Irving Fisher & debt-deflation
Last week Joe Clark gave me a copy of Irving Fisher’s Econometrica article “The Debt-Deflation Theory of Great Depressions” and added a few less than friendly remarks about Fisher. Having just read the article, I can understand why. But instead of pointing out my disagreements, I think it is more valuable to point out what I think are the points of truth in the Fisher thesis.
Fisher suggests that the underlying cause of any large recession (he cites 1837, 1873, 1929) is the twin problems of (1) over-indebtedness; and (2) deflation. These then interact on net worth, profits and trade. Specifically, he points to the problem where de-leveraging (ie paying off debt) leads to deflation, which means the remaining debt is relatively larger… so that paying off debt puts people into more debt. And that makes people sad.
Fisher dramatises the issue (at one point even invoking starvation), but he is right to note the important role of money. As Friedman & Schwartz later showed, it was the sharp tightening of money supply by the Fed (along with the protectionist policies and increased political risk that came from congress) that turned a normal downturn into a “great” depression.
Money is important. Fisher is right to note that a monetary distortion can have devastating consequences on the real economy. This can be explained through Monetarist theory (where people struggle to distinguish between real growth and growth in the money supply) or through Austrian business cycle theory (where monetary distortions leads to mal-investment because different industries react differently to changed interest rates). Either way, a distortion in the money supply leads to a distortion in the real economy.
It is important to understand what “distortion” means here. To abstract from monetary considerations it is necessary to imagine a world with no monetary distortions, so that broad money supply (money base * credit multiplier * velocity) moves exactly in line with total production, so that there is no change in the general price level and all price changes reflect “real” (not “monetary”) changes. Of course, this is an impossible abstraction for many reasons — it is impossible to know exactly how much is being produced, exactly how much broad money exists at any point in time, or how to correctly measure a “general” price level. It is equally fanciful to expect any economic system to be free of monetary distortions. But the idealised abstract is helpful nonetheless.
In this hypothetical world, if the price of apples were to increase we would know that this represents a real change in the world somehow… perhaps there is a craze for apple-pies or maybe there was a storm that negatively affected an apple orchard. The possible cause of “monetary shock” is abstracted out of the equation.
But once we re-introduce monetary shocks into the system, there is forever a threat that a change in price is not a representation of reality, but simply a change in the value of the measuring tool (ie inflation or deflation). When this happens, the resulting change in behaviour can lead to real distortions in the economy. The most common example (provided both in Monetarist and Austrian theory) is when monetary expansion leads to short-term changes in behaviour (over-investment in Monetarist theory, mal-investment in Austrian theory), but these changes are unsustainable in the long-term once price have adjusted and then we have an inevitable correction (ie economic slowdown).
The common theme is that the short-term impact is powerful, but must then be “fixed”. The key difference is that Monetarist theory assumes the fix is a shift in the aggregate behaviour (lower or higher) while Austrian theory notices that the shift can be between sectors (mal-investment). On this point of difference, the Austrians are correct. This is important when we consider the consequence of monetary contractions.
In Monetarist theory, a monetary contraction will lead to under-investment which will eventually be “fixed” through an increase in investment. In Austrian theory, a monetary contraction will lead to under-investment and mal-investment, which will eventually be “fixed” both with some increased investment in the right areas and also some business failures where there was mal-investment. Both agree that monetary contraction can lead to an economic downturn, but the Austrian theory suggests a period of extended pain as the economy later redistributes resources to their “correct” place.
This is all to say why we do not want monetary distortions. Some Austrian economists have taken an aversion to monetary distortions to the extreme, arguing that there should be laws that try to restrict the credit multiplier (ie abolish fractional reserve banking). They argue that the monetary distortions created by changes in the credit multiplier are so severe that they undermine the strength of the economy and so we would be better off without them. They also tend to wrap their arguments up in faulty semantics of “fraud” and neglect the powerful benefit of credit-matching that is achieved through banks. But I think their key mistake is that they over-estimate the sensitivity of the real economy to changes in broad money.
As suggested earlier, the idealised abstract of no monetary distortions is an impossible counter-factual. The truth is that there will always be some monetary distortion in any system, including in the “no credit multiplier” existence. In a gold-currency situation, discovery of new gold will create a distortion. So will the non-discovery of gold if there is an increase in production. And even if those could match, there is no controlling the level of velocity (how often the same piece of money is spent in a time period). And the truth is that it is impossible (without an absurdly totalitarian government) to prevent a credit multiplier.
The truth is that while large monetary distortions cause significant problems in the real economy, small monetary distortions aren’t a big problem. Large distortions can come about through a single large mistake, but they can also come about through an accumulation of minor mistakes when there is an insufficient feedback mechanism.
The important point is that a good monetary system needs to have a good feedback mechanisms to ensure that broad money supply moves roughly in line with production so that there are minimal monetary distortions. This is generally true of gold (if there is “too little” gold then it’s value increases and so there is a greater incentive to find more), and of credit (more productive opportunities means an increased demand, which drives the credit multiplier). It is this sort of self-correcting money that is the goal of “inflation targeting”, and (I believe) would be in the interests of a profit-maximising private currency supplier.
One of the big arguments of monetary policy is which policy will best achieve monetary stability. I won’t comment on that debate now except to say that it certainly isn’t a policy of demand management (ie activist Keynesian policy).
This is all a long way of saying that Fisher is right to worry about how deflation can hurt the real economy, but he was wrong to focus only on deflation instead of looking more broadly at monetary distortions.
There is another point worth quickly noting. If there has been previous inflation, the best policy for the future is price stability at the higher price, not a deflation back to the previous price. The previous monetary distortion is a sunk cost and it is not possible to go back and un-inflate. That will simply be building a new monetary distortion on top of the old monetary distortion.
One example of this is with the 2008 world recession. One of the underlying causes was monetary expansion (mixed with bad regulation and poor risk-assessment) which led to asset inflation in the American housing market. The inevitable correction lead to the world recession. At that point we saw a significant monetary contraction, caused by a drop in the credit multiplier. If the money suppliers had allowed this to occur we would have effectively had two recessions — the first a correction from monetary expansion, and the second as a result of monetary contraction.
The correct response (and the one followed by most money suppliers) was to increase base money supply to offset the drop in the credit multiplier. In other words — cut interest rates. The reason was not (as suggested by Keynesians) to pump-prime the economy… but instead to prevent any further deflationary monetary distortions. This response was a major difference between the “great” depression and recent world recession.