اثر سیاست های پولی بر تولید در EMU3 : یک روش محدودیت ورود به سیستم

This paper examines the effects of monetary policy shocks on output in the three largest euro area economies – Germany, France and Italy (EMU3) – by applying a new VAR identification procedure. The results show that monetary policy innovations are at their most potent in Germany. However, apart from Germany, it remains ambiguous as to whether a rise in interest rates concludes with a fall in output, showing a lack of homogeneity in the responses. Homogeneity in response to a monetary shock is crucial in a one-size-fits-all framework. Nonetheless, the lack of similarity between the responses, which is hypothesised to cause de-synchronised business cycles in optimal currency area literature, is often based on the premise that monetary policy itself is a major source of business cycle fluctuations. This paper concludes that monetary policy innovations play, at most, a modest role in generating fluctuations in output for the EMU3. Consequently, it is less important whether the effects of monetary policy are homogenous.

Economists have long understood that monetary policy can play an useful role in the determination of fluctuations in output and prices. The consensus view that guides European Central Bank (ECB) monetary policy remains that monetary policy has real short-term effects, with the long-run effects restricted to inflation (see Mihov, 2001). A major source of discussion involves the magnitude and impact of the short-run effects and whether much of the business cycle can be explained by monetary policy disturbances. Vector auto-regressions (VAR), pioneered by Sims (1980), led to most of the early empirical work in this area. Although there appears to be little agreement on the correct way of identifying policy shocks, alternative identification assumptions seem to deliver very similar conclusions: (1) short-term interest rates rise; (2) output, employment and money aggregates decline; (3) prices decline with the impact occurring after a delay of at least six quarters; (4) monetary policy shocks account for at most, a modest portion of output and price volatility. Nevertheless, researchers have disagreed on the best means of identifying shocks. The last 10 years has witnessed a considerable effort to identify monetary policy disturbances using parsimoniously restricted time-series models (Canova and Nicoló, 2002).1 There appears to be little consensus on what type of model or the variables should be used as an indicator of monetary policy. There is a long tradition in monetary economics of searching for a single policy variable – perhaps a monetary aggregate or an interest rate – that is more or less controlled by policy and stably related to economic activity (Leeper et al., 1996). Sims (1992) notes that one cannot determine the influence of monetary policy by simply observing changes in interest rates. In inflation targeting regimes, money is often viewed as the nominal anchor of the system, best demonstrated by the fact that sustained price increases cannot occur without an increase in the monetary aggregates. The use of the real short-term interest rate is also tainted by the fact that it mixes monetary and real influences, such as the rate of productivity growth. Christiano et al. (1996) concluded that short-term interest rates are at best a ‘polluted’ measure of the stance taken by monetary policy.2 In contrast McCallum, 1983, Bernanke and Blinder, 1992 and Bernanke and Mihov, 1998 preferred to make the case for short-term interest rates. In partial agreement, Canova and Nicoló (2002) use the slope of the term structure in place of a short-term interest rate. Of course, it is possible that neither short-term interest rate innovations nor money stock innovations are good measures of policy shifts (Leeper et al., 1996).3 In a monetary union, the implications of the use of the real interest rate channel leads to the possible outcome that monetary policy no longer acts as a brake on the business cycle but instead may accentuate regional economic developments. Heterogeneous effects to monetary policy may lead to divergence amongst business cycles, whilst also indicating significant differences in the propagation mechanisms. As Mihov, 2001 and Mojon and Peersman, 2001 point out, differences in the transmission mechanism of monetary policy play a prominent role in the divergence of economies. One aspect being differences in the industrial structure between countries, which implies that the degree of interest sensitivity will vary across countries, as industries tend to differ in their responsiveness to interest-rate changes.4 If the effects of the policy instrument (the short-term interest rate) are asymmetric, a common monetary policy may amplify misalignments as national monetary authorities are unable to respond to idiosyncratic shocks.5 A common monetary policy will smoothen the union cycle, not the national one. Essentially the paper examines two separate issues – Does monetary policy have any real effect and is the reaction to monetary policy shocks same across countries? By investigating the response functions to monetary policy for the constituent countries, it provides evidence on whether the three countries are able to live under a one-size-fits-all monetary policy regime. The ECB currently makes policy decisions on an aggregate perspective, in contrast to setting an interest rate that reflects a country’s weight in the eurozone economy. As a result, if monetary policy has a strong influence on domestic output in terms of its ability to dampen business cycle fluctuations through independent counter-cyclical monetary policy, it would imply that a potentially vital instrument had been lost in reducing output volatility at a national level. It indicates that a potentially large cost could be paid by a constituent country for joining the EMU if idiosyncratic shocks are predominant. The same would also apply for the counterfactual scenario, where monetary policy has little impact on output and prices and, hence, a potentially smaller cost associated with losing monetary policy independence and joining a monetary union. The cost of joining a single currency should not only be viewed in terms of the country in question, but also to the union as a whole, since if a country cannot deal with idiosyncratic shocks it may further economic misalignment in the union, initiating a cost to the whole union. However, there are channels that may overturn misalignment between countries. Firstly, if monetary policy itself is a source of business cycle fluctuations, then coordinated policy will increase the occurrence of common monetary shocks, leading to a larger degree of business cycle fluctuations.6 A common monetary policy would, hence, remove one possible idiosyncratic shock (assuming idiosyncratic shocks are important). Secondly a one-size-fits-all monetary policy regime may promote higher business cycle correlation through anticipated monetary policy (Mihov, 2001). Such arguments of the endogeneity of monetary policy in a monetary union are a fundamental criterion in optimal currency area literature (see Frankel and Rose, 1998).7 Finally, it has been noted that the ECB is handicapped by the existence of structural rigidities (see De Grauwe and Storti, 2005). In particular it is contended that the ECB cannot stabilise output in the same way the Federal Reserve can. If this remains the case, then monetary shocks, if they were to occur, will play a minor role in accentuating output misalignments amongst member countries. A successful monetary union requires the effects of monetary policy to be relatively symmetric across countries. This paper considers the case of EMU3 (Germany, France and Italy), as together they constitute 70% of the eurozone output. If one were to find the responsiveness of the EMU3 to differ greatly in response to monetary policy movements, it would cast serious doubt over whether such differing economies could live under a ‘one-size-fits-all’ monetary policy regime. It is not clear yet how common monetary policy affects the economies in the union, and the data from more than a decade of coordinated policy preceding the euro’s launch must be interpreted with caution (Mihov, 2001). More generally, the results comparing the effects of interest rate shocks across countries of the eurozone have not shown a high degree of consistency across studies, casting some doubts on their robustness (see Smets, 1997, Mojon and Peersman, 2001, Mihov, 2001, van Els et al., 2003 and Angeloni et al., 2003). Thus this paper addresses and performs an important exercise in attempting to identify monetary policy shocks for the EMU3 in a consistent way so that one may compare different country’s responses, given that currency unions will tend to work better if their constituent economies react in a similar way to changes in monetary policy. Given the different data availability across countries, the use of the new Uhlig (2005) sign restriction identification methodology allows for a very similar identification to be achieved across countries despite these data problems. This is because the sign restriction identification strategy identifies shocks using mild restrictions on multiple time-series, i.e. a contractionary monetary policy shock should increase interest rates, appreciate the exchange rate, contract output, reduce inflation and the money supply, in all economies regardless of how they are measured and regardless of data idiosyncrasies i.e. all inflation measures should fall in response to a contractionary monetary shock, and since the sign restrictions put no quantitative restriction on the responses it does not matter which inflation measure is used and so the particular definition used in a country’s inflation measure is of secondary importance. The same argument holds for the other variables. In contrast, identification procedures that rely on only one variable (e.g. the interest rate) will be more influenced by different data definitions across countries. The findings in this paper far from unequivocally establish that a monetary policy shock leads to a short-run output contraction, especially in the case of Italy. Uhlig (2005) found a similar result for the US economy. The recovered monetary policy innovations are then tested against the term structure of interest rates of all three countries, suggesting that this new sign restriction method robustly captures monetary policy innovations for the three countries. However, due to the lack of homogeneity in responses to monetary innovations, a single monetary policy could lead to an uneven distribution of output across the union, leading to questions on the best way of conducting monetary policy in a monetary union. It must be noted that this concern is tainted by the fact that monetary innovations have played a minor role in driving business cycle fluctuations for the constituent economies. This paper is organised as follows. The second section discusses the sign restriction methodology. The data and empirical specification are discussed in Section 3, with the estimation results of monetary shocks in Section 4, followed by the results of non-monetary shocks in Section 5. Section 6 estimates a robust sign restriction methodology, which draws orthogonal impulse vectors. The results from the orthogonal impulse vectors act as a robustness exercise of the estimated results in Sections 4 and 5. The last section concludes the paper.

The VAR literature, using different methods of identifying monetary policy innovations, has led to a broad consensus that a contractionary monetary policy shock leads to a decline in prices, money and an exchange rate appreciation. Using these consensus restrictions within an open economy framework and applying a new sign restriction method due to Uhlig (2005), this paper examines the effects of monetary and non-monetary shocks on output in the context of the three major EMU member countries. The results obtained here are far from conclusive that a tight monetary policy shock leads to a fall in output, except in the case of Germany, in harmony with the results found by Uhlig (2005) for the US economy, providing some support for traditional RBC theory. The differences in magnitude and timing of the dynamic responses of output across the EMU3 indicate differences in the transmission mechanism of monetary policy. As the EMU3 member countries have maintained many of their idiosyncrasies, these have the effect of creating asymmetries in the transmission of common shocks. The results here support the policy conclusions made by De Grauwe and Sénégas (2003) that a common central bank could improve the quality of monetary policy making by using national information about inflation and the output gap, instead of focusing only on the union-wide aggregates. Thus the lack of homogeneity in responses in the short-run could be a source of divergence in business cycles between countries within a monetary union. In this context, our historical and forecast error-variance decomposition results are of the view that monetary policy shocks have not been a significant driver of business cycle fluctuations. Although a common monetary policy removes one possible idiosyncratic shock, this result raises doubts over whether policy makers should be concerned about monetary policy innovations and one-size-fits-all policy in a currency union, since monetary innovations are not a key source accounting for the fluctuations in economic activity for the EMU3.