Friday, April 25, 2014

Joan Robinson on the Weimar Hyperinflation

Joan Robinson’s review of Costantino Bresciani-Turroni’s The Economics of Inflation (The Economic Journal 48.191 [1938]: 507–513) contains Robinson’s views on the causes of the Weimar hyperinflation.

In essence, there are three major explanations of Weimar hyperinflation, as follows:
(1) the standard “quantity theory of money” explanation, which was sometimes called the “English” or “Allied” view. An influential statement of this view was Bresciani-Turroni (1937), and it assumes that money supply is exogenous;

(2) the “balance of payments theory” or what was called the “German” view. This explanation was pursued by Williams (1922), and Graham (1930) is considered by some to hold to some form of it;

(3) a Post Keynesian “balance of payments” theory but emphasising endogenous money and wage and price versus foreign-exchange spirals.

Joan Robinson (1938) formulated this on the basis of the earlier “balance of payments” view.
First, it is necessary to stress what the argument is not about. The argument is not about whether money supply exploded in Weimar Germany or whether the Weimar government ran a huge budget deficit in 1923 that was funded by central bank money creation (or monetising of the deficit).

These things are true.

Nor, as far as I can see, do Post Keynesians dispute that a sustained general increase in prices or hyperinflation requires a growing money stock to sustain it.

But, while a money supply expansion is a necessary condition for this, what is being disputed is whether exogenous money creation was the primary causal factor. The alternative view is that money creation can be seen largely as endogenous and a type of intermediate, though necessary, cause driven by other factors.

The “balance of payments theory” argues that the severe depreciation of the German mark was the fundamental causal factor in the hyperinflation, and that in turn was driven by the burden of the reparations imposed on Germany.

We can see how the Post Keynesian view contradicts the “quantity theory” view by looking at how both sides understand the causality involved in the hyperinflation.

If one accepts the “quantity theory” view and assumes an exogenous money supply, then the chain of causality runs as follows:
exogenous money supply → price inflation → the exchange rate depreciation.
The Post Keynesian view rejects this, and the depreciation of the mark is seen as the fundamental causal mechanism.

So under the Post Keynesian “balance of payments” view, then, the chain of causality runs as follows:
the exchange rate depreciation → price and wage inflation → inflation of money supply.
How does Robinson’s review fit into this?

First, Robinson noted that the facts seem to support the “balance of payments view”: the severe inflation that hit Germany in the later half of 1921 was preceded by the severe depreciation in the exchange value of the mark from May 1921 to November 1921 when the German government began paying large cash reparations payments (Robinson 1938: 507–508).

Of course, one has to demonstrate the causal mechanisms involved in this and look for further empirical evidence in support of it, and Robinson points to this:
“If the impulse comes from the side of the exchange, we should expect the fall in exchange to run ahead of the rise in prices, and the prices of traded goods to run ahead of home prices. This is precisely what occurred, the magnitude and speed of change being in this order: exchange, import prices, export prices, home prices, cost-of-living, wages. Moreover, the geographical diffusion of prices supports the argument, the movement spreading from the great ports and commercial centres to the interior of the country.” (Robinson 1938: 508).
Next, Robinson notes that the proponents of the “quantity theory” explanation of the hyperinflation like Bresciani-Turroni argued that this cannot be the case, because the tendency to balance of payments equilibrium would have prevented this, and that it was the German budget deficit financed by Reichsbank money creation that was the primary cause of the hyperinflation (Robinson 1938: 509).

But Robinson notes that the actual German government budget deficit was falling in 1921 and 1922: in 1920 the deficit was 6 billion gold marks, in 1921 it was 3.7 billion gold marks, and in 1922 2.4 billion (Robinson 1938: 509). Although Robinson concedes it had a role, nevertheless both the “quantity theory” view and the former “balance of payments theory” view had neglected an important missing link in the cause of the hyperinflation:
“The missing item is not far to seek. It is the rise in money wages. Neither exchange depreciation nor a budget deficit can account for inflation by itself. But if the rise in money wages is brought into the story, the part which each plays can be clearly seen. With the collapse of the mark in 1921, import prices rose abruptly, dragging home prices after them. The sudden rise in the cost of living led to urgent demands for higher wages. Unemployment was low …, profits were rising with prices, and the German workers were faced with starvation. Wage rises had to be granted. Rising wages, increasing both home costs and home money incomes, counteracted the effect of exchange depreciation in stimulating exports and restricting imports. Each rise in wages, therefore, precipitated a further fall in the exchange rate, and each fall in the exchange rate called forth a further rise in wages. This process became automatic when wages began to be paid on a cost-of-living basis … . Thus … [Bresciani-Turroni’s] contention that the collapse of the mark cannot have caused the inflation, because the exchange rate will always find an equilibrium level, is deprived of all force as soon as the rise of money wages is allotted its proper role. ….

Without rising money wages, inflation cannot occur, and whatever starts a violent rise in money wages starts inflation.” (Robinson 1938: 510).
To the extent that the Reichsbank met the demand for high-powered money and loans from the private sector as wages and prices soared and there were also both official and unofficial “supplementary currencies” improvised during the course of the hyperinflation to meet demand for money (Robinson 1938: 511), then the explosion in the money supply was endogenous to a great extent.

If one regards the Reichsbank’s money creation to finance the German budget deficits and especially the huge deficit of 1923 as “exogenous” money creation, then the continuance of hyperinflation was dependent on both endogenous and exogenous money creation: this is what “allowed it to continue” (Robinson 1938: 511). Robinson also conceded that the very large German budget deficit of 1923 also had a role in the hyperinflation (Robinson 1938: 510).*

But even these observations do not lessen the role of (1) the exchange rate depreciation as the fundamental and primary causal driver of the hyperinflation and (2) the significant role that endogenous money creation played in the explosion of the money supply.

Addendum
To gauge the role of the German government budget deficits in the period from 1919 to 1923, we can look at this graph below of German government spending and tax revenue (in billions of gold marks) (from Eichengreen 1995: 138, Table 5.1).


As we can see, the hyperinflation was accompanied by a disastrous collapse of tax revenues, especially in 1923 when the deficit widened considerably. Government spending actually fell in 1922 and only rose slightly in 1923. The major cause of the massive budget deficit in 1923, then, was tax revenue collapse, not massive increases in government spending.

Another point is that budget deficits in 1919 and 1920 were very large, but inflation did not spiral out of control in these years even though it did rise appreciably (as seen in the second graph below). If such large budget deficits were the primary cause of hyperinflation, then why did it not begin earlier in 1919 or 1920? Why did it spiral out of control only after the currency depreciation?



Note
* And we can note that Joan Robinson (1938: 510), far from being some unreasonable dogmatist, was perfectly prepared to concede that excessively large budget deficits in periods of very low unemployment can initiate inflations, and that it is “even possible that an increase in the quantity of money might start an inflation” (Robinson 1938: 511), though that was not the primary cause of German hyperinflation.

Further Reading
There is an excellent analysis of the hyperinflation here by Matias Vernengo:

Matias Vernengo, “Inflation, Hyperinflation and Monetarist Perceptions about the Functioning of the Economy,” Naked Keynesianism, Monday, March 24, 2014.

BIBLIOGRAPHY
Bresciani-Turroni, Costantino. 1937. The Economics of Inflation: A Study of Currency Depreciation in Post-War Germany (trans. Millicent E. Sayers), Allen & Unwin, London.

De Grauwe, P. and Polan, M. 2005. “Is Inflation Always and Everywhere a Monetary Phenomenon?,” Scandinavian Journal of Economics 107: 239–259.

Eichengreen, Barry J. 1995. “The Legacy of Hyperinflation,” in Barry J. Eichengreen, Golden Fetters: The Gold Standard and the Great Depression, 1919–1939. Oxford University Press, New York. 125–151.

Graham, Frank Dunstone. 1967 [1930]. Exchange, Prices, and Production in Hyperinflation: Germany, 1920–1923. Russell, New York.

Robinson, Joan. 1938. Review of The Economics of Inflation by C. Bresciani-Turroni, The Economic Journal 48.191: 507–513.

Williams, John H. 1922. “German Foreign Trade and Reparations Payments,” Quarterly Journal of Economics 36: 482–503.

22 comments:

  1. I think that Robinson gives too much credence to the view that the budget deficit was a major causal factor. I'll bet that if you compared the budget deficits in all the major European powers in this period they would all have been very high in 1920-1921. This chart comes to mind (not the deficit but indicates a very high deficit in these years...):

    http://www.economicshelp.org/wp-content/uploads/blog-uploads/2012/09/national-debt-percent-1900-12.png

    So, why no hyperinflation in Britain, France etc?

    ReplyDelete
    Replies
    1. That is a pretty good question.

      Another question: did German tax revenues massively collapse in 1923, and was that a major cause of the budget deficit?

      The data in B. Eichengreen, 1995. “The Legacy of Hyperinflation,” in Barry J. Eichengreen, Golden Fetters: The Gold Standard and the Great Depression, 1919–1939. New York, Oxford University Press. 125–151, at p. 138, Table 5.1, shows that it was a catastrophic collapse in tax revenue that caused the huge German budget deficit in 1923, not a massive increase in government spending.

      Delete
    2. Yes, that is the endogenous case that I made in the comments in the last post. Prices went up for government purchases and the tax collection system went into meltdown.

      When you pile up all the data I think there is only one realistic interpretation, to be honest.

      Delete
    3. I have also add an addendum above with a graph of expenditure and tax revenues.


      Delete
    4. I'm a bit confused how you are using the phrases "exogenous" and 'endogenous" money in the context of a real historical event as opposed to in a model where the phrases have a clear meaning in relation to how the money supply is derived in that particular model.

      The equation of exchange simply shows that when M changes then there must be accompanying changes in P,Q or V. Its says nothing about the causality or the order of these changes.

      In relation to the German hyperinflation you distinguish between the expansion of the money supply "to [meet] the demand for high-powered money and loans from the private sector" which you define as "endogenous"
      and "money creation to finance the German budget deficits" as "exogenous"


      Why is expanding the money supply to accommodate govt deficits any more "exogenous" than expanding the money supply to accommodate the demand for more loans ?
      The mechanism for the expansion are similar (if not identical) in both cases and the monetary authorities have the theoretical ability to accommodate/not-accommodate these things equally.

      Of course political pressure (and the general social situation) may have rendered the decision to stop money creation via either channel impossible to make - but that is a different issue.


      (BTW: You seem to believe that non Post-Keynesian economists must think that the German hyperinflation was caused by an apparent random decision to rapidly expand the money supply "exogenously". Obviously this is total rubbish.)




      Delete
    5. (1) regarding "exogenous" and 'endogenous" money, there is a real debate to be had there.

      Some Post Keynesians may well dispute my idea that the central bank money creation to fund the deficit was exogenous.

      (2) on the direction of causation in the QT, you are just being disingenuous.

      Neoclassicals do normally insist the direction of causation runs from money supply to price level.

      (3) " You seem to believe that non Post-Keynesian economists must think that the German hyperinflation was caused by an apparent random decision to rapidly expand the money supply "exogenously". Obviously this is total rubbish."

      I can only conclude you are being disingenuous here too.

      If there are many mainstream neoclassical economists who accept the "balance of payments" theory of the Weimar hyperinflation with the emphasis on endogenous money, then name a few for me.

      Delete
    6. I have no particular interest in defending so-called Neo-classicals. But when I read various blogs etc written by economists who you would probably call Neo-Classical I rarely see them espouse the views you glibly attribute to them. The truth is much more nuanced than the tribalist view you seem to subscribe to.

      Delete
    7. Could you please provide links Rob Rawlings to substantiate your claims?

      Delete
    8. Just google "Krugman and Inflation" or "Rowe and Inflation" or "Wren-Lewis and inflation" and you will see their views on inflation are not as simplistic as the neo-classical parody that post-Keynesian like to assume.

      Delete
    9. You mean all the guys that we've been heckling for years on this? Doesn't count, sorry!

      Delete
  2. Note from the data presented that the highest deficit was in 1919 and inflation, while bad, was nowhere near hyperinflation territory.

    ReplyDelete
  3. "the depreciation of the mark is seen as the fundamental causal mechanism.... the chain of causality runs as follows: the exchange rate depreciation → price and wage inflation → inflation of money supply."

    What causes the initial depreciation of the mark?

    Like all prices, the exchange rate of the mark was determined by supply and demand. A decrease in the exchange rate occurs when there is (1) a decrease in the demand, or (2) an increase in the supply.

    Are you saying a decrease in the demand for the mark is "the fundamental causal mechanism"?

    ReplyDelete
    Replies
    1. Yes, I can see the game you're playing, Anonymous: a game of "I'll-go-you-one-better" in the causal sequence of explanation.

      That supply and demand dynamics apply to the depreciation of the mark (as people lost confidence in it), however, does not refute the view that the prior unrealistic and unsustainable reparation burden was the fundamental cause of this crises, as incidentally I already said above.

      When the "London ultimatum" in May 1921 was issued and the Allies announced a sum of 132 billion gold marks as the figure for reparations, and an initial payment of 50 billion gold marks to paid immediately, it was politically impossible to raise taxes in Germany and the government faced the real threat of invasion and occupation of Germany and the destruction of the legitimacy of the new Republican government, they were forced to some degree of money printing to buy foreign exchange to pay the burden.

      Nevertheless, the subsequent causal sequence I have given is still valid:

      the exchange rate depreciation → price and wage inflation → further endogenous inflation of money supply

      Delete
    2. You didn't answer my question: what causes the initial depreciation of the mark?

      You do not explain what causes the initial depreciation of the mark. It must be either (1) an increase in the supply of marks or (2) a decrease in the demand for marks.

      Are you saying a decrease in the demand for the mark was "the fundamental causal mechanism"?

      Delete
    3. I have already answered your question.

      Yes, demand for the mark fell and it lost value as the people lost confidence in it and the Reichsbank created more marks to buy foreign exchange. Yes, that can be analyzed in terms of supply and demand.

      But it makes no sense to say that this was just the ultimate cause of the hyperinflation, end of story.

      On the contrary, the fundamental causal mechanism was the unrealistic and unsustainable reparations forced on Germany by the "London ultimatum" in May 1921.

      Delete
    4. You explain the "causal sequence" of the hyperinflation. However, you have not adequately explained what initiated the "causal sequence".

      You claim "the fundamental causal mechanism was ... the 'London ultimatum' in May 1921."

      This must be wrong because the price inflation started before May 1921.

      8 marks bought 1 US dollar in December 1918. 47 marks bought 1 US dollar in December 1919. 73 marks bought 1 US dollar in December 1920.

      Why did the mark depreciate by 90% between December 1918 and December 1920? Was it (1) an increase in the supply of marks, or (2) a decrease in the demand for marks?

      Delete
    5. Nobody's denying the depreciation between 1918 and early 1921.

      However, it's quite clear that this, though serious, did not induce hyperinflation.

      That came after May 1921 as you can very clearly in the chart above.

      What I said is that the 'London ultimatum' in May 1921 induced the severe depreciation in the latter half of 1921 that was clearly the causal factor that set off the hyperinflation.

      Delete
  4. Our host:"Yes, demand for the mark fell and it lost value as the people lost confidence in it and the Reichsbank created more marks to buy foreign exchange."

    Lost confidence? Why? Perhaps because those losing the confidence anticipated the creation? I think you have not answered Anon's point here. While I can well believe other factors make the short term unlike the long term for inflation, it still looks like hyperinflation only happens when the supply of money increases so sharply and drastically it overwhelms the damping and feedback loops that otherwise determine the short term. Which I take to be anon's point.

    ReplyDelete
    Replies
    1. (1) "Perhaps because those losing the confidence anticipated the creation?"

      Well, Ken B, confidence in the mark was affected to some extent by Germany's loss of the war and huge figure for reparations suggested in the aftermath when some depreciation did occur. But there was no hyperinflation then.

      However, there was no final decision on what figure would be paid until May 1921 and confidence did return and the mark stabilised. Read the account here (from Holtfrerich, Carl-Ludwig. The German Inflation 1914-1923: Causes and Effects in International Perspective (trans. Theo Balderston). De Gruyter, New York, 1986).

      As you can see, there is good evidence that by 1920 confidence in the mark had been restored and that confidence and optimism "even remained dominant until the London Ultimate of May 1921".

      And, as I said, the depreciation ended: "the value of money remained almost stable, for the gold mark was worth 14.2 paper marks in April 1920 and 14.9 in March 1921" (Bresciani-Turroni, Costantino. 1937. The Economics of Inflation: A Study of Currency Depreciation in Post-War Germany, p. 56).

      Holtfrerich even notes on p. 193:

      "The importance of expectations and confidence in the currency to the course of the inflation is registered in the fact that between spring 1920 and May 1921 the price level scarcely changed despite considerable expansion of the money supply."

      But when the huge reparations figure was finally announced in May 1921, many people -- international investors and so on -- would have understood that it was going to be impossible for Germany to pay it. That this would cause some kind of severe crisis in the mark's value -- in addition to the extra marks the government created to buy foreign exchange -- seem pretty reasonable to me.

      (2) On point 2, I have stated explicitly above that an increasing money supply is a necessary condition for hyperinflation. I just dispute that it was the primary causal factor here, rather than an intermediate though necessary factor.

      Delete
    2. Exactly. Lord Keynes ignores inflationary expectations.

      Expectations about money creation affects the demand for money. Inflationary expectations cause a decrease in the demand for money. All else equal, inflationary expectations cause the exchange rate to fall.

      The German money supply increased by 1,100% between 1914 and May 1921. People realized that the money printing would never stop. Inflationary expectations changed, and the demand for marks collapsed.

      Delete
    3. (1) "Lord Keynes ignores inflationary expectations. "

      No, I don't. In fact, just read what I wrote above, Anonymous.

      (2) To state my view again: I do not reject the possibility that inflationary expectations played some role, it was just that after May 1921 all these factors were set off by the reparations news.

      And three points quite clearly contradict your "People realized that the money printing would never stop" view:

      (i1) The initial post WWI mark depreciation ended in 1920:

      "the value of money remained almost stable, for the gold mark was worth 14.2 paper marks in April 1920 and 14.9 in March 1921" (Bresciani-Turroni, Costantino. 1937. The Economics of Inflation: A Study of Currency Depreciation in Post-War Germany, p. 56).

      (ii) Holtfrerich p. 193:

      "The importance of expectations and confidence in the currency to the course of the inflation is registered in the fact that between spring 1920 and May 1921 the price level scarcely changed despite considerable expansion of the money supply."

      (iii) it appears that by 1920 confidence in the mark had been restored and that confidence and optimism "even remained dominant until the London Ultimate of May 1921".

      Delete
  5. I have a collection of explanations for hyperinflation:

    http://howfiatdies.blogspot.com/2013/09/hyperinflation-explained-in-many.html

    ReplyDelete