Aftermarket Performance of Closed-End Funds Invested in International versus Domestic Securities

Transcription

1 Aftermarket Performance of Closed-End Funds Invested in International versus Domestic Securities Lena Chua Booth and Hassan ehranian Do funds holding less liquid assets benefit more than others by a closed-end structure? If so, closed-end funds invested in typically less liquid international securities are expected to outperform funds invested in domestic securities. ests that control for fund characteristics and fund type indicate that only smaller or lower priced funds earn significantly higher first-day returns than domestic funds. here is no statistically significant difference in long-term performance, both unadjusted and index-adjusted, between funds invested in international versus domestic securities, except for larger funds. Larger international funds earn significantly lower 36-month index-adjusted returns than domestic funds. [G20, G23] Closed-end funds often trade at a persistent discount, meaning their share prices are lower than their net asset values. Studies that attempt to explain premiums and discounts examine issues such as: the taxation system and timing, ownership structure, liquidity, manager fees and expenses, managerial performance, agency problems between shareholders and managers, market segmentation, investor sentiment, and costs of arbitrage (Dimson and Minio-Kozerski, 1999, provide a summary). Few researchers examine the aftermarket and long-term share performance of closed-end funds from the time they go public, and whether performance is related to fund strategy, Weiss (1989), Peavy (1990), and Levis and homas (1995) focus primarily on initial and short-run returns of closed-end funds and compare them to industrial IPOs. hey use small Lena Chua Booth is an Associate Professor of Einance ai hunderbird, he Garvin School of International Management, in Glendale, AZ Hassan ehranian is a Professor of Einance at Boston College in Chestnut Hill. MA We wouid like to thank the referee for his valuable and helpful comments on prior drafts. Special thanks also to the participants of the 2003 EMA Annual Conference, the discussant Sahiba Chadha, and James Booth for their valuable suggestions. We also would like to thank Lipper Analytical Services for providing part of the data for this study. hanks also to Center for International Business Education and Research at hunderbird for providing research funding to the first author for this project. All remaining errors are ours. samples and document no more than one year of aftermarket share performance. We compare the aftermarket performance of closedend funds invested in domestic and international securities from inception through three years of trading. Focusing on fund performance from inception allows us to examine both initial performance and also the seasoning process of share returns. We also use a larger sample than available previously because of the popularity of closed-end funds in the 1990s. Fama and Jensen (1983a, 1983 b) have speculated that some funds benefit more from a closed-end form than others. hey argue that redeemable shares are not efficient when assets are less liquid and prices are difficult to observe. ransaction costs in liquidating these portfolio securities to meet redemptions can be extremely high. Constant liquidation of assets with prices that are difficult to observe can produce conflicts of interest for investors with different horizons. his implies that portfolios with less liquid assets with difficult-to-observe prices are more efficiently held in closed-end form. Deli and Varma (2002) provide some empirical evidence that funds holding less liquid securities with less transparent prices are more likely to be closed-end than open-end. As investors require a premium for holding illiquid assets (Amihud and Mendelson, 1986), closed-end funds holding less liquid assets would thus be expected to perform better in the aftermarket than funds 24

2 BOOH & EHRANIAN AFERMARKE PERFORMANCE OF CLOSED-END FUNDS 25 holding more liquid assets. As funds invested in intemationai securities are typically less liquid and have prices that are difficult to observe, we would expect them to perform better than funds invested in domestic securities. Our empirical analysis considers a sample of 352 closed-end funds that went public between 1986 and We compare the aftermarket share returns of funds invested in intemationai and domestic securities, controlling for fund characteristics such as offer price, issue proceeds, underwriter ranking, fund type, and investment strategy. We document the initial and longterm share returns for different types of funds (such as stock, bond, and mixed funds) and with different investment strategies (such as those investing in international, emerging market, and country securities), and then compare the performance of funds invested in international and domestic securities across different size and offer price groups. We find significantly higher initial retums for closedend funds invested in intemationai securities than in domestic securities. his result persists across different size funds, in funds with lower offer prices, and in funds with more prominent underwriters. he difference in initial retums for funds with larger issue proceeds or higher offer prices diminishes after controlling for fund characteristics and type of fund. hree-year post-ipo share retums (unadjusted and index-adjusted), however, are not significantly different for funds invested in international versus domestic securities, except for funds with larger issue proceeds. In this case, funds invested in international securities perform significantly more poorly than funds invested in domestic securities after three years of trading, both unadjusted and index-adjusted. For the index-adjusted retums, the difference holds even after controlling for fund characteristics and type of fund. I. Closed-End Funds and Share Performance Closed-end funds might be compared to nonfinancial business corporations, except that fund managers invest in securities of other corporations and manage the investment, while non-financial corporations make real, economic investments. he capitalization of closed-end funds is fixed after the initial offering. New investors can purchase the fund only when current investors sell. hus, a closed-end fund price is a function of the supply of and demand for its shares trading in the secondary market. While almost all industrial firms are listed on Nasdaq when they go public, more than 80% of closed-end fund IPOs are listed on the New York Stock Exchange.' 'Eighty-nine percent of the closed-end fund IPOs in our sample are listed on New York Stock Exchange, 8% on the American Stock Exchange, and 3% on the Nasdaq. Closed-end funds are brought to the market through an IPO process similar to an industrial IPO, but only a small percentage of their shares are purchased by institutional investors. Weiss (1989) reports that only 3.5% ofthe closed-end funds shares were held by institutional investors after the first quarter of trading, compared to 21.8% in industrial IPOs during the same period. A. Closed-End Fund Share s Several authors have examined the behavior of closed-end fund IPOs. hey generally focus on the differences between IPOs of closed-end investment companies and IPOs of non-financial corporations. Studies by Weiss (1989), and Peavy (1990) consider samples of IPOs that came to market in 1986 and his was when closed-end funds began to reemerge after a drought of new funds between 1975 and In sample sizes of between 40 and 64 new funds, they document that initial retums of these funds were on average either zero or negative, compared to the significant positive retums of the IPOs of industrial firms on average. Weiss (1989) finds a positive, about one-half of 1% average initial retum for a sample of 64 closed-end funds going public between 1985 and 1987, but finds a negative longer-term performance. Ritter (1987) reports that IPOs of industrial firms provided a cumulative market-adjusted retum of 18.5% after five months of trading, compared to -12.6% for closed-end funds reported by Weiss (1989). Peavy (1990) finds almost identical results for a very similar sample of funds (51 issues) in the same time period. Peavy and Weiss were the first to document negative performance for closed-end funds in the 1980s. he former suggests that new issues of funds were likely to disappear when investors discovered the poor performance. Levis and homas (1995) provide evidence in the UK that initial retums on a sample of closed-end investment trusts are a positive 1.91%. Over a longer period of 200 trading days, the retums drop to -4.63% unadjusted and -8.48% adjusted. Hanley, Weiss, Lee, and Seguin (1996) attempt to document how closed-end funds drop in price by examining the microstmcture data for the first 100 days of trading. hey find order imbalances in that most of the trades are sell orders; the larger trades, that is, are sell orders, and the smaller trades are buy orders. In the late 1980s, the majority of issues experienced price stabilization in the first several weeks. hey suggest that the substantial price decline after the IPO is a result of professionals selling to less informed investors. his hypothesis is consistent with evidence that only small investors in the US hold these shares in the long run, contrary to the predominantly institutional ownership observed in the UK (see

3 26 JOURNAL OF APPLIED FINANCE SPRING/SUMMER 2005 Dimson and Minio-Kozerski, 1999). Studies using data that end in the mid-1980s have access to data for only a few funds, typically around 50 funds, with around $8 billion total under management. Since 1986, the number and the variety of IPOs of closed-end funds issued in the US have grown dramatically. At least 600 new funds have been created with over $160 billion of funds under management. Growth has been seen in all types of funds, but growth is particularly dramatic in international funds, especially emerging markets, and municipal bond funds. his larger sample of new funds enables more precise measurement of regression parameters. B. Model Predictions he number of shares for a closed-end fund is fixed at the initial public offerings, so investors who want to redeem shares can only do so by selling in the secondary market. Fama and Jensen (1983a, 1983b) suggest that this restriction allows fund managers to: 1) mitigate the transaction costs in liquidating portfolio securities to meet redemptions, and 2) reduce conflicts of interest among investors with different horizons. ransaction costs and conflicts of interest can be substantial when assets of a fund are costly to liquidate. It is inefficient to redeem shares when assets lack liquidity and prices are difficult to observe. his implies that closed-end funds may be more attractive for investment strategies that involve holding less liquid assets, such as international securities. We might thus expect funds invested in international securities to outperform funds invested in US domestic securities, both initially and in the long run.^ II. Sample and Empirical Method Our initial sample comprises 516 closed-end fiinds that went public over , with data provided by Securities Data Corporation (SDC), now owned by homson Financial. All these funds are listed on US stock exchanges. SDC databases of new issues provide information on offer price, date of offering, issue proceeds, first-day initial returns, and fund investment type.' A. Data We classify the funds by their description into four different types: bond, stock, mixed, and municipal bond funds. Bond funds invest only in bonds, whether ^A caveat is that particular fund strategies vary over time depending on market conditions. hus, a closed-end fund sample may be influenced by hot issue considerations that may affect performance regardless of fund strategy. reports first-day returns for only 417 of the 516 funds. corporate or government. Stock funds invest exclusively in stocks and preferred stocks. Mixed funds invest in equity and fixed-income securities. Municipal bond funds invest exclusively in municipal bonds in the US. We also classify the funds into those invested in domestic (US) and foreign securities. Domestic funds invest exclusively in the US market. International funds invest in foreign markets only or in a combination of foreign and US markets.'' We categorize international funds as emerging or developed market funds. Emerging market funds invest exclusively in emerging markets. Developed market funds invest in either developed markets or sometimes in a combination of emerging markets and developed markets. One category of international funds called country funds invests in securities fi-om either a foreign country or a foreign region. Examples are aiwan Fund and Greater China Fund. he initial sample of 516 funds is reduced by the availability of share price data and net asset values (NAVs) after the initial public offerings. he monthly share price and the dividends of the closed-end funds established during are made available by Lipper Analytical Services. For the additional years of data not provided by Lipper, we use the Center for Research in Security Prices (CRSP) to obtain monthly share price returns. We eliminated funds with no trading data available in CRSP or Lipper, or closedend funds without at least three years of share price data after they go public. he final sample consists of 352 funds that went public between 1986 andl998, 95 international funds and 257 domestic funds. hey include 106 bond, 65 stock, 41 mixed, and 140 municipal bond funds. We adopt the measures of lead underwriter reputation developed by Carter and Manaster (1990) and Carter, Dark, and Singh (1998). his widely used measure in IPOs studies is based on relative positions in the tombstone advertisements, with a highest ranking of 9. Closed-end fund IPOs are underwritten mainly by prominent investment banks. he average underwriter ranking in our sample is 8.64 (median 8.83). B. Research iviethod o examine the post-offering price performance of closed-end funds for the different types of funds, we use data that match the SDC, Lipper, and CRSP databases. o adjust returns for market movements, we collect monthly S&P 500 returns from CRSR We also collected the monthly corporate bond returns from he majority of the funds that invest in a combination of foreign and US markets invest a very small percentage of securities in the US market. Hence their returns are driven mainly by the returns of foreign securities.

4 BOOH & EHRANIAN AFERMARKE PERFORMANCE OF CLOSED-END FUNDS 27 Ibbotson Associates Yearbook.' he monthly unadjusted return for each fund month t is calculated as: m 'A (4) Pi, + Di,-Pi,-i (1) where t= l,...,36;i= 1,...,352; P.^ and P.,, are the prices for month t and t-1 and D.^ is the dividend paid by fund i in period t. he long-run aftermarket performance is calculated following a buy-and-hold methodology used by Carter et al. (1998). he unadjusted returns are cumulated over periods starting from the first month after the IPO as follows: (2) where CUR.j. is the cumulative unadjusted return for fund i over months. o account for market movements in the same period, we calculate the cumulative index-adjusted returns for fund i over months (CAR..^,) as follows: where I^ is the returns on the index in month t. Peavy (1990) notes that the portfolio composition of new funds is changing and unknown, and it is unclear how to adjust returns for risk. For this reason, we compare all stock and mixed funds to the S&P 500 and all the municipal and bond funds to the corporate bond index. While the natural choice of benchmarks should be the world index or the native index ofthe specific international fund, we use only US stock and bond index because Chang, Eun, and Kolodny (1995) and Pennathur, Delcoure, and Anderson (2002) show that the returns of international funds tend to behave more like US domestic securities.' he cross-sectional average cumulative indexadjusted return is calculated as follows:,=\ where N is the number of funds with complete data. III. Empirical Results Exhibit 1 provides summary statistics on fund characteristics, underwriter ranking, and aftermarket price performance each year for a sample of 352 closedend fund IPOs that went public over he overall initial-day return is a positive 0.89%. his is statistically significantly different from zero, given the relatively tight distribution of returns around zero (63% of IPOs of closed-end funds have a return of exactly zero on the first day). his result differs from those presented by Weiss (1989) and Peavy (1990) that the first day return is either zero or negative. heir samples consist of closed-end funds that went public from 1986 to 1987; ours is much larger and for a much longer period, including more recent issues. he initial-day return for closed-end fund IPOs is also very different from industrial firm IPOs whose average usually exceeds 15%. i he overall 36-month average cumulative unadjusted share return is 10.78%. he closed-end funds going public in 1995 commanding the highest average cumulative return of %.' 36-mon th unadjusted returns are positive every year except 1998, at only -5.49%. hirty six-month cumulative indexadjusted share return averages are negative each y^ar except 1986 and Most negative returns occurlin 1996 when two stock funds fared poorly compared to the S&P 5OO.«Exhibit 2 provides summary statistics of funjds characteristics and aftermarket performance categorized by type of fund. Of 352 closed-end funds, 140 are US municipal bond funds. Of 95 internatiot al funds, 46 are emerging market funds, and 56 ere country funds. Country funds account for about 6C % of the total closed-end funds that invest internationally. Only 22 stock funds invest exclusive ly in the US. he average offer price is $13.31 for US domes ic funds and $13.56 for international funds. he lowest average offer price is for domestic bond funds at $ 11.^ followed by domestic mixed funds, at $ 'Adjusting returns using CRSP value-weighted and equalweighted returns produces results similar to those adjusted using S&P returns. 'Choice of index does not affect the conclusion, as the end results are similar to those with unadjusted returns. 'Only two closed-end funds went public in 1995, and one earned a cumulative unadjusted return of 228%. he 1995 results do not affect our overall conclusions. 'he 36-month returns extend through 1999 when the S ScP 500 experienced extraordinary returns.

5 28 JOURNAL OF APPLIED FINANCE SPRING/SUMMER 2005 Exhibit 1. Yearly Summary Statistics of Closed-End Fund IPOs from 1986 through 1998 Offer price and proceeds are obtained from Securities Data Corp. Underwriter ranking is a measure of investment banker reputation found in Carter, Dark, and Singh (1998). his measure is based on the relative positions in the ombstone and has a highest ranking of 9. First day return is computed as: [{the closing price of fund on first day of trading - offer price) / offer price] * 100. unadjusted share return is computed as: r [FI (1 + ^/0-1]* 100 where Rj, is the return on month t for fund i, and =36 for the 36-month cumulative return. index-adjusted share return is the: [{ )][U{ )] where I, is the monthly return on the S&P 500 for month t if the fund is a stock or mixed fund, or the return on the corporate bond index if the fund is a bond or municipal bond fund. he fmal sample consists of 352 closed-end funds established during Year All N Offer Price ($) Proceeds ($m) Underwriter Rank First Day month Unadjusted Share month Index- Adjusted Share offer prices for international funds range from $12.72 to $ offer prices do not vary significantly by type of fund, whether domestic or international. proceeds vary more. proceeds are $ million for US domestic funds and $ million for international funds. International bond funds return the highest average proceeds of $ while country funds show the lowest average proceeds of $ underwriter ranking does not vary much across different types of funds. All the funds have average underwriter rankings between 8 and 9. he overall average ranking is Loughran and Ritter (2004) classify a ranking of 8 and above as prestigious. International funds outperform domestic funds in terms of first day aftermarket returns. first day returns are 0.30% for US domestic funds and 2.30% for international funds, and the difference is significant at the 0.01 level. International mixed funds earn 5.37% on the first day, the highest average first-day return among different types of funds. first-day returns for domestic fiinds range from -0.65% to 1.54%, compared to 0.07% to 5.37% for international funds. First-day domestic fund returns are generally lower than returns for industrial IPOs. hirty-six-month unadjusted cumulative share returns are positive for all types of funds but 36-month index-adjusted share returns are all negative. International bond funds fare the best during the sample period: 18.51% cumulative share returns unadjusted, and % cumulative share returns index-adjusted (the least negative). Despite the highest 36-month cumulative unadjusted returns, domestic

6 BOOH & EHRANIAN AFERMARKE PERFORMANCE OF CLOSED-END FUNDS Exhibit 2. Summary Statistics of Closed-End Funds Going Public , by ypes of Fund Offer price and proceeds are obtained from Securities Data Corp, Underwriter ranking is a measure of investmer t banker reputation found in Carter, Dark, and Singh (1998), his measure is based on the relative positions in th ombstone and has a highest ranking of 9, First day return is computed as : [(the closing price of fund at the first day if trading - offer price) / offer price]*] 00. unadjusted share return is computed as: [U (\ + Rit)-I]* \00 where Rj, is the return on month t for fund i, and =36 for the 36-month cumulative return. index-adjusted share return is the: -[11(1 t=l 701* 100 where I, is the monthly return on the S&P 500 for month t if the fund is a stock or mixed fiind, and the return on thie corporate bond index if the fund is a bond or municipal bond fund. Domestic funds invest exclusively in the US market. International funds invest in foreign markets only or a combination of foreign and US markets. Bond funds invest only in bonds, whether corporate or government. Stock funds invest exclusively in stocks and preferred stocks. Mixed funds invest in a variety of securities. Municipals invest exclusively in municipal bonds in the US Emerging market funds invest exclusively in emerging markets. Country funds invest in either a foreign country or a foreign region. he finil sample consists of 352 closed-end funds initiated during the period , ' US Domestic Funds Stock Bond Mixed Municipals Intemational Funds Stock Bond Mixed Emerging Markets Country Funds Overall N Offer Price ($) 13,31 12,73 11,98 12,58 14,21 13,56 12,72 13,15 13, ,75 13,37 Proceeds ($m) 211,97 216,09 260,94 143,58 194,03 234,18 155, ,54 146,12 140,82 217,97 mixed funds do not compare well to their benchmark index. International mixed funds perform the most poorly, with 36-month cumulative unadjusted and index-adjusted share returns of 7,94% and %, respectively. Overall, there is no significant difference in the long-run aftermarket performance of US domestic and intemational funds. his is inconsistent with Patro (2001), who suggests that international closed-end funds outperform US domestic funds,' he unadjusted and index-adjusted share returns for US domestic 'Patro (2001) reports that international closed-end funds do not significantly underperform the world market index while the US domestic funds significantly underperform the US market index, which suggests that international funds perform better than domestic funds. tjnderwriter Rank 8,64 8,37 8,56 8, ,64 8,63 8,57 8,72 8,64 8,62 8,64 First-Day (%) 0,30-0,65 0,32 1,54 0,28 2,30 2,14 0,07 5,37 2,91 2,88 0,89 Unadjuste d Share (%) 9,74 4,25 10,41 19,04 8, ,11 18,51 7,94 16,50 9,60 10,78 Index- Adjustec Share (%) -24, ,34-25,30-21,09-28,73-28,53-14,21-48,94-27,35-30, funds are 9,74% and -24,47% compared to 13,62% and -28,73% for intemational funds. he evidence suggests that the long-run performance of closed-end fund IPOs is very poor, a phenomenon that exists in IPOs of industrial firms as well. Exhibit 3 further examines the unadjusted and indexadjusted retums for funds invested in only domestic securities (domestic funds) and in international securities (international funds). First-day and firstweek unadjusted retums are significantly higher for intemational funds, but by the end of the first year, intemational funds perform significantly more poorly,'" '"Weiss (1989) and Peavy (1990) show that foreign equity funds perform better than domestic equity funds in the first several months following IPOs,

7 30 JOURNAL OF APPLIED FINANCE SPRING/SUMMER 2005 Exhibit 3. Unadjusted and Index-Adjusted Share s for Domestic and International Closed-End Funds Going Public Domestic funds invest exclusively in the US market, Intemational funds invest in foreign markets only or a combination of foreign and US markets. First day return is computed as [(the closing price of fund at the first day of trading - offer price) / offer price. First week return is computed as (the closing price of fund at the first week of trading - offer price) / offer price]*] 00. unadjusted share return is computed as: [n(i + ^;o-i]*ioo where Rj, is the return on month t for fund i, and =36 for the 36-month cumulative return. index-adjusted share return is the: [( )] where I, is the monthly return on the S&P 500 for month t if the flind is a stock or mixed fund, and the return on the corporate bond index if the fund is a bond or municipal bond fund. he final sample consists of 352 closed-end funds initiated during the period , -statistics show whether the means are statistically different from each other. ype of Funds Domestic Funds (N = 257) International Funds (N = 95) t-statistic ype of Funds Domestic Funds (N = 257) International Funds (N = 95) t-statistic Significant at the 0,0t level, Significant at the 0,05 level. Panel A. Unadjusted Share (Standard Deviation in Parentheses) First-Day 0,30 (2,57) 2,30 (8,47) 3,06*** First- Week 0,54 (2,57) 2,91 (11,29) 2,93*** 1-Month 0,77 (10,65) -1,09 (14,43) -1,29 12-Month -1,58 (14,12) -9,53 (23,81) -3,84*** 24-Month 0,24 (19,63) 0,82 (41,88) 0,18 Panel B. ]ndex-adjusted Share (Standard Deviation in Parentheses) 1-Month -0,38 (10,97) -0,121 (14,77) 0,58 12-Month -12,57 (13,44) -17,29 (22,15) -2,42** 24-Month -19,22 (18,88) -21,71 (45,09) -0,73 ^^^^^ 9,74 (24,51) 13,62 (66,53) 0,80-24,47 (29,33) -28,73 (71,58) -0,82 welve-month cutnulative unadjusted returns are -1,58% for domestic funds and -9,53% and intemational funds (the difference is statistically significant at the 0,01 level). he 12-month cumulative index-adjusted share returns exhibit the same pattern; domestic funds earn -12,57%, while international funds earn -17,29%, By the end of the second and third years, there is no significant difference between cumulative returns of domestic and intemational funds. Unadjusted retums go from positive to negative in the first year, and turn positive again in year 2 and 3, he index-adjusted retums are overall negative starting from the first month and they become more negative overtime. hirty-sixmonth returns are the worst, -24,47% for domestic funds and -28,73% for international funds. In Exhibit 4, we revisit the difference between the aftermarket performance of domestic and international funds, categorizing funds according to issue proceed amounts, offer prices, and underwriter rank. We divide sample funds into those with issue proceeds greater than or equal to the median of $ 120 million (large funds), and less than $120 million (small funds). We also divide the sample according to whether the offer price is higher than or equal to the median offer price of $15 (high-priced funds) or less than $15 (low-priced funds). We also categorize the sample into funds that use prestigious underwriters to take them public (underwriter rank greater or equal to 8) and those that use less prestigious underwriters (underwriter rank less than 8), First-day returns for international funds (Panel A) are significantly higher than those for domestic funds, whether funds are large or small. here is no significant difference in first-day returns for high-priced funds or funds underwritten by less prestigious underwriters. For low-priced funds and funds underwritten by

8 BOOH & EHRANIAN AFERMARKE PERFORMANCE OF CLOSED-END FUNDS 31 Exhibit 4. First-Day and s Classified by Proceeds, Offer Prices, and Underwriter Ranl< Domestic funds invest exclusively in the US market. International funds invest in foreign markets only or a combination of foreign and US markets. he median issue proceed and offer price are $120 million and $15 respectively. Underwriter ranking is a measure of investment banker reputation found in Carter, Dark, and Singh (1998). his measure is based on the relative positions in the ombstone and has a highest ranking of 9. Underwriters with a ranking of 8 or above are considered prestigious. First day return is computed as: [{the closing price of fund at the first day of trading-offer price) / offer price]* 100. unadjusted share return is computed as: [i where R;, is the return on month t for fund i, and =36 for the 36-month cumulative return. index-adjusted share return is the: [Yin + Rit)]-[n{l + It)]* \0Q where I, is the monthly return on the S&P 500 for month t if the fund is a stock or mixed fund, and the return on the corporate bond index if the fund is a bond or municipal bond fund. Bond funds invest only in bonds, whether corporate or government. Stock funds invest exclusively in stocks and preferred stocks. Mixed funds invest in a variety of securities. Municipals invest exclusively in municipal bonds in the US International funds invest in foreign markets only or a combination of foreign and US markets. Emerging market funds invest exclusively in emerging markets. Country funds invest in either a foreign country or a foreign region. he final sample consists of 352 closed-end funds initiated during the period statistics show whether the means are statistically different from each other. ype of Funds Domestic Funds International Funds t-statistics Panel A. First Day (Standard Deviation in Parentheses) Proceeds Proceeds Offer Price Offer Price 1Underwriter's >=120m <120m >=$15 <$15 Rank >= (3.05) (1.86) (3.04) (1.51) (2.65) (8.44) (8.57) (7.14) (9.84) (9.16) 1.71* 2.58** *** 3.10*** Panel B. 36-month Unadjusted Share (Standard Deviation in Parentheses) Domestic Funds (21.56) 6.21 (39.35) 9.35 (26.50) (21.06) (24.70) International Funds (24.09) (84.57) (35.74) (92.50) (70.05) t-statistics -2.37** 1.74* Panel C. 36-month Index-Adjusted Share (' %) (Standard Deviation in Parentheses) Domestic Funds International Funds t-statistics SignificantattheO.Ol level. Significant at the 0.05level. Significant at the 0.10 level (17.35) (33.18) -4.53*** (33.19) (89.38) (24.82) (47.78) -3.70*** (27.00) (93.16) (24.34) (73.73) Underwriter's Rank < (1.89) 0.42 (2.64) (15.49) 5.95 (38.29) (17.83) (48.54) 0.29 prestigious underwriters, intematiotial funds continue funds, but the difference is not statistically significant to earn higher first-day retums compared to domestic for the index-adjusted retums. funds. In Panels B and C (36-month cumulative Exhibit 5 presents regression results of the first-day, unadjusted and index-adjusted retums), intemational 36-month unadjusted, and 36-month index-adjusted large funds perform significantly more poorly than returns according to fund characteristics, fund type, domestic large funds: unadjusted and index-adjusted and fund strategies for both large and small funds, retums of 12.83% and % for domestic funds and he first-day retums for large intemational funds are 3.41% and % for intemational funds. Among no longer significantly higher than for domestic funds smaller funds, international funds earn higher after controlling for fund characteristics and fund type, unadjusted 36-month cumulative retums than domestic Similarly, the 36-month unadjusted retums for large

9 32 JOURNAL OF APPLIED FINANCE SPRING/SUMMER 2005 Exhibit 5. Regressions of First-Day, Unadjusted, and Index-Adjusted Share s on Fund Characteristics, Fund ype, and Fund Strategy by Size of Issue Offer price and proceeds are obtained from Securities Data Corp. Underwriter ranking is a measure of investment banker reputation found in Carter, Dark, and Singh (1998). his measure is based on the relative positions in the ombstone and has a highest ranking of 9. First day return is computed as: [{the closing price of fund at the first day of trading - offer price) I offer price]* 100. First week return is computed as: [(the closing price of fund at the first week of trading - offer price) / offer price] *100. International funds invest in foreign markets only or a cotnbination of foreign and US markets. Bond fluids invest only in bonds, whether corporate or government. Stock fimds invest exclusively in stocks and preferred stocks. Mixed funds invest in a variety of securities. Municipals invest exclusively in municipal bonds in the US -statistics corrected for heteroskedasticity using White test are in parentheses. Constant Offer Price Underwriter Rank Bond Stock Mixed International Adjusted R^ First-Day (-0.97) 0.12 (0.85) 0.32 (0.82) (-0.58) 1.12 (0.84) 4.83 (1.34) 1.02 (0.70) Issue Proceeds >= $120m Unadjusted (-0.62) (-0.34) 4.82 (1.19) 2.63 (0.52) (-0.28) 6.25 (0.64) (-L45) Index-Adjusted (-0.83) 1.20 (1.44) 0.03 (0.01) 2.91 (0.77) * (-1.79) (-0.04) ** (-2.58) First-Oay (-1.53) (-0.86) 0.95** (2.27) (-0.13) (-1.25) 0.71 (0.48) 2.87** (2.12) Issue Proceeds < $120m Unadjusted ** (-2.07) 0.47 (0.30) 8.18** (2.24) 9.21 (0.99) 1.07 (0.11) 6.32 (0.67) (1.20) Index-Adjusted ^5.95 (-1.12) (-0.59) 4.36 (1.02) (-0.17) (-1.26) ** (-2.07) (1.38) Significant at the 0.05 level. Significant at the 0.10 level. intertiatiotial funds are no longer significatitly lower than for large domestic funds. Large international funds continue to exhibit significantly lower 36-month indexadjusted returns than domestic funds after controlling for fund characteristics and fund type. Large stock funds earn significantly lower 36-month index adjusted returns than other types of funds. he first-day returns for small international funds are higher than for domestic funds, after controlling for fund characteristics and fund type. Funds underwritten by prestigious investment bankers earn higher first-day and 36-month unadjusted returns than funds underwritten by less prestigious investment bankers. here is no significant difference in 36-month unadjusted and index-adjusted returns for international and domestic funds after controlling for fund characteristics and fund type. Small mixed funds earn significantly lower 36-month index-adjusted returns than other types of funds. Exhibit 6 divides closed-end funds into high-priced and low-priced funds. For the high-priced group (offer price $15 or more), there is no statistically significant difference in the aftermarket performance of international and domestic funds after controlling for fund characteristics and fund type. Underwriter reputation is important in explaining 36-month unadjusted and index-adjusted returns, but not in explaining first- day returns. More prestigious underwriters are associated with significantly better long-term performance for the high-priced funds. he 36-month index-adjusted returns for high priced mixed funds are significantly lower than for the other types of funds; the same returns for the bond funds are significantly higher. For the low-priced group (offer price under $15), underwriter rank is associated with significantly higher first-day and 36-month unadjusted returns, but not 36-month index-adjusted returns. First day returns are significantly higher for international funds than for the domestic funds, but the 36-month returns are not significantly different for the two. In contrast to the high-priced funds where the long-term performance for bond funds are significantly better, bond funds for the low-priced group earn significantly lower 36-month returns, both unadjusted and index adjusted. he regressions in Exhibits 5 and 6 suggest underwriter reputation is important in explaining both the initial and the long-term performance of closedend funds, especially in the case of low-priced funds or small funds. For both small and low-priced funds.

10 BOOH & EHRANIAN AFERMARKE PERFORMANCE OF CLOSED-END FUNDS 33 Exhibit 6. Regressions of First-Day, Unadjusted, and Index-Adjusted Share s on Fund Characteristics, Fund ype, and Fund Strategy by Offer Price Offer price and proceeds are obtained from Securities Data Corp. Underwriter ranking is a measure of investment banker reputation found in Carter, Dark, and Singh (1998). his measure is based on the relative positions in the ombstone and has a highest ranking of 9. First day return is computed as [{the closing price of fund at the first day of trading - offer price) / offer price] *J00. First week return is computed as [(the closing price of fund at the first week of trading - offer price) / offer price] *I00. International funds invest in foreign markets only or a combination of foreign and US markets. Bond funds invest only in bonds, whether corporate or government. Stock funds invest exclusively in stocks and preferred stocks. Mixed funds invest in a variety of securities. Municipals invest exclusively in municipal bonds in the US -statistics corrected for heteroskedasticity using White test are in parentheses. Offer Price >= $15 Offer Price < $15 Constant Ln(Proceeds) Underwriter Rank Bond Stock Mixed International Adjusted R^ First-Day (-0.45) 0.33 (0.77) 0.04 (0.09) 0.11 (0.13) 0.03 (0.03) 3.61 (1.39) (-0.14) Unadjusted ** (-2.39) (-0.83) 13.94** (2.42) 18.56** (2.07) 4.18 (0.38) (1.23) (-0.79) Index-Adjusted *** (-2.64) (-1.12) 13.33** (2.16) 21.73** (2.42) (-1.35) * (-1.77) (-1.04) First-Day -7.90** (-2.09) (-1.00) 1.29** (2.24) (-1.01) (-0.36) 2.23 (0.97) 3.46** (2.42) Unadjusted (-1.55) (-0.70) 9.39* (1.69) *** (-3.24) (-1.00) (-0.52) (0.94) Index-Adjusted (-1.57) (-0.46) 4.66 (0.79) *** (-3.43) (-0.07) (-0.01) (0.90) ***SignificantattheO.OI level. Significant at the 0.05 level. Significant at the 0.10 level. fund strategy is important in explaining first-day returns but not long-term performance after controlling for fund characteristics and fund type. For large funds, international closed-end funds perform significantly more poorly than the domestic funds in the long term. IV. Summary and Conclusions It has been suggested that redeemable shares are not efficient when assets lack liquidity and prices are difficult to observe. his implies that closed-end funds will likely be more attractive when investment strategies involve holding less liquid assets such as international securities. We examine the relation between investment strategy and the share performance of closed-end funds invested in domestic and international securities from inception through three years of trading. We find closed-end funds invested in international securities earn significantly higher first-day returns than funds invested in domestic securities. his result persists across different size funds, funds with lower offer prices, and funds with more prestigious underwriters. here is less of a difference in first-day returns between large or high-priced international and domestic funds after controlling for fund characteristics and fund type. Overall, 36-month post- IPO share returns (both unadjusted and indexadjusted) are not significantly different for funds invested in international securities versus domestic securities. Larger funds invested in international securities perform significantly more poorly than larger funds invested in domestic securities after three years of trading, both unadjusted and index-adjusted. For index-adjusted returns, this difference holds even after controlling for fund characteristics and type of fund.

11 34 JOURNAL OF APPLIED FINANCE SPRING/SUMMER 2005 References Amihud, Yakov and Haim Mendelson, 1986, "Asset Pricing and the Bid-Ask Spread," Journal of Financial Economics 17 (No. 2, December), Carter, R., R. Dark, and A. Singh, 1998, "Underwriter Reputation, Initial s, and the Long-Run Performance of IPO Stocks," Journal of Finance 53 (No. 1, February), Carter, R. and Manaster, S., 1990, "Initial Public Offerings and Underwriter Reputation," Journal of Finance 45 (No. 4, September), Chang, Eric, Cheol Eun, and Richard Kolodny, 1995, "International Diversification through Closed-End Country Funds," Journal of Banking and Finance 19 (No. 7, October), Deli, Daniel and Raj Varma, 2002, "Closed-End versus Open- End: he Choice of Organizational Form," Journal of Corporate Finance 8 (No. 1, January), Dimson, EIroy and Carolina Minio-Kozerski, 1999, "Closed- End Funds: A Survey," Financial Markets, Institutions & Instruments 8 (No. 2, May), Fama, Eugene F. and Michael C. Jensen, 1983a, "Agency Problems and Residual Claims," Journal of Law and Economics 26 (No. 2, June), Fama, Eugene F. and Michael C. Jensen, 1983b, "Separation of Ownership and Control," Journal of Law and Economics 26 (No. 2, June), Hanley, Kathleen, Charles M.C. Lee, and Paul J. Seguin, 1996, "he Marketing of Closed-End Fund IPOs: Evidence from ransactions Data," Journal of Financial Intermediation 5 (No. 2, April), Levis, Mario and Dylan C. homas, 1995, "Investment rust IPOs: Issuing Behavior and Price Performance. Evidence from the London Stock Exchange," Journal of Banking and Finance 19 (No. 8, November), Loughran, im and Jay R. Ritter, 2004, "Why has IPO Underpricing Increased Over ime?" Financial Management 33 (No. 3, Autumn), Patro, Dilip Kumar, 2001, "Measuring Performance of Closed- End International Mutual Funds," Journal of Banking and Finance 25 (No. 9, September), Peavy, John W., 1990, "s on Initial Public Offerings of Closed-End Funds," Review of Financial Studies 3 (No. 4, Winter), Pennathur, A.K., N. Delcoure, and D. Anderson, 2002, "Diversification Benefits of Shares and Closed-End Country Funds," Journal of Financial Research 25 (No. 4, Winter), Ritter, Jay R., 1987, "he Costs of Going Public," Journal of Financial Economics 19 (No. 2, December), Weiss, Kathleen, 1989, "he Post-Offering Price Performance of Closed-End-Funds," Financial Management 18 (No. 3, Autumn),

12