Assume that inflation of 2%, on average, is ideal. This implies that if the price level is 100 this year, it will be 102 next year and 104 (or more precisely, 104.04) in the second year. If inflation is only 1% in one year, a conventional inflation-targeting central bank would aim only to return inflation to a rate of 2% the next. This would leave the price level at 103, lower than its original implied path. In contrast, a central bank that targets the price level wants to make up any lost ground on prices. It would seek to raise inflation to 3% in the second year to get to a target of 104.
In theory price-level targeting is superior to inflation-targeting because it provides more certainty about the long-term purchasing power of money. Central banks always target inflation flexibly. The Bank of Canada and the Bank of England, for example, target a rate of 2% but permit a range of 1% to 3%. That means someone making a 30-year investment must plan for cumulative inflation of as much as 143% or as little as 35%. A credible price-level target eliminates that uncertainty.
So a price level target trumps an inflation target, but it has one glaring problem: it doesn't handle aggregate supply shocks very well. Here is The Economist:
There are questions, too, about how central bankers would deal with a one-time rise in the price level because of a new value-added tax, say, or higher oil prices. The boost to inflation would be temporary, but to the price level, permanent. In theory a central bank would have to wrestle all other prices lower no matter what the cost. It could make an exception, but too many exceptions would dent the bank’s credibility. Conversely, a positive shock such as lower oil prices or higher productivity that pushes prices lower would require the central bank to raise future inflation, driving down real interest rates and maybe risking an asset bubble.
Economists Scott Sumner of Bentley University and David Beckworth of Texas State University are among those who have suggested that the Fed should move gradually toward a new, more rule-bound and predictable monetary policy. The first step would be to signal to the markets that the Fed is willing to do whatever it takes to reach 2 percent average inflation. Over time the Fed would move to stabilize and then slow the growth of nominal GDP, which is the size of the economy as measured in a given year’s dollars. If the nominal GDP target was for 3 percent growth and the economy grew by 2 percent, there would be 1 percent inflation.
That policy would bind the Fed to a rule, thus reducing the uncertainty that recent policy has generated, including the risk that we will get galloping inflation at some point in the future. But it is superior to simply targeting the inflation rate, Beckworth argues, because it incorporates two worthwhile types of flexibility. It allows the price level to move in response to supply shocks: An oil embargo would cause prices to rise, a technological advance would have the opposite effect. And it allows the money supply to move up and down in response to the demand for cash: In periods such as late 2008, when people were holding on to their money, the Fed would have loosened more than it did. But since the rule would have required tighter money during the boom years, the financial crisis might not have been as severe in the first place.