پویایی نرخ ارز و اثرات رفاهی سیاست های پولی در مدل دو کشور با سوگیری محصول خانگی

International spillovers and exchange rate dynamics are examined in a two-country dynamic optimizing model that nests Obstfeld and Rogoff (1995) and allows for home-product bias in consumption patterns. Allowing for home bias changes the results in three ways. Wealth transfers associated with net foreign asset positions induce movements in the real exchange rate and produce large short-run and small long-run deviations from consumption-based purchasing power parity. Interest rates, both real and nominal, can differ across countries; home bias is a necessary but not sufficient condition for Dornbusch (1976) type exchange rate overshooting. And the welfare effects of monetary policy depend not only on world demand but also on the expenditure-switching effect of an exchange rate depreciation; expansionary monetary policy is ‘beggar-thy-neighbor’ if individuals have strong preferences for domestic goods.

International macroeconomic models have traditionally incorporated the presumption of a home-product bias in spending. From the early attempts to add international transactions to the Keynesian model (Machlup, 1943 and Meade, 1951), to the advent of models with perfect capital mobility (Mundell, 1963, Mundell, 1968 and Fleming, 1962), the presumption has been that foreign trade is a small portion of total economic activity. I follow in this tradition by investigating the effects of home bias on the transmission of monetary policy in a two-country dynamic optimizing model. Some models have deviated from the presumption of home bias. In the 1970s, in models of the monetary approach to the balance of payments, traded goods produced in different countries were assumed to be perfect substitutes. More recently, individuals in the two-country, sticky-price dynamic optimizing model of Obstfeld and Rogoff (1995) have identical preferences for all goods. The assumption of identical tastes simplifies the analysis but is somewhat restrictive. Identical goods preferences (and the law of one price) imply that both relative and absolute consumption-based purchasing power parity (PPP) hold at all times. Identical preferences also preclude exchange rate overshooting: interest rates, both real and nominal, are identical across countries, so uncovered interest parity (UIP) implies that after a monetary expansion the nominal exchange rate jumps immediately to its long-run level. Allowing for a home-product bias makes the model consistent with two aspects of observed exchange rate behavior that macroeconomic models are poor at replicating: the extreme volatility of nominal exchange rates and the existence of long-run deviations from PPP. As home bias increases, Dornbusch (1976) type nominal exchange rate overshooting becomes more pronounced. Moreover, when there is home bias, nominal exchange rates are more volatile than fundamentals such as price levels and money supplies, an empirical regularity which real business cycle models have trouble matching (see Chari et al., 1998). It is not surprising in a model with home bias and sticky prices that asymmetric changes in money supplies produce short-run deviations from relative PPP, but they also produce small permanent movements in the real exchange rate, a result that is consistent with empirical findings. 1 In the short run, a change in relative money supplies causes a current account imbalance. Current accounts must be balanced in the long run, but the short-run imbalance results in a permanent net foreign asset position; monetary policy is not neutral in the long run. The interest payments on the permanent net foreign asset position represent (small) wealth transfers. With home bias, the wealth transfer is spent disproportionately on domestically produced goods, inducing (small) permanent movements in the real exchange rate and, hence, small permanent deviations from relative PPP. 2 Allowing for differences in tastes across countries has implications for welfare analysis. When preferences are identical (and initial net foreign assets are zero), a domestic monetary expansion causes an equal increase in utility throughout the world. With home bias, monetary expansion has the standard ‘beggar-thy-neighbor’ effect. The reason is straightforward. The expenditure-increasing effect on utility is identical across countries and is not affected by the degree of home bias, but the expenditure-switching effect, evident only when there is home bias, increases domestic utility at the expense of foreign utility. With strong enough home bias, the switching effect is greater in magnitude than the increasing effect and foreign utility decreases. Only when tastes are identical does the switching effect drop out; in this case utility in both countries depends only (and identically) on changes in world demand. The model presented here belongs to what has been termed the ‘new open economy macroeconomics’ that has grown from Obstfeld and Rogoff’s seminal 1995 paper; see Lane (2001) for an excellent survey. Other models, developed simultaneously, that like this one have the law of one price but obviate the assumption of identical tastes, are Ghironi (1998), which allows for biased spending habits across continents in a US-Europe model, and Hau (2000), which allows for nontraded goods; both use a less general specification for real balances and therefore cannot produce exchange rate overshooting.3 The law of one price does not hold in the pricing-to-market models of Tille (2000) and Betts and Devereux (2000). While there is much empirical evidence against the law of one price (Isard, 1977; Engel and Rogers, 1996), Obstfeld and Rogoff (2000) point out that in pricing-to-market models an exchange rate depreciation counterfactually improves a country’s terms of trade. In any case, many of the conclusions from such models are similar to mine. Finally, home bias has been incorporated into the numerical simulations of real business cycle theorists; Chari et al. (1998) assume home bias, but do not spell out its implications. The paper is organized as follows. In Section 2 the dynamic optimizing model is developed. In Section 3 analytical solutions, obtained by log-linearizing around an initial steady state, are discussed and a numerical example is presented. The welfare implications of home bias are investigated in Section 4. Section 5 concludes.

The model presented in this paper fully nests Obstfeld and Rogoff (1995) while allowing for a home-product bias in consumption. Allowing for home bias enables richer analysis of the welfare implications of monetary policy and has implications for exchange rate determination. To show the implications home bias has for the welfare effects of monetary policy shocks, changes in utility are divided into two components, the shifting effect of increased world demand and the switching effect of an unexpected exchange rate depreciation. The shifting effect is identical across countries and is not affected by the degree of home bias. The switching effect, apparent only when there is home bias, increases Home utility at the expense of Foreign utility. With enough home bias, monetary policy is beggar-thy-neighbor: the switching effect is great enough that Foreign utility falls. Conversely, if individuals prefer imported goods, monetary policy can be beggar-thyself. This division provides a clear explanation of Obstfeld and Rogoff’s welfare result: without home bias there is no switching effect, thus utility in both countries depends only (and identically) on changes in world demand. The model with home-product bias has a number of implications for exchange rate determination. Not surprisingly, in a model with sticky prices and different consumption baskets across countries, there are short-term deviations from PPP. This model also produces small, permanent deviations from PPP. As argued in Krugman (1990), if preferences are biased, wealth transfers affect the real exchange rate. In my model any asymmetric shock results in a temporary current account imbalance, a permanent net foreign asset position, and, hence, permanent wealth transfers and permanent deviations from PPP. Also, with home bias there is Dornbusch (1976) type overshooting and increased volatility of real and nominal exchange rates. The model is decidedly simple in a number of respects. Only monetary shocks are investigated here; the effects of government spending and productivity shocks in the same framework are analyzed in Warnock (1999). The absence of investment and the simplicity of the supply process preclude the model from replicating observed patterns in output. Price dynamics would have to be considerably richer, perhaps in the form of Calvo (1983) pricing, if one were to bring the model to the data. The assumption of mirror-image countries is easily relaxed if one were content with numerical solutions.