Deflation is not worse than inflation

Just over nine years ago, Ben Bernanke gave a famous speech entitled “Deflation: Making Sure ‘It’ Doesn’t Happen Here.” He articulated a view that has since become completely mainstream, namely, that falling prices are far worse than rising prices. Why should this be?

Orthodox macro economists fear deflation (negative changes in the consumer price index) more than inflation (positive changes in the consumer price index) for three reasons. Professor Krugman provided a good summary of the standard argument.

First, falling prices supposedly create expectations of further price declines. This supposedly encourages households and businesses to postpone spending and hoard cash, which could induce a recession. Second, falling prices often—although not always—go hand-in-hand with declines in nominal wages, which means that deflation increases the burden of existing debts. Finally, it is supposedly harder to stop prices from falling than to stop them from rising because the central bank could push nominal interest rates up towards infinity but cannot push them below 0%—hence the so-called “zero bound.”

It turns out, however, that all three of these arguments are either flawed or incomplete. Broad declines in nominal wages and prices, while generally undesirable, are no worse than broad increases in nominal wages and prices.

The first argument is the silliest. Professor Krugman:

When people expect falling prices, they become less willing to spend…After all, when prices are falling, just sitting on cash becomes an investment with a positive real yield.

Where is the evidence that changes in nominal prices affect real activity? Television prices, as measured by the Bureau of Labor Statistics, have fallen at an average annualized rate of 6% every year since 1967:

Yet this has not prevented Americans from buying far more televisions in 2011 than in 1967. According to the Consumer Expenditure Survey at the BLS, U.S. purchases of televisions have increased by a factor of 3 just since the mid-1980s.

Perhaps nominal price declines do not affect the consumption of consumer electronics because of Moore’s Law and cheap Asian manufacturing. What about a service that can only be consumed locally, is immune to technological improvements, and is completely unnecessary to survival? This seems like exactly the sort of situation where deflation would cause reductions in real spending.

Fortunately, there is no need to speculate. A real-life example: at my office, employees can buy two types of discounted movie tickets from HR. The first costs $6.50 and is good for admission to any new film screened in a big theater chain. The second costs $5.00 (a discount of 23%) but can only be used to see movies that have already been out for at least two weekends. Thus, the effective annualized rate of inflation in movie tickets is about -99.9%.

If it were true that nominal prices affected real spending choices, no one would ever pay to see a new release. Yet this is not what happens—many people willingly pay a hefty premium to see a movie on the weekend it hits the theaters. The conclusion is clear: deflation does not reduce real spending, even for purely discretionary items. It beggars the imagination to see how price declines could seriously affect real purchases of necessities like food, clothing, haircuts, healthcare, or shelter (the choice to rent or buy is more complicated because of leverage but everyone needs a place to sleep).

Another way to test this concept is to look at the changes in real GDP per capita against changes in nominal GDP per capita. If real GDP per capita rises when nominal GDP per capita is flat or falling then deflation clearly is not an impediment, in and of itself, to higher standards of living. Turns out history supports this claim:

Incidentally, real GDP per person grew at almost exactly the same pace from 1865-1900 as it did from 1950-present, even though nominal GDP per person grew at an average annualized rate of 5.4%. Clearly there is no stable relationship between the real and nominal variables.

The second reason mainstream economists fear deflation more than inflation is that it creates a vicious cycle of rising real debt burdens. The case was made by Irving Fisher in the early 1930s. Bernanke explained it well:

Although deflation and the zero bound on nominal interest rates create a significant problem for those seeking to borrow, they impose an even greater burden on households and firms that had accumulated substantial debt before the onset of the deflation. This burden arises because, even if debtors are able to refinance their existing obligations at low nominal interest rates, with prices falling they must still repay the principal in dollars of increasing (perhaps rapidly increasing) real value…The financial distress of debtors can, in turn, increase the fragility of the nation’s financial system—for example, by leading to a rapid increase in the share of bank loans that are delinquent or in default.

It is true that a reduction in nominal income increases the real burden posed by a nominally-denominated debt—it is easier to pay a $500,000 mortgage at a given interest rate if you make $150,000 a year rather than $100,000. However, if this nominal pay-cut were anticipated in the initial interest rate charged on the debt, deflation should not affect the debtor’s ability to pay. The change in wages would be offset by a lower interest rate. Likewise, inflation only helps debtors when their nominal incomes rise by more than whatever was priced into their interest rates.

Bernanke’s concern that “even if debtors are able to refinance their existing obligations at low nominal interest rates, with prices falling they must still repay the principal in dollars of increasing (perhaps rapidly increasing) real value” is thus misplaced. Most of the time, debtors can assume that the real value of the principal will fall over time. Creditors also assume this and charge an inflation premium on top of whatever real interest rate they deem appropriate. Debtors are in basically the same position in either environment if the price changes are accurately anticipated.

The issue is not, therefore, inflation vs. deflation but rather large, unexpected changes in the rate of inflation. It is the surprise that is painful, not the actual change in income relative to the debt. This does not mean that surprise deflation is always a bad thing. To the extent that unanticipated changes in the rate of nominal income growth affect real interest costs, surprise deflation can be a helpful way to discourage excessive borrowing when the economy is running hot, just as surprise inflation can facilitate deleveraging. Deflation, therefore, is not worse than inflation but merely a different kind of distortion.

The final argument of the deflation-phobes is that price declines are much more difficult to stop or reverse than price increases. Bernanke provided a good summary in his 2002 speech:

Beyond its adverse effects in financial markets and on borrowers, the zero bound on the nominal interest rate raises another concern—the limitation that it places on conventional monetary policy. Under normal conditions, the Fed and most other central banks implement policy by setting a target for a short-term interest rate—the overnight federal funds rate in the United States—and enforcing that target by buying and selling securities in open capital markets. When the short-term interest rate hits zero, the central bank can no longer ease policy by lowering its usual interest-rate target. Because central banks conventionally conduct monetary policy by manipulating the short-term nominal interest rate, some observers have concluded that when that key rate stands at or near zero, the central bank has “run out of ammunition”—that is, it no longer has the power to expand aggregate demand and hence economic activity.

The entire point of Bernanke’s speech, however, was that this conventional argument was incorrect. He correctly identified the mechanism that could create inflation even when nominal interest rates had hit the floor:

A broad-based tax cut, for example, accommodated by a program of open-market purchases to alleviate any tendency for interest rates to increase, would almost certainly be an effective stimulant to consumption and hence to prices. Even if households decided not to increase consumption but instead re-balanced their portfolios by using their extra cash to acquire real and financial assets, the resulting increase in asset values would lower the cost of capital and improve the balance sheet positions of potential borrowers. A money-financed tax cut is essentially equivalent to Milton Friedman’s famous “helicopter drop” of money.

Of course, in lieu of tax cuts or increases in transfers the government could increase spending on current goods and services or even acquire existing real or financial assets. If the Treasury issued debt to purchase private assets and the Fed then purchased an equal amount of Treasury debt with newly created money, the whole operation would be the economic equivalent of direct open-market operations in private assets.

Note that this is a bit different from what the Fed has actually done since the start of the crisis. If Bernanke practiced what he once preached he would have demanded tax cuts and spending increases to reverse the economy’s decline.

We can see how this worked in the past by looking at the surge in government deficit-spending during WWI and WWII:

And noticing the correlation between those monetized deficits and sharp surges in the price level:

The other thing we can learn from this chart is that price declines do not self-reinforce into a downward spiral of deflationary doom, as some darkly implied. Even the severe price declines associated with the Great Depression came to a halt after a few years. To his credit, Professor Krugman appears to have revised his earlier views:

The inflation-adjusted Phillips curve predicts not just deflation, but accelerating deflation in the face of a really prolonged economic slump. Suppose that the economy is sufficiently depressed that with expected inflation at 3 percent, actual inflation comes out only 1; expectations will actuallyeventually catch up, so that if the economy remains depressed we’d expect inflation to go to -1; but if the economy remains depressed even longer, we’d expect inflation to go to -3, then -5, and so on.

In reality, this doesn’t happen. Prices fell sharply at the beginning of the Great Depression, when the real economy was collapsing; but they began rising again when the economy began to recover, even though there was still a huge negative output gap. Japan has been depressed since before incoming freshmen were born, but its chronic deflation has never turned into a rapid downward spiral.

Having read this far, you may be wondering why these arguments matter. The reason is that there are times when it is desirable to have prices fall (like 1995-2004) just as there are times when it is desirable to have faster increases in nominal wages than normal (like now). Central bankers ignore this at our peril.


About Matthew C. Klein
I write about the economy and financial markets for Bloomberg View. Before that I wrote for The Economist on a fellowship provided by the Marjorie Deane Financial Journalism Foundation. I have worked at the world's largest hedge fund and read every FOMC transcript since May, 1987.

One Response to Deflation is not worse than inflation

  1. Pingback: I still hate Bush... - Page 4

%d bloggers like this: