Martin Wolf claims that the Austrian Business Cycle Theory was more nearly correct than mainstream macreconomic theory in explaining the crisis here. Wolf interprets the Austrian theory to be that:
inflation-targeting is inherently destabilising; that fractional reserve banking creates unmanageable credit booms; and that the resulting global “malinvestment” explains the subsequent financial crash. I have sympathy with this point of view. But Austrians also say - as their predecessors said in the 1930s - that the right response is to let everything rotten be liquidated, while continuing to balance the budget as the economy implodes.
I have become more and more critical of inflation targeting. I strongly disagree with the claim that fractional reserve banking creates unmanageable credit booms and what I believe is the dominant "free banking" faction of Austrian economists agrees. I do think malinvestments should be liquidated, but favor keeping cash expenditures growing at a slow steady rate during the process. I certainly oppose discretionary fiscal policy as a means of keeping cash expenditures on target. On the other hand, I don't favor raising tax rates to balance the budget in a recession. As for cutting government spending--it depends.
Paul Krugman responded by criticizing the Austrian theory here. He rejects the argument that the unemployment during recessions is due to an unfortunately slow and painful need to liquidate malinvestments and shift labor and other resources to more productive uses. He argues that if that were true, then the boom would also generate symmetrical unemployment as labor and other resources where shifted towards the malinvestments.
Krugman expands upon his argument here. He claims that advocates of the Austrian theory sneak in changes in aggregate demand to explain the lack of symmetry. During the boom, the malinvestments develop in an environment of growing aggregate demand. During the recessions, the malinvestments are liquidated in an environment of shrinking aggregate demand. Krugman argues that it is these changes in aggregate demand that are causing the boom and bust. He claims that the unwillingness of Austrians to develop formal models is what allows them to evade this reality.
Tyler Cowen more or less agrees with Krugman here. Cowen is one of those economists who began as a dogmatic Rothbardian and then over time, came to a more eclectic view. Rothbard's version of the Austrian theory is close to what Wolf described above. Currently, the Mises Institute champions his view.
My macroeconomic views developed in much the same way as Cowen, and like Cowen, I think Krugman has a point. Cowen says that the Austrian theory should not be dismissed but that it should be modified to include insights from Keynesian theory, bubble theory, modern financial theory, and real business cycle theory.
Arnold Kling responds to Krugman here. Kling argues that booms are gradual and busts are sudden. That is why there is little unemployment in the boom and high unemployment in the bust.
I think Kling is correct that a slow change in the composition of demand and the allocation of resources can occur with minimal structural unemployment, while a sudden change will result a large increase in unemployment. On the other hand, I don't see how this fits in with the Austrian theory. Why should money creation create a gradual change in the allocation of resources that must be suddenly reversed?
Peter Boettke comments on the discussion here. Boettke suggests that there are commonalities between some aspects of Keynesian and Austrian theory. He also suggests that the existence of heterogeneous capital goods is an important element of the Austrian theory.
I agree with Krugman that the Austrian Theory of the Business Cycle is a bit unclear and blends together a variety of market processes. Institutional assumptions and judgments about empirical magnitudes are smuggled in by simply focusing on one particular chain of cause and effect. I will illustrate this by giving several different accounts of the Austrian theory.
The first begins with an increase in the quantity of money. The ceteris paribus conditions include a given demand to hold money and an unchanged supply of saving and demand for investment. In other words, the natural interest rate doesn't change. The natural interest rate is the level of "the" interest rate that keeps saving--that part of income not spent on consumer goods and services, equal to investment--spending by firms on capital goods. At the same time, it is the level of "the" interest rate that keeps total spending in the economy equal to the productive capacity of the economy.
Given those ceteris paribus conditions, the increase in the quantity of money immediately results in an excess supply of money. As an institutional assumption, the excess money is lent into existence. Perhaps the central bank lowers the interest rate it charges for loans. Perhaps it purchases short term securities, raising their prices, and lowering their yields. Regardless, the short run effect is a reduction in the market interest rate.
The focus here is on monetary disequilibrium. Because money serves as medium of exchange, people accept it in payment even if they have no intention of increasing their money holdings but rather intend to spend the money. The excess supply of money exists as increased real money balances that people are actually holding, but only because they have yet to spend them.
This increase in both nominal and real balances is matched by an increase in both the nominal and real supply of credit. This is due to the particular institutional assumption, that the money was lent into existence. This increase in the real supply of credit only persists as long as there is an increase in the real quantity of money, matching the excess supply of money.
Because saving supply and investment demand are assumed to be unchanged, and so the natural interest rate is also unchanged, the excess supply of money and increase in the real supply of credit causes the market interest rate to fall below the natural interest rate.
The impact of the lower market interest rate on the composition of demand depends on the interest elasticity of investment demand relative to the interest elasticity of saving supply. If investment demand is more interest elastic than saving supply, then investment rises more than saving falls. Since saving is that part of income not spent on consumer goods and services and investment demand is a demand for capital goods, the demand for capital goods grows more rapidly than the demand for consumer goods.
If the interest elasticity of investment demand and saving supply are equal (though opposite, of course,) then the demand for both capital goods and consumer goods and services grow in proportion.
And finally, if investment demand is less interest elastic that saving supply, then the demand for consumer goods and services grows more rapidly than the demand for capital goods. The impact of monetary disequilibrium--even if the money is lent into existence--depends on relative interest elasticities.
In the traditional Austrian story, this is more or less ignored. At this high level of aggregation, the implicit assumption is that the supply of saving is perfectly inelastic so that the decrease in the market interest rate directly increases only investment demand--spending by firms on capital goods.
Of course, these broad aggregates could be broken down further--durable consumer goods, single family homes, software, commercial structures, equipment--or even down to each and every good or service. There is nothing in theory that prevents the demand for restaurant meals from being more interest elastic than other goods. The lower market interest rate provides less future compensation for sacrificing the utility received from current restaurant meals.
Oddly enough, the typical Austrian story assumes that the immediate expansion in demand is determined by where the "new" money is spent rather than the interest elasticity of the demand for particular goods and services. This is an error.
Consider the following example. Households refrain from going out to eat and purchase bonds. Firms sell bonds to fund the purchase of drill press machines. A central bank appears and makes loans at a lower rate of interest to firms buying drill press machines. All of the "new" money is devoted to the purchase of drill press machines. The firms that borrow from the central bank don't purchase bonds from households. What do the household do with the "old" money that they would have used to purchase bonds from the firms? With the usual assumptions regarding the allocation of consumption over time, the households will use at least some of the "old" money to purchase restaurant meals. While the central bank has made no consumer loans, the households simply use more of their income to buy restaurant meals, save less, and purchase fewer bonds. Only if the supply of saving, and so, the demand for restaurant meals, is perfectly interest inelastic, will the impact of expansion of the quantity of money be limited to the particular place the central bank lent the funds.
The institutional assumption that the new money is lent into existence makes the transformation of an excess supply of money into an increase in the supply of credit plausible. After that point, however, determining which money is "new" and which is "old" and tracing the money is pointless. Interest elasticity becomes the key.
The effect of the increase in demand on relative prices and the composition of output depends on the price elasticities of supply and demand for various goods. In the special case where the supply of saving is perfectly inelastic with respect to the interest rate, while the demand for investment does have the usual negative relationship, then the demand for capital goods rises, and there is no immediate impact on the demand for consumer goods and services. With a normal upward sloping supply curve for capital goods, their prices and quantities both rise in response to the increase in demand. Since the demand for consumer goods and services didn't change, the prices of capital goods have risen relative to consumer goods and the composition of aggregate output shifts towards to production of capital goods. (With scarcity, this should shrink the production of consumer goods.)
More generally, if the demand for both consumer and capital goods rises, because both saving and investment are somewhat interest elastic, then what happens to both relative prices and the quantities of the goods depends on price elasticities of supply and demand of the various goods and services.
Suppose that investment demand is more interest elastic than saving supply, so the demand for capital goods rises more than in proportion to the demand for consumer goods and services. If the supply of both consumer and capital goods are perfectly elastic with respect to price, then the quantities of capital and consumer goods and services with both rise in proportion to the increase in their respective demands. The composition of output would shift with an expansion in the production of capital goods relative to consumer goods. Relative prices remain the same.
Suppose instead that both supply and demand curves have their usual upward and downward sloping shapes respectively, and the elasticities of supply and demand are the same for both consumer goods and services and capital goods. In this rather special case, then with the demand for capital goods rising more than in proportion to the demand consumer goods and services, then the prices of capital goods rise relative to the prices of consumer goods and services and the production of capital goods rises relative to the production of consumer goods and services.
On the other hand, suppose that the supply and demand for consumer goods are highly price inelastic, while the supply and demand for capital goods are highly elastic. Even if the demand for capital goods rises more than in proportion to the demand for consumer goods, it is possible that the prices of consumer goods will rise relative to the prices of capital goods.
The composition of the change in output also depends on price elasticities of supply and demand. Suppose that the demand for capital goods rises more than in proportion to the demand for consumer goods. If the supply of consumer goods are highly price elastic and the supply of capital goods is highly price inelastic, then the prices of capital goods could rise relative to the prices of consumer goods, but the production of consumer goods could expand more than in proposition to the production of capital goods.
Simple supply and demand analysis would be adequate if the impact of a change in the interest rate on the relative prices and quantities of two goods making up a small share of the market were being examined. Implicit in the slopes of the supply and demand curves for those two goods would be all the other goods and services that might be produced or consumed.
However, with a macroeconomic analysis, the fundamental economic principle of scarcity must be included. For example, if all goods are divided between consumer goods and services or capital goods, then an expansion in the production of additional capital goods shifts resources--land, labor and existing capital goods--away from the production of current consumer goods and services. Similarly, an expansion of the current production of consumer goods and services requires a shift of existing resources away from the production of new capital goods.
A special situation consistent with scarcity of resources would be if the supply elasticity of both consumer goods and services and capital goods are perfectly inelastic. If the demand for capital goods is more interest elastic than the demand for consumer goods and services, and the price elasticities of demands are the same, then the lower market interest rate will raise the prices of capital goods relative to consumer goods and enhance the profitability of producing capital goods relative to consumer goods. (This result would be reversed if the price elasticity of demand for consumer goods was sufficiently lower than the price elasticity of demand for capital goods.) Of course, because of the assumption that both supplies are perfectly inelastic, there would be no change in the composition of output or the allocation of existing resources between the production of consumer and capital goods.
Now, suppose these elasticities are only perfectly inelastic in the short run. In the long run, entrepreneurs seek profit by reallocating resources between firms and industries. If profits are enhanced in the capital goods industries relative to consumer goods industries, then more resources must be allocated to capital goods industries, expanding their output and reducing their relative prices and profitability. The other side of the coin is that fewer resources are allocated to consumer goods industries, reducing their production and raising their prices and profitability.
It seems quite reasonable, but it is an illusion. The short run price adjustments will cause the excess supply of money to disappear, and so end the increase in the real supply of credit. The nominal supply of credit doesn't decrease. It is that the nominal supply of loans, augmented by the increase in the nominal quantity of money, no longer can purchase an increased amount of consumer and capital goods at higher prices.
In other words, the analytical short cut of imagining a short run equilibrium set of prices of consumer and capital goods consistent with the lower market interest rate, and then considering some kind of output adjustment that equalizes profitability is mistaken.
Assuming that the elasticities of supply are not perfectly inelastic, then what is really involved is an assumption that a market with a greater shortage can pull resources away from a market with a smaller shortage. This seems plausible enough. If the demand for capital goods is more interest elastic than the demand for consumer goods, and the price elasticities of demand and supply in the markets are the same, then an excess supply of money that is lent into existence will create a greater shortage of capital goods than consumer goods. The production of capital goods expands at the expense of the production of consumer goods. As this occurs, the shortage in consumer goods industries becomes larger and the shortage in capital goods industries becomes smaller because of changes in actual quantities relative to quantities demanded.
Of course, the shortages are simultaneously providing incentives for firms producing and selling both capital and consumer goods to raise their prices. It is this process that ends the "boom" and brings on the "recession." As the prices of both capital and consumer goods rise, the real quantity of money falls. As the real quantity of money falls, so does the real supply of credit. Once the real quantity of money has fallen enough so that it has returned to level of real money balances people want to hold, then the real supply of credit is no longer enhanced by the excess supply of money. The real supply of credit has returned to its initial level. The market interest rate, no longer clearing this additional real supply of credit, rises back to the natural interest rate.
If the economy really remains at full employment during this process, then the production of consumer goods decreases in the "boom" even though growing demand has led to shortages of consumer goods. As the economy goes into the recession phase, the production of consumer goods must rise and the production of capital goods must fall.
Because the higher price level reduces the excess supply of money, reduces the real supply of credit, and raises the market interest rate, it is natural to assume that demands for consumer and capital goods fall. And so, the lower demand result in less production and employment. In other words, a recession.
But this is an illusion. The higher market interest rates does reduce demand, but there are shortages, and so the reduction in demand is bringing demand back down to what can actually be produced. The higher interest rates are closing the shortages. While the shortages of the more interest elastic goods close faster (again, assuming equal price elasticities of supply and demand,) the result will be that the shortages for the less interest elastic goods become larger relative to those of the more interest elastic goods. The "recession" phase then should occur in the context of shortages of both capital goods and consumer goods and services, with the production of consumer goods and services expanding through this process of markets with greater shortages pulling resources away from markets with smaller shortages.
Clearly, it would be possible that this process could leave the markets for highly interest elastic goods in surplus. But it isn't necessary that this would occur. And if it did occur, the prices of capital goods, for example, have "overshot," their long run level. However, even if that is true, there remain shortages of consumer goods and services.
Of course, the issue of "overshooting" brings the question of expectations to the fore. The rational expectations equilibrium is for prices to be set to clear these markets. Which would mean that the prices of both consumer goods and capital goods rise in proportion immediately. The real quantity of money remains unchanged at the amount people want to hold. The real supply of credit remains unchanged--the expanded nominal supply of credit purchases the same amount of consumer and capital goods. The market interest rate remains unchanged at the natural interest rate. The production of consumer and capital goods are unaffected.
Returning to Krugman's critique and Kling's response, it seems that this entire disequilibrium process could from beginning to end be associated with somewhat higher structural unemployment. However, in the context of a growing economy, it is difficult to see why having larger shortages in more interest elastic industries and then in less interest elastic industries (again, keeping in mind that price elasticities of demand and supply must be taken into account,) would do little more than causes shifts in which sectors of the economy hire new entrants. It is hard to see how this disequilibrium process could require that highly interest elastic industries absolutely shrink in size, and lay off workers.
Now, if the "tight" full employment assumption is relaxed, so that more capital and consumer goods can be produced when the demand for both rise, then aggregate output rises, and it plausibly rises more in industries with more interest elastic demand As the economy "contracts," then the contraction would be concentrated in those same industries.
Further, if the source of the monetary disequilibrium is not a single increase in the quantity of money, but rather growing quantity of money, the expansion in output could be more presistent. The growing nominal quantity of money generates a growing exess supply of money and a growing real supply of credit. This should be associated with a falling market interest rate and growing shortages of goods with high interest elasticities of demand.
Of course, the rational expectations equilibrium is for the prices of consumer goods and capital goods to immediately begin rising in proportion. The real quantity of money remains the same. The real supply of credit remains the same, even while the nominal supply of credit grows, though able to only purchase unchanged amounts of consumer and capital goods. The market interest rate is unaffected as is the production of consumer and capital goods.
My judgement has always been that this disequilibrium process, especially if there is a "tight" assumption of full employment, is an implausible explanation of recession. Perhaps an excess supply of money, even caused by an acceleration of money growth, might cause a boom that is somewhat imbalanced, but the return to equilibrium will simply be somewhat slower growth in output, especially in those highly interest elastic sectors. A disequilibrium expansion, particularly in some sectors, and then a return to monetary equilibrium is possible, it is just difficult to see how it would be significant.
In my judgement, a much more plausible avenue for a recession would be the opposite sort of monetary disequilibrium, an excess demand for money. If some minimal disruption of production associated with a return to monetary equilibrium triggers a decrease in the quantity of money or an increase in the demand to hold money, then falling cash expenditures on both consumer goods and services and capital goods could result in falling production and employment throughout the economy. As Krugman suggested, the problem would be falling aggregate demand.
Are there other market processes associated with money creation that could result in more significant malinvestment? My next foray into the Austrian Business Cycle Theory will explore an "equilibrium" process that impacts the composition of output.