MS: From Inflation Targeting to Price Level Targeting?

Morgan Stanley - Global Economic Forum

Global
From Inflation Targeting to Price Level Targeting?
July 16, 2009

By Joachim Fels & Spyros Andreopoulos | London

Tougher times for inflation targeters: Over the past two decades, inflation targeting has become the holy grail of modern central banking. Most central banks have adopted some form of inflation target and have set interest rates mainly with a view to stabilising inflation around that target over the medium term. However, the recent experience poses a major challenge for inflation-targeting central banks, for two reasons:

• First, the wild gyrations of commodity prices and the boom-bust credit and economic cycle have caused big swings in inflation. Only a year ago, actual inflation was way above most central banks' targets; now it is far below in most cases. This has contributed to rising uncertainty about the longer-term inflation outlook, as illustrated by more volatile market-implied inflation expectations and by a much greater dispersion of economists' inflation forecasts.

• Second, the asset price bubble and bust has been a useful reminder that stabilising consumer price inflation does not automatically stabilise asset prices. On the contrary, as we have argued repeatedly over the years, by focusing too narrowly on consumer prices, which were kept low for a long time by non-monetary factors such as globalisation, deregulation and IT-led productivity increases, central banks fostered asset price inflation by keeping interest rates too low for too long. Looking ahead, many central banks are thus likely to pay more attention to asset prices in setting monetary policy. This, in turn, may lead to bigger and longer-lasting deviations of inflation from target and thus constitutes a challenge to the credibility of their inflation targets (see "The Morning After", The Global Monetary Analyst, April 1, 2009).

Move to price level targeting would make sense to us: Despite its shortcomings, it would be risky to abandon inflation targeting altogether. A better solution, in our view, would be to modify the existing approach and move from inflation targeting (IT) to price level targeting (PT). Under price level targeting, the central bank aims at a certain path for the price level, with the rate of increase in the price level given by the inflation target. So what's the difference, and why does it matter?

Introducing ‘memory' into inflation targeting: In the case of IT, there are no consequences if the central bank misses its inflation target in one period. Past errors will not be corrected but, at each point in time, the central bank will set the policy rate such that the inflation target is achieved over the next period. Thus, inflation outcomes above or below the average would simply be followed by average inflation. As a result, the price level drifts away from the level implied by the inflation target over time.

With PT, on the other hand, past deviations from the price level target would have to be compensated for in the future in order to return to achieve the target in the medium term. Above-average inflation in one period would have to be followed by below-average inflation in the following period(s).

So PT is like pursuing an inflation target over a longer time horizon, which is why PT is sometimes called ‘average inflation targeting'. And that is how PT could be explained to the public. The central bank's aim would now be to hit the inflation target ‘on average' over a series of years. Overshoots would have to followed by undershoots and vice versa.

PT would help to stabilise long-term inflation expectations: The major advantage of PT is that, if credible, long-term inflation expectations are actually more stable than under IT. During the Gold Standard, which implicitly was a price level-targeting regime, the long-run price level was given by the quantity of gold in the international monetary system. Periods of inflation were followed by periods of deflation because there was a built-in automatic stabiliser, and the price level was stable over the long term. And PT would, if pursued consistently, enhance central bank credibility: monetary policymakers would now be more immediately accountable for past errors - they would have to respond to them directly. Greater accountability would, in turn, increase credibility.

Academic? We think that at the current juncture, this advantage of PT would be particularly useful for central banks, for three reasons:

• Central banks may in future want to - at least occasionally - lean against the wind with respect to asset prices, following the experience with the credit and house price boom-bust. This would probably imply larger and/or longer deviations from the inflation target (see our piece from April 1, 2009, cited above). PT would allow monetary authorities greater flexibility in occasionally responding to asset prices without unanchoring inflation expectations, because they would have committed themselves to correcting deviations from the target at a later stage.

• PT is more helpful at the zero lower bound for interest rates. Under IT, when actual inflation falls below target or even becomes negative, inflation expectations would also fall. But lower inflation expectations would increase real interest rates, while the economy may need stimulus through lower real interest rates. With PT, on the other hand, near-to medium-term inflation expectations would automatically increase, since the public would expect the undershoot in inflation now to be compensated by a later overshoot. Hence inflation expectations act as an automatic stabiliser under PT. (See, for example, Riksbank Deputy Governor Svensson's speech: Monetary Policy with a Zero Interest Rate.)

• PT would also be helpful over the next few years, if inflation were to rise above target but central banks were reluctant to raise interest rates because of fragility in the financial sector or the real economy. With PT, an overshoot of inflation over the target would, rather than being an embarrassment for central banks, be a desired (indeed necessary) correction for the undershoots of the recent past. We looked at the path for the US price level (measured by headline CPI and PCE, respectively) starting from September 2008, the peak of the financial crisis. In the short and medium term (using our US team's forecasts out to 4Q10), the price level is below the level implied by a 2% PCE inflation target. In order to catch up from the 4Q10 level to the level implied by the inflation target by 4Q12, a 3.6% average annual inflation rate between 4Q10 and 4Q12 would be required. Thus, any overshoot would merely be a return to the medium term target level.

Having one's cake and eating IT: To summarise, adoption of PT would potentially allow central banks to deviate from their inflation targets without putting their credibility at stake: they could have their cake and eat it.

Not a popular concept...yet: With so many advantages of PT over IT, why hasn't it been adopted widely already? In fact, there are very few examples of central banks adopting PT or something resembling it. The only country that adopted an explicit price level target was Sweden from 1931 to 1937 (see C. Berg, L. Jonung, "Pioneering Price Level Targeting: The Swedish Experience 1931 - 1937", Journal of Monetary Economics, Vol 43, 1999). Today, the Reserve Bank of Australia probably comes closest to a (somewhat vague) price level target by stipulating that it aims to keep inflation at 2-3% "on average over the cycle". Probably the main reason why PT hasn't been adopted more widely is that a central bank that is targeting a path for the price level would spend approximately half of its time trying to deflate or disinflate the economy. It is easier to let bygones be bygones and start from scratch each period. But here's the catch: with inflation now having fallen significantly below target in many countries, by adopting a price level target with a starting point in, say, the middle of last year, central banks could spend the next few year inflating their economies and sell this as being part and parcel of a credible price level-targeting strategy. Given the fragile state of the global economy and the financial system, this prospect may look quite appealing to some. Stay tuned.

Powered by ScribeFire.