|
Article Excerpt
THE LAST 15 years have recorded sizable and unprecedented current account deficits run by the United States accompanied by a gradual deterioration of the U.S. net international investment position that reached -22% of GDP in the year 2005 and has improved to -17% in 2007. Almost three-quarters of the world's surpluses are absorbed by the U.S. deficit. As documented by Lane and Milesi-Ferretti (2002), these developments have been paralleled by an increase in international financial diversification through instruments of different risk and liquidity characteristics. For the United States both assets and liabilities have increased up to 128% and 145% of GDP. These developments are usually welcomed for the gains that arise because of more integrated financial markets. Still, net negative positions in the international markets matter and global imbalances might have important negative macroeconomic consequences.
This paper studies whether the international monetary system can be affected by the presence of large asymmetries in the positions in international financial markets, i.e., the fact that some countries are large debtors while others are creditors. An important channel that will be explored is the interaction between international risk sharing and the stabilization role of monetary policy. In an important paper Obstfeld and Rogoff (2002) have shown that in a distorted economy with lack of full international risk-sharing self-oriented policies that achieve price stability in each country can replicate the cooperative outcome. The spillovers that monetary policymakers have on the risk-sharing margin are of second-order importance. This paper readdresses this issue in a two-country dynamic model that solves for the optimal cooperative monetary policy when countries have nonzero, but specular, positions in the international financial markets.
The understanding of whether there should be deviations from a policy of price stability at the international level goes parallel with the analysis of the costs of market incompleteness. The main finding is that the welfare costs of incomplete markets are increasing with the cross-country asymmetries in the initial net international positions and in particular they become nonnegligible when the persistence of the shocks increases. In the baseline scenario they are smaller than 0.20% of a permanent increase in steady-state consumption and they increase up to 1% with the persistence of the shocks. In these cases there are also important gains of deviating from a policy of price stability, above 0.2%.
Whereas optimal monetary policy requires a modest increase in the volatility of the producer-price inflation rates, the important adjustment should come through an increase in the volatility of real returns on assets traded. This is mostly reflected in an increase in the volatility of the nominal interest rates in both countries. Indeed, appropriate movements in the asset returns and so valuation effects can correct asymmetries in the business cycle synchronization improving risk sharing. Moreover, optimal monetary policy--in the calibrated example--requires more synchronization of the cross-country nominal interest rates when global imbalances increase. Instead, a policy of price stability commands a mildly positive correlation which is independent of the size of the global imbalances.
The welfare costs of incomplete markets and the optimal monetary policy regime are analyzed using a linear-quadratic solution method in a two-country model in which two bonds, issued in different currencies, are traded. Benigno and Woodford (2006) have shown that linear-quadratic solution methods are appropriate, as a first-order approximation of the optimal solution, for a general class of models. One important exception is the case in which the zero-order approximation is indeterminate, which turns out to be the relevant case for portfolio shares when portfolio choices are considered. In this paper, this problem is resolved by assuming the existence of transaction frictions in trading the two bonds as in Ghironi, Lee, and Rebucci (2006). However, Ghironi, Lee, and Rebucci focus on the case in which only equities are traded and do not consider optimal policies. Without trading frictions, a zero-order approximation will not determine portfolio shares which are instead going to be function of the ratio of second-order moments, as discussed in Devereux and Sutherland (2006). However, in this case, a linear-quadratic solution method, as an approximation of the optimal policy, would not be feasible. (1)
The detailed structure of the work is as follows. Section 1 presents the model. Section 2 discusses the steady state of the model and Section 3 shows the log-linear approximation of the equilbrium conditions. Section 4 presents the welfare criterion while Section 5 computes the welfare costs of incomplete markets under a policy of price stability and the gains obtained by pursuing the optimal cooperative policy, assuming zero initial foreign asset holdings. Finally, Section 6 concludes.
1. THE MODEL
The model belongs to the class of dynamic stochastic general equilibrium models that have been used for the evaluation of monetary policy both in the closed and open-economy literature. The important novelty is the treatment of an incomplete-market asset structure that can be directly compared to the complete-market one used in the literature (see among others Benigno and Benigno 2003, 2006, Corsetti and Pesenti 2006, Devereux and Engel 2003, Kollmann 2002, Obstfeld and Rogoff 2002). Our utility-based welfare criterion can also allow for a direct evaluation of the welfare costs of imperfect risk sharing, with particular emphasis on the different assumptions on the structural parameters of the model, the nature of the shocks--whether supply or demand--and the role of monetary policy. (2)
We consider a world with two countries, home (H) and foreign (F). The population on the segment [0, n] belongs to country H while the population on the segment (n, 1] belongs to country F. In each country, a continuum of differentiated goods is produced with measure equal to the population size. The utility of a generic consumer j belonging to country H is
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII],
where [E.sub.0] denotes the expectation conditional on the information set at date 0, while [beta] is the intertemporal discount factor, with < [beta] < 1.
Households enjoy utility from goods consumption, while they receive disutility from producing goods. The utility function is separable in these two factors. Moreover, in each country, a generic household contributes to the production of all the goods with a separable disutility. (3) g is a preference shock, whereas z is a productivity shock. These shocks are country specific. With starred variables we denote country's F variables.
The consumption index [C.sup.j] is defined as follows:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII], (1)
where [C.sup.j.sub.H] and [C.sup.j.sub.F] are consumption indexes of the continuum of differentiated goods produced, respectively, in countries H and F,
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]. (2)
The lasticity of substitution across goods produced within a country is denoted by [sigma], which is assumed greater than one, while the elasticity of substitution between the bundles [C.sub.H] and [C.sub.F] is [theta].
We assume that all the goods are traded and that the law-of-one-price holds. We further assume that the same composition of the consumption bundle C applies to country F. Given these assumptions, it follows that purchasing power parity holds, i.e., P = [SP.sup.*], [P.sub.H] = [SP.sup.*.sub.H] and [P.sub.F] = [SP.sup.*.sub.F], where S is the nominal exchange rate. Here we define the relative price T, the terms of trade, as T [equivalent to] [P.sub.F]/[P.sub.H].
The household j's demands of a generic good h, produced in country H, and of the generic good f, produced in country F, are
[c.sup.j](h) = [(p(h)/[P.sub.H]).sup.-[sigma]] [([P.sub.H]/P).sup.-[sigma]] [C.sup.j], [c.sup.j](f) = [(p(f)/[P.sub.F]).sup.-[sigma]] [([P.sub.F]/P).sup.-[sigma]] [C.sup.j]. (3)
Aggregating across all households in the world economy, we can write total demands of good h and f as
[y.sup.d](h) = [(p(h)/[P.sub.H]).sup.-[sigma]] [([P.sub.H]/P).sup.-[sigma]] [C.sup.W], [y.sup.d](f) = [(p(f)/[P.sub.F]).sup.-[sigma]] [([P.sub.F]/P).sup.-[sigma]] [C.sup.W]. (4)
where world consumption [C.sup.W] is defined as
[C.sup.W] [equivalent to] [[integral].sup.1.sub.0] [C.sup.j]dj.
We assume that there are two bonds traded internationally: one is denominated in the currency of country H and the other in the currency of country F. Both bonds are risk-free with one-period maturity. Thus, the budget constraint of household j in country H (expressed in real terms with respect to the consumption-based price index) is
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII],
at each date t. [A.sup.j.sub.t] is household j's holding of the risk-free one-period nominal bond, denominated in units of currency F. The nominal interest rate on this bond is [i.sup.*.sub.t]. [B.sup.j.sub.t] is household j's debt issued in units of the risk-free one-period nominal bond denominated in currency H. The nominal interest rate on this bond is [i.sub.t]. We are assuming that households of country H hold assets denominated in foreign currency and issue debt in domestic currency, which reflects the current net international position of the U.S. economy. (4) Most important, we are assuming...
|
