Asset price inflation

Asset price inflation is a rise in the price of an asset that does not reflect a relative change in the price of that asset. It is not a term that is currently widely used or carefully defined, although one sees it in print at various times (Schwartz, 2002; Piazzesi and Schneider, 2009). To formally define asset price inflation, one must define both inflation and asset, neither of which is easy or unambiguous.

In earlier times (pre-1930s), inflation was defined as an increase in the money supply (Bryan, 1997). At that time, in the definition, it was noted that such increases were often accompanied by increases in prices, but the determining factor of inflation was increases in the money supply. As long as the money supply was the numéraire, and was thought of as a physical asset (primarily gold), that served as a reasonable definition. Inflation was the inverse of the price of gold; that is, a fall in the price of gold relative to prices of other things that people bought (both assets and goods).

As money became thought of as separate from gold, that definition of inflation no longer remained clear-cut, but the convention of defining inflation in terms of an increase in the money supply remained. A problem remained, however, as it was unclear what the money supply was: there were many alternative definitions of money, and there was no compelling reason to use one over the other, and thus there was no unambiguous defini- tion of inflation. At that point, inflation started to be defined in terms of an increase in the price of produced goods, not in terms of an increase in the quantity of money.

Precisely what was meant by "goods" was unclear, and over time a number of conven- tions developed as to what goods would be included. Inflation became thought of as the change in the price level of produced goods. Economists developed formal measures of output and price indices, developing well-specified concepts such as real GDP, GDP deflator, CPI, core CPI, and CPE, among others. People's conceptions of inflation fol- lowed these formal measures, and earlier definitions of inflation relating it to the money supply faded away. That led to the way most people think of inflation today, to wit, as an increase in the price level of goods as measured by an inflation index for produced goods. None of these measured concepts was a perfect indicator of changes in the price level of goods, but theoretically they gave a workable measure of the price of a "real" basket of produced goods over time. Initially, economists distinguished relative price changes over time (what one may call real price level changes) from nominal price level changes in the basket of produced goods. They did this by emphasizing that inflation had to be an ongoing change in prices: a one-time change would not count as inflation. That convention faded away, although distinguishing core inflation (which is more likely to be ongoing) from the full measures of inflation (which include temporal relative price fluctuations) relates to that distinction.

What was left out of these "produced goods price" definitions of inflation was assets. Thus, as inflation became associated with changes in the price level of produced goods, the price of assets slowly moved out of the definition of inflation, and what one may call asset price inflation fell from economists' radar screens. One of the reasons for this was theoretical developments in asset pricing theory, and specifically the development of the efficient market hypothesis, which held that the prices of assets reflected their real value. If asset prices reflected their real value, there could be no asset inflation: changes in asset prices were simply intertemporal relative changes in the prices of assets over time.

A theoretical challenge to these conventions came from Alchian and Klein (1973), who argued that the appropriate concept to measure inflation should include asset prices as well as goods prices. Specifically, they argued that measures of inflation should relate to the current cost of expected lifetime consumption, not just to current consumption. To capture the current cost of expected lifetime consumption, the measure of inflation would have to include asset prices as well as goods prices. In fact, it would give a much greater relative weight to asset prices. Pollack (1989) and Shibuya (1992) developed rough measures of such an index in which asset weights in the index were as much as 97 per cent of the relevant measure. This work has not been followed up, and today inflation is almost thought of as changes in the price of produced goods only. The concept of asset price inflation to a large degree disappeared. (Interestingly, however, deflation is dis- cussed in terms of asset prices. Were that not the case, there would be almost no discus- sion of deflation, as an index of goods prices almost never falls significantly.) Thus, the current reality is that we do not have a meaningful measure and a solid understanding of asset price inflation.

In order to have a meaningful concept of asset price inflation, one must have a concept of "real asset", which means that one cannot hold the efficient market hypoth- esis. At an individual asset level, distinguishing whether a change in an asset price is an intertemporal relative change (as the efficient market hypothesis holds) or a bubble is close to impossible. But at the aggregate level of all assets it may be easier, because one would expect to see fewer intertemporal relative price changes. Thus, in his textbook explanations of asset price inflation, Colander (2013) does not focus on the price of any one asset, but rather on a concept that he calls "real wealth", which is the stock equiva- lent to the flow concept of real output. He contrasts real wealth with nominal wealth as a parallel to the contrast between real output and nominal output. Real wealth is the productive capacity of the economic system, while nominal wealth is the money measure of that productive capacity, and the difference in the change of these measures is asset price inflation. Just as real output is differentiated from nominal output by goods price inflation, so too is real wealth differentiated from nominal wealth by asset price inflation.

In steady-state equilibrium real wealth will grow at the same rate as real GDP, assum- ing no major structural changes, so we can get an idea of the degree of asset price infla- tion relative to goods price inflation by comparing nominal wealth to nominal GDP. If nominal wealth grows more than nominal GDP, then there is asset price inflation. If it grows less, there is asset price deflation, where the reference point is the goods price level. As Colander (2013) shows, since the mid 1990s the prices of assets have risen signifi- cantly more than the prices of goods, suggesting that, on average, asset price inflation has exceeded goods price inflation since that period. The rate of asset price inflation has been uneven, and there have been intermittent periods of asset price deflation that have partially offset the net difference. One reason asset price inflation has exceeded goods price inflation is that government policy has encouraged asset price inflation even as it attempts to hold down goods price inflation.

See also:

Asset-based reserve requirements; Basel Agreements; Bubble; Capital requirements; Credit bubble; Debt crisis; Efficient markets theory; Financial bubble; Financial crisis; Financial instability; Housing bubble; Inflation; Macro-prudential policies.