Today I’ll consider the first liquidity trap, which spawned that mutant monster known as Keynesian macroeconomics. But first you need to wade through more of my rantings on politics.
My last post brought up the touchy subject of racism. Big mistake. Then last night I saw a short bloggingheads discussion where Matt Yglesias and Jonathan Chait debated whether and/or when it was acceptable to accuse people you disagree with of being racist or anti-Semitic. I have mixed feelings on this, although my views are a bit closer to those of Chait. There is a tendency to assume that those you disagree with have bad motives. My problem with Yglesias is that I think he fails to see both sides of the coin. Right-wingers are going to make the same sort of judgments about the left.
Here is an example. I think it is fair to say that lots of progressives are strongly anti-racist, and lots of conservatives are mildly anti-anti-racist. What do I mean by anti-anti-racist? Conservatives and progressives disagree about the amount of racism faced by minorities, especially African-Americans. They disagree about whether certain remedies are justified. Conservatives tend to see differences in socio-economic status as reflecting cultural problems within the black community. In general, conservatives tend to minimize the problem of racism in America. These views lead progressives to believe that conservatives are somewhat tribal, and have less empathy for those outside the white “tribe.” Thus progressives interpret the anti-anti-racism of conservatives as hidden support for racism itself.
Now let’s look at anti-communism. Right-wingers like myself are extremely proud of our longstanding anti-communism (just as progressives are proud of their anti-racism.) If we are old (like me) we were anti-communist before it was cool to be anti-communist. When I was younger, anytime a progressive used the term “anti-communist” it was with sarcasm or scorn. In the middle third of the 20th century, most progressive intellectuals made at least one comment about communism that looks very embarrassing today. Often the statements were supportive of figures like Lenin, Stalin, Mao, Ho Chi Minh, etc. Or showed enthusiasm for economic policies undertaken by Mao. Of course the right had its own problems. In the middle third of the 20th century there were a lot of statements about race and civil rights made by conservative intellectuals, at one time or another, that look very embarrassing today.
[Isn’t it great to be a classical liberal! Actually, we also have some skeletons in our closets. Remember that Jefferson owned slaves.]
One thing I have noticed about progressives is that while they pay lip service to having rejected communism, their real passion is for anti-anti-communism. They seem more outraged that a few Hollywood screenwriters lost their jobs in anti-communist witch hunts decades ago, than they do that policies advocated by those screenwriters, such as “to each according to their needs,” led to the starvation of tens of millions of people. One thing I have noticed about conservatives is that while they pay lip service to opposing racism, their real passion is for anti-anti-racism. They seem more outraged about a white firefighter who was passed over for a promotion then they do for all of the tragic history of African Americans (and Native Americans.)
Just so that I won’t be misunderstood, I want to be clear about one thing. I do not believe that most modern progressives secretly favor communism, nor do I believe that most modern conservatives secretly favor Jim Crow laws. I think both groups have absorbed at least some of the lessons of the 20th century. Both have been somewhat enlightened. But as long as each side continues to talk the way they do, then progressives will continue to suspect that conservatives are secret racists, and conservatives will continue to think that progressives are secret Marxists. And speaking of Marxists, what do you make of this recent item in Yglesias’s blog, which discusses economic policies in Nigeria:
Of course to progressive blog readers “business-friendly” doesn’t always come across as a good thing. But when thinking about the developing world, it’s worth keeping in mind Karl Marx’s point that a constructing a functioning capitalist economy is a huge step forward from crushing poverty.
Let’s see if we can deconstruct this interesting quotation.
1. Start with the fact that Yglesias is much too smart to be a Marxist.
2. And yes, I realize that quoting Marx on a single point is not equivalent to endorsing Marxism as a system.
3. Nevertheless, I think it is slightly revealing that Yglesias didn’t just say “If you read publications like The Economist, which pay a lot of attention to economic development, you’ll find that most developing countries are actually better off pursuing free market reforms.” Presumably his progressive readers would not find that argument convincing. But Marx is someone worth paying attention to; after all he was the one that predicted that communist revolutions would happen first in backward countries like Russia and China. (Or did I get that backward?) In any case, Yglesias’ progressive readers know how to read this reference to Marx, as they understand that Marx’s argument is that capitalism is not the final step, but just an unfortunate stop on the path towards . . . no wait, that can’t be right, since 1989 the left has given up all their utopian dreams. Capitalism is the last stop. Nevermind. I promise to stop being suspicious of progressives.
Memo to progressives—this was a joke. Please don’t explain Marxism to me. My point was that it is easy for one side to become suspicious of the other’s motives, regardless of the facts.
What’s my point. I guess Yglesias is free to continue calling conservatives “racists.” But if he does, he shouldn’t get indignant when conservatives accuse progressives of being a bunch of closet Marxists.
During the health care debate I noticed that a lot of progressive bloggers said something to the effect of “Don’t worry about the lack of a public option. This plan will expose the weakness of the private insurance system, leading to its eventual collapse and replacement with a single-payer system.” Eventually Republicans picked up on this meme, and began accusing Obama of trying to bring socialized medicine to America. Of course the progressives were outraged—“There’s no socialism in the bill! The word isn’t even mentioned.”
Each side of the debate should try to be less suspicious of the other side, but also should try to reflect on how its own statements can fan the flames of the suspicion. Have progressives truly learned the lessons of the failure of communism, and have conservatives really tried to think of race from the perspective of minorities? I’m going to try to avoid calling people commies and racists, but I do believe that many progressives and conservatives have not yet fully absorbed the lessons of their deeply embarrassing histories.
I’d also wager that each side knows more embarrassing facts about the other side’s history, than about their own history. We tend to read what we enjoying reading.
A few other points before beginning. Yglesias argues that there is no diversity of opinion on domestic policy among conservatives:
At any rate, Julian Sanchez did a great piece on the incentives pushing toward ideological conformity on the right and the myth of the “Georgetown Cocktail Party Circuit.” I think that the same dynamic Sanchez identifies exists on the left, but it winds up having totally different results because there are actual competing sets of institutions on the left. Everywhere you go there’s pressure to follow the “line” but CAP and EPI have very different lines on teacher compensation. What’s striking about the right is that on domestic and economic policy issues it’s formed a comprehensive set of ideological points from which there’s no escape.
I’d like to make two (seemingly contradictory) points about this comment. At one level it is beyond silly, especially when you consider that progressives like Yglesias regard libertarians as being on “the right.” On issues like monetary policy there is far more diversity among right wing intellectuals than among the left (who tend to be broadly Keynesian.) There is great diversity on supply-side economics (whereas the left are all opposed.) Many on the right think our health care system is wonderful, and must be protected from the evil Democrats. Others (like Robin Hanson and I) think it is massively inefficient, wastes over a trillion dollars, and we need to blow it our and start over. The Republicans allow both pro-life and pro-choice speakers at their conventions, the Dems only tolerate one point of view. When the Massachusetts health plan was adopted the Heritage Institute said nice things about the basic idea, whereas Cato was strongly opposed.
But that last point brings up an uncomfortable truth about the modern GOP, it has become so partisan that a health plan quite similar to the Massachusetts plan is now so beyond the pale that it seems that think tank people are being fired and muzzled (at the AEI) for even considering the Obama plan as a possible starting point for further compromise. Just to be clear, I am so right wing that I strongly oppose even a slightly modified Obama plan—rather I support Brad DeLong’s HSA approach. But it is intellectually dishonest to support an idea when Republicans are in office and completely trash similar ideas when Dems are in power. This is a week where all the worst things Krugman and Yglesias said about the right came true.
Here’s a quotation from our new Massachusetts senator:
“We’re all in favor of the catastrophic care coverage and coverage for children,” Brown told “Good Morning America.” “But what about the backroom deals? What about all the bad things?”
Mitt Romney is spouting similar nonsense. I’m not naive; I understand that if Romney wants to get the 2012 Presidential nomination then he almost has to try to find meaningless distinctions between his plan and Obama’s. But why is that? I think all of this nonsense further supports my earlier point about a disturbing anti-intellectualism in the modern GOP. The message seemed to be “sure we support all of the individual big government elements of Obama’s plan, when considered one by one, but we think the overall plan is dangerous socialism. In the long run that approach will insult the intelligence of voters with even half a brain, and will turn thoughtful independents away from the GOP. David Frum is being ostracized for correctly diagnosing a disease affecting the modern GOP. And I say this as someone who, if a Senator, would not have been able to find common ground with Obama on a compromise health care plan.
America desperately needs a parliamentary system with a centrist party that is pragmatically right wing on economic issues and liberal on social and foreign policy issues. (Something like Germany’s Free Democrats.) I’d guess that at least 25% of the American electorate fit that broad description. Many are well-educated independents. Under a parliamentary system they would form a coalition government with either the Democrats or Republicans. We’d have universal health care, but it probably would have been a better plan. Right now, it’s hardly worth me even going to the polls (although I do out of a perhaps misplaced sense of loyalty to the democratic system I champion.)
6.e Was there a Liquidity Trap in 1932?
At the time the spring 1932 OMPs were widely seen as being almost completely ineffective. To see why, consider some of the macroeconomic data in Table 6.1. In many respects, the period from April to July 1932 was the worst three months of the entire Depression. Commodity prices continued to fall, and both stock prices and industrial production reached their Depression lows in July. Even the M1 money supply declined, despite the massive OMPs. It would be decades before monetary policy was again viewed as an effective stabilization tool.
Table 6.1 Economic Indicators During 1932, monthly.
Month Output WPI Dow Gold MB M1__
Jan. 12.9 67.3 85.88 11,374 7704 21,507
Feb. 12.5 66.3 82.18 11,454 7537 21,310
Mar. 12.4 66.0 81.02 11,535 7539 21,110
Apr. 11.6 65.5 64.49 11,551 7644 20,882
May 11.2 64.4 53.96 11,452 7710 20,531
June 10.8 63.9 50.62 11,384 7788 20,449
July 10.5 64.5 45.47 11,456 7858 20,152
Aug. 10.8 65.2 66.51 11,599 7850 20,189
Sept. 11.5 65.3 67.94 11,730 7897 20,211
Oct. 11.9 64.4 64.22 11,825 7896 20,256
Nov. 11.9 63.9 65.26 11,897 7978 20,555
Dec. 11.6 62.6 61.16 11,933 8028 20,341
I have not mentioned the term “liquidity trap” in this chapter, nor have I provided any evidence for the existence of such a phenomenon. Indeed under the gold market approach to aggregate demand it makes no sense to argue that “a country” is stuck in a liquidity trap. Either policy is constrained by a loss of gold reserves (which is a quite different concept) or the entire world is stuck in a liquidity trap, as would occur if large reductions in the world demand for gold had no impact on the purchasing power of gold.
Regardless of whether the U.S. was actually in a liquidity trap during the early 1930s, it appears that the Keynesian concept of monetary policy ineffectiveness was at least partly based on the perception that the Fed’s spring 1932 open market purchases had failed to revive the economy. Scholars often cite Keynes’ remark that “I know of no example of [absolute liquidity preference] hitherto.” Less well known is that just a few lines later he provides the only real world example of a liquidity trap in the entire General Theory:
“The most striking examples of a complete breakdown of stability in the rate of interest, due to the liquidity function flattening out in one direction or the other, have occurred in very abnormal circumstances. . . . in the United States at certain dates in 1932 there was a . . . financial crisis or crisis of liquidation, when scarcely anyone could be induced to part with holdings of money on any reasonable terms” (pp. 207-08.)
Keynes developed much of the General Theory in 1932 and 1933, and thus it would not be surprising if the failed 1932 OMPs were on his mind when he lost confidence in the efficacy of monetary policy during depressions. Nor was Keynes alone. Conservatives within the powerful financial community saw this as showing the folly of monetary cranks who thought the Depression could be cured by printing money. Even Irving Fisher, who was to the left of Keynes on monetary issues, temporarily lost faith in the efficacy of Fed policy. Whether or not one sees the policy as having been a failure, it is clear that this event played an important role in shaping the public’s views on monetary policy efficacy for the remainder of the 1930s.
Friedman and Schwartz (p. 324) rejected the standard Keynesian view that monetary policy was ineffective during the early 1930s. They argued that the business upswing in the late summer of 1932 was a delayed reaction to the spring OMPs, and that the Fed abandoned the OMPs prematurely. It should be clear from the analysis in this chapter that I have some problems with this view. Monetary policy may impact macroeconomic aggregates with a lag, but it is difficult to reconcile the behavior of stock and commodity prices with the Friedman-Schwartz view. Indeed, one might have expected stock and commodity prices to have risen during the spring OMPs, and then fallen back later in the year when the Fed reverted to a more contractionary policy. Instead, stocks and commodities moved in tandem with shifts in the demand for gold; plunging in the spring when public and private gold hoarding was intense, and then rising when the dollar panic ended and gold demand declined.
If the evidence is not entirely supportive of Friedman and Schwartz’s views on monetary policy effectiveness, the same could be said for the Keynesian view of this episode. The term ‘liquidity trap’ generally refers to a scenario where increases in the money supply fail to boost aggregate demand. But that is not quite what happened in the spring of 1932, as the impact of the open market purchases was mostly negated by gold outflows. During this period the monetary base rose by only about $300,000,000, despite the open market purchases totaling roughly a billion dollars, and M1 and M2 actually declined. Thus the 1932 OMPs provide little support for the view that large increases in the money supply might fail to boost a depressed economy. It is certainly ironic that the only real world example of a liquidity trap in the entire General Theory actually shows something entirely different, the constraints imposed on policymakers by the international gold standard.
It is not surprising that Keynes would have confused absolute liquidity preference with the constraints of the gold standard. Both concepts ultimately rest on policy instruments constrained by a zero lower bound (nominal interest rates in a liquidity trap and the gold reserve ratio under a gold standard regime.) In contrast, if under a fiat money regime the central bank uses either a “quantity of money” or a “price of money” instrument (i.e. exchange rate targeting), then expansionary policy faces no meaningful barriers; there is no upward limit to either the money stock or the nominal price of foreign exchange.
In chapter 4 I argued that Johnson and Mundell’s gold undervaluation hypothesis ultimately relies on a peculiar assumption; that policymakers would have been able to effectively manage gold demand in 1920, but not in 1929. Now we can see one justification for this seeming inconsistency. A coordinated set of devaluations in 1920 (aimed at price stability) would have required central banks to increase their gold demand in the short run, in order to prevent inflation. In contrast, a managed gold standard after 1929 (also aimed at price stability) would have required central banks to lower their demand for gold. Only the 1929 policy counterfactual is in danger of running up against the zero lower bound constraint on gold reserve ratios.
6.f Interpreting 1932
Even though Friedman and Schwartz viewed the 1932 OMPs as being modestly effective, they also saw the policy as being too little, too late. The more important question, and the issue that separates them from critics like Temin and Eichengreen, is whether monetary policymakers in 1932 were constrained by the international gold standard. Although I don’t believe there is any way to definitively answer this question, Friedman and Schwartz’s optimistic view is at least plausible.
The U.S. still held massive gold reserves in 1932, and there is no reason why those reserves shouldn’t have been used more aggressively in an emergency such as the Great Depression. In fact, Timberlake (1993, p. 272) pointed out that “In the true sense of Walter Bagehot’s prescriptions, all the Fed banks’ gold was excess.” Meltzer (2003, p 276) makes a similar point and also observes that the Fed was well aware of the techniques used by the Bank of England during the nineteenth century to protect its gold holdings during a panic. And it’s hard to argue with Timberlake’s maxim (p. 273) that “A proper central bank does not fail because it loses all its gold in a banking crisis. It only fails if it does not.” It is difficult to imagine a more shocking indictment of U.S. monetary policy than the fact that on the day FDR took the U.S. off the gold standard it still held over 37 percent of the world’s monetary gold stock.
But even if it can be shown that the Fed should have done more, the ultimate success or failure of an even more expansionary policy during 1932 would have hinged on the extent to which the expansionary impact of this policy would have been offset by central bank gold hoarding, private gold hoarding, and currency hoarding. If, as seems likely, the gold bloc had essentially completed its replacement of paper assets with gold reserves by mid-1932, then the ability of the Fed to have further increased the U.S monetary base would have depended upon the response of private gold hoarders. Although this response would be difficult to predict, there is no evidence that private hoarding ever reached large enough levels to exhaust U.S monetary gold stocks. But even if the Fed successfully increased the monetary base, there is also the possibility that such a policy would have driven nominal interest rates close to zero, triggering the sort of massive increase in the demand for currency and bank reserves that occurred in the late 1930s, or more recently in Japan and perhaps the US as well.
Keynes viewed a liquidity trap as a situation where further increases in the money supply would have no impact on aggregate demand, or prices. We have no real evidence that such a trap existed in 1932. Instead, the problem was that the gold standard limited the amount by which central banks could increase the base. More recently, a number of economists have argued that monetary injections that are viewed as being temporary might fail to boost aggregate demand, even under a fiat money regime. This sort of ‘expectations trap’ is even more likely to form under an international gold standard regime, where monetary injections can lead to gold outflows. The public may have understood this and thus been skeptical of any proposal to inflate within the confines of a gold standard regime.
Friedman and Schwartz’s central hypothesis is that the Federal Reserve should have, and could have, done much more to prevent the Great Contraction. Although the gold market model developed in this book is not capable of refuting this hypothesis, it does suggest that they may have placed too much emphasis on specific policy steps that might have been more effective under a fiat money regime, including the discount rate increases of October 1931, the OMPs of 1932, and the reserve requirement increases of 1936-37. And more importantly, placed too little emphasis on events that changed expectations of the future path of monetary policy, such as private and central bank gold hoarding, Glass-Steagall, and especially changes in the price of gold during 1933-34.
A recent study Hsieh and Romer (2006) supports Friedman and Schwartz’s view that the Fed policy was not constrained by gold outflows during 1932. They argue that the 1932 OMPs did not lead to fears that the dollar would be devalued, and thus did not cause a run on the dollar. Although I find this conclusion to be plausible, I don’t think the evidence they produce is quite strong enough to refute the alternative view of Temin, Eichengreen, and Bernanke (which is that the gold standard constrained policymakers during 1932.) And later I will suggest that both sides of the debate are asking the wrong question.
The most important piece of evidence cited by Hsieh and Romer is that changes in the (three-month) forward discount on the dollar did not closely track changes in the open market purchase program. The forward discount against the French franc rose sharply after Britain left the gold standard, but then quickly fell back after the Fed’s discount rate increases temporarily restored faith in the dollar. During the first half of 1932 the forward discount on the dollar zigzagged up and down, while generally remaining well above its pre-September 1931 levels, and more importantly well above its levels during the last half of 1932.
Hsieh and Romer (p. 153) focused on the fact that the forward discount fell sharply “following the passage of the Glass-Steagall Act and the first rounds of open market purchases in late February”. Another way of looking at this same data would focus on the fact that the dollar’s forward discount rose after the announcement of Glass-Steagall in early February (which was clearly when the news hit the stock market) and then fell back in late February and March when markets were reassured by the modest size of the early open market purchases. Then the forward discount began rising in April and eventually peaked in early June.
Hsieh and Romer are on firmer ground when they observe that the forward discount on the dollar fell briefly after the OMPs accelerated in mid-April, and then again in mid-June before the end of the program had been announced. Even here one must be careful, however, as it is quite plausible that the market had some ability to anticipate these developments. While Hsieh and Romer rejected the hypothesis that OMPs had led to a run on the dollar, they gave some credence the contemporaneous press reports that agitation in Congress for deficit spending and especially the Goldsborough bill had contributed to the gold outflow.
It is probably true that fluctuations in both the forward discount on the dollar and the stock market correlate more closely to Congressional turmoil than to OMPs. But recall that the Goldsborough bill would have required the Fed to reflate the economy; presumably the sort of policy that modern day critics think the Fed should have adopted. It must give some pause to those critics that even a slight possibility that such a bill might pass apparently contributed to a speculative attack on the dollar.
The biggest problem with Hsieh and Romer’s analysis is their conclusion (p. 172) that during the OMPs there was “virtually no sign of expectations of devaluation.” I would argue exactly the opposite. There were four major runs on the dollar during the Depression, the fall of 1931, the spring of 1932, the winter of 1933, and the fall of 1937. In all four cases the press was full of rumors of the dollar being under stress. In each case there was a massive increase in private gold hoarding (which did not occur prior to mid-1931.) Yes, the forward discounts on the dollar were never very large in any of these crises, but that simply reflects that fact that traders never viewed dollar devaluation as a particularly likely outcome, especially within the next three months. Forward discounts were even low during early 1933, when almost everyone agrees there was a run on the dollar (and the imminent inauguration of FDR made a near term devaluation somewhat more likely than in mid-1932.) And recall that with very low nominal interest rates even a small probability of devaluation could trigger hoarding.
The bigger question is what does this all mean for the policy counterfactuals that are of such interest to economic historians. Bordo, Choudhri, and Schwartz (B/C/S, 2002) run some policy simulations which suggest that the U.S economy was large enough that Fed had sufficient leeway to take major expansionary steps at key dates during the early 1930s. They specifically criticize Eichengreen for linking gold reserves with U.S. monetary aggregates such as M1, i.e. implicitly ignoring the money multiplier. While this is a valid point, taken at face value it also seems to suggest that the monetary base is the relevant monetary policy tool. Of course the base rose substantially in 1932, with little discernable effect on the economy, and fell after March 1933, even as production soared.
B/C/S might reasonably argue that the broader monetary aggregates, not the base, are the appropriate policy indicators, and that these aggregates fell in 1932. Of course this reflected the extreme instability of the money multiplier during the early 1930s. If one is to have any confidence in the B/C/S simulations, however, one has to assume that more expansionary policies would not have led to substantially more private hoarding of either cash or gold. But like Hsieh and Romer, they overlook evidence of massive private gold hoarding during 1932 (and 1931) and thus end up doubting whether there really was a run on the dollar.
B/C/S also implicitly assumed that their policy counterfactuals would not have affected the money multiplier (i.e. would not have led to more cash hoarding.) Maybe so, but one doesn’t have to believe in absolute liquidity preference to note that during the early 1930s runs on the dollar were often associated with banking panics and currency hoarding. I don’t think we can simply assume that base money injections would have led to proportionate increases in the monetary aggregates. And finally, in their conclusion B/C/S argue that adherence to the gold standard was only a bar to recovery in smaller nations. Although I think this is a defensible argument, in the next chapter we will see that an explosive increase in output occurred right after the U.S. abandoned the gold standard, even without much change in the money supply.
As with the Hsieh and Romer study, I find the B/C/S policy conclusions to be plausible. Well-timed monetary injections by the Fed might have been helpful in 1931 and 1932, given their massive gold holdings it was certainly worth a try. But all of these policy counterfactuals may be asking the wrong question. If we assume interwar policymakers to have been well-intentioned proponents of fiscal and monetary stabilization, then surely they would have found a way to prevent the Depression. They probably could have done so without America being forced to devalue, but if not, they would have moved toward a fiat money regime. But the more interesting questions are; were the actual OMPs of 1932 helpful? Did they account for the delayed economic bounce in the late summer? And what explains the severe economic downturn during the spring of 1932? Here I think the supporters of Friedman and Schwartz are on much shakier ground.
It is difficult to know what to make of the market reaction to the spring 1932 OMPs. There were some significant increases in stock prices in response to news of the OMPs, especially the announcement of Glass-Steagall. And yet the stock and commodity markets performance was abysmal during the period of rapid purchases, and then prices soared right after the policy was abandoned. Admittedly a strict application of the efficient market view would favor the former evidence (that the OMPs helped.) But it would be difficult to claim that the markets thought the OMPs would have a significant impact on the course of the Depression. And even if they had less direct impact on the gold outflows than did Congressional agitation for expansionary policies, they might well have been a contributing factor in European fears that America would adopt inflationary policies. Brown (1940, p. 1233) saw these events first hand and argued that:
“The uninterrupted large scale purchases of government securities by the Federal Reserves banks seemed in the eyes of foreigners to be evidence of approaching inflation in the United States. The gold outflow to the continental creditor countries was consequently sharply accelerated as this policy was vigorously pressed forward.”
When the economic recovery began in late summer, it was accompanied by a sharp slowdown in OMPs. The recovery might conceivably have been a delayed response to the spring OMPs (although in 1933 the economy responded immediately to an expansionary monetary policy) but policy lags can hardly account for the sharp fall in commodity and equity prices during the spring, and the equally powerful rally that began in late July. More importantly, the financial markets rallied at almost the exact moment when the private gold hoarding subsided and the dollar crisis ended. With regard to the issue of causality, appendix 6a provides strong evidence that private gold hoarding was far less correlated with the stock market during periods where there were no fears of dollar devaluation, and thus causality almost certainly ran from gold to stocks. Private gold hoarding led to bearish policy expectations, which led to falling stock prices. When dishoarding began, stocks soared. It was exactly the same pattern that occurred in the fall of 1931, and it would happen again in 1937 and 1938.
6.g What About Real Interest Rates?
Thus far I have described the entire Great Contraction without any reference to interest rates playing a causal role. Temin (1976) suggested that because interest rates were declining during 1930, it is unlikely that the concurrent decline in output was caused by a tight monetary policy. He also argued that the deflation of 1930 was largely unanticipated, and hence even (ex ante) real interest rates were probably declining during this period. I find Temin’s assumption about low real interest rates to be plausible, but not his conclusions about the important of monetary policy.
In August and September 1929 the U.S. and Britain raised interest rates to their cyclical peak just as the economy was about to contract, and then lowered rates throughout late 1929 and 1930. This pattern occurs frequently during U.S. business cycles. If it is evidence against money playing a causal role in the 1930 downturn, it is equally so for monetary policy playing any important cyclical role in the U.S. economy. The problem with Temin’s argument is that interest rates (real or nominal) are simply not a very good indicator of the stance of monetary policy. During the Great Contraction it was central bank gold hoarding that best illustrated the contractionary stance of monetary policy.
In contrast to Temin, Cecchetti (1992) argued that investors were able to partially forecast deflation over 3 to 6 month horizons, and in a similar vein Nelson (1991, p. 2) suggested that “while commodity prices fell very rapidly during 1929-1930, cost-of-living indices fell much more slowly. Commentators expected the cost of living to follow commodity prices downward.” Once again, I accept their evidence but question the implication that real interest rates were high.
Barsky (1987) found little or no evidence that price level changes were forecastable under the classical gold standard, and in an earlier paper (1999) I argued that most economists only began paying attention to the Fisher effect after the world moved to the post-WWII fiat money regimes, when the trend rate of inflation moved from the near zero levels of the classical gold standard to the positive and significant levels of the 1960s and 1970s. And we need to be especially careful when looking at real interest rates movements over business cycle frequencies, where many prices are in disequilibrium. For instance, entrepreneurs considering an investment project are presumably interested in some sort of price level estimate showing expected market conditions for products that they plan to sell when their project is completed. But when there is a sudden decrease in aggregate demand, the measured price level may not provide an accurate reading of those market conditions. As an example, measured rents are quite sticky, even during periods when housing prices are falling fast and new apartment buildings stand empty.
Consider the following example: aggregate demand falls unexpectedly by 10 percent. In the short run output falls 5 percent and the (somewhat sticky) price level also falls by 5 percent. Market sensitive commodity and asset prices immediately fall by roughly 10 percent. Investors (correctly) forecast an additional 5 percent fall in the general price level, as sticky prices gradually fall to their long run equilibrium. Assuming no further decrease in aggregate demand, output returns to its natural rate once these prices fully adjust. I would suggest is that the additional 5 percent by which the measured price level is expected to decline (as the economy moves from disequilibrium to equilibrium) is not a meaningful indicator of the expected change in market conditions facing sellers when the quantity of sales is restricted by a sticky price level.
In my scenario the (unexpected) fall in flexible asset prices may provide a better indicator of movements in the price level. Since commodity and asset price movements are generally regarded as being unanticipated, this might at first glance seem to deny any possibility of a Fisher effect under any scenario. But this is not the case, in the U.S. during the 1960s and 1970s persistent inflation was so built into expectations that, even in auction-style markets, the trend rate at which prices changed was partly forecastable.
The preceding thought experiment finds some support in the research of Hamilton (1992), who found evidence from the commodity futures markets that that the deflation of 1930 was unanticipated, and hence that real interest rates were falling. In addition to finding that commodity prices were essentially unforecastable during the Great Depression, Hamilton (1987, p. 167) also estimated that:
“knowledge of the actual course of commodity prices would have reduced the forecast variance for the CPI by 70% . . . and we would seem to have a solid basis for inferring that much of the overall deflation during 1929-1933 was unanticipated.”
While I find the wording to be a bit confusing, Hamilton’s conclusion makes sense if he is (at least implicitly) assuming that predictable changes in sticky prices (the CPI) are not economically meaningful for investors, and should not be incorporated into estimates of real interest rates. Rather, what matters is the expected change in the equilibrium price level (as proxied by commodity prices.)
Temin’s (1976) view of real interest rates was consistent with the earlier Keynesian tradition, which viewed monetary policy as being less important than expenditure instability. The more recent (new Keynesian) view is that monetary policy is generally quite potent, and that real interest rates are the appropriate policy indicator. Papers in the new Keynesian tradition by Romer (1992) and Eggertsson (2005) have argued that ex ante real interest rates were quite high during the early 1930s, in the 10% to 20% range. I have already raised one objection to this view, and Bernanke and James (1991, p. 49) provide an even more persuasive reason to doubt that real interest rates were high during the early 1930s. They pointed out that those countries that left the gold standard early (such as Britain) were able to arrest the decline in prices, but continued to offer the same sort of low nominal yields on safe assets as did countries remaining on the gold standard, such as the U.S. and France.
Given that the gold market approach to the Great Contraction is essentially monetary (broadly defined) the reader may wonder why I have not embraced the high real interest rate hypothesis, which provides a perfectly suitable transmission mechanism between gold hoarding and declining expenditure. In fact, I don’t have any problem with either view of real interest rates during the Depression, as nothing in my gold market analysis hinges on this issue. This is one of those rare cases in economics where either of the opposing views can provide quite plausible monetary transmission mechanisms. If the deflation was anticipated, then high real interest rates depressed investment. If the deflation was unanticipated, then the sticky wage transmission channel provides a plausible link between gold, deflation, and falling output. Rather, I would emphasize that gold hoarding will depress nominal spending, and prices, regardless of what happens to real interest rates. There is no gain from adding interest rates to the story.
 Output is industrial production, the Dow represents mid-month closing prices, and gold is the world monetary gold stock in millions of U.S. dollars. The U.S. monetary base and M1 are taken from Friedman and Schwartz (1963) and are also measured in millions of U.S. dollars.
 The almost universal perception that the spring 1932 OMPs had been a failure did not imply that all forms of monetary policy were viewed as being ineffective in a depression. Few doubted that unorthodox monetary solutions involving fiat money were capable of generating high rates of inflation, but there was little support for this type of experimentation. Eichengreen (p. 315) suggests that the 1932 OMPs led conservative American policymakers to view expansionary monetary policy as “conducive not to economic recovery but to inflation.” But this his based on his (mistaken) view that prices rose during the spring of 1932. In fact, prices fell, and conservatives looked at the liquidity trap issue pretty much the same way as Keynes did—i.e. as a problem of “pushing on a string.”
 Fisher favored switching to a pure fiat money regime. Keynes was unwilling to support such a radical step.
 Laidler (1999, p. 259n) calls the perceived policy failure “one of the key ‘stylized facts’ underlying the evolution of monetary economics in the 1930s and 1940s.”
 Currie (1934 ) was one of the few interwar commentators who understood this.
 Technically this assertion only holds for the price of money. But it is exceedingly unlikely that a central bank could buy up all of the world’s stock of eligible financial assets, without first triggering inflation, if not hyperinflation.
 Meltzer (p. 276) points out that the Bank of England “had suspended the gold reserve requirement and relaxed restrictions on eligible paper for discount” during panics.
 The US situation after October 2008 was complicated by the Fed’s decision to pay interest on reserves at a rate above T-bill yields, which may have inflated commercial bank demand for reserves.
 Hanes (2006) showed that even when U.S. short-term interest rates had fallen close to zero during the mid-1930s, changes in reserve supply continued to impact longer-term interest rates.
 See Sumner (1993b), Krugman (1998) and Eggertsson and Woodford (2003).
 As noted in section 6.d, one type of emergency currency injection during 1932 was explicitly ruled by the Attorney General to have a three year limit.
 Hsieh and Romer also fail to find much evidence for devaluation fears in long term bond yield differentials. But given the uncertain political situation in Europe, one must be careful in making any assumptions about the relative risk of these assets. The U.S. did devalue before France, but the French monetary situation eventually became even more unstable.
 Gold reserves were used to back the monetary base. Because M1 is several times larger than the base, policy counterfactuals that assume M1 is backed by gold would imply an implausibly large need for gold reserves. Bordo (1994) first raised this issue.
 Meltzer (2003) noted that gold standard rules allowed gold reserve regulations to be adjusted in an emergency. And 1929-32 would certainly seem to qualify as an emergency.
 See appendix 6a.