The author explains the large differences in real interest rates across developed economies and shows that differences in the size of economies explain the cross-sectional variation in currency returns. Larger economies have permanently lower real and nominal interest rates than smaller economies, which causes persistent violation of uncovered interest parity (UIP). Stocks in the nontraded sector of larger economies produce lower expected returns compared with those of smaller economies, and the introduction of a currency union lowers the risk-free interest rates and stock returns in the nontraded sector of participating economies.

To examine differences in real interest rates, the author develops an international asset pricing model in which the correlation structure of exchange rates and asset prices arises endogenously in general equilibrium. The model shows that, because very little country-specific risk of larger economies can be diversified internationally, the investors based in larger countries must have a more volatile marginal utility of consumption compared with investors based in smaller countries.

The implication of this insight is that, in a larger economy, assets that make a payment that is partially fixed in terms of the larger country’s consumption bundle are a better hedge against consumption risk and thus must produce lower expected returns. Two such assets are risk-free bonds (e.g., inflation-linked bonds) and stocks in the nontraded sector. Consequently, larger countries should have lower risk-free interest rates, and the stocks in their nontraded sectors should produce lower expected returns.

The author expands the model to larger currency areas and again demonstrates a positive correlation between the real exchange rate of larger currency areas and the marginal utility from traded goods of active households. This finding implies that all the assets that make payments that are partially fixed in terms of the larger currency area’s currency are better hedges against consumption risk. The author concludes that the introduction of a currency union should, therefore, lower risk-free and nominal interest rates and the expected returns on stocks in the nontraded sector within participating countries.

In the author’s international asset pricing model, the stochastic properties of the real exchange rate are an endogenous function of real and monetary shocks. His deviation from standard approaches to modeling exchange rates is that he allows countries to differ in the size of their economies.

The sample consists of quarterly data from Organisation for Economic Co-Operation and Development (OECD) countries from 1980 to 2007. The author focuses on developed countries because of their large share of world wealth, reasonably open financial markets throughout the period, and availability of high-quality data. Consequently, the variation in the size of economies is among OECD members.

The basic prediction of the model is that differences in expected returns on international assets should be explained by differences in the covariances of the returns on these assets with the marginal utility of households that trade in financial markets. The model shows that these differences in covariance are a function of differences in the size of the countries, the size of the currency areas, and the volatility of the shocks affecting each country.

For the model, the author uses proxies for independent variables. The proxy for a country’s size is its economy’s share of OECD GDP. He treats the eurozone as a single economy after the introduction of the euro in 1998. The variance of the bilateral exchange rate to the US dollar is used as a proxy for the volatility of shocks affecting different countries.

The author provides a very simplistic explanation for persistent deviations from UIP. He uses basic economic principles and develops a model that shows that differences in country size and currency union size explain a significant portion of the variation in currency returns. The findings of this study, in conjunction with past or future studies, can be used to document the reasons behind persistent deviations from UIP.