Any opinions, findings, conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of NursingAnswers.net.
The entire mutual fund dataset used is obtained from Bloomberg Terminal located in the Main Library, which provides quite comprehensive monthly gross return data for most mutual fund in Singapore. These observations employed a reinvested income price series which assumes that dividends given by these funds are reinvested. The entire CPF-selected mutual fund list was taken from the latest CPFIS annual report which can be found in the CPF website. To ensure the robustness of these data, the selected funds are restricted to those only available and listed in the Singapore Stock Exchange. Our ﬁnal universe of mutual funds consist of 298 open-end equity mutual funds that exist for at least 60 months between 2000 and 2012. These periods are especially chosen to coincide with the bull and bear phases of the stock market cycle. In addition, the performance of these mutual funds will be segregated into two five year series and one 10 years series, which we take to represent moderate and longer time holding periods.
The whole population of funds are further classiﬁed into three investment categories: Aggressive Growth (180 funds), Balanced Growth (61 funds), and Income Growth (57 funds) for different investment objectives and risk appetite. The given investment category of each fund is directly extracted from its individual fund prospectus. The performance of each investment category is rebalanced at the start of each month to incorporate all fund that have existed throughout that month. In order to have an accurate evaluation of the Singapore Mutual Fund performance using the factor model, it is crucial to choose an appropriate factor mimicking portfolio. As such, the market factor used is the FTSE STI index of total where it includes all reinvested dividends. Excess returns from each factor are calculated by using the one-month Singapore Treasury bill rate which can be obtained from Monetary Authority of Singapore. For size factor, Small capitalization minus Big capitalization (SMB), measured by the difference between the monthly returns on the Morgan Stanley Capital International (MSCI) Singapore market small market capitalization Index and the returns on the FTSE STI index. The value premium, High book to market ratio Minus Low book to market ratio (HML), contrast the difference between monthly returns of the MSCI Singapore value index and the total returns on the MSCI Singapore growth index. These indices are formed by ranking all stocks listed in SGX by its respective book to market ratio. Starting with the highest book-to-market ratio stocks, these are attributed to the value index until 50% of the market capitalisation of the national index is reached. The remaining stocks are attributed to the growth index.
The database used in this research has survivorship biases as it only contains funds that survive throughout the entire sample period. The funds that are CPFIS provides are managed by Mercer investment, when a manager is not performing as expected, the entire return history will be removed. Since the aim of this research is to find out the potential for Singaporeans to invest in CPF included mutual fund, it is not necessary to measure the survivorship of Singapore Mutual Fund data, therefore the return history of these dead funds will not be included to the dataset. Certainly when the data is survivor biased, it draw many important implications about the overall measured performance and its performance persistence. The result would be too optimistic as the performance of the surviving funds to be biased upwards. However, the research find that the performance for both long term and short term funds have only slight mismatch as compared to previous study done by Barras and Keith for both the US and UK mutual fund performance. This will be discussed in the next section. .
The overall market had a huge decline during 2000-2001, the GDP growth rate dropped from 9% to -0.95%. This contrasts with 2002, when the market picked up and become a bullish market with a GDP growth rate of well over 5% in 2002, rising to nearly 10% in 2004. From the beginning of 2002, there was a dramatic change in the environment for mutual fund investment when tax-benefits funds were introduced (Koh, et al., 2010). These created enormous public interest in mutual fund investment, which subsequently brought large cash inflows into the industry.
The results from panel below suggest that, overall, mutual funds generate average raw returns of 34.85% per year, which is below the average return from the market (37.43% per year). Aggressive growth funds produce highest average annual returns at 10.86% compared to balanced growth (7.26% p.a) and income growth (1.38% p.a)
The variability of mutual funds returns over a 12-year period are presented in column 3. Although the returns of mutual funds are lower than the market returns, their returns are also less volatile than investment in the stock market.
Therefore, it can be concluded that during the period 2000 – 2012, mutual fund managers still performed less well than the market, after adjusting for total risk. However, in the bear market mutual funds perform particularly well in relation to the market. Aggressive and balanced funds yield 10.86% and 7.26% per annum respectively. Similarly, the variability of mutual funds is considerably lower than that of the market. In contrast to the period after 2002, when the market was markedly going up, the average raw returns from the market were more than 23% per year. The mutual funds show lower mean returns and standard deviation. The Sharpe ratio of all funds (0.07) in the CPFIS shows better risk adjusted performance than the overall market at 0.06. Aggressive growth funds are the most volatile funds (0.07) among the list of portfolios followed by balanced and income growth funds at 0.02 and 0.06 respectively.
From an investment perspective, the slopes of the three factor model describes a diversified portfolio of passive benchmarks and market risk that mimics the risk exposures of the fund. The regression intercept generates “Alpha” which is the excess return generated by a comparable passive portfolios. A positive intercept (positive alpha) denotes good fund performance and a negative alpha represents poor fund performance.
Table 3 above summarises performance using unconditional Fama French 3-factor model. Similar to results documented in previous literature (Blake & Timmermann, 1998; Koh, et al., 2010), the unconditional estimated alphas for each category are negative, ranging from −0.24% to−0.28% with insignificant p-value varies from 0.05 to 0.31. The alpha in the aggressive growth fund is attributed towards small capitalization and low book to market stocks which are more growth oriented rather than value oriented while the opposite holds for growth and income funds. This suggests that, at both overall and investment policy levels, mutual funds do not have abnormal returns. The alpha of the all-fund portfolio is statistically insignificant below the market by 0.51 percent per month. Looking at the portfolio funds, the average performance of the aggressive growth fund was -0.0027% per month, which performed better than its counterpart fund. (Remove R^2) The adjusted R-squares vary from 89% to 91%. These very high adjusted R-squares are consistent with the literature (e.g. Ferson and Schadt 1996; Sawick and Ong, 2000) and can be interpreted to show that fund managers employ a passive strategy, following the market closely but not performing so well. Since the data used in this study do not include fees and expenses, the underperforming may be due to the expenses and fees, as pointed out in Jensen (1968) and Malkiel (1995). Nevertheless, because the available data are limited, this study is unable to further investigate the performance of mutual funds before deducting fees and expenses.
- Impact of luck on long term performance (skilled funds proportions and skilled percentage needs to be more conservative)
This section begins the empirical analysis by measuring the impact of luck on long term mutual fund performance over range of 12 years, using the monthly return of the three-factor model equation. In the Panel A of Table 4 below shows the estimated proportion of skilled and unskilled funds in the population defined by
πA+, and ˆ
πA-, respectively. Panel B and C shows the proportion of signiﬁcant alpha funds in the right (positive) and left (negative) tail of the t-statistic distribution plot at five different signiﬁcance levels (γ =0.025, 0.05, 0.10, 0.15, 0.20).
In the entire population of 298 funds, it is estimated that there are 32.4% are zero alpha funds. Managers of these funds exhibit stock picking skills that are just sufficient to cover for stocks trading cost and fees. The skilled funds in this population are more than 45% and unskilled turns out to be around 22%. This proportion seems to have inconsistent result with prior literature (Koh, 2010; TSE & Chia, 2000) where the estimated proportion of skilled funds are statistically indistinguishable from zero.
Taking a closer examination in Panel B shows the adjustment of luck is key in understanding the difference between our study and prior research. To elaborate further, there are many signiﬁcant alpha funds in the right tail which peaks at 51.852% of the total population (154 funds) when γ =0.20. When funds are deem significant, it just mean that they have abnormal excess return. This does not mean that the fund managers are truly skilled. Illustrating this point, significant fund in panel B are further decomposes into lucky (
Fγ+) and skilled funds (
Tγ+). Among the 154 funds that are truly skilled, around 6.5% of them are lucky and 45.3% of them are skilled. Panel C shows the left tail of the funds which mainly comprising unskilled funds generally perform poorly over their entire lives, it would be interesting to understand how these funds can survive over such a long period of time. Perhaps, as discussed by (Elton, et al., 2002), such funds exist if they are able to attract a sufﬁcient number of unsophisticated investors, who are also charged higher fees (Berk & Green, 2004).
The main reason for the disparity of performance between previous studies and CPFIS long term performance could be due to the fact that prior research uses parts of the Singapore equity mutual fund to conduct this research, however, the dataset of this research were retrieved from CPFIS selected mutual funds where these portfolios are reviewed frequently to ensure that it meets the CPF board defined screening criteria (CPFIS, 2017). Koh, et al. (2010) have shown that funds that satisfy the screening criteria earned higher average returns and stronger persistence in performance than their excluded fund counterparts. This could certainly explain the increase level of positive skilled funds in the mutual fund industry. To fully address this issue, in the following section, this essay will investigate whether funds exhibit superior true skilled alpha over the short run.
In order to test for mutual fund performance in the short run, the entire mutual fund history was partitioned into two sub-periods of 6 years each. The first sub-period begins with 2000 to 2005 and the second sub-period begins with 2006 to 2011. Each sub-period was treated as a separate portfolio with at least 72 monthly return observations and its respective alpha p-values. The different estimator shown in the table below are computed from the pooled p-values of 330 funds across all 5 year periods.
Results in table above show that this is indeed the case with major distinct difference in the statistics reported in Table 4 of the previous section. Reviewing Panel A of Table 5 shows that a small fraction (4.1% of the population) of funds display skill over the short run. This result seems consistent with various literature (Barras, et al., 2010; Cuthbertson, 2004; Cuthbertson, et al., 2008) stating that the probability of short term exceptional performance in mutual funds is slim, however, it does exist as opposed to long term performance.
The most striking characteristic about the best performing and worst performing funds revealed by the model is the relative high
FDRγ+for the best funds and low
FDRγ-for the poor performing funds. This trend is true for any sampled significance level. For instance, when the significance level γ = 0.1, about only 43 funds (10.3%) have significant positive alpha but given
FDRγ+= 63.632% is much higher than
FDRγ-= 24.44%, only
Tγ+= 3.75% (12 funds) have truly positive alpha. So this finding indicates that 43 funds are significant, however, this does not include false discoveries which result in only 12 funds are truly skilled. As both
FDRγ+increase with γ, the percentage of the truly skilled funds
Tγ+remains stagnant for γ ≤0.1. This indicates that the few skilled funds that truly outperform their 3F benchmarks, while rare, appear to be concentrated in the extreme high t-statistics (extremely low p-value).
As shown in Panel C, the short-term results are similar to the prior-discussed long-term results: the great majority of left tail funds are truly unskilled. For example, for γ =0.05,
FDRγ-is relatively small at 14.28% so of the
Sγ-= 26.5% significant poor performing funds,
Tγ-= 23.18% (78 funds) are truly unskilled. As oppose to the right tail, notice that the proportion of truly unskilled funds
Tγ-increases with γ, indicating that the poor performing funds are fairly evenly spread throughout the right tail of the performance distribution. As explained by (Brown, et al., 2014) , high turnover right situated in the left tail funds seems to suggest that unskilled managers performs regular trading which ultimately hurt their performance. Perhaps in the short run after deducting trading cost and other expenses, the fund return did not appear to outperform the market.
It would be useful for investor to know the return performance of different mutual fund according to their risk and return proﬁles outlined in individual prospectuses. Analysis shown that there are some minor differences in the quantitative result when analysing each mutual fund apply to their separate risk adversity portfolios. Table 4 report results for the aggressive growth, balanced growth and income growth funds where we see that the funds generally provided realized returns and risk consistent with their prospectus proﬁles. For each of the four funds, high
FDRγ+for the best funds and a low
FDRγ-for the worse fund were determined (for all significant levels).
Similar to the test for overall mutual fund sample in the previous sections, this part will be delegated to conduct long term performance test for individual investment objectives. The first column reports the performance of aggressive growth, balanced growth and income growth respectively.
Aggressive growth funds (180 in total) shows similar result to the overall universe of funds, as the long term overall universe of funds, as 65.6% of the funds are statistically significant positive alpha funds and 34.4% of the funds are negative alpha funds. While the vast majority of these funds produce zero alpha (
πo=75.5%), only a small proportion of funds have long term skills (T+ = 2.78%). Around 5 managers are sufficiently skilled to more than compensate for these additional trading cost. Onto Panel C of the same fund, there are about 4.78% of the funds are unskilled which are unable to pick stocks well enough to cover their trading expenses and cost. Column B reports the performance of Growth funds (61 funds) which have pretty much the same result as the short term overall universe of funds. As 76.5% of the funds are zero alpha funds, the rest of the population (23.5%) comprises of skilled funds which are concentrated on the extreme right tail and unskilled funds spreading evenly across the left tail of the cross sectional t-distribution. Finally the last column shows the result for Income Growth funds (57 funds). This category produces the lowest performance: it produces the lowest fraction of significant alpha (only 15.8%) and it also contains the highest proportion of negative alpha fraction (84.2%). These low risk funds invest in dividend stocks and long term fixed deposits which produces returns that are much lower than benchmarks and susceptible to inflation rates. Despite having the lowest expense and trading cost, the result reveal that these managers do not have skill to produce positive performance in the long run. These result are consistent with the investors of both the US (Christopherson, et al., 1998) and UK funds (Blake & Timmermann, 1998), use of the FDR demonstrates that the proportion of skilled funds are limited but the unskilled funds are plentiful and easy to discover.
In previous research, the test for persistence often use fractal portfolio formation that are mostly based on past fund performance (Fletcher, 1995; Henriksson & Merton, 1981). However, there is a possible danger that this kind of funds may consist many lucky funds along with genuinely skilled funds.
Ideally, in this section, we would like capture all funds with truly superior alpha and investigate the performance of these skilled funds over time. As we recall from the previous section that truly skilled funds are situated in the extreme right tail of the cross sectional t-statistics distribution, this mean that by forming a portfolio for the right tail funds, it stand a higher chance to locate superior alpha of the skilled funds (Cuthbertson, et al., 2012; Barras, et al., 2010). By explicitly account for the location of the skilled funds using FDR in the right tail, FDR+, we can determine the expected proportion of lucky funds at difference significance level γ using:
This would mean that setting a low value of FDR+ would ensure that only allowing a small amount of lucky funds in the chosen portfolio. Conversely, forming portfolio using a small FDR+ target has two opposing effect: It increases the expected future performance as the fraction of lucky funds in this portfolio is lowered and it decrease the portfolio diversification because there are only a handful of skilled funds being selected (Barras, et al., 2010).
As we are only interested in looking at forming the best performing funds located in the right tail of the distribution, this persistence test looks at two FDR+ target levels X+, namely, x+ = 10% and 20%. Following the works of Barras (2010), the FDR portfolio is constructed using year-end estimated p-values of each existing funds for the previous five years. These p-values are then use to estimate FDR+ over range of significance level
≥0.6), this is done to precisely determine the location of the target FDR+ along the right tail. Then, choosing only funds with p-values smaller than γ(x+) are included in the portfolio. The constructed portfolio is held for a year, after which the selection procedure is repeated. The excess returns from the portfolio are used to estimate its corresponding three factor “forward-looking” alpha.
In Panel A of table 7 reports the empirical FDRs achieved for the two portfolios (FDR10% and FDR20%) in each formation year as well as the proportion of funds that have been included in the portfolio. As expected, the achieved FDR increases with the FDR target assigned to a portfolio which also means an increase in the fraction of funds to be included in a portfolio. However, the test encountered a phenomenon that is also observed in (Barras, et al., 2010) and (Cuthbertson, et al., 2012) research, the empirical FDR+ does not usually match with the target. As the percentage of significant positive alpha is too low, the empirical FDR+ appear to be high even for very low value of γ. For instance, FDR 10% achieves a target of 41.35% during forecasting year 2005. Noticeably, during 2008 and 2009 Asian Financial Crisis (AFC) the empirical target of 10% and 20% reaches 81.68%, only 1 fund meet the criteria of respective portfolios. A highest number of significant funds (31 funds) are included in the FDR10% portfolio during the 2007 economic boom, in this bubble year all stocks have soared. It would be effortless for fund manager to pick a stock that would beat the market. For FDR 10% portfolio, the target is often exceeded and the number of funds in the portfolio varies between 1 and 31 while the FDR 20% portfolio usually hits the target quite well. Taking a closer look at the result, we notice that there is a notable trend in each rebalancing period, the number of funds increases gradually, plunges in AFC and slowly climbing back up again. This trend seems to indicate that mutual fund reflects efficient market hypothesis, the peak and trough of the financial market are well represented in the behaviour of mutual fund movement (Das, et al., 2015).
In Panel B, the table is a summary of the estimated annualised alpha, along with its bootstrapped p value, the three-factor model loadings, annualised mean excess return, the standard deviation of monthly returns, the portfolio’s Sharpe ratio and the information ratio IR. Recall that the portfolio formation uses forward looking positive alpha, this indicates that past skilled funds retain their positive alpha but the performance for both forward looking alphas is not statistically significant. There is an increase in the information ratio as the FDR target increases due to more funds being included lowering the specific risk of the portfolio. The number of funds in the FDR+ portfolios varies over time but the number of funds being added to these FDR controlled portfolio is generally small and each included fund has a relatively low p-value. The result suggest that there is very low correlation in funds’ performance over time and that investors cannot rely on past performance of fund managers to predict their future performance.
Next, the research attempts to take a further step in answering another objective by looking at how the performance of the two persistent portfolio fare against CPF ordinary account interest rate and the return for Strait Times Index. The table below was tabulated based on the annualised gross median wage for average household in Singapore. As we can recall from the previous section that it is mandatory for every working adult in Singapore to contribute 20% of their monthly salary into the CPF-Ordinary Account (CPF-OA), these figures are presented in column 2 of the panel below.
Browsing through the total returns by these 5 investment options, CPF-OA provides a guaranteed interest rate of 2.5% per annum, with a total return of GBP 1283 for a period of 10 years, leads the other four categories in terms of performance. The specially formed portfolio of genuinely skilled funds (FDR10%) produces positively significant return (about GBP 457 for a 10 year period) that is able to beat the market ETF after expenses and trading cost. For an average Singaporean using CPFIS looking to invest long term into funds managed by truly skilled managers, they will be disappointed as these funds have not been able to outperform the standard interest rate given by the government. They will receive 50% less than simply leaving their monies in CPF ordinary account.
The yearly trend chart was plotted based on the annualised return of investment for the four basic investment scheme that Singaporeans can access to. It is notable that there is a dip in returns from 2008 to 2009 for the STI index due to Asian Financial Crisis. Such behaviour have also been observed in (TSE & Chia, 2000) research, he urged that the return for STI index is far too volatile and it is vulnerable to economic cycles which makes it not advisable for pensioners to invest their hard earned money into these Exchange-Traded Funds. An advice giving to pensioners would be simply leaving their wages into CPF-OA gives a more stable, risk free investment with considerably higher return.
An individual interested in investing mutual fund needs to determine the investment period that one is targetting. The table below shows the performance of long term Singapore mutual funds (10 years) and short term funds (5 years).
Evident from table 10 above, CPFIS-selected Singapore mutual fund behaviour in the 10 years long term have shown to improve tremendously over the short term period of 5 years. The resultant skilled funds has increased from 4.1% (13 funds) to 45.4% (49 funds) and the unskilled funds has decreased from 31.3% to 22.2%. This is largely due to the fact that CPFIS will review their selected mutual funds annually against various selection criteria set by the CPF board, this survivorship bias might disguised the true level of persistence. Funds that have passed the screening critieria have demonstrated better stock picking skills and exhibit stronger persistence in performance that can achieve abnormal profits in the long run. In the short run, on the other hand, only a handful of skilled funds (13 funds) exist in the extreme right tail of the cross-sectional t-statistic distribution.
Comparing CPFIS performance with those documented by (Barras, et al., 2010; Cuthbertson, et al., 2012) find that they have zero skilled funds (T+ = 0) over the whole sample period but evidence of a small proportion of skilled funds (T+ = 2−4%) over ‘short-term’ (non-overlapping 5-year) horizons, concentrated in the extreme right tail. For unskilled funds, performance in all countries are very much compliment to each other. Both studies found strong sizeable proportion of unskilled funds in excess of T=15%, spread throughout the left tail. This suggest that ex post performance in winner and loser funds for both research over short-term horizons are similar but there is slight difference in long run capability.
In terms of measuring annualised excess return, investor aiming for long term investment are found to achieve 3 times more profits than short term trading of 0.12%. Similar trend was found in (Sng, 2007) research who measure the mutual fund performance using bootstrap technique found that mutual fund performance in Singapore has higher return in the long run mainly due to lesser expenses and trading cost involved.
A CPF contributor may be interested to know that long term investment are more susceptible to economic cycle and major financial events, it aim to serve investors who are patient to wait for a better return. Short term investment helps impatient investors to gain quick returns as they exchange mediocre rewards with higher risk and volatility.
One of the objective in this research is to discuss whether different investment objectives plays a role in investor returns on mutual funds. Investors are often advised by ﬁnancial planners to choose funds that match their preferred risk proﬁles, where the pools are broadly classiﬁed as, Aggressive growth (High Risk), Balanced growth (Medium to High Risk) and Income growth (Medium to Low Risk) funds. The individual fund statistics and fund performance (skilled versus unskilled) are presented in the table below.
Aggressive growth funds offer higher expected returns and higher risk, as they invest in growth and value stocks with a high potential for capital appreciation but pay relatively low dividends. They have the highest amount of skilled funds (8 funds) but equally highest among of unskilled funds (12 funds). These high volatility funds are therefore usually deemed most suitable for younger investors comfortable with assuming higher risk. Income growth funds, on the other hand, offer low expected returns by assuming low risk as their portfolio usually consist of money market instruments, defensive and income stocks. Conversely, they received the lowest return with alpha lower than market returns by 2% year on year. However, fund managers of such funds do not try to track the stock market benchmark indices or beating the market. This funds try to protect investors saving into accumulating low risk assets. Most advisers deem these lowest risk mutual funds are more suitable for retirees or the very risk averse. Lastly, balanced funds offer investors moderate capital growth and income at moderate risk, mainly investing in value and blue chip stocks earning mediocre returns. Perhaps these funds are most suitable for investors with average risk preferences (Koh, 2010).
One of the key benefits of investing in mutual fund is that they offer investors risk diversification which they might not otherwise be able to achieve on their own. Risk diversification occurs when investors spread their limited capital over a large number of stocks and financial assets (Koh, 2010). Financial analysts often measure risk diversification by the R-square statistic of regression of excess return of funds on the excess returns of the benchmark index (TSE & Chia, 2000; Thissen, et al., 2002). The result for Singapore Mutual Funds shows that all three portfolios are relative well diversified. The average R-squared range from 80.6 to 90.4, this means that over 80 percent of the variantion is explained by movement in the benchmark indices. These results suggest that CPFIS mutual fund do help investors diversify their portfolio risk, a benefit particularly helpful for members with meager savings (Sng, 2007).
Sustainability in mutual fund performance is one of the main concern for pensioners or CPF contributor to decide whether to invest in mutual funds. This research was designed to examine the consistency or persistence of fund manager performance. That is, whether a fund manager who has performed well in one period can repeat this feat in subsequent periods. The result reveal this method leads to the same conclusion about the existence of managerial skill. The evidence of skill that was uncovered can easily attributable to luck because cross sectional differences in skill are not persistent for as long as 10 years into the future. It is evident that the mutual fund industry is heavily manipulated by the natural fluctuations of the economy between periods of contraction and growth (TSE & Chia, 2000). Many of the literatures uses different approach and different country’s performance to determine persistence in return of mutual fund over time (Tonks, 2005; Koh, et al., 2010; Kosowski, et al., 2006; Hendricks, et al., 1993), however, most of them reached a same general consensus that truly skilled fund managers are tough to locate and they decline substantially over time. Another important information that could be drawn from this research is that as compare to the CPF-OA account, the performance of two FDR portfolio have not shown better return than the interest rate given by the CPF board. However, (Koh, 2010) research showed otherwise, he claimed that mutual fund provided average return that were higher than the risk free rate and default interest rate. However, it is worth noting that a successful investment strategy need not necessarily exhibit persistence. A portfolio can be produce positive long run alphas if they trade infrequently when they have genuine superior information while earning zero alpha at other times.
Singapore’s Central Provident Fund (CPF) permits participants to invest their retirement savings in a wide range of investment instruments if they wish, rather than leaving their savings in CPF accounts to earn interest rate by default. The CPF Board has progressively expanded the menu of instruments that members may elect, though for those who took the challenge to invest themselves, the majority found it difficult to beat the CPF default interest rate, and some have even lost money (TSE & Chia, 2000). Such dismal performance led some members to turn to professional fund managers to help them grow their savings (Koh, 2010). This research examines the performance of CPF-included mutual funds in Singapore to determine whether it is successful in helping investors enhance their retirement savings. In testing the performance of these funds, FDR framework was used to estimate the proportion of zero-alpha funds, skilled and unskilled funds in the population and determine their respective locations in the left and right tails of the cross sectional estimated alpha distribution.
Overall, the performance of mutual fund in Singapore generates average raw returns of 34.85% per year, which is below the average return from the market (37.43% per year). Aggressive growth funds produce highest average annual returns at 10.86% compared to balanced growth (7.26% p.a) and income growth (1.38% p.a) only a minority of these funds outperformed their style-specific benchmarks.
Using the Fama French 3-factor model on all selected funds, the unconditional estimated alphas for each category is negative, ranging from −0.24% to−0.28% and they possessed statistically insignificant p-value. This suggests that, at both overall and investment policy levels, mutual funds do not gain abnormal returns. In addition, the adjusted R-squares vary from 89% to 91% between funds of different risk levels. The result shown are consistent with literature (e.g. Ferson and Schadt 1996; Sawick and Ong, 2000) and can be interpreted to understand that fund managers employ a passive strategy, following the market closely but not performing so well.
CPFIS- selected long term mutual fund have shown vast improvement over short term period of 5 years. The resultant long run skilled funds has increased from 4.1% (13 funds) to 45.4% (49 funds) and the unskilled funds has decreased from 31.3% to 22.2%. In contrast to prior research by (Barras, et al., 2010; Cuthbertson, et al., 2008), our result shows relatively low FDR+ for these funds of 5.1% and 6.7% which indicates that T+ =14.9% to 22.5% of funds (i.e. around 16 to 24 funds out of 108 funds) exhibit long-run skill over the life of each fund. We should reckon to the fact that CPFIS policy of only allow surviving funds into the pool and screens out non-surviving and non-performing funds, government’s effort to safeguard Singaporeans from making poor investment decision. For the worst funds, at a 2.5% (5%) significance level, the FDR− is relatively small at 10.37%. Hence the proportion of unskilled funds are fairly substantial (48 funds). In addition, skilled funds tend to be concentrated in the extreme right tail of the performance distribution while the unskilled funds are dispersed throughout the left tail. Overall observation is that the broad majority of mutual fund managers in Singapore do not exhibit superior stock selection skills. Accordingly, savers should not look to these fund managers to outperform market benchmarks.
When the research examined different investment style in the mutual fund industry, the general pattern of few genuine skilled funds is repeated for all aggressive growth, balanced growth and income growth funds. Aggressive have shown the highest amount of significant positive skilled funds but also equally highest in unskilled funds. Balanced growth and income funds show the higher proportion of truly unskilled funds, as majority of the funds in these two category only earn mediocre returns.
The research goes onto form ex-ante portfolios based on a maximum FDR+ of 10% and 20%., however, these positive alpha funds do not exhibit persistence as their forward looking alphas are not statistically different from zero. When comparing the return of two FDR+ portfolio with default CPF-OA interest rate and Strait Times Index, FDR10% beats the market by more than 1%. Accordingly, CPF-OA consistent and risk free long term growth has shown to be the best performer among the rest, constantly beating the market by more than 200% year on year.
Further works can be done on including bot surviving and non-surviving funds to minimise survivorship bias. When the fund manager is dropped due to poor performance, the measured performance of surviving managers is biased upwards. In particular, past poor performers in a sample with survivorship bias are likely to reverse their performance in the future. Some recommendation can be made for CPFIS that is restricting investment portfolios available to participants is likely to be only a partial step toward enhancing workers ﬁnancial security. That is, people often make serious investment mistakes even when they are offered extremely efﬁcient investment choices. As a consequence, it remains essential to enhance fund competitiveness in an effort to drive down costs, and to educate participants regarding the importance of saving, interest compounding and risk/return attributes of their pension investments. Second, this research have made no allowance for the costs of fund management and turnover rate which proves to be driving the mutual fund industry since returns are low. The findings show that there are some fund managers have the ability to generate consistent abnormal returns above the benchmark portfolios, but whether these abnormal returns outweigh the costs of active fund management is not an issue that have been addressed.
Cite This Work
To export a reference to this article please select a referencing stye below:
Related ServicesView all
DMCA / Removal Request
If you are the original writer of this dissertation and no longer wish to have your work published on the UKDiss.com website then please: