Finance in Europe
Modelling the Markets

A joint CEPR workshop with the Limburg Institute of Financial Economics (LIFE) on international finance took place in Maastricht on 2/3 June. The workshop was organized by Kees Koedijk (Universiteit van Limburg) and Christian Wolff (Universiteit van Limburg and CEPR). It formed part of CEPR's research programme on `Finance in Europe: Markets, Instruments and Institutions', funded by the European Commission's Human Capital and Mobility programme.
The first paper was `Why is There a Home Bias? An Analysis of Foreign Portfolio Equity Ownership in Japan' by Jun-Koo Kang (University of Rhode Island) and René Stulz (Ohio State University). This paper uses data on foreign stock ownership in Japan between 1975–91 to examine the determinants of the home bias in portfolio holdings. Existing models of international portfolio choice that yield a home bias are inconsistent with the evidence provided here: foreign investors are overweight in shares of firms in manufacturing industries, large firms, firms with good accounting performance, firms with low unsystematic risk and firms with low leverage. Consequently, when investors hold foreign shares, they do not hold the market portfolio of foreign countries. Foreign investors do not perform significantly worse than if they had held the Japanese market portfolio, however. After controlling for firm size, there is no evidence that foreign ownership helps understand the cross-sectional variation in expected returns. The paper shows that a model with size-based informational asymmetries and deadweight costs can yield asset allocations consistent with its evidence.
The next paper was `On Casselian PPP, Cointegration, and Exchange Rate Forecasting' by Ronald MacDonald and Ian Marsh (University of Strathclyde). Using an expanded version of the purchasing power parity condition, this paper constructs simultaneous equation models for three key exchange rates, incorporating meaningful long-run equilibrium relationships and complex short-run dynamics. The paper shows that fully dynamic forecasts from these models are capable of significantly outperforming those of a random walk model over horizons as short as three months, and that they are also more accurate than the vast majority of professional forecasts.
`Multiple-country versus Single-country Tests of International Asset Pricing Models' by Maria Vassalou (Universiteit van Limburg) examined the presence of exchange rate and inflation risk premia in the context of three alternative, but nested, international asset pricing models. The paper shows that single country tests are inappropriate for the estimation of exchange rate and inflation risk premia, due to insufficient cross-sectional variation in betas. Exchange rate betas, and also to some extent world inflation betas, possess a strong country factor. As a result, appropriately structured multiple country tests do not suffer from this problem. In contrast to previous studies, the hypothesis of a zero unconditional exchange rate premium is strongly rejected.
In `Selecting Models to Forecast Financial Returns: A New Criterion', by Peter Bossaerts (Tilburg University) and Pierre Hillion (INSEAD), the poor out-of-sample predictability of asset returns was reinvestigated by applying standard model selection criteria on a data set of monthly excess stock returns from 15 countries. The paper's criteria are designed to avoid model overfitting and the source of the remaining bias is evaluated by comparing the results to those from the application of a new selection criterion. This criterion, an extension of Rissanen's Predictive Least Squares criterion, is bias-free if prediction models for asset returns are stationary. In contrast, it substantially overfits if non-stationarity is the cause behind the remaining bias of standard selection criteria. The paper finds the latter to be the case. It underscores the importance of the development of prediction models and tests of asset pricing models that accommodate non-stationarities.
`A Segmented International Capital Markets Model and its Implications for Corporate Finance' by Ian Cooper and Evi Kaplanis (London Business School) extended the literature on international capital market equilibrium and corporate finance. The paper modifies the Stulz (1981) model to include an arbitrary number of assets and countries and an arbitrary structure of investment costs. It then derives optimal capital budgeting rules in segmented international capital markets. The segmentation is generated by the deadweight costs of international portfolio investment. The paper shows how capital budgeting rules and the capital market risk premium depend on the level of costs of cross-border investment. Optimal capital budgeting rules therefore depend on the nationality of the company making the investment. Direct international investment results in improved diversification for investors and this is reflected in discount rates. The net desirability of direct investment depends on three factors. Costs of cross-border investment by investors affect the capital market equilibrium and, therefore, the required rates of return for domestic and foreign companies. Deadweight costs for corporations, such as taxes on remittances, and the comparative operating advantage or disadvantage of foreign companies affect the net cash flow to shareholders of foreign companies relative to domestic companies. The balance of these three factors determines the viability of foreign direct investment.
The last paper was `Time-varying Risk and International Portfolio Diversification with Contagious Bear Markets' by Giorgio De Santis and Bruno Gerard (University of Southern California). This paper estimates and tests a conditional version of the international CAPM. It allows risk premia, betas and correlations to vary through time and tests the cross-section restrictions of the model using a relatively large number of assets. In support of the international CAPM, the paper finds that worldwide risk is priced whereas country-specific risk is not. Furthermore, the price of world risk is time-varying and has a strong January seasonal. When the price of risk is allowed to vary, a January dummy and the world dividend yield are driven out as independently-priced factors. But contrary to the model's prediction, differences in risk premia across countries are explained not only by worldwide risk, but also by a constant country-specific factor. The estimated correlations reveal three main facts: cross-country correlations vary through time; they have been affected only to a limited extent by the process of liberalization of the last decade; and they tend to increase during severe bear markets in the US. International correlations are, however, smaller than correlations among US assets. Thus, investors gain from global diversification even with contagious bear markets.