The demand for euros

https://doi.org/10.1016/j.jmacro.2010.01.001Get rights and content

Abstract

This paper investigates the demand for euros using panel data for 10 euro area countries covering the period from 1999 to 2008. Monetary aggregates are constructed to ensure that money is a national concept by excluding deposits owned by non-residents and including external deposits owned by residents. Initial estimates of a standard money demand function yield income elasticities which are high in comparison with what is typically found in the literature. We next expand the money demand function with four wealth and uncertainty variables, of which housing prices is the most significant one. The inclusion of housing prices in the panel regression reduces the income elasticity to around one. Country-specific developments in housing prices are also able to explain part of the monetary overhang in the euro area since 2005. We conclude that housing price developments within the euro area are relevant to an understanding of the demand for euros, and thus warrant close attention by policymakers at the ECB.

Introduction

Though numerous studies have investigated money demand in the euro area, none has offered an analysis of the demand for euros. The distinction between euro area money demand and the demand for euros relates to the regime shift which the introduction of the euro has brought about. In the empirical literature, euro area money demand studies typically combine pre-1999 with post-1999 data for all or the largest EMU member states.1 As such, they constitute a mixed bag of the demand for the euro’s predecessors and for the new currency. Goodhart (2006) has coined the phrase “pseudo-euro demand-for-money functions” for this line of research.

While understandable from an econometric perspective – time series estimates of money demand benefit from long time series – it is disputable whether the two regimes can be combined. To begin with, non-trivial aggregation issues complicate the construction of historical euro area aggregates (Beyer et al., 2000, Beyer et al., 2001, Bosker, 2006). But also economic theory and empirical evidence caution against the use of pre-1999 data. Rother (1998) employs a cash-in-advance model to argue that changes in the macro-economic environment following monetary unification will lead to shifts in the demand for money. Arnold and De Vries (2000) arrive at similar conclusions on the basis of the monetary model of exchange rates. In general these models imply that the heterogeneity across national money demand functions will be reduced by the introduction of a common monetary policy. Extrapolating money demand relationships from the pre-EMU period to the euro era thus suffers from the Lucas critique.

The recent reduction in monetary heterogeneity across EMU members is evident from looking at a few key monetary variables. A case in point is inflation. Compared to the 1980s – a decade commonly included in euro area money demand estimates – inflation differentials have decreased markedly since the introduction of the euro (Angeloni and Ehrmann, 2004). Adherents to Friedman’s dictum that “…inflation is always and everywhere a monetary phenomenon” would expect a similar change in the behavior of monetary aggregates. Moreover, to the extent that inflation acts like a tax on holding money, the past heterogeneity in inflation rates may have had a differential impact on the demand for money across Europe. That effect has gone now. A comparison of the national non-currency contributions to M3 since the introduction of the euro with their pre-EMU counterparts shows some evidence in favor of increased monetary homogeneity. From the period 1991 to 1998 to the period 1999 to 2008, the average correlation coefficient between national money growth rates increased from 0.12 to 0.22.2 The cross-sectional standard deviation between national money growth rates decreased from an average of 5.7 pre-EMU to an average of 4.6 post-EMU. This decrease took place while average money growth increased (from 6.5% pre-EMU to 8.1% post-EMU), implying an even stronger reduction in the coefficient of variation. Although the post-EMU correlation coefficients are not at the level that we see in established currency unions, see Arnold, 1997, Munoz, 2001, the results indicate a reduction in heterogeneity of monetary aggregates.

Heterogeneity has also decreased in the range of financial instruments which are available to euro area residents and thus in the opportunity costs of holding money. In the past, governments have often limited the availability of alternatives to holding money either by limiting the speed of financial innovation (Germany) or by using capital controls to limit residents’ ability to shift money abroad (France a.o.). Calza and Sousa (2003) note that a different timing of financial innovation and deregulation across euro area countries has spread their effect on money demand. At present, the euro area enjoys complete capital mobility, implying that the same range of financial instruments is available to all euro area residents. The euro area capital markets are also highly integrated, which implies that the opportunity costs of holding money have converged. Last but not least, the ECB sets a single policy rate and manages liquidity on equal terms across the euro area money market. These developments do not imply that euro area money demand has become perfectly homogeneous. Residual cross-country heterogeneity may still arise from differences in banking structures, housing markets, government policies and a lack of convergence in economic growth (Calza and Sousa, 2003).

Empirically the Lucas critique of euro area money demand research has been investigated by comparing the heterogeneity across regional money demand functions within existing monetary unions with the pre-EMU heterogeneity across national money demand functions in Europe (Arnold, 1997, Munoz, 2001). The findings show that the cross-country heterogeneity in European money demand prior to EMU was much higher than that across regions within existing currency unions. Adding these heterogeneous national demand-for-money functions yielded an extremely stable demand for pseudo-euros. At that time, this result lent support to the prominent position of the monetary pillar in the ECB’s initial monetary strategy.

Kontolemis (2002, p. 4) concludes that most studies analyzing aggregate euro-area data suffer from two weaknesses: “(i) the aggregation issue of euro-area data, (ii) the interpretation problems from modeling an aggregate policy reaction function based on data from the different euro-area countries”. Because of the reduction in the heterogeneity in key monetary variables and the imposition of single monetary control, empirical research focusing on post-1999 data will suffer from these weaknesses to a much lesser extent.

To our knowledge, this paper provides the first analysis of European money demand using post-regime change data only. Our dataset covers 10 euro area countries over the period 1999–2008. Not before in modern European economic history have monetary and financial data been more homogeneous across these member countries. We try to overcome the limitations posed by the short time span by using panel data. Adding a cross-sectional dimension to the analysis of money demand goes back some time in the literature (Metzler, 1963, Feige and Swamy, 1974, Gandolfi and Lothian, 1976, Mulligan and Sala-i-Martin, 1992). In recent years an increasing number of papers has employed panel estimation techniques in money demand research (Golinelli and Pastorello, 2002, Sløk, 2002, Mark and Sul, 2003, Dedola et al., 2004, Dreger et al., 2006).

An empirical analysis of the demand for euros has policy relevance for two related reasons. First, several observers have voiced concern about the high rates of money growth since the introduction of the euro (see e.g. Neumann, 2006). The growth in M3 has consistently outpaced the ECB’s reference value of 4.5%. Second, the unease about high money growth has been mirrored in the empirical literature, which has indicated a marked deterioration in the stability of euro area money demand relationships. Dreger and Wolters (2006) note that conventional pseudo-euro demand-for-money functions can be firmly established if data up to 2001 are included. But they tend to break down when the sample is extended beyond 2001. In a subsequent paper, Dreger and Wolters (2008) are able to identify a stable long-run money demand relationship by relaxing the short-run homogeneity restriction between money and prices and by taking into account a rise in the income elasticity after 2001. Goodhart (2006) also dates the start of M3’s “misbehavior” in 2001. This breakdown is no surprise to adherents to the Lucas critique (Giovannini, 1991, Arnold, 1994).

Be that as it may, this still leaves the recent (mis)behavior of monetary aggregates to be explained. Recent research has focused on two explanations for the monetary overhang in the euro area. Goodhart (2006) points at the unusual financial market uncertainty following the bursting of the tech bubble. Heightened uncertainty may translate into an increased demand for nominally safe, monetary assets, the so-called flight to liquidity. The Institut für Weltwirtschaft, 2003, Carstensen, 2004, Greiber and Lemke, 2005 use measures of economic uncertainty and stock market volatility to try and repair the destabilized euro area demand-for-money functions. They find that their uncertainty measures can explain the increase in euro area liquidity during the first years of EMU. This approach is less successful in explaining the monetary surge that took place after 2004 (Goodhart, 2006). A second strand in the literature investigates the role of wealth variables – including housing prices – in euro area money demand (Boone and van den Noord, 2008, Greiber and Setzer, 2007, De Santis et al., 2008). Wealth may influence the demand for money through two channels. According to the substitution channel, higher asset prices will reduce the attractiveness of holding money. In contrast, the income channel predicts a positive relationship between wealth variables and monetary aggregates. This derives from the use of money to store additional wealth and to grease the flow of financial transactions. An intervening variable in the relationship between asset prices and money is bank lending (Greiber and Setzer, 2007). Goodhart (2006) observes that the second monetary surge has been accompanied by an equally rapid increase in bank lending.

This paper contributes to the debate by exploring the demand for euros using a disaggregated dataset excluding pseudo-euros. Apart from providing the degrees of freedom needed for post-1999 estimation, disaggregation also allows for a differential impact of wealth effects across the euro area. Whereas increased financial integration has strengthened the cross-country correlations between equity markets, cross-border arbitrage between national housing markets has remained limited. This suggests that developments in euro area housing markets may have a differential impact on monetary aggregates. In addition to wealth variables we include uncertainty variables in the panel regressions. The remainder of this paper is organized as follows: Section 2 discusses the data and methodological issues. Section 3 reports our empirical findings. Section 4 summarizes the results and concludes.

Section snippets

Construction of monetary aggregates

The construction of national monetary aggregates within a currency union poses a number of statistical challenges. First, our monetary aggregates exclude currency holdings, as euro banknotes and coins are not traceable to individual euro area countries. In this respect, the present study suffers from the same unavoidable defect as previous studies on regional money demand (e.g. Mulligan and Sala-i-Martin, 1992). A second issue relates to the residency of holders of the non-currency components

Benchmark panel estimates

Below we start by estimating a standard demand-for-money function. We thereby focus on the income elasticity. Both theoretically and empirically, economists disagree about the size of the income elasticity.7 The Baumol-Tobin model of the transactions demand for money predicts an income elasticity of 0.5. If transaction costs are positively correlated with the aggregate level of income, the income elasticity would be greater than 0.5. The stochastic

Conclusions

This paper presents estimates of the demand for euros using panel data for the post-1999 period only. This departure from past empirical practice ensures the highest possible level of homogeneity across the panel members and does justice to the regime change in 1999. We find that panel DOLS estimation of a benchmark demand-for-money function yields income elasticities above 1.5, which is high compared to values previously reported in the literature. We extend our benchmark panel model to

Acknowledgements

The authors thank the referees and participants at the November 2008 European Monetary Forum at the Katholieke Universiteit Leuven for helpful comments and suggestions.

References (54)

  • I.J.M. Arnold

    Monetary targeting in the EMU: lessons from the United States

    Kredit und Kapital

    (1997)
  • M. Artis et al.

    Issues in money demand: the case of Europe

    Journal of Common Market Studies

    (2004)
  • A. Beyer et al.

    Reconstructing aggregate euro-zone data

    Journal of Common Market Studies

    (2000)
  • A. Beyer et al.

    Reconstructing historical euro-zone data

    Economic Journal

    (2001)
  • L. Boone et al.

    Wealth Effects on Money Demand in the Euro Area

    Empirical Economics

    (2008)
  • Bosker, E.M., 2003. Eurozone Money Demand: Time Series and Dynamic Panel Results. Research Memorandum WO No....
  • C. Brand et al.

    A money demand system for euro area M3

    Applied Economics

    (2004)
  • Bruggeman, A., 2000. The Stability of EMU-wide Money Demand Functions and the Monetary Policy Strategy of the European...
  • Bruggeman, A., Donati, P., Warne, A., 2003. Is the demand for euro area M3 stable? In: Issing, O. (Ed.), Background...
  • Calza, A., Gerdesmeier, D., Levy, J.V.F., 2001. Euro Area Money Demand: Measuring the Opportunity Costs Appropriately....
  • Calza, A., Sousa, J., 2003. Why Has Broad Money Demand Been More Stable in the Euro Area than in Other Economies? A...
  • Carstensen, K., 2004. Is European Money Demand Still Stable? Kiel Institute of World Economics. Working Paper...
  • G. Coenen et al.

    The demand for M3 in the euro area

    Journal of Applied Econometrics

    (2001)
  • De Santis, R.A., Favero, C.A., Rofia, B., 2008. European Money Demand and International Portfolio Diversification. ECB...
  • Dedola, L., Gaiotti, E., Silipo, L., 2004. Money Demand in the Euro Area: Do National Differences Matter?...
  • Dreger, C., Reimers, H.-E., Roggia, B., 2006. Long-run Money Demand in the New EU Member States with Exchange Rate...
  • Dreger, C., Wolters, J., 2006. Investigating M3 Money Demand in the Euro Area – New Evidence Based on Standard Models....
  • View full text