FIGURE 9.1 The aggregate demand and aggregate supply curves

for identifying potential sources of inflation. Since our interest is in prices rather than output, the issue of whether the aggregate supply curve is vertical or merely upward-sloping is not important: in either case, both expansions of aggregate demand and contractions of aggregate supply raise the price level. Thus there arc many potential sources of inflation. Negative technology shocks, downward shifts in labor supply, upwardly skewed relative-cost shocks, and other factors that shift the aggregate supply curve to the left cause inflation; the same is true of increases in the money stock, downward shifts in money demand, increases in government purchases, and other factors that shift the aggregate demand curve to the right. Since all of these types of shocks occur to some extent, there are many factors that affect inflation.

Nonetheless, when it comes to understanding inflation over the longer term, economists typically emphasize just one factor: growth of the money supply. The reason for this emphasis is that no other factor is likely to lead to persistent increases in the price leveL Repeated increases in prices require either repeated faUs in aggregate supply or repeated rises in aggregate demand. Given technological progress, repeated falls in aggregate

Many shocks affect both curves. A rise in government purchases, for example, may not only shift the aggregate demand curve, but also move the aggregate supply curve through its impact on labor supply. The overall effect of any shock on the price level depends on how it affects both curves.

supply are unlikely. And although there are many factors that can increase aggregate demand, most of them are limited in scope. For example, there cannot be repeated large increases in aggregate demand commg from increases in govemment purchases or reductions in taxes, because there are practical limits on these variables; for instance, we never observe govemment purchases that are larger than total output, or total taxes that are negative. The money supply, in contrast, can grow at almost any rate, and we observe huge variations in money growth-from large and negative during some deflations to immense and positive during hyperinflations.

To see more clearly why money is crucial to inflation, consider the money market. With the specification of money demand from Chapter 5, the condition for equilibrium in the money market is

J = , Y), (9.1)

where M is the money stock, P the price level, z the nominal interest rate, Y real income, and !(•) the demand for real money balances. This condition implies that the price level is given by

Conventional estimates of money demand suggest that the income elasticity of money demand is about 1 and the interest elasticity is about -0.2 (see Goldfeld and Sichel, 1990, for example). Thus for the price level to double over some period of time without a change in the money supply, income must fall roughly in half or the interest rate must rise by a factor of about 32. Altermitively, the demand for real balances at a given interest rale and income must fall in half. All of these possibilities are essentially unheard of. In contrast, a doubling of the money supply, either over several years in a moderate inflation or over a few days al the height of a hyperinflation, is not uncommon.

Thus money growth plays a special role in determining inflation not because money affects prices more directly than other factors do, but because empirically variations in money growth account for most of the variation in the growth of aggregate demand. Figure 9.2 provides powerful confirmation of the importance of money growth lo inflation. The figure plots average inflation against average money growth in the 1980s for a sample of 65 countries; there is a clear and strong relationship between the two variables.

Money Growth and Interest Rates

Since money growth is the main determinant of inflation, it is natural to examine its effects in more detail. As we will see, there are interesting links between the growth of the nominal money stock and the behavior of inflation, real and nominal interest rates, and real balances.

lOOOr

& <u

10

FIGURE 9.2 Statistics)

Mexico

Frante

Switzerland

Germany - + +

Japan

++ UK china

Canada

Argentina +

Peru+" Bolivia

+ Israel

1 10 100

Money supply growth (percent)

Money growth and inflation (data from International Financial

We begin with the case where prices are completely flexible; this is presumably a good description of the long run. As we know from our analysis of fluctuations, this assumption implies that the money supply does not affect real output or the real interest rate. For simplicity, we assume that these are constant at Y and F, respectively.

By definition, the real interest rate is the difference between the nominal interest rate and expected inflation. That is, r = - -*, or

I = r -t

(9.3)

Equation (9.3) is known as the Fisher identity.

Using (9.3) and our assumption that r and F are constant, we can rewrite (9.2) as

I(r + , F)

(9.4)

Assume that initially M and P are growing together at some steady rate (so that M IP is constant), and that equals actual inflation. Now suppose that at some time, time to, there is a permanent increase in money growth. The resulting path of the money stock is shown in the top panel of Figure 9.3. After the change, since M is growing at a new steady rate and r and F are constant by assumption, M/P is constant; that is, (9.4) is satisfied with P growing at the same rate as M and with *" equal to the new rate of monev growth.

But what happens at the time of the change? Since the price level rises faster after the change than before, expected inflation jumps up when the change occurs. Thus the nominal interest rate jumps up, and so the quantitv