An attempt at some deflation clarification
Within economic literature there always seems to be this paradox where sometimes deflation is considered good for an economy and in other times bad. For examples, many including Milton Friedman and Anna Schwartz blame deflation as the primary cause for the severity of the great depression. Whereas others like Michael Bordo and Angela Redish have argued that more often than not, deflation has been correlated with strong economic growth. For instance, they’ve shown that in my homeland of Canada, between 1870-1896, the price level fell a total of 21%, while real GDP was growing respectably at around 2.5% per year through that same time. However, this “paradox” can be solved pretty quick. Both views can be right as soon as we allow for the possibility that there are two distinct types of deflation, each with their own causes and effects. One deflation is bad, the other good.
Coming up with this post, I knew I wanted to write something about money, but I originally wasn’t sure what exactly. I’ve been doing a decent amount of studying on monetary disequilibrium and free banking, and so monetary issues are what I have been thinking about most lately. Then I heard the most recent episode of Macro Musings, where David Beckworth interviews Steve Horwitz about exactly these topics. The writings of these two professors has been a huge influences on my thinking all through my time as a student, so I’ve decided to give an overview of the ideas they discussed, specifically with regards to deflation. A lot of this has already been laid out elsewhere (here for one example), but hopefully this post can work as a nice and clear summary of the two types of deflation, why they happen and what are their effects.
In modern times, the definition of deflation is commonly given as “a fall in the general price level”. However, prices can fall in different ways and for different reasons. As I’ve illustrated in Figure 1, the first type of deflation is caused by a collapse in aggregate demand. This graph is basically taken from David Beckworth’s paper Aggregate Supply-Driven Deflation and Its Implications for Macroeconomic Stability, it’s the type Friedman and Schwartz analyzed during the great depression.
Whether the collapse stems from a decrease in consumption, investment, or Government purchase, the result is a fall in nominal aggregate spending where, all else equal, the price level should be expected to fall as well. It stems from either a reduction in the quantity of money without an offsetting reduction in demand to hold it, or vice versa, by a rise in demand for money without an offsetting increase in the quantity. This excess demand to hold money, creates a scenario where people all want more money than the aggregate amount in existence. And because money trades for real goods and services, the shortage of money means people are trying to sell more labour and goods than is being bought. The result is excess inventories which put downward pressure on prices of goods, services, and wages.
This shows an inward shift of the AD curve due to a decrease in spending. In the short run, the price level will fall from P1 to P2 and output will fall as well, from Y1 to Y2.
If there could be instantaneous movement in prices, or perfect price flexibility, then this wouldn’t be a problem. However, that isn’t the case. Most prices, especially wages are imperfectly flexible, or rigid downward – adjustment takes time. Therefore, there will be economic stagnation as business becomes slack, and this depressed economy only recovers in the long run as prices and wage rates are eventually adjusted to the new monetary equilibrium. Furthermore, there is no reason to believe that price stickiness is uniform throughout the economy, and so prices are likely to adjust to the new equilibrium at different rates. Relative prices will be distorted, sending money and resources away from their best uses.
Because of the negative effects of this demand driven, monetary deflation, economists Bordo and Redish have labelled this type of price level decrease as “bad” deflation, George Selgin has coined it “superfluous” deflation and David Beckworth has labelled it “malign” deflation, and I’m sure lots of other economists have given names to it as well.
This asymmetric short run stickiness is what makes the temporary disequilibrium so painful, but why does it occur? Well, it can be because of wage contracts and unions, but also because of the fact that entrepreneurs cannot be expected to instantly have the knowledge of the cause or the duration of their new lack of revenues. Taking insights from Leland Yeager, entrepreneurs cannot discern whether the sudden lack of business is because customers genuine preferences have been altered, or if there is just a lack of money in the economy. Therefore, when faced with a new and unexpected lack of revenue, some businesses may choose to “ride it out” thinking the lack of spending is only temporary, while others may overreact and layoff too many workers, or close up shop altogether.
Plans of entrepreneurs can be altered in many ways which may not be the optimal choice to achieve the new long run equilibrium. Perhaps the new equilibrium could be achieved sooner if people understood the reason and duration of this lack of revenue that monetary deflation causes, and could then all coordinate the price drop together. However this is not the type of knowledge that prices can be expected to convey. As Yeager points out:
One cannot consistently both suppose that the price system is a communication mechanism— a device for mobilizing and coordinating knowledge dispersed in millions of separate minds—and also suppose that people already have the knowledge that the system is working to convey. Businessmen do not have a quick and easy shortcut to the results of the market process. (Yeager 1986; p.375)
But even if we did assume that some especially attuned entrepreneurs have the knowledge and foresight to understand that an excess demand for money is causing this new lack of spending, there would still be what Yeager describes as the “who-goes-first” problem (ibid p.374-375). This occurs when there is downward pressure on prices and firms therefore need to lower their prices to clear the market, but it is economically disadvantageous for any firm to be the first to do so. In a competitive market, it would not be profitable to lower prices unless costs are lowered first (because, if lowering prices without an accompanying lowering of cost was profitable, competition would have pushed this to happen already). Labourers don’t want to accept a lower wage until goods are cheaper, but it is difficult to make goods cheaper without a lower wage. If the lack of spending is mostly on consumption goods, then the sellers of these goods will want the sellers of intermediate goods to first lower their prices as well and so on. What happens is that everyone is waiting for everyone else (especially labour) to lower its price, in order to try and hold profit constant. This who-goes-first problem is essentially one of coordination, which results in excess inventory sitting on shelves and unnecessary unemployment. In other words, if a business owner lowers the price of the goods he produces without first having a reduction in the cost in the factors of production (as is the case with good deflation), there will be what Horwitz has called a “price-cost squeeze” (Horwitz p.145).
The knowledge problem and the who-goes-first problem, can partially explain what causes the adjustment to be slow and painful, but expectations and nominal contracts can exacerbate this as well. People set up long term contracts under the assumption of certain trends. This falls inline with Irving Fisher’s debt deflation theory. For an example, an honest person who receives a twenty year mortgage, presumably has predicted a steady stream of income to make the necessary payments over time. Unplanned income decreases or unemployment will disrupt these plans. If income decreases and unemployment are systematic, and long term contracts are in markets beyond just mortgages, than this could lead to many debts (which would have been payable in the previous equilibrium), to now be now difficult to pay off. This will cause delinquencies and defaults to excessively rise.
So deflation can be bad, but this is not always the case (Bordo and Redish claim it’s not even the historical norm). Central bankers shouldn’t get their hands on the levers at the first sign of falling prices. Sometimes we should let them fall, for a certain type of deflation makes us better off.
This second type of deflation, what happened in Canada between 1870-1896, is what I’ve depicted in Figure 2 (again, inspired by Beckworth). It shows a deflation that’s caused by positive aggregate supply movement which is not achieved by an increase in the money supply. A permanent outward shift in the aggregate supply curves can only occur with an increase in the natural rate of output, or what we could call an increase in total factor productivity. This would mean that new technology and innovation (or growth in the capital or labour supply), would lower “per unit costs of production” and drive down the price in a competitive market (Beckworth p.367). In this case, all else equal, nominal spending should remain undisturbed.
Figure 2 – Good Deflation
This shows both supply curves shifting outward due to an increase of total factor productivity. Assuming a constant schedule of AD, the price level will fall from P1 to P2, but output will rise from Y1 to Y2.
The quantity of certain goods produced is increased because of improved production techniques, and the consequence will be a fall in the general price level. But this simply reflects the new arrangement of relative prices. It’s a signal that certain goods are less scarce than previous, and it improves the lives of people who demand such goods because they can now get more for less. For these reasons, economists have labelled this type of price level decrease as “good” (Bordo and Redish), “meaningful” (Selgin) and “benign” (Beckworth). I particularly like Beckworth’s health metaphors, but I actually don’t think benign is a strong enough claim here. As Hayek explains, this type deflation is not only benign or “merely not detrimental” but is actually beneficial or “even necessary if disturbances of equilibrium are to be avoided” (Hayek 1929; p.100).
With good deflation, it can’t be stressed enough that it is not a uniform downward pressure on the general price level, but merely a statistical artifact achieved through arbitrary aggregation. For a concrete example, assume China’s economy continues to liberalize, and this causes cheaper imports of car parts to begin flowing into the Czech Republic. This would allow car manufacturers, such as Skoda, to sell cars for a lower price than they did previously. Even if car prices were the only good in the Czech market to drop in price, then by statistical design, the aggregate Czech price level would also drop. In this case, the fall in the general price level is just representing productivity gains of individual firms or markets and not truly showing a pervasive macro phenomenon as the name “a fall in the general price level” may suggest. I think the first form of deflation however, comes much closer representing a pervasive economy wide event, and is the only kind that causes the economic concern.
To sum up, the two forms of deflation can essentially be thought of as stemming from different directions of causality. With good deflation, some businesses find cheaper ways to produce goods and so – as competition lowers the prices of final goods – profit margins can remain constant. Whereas with bad deflation, pressure on prices comes before the improved means of production. Therefore revenue falls faster than cost, and profit margins are reduced throughout the economy. Furthermore, with bad deflation relative prices are distorted as some industries adjust faster to equilibrium than others.
So what’s the takeaway from all this? I’d say it’s that sometimes it makes sense to fear decreases in the price level, but it depends on what causes it. A perfect monetary policy then, would find ways avoid the bad deflation while still allowing for the good. To summarize Horwitz’s conclusion on the podcast: In an ideal world, we could have free banking achieve this. But in the world we live in with central banks, the best we can do would be some sort of NGDP target. An NGDP target, in theory, would stabilize nominal spending (the AD curve) while allowing for productivity shocks to be reflected in the prices level (movements of the AS curves), but whether it could actually work like this in practice is another story.
Beckworth, David. “Aggregate Supply-Driven Deflation and Its Implications for Macroeconomic Stability.” Cato Journal 28.3 (2008): 363-84.
Bordo, Michael, and Angela Redish. “Is Deflation Depressing? Evidence from the Classical Gold Standard.” NBER Working Paper (2003): 1-27.
Fisher, Irving. “The Debt-Deflation Theory of Great Depressions.” Econometrica 1.4 (1933): 337-57.
Friedman, Milton, and Anna J. Schwartz. A Monetary History of the United States, 1867-1960. Princeton: Princeton University, 1963.
Hayek, Friedrich A. 1928. “Intertemporal Price Equilibrium and Movements in the Value of Money,” Money, Capital and Fluctuations: Early Essays. Roy McCloughry, ed.University of Chicago Press, Chicago, 1984.
Horwitz, Steven. Microfoundations and Macroeconomics: An Austrian Perspective. London: Routledge, 2000.
Selgin, George A. Less than Zero: The Case for a Falling Price Level in a Growing Economy. London: Institute of Economic Affairs, 1997.
Yeager, Leland B. “The Significance of Monetary Disequilibrium.” Cato Journal 6.2 (1986): 369-429.