What caused the slump of 1930? (Pt. 4 of 4)

There was no single cause of the Great Depression, rather there was a cause (or causes) of the initial slump, and then another series of events that caused the Depression to deepen in mid-1931.  Below I summarize my views on what caused the initial slump.  But first I’d like to link to three good posts written by liberals (or perhaps I should say 2 liberals and an alleged liberal, as Robert Wright has doubts about Mickey Kaus.)

1.  After my recent comments on the Fed’s decision to raise the discount rate after 2 years of jobs losses, some people argued that the markets misunderstood the Fed’s decision, and that they weren’t “actually tightening.”  When are you people going to stop believing in reality!  Read some post-modernism; reality is passe, it’s all about what people believe, not what they “know.”  And if you don’t believe the post-modernists (and quite frankly I couldn’t blame you if you didn’t) then read Woodford and Eggertsson.  What matters for AD is not the current stance of monetary policy, but rather the expected future path over time.  If a policy action makes people believe policy will be tighter in the future, the effect is contractionary, regardless of what the Fed “intended.”  I read somewhere that the fed funds futures showed a jump in the probability of Fed tightening in late 2010 right after the announcement.  So the effect was contractionary, whatever the Fed’s intentions.

And just so you don’t think my “sterilization” argument is merely the fevered imagination of a right-winger who doesn’t want fiscal stimulus, check out this post from a liberal blogger with impeccable pro-stimulus credentials.  Perhaps Yglesias isn’t quite as dogmatic as I am, but he certainly shares my concern that there is a risk the Fed will simply offset the impact of any fiscal stimulus.

2.  I have always believed that the impact of monetary policy on the price level is best illuminated by looking at changes in the median goods price (defined as the rate where half of all goods prices rise faster, and half rise slower.)  In theory, you’d expect the relative price of roughly half of all goods to rise, and half to fall, in response to real shocks.  If the relative price of a major good (such as energy) rises sharply, then the overall CPI may rise faster than the price of the median good.  Even in this case, however, the median good should still be picking up the effect of monetary policy, which impacts the nominal price of all goods.  Check out the median good price index in this Krugman post.  I’m not quite as pessimistic as Krugman (I still think we’ll avoid the Japanese scenario) but the fall in the median good inflation rate since mid-2008 is worrisome.

3.  Mickey Kaus is the first blogger I ever read, and I still like his style.  He has the right attitude toward all the political follies–bemusement.  I’m so dumb it took me months to figure out that his “editor” was actually himself.  I’d love to use that technique, but feel it would be plagiarism.  (Ed. — but blogging techniques are not patented.)  Here he shows why liberals should be concerned about Davis-Bacon.

Here’s the end of chapter 4, next week I’ll start the chapter on 1931:

4.i   What Caused the Depression of 1930?

In this chapter we have seen that the first year of the Depression was associated with a sharp rise in the world gold reserve ratio.  As theory would predict, this increase depressed aggregate demand and produced deflation throughout much of the world.  Surprisingly, there were no significant problems with the banking system despite the fact that the slump was already considerably worse than the 2008-09 recession.  Relative to the banking system of the early 21st century, banks were conservatively managed in the 1920s,  at least relative to the banking system of the early 21st century, due to the lack of government deposit insurance and “too big to fail” policies.

Of course correlation doesn’t prove causation, so we also need to consider other evidence that a gold reserve ratio triggered the contraction.  If we were to look at the entire 20th century, the contraction that most resembled 1929-30 was the depression of 1920-21.  In both cases the monthly indices for both wholesale prices and industrial production showed sharp declines.  Does our explanation of the onset of the Great Contraction shed any light on that earlier depression?

The Fed began raising interest rates in early 1920 in order to stop inflation, and between October 1920 and October 1921 the U.S. gold-reserve ratio soared by more than 53 percent.  With the U.S. then holding more than one third of the world monetary gold stock, this would have been sufficient, ceteris paribus, to reduce the world price level by at least 18 percent.  The actual U.S. GNP deflator declined by 16 percent between late-1920 and late-1921, and wholesale prices and industrial production fell even more sharply.  The contractionary Fed policy also induced a large inflow of gold to the U.S.  But because the increase in the U.S. gold stock was roughly equal to the world increase, foreign gold stocks remained relatively stable.  Even so, the sharp deflation (which was even greater in many European countries) can be plausibly attributed to U.S. policy.  Although most European powers were not on the gold standard at that time, they intended to eventually return to the standard and were reluctant to allow their currencies to depreciate against the dollar (i.e. gold.)  They were also reluctant to see a reduction in their gold stocks, which had already been depleted during the war.  To prevent gold outflows they were forced to follow the deflationary policies of the Fed.

If the preceding explanation can explain the similarities between 1929-30 and 1920-21, can it also explain the differences?  In particular, why did 1920-21 not see anything comparable to the 1929 stock market crash, and why wasn’t the 1920-21 downturn followed by a ‘Great Depression’?  The answer to the first question is fairly clear.  The immediate post-WWI situation was highly unstable, both domestically and internationally.  In 1920 there was as yet no reason to anticipate a ‘New Era’ where central banks would be able to deliver macroeconomic stability, and thus stock prices never approached the lofty levels of 1929.  Also recall that in 1930 stock prices had only fallen to the levels of the relatively prosperous mid-1920s, it wasn’t until late 1931 that the Great Depression was fully priced into the U.S. equity markets.  It would have been reasonable for investors to view 1921 as a painful but brief transition toward “normalcy.”  In contrast, by 1929 investors (incorrectly) believed that after seven years of relative stability, all of the post-war adjustments had already occurred.

The instability of the early postwar period may also explain why the 1920-21 depression was relatively brief.  There were no exchange rate crises comparable to those of 1931, because in 1921 most of the major currencies were still floating.  And because the U.S. did not experience a major banking panic during 1921, none of the major factors that deepened the Great Depression after late 1930 were at work during the 1920-21 contraction.  Conservative often point to Hoover’s high wage policy, which contrasts with Harding’s more laissez-faire approach.  This difference cannot account for the sharp fall in nominal GDP after 1929, but may explain why the 1921 depression saw a bit more price deflation, whereas in 1930 a larger share of the decline in nominal GDP was comprised of falling output.

Another difference between 1920-21 and 1929-30 is that in the latter period the causes of the monetary policy shift were much more international in scope.  Previous narratives of the Depression often focus on problems in one country being transmitted to the rest of the world through gold flows.  In contrast, the gold market approach used here is not reliant on Hume’s price-specie-flow transmission mechanism.  Simultaneous increases in central bank demand for monetary gold in many different countries caused the real value of gold to rise sharply during 1930.  This was equivalent to a change in the terms of trade between gold and other commodities traded in international markets, i.e. deflation of tradable goods prices in all countries adhering to the gold standard.  And this deflation occurred immediately in the commodity markets, even before exchange rate pressures had time to impact domestic money supplies.

Of course the Great Depression did not begin at precisely the same time in every country.  In France the undervaluation of the franc led to rising prices for non-tradable goods, and for a time this offset the effect of deflation in the international commodity markets.  But in a more fundamental sense both the causes and the consequences of the tight money policy during 1929-30 were international in scope from the very beginning.  Even in France, the wholesale price index was already falling by late 1929.

Eichengreen placed great emphasis on the fact that the Depression began earlier in some countries on the periphery of the world economy.  He argued (p. 392) that the tight U.S. monetary policy of 1928 forced those countries experiencing balance of payment deficits to also tighten their monetary policies.  But I don’t think it makes sense to see this as the beginning of the Depression.  Rather, the strong supply-side of the U.S. economy raised the real exchange rate for the dollar, which forced mild deflation on those less robust economies that also had tied their currencies to gold.  Wall Street had no particular reason to be concerned about this pattern, which was quite similar to the problems faced by countries with currencies tied to the dollar in the late 1990s, when there was another supply-side boom in the U.S.  Some countries slumped before the U.S., and some after, but Table 4.1 shows that the world price level broke sharply in late 1929.

The use of gold flows to interpret monetary policy has two major drawbacks.  While the inflow of gold into a country suggests that that country is following a contractionary policy, (relative to the rest of the world), it does not provide a way of quantifying that policy’s impact.  Furthermore, if all major countries simultaneously conduct a contractionary policy (as during 1929-30) then gold flows may be unable to detect the policy shift.  The deflationary nature of this policy shift would, however, be evident from the sharp increase in the world’s gold-reserve ratio.

Friedman and Schwartz cited the large gold flows into the U.S. during 1929-1931 as evidence that contractionary monetary policies in the U.S. helped transmit the Great Depression to the rest of the world.  Fremling (1985) disputed this contention, noting that the U.S. gold inflows, while substantial, were smaller than the net increase in the world’s gold stocks and did not lead to a decrease in foreign gold stocks.  Therefore U.S. policy did not force foreign central banks to contract their money supplies.  Both of these views oversimplify a complex phenomenon.  U.S. policy had the largest (deflationary) impact on the world price level during 1929-30, French policy had the greatest impact over a longer time horizon, and British policy was most responsible for the sudden tightening of world monetary policy in late-1929 (although they had little choice because of their low gold reserves.)

How one partitions “blame,” depends on how seriously one takes the concept of the “rules of the game,” i.e. a stable gold reserve ratio.  In my view the system was fundamentally flawed, which makes it difficult assign blame.  Nevertheless, during the 1920s there was an understanding that a shortage of gold posed a deflationary threat to the world economy.  Given the risk of deflation, I think it is fair to say that the large increase in U.S and French gold stocks was “unhelpful.”

Although I have argued that monetary tightening was the proximate cause of the onset of the Great Contraction, and may also have contributed to the 1929 stock market crash, I do not believe that the Depression had any single ‘root’ cause; in the sense of a cause deeply embedded in preceding events.  Of course there are many difficult philosophical problems associated with establishing ‘causation,’ some of which will be discussed in chapter 14.  For the moment, consider the following five policy counterfactuals:

1. Through deflation or currency devaluation, policymakers avoided undervaluing gold in the immediate aftermath of WWI. (The Johnson/Mundell solution)

2. Policymakers allowed the purchasing power of gold to rise gradually during the 1920s, through a policy of mild deflation.  (The Austrian solution)

3.  Beginning in 1929, the major central banks cooperated to economize on the holding of gold reserves.  (The Bretton Woods solution.)

4.  Beginning in 1929, the Fed adopted a 3 percent M2 growth target.  (The monetarist solution.)

5.   Beginning in 1929, the U.S switched to a fiat money regime and began targeting the inflation rate at 2 percent.  (The new Keynesian solution.)

It is quite plausible that any of these policy counterfactuals would have prevented the sharp decline in aggregate demand after 1929.  In that sense, the failure to adopt any one of these policies could be viewed as a “root cause” of the Depression.  The ‘root causes’ approach, however, is hard to reconcile with modern finance theory.  By 1929, investors presumably knew that these policy counterfactuals had not occurred, or were not likely to occur.  If the root causes of the Depression were already in place by September 1929, then stock prices should never have risen to such lofty levels.  Either the Great Depression was not forecastable in September 1929, or financial markets are not efficient.

Viewed from an “efficient markets” perspective, a true causal factor would be an event that would generate unanticipated movements in stock and commodity prices.  One might view the 1929 macroeconomy as a highly complex and potentially unstable system that was impacted by a “perfect storm” of unforecastable events.  Even worse, those political and economic shocks may have been at least partially interrelated””raising the prospect of small shocks having enormous repercussions””the famous “butterfly effect.”[1]  Events such as the Great Contraction don’t happen very often””indeed only once in modern world history””so it’s not hard to see why investors in prosperous 1929 would have had difficulty foreseeing this complex oncoming disaster.

[1] The Depression might never have been “Great” without the ripple effects flowing from the banking panics which began in late 1930.  And these panics might have been avoided had the economy been a bit strong in the summer and fall of 1930.  Thus events such as the death of Governor Strong, or Hoover’s decision to sign Smoot-Hawley, could well have been pivotal turning points, even if their direct effects were quite modest.

If we look at Mundell and Johnson’s gold-undervaluation hypothesis we can see some of the conceptual difficulties associated with determining causality.  They argued that World War I (which roughly doubled the U.S. price level) had left gold severely undervalued.  Even after the deflation of 1921, wholesale prices in the U.S. were about 50 percent above their pre-war level.  And because the interwar gold exchange standard economized on the monetary use of gold, prices remained far above the pre-war level for the remainder of the 1920s.  According to Mundell and Johnson, the Great Deflation of 1929-32 was caused by the preceding (unsustainable) undervaluation of gold.  It was almost inevitable that the value of gold would eventually return to its long run equilibrium value, and when that happened price levels would fall in terms of any currency tied to gold.  This hypothesis (which does utilize the gold market approach to monetary economics) has the virtue of being correct about the weaknesses in the interwar gold standard.  After 1929 the gold exchange standard did gradually evolve into something closer to the pre-war system, as central banks replaced foreign exchange with gold reserves.  And prices did fall to roughly their pre-war levels.

If we take a closer look at the Mundell-Johnson hypothesis, however, a key issue is whether or not the gold undervaluation was sustainable.  Prior to the crash, investors might have believed that central banks would be able to continue the policies that had successfully stabilized the U.S. price level during 1922-29, i.e. would continue to economize on gold reserves with a “gold exchange standard”.  In retrospect they would have been wrong, but their expectations were not obviously unreasonable””the U.S. economy had continued to boom during 1927-29 even as the French were hoarding large quantities of gold.  It wasn’t obvious that this process could not continue.  Indeed Bordo and Eichengreen (1998) showed that with more enlightened monetary policies the world monetary gold stocks were sufficient to underpin the international gold standard for several more decades. From this perspective, one could just as easily argue that the root cause of the Great Depression was not the undervaluation of gold, but rather the failure of central banks to maintain the policies they had so effectively pursued during the late 1920s.

Nor is it obvious that a series of well-timed currency devaluations in the early 1920s could have solved the gold undervaluation problem.  It wasn’t so much that gold was undervalued in 1920 (the market was in short run equilibrium) but rather that one might reasonably have expected the value of gold to have gradually risen during the 1920s and 1930s, as the gold standard was restored and monetary gold demand increased.  Thus a policy of coordinated devaluations in 1920 would have had to have been accompanied by massive gold sterilization, in order to prevent a burst of inflation from again undervaluing gold in the short run.  If this was done, then gold would have been gradually desterilized as countries like France built up their gold stocks.  But if policymakers were capable of that sort of sophisticated inflation targeting in 1920, why couldn’t they have done the same in 1929?  We will consider one answer to this question when we examine liquidity traps in chapter 6, which imply an asymmetry in the gold standard that supports the Johnson/Mundell view.

Similar objections could be raised to any of the other four policy counterfactuals.  Nevertheless, we can combine all five policy counterfactuals into a plausible composite hypothesis; where the Depression was caused by the failure to set post-war exchange rates at appropriate levels, and given that decision, the failure of the Fed to allow a gradual deflation after 1921, and given that they had already rejected these two policy options, the failure of the Fed (and other central banks) to maintain high prices after 1929.  Then the policy switch which led to high gold reserve ratios after late 1929 (and which also surprised investors), becomes the proximate cause of the Depression, and the deeper “root causes” depend on one’s views of which policy alternatives would have been both effective and politically feasible.

4.j  Summary     

Many economists have argued that monetary policy was too tight during 1928-29, and that the slump that began in late 1929 was a lagged response to the tightening.  Others argued that policy was too loose; feeding a stock market bubble that would inevitably burst.  As far as I know this is the only study to argue that there were no significant policy failures in the period preceding the Wall Street crash of 1929.[1]  Rather, policy only began to depress aggregate demand in late 1929.[2]  A switch toward tighter monetary policy in Britain and the U.S. led to a sudden increase in the world gold ratio, which was the proximate cause of the subsequent (worldwide) deflation.  The key U.S. policy error was the failure to accommodate Britain’s need to rebuild gold reserves in 1930, as they had in 1927.      French policy also continued to exert deflationary pressure on the world economy.  There are at least five pieces of evidence for this interpretation.

[1] Schuker (2003) is a partial exception to this generalization.  He argues that the gold exchange standard of the 1920s was not fundamentally flawed, and that the subsequent problems reflected “contingent political outcomes.”

[2] This is not to say that given what happened later, a different policy in 1928-29 would not have been preferable.  We will consider some other counterfactuals in chapter 14.

1.  Only the gold reserve ratio changed sharply after October 1929, there were no major changes in the other two factors influencing the international gold market; growth in real currency demand increased by less than 3 percent and growth in the supply of monetary gold actually accelerated slightly (which is inflationary.)  Indeed these other two factors essentially neutralized each other.

2.  Economic theory predicts that a sharp increase in the real demand for gold will increase the value of gold, which is deflationary for any country using gold as a medium of account.

3.  Between 1920 and 1921 there was an (even sharper) increase in the world gold reserve ratio, and there was even greater world-wide deflation.

4.  Stocks markets in many countries fell sharply about the time when this policy switch occurred.

5.  We found some evidence that, throughout 1929 and 1930, the U.S. stock market reacted adversely to signs that Hoover would not be an effective leader and also responded negatively to signs that international cooperation would not be forthcoming.

The last point is obviously very tentative, and we need much more evidence to have any confidence in the hypothesis that a lack of international policy coordination affected monetary policy expectations in the U.S.  Fortunately, the year 1931 will provide a mountain of such evidence.




4 Responses to “What caused the slump of 1930? (Pt. 4 of 4)”

  1. Gravatar of Steve Steve
    22. February 2010 at 09:52

    Dear Scott,

    I am very interested in your explanations of the GD. However, being not very familiar with the Gold Standard and its implications, your gold-focused explanations are hard to understand, especially in the beginning (Payback time, part one). Couldn’t you do a post explaining a little background on this? I am sure it is all obvious to you, but why is it again a problem if the gold reserve ratio increases?! Those things…


  2. Gravatar of Scott Sumner Scott Sumner
    23. February 2010 at 05:55

    Steve, This is what happens when I leave out the first 3 chapters. I agree it is counterintuitive. Under the gold standard gold is the medium of account. Thus inflation is by definition a fall in the value, or purchasing power of gold. Deflation is a rise in the value of gold. If price levels fall 2%, then each gold ounce buys 2% more goods and services than before.

    So why did gold increase in value? Commodity values are determined by supply and demand. In the case of gold, the supply of newly-mined gold is fairly stable. So the deflation had to be caused by an increase in demand. This could be due to more private demand and/or more central bank demand. The data suggests that central bank demand was the key problem. Central banks held gold to back up currency. These were called gold reserves. If a central bank went from 40% backing to 60% backing (which happened in the early 1930s in some countries) this higher gold ratio meant that central banks demanded more gold. This caused the value of gold to rise in world markets, and led to deflation. I hope this helps.

    BTW, Hoarding of currency in banking panics was also deflationary, as even with a stable gold reserve ratio, more currency being hoarded meant that central banks needed (demanded) more gold to back up that extra hoarded currency.

  3. Gravatar of Steve Steve
    24. February 2010 at 03:05

    Scott, thanks a lot, that does help. Now I will have a look again at your posts about the GD. I printed them, 40 pages. Maybe you could include a pdf version of each post for easier printing, I guess a lot of people find it hard to read that much on a computer screen…

  4. Gravatar of scott sumner scott sumner
    24. February 2010 at 06:51

    Steve, I’ll look into that.

Leave a Reply