How to extract information from the monetary pillar

By: Georg Rich, Rich International Consulting

The ECB employs a two-pillar approach to setting monetary policy. Under the first pillar, it monitors a wide range of data and relies on various econometric models in order to forecast real growth and inflation of the euro area. Under the second pillar, it pays specific attention to monetary developments, which, in its view, play a crucial role in determining the future course of inflation.

While the first pillar conforms to the approaches followed by other leading central banks, the ECB’s emphasis on the second pillar is unusual because money has largely gone out of fashion among monetary policy makers. Nowadays, most central bankers share the view espoused by monetarists three or four decades ago that inflation is largely a monetary phenomenon. Nevertheless, they refuse to place a great deal of emphasis on money as a policy indicator. In many countries, the relationship between inflation and money growth is not stable enough to allow policy makers to forecast future price movements from monetary developments alone. For this reason, it is not surprising that the ECB’s second pillar has aroused considerable controversy among observers of European monetary policy.

 

 

 

The development of the euro-area money stock M3 since the introduction of the common currency tends to nourish the doubts about the usefulness of the second pillar. The ECB has fixed a reference value of 4.5 percent for M3 growth. Provided real GDP is near its potential level, the reference value equals the rate of increase in M3 that the ECB considers to be consistent with its objective of keeping HICP inflation at slightly less than 2 percent. As indicated by Fig. 1, M3 growth since 2001 has almost consistently exceeded its reference value, frequently by a substantial margin. Despite the excessive growth in M3 during much of the period 2001-2006, inflation has not increased greatly, contrary to what one would expect if the deviations from the reference value were to contain useful policy information.

In a Background Paper submitted to the EMU Monitor press conference of 29 November 2005, I discussed various difficulties associated with interpreting money growth. An important condition for money growth to serve as a reliable indicator of future inflation is the existence of a reasonably stable relationship between money demand and its key determinants such as prices, real GDP and interest rates. Another problem arises from the fact that even strong fluctuations in money growth need not trigger corresponding movements in inflation. Suppose that the ECB is successful in achieving its objective of keeping inflation slightly below 2 percent. It correctly reacts to any shock threatening to jeopardize its inflation objective by adjusting monetary policy, which in turn alters money growth. In these circumstances, the ECB would waste its valuable research resources by trying to forecast changes in inflation from movements in past money growth. Since inflation would remain more or less constant at slightly less than 2 percent, while money growth would fluctuate considerably, the ECB could not possibly detect a statistically significant relationship between these two variables. Consequently, provided the ECB manages to achieve its objective, forecasts of inflation from past money growth would be likely to understate the importance of money as a driving force behind future price movements.

In view of the difficulties arising from the ECB’s second-pillar analysis, it would be desirable to develop procedures avoiding the pitfalls in forecasting future inflation from current money growth. In this note, I present an alternative based on an econometric analysis of demand for euro-area M3. I assume that the ECB proceeds as follows: It estimates at regular intervals long-run money demand functions, relating the level of M3, deflated by the HICP consumer price index, to the level of real GDP and various interest rate variables, notably to the differential between the three-month euribor rate and the ten-year bond yield. I choose these explanatory variables because they have been shown to be important determinants of the demand for M3 in previous studies.

 

From the estimated money demand functions, the ECB derives reference lines describing the evolution of the nominal value of M3, likely to be consistent with its objective of keeping inflation slightly below 2 percent. Moreover, I take account of the fact that the ECB is prepared to accommodate the increase in the demand for M3 arising from potential real growth in the euro-area economy.

To determine its reference lines, the ECB, say, at the end of 2001 estimates a money demand equation for the sample period 1992Q1 to 2001Q4. It then constructs a reference line for nominal M3, covering the subsequent period from 2002Q1 to 2003Q1. To this end the ECB plugs into the money demand equation an expansion path for the HICP index consistent with its inflation objective. For simplicity, the path for the HICP index is derived from the assumption that the ECB is willing to tolerate an increase in consumer prices of exactly 2 percent from 2001Q4 to 2003Q1. The ECB also inserts in the money demand equation an expansion path for potential real GDP, which is assumed to be equal to the log-linear trend of the actual values. Finally, the ECB takes account of the interest rates by plugging into the money demand function the components of the interest rate differential and the bond yield attributable to its policy actions, as measured by its refinance rate. These components are determined with the help of regression equations, relating each the interest differential and the bond yield to the refinance rate.

Any adjustments in the refinance rate during the periods covered by the reference lines are taken into account and affect their locations and shapes. Consequently, given the estimated money demand equations, the locations and shapes of the reference lines depend not only on the ECB’s inflation objective and its assumption about potential growth, but also on its policy rate of interest. In this note, the ECB is assumed (what I am sure it is doing in practice) to set its refinance rate at levels that – in its view – will help to achieve or safeguard its inflation objective in the longer run. Thus, the procedure presented her does not yield any information on the appropriate levels of the ECB’s refinance rate. Instead, the ECB must rely on its first pillar to determine the levels of its refinance rate that are in sympathy with its inflation objective. Since my proposed procedure exploits information drawn from the first pillar, it cannot serve as a stand-alone approach to setting monetary policy. However, it can be employed to cross check the results of the ECB’s first-pillar analysis.

The reference lines emerging from my procedure are displayed in Fig. 2. They cover the period 2001 to 2007 and extend from the first quarter of each year to the same quarter of the following year. The reference lines for the years 2001-2006 are based on the refinance rates actually set by the ECB. To derive the reference line for 2007, I assume that the ECB will keep its refinance rate at the level of 3.5 percent, that is, at the level at which it set its rate on 13 December 2006. The estimated regression equations are not shown here but are updated and extended versions of those presented in my 2005 Background Paper.

 

 

 

For the period 2001-2006, the reference lines may be compared with the actual development of M3. Major deviations in M3 growth from its reference lines imply that the signals extracted from money growth are at variance with those obtained from the ECB’s first pillar, as the actual development of M3 is inconsistent with the inflation objective and the assumption about potential growth.

This comparison suggests that the ECB should have varied its refinance rate more strongly than it actually did. It should have curbed the marked acceleration of money growth toward the end of 2000 by hiking the refinance rate above the level of 4.75 percent reached in October of that year. In this way, it might have been able to keep inflation nearer its objective. The subsequent cuts in the refinance rate to a low of 2 percent in June 2003 were well justified because M3 growth was near its reference lines in 2004 and 2005.

Of course, the signal-extraction procedure presented in this note is sensible only if money demand is reasonably stable and reliable estimates of the elasticity with regard to real GDP and interest rates can be obtained. My econometric estimates cast considerable doubt on the stability of euro-area money demand. The upward drift in the reference lines until 2004 partly reflects an increase in the estimated income elasticity of money demand. Thereafter, the estimated values of the income elasticity fell again, which explains, at least in part, the subsequent downward shift in the reference lines. Another finding is even more troublesome: The estimated income elasticity for the most recent sample period extending from 1997Q1 to 2006Q3 is statistically insignificant. Furthermore, the estimated parameter for the interest rate differential – though statistically significant – has drifted up substantially throughout the period covered by the econometric analysis.

For this reason, it would be unwise to read too much into the reference line derived for 2007. Perhaps it is possible to draw one conclusion from my analysis: If at all, current M3 growth lies above the reference line. Therefore, current monetary trends suggest that there is room for further tightening of European monetary policy. However, another conclusion is rather more striking: Considering the doubts about the stability of money demand, at least in the most recent sample period, the ECB, right now, would be well advised not to pay too much attention to the current high rate of M3 growth.

In a subsequent note I will also apply my analysis to the aggregate M2. Preliminary results suggest that reference lines derived for M2 may have greater informational content than those for M3.

 

 

---