Mises Daily Articles

Home | Mises Library | Why Money Supply Matters

Why Money Supply Matters

Tags The FedMoney and BanksMonetary Theory

11/08/2005Thorsten Polleit

For governments in general and the US government in particular, Ludwig von Mises had a policy recommendation: do not increase the stock of money any further. He made this point in "Monetary Reconstruction" (written in 1952 and published in 1953): "The first step must be a radical and unconditional abandonment of any further inflation. The total amount of dollar bills, whatever their name or legal characteristic may be, must not be increased by further issuance."[1]

Why did Mises take such a position, one which would presumably provoke outright opposition from many of today's mainstream economists?

The answer lies in two interrelated issues. First, to Mises inflation was the source of many evils because inflation would mislead businessmen in their investments (by creating an illusory boom), cause waste of scarce resources and would inevitably lead to economic crises.[2] Ensuing recessions, even depressions, would provoke public opinion to call for (even more) government market intervention, thereby undermining the very ideal of a free society. Second, Mises considered inflation — and therefore economic crises — to be a monetary phenomenon, brought about by governments' increase in the quantity of money: "There is no means of avoiding the final collapse of a boom brought about by [bank] credit [and therefore money] expansion."[3] As a result, it appears all too consequential that Mises — a steadfast defender of the free society — recommended an end to the expansion of the stock of money.

It might be a relief to followers of Mises's economic program that keeping inflation low and at a stable level has become a widely accepted societal objective. However, a major issue of concern remains. Central banks under the government-sponsored paper money system keep expanding the money supply. To make things worse, central banks, in line with many economists, currently appear to pay relatively little, if any, attention to money supply growth: "Most people believe that economics is about money. Yet most economists hold conversations in which the word 'money' appears hardly at all. Surprisingly, that appears true even for central bankers."[4] In fact, most would consider money growth important in the long run but not in the short term. The following four aspects might explain why monetary policy has increasingly become "moneyless."

First, many studies that report a close long-run relation between money and inflation make use of cross-country data, and it is sometimes said that the finding relies heavily on the presence of countries with high money growth and inflation.[5] It would be much less clear that a close relationship exists within countries with relatively small changes in money growth such as in the western developed countries.

Second, central banks make decisions on a rather frequent basis. For instance, the Governing Council of the European Central Bank (ECB) meets every four weeks to decide on interest rates, the US Federal Reserve every six weeks. So even if a close relationship between money growth and inflation exists over the long run, that relationship might not be visible over short time horizons such as a month or a quarter. Consequently, a close relationship between money growth and inflation that exists only over long time horizons might be said to be of little use to policy makers trying to control inflation over the next months or quarters rather than years.

Third, central banking, like any field of policy, might be subject to fashions and even theory fads. For instance, in theory it has become a state-of-the-art concept to model monetary policy without money.[6] As a result, monetary policy has increasingly started "looking at everything," such as changes in output, oil prices and employment, because responding to these variables is widely seen to be more appropriate in keeping inflation in check than is reacting to money supply signals.

And fourth, many central banks reportedly found demand-for-money functions to be unstable; that is, a reliable stable relation between the stock of money, output, interest rates, and prices was no longer said to exist. Without a stable demand function, money would no longer be seen as a proper yardstick for guiding monetary policy decisions. Bank of Canada's former Governor Gerry Bouey is frequently quoted as saying that central banks didn't abandon monetary aggregates, monetary aggregates abandoned central banks.[7]

Against this backdrop, the question is: how long must money grow ("strongly") before it should be of concern to central banks? Or to put it differently, what is the shortest period of time over which money growth seems to be reliably associated with price inflation?[8] To find an answer to these questions, one has to start with a theory. A relation between money growth and inflation can be established through the equation of exchange. It can be written as follows:

(1) M * V = Y * P

Equation 1 states that the stock of money (M), multiplied by the number of times a money unit is used for financing purposes (V), equals real output (Y) multiplied with the price level (P). For an easier handling, equation 1 can be restated in growth rates:

(2) Δ m + Δ v = Δ y + Δ p,

where lowercase letters stand for logarithms and Δ for growth rates (that is, differences of the variables under review).[9] Assuming that (i) the change in velocity is zero and (ii) output growth tends to oscillate around a stable level (that is the economy's "trend growth"), money supply growth can be expected to show a relation to inflation:

(3) Δ p ˜ Δ m.[10]

Of course one should not expect that changes in money supply have an immediate effect on inflation. Taking into account adjustment processes, it takes some time for the change in money supply to make itself felt in the economy. So in the following the relationship between money growth and consumer price inflation (central banks' actual "target" variable) shall be examined across three time periods: two, four, and six years. To examine the question of whether the relationship between money growth and inflation is notably close over any of these time horizons, and, if it is, how clearly that relationship holds up over shorter time horizons.

For the period 1971 to 2005 in the euro area, money growth (measured as annual change in the stock of M3) and consumer price inflation exhibit a rather obvious relation, that is higher money supply growth is accompanied by higher inflation and vice versa. In view of simple correlation coefficients, which measure the degree of co-movement between two time series, the relation seems to be most pronounced when using gliding 6-year averages of growth rates.

Source: ECB, Thomson Financial, Bloomberg. Period: January 1971 to July 2005, monthly data; own calculations. — The simple correlation coefficient for contemporaneous relation is .78, for 2-year averages 0.83, 4-year averages .90 and 6-year averages .93.

In Japan, a quite similar relation between money growth and inflation can be observed. As in the euro area, the relationship between money and inflation appears to be positive. That is, a rise (decline) in money growth is accompanied with an increase (fall) in inflation. Also, the relation increases with the length of the averaging period. The strongest co-movement is for a 6-year average.

Source: ECB, Thomson Financial, Bloomberg. — Period: January 1971 to July 2005; own calculations. — The simple correlation coefficient for contemporaneous relation is .59, for 2-year averages .72, 4-year averages .87 and 6-year averages .90.

Finally, the relation between US money growth (in the form of the stock of M2) and consumer price inflation is also positive, albeit generally lower than in the euro area and Japan. Again, the correlation coefficient rises with the length of the averaging period. However, the relation between money growth and price inflation seems to be somewhat lower than in the euro area and Japan, and has become somewhat weaker since the middle of the 1990s.

Source: ECB, Thomson Financial, Bloomberg. — Period: January 1971 to July 2005; own calculations. — The simple correlation coefficient for contemporaneous relation is .20, for 2-year averages .36, 4-year averages .60 and 6-year averages .71.

The results of these admittedly rather simple analyses suggest that a relatively close relationship between money growth and consumer price inflation seems to exist at least over long time horizons in all currency areas. This finding could serve as a reminder that ignoring money growth for too long a period may be unwise when central banks aim at keeping consumer price inflation in check.

However, would these findings really justify central banks' practice of considering money supply changes important only in the long run but not in the short term? An answer to that question might be found in the observation that the relation between money supply growth and consumer price inflation appears to have become somewhat weaker since the middle of the 1990s. Could it be that consumer prices no longer represent a proper measure of money changes' inflationary impact?

Hypothetically, a weakening relation between money growth and consumer price inflation might be explained by the possibility that since around the second half of the 1990s ("excessive") money growth might have increasingly been affecting asset prices — including bonds, stocks, real estate and housing — rather than final production prices. That said, the potential existence of "asset price inflation" might have weakened the relation between money and consumer price inflation.

The graph on the left hand side shows nominal wealth growth (measured as the annual change in GDP plus the stock market capitalization) and the expansion of the stock of "M3 adjusted" (measured as the stock of US M3 minus small time deposits) for the period 1980-Q1 to 2005-Q2. Making allowance for some "noise" in the data, both series are clearly positively correlated. Moreover, the underlying growth trend of M3 adjusted seems to be associated with the underlying trend of nominal wealth growth.

Turning to the euro area, a rather similar picture emerges. An increase (decline) in the growth of the stock of money euro M1 is accompanied by a change in nominal wealth in the same direction throughout the period under review. Interestingly, money expansion seems to lead changes in nominal wealth since around the end of the 1990s; that is, a change in nominal wealth seems to have followed changes in money supply.

The insights above could suggest that the consequences of money growth — and in particular "excessive" money growth — are not confined to final product prices (measured by consumer price indices or output deflators), but also increasingly affect asset prices, which in turn are usually not included in consumer price indices. If that is so, two interrelated conclusions could be drawn.

 A history of calamity: $25

First, it might no longer be appropriate for central banks to focus their efforts on "targeting" consumer prices when the overall objective is to preserve the value of the currency. If monetary policy really wants to stick to common practice (that is, following an "index standard" as proposed by Irving Fisher), what would be needed is a much broader measure of the economy's price level, including the prices of current production as well as asset prices.

Second, central banks might have to take money supply signals — including the more "short-term" ones — much more seriously when setting interest rates when the overall objective is preserving the value of the currency. In fact, money supply signals might actually be far more important for inflation — even in the short-term — than current central bank practice suggests.

Let us assume the above conclusions were put into practice. Monetary policy would expand money supply to keep a broadly defined price index stable. Would that eliminate Mises's concern? Unfortunately the answer would not be affirmative, most importantly because according to Mises the expansion of the money stock would inevitably distort the structure of relative prices — even if a (broadly defined) price index would not change — and consequently lead to misallocations and, sooner or later, precipitate economic crisis. And it is the very consequence of such a crisis that Mises identified as a major problem:

"But still more disastrous are [the crisis's] moral ravages. It makes people despondent and dispirited. […] The individual is always ready to ascribe his good luck to his own efficiency and to take it as a well-deserved reward for his talent, application, and probity. But reverses of fortune he always charges to other people, and most of all to the absurdity of social and political institutions. He does not blame the authorities for having fostered the boom. He reviles them for the inevitable collapse. In the opinion of the public, more inflation and more credit expansion are the only remedy against the evils which inflation and credit expansion have brought about."[11]

That said, to Mises even a monetary policy that would pursue a pre-determined rate of money supply expansion (as proposed for example by Milton Friedman's k-percent rule) for stabilizing a broadly defined price index would remain a potential source of crisis which, in turn, bears the risk of undermining the value of the currency. This explains why Mises, in an effort to reduce that very risk to the ideal of a free society, argued for stopping the expansion of the money supply, thereby arguing for a concept quite different from today's state-of-the-art monetary policy.

Thorsten Polleit is Honorary Professor at HfB — Business School of Finance & Management, Frankfurt, e-mail: thorsten.polleit@barcap.com.

[1] Mises, L. von (1981), Monetary Reconstruction (1953, 1952), in: The Theory of Money and Credit, Chapter 23, Liberty Fund, Indianapolis, p. 491.

[2] Mises, L. von (1996), Human Action, 4th Ed., Fox & Wilkes, San Francisco, pp. 426.

[3] Ibid, p. 572.

[4] King, M. (2002), No money, no inflation — the role of money in the economy, in: Bank of England Quarterly Bulletin, Summer, p. 173. At a conference in July 1992, John B. Taylor said: (…) "interest rates are likely to remain the preferred operating instrument of monetary policy," but he also advised: "The evidence that the large swings in inflation are related to money growth indicates, however, that money should continue to play an important role in monetary policy formulation in the future." Taylor, J. B. (1992), The Great Inflation, the Great Disinflation, and Policies for Future Price Stability, in: A. Blundell-Wignall (ed.), Inflation, Disinflation, and Monetary Policy, Ambassador Press, pp. 9 — 31.

[5] See, for instance, Herwartz, H., Reimers, H.-E. (2001), Long-run links among money, prices and output: world-wide evidence, Discussion paper 14/1, Deutsche Bundesbank, September. Also McCandless, G. T., Weber, W. E. (1995), Some Monetary Facts, Federal Reserve Bank of Minneapolis, Quarterly Review, Summer, 19, 2-11; Romer, D. (1996), Advanced Macroeconomics, McGraw-Hill, New York; Lucas, R. E. (1996), Nobel Lecture: Monetary Neutrality, Journal of Political Economy, Vol. 104, pp. 661-682.

[6] Today's standard model of monetary policy has aggregate demand responding directly to an interest rate under the central bank's control, and ignores the role played by the quantity of money in the transmission mechanism. For a critical review see, for instance, McCallum, B. T. (2001), Monetary Policy Analysis in Models Without Money, in: Federal Reserve Bank of St Louis, July/August, pp. 145 — 160.

[7] These comment referred to the monetary aggregate M1 in Canada. Canada, House of Commons (1983), Standing Committee on Finance, Trade and Economic Affairs, Minutes of Proceedings and Evidence, No. 134, 28 March, p. 12.

[8] In the following, I draw heavily on the work of Fitzgerald, T. J. (1999), Money Growth and Inflation: How Long is the Long-Run?, in: Federal Reserve Bank of Cleveland, 1 August.

[9] The calculus-minded reader knows that d(lnX)dX = 1/X or d(lnX) = dX/X, that is for infinitesimal small changes a change in lnX is equal to the relative or proportional change in X.

[10] Recall that the simple theory of inflation actually holds that money growth in excess of real output growth should be more closely connected with inflation.

[11] Mises, L. von (1996), pp. 576.



Thorsten Polleit

Dr. Thorsten Polleit is Chief Economist of Degussa and Honorary Professor at the University of Bayreuth. He also acts as an investment advisor.