انتقال بین المللی از تکان های وارد شده سیاست پولی آمریکا: شواهد از VAR's

This paper documents data-oriented, detailed evidence on the international transmission of U.S. monetary policy shocks for the flexible exchange rate period using VAR models. First, U.S. expansionary monetary policy shocks lead to booms in the non-U.S., G-6 countries. In this transmission, changes in trade balance seem to play a minor role while a decrease in the world real interest rate seems important. Second, U.S. expansionary monetary policy shocks worsen the U.S. trade balance in about a year, but subsequently it improves. Overall, the basic versions of Mundell–Flemming–Dornbusch (MFD) and the sticky price (or sticky wage) intertemporal models do not seem to be consistent with the details of the transmission mechanism, and some extended versions seem to be necessary to fit the details.

This paper examines the mechanism by which U.S. monetary policy shocks are internationally transmitted in the flexible exchange rate regime. Does a monetary expansion in the U.S. lead to recessions or booms in other countries? Does a monetary expansion improve or worsen the trade balance (or the current account)? These questions have long been discussed, but remain controversial.1 The ambiguity in the effects (predicted by theoretical models) comes from two sources. The first is ambiguous predictions by each model. For example, the traditional Mundell–Flemming–Dornbusch (MFD) model alone has ambiguous predictions on the effects.2 The ambiguity in the international monetary transmission mechanism has been amplified by the development of a new framework. Applications of the intertemporal model (equipped with sticky price or/and sticky wage) to the international monetary transmission mechanism, for example, Svensson and Van Wijnbergen (1989) and Obstfeld and Rogoff (1995), provide a different perspective. Their basic predictions are sometimes different from those of the MFD model. Furthermore, even when both frameworks suggest similar predictions, the detailed transmission mechanism differs in some cases.3 To resolve the ambiguity of the two sources, this paper documents empirical evidence. First, I examine the effects of monetary policy shocks on the variables of the primary interests, such as trade balance and foreign output. Second, I infer the exact transmission mechanism by examining the effects on related variables such as terms of trade, real interest rates, and so on. The first type of evidence is directly related to policymaking. For example, if a monetary expansion leads to an improved trade balance, then a country with trade deficits may use a monetary expansion to improve it. The second type of evidence can shed a light on the recent extensive search for the correct theoretical model for international monetary policy analyses.4 I focus on the two main consequences of international transmission of monetary policy shocks on which ambiguities are clearly present: (1) the effects on the trade balance (or the current account) and (2) the effects on foreign output. First, the basic MFD model predicts that a monetary expansion leads to terms of trade deterioration or real exchange rate depreciation, which leads to an improvement in the trade balance (the expenditure-switching effect). However, an increase in domestic income following a monetary expansion increases domestic import demand, which may worsen the trade balance (the income-absorption effect). The intertemporal model emphasizes the forward-looking intertemporal decisions of economic agents. A monetary expansion leads to a temporary increase in income, so that the current account may improve as a result of consumption smoothing. However, the current account may worsen if investments increase substantially (due to a fall of the real interest rate).5 Second, regarding foreign output, the basic prediction of the MFD model is based on trade balance movements. A domestic monetary expansion leads to a worsening of the foreign trade balance through the expenditure-switching effect and a decrease in foreign output (beggar-thy-neighbor policy). This adverse effect on foreign output may be reversed if the foreign trade balance improves through the income absorption effect.6 The intertemporal model also comprises the expenditure-switching effect, so that the foreign output may decrease. However, a world wide drop in real interest rate (if the home country is a large open economy like the U.S. and the world capital market is integrated to some extent) may increase the world aggregate demand for current goods, including foreign current goods. As a result, foreign output may grow.7 There have been many empirical investigations on the international monetary transmission mechanism. First, there are simulation experiments that use large-scale structural models which employ different versions of the MFD models (for example, studies in Bryant et al., 1988, and Taylor (1993)) or that use calibrated dynamic stochastic general equilibrium models (for example, Cardia (1991), Kollman (1997) and Kollman (1999), Betts and Devereux (2000a) and Betts and Devereux (2000b), Chari et al. (1998), and McKibbin and Sachs (1991)). However, most models in these studies are based on very large number of identifying restrictions so that the evidence may not serve as data-oriented empirical evidence. Furthermore, they are based on specific theoretical models, which may not serve as the gauge for the relative success and failure of different theoretical models. Second, some studies use empirical models that employ minimal identifying restrictions and that do not depend on specific theoretical models (for example, Lastrapes (1992), Eichenbaum and Evans (1995), Grilli and Roubini (1995), Kim and Roubini (2000), and Clarida and Gali (1994)). However, these studies are limited in that they investigate only certain features of the data, for example, exchange rate dynamics.8 This paper documents empirical evidence using the empirical model that employs minimal identifying restrictions and does not depend much on a specific theoretical model. It further tries to infer the detailed international monetary transmission mechanism by investigating a broad set of variables. To achieve these objectives, I employ newly developed models from the VAR literature on identifying monetary policy shocks, for example, Christiano et al. (1996) and Kim (1999). I employ the “marginal” method and add each international/foreign variable one by one to the basic model. In this way, I am able to obtain more precise estimates, and I am able to examine the effects on a large number of variables without confronting possible arbitrariness and the complexity of modeling international interdependence. I further examine the robustness of results by adding two or three variables simultaneously and allowing some international interactions. On the other hand, I experiment with several identification schemes, including both recursive and nonrecursive schemes since some past studies suggested that different identification schemes produce different results in some cases. I also experiment with different data frequency. Overall, I argue that I provide some initial, data-oriented, empirical evidence on the detailed international monetary transmission mechanism. The major finding of the paper is that U.S. monetary expansion has a positive spillover effect on non-U.S., G-6 output. This positive spillover effect seems to occur through the world capital market. A monetary expansion of the large open economy (the U.S.) decreases the world real interest rate and seem to stimulate the world aggregate demand on current goods and services of both U.S. and non-U.S. countries, which was theoretically suggested by some intertemporal models such as Svensson and Van Wijnbergen (1989) and Obstfeld and Rogoff (1995). Though a U.S. monetary expansion leads to medium-run and long-run improvements in the U.S. trade balance (and a possible worsening of the foreign trade balance) as predicted by the basic MFD model, the magnitude of the trade balance changes seems to be too small to ensure the beggar-thy-neighbor aspect of monetary expansion. There are some other interesting findings. U.S. monetary expansion leads to a short-run worsening of the trade balance but a persistent medium- and long-run improvement of the trade balance. Further investigation on other related variables, such as exchange rates and terms of trade, suggests that empirical evidence is consistent with the MFD model in which the income-absorption effect dominates in the short run but the expenditure-switching effect dominates in the medium and long run. Regarding the intertemporal model, a simple form of the intertemporal model without investments (such as Svensson and Van Wijnbergen (1989) and Obstfeld and Rogoff (1995)) falls short of explaining the dynamics of the short-run trade balance (or current account). Investment, as well as saving, is important in explaining the short-run dynamics. Finally, an interesting result on policy endogeneity is provided. Previous researchers, such as Grilli and Roubini (1995), suggested that non-U.S., G-7 monetary policy strongly follows the U.S. monetary policy. However, after controlling for inflationary or supply shocks that were not controlled in past studies, I find that the reaction of non-U.S. monetary authorities endogenous to U.S. monetary policy does not seem to be very strong (except for Canada). Section 2 explains the VAR modeling and the empirical methodology. Section 3 discusses the effects on the trade balance. Section 4 examines the effects on foreign output. Section 5 discusses extended experiments and the robustness of results. Section 6 summarizes the results and concludes.

6. Conclusion I examine the international transmission mechanism of U.S. monetary policy shocks using VAR models. First, a U.S. monetary expansion worsens the trade balance in the short-run, but improves trade balances persistently in the medium and the long run. Impulse responses of other related variables are consistent with the MFD model in which the short-run income-absorption and the long-run expenditure-switching effects are present. The intertemporal model in which investment as well as saving plays an important role in the short run is also consistent with the results. Second, a U.S. monetary expansion leads to booms in non-U.S., G-6 countries. Changes in the trade balance seems to be too small to explain the foreign booms while the increase in world aggregate demand (through the world real interest rate changes) seems to be the important channel in the transmission. In this regard, the basic MFD model, which emphasizes the role of trade balance, does not seem to be consistent with the empirical evidence, while some versions of intertemporal model emphasizing the changes in the world real interest rate seem to be consistent with the evidence. Overall, the basic versions of Mundell–Flemming–Dornbusch (MFD) and the sticky price (or sticky wage) intertemporal models do not seem to be consistent with the evidence about all of the details of the transmission mechanism, and some extended version seems necessary to match the details. Future research on theoretical models, which specifies detailed international monetary transmission and which can explain the detailed empirical regularities, including those reported in this paper offers promise. On the other hand, following the initial data-oriented evidence on international monetary transmission reported in this paper, further documentation on the data-oriented evidence is necessary. It is worthwhile to construct large-scale, multi-country, (Bayesian) VAR models to consider possible international interdependence, which may not be captured in the small-scale VAR models used in this paper, in addition to the further documentation using the small scale VAR models.