بهره وری جریان اطلاعات بین المللی: شواهدی از ETF و قیمت CEF

While similar in their trading and organization, closed-end funds (CEFs) and exchange-traded funds (ETFs) differ in their liquidity and ease of arbitrage. We compare their price transmission dynamics using a sample of funds that invest in foreign securities and are most likely to show the deficiencies in the manner in which they process information. Our analysis shows that ETF returns are more closely related to their portfolio returns than are CEF returns. However, both fund types underreact to portfolio returns but overreact to domestic stock market returns. A simple trading strategy using these results is profitable with roundtrip trading costs less than 1.38% for CEFs and 0.71% for ETFs.

As globalization advances, the flow of information across countries becomes increasingly important. This is particularly true in securities markets, and the process by which stocks incorporate this international information flow is complicated. Since the world's securities markets are not open at the same time, price changes in one market are not reflected in another market's trading until it opens. Furthermore, arbitrage, which aligns prices of similar goods in different markets, is difficult to conduct across countries and regulators. Previous studies have analyzed the international flow of information in the market for American Depository Receipts (ADRs). Kato, Linn, and Schallheim (1991) find that the ADR market does not offer persistent arbitrage opportunities. In contrast, Kim, Szakmary, and Mathur (2000) find that ADRs overreact to changes in the U.S. stock market and underreact to changes in the underlying stock prices. While these studies of ADRs offer insight into the international linkages of stock markets, several characteristics of these securities make them unique examples of how prices are transmitted across nations. First, large price differences in the ADR market can be arbitraged away by the creation or cancellation of these securities. This mechanism for the arbitrage of prices across countries does not exist for the vast majority of securities. Second, it is difficult to tell if investors are trading ADRs to profit from trends in the company or the country. Finally, institutional investors, who are generally considered to be relatively sophisticated, often engage in a substantial portion of the trading in these securities. Thus, the prices of ADRs may not be as sensitive to the sentiments of individual investors. While prior studies have examined the price transmission dynamics of ADRs, little is known about how daily price changes from foreign markets affect the values of closed-end country funds (CEFs) and exchange-traded funds (ETFs), which have different characteristics than ADRs. Unlike ADRs, the shares of CEFs cannot be easily arbitraged. As these funds hold a diversified portfolio of assets, the trading activity in these funds is less likely to reflect company-specific information than the trading in ADRs. While previous studies on country funds have used weekly data on NAVs, this weekly data is not of sufficient frequency to provided insight into the short-term transmission of information across markets. An ETF is another type of investment company that is commonly used for international diversification. Like CEFs, ETFs own a portfolio of securities and trade on an exchange like ADRs. However, ETFs can be arbitraged through the fund distributor using an in-kind process. This process may cause ETF prices to quickly reflect the returns in foreign markets. We add to the literature by analyzing the transmission of price changes between ETF and CEF markets and their respective underlying portfolios. Our first objective is to use a vector autoregression model (VAR) to examine the daily dynamics of how fund prices respond to changes in underlying values. The impulse response functions from the VAR model reveal that shocks to NAV affect ETF prices for only two days at most but influence over three-fourths of the CEF prices for at least three days. Our second objective is to determine the sensitivity of fund prices to daily domestic stock returns. We find that the vast majority of CEF and ETF returns have a negative relation with either the one-day or two-day lagged domestic return. Finally, we examine whether ETFs process information more efficiently than CEFs. Despite the relative ease with which ETFs can be arbitraged, their returns underreact to portfolio returns and overreact to domestic stock returns like CEF and ADR returns. We examine the profitability of a trading strategy using these relationships and find that trading costs in ETFs need to be about half the level of transaction costs for CEFs to eliminate profitability.

Many investors desire exposure to a diversified portfolio of foreign stocks while limiting their trading to products on a domestic stock exchange. The two main vehicles for achieving this goal are CEFs and ETFs, but these investment products differ in their liquidity and ability to be arbitraged. Trading in ETFs is generally more active than in CEFs, so the trading costs of CEFs are likely higher. In contrast to CEFs, institutional investors can easily arbitrage ETF shares and minimize any differences between share prices and NAVs. To see how these differences affect the market for these funds, this study investigates the transmission dynamics between fund prices and the underlying portfolio values, exchange rates, and index returns in the foreign and domestic markets. The first conclusion from our analysis is that ETF returns are more closely related to their portfolio returns than are CEF returns. Innovations in the NAV explain 78% of the 5-day ahead forecast error variance for ETF share prices but only 54% of the forecast error variance for CEFs. The impulse response functions show that shocks to the NAV have a positive effect on fund prices and persist for at least three of the five days following the shock for 79% of the CEFs. The impulse response functions for the ETFs reveal that none of the ETF fund prices are affected by a NAV shock for more than two of the five days following the shock. The discrepancy in the way that ETFs and CEFs process information may be due to differences in the ease of arbitrage, market liquidity, or even ownership clienteles. Our second insight into the dynamics of fund prices is revealed from our regression analysis of fund prices. Both CEF and ETF prices underreact to NAV returns but overreact to domestic stock market returns. For 95% of the CEFs and 81% of the ETFs, share prices are positively related to the one-day lagged NAV returns. The two-day lagged NAV returns are a positive influence on 74% of the CEF prices and 13% of the ETF prices. As the prices are slow to reflect changes in NAV, we conclude the shares underreact to changes in underlying value. Both the CEFs and ETFs have a positive relation with the contemporaneous return on the S&P 500 Index. However, over two-thirds of CEF returns and 94% of the ETF returns have a negative relation with either the one-day or two-day lagged domestic return. This result is consistent with previous findings for the ADR market. As fund prices take days to fully reflect changes in NAV and the domestic stock market, we examine the profitability of a simple trading strategy and provide a third contribution to the literature. Using coefficients from only the lagged variables in our regression analysis, the strategy would be unprofitable with roundtrip trading costs of at least 1.38% for CEFs and 0.71% for ETFs. In the ADR market, roundtrip trading costs need to be 0.27% to eliminate the profit from a simple trading strategy (Kim et al., 2000). By some measures, the market for CEFs and ETFs that invest in foreign stocks is less efficient than the market for ADRs. Such inefficiencies may be particularly costly since these funds attract short-term investors who do not want to be restricted by the limited trading opportunities associated with mutual funds.