نرخ ارز - یک جاذب شوک و یا منبع شوک؟ یک مطالعه از چهار اقتصاد باز

The paper provides SVAR estimates for the UK, Canada, Sweden and Denmark, making explicit a monetary policy reaction function and taking account of exchange rate targeting practices. It examines the role of the exchange rate as shock-absorber as opposed to a source of its own, and destabilizing, shocks. We find that in all countries but the UK, real shocks are predominantly symmetric relative to the neighbor, implying little need for the exchange rate to act as shock-absorber. The shocks arising in the exchange market appear to have played a more important role in Denmark than in the other countries.

Traditional optimum currency area (OCA) analysis gives a special role to the exchange rate as a potential shock-absorber and stabilization mechanism; in effect, the “OCA null” is that exchange rates behave in this way. On the other hand, foreign exchange markets can actually contribute towards destabilization of the economy, as for example in times of exchange rate crisis. The role that is effectively played by the exchange rate is thus a critical issue. For example, commenting on the position of the UK vis-à-vis EMU, Buiter (2000) has suggested that a principal benefit of EMU membership for the UK might well be to escape from the destabilizing effects of the sterling foreign exchange market: “Under a high degree of international financial integration, market-determined exchange rates are primarily a source of shocks and instability” (Buiter, 2000, p. 213). An article IV consultation document of the International Monetary Fund (1999) has also drawn attention to the fact that output fluctuations in the UK have in recent decades been relatively large; and that important roles for the interest rate and the exchange rate in generating those fluctuations can be identified. More recently Kontolemis and Samiei (2000) expanded on these points in an IMF working paper. This paper proceeds to examine the issue raised by Buiter and the IMF empirically, using an SVAR approach. In order to provide perspective, however, the scope of the analysis is deliberately not restricted to the UK. Rather, we have chosen to examine the position of four countries which, like the UK, have large neighbors with whom they trade a great deal, and with whom monetary union (or a quasi-union arrangement) is an important policy option. The four economies concerned are those of the UK, Canada, Sweden and Denmark. The UK, Sweden and Denmark all face the important option of participating in EMU. For Canada, whilst no formal option exists of joining the monetary union of the US, “quasi-union” options exist; for example, Canada could adopt a Currency Board arrangement or allow her economy to become (US-)dollarized. There has been a flurry of recent discussion of this type of option (see, e.g., Buiter, 1999, Courchene and Harris, 1999 and Laidler and Poschmann, 1998). As illustrated below (Table 1), of these four countries the UK is the largest in relation to its ‘neighbor’ (whether this is taken to be Germany or the euro area). Canada and Sweden are of similar size relative to their neighbor (when Sweden's neighbor is taken to be Germany), and Denmark is the smallest. When the relative size of Denmark and Sweden is measured against the euro area, both countries are considerably smaller than Canada. Table 1. Country comparison: size and openness Size relative to neighbora Export share in GDP Import share in GDP Share of exports to neighbora Share of imports from neighbora Canada 0.09 40.7 39.0 0.83 0.67 Denmark 0.07/0.02 36.0 32.0 0.22/0.44 0.22/0.51 Sweden 0.10/0.03 43.8 36.8 0.11/0.38 0.19/0.49 UK 0.66/0.20 28.7 29.2 0.11/0.46 0.12/0.44 Sources: OECD Main Economic Indicators, IMF Direction of Trade Statistics, own calculations. a Neighbor is US for Canada, Germany/euro area for the other countries. Relative size is measured on the basis of PPP-adjusted GDP. All data refer to 1997. Table options Using an SVAR approach to examine the topic in question seems the most obvious way to proceed. The methodology directly concerns itself with the stochastic behavior of economies, which is the central issue here. Moreover, as detailed below, there are some antecedent studies with which our results can be compared. There is also a small but growing literature on SVAR estimation of open economies which provides some useful pointers for the current study. We find that in all countries but the UK, real shocks are predominantly symmetric relative to the neighbor, such that there is little need for the exchange rate to act as shock-absorber. The shocks arising in the exchange market appear to have played a more important role in Denmark than in the other countries. In the remainder of the paper, we will first introduce the research strategy with which we will analyze the issues at hand, highlighting where we depart from earlier contributions in the field. Section 3 describes the econometric methodology that will be applied, before Section 4 presents the results obtained. In Section 5, we will analyze the robustness of our results. Section 6 concludes.

The aim of this paper has been to cast light on the role of the exchange rate, as either stabilizing or distorting the economy. We have examined this question in the context provided by four open economies – the UK, Sweden, Denmark and Canada – each of which faces an important policy option of monetary union (or a similar arrangement) with a big neighbor. In contrast to several earlier contributions to the literature, we chose to tackle this question with a model that is not formulated in relative terms to the neighboring country and trading partner. This specification has the advantage that it enables us to gauge whether shocks are mainly symmetric or idiosyncratic in nature. Only if shocks are predominantly asymmetric is the exchange rate needed to absorb such shocks. We therefore analyze the shock-absorbing role of the exchange rate only in cases where this is a relevant task. We went on to examine whether shocks are bred by the exchange market. If such shocks are important determinants of the exchange rate and have the potential to distort the economy, the exchange rate can destabilize the economy. Finally, we discussed whether, in fact, monetary policy has significant effects on real output and prices and at what horizons. This analysis was based on a specification which draws on the SVAR literature on monetary policy in an open economy, allowing explicitly for exchange rate targeting practices. We conclude that on this basis, in the case of Sweden, where shocks appear predominantly symmetric, the exchange rate appears to “dance to its own tune” and monetary policy has little effect on output, the role of the exchange rate in shock absorption has been very limited. This result is slightly modified with an alternative identification scheme, where the exchange rate is estimated to have a somewhat larger role as shock-absorber in the case of supply shocks. For Denmark, a similar verdict is available. The reasoning is slightly different. Shocks appear mainly symmetric, again. The exchange rate is quite strongly driven by monetary policy innovations, and monetary policy itself responds to demand and supply shocks. However, the evidence is that monetary policy in Denmark has no effect on output. In the case of Canada, whilst shocks are mainly symmetric, monetary policy has real effects, and the exchange rate is not mainly a source of shocks. A stabilizing role is feasible for monetary policy and the exchange rate. For the UK, there are inconsistent indications: shocks are asymmetric and monetary policy is (weakly) effective for output. On the other hand, a large component of variation in the exchange rate is due to exchange market disturbances themselves: demand and supply shocks are negligibly involved. The role of the independence of monetary policy and the exchange rate for shock absorption is therefore not clear-cut. What do these results suggest about the monetary union option for these countries? On the face of it, our scorecard might suggest that monetary union is easier to recommend for Sweden and Denmark than it is for Canada and, still more, the UK. But it is hard to predict how far losing the exchange rate and monetary policy as independent tools for macroeconomic policy would change the stability of the economies, since in monetary union other factors will come into play, as, for example, an increase in the volume of trade. To the extent that the OCA criteria are endogenous (as suggested, e.g., by Frankel and Rose (1998) and, implicitly, by Kontolemis and Samiei (2000)) a simple extrapolation of economic structures into the new regime is not necessarily meaningful. Nonetheless, it is useful for those countries considering a monetary union option to assess the workings of their stabilizers under the current regime. In this respect, the present paper provides useful guidance. Of course, the results obtained in any SVAR exercise are only as good as the identification scheme that is employed to ‘make sense’ of the residuals. The restrictions involved here are, largely, conventional in that they have been adopted in earlier studies of monetary policy effectiveness; and the over-identifying restrictions suggested by theoretical priors are all largely satisfied in our results. We have provided a wide range of sensitivity analyses, varying the underlying assumptions of our model. None of our main conclusions is perceptibly affected.