Essays on international finance.
Chapter one, analyzes the effect of both direct and indirect barriers on international asset pricing using evidence from the price-net asset value ratio of closed-end country funds. I find that, contrary to conventional beliefs, capital controls (direct barriers) do not necessarily result in premium in the country fund prices. In fact, in the presence of strong capital controls, significant discounts are not only possible, but frequent. I show that the premium or discount largely depends on indirect investment barriers, which include economic and political instability, liquidity risk, availability of market information, reliability of accounting standards and investor protection. These results suggest that both direct and indirect barriers have strong effects on international capital market integration. Chapter two, develops a new approach to measure the market integration of a number of developed and emerging countries with the United States. The approach used here avoids the need to rely on a specific asset pricing model by exploiting the price differential between closed-end fund prices and their net asset values. It also avoids having to identify and quantify all direct and indirect barriers by establishing a link between investment barriers, international equity flows and country fund premiums. The measure is applied to seventeen countries, seven of which are developed and ten emerging. Overall emerging markets tend to be more segmented than developed markets, but they exhibit a clear trend towards integration through time. However, emerging markets like Mexico and Portugal appear to be well integrated with the US. Chapter three incorporates transaction costs in the estimation of volatility bounds on Intertemporal Marginal Rates of Substitution (IMRS) implied by excess returns in international investments. We find that even small transaction costs of an average magnitude of less than 0.1% in the foreign exchange returns, are enough to dramatically reduce the required volatility of IMRS implied by the data. Similar results prevail for the equity excess returns. The estimated bounds are consistent with the bounds implied by typical representative agent asset pricing models. Furthermore, the empirical results suggest that incorporating conditioning information is not important when transaction costs are also incorporated.