The Futility of Inflation-Targeting

There has been much talk recently from some commentators of how the Fed risks "going overboard" in its interest rate hike campaign, for example in this Marketwatch story.

If by "going overboard" they mean sending the U.S. economy into a recession, I think they're right. However, if by "going overboard" they mean doing more than necessary to bring down price inflation, they are not right.

But this reflects a problem really inherent in the inflation-targeting dogma that now most central banks have adopted. Even assuming for the sake of the argument that successful stabilization of official consumer price inflation really would stabilize the economy (a assumption which can be challenged for several reasons), there is no reason to believe that central banks can stabilize consumer price inflation. The reason for that is that there is for several reasons a significant time lag between monetary policy and its impact on consumer prices.

This reflects in part a time lag between monetary policy decisions and monetary inflation (a.k.a. money supply increases), in part it reflects the time lag between monetary inflation and price inflation and in part it reflects the distortions of price inflation that the two most popular gauges of inflation, the CPI and the PCE deflator, contain.

The latter factor means primarily the issue of how the cost of housing is measured. When the Fed lowers interest rates this will in the short term lower the actual cost of living in owner occupied housing, something which will all other things being equal reduce demand for rented housing something which in turn will depress rents and therefore by extension will lower the "owner's equivalent rent" component of official inflation indices. In the long run, the interest rate cuts will however not lower housing costs because either interest rates will be raised again or if the cuts are permanent, housing prices will be adjusted upwards. Indeed, because lower interest rates will raise credit demand and because higher demand for credit will increase the money supply, the long run effect of interest rate cuts will be to raise housing costs.

But since the short-term effect of the cuts is to lower the housing component of official inflation indices this will cause central banks to overreact to fight off what central bankers in general and Ben "Helicopter" Bernanke in particular has depicted as the worst evil conceivable – price deflation (a.k.a. falling cost of living). This was of course exactly what happened in 2003, when the short-term effect of the previous interest rate cuts had depressed rents so much that official "core" inflation indices approached 1%, something which made the Fed lower interest rates all the way down to 1%.

Now, we are seeing the opposite situation. In part because of the lagged effect of previous super-easy monetary policy have made housing prices soar to unprecedented levels relative to both income and rents and in part because interest rates have now returned to more normal levels, owner-occupied housing has become increasingly unaffordable. That has increased demand for rented housing something which in turn has pushed up rents and by extension the "owner's equivalent rent" component of official inflation indices.

But this means that central banks will in practice always aggravate – not reduce – swings in actual housing costs, as during downturns we will experience the lagged depressing effect on house prices from previous tight policies and the lower mortgage interest rates from current easy policies while during cyclical peaks (like now) we will face the lagged boosting effect on house prices in combination with rising mortgage interest rates.

Just how great the total time lag between monetary policy decisions and official price inflation numbers (taking into account all three aforementioned time lags) are is unclear and will like most quantitative relations vary somewhat between different countries and different time periods, but most economists believe that it is generally about 1 to 2 years – or more. For that reason, it really makes no sense from an inflation targeting perspective to adjust policy according to the latest monthly data.

However, the problem is that while the latest monthly data may be of little or no relevance, it is really the only information central banks have. No one – not even the best economists – can today say whether the core PCE deflator in June 2008 will have increased 1.7% or 2.7% compared to June 2007 – or even say whether the core PCE deflator in June 2007 will have increased 1.7% or 2.7% compared to June 2006. Central banks have simply no idea what will happen in the future, meaning they cannot target future inflation either.

Central banks who are guided by the fashionable dogma of inflation targeting thus either face the choice of trying to target a number they know, but cannot affect or trying to target a number they can affect but have no knowledge about.

It is for this reason it was even ironically the case that in recent years, Denmark who pegs its currency to the euro have been more successful in achieving a 2% inflation rate than Sweden who have a floating exchange rate and who are guided by the inflation targeting dogma. This despite the fact that Denmark haven't even attempted to achieve a certain inflation rate, illustrating the futility of inflation-targeting.

June 27, 2006