Happy new year everyone! The Wiener Philharmoniker are going through their New Year's tradition while I'm typing this. I just can't believe it's Nikolaus Harnoncourt conducting Strauss (any of them); I still associate him primarily with earlier composers. There is a guest appearance by Johannes Brahms this year in the form of his Hungarian Dances, which makes the Harnoncourt association somewhat more plausible. But in any case, with the effects of the fireworks dissipated and a whole new year to trudge through, how can I resist writing about monetary policy in the Eurozone?

I have written before about the stresses in the European Monetary Union, focusing mostly on the divergent inflation rates within the member states. Looking at real interest rates deflated by the harmonized inflation indices there is a huge difference between large member states of EMU such as Germany and Italy of almost 200 basis points. Trying to refine this analysis a bit further, I've been playing with Taylor-Rule implied interest rates.

The Taylor Rule was first formulated by economist John Taylor in 1993 with European Monetary Union in mind. The rule is fairly simple: it says that the nominal interest rate is determined by four things: the real equilibrium interest rate, recent inflation, the central bank's inflation target and the output gap. Conecptually it's one of those things that make good sense, but in practice it's hard to get much out of it. The problem is that out of the four quantities you need to calculate the nominal interest rate, two are hard to quantify. The real equilibrium interest rate is usually approximated by taking a long-run average of real interest rates in the past, which of course have been influenced by monetary policy. Inept monetary policy in the past will lead to higher real interest rates as a "credibility premium" that the market demands. Even hard to estimate (especially contemporaneously) is the output gap. This is the difference between actual growth and potential growth in the economy. Actual growth is simple to measure, but how do you come up with potential growth? This is usually again calculated by taking a long-term trend growth rate for the economy. Neither of these two approximations is perfect (or even close to it), but the Taylor has gained widespread currency as a rule of thumb for assessing monetary policy. Many of the practical problems are explored in this paper by Robert Hetzel of the Federal Reserve Bank in Richmond.

In the case of European Monetary Union, the Taylor Rule does a remarkably good job (pdf) in explaining central bank policy. Graph 2 on page 7 of the linked document shows the Taylor Rule interest rate set against the actual interest rate prevailing at the time. In order to see what the Taylor Rule is telling us now about the proper level of interest rates for the member countries of the European Monetary Union, I used the latest OECD data, which includes estimated output gaps, nominal interest rates and inflation rates. From this I calculated the average real interest rate over the 1994-2001 period and used that as the equilibrium value in the Taylor Rule, except for the EMU countries where I used the 3.55% value calculated in the BIS study linked above. As an inflation target I used 2% for EMU (which is the ECB's "reference value"), and 2.5% for the UK and the US. The Federal Reserve does not have an official inflation target though. The results are very interesting. This graph shows the Taylor Rule implied nominal short rates for most of the OECD countries, while this graph shows just the EMU data.

There are two immediate observations: first, the Taylor Rule comes up with much higher interest rates than the current ones in most countries and second, there is a very substantial dispersion in Taylor Rule rates for the EMU countries. The first phenomenon can be explained by the fragile state of the world economy. The threat of deflation is bigger than the threat of inflation, so the Federal Reserve is erring very much on the side of loose monetary policy. Curing inflation is easier than curing deflation. The ECB has grudgingly followed the Fed's lead, although the widespread perception still is that the ECB is still fighting the last war rather than the current one. This explains the very low interest rates that we're currently seeing. The equilibrium real interest rates that went into the calculation are higher than the currently prevailing nominal interest rates for both EMU and the US. Obviously the judgment now is that such high levels of real interest rates are inappropriate. Reducing the equilibrium real interest rate is a translation: all levels of Taylor Rule rates shift down by the same amount. If one were to assume that real rates should be 3% lower than the equilibrium used in the calculation, one should subtract 3% from the values shown in the graphs.

But as for the second phenomenon, why is there such a huge divergence within EMU countries when the BIS study showed that the Taylor Rule did such a good job explaining monetary policy? The reason is that the BIS study used a weighted average of EMU countries' economic data. This means that within that average, each country was able to fine-tune its monetary policy to the prevailing domestic circumstances. Having a single interest rate for all 12 EMU countries means that this fine-tuning of monetary policy to local output gaps and inflation is no longer possible. Thus the single interest rate set by the ECB is wildly inappropriate for a number of countries. The only ones within a 25 basis point range of the Eurozone Taylor Rule rate are Italy and the Netherlands. As noted above, this dispersion is not affected by assuming a different equilibrium real interest rate for the Eurozone. The dispersion would be affected if one were to calculate an equilibrium real interest rate for each of the countries, resulting in a lower rate for Germany and higher for Italy (a gap of 110 basis points). In other words, the Taylor Rule dispersion would be even bigger in that case, from 151 basis points to 261 basis points. Either one is a huge difference in monetary policy.

The Taylor Rule is not explaining current monetary policy very well, even if it has a decent track record in the Eurozone. Real interest rate are now much lower due to the balance of risks in the world economy. Whatever framework one uses, be it simple real interest rates or a Taylor Rule, the divergence within the EMU countries is still substantial and shows that having one interest rate for all the EMU countries is far from an optimal situation.