A thesis submitted to the University of Mysore, Mysore, through the Department of Studies in Economics and Cooperation, University of Mysore, Mysore, For the Award of the Degree of Doctor of Philosophy in Economics By T.S. SOMASHEKAR Under the supervision of Dr. S. Madheswaran Institute for Social and Economic Change, Baogalore December 2008 DEDICATED TO MY WIFE, KAVITHA FOR ALL THE MORAL SUPPORT DECLARATION I hereby declare that the thesis entitled "Performance and Regulation of Mutual Funds in India: An Economic Analysis" is a result of research work carried out by me at the Institute for Social and Economic Change (ISEC), Bangalore, under the guidance of Dr.S.Madheswaran, Professor, Centre for Economic Studies and Policy, ISEC, Bangalore. I further declare that it has not been previously submitted, in part or full, to this or any other university for any degree. Due acknowledgements have been made whenever anything has been borrowed or citied from other sources. Research Scholar Centre for Economic Studies and Policy (CESP) Institute for Social and Economic Change Bangalore - 560 072 India INSTITUTE FOR SOCIAL AND ECONOMIC CHANGE Dr V K R V Rao Road, Nagarabhavi P.O., Bangalore - 560072 Fax: 080-23217008 website: www.isec.ac.in Email: madhes@isec.ac.in An all India institute for inter.<fisciplinary research and training in the Social Sciences CERTIFICATE This is to certify that the thesis entitled RPeriormance and Regulation of Mutual Funds in India: An Economic Analysis" . submitted by T.S.Somoshekar in fulfillment of the requirements for the award of the degree of DOCTOR OF PHILOSOPHY in ECONOMICS is an original work. I also certify that it has not been previously submitted for the award of any degree or diploma or associate fellowship of the University of Mysore or any other University. S. y.JL.....y
Supervisor Professor Centre for Economic Studies and Policy Institute for Social and Economic Change Bangalore -560 072 India ACKNOWLEDGEMENT It is with a sense of achievement and pride that I complete my thesis. It has been made possible by the unstinting support of many people. These few words only convey my formal thanks but cannot convey the depth of my gratitude to them. I have to begin by thanking my supervisor Prof Dr. S Madheswaran, under whose able guidance this work has been carried out. His constructive criticism, patience in going through my work, simplifying suggestions, constant encouragement and rock solid moral support has gone a long way in helping me complete my thesis. I am grateful to him for this. I would like to express my thanks to my Doctoral committee members ProfMR.Narayana and ProfMeenakshi of ISEC, for sparing their valuable time in going through my draft and providing constructive invaluable suggestions and moral support. Sincerely acknowledge their help. I am grateful to the previous directors, ProfMGovinda Rao, ProfGopal K.Kadekodi, ProfNJayaram and the present director ProfDeshpande of the Institute for Social and Economic Change (ISEC), for their support and encouragement. I am deeply indebted to the Institute for Law and Economics, University of Hamburg for selecting me and providing me an opportunity to pursue my doctoral research. I express my sincere gratitude to my guide at the Institute, Prof A Domadoran, Indian Institute for Management, Bangalore for their patient guidance in my quantitative analysis. I also sincerely acknowledge the valuable comments and suggestion of ProfShashank Bhide, ProfMRamachandran, Ms. B. P. Vani, Assistant Professor, ISEC, ProfNSSNara The important role of a fund manager Mr.Harsha, in UTI Mutual Fund Ltd, in terms of fund manager realities has been of immense help. lowe him my gratitude. My Sincere gratitudes are due to ProfS N Sangita, Dr. Mahadeva, Dr.K.G.Gayathri Devi, Dr.MD.Usha Devi and Dr.Anand Inbanathan for their encouragement and held during the course of all my biannual seminars in ISEC. I am also thankful to other faculty members of ISEC for their encouragements. My special thanks to Mr.K.s.Narayanafor helping me with the editorial work and tremendous support at different stages of my thesis work. To Mrs.Susheela the efficient typist lowe a lot of gratitude. 1 acknowledge the assistance received from the computer staff Mr. Krishna Chandran and Mr.Satish kamat and also the ISEC library staff, Mr. Sharma, Mrs. Lila, Kalyanappa, Suresh and Rudresh. Further, 1 express my sincere gratitude to the University of Mysore, faculty members at Department of Economics and staff for aI/owing me to register for Ph.D and their valuable suggestions and kind cooperation. 1 take this opportunity to thank my fellow Ph.D students and friends Badrinarayan Rath, Mahesh, Nisha, Subir and Venugopal, who were always there in times of need. I am deeply indebted to my parents and my wife for aI/ the moral support and encouragement in my academic pursuits. T.S.Somashekar CONTENTS Chapter I - Introduction 1.1. The concept of Mutual Funds 1.2. Organization of Mutual Funds in India 1.3. Types of Mutual Funds 1.4. The Mutual Funds Industry in India 1.5. Statement of the Problem 1.6. Review ofTheoretical Studies 1.7. Review of Empirical Studies 1.8. Objectives of the Study 1.9. Analy1ical Tools 1.10. Relevance of the study 1.11. Scope of the Study 1.12. Data Description, Sources and period of Study 1.13. Chapter Scheme of the Study Chapter II - Review of Literature 2.1. Introduction 2.2. Theoretical Review 2.2.1. Theories of Regulation 2.2.2. TheOries of Risk Measurement 2.3. Empirical Review 2.3.1. Review of Studies on Mutual Funds performance 2.3.2. Empirical studies on Indian Mutual Funds 2.3.3. Empirical Studies on Fund Size and Returns 2.3.4. Fund Style and Fund Returns 2.3.5. Empirical studies on Indian Mutual Funds Page No. 1 - 28 29-47 Chapter III - Performance of Equity Funds using Variance and Semi-Variance of Risk 3.1. Introduction 3.2. Empirical framewor1< and Data description 3.2.1. Empirical framewor1< 3.2.2. Data 3.3. Results and Analysis 3.4. Conclusions Chapter IV - Regulatory Constraints And Fund Performance - Comparing Alternate Regulatory Regimes 4.1. Introduction 4.2. Comparison of the Regulatory structure of SEBI (Regulations) and UTI Act (1963) 4.2.1. UTI Act of 1963 and ~ s constraints on fund behaviour 4.2.2. SEBI Regulations and its constraints on fund behaviour 4.3. Empirical framework, data description and methodology 4.3.1 Empirical framewor1< 4.3.2. Data 4.4. Results and Analysis 4.5. Chapter Summary Page No. 48-58 59-86 Chapter V - Cost-Management Fee Regulation and Economies 87 -107 of Scale 5.1. Introduction. 5.2. Regulation of costs and management fees 5.3. Empirical framewor1<, Data description and Methodology. 5.3.1. Data 5.3.2. Empirical Framewor1< 5.4. Results and analysis 5.5. Cost composition: Management fees, mar1<eting expenses and other expenses 5.S. Chapter Summary Chapter VI - Fund Size and Fund Returns 6.1. Introduction 6.2. Empirical framework and data 6.2.1. Data 6.2.2. Empirical framework and model 6.3. Resuijs 6.4. Chapter Summary Chapter VII - Summary of Findings and Policy Suggestions 7.1. Summary 7.1.1. Relative Fund Performance Using Altemate Risk Weighted Returns: Mean-Variance and Lower Partial Movement 7.1.2. Regulatory Constraints and Fund Performance- Comparing Alternate Regulatory Regimes 7.1.3. Cost- Management Fee Regulation and Economies of Scale 7.1.4. Fund Size and Fund Returns 7.2. Lim itations of the Study 7.3. Further Scope for Study 7.4. Policy Suggestions and Concluding Observations References Page No. 108 - 121 122 - 132 133 -138 III List of Abbreviations NAV Net Asset Value AUM Assets Under Management SEBI Securities and Exchange Board of India GIPS Global Investment Performance Standards TWR Time Weighted Return EXR Excess Return SR Sharpe Ratio IR Information Ratio TE Tracking Error SOR Sortino Ratio MAR Minimum Acceptable Return CAPM Capital Asset Pricing Model UTI Unit T rust of India OEG Open end Equity Diversified Growth Funds CIU Collective Investment Vehicle CIS Collective Investment Schemes AMC Asset Management Company VCF Venture Capital Fund IPO Initial Public Offer TM Treynor Mazuy ELSS Equity Linked Savings Scheme VCF Venture Capital Funds RBI Reserve Bank of India 137 AMC Asset Management Company FT Franklin Templeton BSE Bombay Stock Exchange ETF Exchange Traded Fund's US Unit Schemes LlC Life Insurance Corporation of India GIC General Insurance Corporation of India \38 PgNo. Chapter I - Introduction Table 1.1: Comparative AUM as a percent of GDP (CMP) in 2004 11 Chapter III - Performance of Equity Funds Using Mean-Variance And Lower Partial Movement Table 3.1: Sample of Equity Diversified Funds 52 Table 3.2: Sharpe Ratio ranking of funds 53 Table 3.3: Treynor Index Rankings fund 54 Table 3.4: Summary statistics of the normality test for returns of sampled 56 funds Table 3.5: Ranking of funds based on the Sortino Ratio 57 Table 3.6: Spearman Rank correlation coefficient 58 Chapter IV - Regulatory Constraints And Fund Performance- Comparing Alternate Regulatory Regimes Table 4.1: Differences in the Regulatory systems 70 Table 4.2: Sample of funds 75 Table 4.3: Sharpe ratio rankings for sampled funds 77 Table 4.4: Information ratio (IR) rankings for sample funds 79 Table 4.5: Poolability test results for SEBI regulated funds 80 Table 4.6: Hausman Test for SEBI regulated funds 81 Table 4.7: Poolability test results for UTI funds 81 Table 4.8: Average Jensen alpha for funds regulated by SEBI 82 Table 4.9: Average Jensen alpha for UTI funds 83 Table 4.10: Average Assets and Management fees of a pure equity funds 84 Table 4.11: Average Jensen alpha (with average managementtee 85 included) for SEBI regulated funds Chapter V. Cost Management Fee Regulation and Economies of Scale Table 5.1: Descriptive statistics for Equity Funds 92 Table 5.2: Descriptive statistics for debt funds 92 Table 5.3: Cost function and explanatory variables 94 Table 5.4: Time series, cross section averages of (yearly) correlations 95 between variables of interest for equity funds. Table 5.5: Time series, cross section averages of (yearly) correlations 96 between variables of interest for debt funds. Table 5.6: Correlated Random Effects - Hausman Test 99 Table 5.7: Regression results without the control variable for equity funds. 99 Table 5.8: Regression results with control variable for equity funds. 99 Table 5.9: Correlated Random Effects - Hausman Test 101 Table 5.10: Regression results without the control variable for debt funds 101 Table 5.11: Regression results with the control variable for debt funds 101 Table 5.12: Time series cross section averages of expense components of 103 equity funds Table 5.13: Time series cross section averages of expense components of 105 debt funds. Chapter VI Fund Size and Fund Returns Table 6.1: Fund style deSCription Table 6.2: Time Series cross sectional averages and standard-deviations Table 6.3: Time series averages of the cross sectional correlations between the various fund characteristics Table 6.4: Means and standard deviations for the monthly fund returns Table 6.5: Mutual fund Beta's and constants across different asset classes with CNX500 as the benchmark index. Table 6.6: Correlated Random Effects - Hausman test 110 111 112 113 116 117 Table 6.7: Regression of Beta and market adjusted fund perfonnance on 117 fund size and other characteristics over one period lagged effects Table 6.8: Regression of market adjusted returns (in bear & bull phases of the market) on lagged fund size and other fund characteristics. Table 6.9: Regression of beta adjusted returns (in bear & bull phases of the market) on lagged fund size and other fund characteristics. 119 119 Chapter I - Introduction Chart 1.1: Concept of Mutual Funds Chart 1.2: Organization of Indian mutual funds Chart 1.3: Types of Mutual Funds Chapter III - Performance of Equity Funds Using Mean-Variance And Lower Partial Movement 02 05 06 Chart 3.1 Trends in CNX500 for the sample period. 51 Chart 3.2 Presents a test for normality in distribution of returns for the LlC 55 Tax Plan Chapter V - Cost Management Fee Regulation and Economies of Scale Chart 5.1 Regulatory ceilings for Expense Ratio and Management Fee (as 90 percent of AUM) for equity and equity oriented balanced funds CHAPTER -I INTRODUCTION CHAPTER ONE INTRODUCTION 1.1. The concept of Mutual Funds Mutual funds are that collect money from several sources - individuals or institutions by issuing 'units', invest them on their behalf with predetermined investment objectives and manage the same all for a fee. They invest the money across a range of financial instruments falling into two broad categories - and debt. Individual people and no doubt, can and do invest in equity and debt instruments by themselves but this requires time and skill on both of which there are constraints. Mutual funds emerged as professional financial intermediaries bridging the time and skill constraint. They have a team of skilled people who identify the right stocks and debt instruments and construct a portfolio that promises to deliver the best possible 'constrained' returns at the minimum possible cost. In effect, it involves outsourcing the management of money. More the benefits of investing in and debt instruments are supposedly much better if done through mutual funds. This is because of the following reasons: Firstly, fund managers are more skilled. They are trained to identify the best investment options and to assess the portfolio on a continual basis; secondly, they are able to invest in a diversified portfolio consisting of 15-20 different stocks or bonds or a combination of them. For an individual such diversifICation reduces the risk but can demand a lot of effort and cost. Each purchase or sale a cost in terms of brokerage or transactional charges such as demat account fees in India. The need to possibly sell 'poor' stockslbonds and I buy 'good' stocksJ bonds demands constant tracking of news and performance of each company they have invested in. Mutual funds are able to maintain and track a diversified portfolio on a constant basis wijh lesser costs. This is because of the pecuniary economies that they enjoy when it comes to trading and other transaction costs; thirdly, funds also provide good liquidity. An investor can sell her/his mutual fund investments and receive payment on the same day with minimal transaction costs as compared to dealing with individual securities, this totals to superior portfolio returns with minimal cost and better liquidity. This can be represented with the following flow chart: Chart 1,1 lund Securtties Source: Association of Mutual Funds in India (AMFI) 2 In India one can gain additional benefrt by investing through mutual funds- tax savings. Investment in certain types of funds such as Equity Linked Tax Savings Schemes (ELSS) allows for certain amount of income tax benefits.' 1.2. Organization of Mutual Funds in India Mutual Funds are one form of Collective Investment Vehicles (CIV's) in India. The other forms being Collective Investment Schemes (CIS's) and Venture Capital Funds (VCF's).The organization of mutual funds in India (excepting for Unit Trust of India)2 is dictated by the Securities and Exchange Board of India - SEBI (Mutual Funds) Regulations, 1996, (henceforth termed as 'regulations'). Bank-owned mutual funds are also supeNised by the ReseNe Bank of India (RBI). This does not overlap with SEBl's supeNision. Besides, the Indian Companies Act of 1956 and Indian Trust Act of 1882 also govem funds. The SEBI regulations stipulate a three tier structure. The constituents of a mutual fund include the Sponsors, the Mutual Fund, the Trustees and the Asset Management Company (AMC). The Sponsor as an individual or along with another corporate body initiates the process by approaching the SEBI for registration of a mutual fund. There are certain eligibility criteria which the sponsor has to fulfill [as laid out in Chapter II, clause 7 of the regulations). Broadly speaking it requires a sound financial track record over the past fIVe years, a sound reputation with respect to integriity and a minimum of 40 percent stake in the AMC by the sponsor. For instance, Tata Mutual I Section 80 C of the Indian Income Tax Act allows for Income Tax exemptions uplO a maximum of Rs.lOO,OOO. 2 Unit Trust of India (UTI) is governed by the Government of India UTI Act of 1963 Fund set up in 1995 is sponsored by Tata Sons Limited and Tala Investment Corporation Lim ited. 3 The Mutual fund is itself set up in the form of a trust under the Indian Trust Act of 1882. The instrument of the trust is executed by the sponsor in favour of trustees and is registered under the Indian Registration Act, 1908. The investor subscribes to the 'units' of the fund and, the collected funds/assets are held by the trstee for the benefrt of the investor. The Sponsor then appoints the Trustees, AMC and Custodian [Chapter II, clause 7 (e), (I), (g) of the regulations]. A Trustee holds the fiduciary responsibility of protecting the interest of the investor. The trustees themselves are to be of impeccable personal credentials [Chapter II, clause 16 (2)]. Two thirds of them have to be independent persons and should not be associated with the sponsors in any manner'. No employee of the AMC is to be a part of the Trustee. The Trustee has the duty of ensuring that the AMC carries out its activities in accordance with the regulations and prevent conflict of interest between the investors and the AMC. The trustee could be either a group of individuals or a Trust Company. Most funds prefer a board of trustees. The Trustee for Tala Mutual Fund is Tala Trustee Company Pvt. Ltd. The AMC consists of the fund managers who manage the investments Regulations are laid down [Chapter IV] with regard to their eligibility and obligations. The AMC takes decisions with respect to investment/sales, computing asset values, declaring dividend and providing investor information, regularty. An AMC cannot 3 www.tatamutualfund.comlrisk-factors.asp 4 Earlier only 50 percent were required to be independent members. This was amended in 2006 by the SEBI (mutal funds) (Fifth Amendment) regulations. act for any other fund. The AMC for Tata mutual fund is Tata Asset Management Ltd. In addition to the above three principal constijuents there is the custodian (Chapter IV, clause 26) predominant duties includes stock keeping of securities and settlement between funds. A custodian can service more than one fund but not a fund promoted by a sponsor who has 50 percent stake or more in the custodian. For Tata Mutual Fund, ABN AMRO Bank N.U and Deutche Bank are the custodians. Apart form these there are the depositories, transfer agents and distributors who complete the organizational chain for mutual funds in India. chart: The organization of Indian mutual funds is represented in the following flow Chart 1.2 Sponsors I The organization conceived as above has ijs emphasis on eliminating moral hazards that could arise out of post contractual opportunistic behaviour on the part of fund managers. It aims at ensuring arms length transactions between the sponsor and the AMC. 5 1.3. Types of Mutual Funds A mutual fund, say, Tata Mutual Fund, can have several 'funds' [called 'schemes' in India) under its management. These different funds can be categorized by structure, investment objective and others. It would be well illustrated by the following flow chart: ! I A. Structure I ! 1 open close Internal ended ended funds funds funds Chart 1.3 Types of Mutual Funds B. Investment Objective ! ! ! Growth Income Balanced funds funds Exchange Traded funds funds Tax saving funds Index funds Source: Associalion of Mutual Funds in India (AMFJ) MOnlY market fund. c. Others I Special Schemes Sector specific funds An 'Open end' fund is available for purchase or redemption on continuous basis at the day's closing Net Asset Value (NAV). This gives liquidity to investments. 6 A 'Close end' fund is open for investment only during the Initial Public Offer (IPO) after which the investment is locked in until the maturity date which could be between 3-7yrs. The investor can, however, sell or buy the shares of the funds on the stock exchange where the shares are listed. Interval funds combine the characteristics of both 'open end' funds. They can be bought or redeemed by the investor at predetermined times, say once in six or twelve months. 'Growth' oriented funds aim at providing capital appreciation. They tend to invest primarily in equities. 'Income' funds aim at providing regular income to investors. They generally invest a major portion of their assets in fixed income earning instruments such as govemment securities, corporate bonds and money market instruments. Their retums are determined by fluctuations in interest rates. A 'Balanced fund' tries to provide both capital appreciation and regular income. They invest in both equities and fixed income securities. They specify the maximum equity exposure in the prospectus and is normally 60 percent; of late other types of balanced funds such as "Asset Allocation funds and 'Arbitrage funds' have also emerged. Asset allocation funds, such as the Franklin Templeton (FT) PIE ratio fund, allocate funds to equity or debt depending on the dynamic situation. They tend to increase exposure to equity during a market downturn and move out during market peaks. The FT PIE ratio fund uses the market PIE ratio to determine the degree of equity exposure. 7 Arbitrage funds are funds that try to capitalize on the arbitrage opportunities that arise out of pricing mismatch of stocks in the equity and derivative (Mures and options) segments of the stock market (Value Research Inc). They invest predominantly in equities 'Money Marker. Funds invest only in short term debt such as call money, treasury bills and commercial paper. In the case of these funds the Net Asset Value is simply the interest accrued on these investments on a daily basis. Their NAV does not fall below the initial investment value, unlike bond funds which are marked to market. Tax saving funds give an investor tax benefits under section 80 C of the Income Tax Act. Such funds also termed as Equity Linked Saving Schemes (ELSS), have a lock in period of three years. By investing in such funds a person can avail of a maximum of rupees one hundred thousand in tax deductions. ELSSs are normally diversified equity funds. Index funds invest in securities of a particular index such as the Bombay Stock Exchange (BSE) sensex in the same proposition. They provide returns which are close to that of the benchmark index with similar risks as well. It is a passive investment approach with lower costs. Sector specific funds focus their investments on specific sectors which the fund manager feels would do well. For instance, Franklin FMCG fund invests only in shares of companies that produce fast moving consumer goods. Exchange Traded Fund's (ETF) are relatively a new concept in India. Such funds are essentially index funds that are listed and traded on the stock markets. There are also commodities ETFs such as Reliance hold ETF. 8 1.4. The Mutual Funds Industry in India The beginning of mutual funds in India was laid by the enactment of the Unit Trust of India (UTI) Act in 1963. The objective was to provide investors from the middle and lower income groups with a route to invest in the equity market. It was also meant to encourage savings. UTI brought out its first fund, Unit Scheme (US) 64 in 1964. It called an amount of Rs.246.7 millionS. UTI remained a monopoly in the mutual fund industry till 1987. By then US 64 had grown to Rs.32.69 billion and the overall asset base of UTI was RS.67.38 billion with 25 different schemes. In 1987 other public sector banks were allowed to offer mutual funds. The State Bank of India (SBI) set up the SBI Mutual Fund and Canara Bank Mutual Fund. Other public sector banks such as Bank of India, Punjab National Bank, Indian Bank entered the fray by 1990. Two public sector insurance companies - Life Insurance Corporation of India (LlC) and General Insurance Corporation of India (GIC) also started their own mutual fund companies. But during this period only public sector companies were permitted to enter the mutual fund market. The collective assets under management continued to grow and by the end of 1993 it was Rs.470 billion with UTI alone accounting for RS.390 billion>' There were 44.7 million investors in mutual funds". 1992-93 saw the beginning of economic reforms in India. The reforms aimed at reducing government control over the economy and allowing for greater play for the private sector besides others. In keeping with tihis direction the private , Mutual Funds in India. 19641994, pp 12 UTI Annual report 198788 7 Mutual Funds in India, 1964-94, UTIICM pp 13 8 Mutual Funds in India, 1%494 pp 13 9 sector was allowed to enter the mutual fund industry in 1993. In keeping with this direction the private sector was allowed to enter the mutual fund industry in 1993. In the same year the first mutual fund regulations 1993 SEBI (mutual fund) Regulations came into being. This was later substituted by a more comprehensive set of regulations - SEBI (mutual fund) Regulations 1996. However, UTI did not come under these regulations and continued to be governed under the UTI Act of 1963. By 2003 the total assets under management (AUM) had increased to Rs.1 ,218 billion w ~ h 33 mutual fund families and 401 funds. UTI alone accounted for Rs.445 billion of the total AUM 9 . In 2003 the public sector UTI, which had faced serious problems in the late 90's and again during 2002, was s p l ~ into two entities. One was the specified undertaking of UTI which managed US 64, assured retum schemes and others which totaled to Rs.298.4 billion and the other was UTI Mutual Fund Ltd'o. The latter came under the regulations of SEBI. Since 2003 the mutual fund industry has also seen a spate of mergers. Hence this period was marked by consolidation. By March 2007 the total AUM excluding UTI touched Rs.3,591 billion showing a phenomenal growth of 47 percent year-on-year since 2003". During this period only Russia and China did better than India w ~ h AUM growth rates of 97 percent and 67 percent, respectively". , AMFI 10 AMFI monthly March 2007 " AMFI 12 Indian Asset Management: Achieving Broad Based Growth - Mc Kinsey India 10 1.5. Statement of the Problem The financial savings of the households in India and the savings of the private corporate sector form the main source of funds for the mutual fund industry. The gross financial assets of Indian households increased from Rs.l09.6 billion (10.4 percent of the GOP at CMP) in 1993-94 to Rs.4,176.8 billion (14.85 percent of GOP at CMP) in 2003-0413. The gross financial assets include currency held, bank and non-bank deposits, life insurance, provident and pension funds, claims on the govemment, shares and debentures, investments in Unit Trust of India and the net trade debt. Bank deposits comprised of 42.83 percent of the total financial assets and shares and debentures (which include mutual fund investments) forming a small 1.81 percent (EPW). The total AUM at the end of March 2004, Rs.l,396.16 billion was 4.96 percent of the GOP (CMP). A comparison of India with other countries is given below: Table 1.1 Comparative AUM as a percent of GOP (CMP) in 2004 Country AUM in $.Billion U.S 7,414.4 Brazil 171.6 Korea 121.66 India 29.8 Russia 0.85 Data from Investment Company Institute 2004, fact book GDP data from IMF world economic outlook databases %ofGOPCMP 67.5 34.4 20.08 5.2 0.2 The above table shows the vast potential for growth in the mutual funds industry. A minuscule 5 percent of the GOP (CMP) was invested in mutual funds as "Economic and political weekly, Oct. 09,2004, pp 4487 11 compared to 67.5 percent in the U.S. The comparison is to show what the growth potentials are. In many ways the mutual fund industry can be termed to be still in its infancy. A SEBI survey of Indian Investors 14 for the period April 01, 1999 to March 31, 2001 revealed the low household penetration rate for mutual funds. The survey found that 7.4 percent of Indian households (13.7 percent of urban and 3.8 percent of rural households) had invested in mutual funds. Even among the urban households most investors were from the largest cities with a population of 5 million or more. The facts point to a low household penetration rate by mutual funds and also a very narrow urban bias. For individual investors direct investment in equity was a risky proposition and an important deterring factor as per the survey. Mutual funds have potential to offer a safer route to the vast untapped households that still seem to prefer bank savings. But to use the mutual fund route there are other concerns which need to be addressed. Firstly funds have to deliver in terms of performance. Comparisons are bound to arise particularly between fund return and benchmark indices are to be expected. Out performing benchmarks with lower costs is an important factor to attract more savings into mutual funds. Apart from performance there is the issue of moral hazards. Once a contract has been entered into with the fund house there are risks of conflicting interests. These risks could be broadly classified into portfolio selection risks and management process risks. The former involves 'adverse portfolio selection' which contradicts the objective of the fund as mentioned in the prospectus. This could mean higher risk of the fund portfolio and l+rhrust on Risk Containment. Opinion, Business Line, 2 ~ October 2000 12 or lower risk-weighted returns. There could also be excessive churning of the portfolio leading to more expenses that would be deducted from the AUM. Management process risks involve risks arising due to errors in execution of transactions losses due to counter-party default. It is to protect investors against such risks that SEBl's (mutual fund) Regulations of 1993 were framed. The same was substantially amended in 1996. Given the growth of the mutual fund industry, it's present and potential importance as a vehicle of financial saving for Indian households, and the development of regulations to govern fund behaviour we feel it is important to assess the role of regulation in adding value for the investor. Have the regulations ensured due diligence, transparency and sound portfolio selection? Have the dynamics of the fund industry led to the necessity for change in regulations? Is the present form of fund performance information dissemination adequate? These are some of the questions that the present study attempts to answer. This is sought to be done by examining the ability of regulations to: ensure proper performance disclosure; better returns from funds; control costs of operation; prevent excessive management fees and be proactive in tackling new issues. 1.6. Review of Theoretical Studies A brief outline of the theoretical literature surveyed is presented here to fix the basis for the study of risk involved in mutual fund returns and the need for regulating mutual funds. The study firstly reviews the theOries of regulation. The works of Stigler (1971) and Posner (1969) discuss the general theoretical approaches to regulation. \3 The justification of mutual fund regulations stems from asymmetric information leading to possible investor - manager conflict of interest. To control for such behaviour in 'public interest' the regulator might seek to use different approaches including direct price control and or disclosure norms. The study reviews literature relating to measure of risk in the case of equity returns. It looks at both the Mean-Variance (M-V) approach and the lower partial movement (LPM), while the M-V approach assumes that preferences are based only on the mean returns and the variance of the fund portfolio. It is appropriate only if the returns are normally distributed. During the 70s a semi variance measure of risk known as LPM was developed. It gave a better approach towards measuring the risk-return combination. Writings on the single factor capital asset pricing model (CAPM) which uses the M-V approach to risk and aHemative multi-factor models used to measure portfolio performance are also reviewed. In the process it looks at the efficient market hypothesis and the theoretical impossibility of earning more returns than an informationally efficient market. The study then moves on to theories of regulations and the necessity for the same to tackle Principal-Agent problems that arise due to information asymmetries. Regulation ought to be dynamic changing according to market conditions as the fund industry gets competitive the necessity of regulations might wane. In fact it might be a bane as regulations have an impact on performance and increase costs. Posner speaks of the costs of regulations generally. 14 To sum up, the basis of regulating mutual funds appears to be from a public interest perspective to control hazards to investors arising from mar1<et imperfections. Further the need to have appropriate risk-reward measures emerges clearly. 1.7. Review of Empirical Studies Here, again, a brief outline of the empirical literature surveyed is presented. The aim is to show the developments in methodology and the gaps in studies conceming India. Empirical studies pertaining to mutual fund performance can be grouped into single factor CAPM which uses a single benchmar1< and multiple factor CAPM. Some of the predominant single factor CAPM studies were those of Sharpe (1966), Treynor and Mazuy (1966), Jensen.M (1968). But, Ross (1976) argued that systematic risk need not be explained by a single factor and that there could be 'K' factors. Hence the basis for multi factor models for assessing performance. Several factors such as retum of small cap stocks, large cap stocks, midcap stocks, growth stocks, value stocks and momentum are included in various studies. For instance, Fama and French (1993) used multifactor model with the mar1<et retum, the return of small less big stocks (SMB) and the return of high booklmar1<et, less low booklmar1<et stocks (HML) as three important factors. Using multiple factors eliminates the error that arises out of the assumption of homogeneity of assets held by different portfolios or funds. The multiple factor models recalculate the Jansen alpha which measured superior performance. 15 Apart from these two broad approaches to performance, empirical worll on other special factors influencing performance analysis is reviewed. Studies on the influence of size on performance such as Grinblatt and Titman (1989), Indro et al (1999) and Chen at al (2003) are some who examine the role of size and diseconomies of scale/all in the U.S context). The results are however mixed with no clear consensus on diseconomies of scale. With respect to India there has been no attempt to study the possibilities of size affecting returns. Another factor which has been found important is the style of a fund. Sharpe (1992), Grinblatt et al (1995), and Bogle.J (1998), discuss the role of style. Style describes the asset class of the portfolio of a fund. This explains a large part of the funds return variability. Studies evolved from the holdings based style analysis (HBSA) method (categorization of funds on the basis of average marXet capitalization and average price-to-eamings of the fund portfolio) to the returns based style analysis (BSA) classifICation (compares fund returns to returns of a number of selected indexes) While taking samples of funds for assessing their performance the survivor bias has to be considered. Funds tend to close or merge with others, at times this covers up for poor performance. So choosing funds which survive might tend to bias performance analysis. Several funds in India had been terminated or merged and hence this factor has to be considered while assessing performance. Since the overwhelming emphasis of regulations in India is on direct controls on fees and expenses that are charged to fund, investors' empirical studieS relating to this issue have been reviewed. In the relationship between fund expenses and performance Sharpe (1966) and Ippolito (1989) throw up different conclusions. 16 On the relationship between fund size and expenses or economies of scale and scope Baumol et al (1989), Rea (1999), Latzko (1998) establish economics of scale and Baumol also finds economics of scope. The Indian context w ~ h regulatory caps on fund manager fees and expenses and ~ s impact on constraining fund expenses has not been dealt with so far. Studies on Indian mutual fund performance by Sahadevan and Thiropol Raju (1997) and Sadhak (1997) focus on general trends in the mutual fund performance, regulation and expenses from 1990-91 to 1996. Their focus was not on the evolved risk return analysis. Madhu S Panigrahi (1996) and Bijan Roy et al (2003) have attempted risk-return analysis and using traditional CAPM and conditional performance evaluation techniques respectively. The review of empirical literature points out to the role of benchmarks and style of funds in deciding fund performance. Studies also cover fund behaviour in terms of management fee and expenses. These help point to the general issues of focus concerning mutual funds. The Indian studies, it is found, tend to focus on using evolving techniques to study of fund performance. But there is no attempt to study the impact of regulations on fund behaviour in terms of expenses, fees and performance. This gap in Indian studies needs to be seriously considered. It gives us the motivation for our study. We need to study the past implications of fund regulations before we go ahead with further changes in the same. Have the costs of regulations eXCeeded the benefrts? Can we improve regulations to ensure better governance? Do we notice tendencies of price competition between funds? Do they deliver more than what regulations demand in terms of cost charged to investors? These questions provoke our study. 1.8. Objectives of the Study In the light of the above discussion the specific objectives of the study are as follows: ~ To examine the performance of actively managed equity funds using the variance and sem i-variance approaches to risk. ~ To analyze the impact of regulations on performance by comparing funds under weaker regulations (UTI) and funds under stronger regulations (all other funds that come under SEBI regulations). ~ To examine the impact of regulations on fund manager fees and overall expenses and determine if the cost reductions with size are forced by regulations or due to actual economies of scale. ~ To examine new trends in the mutual fund industry. specifICally, the growth in size and its impact on performance in different market conditions. 1.9. Analytical Tools Here, we give a survey of the important general analytical tools used to study various objectives. For studying fund performance we first have to compute retums over a period. The approach in this study is to use monthly returns of funds. The monthly returns are computed using the Net Asset Values (NAV). The NAV is the value of the assets under management (AUM) minus the operating expenses, divided by the 18 lumber of shares or u n ~ s of the mutual fund. This value is declared at the end of each business day. While measuring performance certain standards have to be maintained. Among the better known standards is the Global Investment Performance Standards (GIPS)*. We summarize the most important ones for comparing funds a) Com paris ions should be over the same period b) Funds ~ h the same broad objectives should be compared. c) Fund performance must be calculated in the same manner for all the funds. d) At least frve years performance or from inception (if later) should be used as the track record. Hence to compare funds performance the above standards are maintained and a Time-Weighted Return (TWR) is used. (1.1) NAV T ., is the fund retum at the beginning of the month and NAV T is the fund return at the end of the month. This is relevant for funds which do not give dividends or any other periodical benefits. In case any dividends are given they must be added back to give true comparision. After using the above method to calculate time series monthly returns data, the risk is calculated for each sampled fund to arrive at risk weighted returns. Two approaches are used to calculate risk. The most popular measure is the standard deviation (0) of retums or volatility of returns. 19 (1.2) U= ,-, I-I Rp is the portfolio or funds retum in month i and R is the average of the fund's monthly retums. Monthly retums are calculated using (1.1). l' is the number of months involved in the study. Using the standard deviation measure of risk the following measures are computed. a) The Sharpe Ratio (SR) The Sharpe ratio was developed by Bill Sharpe (1966) to measure risk- adjusted performance. It is calculated by taking the portfoliolfund retums Rp net of R, is the risk free rate of retum and dividing it by the standard deviation of the fund retums. This gives the return above the risk free rate of retum per unit of the risk involved. SR = (1.3) The higher the SR, the better is the fund's return per unit of risk. b) The Information Ratio (IR) Here the fund's monthly return Rp are taken net of the selected benchmark retums. This gives the funds excess returns compared to the benchmark. The standard deviation of these excess retums (EXR) is called Tracking Error (TE). Dividing the average EXR by the TE gives us the Information Ratio. IR= EXR TE (1.4) 20 This ratio gives the risk-return generated by the fund manager's ability to use information to deviate from the benchmark. The higher the IR the better is the fund. The above two measures were based on the standard deviation as a measure of risk. An alternate method of measuring risk is the semi-variance. This gives us the dispersion of all the returns that fall below a certain acceptable return say, the risk free rate return. The Sortino Ratio (SOR), (Sortino, 1991) is a ratio that measures risk- weighted returns using the downside risk or volatili1y of returns below a certain minimum acceptable return (MAR). This ratio is give below.
"N . (Rp < Rnrin- R nrin)'
(1.5) N While the numerator is similar to the Sharpe ratio the denominator is the standard deviation of returns below a certain minimum (R mln ). N is the number of observations. For the next objective on relevance of regulations for fund performance we try to arrive at the regulatory impact on performance through a comparision method. The performance of funds regulated by 'weak' UTI Act Regulations and those regulated by 'strong' SEBI regulations are compared. They are compared on the basis of the Sortino and Information ratios. We then compare their portfolio selection skills as measured by the Jensen alpha (Jensen, 1968). This is computed using a single factor CAPM model. The 21 alpha is computed separately for SEBI regulated funds and UTI funds which did not faqll under SEBI regulations. Since all funds are identical in portfolio (equity diversified) and objective (growth), they satiSfy the comparability criterion of GIPS. Limnations of data prevented a study of their possible investment styles. A time series and cross section data set is used. (1.6) Rip! is the return of fund 'I' in the period '1'. R. is the risk free rate return in the period 'j'. Rm is the monthly market return. The 'a" gives the average alpha for SEBI regulated or UTI funds depending on the funds considered for the regression. The Jensen's alpha helps to understand if the fund is earning returns that justify its risk. If the alpha is positive, ~ gives an indication of superior stock picking skills of the fund manager. Beta (13) is a measure of the systematic risk or the degree of the variance in the portfolio that can be explained by the benchmark index. The difference in the alpha of the two fund groups gives us an indication of the behaviour induced by regulations. For the third objective which studies the role of regulations in contrOlling fund operating expenses a log-Iog cost function is used to detect the prevalence of economies beyond regulatory levels. The data are of unbalanced panel format. ; = 1, ... , N Lcost. is the log of operating costs of fund 'I' in period 'I'. Laa. is the log of average assets of fund 'I' in period 'I'. (1.7) 22 (Iaa)' is the quadratic term that is used to capture economies of scale. Xi is a vector of fund characteristics that can affect costs. It includes family asset size (LTAA), total load (TlOAD), turnover (T), and regulatory cost ceiling (RCOST) as a control variable. Use of log-log form helps us to arrive at elasticity estimates and also helps us to reduce multicollinearity. For the last objective on size and performance, first a single factor CAPM model is used to compute the for funds of different size categories. (1.8) The returns of fund belonging to asset class i= 1, .. ,4 for the period 't' is given by Rip!. The benchmar1< index returns, R CNX500 is the returns on the National Stock Exchange CNX500 index. Using the average Ws computed for different asset sizes, the beta weighted returns are computed. (1.9) These beta adjusted returns-SRp. (risk weighted returns) and Rp are then regressed on fund size controlling for various fund characteristics (styles) the help of dummy variables. A one period lagged model is used for most variables. (B R,,) = C + a, LN( aa,,_,) + a,LN(T AA,,-,) + a,jlow,,_, + a.age,,_, + as/load" +a 6 D 1I1 +&" LN (aa,.,) is the natural log of asset size of the i"' fund in the previous month, t-l. (1.10) 23 LN (TAA._,) is the natural log of total family assets of the fund family to which the i" fund belongs minus the i'" fund's assets. Other variables include 'flow' (growth in fund size due to net sale of fund units to investors). age of the fund, total entry and exit load and two dummies that represent growthlvalue stock orientation (0,) and large cap/midcap funds (0,). The same process is repeated in a disaggregated manner for fund performances in bull (returns greater than benchmark index) and bear phases (returns lesser than the benchmark index). This gives a performance under different market conditions and its variation with fund size. 1.10. Relevance of the study The relevance of the study arises from four important factors. First, even though there are studies which use the single factor CAPM to study performance they fail to consider comparison of performance under risk measures. Secondly, the ability of regulations in protecting investors through better disclosure norms and organizational structures has not been addressed at all. Such a study can point to the need for regulatory strength if needed. Thirdly, the ability of regulations in constraining costs needs to be examined while regulations have capped fund operating expenses it remains to be seen if these are really constraining or well above what economics of scale and competition could lead to. Lastly, studies have not covered the issue of growing size of Mutual funds in India and the possible negative effects of the same on fund performance. 24 examine need for new ones in the light of changing circumstances. 1.11. Scope of the Study The primary focuses in terms of performance are the equity diversified funds. The study of performance and regulation of mutual funds is from the period May 1995 to October 2007. The study period has been extended to 2007 to incorporate the study of fund size and its impact on retums. It is a study based on time series and cross sectional data or panel data. Different objectives have different time periods of study. For the first objective of altemate approaches to risk, the time period is from June 2000 to December 2006. This time period is purely to enhance the sample of open end equity diversified growth (OEG) funds The second objective which involves examining the role of regulations in ensuring better performance covers the time period May 1995 to September 2000. This is because we have two different regulations UTI Act of 1963 and SEBI Regulations existing simuHaneously during their period. It gives an opportunity to compare and arrive at conclusions and hence overcome the counter factual problem of deciding how performance might have been w ~ h weaker regulations. The third objective covering cost regulations and economics of scale is for the period April 1999 to March 2007, covering eight fiscal years. The start from April 1999 is guided by the availability of annual reports to gather disaggregated cost details. It is extended till March 2007. Firstly, to get complete financial years and secondly, to get more reliable estimates as n is unbalanced panel data. This 25 study covers the open ended equity (diversified) growth funds and the open ended medium and long term debt oriented funds. The fourth objective adds relevance in rapidly changing circumstances such as size of individual funds and also style of funds. This study covers the period August 2002 to October 2007 as monthly asset size of funds is available only since August 2002. It reaches up to October 2007 to enable a minimum of 60 months data on one month a-lagged basis. This study covers only open ended equity (diversified) growth funds as these funds witnessed rapid growth during this period. 1.12. Data DeSCription, Sources and Period of Study The study's main focus for analyzing performance and its variation with alternate risk measures. style and size are Open Ended Equity Diversified Growth (OEDG) funds. With respect to the influence of regulations on performance both closed and open ended equity diversified funds have been considered. For the study on cost regulations. both OEDG funds and open ended growth debt funds with medium and long term debt orientation have been considered. For the first objective, alternate risk approach to fund performance. the time period taken is June 2000 to December 2006. The sample consists of 28 OEDG funds from 13 different fund houses. Data on the NAV and the benchmark index Bombay Stock Exchange (BSE) sensex has been sourced from the data base NAV India. The 91-day t-bill return is taken as the risk free rate of return uniformly throughout the study and data are sourced from the Reserve Bank of India (RBI). For the second objective on regulations and performance the time period considered is May 1995 to September 2000. A sample of ten equity diversified 26 growth funds. three from UTI and the rest under SEBI regulations have been chosen. Data on UTI funds have been sourced from UTI. data on JM equity are from NAV India and for the rest it is from their respective mutual fund websites. The third objective examining cost regulations and economics of scale the time period is from April 2000 to March 2007. It involves time series cross section data or panel data in unbalanced format. It includes both OEDG and debt funds with the maximum number touching 98 funds in 2002 -2003 and the minimum being 33 in (1999-2000). Data have been sourced from annual reports of mutual funds. The fourth objective on size. style and returns covers the period August 2002 to October 2007 and the data are sourced from AMFI. Value Research India and individual mutual funds. It covers OEDG funds. It is of unbalanced panel format as asset size for two months is unavailable. 1.13. Chapter Scheme of the Study The present study is organized in seven chapters: The first chapter provides the background of the study and statement of the problem. This chapter also includes major researchable issues. objectives. analytical tools. relevance of the study. scope of the study and the database. The second chapter presents a detailed review of literature and identifies the existing research gaps. In the third chapter, a comparison of Mean-Variance (M-V) and lower partial movement approach to risk for performance ranking has been carried out. This helps us to understand the relevance of the widely used M-V measure. 27 ~ Regulations and performance are treated in the fourth chapter helping us to understand if regulations have hampered performance on a comparative basis. ~ The f"dth chapter examines the impact of regulations on the operations costs charged by funds to the fund investors. It throws light on whether regulatory cost ceilings are really a constraint and how they can be improved upon. ~ The sixth chapter critically assesses the impact of growing fund size in differing market conditions after accounting for differing fund styles. It helps to understand if diseconomies have crept in and whether fund size should be regulated or left to competition. ~ The concluding seventh chapter presents major findings, limitations, further scope for study and policy suggestions emerging from the study. 28 CHAPTER -II REVIEW OF LITERATURE 2.1. Introduction CHAPTER -II REVIEW OF LITERATURE The review of literature is to guide us in the methodologies to be used, estimation procedures and interpretation of results. This chapter, therefore, focuses on both theoretical and empirical literature to understand the need for regulation, the fonn of regulation, approaches to risk and perfonnance assessment for funds and also estimating cost functions. Mutual funds play a crucial role in reducing risk and transaction cost while investing in the stock markets. They offer a more efficient route of investing. In the process of encouraging more investments they help in realizing true prices of securities. This in tum helps attract investments through the initial public offer route and unleash the savings of Indian households. This chapter is divided into two broad sections. The first section reviews theories of regulation, risk and perfonnance measurement. The second section is divided into five subsections. The first subsection reviews altemate risk measure studies. The second subsection examines the various approaches to fund perfonnance measurement. In the third, we deal with studies relating to economics of scale in mutual fund expenses. The fourth subsection reviews papers on the impact of size of assets on mutual fund perfonnance. Given the importance of fund investment style in detennining fund returns, in the fifth subsection we review empirical literature on fund style. In the last subsection, we focus on a few important studies covering the Indian mutual funds. 29 2.2. Theoretical Review 2.2.1. Theories of Regulation We move from a controlled economy in India to one that is regulated. The line between control and regulation has to be distinct and, therefore, the basis, extent and mode of regulation has to be made clear. Regulation essentially seeks to control price, sale and production decisions of finns such that private interests align with the larger 'public interest'. The public interest approach to regulation can be dated back to Arthur Pigou. Pigou 15 dealt with externalities as a source of market failure. The costs of negative externalities such as pollution is not internalized reflect true costs. Markets prove inefficient and hence regulatory control is justified. The purpose of regulation is to improve public welfare (public interest) by correcting market failures: Government intervention improves welfare. Posner (1969) discusses exhaustively the regulating monopolies, although dealing with the issue of regulating 'natural monopolies' Dr more specifICally 'utilities' refonns questions the traditional basis and of regulating monopolies. The traditional 'dead weight loss' of monopoly profit maximizing price is questioned. He maintains that price need not be to maximize short tenn profits. He points out other managerial objectives which may lead to lower price. Preventing potential entrants from entering developing good reputation are two such reasons. Besides he questions the rationale for comparing monopoly pricing with competitive price when 15 The New Pal grave Dictionary of Economics and the Law Vol). pp 271 30 the general industrial tendency is oligopolistic whose prices themselves are uncertain. He points out that supra competitive profits are themselves an incentive for the monopolist to contain costs and innovate. Taxing monopoly profits is a far better anernative to direct regulatory controls for Posner. He dismisses other effects of monopoly such unresponsive behaviour and inferior quality as likely to cause loss of profit and lower costs which attracts entry. He emphasis on the distorlionary effects of rate of retum regulation which lead to inferior services besides, the problems of regulatory lags. He calls for a regulatory system based on a cost benefit analysis that includes both direct and indirect costs of regulations. Stigler (1971) feels that the demand for regulation is not often for 'public benefit' but rather for the benefit of the industry in question. The states coercive power allows it to tax, control entry, effect make policies which affect complements or substitutes or even fix prices. Such powen; can be 'bought' by 'industries in' return for campaign funds to restrict competition or ensure profrts. Stigler points out that such regulations are actually welfare-reducing as the benefits ineffICient policies are possible only because in a democracy voting on each policy is costly and hence not done and also because not all voters who vote might have an 'interest' in the issue. Tullock (1975) downplays the need for regulation and believes that the costs of government failures or regUlatory failures are larger than the cost of market failures. The other important reason for market failure and ground for regulation is that of information asymmetry and bounded rationality. Akerlof (1970) pioneered Ihe study on information asymmetry and showed how imperfect information would 31 lead to adverse selection and the ultimate collapse of the martlet through low quality sellers crowding out good quality sellers. Schwartz and Wilde (1979) deal with the necessity for govemment intervention in markets with imperfect information. Governments intervene in markets when the percent of uninformed consumers in the market is sufficient to do so. But this is expensive and besides one does not know what level of information is considered adequate for a consumer, besides the focus should be on the martlet and not the individual. If there exist suffICient number of informed consumers the firm has every incentive to behave competitively. The guiding variable, they point out, is whether non-competitive behaviour has occurred in the market. Even then they suggest providing information is a better method than price control. A mutual fund investor may be operating with imperfect informational. She is incapable of observing the fund manager and the effort levels to produce best retums. It could induce post contractual opportunistic behaviour which opens the investor to greater risks and lower retums. There is also the possibility of the investor not comprehending the information related to the mutual fund. Both of these are grounds for regulatory intervention. Frank and Mayer (1989) point out that information asymmetries can lead to organizational failures which include fraud by employees, misutilization of funds, reckless investments and excessive chuming of portfolio. Regulation of mutual funds normally cover mandatory disclosure of relevant information, fixing management fees and expense limits, guidelines for valuing the 32 portfolio and conducting shareholder transactions, control of false and misleading information, investment norms and corporate governance structure. 2.2.2. Theories of Risk Measurement The purpose of this section is to identify the alternate approaches to risk. In the tradnional portfolio models an investor is assumed to maximize expected utility of the portfolio. This is proposnion of securities such as to minimize risk while holding the mean return constant at a level. Markowitz (1952) was the first to propose the mean-variance analysis of portfolio decisions. He discussed the concept of effICiently diversified portfolios which maximized expected returns for a given amount of risk measured by variance. In 1959, Markowitz elaborated further and provided means for calculating efficient portfolios when the means, variances and covariances of retum of the securities were available. Tobin (1956) was more interested in the implications for the macro economy ~ n l i k e Markowitz who focused on the rational investor. Tobin used the m ean- iariance approach to study the choice between safe liquid assets and risk assets. Tobin points out that the fundamental basis of the M-V approach was that ;ecurities were imperfect substitutes for one another'. In the same article Tobin ;howed that the optimal portfolio for any investor is a combination of the optimal JOrtfolio of risky assets and a portfolio of risk free securities. This speaks of nvestors' ability to separate the decision about investment proportions among risky Issets separate from the decision with respect to investment proportion between 'iskyand risk-free assets. This property is termed as monetary-separation. 33 normally distnbuted. Funner vanance penalIZes nOl omy me PO"5IUIllIY UI yel y ou .. retums but also the possibility of very high returns. The real investor does not perceive retums above a certain acceptable minimum as a risk. It is, therefore, the downside risk that is more important and the true risk. Markowitz (1959) himself, proposed the 'semi-variance' rather than variance as a measure of risk as it captured only adverse deviations. Semi variance can also be described as below-mean target semi variance or in other words returns below the mean return is taken for measuring risk. Hogan and Warren (1974) developed a CAPM using a below target semi variance risk measure also termed as the ES-CAPM. This corrects for the symmetric distribution approach of the traditional CAPM. Bawa (1975) generalizes the semi-variance measure of risk to reflect a less restrictive class of decreasing absolute risk averse utility function know as the Lower Partial Movement (LPM). The M-V approach was restrictive as it ensured an optimal portfolio only if the utility function was quadratic. Bawa and Lindenberg (1977) proved that 'separation' holds for the LPM risk measure when the target was the risk-free rate (for fixed target). Where variance measures the expected squared deviation around the mean, LPM's measures the expected value of the alto power of the deviation below a target (when a > 1). When a=2, the LPM is also termed as target semi variance (TSV). Fishbum (1977) introduced the general LPM risk measure setting a=2 to give the TSV. Since the TSV only punishes returns below the target it fit investor 34 preferences. The induced effICient set of his model was found to satisfy the stochastic dominance criteria (which consider the whole probability distribution of retums and not just the M-V a measure). Stochastic dominance is better tor judging the performance of portfolios, because it does not make any assumptions about the probability distribution of security retums and is based on a general utility function. The above literature points to the development in measurement of portfolio. The LPM with a target return based on a more general utility function appears better in representing investors' preference on a risk return space. 2.3. Empirical Review 2.3.1. Review of Studies on Mutual Funds perfonnance This section aims to review literature on mutual fund performance - more specifically on equity funds. The performance of mutual funds has centered around the appropriate choice of benchmaril for comparing fund performance and the appropriate model to be used. Most studies focus on whether fund's have managed to outperform the benchmaril. But studies have also covered diverse areas such as persistence in fund performance, marilet timing. and bull and bear marilet performance. Initial studies on performance such as those by Sharpe. Treynor and Jensen use the CAPM model with a single benchmaril tor assessing performance. Sharpe (1966), was among the earliest to use the CAPM to assess mutual funds performance. He assumed that expected return E(R.). of a fund and its risk (0.) are linearly related. This can be explained with the help of the following equation. 35 RF is the risk free rate and fl is the risk premium. This is based on CAPM assumptions which include the following". Zero transactions costs Assets are infinitely divisible Absence of personal income tax Individual cannot affect the price of the stock I nvestors make their choices solely based on M-V of the portfolio Unlim ned short sales Unlimned borrowing and lending at the risk-less rate Homogeneny of investors in terms of the relevant period and in terms of what they consider as necessary inputs for the portfolio decision Sharpe found that funds under performed the Dow Jones index by 40 basis points. Further better performing funds had lower expense ratios. According to equation (2.1) an investor who can borrow or lend at the risk free rate and invest in a portfolio with E(Rp) and Op, can by allocating funds between the risky and risk-free asset reach the Capital Market Line (CML): E(R) = RF + lRp';pRF F- (2.2) The optimal portfolio is the one which delivers the best reward to variability ratio. (2.3) 36 Equation 2,3 is also popularly known as the Sharpe Ratio (SR), Sharpe examined the perfonnance of 34 open end funds for the period 1954-1963 in the US, His study revealed that the SR measure varied from 0.78 to 0.43, This variation can be attributed to managerial ability. Treynor and Mazuy (1966) attempted to study a fund manager's ability to time the market. They used a quadratic to check for this. The quadratic Traynor- Mazuy TM model as it is popularly called is as follows: (2.4) r pe is the fund excess return over the risk free rate in period t, a, is the estimated selectivity performance, b, is the portfolios systematic risk. rmt is the excess return of the market over the risk free rate in period t. b 2 is the indicator of market timing perfonnance and epe is the residual. A positive value for b 2 is argued as an indicator of good timing ability as it allows the characteristic line to be upwardly concave, They found no evidence of this in their study of 57 funds for the period 1953-1962. Jensen (1968) was the first to introduce a model that captured fund performance in relation to a benchmark, The model is again based on the CAPM and is given as fOllows: (2.5) Where R ~ is the expected return of the jth portfolio, R, is the risk free rate of retum, ~ j is the systematic risk, and R mt is the market return all for the time period t 37 (1=1 ......... , N). The constant 'n' is the Jensen's alpha. A positive 'a' indicates an ability to generate superior stock picking. Using this model Jensen studied 115 funds for the time period 1945-59. A statistically SignifICant number of funds had a negative alpha and the average alpha was minus 110 basis points. Both these studies by Sharpe and Jensen seemed to confirm the efficient market thesis (EMT)'. The EMT believes that since security prices reflect all available information it is impossible to beat the market through active portfolio market'. Post expenses active management turns out to be waste of money on fund management fees and expenses. In contradiction to these two studies Carlson (1970) calculated the Jensen alpha and Sharpe ratio for 86 funds over 1948-67. He found an average positive alpha of 60 baSis points and further found the Sharpe ratio of funds at 0.57 to be higher than that for the Dow Jones with 0.43. More importantly Carlson funds that performance measurement is dependent on the fund type the benchmark index and the time period considered. Mc Donald (1974) studied the performance of 123 funds using monthly data for the period 1960-69. His study was based on a CAPM model four measures- monthly mean excess return; reward-ta-volatility Ratio, Jensen's alpha and reward to variability ratio were calculated. He concluded that the average fund performance was not SignifICantly different from the market and given fees and expenses they were slightly better. Mains (1977) carried out Jensen's study for 70 of the funds over the same period. He carried out the study with annual and monthly returns. While annual 38 data gave an average alpha of minus 62 basis points the monthly data gave a nine basis points. Mains argued that monthly data was more efficient. These studies show that fund managers do out perform a passive benchmark index portfolio while several studies also devote to the market timing ability of funds but these are excluded as they are not the objective in their study. Importantly these studies also show that performance assessment is dependant upon the time period, the type of funds studied and more importantly the index for comparison. Increasingly the single factor CAPM came under criticism. The single risk factor of the CAPM has led to substantial research and that a muHifactor model would serve us better. Roll and Ross (1979) proposed the Arbrrrage Price Theory' as an aHemative to the single factor CAPM. They found that, between 1962 and 1972, there were at least four other factors which could explain risk. But their test, they themselves admitted was a weak one as no economic justifICation could be given for these factors. Fama and French (1992, 93) argued that the risks of a portfolio cannot be captured by a single beta. Portfolio risks are muHi-dimensional for them. When they studied the US stock returns, for the period 1963-90, they found that the CAPM beta could not explain the entire return variance. They proposed a three factor model to better capture the risk. Apart from the market risk they included the Small cap stock returns -Minus-Big cap stock returns (SMB) factor to capture the 'size risk' and, also the High-Minus-Low (HML) factor to capture the 'value risk'. They 16 Ross (1916), had analyzed the limitations of the single factor CAPM and proposed the APT. 39 find that historically small cap stocks have delivered better average returns (in the US market) and hence, can be an important explanatory factor for determining the sources of performance of a fund. The value risk is captured by taking the difference between the returns of the high book-ta-market (B/M) stocks minus Low book-to-market stocks. Here they find that value stocks deliver better returns when the market corrects for ij's over reaction. Their model proposed increased the R- square or the explanation for fund performance compared the single factor CAPM models. This was done by essentially accounting for the portfolio differences between funds or the style of funds. The posijive alpha probably due to the style and not necessarily due to POrtfolio selectivity and hence, for a truer measure of alpha we need these addijional factors according to them. Several tests were carried for the Fama-French model. Kothari et al (1995) find that there is no 'value' effect but certainly a size effect. We take these factors into account when discussing the impact of size of fund on retums. But we do not consider the multifactor model proposed by them rather we use the single factor CAPM but separate the funds on the basis of the portfolio and use a dummy to capture the differential effects of fund style on fund alpha. 2.3.2. Review of Empirical Studies on Mutual Fund Costs and Fees ~ h e r e have been several studies on the existence of economies of scale and scope In the financial sector in general and mutual fund industry in specific. 40 Murray and White (1983), find evidence of both economies of scale and scope in the British Columbia credit unions and point out that regulation limiting the growth and diversification of these unions could lead to higher operating costs. Baumol et al (1989) ,using translog cost functions, find economies of scale present for mutual fund complexes in the U.S .They however find such economies weakened when assets per account ( with the assumption that the number of accounts grow in the same proportion as assets) are taken as hedonic variables. Further they also find significant economies of scope within the fund complexes implying that buying into funds which are a part of a large complex should give more net returns other things remaining the same. Their findings clearly shows that not all funds can be treated alike regarding the assumption of persistence of economies of scale as hedonic variables can cause differences. A study of US funds by Rea, Brian and Reid (1999), for the year 1998 shows, again, strong inverse relationship between operating expense ratios and asset size of similar funds. Further, funds with asset increases over time have also shown decreases in expense ratios depending on the extent of increase. While the presence of economies of scale seems to be the general norm, Latzko (1998), tries to go beyond this and establish the source of economies of scale. It could be due to management fee reduction or due to 'other administrative expenses'H. Economies of scale are found to originate substantially from 'other administrative expenses'. Though economies of scale were found across the entire range of asset size, the rapidity of decrease in average costs is found to slow down t 1 Transaction costs of shareholder services and record keeping. 41 :er a critical asset size prompting one to conclude that smaller funds seem to derive more marginal benefrts of economies of scale than the larger funds. The study of fund's economies becomes all the more important in the light of the fact that fund size could have a negative importance on performance. Chen et al (2005) show that organizational diseconomies (due to hierarchy costs) " are particularly pronounced for funds that play small cap stocks. Regarding price competition amongst funds there exist divergent views. Elton, Gruber and Busse (2004), find no substantial importance given to even expenses by investors when choosing S&P index funds. This could be possible in a time when most funds are performing well thereby causing investors to neglect the expenses. In a survey conducted by Capon, Fitzsimons, and Prince (1996) only one fourths of investors give importance to management fees. On the other hand, Christofferson (2001) shows that money market fund managers have been willing to be flexible with respect to management fees given the convex relation between fund inflows and lagged performance. A similar convex relation has been presented by Chevalier and Ellison (1997) in equity funds. Brian Reid (2005) points out that small fund's in the U.S, in order to compete with larger funds which have lower operational average costs, have tended to voluntarily waive of a portion of their fees. He finds that over 90 percent of new cash flows went to funds with expenses below the median funds expenses. 1M Chen el aI. describe lhese costs due to fund managing te-am members over.investing in oonvincing their colleagues of the research behind their stock choice. 42 2.3.3 Empirical Studies on Fund Size and Returns While there have been a few im portant contributions in the area of size and performance they are not specifIC to India. Grinblatt and Sheridan Titman (1989) find no clear trends of decreasing returns for larger funds in the US between 1974 and 1984. Perold and Solomon (1991), on the other hand, found diseconomies of scale arising out of increased price impact of large transactions. Such impact costs are found to be larger for larger trades by Gallagher and Looi (2003) as well for a sample of institutions in the Australian markets. Chen et al (2003), after accounting for various benchmarks find that return :lee lined with lagged fund size for the period 1962 to 1999 in the US. The main reasons for this are found to be investments in small illiquid stocks and also )rganizational diseconomies. On the contrary, Gallagher and Martin (2005) find no statistically significant :lifferenee in the returns of large and small funds in Australia. The opinion among Indian fund managers on the issue of size is diverse. Some feel that there are ample opportunities for investment while others feel that nvestment choices are drying up. Z.3.4. Fund Style and Fund Returns nvestment styles of mutual funds have an important role to play in determining 'und retums. Some equity funds could choose to invest predominantly in growth 43 stocks. These are stocks that have a high price-to-earning (PIE) ratio and do not generally pay dividends. Such styles of investment can increase volatility. On the other hand there is the value style of investment. Funds having the value style tend to have more exposure to cheap stocks. Such stocks tend to have low PIE ratio .and yield high dividends. Such stocks are less volatile. AHernatively there is the large cap, small cap investment style. Fama and French (1992) had pointed out the tendency of small cap stocks giving better returns than large cap stocks. But small-cap stocks are also riskier as their prices are more volatile. Sharpe (1992) explains how a substantial part of the return variability of a portfolio depends on the distribution of the portfolio across a number of asset classes. He uses 12 asset classes represented by an appropriate index, to study the influence of 'Style' on the performance of 395 funds for the period 1985-89. He employs the quadratic programming methodology to determine funds exposure to different asset classes. He finds that growth equity finds tend to predominantly focus on growth stocks. Grinblatt, Titman and Wermers (1995) analyzed the quarterly holding of 155 mutual funds for the period 1975-1984. Using muHiple cross-sectional regressions of fund performance on fund characteristics they found that 77 percent of mutual funds tended to be momentum investors. This meant that funds tended to buy past winners and sell past losers. Momentum investing gave funds better returns than contrarian investors and the index. Bogle (1998) uses the mine box mutual fund rating system. He explains that US fund style evolved from a homogenous form in the 1970s to amore heterogeneous form during the 1990s. He very significantly finds that expenses 44 are the prime differentiator between actively managed equity fund returns. Further he finds that index funds delivered better average risk weighted returns compared to actively managed equity funds. The importance of fund style in explaining fund returns cannot be denied. Indian mutual funds although not as diverse, in terms of style have, in the recent times, adopted GrowthNalue and Large/Mid or small cap oriented styles. We, therefore, factor this in our study its impact on mutual fund size and performance. 2,3.5. Empirical studies on Indian Mutual Funds While empirical studies of mutual funds in India are not comparable to sheer volume of studies conducted in the US it is picking up with the growth of the industry and more importantly with the availability of more frequent data. Sahadevan and Thiripalraju (1997) attempted to compare the performance of funds using total return, consistency and volatility. They did not attempt to use any CAPM single or multifactor models. Their study covered private and public sector mutual funds for the period 1995-96. The benchmark used was the SSE National Index terms of absolute returns. Out of 32 public sector funds, 11 outperformed the index. In the case of private sector mutual funds time out of the ten studied outperformed the index. The study of course did not intend to go into deeper risk return analysis and compared a host of different types of funds to a Single index. Panigrahi (1996) in his study of Indian Mutual funds selected a sample of four growth funds for the time period October 93 to December 95 and divided them into two periods to capture their performance in boom and bear phases separately. 45 The general conclusion reached was that funds had lesser risk than the market index, BSE 200 or the RBI ordinary share price index and they delivered near market returns under normal condijions. He uses a log-log model regressing the log of (monthly average) of NAV on the log of the market index (monthly tune rates of growth) to arrive at the R2 or diversification and risk compared to the market. The study of course, has a very limited objective, sample and time period. Roy and Deb (2003), use the conditional performance evaluation technique to study fund timing and performance. They pointed out that the tradition models of Treynor-Mazuy had shortcomings as they based their assessment on historical average returns without considering the role of new information and the time varying element of returns. Taking a sample of 89 funds (consisting of equity diversified and balanced funds) they test for alpha over the time period January, 1999 to July, 2003. Using lagged information variables they find that the alpha deteriorates. But their study is weak with poor statistical signifICance. Indian empirical studies tend to focus on the application of evolving methods to the study of mutual fund performance. However, to the best of our knowledge, there have been no studies that tried to deal with the interplay between regulations and fund behaviour. Issue such as the influence of regulations on performance and its ability to act as a genuine constraint on expenses are left untouched. This study makes an attempt to fill this gap. 46 CHAPTER - III PERFORMANCE OF EQUITY FUNDS USING VARIANCE AND SEMI-VARIANCE OF RISK CHAPTER - III Performance of Equity Funds using Variance and Semi-Variance of Risk 3.1. Introduction The need for assessing the perfonnance of mutual funds cannot be over emphasized. Fund manager's capability of generating better risk-weighted retums than the comparable benchmark infuses confidence in the minds of the investors. It encourages savings as people have more variety of savings options rather than the tradnion bank saving accounts or the more risky and costlier route of investing directly in the stock markets. The more such savings are used to discover the optimal price of stocks the easier n is for companies to raise funds through the equity option. While assessing performance by which we mean the returns net-expenses generated by the funds, we focus primarily on the actively managed diversified pure equity funds. Further, while measuring such performance we deviate from the widely reported measures (in the Indian context) used to measure performance by using the lower partial movement to capture better risk weighted performance. The entire chapter is divided into four parts. The second part is devoted to _the empirical frame work. Here we discuss the methodology of assessing performance and risk and also the data. In the third part we give the results and t nalySiS. Here we rank funds on the basis of their performance using both easures of risk and see if the ranking varies significantly. In the last part we erive our conclusions. The objective of this chapter can be stated as follows: 47 To study the performance of pure equity diversified funds using alternate measures of risk and to check if they give signifICantly different results in terms of comparative ranks of funds. 3.2. Empirical framework and Data description 3.2.1. Empirical framework The monthly returns of funds are computed by using the month beginning and month end Net Asset Value (NAV) as given below: TWR =[NAV, -NAV,_I ]XIOO NAV H (3.1) TWR is the time weighted return NAV,., is the NAV at the beginning of the month and NAV, is the month end NAV. On computing absolute returns we then move ahead to find the risk aSSOCiated with these returns and the consequent risk weighted return. Two alternate approaches are used here. The first is the more popularly used and reported measures of Sharpe RatiO and Treynor Index. These measures are based on the assumption that returns distribution is symmetric and, therefore, the conclusion that mean-variance of a portfolio is enough to derive the risk-weighted retums. The standard deviation of portfolio returns: (3.2) 48 T is the number of monthly retums over a given time period, Rp is the arithmetic mean of the returns and Rp is the rnonthly return. Using the standard deviation as a measure of risk the following risk- weighted measures are computed Sharpe Ratio (SR) Rp-R (SR) = J (3,3) CY RP The top five funds have a Sharpe Ratio that ranges frorn 0.35 to 0.38. This is the average return per month per unit of risk measured by the standard deviation of the fund portfolio returns, the higher the Sharpe Ratio the better. The Sharpe Ratio has a signifICant advantage as a measure of relative performance as it is based on the standard deviation of the funds return and not relative to an index. When relating performance to an index to arrive at the fund alpha the r-squared has to be reasonably high. This depends on the index selected and, therefore, the possibility of wrong index bias. Treynor Index (TI) TI = _R.!:...p _-_R.:....F f3 port/aim (3.4) The TI is similar to SR excepting for the fact that it uses beta instead of standard deviation of the portfolio or fund returns. is the funds beta or the systematic risk. This ratio gives a funds average excess return per unit of market risk. This measure is rnore relevant for well diversified portfolios. The Treynor and 49 Sharpe Ratios would be the same if the total risk of the portfolio is the same as the systematic risk. The Treynor Index is based on the assumption that all unsystematic risk is diversified completely. If this holds true then the total risk as measured by the fund returns standard deviation is the same as the systematic risk as measured by the beta. If these two risks do not converge then we can expect a different form of ranking. Rp-R (SR) = f (3.5) (1'RP In order to check if the ranks based on the Sharpe Ratio. Treynor Index and Sortino are different we use Spearmans Rank correlation coefficient. This is given by (3.6) d;=Y;- X; (difference in ranks) n=sam pie size 3.2.2. Data The wider population for our sample consists of the actively managed funds which were pure equity oriented (100 percent equity) and diversified in portfolios. The time period considered for this study is September 2000 to December 2006. In total it consists of 76 monthly observations. This period gives us a good mix of bearish and bullish phases in the stock market. The below chart gives the time series trends for the index CNX500. 50 .. 4000 3500 3000 2500 -3 2000 -= 1500 1000 500 o Chart 3.1 Trends In CNX500 for the sample period. CNX 500 trend .. .. - - ~ l ' O . . 11 . /1 1.1 /'I r II. .IN' I-SerieS11 / V .... ~ /' ~ .. :t .. . , : . ~ -,::. " . ~ iii ~ iii ~ ~ ~ iii Dale The CNX500 shows a bearish trend between September 2000 and May 2003. During this period the index moved in a narrow range before breaking out on a high growth pattem till the end of the sample period. The CNX 500 is also the chosen benchmark index as ~ represents almost 93 percent of the market capijalization of the National Stock Exchange (NSE). The sampled fund consists of 45 funds. Of these, 29 funds are open end equijy diversified funds and 16 are equity linked tax saving schemes (ElSS). Tax savings funds have a lock-in period and are ciose-end funds but are diversified in portfolio. Table 3.1 gives the list of sampled funds. 51 Table 3.1 Sample of Equity Diversilled Funds Fund type Fund type 1. Birla advantage fund ED 24. lie Equity ED 2. Birla equity fund ED 25. Lie growth ED 3. Birla sunlife mnc ED 26. Lie taxplan ELSS 4. Birla sunlife tax relief 96 ELSS 27. Magnum tax gain ELSS 5. Birla tax plan ELSS 28. Morgan stanley ED 6. Bobelss96 ELSS 29. Principal personal tax saver ELSS 7. Bobelss97 ELSS 30. Principal tax savings ELSS 8. Can expo ED 31. Pru icici growth ED 9. Banbank equity ELSS 32. Pru icici power ED 10. DSPML equity ED 33. Pru icici tax savings ELSS 11. DSPML opportunrties ED 34. Reliance growth fund ED 12. Escorts tax plan ELSS 35. Reliance vision ED 13. Franklin India bluechip ED 36. Sahara tax gain ELSS 14. Franklin India prima ED 37. SSI contra fund ED 15. Franklin India prima plus ED 38. SSI equity fund ED 16. Franklin tax saver ELSS 39. Sundaram SNP paribas growth ED 17. HDFC capital builder ED 40. Tata growth ED 18. HDFC equrty ED 41. Tata pure equity ED 19. HDFC tax saver ELSS 42. Tata tax saving ELSS 20. HDFC top 200 ED 43. Taurus bonanza exclusive ED 21. Ingvysya select stocks ED 44. Taurus discovery stock ED 22. JM equity ED 45. Taurus starshare ED 23. Kotak 30 ED ED - EqUIty Dwersified: ELSS Equity Linked Savings Scheme 3.3. Results and Analysis We first consider the Sharpe Ratio rankings. The table 3.2 provides ranking of fund performance on the basis of this ratio. 52 Table 3.2 Sharpe Ratio ranking of funda Rk Fund Type Sharpe Rk Fund Type Sh.rpe 1. Reliance vision ED 0.38 24. Franklin tax saver ELSS 0.25 2. Reliance growth fund ED 0.37 25. Kotak 30 ED 0.2.5 3. Franklin India prima ED 0.35 0.24 26. Pru icici growth ED CNX50 4. HDFC tax saver ELSS 0.35 27. Bina sun life tax relief ELSS 0.23 96 5. HDFC equity ED 0.35 28. T ata tax saving ELSS 0.2.3 6. HDFC top 2.00 ED 0.34 29. Magnum tax gain ELSS 0.23 7. Taurus bonanza exclusive ED 0.33 30. Morgan stanley ED 0.23 8. Franklin India prima plus ED 0.31 31. Taurus starshare ED 0.22 9. Tata growth ED 0.30 32. JM equity ED 0.2.1 10. Pru icici tax savings ELSS 0.30 33. BobELSS96 ELSS 0.20 11. Franklin India bluechip ED 0.29 34. Bina advantage fund ED 0.17 12. Pru icici power ED 0.29 35. Can expo ED 0.17 13. DSPML opportunities ED 0.29 36. Lie growth ED 0.16 14. Tata pure equity ED 0.28 37. Lic equrty ED 0.15 15. SBI contra fund ED 0.28 38. Sahara tax gain ELSS 0.14 16. HDFC capital builder ED 0.28 39. Bina sun life mne ED 0.14 17 Sundaram BNP paribas ED 0.28 40. Lic taxplan ELSS 0.13 growth 18. Principal tax savings ELSS 0.27 41. Can bank equity ElSS 0.12 19. BobELSS97 ELSS 0.26 42. Taurus discovery stock ED 0.12 20. Escorts tax plan ElSS 0.26 43. SBI equity fund ED 0.10 21 DSPML equity ED 0.26 44. Ingvysya select stocks ED 0.09 22 PrinCipal personal tax saver ELSS 0.26 45. Birta equity fund ED 0.00 23. Bina tax plan ELSS 0.26 ED - EqUity Diversified: ELSS ~ Eqully Linked Savings Scheme The Sharpe Ratio values range from 0 to 0.38. The top five funds are all above 0.34. While one can compare funds using the Sharpe Ratio nothing much can be said about the absolute size. It cannot be said as to whether 0.35 is a good measure. The higher the Sharpe measures the better. When comparing funds performance using Sharpe Ratio, with that of the broad based index, the CNX500, 53 we find that 36 funds do better. The Sharpe Ratio for CNX500 is 0.24. 19 funds fail to give better risk-weighted returns than a passive investment in the index. Therefore, not all the funds have been able to show that performance was worth the expenses or management fee paid for fund management. We then consider the Treynor rankings. The same is given in table 3.3 Table 3.3 Treynor Index Rankings Rk Fund Typo nR Rk Fund Typo nR 1. Reliance growlh fund ED 3.34 24. Morgan stanley ED 2.01 2. Reliance vision ED 3.31 25. DSPML equity ED 2.00 3. Franklin India Prima ED 3.26 26. Taurus starshare ED 1.98 4. HDFC tax saver ELSS 2.93 27. Franklin tax saver ELSS 1.97 5. Taurus bonanza exclusive ED 2.85 28. Magnum tax gain ElSS 1.96 6 Tata growlh ED 2.80 29. Tata tax saving ElSS 1.96 7. SBI contra fund ED 2.80 30. Birla sunlife tax relief ElSS 1.91 96 8. HDFC equity ED 2.76 31. Pru icici growlh ED 1.87 9. HDFC top 200 ED 2.69 32. JM equity ED 1.75 10. Pru icici tax savings ElSS 2.56 33. Birla advantage fund ED 1.52 11 . HDFC capital builder ED 2.55 34. Can expo ED 1.46 12. Franklin India prima plus ED 2.47 35. Lic growth ED 1.43 13. Franklin India bluechip ED 2.46 36. Lic equity ED 1.39 14. Pru icici power ED 2.37 37. Birla sun life mnc ED 1.32 15. BobEquity Linked Savings ElSS 2.35 38. Taurus discovery stocl< ED 1.22 Scheme97 16. Birla tax plan ElSS 2.35 39. Sahara tax gain ElSS 1.19 17. Tata pure equity ED 2.31 40. Lic taxplan ElSS 1.16 18 DSPML opporlunities ED 2.22 41. Canbank equity ElSS 1.09 19. Principal tax savings ElSS 2.17 42. S61 equijy fund ED 0.88 20. Sundaram 6NP paribas ED 2.17 43. Ingvysya select stocl<s ED 0.84 growth 21, Principal personal tax saver ElSS 2.16 44. BobEquity Linked ElSS 0.61 Savin gs Scheme96 22. Escorts tax plan ElSS 2:08 45. Birla equity fund ED 0.15 23. Kotak 30 ED 2.08 54 The beta weighted returns range from 0.15 to 3.34. The top five funds under this measure are the same as the top five funds under the Sharpe Ratio except for Taurus Bonanza Exclusive which displaces HDFC equity at the fifth place. The bottom fIVe funds, according to the Sharpe and Treynor Index, are qune different. The above two measures are based on the symmetry of distribution assumption. In order to overcome this we consider the Sortino ratio. Before we do compute the Sortino ratio we present some evidences on the possibilities of non-symmetrical distribution. Chart 3.2 presents a test for nonnality in distribution of returns for the LlC Tax Plan fund 20 Series: LICTP Sample 2000M09 2006M12 16 Observations 76 Mean 1.224211 12 Median 2.940000 Maximum 17.67000 Minimum -26.47000 8 Std. Dev. 7.958188 Skewress -1.268177 KlXIosis 5.500905 4 Jarque-Bera 40.17744 Probability 0.000000 0 The fund returns are negatively Skewed (-1.27). It also has a positive Kurtosis of 5.5 which indicates that the returns distribution is leptokurtic. A kurtosis value above 3 gives a peaked leptokurtic distribution. Importantly, the Jarque bera PB) test-statistic, a test for normality of distribution indicates a non-normal distribution. The null hypothesis of this test is HD=normal distribution. The theoretical JB value at 90 percent confidence level is 4.61 while the actual is 40.17 and highly signifICant. Thus, the null hypothesis is rejected and we conclude that the returns are non-normally distributed. We carry out a similar exercise for all the funds and the summary statistics are given in table 3.4 Table 3.4 summary statistics of the nonnality test for returns of sampled funds Parameters Number of Percent of fund total Skewness <0 46 100 Skewness >0 a 0 Kurtosis >3 21 46 Jarque Bera > 4.61 15 33 All funds have a negative skew in their returns and 46 percent of them are leptokurtic. The Jarque bera statistic gives a SignifICantly higher JB than the Iheoreticallevels for 33 percent of the funds. Based on our analysis of the distribution we feel that the use of the Sortino Ratio would be pertinent to better capture risk-weighted returns. 56 Tabla 3.5 Ranking of funds basad on the Sortino Ratio Rk Fund Type SR Rk Fund Type SR 1. HDFC tax saver ElSS 0.66 24. escorts tax plan ElSS 0.41 2. reliance growth fund ED 0.64 25. bobELSS97 ElSS 0.41 3. HDFC top 200 ED 0.63 26. taurus starshare ED 0.40 4. HDFC equity ED 0.62 27. pnncipal personal tax saver ElSS 0.39 5. Reliance vision ED 0.61 28. bir1a sunlife tax relief 96 ElSS 0.36 6. franklin India prima ED 0.60 29. magnum tax gain ElSS 0.37 7. tata pure equity ED 0.54 30. bobELSS96 ElSS 0.36 8. pru icici power ED 0.54 31. JM equity ED 0.35 9. SBI contra fund ED 0.54 32. Kotak 30 ED 0.34 10. DSPML opportunities ED 0.51 33. Morgan stanley ED 0.31 11. taurus bonanza exclusive ED 0.51 34. sahara tax gain ElSS 0.27 12. principal tax savings ElSS 0.49 35. Bir1a advantage fund ED 0.26 13. franklin India pnma plus ED 0.49 36. Can expo ED 0.25 14 franklin India bluechip ED 0.48 37. lie equity ED 0.24 15 HDFC caprtal builder ED 0.48 38. lie growth ED 0.24 16. pru icici tax savings ElSS 0.46 39. taurus discovery stock ED 0.20 7. sundaram BNP paribas ED 0.48 40. bir1a sun life mnc ED 0.20 growth 8. Bir1a tax plan ElSS 0.44 41. canbank equity ElSS 0.19 9. tata growth ED 0.44 42. SBI equrty fund ED 0.16 o. pru icici growth ED 0.42 43. ingvysya select stocks ED 0.13 T ata tax saving ElSS 0.41 44. lie taxplan ElSS 0.12 2 franklin tax saver ElSS 0.41 45. Biria equity fund ED 0.01 3. DSPML equity ED 0.41 '5D - Equity Diversified; ELSS - Equity Linked StNings Scheme Comparing the top five funds on the basis of the Sortino ratio and Sharpe Ratio we find that there is only one change. HDFC Top 200 funds a place in the top fIVe for Sortino but not Franklin Prima fund which was top fIVer performer according to Sharpe Ratio. Similarly the bottom five also have only one change when both these rankings are compared. 57 Table 3.6 Spearman Rank correlation coefficient Sharpe Treynor Sortino Ratio Index Ratio Sharpe Ratio 1 Treynor Index 0.96 1 (0.00) Sortino Ratio 0.95 0.91 1 (0.00) (0.00) On studying the rank correlation coefficients we find a very high positive correlation between the Sharpe Ratio and the Treynor Index and Sortino Ratio. The correlations are also highly significant. This implies that rankings or relative performance of the sample funds do not change significantly when we use aHernate measures of risk. 3.4. Conclusions In this chapter we set out to assess the performance of funds in terms of simple and widely used measures, the Sharpe Ratio and Treynor Index what we find is that almost 42 percent of the funds do not deliver better risk weighted retums when compared to the broad based CNX500 index. This places a question mark on the ability of a Significant number of fund managers. The expenses and management fees are not justified by the retums. We also set out to examine if using aHemate risk measures as demanded by non-symmetrical retums distribution, would affect the rankings based on Sharpe, Treynor and Sortino measures. We find that the ranking are not significantly different. The use of Sortino measure adds no substantial value to what the Sharpe Ratio -ives us. 58 CHAPTER-IV REGULATORY CONSTRAINTS AND FUND PERFORMANCE - COMPARING ALTERNATE REGULATORY REGIMES CHAPTER-IV REGULATORY CONSTRAINTS AND FUND PERFORMANCE- COMPARING ALTERNATE REGULATORY REGIMES 4.1. Introduction The importance of assessing the impact of regulatory constraints on the performance of mutual funds cannot be underemphasized. Mutual funds need to be regulated to overcome the investor-manager conflict of interests that might arise. Such post-contractual opportunistic behaviour can compromise the interest of the investors and expose them to risks such as poor portfolio selectivity, excessive churning leading to higher costs, higher portfolio risks than what was indicated in the offer document and other organizational risks such as delayed settlement of transactions. Hence, the regulator steps in to over corne such market failures with the objective of ensuring ideal type market conditions as would have prevailed under perfectly competitive markets. Given the rapid growth of the mutual fund industry in terms of the number of fund houses, number of funds and the variety of funds it is assumed that it approximates a contestabl.e market which Baumol et al (1982) discussed. So, the need for anticompetitive type of regulations is not envisioned as necessary. But, however, regulations are not costless. They involve both a direct (compliance) cost and indirect (negative impact on performance) cost. Hence, it is important to understand if regUlators have been excessive in their constraints. It is diffICult for a regulator to get the exact degree of control to be imposed and hence the dangers of under or over-regulating are possible. Under-regulation can mean that the investor is exposed to unexpected risks and over-regulation means that the 59 investor is eXPOsed to lower retums. To get the constraints perfectly right is not the issue as such precise calculations of regulatory cost-benefit analysis is not easily done and perhaps not possible. A regulatory regime must aim at getting its regulations right with minimal error margins. In this chapter we examine how to possibly capture the impact of regulations on mutual fund performance. In other words, we are interested in finding whether regulations have imposed more costs than benefits and, therefore, compromised retums. Such a process is fraught with the diffICUlty of engaging in counterfactual questions. One is posed with the problem of understanding what the possible performance would have been in the absence of regulations. This presents us with an almost impossible task. But, fortunately, the situation eases up as we have a comparative perspective possible. During the period 1993 - 2003 we had two aHemate regulatory regimes for mutual funds 1 . The UTI and its funds came under the regulations of UTI Act of 1963 and all other mutual funds came under the SEBI regulations. SEBI regulations were definitely more constraining than the UTI regulations. Hence, we have a set of funds with stronger regulations and another with weaker regulations. What we attempt to do here is to compare the performance of these two sets of funds and examine if their performance differed in terms of various relevant ratios. To make them comparable only similar funds are considered. In a competitive market funds cannot exist if they consistently under perform and hence funds would ordinarily strive to ensure that they deliver retums which are comparable to their peers. In the case of UTI, it being a public sector " In January 2003, UTI was split into two fund houses. All assured return schemes and the US-64 were retained with the UTI and the rest of the schemes were transferred to a new fund called UTI Mutual Fund Pvt. Ltd. The latter came under SEBI Regulations. The coexistence of two separate I regulatory regimes came to an end with this split. 60 mutual fund with supposed govemment backing, one could argue that the incentive to perform would probably not be there. But one fails to see the point that the UTI had become a very popular investment mechanism for millions of investors and that the govemment could afford to see it under perform only to face political risks. In act when the most popular fund of UTI, US- 64, began to face performance issues during the 1990s , they dug deep into their reserves to deliver good retums to their investors despite not being an assured retums scheme 2o Such was the pressure to ensure that investors did not go disappointed. So it would be flippant to assume that the pubic sector fund house had no incentive to perform. But, at the same time, the organizational aspirations need not necessarily be realized due to the potential moral hazards. Being under lesser constraints and weaker disclosure norms the temptation of getting away with hazardous behaviour is strong. The primary means of doing this could be through choice of investments based on u ~ e r i o r motives rather than retums based motives. This should mean poor stock selectivity and should represent an inferior alpha as compared to funds under stronger SEBI regulations. This chapter is divided into three broad parts. The second part, after the introduction, deals with regulatory constraints under the UTI regulations and SEBI regulations. After establishing clearly the differences between the two the empirical framework is discussed. This deals with the data, the methodology and the resuHs. In the third section conclUSions and recommendations are spelt out. 20 The concept of 'assured return' schemes is strange for mutual funds. It was popular during the early 1990s before SEBI regulations came into effect. SEB! regulations of 1996 called for such schemes to guarantee both the capital and the returns. Many funds which assured returns ran into rough weather. For example Canstar fund of Can bank Mutual Fund (now Canrobeco Mutual Fund) and GIC Big Value Fund of the General Insurance Company (GIC) Mutual Fund could not deliver the promised returns. For more on this see SEBI Annual Report 1996-97. 61 4.2. Comparison of the Regulatory structure of SEBI (Regulations) and UTI Act (1963) In this chapter we first deal with the regulatory differences between the UTI Act and the SEBI regulations. We first deal with the UTI Act of 1963 and then proceed to the SEBI Regulations. 4.2.1. UTI Act of 1963 and its constraints on fund behaviour UTI was created by an act of the Union government of India and hence is a Union government undertaking. It came into being on the first of February, 1954, by virtue of the UTI Act of 1963. The preamble of the Act declares that the objective of the fund was to promote savings and investment, of the Indian publiC, through participation in the securities market". The initial capital of the UTI was Rupees 50 million contributed by the Reserve Bank of India (RBI), the State Bank of India (SBI), the Life Insurance Corporation of India (LlC) and other scheduled banks and financial institutions. In 1975 the Industrial Development Bank of India (lOBI) took over the RBI stake and along with it, the rights and responsibilities as well. 22 In 1954 it launched its first 'scheme' or fund called the Unit Scheme - 54 (US-54). It was intended to be an open end, non-assured income fund with a balanced portfolio of equity and debt. In 1993 a three member committee with NVaghul as the Chairman lxamined the structure of mutual fund regulations. It had recommended that all the lxisting schemes of the UTI be divided into two categories namely, mutual fund ype schemes and non-mutual fund type schemes. All mutual fund type schemes I Verma, page 103 2 Malegam committee report on the UTI. 62 were to have an asset management company and come under SEBI Regulations. Following these recommendations UTI placed all new schemes, launched after July, 1994 under SEBI Regulations. The regulatory structure of UTI can be discussed in terms of its organizational structure, its investment constraints and its disclosure norms. In terms of its organizational structure, there was a combination of all the roles, the sponsor, the manager and the trustee in one body. The board of trustees was entrusted with the responsibility of superintendence, direction and management of the affairs of the UTI. Each of the trustees was appointed by the sponsors themselves with the lOBI, by virtue of being the largest stakeholder, having the largest number of trustees of four. The RBI appointed one and the chairman of UTI was appointed by the lOBI in consultation with the Union government of India. This meant that there was no arms length relationship between the sponsor, the trustee and the funds manager. The corporate governance structure did not do much to discourage opportunistic behaviour. UTI's investment objectives allowed it to borrow to meet dividend repayment obligations with a limit of 10 percent of the fund's NAV. It could also act like a lending agency and lend to corporations and cooperative societies engaged in industrial activities. No schemes of UTI were allowed to invest more than 5 percent of the NAV in the equity of a specific company and not more than 15 percent in the securities of a company. UTI could also invest in the securities of an unlisted but soon to be listed company to the extent of 20 percent of its NAV. New company investments can go up to 1 0 percent of the securities issued or 30 percent of the aggregate assets. 63 In tenns of disclosure nonns, UTI had to state the objectives clearly in the offer document and reveal scheme perfonnance once a year. But most importantly UTI had no obligation to reveal its portfolio. UTI had no obligations in tenns of controlling expense ratios but charged no management fees. While there are other issues related to the flagship scheme US-54, such as nondisclosure of the NAV, we shall restrict ourselves to these fundamental regulatory aspects. Right away one notices that UTI had neither the corporate governance structure nor the transparency in tenns of portfolio disclosure to ensure that moral hazards would be contained. It seemed a perfect recipe for opportunistic activities through dubious investments. Ghosh (1999) in his review of the Parekh committee report 23 points out how the portfolio structure of US-64 was substantially altered to become more equity heavy ( 63:37 equity- debt composition) in 1998 compared to 1986 (21:79) without infonning the investing public. Faced with decreasing returns from debt UTI had no option but to tum to equity for better returns as the US -64 sought to maintain returns and had been dipping into its reserves to continue paying dividends unrelated to perfonnance. Trustees had allowed the situation to deteriorate when they were to be the protectors of investor's interest. The Parekh committee believed that the dominance of the public sector sponsor nominated trustees allowed the government to abuse the scheme and use it to support the government's disinvestment of public sector undertakings. The committee found 2l The Deepak Parekh committee was set up in the wake of the crises which faced the US-64 in 1997- 98. The committee recommended that independent trustees be appointed and that each fund should have a separate fund manager. For a detailed commentary on the happenings see The Hindu. Business, January 10,2002. 64 that investment decisions were too concentrated in the hands of the top management and that there was no independent fund management team. While these observations were general to UTI, the main focus was the US- 64. These distortions can be attributed primarily to the lack of proper corporate governance structures and also lack of transparency. Investments were made with little regard for generating best retums. 4.2.2. SEBI Regulations and its constraints on fund behaviour SEBI (Mutual Fund) Regulations of 1993 replaced the guidelines issued in 1990 by the government of India. This in turn, was replaced by a comprehensive set of regulations in 1996. These regulations laid out the corporate governance structure, the investment constraints, expense caps and disclosure requirements. The regulations have undergone continuous revisions as and when weaknesses were seen. A person or group of persons (sponsors) interested in setting up a fund has to apply to the SEBI. The latter, in turn, goes through the credibility of the sponsors and decides on the registration after certain minimum conditions are fulfilled. The mutual fund itself is to be constituted in the form of a trust (Indian Trust Act) to be executed by the trustees in favour of the investors. The trustees themselves are appointed by the mutual fund and two thirds of them are to be independent not aSSOCiated with the sponsor in any manner. The trustee members are to be people of established integrity. The trustee or the sponsor then appoints the AMC. The AMC itself must have proven track record and reputation for fairness. The tnustees ensure that the AMC functions in accordance with the interests on the investors 65 and within the framework. of the regulations. Hence, SEBl's regulations create a corporate governance structure with three distinct arms and further guidelines ensure arms length transactions between the three. Funds investment regulations are contained in SEBI regulations schedule VII. Some of the more important ones are spell out below: 1. No mutual fund under all its schemes should own more than ten per cent of any company's paid up capital carrying voting rights. 2. Inter-scheme transfers of securities within a fund house are to be done at market prices and on a spot settlement basis. 3. The initial issue expenses in respect of any scheme may not exceed six per cent of the funds raised under that scheme. 4. No short sales are permitted. 5. No investments in securities of unlisted companies which are associated with or part of the sponsors group of companies. 6. No mutual fund scheme shall invest more than 10 per cent of its NAV in the equity shares or equity related instruments of any company. This limit is exempted for investments in index fund or sector or industry specific scheme. These are some of the investment restrictions and are comparable in some ways with that of the UTI except that there were no specifIC inter-scheme transaction specific regulations for the latter. The issue of sponsor related companies was also not touched by UTI regulations. In order to contain the expenses that could be charged to the investors SEBI decided to go in for direct control. These regulations are contained in chapter 66 VII (General obligations) and Schedule X (initial issue expenses). These are briefly outlined as below: 1. The Asset Management Company is allowed to charge investment and advisory fees limited to (i) One and a quarter of one per cent of the weekly average net assets outstanding in each accounting year for the scheme concerned, as long as the net assets do not exceed RS.l000 million, and (ii) One per cent of the excess amount over RS.l000 million, where net assets so calculated exceed RS.l000 million. For schemes launched on a no load basis, the asset management company is allowed an additional management fee not exceeding 1% of the weekly average net assets outstanding in each financial year. 2. The AMC is allowed to charge initial issue expenses and recurring expenses. These cover the following expenses. Marketing and selling expenses including agents commission Brokerage and transaction cost. Registrar services for transfer of units sold or redeemed. Fees and expenses of trustees. Audit fees. Custodian fees. Costs related to investor communication. Cost of providing account statements and dividend/redemption cheques and warrants. Insurance premium paid by the fund. 67 Costs of statutory advertisements (included through an amendment in January, 1998). Other costs approved by the board (included through an amendment in January, 1998) Besides spelling out the various costs that could be included in the overall operating expenses which included the advisory or management fee, SEBI also placed a cap on the same. The operating expense ratio (ER) as a percent of assets managed is capped in the following manner: (i) On the first Rs. 1 000 million of the average weekly net assets 2.5% (ii) On the next RS.3000 million of the average weekly net assets 2.25% (iii) On the next RS.3000 million of the average weekly net assets 2.0% (iv) On the balance on the assets 1.75%. For debt funds such recurring expenses are to be lesser by at least 0.25% of the weekly average net assets outstanding in each financial year. Hence SEBI has ventured beyond other regulatory systems across the globe by placing ceilings on the amount of expenses and fees and not allowing forces to detennine the same. What is notable is that the management fee is tied to the asset size. This could be construed as indirectly related to perfonnance as the asset size of an open end fund also depends on the relative return generated. We would cali this as 'weak fonn 'of perfonnance related fees. It assumes that fund flow depends on fund perfonnance. So the Regulations have also provided a carrot to perfonn better. But SEBI has ventured beyond other regulatory regimes across the globe by placing 68 ceilings on the amount of expenses and fees and has not left the same to competition. The schedule X deals with expenses of funds with and without entry and exrt loads. Mutual funds are allowed 'load' funds or 'no-load' fund or a mixture of them. For open-ended schemes floated on a 'Ioad' basis, the initial issue expenses can be amortized over a period not exceeding five years. But if any Issue expenses are incurred during the life of open-ended schemes they cannot be amortized. If the fund has no entry load then the AMC is pennitted to charge an additional fee not exceeding 1 percent of the NAV. SEBI has also gradually strengthened the disclosure rules for funds. These are contained again in chapter VII (General Obligations). These obligations require semi-annual disclosure of portfolio and accounts in a widely circulated English Daily within one month of the half year ending. Overall SEBI regulations have tried a mix of better corporate governance structure, infonnation flow and direct controls over expenses and fees to ensure that the investor is given a fair deal. These two regulatory regimes are compared in table 4.1 given below. 69 Table 4.1 - Differences in the Regulatory systems Regulatory requirements uri Act SEBI Regulations Implications if Followed Three tier structure of No yes Improves due diligence corporate governance and reduces moral Including lhe sponsor, hazards. Asset Management company and the Truslee Ceilings on management No management yes Can forcibly reduce fees and operating costs. fees and cost costs and especially ceiling. useful W market lacks competition. Investment constraints on Limits mentioned Limited to 5% maximum Can reduce moral investment in a particular but lack internal for a specific company Hazards and ensure company checks to ensure and not permitted to belter portfolio it is fOllowed. keep cash holdings selectivity and returns. beyond 15 days. Borrowing and lending Could borrow and Limited borrowing Can be a problem in constraints. lend. permilted. case of heavy redemptions causing sales at low prices. Disclosure norms. No obligation to Specific company Disclosure of holdings disclose portfolio. holdings disclosed reduces moral hazards Initially annually and but can lead to later semiannually. imitating behavior W more frequent. The above differences in regulations lead us to two conclusions. Firstly, UTI investors have greater risks due arising out of principal - agent problems. The chances of poor portfolio selectivity and consequently poorer returns are greater when compared to funds regulated by SEBI regulations. But at the same time UTI is a non-profrt public sector with no management fees charged and possibly expenses also being low. These lower expenses could make up for the slack in performance when the NAV is considered. Further in the case of funds regulated by SEBI there could be a greater assurance in terms of portfolio selectivity but this could come at a certain cost. The cost of regulations is both direct and indirect. The direct costs are the compliance costs. The compliance costs include costs of the trustee board, shareholder intimations and disclosure costs (personnel, printing and 70 mailing) and SEBI fees 24 The second type of cost is indirect and often ignored. They involve the possible negative effect of regulatory constraints such as investment limns and portfolio disclosures. Investment limns reduce flexibility with respect to profrtability investments. It forces a certain minimum amount of diversifICation on a fund. For instance, a diversified equity fund has to invest in a minimum of 10 secumies as no equity investment in a single company can cross 10 percent of the assets. This might sound reasonable enough but never-the-Iess is a restraint even if a deviation is profrtable for the fund's investors. This also has to be maintained keeping in mind that different security prices could move in different directions changing their percent composnien in the portfolio. This requires constant churning of portfolio and hence additional trading costs. The disclosure of portfolio also has a cost in that it allows traders to 'front-run'2S the fund. This means that the traders or other investors can guess fund behaviour through fund flow and push up the price of securities that the fund intends to buy and push down the price of those that the fund intends to sell. This reduces the realizable retums of the fund. It can also be termed as 'free-riding' on the research of the fund by outsiders. Hence, what we see is that the stronger SEBI regulations benefrt in tenms of investor security but can have possible costs. Does this regulatory system deliver better when compared to the weaker regulations of UTI? This is the question that we seek to answer. In the next section we outline the empirical framework. 14 Payal Malik (20M), in a survey finds that the present SEBI regulations are perceived to be of the high cost-high benefit type by Mutual Fund managers. She estimates the compliance costs to be an average of 5.35 percent of the overall operating expenses of the AMC's. It is comparable to the US but much higher compared to similar markets such as Thailand. "See Wermers (2001) for a detailed study of the same. 71 4.3. Empirical framework, data description and methodology 4.3.1 Empirical framework In order to compare the performance of funds we first calculate the monthly retums. The same has to be done for a minimum of 60 months for a valid risk-return analysis. Monthly returns are computed using the formula: returns NAV T - NA V,-I NAV,_, (4.1) Where NAV._. is the month end NAV and NAV._. in the month beginning NAV, the monthly retums are then used to compute various risk-return ratios. They are given below. Sharpe Ratio (4.2) Rp is the average fund return. RF is the risk free return. The risk free return here is taken as the bank interest on one year term deposits. Here the one year term deposit is used as the risk free rate of interest as the general public did not have access to treasury bills during this period and public sector banks were considered risk free. 0" p is the standard deviation of the portfolio. The results of this will help us to understand the risk-retum performance of various funds. Information Ratio (IR) (4.3) 72 Where Rm is the return on the benchmark - in this case the BSE-sensex><' and 0'" is the standard deviation of the excess return or returns of the portfolio in excess of the market. This ratio focuses on the risk return generated by the managers' ability to use information to deviate from the benchmark, the higher the better. The standard IR measure, however, runs into problems if there are negative excess returns. Hence, we use an alternate measure called as the 'Modified Information Ratio' (MIR). This is given as below. MIR = (4.4) The only change is that the denominator, the standard deviation of the excess returns, is modified by adding an exponent. The exponent is excess return divided by the absolute value of the excess return. When the excess return is positive the standard IR and MIR the same. When excess return is negative, the IR and MIR can be very different. Jensen's Alpha (4.5) Equation 3.3.1.5 helps us to arrive at the Jensen alpha. The IIp gives the systematic risk and Op gives the fund managers superior stock picking capability. A positive alpha indicates a fund manager's superior stock picking talent. This regression is run once each for the entire SEBI regulated funds together and for 16 The benchmark has been chosen on the basis of the fact that it was the most popular reference index which funds themselves opted to be compared against. The author acknowledges literature discussions on the sensitivity of measures to the chosen benchmark and feels the BSE sensex is a self selected option of funds. 73 UTI funds together in a panel data fonnat. It helps us to arrive at the differences in the average alpha for funds under SEBI regulation and UTI 1963 regulations. Existence of opportunistic behavior should resutt in an inferior average alpha for UTI funds. The following equation gives the panel fonnat of the regression. (4.6) This helps us to capture the average alpha or stock picking skills of UTI funds as compared to the average for the rest. We expect that if there was opportunistic behavior in stock picking then, the average alpha for UTI funds should be definitely lesser and possibly even negative. Further, Since UTI charges no management fee the comparisons might be slightly distorted in its favor. To compare like-with-like we add a minimum management fee of one percent per annum for the private funds. This is lesser than the 1.25 percent allowed by SEBI for the first 1000 million in assets. For monthly returns this works out to an addition of 1/12 percent. This will give as even truer picture of comparison when it is compared with the same regression resutts for the SEBI regulated funds. 4.3.2. Data A sample of ten open end diversified equity funds with growth objective has been selected. The period of the study is May, 1995 to October, 2000. It gives us 66 monthly returns for each fund which is sufficient for risk weighted returns comparison. The time period was selected to give us a comparison of funds perfonnance during the existence of the two different regulatory structures. The criteria for selection of funds were that they should have started before 1996 for the funds that came under SEBI regulations and for the UTI, the funds 74 should have been started before 1994. This is because all funds of UTI that were started after 1994 were 'voluntarily' brought under SEBI regulations. Hence these have been excluded as they would not serve the purpose comparing the impact of regulations. Further, the term 'voluntary' is a vague concept with no definite obligations and place such funds neither under SEBI nor under UTI with definiteness. It was imperative that the funds selected be alike in terms of their investment objective. They are all alike in terms of investment objective as they focus on a diversified equity portfolio with a growth objective. The sample size is however, limited as data availability in a continual manner was a serious constraint. Most funds reported their NAV on a weekly basis which made issues more complicated as the NAV's used for monthly return calculations had match in terms of the dates. NAV data have been sourced from UTI, the database NAV India" and individual fund house websites. The sample is described in the following table. Table 4.2 - Sample of funds Fund Fund House Inception 1. Master Gain UTI April 1992 2. Master Growth UTI February 1993 3. Master Plus UTI December 1991 4. HDFC Prudence' HDFC February . 1994 5. HDFC Capital Builder' HDFC January 1994 6. HDFC Equity' HDFC December 1994 7. Franklin Blue Chip" Franklin Templeton November 1993 8. Franklin Prima" Franklin Templeton November 1993 9. Franklin Prima Plus" Franklin Templeton September 1994 10. JM Equity JM December 1994 All these funds W9f9 initially a part of 2d" Century Asset Management Company taken owr by Zurich India Mutual Fund. Zurich India was itself taken over by HDFC Mutual Fund in 2003 " The database NA V Indian is mutual funds specific database which is venture of Capital Market Limited. 75 - These funds we'" taken over from Pioneer ITf Mutual Fund by F"'nlcHn Templeton mutual fund in 2002. Data for UTI is sourced from UTI. Data for Franklin Templeton funds is sourced from the Funds website and for other funds data is from /olAV India data bass. The sample contains three funds of UTI, three of HDFC taken over from Zurich, three from Franklin Templeton taken over from Pioneer ITI mutual fund and one from JM fund house. In total, we have ten funds selected from four fund houses. Very consciously the flagship fund of UTI, US-64 has not been chosen in the sample. This fund is an income fund and was the subject of intense controversy when ~ collapsed in the late 1990s. It was probably the best example for demonstrating moral hazards due to poor transparency or information disclosure. But, at the same time, this fund was also most abused by the government to induce deliberate interventions in the market and also support government public sector disinvestments. This fund is there avoided as ~ is not a pure equity fund and secondly its portfolio distortions are not entirely due to internal opportunistic behaviour but due to extemal constraints in the form of govemment coercion. 4.4. Results and AnalYSis We start by taking the performance of funds in terms of the equations which define Sharpe ratio, Information ratio and Jensen's alpha and then rank the funds in terms of each measure. These results are given in the table 4.3 with explanations following. 76 Table 4.3 - Sharpe ratio rankings for sampled funds Rank Fund Name (rp - r f ) Up Sharpe Ratio 1 Franklin Prima plus 1.06 9.12 0.12 2 Franklin Blue Chip 0.96 8.59 0.11 3 HDFC Equity 0.79 8.22 0.09 4 HDFC Prudence 0.39 4.9 0.08 5 Master plus 0.14 7.53 0.02 6 Franklin Prima 0.17 10.62 0.02 7 Index -0.01 8.64 -0.001 8 Master growth -0.1 8.09 -0.01 9 JM equity -0.14 9.34 -0.02 10 HDFC capital builder -0.15 7.52 -0.02 11 Master gain -0.43 7.66 -0.06 All computatIons by the authors Among the funds we see that two of UTI funds have negative Sharpe ratio which indicates that the return per unit of risk defined as standard deviation of the fund returns is negative. Two other private sector funds regulated by SEBI also demonstrate such negative Sharpe ratios. These four with negative Sharpe ratios also have a lower ranking than the index. This demonstrates that a passive investment in the index would have delivered superior returns. While all other private sector funds have demonstrated positive returns per unit of risk one from the UTI stable - Master plus fund joins them. This fund has a rank of 5 with a very small Sharpe ratio of 0.02. There can be nothing extremely positive that can be said about UTI funds in terms of this measure. While giving 77 these ranks one has to keep in mind that UTI had substantially lower expense ratios (ER) as compared to all other funds. This meant that the deductions from the unn asset value in terms of expenses would be smaller as compared to other funds. The ER of UTI is lower due to two reasons. Firstly, due to the possible economies of scale that ~ enjoyed as it was substantially larger than all other funds. Secondly, UTI did not charge management fees as compared to the 1 .25 - 1 percent charged by other funds. In fact, UTI cross subsidized the expenses of some domestic funds such as US-64 through the earnings/fees generated by its funds meant for non- Indians. The hypothesis that UTI enjoyed economies of scale is untested but it certainly did not charge management fees which gave it an edge in terms of performance after expenses were deducted. So the performance of UTI funds appears all the more unimpressive. Table 4.4 gives us the IR and MIR rankings. Since IR gives the returns of a fund as compared to a benchmark (excess return) per unit of tracking error (standard deviation of the excess return) it is different from the Sharpe ratio which gives returns in excess of the risk free return divided by the portfolio returns standard deviation. The MIR is also calculated to avoid errors which might creep in due to negative excess return. 78 Table 4.4 - Infonnation ratio (IR) rankings for sample funds Rank Name R" CT" IR MIR 1 Franklin Prima plus 1.07 9.12 0.12 0.12 2 Franklin Blue Chip 0.97 B.59 0.11 0.11 3 HDFC EqUity 0.8 8.22 0.09 0.09 4 HDFC Prudence 0.4 4.9 0.08 0.08 5 Master plus 0.15 7.53 0.02 0.02 6 Franklin Prima 0.18 10.62 0.02 0.02 7 Master gain -0.43 7.66 -0.05 -27.99 8 HDFC capijal bUilder -0.14 7.52 -0.02 -33.40 9 Master growth -0.09 8.09 -0.01 -36.18 10 JM equity -0.13 9.34 -0.02 -41.66 All calcu lations by the authors The rankings of funds in terms of I R are the same as that for the Sharpe ratio. But once we use the MIR the ran kings change for funds with negative excess returns. But overall it is the same story for UTI funds. The IR or the MIR give us the return tQ the risk taken by the fund manager for deviating from the market risk or iI is the riSk-return for only the fund managers selectivity skills and not for the returns due the general market movement. In this case Master plus is the only UTI fund to shOW a positive return for the fund manager's selectivity skills but is a very small 0.02. The other two funds of UTI have negative returns to show for deviating from the martel. Two other private sector funds also show negative MIR. A fund house like UTI ought to demonstrate better selectivity skills given iIs considerable experience in the Indian markets. But the new come private funds have done better. It is difficult to accept that this is due to inferior management and only raises 79 doubts lack of transparency leading to stock selection not based on maximizing retums. In order to carTY out the empirical analysis, we first need to test for poolabilty of data. The pool ability test which uses the F- statistic is carried out for both the fund groups separately. The F- statistic is computed as below. F=(S,-S\)/[(n-l)(K+l)] _ F[(n-I)(K+l),n(T-K-I)] (4.7) S\ l[nT -n(K + I)] 8 3 -8, is the difference between the collective residual sum of squares and the sum of all individual regressions residual sum of squares. N is the number of cross sections. T is total number of observations and k is the number of regressors. This test aims to see if there is any difference between the slope coefficient of the cross sections. The null hypothesis is that the slope coefficients are the same for all. Table 4.5 - Poolabiltly test results for SEBI regulated funds 53 15513.79 5, 14551.66 N 7 K 2 T 448 5 3 -5, 962.1319 nr 53.45177 dr 4.67148 F 11.44215 80 The Ho is rejected at 5% level of signifICance as the tabulated F(18,3115) = 1.61. The calculated F > tabulated F. The test shows that the F-stat is SignifICant or the data is not poolable. Hence we go in for the Hausman specification test to establish whether we need to go in for a fixed effects or a random effects model. Table 4.6 - Hausman Test for SEBI regulated funds Variable Coefficients (fixed) 0.708 chi2(1) - 0.00. Prob>chi2 = 1.0000 Coefficient (random) 0.708 Difference 5.55e-16 Standard error 0.0028 The Hausman specifICation test shows that the model is a random effect model. The null hypothesis, Ho, is random effect and it is not rejected in this test at 5% level of significance.we similarly carry out the poolability test for UTI funds. This is given as below. Table 4.7 - Poolability Test for UTI Funds s3 4370.355 sl 4356.656 N 3 K 2 t 192 s3-s1 13.69893 nr 2.283155 dr 7.683697 F 0.297143 The Ho is not rejected at 5% level of signifICance. The tabulated F(6,567)= 2.12. So calculated F< tabulated F. The test is showing that the F-stat is insignificant or the data is poolable. Hence in this case we go in for OL5 pooled regression. 81 We next move on to computing the average Jensen alpha for the two set of funds using equation 4.7 for each separately. We use a random effects model for SEBI regulated funds and a OLS model for UTI funds. Table 4.9 gives us the results for funds regulated by SEBI. Table 4.8 - Average Jensen alpha for funds regulated by SEBI Variable C B Coefficient 0.45 0.71 Adjusted R-squamd 0.52 Std-error 0.27 0.03 t .. tatistic 1.60 21.8 Probability 0.10 0.00 Table 4.8 gives us a positive alpha of 0.45 Significant at the ten percent levels. This means that SEBI regulated funds have an average positive stock selection skills. The beta is 0.71 which shows the degree to which their retums are explained by the chosen benchmark. Their deviation from the index through altemate stock composition shows positive rewards. In a competitive market funds are constrained to spend money on research to identify stocks that could deliver better returns than the index. Such performance is rewarded by investor through further inflows of money/investment into the open end equity fund. Enhancing the assets managed in tum is an incentive for the fund manager as the management fee is fixed as percent of the same, higher the assets larger the absolute fee. Though there was no use of performance based fees during this period the fact that absolute fees were linked to asset size and that there was no indexing of the fees in any manner to inflation would itself to a strong incentive to enhance performance. The manager's interests are therefore inter-twined with that of the investors. This is proven by the average positive alpha. 82 Table 4.9 - Average Jensen alpha for UTI funds Variable c Coefficient -0.12 0.71 Adjusted R-squared 0.62 Std-error 0.35 0.04 t .. tatl.tlc -0.36 17.49 Probability 0.72 0.00 Table 4.9 gives the results for UTI funds. Our hypothesis of inferior stock selectivity is borne true by an average alpha which has a negative sign (as expected) but is not signifICant. The low significance part could be due to the small sample. But even with this small sample we have no signifICant tock selectivity skills demonstrated. The average beta of 0.71 shows that the UTI funds did not have the same risk as the index, but their deviation from the index composition has not been met with any reward. This leads us to the issue of possible clash of interests and opportunistic behaviour. It could be argued that this lack of stock selection skills is due to purely bad decisions made in good faith or, in other words, the fund manager is inefficient. If this were the case then one might have expected change in fund managers. UTI had no specific manager, though they continued despite poor performance. But poor selectivity and good faith is hard to accept when you expect a fund manager in UTI to have considerable experience and, . therefore, better knowledge of Indian companies and potential good stocks as compared to newer fund houses. We would prefer to conclude that the difference in alpha is due choice of portfolio-a portfolio the fund did not have to disclose based on signifICant 'self-interest' and not 'investor-interest'. The difference in the alpha of 0.45 may contain some element of misjudgment as well but is too large. 83 The comparison so far has been between UTI funds, which charged no management fee, and the SEBI regulated funds which were allowed to charge management fees between 1 and 1.25 percent of AUM. Therefore, our comparison, so far, would tend to bias the results in favour of UTI funds. The NAV of SEBI regulated funds would be lesser by the extent of the management fee, whereas, UTI funds operated on a non-profit motive. In fact the US-64 was cross-subsidized by income earned from floated for non-Indian citizens. Such being the case, we think it might be pertinent to add back the management fee to the return of other funds and compare the results again. But this involved certain difflCuHies. Data on actual management fees charged were not available for these funds. The only aHernative was to usae an average for the industry. The Mutual Fund Yearbook 2
gives us aggregate data with respect to expense categories and asset size for pure equity funds. From this it was possible to calculate the average percent of management fees charged by the SEBI regulated pure equity funds. Table 4.10 gives these numbers. Table 4.10 - Average Assets and Management fees ofa pure equity funds 1996 1997 1998 1999 2000 Assets Size (Rs.Miliions) 496883 394899 395019 487253 91108 Management Fee (Rs.Miliions) 5783 4639 4353 3985 6724 Management Fee (% of assets) 1.14 1.17 1.09 0.8 0.74 All computations by the authors 21 The Mutual Fund Yearbook is a joint publication of Association of Mutual Funds of India (AMFI) and UTI Institute for Capital Markets (UTIICM). It contains aggregative data pertaining to assets, income and, expenditure of di fferent categories of funds. 84 From the above table we get an average management fee of 0.97 or close to 1 percent of the assets. We divide this by 12 and add the resultant to each monthly retum of the private funds. Using these retums we then run regression 4.8 again. Table 4.11 Average Jensen alpha (with average management fee included) for SEBI regulated funds Variable Coefficient Std-error t..atatistic Probability c 0.53 0.26 1.9 0.05 0.71 0.03 21.79 0.00 Adjusted squared 0.62 Table 4.11 gives us the new results for private SEBI regulated funds with the management fee added to returns. We, obviously, get a higher alpha of 0.53 and also now signifICant at the five percent levels. These results help to further consolidate our hypothesis that SEBI regulated funds had a better incentive to ensure due diligence in portfolio selection and hence, perform better. 4.5 Chapter Summary This chapter set out to examine if SEBI regulations were too costly in terms of direct and indirect costs of regulations. Too costly here is implied to mean that the net benefits of preventing fraud on the investor through self interested opportunistiC portfolio selection were less than the costs. This was sought to be studied by comparing them w ~ h a set of UTI funds which did not have the sort of regulations that encourages due diligence or prevents opportunistic behaviour. 85 Our findings show that regulations have delivered in terms of ensuring far superior selectivity skills in portfolio and hence also return. For SEBI regulated funds the average Jensen alpha is 0.45 without adding management fee back to the returns. With the management fee included, the Jensen alpha is higher at 0.53. For UTI funds the average Jensen alpha is negative but not significant. On an average, we could say that regulations can take substantial credit for the average 0.45 to 0.53 higher selectivity skills. A combination of incenlivising management fee tied to asset size) and disclosure norms on the part of SEBI regulations and competition between funds can stake claim to this. The net benefit of regulations can be said to be positive. To go further, we feel regulatory disclosure can be made more periodical given its benefits. At present portfolios are disclosed twice a year. This still provides for substantial leeway in terms of hazardous portfolio selection which in tum can compromise on the risk level of the fund and also the return. Besides, we also feel that the Regulations need to move away from the 'weak form' of performance fees to a more stronger form. There is no certainty that fund flows are primarily driven by relative performance. In fact marketing expenses could influence the investor's choice more than the relative performance. The move to a 'stronger form' performance fee has to aHematives. The choice is either a symmetric or fulcrum fee as allowed in the US or a non-symmetric fee. The literature on this is divided there is no clear indication as to which is better. From a gradualist perspective SEBI could start with a non-symmetric fee which has a base fee lower than the present levels and a performance fee that rewards good relative performance. Relative performance can be measured with respect to the benchmark index that the fUnd opts for com parison in its offer document. 86 CHAPTER-V COST - llANAGEMENT FEE REGULATION AND ECONOMICS OF SCALE CHAPTER-V Cost- Management Fee Regulation and Economies of Scale 5.1. Introduction In this chapter we examine the direct regulations of expenses of mutual funds. The manner in which the operating expenses or expense ratio (ER) and management fees are capped by SEBI regulations has already been outlined in Chapter 4. SEBI has ventured beyond other regulatory regimes for mutual funds in specifying the percentage limits for ER and management fees. SEBl's i m p l i c ~ assumption appears to be that left to the market mutual funds would tend to overcharge the investor in terms of expenses and fees charged for services. This is possible in a market which lacks competition or lacks a strong competition law to restrain anticompetitive behaviour. In the initial phases of regulation, when the mutual fund market was in ~ s infancy, it was probably the right thing to do. What is interesting about SEBl's regulation of cost is that it places a progressively lower ceiling as the asset size increases. Again this seems to be based on the i m p l i c ~ assumption that mutual funds would have economics of scale and therefore a progressively lower average expense ceiling is viable. In trying to set ceilings SEBI risks two types of errors. Firstly, the ceiling could be too high, therefore, it could allow funds to charge higher than needed. Secondly it could be too less prompting funds to either sacrifICe their fees to cover expenses or resort to means of cutting costs in manners which could affect returns, which might be detrimental to the investors. For instance a fund might decide to cut back on research expenses. 87 As the size of a fund increases economies of scale in administration expenses is possible. The benefits of lower average costs would be passed on to shareholders in an industry that is competitive. This would lower expense ratios. For fund complexes with a large number of funds, there could be further economies possible in sharing of research expenses, general advertisement expenses and marketing infrastructure expenses. Since fund managers derive their profrts from management fees they would certainly strive to increase the asset size under their management. For this they have to compete with other funds for investor's pool of savings. It is possible that funds might also decrease their management fees voluntarily to attract more funds. Theoretically, this is viable up to the point where the marginal benefrt of fee reduction (in tenns of asset increase) is equal to the marginal cost (reduction in total fees). Reduction in fees WOUld, however depend on investor preferences between price and non-price factors in choosing funds to invest in. Non-price competition through product diversifICation, advertisements and after sales services can help to reduce elasticity of demand with respect to fees. But given that most services are mandatory, price competition may still be important. The fund industry has grown substantially in tenns of the number of funds, variety of funds and also number of fund houses. The number of fund complexes has risen from 10 to 29 during 1993 - 2006, with a total of 449 funds and 55 new funds added in 2005 - 2006 29 The types of funds have also been getting increasingly diverse trying to cater to the risk profile of potential clients. This in itself can be a pointer to increasing competition in the mutual fund industry and, therefore, if there are any greater economies one would expect funds to lower ER 29 Figures sourced from Association of Mutual Funds of India (AMFI) monthly, March 2006. 88 for investors. Further compe@on can also induce lower ER through lower management fees even while keeping other expenses at the regulatory level. This chapter has two objectives. First it intends to compare cost regulations with actual costs and see if there are greater economies of scale greater than what SEBI assumes for funds. Next, it also aims to see if funds compete in terms of price (management fee). This chapter is divided into five parts. In the second part we quickly outline the regulations and proceed to the empirical outline in the third. In the fourth part we present the results and analysis. The summary and conclusion is provided in the last part. 5.2. Regulation of costs and management fees In administering the funds, vested with them by investors, mutual funds incur operating expenses (ER). These expenses are deducted from the assets and therefore, are crucial in determining net returns to investors. The regulatory authority, Securities and Exchange Board of India (SEBI)30, has placed limits on the ER that can be incurred and any excess expense has to be made good by the fund sponsors without affecting the investors. One of the most important components of operating expense is the management fee. The management fee has been capped and the cap varies for asset size. For instance all funds can charge a fee of 1.25 percent up to one billion lO Clause 52, chapter VII , SEBI Mutual Funds (Regulations) Act of 1996. 89 rupees in weekly average assets and 1 percent on assets beyond this level. Perfonnance based fees are not common and at the point of writing there was only one fund which had such a fee" . The other operating expenses include expenses on marketing and selling, brokerage, investor communication, registrar services, trustee fee, audit fee, custodian charges, and SEBI fees. The overall expenses (including management fees) are also capped and cannot exceed 2.5 percent for the first one billion Indian rupees, 2.25 percent for the next 3 billion, 2 percent for the next 3 billion and for the balance 1.75 percent for all equity funds .This limn in itself would cause the average expenses to decline as asset size increases. For funds investing in fixed returns securities the expense ratio has to be lesser by 0.25 percent at all levels. Expenses in excess of those fixed by regulations have to be borne by the sponsor. chart 5.1 gives a graphic description of the regulatory ceilings for equity, debt and index funds. Chart 5.1 Regulatory ceilings for Expense Ratio and Management Fee (as percent of AUM) for equity and equity oriented balanced funds Regulatory ceilings for Expense ratio and Management Fee (as a percent of assets) I Sahara Wealth plus Fund was launched in July 2005. It offers a variable pricing option which Ipmds on whether the net portfol io returns is above zero and greater than the benchmark. 90 5.3. Empirical framework, Data description and Methodology 5.3.1. Data The time period considered is April, 2000 (financial year 2000-01) to March, 2007 (financial year 2006-07) giving a total of seven years. Data on the relevant variables such as expense ratios, management fees, marketing expenses, turnover, have all been primarily sourced from the annual reports of various fund complexes. The data is unbalanced for reasons of lack of access to continuous reports. lack of complete information in some annual reports and also because some funds started after 2000 and some funds were terminated before March, 2007. The selection of equity funds was based on SEBI regulations for expenses. For actively managed funds since equity funds and balanced funds were permitted the same expense levels, open ended diversified equity funds and equity oriented (more than sixty percent of their assets in equity) balanced funds were chosen as the target population. Tax saving equity funds were not included as they are close ended funds with expenses in the first three years being inflated to cover initial issue charges. Similarly index funds have also not been included as they do not faU under "actively managed" funds criteria. The choice of funds has been random but restricted to those with an existence of at least three years .Only full years of operation were considered and funds which started mid year or were terminated mid year were excluded as their annualized costs would not reflect what actual expenses could have been (say due to waiver of management fee). 91 Similarly for debt funds, long term and medium term debt funds have been considered with the sample again depending on a minimum of three years of existence for the fund and also availability of data. The sampled funds, their number and average assets, across the years considered are given below. Table 5.1: Descriptive statistics for Equity Funds Yea. 1999_ 2000-01 2001-02 2000 2002-03 2003-04 2004-05 2005-05 2005-07 No. 21 39 53 74 73 69 67 69 AA" 1471.62 1377.4 925.5 822.7 1362.8 2265 2990 5038.8 (1388.4) (1777.5) (1235.5) (951.7) (1779.9) (345) (4241.4) (6974.9) Asset size is in Rupees mil/;ons. Figures in parentheses ore standard devlOlicms. Table 5.1 shows the average assets for equity funds and the average number of funds that were sampled across the time period. Average asset size has shown a significant increase since 2004 ever since the Indian stock markets entered into a bullish phase. The standard deviations (in parenthesis) show that the spread of fund in terms of size has also tended to increase similarly. Table 5.2: Descriptive statistics for debt funds VH. 2000 2001 2002 2003 2004 2005 2006 ZOO7 No. 11 16 17 24 23 24 23 24 4687 5084.8 7005 10882 10402.7 4585.9 2028 1411.6 . AA (3792.7) (1643) (1315.9) (4912.9) (5243.7) (7268) (14578) (11676.6) .. Average asset see in Rupees millions. Figures in parentheses are standard deviations 92 Table 5.2 shows the average assets (standard deviation) and number of funds in the debt category. The size of debt on an average is much larger than that for equity funds. Interestingly the average debt fund size [even if one considers the standard deviations) dropped since 2004, quite contrary to what was seen for the equity funds. There appears to have been a redistribution of assets into equity funds due to high stock market returns and also due to lower interest returns on debt instruments. 5.3.2. Empirical Framework A priori. theoretically the total cost is expected to be dependant upon the size of the fund, the size of the fund family, the entry and exit load and the turnover of the funds portfolio. Cost components such as management fees, and other expenses including brokerage charges, research and marketing expenses are predominantly influenced by asset size. But other characteristics such as total load (entry load and exit load) charged by funds total family average assets and also the turnover of the fund are also important. Entry load serves to defray marketing expenses for new subscribers and exit load attempts to discourage early redemptions which can cause unwanted portfolio churning. Hence, total load should have a negative influence on costs. Family size can help reduce average fixed costs per fund and also reduce the average research expenses for a fund as it is common knowledge/expense whose cost is assumed to be shared between all the funds of a family. Turnover of the portfolio has a positive impact on costs as larger the churning of the portfolio larger the absolute amount of brokerage fees to be paid. These variables are described in table 5.3. 93 Table S.3 - Cost function and explanatory variables Cost.. trotal Includes the sum of managerial fees. marketing pperating expenses and other expenses such as brokerage, p>sts shareholder transactions, registrar fees, etc of the ith fund in year. LAAn ,...og of Average of weekly assets under management of the I"Iverage i lh fund in year. !assets LAASQ. ,...og of Average of weekly assets under management of the fiverage i lh fund in to. year. !assets
LTAA. r-og of It is the sum of the average weekly assets of all the amily funds under management of the fund complex to !assets which the fund belongs to in the to. year. TloIIdtt trotalload Sum of the entry and exij loads of the i" fund in to. year. It is expressed as percentage of investment made and/or redemption. T frumover The minimum of sales or purchases as a percent of average weekly assets By accounting for all these factors the genuine relationship between operating costs and asset size of a fund is sought to be derived. To further check for the relevance of the explanatory variables a simple correlation is presented in table 5.4. We notice that all the theoretically assumed important variables have a II8tiStiC8HY significant relation with the dependant variable Leost. However, the coeffjcients of turnover and \load have unexpected signs. One reason for this could be that funds which have experienced higher marketing expenses and redemptions tInd to have shifted to larger loads and hence the positive correlation. Higher UrIO'I8I' leading to lower costs could be due to the fact that larger fundS selected 94 are predominantly large cap-growth stock oriented and such funds normally evade from too much of churning of the portfolio as it has adverse impact on returns. As expected log average assets (Iaa) and log of family average assets (ltaa) have a reasonably strong positive correlation (O.41).Total load has a small positive correlations the strongest being with family assets. Turnover has small negative correlations with average assets and a small positive correlation with family asset size. It helps to control for these factors before examining the relation between cost and asset size. Table 5.4: Time series, cross section averages of (yeariy) conelations between variables of interest for equity funds Lc:ost AA La Laaaq 1.00 0.99 0.99 0.94 Leost (-10.33) (-508.21) (-31.8) AA 1.00 0.71 0.84 (10.2) (9.7) 1.00 0.96 LAA- (-26.9) LAASQ*** 1.00 LTAA- TLOAD" Turnover AA represents (lIIerage assel sc:e oflhe mdlVlduolfund. LAA is the nalural log of assels . LAASQ is LAA squared @ LTAA is Ihe log offamily assel si=e. Ltao 0.39 (-83.85) 0.40 (-20.16) 0.41 (-0.42) 0.48 (20.16) 1.00 @@ TLOAD is Ihe 10101 load which is the sum of enlry and exilload. Figures in parentheses are I-values. All compulaJians by lire aulhors. Tload Turnover 0.07 -0.09 (-12.91) (-12.59) 0.07 -0.13 (10.26) (1.08) 0.08 -0.08 (23.1) (-12.4) 0.05 -0.12 (29.96) (-11.32) 0.04 0.14 (74.48) (-12.22) 1.00 0.01 (12.53) 1.00 95 In the case of debt funds the family size and average assets are weakly negatively correlated. Turnover, however, has a reasonably strong posnive correlation wnh family size and a negative correlation with asset size. larger funds tend to have less churning but larger families increase the same. In the case of larger families this could probably be due to larger inter-fund transactions. Total load while having a small positive correlation wnh asset Size has a reasonable negative correlation with family size. Hence in the case of debt funds as well all the above mentioned variables are retained. Table 5.5: Time series, cross section averages of (yearly) correlations between va riables of interest for debt funds Least AA Lao LMoq Llaa noad Turnover Lcost 1.00 AA 0.79 1.00 (-9.19) LAA 0.99 0.81 1.00 (-242.59) (9.13) LAASQ 0.98 0.87 0.99 1.00 (-30.64) (8.8) (-26.9) LTAA -0.07 0.Q1 -0.07 -0.06 1.00 (-52.2) (9.06) (-22.52) (21.5) T 0.1 -0.21 -0.14 -0.16 0.23 1.00 (10.35) (-2.09) (-11.45) (-11 ) (-11.4) TLOAD -0.19 0.02 0.03 0.04 -0.18 -0.33 1.00 (-11.5) (9.2) (49.67) (28.93) (85.79) (11.53) Once again all the theoretical explanatory variables have a statistically significant correlation with the dependent variable lcost and all the correlations are of the expected sign. Importantly the correlations with the dependent variable are larger than or equal to the correlations between the explanatory variables. This is 96 useful as it reduces the role of multicollinearity. Although this was not so the case for equity funds given the significance of these explanatory variables log of the variables is used to reduce multicollinearity and the variables are retained. The functional form of the model is therefore given as below: C = C{AA,TAA,TLOAD ,T} (5.1) The empirical methodology consists of using a fixed effects approach to detenn ine the impact of fund size on fund operating cost. Funds are divided into two broad categories - equity and debt. They are taken separately give their different cost regulations and investment portfolios. In order to reduce the multicollinearity problem and to identity presence of economies of scale with varying asset size a log-log model is used. L cost it = a l + a 2 1aa it + a 3 (laa;Y + L j aJx jit + e (5.2) i= 1 ..... N Lcost" is the natural log of operating costs of the ith fund in the year t. Laa., is the natural log of the average assets of the ith fund in the year t. (Lao)' is the square of the log of average assets to capture economies of scale. X. is the vector of fund characteristics that can affect costs. It includes family J asset size (LT AA). total load (TLOAD). turnover (T). and regulatory cost ceiling (RCOST) as a control variable. For both the equity and debt funds the regressions are carried out in two stages. First. without the control variable and then. with the control variable. This helps us to understand the impact of regulatory cost ceilings. 97 5.4. Results and analysis Guided by the Hausman specifICation test provided in table 5.6 a fixed effects model is used for the unbalanced panel data on equity funds. At first the regression is carried out wnhout the control variable of regulatory costs. The resuhs of the same are presented in table 5.7. The results are as expected with the coefficient of laa having a posnive sign and the coeffICient of laasq having a negative sign. None of the other variables have any explanatory power. This indicates that there are economies of scale present and that total costs are increasing at a decreaSing rate. This, however, as pointed out could be merely due to forced economies created by regulations. In order to verify this, the regression is run again with the control variable of regulatory costs. These results are presented in table 5.8. The results show that the coefficient of Laasq is no longer signifICant and has a posnive sign. The control variable has the expected negative sign and has absorbed all the possible economies of scale. Hence, regression results with regulatory cost ceilings used as control variables reveal no true economies of scale. All economies are strictly due to the extent caused by regulations. Mutual funds either have no scope for further reduction in costs or even if there is a possibility they do not show any inclination to do so. Total load is statistically inSignifICant and turnover of portfolio has no signifICant impact on cost. Importantly the resuhs show that overall size of the fund complex does not contribute to reductions in fund expenses. SEBI regulations require that non-fund specifiC expenses be distributed across all funds in the family in proportion to the asset size. These could include 98 research expenses. These expenses have no impact on cost as the family expands. Table 5.6: Correlated Random Effects - Hausman Test Tn'Summary Cross-section random Chi-Sq. Statistic Chi-Sq. d.f. 30.966754 6 Prob. 0.0000 Table 5.7: Regression results without the control variable for equity funds. Variable Coefficient Std. Error t-Statistic Prob. C -3.809811 0.061342 -62.10753 0.0000 lAA 1.027017 0.012975 79.15149 0.0000 lAASQ -0.005694 0.001466 -3.885183 0.0001 LTM 0.004396 0.007591 0.579037 0.5632 TURN 6 . 1 1 E ~ 5 2.46E-05 2.480224 0.0140 TLOAD -0.008418 0.007957 -1.057905 0.2914 R-squamd-O.99 Table 5.8: Regression results with control variable for equity funds. Variable Coefficient Std. Error t-Statistic Prob. C -3.802441 0.060688 -62.65582 0.0000 lAA 1.013185 0.014033 72.19810 0.0000 lAASQ -0.001639 0.002213 -0.740595 0.4598 LTM 0.002288 0.007551 0.303054 0.7622 TURN 5.98E-05 2.44E-05 2.456575 0.0149 TLOAD -0.008071 0.007864 -1.026405 0.3059 R squamd = 0.99 In the case of debt funds again a fixed effects model is applied justified by the Hausman specifICation test in table 5.9. Once again the analysiS is done with 99 and without the control variable. Table 5.10 gives the results without the regulatory cost control variable. As expected, the coeffICient of laa has a positive sign and the coefficient of laasq has a negative sign. Both are highly significant. This implies that debt funds experience economies of scale. But, interestingly, family size seems to have a positive impact on the cost of operation of individual funds. This implies that family size somehow induces diseconomies for debt funds but no similar relation was seen for equity funds. The economies of scale seen could again be due to the forced hand of regulatory cost ceilings. So the process is repeated with the control variable. The resuHs are given in table 5.11 . Table 5.11 shows that the coefficient of laasq retains the negative sign and continues to be significant at the five percent level. What is interesting is that the regulatory cost ceiling is not a significant determinant of the overall economies of scale. Hence, debt funds tend to reveal true economies of scale or economies beyond what is stipulated by regulations. Family asset size continues to have a small impact on the cost of individual debt funds. Total load and turnover do not seem to be important factors in explaining the cost trends. Other factors, such as retums, seem more important in determining costs. The positive impact of family size on debt fund cost is a bit of a puzzle given SEBl's requirement that common costs be distributed on a proportional basis across all funds in the fund family. The possibility of larger families tending to be more immune to cost considerations has less bearing for debt funds which faced a run on their assets given superior equity returns and also decreasing interest rates in India. 100 Table 5.9 : Correlated Random Effects - Hausman Test Test Summary Cross-section random Chl-Sq. Statistic 17.108111 Chi-Sq. d.f. Prob. 6 0.0089 Table 5.10: Regression results without the control variable for debt funds Variable Coefficient Std. Error t-Stati.tlc Prob. C -4.700793 0.169473 -27.73766 0.0000 lAA 1.186152 0.049506 23.95966 0.0000 lAASQ -0.019211 0.004410 -4.356013 0.0000 LTAA 0.022510 0.009575 2.350827 0.0204 TURN 4.67E-06 1.58E-05 0.295249 0.7683 TLOAD 0.000161 0.064823 0.002485 0.9980 R squared - 0.99 Table 5.11: Regression results with the control variable for debt funds Variable Coefficient Std. Error t-Stati.tic Prob. e -4.825338 0.224297 -21.51317 0.0000 lAA 1.246100 0.086270 14.44423 0.0000 lAASQ -0.026089 0.009226 -2.827720 0.0055 LTAA 0.021454 0.009667 2.219334 0.0284 TURN 5.45E-06 1.59E-05 0.343743 0.7317 TLOAD 0.006915 0.065386 0.105759 0.9160 RCOST 0.002230 0.002626 0.849002 0.3976 R - squBf9d 0.99 101 These results make it clear that regulatory ceilings have contained costs for equity funds but have no similar impact for debt funds. In order to understand the sources of such constraints (or the lack of it), for equity (debt funds) a disaggregated look at the composition is attempted. The following section attempts to do the same and also examine fund behaviour with respect to management fees and marketing expenses. Expenses can be split into management fees, marketing expenses and other expenses. The decomposition is done so deliberately in order to have a glimpse of management fee as a price variable and marketing expenses as a non- price instrument of competition. These two variables are much within the control of funds and can be decreased or increased depending on competition levels. Other expenses include costs of investor communication, mandatory fees, and auditing, trustee, brokerage, and other transaction costs. These are the expenses which we would expect to see economies or diseconomies if any. Table 5.12 gives the disaggregated cost structure of equity funds. Small funds with assets less than 1 billion rupees tend to charge expense ratios, which are much below the ceilings on an average. While this is the case for the next two higher asset classes to a smaller extent, the gap disappears for the largest asset class. Even though expense ratios show a falling trend, they tend to collide with the ceiling which is the reason why there were no economies of scale after controlling for regulatory ceilings. 102 Table 5.12: Time series cross section averages of expense components of equity funds variables <-1 AA- 0.43 (0.28) AER- 2.29 (0.31) RER' 2.5 (0.00) AMFEE$$ 1.10 (0.26) RMFEeO 1.25 (0.00) ME .... 0.59 (0.57) RETURNS 52.31 (53.76) OE 0.60 Average assets In rupees billiOns .. Average of actual expense ratios 1 -4 4-7 2.22 5.07 (0.94) (0.84) 2.32 2.20 (0.16) (0.22) 2.39 2.27 (1.11) (0.05) 1.09 1.01 (0.15) (0.21) 1.13 1.06 (0.06) (0.02) 0.68 0.77 (0.30) (0.20) 52.32 50.52 (52.57) (57.26) 0.50 0.42 ."Average of the 10101 operating expenses minus the sum of manage menifee and marketing expenses. $ Regulatory expense ratio ceilings $$ Actual management fee charged @ Regulatory ceilingfor management fee (as a percent of assets) @@ Marketing expenses as a percent of average assets. >7 13.7 (6.56) 2.04 (0.13) 2.00 (0.22) 1.02 (0.08) 1.02 (0.01) 0.54 (0.23) 41.97 (41.67) 0.46 On examining the management fees we notice a distinct tendency among the smaller funds to waive off a portion of the allowed management fees. The gap between the allowed and the actual management fee tends to close until they are the same for the largest asset size. The largest fund class charges the maximum permissible. Most, but not all of the below-the-ceiling expenses can be explained by the lower management fee. Hence, it can be argued that smaller funds try to be more cost competitive by cutting down on management fees chargeable. 103 Marketing expenses as a percent of assets tends to increase as the asset
size increases until it dips for the largest asset category. Marketing expenses are not a transparent category and one can only assume that it is used to cover advertisement expenses or brokerage expenses. It is akin to the 12b-l expenses of the US mutual funds. Such expenses are supposed to attract more investments, increase the size of the fund and thereby lead to cost savings. But new investors also pay an entry load to defray such expenses. Besides, the marketing expenses appear to be negatively correlated with annual average returns. Funds seem to be spending more on brokerage and advertisements to cover for relative underperformance. Given these observations it is a surprise that marketing expenses remain so opaque and unexplained for so long. The fact that the trend in the gap between allowed and actual management fee and the doubtable marketing expenses it appears that funds cost consciousness decreases with size. Other expenses, where economies were expected, predictably dip but increases again at the last asset class. The dip in marketing expenses for the largest class appears to be due to the regulatory constraint along with diseconomies creeping in. Given that diseconomies creep in for the largest asset class the justification for marketing expenses becomes weaker as it would be difficult to meet regulatory ceilings unless they cut management fees. Regulatory ceilings seem to be most effective as a cap for the largest asset class witnessing diseconomies and exhausting behaviour in terms of management fees charged and marketing expenses incurred. But perhaps the ceilings can be lower for all the asset classes given the questionable nature of marketing expenses and their use. 104 For debt funds the actual expense ratio is always lower than the maximum though the gap reduces with size. Managerial fees are waived off at all asset levels with the tendency again reducing with size. Again they appear to be the most important factor for keeping expenses below the ceiling. Marketing expenses show an increasing trend and appear to be justified by the drop in average 'other expenses' though not so much in actual expense ratios which ultimately matters. Once again smaller funds show the willingness to cut costs through waiving management fees and spending less on marketing. This appears to be guided by the need to provide competitive returns. On an average, smaller funds have actually given lower returns and this we feel could explain their attempt to cut costs and give acceptable net returns. But the other expenses' dropping sharply is a sign that substantial economies are present even in the largest asset category. This proves the earlier regression results showing true economies for debt funds. The source of this is not just due to containing management fees but also due to sharp drop in other expenses which strengthen the argument that true economies prevail. Table 5.13: TIme series cross section averages of expense components of debt funds. Variables <1 1-4 4-7 >7 AA 0.59 2.37 5.29 15.56 (0.23) (0.82) (0.88) (8.63) AER 1.68 1.76 1.59 1.67 (0.42) (0.35) (0.32) (0.12) RER 2.25 2.02 1.64 1.72 (0.00) (0.31) (0.38) (0.1) AMFEE 0.88 1 0.88 0.94 (0.32) (0.19) (0.20) (0.20) RMFEE 1.25 1.12 1.05 1.02 (0.00) (0.05) (0.01) (0.01 ) ME 0.42 0.47 0.47 0.57 (0.25) (0.27) (0.27) (0.17) R 7.32 9.11 10.99 10.15 (8.26) (7.07) (5.9) (5.37) OE 0.38 0.29 0.24 0.06 105 5.5. Chapter Summary This chapter has set out to examine if regulatory ceilings are too high or too low. The WOrd "too" is used as one cannot expect a perfect ceiling but at least one which would approximate the actual in a competitive market. Equity funds show no 1rue economies'. Family size does not cause expenses to decline either. Smaller funds tend to behave competitively by waiving fees and keeping expenses below regulatory ceilings. But, as asset size increases, they tend to exhaust most of the allowed expenses. Increase in marketing expenses appears to be a prime reason for the tendency to close the gap with the expense cap. The use of marketing expenses with entry loads and the tendency to increase the former to possibly cover for poorer comparative performance is an issue for concern. Large funds also seem to be experiencing diseconomies with an upturn in 'other expenses' (OEl. Coupling this with lower returns, one is forced to think if a size cap might be needed. Regulations prove to be a relevant constraint for contrOlling costs in the case of larger funds particularly as it helps to curb marketing expenses to accommodate increasing OE within an overall expense cap. Debt funds maintain expense ratios persistently below the ceiling and also show the presence of true economies. The lower expenses are aided by waiving a portion of management fees and also through persistently decreasing OE. The latter is most important for examining economies of scale and upholds the findings of true economies in debt funds. Their main reason for keeping fees lower could be the trend of decreasing returns on debt and good returns in the equity market. This could lead to a refocus of fund manager's interest in debt funds as equity fund 106 management becomes more attractive in terms of fees. Interestingly family size seems to have a p o s ~ i v e impact on debt fund costs. 107 CHAPTER-VI FUND SIZE AND FUND RETURNS CHAPTER- VI Fund Size and Fund Returns 6.1 Introduction The mutual fund industry in India has witnessed a sharp growth in recent times. The growth was almost 50% during the financial year April 2006 to March 2007. The industry size as on October 2007 was close to rupees 5300 billion. Households now save more than 5% of the savings in mutual funds as compared to 0.4% in 2002-03(Reserve Bank of India). Given their importance as instruments of saving and their explosive growth recently we feel it would be important to examine the impact of size of individual funds and fund families on the performance of funds. The importance of such a study becomes all the more apparent considering the voluntary caps imposed on the asset size of individual funds". Negative influences on returns take place as open ended funds faced w ~ h increased inflow may have a problem in terms of both timing and stock selection. Larger funds may also find ~ difficult to be as fleet footed as smaller funds as their transactions in the market are bound to be larger attracting adverse price movements. Persistence of fund performance depends on the ability of funds to continue to display superior micro and macro skills. This gets harder as successful funds attract more fund inflows thereby increasing fund size. 32 For instance Reliance Equityfund decided to cap the amount o/money it raised through an IPO in 2006 (Rs.58 billion made iI the largest fund) because iI was worried about the difficulty of generating good returns with a very large asset si:e. Few olher funds followed pursuit capping their asset s;;e at much lower levels. 108 While this is a cause for concem to the investors in funds. fund managers have every incentive to allow for increased asset size as their fees are a fIXed percentage of the assets. But to do so despite the possibility of declining or lower marginal return constitutes an agency problem detrimental to the investors. It has been pointed out that larger funds generate econom ies of scale and therefore reduCed operating expenses. This could be a valid factor if the retums net of expenses turn out to be higher for larger funds. In the case of this in itself will cause average expense ratios to drop. It would be of interest. therefore. to examine if the net returns justify investment in larger funds. The objective of this chapter is to understand the implications of size of AUM and its impact on performance of open end equity diversified funds. This chapter is divided into four parts. The second part deals with the empirical frame work which discusses the data and the empirical model. The third part discusses the resuijs and the last covers the summary and conclusions. 6.2 Empirical framework and data 6.2.1 Data The lime period considered for the study is August 2002 up to October 2007. This gives altogether 63 months of data. Data on fund net asset value have been sourced from the Association of Mutual Funds of India (AMFI)33. The sample of funds includes 58 open end equity diversified growth funds selected purely on lJ Expenses again h(]lle been capped by SEBI at 2.5 percent of the assets for the first one billion rupees. 2.25 percent for the next 3 billion. 2 percent for the next 3 billion and I. 75 percent for the rest. AMFI;s a vo/unJary association o/mutua/funds in India which seeks to act as a selfregulalory body. 109 the basis of availability of data on monthly asset size from August 2002. The sample data has been sourced from Mutual Fund Insight, a monthly published by Value Research Inc". Since fund retums are due to variations in style such as Large Cap, Mid Cap, Small Cap and Growth stocks, value stocks or blend stocks orientation an effort has been made to track changes in the style of funds over the period. Data on fund style has again been sourced from Mutual Fund Insight. The style of a fund can be described with the help of the following matrix used by Value Research Inc. Table 6.1 Fund style description Growth Blend Value Large Cap A B C Mid cap 0 E F Small cap G H I A fund in the cell A is a fund that invests in Large Cap stocks and stocks which have a high growth potential as reflected by a high Price/Earnings ratio (PE). A fund in cell B invests in large cap stocks but has a mix of growth and value stocks (stocks of companies which have a high book value but low PE and give good dividends as well). A fund in cell C reflects a fund which invests in Large Cap value stocks. Similarly we have Mid Cap and small cap oriented funds that invest in growth, blend or value stocks. While some funds are fixed in terms of their style, J< Value Research Inc. is an independent research company devoted towards the study of mutua! funds. 110 some others are f1exi cap and move between large to small cap stocks. To track such changes we use monthly data on style. Data on expense ratio (percent of recurring expenses to asset size) is sourced from the annual reports of various fund families. Since age generates experience and knowledge of the market it is also included and all that was needed was the inception data of the fund which was available in Mutual Fund Insight. Data on the benchmark S&PCNX500 was sourced from the National Stock Exchange (NSE) website. The CNX500 is a broad based stock market index representing 96% of the total turnover ofthe NSE. The risk free rate used is the three month implicit yield of the 91 day Treasury bill given in the RBI monthly bulletin. Family asset size per month is derived from AMFI. Total load is the sum of the entry load and exit load and data is from Mutual Fund Insight. Flow represents the net inflow of open funds [sales over redemption]. Table 6.2 TIme Series cross sectional averages and standard-deviaUons Mutual fund size On rUI!!:" billion] Upto 1 14 47 7 and more All fund. Assets 0.39 2.21 5.38 15.2 4.08 (Rs.Billion) (0.27) (0.87) (0.86) (8.03) (6.32) Family Asset 66.87 146.52 185.05 246.68 134.48 (Rs.Billion) (83.28) (99) (106.89) (121.17) (118.32) Age 7.82 8.19 9.17 11.09 8.64 (Years) (3.04) (3.03) (254) (2.33) (3.11 ) Expense Ratio 0.175 0.19 0.178 0.163 0.178 (% per annum) (0.04) (0.028) (0.023) (0.02) (0.034) Flow 0.404 6.8 10.11 26.05 7.76 (% per annum) (17.34) (69.22) (102.03) (228.2) (107.86) TLoad 2.08 2.28 2.30 2.57 2.25 1%1 (0.96) (051) (0.37) (0.59) (0.75) 111 Table 6.2 gives the summary statistics of the sample. The means and standard deviations (in parenthesis) of the variables of interest are mentioned. The spread of the asset size is large and obvious as reflected in the standard deviations. The average family size is larger, for larger funds. The same trend is noticed with respect to age as well. Larger funds have been in existence for a longer period of time. Their average age in years is 11.09 as compared to the average of 8.64 for all funds. The expense ratio is the annual expense ratio divided by 12 to give a monthly average. It shows a declining trend with the smallest asset class bucking the trend and actually having an expense ratio higher only compared to the 4 billion class. Average flow and total load are found to increase with fund size. The largest fund size has the highest average load. This could actually be because larger funds face market impact problems while dealing with large scale redemptions. Table 6.3 - Time series averages of the cross sectional correlations between the various fund characteristics log log Age Total Flow Asset Family load Asset Log Asset 0.64 0.32 0.28 0.07 Log Family 1 0.07 0.10 0.02 Asset Age 1 0.28 0.01 Total Load 1 0.003 Flow 1 Table 6.3 on examining the cross sectional correlations between the fund characteristics, we find that log asset has a reasonably strong positive correlation 112 with log family size. age and total load. It therefore becomes obvious that these fund characteristics have to be controlled for before estimating cross sectional relationship between asset size and retums. The sample has been chosen keeping in mind the survivorship bias. Funds which were terminated or merged mid-way through the sample were also included which gave the data an unbalanced panel structure but addressed the survivorship bias problem. Otherwise the correlations between age and fund size could have been stronger. Table 6.4 - Means and standard deviations for the monthly fund retums Mutual fund asset Bize [in rUe"! billign] <1 1-4 4-7 7> Net Fund 3.03% 3.57% 3.86% 4.25% Returns (7.38%) (6.88%) (6.84%) (6.17%) Gross 3.24% 3.76% 4.04% 4.41% Fund Returns (7.42%) (6.88%) (6.84%) (6.17%) Table 6.4 reveals that mean returns show an increasing trend with asset size. This is so for both net and gross fund returns. On the face of it. without controlling for other fund characteristics there seems to be no decreasing returns to scale. But simple fund returns alone may not reflect the relation between size and returns. It is important to have an idea of the ability of the fund to outperform the market on a risk adjusted basis or absolute basis controlling for other fund characteristics including style. The ability to do so should get tougher with size. 113 6.2.2 Empirical framework and model Fund returns were computed by using the following formula: (6.1) NA V. - end ofthe month net asset value (NAV) NAV._. - beginning of the month Similarly monthly returns are computed for the benchmark index - CNX500. Funds are divided into the asset classes 0-1 billion. 1-4 billion, 4-7 billion and 7 billion as this is the division created by the regulator SEBI, for purposes of benchmarking expenses. Individual regressions are run for each of the four asset categories using the following specifications to derive the Beta or the market benchmark correlation (systemic risk). A single factor CAPM is used to derive the beta 3S
(6.2) R n;p is the return of the fund i in asset class n. R,the Risk free return (or 91day t bill return) sourced from the RBI Pn is the systemic risk for the specifIC asset category considered. n represents the four asset classes (up to Rs. 1 billion, 1-3 billion, 3-7 billion and above seven 7 billion) Next Gross Beta adjusted returns and Gross market adjusted returns are computed as follows: JS We continue with the single factor model as we later segregate funds on the basis of the fund portfolio or style. This should allow us to differentiate performance of funds which could be different depending on whether they are value stock Igrowth focused and small capllarge capitalized stock orientation. 114 Beta adjusted returns = Returns. _ j3" (CNX500 returns). (6.3) This helps us to arrive at the risk adjusted excess returns of the fund as compared to the index, CNX500. Market adjusted return is essentially the difference between the fund return and the index return. This aims to check for the influence of size on the ability of the fund to outperfonn the market without any adjustment for risk. The ability to outperfonn the market should be much more diffICult as size increases. The influence of various factors including size on these adjusted returns can be represented in the following functional fonn. adjrelurn (R a) = R(assel ,jamilyasse I, flow, age ,lolaUoad ,fundstyle ) (6.4) On deriving these two adjusted returns the regression specified by equation 6.2.2.4 is used for both these returns separately. Retums.=c+ 0, In (asset . ,) + 02 In (Family Asset..,) + 03 Flow . , + 14 age .. , + 05 (6.5) i'" 1, .. , N funds t=1, .. , rn months Monthly Flow is computed as follows: '1 _ [A A ] I Re I ~ s s e Is,.! j F,ow I - ssets I - ssets I-I - t 100 (6.6) Where t-1 is the beginning of the month and t is the end of the month. The In1pact of fund size and family size on fund returns, both gross (inclusive of expenses) and net is sought to be derived by controlling for fund specific qualitative 115 factors such as age, flow, total load and style. Style here is represented by two dummies 0, and O 2 . 1 if growth stocks oriented 0,= o if blend or value stock oriented 1 if large cap O 2 = o if mid or small cap oriented 6.3 RESULTS Table 6.5 - Mutual fund Beta's and constants across different asset classes with CNX500 as the benchmark index. A.set class [in ruoees billion] <1 14 4-7 7> Constant 0.74 1.06 0.85 0.82 (0.00) (0.00) (0.00) (0.00) P 0.73 0.71 0.84 0.77 (0.00) (0.00) (0.00) (0.00) R' 0.54 0.54 0.70 0.63 Figures in parenthesis represent the p-values Table 6.5 shows the results of the regression using equation 6.3.2. The Beta's are persistently below one for all funds reflecting a risk level lower than that of the index. Most funds keep very little cash and stay invested in stocks hence the lower beta cannot be attributed due to investment in short term debt. The constant is the highest for the middle two asset classes reflecting superior stock Picking capacity. Hence the average betas being less than one cannot be explained by cash or debt holding. Using these beta's beta adjusted returns are derived (Table 6.7) and the regression (Table 6.8) is carried out. A random effects model is used for beta 116 adjusted retu rns and a fIXed effects model is used for the marKet adjusted returns, the choice be' In9 based on the Hausman specifICation results. The Hausman specifi . lCalion results are provided in table 6.6 below. Table 6.6 - Correlated Random Effects - Hausman tast Test summary Cross sectlon random effects for beta adjusted returns Cross sectlon random effects for market adjusted returns Chi-sq statistic 7.71 12.14 Chl-sq d.f Prob 7 0.34 7 0.09 Table 6.7 - Regression of Beta and market adjusted fund performance on fund size and other characteristics over one period lagged effects Beta adjusted Marl<et adjusted returns returns Intercept 0.68 (0.40) 2.95 (0.02) Log Fund Size 0.26 (0.01) -0.18 (0.25) Lag Family I Fund -0.11 (0.26) -0.22 (0.25) Size Age -0.06 (0.15) 0.01 (0.90) Total load -0.24 (0.13) -0.02 (0.91) Flow 0.00 (0.37) 0.00 (0.00) 0, (Growth Stock) -0.15 (0.59) -0.02 (0.91) O 2 (Large cap) -0.42 (0.09) -0.48 (0.02) Figures In parentheSIS represent the p-values In Table III (3) for beta adjusted returns fund size has a p o s ~ i v e impact on retums and large cap funds have underperfonned medium and small cap funds in tennS of risk adjusted excess perfonnance. None of the other variable seems to have an impact on beta adjusted returns. For marKet adjusted returns none of the 117 variables are significant except for the dummy D2 representing large cap funds. Large cap funds have again underperformed in terms of excess absolute returns over and above the index.3637 This indicates that larger funds actually deliver higher risk weighted excess returns. This, however, could be because of the sample giving the average of bull and bear phase performance in a sample time with larger nurnber of bull periods. In order to further explore the performance of large funds over bearish and bullish phases, of the market, the sample months are s p l ~ into two periods where return of CNXSOO is either greater or less than zero. CNXSOO returns above zero are considered as bullish markets and returns below zero are taken as bear markets. Regression (6.3.4) is repeated again for Beta and market adjusted returns for bullish and bearish market periods separately. The tables 6.3.4 and 6.3.S give the results. This decomposition shows that larger funds show negative correlation between asset size and returns during bearish phases and vice versa during bullish or positive returns phases of the CNXSOO.This resuH proves that larger funds fund it difficuHy to pull out of or face larger impact costs in falling markets as their larger transaction cause negative price impact. 36 However. it must be noted that the fixed effects approach is subject to a regression-to-the-mean bias which can result in showing an inverse relation between performance and size. The random effects approach when used gave a positive significant relation for market adjusted returns with size and a significant negative relation with load and age. The rest of the variables were of the same sign and did not differ substantially in significance or impact. 118 Table 6.B: Regression of market adjusted retums (In baar & bull phases of the market) on lagged fund size and other fund characteristics. Bear Phase Bull Ph.s. (671 fund months) (2251 fund months) Intercept 3.99 (0.01) 0.86 (0.38) Log asset u-. -0.42 (0.03) 0.29 (0.02) Log Family ,,_. 0.33 (0.09) -0.2 (0.13) Asset ','_' Age ",., -0.15 (0.09) -0,02 (0.68) Tolal load ,,_. -0,74 (0.01) -0.23 (0.18) Flowu_. 0.00 (0.00) 0.00 (0.00) D. -0.09 (0.83) 0.07 (0.8) D, 0.58 (0.19) -0.69 (0.01 ) Figures in parenthesis represent the ~ v a l U 9 s For beta adjusted returns based on the Hausman test a random effects model and a fixed effects model are applied for bullish and bearish markets respectively. The test results are not present for brevity, as ~ follows the same pattem as for earlier regressions for the complete period. Table 6.9 - Regression of bala adjusted returns (in baar & bull phases of the market) I dfu d' d th f d h t' Ii on agge n size an 0 er un c arac ens es. Bear Phase Bull Phase (671 fund months) (2251 fund months) Intercept 9.2 (0.00) -1.56 (0.32) Log asset., . -0.88 (0.01 ) 0.38 (0.01 ) Log Family,,_. 0.87 (0.04) -0.17 (0.22) Asse!;.._. Age"._. -0.57 (0.01) -0.02 (0.70) T olal load, .. , -1.17 (0.00) -0.09 (0.62) Flowu_. 0.00 (0.96) 0.00 (0.00) D. 0.06 (0.91) -0.31 (0.22) D, -0.52 (0.41) -0.67 (0.03) Figures In parenthesIs represflnt the ,rvalues 119 When the test is repeated for beta adjusted returns size. age and total load show negative impact on retums. Older funds and funds with higher total loads do worse during such phases. But interestingly funds in larger families seem to do better. This could perhaps be due to larger fund houses spending more on research and achieving superior stock selectivity. The variable of interest. fund size behaves in a manner predicted and larger funds do worse in bearish markets. In bullish markets larger funds do better as indicated by the positive and statistically Significant coeffICient. None of the other variables are significant excepting for the dummy D3 which shows that large cap stocks oriented funds do not do as well as mid and small cap stock oriented ones 38 . This shows that better performance in the bullish markets seems to help overcome underperformance in bearish markets for the larger funds. 6.4. Chapter Summary Most Indian mutual funds cannot be compared in terms of size to their counterparts in the US or other developed markets. Yet there are seeds of diseconomies that seem to be arising akin to that in the US funds. Though the average performance of funds both beta adjusted and market adjusted do not reveal diseconomies. Larger funds seem to under perform to a larger extent during bearish phases of the market after performances are adjusted for fund characteristics. Given more positive or bullish phase months it appears that larger funds perform better only on an average. Further it appears that larger families generate positive influences on adjusted fund returns during weaker market 38 Again it has to be pointed out that the fixed effects regression-to-the-mean-bias is a factor as in a random effects situation the coefficient of log asset is significant at five per cent levels but is around- 0.48, a much lower negative relation. The positive impact of family size is also lower at 0.39. 120 periods. This aspect has not been dealt with in greater detail. But this could be due to the ability of large fund houses to maintain in-house research teams which can add value to retums generation. Growth of individual fund size is, therefore, a cause for concern and justifies the recent moves on the part of certain funds to curtail size. But should SEBI step in to mandatorily reduce size? We again believe that the tendency of funds to voluntarily curtail size shows that concern over generating competitive returns has caused funds themselves to take suitable action. The role of SEBI could come in when a fund desires to limit the size of its assets when it has grown over a period of time due to inflow of funds and increase in stock prices. 121 CHAPTER - VII SUlULARY OF FINDINGS AND POLICY SUGGESTIONS CHAPTER - VII SUMMARY OF FINDINGS AND POLICY SUGGESTIONS 7.1. Summary Mutual funds in India are emerging as an important instrument of household savings. II has the potential to rival bank deposits for attracting household savings. Given its present and potential importance a study of the performance of mutual funds and the role of regulations in preventing investor-manager conflict of interests becomes crucial. The few studies which have dealt with fund performance, in terms of returns net of expenses, have not considered the role of regulations in influencing fund behaviour. We started the study by examining some popular measures used to convey risk weighted return. The Sharpe ratio and Treynor index are among the widely reported in various mutual fund annual reports and mutual fund specifIC journals such as Mutual Fund Insight. Since these measures do not reflect the tnuer risk of returns below mean or minimum acceptable returns we felt n would be important to arrive at a better measure and see if it affected the ranking of funds. We then sought to examine if regulations have influenced performance. Further since regulations also control operating expenses and management fee of funds, the study focuses on the effectiveness and need for such controls. Lastly it studies the increasing size if individual funds in terms of assets managed and the possibilities for new areas of potential regulation. In order to pursue this study the discussion through the thesis has been divided into seven chapters. 122 Chapter One is an introductory chapter which underlines the evolution of mutual funds in India. It also discusses the regulatory structure for this industry and the importance of mutual funds. Chapter two focuses on review of literature pertaining to the various objectives of the study. There is a substantial amount of literature dealing with the performance of mutual funds more specifically in the United States of America. The literature has evolved from the focus on 'pure performance' to various factors that influence the same. The need for considering alternate approaches to riSk, importance of style and size in influencing performance and the relationship between asset size and operating costs are discussed. The same is considered for India and the gap in research is obvious. There exists practically no work which tries to establish the net benefit of regulations and its effectiveness of the same in controlling costs. Further, the issue of fund size and diseconomies in retums is also very important but, not dealt with adequately. The review of literature therefore points the evolving methodologies and issues and the lack of such studies in India. After identifying these gaps the study has framed its objectives and analyzed the same. The resuHs are summarized below: 7,1.1. Relative Fund Perfonnance Using Alternate Risk Weighted Returns: Mean-Variance and Lower Partial Movement In chapter three an attempt has been made to study fund performance using two different approaches to risk and the consequences of the same for fund performance ranking. The first measure of risk which is widely used is the standard deviation of returns. This approach gives us the popular risk weighted measures, Sharpe ratio and Treynor index. Since these measures are based on standard deviation measure of risk they are susceptible to errors due to non-symmetric 123 distribution of retums. Hence we compute the Sortino ratio which uses the returns below a certain minimal acceptable retum as a truer measure of risk. The results are briefly summarized as follows. The chapter first tries to rank funds according to their performance. It then uses the Sharpe ratio and Treynor index which are all based on the M-Vapproach to portfolio studies. We find that 58 percent of the funds considered are able to out perform the broad based index. the CNX500. That leaves us with almost half the number of funds under performing the index. This is not a good sign as 48 percent of the funds are not shOWing the benefrts that merit a management fee and other expenses. Next the chapter goes on to compare ranking of funds based on the Sortino ratio which uses semi-variance as a measure of risk and the 91 days treasury bill retums as the minimal acceptable return. We compare rankings based on the three , measures using the Spearman rank correlation coefficient. We find no signifICant difference between them and hence conclude that the present widely used ratios, Sharpe ratio and Treynor index give proper information. 7.1.2. Regulatory Constraints and Fund Perfonnance- Comparing Alternate Regulatory Regimes Chapter four makes an effort to check if regulatory constraints, and its accompanying costs, have undermined the benefits. Since it involves the impact of regulations on net performance (net of all direct and indirect costs) we needed to compare fund performance to a situation with weaker regulations. The c0- existence of a weaker set of regulations, the UTI Act of 1963 which govemed UTI funds, allowed as a comparative regulatory benchmark. Funds regulated by SEBI 124 were compared to those of UTI to derive an average differential in terms of return and portfolio selectivity skills. Our assumption was that superior regulations would lead to greater due diligence and portfolio selectivity and hence deliver better net returns. Our findings are as follows. By and large SEBI regulated funds have outperformed UTI funds in terms of risk weighted returns for the period June 1996 to September 2000. Out of the three UTI funds, two under perform the index and so also two private funds. This holds true for performance measured in terms of the Sharpe and Modified Information Ratio. We then proceeded to check for the stock selectivity skills of UTI fund with the SEBI regulated fund. One would expect that the UTI, with its vast experience and knowledge of the Indian markets would exhibit superior stock selectivity. We hypothesize that the benefits of this could be undermined by moral hazards in portfolio selection. This hypothesis is based on the poor corporate govemance structure and weak disclosure norms for certain UTI funds. Hence, we feel they would display poorer Jensen alpha's compared to the SEBI regulated funds. Our study goes on to find that this poor selectivity is indeed the case for UTI funds. Their average 'alpha' is negative and not statistical signifICant. This compared to the private SEBI regulated funds which have a signiflCBnt positive alpha of 0.45. Since UTI has no management fee we add average industry management fees to the private funds returns and get an even higher and statistically signifICant alpha of 0.53. This difference we admit could be due to fund manager skills as well as regulatory effects. But given the vast experience of UTI fun rnanagers or administration one cannot expect them to be inferior the new comer private funds. 125 We believe ~ is the moral hazards of poor information disclosure along with the other weak regulations which induced poor portfolio and hence poor returns. SEBI's regulations have been able to induce better portfolio selectivity through frequent disclosures and better corporate governance structures. Along with the cap on expenses and management fees SEBI regulations have ensured better net performance. Based on our findings we feel encouraged to recommend certain improvements in regulations to provide better incentives to perform. Firstly, the number of independent trustee members can be increased. It was increased from half the number to two-thirds in 1997. We believe this can go the whole way and all can be independent members. Further it is felt that the present management fee is 'weak-form performance fee'. The management fee is linked to asset size. Hence a fund manager would earn more if the open end equity fund attracts more investments. This attraction, n is presumed, would be via better relative performance. But ~ has not been clearly established if performance is the most important factor in attracting investments. In fact marketing expenses or brokerage paid could prove more important in influencing potential investors. Hence we would recommend stronger-form of performance fee. This could initially be a non-symmetric fee with a base level fixed fee and a bonus fee for superior performance. The base fee can be lowered from the present allowed fixed rates. 126 7.1.3. Cost- Management Fee Regulation and Economies of Scale In chapter five we examine SEBI's cost and management fee regulations and the economies of scale in India mutual funds. The study tried to assess the effectiveness of SEBI's expense ratio cap in controlling fund expenses. It also examined fund behaviour in terms of actual management fee charged. The sample covered a varying number of equity and balanced funds under the 'actively managed' fund category and debt funds, both over the period April 2000 to March 2007. The main findings of this can be summarized as follows: In the case of equity funds and balanced (equity oriented) funds we find that mutual funds tend to charge lower than ceiling management fees and expenses at lower asset levels but as the grow in size this tendency disappears. As size increases average of other expenses (which reflects economies of scale) tends to drop but starts increasing again for the highest asset class of RS.1 billion and more. Hence we see diseconomies creeping in for mutual funds. In terms of actual costs charged these exists no economies of scale with size that are greater than that assumed by regulations. However on closer examination it is found that larger funds tend to charge the maximum expenses allowed not due to an increase in other expenses but also due to the unclear category of marketing expenses. In the case of debt funds we find economies if scale which are greater than what is granted by the regulatory ceiling. There is a perpetual drop in the category of 'other expenses'. This however appears to be reaching a limit at an average of six basis point percent of the assets. Further debt funds charge lower managerial fees and expense ratios across all asset classes compared to the regulatory ceilings. This appears to be due to the need for generating better returns in a 127 macro economic environment that saw dropping interest rates and simultaneous high returns on the equity mar1<ets. Hence competijion from equity funds seems to have compelled them to keep costs low through sacrifICing managerial fees. The other important finding is thai family size seems to have no impact in reducing costs any particular fund in the equity and balanCed fund category. Sharing of mar1<eting and selling infrastructure wages of personnel involved and also sharing the benefits and costs of research should have generated some gains. But there are no economies generated through family size. On the contrary we see a positive impact of family asset size for debt funds. This is a puzzling phenomenon which needs to be further studied as normally common costs of fund Operation have to be shared proportionally as per SEBI regulations. We also find that the total of entry and exit load and Ihe lum over of portfolio are no significant explanations for cost for both categories of funds. These results lead us to the conclusion that policy caps have restrained fund expenses at the larger level and that policy needs to examine the need to charge 'mar1<eting expenses'. This accounts for a significant proportion of the expense ratio, close to 25 percent for equity and 35 percent for debt funds. This seems to have no clarity and is essentially involves spending more to promote a fund at the expense of existing share holders. It needs to be clear as to whether this is in any way culling on other expenses which investors might have to bear. Further we notice a good trend in that small equity and balanced funds tend to compete by keeping expenses and fees lower than allowed. This is an indication 128 that the market in moving towards a more mature phase with competition appearing robust. 7.1.4. Fund Size and Fund Returns The sixth chapter addresses the new issue of growing size of mutual funds. The study focuses on open end equity diversified funds over the period August 2002 to October 2007. The summary of the result are as follows. We find that there is no impact of size on both beta and market adjusted returns. But on decomposing the period into bearish and bullish phases of the market we find that larger funds tend to under perfonn, in tenns of market adjusted returns, during the bear markets. The better perfonnance during the bullish phases of the market seems to have helped deliver an overall better return. Once again the fact that mutual funds are voluntarily seeking to curtail size indicates the concern over size affecting relative perfonnance. This we feel, is again a consequence of competitive pressures. So even large funds, while continuing to charge the maximum possible expenses and fees allowed, do seem to be concerned over their competitive perfonnance. But this is not a tendency among all large funds and it appears to be the beginning of a trend. We feel our study has two possible limitations. We discuss them below 7.2. Limitations ofthe Study The study of regulatory impact on due diligence portfolio selection and returns depends on the single factor model alpha. This could be questioned on two grounds. The alpha difference could be due to fund style or due to superior fund managers. We feel this is not a serious limitation as UTI fund 129 managers should on the basis of their experience be more skilled than the new fund managers. Further the diversifICation of fund styles was a phenomena more witnessed after 2000. Our study of economies of scale is hampered by the lack of hedonic variable such as number of investors in afund and the division of a fund asset into retail and institutional investors. The cost of fund management could be higher when the average asset per investor is lower due to diseconomies. However such data is not available to the desirable extent 7.3. Further Scope for Study We feel that with increasing frequency of data the following studies would be important The relevance of fund perionnance for attracting new investors. The relevance of expenses for the ordinary retail investor in choosing a fund. The role of marketing expenses in attracting fresh investor. 7.4. Policy Suggestions and Concluding Observations. In the light of the above results, this study makes the following comments on and suggestions for regulatory refonn. Despite the widespread studies on the limitations of the Mean-Variance approach to risk we find that measure based on this serve the purpose well. 130 There is no need for emphasizing on the downside risk measure, Sortino ratio, as there is no significant change in rankings of equity diversified funds in term of performance. Use of many terminologies could lead to the possibility of confusing the average retail investor. Regulations have been able to induce due diligence in portfolio selection. This we believe has been the most important factor in explaining the superior average returns of equity diversified funds. We feel encouraged to , therefore, recommend further strengthening of regulations which have limited potential for imposing costs. The number of independent trustees can be increased to 100 percent from the present two thirds. The POrtfolio disclosure frequency can be increased to a monthly basis as we feel there are substantial gains to be achieved through additional transparency. SEBI could now consider the possibility of a stronger -form performance fee. The present asset size based fee assumes that asset size growth depends on relative performance. This is not a certainty. Hence we Would recommend a non-symmetric performance fee with a lower base fee, compared to the present permitted fees, and a bonus for outperforming the relevant benchmark index. SEBI needs to examine the marketing cost aspect of expenses. At present there is no clarity on this and it seems to promote asset growth at the cost of existing investors. This could further lower expenses across aU asset sizes. SEBI also needs to look at the possibility of funds aiming to curtail their asset size. It needs to see that such a process does not harm the interest of 131 the investor. No caps on asset size are required as funds are voluntarily restricting size due to com petitive pressures. 132 REFERENCES REFERENCES Akerlof G . Ma k ' eorge A., (1970). "The Market for 'Lemons': Quality UncertaInty and the r et Mechanism," The Quarterly Journal of Economics, Vol. 84(3), pp 488-500. AMFI httP'11 Monthly Reports (various) available at . WWW.amfiindia.com/showhtml.asp?page=amfimonthly Baltagi Bad' H nd Wit ' 1., (2001), Econometric Analysis of Panel Data, 2 edition, John ey, New York ;::a. V S., (1975), "Optimal rules for ordering uncertain prospects", Journal of nc'al Economics 2, pp 95-121 ::;i V S, E.B.Lindenberg., (1977), "Capital Market equilibrium in a mean-lower 1a movement frameworks", Journal offinancial Economics 5, pp 189-200 :::mke l ,William J., John C. Panzar, & Robert D. Willig., (1982). Contestable r ts and the Tbeory of Industry Structure, Hardcourt College
William J. Stephen M.Goldfeld, Lilli A. Gordon and Michael F.Koehn.,
D_ 9), Tbe Economics of Mutual Fund Markets: Competition versus ....-&ulation, Kluwer, London Dogie, J., (1998), "The Implications of Style Analysis of Mutual Fund Performance Evaluation," The Journal of Portfolio Management. Summer, 34-42. fkown, S. and W.Goetzmann, (1997), "Mutual Fund Styles," Journal of Financial 43, 373-399 e,.pon, Noel, Gavan J. Fitzsimons and Russ Alan Prince (1996)., "An Individual [.evel Analysis of the Mutual Fund Purchase Decision," Journal of Financial fe!rv;ces Research. 10, pp 59-82. baantberlain, G., (1983), "A characterization of the distributions that imply mean- utility functions", Journal of Economic Theory, 29, pp 185-20 John M.R., Edelen, Roger M. and Kadlec, Gregory B., (1999). !Transaction-cost Expenditures and the Relative Performance of Mutual Funds" IIFIC Working Paper No. 00-02. Available at SSRN: , p:/lssrn.comlabstract=218108 t' Eric C., and Wilbur G. Lewellen., (1985), "An Arbitrage Pricing Approach l .. E .. valuating Mutual Fund Performance", Journal of Financial Research, 8(\), pp ,.30 133 Chen, Nai-Fu, Richard R. Roll, and Stephen A. Ross., (1986), "Economic Forces and the Stock Market", Journal of Business, 59(3), pp 383-404 Chen,J., H.Hong, M.Huang, J.D.Kubik., (2004), "Does Fund Size Erode Mutual Fund Performance"? The Role of Liquidity and Organisation", The American Economic Review, 94(5), pp 1276-1302 Chevalier, J.A. and G.D. Ellison., (1997), "Risk Taking by Mutual Funds as a Response to Incentives," Journal of Political Economy 105, pp 1167-1200 Christofferson, S.E.K., (2001), "Why Do Money Fund Managers Voluntarily Waive their Fees?" Journal of Finance 56, pp 1117-1140 Connor, Gregory, and Robert A. Korajczyk., (1986). "Performance Measurement with the Arbitrage Pricing Theory: A New Framework for Analysis", Journal of Financial Economics, 15(3), pp 373-394 ----. (1988), Risk and Return in an Equilibrium APT: Application of a New Test Methodology, Journal of Financial Economics, 21(2), pp 255-290 ---. (1993), A Test for the Number of Factors in an Approximate Factor Model, Journal of Finance, 48(4), pp 1263-1291 DiBartolomeo, D. and E. Witkowski, 1997, Evidence Based on Style Analysis," (September/October), 32-43. "Mutual Fund Misclassification: Financial Analysts Journal, Dybvig, Philip, and Stephen A. Ross, 1985, "Yes, the APT is Testable", Journal of Finance, 40(4), pp 1173-1188 Elton, E.J, MartinJ Gruber.,(l991), Modern Portfolio Tbeory and Investment Analysis, 4th edition, New York :John Wiley and Sons, Inc., Elton, E. J., Martin. J. Gruber, and J. A. Busse, 2004, "Are investors rational? Choices among Index Funds", The Journal of Finance 59, pp 261-288 Fant, L. and E. O'Neal, 1999, "Do You Need More Than One Manager for Given Equity Style? Evidence from Mutual Funds," The Journal of Portfolio Management, Summer, 68-75. Fishburn, P.C., (1977), "A standard measure of risk and risk-value models", The American Economic Review, Yol.07 (2), pp. 116-126. Frank,lF., and J.C.Francis., (1979), "Mutual Fund Systematic Risk for Bull and Bear markets: An Empirical Examination", The Journal of Finance, 34(5),pp 1243- 1250 134 and Looi,A.,(2003), "An Examination of the Market Impact Costs of htt Australian Equity Managers ". SIRCA available at p. WWw.melbournecentre.com.au!project2_2.pdf and K.M.Martin,(2005), "Size and Investment Performance: a search Note", Abacus 4I{I), pp 55-65 Gallo J d L . . Eq .' . an . Lockwood, (I997), "Benefits of Proper Style ClasslficahOn of Ulty Portfolio Managers," The Journal of Portfolio Management, 23,47-55 :;;;-- {I 999), "Fund Management Changes and Equity Style Shifts," The "'mal of Portfolio Management, 55,44-52. and. S.Titm.an., (1989), "Mutual Fund Performance: An Analysis of rterly Portfoho Holdmgs", The Journal of Business, 62 , pp 393-416 Perf! "A Study of Monthly Fund. Returns a.nd A onnance EvaluatIOn Techmques", Journal of Financial and Quanlllative /1Qlysis 9(3) pp.419-444 M., S. Titman and R. Wermers, (1995), "Momentum Investment A Portfolio Performance, and Herding: A Study of Mutual Fund Behavior," mencan Economic Review, 85, 1088-1 \05 HUbennan, Gur., (I982), "A Simple Approach to Arbitrage Pricing Theory", JOurnal of Economic Theory, 28, pp 183-191. C.Jiang, M.Hu, and W.Lee., (1999), "Mutual Fund Performance: Does Size Matter?", Financial Analysts Journal, 55, pp 74-87. Ingersoll, Jonathan., (I984), "Some Results in the Theory of Arbitrage Pricing", JOurnal of Finance, 39, pp \021-1039. Jarrow, Robert, and Andrew Rudd., {I 983), "A Comparison of the APT and CAPM: A Note", Journal of Banking and Finance, 7(3), pp 295-303 . .Jia, J., J.S.Dyer (I996) "On Mutual Fund Investment Styles", Management Science 42 (12), pp 1691-1705 ICon, Stanley J. (I983), "The Market- Timing Performance of Mutual Fund Managers". Journal of Business Vol. 56(3), pp 323-347 Latzko, David A. (1998), "Economies of Scale in Mutual Fund Administration", . Journal of Financial Research, 22(3), pp 331-339 Lee. Cheng-Few and Rahman Shafiqur., (I990), "Market Timing, Selectivity and Mutual Fund performance: An Empirical Investigation", Journal of Business Vol. 63 (2). pp. 261-278. 135 Madhu S Panigrahi., (1996), "Mutual Funds: Growth, Performance and Prospects", Economic and Political Weekly, Vol 31(12) pp.765-775 Malegam Committee Report., (2001), available at www.capitalmarket.coml mutuaVutireport.htrn Malik, Payal., (2004), "Compliance cost of Financial Regulation," Economic and Political Weekly, VoI.39(28), pp 3083-3085 Markowitz, H.M,,(1952), "Portfolio selection", The Journal of Finance, Vol.7 (I), pp.77-91 Murray, John D. and Robert W. White.,(1983), "Economies of Scale and Scope in Multiproduct Financial Institutions: A Study of British Columbia Credit Unions", The Journal of Finance, 38(3), pp 887-902 Mc Cann, Craig., (1999) "Churning", Journal of Legal Economics Vol. 9(49) also available at http://www.slcg.com/pdf/ workingpapers! McCann_on_Churning.pdf Murty G.R.K., (2003), "Indian Derivatives Market, Issues at Stake", Treasury Management Najand, M. and L. Prather, (1999), "The Risk Level Discriminatory Power of Mutual Fund Investment Objectives: Additional Evidence," Journal of Financial Markets,2, pp 307-328 Newman, Peter (Ed), (1998), The New Pal grave Dictionary of Economics and the Law Vol I, UK Mac Millan Nyholm Ken.,(2002), "Estimating The Probability of Informed Trading", Financial Research, Vol 25 (4) pp. 485-505. Peltzman, S., (1976), "Toward a more general theory of regulation", Journal of Law and Economics, 19, pp 807-820. Perold, A., Salomon, R., (1991), "The Right Amount of Assets under Management", Financial Analysts Journal, Vol. 47(3), pp 31-39 Posner, Richard A., (1969), "Natural monopoly and its regulation", Stanford Law Review 21, pp 548-M3. =-.,...",......,, ___ (,1971) "Taxation by Regulation," Bell Journal of Economics, The RAND Corporation, vol. 2(1), pp 22-50 .,,-___ ..,..(1974). "Theories of Economic Regulation," Bell Journal of Economics, The RAND Corporation, vol. 5(2), pp 335-358 136 -;;-;-:-:---;--=-_,(1975). "The social costs of monopoly and regulation", Journal 0/ Political Economy 83(August): 807-827 Rea, John D. Brian K.Reid, and Kimberler W.Miller., (1999), "Operating Expense Ratios, Assets and Economies of Scale in Equity Mutual Funds", Perspective 5(5), Investment Company Institute. Reid, Brian., (2005), "Competition in the Mutual Fund Commentary", Investment Company Institute www.icLorglpdf/rc_competition.pdf Business, Research available at Roll, Richard R., and Stephen A. Ross.,(1980), "An Empirical Investigation of the Arbitrage Pricing Theory", Journalo/Finance, 35, pp 1073-1104 Ross, Stephen A.,(1976), "The Arbitrage Theory of Capital Asset Pricing", Journal of Economic Theory, 13, pp 341-360 := __ ----;--;-::-(1978), "Mutual Fund separating distributions", Journal of Economic Theory, voU7, pp.254-286 Roy, Bijan and Deb, Saikat Sovan., (2003), "The Conditional Performance of Indian Mutual Funds: An Empirical Study" Available at SSRN: http://ssrn.com/abstract=593723 Sadhak H. (1997), "Mutual Funds in India. Marketing Strategies and Investment Practices", Response Boolcs (New Delhi) Sahadevan, K G and Thiripalraju (1997), "Mutual Funds: Data Interpretation and Analysis", Prentice Hall (India) SEBI Annual Report, (various) available at www.sebi.gov.inlSectIndex.jsp?sub_sec_id=90 SEC., (2000) Report on Mutual funds fees and expenses, available at www.sec.gov/newslstudieslfeestudy.htm Shanken, Jay, 1982, The Arbitrage Pricing Theory? It is Testable", Journal of Finance,37, ppI129-1140. ---;:-=-:-__ -:-=-.(1992), "The Current State of the Arbitrage Pricing Theory". Journal of Finance, 47(4), pp 1569-1574 Sharpe, William F., (1966). "Mutual Fund Performance" Journal of Business, January 1966, pp. 119-138 ::-:-___ --::-:--(1992), "Asset Allocation: Management Style and Performance Management," Journal of Portfolio Management, 18, 7-19. 137 -,--:-___ "The Sharpe Ratio", The Journal of Portfolio Management Vol21 (I) Pg.49-58 Siggelkow, Nicolaj .,( 1999), "Expense Shifting: An Costs in the Mutual Fund Industry", http://ssm.com/abstract=147274 Empirical Study of Agency Available at SSRN: Starks, Laura. T., (1987), "Performance Incentive Fees: An Agency Theoretic Approach", Journal of Financial and Quantitative Analysis Vol.22 (I) pp 17-32 Stigler, G.J.,(1971). "The theory of economic regulation", Bell Journal of Economics and Management Science 2, pp 3-21 Tobin, J. (1958), "Liquidity preference as behaviour towards risk", Review of Economic Studies. Vo1.25, pp. 65-86 Trzcinka, Charles A., (1986), "On the Number of Factors in the Arbitrage Pricing Model", Journal of Finance, 41(2), pp 347-368 -=--=--:-...,....,._( 1995), "Equity Style Classifications: Comment," The Journal of Portfolio Management. spring, 44-46 Tullock, G. (1967). "The welfare costs of tariffs, monopolies, and theft", Western Economic JournaI5(June), pp 224-232 -,-,-_---,_(1975), "The transitional gains trap". Bell Journal of Economics 6 (Autumn), pp 671-678. \38