Saturday, March 22, 2008

2008 = 1929?

It has become fashionable lately to see parallels between the current financial turmoil and what happened during the Great Depression. See, for example, Paul Krugman's latest column.

Let me remind everyone of one important difference: Deflation was a large part of the story of the 1930s, and that does not seem like a significant risk today. Here is a relevant passage from my favorite intermediate macro textbook (sorry, I cannot post the referenced figure):

The Money Hypothesis Again: The Effects of Falling Prices

From 1929 to 1933 the price level fell 25 percent. Many economists blame this deflation for the severity of the Great Depression. They argue that the deflation may have turned what in 1931 was a typical economic downturn into an unprecedented period of high unemployment and depressed income. If correct, this argument gives new life to the money hypothesis. Because the falling money supply was, plausibly, responsible for the falling price level, it could have been responsible for the severity of the Depression. To evaluate this argument, we must discuss how changes in the price level affect income in the IS‑LM model.

The Stabilizing Effects of Deflation

In the IS‑LM model we have developed so far, falling prices raise income. For any given supply of money M, a lower price level implies higher real money balances M/P. An increase in real money balances causes an expansionary shift in the LM curve, which leads to higher income.

Another channel through which falling prices expand income is called the Pigou effect. Arthur Pigou, a prominent classical economist in the 1930s, pointed out that real money balances are part of households' wealth. As prices fall and real money balances rise, consumers should feel wealthier and spend more. This increase in consumer spending should cause an expansionary shift in the IS curve, also leading to higher income.

These two reasons led some economists in the 1930s to believe that falling prices would help stabilize the economy. That is, they thought that a decline in the price level would automatically push the economy back toward full employment. Yet other economists were less confident in the economy's ability to correct itself. They pointed to other effects of falling prices, to which we now turn.

The Destabilizing Effects of Deflation

Economists have proposed two theories to explain how falling prices could depress income rather than raise it. The first, called the debt‑deflation theory, describes the effects of unexpected falls in the price level. The second explains the effects of expected deflation.

The debt‑deflation theory begins with an observation from Chapter 4: unanticipated changes in the price level redistribute wealth between debtors and creditors. If a debtor owes a creditor $1,000, then the real amount of this debt is $1,000/P, where P is the price level. A fall in the price level raises the real amount of this debt‑‑the amount of purchasing power the debtor must repay the creditor. Therefore, an unexpected deflation enriches creditors and impoverishes debtors.

The debt‑deflation theory then posits that this redistribution of wealth affects spending on goods and services. In response to the redistribution from debtors to creditors, debtors spend less and creditors spend more. If these two groups have equal spending propensities, there is no aggregate impact. But it seems reasonable to assume that debtors have higher propensities to spend than creditors‑‑perhaps that is why the debtors are in debt in the first place. In this case, debtors reduce their spending by more than creditors raise theirs. The net effect is a reduction in spending, a contractionary shift in the IS curve, and lower national income.

To understand how expected changes in prices can affect income, we need to add a new variable to the IS‑LM model. Our discussion of the model so far has not distinguished between the nominal and real interest rates. Yet we know from previous chapters that investment depends on the real interest rate and that money demand depends on the nominal interest rate. If i is the nominal interest rate and πe is expected inflation, then the ex ante real interest rate is i - πe. We can now write the IS‑LM model as

IS: Y = C(Y - T) + I(i - πe) + G
LM: M/P = L(i, Y)

Expected inflation enters as a variable in the IS curve. Thus, changes in expected inflation shift the IS curve.

Let's use this extended IS‑LM model to examine how changes in expected inflation influence the level of income. We begin by assuming that everyone expects the price level to remain the same. In this case, there is no expected inflation (πe = 0), and these two equations produce the familiar IS‑LM model. Figure 11-8 depicts this initial situation with the LM curve and the IS curve labeled IS1. The intersection of these two curves determines the nominal and real interest rates, which for now are the same.

Now suppose that everyone suddenly expects that the price level will fall in the future, so that πe becomes negative. The real interest rate is now higher at any given nominal interest rate. This increase in the real interest rate depresses planned investment spending, shifting the IS curve from IS1 to IS2. (The vertical distance of the downward shift exactly equals the expected deflation.) Thus, an expected deflation leads to a reduction in national income from Y1 to Y2. The nominal interest rate falls from i1 to i2, while the real interest rate rises from r1 to r2.

Here is the story behind this figure. When firms come to expect deflation, they become reluctant to borrow to buy investment goods because they believe they will have to repay these loans later in more valuable dollars. The fall in investment depresses planned expenditure, which in turn depresses income. The fall in income reduces the demand for money, and this reduces the nominal interest rate that equilibrates the money market. The nominal interest rate falls by less than the expected deflation, so the real interest rate rises.

Note that there is a common thread in these two stories of destabilizing deflation. In both, falling prices depress national income by causing a contractionary shift in the IS curve. Because a deflation of the size observed from 1929 to 1933 is unlikely except in the presence of a major contraction in the money supply, these two explanations give some of the responsibility for the Depression‑‑especially its severity‑‑to the Fed. In other words, if falling prices are destabilizing, then a contraction in the money supply can lead to a fall in income, even without a decrease in real money balances or a rise in nominal interest rates.

Could the Depression Happen Again?

Economists study the Depression both because of its intrinsic interest as a major economic event and to provide guidance to policymakers so that it will not happen again. To state with confidence whether this event could recur, we would need to know why it happened. Because there is not yet agreement on the causes of the Great Depression, it is impossible to rule out with certainty another depression of this magnitude.

Yet most economists believe that the mistakes that led to the Great Depression are unlikely to be repeated. The Fed seems unlikely to allow the money supply to fall by one-fourth. Many economists believe that the deflation of the early 1930s was responsible for the depth and length of the Depression. And it seems likely that such a prolonged deflation was possible only in the presence of a falling money supply.