Ever since its introduction at the beginning of 1999, the external value of the euro has been decreasing, reaching an all-time low late October 2000, not just relative to the U.S. dollar but on a range of crossrates as well. At that point, the ECB felt obliged to intervene in foreign exchange markets. Following the Presidential election and in the context of rapidly weakening economic indicators in the U.S., however, market sentiment vis-a-vis the euro improved markedly in December and January, allowing the new currency to appreciate against the dollar by some 15 percent (from 0.83 to 0.95) before losing ground again since March 2001.

To evaluate whether the euro has truly bottomed out and may now be able, at long last, to establish a sustained upward trend, one first has to understand, what had caused its steep two-year long dive. Are these causes no longer present? Are they likely to be reversed now? Or are they more likely to re-assert themselves?

In Euroland, excess liquidity had grown steadily since the introduction of the euro, as illustrated by an accelerating, credit-driven trend growth of M3 up to the third quarter of 2000. After the ECB had actively eased its already loose policy in April 1999, perceptions gained ground in the market that, while real economic growth in Euroland has recovered to historically high rates of 3 1/2 percent, short-term interest rates might continue to be kept at historically low levels. The yield curve in the bond market therefore steepened through much of 2000, reflecting the expected degree of excess liquidity. In response to these spreading perceptions, the European Central Bank (ECB) reluctantly raised its official interest rates in seven steps by 200 basispoints between November 4, 1999 and October 5, 2000, before lowering it again by a shade in May 2001 and one more time in August 2001. The key repo rate now stands at 4.25 %. Through November and December 2000 this interest rate level proved sufficiently high to turn the euro around, because expected interest rates in the U.S. began to ease at the same time and have now fallen to below 3.5 % on futures contracts maturing in the fourth quarter of 2001.

The effect of excess liquidity on capital outflows from Euroland and thereby on the euro's exchange rate had been amplified by the fact that international borrowers stormed European credit markets and raised funds in advance of their short-term needs to take timely advantage of still low borrowing costs and of unprecedented liquidity levels offered by the monetary unification of hitherto fragmented European markets. Euroland's newborn corporate bond market saw the biggest slice of the action. Once international borrowers had raised their funds, most of these were swapped into US dollars right away. American borrowers played an important role in this game, but quantitatively even more important were European borrowers themselves, who no longer borrow as much in U.S. markets (then swapping the proceeds into their home currency) as they used to prior to 1999, because financing terms on their European home markets have now become just as competitive. The net effect was a structural decline of capital imports (or, borrowing overseas) into Euroland.

Moving forward, foreign borrowing in booming European financial markets is likely to continue, perhaps not quite at the frenzy pace of the last two years, since many potential borrowers must feel relatively well-funded by now, but nevertheless. More interesting should be the question how much excess liquidity is still left in Euroland. In other words, is there still a gap between required short-term interest rate levels and actual interest rates which would need to be closed, and if so, will it be closed or not?

One way to determine, where policy-controlled interest rates "should" be is provided by the so-called Taylor rule. This analytical rule is derived from the historical relationship between inflation and real economic activity on the one hand and policy-controlled interest rates on the other hand. (More precisely, Taylor rule models attempt to predict policy-controlled interest rates as a function of the difference between actual inflation and an inflation norm and of the level of the gap between actual and potential output. The seminally original research work to this effect can be found in John B.Taylor, "Discretion versus policy rules in practice", Carnegie-Rochester Conference Series on Public Policy, Vol.39, 1993, pp. 195-214. For two topical and still reasonably up-to-date applications of this approach, see in the U.S. context the same author's "Homer Jones Lecture" at the Federal Reserve Bank of St.Louis in 1998, which you can download here in PDF, and in an EMU context Stefan Gerlach and Gert Schnabel, "The Taylor rule and interest rates in the EMU area: a note", BIS, August 1999, which you can download here in PDF as well.) We, too, work with Taylor rule matrices where the top row shows a realistic range of inflation rates and the front column includes a realistic range of unemployment rates (output gap estimates are in theory superior, but in practice quantitatively less reliable than employment gap measures). For each combination of inflation and unemployment, our matrix shows the "fitting" policy-controlled interest rate. Here, our underlying (published, but password-protected) econometric relationships which generate the "required" interest rate are not so much derived from past history as from currently prevailing policy simulation rules, including, among other things, a stylized semi-elasticity of unemployment with respect to wages and an inflation tolerance threshold.

It is well-known that the link between monetary policy and measured overall unemployment is relatively weak in Euroland, because the link between unemployment and short-term changes in the overall business cycle is relatively weak. In other words, the cyclical component of unemployment is relatively small. We therefore focus, alternatively, on a small number of regional unemployment rates all across Euroland (yet, in different countries) which have historically been among all regional unemployment rates the cyclically most responsive ones. Among the latter, we include both regions with particularly low unemployment on average of the cycle and regions with particularly high unemployment rates on average of the cycle. Our inflation indicator is unit labor costs using a trend measure of productivity growth. We thereby obtain a "required" short-term interest rate, which changes each month as new data for unemployment and wage inflation come in. If the actual short-term rate is and is expected to remain below the required rate, we would say the degree of excess liquidity in the economy is expected to grow.

We run the same kind of Taylor rule matrix for the United States, also using only the cyclically most responsive regional unemployment rates instead of the more sluggish national average measure. We then compare the two gaps between required and actual interest rates in Euroland and in the United States, and where the gap is arithmetically larger, we would expect the exchange rate to fall.

This framework can also be applied to analyze the implications of the current downturn of the US economy for the euro exchange rate. Much will depend on whether Chicago futures markets expect US interest rates to drop below or to stay above the thus defined required rate at predicted levels of unemployment and inflation.

Our password-protected dollar-euro area

- [obtain a Trial password for exploratory access] -