This paper shows that the simple fact that some economies are larger than others has important implications for international asset returns, and that acknowledging this fact may help explain a number of otherwise puzzling features of the data. In particular, large economies should have lower real and nominal interest rates than smaller economies, and uncovered interest parity should fail between the currencies of any two economies that are not of the same size.
The intuition for this result is simple: Think of a world which consists of two economies; a large economy (call it the US) and a small economy (call it Australia). A risk&free bond from the perspective of a US consumer is an asset that promises to deliver one unit of the US consumption bundle; and a risk free bond from the perspective of an Australian consumer is an asset that promises to deliver one unit of the Australian consumption bundle. The risk free bonds of the two economies are thus country specific consumption insurance: If I issue a risk free bond denominated in an economyqs consumption bundle, I am offering insurance against shocks that affect consumers who live in that economy. Intuitively, it has to be more expensive to insure against shocks that affect a larger fraction of the world population at the same time. The US risk free bond must therefore be more expensive than the Australian risk free bond, which implies that it pays a lower expected return and that the US risk free rate is lower than the Australian risk free rate.
The logic underlying this result is very powerful: It holds for risk due to real shocks (rlarge countries have lower interest ratess) and for risk due to nominal shocks (rlarge currency areas have lower interest ratess). Moreover, the same forces that make it more expensive to provide insurance against shocks that affect US consumers, make it attractive to hold stocks in the non traded sector of the US economy: Stocks in the non traded sector of large economies provide implicit insurance against shocks affecting their consumers and should thus pay lower expected returns than those of smaller economies.
Real Model In the baseline model, asset markets are complete and households in each country receive stochastic endowments of a traded and of a non traded consumption good. Since only the traded good can be shipped internationally, all international assets have to be settled in terms of this traded good. It follows that assets in this economy are priced with the marginal utility of traded consumption, which is the same for all households in equilibrium. This marginal utility is high when times are bad, i.e. when the average world endowments of either traded or non&traded goods are low. Investors are thus willing to pay more for assets that pay off well in terms of traded goods when world endowments in the two sectors are low. The crucial insight is that a given percentage change in the endowment of a large country has a larger impact on marginal utility than the same change in the endowment of a small country: Bad times in a large country are more likely bad times for the average world investor than bad times in a small country.
Download
Country Size, Currency Unions, and International Asset Returns
