基金屬性如何影響投資人買賣－以台灣境內股票型基金為例 - 政大學術集成

全文

(1)國立政治大學企業管理（MBA 學位學程）研究所碩士學位論文. How Fund’s Attributes Influence Investor’s Buy-and-Sell. 治 Equity Fund Markets – Evidence from Taiwan政 Domiciled 大. 立. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. e. i. i Un. v. g c h 博士指導教授：n屠美亞研究生：廖庭鋒撰. 中華民國 104 年 12 月.

(2) 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. i Un. v.

(3) Acknowledgement I would like to express my deepest gratitude to my advisor Prof. Twu, Mia, whose patience, support and guidance made my thesis work possible. Dr. Twu has always inspired me to make a difference rather than settling for what is. I thank my advisor for the countless hours of meetings, revisions, and advice on my thesis.. I thank the committee for their valuable comments and suggestions to improve my thesis. I would also like to thank my friends, ex-colleagues, co-workers for the. 治政 assistance, clarification and for their time on my thesis.大立 ‧ 國. 學. Last, and most definitely importantly, my family is always there to give me the. ‧. support and the encouragement when I go through my MBA study.. n. er. io. sit. y. Nat. al. Ch. engchi. I. i Un. v.

(4) Abstracts The thesis focuses on how embedded fee structure components (i.e. brokerage commission, transaction fee, management fee, custodian fee), fund turnover ratios, past fund performance affect Investor’s Buy and Sell in Taiwan onshore equity fund markets. To facilitate the analysis, two models are developed: multi-variable regression model and panel fixed-effect model. Our dataset consists of 19,906 Taiwan domiciled equity fund monthly data from January 2010 to December 2014.. 治政 For the fee structure, first, brokerage commission 大 is significantly positive to 立 both investor’s buy and sell. Second, transaction fee is negative to investor’s buy but ‧ 國. 學. positive to investor’s sell. Third, management fee is pronounced to investor’s buy,. ‧. which may result from three facets: fund product signal, revenue-sharing mechanism,. sit. y. Nat. and performance-related compensation. Forth, custodian fee has little predictive. io. er. power although the arguably results are explored from the two models.. al. n. iv n C For the turnover ratios, we h show that investorsUprefer buying funds with lower engchi turnover ratios. For the past fund performance, we also reveal that the lure of past performance-chasing is significantly evident for both investor’s buy and sell.. Through our studies, we also find several explanations from the viewpoints of irrationality of behavioral finance to illuminate fund investor’s buy and sell – representative heuristic, disposition effect, familiarity effect and home bias.. II.

(5) Contents and Lists of Tables Chapter 1 Introduction 1.1 Research Background 1.2 Research Motive 1.3 Research Objective. 1. Chapter 2 Literature Review 2.1 About Mutual Fund Industry 2.1.1 Mutual Fund Market Structure 2.1.2 Fund Companies Pay to Play 2.1.3 Stories behind Mutual Fund Fee’s Expenses 2.2 Fee Structures on Investor Welfare. 7. 1 2. 4. 7 7 9 10. 政治大 2.3 Turnover of an Equity Mutual Fund 立 2.4 Mutual Fund Performance Explanation. n. engchi U. y. sit er. ‧ 國. io. Ch. 14. ‧. Nat. Chapter 3 Methodology 3.1 Data and Sample Selection 3.2 Definition and Explanation of Variables 3.2.1 Fee Description 3.2.2 Control Variables 3.3 Research Methods 3.4 Model Specifications 3.5 Limitations. al. 13. 學. 2.5 Hypotheses Development. 11. v ni. 16. 20 20 21 21 24 25 28 30. Chapter 4 Empirical Results 4.1 Summary of Descriptive Statistics 4.2 Correlation Coefficient 4.3 Regression Results and Findings. 31. Chapter 5 Summary and Conclusions 5.1 Summary and Conclusions 5.2 Recommendations for Further Research. 47 47 50. Reference. 51. III. 31 35 36.

(6) List of Figures & Tables Figure 1 – Taiwan Onshore Fund Size Movement Figure 2 – The Research Process and Flow Figure 3 – The Flow of Hausman’s Test. 2 5 27. Table 1 – Taiwan Onshore Funds as of the End of December 2014 Table 2 – Summary of Descriptive Statistics for All Sample Funds Table 3 – Top 10 Funds with Yearly Cumulative Selling Turnover Ratios Table 4 – Correlation Coefficients among Six Independent Variables Table 5 – The results of the multiple regression model Dependent Variable = Buy (corresponding to Model A). 政治大 Table 6 – The results of the multiple regression model 立 Dependent Variable = Sell (corresponding to Model C). 6 32 33 35 44 45 45. Table 8 – The results of the two-way fixed effect model Dependent Variable = Sell (corresponding to Model D). 46. ‧. ‧ 國. 學. Table 7 – The results of the two-way fixed effect model Dependent Variable = Buy (corresponding to Model B). n. er. io. sit. y. Nat. al. Ch. engchi. IV. i Un. v.

(7) Chapter 1 Introduction 1.1 Research Background The asset management industry in Taiwan was born with the first fund company, Mega International Investment Trust founded in 1983. Since then, a number of Taiwan onshore funds have grown to 646 till the end of December 2014. Taiwan onshore mutual fund market by the end of 2014 totaled NTD 1,975 billion, compared to NTD 1,967 billion in 2013 according to Securities Investment Trust & Consulting Association of the R.O.C. (SITCA) data (hereafter ‘SITCA’). Taiwan onshore fund. 治政 sizes have been growing slowly in the past ten years,大 form 1,963 billion in 2005 to 立 1,975 in 2014. The biggest slump in Taiwan domiciled fund size was in 2008 because ‧ 國. 學. of the financial crisis, but overall onshore fund sizes started to recover from 2011.. ‧. Figure 1 shows the trend and the movements of the Taiwan onshore fund sizes as well. io. sit. y. Nat. as the number of the funds registered in Taiwan.. n. al. er. In total, there are 646 onshore mutual fund products in Taiwan; money market. Ch. i Un. v. funds account for 38.44% of all onshore fund sizes, followed by equity funds 27.21%. engchi. and fixed-income funds 13.54%. Details are exhibited in Table 1 according to SITCA’s data1.. 1. In comparison, the U.S. mutual fund industry accounts for nearly $16 trillion in assets, boosted. primarily by appreciation in stock and bond prices, and remains the largest mutual fund market in the world at year-end 2014, quoted from the report of the Investment Company Institute (2015).. 1.

(8) Data source: Securities Investment Trust & Consulting Association of the R.O.C.. 政治大. Figure 1 Taiwan Onshore Fund Sizes Movement. 立. ‧ 國. 學 ‧. 1.2 Research Motive. sit. y. Nat. From March 2, 2011, Taiwan’s Financial Supervisory Commission requires. n. al. er. io. existing funds to disclose all commission fees they award to bank distributors.. i Un. v. According to the Financial Time by Glori Ye, the move aims to improve the. Ch. engchi. transparency of distribution so as to support onshore fund management companies with more competitive edges, because offshore fund companies have been paying higher commission to banks than onshore managers. In February 2013, Financial Supervisory Commission, R.O.C. adopts the 'Plan to Encourage Stronger Business Ties in Taiwan for Offshore Funds,' which targets to encourage offshore funds companies to set up business locations in Taiwan. From these moves, it’s apparent that Taiwan regulatory bodies are eager to support Taiwan onshore mutual fund markets so as to create competitive advantages for Taiwan domiciled funds. So, these moves also draw our attention to the attributes of Taiwan domiciled funds on fund investors so. 2.

(9) that onshore funds can be more competitive while fund companies can further understand investor’s investment appetite. Compared to NTD 3,321 billion AUM of offshore fund size by the end of December 2014, Taiwan regulators aim to promote the development of substantive onshore fund management activities steadily, which accounts for the NTD 1,975 billion AUM onshore fund size.. There are many papers discussing the mutual fund fee structure, turnover ratio, and fund performance separately. For fee structure, Ruckman (2003) studies the expense ratios of North American mutual funds; Khorana, Servaes, and Tufano (2008). 政治大. analyze mutual fund fees charged by 46,580 mutual fund classes offered for sales in. 立. 18 countries. For fund turnover ratio, Ippolito (1989), Grinblatt and Titman (1994),. ‧ 國. 學. Droms and Walker (2001), Peterson and Riepe (2007) discuss whether the turnover ratio of a mutual fund is observed and whether or not it matters. As for the mutual. ‧. fund performance related, the classics in mutual fund performance evaluation are. y. Nat. io. sit. Treynor (1965), Sharpe (1966), Jensen (1968); Henriksson (1984) and Chang and. n. al. er. Lewellen (1984) document the net returns to fund investors from Sharpe-Lintner. Ch. i Un. v. market line while Connor and Korajczyk (1986) measure fund performance by APT. engchi. frameworks, and Ippolito (1989) analyzes the fund returns from Capital Asset Pricing Model (CAPM) market line.. As well as these, there are plenty of studies on how fee structures influence the mutual fund returns, do fees empirically realize its economies of scale, and whether or not the turnover ratio of the equity mutual funds matters. Still, much discussion linked to mutual fund returns are brought out, and those factors include fund manager tenure, performance persistence, market timing and selectivity, and super star funds.. 3.

(10) However, there are a few papers discussing fund attributes from investor’s perspectives. Given the growing size of Taiwan domiciled funds and the growing importance and protection for mutual fund investors, it's crucial to gain a deeper understanding from investor's angles. Specifically, the topics mentioned above from mutual fund investor’s viewpoints are scare, especially when these factors are taken into consideration at the same time. In fact, there are many factors influencing mutual fund investor's subscription and redemption, subjectively and objectively. We see that one area of emerging interest of among fund companies, regulators and fund investors is the fund investors' buy and sell activities.. 立. 政治大. ‧ 國. 學. 1.3 Research Objective. ‧. In this thesis, we aim to offer a detailed overview–with special emphasis on. sit. y. Nat. the three groups of the factors in mutual funds, including the embedded mutual fund. io. er. fee structures, turnover ratios, and past performance returns–of onshore equity. al. mutual fund markets in Taiwan and we attempt to shed light on how these factors. n. iv n C influence investor's buy and sell. hBy adopting quantitative methods in testing the engchi U research hypotheses, we tend to gain a deeper understanding from investor's angles.. Specifically, we aim to see the results of investor’s buy (fund inflows) and sell (fund outflows) respectively since the mix of the gross fund flows would likely make the results biased.. We believe the six independent variables we choose are the crucial fund attributes and these attributes are key consideration for investor’s buy and sell. We show the research process and flow in the Figure 2 below. From mutual fund investors’ angle, we look forward to extending and seeing that this thesis is helpful in fund 4.

(11) companies' fund distribution, fund strategy, and fund product development while it also helps regulatory bodies keeping pace with the fast-evolving mutual fund markets from Taiwan domiciled funds.. Introduction ￭. Research Background. ￭. Research Motive Research Objective. ￭. Nat. y. ￭. ‧. ‧ 國. ￭. 學. ￭. 立. About Mutual Fund Industry Fee Structures on Investor Welfare Turnover of an Equity Mutual Fund Mutual Fund Performance Explanation. io. sit. Hypotheses Development. n. al. Data & Sample Selection ￭ Definition of Variables ￭ Chosen Control Variables. Methodology. Ch. engchi. er. ￭. 政治大. Literature Review. i Un. v. Research Methods & Models Specifications ￭ Multivariate Linear Regression Model ￭. Panel Fixed-Effect Model. Empirical Results & Findings Conclusions & Recommendations for Further Research Figure 2 The Research Process and Flow. 5.

(12) Table 1 Taiwan Onshore Funds as of the End of December 2014 Groups. Types. No.. AUM in NTD. Domestic Equity Fund. 151. 238,460,758,828. International Equity Fund. 189. 298,861,688,099. Domestic Balanced Fund. 25. 23,767,483,109. International Balanced Fund. 19. 42,811,253,841. Quasi Money Market Fund. 0. 0. Domestic Fixed-income Fund. 1. 3,400,392,928. 48. 69,725,959,042. 1. 337,299,350. High Yield Bond Fund. 36. 193,898,910,938. Domestic Money Market Fund. 治 International政 Money Market Fund 大 Domestic Fund of Funds 立. 45. 741,434,611,129. 8. 17,719,871,460. 0. 0. International Fund of Funds - Equity Fund. 12. 14,667,159,406. 29. 61,613,177,323. 25. 59,946,835,358. Equity Fund. Balanced Fund Quasi Money Market Fund. International Fixed-income Fund Fixed-income Fund Financial Asset Securitization Fund. Principal Guaranteed Fund. al. n. International. Exchange Traded Fund. n U i e h nFund gc International Index. Ch. Domestic Index Fund Index Fund Total(NTD). 8. 15,212,935,877. 12. 19,202,937,442. 17. 89,973,600,155. 6. 71,035,689,909. 3. 3,125,079,077. 8. 8,615,061,440. 646. 1,974,798,827,055. er. io. Domestic Exchange Traded Fund. Exchange Traded Fund. 988,122,344. sit. REITs Fund. 3. y. International Fund of Funds - Others. Nat. REITs Fund. International Fund of Funds - Balanced Fund. ‧. Principal Guaranteed Fund. International Fund of Funds - Bond Fund. 學. Fund of Funds. ‧ 國. Money Market Fund. iv. Data source: Securities Investment Trust & Consulting Association of the R.O.C.. 6.

(13) Chapter 2 Literature Review At the beginning, we introduce the mutual fund industry, primarily focusing on the three parts: mutual fund market structure, the unspoken rules about paying to play for fund companies, and then the stories behind mutual fund fee’s expenses. After that, the literature related to our thesis may be divided into three main groups. First, there is the considerable body of work that has focused on the fee structure of the mutual fund characteristics. Second, there is the literature that looks at turnover ratios linking to the fee structure context or mutual fund performance. Finally, a set of papers takes a. 政治大. more ‘mutual fund performance focus’ to examine the possible independent variables.. 立. ‧ 國. 學. However, the empirical literature on the relationship among mutual fund flows, including buy and sell (dependent variables), and the mutual fund fee structure (i.e.,. ‧. brokerage commission, transaction fee, management fee, and custodian fee), turnover. Nat. sit. y. ratios, and mutual fund past returns (the six independent variables), is somewhat. n. al. er. io. limited. This thesis aims to make the empirical contributions from mutual fund. i Un. v. investor’s perspective (i.e. investor’s buy and sell) using Taiwan onshore equity fund data.. Ch. engchi. 2.1 About Mutual Fund Industry 2.1.1 Mutual Fund Market Structure In the mutual fund industry, it's easy to enter with many fund houses. Still each fund house provides unique product features, quality, and service, which makes fund houses far from price competition. Therefore, competition in mutual fund industry does not focus on fees charged to shareholders. Instead, mutual fund industry 7.

(14) fundamentally exhibits characteristics of monopolistic competition and the mutual fund market structure basically prevails and competes on product attributes other than expense ratios. Sirri and Tufano (1998) define ‘fund products’ as close, but imperfect substitutes because of its unique features. A General Accounting Office (2000) report notes that mutual funds barely endeavor to compete on cost. Luo (2002) further states that if the mutual fund industry is not perfectly competitive, the fund company has market power to set the fees above its costs, besides cost-related charges including the performance, age, size, and cash ratio of the fund. Khorana and Servaes (2004) support empirically that product differentiation affects market share in the mutual. 治政 fund industry by using the universe of all U.S. fund families 大 covering the 1979-1998 立 periods. Further, Wahal and Wang (2011) explain that if we regard the mutual fund ‧ 國. 學. portfolio as the ‘‘product,’’ then we could think of fund fees as ‘‘prices’’ charged for. ‧. that product and stock holdings as key ‘‘inputs.’’ Houge and Wellman (2007) explains. sit. y. Nat. that fund companies attempt to differentiate their funds, part investment vehicle, part. io. al. n. investment philosophy and policy, and quality of service.. Ch. engchi. er. consumer product, by proactively marketing and branding historical performance,. i Un. v. Wang and Venezia (2012) show the empirical U-shape structure-performance relation for Taiwan mutual fund market structure. The authors suggest that as long as a fund’s market share becomes too larger or too smaller, the negative influence on fund performance of market concentration will get stronger. The situation causes that mutual funds endowed with too weak or too strong market power can erode their performance. Therefore, keeping the modest scale in proportion to rivals in the Taiwan mutual fund industry would be more advantageous to its fund performance.. Besides monopolistic competition, non-price competition, and evidenced 8.

(15) U-shape market structure in Taiwan, another highlight is economies of scale in the mutual fund industry. Ferris and Chance (1987), McLeod and Malhotra (1994), Malhotra and McLeod (1997), and Latzko (1999) show a negative relation between the expense ratio and fund assets, the situation explained by economies of scale in mutual fund administration.. 2.1.2 Fund Companies Pay to Play Between mutual fund companies and distributors like banks, securities firms,. 政治大 revenue-sharing arrangements. Mutual fund companies usually negotiate fixed or 立. insurance companies and direct sales, there exists costly practice for pay to play, or. ‧ 國. 學. floating dollar amounts of fees to distributors and channels in order to promote asset under management (AUM) of the mutual fund companies. However, these fees could. ‧. provide a conflict of interest for financial consultants in recommending suitable. Nat. sit. y. mutual funds to their clients. Besides, part of the fees that mutual fund companies pay. n. al. er. io. to distributors are in order to win the ticket to have their funds available to investors. i Un. v. via distributor’s platforms, especially banking’s platforms. Another hidden problem is. Ch. engchi. onshore funds registered in Taiwan usually cannot compete with offshore funds registered outside Taiwan in fee negotiation with distributors; this is because offshore fund companies do not need to build local infrastructure and networks in Taiwan and therefore face lower day to day operation costs and hence shift the saving as fees to promote their AUM and preferred funds. Due to the higher commission fees from offshore funds, distributors desire to sell offshore funds.. With such background, from March 2, 2011, Taiwan’s Financial Supervisory Commission (FSC) has required existing funds to disclose all commission fees they award to bank distributors. This move aims to improve the full disclosure and full 9.

(16) transparency of fund distribution, targeting to ease up on commission fees and to mitigate the situation that offshore funds pay higher commission fees for distributors than onshore funds.. On September 16, 2015, Taiwan’s Financial Supervisory Commission further hints the Taiwan online fund supermarket platform is under way, which could be similar to Fund Online Korea. According to the overview of Fund Online Korea, founded as a public financial investment company, it is an advanced fund distribution channel that enables customers to invest in all funds online at reasonable costs. If. 政治大. launched, most public funds could be tradable in the online fund supermarket. Most. 立. importantly, the situation of pay to play, or revenue-sharing arrangements could be. ‧. ‧ 國. 學. re-arranged.. 2.1.3 Stories behind Mutual Fund Fee’s Expenses. sit. y. Nat. io. er. Admati and Pfleiderer (1998) note that since fund investors do not observe the information directly, fund companies ‘sell’ information by delivering mutual funds. al. n. iv n C each investor. to the h e n gAccording chi U. and charging fees from. 2015 report from the. Investment Company Institute (ICI), expenses incurred by equity mutual fund investors have declined substantially, from the asset-weighted averages of 0.99% in 2000 to 0.70% in 2014. As suggested by Porter (1980), high market share and cumulative experience, together with a strategy of differentiation, leads to lower costs and lower fees.. In the article 'Expense ratio tells just part of story' by Moroney, Sweet, Carlson, Wright, and Hopfinger (2007), it highlights the actual costs from a fund can be much higher than the apparent expense ratios since expense ratios do not reflect trading 10.

(17) costs, which include front-end load fee/ back-end load fee and the indirect costs of trading. And these fees could be potentially huge drags on fund performance. The article further explains for the average equity funds, they incur trading costs coming from bid-ask spreads when trading underlying assets and facing unloading a large stakes or buying heavy trades.. John (2004) makes a commentary regarding that are mutual fund expenses too high? John summarizes that value funds have lower expense ratios than growth funds and that large-cap value funds have the lowest expenses across all investment styles.. 治政 He also mentions that older equity funds with aging 大 incur lower expense ratios, and 立 that international equity funds have higher expenses than domestic funds. Khorana ‧ 國. 學. and Servaes (2004) highlight the positive correlation between market share of a fund. ‧. company and its fees charged for marketing and distribution of funds. They further. sit. y. Nat. show that fund families with lower fees can increase their fees at a reasonable. io. er. mark-up, and most importantly, those fund families can still retain its market shares,. al. or assets under management. Therefore, fees for marketing and distribution can be. n. iv n C justified if fund investors have taken into account before deciding to invest h these e n gcosts chi U which funds.. 2.2 Fee Structures on Investor Welfare About fee structures on investors, we show previous studies four-fold: mutual fund fee-performance relation, the fee structure practice, fee-setting, and the latest evolution of fee structure from Morningstar report.. For mutual fund fee and fund performance related, Otten and Bams (2002) 11.

(18) document a negative influence of mutual fund fees on fund performance from five European markets. However, Goetzmann (1995) and Dowen and Mann (2004) find no relation between expense ratios and performing funds. Haslem, Baker, and Smith (2007) document an interesting note that directors seldom replace fund managers for poor performance, but they routinely approve increases in fund expenses. On the other hand, Khorana and Servaes (2004) concludes that fund managers and fund investors face conflicts of interest when the former would like to maximize mutual fund fee related revenues while the later prioritizes the fund investment returns.. 治政 For fee structure practice, Chordia (1996) 大 provides empirical evidences 立 suggesting: (i) In face of redemption, Redemption fees take more effects at investor’s ‧ 國. 學. redemption than Front-end load fees, (ii) In terms of fund type, aggressive growth. ‧. funds with smaller and less liquid assets are more sensitive to cash flows volatility in. sit. y. Nat. case of meeting investor’s liquidity, and therefore load fees could be a good tool at. io. al. n. front-end loads generally result in lower expense ratios.. Ch. engchi. er. dissuading redemptions. Dellva and Olson (1998) conclude that funds sold with. i Un. v. On another issue, Luo (2002) shows that besides the cost-related characteristics of the fund, the competitiveness of the market impacts the fee-setting significantly. Two components to determine fee-setting are highlighted: the marginal cost and the mark-up. The author shows that the traditional cost-related variables (the size, age, cash ratio, and performance of the fund) play only a small part of the fee determination; Luo (2002) empirically and quantitatively models that the majority of the fee is the mark-up, added to its marginal cost by the fund, which possesses market power due to the monopolistic competition market structure. Christoffersen and Musto (2002) document the fee-setting from the demand elasticity and propose that 12.

(19) mutual fund fees are set to respond to various elastic demands. Still, Das and Sundaram (2002) propose the three roles behind the fee structure as the following: (i) It influences trading behavior and portfolio choice by affecting investment adviser incentives since fee structure could be sources of compensation, (ii) It determines the distribution of returns between investor and adviser and therefore fee structure serves as a risk-sharing function, and (iii) It may be used as a device for signaling fund ability since fee structure affects how funds compete with each other.. In 2015 Morningstar Global Fund Investor Experience Report, it mentions the. 治政 evolution of fee structure from bundled to unbundled 大 fee arrangements, which means 立 a sales commission paid to a financial consultants evolves and shifts from fund ‧ 國. 學. companies to fund investors. From 2013, the U.S. has become the main market for. ‧. this unbundled fee arrangement. Following by the U.S., unbundled fee model. sit. y. Nat. occupies the markets in Australia, South Africa, and the Netherlands, together with. io. n. al. er. gradually growing markets in New Zealand, U.K., Denmark, and Switzerland.. Ch. i. i Un. v. n g c h Fund 2.3 Turnover of an Equitye Mutual There are two areas for mutual fund turnover ratios related studies: fund turnover— performance relation, and fund turnover—fund size relation.. For fund turnover—performance relation, Grinblatt and Titman (1994) conclude that turnover ratio significantly links to the ability of the fund managers for earning abnormal returns. However, Droms and Walker (2001) examine the persistence of fund characteristics, showing that turnover ratios for mutual funds that operated over the full two decades from 1971 to 1990 are not observed. This result 13.

(20) supports Ippolito’s (1989) study, which states no statistical relationship between turnover and performance is identified, which may be because turnover does not occur equally and evenly throughout the whole year. Still, Peterson and Riepe (2007) in the article ‘Does the Turnover of an Equity Mutual Fund Matter?’ conclude that the relationship between turnover and portfolio performance is more nuanced than conventional wisdom would suggest. This contradicts the traditional view that high turnover rates of portfolio erode the performance of an equity mutual fund. This also means the traditional view would not sustain: the explicit hypothesis says that high turnover creates more transactions costs and these costs act as a drag on fund. 治政 performance, and the implicit hypothesis hints that portfolio 大 turnover generates only 立 costs and no alpha. ‧ 國. 學 ‧. Another interesting study is fund turnover—fund size relation. Chan and. sit. y. Nat. Lakonishok (1995) show that the total trading costs are higher for high turnover funds. io. er. since high turnover funds prefer employing short-term trading strategies (i.e.,. al. momentum strategy), and therefore they have relatively high demands for immediacy. n. iv n C in their trades. With liquidity needs, that the adverse effect of fund hiteisnnotg surprising chi U. size on performance is greater among high turnover funds. Yan (2008) further proves that the adverse effect of scale on performance is more pronounced among high turnover funds.. 2.4 Mutual Fund Performance Explanation The most pronounced studies are fund flow-performance relation. Warther (1995) documents a strong positive correlation between monthly fund flow and market returns. Elton, Gruber, and Blake (1996) demonstrate the findings that the past 14.

(21) risk-adjusted performance carries the information about the future. Still, these studies also hints the asymmetric fund flows. Goetzmann and Peles (1997) show a significant relation between flows and past returns only for the top quartile of past returns, which is evident for the asymmetric fund flows.. Another interesting area is regarding the concept of star fund managers noted by Ippolito (1992) and Sirri and Tufano (1998). They show that if a fund company contains one or more star performers, it can expand or retain its asset under management since funds with previously exceptional performance appeal to more. 治政 fund investors for consecutive years. To enhance market 大share, or AuM, Khorana and 立 Servaes (2004) propose that the presence of a star performer in the family like ‧ 國. 學. Morningstar performance ranking could enhance its fund market share. Contrary to. ‧. this claim, some studies (Friend, Blume, and Crockett, 1970; Grinblatt and Titman,. er. io. sit. y. Nat. 1989, 1992) find no superior mutual fund performance by fund managers.. al. For the relation between the fund performance and fund expenses, Carhart. n. iv n C (1997) shows the inverse relation h between net returns e n g c h i Uand expense levels, especially. for those actively managed funds. This is in line with the results of Elton, Gruber, and Blake (1996) that the funds with poor performance carry the higher expense ratios. However, Goldstein and Krutov (2000) show the evidences that whether or not the fees matter, it depends on the historical fund performance. As long as fund investors satisfy their investment returns, the fee cost paid to the fund is not critical.. 15.

(22) 2.5 Hypotheses Development This thesis spotlights on ‘How fund’s attributes influence investor’s buy-and-sell.’ To address this issue, this study presents a series of alternative hypotheses, which are grouped into three parts, including mutual fund fee structures, turnover ratios, and performance returns. The alternative hypotheses are constructed and explained as follow:. The literature on the relation between fund expenses and fund flows exhibits. 政治大. mixed results. Sirri and Tufano (1998) document a negative relation between total. 立. fund expenses and fund flows while But Barber and Odean (2000) examines the. ‧ 國. 學. impacts of various operating fees on mutual fund flows and finds no relation between the two. For one explanation, Khorana and Servaes (2004) propose the mutual fund’s. ‧. fee is a result of the mark-up, added to its marginal cost by the fund. Taking mutual. y. Nat. io. sit. fund returns into consideration, Goetzmann (1995) find that high expenses cannot. n. al. er. explain those poor performing funds. Dowen and Mann (2004) find no relation. Ch. i Un. v. between expense ratios and returns for equity funds. However, from investor’s. engchi. perspective, we support the results from Malkiel (1995), who concludes that purchasing low expense funds is better for fund investors; we also believe that fund fees as the costs would drag fund performance, which echoes Carhart (1997)’s paper showing that fund fees have a significant impact on fund performance. Therefore, we argue that if total fee is too high, investor’s buy will be retreating while sell will be accelerating.. In details, funds trade through brokers and brokerage commission applies per trade—purchases or sales. Ideally no matter whether it is fund inflow or outflow, fund 16.

(23) managers need to adjust its portfolios based on the new subscription and redemption coming into the funds and thereafter brokerage firms execute trades by the instructions. Thus, we expect this variable, or brokerage commission, to be positive. This is in line with the so-called hidden costs (Prior, 2010; Gao and Livingston 2013). Transaction fee includes all taxes and related costs, which may be due on the assets and the income of the fund. Especially, transactions in less liquid markets (e.g. emerging markets or frontier markets) will incur greater costs than those in more liquid markets. From buyer’s view, transaction fee could deter the investor’s subscription if it is too costly but for sellers, it may just be the costs if it’s decided to. 治政 sell the assets. Another guess is we wonder investors 大would be more interested in 立 those familiar markets and transaction fee reflects the truth of the higher costs when ‧ 國. 學. funds enter in those unfamiliar markets. Under such circumstances, we would expect. ‧. transaction fee to be negative for investor’s buy but positive for investor’s sell. For. sit. y. Nat. management fee, Chevalier and Ellison (1997) conclude that if the fund performance. io. er. is high, total fee revenue increases as fund grows. Deli (2002) offers the consistent. al. argument that fees are set to provide proper performance incentives for fund managers.. n. iv n C However, we view this much linksh to the management e n g c h i Ufee because a fund with a high management fee might take a more aggressive investment approach to gain the returns so as to satisfy mutual fund investors. We therefore expect management fee to be. positive for investor’s buy and sell. The custodian’s business is the provision of custodial and trustee services for collective investment schemes and fund portfolios. From investor’s angle, we would not expect it have effects on investor’s buy and sell. Based on these reasoning, the mixed set of hypotheses we test for the fee structure on the investor’s buy and sell as follow:. 17.

(24) H1-1.1: Mutual fund’s fees – Brokerage Commission –are positive to investor’s Buy H1-1.2: Mutual fund’s fees –Brokerage Commission –are positive to investor’s Sell H1-2.1: Mutual fund’s fees – Transaction Fee –are negative to investor’s Buy H1-2.2: Mutual fund’s fees –Transaction Fee –are positive to investor’s Sell H1-3.1: Mutual fund’s fees – Management Fee –are positive to investor’s Buy H1-3.2: Mutual fund’s fees –Management Fee –are negative to investor’s Sell H1-4.1: Mutual fund’s fees – Custodian Fee –are neutral to investor’s Buy H1-4.2: Mutual fund’s fees –Custodian Fee –are neutral to investor’s Sell. Ippolito’s (1989) and Droms and Walker (2001) document no persistence for. 政治大. the turnover rates of mutual funds. Peterson et al. (2001) and Peterson and Riepe. 立. (2007) conclude that turnover rates have little to do with mutual fund performance.. ‧ 國. 學. Based on the claims, it hints the fund portfolio turnover generates alpha besides the incurring costs. It also hints that turnover by skilled managers is a plus while turnover. ‧. by unskilled managers is a minus. From our view, although higher alpha funds are. y. Nat. io. sit. able to offset some negative effects from turnover, it is not costless. Turnover of the. n. al. er. assets incurs costs in order to trade on information or to meet certain investment. Ch. i Un. v. criteria like hedging ratios or derivatives limits. Besides, fund sizes of Taiwan equity. engchi. domiciled funds are not as big as those of offshore funds; it means a big subscription or redemption will need a fund manager to adjust its portfolio timely. Therefore, purchase turnover and sell turnover may increase with incurring costs but generating alpha uncertainly and arguably. However, there is no direct evidence on our thoughts in Taiwan domiciled equity fund market. Therefore, contrary to the previous studies, we hypothesize and test a negative relation between fund turnover ratios and investor’s buy and sell. H2-1: Turnover rates of mutual funds have negative effects on investor’s Buy H2-2: Turnover rates of mutual funds have negative effects on investor’s Sell 18.

(25) Hendricks, Patel, and Zeckhauser (1993), and Goetzmann and Ibbotson (1994) document strong evidence that the past mutual fund performance predicts future performance. Some papers support the evidence of the positive correlation between the past fund performance and the signal about the future information, and even the fund flows (Patel et al., 1994; Warther, 1995; Elton, Gruber, and Blake, 1996; Chevalier and Ellison, 1997; Goetzmann and Peles, 1997; Sirri and Tufano, 1998; Wermers, 2000). Consistent with the study, we hypothesize that the historical fund performance could stimulate fund flows, from investor’s viewpoints. If a fund persists in a higher performance return, it would stimulate the greater buy-and-sell for mutual. 治政 fund investors to chase profit-taking. Thus, we expect 大 this variable to be positive. That 立 is, fund flows of investor’s buy-and-sell are positive related to underlying fund. Nat. er. io. sit. H3-2: Past performance stimulates investor’s Sell. y. H3-1: Past performance stimulates investor’s Buy. ‧. ‧ 國. 學. performance.. al. Therefore the hypotheses 1 to 3 are formulated to test whether or not the. n. iv n C factors identified from previous findings investor’s subscription U h e n gactually c h i influence and/or redemption. These factors include “expense ratios and its breakdown, including brokerage commission, transaction fee, management fee, and custodian fee,” “turnover,” and “fund performance.”. 19.

(26) Chapter 3 Methodology 3.1 Data and Sample Selection This thesis uses a unique combination of data sets. We obtain Taiwan onshore mutual fund data from Securities Investment Trust & Consulting Association of the R.O.C. (SITCA). The SITCA database contains the thorough onshore open-end mutual funds data dating from March 2001 onwards and data is available at the individual fund level. The data required for the thesis is combined from four SITCA’s. 政治大 report, and fund performance report (Morningstar version). Details are as below: The 立 data sources including fund summary report, fund expense ratio report, fund turnover. ‧ 國. 學. information includes fund code, fund ID, fund size, market share, net asset value (NAV), issued shares, subscription amount, redemption amount, commission fee,. ‧. transaction fee, management fee, custodian fee, total fee, turnover ratio, performance. n. al. er. io. sit. y. Nat. return-1y.. i Un. v. In our empirical analysis, because of incomplete data on some variables or. Ch. engchi. missing data on some funds, we restrict our sample to 19,906 equity fund data. We sort out the data on the monthly basis, ranging from January 2010 to December 2014. Although this spans a shorter time-period for five years, we have more frequent observations for fund samples on a monthly basis. Still, the thesis focus on onshore equity fund analysis, and hence we only include the categories AA1 and AA2, which is specified by SITCA meaning the fund types of Domestic Equity Fund, and International Equity Fund.. As for the macro data, we use them as the control variables, which will be. 20.

(27) explained later. The macro data is collected from Central Bank of the Republic of China (Taiwan), including 4 variables and using month-end data on a monthly basis from January 2010 to December 2014: Monetary Aggregate M1B (yoy), CPI Change Rate (yoy), 31- 90 Day Commercial Paper Interest Rate, and Taiwan 10-Year Bond Yield.. 3.2 Definition and Explanation of Variables 3.2.1 Fee Description. 立. 政治大. Brokerage Commissions and Transaction Fees. ‧ 國. 學. Brokerage commissions go toward paying brokerage firms commissions for executing trades. After a brokerage trade has been executed, the securities broker may. ‧. collect a processing fee from the funds. Transaction fee goes toward paying taxes. y. Nat. n. al. er. io. mutual fund.. sit. related incurred from dealing with purchases and sales of underlying assets from a. Management Fees. Ch. engchi. i Un. v. John (2004) defines the management fees by far the largest component of mutual fund expenses, which represent investment advisory charges for portfolio management, but may also confusingly include other administrative expenses as well. Khorana, Servaes, Tufano (2009) point that management fees represent the charges levied each year by funds for management services, which always cover investment management services, while some even include payments for administration and distribution.. 21.

(28) Custodian Fees According to the Regulations Governing Securities Investment Trust Funds, amended on September 25, 2009, Article 16 says the custodian institution shall establish and maintain separate accounts for the securities investment trust funds under its custody. Custodian is obligated to safeguard the assets of the fund and it is acting pursuant to specific proper instructions and providing custodial and trustee services. Charles and Dennis (2005) in the paper titled 'An Analysis of Mutual Fund Custodial Fees' mention that the custodial fee is usually subject to arms-length. 政治大 different from distribution fees paid to a related party. In the above paper it shows the 立 negotiation with third parties, and is typically paid directly from fund assets, which is. ‧ 國. 學. evidence that the expense ratio (excluding custodial fees) is negatively associated with custodial fees. One possible reason for this result is that firms exposed to higher. ‧. non-custodial expenses might pressure the custodians to provide services at a lower. Nat. n. al. er. io. sit. y. price to avoid having their expense ratios become unavoidable high.. Turnover Ratio. Ch. engchi. i Un. v. Based on the definition of Morningstar, turnover ratio measures a fund's trading activity, which is computed by taking the lesser of buying or selling (excluding all securities with maturities of less than one year) and dividing by average monthly net assets. The ratio means a mutual fund holdings that have been turned over, changed, or replaced with other holdings in a given time period. It tells you the roughly idea of the fund’s strategy like a buy-and-hold strategy or it experiences considerable purchases and sales of its holdings. Because every fund's trading activity involves a commission, this means that high turnover ratios often incur high expense ratios. 22.

(29) Lavin and Magner (2014) posit three dimensions determining the turnover of a mutual fund. They are efficiency, agency problems, and behavioral biases. For efficiency, because of economies of scale and scope, financial intermediaries are able to reduce its transaction costs by diversifying their portfolios, preserving added liquidity, and even monitoring the underlying investments. In the same way, Hellwig (1991) suggests that mutual fund products can be seen as coalitions of investors who exploit economies of scale and scope in transaction technologies. As for agency problems, the incentives for fund managers can induce them to execute an excessive number of trading but excessive transaction costs will reduce shareholders’ return on. 治政 mutual fund investments. Lastly, for behavioral biases, 大or overconfidence, it arises 立 because of fund managers’ overestimation of their ability and judgement for the ‧ 國. 學. underlying stocks and therefore they overreact to buying and selling of financial. ‧. assets (Daniel, Hirshleifer, and Subrahmanyam, 1998). The mentioned three. er. io. sit. y. Nat. dimensions explain why the turnover rates are high for some mutual funds.. n. Past Fund Performance a (one year) In the thesis, we. iv l C n U h e year use one h i performance n g cfund. calculated by the. Morningstar's Total Return. As defined by the Morningstar Investing Glossary 2 , Morningstar's calculation of total return, expressed in percentage terms, is determined by taking the change in price, reinvesting, if applicable, all income and capital gains distributions during the period, and dividing by the starting price. For the calculation, it doesn't take sales charges (such as front-end loads, contingent deferred sales charge, or redemption fees) into consideration.. 2. Definition of 'Total Return' by Morningstar Investing Glossary http://www.morningstar.com/InvGlossary/total-return.aspx 23.

(30) 3.2.2 Control variables In our regression analysis, we add five macro-economic factors as the control variables to further explain additional information if any independent causal effects on the dependent variables. They are Monetary Aggregate M1B (yoy), CPI Change Rate (yoy), 31- 90 Day Commercial Paper Interest Rate, and Taiwan 10-Year Bond Yield, respectively. Why we choose these control variables lies in the two reasons. First, these macro-economic variables are highly related to the business cycle in Taiwan. Second, some macro-economic factors reflect the opportunity costs, or the value forgone in order to buy or sell the mutual funds. By doing so, the results may be. 治政 prevented from the misleading or biased estimation. The 大definition below is from the 立 Central Bank of the Republic of China (Taiwan) and Bloomberg websites. 3. ‧. ‧ 國. 學. Monetary Aggregate M1B (yoy). sit. y. Nat. M1B = M1A + Passbook savings deposits of Individuals and non-profit. n. al. er. io. organizations in banks and community financial institutions. (So far, only individuals. v. and non-profit organizations are allowed to open Passbook Saving accounts).. Ch. engchi. i Un. CPI Change Rate (yoy) According to the definition from the World Bank, Consumer Price Index reflects changes in the cost to the average consumer of acquiring a basket of goods and services that may be fixed or changed at specified intervals, such as yearly. Consumer Price Index CPI in Taiwan is reported by the Directorate-General of Budget, Accounting and Statistics, Taiwan. 3. The definition is from ‘Definitions of Key Financial Terms’ on the website of the Central Bank of the Republic of China (Taiwan). 24.

(31) 31- 90 Day Commercial Paper Interest Rate Commercial papers can be divided into type I (CPI) and type II (CPII). The CP1 is derived from an actual transaction while the CP2 is issued by legally registered private corporations or public enterprises for the purpose of fund raising.. Taiwan 10-Year Bond Yield Central government bonds (CGBs) are negotiable debt obligations issued by the central government with a maturity of more than one year. CGBs are commonly. 治政 year since the issue date, and are redeemed at par value大 at maturity. 立. issued with maturities of 2, 5, 10 and 20 years. They make coupon payments once a. ‧. ‧ 國. 學. TWSE. According to the definition on Bloomberg, the TWSE, or TAIEX, Index is. sit. y. Nat. capitalization-weighted index of all listed common shares traded on the Taiwan Stock. n. al. er. io. Exchange. The index has a base value of 100 based on its 1966 level. The index is launched on January 5, 1967.. Ch. engchi. i Un. v. 3.3 Research Methods We employ an ordinary least squares multivariate regression model and the panel fixed-effect model to determine whether an association or reaction exists between dependent variables (mutual fund investor’s subscription and redemption) and independent variables (expense ratio, fund asset turnover, and fund performance) in Taiwan onshore equity fund market.. 25.

(32) Multivariate Linear Regression Model In fact, there are many mutual fund related studies using the method of ordinary least squares (OLS) regression analysis. The most prominent analysis is to examine the economies of scale in mutual fund expenses, including Ferris and Chance (1987), Malhotra and McLeod (1997), Latzko (1999), and Rea, Reid, and Millar (1999). Still, the relation between mutual fund fees and mutual fund performance is analyzed, including Droms and Walker (1996), Costa and Porter (2003) and so on.. Panel Fixed-Effect Model. 治政大 the advantages of panel Over two decades ago, Hsiao (1986) has summarized 立. data4. According to Oscar Torres-Reyna (2007)5, Panel data is a dataset in which the. ‧ 國. 學. behavior of entities is observed across time, which allows us to control for variables. ‧. we cannot observe or measure like difference in business practices across companies,. sit. y. Nat. or variables that change over time but not across entities. That is, it accounts for. io. er. individual heterogeneity when heterogeneity is constant over time and correlated with independent variables. From Richard Williams (2015)6, fixed effects models control. al. n. iv n C for, or partial out, the effects of time-invariant h e n g c hvariables i U with time-invariant effects.. With panel data you can include variables at different levels of analysis (i.e. students, schools, districts, states) suitable for multilevel or hierarchical modeling.. In order to choose between Random and Fixed effects model, we conduct the Hausman’s test (Greene, 2008; Gujarati and Porter, 2009). For the Hausman test, Tom 4. Cheng Hsiao. (1986) in the ‘Analysis of Panel Data’ says that ‘Besides the advantage that panel data allows to construct and test more complicated behavioral models than purely cross-sectional or time-series data, the use of panel data also provides a means of resolving or reducing the magnitude of a key econometric problem that often arises in empirical studies, namely, the often heard assertion that the real reason one finds (or does not find) certain effects is the presence of omitted (mis-measured or unobserved) variables that are correlated with explanatory variables.’ 5 Oscar Torres-Reyna. (2007), Princeton University, http://dss.princeton.edu/training/ 6 Richard Williams. (2015), University of Notre Dame, http://www3.nd.edu/~rwilliam/ 26.

(33) and Drew (2012)7 document that it aims to detect violation of the random effects modeling assumption that the explanatory variables are orthogonal to the unit effects. They further suggest that if no correlation between the independent variables and the unit effects exist, then estimates of β in the fixed effects model (β̂FE) should be similar to estimates of β in the random effects model (β̂RE). The Hausman test statistic H is a measure of the difference between the two estimates: H = (β̂RE - β̂FE)'[Var(β̂FE) – Var(β̂RE)]-1(β̂RE - β̂FE) Under the null hypothesis of orthogonality (the unique errors are not correlated. 治政 with the regressors), H is distributed chi-square with degrees 大 of freedom equal to the 立 number of regressors in the model. A finding that p < 0.05 means that the two models ‧ 國. 學. are different enough to reject the null hypothesis. Since our Hausman’s test shows. ‧. p-value < 0.0001, the fixed effect model is more appropriate as shown in the Figure 3.. n. er. io. sit. y. Nat. al. Ch. engchi. i Un. v. Figure 3 The Flow of Hausman’s Test. These processes are particularly useful. We use an ordinary least squares multivariate regression first, and then run the panel data model with two-way fixed effects to support the relation between dependent variables and independent variables.. 7. Tom S. Clark and Drew A. Linzer. (2012). 'Should I Use Fixed or Random Effects?' 27.

(34) 3.4 Model Specifications Since our thesis focuses on mutual fund investor’s buy and sell, we define the two dependent variables – Buy and Sell. As for the independent variables, there are two sets:. We run multivariate linear regression model, and two-way fixed-effect model respectively. As long as it’s the ordinary least squares regression, we add the five control variables as defined in the section ‘3.2.2 Control variables.’ But if it is the two-way fixed-effect model, we only retain the independent variables we are. 治政 interested in. Therefore, for the following models A, B,大 C, and D, Model A and Model 立 C are for the multivariate linear regression model, and Model B and Model D are for ‧ 國. 學. the panel fixed-effect model.. ‧ y. Nat. Model A. sit. n. al. er. io. 𝑩𝒖𝒚 = α0 + b1 𝐁𝐫𝐨𝐤𝐞𝐫𝐚𝐠𝐞 𝐂𝐨𝐦𝐦𝐢𝐬𝐬𝐢𝐨𝐧 + b2 𝐓𝐫𝐚𝐧𝐬𝐚𝐜𝐭𝐢𝐨𝐧 𝐅𝐞𝐞 + b3 𝐌𝐚𝐧𝐚𝐠𝐞𝐦𝐞𝐧𝐭 𝐅𝐞𝐞 + b4 𝐂𝐮𝐬𝐭𝐨𝐝𝐢𝐚𝐧 𝐅𝐞𝐞 + b5 𝐓𝐮𝐫𝐧𝐨𝐯𝐞𝐫 𝐃𝐮𝐦𝐦𝐲 + b6 𝐋𝐚𝐠 𝐏𝐞𝐫𝐟𝐨𝐫𝐦𝐚𝐧𝐜𝐞 + b7 M1B + b8 CPI + b9 CP rate + b10 BD rate + b11 𝐿𝑛TWSE + e. Ch. engchi. i Un. v. Model B 𝑩𝒖𝒚 = α0 + b1 𝐁𝐫𝐨𝐤𝐞𝐫𝐚𝐠𝐞 𝐂𝐨𝐦𝐦𝐢𝐬𝐬𝐢𝐨𝐧 + b2 𝐓𝐫𝐚𝐧𝐬𝐚𝐜𝐭𝐢𝐨𝐧 𝐅𝐞𝐞 + b3 𝐌𝐚𝐧𝐚𝐠𝐞𝐦𝐞𝐧𝐭 𝐅𝐞𝐞 + b4 𝐂𝐮𝐬𝐭𝐨𝐝𝐢𝐚𝐧 𝐅𝐞𝐞 + b5 𝐓𝐮𝐫𝐧𝐨𝐯𝐞𝐫 𝐃𝐮𝐦𝐦𝐲 + b6 𝐋𝐚𝐠 𝐏𝐞𝐫𝐟𝐨𝐫𝐦𝐚𝐧𝐜𝐞 + e. Model C 𝑺𝒆𝒍𝒍 = α0 + b1 𝐁𝐫𝐨𝐤𝐞𝐫𝐚𝐠𝐞 𝐂𝐨𝐦𝐦𝐢𝐬𝐬𝐢𝐨𝐧 + b2 𝐓𝐫𝐚𝐧𝐬𝐚𝐜𝐭𝐢𝐨𝐧 𝐅𝐞𝐞 + b3 𝐌𝐚𝐧𝐚𝐠𝐞𝐦𝐞𝐧𝐭 𝐅𝐞𝐞 + b4 𝐂𝐮𝐬𝐭𝐨𝐝𝐢𝐚𝐧 𝐅𝐞𝐞 + b5 𝐓𝐮𝐫𝐧𝐨𝐯𝐞𝐫 𝐃𝐮𝐦𝐦𝐲 + b6 𝐋𝐚𝐠 𝐏𝐞𝐫𝐟𝐨𝐫𝐦𝐚𝐧𝐜𝐞 + b7 M1B + b8 CPI + b9 CP rate + b10 BD rate + b11 𝐿𝑛TWSE + e. 28.

(35) Model D 𝑺𝒆𝒍𝒍 = α0 + b1 𝐁𝐫𝐨𝐤𝐞𝐫𝐚𝐠𝐞 𝐂𝐨𝐦𝐦𝐢𝐬𝐬𝐢𝐨𝐧 + b2 𝐓𝐫𝐚𝐧𝐬𝐚𝐜𝐭𝐢𝐨𝐧 𝐅𝐞𝐞 + b3 𝐌𝐚𝐧𝐚𝐠𝐞𝐦𝐞𝐧𝐭 𝐅𝐞𝐞 + b4 𝐂𝐮𝐬𝐭𝐨𝐝𝐢𝐚𝐧 𝐅𝐞𝐞 + b5 𝐓𝐮𝐫𝐧𝐨𝐯𝐞𝐫 𝐃𝐮𝐦𝐦𝐲 + b6 𝐋𝐚𝐠 𝐏𝐞𝐫𝐟𝐨𝐫𝐦𝐚𝐧𝐜𝐞 + e Where: α is the constant term and these variables’ data is calculated by the end of the month (t) for each fund in our scope. Besides, in order to avoid the endogenic problems, we use the turnover ‘dummy’ and set the lag ‘Fund Performance’ (t-1) in the variable’s calculation. Here is to clarify the variables in the models:. 政治大. Buyi,t = Fund Inflow (Subscription)i,t / Fund Size i, t-1. 立. Selli,t = Fund Outflow (Redemption) i,t / Fund Sizei, t-1. ‧ 國. 學 ‧. Brokerage Commissioni,t = Total Brokerage Commissioni,t / Fund Sizei,t Transaction Feei,t = Total Transaction Feei,t / Fund Sizei,t. n. al. Ch. engchi. er. io. Custodian Feei,t = Total Custodian Feei,t / Fund Sizei,t. sit. y. Nat. Management Feei,t = Total Management Feei,t / Fund Sizei,t. i Un. v. Turnover Dummyi,t is defined as: Turnover ratio of the funds is reported in percentages including purchase turnover and sell turnover, calculated on the monthly basis by SITCA. The purchase turnover and sell turnover are the fund’s purchase and sales divided by the fund’s total net asset value, respectively. We use the dummy variable marked '1' when the turnover ratio of the underlying mutual fund is greater than the turnover ratio of the market average or greater than the turnover ratio in the same category in the month. Lag Performance i,t = Fund Performance (One year)i, t-1. 29.

(36) 3.5 Limitations The main concern of conducting the thesis is that other potential factors could influence investor’s subscription and redemption. Especially, financial consultants from various channels play an important role on investor’s decision making. As well as this, marketing campaigns from banking distributions may have some effects on investor’s subscription and redemption. Still, another concern is that not all funds are available on the overall distribution, which may limit investor’s access to some funds’ purchasing.. 立. 政治大. Besides our chosen four embedded fee variables, another key fee for fund. ‧ 國. 學. investors is the front-load fee or back-end load fee. Although this fee is paid to the distribution channels by fund investors, it plays an important role for investor’s buy. ‧. and sell. However, due to the data availability, the front-load fee or back-end load fee. y. Nat. n. er. io. al. sit. is not taken into consideration in our thesis.. Ch. engchi. 30. i Un. v.

(37) Chapter 4 Empirical Results 4.1 Summary of Descriptive Statistics Table 2 reports summary statistics. The data is obtained from SITCA, ranging from January 2010 to December 2014 on a monthly basis. The average asset size of fund size, or the annual total net assets under management, is NTD 1,7billion (for fund types of AA1 and AA2). The biggest fund size is NTD 32.6billion for Allianz Global Investors Global Eco Trends Fund in March 2010 and the smallest fund size in. 政治大. the period is Mirae Asset Vigorous Small & Mid. Cap. Fund, with the fund size NTD. 立. 16million in September 2012. Besides, the monthly average size for investor’s. ‧ 國. 學. purchase is NTD 62million while the average size for investor’s redemption is up to NTD 84million.. ‧ y. Nat. io. sit. As summarized in Table 2, of the 19,906 onshore Taiwan equity domiciled. n. al. er. fund data in the period, management fee ratio of Taiwan equity mutual funds is. Ch. i Un. v. around 1.68% (0.14 %*12) of the net asset value yearly, which accounts for roughly. engchi. 56% of the total expense in the mutual fund. Brokerage fee, or the brokerage commissions, the payments made to brokerage firms for executing trades, is 0.04% monthly or 0.48% (0.04%*12) yearly on the average while transaction fee is also 0.48% (0.04%*12) on a yearly basis, the two accounting for 32% of the total expense roughly. Across the entire population of sample data, management fee, proportional to AUM, averages 0.14% monthly, or 1.68% yearly. Expense ratio is defined as the total fee divided by the net assets value of the mutual fund, which is 3% (0.25%*12) yearly on the average.. 31.

(38) Table 2 Summary of Descriptive Statistics for All Sample Funds (Monthly Data) N. Mean. Median. Std. Dev. Max.. Min.. Brokerage Fee. 19,906. 0.04%. 0.03%. 0.04%. 0.92%. 0.00%. Transaction Fee. 19,906. 0.04%. 0.03%. 0.05%. 1.33%. 0.00%. Management Fee. 19,906. 0.14%. 0.14%. 0.03%. 2.10%. 0.02%. Custodian Fee. 19,906. 0.02%. 0.02%. 0.01%. 0.21%. 0.00%. Expense Ratio. 19,906. 0.25%. 0.23%. 0.10%. 4.41%. 0.28%. Performance_1y. 19,906. 7.09%. 6.16%. 17.83%. 133.75%. -49.40%. N. Mean. Median. Std. dev. Max.. Min.. Size. 19,906. 1,738,025,465. 1,022,789,723. 2167302138. 32,648,238,777. 16,048,788. Buy. 19,906. 62,002,684. 19,199,872. 5,673,944,000. 0. Sell. 19,906. 83,968,042. 3,508,828,148. 0. 立. 166,117,259 治政大 34,860,792 146,435,576. ‧ 國. 學. In order to take a deeper look about which funds with higher turnover ratios, we sort out the data for the top 10 funds with the highest yearly cumulative selling. ‧. turnover ratios showing in the Table 3. Although a few funds are not within our scope. y. Nat. sit. (AB1 and AB2), most of the funds are the types of Domestic Equity Fund (AA1) and. n. al. er. io. International Equity Fund (AA2). To highlight, it’s also interesting to find that the. Ch. i Un. v. funds with the highest selling turnover ratios consist of investing in Taiwan or Greater China markets mainly.. engchi. 32.

(39) Table 3 Top 10 Funds with Yearly Cumulative Selling Turnover Ratios 2014 -. Category. Fund Name. Selling Turnover. Ranking. Union Golden Balanced Fund. AB1. 495.46%. 1. Fuh Hwa Greater China Fund. AA2. 459.52%. 2. Capital Emerging Market Domestic Demand Fund. AA2. 436.16%. 3. Capital Emerging Markets Fund. AA2. 414.20%. 4. Capital China New Opportunities Fund. AA2. 412.50%. 5. Fuh Hwa Greater China Mid & Small Cap Fund. AA2. 404.40%. 6. FSITC Innovation Fund. AA1. 403.20%. 7. Jih Sun Jih Sun Fund. AA1. 358.17%. 8. Paradigm Small Capital Fund. AA1. 357.50%. 9. Capital Asia-Pacific Mega-Trend Balance Fund. AB2. 352.98%. 10. Selling Turnover. Ranking. 434.72%. 1. 434.29%. 2. 428.56%. 3. 2013 -. 政治大. 立. Fund Name. Category. Jih Sun Hi-Tech Fund. AA1. Mega Global Fund. AA2. ‧. 4. y. ‧ 國. AB1. 學. Union Golden Balanced Fund. 383.43%. 6. 357.20%. 7. 355.42%. 8. 348.04%. 9. 339.91%. 10. Category. Selling Turnover. Ranking. KGI Resource Countries Fund. AA2. 499.10%. 1. Mega First Fund. AA1. 457.21%. 2. Truswell Taiwan A-Plus Fund. AA1. 454.64%. 3. Jih Sun Hi-Tech Fund. AA1. 454.32%. 4. FSITC Great China Fund. AA1. 442.81%. 5. FSITC OTC Fund. AA1. 441.35%. 6. Fuh Hwa Asia Pacific Equity Fund. AA2. 441.26%. 7. Mega Global Consumer Fund. AA2. 436.18%. 8. Jih Sun Neo Taiwan Enterprises Fund. AA1. 436.12%. 9. Schroder Taiwan Equity Alpha Fund. AA1. 425.80%. 10. AA2. Capital Asia-Pacific Mega-Trend Balance Fund. AB2. Yuanta New China Fund. AA2. io. al. CTBC Taiwan Small-cap Fund Union China Fund. 2012 -. AA1. n. Polaris 2001 Fund. Ch. Un e n g c h i AA1 AA1. Fund Name. 33. 424.76% 399.42%. 5. sit. Capital Emerging Markets Fund. er. AA2. Nat. Fuh Hwa Greater China Fund. iv.

(40) Category. Selling Turnover. Ranking. Ontario High Technology Fund. AA1. 1222.60%. 1. Union China Fund. AA1. 757.56%. 2. Polaris 2001 Fund. AA1. 636.67%. 3. Fubon Technology Fund. AA1. 615.14%. 4. Truswell Taiwan A-Plus Fund. AA1. 591.30%. 5. Union Knowledge Industry Fund. AA1. 579.90%. 6. Taishin Mainstream Fund. AA1. 573.74%. 7. Truswell Stable Growth Fund. AB1. 547.75%. 8. Truswell Taiwan Small-Cap Fund. AA1. 542.58%. 9. Mega New Emerging Enterprise Fund. AA1. 523.81%. 10. Category. Selling Turnover. Ranking. 2277.79%. 1. 806.77%. 2. 679.07%. 3. 633.41%. 4. 2011 -. 2010-. Fund Name. Fund Name. 政治 AA1大 AA1. Ontario High Technology Fund. 立. KGI Economic Moat Fund. AA1. 學. ‧ 國. KGI Big Five Taiwan Fund. Truswell Taiwan Small-Cap Fund. AA1. 625.57%. 5. Truswell Eastern Key Fund. AA2. 615.33%. 6. Hua Nan Yung Chong Fund. AA1. 614.67%. 7. 612.10%. AA1. io. Truswell Taiwan A-Plus Fund. n. al. AA1. Ch. engchi. 34. y. TLG Apollo Fund. 8. sit. AB1. 605.99%. 9. 603.72%. 10. er. Polaris Win-Win New Balanced Fund. ‧. AB2. Nat. Truswell Global Balanced Fund. i Un. v.

(41) 4.2 Correlation Coefficient With the six chosen variables, the correlation coefficients are presented to determine whether multicollinearity exists in Table 4. Table 4 Correlation Coefficients among Six Independent Variables Buy. 0.476**. 1. 0.072**. .143**. Fee. Fee. Fee. Dummy. Performance. 立. ‧ 國. 政治大 1. .698**. 1. 0.081**. .075**. .149**. .018*. 1. -0.004. -.305**. .617**. Nat. .084**. .039**. io. 0.017*. al. n. Fee. Turnover. Fee. 0.014. Fee. Custodian. Past. ‧. Management. Turnover. 學. Fee. Custodian. 1. Fee. Transaction. Management. y. Brokerage. Transaction. Ch. -0.057**. -.118**. .498**. 0.216**. .185**. -.042**. sit. Sell. 1. Brokerage. er. Buy. Sell. i Un. e n.516** g c h i -.099**. 1. v. -.266**. 1. -.101**. -.057**. Dummy. Past. 0.009. -.075**. Performance. ** Significant at 1%, two-tailed confidence level at 99% ** Significant at 5%, two-tailed confidence level at 95%. 35. 1.

(42) 4.3 Regression Results and Findings Table 5 to Table 8 summarize the associated t-statistics for brokerage commission, transaction fee, custodian fee, turnover dummy and lagged fund performance for investor’s buy and sell over January 2010 to December 2014. Comparing the two models, we find consistent evidences for the predictive power of brokerage commission, transaction fee, turnover dummy and past fund performance while custodian fee is arguably inconsistent. Another disputable variable is the management fee but based on the fund industry practice, we would support the results. 政治大. of panel fixed effect model. We show our empirical regression results below and look. 立. into each independent variable by analysis, survey, and interview.. ‧ 國. 學 ‧. Brokerage Commission. For our hypotheses H1-1.1 and H1-1.2, our result shows positive relation between. y. Nat. er. io. sit. brokerage fee and fund flows, either fund inflows or outflows, with p<0.000 as exhibited in Table 5 to Table 8. One simple explanation is related to the trading. n. al. Ch. i Un. v. frequency of buying and selling more securities. Funds incur commissions as the. engchi. results of inflow and outflow to them (liquidity trades). This supports the results of Rabarison (2015) that commissions paid by mutual fund managers are strongly influenced by investor flows. In the same vein, Lückoff (2011) shows daily gross inflows and gross outflows can only balance and offset each on the same day without incurring more costs. This is in line with the results of Chordia (1996) that fund investors trade-off diversified equity positions by investing mutual funds but absorb these direct costs for liquidity. Therefore, funds with more volatile daily flows tend to experience higher brokerage fees unless the fund size is big enough to meet clientele’s liquidity. 36.

(43) In practice for Taiwan domiciled funds, some fund companies design ‘Short-Term Trading rules’ stating the required holding period to discourage and punish excessive or short-term trading and to charge investor’s extra fees for such cases. This is also to avoid that the fees are born equally by all fund investors. In comparison, most offshore funds charge the short-term trading fees in the same way; however, offshore funds also use ‘Anti Dilution Levy’ or ‘Swinging Single Pricing’ in face of huge inflows or outflows on the day. But this practice is seldom taken by Taiwan domiciled funds.. 立. Transaction Fee. 政治大. ‧ 國. 學. The hypotheses tested (H1-2.1 and H1-2.2) based on predictions is that transaction fee is negative to investor’s Buy but positive to investor’s Sell. For our empirical. ‧. results in response to it, investor’s buy is sensitively negative to transaction related. sit. y. Nat. fees (p<0.000) in Table 5 and Table 7 while investor’s sell is not significant in face of. er. io. transaction fees by the results of multiple regression (p = 0.081) in Table 6 but it is. al. n. iv n C h e n g candhPoterba 8. The related study is done by Bergstresser i U (2002), who find that gross. significantly positive by the results of the panel fixed effect model (p<0.000) in Table. inflow is sensitive to tax burden. We therefore think that transaction related fees on mutual funds would finally reflect on and affect the fund price and then investors buy and sell. Lang (2014) highlights the main focus should lie on the tax rate charged to a fund company, irrespective of any transaction costs, since in face of higher tax costs, fund companies are likely to increase the charge of the fees on funds so that fund companies can keep their profit margin. This suggests a positive relation between fees and taxes. Based on this, we argue transaction fee and investor’s Buy are expected to be related negative while transaction fee and investor’s Sell are expected to be. 37.

(44) positive.. Management Fee In response to our hypotheses H1-3.1 and H1-3.2, there’s a strongly positive relation between management fee and investor’s Buy, with p<0.000 in Table 5 and Table 7. As for investor’s sell, management fee shows seemingly conflicting results by the multiple regression (positive) in Table 6 and the panel fixed effect model (negative) in Table 8. However, we cast doubts on the positive relation between management fee. 政治大 positive relation between management fee and investor’s buy while negative relation 立. and investor’s Sell from the multiple regression result. We would conclude the. ‧ 國. 學. between management fee and investor’s sell as following explanations. Management fee represents one of the main revenues for fund companies, which accounts for. ‧. roughly 56% of the total expense in the mutual fund based on our sample data. Our. sit. y. Nat. results point to three possible explanations for why mutual funds with significantly. n. al. er. io. higher management fees are positive to investor’s inflows: fund product signal,. v. revenue-sharing mechanism, and performance-related compensation.. Ch. engchi. i Un. Fund product signal reflects significant research costs, more monitoring efforts, better mutual fund managers’ stock-picking abilities, or superior fund strategy. For this reason, we find the evidences from some funds with higher management fee expenses disclosed in the fund’s prospectus/factsheet fee table. For examples, Nomura Middle East & Africa Fund charges 2% management fee annually and explains that the political climate in some countries is unstable and research/ monitoring is not easily achievable. Hence, more research resources and greater monitoring efforts are required. JPMorgan (Taiwan) Japan Brilliance Fund also charges 2% management fee annually, highlighting the absolute return strategy, even when share markets face 38.

(45) volatile, flat or falling situations. Still, Allianz Global Investors Global Resources Trends claims the better bottom-up skills with 2% management fee annually.. The revenue-sharing mechanism is related to the distribution channels. The most common form is trailing commissions, which is the embedded compensation for the distribution channels and it is from the management fees. Trailing commission is one of the incentives for financial consultants, who do efforts to evaluate various fund products and assist clients in selecting the funds that meet clients’ needs. By the revenue-sharing, fund companies pay the distribution channels for marketing and. 政治大. promoting support, training and development, consulting & advisory services, and. 立. selling funds. Revenue-sharing is based on the total sales of the total client assets in. ‧ 國. 學. the fund products and this incentive is calculated by asset under management (AuM) or fund inflows. Therefore, the more asset under management and/or fund inflows. ‧. come, the more revenue-sharing generates.. sit. y. Nat. n. al. er. io. The performance-related compensation is usually linked to the fund size. The. i Un. v. rationale for the performance-related compensation is that it aligns the interests of. Ch. engchi. both fund managers and fund investors, and that it plays an incentive for fund managers to generate positive returns to attract more fund inflows. As long as the fund size reaches the threshold, the fund manager’s compensation would be marked-up, which works like progressive mechanism reflecting the actual fund inflows generated from exceeding a fund size (or asset under management). This kind of compensation usually means that investment funds are managed or consulted by third-party investment advisers. Hence, the bigger a fund size is, the more performance-related compensation wins. Our empirical evidence is also in line with the result of Berk and Green (2004), which demonstrates that mutual fund manager costs are an increasing. 39.

(46) function of the amount of funds under management.. Based on the above three explanations – fund product signal, revenue-sharing mechanism, and performance-related compensation, while considering fund's individual-specific effect and time-specific effect, we argue that ongoing charges of management fee is positive to investor’s buy but negative to investor’s sell.. Custodian Fee Our hypotheses tested (H1-2.1 and H1-2.2) seem to be arguably inconsistent as. 政治大. shown in Table 5 to Table 8. The independent custodian aims at the double mission of. 立. safekeeping the fund assets and making sure fund portfolios in compliance with the. ‧ 國. 學. law. Ideally, this indirect cost should not have a direct effect on investor’s buy and sell. ‧. as we hypothesize. However, the multiple regression shows the results that custodian fee is significantly negative to both investor’s buy and sell with p<0.000. On the flip. y. Nat. er. io. sit. side, the panel fixed effect model tells that custodian fee should be irrelevant to influence investor’s Buy (p=0.703) and Sell (p=0.708). Considering the practices in. n. al. Ch. i Un. v. Taiwan fund market, we argue that, from the multiple regression, the evidence on. engchi. custodian fee and investor’s buy and sell presents a somewhat nonstandard and unexpected portrait. Overall, our empirical analysis does not find consistent evidence that custodian fee matters for mutual fund investors.. Turnover Dummy Our alternative hypotheses (H2-1 and H2-2) with p<0.000 document that turnover ratios of mutual funds have both negative effects on investor’s buy and sell in Taiwan domiciled equity fund market as evidenced in Table 5 to Table 8. This corresponds to the fact that investors prefer to purchase the funds with buy-and-hold 40.

(47) strategy with lower turnover ratios.. If we look into the turnover ratio, something interesting could be found that the yearly cumulative top 10 selling turnover ratios consist of the most mutual funds investing in Greater China or Taiwan markets as exhibited in Table 3. One explanation is familiarity effect or home bias. This is maybe because of language, culture, and distance so that fund managers prefer to trade familiar companies, which echoes the results of Grinblatt and Keloharju (2001) in Finnish markets and Huberman (2001) in the United States markets.. 政治大 In order to have further checking for the turnover ratio on investor’s buy and 立. ‧ 國. 學. sell, we extract the data with turnover ratio dummy marked ‘1’ and run the simple regression below for the analysis:. ‧. Performancet =α+ Selling Turnovert + e. sit. y. Nat. n. al. er. io. The results show the t-statistics is -9.094 with p<0.000. This documents the. i Un. v. evidence that, for the funds with turnover ratios above the market average, mutual. Ch. engchi. funds have not been able to generate enough returns to compensate their costs. However, despite the above negative results of funds with selling turnover affecting performance returns, there exists the possibility that a few funds may be skilled enough to consistently perform well so their turnover costs are compensated as highlighted by Rubio (1995).. Past Fund Performance The alternative hypotheses (H3-1 and H3-2) state that mutual fund past performance does stimulate mutual fund investor's buy and sell. Table 5 to Table 8 41.