Malaysia Mutual Fund Performance 9781846632594, 9781846632587

Citation preview

mf cover (i).qxd

09/01/2007

11:36

Page 1

ISSN 0307-4358

Volume 33 Number 2 2007

Managerial Finance Malaysia mutual fund performance Guest Editor: Soo-Wah Low

www.emeraldinsight.com

Managerial Finance

ISSN 0307-4358 Volume 33 Number 2 2007

Malaysia mutual fund performance Guest Editor Soo-Wah Low

Access this journal online _________________________

87

Editorial advisory board___________________________

88

The price linkages between Malaysian unit trust funds and the stock market: short run and long run interrelationships Soo-Wah Low and Noor Azlan Ghazali _____________________________

89

Malaysian unit trust aggregate performance Fauziah Md. Taib and Mansor Isa ________________________________

102

An integrated framework for style analysis: how is it useful to Malaysian equity trust investors? Wee-Yeap Lau_________________________________________________

122

Investigation of performance of Malaysian Islamic unit trust funds: comparison with conventional unit trust funds Fikriyah Abdullah, Taufiq Hassan and Shamsher Mohamad ___________

Access this journal electronically The current and past volumes of this journal are available at:

www.emeraldinsight.com/0307-4358.htm You can also search more than 150 additional Emerald journals in Emerald Management Xtra (www.emeraldinsight.com) See page following contents for full details of what your access includes.

142

CONTENTS

CONTENTS continued

Malaysian unit trust funds’ performance during up and down market conditions: a comparison of market benchmark Soo-Wah Low _________________________________________________

154

Call for papers ___________________________________

167

www.emeraldinsight.com/mf.htm As a subscriber to this journal, you can benefit from instant, electronic access to this title via Emerald Management Xtra. Your access includes a variety of features that increase the value of your journal subscription.

Structured abstracts Emerald structured abstracts provide consistent, clear and informative summaries of the content of the articles, allowing faster evaluation of papers.

How to access this journal electronically

Additional complimentary services available

To benefit from electronic access to this journal, please contact [email protected] A set of login details will then be provided to you. Should you wish to access via IP, please provide these details in your e-mail. Once registration is completed, your institution will have instant access to all articles through the journal’s Table of Contents page at www.emeraldinsight.com/0307-4358.htm More information about the journal is also available at www.emeraldinsight.com/ mf.htm

Your access includes a variety of features that add to the functionality and value of your journal subscription:

Our liberal institution-wide licence allows everyone within your institution to access your journal electronically, making your subscription more cost-effective. Our web site has been designed to provide you with a comprehensive, simple system that needs only minimum administration. Access is available via IP authentication or username and password.

E-mail alert services These services allow you to be kept up to date with the latest additions to the journal via e-mail, as soon as new material enters the database. Further information about the services available can be found at www.emeraldinsight.com/alerts

Emerald online training services Visit www.emeraldinsight.com/training and take an Emerald online tour to help you get the most from your subscription.

Key features of Emerald electronic journals Automatic permission to make up to 25 copies of individual articles This facility can be used for training purposes, course notes, seminars etc. This only applies to articles of which Emerald owns copyright. For further details visit www.emeraldinsight.com/ copyright Online publishing and archiving As well as current volumes of the journal, you can also gain access to past volumes on the internet via Emerald Management Xtra. You can browse or search these databases for relevant articles. Key readings This feature provides abstracts of related articles chosen by the journal editor, selected to provide readers with current awareness of interesting articles from other publications in the field. Non-article content Material in our journals such as product information, industry trends, company news, conferences, etc. is available online and can be accessed by users. Reference linking Direct links from the journal article references to abstracts of the most influential articles cited. Where possible, this link is to the full text of the article. E-mail an article Allows users to e-mail links to relevant and interesting articles to another computer for later use, reference or printing purposes.

Xtra resources and collections When you register your journal subscription online, you will gain access to Xtra resources for Librarians, Faculty, Authors, Researchers, Deans and Managers. In addition you can access Emerald Collections, which include case studies, book reviews, guru interviews and literature reviews.

Emerald Research Connections An online meeting place for the research community where researchers present their own work and interests and seek other researchers for future projects. Register yourself or search our database of researchers at www.emeraldinsight.com/ connections

Choice of access Electronic access to this journal is available via a number of channels. Our web site www.emeraldinsight.com is the recommended means of electronic access, as it provides fully searchable and value added access to the complete content of the journal. However, you can also access and search the article content of this journal through the following journal delivery services: EBSCOHost Electronic Journals Service ejournals.ebsco.com Informatics J-Gate www.j-gate.informindia.co.in Ingenta www.ingenta.com Minerva Electronic Online Services www.minerva.at OCLC FirstSearch www.oclc.org/firstsearch SilverLinker www.ovid.com SwetsWise www.swetswise.com

Emerald Customer Support For customer support and technical help contact: E-mail [email protected] Web www.emeraldinsight.com/customercharter Tel +44 (0) 1274 785278 Fax +44 (0) 1274 785201

MF 33,2

88

EDITORIAL ADVISORY BOARD

Professor Kofi A. Amoateng NC Central University, Durham, USA

Professor Moses L. Pava Yeshiva University, New York, USA

Professor Felix Ayadi Jesse E. Jones School of Business, Texas Southern University, USA Professor Mohamed E. Bayou The University of Michigan-Dearborn, USA Dr Andre de Korvin University of Houston-Downtown, USA Dr Colin J. Dodds Saint Mary’s University, Halifax, Nova Scotia, Canada Professor John Doukas Old Dominion University, Norfolk, Virginia, USA

Professor George C. Philipatos The University of Tennessee, Knoxville, Tennessee, USA Professor David Rayome Northern Michigan University, USA Professor Alan Reinstein Wayne State University, Detroit, Michigan, USA Professor Ahmed Riahi-Belkaoui The University of Illinois at Chicago, USA Professor Mauricio Rodriguez Texas Christian University, Fort Worth, Texas, USA

Professor Uric Dufrene Indiana University Southeast, New Albany, Indiana, USA

Professor Salil K. Sarkar Henderson State University, Arkadelphia, Arkansas, USA

Professor Ali M. Fatemi De Paul University, Chicago, Illinois, USA

Professor Atul A. Saxena Mercer University, Georgia, USA

Professor Iftekhar Hasan New Jersey Institute of Technology, USA

Professor Philip H. Siegel Monmouth University, New Jersey, USA

Professor Suk H. Kim University of Detroit Mercy, Detroit, USA

Professor Kevin J. Sigler The University of North Carolina at Wilmington, USA

Professor John Leavins University of Houston-Downtown, USA Professor R. Charles Moyer Wake Forest University, Winston-Salem, North Carolina, USA Dr Khursheed Omer University of Houston-Downtown, USA

Managerial Finance Vol. 33 No. 2, 2007 p. 88 # Emerald Group Publishing Limited 0307-4358

Professor Gordon Wills International Management Centres, UK Professor Stephen A. Zeff Rice University, Texas, USA

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0307-4358.htm

The price linkages between Malaysian unit trust funds and the stock market Short run and long run interrelationships

Unit trust funds and the stock market 89

Soo-Wah Low and Noor Azlan Ghazali School of Business Management, Faculty of Economics and Business, Universiti Kebangsaan Malaysia, Selangor, Malaysia Abstract Purpose – The primary objective of the paper is to examine the short and long run price linkages between Malaysian unit trust funds and the stock market index as proxied by the Kuala Lumpur composite index (KLCI) over the period 1996-2000. Design/methodology/approach – Cointegration analyses are used to identify the long run relationship between unit trust funds and the stock market index while Granger causality tests are used to measure the short run price linkages. Findings – Cointegration results show that the long run pricing performance of the unit trust funds differs significantly from that of the KLCI. Interestingly, the findings also reveal that two index funds are found not to be cointegrated with the stock market index. In the short run, one-way Granger causality test shows that changes in the KLCI Granger causes changes in the unit trust funds. This suggests that fund managers are responding to the past changes in the stock market index over the short run. Research limitations/implications – The findings of non-cointegration between passively managed funds and the KLCI are restricted to only two index funds in the sample among other actively managed funds. Since there were not enough index funds available over the study period, future research should include more index funds in the analysis. Practical implications – In the short run, investors may gather information on the changes in their portfolio composition by observing the movement in the KLCI. Originality/value – The paper represents the first evidence on the pricing relationships between unit trust funds and the local stock market index and the findings are important to investors in terms of their investment strategies. Keywords Price positioning, Unit trusts, Process analysis, Stock markets, Malaysia Paper type Research paper

Introduction Mutual funds or more popularly known as unit trust funds in Malaysia, have experienced considerable growth over the last decade in terms of the number of funds offered and the volume of capital managed by the unit trust management companies. As at 31 December 1995, the percentage of the net asset value of the funds to the Bursa Malaysia’s market capitalization was 7.80 per cent and this figure grew to approximately 12.10 per cent as at 31 December 2004[1]. A unit trust fund is an indirect investment product created to serve as an alternative to direct stock market investment for investors. It is, therefore, reasonable to expect that fund prices have some degree of responsiveness to the direct equity market as proxied by the Kuala Lumpur composite index (KLCI). The primary objective of this paper is to examine the possible pattern of cointegration and causal relationships between unit trust funds and the stock market index in Malaysia. The degree to which fund prices are related to the stock market index has several important implications for investors with regard to their investment strategies.

Managerial Finance Vol. 33 No. 2, 2007 pp. 89-101 # Emerald Group Publishing Limited 0307-4358 DOI 10.1108/03074350710715827

MF 33,2

90

On the one hand, the presence of cointegration or long run equilibrium relationship between unit trust funds and the stock market index suggests that investors could regard investing in unit trust funds as an alternative or a substitute for direct investing in the stock market (see Ben-Zion et al., 1996; Matallin and Nieto, 2002). The existence of cointegration relationship implies that unit trust funds are replicating the stock market index over the long run. On the other hand, the lack of cointegration or long run equilibrium relationship between unit trust funds and the stock market index suggests that unit trust funds do not show parallel movement with the market index in the long run. In other words, there are sufficient long run differences between unit trust funds and the stock market index to provide further evidence for the existence of active fund management activities among the fund managers. The paper is organized as follows. Section II provides brief discussion on previous research. Section III explains the data and methodology employed. The findings are discussed in Section IV and concluding remarks are offered in Section V. Previous research The literature on the investment performance of mutual funds is extensive and spans several decades. Many of these empirical studies make a comparison between the fund’s return with that of the market. Such comparison allows investors to gauge the differences in the performance between actively managed funds and a passively managed portfolio, i.e. the market index. In general, most previous studies with few exceptions have found either negative performance or no performance at the aggregate level for the average mutual funds. Some of the more important studies showing a lack of superior performance by fund managers are Sharpe (1966), Jensen (1968), McDonald (1974), Chang and Lewellen (1984), Cumby and Glen (1990) and Droms and Walker (1994) among others. Within the studies on fund performance, the strand of literature, which focuses on timing and selectivity performance of mutual funds have reached mixed conclusions. Studies on this aspect of the research are provided by Merton (1981), Henriksson and Merton (1981), Chang and Lewellen (1984), Henriksson (1984), Cumby and Glen (1990), Chen et al. (1992), Coggin et al. (1993), Kao et al. (1998) and Rao (2000) among others. The issue of whether active management of mutual funds through market timing and security selections, generates superior or inferior abnormal returns is not within the scope of this study[2]. This paper aims to examine the degree to which mutual funds are related to the market index regardless of whether these funds are actively or passively managed. There are several studies that examine the relationship between mutual funds and local stock market indices. Bailey and Lim (1992) find significant correlations between the returns of country funds and the returns of the market index. However, they find that the pricing of country funds reflects more of the domestic US stocks than of the foreign equities in which these funds are invested. While correlation analysis gauges the degree of co-movement between the fund and the market, cointegration analysis tests whether there is a tendency for the two variables to move together with time towards a long run equilibrium state. Chang et al. (1995) find that cointegration exists between a country fund’s share value (determined in the US market) and its NAV (determined in the home market) for majority of the closed-end country funds from North America and Europe. However, they find no long run equilibrium relationship between the two variables for Asian emerging markets (i.e. India, Korea, Malaysia, Taiwan and Thailand) as well as for Brazil and Spain.

Another study by Allen and Macdonald (1995) also include Malaysia as part of their sample countries. In their study on international equity diversification, Allen and Macdonald (1995) find no evidence of cointegration between Australia and the following financial markets: Austria, Belgium, Italy, Japan, Norway, Malaysia, Singapore, Spain, Sweden and USA. Their findings suggest that Australian investors could benefit by diversifying into these countries. Ben-Zion et al. (1996) use cointegration and Granger causality tests to examine the price linkages between three country funds listed in the USA and their respective local indices (Germany, Japan and UK funds). Their study finds that none of the country funds are cointegrated with the local stock markets. The Granger causality tests show that dual causality does exist between the three funds and their own local market indices. In a more recent study, Matallin and Nieto (2002) use cointegration analysis to investigate the relationship between mutual funds and the Spanish stock market portfolio, Ibex 35. Their findings indicate that only 11 out of 63 funds analyzed are cointegrated with the market index or have a long run equilibrium relationship with Ibex 35. According to the authors, the findings of a small number of mutual funds, which are cointegrated with the market index could be due to the fact that these funds are actively managed funds. This is because active management of mutual funds through market timing and or security selection activities could cause the mutual funds to deviate from the evolution of the market index in the long run. In addition, the authors added that technical limitations associated with the creation of index funds which mirror the market portfolio could also explain why there is only a small number of funds which show parallel movement with the Ibex 35. In the Malaysian context, although there have been many empirical studies related to the investment performance of unit trust funds, most studies focused the analyses on the overall fund performance. Examples of such studies are provided by Shamsher and Annuar (1995), Tan (1995) and Leong and Aw (1997). Collectively, empirical findings on the overall fund performance indicate that on the average, unit trust funds in Malaysia perform worse than the market. The few studies that separate the overall fund performance into selectivity and market timing components are provided by Annuar et al. (1997), Low and Noor A. Ghazali (2005) and Low (2005). These studies show that on average, the timing performance of fund managers is negative. While there are many studies that investigate the investment performance of unit trust funds in Malaysia, so far no study has examined the price linkages between unit trust funds and the local stock market index. The purpose of the present paper is to add further evidence on these pricing relationships in the Malaysian market. The issue of whether or not fund prices and market index are cointegrated has important implication for investors in terms of their investment strategies. Many investors especially those who believe that there are inefficiencies in the equity market would purchase actively managed funds to capitalize on these inefficiencies in order to outperform the market. Such investors who opt for active management of their funds do so largely because of the belief that active strategies by fund managers through market timing and security selection offer better opportunity of adding value to their portfolios. In other words, these investors believe that the active management of portfolios by fund managers could cause the funds to deviate from the evolution of the market index with the hope of achieving better performance than the market index. Similarly, funds which are passively managed, for example, index funds are more likely to show evidence of parallel movement with the stock market index. This is

Unit trust funds and the stock market 91

MF 33,2

92

because by definition an index fund is fund whose portfolio matches that of a market index and therefore both index fund and the market index should move in tandem with each other. Investors who subscribe to market efficiency purchase index funds to at least keep up with the market performance since they believe that trying to outperform the market averages over the long run is a futile effort. As noted by Elton et al. (1996), due to the nature of passive management, index funds are available to investors at a relatively lower cost than those actively managed portfolios. It is, therefore, important to identify the nature of the existing relationship between unit trust funds and the market index so that investors can work out their investment strategy accordingly. Data and methodology The data employed in this study consists of monthly prices for 35 unit trust funds and one stock market portfolio as proxied by the KLCI (formerly known as the Kuala Lumpur Stock Exchange (KLSE) Composite Index). The data is obtained from the local newspapers, Investors’ Digest (January 1996-December 2000) published by the KLSE, fund prospectus and annual reports of the fund management companies. This study covers the period January 1996-December 2000. Appendix Table AI provides data description on the sample funds used in the study. A prior condition for causality and cointegration tests is that the time series or variables are stationary. We perform the unit root tests based on the Augmented Dickey–Fuller (ADF) test (Dickey and Fuller (1979; 1981)) to test for the presence of unit roots in both the price levels and first difference of funds and the stock market index. The ADF test is based on the following ordinary least squares (OLS) estimation:
xt ¼
þ T þ axt
1 þ

k X

’i
t
i þ t

ð1Þ

i¼1

where,
is the first-difference operator; xt is the time series or variable tested for stationarity; T is a linear time trend; and t is a covariance stationary random error. The appropriate number of lagged difference (K) is determined by Akaike Information Criteria (AIC) due to Akaike (1970). Optimal choice of lag length removes autocorrelations in the error term. The null hypothesis of unit root, |a| ¼ 1, is tested against alternative of stationarity, |a|< 1. The critical value for the test was developed by MacKinnon (1991). If a time series is stationary, any shock to the variable will temporarily or momentarily draw the variable away from its long run mean values. However, if the series is non-stationary, the deviation from the long run mean values will be permanent. By definition, a series or variable which is I(0) is said to be stationary at the level form. The problem of non-stationarity can be eliminated by taking differences in the series. Therefore, if the series is characterized by I(d ), that is, integrated of order d, it means that the series need to be differenced d times before becoming a stationary series. For example, if the series tested is I(1), it means first-differencing is required to achieve stationarity in the series. Evidence that the series tested are I(1) form a prior condition for cointegration relationships to exist. Engle and Granger (1987), here onward EG, note that even though economic or financial time series may be described as a random walk process, it is possible that the linear combinations of the series or variables would over time converges to an equilibrium. If two series are non-stationary in their level forms, that is I(1) and the

series are integrated of the same order, d, (for example, both series are I(1)) and if the error term from regressing one series on the other is stationary, then the series are said to be cointegrated. Thus, cointegration exists if two variables are individually I(1) and the error term from the linear regression between the two variables is I(0). In this study, the two time series tested are the price levels of unit trust funds and the market index. Cointegration between unit trust funds and the market index suggests that over the long run, they move in tandem with each other although the behavior of unit trust funds could be different from that of the index in the short run. To test for possible existence of cointegration, we performed ADF test on the error term, "t from the following linear combinations between the unit trust funds and the market index: FUNDt ¼  þ  KLCIt þ "t

ð2Þ

Each fund is tested for possible existence of cointegration between the fund and the index and this required us performing 35 cointegration tests (one per unit trust fund). Logic dictates that the independent variable should be the stock market index since the behaviors of the unit trust funds are driven by the changing patterns in the stock market index. On the one hand, based on the cointegration analysis, if the error series, "t is found to be stationary (i.e. I(0)), then unit trust funds and the stock market index are said to be cointegrated. When the two variables are cointegrated, they have long run equilibrium relationship although in the short run, the two variables may deviate momentarily from each other. Over time, one of the variables would react to the deviation and make the necessary adjustments to maintain a long run equilibrium with the other variable. In our case, the presence of cointegration relationship suggests that unit trust funds will be subjected to its deviation from long run path dictated by the market index. The cointegration relationship can be represented by an error correction model (ECM) in the following equation:
FUNDt ¼ "t
1 þ

p X

j
FUNDt
j þ

j¼1

q X

j
KLCIt
j

þ t

ð3Þ

j¼1

On the other hand, if the error series, "t in Equation (2) is found to be non-stationary, then there exist no long run equilibrium relationship between unit trust funds and the stock market index. For that reason, we employ Granger causality tests to investigate the possible short term relationship between unit trust funds and the stock market index. In view of the fact that fund managers do include some of the index-based component stocks into their portfolio, it only make sense that the changes in the stock market index bring about changes in unit trust funds. Hence, we test for one-way causality by running a regression of unit trust funds on past values of unit trust funds and the stock market index as shown in the following equation:
FUNDt ¼ 0 þ

J X j¼1

j
FUNDt
j þ

K X

k
KLCIt
k þ t

ð4Þ

k¼1

The optimal lag structures are chosen based on the Akaike’s Information Criteria (AIC) due to Akaike (1970). The F-test of the joint significance of the estimated lagged coefficients of k provides insights into the short run relationship. Statistically significant F-statistic would indicate that
KLCI Granger causes

Unit trust funds and the stock market 93

MF 33,2

94


FUND, implying that fund managers are responding to the past changes in the stock market index over the short run. This would be precisely the behavior to be expected in the short run since most funds are structured partially with some index-based component stocks. Results and discussion Table I presents the ADF unit root tests for 35 unit trust funds and the stock market index in both the price levels and first-differenced forms. The ADF test statistics for

No.

Fund name

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35

ASN M Investment SBB Double Growth CT Prime KL City Sapphire KL City Ruby KL City Emerald Public Index Public Regular Saving ASM 2 Index ASM KMBY 3 ASM KMBY 4 ASM KMBY 5 ASM KMBY 6 ASM KMBY 7 ASM KMBY 8 ASM KMBY 10 Mayban Unit Trusts MBf Growth RHB Dynamic RHB Capital M Berjaya M Equity SBB Emerging Co. Growth SBB Savings SBB High Growth HLG Growth Public Industry Public Aggressive Growth ASM KMB Growth Pacific Premier ASM TP Balanced Public Balanced Mayban Balanced MBf Balanced

Price level

First difference


1.494
1.958
2.153
1.857
1.553
1.306
1.285
3.180
2.216
1.781
1.663
1.750
1.865
1.777
1.938
1.752
1.587
2.158
1.524
2.228
1.754
1.903
1.672
3.109
2.001
2.376
2.303
1.847
1.724
1.746
1.793
1.915
1.334
1.652
1.537


4.527*
5.187*
4.833*
5.274*
6.203*
8.043*
6.746*
7.205*
7.264*
6.582*
6.201*
5.933*
5.728*
6.241*
6.600*
7.076*
5.999*
3.853*
6.762*
6.245*
6.438*
6.295*
5.914*
3.824*
5.584*
4.527*
3.660*
6.807*
6.607*
6.303*
6.657*
7.332*
3.982*
4.401*
7.316*

Table I. Stock market index Augmented Dickey– Kuala Lumpur Composite Index (KLCI)
1.669
5.7915* Fuller (ADF) unit root tests – level and first differences of fund prices Notes: * Indicates significance at the 5 per cent level or lower; the critical values of the ADF tests are developed by MacKinnon (1991); the critical value for 5 per cent level of significance is and stock market 3.447 for ADF model with constant and trend parameters index (KLCI)

the price levels series are not significant for all of the 35 unit trust funds and the stock market index. However, the ADF test statistics are significant in first-differenced form at the 5 per cent level for the stock market index and for the 35 unit trust funds. This suggests that both the funds and index series are I(1) that is, both the variables are first difference stationary. Table II shows the results of the 35 cointegration tests performed on each of the unit trust fund and the stock market index. The findings obtained from the cointegration tests show that none of the funds are cointegrated with the stock market index, KLCI. The ADF statistics testing for stationarity of the error series, "t from Equation (2) are not significant for all of the 35 funds. The absence of cointegration suggests that long run equilibrium relationship does not exist between unit trust funds and the stock market index. This implies that, in the long run, the prices for unit trust and the stock

Dependent variable fund name ASN M Investment SBB Double Growth CT Prime KL City Sapphire KL City Ruby KL City Emerald Public Index Public Regular Saving ASM 2 Index ASM KMBY 3 ASM KMBY 4 ASM KMBY 5 ASM KMBY 6 ASM KMBY 7 ASM KMBY 8 ASM KMBY 10 Mayban Unit Trusts MBf Growth RHB Dynamic RHB Capital M Berjaya M Equity SBB Emerging Co. Growth SBB Savings SBB High Growth HLG Growth Public Industry Public Aggressive Growth ASM KMB Growth Pacific Premier ASM TP Balanced Public Balanced Mayban Balanced MBf Balanced

Independent variable stock market index

ADF test for error term

KLCI KLCI KLCI KLCI KLCI KLCI KLCI KLCI KLCI KLCI KLCI KLCI KLCI KLCI KLCI KLCI KLCI KLCI KLCI KLCI KLCI KLCI KLCI KLCI KLCI KLCI KLCI KLCI KLCI KLCI KLCI KLCI KLCI KLCI KLCI


2.104
1.733
1.563
2.206
2.705
2.263
2.644
3.318
2.987
2.112
2.259
2.349
2.660
2.139
2.557
2.618
2.541
2.422
2.255
2.979
3.224
3.054
2.339
1.760
1.319
2.480
2.568
1.897
1.934
2.324
3.073
0.210
3.219
2.825
2.472

Unit trust funds and the stock market 95

Table II. Evidence of cointegration between fund prices and stock market index (KLCI)

MF 33,2

96

market index do not converge and there are sufficient long run differences between the funds and the stock market index. For actively managed funds, this result is attributable, in part, to the fact that the active management of the funds through market timing and security selection prevents the funds from having a parallel movement with the stock market index in the long run. It is also relatively common for fund managers to restructure the composition of their portfolio as the market condition changes. Furthermore, in view of the fact that fund managers have to invest according to the stated investment guidelines, it is therefore important for them to rebalance their portfolio holdings with the aim of maintaining a strategic asset allocation strategy in the long run. Hence, the active management of funds, together with the rebalancing activities of fund managers could help explain why a long run equilibrium relationship is not observable between unit trust funds and the stock market index. However, it is interesting to note from Table II that, there are two index funds (Public Index and ASM 2 Index) which are found not to be cointegrated with the market index[3]. The absence of a long run relationship between index funds and the stock market index is certainly surprising. This is because, by definition, the performance of index fund should mimic that of the stock market index and both of the portfolios should move in tandem with each other. The finding implies that investment in index fund is not equivalent to direct investment in securities represented in the stock market index. Investors who seek passive investing approach by purchasing index funds should, therefore, reassess their investment strategies. The results highlight the fact that in a changing market environment, even the most prudently constructed portfolios which are constructed completely with index-based component stocks, fail to show parallel movement with the stock market index in the long run. As pointed out by Matallin and Nieto (2002), such finding may be attributed to the technical limitations associated with the construction of index funds which are intended to mimic the stock market index. Nevertheless, in interpreting the results, it must be kept in mind that this study uses KLCI as a proxy for the stock market index, without identifying the specific stock market index which the two index funds are mimicking. Thus, the findings of noncointegration between index funds and the stock market index, may be attributed to the fact that the two index funds are actually replicating the investment in a particular stock market index other than the KLCI. Furthermore, even if the KLCI in actual fact represents an appropriate stock market index for both of the index funds, if the investment percentages of the funds are constructed differently from those of the KLCI, this could also prevent the two index funds from showing a parallel movement with the KLCI in the long run. Table III presents the Granger causality test results for unit trust funds and the stock market index. Among the 35 funds studied, 13 funds (representing 37.1 per cent of the total) are found to have causal links with the stock market index at the 5 per cent significance level or lower. Six funds exhibit short run causality at the 10 per cent level of significance. The findings that changes in the KLCI Granger causes the changes in the unit trust funds suggest that the transmission of information as contended earlier on, is flowing from the market index to the unit trust funds[4]. The results suggest that unit trust funds are related to the stock market index in the short run and this implies that fund managers are responding to the changes in the KLCI by adjusting the composition of their portfolio holdings.

No.

Fund name

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35

ASN M Investment SBB Double Growth CT Prime KL City Sapphire KL City Ruby KL City Emerald Public Index Public Regular Saving ASM 2 Index ASM KMBY 3 ASM KMBY 4 ASM KMBY 5 ASM KMBY 6 ASM KMBY 7 ASM KMBY 8 ASM KMBY 10 Mayban Unit Trusts MBf Growth RHB Dynamic RHB Capital M Berjaya M Equity SBB Emerging Co. Growth SBB Savings SBB High Growth HLG Growth Public Industry Public Aggressive Growth ASM KMB Growth Pacific Premier ASM TP Balanced Public Balanced Mayban Balanced MBf Balanced


KLCI does not cause
Fund F-statistics (Lags J, K) 1.571 0.230 0.378 0.547 4.836 6.708 3.450 2.608 4.442 0.587 0.435 0.606 0.489 1.915 1.159 1.321 6.708 5.090 3.257 5.515 6.469 4.446 1.485 0.766 0.217 1.031 3.919 2.692 6.064 1.779 4.144 3.579 4.095 2.807 2.215

(2,1) (2,1) (2,1) (1,1) (2,1)* (1,1)* (2,1)** (1,10)* (1,1)* (6,3) (2,12) (5,1) (2,12) (5,1) (6,12) (3,1) (7,1)* (1,2)* (3,1)** (4,1)* (9,1)* (2,1)* (1,4) (4,1) (2,1) (3,6) (4,1)** (7,1)** (1,1)* (1,1) (2,1)* (1,1)** (1,4)* (2,11)* (3,3)**

Notes: * Indicates significance at the 5 per cent; ** indicates significance at the 10 per cent; AIC was used to determine the optimal lag lengths (included in parentheses)

Most fund managers do include blue chips stocks as part of their portfolio holdings and these stocks are very likely to be index-based component stocks. Hence, the inclusion of index-based component stocks into the unit trust portfolio could perhaps explain the short run relationship that exists between the fund and the stock market index. In view of the fact that actively managed funds are constructed partially with some index-based component stocks, it is therefore not surprising that changes in the fund portfolios are determined by the changes in the stock market index in the short run. To summarize, in terms of the interrelationships between unit trust funds and the stock market index, this study finds that while in the short run, changes in the KLCI does influence the price changes of unit trust funds, in the long run, the two variables

Unit trust funds and the stock market 97

Table III. Granger causality tests

MF 33,2

98

are not cointegrated or do not move in tandem with each other. This suggests that the two variables may deviate from each other over time and they are not bounded by a long run equilibrium relationship. Summary and conclusion This paper examines the price linkages between unit trust funds and the stock market index during the period 1996-2000. We perform ADF unit root tests on both the price levels and first difference of funds and the stock market index. The results show that both the fund and index prices are first difference stationary. Cointegration analyses are used to determine the existence of long run relationship between the unit trust funds and the stock market index while Granger causality tests are used to measure the short run price linkages. The results of the cointegration tests reveal that there is no long run equilibrium relationship between unit trust funds and the KLCI. This suggests that in the long run, the performance of unit trust funds can diverge significantly from that of the stock market index. Since fund managers are obliged to adhere to their investment policies with the aim of maintaining a long term allocation strategy, this could possibly explain why unit trust funds and the stock market index do not have a long run equilibrium relationship. Interestingly, the findings also show that index funds and the stock market index do not share a common trend in the long run. This raises questions about the capability of index funds to replicate the stock market index because such finding suggests that investing in index funds is not equivalent to direct investing in the market index. While in the long run both the prices of unit trusts and the stock market index do not converge, in the short run, we find that the prices of unit trust funds are related to the stock market index. Since actively managed funds are constructed with some index-based component stocks, the short run relationship that exists between the fund portfolios and the market index is implying that managers are in fact responding to the past changes in the KLCI by changing the securities components of their portfolio holdings. Given such results, investors may gain insights into the activities of the fund managers by observing the movement in the KLCI. Notes 1. Source: Securities Commission, Malaysia. 2. Most previous evidence on timing and selectivity is based on findings from developed countries. In the Malaysian context, there is remarkably little evidence which make the distinction between performance due to selectivity and market timing abilities of fund managers. The few studies are provided by Annuar et al. (1997), Low and Noor A. Ghazali (2005) and Low (2005). 3. There were not enough index funds available over the study period to allow a separate cointegration analysis for index funds. 4. We also performed the causality tests to determine if unit trust funds Granger cause the market index, KLCI. As we would expect, the results were not statistically significant. References Akaike, H. (1970), ‘‘Autogressive model fitting for control’’, Annals of the Institute of Statistical Mathematics, Vol. 22, pp. 163-80. Allen, D.E. and Macdonald, G. (1995), ‘‘The long-run gains from international equity diversification: Australian evidence from cointegration tests’’, Applied Financial Economics, Vol. 5, pp. 33-42.

Annuar, M.N., Shamsher, M. and Ngu, M.H. (1997), ‘‘Selectivity and timing: evidence from the performance of Malaysian unit trusts’’, Pertanika Journal of Social Science and Humanities, Vol. 5, pp. 45-57. Bailey, W. and Lim, J. (1992), ‘‘Evaluating the diversification benefits of the new country funds’’, Journal of Portfolio Management, Vol. 18, pp. 74-80. Ben-Zion, U., Choi, J.J. and Hauser, S. (1996), ‘‘The price linkages between country funds and national stock markets: evidence from cointegration and causality tests of Germany, Japan and UK Funds’’, Journal of Business Finance and Accounting, Vol. 23, pp. 1005-17. Chang, E., Eun, C.S. and Kolodny, R. (1995), ‘‘International diversification through closed-end country funds’’, Journal of Banking and Finance, Vol. 19, pp. 1237-63. Chang, E.C. and Lewellen, W.G. (1984), ‘‘Market timing and mutual fund investment performance’’, Journal of Business, Vol. 57, pp. 57-72. Chen, C.R., Lee, C.F., Rahman, S. and Chan, A. (1992), ‘‘A cross-sectional analysis of mutual funds’ market timing and security selection skill’’, Journal of Business Finance and Accounting, Vol. 19, pp. 659-75. Coggin, T.D., Fabozzi, F.J. and Rahman, S. (1993), ‘‘The investment performance of US equity pension fund managers: an empirical investigation’’, Journal of Finance, Vol. 48, pp. 1039-55. Cumby, R.E. and Glen, J.D. (1990), ‘‘Evaluating the performance of international mutual funds’’, Journal of Finance, Vol. 45, pp. 497-521. Dickey, F. and Fuller, W.A. (1979), ‘‘Distribution of the estimates for autogressive time series with a unit root’’, Journal of American Statistical Association, Vol. 74. pp. 427-31. Dickey, F. and Fuller, W.A. (1981), ‘‘Likelihood ratio statistics for autoregressive time series with a unit root’’, Econometrica, Vol. 49, pp. 1057-72. Droms, W.G. and Walker, D.A. (1994), ‘‘Investment performance of international mutual funds’’, Journal of Financial Research, Vol. 17, pp. 1-14. Elton, E.J., Gruber, M.J. and Blake, C.R. (1996), ‘‘The persistence of risk-adjusted mutual fund performance’’, Journal of Business, Vol. 69. pp. 133-57. Engle, R.F. and Granger, C.W. (1987), ‘‘Cointegration and error correction: representation, estimation and testing’’, Econometrica, Vol. 55 pp. 251-76. Henriksson, R.D. (1984), ‘‘Market timing and mutual fund performance: an empirical investigation’’, Journal of Business, Vol. 57, pp. 73-96. Henriksson, R.D. and Merton, R.C. (1981), ‘‘On market timing and investment performance II. Statistical procedures for evaluating forecasting skills’’, Journal of Business, Vol. 54, pp. 513-33. Investors’ Digest ( January 1996-December 2000), Monthly publication of the Kuala Lumpur Stock Exchange, various issues. Jensen, M.C. (1968), ‘‘The performance of mutual funds in the period 1945-1964’’, Journal of Finance, Vol. 23, pp. 389-416. Kao, G.W., Cheng, L.T.W. and Chan, K.C. (1998), ‘‘International mutual fund selectivity and market timing during up and down market conditions’’, Financial Review, Vol. 33, pp. 127-44. Leong, K.H. and Aw, M.W. (1997), ‘‘Measuring unit trust fund performance using different benchmarks’’, Capital Market Review, Vol. 5, pp. 27-44. Low, S.W. (2005), ‘‘Malaysian mutual fund performance during up and down market conditions: a comparison of market benchmark’’, working paper, Universiti Kebangsaan Malaysia.

Unit trust funds and the stock market 99

MF 33,2

100

Low, S.W. and Noor A. Ghazali, (2005), ‘‘An evaluation of the market timing and security selection performance of mutual funds: the case of Malaysia’’, International Journal of Management Studies, Vol. 12, pp. 215-33. MacKinnon, J.G. (1991), ‘‘Critical values for cointegration tests in long-run econometric relationships, readings in cointegration’’, in Engle, R.F. and Granger, W.J. (Eds), Oxford Press, New York, NY, pp. 266-76. Matallin, J.C. and Nieto, L. (2002), ‘‘Mutual funds as an alternative to direct stock investment: a cointegration approach’’, Applied Financial Economics, Vol. 12, pp. 743-50. McDonald, J.G. (1974), ‘‘Objectives and performance of mutual funds’’, Journal of Financial and Quantitative Analysis, Vol. 9, pp. 311-33. Merton, R.C. (1981), ‘‘On market timing and investment performance. I. An equilibrium theory of value for market forecasts’’, Journal of Business, Vol. 54, pp. 363-406. Rao, S.P.U. (2000), ‘‘Market timing and mutual fund performance’’, American Business Review, Vol. 18, pp. 75-79. Shamsher, M. and Annuar, M.N. (1995), ‘‘The performance of unit trusts in Malaysia: some evidence’’, Capital Market Review, Vol. 3, pp. 51-69. Sharpe, W.F. (1966), ‘‘Mutual fund performance’’, Journal of Business, Vol. 39, pp. 119-38. Tan, H.C. (1995), ‘‘The investment performance of unit trust funds in Malaysia’’, Capital Market Review, Vol. 3, pp. 21-50. Further reading Bank Negara Malaysia (January 1996-December 2000), Monthly statistical bulletin, various issues. Appendix

Table AI. List of unit trust funds used in the sample

No.

Fund name

Fund type

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22

ASN M Investment SBB Double Growth CT Prime KL City Sapphire KL City Ruby KL City Emerald Public Index Public Regular Saving ASM 2 Index ASM KMBY 3 ASM KMBY 4 ASM KMBY 5 ASM KMBY 6 ASM KMBY 7 ASM KMBY 8 ASM KMBY 10 Mayban Unit Trusts MBf Growth RHB Dynamic RHB Capital M Berjaya

Federal Income Growth Income Income Income Income Index Income Index Income Income Income Income Income Income Income Income Growth Income Growth Growth (Continued )

No.

Fund name

Fund type

23 24 25 26 27 28 29 30 31 32 33 34 35

M Equity SBB Emerging Co. Growth SBB Savings SBB High Growth HLG Growth Public Industry Public Aggressive Growth ASM KMB Growth Pacific Premier ASM TP Balanced Public Balanced Mayban Balanced MBf Balanced

Small companies Small companies Balanced Growth Growth Income Growth Growth Income Balanced Balanced Balanced Balanced

Source: Federation of Malaysian Unit Trust Managers (FMUTM), available at: www.fmutm.com.my

Corresponding author Soo-Wah Low can be contacted at: [email protected]

To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

Unit trust funds and the stock market 101

Table AI.

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0307-4358.htm

MF 33,2

Malaysian unit trust aggregate performance Fauziah Md. Taib

102

School of Management, Universiti Sains Malaysia, Penang, Malaysia, and

Mansor Isa Faculty of Business and Accountancy, Universiti Malaya, Kuala Lumpur, Malaysia Abstract Purpose – This paper seeks to focus on examining unit trust performance in Malaysia over the period 1991-2001. Design/methodology/approach – The broad based study covers full economic cycles using 7 different performance measures: raw return, market adjusted return, Jensen’s alpha, adjusted Jensen’s alpha, Sharpe Index, adjusted Sharpe Index, and Treynor Index. Findings – The results show that on average the performance of Malaysian unit trust falls below market portfolio and risk free returns. However, the variance of unit trust monthly returns is less than the market. Performance by type of funds indicates that bond funds show relatively superior performance, over and above the market and equity unit trusts. This is due to the high interest rate kept during the crisis period. Findings also suggest that there is no persistency in performance as there is no significant inter-temporal correlation between past and current performance. Research limitations/implications – The issue of inferior performance needs further investigations to adjust for great importance placed on maintaining consistent dividend distribution. In addition, ill-managed funds must be separately analysed to see if limited budget, less qualified managers, use of limited information and less sophisticated software could explain the poor performance. Practical implications – A very useful source of information for potential investors and portfolio management companies looking for opportunities to invest. Originality/value – The paper contributes to the present body of knowledge by offering broad based performance evidence from an emerging market with strong government back up for unit trusts investment. Keywords Unit trusts, Performance monitoring, Fund management, Malaysia Paper type Research paper

Managerial Finance Vol. 33 No. 2, 2007 pp. 102-121 # Emerald Group Publishing Limited 0307-4358 DOI 10.1108/03074350710715836

Introduction Having experienced dramatic growth during the period of 1990-1996, the East Asian countries have seen the emergence of mutual funds as one of the important vehicles of investment. In Malaysia for instance, the net assets value of unit trust funds was just USD 4.14 billion (RM15.72 billion) in 1992 but grew to USD 14.14 billion (RM53.74 billion) in 2002. While many of the earlier studies have comprehensively looked at the performance of mutual funds in the developed countries, more is yet to be understood from the emerging markets. A review of the previous literature mostly originating from the developed country seems to indicate there is a changing trend in the mutual funds performance before and after 1980s. Most of the earlier studies show the absence of persistency in funds performance and the general conclusion is that fund managers have no superior The authors wish to thank ASEAN University Network for the funding of this research. The views expressed and responsibilities for any errors are, however, those of the authors.

forecasting ability (Sharpe, 1966; Jensen, 1968; Firth, 1977). However, evidence after the 1980s suggests otherwise. Managers appear to hold hot hand and there is persistency in mutual funds performance (Hendricks et al., 1993; Brown and Goetzman, 1995; Wermers, 1996). The change in trend might signify that there is a possibility that the fund managers in developed countries have significantly improved their portfolio techniques. Evidence from Malaysia with regard to unit trust performance is very limited. Most of the studies use small sample sizes and the results are inconclusive. The earliest study with only 12 unit trusts as a sample reports that the performance is well above market return and quite consistent over the period 1974-1984 (Chua, 1985). Later studies, suggest otherwise, that is unit trusts produce lower returns than the market portfolio (Ewe, 1994; Shamsher and Annuar, 1995). Given the strong interest by the Malaysian government in this area as evidenced by active promotion and offering of National Unit Trust to the public, questions remain whether promoting unit trust as a form of safe investment in Malaysia is warranted in terms of risk and return. This study is set to present another look at the unit trusts performance with a bigger sample size and covering a longer time frame to assess the current status of the industry. The rest of the paper is organized as follows. Malaysian unit trust presents the background of the Malaysian unit trust industry and discusses the relevant Malaysian empirical studies. Methodology describes the sample and the methodology. Results and conclusion are presented in sections 4 and 5, respectively. Malaysian unit trust The unit trust industry in Malaysia was first established by British investors in 1959 with the introduction of the Malayan Unit Trust Ltd. It is called unit trust instead of mutual fund because the ownership of the fund is divided into units of entitlement. Initially, the growth of the unit trust in Malaysia was very slow due to lack of public interest. The turning point for the industry was when the Malaysian government decided to enter into the industry by launching a government sponsored unit trust known as Amanah Saham Nasional (ASN). The initial intention of launching the unit trust was to help improve the indigenous Malays’ (Bumiputera’s) social-economic status. Since then, the growth of the unit trusts (government and private funds) has been tremendous particularly during the period of 1991-1996, although the rapid development of the industry was slightly hindered by the 1997 Asian Financial Crisis. The rapid growth of the unit trust industry could be observed from the number of management companies from 13 in 1992 to almost triple the size to 37 in 2002. Similarly, the number of funds approved has also increased to 195 from 39 for the same period. A number of factors have jointly contributed to the rapid expansion of the industry and those include strong economic and good stock market performance, expansion of the local stock market and success of the privatization companies. Previous studies of the Malaysian unit trusts performance have yielded conflicting results. Chua (1985) finds that unit trust funds in Malaysia perform better than the market during his study period, 1974-1984. He concludes that the performance of unit trusts is fairly consistent and fund managers have diversified and performed risk control reasonably well. In addition, Chua finds that government sponsored funds perform better than private funds. This may be due to certain investment ‘‘privileges’’ accessible to only government-sponsored funds.

Malaysian unit trust aggregate performance 103

MF 33,2

104

Later studies, however, find results that are inconsistent with Chua’s findings. These studies include Ewe (1994), Shamsher and Annuar (1995), and Tan (1995). Shamsher and Annuar (1995), focus their study on the performance of 54 unit trusts covering the period of late 80s to early 90s. They find that the returns on investment in unit trust are well below the risk free and market returns. Furthermore, the results indicate that not only the degree of portfolios diversification is below expectation but the actual returns and risk characteristics of funds are also inconsistent with their stated objectives. Tan (1995) analyzed performance of 12 unit trusts over a 10-year period, 1984-1993. He concludes that unit trusts in general perform worse than the market portfolio. Consistent with Chua’s findings, Tan also concludes that government sponsored funds perform better than private funds. It therefore, seems that there exists a shift of performance over time. Fund managers seem to be doing well in the 70s, but have performance problems in the 80s. A study by Fauziah et al. (2002) seems to indicate that the poor performance of unit trust managers continues into later decades. Their study is based on 78 unit trust fund companies that exist during 1990-1999. They find that unit trust returns are not significantly above risk free and market returns. They discover that unit trusts do not exhibit consistent investment performance over time and there is no evidence to indicate that Malaysian fund managers have superior forecasting ability. However, their study lacks rigorous statistical testing and the conclusions reached are mainly based on individual fund observations and analysis. Caution should also be exercised in interpreting their results as the period of study includes the 1997-98 financial crisis. It is possible that the severity of the crisis has influenced the statistical results. The results of Ong (2000) completely contradict those of Taib et al. (2000). Ong studies the performance of 53 unit trusts (37 private and 16 government-sponsored) before and during the 1997-98 financial crisis. He finds that Malaysian unit trusts were able to out-perform the market before and during financial crisis. In fact, the performance of unit trust during the crisis is ‘‘better’’ than before crisis. Also, consistent with previous studies, he finds that government-sponsored funds perform better than private funds before the crisis. But the situation reverses during the crisis. He also finds that fund size has no influence on performance. The mixed evidence on fund performance in Malaysia is far from being resolved. Chong and Kho (2002) studied the persistence in the performance of 63 Malaysian equity unit trusts over the period 1991-2000 using different methods of analysis and different performance measures. Their findings are indeed a bag of mixed results. They find no performance persistence when using the cross-sectional regressions of residual returns. However, they find strong performance reversals for mean-adjusted residuals. And yet the time-series regressions indicate non-existence of performance persistence. Upon closer observation it was found that all of the findings except Taib et al. (2002) are based on either relatively small sample size or covering a limited time period. This does restrict the applicability of the findings and make it difficult for investors to draw inferences. The current study is yet another attempt to study fund performance but with more a comprehensive database. Our focus is on performance against a market benchmark, and its persistence over time. With a more comprehensive data set and using various performance measures, it is hoped that our results will provide a better understanding of the performance situation. Early studies of unit trust funds in developed markets show that unit trusts do not outperform the market and managers do not have superior ability to consistently beat the market (Sharpe, 1966; Jensen, 1968; Firth, 1977). Indirectly, the evidence indicates

that the market is remarkably efficient. Thus, the notion of studying past stock prices provide no helpful information in predicting future price movements hold. Studies in the 80s, however, have discovered that fund managers are able to outperform the market. This is in contrast to the general findings of earlier studies. Does it mean that market was efficient in the 70s and became inefficient in the 80s? It seems that fund managers are able to predict stock prices based on several fundamental variables such as initial dividend yields, market capitalization, price earning ratios, and price to book value ratios. Henriksson (1984) and Chang and Lewellen (1984) find that during the 70s, net returns of funds lie along the SharpeLintner market line. This implies that fund managers have access to enough private information to offset their expenses. The finding is later supported by Ippolito (1989) in a study on 143 mutual funds in the USA over the period of 1965-1984. Ippolito shows that mutual funds with high turnover, fees and expenses are able to earn higher returns to offset the higher charges. These results are consistent with the notion that mutual funds are efficient in their trading and information gathering activities. Recent studies by Grinblatt and Titman (1992), Hendericks et al. (1993), Goetzmann and Ibbotson (1994), Malkiel (1995), Gruber (1996), show that fund managers are able to outperform the market and the hot hand phenomenon does exist in the US market. Contrary to those of the US evidence, mutual fund studies in Australia generally find no evidence of persistency in performance. These studies include Robson (1986), Vos et al. (1995) and Hallahan (1997). Hallahan for example, using three different method for identifying performance predictability and four performance measures concludes that his overall results suggest that past performance of a fund is an unreliable guide to future performance. The inconsistency in the findings between USA and Australia certainly begs further investigation. One possible area of investigation is market specific characteristics, such as the differences in regulatory environment. A study in the European market by Pleschiutschnig (1999) reveals that persistency in performance is present amongst the European mutual funds. This finding is consistent with those of the US studies. Methodology As of 31 December 2001, there are 164 approved unit trusts in Malaysia, out of which 110 funds are included in this study. The remaining 54 funds are excluded due to various reasons: 28 are not yet launched, three are closed ended funds and 23 are newly launched funds (less than one year). Out of these the price of two of the funds managed by Permodalan Nasional Berhad (a government linked company) namely Amanah Saham Bumiputra (ASB) and Amanah Saham Wawasan 2020 (ASW2020), are nonfloating and thus excluded from this study. In essence, every fund that existed during the period was included to lessen the problem of survivorship bias. All funds listed for analysis are open-ended funds. Out of this sample, 91 fall under equity type of fund, 12 are balanced funds while the remaining seven are bond funds. Table I shows the sample size, while Table II shows the sample size by type of funds and sub periods. Various monthly data pertaining to the unit trust performance covering from January 1990 to December 2001 is collected from The Star and The Edge Malaysia newspapers. Information regarding dividend is obtained either from annual reports or collected directly from the fund managers. One-month Kuala Lumpur Inter-bank Offer Rate (KLIBOR) is used as a proxy for risk free while monthly return of Kuala Lumpur Composite Index (KLCI) is taken to serve as a benchmark for the market portfolio.

Malaysian unit trust aggregate performance 105

MF 33,2

106

Dividend yield is omitted from the calculation of market return, as it will not have significant influence on beta value (Sharpe and Cooper, 1972). In this study, we use three standard performance measures namely Jensen’s Alpha, Sharpe and Treynor Index and four variants of similar measures including raw returns to see if results are biased according to how it is being measured. The formula to compute the performance of unit trust, market index and risk free returns are as follows: NAVj
t þ Dj
t NAVj
t1 It ¼ loge It1
 1 þ Rf ;t ¼ loge 12

Rj;t ¼ loge Rm;t Rfm;t

ð1Þ ð2Þ ð3Þ

where Rf,t ¼ monthly continuously compounded rate of return of the jth unit trust during month t; NAVj,t ¼ net asset value for unit trust j at the end of month t ; Dj,t ¼ dividend per unit paid by unit trust j during month t; Rm,t ¼ estimated monthly continuously compounded rate of return on market portfolio m for month t ; It ¼ level of the KLCI at the end of month t; Rf,t ¼ KLIBOR for one month (quoted in yearly rate); and Rfm,t ¼ KLIBOR for one month (quoted in monthly rate). Sample selection

Table I. Malaysian unit trust funds

Total approved funds (as at 31 December 2001) Less: Non-floating funds Newly launched funds (less than one year) Not yet launch funds

164 (3) (23) (28)

Total number of unit trusts in sample

110

Types of funds/ sub period

Table II. Malaysian unit trust funds: sample size

Number

Equity Balance Bond Total

Overall period (1990-2001)

Before crisis (1990-1996)

During crisis (1997-1998)

After crisis (1999-2001)

Complete data (1990-2001)a

91 12 7 110

51 7 4 62

63 11 4 78

91 12 7 110

23 1 1 25

Note: aThe surviving unit trusts since 1990

Jensen (1968) shows that the capital asset pricing model (CAPM) holds for any arbitrary length of time as long as the returns are expressed in terms of the proper compounding length of interval. Jensen asserts that the natural logarithm form of return provides a very good approximate for calculating returns. Consequently, in an effort to avoid huge fluctuations in prices that might distort our data, we employ the compounded rate of return. Equations (1)-(3) are used to calculate the rates of return based on a continuous compounding method that was adopted by Jensen (1968). Jensen further suggests that loading charges could be excluded from the calculation of the funds’ rates of return when conducting an evaluation of the forecasting ability of fund managers. In addition, we omit the dividend yield of the market portfolio from our analysis, since, as mentioned earlier, Sharpe and Cooper (1972) suggest that the value of betas would not change significantly. Performance against benchmarks Adjusted Jensen’s alpha Aj ¼ ðRjt  Rft Þ 
j ½Rmt  Rft  þ "j

ð4Þ

where Aj ¼ Jensen’s alpha for unit trust j; Rjt ¼ unit trust j returns at time t ;
j ¼ beta of unit trust j; Rmt ¼ market return at time t; Rft ¼ risk free rate at time t; and "j ¼ unit trust j error term at time t. This model (equation (4)) was developed by Jensen in 1968. The Aj (intercept) represents the average incremental rate of returns on the portfolio per unit of time. Aj will be positive if the fund managers have forecasting ability and vice versa. The model assumes random error is equal to zero and a randomly constructed portfolio is expected to have zero Aj. For comparison with other portfolios, Jensen’s Alpha is adjusted for systematic risk (equation (6)). Adjusted Jensen’s alpha (AJA) ¼

Aj
j

ð5Þ

Adjusted Sharpe Index. The Sharpe’s Index is given by the following formula: SI ¼

Risk premium Raj  Raf ¼ Total risk j

where Raj ¼ average monthly return of unit trust j; Raf ¼ average risk free premium; and j ¼ standard deviation of unit trust j return.

ð6Þ

Malaysian unit trust aggregate performance 107

MF 33,2

Sharpe Index (equation (6)) can be defined as a ratio of the mean excess return over the standard deviation of unit trust j. This model uses total risk instead of systematic risk employed by the Jensen measure. Miller and Gehr (1978) find that the model produces biased outcome. Jobson and Karkie (1981) correct the Sharpe Index by adding the number of return intervals (K) in the model. This corrected model is better known as the ASI (equation (7)).

108 ASI ¼

ðSIÞK K þ 0:75

ð7Þ

where K ¼ number of return intervals in the evaluation periods; and SI ¼ Sharpe Index. Treynor Index. This model differs from Sharpe Index because Treynor Index (TI) uses systematic risk instead of standard deviation of unit trusts return. Treynor Index is defined as a ratio of mean excess returns over systematic risk of unit trust (equation (8)). TI ¼

ðRj  Rft Þ
j

ð8Þ

where Rj ¼ average monthly return of unit trust j, Rft ¼ average rate of return of risk-free asset; and
j ¼ systematic risk of unit trust j. Persistency tests. Besides determining whether a particular fund is able to perform better than the market portfolio for a given level of risk, it is of great consequence for investors to know whether a good performing fund can continue to repeat its superior performance over time. In response, we rank the unit trusts using the ASI on an annual basis for the period 1990-2001. The Spearman–Rank correlation coefficient (Rs) as shown in equation (9) is used to determine the consistency of performance over time. The test of significance of Rs is then performed by using the t-statistic given by equation (10). P 6 d2 ð9Þ Rs ¼ 1  nðn2  1Þ To ascertain whether the funds have stable rankings of beta values over time, the Spearman Rank correlation coefficients together with the t-statistic are calculated for each pair of years during the period 1990-2001. Equations (9) and (10) are used. T¼

Rs ðn  2Þ0:5 ð1  Rs2 Þ0:5

ð10Þ

To allow for a thorough investigation of unit trust performance during different economic conditions, we first examine the issue for the overall period, 1990-2001.

We then partition the data into several sub-periods which is the bull or before the crisis period (1990-1996), during the crisis period (1997-1998) and after the crisis period (19992001). The examination is not quite complete if we do not consider the persistency of performance. To carry out the tests, one would need data that covers a long span of period. This leads us to having another set of data, which consists of companies that have been in existence since 1990. For lack of better term, we name this set of data as complete data sample. Consistency test Different performance measures may provide different performance rankings. Since our study uses many performance measures to study funds performance, it may be unclear or even confusing if different measures produces different performance rankings. In this section we perform a non-parametric consistency test for the consistency in the rankings based on different benchmarks. The statistics used is the Kendall’s coefficient of concordance (W ), where W expresses the degree of association among a number of sets of rankings. Kendall’s W statistic ranges between 0 (no agreement) and 1 (perfect agreement). Under the null hypothesis of no consistency in rankings the W statistic would be zero or insignificant. If W is significant, there exists an agreement in the rankings produced by different measures. Results Analysis of raw and market adjusted returns Table III summarizes the descriptive statistics of unit trust funds and its market adjusted returns, market returns (KLCI) and risk-free returns (KLIBOR) over various sub-periods of the study. Statistical tests of significance of the return performance are presented in Table IV. In this sense the two tables should be read together. Panel (a) of Table III shows mean returns of unit trusts, risk-free asset and market index over the entire 12-year period of study, 1990-2001. It could be expected that the result in this period may be greatly influenced by the 1997-98 financial crisis. It can be seen that risk-free asset provides the highest return compared to unit trusts or the market index. Market risk premium is a negative 0.53 per month or 6.36 per annum, highly significant at 1 per cent level. Average return on unit trust is 0.18 per cent per month, or 2.16 per cent per annum. This is worse than average market return of 0.08 per month or 0.96 per year, but the difference, however, is not significant. The total risk as measured by standard deviation of unit trust returns is less than the market. Panel (b) shows return performance over the sub-period before the financial crisis. It shows that unit trusts on average yield a positive return of 0.84 per cent per month or 10.08 per cent per year. But this is still less than the market return of 0.99 per month or 11.88 per cent per year. The negative market adjusted return is significant at 5 per cent level. But again, the standard deviation of unit trust returns is smaller than that of the market. Panel (c) shows the return performance during the crisis years of 1997-98. In this sub-period, risk-free asset performs the best. Unit trusts show better performance than the market. The market adjusted return is 0.81 per month or 9.72 per year that is significant at 1 per cent level. During this period the market risk premium is a huge 45 per cent per year. Unit trust performance after the crisis is shown in Panel (d). In the aftermath of the crisis, it seems that unit trusts are still trailing the market, giving a significant negative adjusted return. Panel (e) considers only the funds that have complete data from the

Malaysian unit trust aggregate performance 109

MF 33,2

110

(a) Overall data (1990-2001) Unit trust (raw return) UT market adjusted return Market (KLCI) Risk free (b) Before crisis (1990-1996) Unit trust UT market adjusted return Market (KLCI) Risk free (c) During crisis (1997-1998) Unit trust UT market adjusted return Market (KLCI) Risk free (d) After crisis (1999-2001) Unit trust UT market adjusted return Market (KLCI) Risk free (e) Complete data (1990-2001) Unit trust UT market adjusted return Market (KLCI) Risk free

Table III. Descriptive statistics of unit trusts, market portfolio and risk free returns

N

Minimum

Maximum

Mean

S.D.

8,374

1.5591 1.6263 0.3939 0.0011

0.5664 0.4984 0.3626 0.0091

0.0018 0.0010 0.0008 0.0045

0.0827 0.0685 0.1031 0.0022

3,072

0.2975 0.3008 0.1639 0.0036

0.2973 0.1933 0.2367 0.0069

0.0084 0.0015 0.0099 0.0056

0.0548 0.0340 0.0627 0.0009

1,692

0.4661 0.3626 0.3939 0.0052

0.3735 0.4047 0.3626 0.0091

0.0224 0.0081 0.0305 0.0072

0.1145 0.1093 0.1712 0.0013

3,610

1.5591 1.6263 0.1153 0.0011

0.5664 0.4984 0.2936 0.0051

0.0008 0.0048 0.0040 0.0023

0.0830 0.0651 0.0841 0.0008

3,575

0.9217 0.9562 0.3939 0.0011

0.5508 0.4984 0.3626 0.0091

0.0004 0.0019 0.0015 0.0051

0.0827 0.0589 0.0965 0.0019

Notes: Kuala Lumpur Inter Bank Offer Rate (KLIBOR) and Kuala Lumpur Composite Index (KLCI) are used as proxies for risk free and market portfolio, respectively. Net Assets Value (NAV) of unit trust at the end of each month is used to measure returns. Monthly continuous compounded rate of returns equal to Log NAV at the end of month t plus dividend at month t and divide by NAV at the end of month t  1 [Rj ¼ loge((NAVjt þ Djt)/NAVjt  1)]. As for risk free the used the following formula to calculate monthly rate of returns, Rfm ¼ loge[(1 þ KLIBOR)/12]

beginning until the end of the period of study. Here again, as in most of the sub-period analysis, return on unit trusts is less than the market return, giving an overall negative adjusted return of 0.19 per cent per month or 2.28 per cent per year. It can be concluded that initial analysis of returns indicates that unit trust provides, on average, a negative raw returns to unit holders over the period 1990-2001. However, over the same period, the market in general is providing a positive return. Not only the funds are not able to outperform the market, our evidence suggests they perform worse than the market. However, it may be comforting to note that the variance of unit trust monthly returns is less than that of the market. Tables V and VI show return analysed by sub periods and types of funds. Equity funds are those having equity investment of more than 50 per cent of the funds. A balance funds would be dividing investments more or less equally between equity and bonds. A bond fund is one having bond investments more than 50 per cent of the fund. Table VI presents the significance tests of the market-adjusted returns of the unit

N

Mean difference

t-statistic

8,374 8,374 8,374

0.0063 0.0010 0.0053

6.970* 1.322 4.709*

(b) Before financial crisis (1990-1996) Unit trust – risk free 3,072 Unit trust – market 3,072 Risk free – market 3,072

0.0027 0.0015 0.0042

2.776* 2.443** 3.752*

(c) During financial crisis (1997-1998) Unit trust – risk free 1,692 Unit trust – market 1,692 Risk free – market 1,692

0.0296 0.0081 0.0378

(d) After financial crisis (1999-2001) Unit trust – risk free 3,610 Unit trust – market 3,610 Risk free – market 3,610

0.0031 0.0048 0.0017

2.237** 4.461* 1.247

(e) Complete data sample (1990-2001) Unit trust – risk free 3,575 Unit trust – market 3,575 Risk free – market 3,575

0.0055 0.0019 0.0036

3.941* 1.927* 2.201**

(a) Overall sample (1990-2001) Unit trust – risk free Unit trust – market Risk free – market

Malaysian unit trust aggregate performance 111

10.629* 3.063* 9.055*

Table IV. Notes: Paired T-tests were conducted to test whether unit trusts and market portfolio perform significantly different from each other. The calculation of average return is based on monthly observations for the specific period investigated. * indicates statistical significance at 1 per cent; ** indicates statistical significance at 5 per cent; *** indicates significance at 10 per cent

trusts and market excess (over the risk free rate) returns. The two tables should be read together. Table V shows that, for equity fund, both the raw returns and market-adjusted returns are negative in four out of the five sub periods, all of which are significant at either 1 per cent or 5 per cent level. The only sub period showing a positive adjusted return is during the crisis. The balance funds seem to depict similar trend, but none is significant at 5 per cent level. The bond funds, on the other hand, show more positive sub period returns than their equity and balance counter parts. This is to be expected because during the times of financial crisis, the interest on bonds remains high while equity prices plunged. However, none of the bond adjusted returns is significant at 5 per cent level. Overall results of performance measures Table VII presents various performance measures of unit trust that include raw return, market adjusted return, Jensen Alpha, Sharpe Index and Treynor Index, by sub periods and by types of funds. Again, for equity funds and balance funds, the overall sample indicate negative returns. The same is true for other sub periods and for complete data sample. Only bond funds seem to show superior performance in all sub periods. The table also shows the R-square from the regression of the fund returns against the market returns. As expected, the results show that, for all sub periods, equity funds

Paired comparison between risk free, market portfolio and unit trust

424 105 96 223 143

Bond Overall data Before crisis During crisis After crisis Complete data 0.5834 0.1335 0.1372 0.0850 0.1372

0.5834 0.1552 0.3202 0.5834 0.2982

1.5591 0.2975 0.4661 1.5591 0.9217

0.3735 0.1719 0.1801 0.0734 0.1801

0.3735 0.2357 0.3735 0.3075 0.2357

0.5664 0.2973 0.3610 0.5664 0.5508

0.0004 0.0075 0.0043 0.0035 0.0054

0.0027 0.0075 0.0168 0.0007 0.0005

0.0021 0.0085 0.0251 0.0011 0.0007

Unit trust return Maximum Mean

0.0259 0.0397 0.0278 0.0142 0.0406

0.0752 0.0477 0.1003 0.0714 0.0680

0.0858 0.0558 0.1199 0.0875 0.0847

SD

0.5411 0.1575 0.3626 0.2936 0.3344

0.5411 0.2028 0.3432 0.5411 0.2844

1.6263 0.3008 0.3626 1.6263 0.9562

Minimum

0.4047 0.1275 0.4047 0.1332 0.3939

0.3939 0.1099 0.3939 0.1023 0.3596

0.4984 0.1933 0.4043 0.4984 0.4984

0.0030 0.0045 0.0361 0.0014 0.0039

0.0003 0.0021 0.0127 0.0059 0.0010

0.0016 0.0013 0.0055 0.0052 0.0022

Market adjusted return Maximum Mean

0.1014 0.0461 0.1630 0.0823 0.0851

0.0777 0.0337 0.1242 0.0619 0.0598

0.0647 0.0335 0.1018 0.0641 0.0575

SD

Notes: Unit trust return is monthly continuous compounded rate of returns equal to log NAV at the end of month t plus dividend at month t and divide by NAV at the end of month t  1 [Rj ¼ loge((NAVjt þ Djt)/NAVjt  1)]. Unit trust market adjusted returns is computed by subtracting market return (KLSE CI) from the unit trust actual return. The calculation of average return is based on monthly observations for the specific period investigated

860 221 212 427 3,432

Balance Overall data Before crisis During crisis After crisis Complete data

Table V. Unit trusts summary of descriptive statistics by types of fund

7,090 2,746 1,384 2,960 3,289

Minimum

112

Equity Overall data Before crisis During crisis After crisis Complete data

N

MF 33,2

N

Mean difference

t-statistic

0.0066 0.0160 0.0070 0.0003 0.0005 0.0078

6.506* 2.078** 2.711* 0.121 0.361 1.586

0.0029 0.0013 0.0019 0.0021 0.0018 0.0045

2.676* 2.098** 0.578 0.914 0.464 0.997

0.0057 0.0022 0.0046 0.0010 0.0003 0.0039

3.889* 2.183** 0.805 0.204 0.090 0.543

0.0323 0.0055 0.0240 0.0127 0.0029 0.0361

10.020* 2.012** 3.482* 1.488 1.018 2.170**

Overall data (1990-2001) Equity Unit trust Unit trust Balance Unit trust Unit trust Bond Unit trust Unit trust

– – – – – –

risk free market risk free market risk free market

7,090

Before crisis (1990-1996) Equity Unit trust Unit trust Balance Unit trust Unit trust Bond Unit trust Unit trust

– – – – – –

risk free market risk free market risk free market

2,746

Complete data (1990-2001) Equity Unit trust – Unit trust – Balance Unit trust – Unit trust – Bond Unit trust – Unit trust –

risk free market risk free market risk free market

3,289

During crisis (1997-1998) Equity Unit trust Unit trust Balance Unit trust Unit trust Bond Unit trust Unit trust

– – – – – –

risk free market risk free market risk free market

1,384

After crisis (1999-2001) Equity Unit trust Unit trust Balance Unit trust Unit trust Bond Unit trust Unit trust

– – – – – –

risk free market risk free market risk free market

2,960

860 424

221 105

143 143

212 96

427 223

0.0034 0.0051 0.0031 0.0059 0.0013 0.0014

Malaysian unit trust aggregate performance 113

2.128** 4.376* 0.892 1.961*** 1.338 0.258

Notes: Paired T-tests were conducted to test whether unit trusts and market portfolio perform significantly different from each other. The calculation of average return is based on monthly observations for the specific period investigated. * indicates statistical significance at 1 per cent; ** indicates statistical significance at 5 per cent; *** indicates statistical significance at 10 per cent

have the highest R-square and bond funds have the lowest. This logically implies that equity funds are the most diversified, followed by the balanced funds and then the bond funds. Table VIII shows unit trust performance against the market portfolio. The beta value for unit trust ranges from a low of 0.5023 during the crisis years to a high of 0.6677 during the years after the crisis. The complete data sample shows a beta value of 0.6812. Our results in this table are consistent with the previous tables in terms

Table VI. Paired comparison between types of unit trust, risk free and market portfolio

Table VII. Unit trusts performance by types of funds and various performance measures 0.0024 0.0045 0.0105 0.0062 0.0155 0.0361 0.0044 0.0057 0.0024 0.0022 0.0010 0.0039

Before crisis (1990-1996) Equity 51 0.6883 Balance 7 0.4781 Bond 4 0.3114

During crisis (1997-1998) Equity 63 0.7079 Balance 11 0.6687 Bond 4 0.0909

After crisis (1999-2001) Equity 91 0.6688 Balance 12 0.6213 Bond 7 0.0210

Complete data (1990-2001) Equity 23 0.6747 Balance 1 0.6286 Bond 1 0.2354 0.0007 0.0005 0.0054

0.0012 0.0007 0.0035

0.0217 0.0095 0.0043

0.0082 0.0093 0.0083

0.0026 0.0026 0.0044

UT return

0.0032 0.0026 0.0010

0.0043 0.0040 0.0013

0.0080 0.0023 0.0013

0.0010 0.0002 0.0021

0.0029 0.0035 0.0011

Jensen’s alpha

0.0046 0.0047 0.0051

0.0023 0.0071 0.1027

0.0045 0.0337 0.1550

0.0013 0.0013 0.0286

0.0008 0.0080 0.0210

Adjusted Jensen’s alpha

0.0081 0.0082 0.0015

0.0017 0.0038 0.1036

0.0567 0.0042 03518

0.0037 0.0094 0.0423

0.0933 0.0189 2.0701

Treynor Index

0.0671 0.0674 0.0075

0.0331 0.0233 0.1125

0.2247 0.0688 0.1544

0.0546 0.1812 0.3028

0.0998 0.1125 0.4032

Sharpe Index

0.0667 0.0671 0.0075

0.0324 0.0228 0.1096

0.2200 0.0756 0.1497

0.0529 0.1608 0.2714

0.0942 0.1109 0.3945

ASI

Notes: Sharpe Index (SI) defined as average excess returns divided by unit trust j standard deviation [(Rj  Rft)/j]. Adjusted Sharpe Index (ASI) is SI corrected for the number of observation [ASI ¼ (SI  K)/(K þ 0.75)]. Treynor Index (TI) used the systematic risk the equation [TI ¼ ((Rj  Rft)/
j)]. Adjusted Jensen’s alpha (AJAI) is the ratio of Jensen’s alpha () over systematic risk (AJAI ¼ j/
j). All monthly returns are pooled before an average return for each unit trust is computed (pooled monthly by unit trusts)

0.0012 0.0005 0.0084

UT market adjusted return

Overall data (1990-2001) Equity 91 0.6536 Balance 12 0.5653 Bond 7 0.0542

R2

114

N

MF 33,2

Beta

R2

Jensen’s alpha

Adjusted Jensen’s alpha

Treynor Index

Sharpe Index

ASI

0.0027 0 36.36

0.0011 0 35.45

0.2110 0.0055 24.55

0.1205 0.0564 28.18

0.1151 0.0554 29.09

Overall data (1990-2001) Unit trusts 0.5804 Market portfolio 1 % of funds > market

0.6058 1

Before crisis (1990-1996) Unit trusts 0.6260 Market portfolio 1 % of funds > market

0.6402 1

0.0007 0 46.77

0.0009 0 45.16

0.0068 0.0058 45.16

0.0849 0.1366 37.10

0.0792 0.1251 37.10

During crisis (1997-1998) Unit trusts 0.5023 Market portfolio 1 % of funds > market

0.6707 1

0.0069 0 23.08

0.0005 0 24.36

0.0633 0.0350 23.08

0.1991 0.2036 32.05

0.1960 0.1973 32.05

After crisis (1999-2001) Unit trusts 0.6677 Market portfolio 1 % of funds > market

0.6224 1

0.0039 0 32.73

0.0038 0 31.82

0.0048 0.0011 31.82

0.0228 0.0083 33.64

0.0223 0.0084 33.64

Complete data (1990-2001) Unit trusts 0.6812 Market portfolio 1 % of funds > market

0.6553 1

0.0030 0 32.00

0.0042 0 32.00

0.0077 0.0036 32.00

0.0641 0.0368 32.00

0.0638 0.0366 32.00

Notes: Beta represents the level of portfolio risk while R-square indicates the level of diversification. Sharpe index (SI) defined as average excess returns divided by unit trust j standard deviation [(Rj  Rft)/j]. Adjusted Sharpe Index (ASI) is a revised version of SI after corrected for number of observation [ASI ¼ (SI  K)/(K þ 0.75)]. Treynor Index (TI) on the other hand used systematic risk instead total risk in the equation [TI ¼ ((Rj  Rft)/
j)]. While adjusted Jensen’s alpha (AJAI) is simply a ratio of Jensen’s alpha () over systematic risk (AJAI ¼ j/
j)

of a proliferation of negative performance measures. Table VII also shows the percentage of the funds performing above the market using the various performance measures. For the overall data, it can be seen that 36.36 per cent of the funds perform better than the market in terms of Jensen’s alpha. Averaging across measures, for the overall sample, 30.73 per cent of the unit trust out-performs the market. Before the crisis, unit trusts perform relatively better than other sub periods, but when compared to the market benchmark, only about 42.26 per cent perform better than the market. The overall percentage of unit trust outperforming the market is only 32.93 per cent or about one-third. Consistency test Consistency test was performed in order to know whether or not different performance measures used in this study provides consistent ranking of performance. It would be confusing if one measure yields a ranking that is completely different from another. We use a non-parametric test known as Kendall’s coefficient of concordance (W-statistic). W-statistic ranges between 0 (no correlation of ranking) to 1.0 (perfect agreement). Table IX shows results of consistency test. It shows that for all years, except for 1998,

Malaysian unit trust aggregate performance 115

Table VIII. Unit trusts performance against market performance

Table IX. Consistency evaluation of unit trusts performance among all performance measures 2.70 2.42 3.84 3.00 3.04 11.359* 0.114**

4.42 4.30

2.76 1.40

2.12 71.126* 0.711*

1991

2.96 17.152* 0.172*

3.92 2.92

3.12 2.08

1992

4.00 95.200* 0.952*

3.00 5.00

1.40 1.60

1993

2.16 71.911* 0.719*

2.84 1.32

4.38 4.30

1994

2.32 51.149* 0.511*

2.72 1.64

4.24 4.08

1995

4.00 93.856* 0.939*

2.96 5.00

1.48 1.56

1996

2.00 95.703* 0.957*

3.00 1.00

4.66 4.34

1997

2.74 22.068 0.221

2.20 2.78

3.08 4.20

1998

4.00 94.084* 0.941*

2.96 5.00

1.46 1.58

1999

116 2.00 95.479* 0.955*

3.00 1.00

4.38 4.62

2000

2.56 23.040* 0.230*

3.82 2.12

3.76 2.74

2001

Notes: The consistency of unit trust performance based on five different measures is tested using Kendall’s coefficient of concordance (W) which ranges between 0 (no agreement) to 1 (perfect agreement). This test is conducted on a complete data sample of 25 unit trusts that have complete data set from 1990-2001. The consistency of ranking produced under each of the performance measure is measured through W. All yearly returns throughout the specific period investigated are pooled before an average return for each unit trust is computed (pooled yearly by unit trusts). * indicates statistical significance at 0.01 level; ** indicates statistical significance at 0.05 level; *** indicates statistical significance at 0.1 level

Jensen alpha Adjacent Jensen alpha Treynor Index Sharpe Index Adjacent Sharpe Index Chi-square Kendall’s W

1990

MF 33,2

W-statistic is significant at 1 per cent level. The test indicates that the ranking provided by different measures are significantly consistent. Persistency of fund performance Studies by Hendericks et al. (1993), and Brown and Goetzmann (1995) find evidence of short-term persistence in mutual fund performance. Thus, it is of great interest to find out whether such persistence exists in the Malaysian unit trusts. For this analysis we use the complete data sample, the 25 unit trusts that have continuous monthly data from the beginning until the end of our study period. Spearman non-parametric rank order test is used to examine the correlation of the current year’s ranking with the ranking of the previous year. Table X shows the results of the one-period rank order correlation. For the year on year correlation, it seems that for the yearly ranking of 1990 and 1991, four out of the five performance measures indicate significant relationship in the rank ordering at the 5 per cent level. However, the correlation is negative, indicating that the ranking is related in the opposite direction. This means that if a fund does well in 1990 it will not do well in 1991. For all other years, the correlation is not significant. The bottom part of Table X shows the correlation between sub periods. It shows that largely the correlation is insignificant. The results from this table indicate very clearly that the hot-hand hypothesis is not supported by the Malaysian data. Investors should, therefore, be wary of using historical performance to decide on their unit trust investments.

Yearly ranking 1990 and 1991 1991 and 1992 1992 and 1993 1993 and 1994 1994 and 1995 1995 and 1996 1996 and 1997 1997 and 1998 1998 and 1999 1999 and 2000 2000 and 2001 Mean yearly ranking 1990-1996 and 1997-1998 1997-1998 and 1999-2001 1990-1996 and 1997-2001

Jensen’s alpha

Adjusted Jensen

Treynor Index

Sharpe Index

ASI

0.453* 0.223 0.008 0.010 0.158 0.389 0.509 0.213 0.252 0.310 0.055

0.383 119 150 0.010 0.152 0.144 0.471 0.077 0.116 0.048 010

0.454* 0.082 0.097 0.114 0.083 0.185 0.514* 0.079 0.133 0.074 0.109

0.426* 0.179 0.013 0.156 0.070 0.312 0.371 0.304 0.168 0.125 0.204

0.428* 0.179 0.013 0.156 0.070 0.312 0.371 0.304 0.168 0.125 0.204

0.240 0.169 0.176

0.356 0.187 0.073

0.353 0.187 0.083

0.521** 0.298 0.258

0.225 0.182 0.141

Notes: The persistency of unit trust performance from year to year based on five different measures is tested using Spearman Rank Correlation, which ranges between 1 (perfect disagreement) to 1 (perfect agreement). This test is conducted on a complete data sample of 25 unit trusts that have complete data set from 1990 to 2001. All yearly returns throughout the specific period investigated are pooled before an average return for each unit trust is computed (pooled yearly by unit trusts). * indicates statistical significance at 0.05 level; ** indicates statistical significance at 0.01 level; *** indicates statistical significance at 0.1 level

Malaysian unit trust aggregate performance 117

Table X. Persistence in unit trusts performance based on Spearman Rank Correlation

MF 33,2

118

Table XI. Summary of results

Discussion Our study employed seven different performance measures that include: raw return, market adjusted return, Jensen’s alpha, adjusted Jensen’s alpha, Sharpe Index, ASI and Treynor Index. We analyze not only the whole period sample, but also shorter sub period to capture the dynamics of the funds as we travel through time from the 1991 to 2001. It should also be noted that this period consists various sub periods with different economic conditions. In the beginning it was a period of high growth and very bullish stock market (1991-1993). This is followed by a period of ‘‘normal’’ economic growth (1994-1996). Then the country experiences a severe financial crisis in the year 19971998. This is then followed by recovery years of 1999-2001. Because of the different short-term characteristics of the economic situation, our results may have been strongly influenced by the severe financial crisis. Hence extreme caution needs to be exercised in interpreting the results. Table XI summarizes our major findings. Our statistical results indicate that unit trusts have not performed well over the period of study. The average raw return was positive for the sub period before the crisis. It was negative for all other sub periods and other sub samples. The market adjusted returns, on the other hand yields positive results only during the crisis, and this is due to the fact that market experienced huge loss in value. In all other sub periods and sub samples, market adjusted returns are all negative. We may therefore conclude that as far as returns are concern, unit trust returns are no better than the market over the period of our study. But when analyze by types of funds we find that equity and balance funds behave in a similar fashion as described above. The bond funds however, show very superior performance, over and above the market and equity unit trusts. This is of course due to the high interest rate kept during the crisis period and this has tremendously benefited the bond funds. Table XI also shows three performance risk adjusted performance measures: Jensen’s alpha, Treynor Index and Sharpe Index. The table shows that Jensen’s alpha are all negative for all sub periods and sub samples, while the Treynor and Sharpe indices are all negative except for the sub period before the crisis. The inferior performance of the risk adjusted performance measures seems to be in complete agreement with returns analysis. The table also indicates, on average, that less than a third of the funds perform above market. The low beta values indicate that funds may be less aggressive in forming their portfolios. Beta values between 50 and 68 seem to show rather conservative attitude in investing. Needles to say, our findings are not very encouraging in terms of promoting unit trust investments. It would be unwise to parade our results to the public and advise investors to shun unit trusts. Yet, it is hard to reconcile the fact that professional

Raw return Market adjusted return Jensen’s alpha Treynor Index Sharpe Index % above market Beta R-square

Before crisis

During crisis

After crisis

All data

Complete data

þ 

 þ

 

 

 

 þ þ 42.26 0.63 0.64

   26.92 0.50 0.67

   32.73 0.67 0.62

   30.73 0.58 0.61

   32.00 0.68 0.66

managers entrusted to manage public’s money, and having all the information at their disposal, turn in results that are inferior to the general market performance. It seems the managers should be well advised to just manage index funds. We believe that the issue of inferior performance of unit trust needs further investigations. One area of investigation would be to recalculate returns by including dividend distributions. It is known that some unit trusts in Malaysia place great importance on maintaining consistent dividend distribution. Dividend yields may become very important during the crisis years and its aftermath when the funds’ NAV have greatly shrunk. A second area of investigation is to segregate well-managed funds from ill-managed funds. A casual survey of management companies in the 80s revealed that some of the funds are stand-alone companies. It was later made mandatory that unit trust management companies need to be subsidiaries of banking or financial institutions. It may well be during the period of study, there exist ill managed funds due to lack of resources and expertise of the managers. They operated on limited budget, were forced to hire less qualified managers, possessed limited information and used less sophisticated software. When averaged out these poorly run funds may have negatively influenced the statistical analysis. A third issue for investigation would be the portfolio management of the funds. This involves whether or not there exists an investment policy for each funds, and if there is, whether the policy is strictly adhered to. The issue may be especially important for small funds with limited resources and expertise. Lastly, it would be interesting to see if the results are sensitive towards the different return of benchmarks for each different type of mutual fund. This could also help to explain the superiority of bond funds performance against other types of funds. As for the hot hand hypothesis, our results reveal non-existence of meaningful correlation between current performance and past performance. This is clearly shown by the non-significance of correlation coefficients between the two successive periods of performance ordering, using various performance measures. A second persistency test uses the multiple regression analysis of current performance against its lag values yield similar results. Our interpretation of these results is as follows. Persistency tests are derived from the performance test. Since our performance test needs further refinements as mentioned above, the conclusions of persistency test using our performance measures in this study may be taken as only tentative. Conclusion This paper focuses on examining unit trust performance in Malaysia over the period 1991 until 2001. Performance is analyzed from two main perspectives. The first perspective is investigating return performance of unit trust and measuring it against an appropriate benchmark. The second perspective is testing whether or not there is significant intertemporal correlation of performance. This is commonly referred to as testing the hot-hand hypothesis in mutual fund literature. Our findings show that unit trusts have not performed well over the period of study. In most of the instances, unit trust trail behind the performance of the market portfolio except during the crisis period when the unit trust market adjusted returns yields positive results. This is due to the fact that market experienced huge loss in value. The conclusion remains even after taking into account the risk adjusted performance measures. Of the various types of unit trust under investigation, we find the bond funds to show very superior performance over and above the market and equity unit trusts. The high interest rate kept throughout majority of the period particularly during the crisis period has tremendously benefited the bond funds. For the second

Malaysian unit trust aggregate performance 119

MF 33,2

120

part of the study, we find no meaningful inter-temporal correlation between current performance and past performance. Since the persistency tests are derived from the performance test, we realize that the offered conclusion is necessarily tentative pending further performance test refinements. References Brown, S.J. and Goetzmann, W.N. (1995), ‘‘Performance persistence’’, Journal of Finance, Vol. 50, pp. 679-98. Chang, E.C. and Lewellen, W.G. (1984), ‘‘Market timing and mutual fund investment performance’’, Journal of Business, Vol. 57, pp. 57-72. Chua, C.P. (1985), ‘‘The investment performance of unit trusts in Malaysia’’, unpublished MBA thesis, Faculty of Economics and Administration, University of Malaya, Kuala Lumpur. Ewe, S.J. (1994), ‘‘The performance of Malaysian unit trusts in the period 1988-1992’’, unpublished MBA thesis, School of Management, Universiti Sains Malaysia. Firth, M. (1977), ‘‘The investment performance of unit trusts in the period 1965-1975’’, Journal of Political Economy, Vol. 81, pp. 607-36. Goetzmann, W.N. and Ibbotson, R. (1994), ‘‘Do winners repeat? Patterns in mutual fund behavior’’, Journal of Portfolio Management, Winter, pp. 9-18. Grinblatt, M. and Titman, S. (1992), ‘‘The persistence of mutual fund performance’’, Journal of Finance, Vol. 42, pp. 1977-84. Gruber, M. (1996), ‘‘Another puzzle: the growth in actively managed mutual funds’’, Journal of Finance, Vol. 51, pp. 783-807. Hallahan, T.A. (1997), ‘‘Persistence in the fund portfolio performance and the information content of portfolio performance history: an examination of rollover funds’’, unpublished MBA thesis, Royal Melbourne Institute of Technology. Hendericks, D., Patel, J. and Zeckhauser, R. (1993), ‘‘Hot hands in mutual funds short-run persistence of relative performance (1974-1988)’’, Journal of Finance, Vol. 48, pp. 93-130. Henriksson, R.T. (1984), ‘‘Market timing and mutual fund performance: an empirical investigation’’, Journal of Business, Vol. 57, pp. 73-96. Ippolito, R.A. (1989), ‘‘Efficiency with costly information: a study of mutual fund performance, 1965-1984’’, Quarterly Journal of Economics, Vol. CIV, pp.1-23. Jensen, M.C. (1968), ‘‘The performance of mutual funds in the period 1945-1964’’, Journal of Finance, Vol. 23, pp. 389-416. Jobson, J.D. and Karkie, B.M. (1981), ‘‘Performance hypothesis testing with the Sharpe and Treynor measures’’, Journal of Finance, Vol. 34. Malkiel, B. (1995), ‘‘Returns from investing in equity mutual funds 1971-1991’’, Journal of Finance, Vol. 50, pp. 549-73. Miller, R.E. and Gehr, A.K. (1978), ‘‘Sample size bias and sharpe’s performance measure: a note’’, Journal of Financial and Quantitative Analysis, Vol. 13 No. 5. Ong, K.W. (2000), ‘‘Performance of unit trust in Malaysia in the period of 1995-1999’’, unpublished MBA thesis, University of Malaya. Pleschiutschnig, G.A. (1999), ‘‘Performance persistence: evidence for the European mutual fund market’’, unpublished PhD thesis, University of St. Gallen. Robson, G.N. (1986), ‘‘The investment performance of unit trusts and mutual funds in Australia for the period 1969-1978’’, Accounting and Finance, Vol. 26, pp. 55-79. Shamsher, M. and Annuar, M.N. (1995), ‘‘The performance of unit trusts in Malaysia: some evidence’’, Capital Markets Review, Vol. 3 No. 2.

Sharpe, W.F. (1966), ‘‘Mutual fund performance’’, Journal of Business, Vol. 39, pp. 119-38. Taib, F., Shahnon. S. and Lee, H.L. (2002), ‘‘Malaysian unit trusts performance’’, Proceeding of 4th Annual Malaysian Finance Association Symposium 2001, Penang, pp. 451-74. Tan H.C. (1995), ‘‘The investment performance of unit trust funds in Malaysia’’, unpublished MBA thesis, University of Malaya. Vos, E.d., Brown, P. and Christie, S. (1995), ‘‘A test of persistence in the performance of New Zealand and Australia equity mutual funds’’, Accounting Research Journal, pp. 19-34. Wermers, R. (1996), ‘‘Momentum investment strategies of mutual funds, performance, persistence, and survivorship bias’’, working paper, Graduate School of Business and Administration, University of Colorado. Further reading Koh, F. and Koh S.K. (1987), ‘‘An empirical analysis of the performance of unit trusts in Singapore’’, Securities Industry Review, Vol. 13 No. 2, October. Lee, H.H. (2000), ‘‘Performance of unit trusts’’, unpublished MBA thesis, School of Management, Universiti Sains Malaysia. Sharpe, W.F. (1964), ‘‘Capital asset prices: a theory of market equilibrium under condition of ris’’, Journal of Finance, Vol. 16, pp. 425-138. Corresponding author Fauziah Md. Taib can be contacted at: [email protected]

To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

Malaysian unit trust aggregate performance 121

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0307-4358.htm

MF 33,2

An integrated framework for style analysis: how is it useful to Malaysian equity trust investors? Wee-Yeap Lau

122

Graduate School of Economics, Osaka University, Osaka, Japan and Faculty of Economics and Administration, University of Malaya, Kuala Lumpur, Malaysia Abstract Purpose – The Asian financial crisis revealed the weaknesses of Malaysian fund management industry. The financial losses incurred in the crisis brought forward the issue of asymmetric information between the fund managers and investors i.e. the mismatch between the investors’ investment objectives and funds’ profiles. This paper aims to demonstrate the usefulness of the new framework in mitigating the above problem. Design/methodology/approach – This paper advocates an integrated framework of style analysis, i.e. using strong-form of style analysis by Sharpe – a multi-indexed benchmark for estimating the investment style of the unit trust funds, and using the weak-form of style analysis by Amenc, Sfeir and Martellini, to measure the risk-adjusted performance of the funds with alpha and selection return. Findings – This study concludes that: First, the inclusion of asset classes with negative correlation coefficient enhances the performance of funds. Second, funds with relatively high degree of style (above 70 per cent) that hold large-cap stocks together with high portion of liquid asset class (6-35 per cent) tend to have higher alpha, translating into higher information ratio. Third, index funds have the lowest information ratio, implying that these funds are not actively managed compared with others. Research limitations/implications – This paper discovers a possible trend of misclassification as the degree of styles for index funds does not differ from the other fund types. Future researchers can look into the issue of misclassification of fund objectives for the existing unit trust funds. Practical implications – This paper highlights the importance of the equity style management in bridging the gap between emerging capital markets and developed markets. One of the conclusions highlighted in the paper is the creation of new indices for different asset classes. Asset management companies are also advised to improve their disclosure in their annual reports to mitigate the issue of asymmetric information between fund managers and investors. Originality/value – To the author’s knowledge, this empirical work using integrated framework style analysis is the first of its kind on Malaysian unit trust funds. This paper is particularly useful to regulators of emerging capital markets and asset management companies. Keywords Unit trusts, Fund management, Equity capital, Assets management, Malaysia Paper type Research paper

Managerial Finance Vol. 33 No. 2, 2007 pp. 122-141 # Emerald Group Publishing Limited 0307-4358 DOI 10.1108/03074350710715845

Introduction Unit trust funds are investment products created by asset management companies, to pool resources from individual investors and invest in diversified portfolio of securities, with the purpose of adding value to their financial wealth in the future period. While the earliest unit trust fund was created in the newly independent Malaya around the year 1959, followed by the introduction of subsequent funds in 1966, 1967 and 1977, the development of fund management industry in 60s and 70s had been retarded, due to lack of push and pull factors from institutional settings. The launching of Amanah Saham National (ASN) unit trust funds in 1981 has provided an impetus for new growth in fund management industry (Shamser et al., 1995). As at 31 December 1997, there were 57 private funds and 27 government-sponsored funds with RM34

billion of assets. In addition, the net asset value (NAV) of all the unit trust funds constituted about nine percent of the total market capitalization (Federation of Malaysian Unit Trust Manager, 1998). The entire fund management industry is relatively young by its size, NAV and product development compared to other developed markets. In the aftermath of the Asian financial crisis in 1997, the market capitalization of Kuala Lumpur Stock Exchange ( KLSE) lost 53.42 per cent when comparing the 1997 and 1996 Year Ends ( Figure 1). In the same period, the size of the unit trust fund industry was reduced from RM60 billion to RM34 billion, or a loss of 44.01 per cent of its NAV. It was not until in 2002 where the NAV managed to resume to RM54 billion (Table I), but still, thousands of investors suffered financial losses and incapacitated to make important financial decision as their funds would be sold at losses during this turnaround period if they chose to. Most of NAVs of the funds were below their pre-crisis price[1]. In addition, the downward trend during the onset of Asian financial crisis and the recovery phase experienced in late 90s reveal a non-proportional relationship[2] between the NAV and the KLSE market capitalization. For instance, when KLSE made a recovery of 48 per cent in the 1999 year-end, the NAV of the unit trust industry managed to recover as much as 12 per cent. In 2000, when KLSE made a loss of 20 per cent, the NAV of the unit trust industry seemed to be stagnant at the level of 1999 year-end (Figure 1)[3]. This relationship underscores the importance of investigating asset class allocation of individual funds. Investors should be made aware of the asset allocation made by their fund managers and the exposure level permitted as stated in the respective investment policy. However, a closer examination across the annual reports of the unit trust funds reveals less than specific or ambiguous asset allocation policy[4]. Could investors unravel the asset allocations of their funds given the limited

An integrated framework for style analysis 123

Year-on-Year Changes in NAV and KLSE Market Capitalization

60.00

47.57 42.63 40.00 35.87

Year

20.00

15.37

11.70

9.35 0.09

0.00 1996

1997

–0.34 1998

1999

–20.00

2000

13.41 4.64

2001

3.58 2002

–19.60

–40.00 –44.01

–60.00

–53.42

Percentage NAV of trust funds

KLSE Market Capitalization

Sources: Author’s calculation based on PNB (2001) and Federation of Malaysia Unit Trust Managers (n.d.)

Figure 1. Year-on-year changes in NAV (unit trust funds) and KLSE market capitalization, 1996-2002

MF 33,2

124

Table I. Statistics on the Malaysian unit trust industry and KLSE

1995 Industry Units in circulation (billion units) No. of accounts (’000) NAV (RM billion) KLSE KLSE composite index Market capitalization (RM billion) NAV to market capitalization (%)

1996

1997

1998

1999

2000

2001

2002

31.94 38.94 42.25 46.54 52.63 63.85 71.39 84.53 6,850 7,964 8,263 8,588 8,910 9,582 9,990 10,175 44.13 59.96 33.57 38.73 43.26 43.30 47.35 53.70 995.17 1237.96

594.44

586.13

812.33

679.64

696.09

646.32

565.63

806.77

375.8

374.52

552.69

444.35

464.99

481.62

7.80

7.43

10.34

7.83

9.74

10.18

11.15

8.93

Sources: PNB (2001) and Federation of Malaysian Unit Trust Managers (n.d.)

information available? Alternatively, could they estimate the style of their equity funds? Albeit the existence of non-proportional relationship between NAV and the KLSE market capitalization, majority of the studies conducted with respect to the performance measurement of Malaysian unit trust funds have utilized market benchmarks either Kuala Lumpur Composite Index (KLCI) or EMAS Indices (Chua, 1985; Ewe, 1994; Shamsher and Annuar, 1995; Leong and Aw, 1997; Ch’ng and Kok, 1998). If unit trust funds were to invest in various investment vehicles, it would be more appropriate to examine the performance of mutual funds against a multi-indexed benchmark. Accordingly, not only the difference in terms of performance among the different investment styles can be observed, the possibility of mismatch between the fund objectives and the self-defined investment style by fund managers could also be detected. This paper advocates for an integrated framework of style analysis i.e. using strong-form of style analysis by Sharpe (1988; 1992) – a multi-indexed benchmark for estimating the investment style of the unit trust funds, and using the weak-form of style analysis by Amenc et al. (2002), to measure the risk-adjusted performance of the funds with risk-adjusted alpha, selection return and market timing, in addition to the verification of the style consistency and calculation of the information ratio. In the concluding section, with respect to the findings obtained from this study, this paper evaluates the implementation of style analysis and recommends ideas for further improvement in the context of an emerging capital market. Literature review Equity style classification With the advent of the concept of a fund’s ‘‘effective asset mix’’ and ‘‘attribution analysis’’ by Sharpe (1988; 1992), there have been a number of proponents for style analysis with each of them demonstrated the usefulness of this analysis with respect to equity style classification (Tierney and Winston, 1991; Bailey, 1992; Bailey and Tierney, 1993; Coggin, 1998). This analysis has also been used to link the investment returns and asset allocation policies in some of the recent research (Brinson et al., 1991; Ibbotson and Kaplan, 2000).

Tierney and Winston (1991) supported the use of return-based style analysis to analyze the asset mix of a portfolio manager. Using a four equity style portfolios produced by Wilshire Asset Management as generic portfolio for style-point analysis, they concluded that creation of a custom benchmark is the best way to address the style issue. Christopherson (1995) linked the crucial relationship among past return patterns, portfolio characteristics and future returns and pointed out that the reason for studying investment style was not so much concerned with the past returns, but to anticipate future returns. TerHorst et al. (2004) stated that while the estimated portfolio may indeed differs from actual portfolio holdings, but ‘‘ . . . if the aim is to predict future fund returns, factors exposures seem to be more relevant than actual portfolio holdings, and return-style based style analysis performs better than holding-based style-analysis’’[5]. It is inevitable for the problem of asymmetric information between fund manager and investors to exist as timely mutual fund holdings are not readily updated even in the developed market as discussed by Lucas and Riepe (1996). Furthermore, they identified style analysis to be a useful tool for investors to comprehend a trust fund’s investment policy and objective. In a number of subsequent studies, in the course of identifying a system of classification for equity trust funds, the researchers have also presented the evidence of mis-classifications if self-reported investment objectives were to be compared to the estimated styles (DiBartolomeo and Witkowski, 1997; Brown and Goetzmann, 1997; Kim et al., 2000). In one of the recent studies, Amenc et al. (2002) have proposed an integrated framework for assessing the risk-adjusted performance of mutual fund managers. This methodology is designed to be consistent with modern portfolio theory and constraints imposed by practical implementation of investment management where a variety of styles have to be accounted for. Mutual funds in Malaysia Chua (1985) with exclusive samples of 12 Malaysian mutual funds between 1974-1984, concluded that funds outperformed the market proxy and performance was fairly consistent over time. High performance funds tend to relate to those with low expense ratio, low asset size and low portfolio turnover. In a subsequent study, Ewe (1994) with sample of 37 funds and a period between 1988-1992, with test of performance by Jensen’s Alpha Measure and Sharpe Index Measure, reported that while risk adjusted returns overall were less than those of stock market implying that the managers had low forecasting ability. Shamsher and Annuar (1995) found a similar result with Ewe (1994), where the returns on investment in 54 funds for the period 1988-1992 were below risk-free and market returns. Besides the performance is inconsistent over time, the degree of diversification of the portfolios was below expectation. In addition, the studies conducted with respect to the performance measurement of Malaysian unit trust funds have utilized market benchmarks such as KLCI and EMAS Index (Leong and Aw, 1997; Ch’ng and Kok, 1998). These researchers have advocated for more than one kind of market benchmarks for performance measurement. All the prior studies before 1997 have concentrated on using the broad market index i.e. KLCI as the single yardstick. In another study by Shamsher and Annuar (2001), with a sample size of 41 nongovernment based mutual funds from 1995 to 1999, they reported that based on riskadjusted returns basis, both active and passive funds performed equally well, but

An integrated framework for style analysis 125

MF 33,2

126

underperformed the market portfolio. They concluded that choice of active or passive funds was irrelevant given equal performance, but growth funds should be prioritized over income if investors preferred actively managed funds over passive funds and vice versa. Using the return-based style analysis with a sample size of 42 funds from February 1996 to January 2001, Lau (2002) noted that, in addition to the usual market benchmark comparison, the performance of funds can also be compared against their respective peer groups. It was also noted that the level of passive management for index funds were indistinguishable from other types of fund. Data Data selection The data comprises of 60 month-end NAV of the equity funds listed on daily newspapers. NAV is selected as the measure of a unit trust fund’s value as it reflects the actual amount of funds fund managers have to invest with. Table II shows the sample of 43 funds of which can be divided into six groups of fund types. Data description As the methodology of style analysis requires at least 60 consecutive monthly returns of funds, a sample period from December 1996 through December 2001 is chosen. Dependent variables The continuous compounding return for the fund is used as the dependent variable. It is calculated as
 Pj;t Rj;t ¼ ln Pj;t
1
 Im;t Rm;t ¼ ln Im;t
1 Rf ;t ¼ lnð1 þ r f ;t Þ where Rj,t, the continuous compounded return for j unit trust fund at time t; Rm,t, the continuous compounded return for m benchmark portfolio for the month t; Rf,t, the continuous compounding risk free rate of interest for month t; Pj,t, the NAV for j unit trust fund at time t ; Im,t, the asset class index at the end of month t ; rf,t, the discount Classification of funds

Table II. Composition of sample data

Income Growth Balanced Small companies Index Federal Total

No. of funds

Percentage

25 8 4 3 2 1 43

58.1 18.6 9.3 7.0 4.7 2.3 100

Source: Federation of Malaysian Unit Trust Managers (n.d.)

rate of the 90-day T-Bill for month t as the proxy for the risk free rate of interest; and ln, the natural logarithm. Independent variables Independent variables are returns series of asset classes invested by fund managers. The asset classes that represent the investment universe are shown in Table III. These asset classes are chosen after careful examination on literatures such as Choong (2001) and fund prospectuses. Out of 43 funds in our sample, three funds that also invest in foreign stocks have 6 asset classes as their independent variables. As stated by Sharpe (1992) ‘‘ . . . while not strictly necessary, it is desirable that such asset classes should be mutually exclusive, exhaustive and have returns that ‘‘differ’’, . . . and the asset classes returns should either have low correlations with one another or, in cases in which correlations are high, different level of standard deviations’’. While style analysis in equation (1) has attempted to capture the investment universe i.e. to include all possible investment products in the model, careful consideration has been taken to ensure that asset classes chosen are not correlated to one another. However, as shown in Table IV, it is found that one pair of correlation coefficients i.e. the second board[6] and EMAS[7], has rather high correlation of 0.83. An examination on their standard deviations reveals that their respective values are different i.e. the standard deviation of EMAS Index is 12.14 per cent while second board is 17.48 per cent. As such, this fulfills the above requirement.

An integrated framework for style analysis 127

Methodology Return-based style analysis As in Sharpe (1992), this study initially introduces the generic factor model in equation (1) before adapting it into style analysis in equation (2). h i ~ i ¼ bi1 F ~1 þ bi2 F ~2 þ bik F ~k þ    þ bin F ~n þ ~ei R ð1Þ

Asset class

Description

Large capitalization stocks Medium capitalization stocks Cash

Represented by EMAS Index, an all-share index covers investment in equities listed at KLSE main board Represented by second board index, an all-share index covers investment in equities and securities listed at KLSE’s second board A proxy for short-term Ringgit money market instruments. Represented by Kuala Lumpur Inter Bank Offer Rate (KLIBOR). KLIBOR 1-month deposit rate is used Represented by MGS-bond all tenure Indexa, which account for MGS with value above RM100 million on issues for maturity greater than one year Represented by RAM Listed Bond Indexa, which account for all bonds and loan stocks listed on KLSE a term to maturity of more than one year. A proxy for listed private debt securities Represented by MSCI World Indexb. A proxy for all international stocks index

Government bonds Corporate bonds International stocks

Sources: aRating Agency Malaysia (RAM)-Quantshop, 2004; (accessed 5 May 2005)

b

available at: www.msci.com

Table III. Asset class indices

MF 33,2

128

~ i, return of fund i; F ~ k, return of factor k for fund i; bik, sensitivity of fund i to where R factor k; and ~ei, non-factor return of asset i of mean zero with the assumption that the non-factor returns are uncorrelated
eiej ¼ 0. Style analysis is the use of constrained quadratic programming for solving the asset allocation problem. This approach incorporates two specific constraints: first, the coefficients must sum to 100 per cent and second, coefficients must be positive. Negative coefficients can be interpreted as short positions in asset classes. This type of strategy is rarely used by the funds examined, and prohibiting these coefficients provides better, more usable results[8]. The factor is rewritten as ~1 þ bi2 F ~2 þ bik F ~k þ    þ bin F ~n  ~ i
½bi1 F ~ei ¼ R ð2Þ ~ i, return of fund i; F ~ k, return of factor k for fund i; and bik, where ~ei, selection; R sensitivity of fund i to factor k. To obtain the style, minimize variance of residual return ~ei Subject to constraints n X bik ¼ 1 for any fund i and asset class k and 0 < bik < 1 j¼1

With the two specific constraints, the coefficients tabulated in equation (2) will resemble the weights within a portfolio and conveniently displayed as part of the portfolio. The asset class indices in Table III which represents the factors in equation (1) and the sensitivity of each of the fund’s return series to each of the asset class index factors is used to construct a passive benchmark portfolio return series for performance measurement. In other words, the return of funds will be measured against the style-based, passive benchmark contained as second, bracketed terms in the right hand side of equation (2). Upon obtaining results from the quadratic programming in equation (2), the proportion of variance ‘‘explained’’ by the selected asset classes, for fund i can be obtained as below: R2 ¼ 1


Varð~eÞ

ð3Þ

~Þ VarðR

The second term of the right-hand side of the above equation represents the proportion of variance ‘‘unexplained’’ or due to active management (selection). In other words, the

Correlation with

Table IV. Mean, standard deviation, correlation coefficients between the returns of asset classes

Asset class

Mean

Standard deviation

EMAS SB KLIBOR MGS LBI MSCI World


5.28
1.02 0.44 0.80 2.07 0.62

12.14 17.48 0.23 1.25 13.83 4.72

FMAS

SB

KLIBOR

MGS

LBI

MSCI World

1.00 0.83
0.21 0.20 0.11 0.44

1.00
0.21 0.22 0.11 0.38

1.00
0.14
0.17 0.14

1.00 0.02
0.22

1.00 0.00

1.00

return of unit trust fund is decomposed into return on a set of asset classes and residual return. The former is attributed to style and represented by the R-square while the latter is attributed to selection. In order to take into account the added (or subtracted) value provided by a fund i.e. its benchmark and the added risk, the monthly mean selection return is divided by the standard deviation of monthly selection returns. This calculation gives a monthly selection Sharpe ratio ( MSSR) as stated in equation (4). The selection Sharpe ratio (SelSR) which denotes the valued added (subtracted) through active management per unit of added risk is the annualized MSSR, obtained by multiplying MSSR with the square root of 12 as shown in equation (5). Eð~ei Þ MSSR ¼ ð4Þ
~ei pffiffiffiffiffi SelSR ¼ MSSR  12 ð5Þ The monthly mean selection returns can be measured for its statistical significance using a t-statistic. The null hypothesis is stated as selection return equals to zero. t¼

ðrs
Þ pffiffiffi s= n

where rs, the monthly mean selection returns; , zero, the null hypothesis; s, the standard deviation of monthly selection return; and n, the number of observations. Weak-form style analysis As noted by Amenc et al. (2002), the strong-form style analysis with two specific constraints in equation (2) does not provide for an understanding of abnormal return measurement. Henceforth, the risk-adjustment can be addressed by using an index multi-factor model such as: h i ~ i
rf ¼ i þ bi1 ðF ~1
rf Þ þ bi2 ðF ~2
rf Þ þ bik ðF ~k
rf Þ þ    þ bin ðF ~ n
rf Þ þ ei R ð6Þ ~ i
rf, the excess return of fund i
3-month T-Bill yield; i, the estimated where R ~ i
rf); F ~ k
rf, the excess return (over intercept term which is not explained by (R 3-month T-Bill yield) of factor k for fund i; bik, sensitivity of fund i to factor k; and ei, residual term with mean zero. Equation (6) is the weak-form of style analysis i.e. excluding the positivity and portfolio constraints, and including a constant term in the unconstrained regression. The intercept term or alpha indicates the portion of excess return of fund i which is not explained by the passive, style-based portfolio. The alpha return is the total portion of the fund’s attributed to the active management, which includes investment strategies such as security selection, market timing, industrial sector rotation and currency management. However, Lucas and Riepe (1996) indicated that total alpha in the context of style analysis comprises of selection (i.e. value added activities by fund managers against to the customized, style-based benchmark) and market timing. The residual term is assumed to have a mean of zero and be uncorrelated with the asset classes or factor ks.

An integrated framework for style analysis 129

MF 33,2

130

In addition, the coefficient of determination of R-square indicates the proportion of trust fund’s variance ‘‘explained’’ by the weighted return of the passive, style-based portfolio of selected asset classes. The coefficient of determination provides another measure of checking the style consistency i.e. a high R-square indicates that the trust fund’s return tracks the passive, style-based portfolio of benchmark returns closely throughout the 60-month periods understudied or vice versa. Performance measurement Total alpha return is the total portion of the fund’s return that are not explained by the style analysis portfolio. The result from regression in equation (6) is used to calculate the total alpha return as below: ~ ta ¼ R ~i
R ~ sa R ð7Þ ~ i, the monthly ~ ta, the monthly total alpha return for the fund in period t; R where R ~ return of unit trust fund in period t; and Rsa, the monthly style analysis portfolio return in period t. Selection return is defined as the measure of the value added by the fund relative to its customized benchmark. Selection return is the residual risk from the unconstrained regression in equation (6). The selection return for each trust fund is defined as below: ~s ¼ R ~i
R ~b R

ð8Þ

~ i, the monthly return of ~ s, the monthly selection return of the fund in period t ; R where R ~ b, the monthly return on the customized benchmark in unit trust fund in period t ; and R period t. Market timing return is used to measure the value added through active market timing or sector rotation. It could be calculated as ~ mt ¼ R ~b
R ~ sa R

ð9Þ

~ b, the monthly ~ mt, the monthly market timing return of the fund in period t; R where R ~ sa, the monthly style analysis return on the customized benchmark in period t; and R portfolio return in period t. Alternatively, it could be derived from equations (7)-(9) that the total alpha is the sum of market timing return and selection return. Information ratio With the weak-form style analysis of unconstrained regression in equation (7), an additional performance measurement known as Information ratio (IR) could be obtained. IR is the annualized ratio of residual return to residual risk. It is the ratio of alpha to the standard deviation of residual returns, annualized[9]. IR ¼

i
ei

ð10Þ

The information ratio can be measured for its statistical significance using a t-statistic. The null hypothesis is stated as alpha or excess return equals to zero[10].

i pffiffiffiffi
ei = T IR ¼ pffiffiffiffi 1= T pffiffiffiffi ¼ T ðIRÞ

An integrated framework for style analysis

t-statistic ¼

ð11Þ

131

where T ¼ number of monthly observations. Style consistency To verify the style consistency of each these funds i.e. whether the fund managers are consistent in their strategies, this paper divides the period of returns series between January 1997 and December 2001 into two sub-periods, and testing the below hypothesis. Hypothesis null: There exists style consistency between the two sub-periods. Hypothesis alternative: There is no style consistency between the two sub-periods. Initially, the sample is regressed throughout the period from January 1997 to December 2001. Subsequently, the 60-month sample period is divided into two sub-periods of 30 months and regressed separately i.e. the 1st sub-period is from January 1997 to June 1999, and the second sub-period is from July 1999 to December 2001. The Chow test F¼

ðSSER
SSEU Þ=J SSEU =ðT
KÞ

ð12Þ

where SSER ¼ the restricted sum of squared errors; SSEU ¼ the unrestricted sum of squared errors, of which SSEu ¼ SSE1 þ SSE2; SSE1 ¼ the sum of squared residuals from the estimation of sub-period 1; SSE2 ¼ the sum of squared residuals from the estimation of sub-period 2; T ¼ total number of the sample; J ¼ degree of freedom of the numerator; and T
K ¼ degree of freedom of denominator where K ¼ 2J. Results Strong-form style analysis The results of return-based or strong-form style analysis are shown in Table V. Across the different fund types, it could be observed that income, growth and federal funds have substantial holdings in large-cap stocks (about 63-65 per cent), followed by index funds (about 61 per cent) and small companies funds (about 59 per cent). Balanced funds hold the least in large-cap stocks (about 50 per cent). This finding concurs with the notion that Malaysian growth and index fund managers hold large-cap stocks in anticipation of capital gains. On the other hand, federal fund managers hold large-cap stocks as most of the government linked companies are mainly listed in the main board.

Affin equity AM total return M Berjaya M Investment ASM 3 ASM 4 ASM 5 ASM 6 ASM 7 ASM 8 ASM 11 ASM premier ASM ptnb CT trust CT prime Mayban UT Pacific premier BSN Public savings Public growth Public industry Public regular savings RHB dynamic Premium capital TA growth Income fund

ASM dana growth SBB double growth SSB high growth HLG growth MBF growth Public aggressive growth

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

1 2 3 4 5 6

Table V. Results of the estimation: the degree of styles and selection, asset classes holdings by different funds and selection Sharpe ratio

Fund

67.13 75.26 59.49 73.65 85.39 72.27

85.65 51.41 92.31 92.34 64.15 52.42 73.43 51.01 65.47 56.41 74.29 82.40 86.43 81.10 85.89 77.99 81.56 72.68 60.80 62.25 53.22 46.08 88.14 75.67 74.60 71.51

Style

32.87 24.74 40.51 26.35 14.61 27.73

14.35 48.59 7.69 7.66 35.85 47.58 26.57 48.99 34.54 43.59 25.71 17.60 13.57 18.90 14.11 22.01 18.44 27.32 39.20 37.75 46.78 53.92 11.86 24.33 25.40 28.49

Selection

50.05 61.46 61.40 57.42 91.53 56.91

89.01 69.61 92.93 88.04 59.60 66.12 53.79 47.80 46.17 61.71 91.52 66.57 90.07 74.30 94.18 49.97 57.22 84.24 21.76 49.22 45.11 36.63 60.97 62.20 57.96 64.67 7.12 0.00 0.00 2.40 0.00 0.00

0.00 0.00 0.00 0.00 5.14 7.17 12.91 5.64 7.82 9.82 1.18 4.85 0.11 5.53 5.82 2.85 9.11 0.57 14.59 0.00 2.16 0.00 0.00 0.81 5.82 4.08

Mediumcap

42.23 0.00 20.57 5.45 0.00 16.49

10.99 0.00 5.55 11.12 8.47 26.71 0.00 13.86 0.00 0.00 0.00 0.00 0.00 0.00 0.00 24.34 27.22 14.91 62.42 50.63 37.60 39.53 26.70 35.73 0.00 15.83

KLIBOR

0.00 28.48 18.03 33.86 6.90 26.60

0.00 23.70 0.00 0.00 26.79 0.00 33.30 31.47 45.24 27.36 6.77 28.30 9.82 20.16 0.00 20.53 3.16 0.00 0.00 0.00 12.10 22.91 12.33 0.00 35.11 14.36

Govt bonds

0.59 0.08 0.00 0.87 1.57 0.00

0.00 6.70 1.52 0.84 0.00 0.00 0.00 1.22 0.77 1.10 0.52 0.28 0.00 0.00 0.00 2.31 3.30 0.28 1.23 0.15 3.03 0.93 0.00 1.25 1.11 1.06

Corp bonds

9.98

MSCI World 0.21 0.02 0.91 0.26
2.36*
1.52
1.94**
2.09*
2.54*
2.09*
0.40
2.34*
1.06
1.23
0.44
2.77***
0.86
0.54
1.78**
1.84**
1.55
1.92*
0.72
0.50
1.37
0.99
0.32 0.11
0.43
0.48
1.30


0.47
0.17 0.09
0.22
0.26
0.52

1-Statistic (sel return)

0.13 0.018 0.462 0.111
0.841
0.82
0.73
0.83
0.81
0.88
0.29
0.71
0.45
0.54
0.27
0.71
0.36
0.36
0.60
0.67
0.67
0.70
0.22
0.24
0.56
0.46

Monthly mean sel return (%)


0.13
0.04 0.01
0.06
0.06
0.17

0.03 0.002 0.12 0.03
0.30
0.20
0.25
0.27
0.33
0.27
0.05
0.30
0.14
0.16
0.06
0.36
0.11
0.07
0.23
0.24
0.20
0.25
0.09
0.06
0.18

Monthly sel Sharpe ratio

132

No

Largecap


0.44
0.14 0.05
0.19
0.22
0.58 (Continued)

0.09 0.01 0.41 0.12
1.05
0.68
0.87
0.94
1.14
0.93
0.18
1.05
0.48
0.55
0.20
1.24
0.38
0.24
0.79
0.82
0.69
0.86
0.32
0.22
0.61

Selection Sharpe ratio

MF 33,2

Style 89.28 77.39 74.98 67.36 41.61 84.92 61.17 63.76 88.10 92.39 59.11 79.87 80.21 77.66 78.94 75.28

Fund

RHB capital OSK-UOB equity Growth fund

SBB savings fund Mayban balanced MBF balanced Public balanced Balanced fund

M Progress M Equitya SBB emerging co Small companies fund

ASM 2 Public index Index fund

ASN Federal fund

24.72

19.79 22.34 21.06

11.90 7.61 40.89 20.13

32.64 58.39 16.08 38.84 36.24

10.72 22.61 25.02

Selection

63.51

71.26 50.54 60.90

54.73 90.97 30.11 58.61

45.54 23.84 91.52 35.53 49.11

66.95 62.43 63.52

0.00

0.00 0.00 0.00

11.52 0.00 20.12 10.55

0.00 0.00 0.00 0.69 0.17

0.00 0.00 1.19

Mediumcap

34.24

0.00 19.36 9.68

33.72 9.03 16.07 19.61

0.00 70.60 0.00 54.33 31.23

0.00 0.00 10.59

KLIBOR

0.00

28.74 30.09 29.41

0.00 0.00 23.68 7.89

53.81 5.56 6.91 9.45 18.93

33.05 37.36 23.04

Govt bonds

2.25

0.00 0.00 0.00

0.00 0.00 0.00 0.00

0.65 0.00 1.57 0.00 0.56

0.00 0.21 0.42

Corp bonds

10.00 3.33

0.00

0.00

1.25

MSCI World


0.04 1.16**
0.15
1.63
0.74**
1.45


0.60
0.53
0.57
0.56


1.09
2.25*
0.69
2.20*


1.09
1.26

1-Statistic (sel return)


0.01 0.63
0.11 0.17


0.43
0.63
0.34
0.63
0.51


0.33
0.50
0.30

Monthly mean sel return (%)

Selection Sharpe ratio
0.49
0.56
0.49
1.01
0.31
0.98
0.02 0.53
0.07
0.73
0.78
0.65

Monthly sel Sharpe ratio
0.14
0.16
0.14
0.29
0.09
0.28
0.005 0.15
0.02
0.21
0.23
0.19

Notes: * Denotes level of significance at 5 per cent; ** denotes level of significance at 1 per cent; *** denotes level of significance at 10 per cent level, respectively; athe number of return series of M Equity fund is 58 (from December 1996 to October 2001)

1

1 2

1 2 3

1 2 3 4

7 8

No

Largecap

An integrated framework for style analysis 133

Table V.

MF 33,2

134

Among the fund types, as the name implied, small companies funds hold the most medium-cap stocks of 11 per cent, followed by income funds (about 5 per cent) and growth funds (about 2 per cent). In addition to the notion of relatively risky asset class, medium-cap stocks are often considered as second-line stocks that are more speculative in nature. This type of stocks has the tendency to follow the sentiments of the main board. As such, it is observed that medium-cap stocks are not the main focus of growth and index fund managers. Overall, the best performance of all the funds during this period being small companies funds, which have the highest monthly mean selection return of 0.17 per cent, followed by growth funds (
0.30 per cent), income funds (
0.47 per cent), balanced funds (
0.51 per cent), federal fund (
0.56 per cent) and index funds (
0.57 per cent). It is also interesting to note that the degree of styles of index fund i.e. 79 per cent is not dissimilar with other fund types such as small company funds (around 80 per cent). It would be expected that index funds to have a relatively high degree of style and lower degree of selection given the nature of the funds. This phenomenon could be further investigated to find out if there is any mis-classification of funds in the sample. Of all the 43 funds in the sample, two funds with a relatively high selection Sharpe ratio i.e. 0.41 for M Berjaya and 0.53 for M Equity funds, have some interesting characteristics. Both have substantial large-cap stocks (around 91-93 per cent) and cash (around 6-10 per cent). Investing in money market instruments like T-bills provide for consistent income as the buffer during economic uncertainty, while investing in the large-cap stocks provides for unexpected capital gain should the market has short recoveries in the business cycle. Weak-form style analysis In Table VI, the results of the weak-form style analysis are shown. It is observed for the 5-year period from December 1996 to December 2001, federal fund has the highest alpha of 0.11, followed by small companies funds of 0.04, growth and balanced funds of
0.11, income funds of
0.13 and index fund of
0.17. As the sample period of this study collides with the period of the Asian financial crisis, only qualified conclusion on performance of mutual funds can be drawn. However, for those funds with positive alpha, seven funds hold a combination of large-cap stocks (55-93 per cent) and cash instruments (5-35 per cent) and the other two funds hold a combination of large-cap stocks and government bonds (28-36 per cent). The negative correlation coefficient between the large-cap stocks and cash provides the reason for better performance when comparing the former combination of asset classes against the latter. In terms of information ratio, federal fund has the highest information ratio of 0.63, followed by small companies funds of 0.44, income funds of
0.71, growth funds of
0.74, balanced funds of
0.80, and index funds of
1.20. Three conclusions are notable. First, our result confirms that federal fund manager’s annualized excess return over the benchmark is positive and significant. Second, small companies funds with medium-cap stocks portfolios have made excess return with short fluctuation within the economic cycles. Third, among the sample, the lowest information ratio of index funds confirm the notion that index fund managers do not exercise active management in relative frequency as compared to other funds. Funds with positive alpha also have high information ratios. Among the outstanding funds are M Equity with information ratio of 0.99, followed by ASN of 0.63, M Berjaya of 0.53 and M Progress of 0.51, with level of significance at one per cent

0.78 0.47 0.63 0.92 0.17 0.74 0.17 0.16 0.39 0.32 0.42 0.07 0.05 0.15 0.50 0.24 0.50 0.82 0.43 0.95 0.13 0.40 0.37 0.86 0.97 0.54 0.80 0.73 0.52 0.89 0.11

0.05
0.21 0.06 0.01
0.20
0.06
0.25
0.31
0.16
0.27
0.19
0.26
0.28
0.24
0.13
0.15
0.10 0.06
0.12
0.01
0.28
0.15
0.09 0.02 0.01
0.13
0.11 0.02
0.08
0.11
0.02
0.24

Fund

Affin equity AM total return M Berjaya M Investment ASM 3 ASM 4 ASM 5 ASM 6 ASM 7 ASM 8 ASM 11 ASM premier ASM ptnb CT trust CT prime Mayban UT Pacific premier BSN Public savings Public growth Public industry Public regular savings RHB dynamic Premium capital TA growth Income fund

ASM dana growth SBB double growth SSB high growth HLG growth MBF growth Public aggressive growth

No

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

1 2 3 4 5 6

p-value

Alphas (annual)

67.26 76.01 60.09 73.88 85.63 73.06

86.31 54.47 93.37 93.34 83.74 86.79 73.67 51.49 66.40 56.62 74.58 82.68 87.53 82.06 87.19 78.07 81.60 73.84 61.00 65.37 54.12 46.45 88.85 76.45 75.01 74.44

R-square

6.52* 3.99* 3.72** 2.44** 0.84 14.08*

6.18* 5.14* 6.19* 1.02 9.87* 14.59* 9.76* 1.63 4.30* 0.76 0.44 0.71 2.80** 4.19* 2.59*** 2.62** 2.38*** 2.58** 2.53*** 4.74* 2.36*** 7.80* 3.89** 3.01** 1.60

Fi,j

0.18 0.16 0.22 0.16 0.17 0.16

0.16 0.29 0.11 0.10 0.15 0.17 0.19 0.24 0.18 0.29 0.24 0.15 0.15 0.17 0.19 0.13 0.14 0.22 0.14 0.16 0.20 0.18 0.10 0.16 0.17 0.17


0.11
0.50
0.05
0.09
0.35
0.22
0.43
0.55
0.34
0.56
0.43
0.41
0.43
0.41
0.32
0.28
0.24
0.17
0.26
0.17
0.48
0.32
0.19
0.14
0.16
0.31
0.29
0.14
0.30
0.27
0.20
0.40

Selection return (annual)

Market timing (annual) t-Statistic (IR) 2.34**
5.53* 4.11* 0.90
10.51*
2.72*
10.17*
10.03*
6.69*
7.31*
6.15*
13.56*
14.14*
10.82*
5.30*
9.22*
5.31* 1.93***
6.26*
0.45
11.01*
6.44*
7.20* 0.78 0.32
4.82* 1.18
2.74*
5.10*
1.13
12.11* (Continued)

Information ratio (IR) 0.30
0.71 0.53 0.12
1.36
0.35
1.31
1.30
0.86
0.94
0.79
1.75
1.83
1.40
0.68
1.19
0.69 0.25
0.81
0.06
1.42
0.83
0.93 0.10 0.04
0.71
0.62 0.15
0.35
0.66
0.15
1.56

An integrated framework for style analysis 135

Results of weak-form style analysis: alpha coefficient, R-square, market timing, selection return, IR and F statistics

Table VI.

SBB Savings fund Mayban balanced MBF balanced Public balanced Balanced fund

M Progress M Equitya SBB Emerging Co Small companies fund

ASM 2 Public index Index fund

ASN Federal fund

1 2 3 4

1 2 3

1 2

1

0.54

0.15 0.33


0.22
0.11
0.17 0.11

0.65 0.40 0.77

0.28 0.66 0.16 0.39


0.16
0.06
0.13
0.11
0.11 0.06 0.11
0.04 0.04

0.15 0.20


0.15
0.19
0.11

p-value

82.70

81.12 80.17 80.65

88.43 93.51 60.25 80.73

67.96 42.80 86.79 62.98 65.13

90.17 78.84 75.62

R-square

1.17

5.39* 4.58*

1.97 3.59** 2.69**

2.44* 7.21* 2.56* 5.37*

3.97** 4.18*

Fi,j

0.16 0.11 0.14 0.17


0.06

0.11 0.11 0.22 0.14


0.05
0.001
0.25
0.10
0.37
0.23
0.30

0.14 0.12 0.16 0.13 0.14

0.10 0.15 0.16


0.30
0.18
0.29
0.23
0.25


0.24
0.34
0.27

Selection return (annual)

0.63


1.40
0.99
1.20

0.51 0.99
0.16 0.44


1.08
0.46
0.79
0.87
0.80


1.46
1.26
0.74

Information ratio (IR)

4.84*


10.87*
7.67*

3.94* 7.54*
1.27


8.36*
3.55*
6.13*
6.72*


11.31*
9.74*

t-Statistic (IR)

Notes: * Denotes level of significance at 10 per cent; ** denotes level of significance at 5 per cent; *** denotes level of significance at 1 per cent level; athe number of return series of M Equity fund is 58 (from December 1996 to October 2001)

RHB capital OSK-UOB Equity Growth-fund

7 8

Table VI.

Fund

Market timing (annual)

136

No

Alphas (annual)

MF 33,2

level, respectively. In addition, 34 funds exhibit the information ratios which fall within 99 per cent of confidence intervals, indicating statistical significance. Overall the coefficient of determination or R-square of 82.70 per cent for the federal fund is the highest of all, followed by small companies funds and index funds (80-81 per cent), growth funds (about 76 per cent), income funds (about 74 per cent) and balanced funds (65 per cent). It can be concluded that the style benchmarks of federal fund, small companies and index funds used are suitable in explaining the underlying asset classes. .

Style consistency To verify the style consistency, the 60-month period is subdivided into two sub-periods. First sub-period started from January 1997 to June 1999. The second sub-period started from July 1999 to December 2001, of which both sub-periods have 30-month of return series (except M Equity which has only 29-month of return series in each sub-period). It could be observed from Table VI, the F-test show that majority of the funds have significant differences in their residual variances from the two sub-periods. However, the annual alphas for most of the funds do not show significant results for the whole 60-month period or 58-month period (for M Equity fund) of return series. These results identify significant style changes from the two sub-periods. This could be due to several reasons, such as the need to rebalance the portfolio according to economic changes, the imposition of capital control (from 2 September 1998) and the slight withdrawal of the restrictions of capital control in the following months in 1999 and 2000. However, as the alphas for majority of the funds are not statistically significant for the whole sample period, it could not be concluded that style changes resulted in negative performance for the 60-month period with any degree of statistical confidence. Conclusion While this study concurs with Chua (1985) that funds with passive strategy or high degree of style are likely to be the winner, this study draws additional conclusions. First, the inclusion of asset classes with negative correlation coefficient such as large-cap stocks with cash enhances the performance of funds. Second, besides having a relatively high degree of style (above 70 per cent), funds that hold large-cap stocks together a relatively high portion of liquid asset class (6-35 per cent) tend to have higher alpha, translating into higher information ratio. It could be implied that liquid asset class enables fund managers to invest in stocks that improve their values in economic cycles. It is observed that the integrated style analysis framework has enhanced the understanding of Malaysian unit trust funds, particularly the estimated style of different fund types and the behaviour of different groups of fund managers, which could be useful to equity trust investors, fund managers and the regulators. It is obvious that most of the funds concentrated on large-cap stocks, followed by mediumcap and government bonds. In view of this trend, it would be beneficial to equity investors if the authority would create sub-indices in the large-cap, in order to further enhance product differentiation in the financial markets[11]. In addition, in the context of an emerging market, there is a need for market movers and regulators to enhance the development of Malaysian capital market, such as the creation of more asset classes for example listed property, international stocks and bonds, and parallel to product development, appropriate asset class indices are also being created. This would certainly enable the usage of style analysis to gauge the adherence of plan sponsors and money managers to their investment policy set-forth in the fund prospectuses. In addition, the development of new investment product would

An integrated framework for style analysis 137

MF 33,2

138

enhance the liquidity of the markets, besides providing investment opportunities for different types of investors. Through the analysis, the degree of styles has indicated a possible trend of misclassification as the degree of styles for index funds do not differ from the other fund types. On the contrary, it could be said that the styles of other fund types are converging towards index funds. There are two implications here. Firstly, further research could be conducted on the Malaysian equity funds as to ascertain whether mis-classification of fund exists and to the extent of its economic impact to investors. Secondly, all fund managers are crowding towards large-cap stocks resulting in a lack of product differentiation. Parallel to the development of Malaysia capital market, sub-indices should be created along with introduction of investment products as discussed in the previous part. Albeit that most of the funds do not outperform the passive style benchmarks, the weak-form of style analysis has enhanced the understanding of various performance measurements. In addition, by understanding the estimated style of funds, investors could plan their optimal portfolio mix[12], rebalancing or switching to funds that fulfill their investment objectives. There is greater responsibility of asset management companies to provide a full disclosure, if not, an up-to-date information of their asset allocation in annual reports and fund prospectuses. The spate of new funds being launched in the recent years by the Malaysian fund management industry, and the lessons from the Asian financial crisis, pose a greater need for Malaysian fund managers and the regulator – Securities Commission (SC), likewise their counterparts in the developed markets, to place a greater focus on equity style management and risk management to benefit the unit trust investors. In view of eventful financial liberalization of Malaysian capital market by 2007 within the context of ASEAN Free Trade Area (AFTA), it certainly takes a concerted effort from all the market participants to enhance the unit trust industry towards its long-term objectives of having 40 per cent of market capitalization in Malaysian capital market by the year 2020. Notes 1. Refer The Edge (2003), 13 January, p. 37. 2. When the changes of NAV and KLSE market capitalization are plotted in a payoff diagram, the non-proportional relationship can be clearly observed. 3. From Figure 1, it could observed that for the year 2001 and 2002, the year-on-year changes on NAV exceeds the KLSE market capitalization, it could be argued that this is due to new subscriptions in unit trust funds, rather than the recovery of asset value from financial crisis. 4. To cite one example. Refer pages 33 and 35 of the annual report dated 6th December 2001 of unit trust fund PHB. Among the asset allocation by sector as at 30 September, 2001 shown in pie chart, it has listed cash (39 per cent), industrial (6 per cent), trading (7 per cent), finance (4 per cent), other assets (5 per cent), others (19 per cent), unquoted instrument (16 per cent) and other liabilities (4 per cent), of which is lack of accompanying notes. This type of ambiguous disclosure raises of issue of whether this fund has actually violated its own investment limit which it must not exceed 10 per cent of its NAV of fund in investments in unlisted securities. Users of this annual report are left to their own interpretation of the meaning of ‘‘other assets’’ and ‘‘others’’ which accounts for about 25 per cent of the total NAV. 5. Refer TerHorst et al. (2004), p. 30, para 4. 6. The companies listed in second board must have a minimum paid-up capital of RM40 million with stock of RM1 per share. The board was launched in 1991, an all-share

7.

8. 9.

10. 11.

12.

index which is weighted by market capitalization with base date on 31 December 1990. For other listing requirements, refer Kuala Lumpur Stock Exchange (2001) information handbook, p. 56. EMAS Index is the abbreviation of Exchange Main Board All-Shares Index. The board is weighted by market capitalization, with base date on 1 January 1994 and an assigned index value of 100, and a total of 269 companies listed on base date. The positivity constraint of style analysis here appears to have no contradiction to the application to Malaysian mutual fund industry as short-selling is not an approved practice. Refer Kahn and Rudd (2003) in pp. 261-2 of Coggin and Fabozzi (Eds) (2003). Also refer to Goodwin (1998), p. 35, equations (6) and (7). The monthly alpha estimates are annualized by (1 þ )12
1. Refer Goodwin (1998), p. 37, equation (13). The author would like to draw the Australian experience where indices such as S&P/ ASX100, S&P/ASX200, etc. have been created to cater for the needs of different groups of investors. Available at: www.asx.com.au, accessed 5 March 2004. Refer Lucas and Riepe (1996) for discussion on how investors could design an optimal mix of holdings of different asset classes once the estimated styles of their funds are known.

References Amenc, N., Sfeir, D. and Martellini, L. (2002), ‘‘An integrated framework for style analysis and performance measurement’’, EDHEC Risk and Asset Management Research Centre, France, available at: www.edhec-risk.com/performance_and_style_analysis/EDHEC%20Publications/ RISKReview1048588983078116286/attachments/edhec_perf_analysis.pdf (accessed 30 April 2004). Bailey, J.V. (1992), ‘‘Are manager universes acceptable performance benchmarks?’’, Journal of Portfolio Management, Vol. 18 No. 3, pp. 9-13. Bailey, J.V. and Tierney, D.E. (1993), ‘‘Gaming manager benchmark’’, Journal of Portfolio Management, Vol. 19 No. 4, pp. 37-40. Brinson, G.P., Singer, B.D. and Beebower, G.P. (1991), ‘‘Determinants of portfolio performance II: an update’’, Financial Analysts Journal, Vol. 47 No. 3, pp. 40-8. Brown, S.J. and Goetzmann W.N. (1997), ‘‘Mutual fund styles’’, Journal of Financial Economics, Vol. 43 No. 3, pp. 373-99. Ch’ng, T.L. and Kok, K.L. (1998), ‘‘Performance of unit trusts in an emerging market: a case study of Malaysia’’, Capital Market Review, Vol. 6 No. 1 and 2, pp. 1-17. Choong, D. (2001), Investor’s Guide to Malaysian Unit Trusts, Sage Information Services, Kuala Lumpur. Christopherson, J.A. (1995), ‘‘Equity style classification’’, Journal of Portfolio Management, Vol. 21 No. 3, pp. 32-43. Chua, C.P. (1985), ‘‘The investment performance of unit trusts in Malaysia’’, unpublished MBA thesis, School of Management, University Malaya, Kuala Lumpur. Coggin, T.D. (1998), ‘‘Long-term memory in equity style indexes’’, Journal of Portfolio Management, Vol. 24 No. 2, pp. 37-46. Coggin, T.D. and Fabozzi, F.J. (Eds) (2003), The Handbook of Equity Style Management, 3rd ed., John Wiley & Sons, Inc., New York, NY. DiBartolomeo, D. and Witkowski, E. (1997), ‘‘Mutual fund mis-classification: evidence based on style analysis’’, Financial Analysts Journal, Vol. 53 No. 5, pp. 32-43. Ewe, S.J. (1994), ‘‘The performance of Malayan unit trusts in the period 1988-1992’’, unpublished MBA thesis, School of Management, University Sains Malaysia, Penang.

An integrated framework for style analysis 139

MF 33,2

140

Federation of Malaysian Unit Trust Managers (n.d.), available at: www.fmutm.com.my (accessed 5 March 2004). Federation of Malaysian Unit Trust Managers (1998), Understanding Malaysian Unit Trusts, 1st ed., FMUTM, Kuala Lumpur. Goodwin, T.H. (1998), ‘‘The information ratio’’, Financial Analysts Journal, Vol. 54 No. 4, pp. 34-43. Ibbotson, R.G. and Kaplan, P.D. (2000), ‘‘Does asset allocation policy explain 40, 90 or 100 per cent of performance’’, Financial Analysts Journal, Vol. 56 No. 1, pp. 26-33. Kahn, R.N. and Rudd, A. (2003), ‘‘The persistence of equity style performance: evidence from mutual fund data’’, in Coggin, T.D. and Fabozzi, F.J. (Eds), The Handbook of Equity Style Management, John Wiley & Sons, Inc., New York, NY, pp. 259-71. Kim, M., Shukla, R. and Tomas, M. (2000), ‘‘Mutual fund objective misclassification’’, Journal of Economics and Business, Vol. 52 No. 4, pp. 309-23. Kuala Lumpur Stock Exchange (2001), Information Handbook, Kuala Lumpur Stock Exchange, Kuala Lumpur. Lau, W.Y. (2002), ‘‘Does asset allocation explain the styles and performance of unit trust funds: a style analysis with evidence from Malaysia’’, Journal of Malaysian Studies, Vol. XX No. 2, pp. 1-32. Leong, K.H. and Aw, M.W. (1997), ‘‘Measuring unit trust fund performance using different benchmarks’’, Capital Market Review, Vol. 5 No. 2, pp. 27-44. Lucas, L. and Riepe, M.W. (1996), The Role of Return-Based Style Analysis: Understanding, Implementing, and Interpreting the Techniques, Ibbotson Associates, Inc., USA. Shamsher, M. and Annuar, M.N. (1995), ‘‘The performance of unit trusts in Malaysia: some evidence’’, Capital Market Review, Vol. 3 No. 2, pp. 51-69. Shamsher, M. and Annuar, M.N. (2001), ‘‘Investment in unit trusts: choosing active or passive funds’’, The Company Secretary, MAICSA, Kuala Lumpur, No. 3, pp. 16-7. Sharpe, W.F. (1988), ‘‘Determining a fund’s effective asset mix’’, Investment Management Review, pp. 59-69. Sharpe, W.F. (1992), ‘‘Asset allocation: management style and performance measurement’’, Journal of Portfolio Management, Vol. 18 No. 2, pp. 7-19. TerHorst, J.R., Nijman, T.E. and DeRoon, F.A. (2004), ‘‘Evaluating style analysis’’, Journal of Empirical Finance, Vol. 11 No. 1, pp. 29-53. The Edge (2003), The Edge Communications Sdn Bhd, Kuala Lumpur, 13 January. Tierney, D. and Winston, K. (1991), ‘‘Using generic benchmarks to present manager styles’’, Journal of Portfolio Management, Vol. 17 No. 4, pp. 33-6. Further reading Brinson, G.P., Hood L.R. and Beebower, G.P. (1986), ‘‘Determinants of portfolio performance’’, Financial Analysts Journal, Vol. 42 No. 4, pp. 39-48. Dhesi, D. (2001), ‘‘Unit trust sector targeted to hit 40% KLSE market cap by 2020’’, The Star, Star Publications Ltd., Kuala Lumpur, 6 December. Lehman, B.N. and Modest, D.M. (1987), ‘‘Mutual fund performance evaluation: a comparison of benchmarks and benchmarks comparisons’’, Journal of Finance, Vol. XLII No. 2, pp. 233-65. Norashikin, A.H. (2000), Guide to The Malaysian Bond Market, Rating Agency Malaysia Berhad, Kuala Lumpur. Permodalan Nasional Berhad (2001), Malaysia Unit Trust Directory 2001, Permodalan Nasional Berhad, Kuala Lumpur.

Appendix No. Plan sponsors

Fund

Launch date Fund type

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43

Affin Equity ASN AM First M Progress M Berjaya M Equity M Investment ASM 2 Index ASM 3 ASM 4 ASM 5 ASM 6 ASM 7 ASM Growth ASM 8 ASM 11 ASM premier ASM ptnb Double Growth Emerging Companies Savings Fund High Growth Fund CT Trust CT Prime HLG Growth Mayban Unit Trust Mayban Balanced MBF Balanced MBF Growth Pacific Premier BSN Public Savings Public Growth Public Index Public Industry Public Aggressive Growth Public Regular Savings Public Balanced RHB Dynamic RHB Capital Premium Capital OSK-UOB Equity TA Growth

93.04.29 81.04.20 89.01.10 70.06.01 76.05.05 82.02.20 96.07.18 69.02.19 69.11.01 70.02.02 71.09.03 72.05.05 72.12.28 72.12.28 75.07.17 79.10.28 95.06.12 95.08.28 91.05.15 94.05.10 95.08.05 95.09.28 89.08.19 91.05.14 95.09.08 92.03.26 94.09.19 91.05.01 95.06.01 95.08.10 95.01.12 81.03.29 84.12.11 92.03.02 93.11.18 94.04.25 94.04.25 92.09.15 92.09.15 95.04.12 95.08.01 96.08.08 96.07.01

Affin Trust ASNB Arab Malaysian Asia Unit Trust Asia Unit Trust Asia Unit Trust Asia Unit Trust Amanah Saham Mara Amanah Saham Mara Amanah Saham Mara Amanah Saham Mara Amanah Saham Mara Amanah Saham Mara Amanah Saham Mara Amanah Saham Mara Amanah Saham Mara Amanah Saham Mara Amanah Saham Mara SBB SBB SBB SBB Commerce Trust Commerce Trust HLG Mayban Mayban MBF MBF Pacific Mutual BSN Public Mutual Public Mutual Public Mutual Public Mutual Public Mutual Public Mutual Public Mutual RHB RHB SBB OSK-UOB TA Unit Trust

Units (Mil)

Income Federal Income Small companies Income Small companies Income Index Income Income Income Income Income Growth Income Income Income Income Growth Small companies Balanced Growth Income Income Growth Income Balanced Balanced Growth Income Income Income Income Index Income Growth Income Balanced Income Growth Income Growth Income

Source: FMUTM. Available at: www.fmutm.com.my (accessed 5 March 2004)

Corresponding author Wee-Yeap Lau can be contacted at: [email protected] To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

300 2,500 500 300 50 50 300 20 20 20 20 20 20 20 20 20 350 50 550 700 500 1,000 300 300 300 500 1,000 750 300 500 500 500 1,000 500 1,000 500 1,500 1,000 750 500 500 750 350

An integrated framework for style analysis 141

Table AI. List of unit trust funds in the sample

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0307-4358.htm

MF 33,2

Investigation of performance of Malaysian Islamic unit trust funds

142

Comparison with conventional unit trust funds Fikriyah Abdullah School of Finance and Accounting, Universiti Utara Malaysia, Kedah, Malaysia

Taufiq Hassan Department of Accounting and Finance, Faculty of Economics and Management, Universiti Putra Malaysia, Serdang, Selangor, Malaysia, and

Shamsher Mohamad Faculty of Economics and Management, Graduate School of Management, Universiti Putra Malaysia, Serdang, Selangor, Malaysia Abstract Purpose – One of the implications of Islamic investment principles is the availability of Islamic financial instruments in the financial market. The main aim of this research is to observe the differences in terms of performance between Islamic and conventional mutual fund in the context of Malaysian capital market. Design/methodology/approach – To achieve the major objectives of this paper standard methods wereused for evaluating the mutual funds performance, for example, Sharpe index and adjusted Sharpe index, Jensen Alpha, Timing and selectivity ability. The scope of the paper is to measure the relative quantitative performance of funds which was managed based on two different approaches. Findings – The basic finding of the paper is that Islamic funds performed better than the conventional funds during bearish economic trends while, conventional funds showed better performance than Islamic funds during bullish economic conditions. In addition to that finding, both conventional and Islamic funds were unable to achieve at least 50 per cent market diversification levels, though conventional funds are found to have a marginally better diversification level than the Islamic funds. The results also suggest that fund managers are unable to correctly identify good bargain stocks and to forecast the price movements of the general market. Research limitations/implications – The main limitation is that the samples of conventional and Islamic mutual funds were from one developing market. The findings could be better validated if the sample included the mutual funds from other developed and developing economies, where both Islamic and conventional funds are available. Practical implications – The findings suggest that having Islamic mutual funds in an investment portfolio helps to hedge the downside risk in an adverse economic situation. Originality/value – So far there is no published evidence on the relative performance of Islamic and conventional mutual funds in Malaysia as well as other developing countries. Therefore, this paper adds new knowledge to the mutual funds literature. Keywords Performance management, Financial risk, Islam, Fund management, Unit trusts Paper type Research paper Managerial Finance Vol. 33 No. 2, 2007 pp. 142-153 # Emerald Group Publishing Limited 0307-4358 DOI 10.1108/03074350710715854

Introduction Mutual fund or better known as unit trust fund in Malaysia is an investment vehicle created by asset management companies specializing in pooling savings from both retail and institutional investors. Individual investors seeking liquidity, portfolio diversification

and investment expertise are increasingly choosing unit trust funds as their investment vehicle. However, these investors do differ in their preferences based on their risk threshold, liquidity needs and their needs to comply with religious requirement. Over the past decades, Malaysian capital market has been growing at a very fast pace and this has been supportive of the various complex needs of the country. In view of the fact that capital market is functioning based on interest, it is therefore not surprising that the operation of capital market does not conform to the Shari’ah principles (Islamic law as revealed in the Quran and Sunnah). Hence, it is difficult for Muslim investors representing more than 50 per cent of the total population, to participate freely in the Malaysian capital market. The increasing demand for alternative investment vehicles, which conform to the Shari’ah principle has prompted the Malaysian government to introduce various measures with the aim of boosting the Islamic capital market (ICM). For examples, among other measures, an Islamic bank was established in 1980 and the Kuala Lumpur Shari’ah Index was introduced in 1999 with the objective of fulfilling the investment needs of Muslim investors. Based on the Shari’ah principle, transactions taking place in the capital market should be free from prohibited activities or elements such as usury (riba), gambling (maisir) and ambiguity (gharar). Islamic investing can be defined as investment in financial services and other investment products, which adhere to the principles established by the Shari’ah[1]. These principles require that: (1) Investment must be made in ethical sectors. In other words, profits cannot be generated from prohibited activities such as alcohol production, gambling, pornography etc. In addition, investing in interest (riba)-based financial institutions are not allowed. (2) All wealth creation should result from a partnership between an investor and the user of capital in which rewards and risks are shared. Returns in invested capital should be earned rather than be pre-determined. One of the implications of Islamic investment principles is in the availability of Islamic financial instruments in the financial market. The Shari’ah’s prohibition against riba (interest) and some Fiqhi (Islamic Jurisprudence) issues in the interpretation of gharar (excessive risk) suggests that many of the instrument products, which are available to conventional funds are not available to Islamic funds. In the case of financial services, for instance Islamic banking, it is not much of a hassle for a conventional bank to open up Islamic banking window as an Islamic investment arm. However, in the case of Islamic investment instruments for example, Islamic unit trust funds, fund managers do have some limitations in selecting stocks to form part of their portfolio. Even though most of the banks listed in Bursa Malaysia have Islamic banking arms, due to the Shari’ah guidelines, fund managers are unable to include banking stocks in their portfolio. Furthermore, due to the absence of Islamic money market, Islamic unit trust funds depend solely on the equity market for investment. For conventional equity unit trust funds, fund managers do not invest solely in the equity market. Rather, a fraction of their investment goes to the money market, which comprises of risk free investments. Table I presents the investment proportions of conventional equity funds in Malaysia over the seven-year period, from 1995-2001. Before the outbreak of the financial crisis in 1997, the Malaysian economy had registered high growth for a decade, averaging at 8.5 per cent per annum. Due to high GDP growth rate and inflows of foreign capital, cyclical sectors such as banking sector, property and entertainment sectors had recorded high performance in the stock market[2]. Unlike Islamic unit trust funds, conventional funds do not have any

Malaysian Islamic unit trust funds 143

MF 33,2

144

restrictions to incorporate these high performing stocks into their portfolio during good economic conditions. Similarly, during financial crisis in 1997, stocks in the cyclical sectors recorded the worst performance and consequently, this was reflected in the poor performance of conventional unit trust funds. However, given that Islamic funds have limited investment choices, their performances were, therefore, not subjected to the cyclical effects of the economy. While conventional funds are subjected to the capital market laws, Islamic funds are subjected not only to the capital market laws but also to the Quranic law of economics. Based on this strategic difference, the purpose of this paper is to provide an understanding on the performances of these two diverse unit trust funds in the Malaysian capital market, namely conventional and Islamic funds. Islamic funds are relatively new as compared to the conventional funds. In the Malaysian context, so far, there is no published evidence on the comparison of performance between Islamic funds and conventional funds. This study intends to fill the gap in the literature by providing further empirical evidence on the Malaysian unit trust funds. The rest of the paper unfolds as follows: Review of previous studies provides a brief review of the relevant literature. Data and methodology outlines the methodology and the measures of performance used. Results and discussion discusses the findings from the empirical analysis and Results and discussion provides concluding remarks. Review of previous studies The literature on the performance of mutual funds has long standing issues. The issues addressed by previous studies include the risk-return performance, selection and market timing abilities of fund managers and the level of diversification of mutual funds. McDonald (1974) estimates the Sharpe, Treynor and Jensen measures for 123 mutual funds using monthly data for the period between 1960 and 1969. The findings show that majority of the funds did not perform as well as the New York Stock Exchange (NYSE) index. Kon and Jen (1979) examines the non-stationarity of the market-related risk for mutual funds over time. They separate their samples into different risk regimes and then run the standard regression equation for each such regime. Using a sample of 49 mutual funds and their net monthly returns from January 1960 to December 1971, they find the existence of multiple levels of beta for 37 funds. This implies that there are a large number of funds engaging in market timing activities. Kon (1983) extends his analysis to examine both market timing and selectivity performance. He finds that 14 funds have positive overall timing performance but none is statistically significant at a reasonable level. Five out of 23 funds show statistically significant overall selectivity performance.

Table I. Average investment among sectors for conventional funds: 1995-2001

Industrial products (%) Consumer products (%) Trading/services (%) Finance (%) Property (%) Construction (%) Plantation (%) Others (%) Total equities (%)

1995

1996

1997

1998

1999

2000

2001

9.65 6.53 19.99 16.00 8.15 2.48 10.95 2.04 75.78

12.87 5.72 24.69 14.08 7.16 8.12 5.06 3.69 76.45

13.42 7.23 20.68 15.12 3.51 4.07 15.09 9.99 89.07

8.82 4.61 20.91 11.24 2.81 5.43 5.49 6.06 64.57

10.58 6.13 22.12 12.52 2.74 8.00 4.18 5.49 71.07

4.80 7.97 22.58 46.86 4.05 0.00 9.60 0 95.86

8.9 7.54 22.9 17.44 2.21 6.51 5.65 7.83 78.98

In their study, Chen et al. (1992) examine a sample of 93 mutual funds covering a period from January 1977 through March 1984 by using a quadratic market model in conjunction with a systematically varying parameter regression method. The results indicate that fund managers do not possess market timing abilities. Furthermore, they find a trade-off between market timing and security selection abilities. Annuar et al. (1997) use the Treynor and Mazuy model to examine the selectivity and timing performance of 31 unit trust funds in Malaysia for the period of July 1990 through August 1995. On average, the selectivity performance of the funds is positive and the timing performance is negative. The study finds a positive correlation between selectivity and timing performances. The results also show that the funds have not achieved the expected level of diversification and the risk-return characteristics of the unit trust funds are generally inconsistent with their stated objectives. Shamsher et al. (2000) conduct a study on the performance of 41 actively and passively managed funds in Malaysia covering the period from 1995 through 1999. The performance measures used are the Sharpe’s index, the Treynor’s index and the Jensen’s index. The findings reveal no significant differences in the performance of actively and passively managed funds. Moreover, the returns of these funds are lower than the returns of the market portfolio. The diversification levels of these two funds are less than 50 per cent of the diversification level of the market index as proxied by the Kuala Lumpur Composite Index (KLCI). The selection skills of active fund managers are no better than that of the passive fund managers. The market timing abilities of managers are found to be poor for both the actively and passively managed funds. Data and methodology This study focuses on equity-based unit trust funds and the sample consists of both Islamic and conventional funds. The conventional funds are further divided into governmental and non-governmental funds, which include growth funds, income funds and balance funds. The sample consists of 65 funds, 14 of which are Islamic funds. Monthly returns adjusted for dividends and bonuses distributed to unit holders are computed for the 10-year period starting from January 1992 through December 2001. To serve as a market benchmark, the returns on the KLCI i.e. the KLCI (formerly known as the KLSE Composite Index) are used as a proxy for the returns on the market portfolio and the risk free rate is proxied by the three-month Treasury Bills[3]. In view of the fact that the Malaysian economic conditions changed dramatically as a result of the financial crisis in 1997, we divide the study period into three different periods to ascertain the impact of the economic conditions on the performance of unit trust funds. The three different periods are: pre-(1992-1996), during (1997-1998) and post-(1999-2001) financial crisis. Measurement of performance The returns on the unit trust funds are derived from two components namely income and the capital gain. The rate of returns for each fund is calculated as follows: Rp ¼

NAVt
NAVt
1 þ Dt NAVt
1

where Rp ¼ Total return of a portfolio (individual fund); NAVt ¼ Net Asset Value at time t;

ð1Þ

Malaysian Islamic unit trust funds 145

MF 33,2

146

NAVt
1 ¼ Net Asset Value one period before time t; and Dt ¼ Dividend or cash disbursement at time t. In this study, we employ three standard methods namely the Treynor’s Index, the Sharpe Index and the Jensen Index and Adjusted Jensen Index to evaluate the performance of unit trust funds. Due to the biasness in the estimation of the standard deviation, the Sharpe’s Index has been modified by Jobson and Korkie (1981) to become the Adjusted Sharpe Index. This study also employs Adjusted Sharpe Index, which is expressed as follows: ASI ¼

SI  no. of observations no. of observations þ 0:75

ð2Þ

Modigliani and Modigliani (1997) propose an alternative risk adjusted measure which is easier to understand. The measure expresses a fund’s performance relative to the market in percentage terms, which is given as follows: MM ¼

Rp
Rf 
m
t

ð3Þ

where MM ¼ Modigliani measure; Rp ¼ ex post adjusted returns on the mutual funds over the measurement period; Rf ¼ risk-free rate of return on corresponding period on a government security;
t ¼ standard deviation of returns of the mutual funds; and
m ¼ standard deviation of market (index) excess return. An additional performance measure known as Information ratio, is defined as follows: Information ratio ¼

Fund return
Benchmark return Standard deviation (fund return
Benchmark returnÞ

ð4Þ

Measurement of risk, selectivity and timing The total risk on investments is measured using the standard deviation. Unlike the standard deviation, which is an absolute measure of variability, the coefficient of variation is a relative measure of variability. The coefficient of variation (CoV) ratio which measures the amount of risk assumed per unit of average returns, is expressed as follows: CoV ¼


i EðRi Þ

ð5Þ

where
i ¼ standard deviation (total risk) of asset i; and E(Ri) ¼ average return of asset i. The stock selection and market timing performances of each fund are estimated using the Treynor and Mazuy (1966) model and the equation is as follows: Rp ¼ p þ p ðRm Þ þ ðRm Þ2 þ  where

ð6Þ

Rp ¼ dividend-adjusted return on portfolio per cent minus the yield on 91-day Treasury Bill’s rate; p ¼ coefficient that indicates estimated selectivity skill;

Malaysian Islamic unit trust funds

p ¼ beta risk of unit trust; Rm ¼ observed return on the KLSE Composite Index minus Rf (risk-free rate);

147

 ¼ coefficient that indicates market-timing skill; and  ¼ residual excess return on portfolio per cent. A positive and significant  and  indicate superior selectivity and market-timing skills, respectively. Finally, heteroscedasticity and serial correlation problems, which are common in any regression based model, are corrected by using White’s (1980) correction test and Newey-West’s (1987) correction test. Results and discussion Non risk-adjusted returns of unit trust funds Table II summarizes the non risk-adjusted returns for both Islamic and conventional funds, which include governmental and non-governmental funds. Also reported are the returns of the market portfolio as proxied by the KLCI. During the pre-crisis period, among the portfolios, KLCI achieves the highest level of average monthly returns. The average monthly returns for Islamic funds, conventional funds, be they governmental or non-governmental funds are below the average monthly return of the KLCI. This suggest that unit trust funds under-perform the market on a non risk-adjusted basis. During the crisis, the stock market was badly affected by the economic downturn and this was reflected by a negative average monthly return of
1.98 per cent for the KLCI which recorded the lowest level of returns among the portfolios. The performances of Islamic and non-governmental funds appear to be better during the crisis period than during the pre-crisis period. For conventional and governmental funds, their average monthly returns during the crisis period are lower than during the pre-crisis period. The results for the post-crisis period reveal that the market portfolio, KLCI experiences the highest level of average monthly returns. The average monthly returns for all of the unit trust funds during the post-crisis period are better than those recorded during the pre-crisis period. Over the 10-year period, the average monthly return of the market is 0.81 per cent. The average monthly returns for Islamic, conventional and non-governmental funds are 0.72, 0.73 and 0.54 per cent, respectively. This suggests that most of the funds, under-perform the KLCI. Exception to this is the government funds which record an

Islamic funds t-statistic Conventional funds t-statistic Governmental funds t-statistic Non-governmental funds t-statistic Market (KLCI)

Pre-crisis

During crisis

Post-crisis

Overall

0.000781 (
0.543283) 0.006926 (
0.496229) 0.011311 (
0.052740) 0.002673 (
0.802749) 0.012147

0.024430 (0.860750) 0.003064 (0.662587) 0.005057 (0.509891) 0.009267 (0.863347)
0.019811

0.004767 (
0.242770) 0.007751 (
0.317379) 0.012837 (
0.050591) 0.004742 (
0.443947) 0.014647

0.007197 (
0.029178) 0.007309 (
0.048464) 0.011949 (0.143911) 0.005359 (
0.159691) 0.008103

Table II. Non risk-adjusted returns of unit trust funds over market portfolio

MF 33,2

148

average monthly returns of 1.19 per cent. One plausible reason for this is that these government funds are closely monitored by government agencies which are responsible for ensuring good performance of the funds. Therefore, these government funds may have better opportunities of investing in government-backed projects as compared to any other funds. Furthermore, government-backed projects are relatively more secure and stand a better chance of generating good returns. Risk-adjusted average monthly returns of unit trust funds Table III presents the comparative performance analysis over a 10-year period for Islamic funds vs conventional funds as well as for governmental vs non-governmental funds. It is shown that conventional funds perform better than Islamic funds during the pre-crisis period using various types of performance measures: Adjusted Sharpe Index, Treynor Index, Adjusted Jensen Index, Modigliani Measures and the Information Ratio. However, the opposite is also true during the crisis and post-crisis periods. During the crisis period, regardless of the type of return measurements used, Islamic

Period

Funds

Pre-crisis

Islamic funds Conventional funds t-statistic Governmental funds Non-governmental funds t-statistic Islamic funds Conventional funds t-statistic Governmental funds Non-governmental funds t-statistic Islamic funds Conventional funds t-statistic Governmental funds Non-governmental funds t-statistic Islamic funds Conventional funds t-statistic Governmental funds Non-governmental funds t-statistic

During crisis

Post-crisis

Overall

Table III. Risk adjusted monthly returns of different classes of unit trust funds

Treynor’s Index

Adjusted Jensen’s Alpha Index

Modigliani Measures

Information Ratio


0.1047
0.0196 (
1.841)* 0.0225


0.0086 0.0028 (
1.068) 0.0109


0.0158
0.0044 (
1.068) 0.0037


0.0073
0.0014 (
1.837)* 0.0015


0.3592
0.1296 (
1.488)
0.0144


0.0673 (2.495)** 0.0025
0.1523 (2.598)**
0.1798


0.0049 (1.904)* 0.0031
0.0494 (1.366)
0.037


0.0121 (1.904)* 0.0284
0.0241 (1.3657)
0.0117


0.0047 (2.496)** 0.0013
0.0268 (2.575)**
0.0319


0.2091 (2.224)** 0.2459 0.2458 (0.001) 0.1998


0.0879 (
1.650) 0.0393 0.0183 (0.766) 0.0035


0.0377 (0.017) 0.0209 0.0025 (0.216) 0.0264


0.0123 (0.017) 0.0088
0.0095 (0.216) 0.0143


0.0151 (
1.645) 0.0181 0.0125 (1.235) 0.0099

0.2677 (
0.285)
0.0685
0.0457 (
0.422)
0.0109

0.0308 (
1.105)
0.0053
0.0324 (1.169)
0.0207


0.0017 (0.367)
0.0998
0.0088 (
0.037) 0.0073


0.0138 (0.367)
0.0138
0.0127 (
0.036) 0.0036

0.0153 (
1.326)
0.00057
0.0038 (1.184)
0.0024


0.0673 (1.274)
0.0084
0.0084 (
3.9E-05) 0.0345


0.029 (0.390)


0.0159 (0.852)


0.0197 (0.852)


0.0034 (0.391)


0.029
1.03

Adjusted Sharpe’s Index

Notes: * Indicates significance at the 10 per cent; ** indicates significance at the 5 per cent

funds record positive average monthly while the conventional funds show negative average monthly returns. In the post-crisis period, the performance of Islamic funds is also better than that of the conventional funds. An intuitive explanation for this is that since Islamic funds are restricted to invest in products, which comply with the Shari’ah principles, the investment choices of Islamic funds are relatively limited in scope when compared to the investment choices available for conventional funds. Conventional funds tend to perform better than the Islamic funds during bullish market trend because conventional funds are able to invest in any stocks including those with high risks exposure. Accordingly, the riskreturn trade off suggests that the returns from investing in risky investments should be high in order to compensate investors for the high level of risks assumed. The opposite is also true during bearish market trend. Since the investment of Islamic funds excludes usury, gambling and ambiguity or uncertainty elements, Islamic funds have a lower degree of risk exposure than conventional funds and therefore are able to minimize their overall risk level. This contributes to making the investment of Islamic funds relatively less volatile during the crisis period and thus, resulting in a better return during bearish economic conditions. Nevertheless, the differences in the performance between Islamic and conventional funds are marginally significant. Similarly, a comparative analysis for the governmental and non-governmental funds reveals that governmental funds perform better than non-governmental funds during the pre-crisis period and the differences are statistically significant. However, during and after the crisis periods, the differences in the performance between these two funds are no longer statistically significant. This could be due to the fact that ‘‘special attention’’ is given by the Malaysian government in its effort to minimize political risk and to ensure that all conventional funds be they governmental or nongovernmental, are equally competitive in the market. Differences in risks of unit trust funds Beta, standard deviation and coefficient of variation are used to measure systematic risk, total risk and risk per unit of returns. As shown in Table IV, over the 10-year period, the beta values of Islamic funds and conventional funds are 0.251 and 0.383, respectively. This suggests that Islamic funds are less sensitive to changes in the market as compared to conventional funds. Such findings are not surprising given that Islamic funds are restricted to invest in Shari’ah-approved stocks only. Within the conventional funds, governmental funds appear to show a higher level of systematic risks than the nongovernmental funds. The beta values for both governmental and non-governmental funds are 0.4877 and 0.2998, respectively. This implies that changes in the market will have greater impact on the governmental funds than on the non-governmental funds. Table IV also reveals that the investment returns of Islamic funds are less variable and thus less risky than the conventional funds. Over the 10-year period, the standard deviation of returns of 0.1295 for the Islamic funds is lower than that of the conventional funds which is 0.1566. The findings also indicate that the variability of returns for governmental funds is much higher than that of the non-governmental funds. The standard deviation of returns for governmental and non-governmental funds are 0.2100 and 0.1262, respectively. In terms of risk per unit of return, as shown in Table IV, Islamic funds have a lower value than conventional funds. The coefficient of variations for Islamic and conventional funds are 17.9972 and 21.4178, respectively. On the other hand, although the standard deviation of returns for governmental funds is higher than that of the

Malaysian Islamic unit trust funds 149

MF 33,2

150

Table IV. Differences in risks of unit trust funds as measured by beta

Beta Funds Islamic funds Conventional funds t-statistic Governmental funds Non-governmental funds t-statistic Standard deviation Islamic funds Conventional funds t-statistic Governmental funds Non-governmental funds t-statistic Coefficient of variation Islamic funds Conventional funds t-statistic Governmental funds Non-governmental funds t-statistic

Pre-crisis

During crisis

Post-crisis

Overall

0.575769 0.679004 (
0.866223) 0.730952 0.638741 (1.132801)

0.374707 0.647849 (
1.443814) 0.819279 0.477680 (2.036165)*

0.141840 0.168516 (
0.616671) 0.171547 0.159892 (0.31649)

0.251128 0.383173 (
1.901845) 0.487663 0.299827 (3.128808)**

0.050421 0.092819 (
0.935872) 0.115341 0.067686 (1.275648)

0.237621 0.221600 (0.114755) 0.263680 0.207518 (0.446388)

0.072436 0.108288 (
0.71171) 0.150526 0.079931 (1.574359)

0.129528 0.156552 (
0.399858) 0.210029 0.126238 (1.390866)

64.524025 13.400770 (1.44431) 10.197726 25.22635 (
1.033506)

9.726549 72.313126 (0.892007) 52.146350 22.393470 (0.810646)

15.196741 13.970884 (0.59148) 11.725764 16.856388 (0.77804)

17.9972 21.417863 (
0.739998) 17.576680 23.557523 (0.64716)

Notes: * Indicates significance at the 10 per cent; ** indicates significance at the 5 per cent

non-governmental funds, the relative variation of governmental funds is smaller than that of the non-governmental funds. However, the differences in risks between Islamic and conventional funds as well as between the governmental and non-governmental funds are only marginally significant. Diversification level of unit trust funds The coefficient of determination, R2 is used to measure the degree of diversification of the fund relative to the diversification of the market portfolio. The results in Table V shows that conventional funds have a marginally better diversification level than the Islamic funds. Since conventional funds do not have much restrictions in terms of their investment choices, it is therefore not surprising that the R2 of 0.2854 for conventional funds is higher than the R2 of 0.2618 for the Islamic funds. Meanwhile, during the same period,

Funds

Table V. Diversification level (R2) of unit trust funds

Islamic funds Conventional funds t-statistic Governmental funds Non-governmental funds t-statistic

Income 0.235441 0.264147 (
0.38706) 0.347976 0.228837 (0.1900223)

Objective Balance 0.266965 0.242193 (0.336758) 0.234743 0.254262 (
0.363497)

Growth

Overall

0.265875 0.324325 (
0.71418) 0.331997 0.305018
0.31

0.261839 0.285356 (
0.45434) 0.276533 0.281843 (
0.113327)

there are hardly any differences in the diversification level between governmental and non-governmental funds. This could be due to the fact that both of the funds have almost equal investment opportunities. Nevertheless, the differences in the levels of diversification for the funds are not statistically significant. The findings of low diversification level for the funds are consistent with the results of previous studies such as Koh et al. (1987), Shamsher and Annuar (1995) and Shamsher et al. (2000). Selection performance of unit trust funds The results in Table VI indicate that the selection abilities of Islamic and conventional fund managers are negative and statistically significant. Although both fund managers have poor selectivity performance, the selection ability of conventional fund managers seems to be better than that of the Islamic fund managers. The selection abilities of governmental and non-governmental funds are also negative, but only the nongovernmental fund managers show statistically significant negative selection ability.

Malaysian Islamic unit trust funds 151

Market timing performance of unit trust funds Table VII presents findings on the market timing abilities of fund managers. For the overall period, the negative timing coefficients for Islamic and conventional fund a Funds Islamic funds Conventional funds Governmental funds Non-governmental funds

Pre-crisis

During crisis

Post-crisis

Overall


0.013337 (
2.8908)*
0.013879 (
2.3077)**
0.007537 (
0.7662)
0.020130 (
3.3202)*


0.043725 (
0.7160) 0.012405 (0.3871) 0.014876 (0.2834)
0.004075 (
0.1322)


0.028864 (
2.3133)**
0.025880 (
1.7145)***
0.014273 (
0.6611)
0.031613 (
2.4446)**


0.027293 (
3.2724)*
0.016079 (
2.2287)**
0.006831 (
0.6550)
0.024083 (
3.7241)**

Notes: * Indicates significance at the 1 per cent; ** indicates significance at the 5 per cent; *** indicates significance at the 10 per cent; aHeteroscedasticity and serial correlation problems are corrected by using White’s correction test (1980) and Newey-West’s correction test (1987) when necessary

Table VI. Selection performance () of unit trust funds

 Funds Islamic funds Conventional funds Governmental funds Non-governmental funds

Pre-crisis

During crisis

Post-crisis

Overall


0.554124 (
1.4531)
0.330591 (
0.6651)
0.353337 (
0.4346)
0.393200 (
0.7846)

1.246815 (0.8254)
0.532492 (
0.6719)
0.034723 (
0.0267)
0.270850 (
0.3552)

0.200465 (0.7684) 0.256704 (0.8133)
0.017918 (
0.0397) 0.354440 (1.3108)


0.045612 (
0.2096)
0.427398 (
2.2706)*
0.613115 (
2.2535)*
0.216750 (
1.2847)

Note: * Indicates significance at the 5 per cent

Table VII. Market timing ability () of mutual funds managers

MF 33,2

152

managers suggest that both managers do not possess good market timing abilities. However, the timing coefficient for Islamic funds is not statistically significant. This could be due to the fact that the investment choices of Islamic fund managers are less dependent on the fluctuation of the economy since their investments are not cyclical in nature. Similarly, over the 10-year period, the governmental and non-governmental fund managers also show poor market timing abilities but the timing estimate is only significant for the governmental funds. The results of negative timing performance are consistent with most of the previous findings for examples: Henriksson (1984), Chen et al. (1992), Coggin et al. (1993) and Annuar et al. (1997). Conclusion and implication The primary focus of this study is to ascertain the relative performance of Islamic and conventional funds across different economic conditions. The study further divides the conventional funds into governmental and non-governmental funds. In this regard, we conduct a comparative performance analysis between Islamic and conventional funds as well as between governmental and non-governmental funds over three different economic periods namely pre, during and post economic crisis. The results show that performances of the funds are marginally below than that of the market. A relative measure of variability indicate that Islamic funds are less risky than conventional funds. Similarly, governmental funds are found to be less risky than nongovernmental funds. Both Islamic and conventional funds have diversification levels which are less than 50 per cent of the diversification level of the market portfolio. Although governmental funds have a marginally better diversification level than the non-governmental funds, their diversification levels are also below 50 per cent. Poor selection and market timing abilities are documented for all classes of funds – Islamic, conventional, government and non-government funds. Our results also indicate that Islamic funds perform better than conventional funds during bearish economic trend i.e. during the crisis period. However, during bullish trend defined as the pre-crisis period, the performance of conventional funds is better than that of the Islamic funds. This implies that Islamic funds can be used as a hedging instrument during any financial meltdown or economic slowdown. The findings of the study may have important implications for investors and regulatory agencies of the unit trust industry. In view of the fact that conventional funds perform better than Islamic funds during good economic period and vice-versa during bad economic period, this provides good justification for the market regulators to further enhance the ICM in Malaysia. Notes 1. List of Securities Approved by Shari’ah Advisory Council of the Securities Commission, 26 October 2001. Securities Commission. 2. Based on the Shari’ah guideline, investment in the property sector, which normally has a debt level of more than 50 per cent, is not permitted. 3. The continuous monthly risk-free return is calculated using the following equation: rft ¼ (ln[1 þ rft]/12). References Annuar, M.N., Shamsher, M. and Ngu, M.H. (1997), ‘‘Selectivity and timing: evidence from the performance of Malaysian unit trusts’’, Pertanika Journal of Social Science and Humanities, Vol. 5, pp. 45-57.

Chen, C.R., Cheng F.L., Rahman, S. and Chan, A. (1992), ‘‘A cross sectional analysis of mutual fund’s market timing and security selection skill’’, Journal of Business Finance and Accounting, Vol. 19, pp. 659-74. Coggin, D.T., Fabozzi, F.J. and Rahman, S. (1993), ‘‘The investment performance of US equity pension fund managers: an empirical investigation’’, Journal of Finance, Vol. 48, pp. 1039-55. Henriksson, R. (1984), ‘‘Market timing and mutual fund performance: an empirical investigation’’, Journal of Business, Vol. 57, pp. 73-96. Jobson, J. and Korkie, R. (1981), ‘‘Performance hypothesis testing with the sharpe and treynor measures’’, Journal of Finance, Vol. 36 No. 4, pp. 889-908. Koh, F., Koh, S.K. and Cheng, T.C. (1987), ‘‘An empirical analysis of the performance of unit trusts in singapore’’, Securities Industry Review, Vol. 13, October, pp. 1-14. Kon, S.J. (1983), ‘‘The market timing performance of mutual fund managers’’, Journal of Business, Vol. 56, pp. 321-47. Kon, S.J. and Jen, F.C. (1979), ‘‘The investment performance of mutual funds: an empirical investigation of timing, selectivity and market efficiency’’, Journal of Business, Vol. 63, pp. 261-78. McDonald, J.G. (1974), ‘‘Objectives and performance of mutual funds’’, Journal of Finance and Quantitative Analysis, Vol. 13, pp. 311-33. Modigliani, F. and Modigliani, L. (1997), ‘‘Risk adjusted performance’’, Journal of Portfolio Management, Vol. 23, pp. 45-54. Newey, W.K. and West, K.D. (1987), ‘‘A simple positive semi-definite heteroskedasticity and autocorrelation consistent covariance matrix’’, Econometrica, Vol. 55, pp. 703-8. Shamsher, M. and Annuar, M.N. (1995), ‘‘The performance of unit trusts in malaysia: some evidence’’, Capital Market Review, Vol. 3, pp. 51-64. Shamsher, M., Annuar, M.N. and Taufiq, H. (2000), ‘‘Investment in unit trusts: performance of active and passive funds’’, Proceedings of FEP Seminar 2000: Issues in Accounting and Finance 2, Universiti Putra Malaysia Press, Serdang, pp. 129-41. Treynor, J.L. and Mazuy, K.K. (1966), ‘‘Can mutual funds outguess the market?’’, Harvard Business Review, Vol. 44, pp. 131-6. White, H. (1980), ‘‘A heteroscedasticity consistent covariance matrix estimator and a direct test of heteroscedasticity’’, Econometrica, Vol. 48, pp. 817-8. Futher reading Jobson, J. and Korkie, R. (1982), ‘‘Potential performance and test of portfolio efficiency’’, Journal of Financial Economics, Vol. 10, pp. 433-66. Lee, C.F. and Rahman, S. (1990), ‘‘Market timing, selectivity and a mutual fund performance: an empirical examination’’, Journal of Business, Vol. 63, pp. 261-78. Corresponding author Fikriyah Abdullah can be contacted at: [email protected]

To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

Malaysian Islamic unit trust funds 153

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0307-4358.htm

MF 33,2

Malaysian unit trust funds’ performance during up and down market conditions

154

A comparison of market benchmark Soo-Wah Low School of Business Management, Faculty of Economics and Business, Selangor, Malaysia Abstract Purpose – The paper seeks to examine whether selectivity and timing performance of fund manager is sensitive to the choice of market benchmarks. The two benchmarks used are the Kuala Lumpur Composite Index (KLCI) and the Exchange Main Board All-Share (EMAS) Index. Design/methodology/approach – The paper seeks to employed Jensen’s model to estimate the overall fund performance and Henriksson and Merton’s model to separate the fund manager’s investment performance into the selectivity and market-timing components. Findings – The findings indicate that, on average, the funds display negative overall performance with either the KLCI or the EMAS Index. In addition, there is little variation in the manager’s markettiming and selectivity performance across alternative market benchmarks. It is also reported that a manager’s poor timing ability contributes significantly to the fund’s negative overall performance. Research limitations/implications – The paper employed just two market benchmarks. Inclusion of more market benchmarks in future research may provide further support for the existing findings. Practical implications – Regardless of the market benchmarks used, the results imply that fund managers should seriously reassess their market timing efforts, given that their predictions are very often in the wrong direction than in the right direction. Such findings suggest that no economic benefit accrues to the average fund manager involved in market-timing activities. Originality/value – The paper provides first evidence on the sensitivity of a fund manager’s separate investment components (timing and selectivity) to different specification of the market benchmarks. Keywords Unit trusts, Fund management benchmarking, Securities, Market outing, Malaysia Paper type Research paper

Managerial Finance Vol. 33 No. 2, 2007 pp. 154-166 # Emerald Group Publishing Limited 0307-4358 DOI 10.1108/03074350710715863

Introduction The investment performance of mutual fund has attracted considerable research in the literature of finance and there has been much controversy about the ability of fund managers to outperform the market. Some of the more important early research by Jensen (1968; 1969), Sharpe (1966) and Carlson (1970) provided evidence that goes against the fund managers. Their findings show that not only fund managers have trouble outperforming the market but they even performed at a level inferior to that of the market. Several other studies which also suggest a lack of superior performance are provided by McDonald (1974), Firth (1977), Lehmann and Modest (1987), Cumby and Glen (1990), and Droms and Walker (1994). A more recent study by Malkiel (1995) shows that in aggregate mutual fund performed worse than the market even before taking into account the fund expenses. Although some later studies by Eun et al. (1991) and Kao et al. (1998) have found results which are in favor of the fund managers, the general academic opinion indicates that on average, funds exhibit either negative or no abnormal performance[1]. The surge in the demand for index funds over the past decade is a good example to show that there are fund managers and investors who also

share the belief that on average, actively managed funds cannot outperform the market. A number of studies have decomposed the overall fund performance into timing and selectivity components. Early research on timing performance indicate that fund managers have poor market timing skill (see Merton 1981; Henriksson and Merton, 1981; Veit and Cheney, 1982; Kon, 1983; Chang and Lewellen, 1984; Henriksson, 1984). Similarly, more recent empirical findings by Kao et al. (1998), Volkman (1999) and Rao (2000) also suggest that, on average, fund managers are unable to time major market movements successfully. Similar findings indicating that fund managers have negative timing ability are provided by Grinblatt and Titman (1989b), Cumby and Glen (1990), Connor and Koraiczyk (1991), Chen et al. (1992), and Coggin et al. (1993). Nevertheless, research by Kon (1983), Lehmann and Modest (1987) and Lee and Rahman (1990) show that at the individual fund level, there are some funds that do exhibit either good selectivity performance or superior timing performance. Although the evidence of selectivity is present for some fund managers, Chen et al. (1992) found that the evidence is generally weak after taking into account management fees. Findings by Eun et al. (1991) reveal weak evidence of good market timing performance among international fund managers[2]. In addition to these findings, many studies also report negative correlation between the manager’s selectivity and timing estimates (see Henriksson, 1984; Chang and Lewellen, 1984; Volkman, 1999; among others). Despite the strong evidence indicating that actively managed funds cannot beat the market on a risk adjusted basis, fund managers who are able to outperform the market argue that many academic studies do not accurately measure the investment performance of mutual funds. For example, in measuring the investment performance of managed funds, many studies have typically employed performance methods that make a comparison between the return of the fund and the return of a market benchmark, which may not be appropriate[3]. This is particularly true if the chosen market benchmark does not reflect the portfolio holdings of the evaluated fund or the investment style of the fund manager. The most widely used performance evaluation method is based on the Jensen’s (1968; 1969) alpha which is the intercept of a regression of the excess return of the fund (fund return minus risk free rate) on the excess return of a market benchmark. Since there are many benchmarks available, the problem of how to pick the most appropriate market benchmark remains largely an unresolved issue. Theoretically, the chosen market benchmark for measuring fund performance should reflect the investment characteristics of the evaluated mutual fund. Therefore, improperly chosen proxy for the market benchmark can have a dramatic effect on inferences about fund performance. Roll (1978) and other researchers have noted that the Jensen measure can be sensitive to different specification of the benchmark portfolio. There are several studies that investigate the empirical sensitivity of mutual fund performance to alternative market benchmarks. Lehmann and Modest (1987) and Grinblatt and Titman (1989a; 1993) found that inferences about fund performance are sensitive to the chosen benchmark portfolios. Lehmann and Modest (1987) examined selectivity using the Jensen-like measure based on CAPM and APT models and found substantial differences in the performance results between benchmarks. Market timing performance was ignored in their analysis. Coggin et al. (1993) demonstrate that regardless of the choice of benchmark portfolio, on average, fund managers exhibit positive selectivity and negative timing performance. However, both the selectivity and

Malaysian unit trust funds’ performance 155

MF 33,2

156

timing measures do appear to be somewhat sensitive to the choice of the benchmark portfolio when fund managers are classified by investment styles. Since fund managers also invest in non-index asset, previous studies have also highlighted the importance of taking into account the existence of such assets in the portfolio holding of fund managers. The research of Bello and Janjigian (1997) indicate that unless proper market benchmarks are chosen, the existence of non S&P assets in the mutual fund holding can lead to erroneous conclusions regarding fund performance. The study of Elton et al. (1993) also show similar findings and they corrected the problem by including a bond index and a non S&P 500 equity index in their analysis. Zimmermann and Wetter (1992) in their study of five Swiss stock indices, find that performance measures are very sensitive to different specification of the benchmark index. The findings of Grinblatt and Titman (1994) show that inferences about fund performance can be dramatically affected by the chosen benchmark. They indicated that the results may be due to firm size differences in the mutual fund portfolios. Other studies by Brown and Brown (1987), Daniel et al. (1997) and Block and French (2002) stressed the importance of considering the portfolio weighting and portfolio composition when measuring fund performance. In the Malaysian context, the performance of mutual funds or more popularly known as unit trust funds as reported by Shamsher and Annuar (1995), Tan (1995), Leong and Aw (1997), Annuar et al. (1997) and Low and Noor A. Ghazali (2005) concluded that on average, funds were unable to beat the market. Similar findings were reported by Chua et al. (1985) and Koh et al. (1987) for the performance of unit trust funds in Singapore. In evaluating fund performance, many studies in Malaysia have used the Kuala Lumpur composite index (KLCI) as the market benchmark. The one and only study that examined the sensitivity of fund performance to different benchmark portfolios is provided by Leong and Aw (1997). The two benchmarks used are the KLCI and the (EMAS) Exchange Main Board All-Share Index. Their findings show that when the EMAS Index is used, more funds exhibit better performance than the market based on risk adjusted performance measures. In addition, EMAS Index is also shown to produce higher R-squared than the KLCI. Their findings suggest that the choice of market benchmark is important in measuring the investment performance of Malaysian mutual funds. However, their work focused on the overall fund performance and they did not evaluate the separate contribution of market timing and selectivity to the fund’s overall return. This paper seeks to empirically investigate whether selectivity and timing performance of fund manager are potentially sensitive to the choice of market benchmark. In particular, using both the KLCI and EMAS Index, this paper examines to what extent do alternative market benchmarks affect the timing and selectivity performance of the Malaysian unit trust funds. Both the market benchmarks, KLCI and EMAS Index are weighted by market capitalization but they differ in terms of the underlying portfolio of stocks. The most widely used and popular index among the financial community is the KLCI and it comprises of 100 blue chip stocks listed on the Main Board of the Kuala Lumpur Stock Exchange (KLSE), currently the Bursa Malaysia. On the other hand, the EMAS Index comprises of all stocks listed on the Main Board of the KLSE, and as of 31 March 2003, the index consists of 569 stocks. Given that the portfolio holdings of fund managers also include other blue chip stocks that are not included in the KLCI, it would be interesting to examine whether inferences about the performance components of unit trust funds are affected by the choice of using a broad market index (EMAS Index) vs a narrowly constructed index (KLCI).

As yet, in the Malaysian context, there is no study that examine the sensitivity of timing and selectivity performance to the choice of market benchmark. While Leong and Aw (1997) provide evidence on the sensitivity of fund performance to alternative market indices, their study gives no attention to the sensitivity of the investment components to different specification of the market benchmarks. The objective of this paper is to provide such empirical evidence and the paper is organized as follows. Section II presents the data and methodology employed in the study. Section III discusses the findings of the study. Concluding remarks are offered in Section IV. Data and methodology The data in this study is made up of 40 Malaysian unit trust funds for which monthly price records and distribution information were available for the 5-year period, January 1996 through December 2000. The data were obtained from the local newspapers, fund prospectus and annual report of the fund management companies. The monthly return for each fund is calculated as follows: Rt ¼ ðNAVt
NAVt
1 þ DISTt Þ=NAVt
1 where NAV is the net asset value and DIST is the income and capital gain distributions of the fund[4]. The sample of 40 funds includes 25 income funds, ten growth funds and four balanced funds. Monthly returns on the KLCI and EMAS Index served as benchmarks to proxy for the market’s returns and the information is obtained from the Investors’ Digest (January 1996-December 2000) published by the KLSE. The proxy for risk free rate is a three-month Treasury bill rate gathered from the Monthly Statistical Bulletin, published by the Bank Negara Malaysia (January 1996-December 2000) (Central Bank of Malaysia). Since the reported Treasury bill rate is an annualized holding period yield on a three-month Treasury bill, this rate is converted to a monthly equivalent, consistent with the monthly returns of the unit trust funds and the market’s returns[5]. In most previous studies of fund performance, it has become a standard practice to use performance measures that calculate performance relative to a benchmark portfolio. The most widely employed method is based on the Jensen’s model (1968; 1969) with the following regression model: Rpt
Rft ¼
J þ p ðRmt
Rft Þ þ pt

ð1Þ

where Rpt ¼ rate of return on fund; Rft ¼ rate of return on a three-month Treasury bill; Rmt ¼ rate of return on the market benchmark as proxied by the KLCI or the EMAS Index;
J ¼ Jensen’s performance coefficient, indicating the risk-adjusted performance of the fund; p ¼ stimate for the systematic risk level of the fund; pt ¼ the random error term. The Jensen’s model assumes that the systematic risk of the fund is constant over time. Such assumption has ignored the fact that fund manager can engage in market timing activities. As such, the Jensen’s model attributes the fund’s performance exclusively to

Malaysian unit trust funds’ performance 157

MF 33,2

158

the security selection ability of the fund manager. As a matter of fact, the performance of a fund can be the result of the manager’s selection ability, market timing ability or a combination of the two abilities. In reality, fund manager is able to adjust the risk level of the fund by changing the fund’s asset holdings when the market condition changes. For instance, if the fund manager predicts that the market return will rise in the coming period, he will restructure the composition of his fund by holding high risk securities in order to maximize the return of the fund. Conversely, if the market return is forecasted to fall in the coming period, the manager will position his fund accordingly by holding low risk securities in order to minimize the loss. Thus, given the existence of market timing activities among fund managers, the systematic risk of the fund will no longer be constant[6]. While popular, the Jensen model has a major drawback, that is, the potential for misinterpretation may arise when the fund manager engages in market timing activities. In particular, if the manager is a good market timer and his ability is not accounted for in the Jensen’s model, the resulting performance estimate will overestimate the selection ability of the manager. Similarly, a poor market timing manager will cause a downward bias to the selectivity performance of the fund[7]. In addressing the possibility that fund manager may engage in market timing activities, Henriksson and Merton (1981) developed a model which simultaneously test for the presence of market timing and security selection abilities of a fund manager. Their model provides a more complete evaluation of the fund performance than did the Jensen’s model. In particular, the Henriksson and Merton’s approach can remove certain biases in the Jensen’s performance estimate when the fund manager’s market timing ability is not taken into account. The model can be stated as follows[8]: Rpt
Rft ¼
S þ UP Xut þ DOWN Xdt þ pt

ð2Þ

where Xut ¼ max [0, Rmt
Rft] is the market risk premiums during up market conditions; Xdt ¼ min [0, Rmt
Rft] is the market risk premium during down market conditions;
S ¼ component of the fund’s return attributable to the manager’s security selection ability;  UP ¼ systematic risk of the fund during up market conditions;  DOWN ¼ systematic risk of the fund during down market conditions; and pt ¼ the random error term. The Rmt is the rate of return on the market benchmark as proxied by the KLCI or the EMAS Index and the Rft is the rate of return on a three-month Treasury bill. Two target risk levels are assumed in the Henriksson and Merton’s model, depending on the prediction whether or not the return of the market benchmark exceeds the return of a risk free security. An up market condition refers to a situation when the market return exceeds the risk free rate, and a down market condition is a situation when the market return is less than the risk free rate. Equation (2) regresses the excess return on the fund on the bull and bear market risk premiums (Xut and Xdt, respectively) to estimate the up and down market beta for the fund. The market timing ability of the manager is measured by the change in the risk level of the fund from an up to a down market condition. A good market timing manager should have an up market beta that is significantly greater than the down market beta. The intercept
3 provides an estimate for the selectivity ability of the manager after filtering out his market timing ability.

This paper employed the Jensen’s model to estimate the overall fund performance and analyze the sensitivity of the overall fund’s return to alternative specification of the market benchmark. Since the investment performance of a fund can be driven by the manager’s selection ability, timing ability or a combination of both abilities, it is the objective of this paper to investigate whether inferences about the manager’s selectivity and timing performance are affected by the chosen market benchmark. The Henriksson and Merton’s (1981) model is employed to identify the separate contribution of selectivity and timing components to the overall fund performance.

Malaysian unit trust funds’ performance 159

Empirical results and discussion Table I shows the average monthly returns for the fund and market benchmarks during the five-year period, 1996-2000. The results indicate that the KLCI has the highest mean return and it is the only portfolio with a positive average return. The full sample and sub-sample of fund have a lower mean return than the Malaysian stock market, proxied by the KLCI or the EMAS Index. The average monthly return of the KLCI and the EMAS Index are 0.037 and
0.250 per cent, respectively. The full sample of fund shows an average monthly return of
0.074 per cent. Table I also reports the systematic risk measures for the funds using different market benchmarks. The results indicate that the average betas for the funds are generally similar irrespective of the market benchmarks used. This suggests that alternative market benchmark has minimal impact on the systematic risk of the fund. While the funds have similar average betas using the two market benchmarks, the results report consistently higher value of R-squared when the EMAS Index is used. R-squared measures the percentage of the fund’s movements that can be explained by movements in the market benchmark and it can be used to gauge the reliability of the beta coefficient. A high R-squared suggests that beta is a reliable measure of the fund’s volatility in relation to its respective market benchmark. The findings show that EMAS Index produces R-squared that range from the low of 71.4 per cent to the high of 80.5 per cent whereas the KLCI generates R-squared values ranging from 66.7 to 74.8 per cent. Based on the R-squared figures, it can be inferred that EMAS Index explains the fund’s return movement better than the KLCI. That is, EMAS Index seems to reflect the investment performance of the funds better than the widely used KLCI. Since the portfolio holdings of fund manager also include other blue chip stocks not included in the KLCI, it is, therefore, not surprising that the EMAS Index can capture the investment attributes of the funds better than the KLCI. Table II presents the Pearson correlation coefficients between the monthly returns of the funds and the two market benchmarks. The figures reveal that the correlation

Portfolio Market KLCI EMAS index Fund Full sample Income Growth Balanced

Mean monthly return

KLCI Average 

R2

EMAS index Average  R2

0.00037
0.00250

1 –

1 –

– 1

– 1


0.00740
0.00959
0.00268
0.00994

0.64965 0.68378 0.63145 0.48632

0.74650 0.74803 0.77210 0.66655

0.64208 0.67422 0.62554 0.48666

0.78277 0.78376 0.80541 0.71433

Table I. Summary statistics using KLCI and EMAS index as market benchmarks

MF 33,2

160

Table II. Correlation coefficients between monthly returns: unit trust funds and market benchmarks

Table III. Mean values of performance estimates across market benchmarks

between the KLCI and the EMAS Index is very high, that is 0.98. This very high correlation suggests that both the market indices can serve as substitutes for one another. However, since the reliability of beta coefficient is gauged by R-squared, and given that EMAS Index produces higher R-squared than the KLCI, the beta values generated using EMAS Index are more reliable than those obtained using the KLCI. Over the five-year period, the correlation of the KLCI to income fund was about 94, 92 and 80 per cent correlation to growth and balanced fund, respectively. The EMAS Index had a 96 per cent correlation to income fund, a 94 per cent correlation to growth fund and a 85 per cent correlation to balanced fund. Both the KLCI and EMAS Index correlate most with income fund and least with balanced fund. While the correlation structure of the market benchmarks suggests that both the KLCI and the EMAS Index can be regarded as close substitutes, as reported earlier in Table I, their mean monthly return figures are vastly different from each other with the KLCI and EMAS Index having a mean monthly returns of 0.037 and
0.25 per cent, respectively. Furthermore, the correlation results in Table II suggest that EMAS Index correlates better with the fund’s return for both the full sample and sub-sample of funds. Such finding is consistent with the R-squared figures reported in Table I, implying that EMAS Index can explain the movements of the fund’s returns better than the KLCI. Table III reports the mean values of performance estimates from Equation (1) which ignore the market timing component and Equation (2) which considers market timing and security selection simultaneously using both the KLCI and EMAS Index. Consistent with Leong and Aw (1997), this study finds that the funds display negative overall performance, on average, with either the KLCI or the EMAS Index. The overall

KLCI EMAS index

Full sample

Income

Growth

Balanced

KLCI

EMAS index

0.9198* 0.9525*

0.9369* 0.9643*

0.9195* 0.9425*

0.8020* 0.8483*

1 0.9831*

0.9831* 1

Notes: * Denotes statistical significance at the 0.01 level; ** denotes statistical significance at the 0.05 level

Type of fund

Overall performance

Selectivity performance

Up market beta

Down market beta




Panel A: KLCI Full sample Income Growth Balanced


0.00671*
0.00885*
0.00160
0.00738**


0.00071
0.00078 0.00080
0.00439

0.59024* 0.60218* 0.61625* 0.44409

0.73908* 0.79681* 0.68033* 0.53978**


0.14833*
0.19463*
0.06408
0.09569

0.00108 0.00076 0.00282
0.00175

0.58324* 0.59803* 0.60184* 0.43961

0.72219* 0.77497* 0.66754* 0.54263**


0.13896*
0.17694*
0.0657
0.10302

Panel B: EMAS index Full sample
0.00489* Income
0.00675* Growth
0.00009 Balanced
0.00643

Notes: * Denotes statistical significance at the 0.01 level; ** denotes statistical significance at the 0.05 level

performance refers to the Jensen measure and it reflects the performance of the fund when timing activities of the fund manager are not considered. Accordingly, the observed overall performance of the fund consists entirely of the fund manager’s selectivity performance. The results indicate that the full sample of funds exhibit a significant mean risk adjusted returns of
0.671 and
0.489 per cent using the KLCI and EMAS Index, respectively. Since timing activity is ignored in the Equation, these figures attributed the fund’s performance exclusively to the selection ability of the fund manager. In the Henriksson–Merton’s (1981) model where timing and selectivity are considered simultaneously, the market timing ability is captured in the difference of the up market beta over the down market beta (
 ¼ up market beta
down market beta). A fund manager is said to have good (bad) market timing ability if the up-market beta of the fund is significantly greater (lesser) than the down-market beta. The market timing estimates for both the full sample and the income fund show significant negative mean timing values across the two alternative market benchmarks. For the full sample of funds, the average up-market beta is approximately 20 per cent smaller than its down-market beta when the KLCI is used (0.59024 vs 0.73908). The figure for beta differential of up market beta over down market beta is 19 per cent lower when the EMAS Index is used. The significant negative timing estimate for the full sample and income fund together with the insignificant selectivity estimate suggest that the overall negative fund performance is driven by the poor timing ability of the fund manager. Such findings indicate that regardless of the market benchmarks used, on average, fund manager do not possess good market timing ability. The implication of these results is that, at least on average, the specification of the market benchmarks does not appear to make a difference on the selectivity and market timing performance of the fund manager. Table IV reports the frequency counts of positive and negative regression estimates obtained from Equations (1) and (2) using both the KLCI and EMAS Index. At the individual fund level, in the model, which ignores market timing, the frequency counts show that while the number of positive and negative selectivity estimates differs slightly across the two market benchmarks, the number of significant positive and negative estimates remains the same. Such findings indicate that in the model that ignores market timing activities of fund manager, the funds show similar performance using the KLCI and EMAS Index.

Model

Market benchmark: KLCI Selectivity Timing Positive Negative Positive Negative

Timing ignored Full sample Income Growth Balanced

(Equation 7(0) 2(0) 5(0) 0(0)

(1)) 33(8) 23(8) 6(0) 4(1)

Timing considered (Equation(2)) Full sample 22(2) 18(4) Income 13(2) 12(2) Growth 7(0) 4(1) Balanced 2(0) 2(1)

– – – – 11(2) 6(1) 4(1) 1(0)

– – – – 29(15) 19(11) 7(3) 3(1)

Malaysian unit trust funds’ performance 161

Market benchmark: EMAS Index Selectivity Timing Positive Negative Positive Negative

9(0) 4(0) 5(0) 0(0) 22(2) 13(2) 7(0) 2(0)

31(8) 21(8) 6(0) 4(1) 18(4) 12(3) 4(0) 2(1)

– – – – 11(4) 6(3) 4(1) 1(0)

– – – – 29(14) 19(11) 7(2) 3(1)

Table IV. Frequency counts of positive and negative regression estimates across market benchmarks

MF 33,2

162

Table V. Pair-wise Pearson correlation coefficients between performance estimates using different market benchmarks

Based on the Henriksson and Merton’s (1981) model which consider timing and selectivity simultaneously, the frequency counts indicate that individually, some funds show evidence of significant abilities to select undervalued securities and time market movements. For the full sample of fund, the number of significant positive and negative estimates for the selectivity performance is exactly the same using the KLCI and EMAS Index. For the full sample of fund, a total of six funds exhibit significant selectivity performance using both the KLCI and EMAS Index, with four funds having negative selectivity performance and two funds with positive selectivity performance. For the income fund, there are slightly more funds having negative selectivity performance when EMAS Index is used (3/25 ¼ 12 per cent). With the KLCI, 8 per cent (2/25) of the income fund exhibits negative selectivity estimates. At the individual fund level, the frequency counts for negative timing estimate are much higher than the positive counts, that is, 29 out of the 40 funds studied. This is true for both the KLCI and EMAS Index. These results imply that fund managers should seriously reassess their market timing efforts because it seems that their predictions are very often in the wrong direction than in the right direction. The KLCI is shown to produce a slightly higher number of significant negative timing estimate (15/40 ¼ 37.5 per cent) than does the EMAS Index (14/40 ¼ 35 per cent). The results reveal weak evidence of positive timing skills among fund managers. When the EMAS Index is used, 10 per cent of the funds (4/40) reported significant positive timing ability. On the other hand, when the KLCI is used, only 5 per cent of the funds have significant positive timing performance. For the income fund, 12 per cent (3/25) of the funds has positive timing ability using the EMAS Index and 4 per cent of the funds has positive timing performance when the KLCI is used. These findings suggest that when the funds are evaluated from the perspective of the EMAS Index, the funds seem to exhibit more significant positive timing performance and less incidence of negative timing performance. The findings for the growth and balanced funds are qualitatively similar. Table V presents the pair-wise Pearson correlation coefficients between the performance estimates using different market benchmarks. As reported, irrespective of the market benchmarks used, the overall fund performance is not significantly correlated with selectivity but is significantly positively correlated with the timing performance. Overall performance and selectivity

Overall performance and timing

Timing and selectivity

Panel A: KLCI Full sample Income Growth Balanced

0.0203 0.0552
0.3357
0.0632

0.4917* 0.4657* 0.5788 0.1801


0.8557*
0.8568*
0.9603*
0.9732**

Panel B: EMAS index Full sample Income Growth Balanced

0.0461 0.0530
0.3619 0.1077

0.5140* 0.4925* 0.6628** 0.1777


0.8309*
0.8402*
0.9374*
0.9592**

Notes: * Denotes statistical significance at the 0.01 level; ** denotes statistical significance at the 0.05 level

Consistent with the regression results reported earlier, such findings suggest that the selection ability of the fund manager is not related to the overall fund performance. Instead, the market timing ability of the manager is found to be positively linked to the overall performance of the fund. In addition, the results also indicate a significant strong negative correlation between the timing and selectivity estimates, suggesting that managers with good selection ability tend to be poor market timers. Such findings are consistent with those of previous studies, suggesting the presence of activity specialization among fund managers. Conclusion This paper investigate the sensitivity of the performance components of unit trust funds (market timing and selectivity) to alternative specification of the market benchmarks. The two benchmarks used are the KLCI and the EMAS Index. The paper employed a model developed by Henriksson and Merton (1981) to simultaneously test for the presence of market timing and or security selection ability of fund manager. The traditional Jensen’s (1968; 1969) model is used to estimate the fund’s overall performance. The findings reveal that, regardless of the market benchmarks used, on average, the funds display negative overall performance. However, it is reported that both the full sample and the subsamples of fund track the movements in the EMAS Index better than the widely used KLCI. In addition, the correlation between the fund’s return and the return of the EMAS Index is also shown to be stronger than the one with the KLCI. When the performance is broken down into market timing and selectivity components, the findings suggest that, on average, inferences about fund manager’s market timing and selectivity performance are not significantly affected by the chosen market benchmarks. It is reported that the manager’s poor timing ability contributes significantly to the fund’s negative overall performance. Such results suggest no economic benefits accrues to the average fund manager involving in market timing activities. Although findings at the individual fund level show some evidence of positive timing performance, the evidence is generally weak. In conclusion, the overall findings on the separate components of fund performance (timing and selectivity) indicate that using the KLCI or the EMAS Index per se does not seem to matter much. Notes 1. Grinblatt and Titman (1989a) indicated that although some mutual funds exhibit above normal performance based on gross returns, they fail to do so after considering all expenses. For further review on the performance of mutual fund, see Ippolito (1993). 2. Other studies providing evidence of superior selection performance by fund managers are Chang and Lewellen (1984), Henriksson (1984), Bello and Janjigian (1997) and Kao et al. (1998). 3. Grinblatt and Titman (1993) suggested an approach to performance evaluation that does not require the use of a benchmark portfolio. Instead, they utilized information about the portfolio holdings of the evaluated mutual fund. In particular, their approach can eliminate problems associated with inefficient benchmark portfolios. 4. Because the monthly return of the fund is calculated based on the sum of distributions and the change in net asset values over time, the rate of return therefore reflects net return after the deductions of operating expenses, fees and transaction costs. The return is however gross of sales fees (load charges). According to Jensen (1968), since the primary focus of the study is to assess the fund performance in terms of the manager’s

Malaysian unit trust funds’ performance 163

MF 33,2

5. 6.

164

7.

8.

forecasting ability and not to measure the fund performance from the viewpoint of an investor, the load charges were excluded from the return calculations. We estimated the monthly equivalents of the annualized yield as a geometric mean, that is (1 + Annualized Yield)1/12
1. Examples of empirical evidence on the non-stationarity of the mutual fund’s risk over time are provided by Kon and Jen (1978; 1979), Fabozzi and Francis (1979), Miller and Gressis (1980), among others. The results of Chang and Lewellen (1984), Henriksson (1984) and Lee and Rahman (1990) show that when timing is ignored, the selectivity performance of fund manager has the tendency to display lower value. Their findings support Grant’s (1977) assertion that the presence of market timing activities will affect the results of performance tests that concentrate only on the manager’s security selection ability. The Henriksson and Merton’s (1981) model is derived from the basic model of market timing developed by Merton (1981) and is given by the following regression equation: Rpt
Rft ¼
S þ  1Xt þ  2Yt þ pt where Xt ¼ Rm
Rft, Yt = max[0,
(Rm
Rft)]. While the equation provides an estimate for the timing performance  2, it does not explicitly show the separate risk levels of the fund during up and down market conditions. Through a linear transformation, Henriksson and Merton (1981) show that the separate risk levels can be identified in an alternative version of the basic model, as shown by Equation (2).

References Annuar, M.N., Shamsher, M. and Ngu, M.H. (1997), ‘‘Selectivity and timing: evidence from the performance of Malaysian unit trusts’’, Pertanika Journal of Social Science and Humanities, Vol. 5 No. 1, pp. 45-57. Bank Negara Malaysia (January 1996-December 2000), Monthly statistical bulletin, various issues. Bello, Z.Y. and Janjigian, V. (1997), ‘‘A reexamination of the market-timing and security-selection performance of mutual funds’’, Financial Analysts Journal, Vol. 53, pp. 24-30. Block, S.B. and French, D.W. (2002), ‘‘The effect of portfolio weighting on investment performance evaluation: the case of actively managed mutual funds’’, Journal of Economics and Finance, Vol. 26, pp. 16-30. Brown, K.C. and Brown, G.D. (1987), ‘‘Does the composition of the market portfolio really matter?’’, Journal of Portfolio Management, Vol. 13, pp. 26-32. Carlson, R.S. (1970), ‘‘Aggregate performance of mutual funds 1948-1967’’, Journal of Financial and Quantitative Analysis, Vol. 5, pp. 1-32. Chang, E.C. and Lewellen, W.G. (1984), ‘‘Market timing and mutual fund investment performance’’, Journal of Business, Vol. 57, pp. 57-72. Chen, C.R., Lee, C.F., Rahman, S. and Chan, A. (1992), ‘‘A cross-sectional analysis of mutual funds’ market timing and security selection skill’’, Journal of Business Finance & Accounting, Vol. 19, pp. 659-75. Chua, J.E., Koh, S.K. and Koh, F. (1985), ‘‘Unit trusts in Singapore’’, Securities Industry Review, Vol. 11, pp. 70-88. Coggin, T.D., Fabozzi, F.J. and Rahman, S. (1993), ‘‘The investment performance of U.S. equity pension fund managers: an empirical investigation’’, Journal of Finance, Vol. 48, pp. 1039-55. Connor, G. and Korajczyk, R.A. (1991), ‘‘The attributes, behavior, and performance of U.S. mutual funds’’, Review of Quantitative Finance and Accounting, Vol. 1, pp. 5-26.

Cumby, R.E. and Glen, J.D. (1990), ‘‘Evaluating the performance of the international mutual funds’’, Journal of Finance, Vol. 45, pp. 497-521. Daniel, K., Grinblatt, M., Titman, S. and Wermers, R. (1997), ‘‘Measuring mutual fund performance with characteristic-based benchmarks’’, Journal of Finance, Vol. 52, pp. 1035-57. Droms, W.G. and Walker, D.A. (1994), ‘‘Investment performance of international mutual funds’’, Journal of Financial Research, Vol. 17, pp. 1-14. Elton, E.J., Gruber, M.J., Das, S. and Hlavka, M. (1993), ‘‘Efficiency with costly information: a reinterpretation of evidence from managed portfolio’’, Review of Financial Studies, Vol. 6, pp. 1-22. Eun, C.S., Kolodny, R. and Resnick, B.G. (1991), ‘‘U.S.-based international mutual funds: a performance evaluation’’, Journal of Portfolio Management, Vol. 17, pp. 88-94. Fabozzi, F.J. and Francis, J.C. (1979), ‘‘Mutual fund systematic risk for bull and bear markets’’, Journal of Finance, Vol. 34, pp. 1243-50. Firth, M.A. (1977), ‘‘The investment performance of unit trusts in the period 1965-1975’’, Journal of Money, Credit and Banking, Vol. 9, pp. 597-604. Grant, D. (1977), ‘‘Portfolio performance and the cost of timing decisions’’, Journal of Finance, Vol. 32, pp. 837-46. Grinblatt, M. and Titman, S. (1989a), ‘‘Mutual fund performance: an analysis of quarterly portfolio holdings’’, Journal of Business, Vol. 62, pp. 393-416. Grinblatt, M. and Titman, S. (1989b), ‘‘A comparison of measures of mutual fund performance on a sample of monthly mutual fund returns’’, Journal of Business, Vol. 62, pp. 383-416. Grinblatt, M. and Titman, S. (1993), ‘‘Performance measurement without benchmarks: an examination of mutual fund returns’’, Journal of Business, Vol. 66, pp. 47-68. Grinblatt, M. and Titman, S. (1994), ‘‘A study of monthly mutual fund returns and performance evaluation techniques’’, Journal of Financial and Quantitative Analysis, Vol. 29, pp. 419-44. Henriksson, R.D. (1984), ‘‘Market timing and mutual fund performance: an empirical investigation’’, Journal of Business, Vol. 57, pp. 73-96. Henriksson, R.D. and Merton, R.C. (1981), ‘‘On Market timing and investment performance. II. Statistical procedures for evaluating forecasting skills’’, Journal of Business, Vol. 54, pp. 513-33. Investors’ Digest (January 1996-December 2000), Monthly publication of the Kuala Lumpur Stock Exchange, various issues. Ippolito, R.A. (1993), ‘‘On studies of mutual fund performance, 1962-1991’’, Financial Analysts Journal, Vol. 49, pp. 42-51. Jensen, M.C. (1968), ‘‘The performance of mutual funds in the period 1945-1964’’, Journal of Finance, Vol. 23, pp. 389-416. Jensen, M.C. (1969), ‘‘Risk, the pricing of capital assets, and the evaluation of investment portfolios’’, Journal of Business, Vol. 42, pp. 167-247. Kao, G.W., Cheng, L.T.W. and Chan, K.C. (1998), ‘‘International mutual fund selectivity and market timing during up and down market conditions’’, Financial Review, Vol. 33, pp. 127-44. Koh, F., Koh, S.K. and Cheng, T.C. (1987), ‘‘An empirical analysis of the performance of unit trusts in singapore’’, Securities Industry Review, Vol. 13, pp. 1-13. Kon, S.J. (1983), ‘‘The market-timing performance of mutual fund managers’’, Journal of Business, Vol. 56, pp. 323-47.

Malaysian unit trust funds’ performance 165

MF 33,2

166

Kon, S.J. and Jen, F.C. (1978), ‘‘Estimation of time-varying systematic risk and performance for mutual fund portfolios: an application of switching regression’’, Journal of Finance, Vol. 33, pp. 457-75 Kon, S.J. and Jen, F.C. (1979), ‘‘The investment performance of mutual funds: an empirical investigation of timing, selectivity, and market efficiency’’, Journal of Business, Vol. 52, pp. 263-89. Lee, C.F. and Rahman, S. (1990), ‘‘Market timing, selectivity, and mutual fund performance: an empirical investigation’’, Journal of Business, Vol. 63, pp. 261-78. Lehmann, B.N. and Modest, D.M. (1987), ‘‘Mutual fund performance evaluation: a comparison of benchmarks and benchmark comparisons’’, Journal of Finance, Vol. 42, pp. 233-65. Leong, K.H. and Aw, M.W. (1997), ‘‘Measuring unit trust fund performance using different benchmarks’’, Capital Market Review, Vol. 5, pp. 27-44. Low, S.W. and Noor A. Ghazali (2005), ‘‘An evaluation of the market-timing and securityselection performance of mutual funds: the case of Malaysia’’, International Journal of Management Studies, Vol. 12 No. 1, pp. 97-115. Malkiel, B.G. (1995), ‘‘Returns from investing in equity mutual funds 1971-1991’’, Journal of Finance, Vol. 50, pp. 549-72. McDonald, J.G. (1974), ‘‘Objectives and performance of mutual funds’’, Journal of Financial and Quantitative Analysis, Vol. 9, pp. 311-33. Merton, R.C. (1981), ‘‘On market timing and investment performance. I. An equilibrium theory of value for market forecasts’’, Journal of Business, Vol. 54, pp. 363-406. Miller, T.W. and Gressis, N. (1980), ‘‘Nonstationarity and evaluation of mutual fund performance’’, Journal of Financial and Quantitative Analysis, Vol. 15, pp. 639-54. Rao, S.P.U. (2000), ‘‘Market timing and mutual fund performance’’, American Business Review, Vol. 18, pp. 75-9. Roll, R. (1978), ‘‘Ambiguity when performance is measured by the securities market line’’, Journal of Finance, Vol. 33, pp. 1051-69. Shamsher, M. and Annuar M.N. (1995), ‘‘The performance of unit trusts in Malaysia: some evidence’’, Capital Market Review, Vol. 3, pp. 51-69. Sharpe, W.F. (1966), ‘‘Mutual fund performance’’, Journal of Business, Vol. 39, pp. 119-38. Tan, H.C. (1995), ‘‘The investment performance of unit trust funds in Malaysia’’, Capital Market Review, Vol. 3, pp. 21-50. Veit, E.T. and Cheney, J.M. (1982), ‘‘Are mutual funds market timers?’’, Journal of Portfolio Management, Vol. 8, pp. 35-42. Volkman, D.A. (1999), ‘‘Market volatility and perverse timing performance of mutual fund managers’’, Journal of Financial Research, Vol. 22, pp. 449-70. Zimmermann, H. and Wetter, Z. (1992), ‘‘On detecting selection and timing ability: the case of stock market indexes’’, Financial Analysts Journal, Vol. 48, pp. 80-3. Corresponding author Soo-Wah Low can be contacted at: [email protected]

To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

Call for papers International Journal of Managerial Finance Special issue: Corporate Governance Guest Editors: Igor Filatotchev, Kings College, London, UK, and Elisabeth Dedman, Manchester Business School, UK Deadline: May 31, 2007 The scandals at Enron, WorldCom and Parmalat have placed the corporate governance systems of modern corporations around the world under closer scrutiny than ever. Lapses in the personal and professional integrity of accounting firms and their corporate clients have undermined confidence in capital markets and led to substantial erosion of trust in institutions of modern capitalism. As a result, investors and regulators are forcing companies to improve disclosure policies, to rethink their relationships with auditors and strengthen corporate boards as part of a wide ranging reform of corporate governance. Although scholars have developed a substantial body of research on various governance-related topics, far more research is needed. International Journal of Managerial Finance invites paper submissions for a special issue devoted to corporate governance. The issue aims to strengthen our understanding of corporate governance including such aspects as performance outcomes of different governance configurations, boardroom dynamics, the roles of large-block shareholders, managerial processes, etc. We particularly welcome research that analyzes these issues in an international context.

.

The impact of different corporate governance mechanisms on firms’ cost of capital assessments of recent corporate governance reforms on firm and market behaviour.

.

The role of institutional investors and large-block holders in different regimes.

.

The changing role of the auditor as an external monitor.

.

Harmonisation of corporate governance across the EU.

.

Corporate governance effects in different international settings.

.

The relationship between corporate governance and measures of firm performance.

.

The role of corporate governance in emerging economies.

.

The impact of the EU takeover directive on the market for corporate control.

.

Management compensation.

.

Shareholder protection issues in the US vs the UK.

Submission 31 May 2007

Topics Topics suitable for this special issue include, but are not limited to, the following:

Decisions early September 2007 for presentation at a paper development conference in mid-October 2007 and publication thereafter.