Tom___

    If I understand you correctly, the argument you’re trying to make is that the Fed’s discount rate cut in 1921 was not a good indicator of the stance of its overall monetary policy. And your evidence for this is that the Fed was steadily selling most kinds of assets other than gold during the rate-cutting period, thus reducing base money supply, Therefore, the Fed was in fact tightening during the period that it was cutting the discount rate. That seems to me a plausible thesis and worth investigating.

    You also make the point out that easy money is not the same thing as easy monetary policy. In an economic slump, it becomes easier to raise funds at least in nominal interest rate terms even if monetary policy doesn’t change. Without forgetting the importance of real rates, agreed.

    I think to complete this argument you need to look at those market interest rates, especially the market interbank rates, the deposit rates and any other data available that would show banks’ actual cost of raising funds during the period in question. If you can demonstrate that the market rates fell ahead of the Fed’s discount rate cuts, and the Fed’s cuts didn’t affect the banks’ actual cost of raising funds because the discount rate always remained well above the market cost of raising funds, then you will have definitely proved that the discount rate cuts were not real loosening. If on the other hand the 2 point discount rate cut produced an obvious reduction in banks’ cost of raising funds, then I would say that was evidently a loosening, and you will be left with at most a softer argument that the Fed’s loosening wasn’t important to the recovery.

    I would avoid bringing in Sumner’s kooky views. He’s not merely pointing out the difference between easy money and easy policy, he’s attempting to radically redefine easy vs tight money as a sort of phlogistonian residual between actual and desired NGDP.

  • George Selgin

    "You also make the point out that easy money is not the same thing as easy monetary policy."

    I'm sure I don't make this point; in fact I've always treated "easy money" as shorthand for "easy monetary policy." What I do say is that low rates aren't proof of easy money.

    Regarding market interest rates, they were in fact falling ahead of the Fed's discount rate. Compare the series in the second chart here:

    With the one about half-way down here:

    Finally, I must be a bit kookie myself, because I consider Scott Sumner's views entirely reasonable.

  • ___Tom___

    I found this data which confirms the monetary base continued to shrink throughout the rate-cutting period (change the end year to 1922 to zoom in):,SBASENS,

    So that seems to demonstrate that the discount window couldn't have been used in significant volumes. The charts you pointed me to don't show in a completely cut and dry way that the discount rate cuts were reactions not motivations but they do make clear that market rates began falling well before the Fed started cutting and that they continued to fall during the pause between the cut to 4.5 and the cut to 4.0. So I'd say overall you've got a strong case that you should work up in more detail (eg with versions of those charts focusing on the few years in question).

    I guess I misunderstood "a low policy rate doesn't always mean easy money" to mean easy policy doesn't always mean easy money, but you are using a different lingo. I think it makes more sense to use easy/tight money to refer to objective cost-of-funds market conditions and easy/tight policy as the set of policies that manipulate those conditions, but as long as you make yourself clear, whatever works.

    I would again strongly urge you not to muddle your case by bringing in Sumner's views. In his framework any significant shortfall in NGDP from trend is defined as "tight money." Using that definition, easy/tight money becomes a phlogiston, incorporating any and every reason that NGDP doesn't stay on a stable trend. That's just analytically obfuscatory. Really, if you try to bring in the 1920-1922 NGDP trend as evidence of tight money, you will be wrecking a serious argument and steering into nutterville.

    Since we're on the topic, what most annoys me about Sumner is how he presents himself as a center-right skeptic of fiscal policy. How anyone who has lived through the QE era still thinks public authority can hold NGDP to a stable trend without a very aggressive fiscal policy is really beyond me. The reality is that central bank NGDP targeting could only work if the central bank became a fiscal policy maker, spending aggressively in the real economy whenever private spending slows. As a policy it's actually well to the left of Krugman, arguably to the left even of MMT.

  • Unsal Cetin


    I think that Rothbard's "The Panic of 1819" is very much like Grant's "The Forgotten Depression". Would you aggree with that?

    • George Selgin

      Hard to say, Unsal–depends on what parallels you have in mind.

      • Unsal Cetin

        Well, as much as I can understand, the panic of 1819 ended quickly because there was no intervention. Like Forgotten Depression of 1920-21. (I did read Rothbard's book, but not Grant's yet. However, Grant's book will be in my hands next week.)

  • DMXRoid

    It's a little disingenuous for Rosser to cite Friedman & Schwartz as supporting his position. Their arg isn't that the Fed tightened money and therefore, depression, it's that it tightened it _too late_ to stave off the cyclical downturn that resulted from wartime inflation, and that the late action was worse than no action at all. They're very explicit in their conclusion that, had the Fed started its tightening policy in 1919 or early 1920, the outcome would have been much different, as the economy was slowly weaned off of the loose money of the war years. Whether or not the Fed could have actually softened the downturn is besides the point, at no point do F&S argue that tight money _caused_ the '20-21 depression.

  • noah

    GS says "the expansion that took place was due solely to gold inflows… rather than by activist Fed policy."

    Rothbard says for the period June 1921–July 1922 "a superficial glance would
    lead one to believe that the main inflationary factor was the heavy
    gold inflow (but that)if the government had remained completely passive member bank reserves would have declined by
    $303 million. Instead, the government actively pumped in $462 million of new reserves, yielding a net increase of $157 million."

    There are several times in "America's Great Depression" that Rothbard says the Fed began actively pursuing credit expansion in 1921. Was he wrong?

    True enough, in your reply to LK you say the open market "operation in question was of minor importance" in terms of relative magnitude. But that does not seem to negate the fact that the Fed was active, not passive.

    It seems akin to saying since the firemen poured so little water onto the fire, relative to the size of the inferno, that the fire was allowed to burn without interference. But that would be factually inaccurate, would it not?

    I remain unconvinced (based on my reading of Rothbard) that monetary policy as of mid-1921 was more contractionary than expansionary. That market interest rates were falling ahead of the Fed's discount rate does not in itself demonstrate that the Fed was a passive observer. The political pressure was to begin expansive policy, and Rothbard indicates that the Fed yielded to this pressure, not after the slump, but during it.

    • George Selgin

      It's evident that Rothbard's remarks refer to the fed's open-market purchases only. Those purchases were more than offset by reduced Fed discounting (see the figure above). Like I said, total Fed credit (as reflected in its interest-earning asset holdings) declined during the period in question.

      • noah

        I think Rothbard's point is that a simple view of total Fed credit does not fully tell the story of the Fed's efforts to inflate, and thus prevent further deflation and liquidation.

        He is claiming that increases in Fed credit were largely "controlled" and that offsetting decreases were "uncontrolled" — which means that a net decrease does not mean the Fed stepped back and let markets do their thing. It means Fed attempts to inflate were offset by factors beyond their control.

        He says: "The finer we break down the record, therefore, the greater the extent both of controlled increases by the government, and of uncontrolled declines prompted by the banks.
        (T)o lump all Bills Discounted as controlled
        and let it go at that… would give the Federal
        Reserve an undeserved accolade for reducing member bank debts."

        The implication is, even further and deeper liquidation was to a degree prevented by Fed policy. The inflationary debt overhang from WWI was never allowed to fully clear, continued to be partly monetized by the Fed as of 1921, and contributed to the malinvestment and credit/debt burden of the next cycle. That is David Stockman's take, too, as I read it.

        The larger point is, once the Fed becomes "activist" during WWI, at what point can we say it steps back, and makes little or no attempt to influence the economy, to the satisfaction of either certain banking/business interests or politicians ? Are interest rates, gold flows (and sterilization), money supply, etc. ever really again left to their own devices? Federal Reserve policies influenced the movement of all these, and they did so in conjunction with, and in reaction to, other central banks (especially leading up to, and after, Genoa in 1922).

        Rothbard's claim that Fed inflation of the 1920s began in 1921 runs counter to Grant's depiction of this period as America's last slump to cure itself. It seems one could also depict it as our first slump to get the benefit of central bank medication, notwithstanding the fact that the medication and its mechanism was primitive, weak, and not fully understood at the time.

  • noah

    Here's another take, from a "productivity norm" angle that you (and Grant) might appreciate.

    If we accept that great gains in productivity should mean a falling price level, we have no problem seeing that the price stabilization of the 1920s failed to account for this: monetary inflation meant prices that "should" have been falling remained fairly stable, hiding the inflation.

    Certainly the decade of the 1920s showed great gains in productivity… but can't we say the same for the several decades prior, if not even more so? Wouldn't Henry Ford's greatest efficiency and technological advances, for example, been BEFORE the 1920s?

    So when we talk about where prices "should" have been after the 1920-21 deflation, from a productivity point of view, wouldn't we expect them to be BELOW pre-war levels rather than being stabilized well above pre-war levels?

    If the natural market price mechanism was largely "un-natural" in the post-1914 years right through WWII and beyond, I don't see how we can grant (no pun intended) a two-year exception and say it worked "naturally" in that specific time period.