CHAPTER-7 PORTFOLIO PERFORMANCE ATTRIBUTION

Transcription

1 i/y CHAPTER-7 PORTFOLIO PERFORMANCE ATTRIBUTION 7.1.INTRODUCTION The measures of risk-adjusted performance discussed in the last chapter are primarily oriented to analysis of the overall performance of funds. It is sometimes useful to develop a more refined breakdown and assess the components, or sources of performance. Portfolio managers and their clients often want to know why a portfolio has performed better or worse than the market or other portfolios. Differences in performance are due to difference in security selection, or market timing, or both. Eugene Fama provided an analytical framework that elaborates on the three previously discussed riskadjusted return methods (Sharpe s, Treynor s, and Jensen s), and allows a more detailed breakdown of funds performance. The current chapter tries to examine the components and sources of performance of the sampled mutual funds and to attribute it to specific activities of portfolio managers METHODOLOGY The difference between the actual performance and risk-free return is the overall performance, (also called excess return). The overall performance is contributed by mutual funds managers ability to pick the best securities at a given level of risk (selectivity). Hence, that part of overall return that is attributed to return on selectivity measures how well the chosen portfolio did relative to a naively selected portfolio with the same level of risk. The remaining part of the overall performance is due to fund managers decision to take on a positive amount of risk. Thus,

2 180 Overall performance=rp-rf Where, RP is actual (realised) return on the portfolio, and Rf is risk-free rate of return. Rp-RF = [Rp-Rm(P)]+ [Rm(P)-Rp] Or Overall = Selectivity +Risk. Performance. Where, Rm is the return on market portfolio, PP is the risk measured by beta or volatility of portfolio under consideration, and Rm(pp) is the return on market portfolio with the risk of chosen portfolio. Further, the overall performance attributed to selectivity can be evaluated in terms of breakdown of selectivity into Diversification and net selectivity. Intuitively, to some extent the portfolio decision may have involved putting more eggs into one or few baskets. That is, the investment manager places his bets on a few securities that he thinks are winners. Thus, diversification measures extra portfolio returns that the winners have to produce in order to

3 181 make concentration of resources in them worthwhile. Diversification is always non-negative, so the net selectivity is equal to or less than selectivity. If net selectivity is not positive, it means that the manager has taken on diversifiable risk that his winners have not compensated for in terms of extra returns. Therefore: [Rp-Rm(p)] = Net selectivity + [Rm(5( Rm))]-[Rm(Pp)], {Selectivity = Net selectivity -(-Diversification} Or, Net Selectivity = [Rp-Rm(p)] - [Rm(o ("Rm))]-Rm(P) =(Selectivity) - (Diversification) Where, Rm [a ('Rm)] is the return on the market portfolio M that has return dispersion (orp) equivalent to that of the actual portfolio chosen. In other words, diversification measures the extent to which a ("Rp)>Pp.The Portfolio manager may decide to take on some portfolio dispersion, which might have been diversified away, in a belief that some securities would pay more if additional resources are deployed therein. This is based on the premise that in the case of perfect diversification, the only risk involved is represented by its volatility. The residual of performance attributed to selectivity is designed as performance on net selectivity. The aforesaid break down of selectivity performance given above is then only relevant and useful consideration in the study. The rest discussion is about

4 182 ingredients of overall performance, i.e. risk. If the investor has a target risk level (pt) for his portfolio, the part of the overall performance due to risk can be allocated to the investor and the portfolio manager as follows: [(Rm -Rf )Pp] = [Rm(Pp) -Rm(Pt)] + [Rm (Pt) - Rf] Or Risk = Manager s Risk + Investor s Risk. Where, Rm(Pt) is the return on the naively selected portfolio with the target level of market risk. Thus, manager s risk is that part of the overall performance and risk which is due to the manager s decision to take on a level of risk pp different from the investor s target risk level pt. Investor s risk, on the other hand is that part of the overall performance that results from the fact that investor s target level of risk is positive. These risk components depend to a large extent upon the investor s target level of risk, which is not easy to determine in the absence of any communication from the investors in this regard. Appropriate presumption on this count has been made. Manager s risk might result from a timing decision. That is due to having chosen a portfolio with a level of risk higher or lower than the target level because he might have felt that risky portfolio in general do abnormally poor during the bear phase and good in the bull phase of the market. Therefore, a more precise measure of the results of such a timing decision can be obtained as follows:

5 18J Manager s Timing [Rm (pp)-rf] = [Rm (pp)- E (Rm (pp))] - [RM (Pi)-E (RM (Pi))] (Total Timing) (Market conditions) + [E (Rm (Pp)> E (Rm (pt))] - [Rm (Pt)-Rf] (Manager s Expected Risk) (Investor s Risk). The first three terms in the above relationship sum up the manager s risk. Performance on manager s expected risk is the incremental expected return from managers decision to take on non-target level of risk. The performance on market Conditions is the difference between the return on naively selected portfolio with target level of risk and the expected return on the portfolio. It answers the question: By how much did market deviate from expectations at the target level of risk? Total Timing performance is the difference between the ex-post return on the naively selected portfolio with risk pp and the ex-post Expected return. It is positive when RM>E (Rm) and negative when RM<E(RM). The performance difference between total timing and Market Conditions is the Manger s timing, which measures excess performance on total timing over timing performance that might have been generated by choosing naive portfolio with target level risk. Performance on manager s timing is only positive when sign of the difference between pp and pt is the same as the sign of difference between RM and E(RM)-That is, it is not so obtained when the chose level of market risk is above(be!ow)the target level and RM is above

6 184 (below) E (Rm). It is (managers timing performance), thus somewhat more sensitive than performance on total timing. Thus, overall mutual fund performance can be attributed to selectivity and risk. The performance contribution of selectivity is decomposed into net selectivity and diversification, whereas performance attributed to risk is contributed by the manager s risk and the investor s risk. Further, performance by manager s risk bearing is contributed by total timing, market conditions, and manager s expected return. The differential return of total timing and the market conditions is assumed to have been contributed by the manager s timing decisions. The performance composition of sampled mutual funds in the aforesaid framework is analysed in relation to the three fund characteristics in the following section. The relative contribution of each of the said activities of portfolio managers will facilitate a more comprehensive assessment of performance of the managed portfolios. The table below shows the sampled funds classified into different components of performance and their respective measures.

7 ISD s. No. Scheme Name Table 7.1 COMPONENTS OF PERFORMANCE Investment objective Excess return (Rp-Rf) Risk Premium (Rm-RF) Pp Selectivity Diversification Rm(o CRm)))- Net Selectivity Rm(P) A B C=(A-B) D E=(C-D) 1 Alliance Income fund Income BOB growth 95 Growth BOI BOINEG Growth CanbankCANGANGA Balanced Canbank CANGLOBAL Balanced Canbank CANPEP 92 Balanced Canbank CANPREMIUM Balanced * Canbank CANTRIPPLE Balanced DSP Merrill Lynch Bond Fund Income Escorts Income fund Income GIC Balanced fund Balanced GIC Growth plus II Growth GIC Fortune 94 Growth * IDBI-PRIN CIPAL equity fund Growth * IDBI-PRINCIPAL Deposit Income * fond 16 Indianbank IND NAVRATNA Balanced * JM Balanced fund Balanced * LIC Dhansahayog-A Growth * LIC Dhansahayog-B Growth * LIC Dhansahayog-C Growth * Morgan Stanley Growth Growth Prudential ICICI growth plan Income Prudential ICICI income plan Income * Prudential ICICI Premier Balanced Reliance Vision Growth SBI MMIS 98(1) Income * SBI MMIS 97 Income * Tata Balanced fund Balanced UTI GCGIP Balanced * UTI Master Gain 92 Growth UTI Master Share Growth UTI Master Value Unit Growth UTI MIP 98(IV) Income UTI Unit Scheme Growth * (growth) 35 UTI Unit Scheme Income (income) 36 Zurich India High Interest fund Income SAMPLE AVERAGE BSE SENSEX Source. Original * - Not doing well (Risk premium is below the market risk premium)

8 18b 7.3.0VERALL PERFORMANCE Overall performance (Excess return) is the difference between return on the chosen portfolio and the risk-less asset. The risk-less return, as pointed out earlier, is the average annual discount rate (return) on 91-day Treasury bills of the Government of India. The overall performance here under reference is same as the excess performance in the risk and return analysis. This measure of mutual fund performance reflects on nature and adequacy of the portfolio returns in a broader perspective across the three characteristics of sampled mutual fund schemes. Even a cursory look on this performance measure enables the analyst to comprehend the level of performance. Due care is, however, required to be exercised before passing any judgment on portfolio performance since it (overall performance) visualizes performance in terms of associated returns alone. The overall performance of sample mutual funds is report in table Table Overall Performance of sampled Mutual Fund Schemes Overall Performance Number of Schemes Percentage < > < > < > Total Overall performance (benchmark)=10.06 percent, Sample = 9.50 percent. It may be seen that majority (50.0 per cent) of sampled schemes have incurred an overall performance lower than that of broader market portfolio (10.06 per cent). Whereas some individual funds have experienced superior performance in overall perspective, on the whole, overall performance has been poor (9.50 per cent).

9 PERFORMANCE ON SELECTIVITY Selectivity measures how well a chosen portfolio does relative to market portfolio with same level of risk (systematic). Selectivity is positive if mutual fund under consideration produces a higher return than the market portfolio with same level of risk (Fama, 1972). It is only then that the management activity relating to stock selection is appreciated. The portfolio managers are preoccupied most of the time in identification and selection of undervalued securities having potential for capital appreciation. As portfolio return is attributed to this activity to a large extent, it becomes imperative to examine this part of overall performance of sampled mutual funds. Table summarises the findings from the sampled funds.. Table Selectivity Performance of Sampled Mutual Fund Schemes (per cent per annum) Selectivity Number of Percentage Performance Schemes < >-5.0 < >0.0 < > Total Average (Sample) = 0.41 It can be noticed that a majority (58.34 percent)of the sampled fund schemes have experienced positive performance on selectivity, while the remaining percent experienced negative performance. It implies that selectivity of the fund managers have been fruitful in the case of percent of the

10 188 schemes. The average selectivity (0.41 percent) of the sampled mutual funds is also positive confirming the statement percent of the sample registered a superior positive performance (>5.00 percent). Mutual fund managers seem to have been rewarded collectively even though some individual fund managers have been penalised which implies their failure to identify superior stocks. 7.5.PERFORMANCE ON RISK. Overall performance is also contributed to by risk in addition to the selectivity. Performance on risk measures return from decision to take on positive risk. It is determined by the level of risk chosen (J3P) relative to return on market portfolio and return on risk-free asset, RF. The mutual fund managers bear risk in expectation that it (risk bearing) contributes to overall portfolio return. Otherwise, they could have invested all their resources in risk-less assets. Their activities in this regard do not always bear fruits, sometimes leading to huge losses. Yet both investors and investment mangers are tempted to bear risk. Table brings out a summary of findings on how sampled funds performed in relation to risk. Table Performance on Risk of sampled Mutual fund Schemes (per cent per annum) Performance Number Percentage Schemes of Schemes < > 8.34 < > 9.34< > Total Average (Sample)=9.34

11 Sampled funds according to this study have fared well on account of risk bearing activity of the fond managers. The risk-attributed performance differs within a narrow range across various fond characteristics. Majority of the sampled funds (58.33 percent) registered a performance above average. As regards fond characteristics, close-end funds have outperformed open-end funds, and growth funds outperformed income and balanced funds. For the sponsorship categories, private sector funds led followed by bank sponsored, OFI, and UTI, in that order. It thus shows a failure on the fond managers expectations, hence an increased agony to the investors. 7.6.PERFORMANCE ON DIVERSIFICATION The portfolio performance attributed to selectivity may be further attributed to diversification and Net Selectivity. Diversification is a process for risk reduction and return magnification through portfolio investment across different securities. Performance attributed to diversification is what an investor should get if only the non-diversifiable (systematic) risk is taken into account The methodological support in this regard as explained earlier in this chapter is used to obtain the results plotted on table Table Performance on Diversification of sampled Mutual Fund Schemes Diversification Number of Schemes Percentage < > 1.90 < >2.90 < > Total Average Performance on Diversification = 1.90

12 of the sampled mutual funds have done well to lose lower than average return in the process of portfolio diversification. As many as percent of the sampled funds have lost up to three percent returns in this pursuit while the remaining percent have lost more than three percent return. 7.7.PERFORMANCE ON NET SELECTIVITY. Net selectivity measures funds performance adjusted for its imperfect diversification The residual performance on selectivity (after accounting for diversification) is what is accredited to net selectivity. Performance on net selectivity is equal to (or less than) that on the selectivity. Portfolio managers efforts are paid for if performance is positive on this count. If net selectivity performance is not positive, it implies that a manager has taken diversifiable risk that his winners have not compensated for in terms of extra return. On this issue, Lee and Rahman (1990) concluded that a mutual fund with no forecasting skill might consider a totally passive portfolio management strategy and just provides diversification services to their shareholders. Whereas the positive performance on net selectivity indicates a substantial improvement in performance of mutual funds managers. An examination carried out in this study gave results, which have been plotted in table Table Net Selectivity Performance of sampled Mutual fund Schemes Performance on Number of Schemes Percentage Net Selectivity < >- 5.0 < >0.0 < >5.0 - Total Average (Sample) = 71

13 From the table it can be seen that in only percent of the sampled funds, fund managers lost returns in an apparent bid to generate superior returns by picking under priced stocks. The remaining majority per cent has generated positive returns through selectivity activities. No funds generated positive return in excess of five percent. The incidence of negative returns was lower (>-5.0 percent) in the case of percent and higher in the case of the remainingl6.66 percent (<-5.0 percent) of the funds in terms of Net selectivity. The analysis thus shows evidence of superior stock selectivity across percent of the sample. The managers in these funds were compensated for the risk. 7.8.PERFORMANCE ON INVESTOR S RISK The risk-led part of overall performance is decomposed into the one attributed to manager s risk and the other attributed to investor s risk. The manager s risk is that part of overall performance of risk which is due to the manager s decision to take on a level of risk (pp) different from the investor s target level risk (pt). Investor s risk, on the other hand, is that part of the overall performance that results from the investor s target level of risk being positive. To identify the investor s target level of risk (pp) is indeed too hard a thing to do. This can be overcome with some additional methodological assumptions. Since mutual fund investment is treated as safe by the investors, risk perception of mutual fund investors is less than market benchmark portfolio. In view of this, average investment risk of sampled funds is considered as investor s target level of risk. This target level of risk has remained uniform regardless of the three fund characteristics. As a result, the analysis has demonstrated that out of portfolio performance attributed to risk, 8.48 per cent funds have lost on investor s risk under the aforesaid assumption across all fund characteristics. This is primarily due to lower return on benchmark portfolio in relation to the risk-less rate. The residual risk attributed performance is contributed by the manager s risk, which is discussed below.

14 PERFORMANCE OF MANAGER S RISK As pointed out earlier, manager s risk is that part of overall performance and of risk which is due to the manager s decision to take on a level of risk ((3T) different from the investors target level risk ((3P). The manager may expose his portfolio to a higher risk class than the investors expected level in anticipation of positive returns. In order to assess whether the fond managers were successful in delivering goods to the investors PERFORMANCE ON TOTAL TIMING Manager s risk results in part from timing decision. A manager may chose a portfolio with a risk level higher or lower than the target level because a risky portfolio would do abnormally well during the relevant time period, given a average return on market portfolio. The results examined on this aspect show that mutual funds have reported negative performance within the range of percent to percent on annualized basis. The calculations in this regard are based on 12 per cent investor s return. This is decided in relation to risk-less rate and expected risk premium keeping in view the return on deposits in the nationalised banks that might have been earned, instead of investing in mutual funds. Since the negative returns in this regard have moved in such a narrow band, it was not considered appropriate to examine it further in relation to the three fond characteristics. Notably, depressed market sentiments were taken as primarily responsible for such a high incidence of negative performance on total timings. This performance is further influenced by market conditions examined below PERFORMANCE ON MARKET CONDITIONS Portfolio performance is also influenced by the prevailing conditions in the securities markets. In the boom period securities tend to yield superior returns and contrary holds in a depressed securities market. The performance on this count enables us to

15 answer: how much a market deviates from expectations at the target level of risk? Performance attributed to market conditions is the difference between return on naively selected portfolio with target level of risk and expected return on the portfolio under consideration. Investigation have revealed that all sampled funds have lost returns uniformly across all fund characteristics to the extent of minus per cent due to depressed conditions in the securities market in the relevant period. This high incidence of negative performance has also been due to investors high expectations. In view of uniform performance across all fund categories, it has not been considered worth the efforts required to examine portfolio performance in relation to chosen fund characteristics PERFORMANCE ON MANAGER S TIMING The differential performance due to total timing and market conditions is attributed to manager s timing efforts. It measures excess of total timing performance over timing performance, which may have been generated by choosing naively portfolio with a target level of risk. Thus, it is that component of return attributed to total timing decisions which is independent of market conditions. In other words, it measures return generated by manager s efforts to timing the portfolio investment in view of prevailing market conditions PERFORMANCE ON MANAGER S EXPECTED RISK Manager s expected risk is the incremental expected return of portfolio managers decision to take on non-target level of risk. It measures the return generated by fund managers by exposing their portfolios to a higher or lower risk level than what the investors may have expected. It is done by portfolio managers to generate positive returns.

16 iy CONCLUSION The overall performance of sampled funds was not considered satisfactory in view of poor performance by market portfolio during period of study. This position remained the same across three fund characteristics. The positive portfolio performance on selectivity was influenced by these characteristics. The private sector, including FII sponsored mutual fund managers failed to deliver goods to investors. Risk attributed portfolio performance was not found to be impressive as fund managers lost returns in a bid to magnify the same. The results indicated absence of any impact of chosen fund characteristics on this component of overall investment performance. Portfolio returns were also not found magnified by the process of diversification by mutual fond managers. Their diversification activity resulted in loss of performance that increased the mutual fond investors woos across all identified fond characteristics. However, incidence of this lost return varied across such categorization of sampled funds. Likewise, performance on net selectivity also contributed in portfolio return reduction, which was significantly influenced by fond characteristics. The sampled funds demonstrated negative performance on investors risk bearing so was on performance due to managers risk bearing efforts. In addition, portfolio managers also unsuccessfully attempted to time market as was evident by negative performance. Performance also suffered due to prevailing market conditions. Investors discontentment with mutual fond investment was farther enhanced by negative returns managers timing efforts. However, managers average expected risk-bearing performance was found to be positive and the portfolio characteristics also affected performance.

17 REFERENCES 1. Fama, E., Efficient Capital Markets: A review of Theory and Empirical Work, Journal of Finance,25(1970): Fama, Eugene F., Components of investment Performance, Journal of Finance, 27(1972): Jensen, M.C., The performance of Mutual funds in the period , Journal of Finance, 23(1968): Chang, Eric C. and Lewellen, Wilbur G., Market Timing and Mutual Funds Investment Performance, Journal of Business, 57(1984): Kon, Stanley J., Market Timing Performance of Mutual Fund Managers, Journal of Business, 56(19830: Williamson, Peter J., Measurement and Forecasting of Mutual Fund Performance: Choosing on Investment Strategy Financial Analysts Journal,28(1972): Lee, Cheng-Few and Rahman, Shaflqur., Market Timing,Selectivity.and Mutual Fund Performance: An Empirical Investigation, Journal of Business,63(l 990) : Ankrim, Ernest M., Risk adjusted Performance Attribution, Financial Analysts Journal,48(1992): Brinson, Gray P., Hood, Rudolph L. And Beebower, Gilbert L. Determinants of Portfolio Performance, Financial Analysts Journal, 51(1995): Merton, R.C., On Market Timing and Investment Performance- I: An Equilibrium Theory of Value fore Market Forecasts, Journal of Business,54(1981): Merton, R.C. and Henriksson, Roy D. On Market Timing and Investment Performance-II, Statistical Procedures for Evaluating Forecasting Skills, Journal of Business, 54(1981):

18 Henriksson, Roy D., Market Timing and Mutual Fund Performance: An Empirical Investigation, Journal of Business, 57(1984): Kane, Alex and Marks, S.G., Performance Evaluation of Market Timer: Theory and Evidence, Journal of Financial and Quantitative Analysis, 23(1988):