research

Martin Schneider, June 2002.

Abstract: This paper studies international portfolio flows of US investors to examine the

information structure of international equity markets. Based on a model of portfolio choice with both public

and private information, we propose new empirical measures of trades due to private information.

We show that these trades help explain the cross section of international equity returns, after controlling

for public information. We find that such trades are highly correlated across countries. In particular, a common

"global" factor accounts for about half of the variation in trades due to private information and can also

be used to predict returns in many countries. The finding that a substantial portion of trades due to private

information across countries contains the same common information challenges the conventional view that

domestic investors have better private information about their home market than foreign investors.

Abstract: This paper analyses the behavior of the foreign exchange risk premium using longhorizon

regressions. A long‑horizon analysis may provide new evidence about three key issues concerning the foreign

exchange risk premium. The first issue is the relationship between expected foreign exchange returns and

expected returns on other assets.Long‑horizon regressions are able to explain a large portion of the variation

of foreign exchange returns using instruments that have been shown to predict domestic asset returns. I

undertake a careful small‑sample study to examine the size of the statistics and provide evidence of increased

power. The second issue is the high variability of the foreign exchange risk premium. Using the long‑horizon

regression results, I show that Fama's (1984) finding that the variability of the risk premium is greater than

that of the expected change in the spot rate holds even for horizons extending out to four years and is,

therefore, not the result of market frictions which would bind only in the short run. The third issue is the

economic model that can explain foreign exchange risk premia. I show that both the length of the holding

period and the inclusion of global and local risk factors are important for tests of latent variable models.

Abstract: How investors should hedge the currency risk component of their international stock portfolios

is the subject of much debate. In this paper, we add to the existing literature in a number of ways. First, we

focus on conditional portfolio and hedging decisions by using a recursive vector autoregression model to

identify a shock arising in the money, equity or currency markets. We treat all of the international markets

as integrated and examine how a shock in one asset market is propagated across several goods and asset

markets. The results reveal that money and equity market shocks cause long‑run movements in both equity

and currency returns. Second, we examine how an investor can use this joint predictability to design optimal

international stock portfolios and currency hedges. The approach is Bayesian so the investor accounts for

the large amount of parameter uncertainty while incorporating his beliefs about international asset return

predictability. Despite having priors that are strongly weighted against the predictability of both equity and

currency returns, asset market shocks cause the investor to shift his portfolio allocation and hedging strategy in

order to maximize his expected utility. Third, we isolate the economic significance of the predictability of

currency returns from that of equity returns by having the investor manage a currency overlay portfolio.

We use this to extend the results of Eichenbaum and Evans (1995) on long‑run currency returns and show

that the investor values his ability to condition on equity market shocks as much as he values conditioning on

money market shocks

03/03/2006