The Investment Performance of Dutch Pension Funds
An empirical investigation on the impact of market timing and security selection
on investment performance
Nihaila Dometilia Student number: 1059718 missdometilia@yahoo.com
Master Thesis Portfolio management Faculty of Economics
University of Groningen
Nihaila Dometilia
Supervisor: Dr. A. Plantinga
Acknowledgements
First of all, I want to thank my supervisor, Auke Plantinga, for the many useful suggestions he provided regarding this thesis, together with the exceptional insight with regard to the techniques used and the helpful comments.
Furthermore, I want to express my appreciation for all the researchers who contributed to investment performance analysis, especially to Gary Brinson, Randolph Hood and Gilbert Beebower for their work that inspired this thesis.
Executive Summary
This thesis examines the contribution of the different aspects of the investment management process to the investment performance of a sample of Dutch pension funds, using the framework of Brinson, Hood and Beebower (1986). The components of the investment management process are: investment policy and investment strategy (market timing and security selection).
Following the framework of Brinson, Hood and Beebower, a passive benchmark portfolio (also called investment policy portfolio) is created representing each pension fund’s internal benchmark asset classes, weighted by their internal benchmark allocations. The contribution of market timing and security selection to the pension fund investment performance is calculated as the difference between the actual return and the return on the investment policy portfolio. The actual return is the result of the combination of investment policy and investment strategy. Any significant departure from investment policy represents an exercise of investment strategy by the pension fund investment manager.
Using a data set on 35 Dutch pension funds over the period 2001-2005 this thesis shows that investment policy, dominates market timing and security selection. The average total return on the investment policy portfolio over the period was 22.63 percent, versus 24.39 percent for the actual average total return. Investment policy alone explained on average 91.6 percent of the variation in total return. On average the Dutch pension funds did better then their benchmark over the 5-year sample period. The average pension fund gained 1.76 percent per year from active management. Even though, the effects of active management on the individual pension funds ranged from a low of -12.37 percent per year to a high of 14.96 per cent per year.
Table of content
Chapter 1 General Background ...4
1.1 Introduction ...4
1.2 Dutch Pension funds in general...7
Chapter 2 Literature review...15
2.1 Models of selectivity and timing and mutual fund performance...15
2.2 Models of selectivity and timing and pension fund performance ...18
Chapter 3 Methodology...20
3.1 The framework of Brinson, Hood and Beebower ...21
Chapter 4 Data...23
4.1 Characteristics of sample ...23
4.2 Important economic and stock market developments during sample period ...27
Chapter 5 Results ...32
Chapter 6: Conclusion ...38
Appendix 1 ...39
Chapter 1 General Background
1.1 Introduction
The investment performance of pension funds, unlike mutual funds has not been extensively examined in the finance literature. The purpose of this thesis is to empirically examine the effects of the components of the investment management process, namely: investment policy and investment strategy (market timing and security selection), on Dutch pension funds total portfolio return by using the attribution model of Brinson, Hood and Beebower (1986). The goal is to determine, from past investment data on Dutch pension funds, which investment decisions had the greatest impact on the size of total return and on the variability of that return.
Brinson, Hood and Beebower (1986), further in this thesis referred to as BHB, define investment policy to be a combination of the choice of asset classes and their normal, or long-term weights. Good (1984) better defines investment policy as the asset-allocation plan that best meets the fund's long-term objectives in the absence of active management. Market timing refers to the act of attempting to predict the future direction of the market. If the investment manager believes he can forecast the market return, he will adjust his investment mix weight away from normal in an attempt to capture excess returns from short-term fluctuations in asset class prices. If successful, he will earn abnormal returns relative to an appropriate benchmark.
BHB describes security selection as involving the identification of individual securities within an asset class to achieve superior returns relative to that asset class. Applying the Capital Asset Pricing Model, introduced by Treynor (1961), Sharpe (1964) and Lintner (1965), the investment manager attempts to identify securities with expected returns that lie significantly off the security market line. The manager will then invest in the securities which offer an abnormally high risk premium.
investor at some or at all times. Beasily et al. (2003) also recognize two broad categories of portfolio management. Passive investment management is adopted by investors who believe that financial markets are efficient. Consequently, passive managers believe that it is impossible to consistently beat the market. Their main activity is attempting to achieve the same returns and risk of a certain benchmark. Active investment management on the other hand relies on the belief that skilful investors are able to out-perform the market. An active investment manager applies market timing and security selection in its attempt to out-perform the market. A mixed strategy of passive and active investment is also possible. Both strategies have their advantages and disadvantages when compared.
Grinold and Kahn (2000) and Hodgson, Breban, Ford, Streatfield and Urwin (2000) discussed some of these advantages and disadvantages of which a few will be mentioned in this thesis. One of the advantages of active investment management is that experienced managers may profit from inefficiencies in the market, creating a possibility of superior returns in excess of the fees paid. Another advantage of active investment management is that managers can hedge against the risk that the market index becomes an easy target to beat, limiting peer group risk. Active investment management disadvantages include high transaction and management cost, and exposure to market and company risk. Also during the decision making process of securities selection, many mistakes are possible. Successful security selection and market timing is difficult.
benchmark index exists. Consequently, although the major markets are well covered and there are some indices that can be customized, some smaller, illiquid markets may not be adequately covered1.
The active management opportunities are possible because of the assumption that financial markets are not fully efficient. The efficient market hypothesis states that asset prices in financial markets should reflect all available information, which means that prices should always be consistent with market fundamentals (Fama, 1991). According to the efficient market hypothesis, active investment management activities can not be successful. The only rational choice for a pension fund investment manager is to invest in a passively managed market index. For this reason, in an efficient market, no pension fund investment manager would pay active management fees or rationally invest in an investment program which cannot outperform a market index. However, there are reasons to suspect that there is a very large active pension fund management business in the Netherlands.
This thesis presents an empirical examination of the investment policy return and the selectivity and market timing performance. Researchers have long been attempting to develop methods to distinguish between the ability to time the market and the ability to pick stocks. In their paper “Determinants of Portfolio Performance” (1986), which set a framework for decomposing total portfolio returns, BHB found that on average, investment policy is the primary determinant of portfolio return variability, with security selection and market timing playing minor roles. They examined the quarterly investment returns of 91 large pension plans over a 10 year period (1974-1983), concluding that investment policy explained an average 93.6 percent of the variation in total plan returns. An update to the study in 1991, which used data from 1977 to 1987, similarly found that on average 91.5 percent of returns variability could be explained by investment policy decisions. This paper will play a crucial role in this thesis.
The following chapter will present the relevant empirical works to date. In Chapter 3 the research methodology is discussed. Chapter 4 presents the data and chapter 5 provides an analysis of the empirical results. Finally, Chapter 6 is a conclusion of the empirical results. Before entering upon
1
these detailed issues, some general background information on Dutch pension funds will be discussed in the next sub-chapter.
1.2 Dutch Pension funds in general2
The investment dictionary, Investopedia.com, defines pension fund as a fund established by an employer to facilitate and organize the investment of employees' retirement funds contributed by the employer and employees. The pension fund is a common asset pool meant to generate stable growth over the long term, and provide pensions for employees when they reach the end of their working years and go into retirement. It constitutes the second pillar of the Dutch pension system.
The Dutch retirement pension system is built on three pillars. The first pillar is the basic pension for every person over 65, under the Old Age Pensions Act. The original Dutch term is Algemene
Ouderdomswet; AOW. This benefit is funded by the government according to the pay-as-you-go method. AOW pensions are paid out of current workers’ contribution income and taxes, and no assets are set aside. The second pillar consists of employees’ compulsory membership of occupational, corporate or industry-wide pension schemes (labor unions and government may also fund pensions), with employees saving for a pension in addition to the AOW benefit, according to a funded system. In a funded pension arrangement, the plan sponsor or employer, and employee must make contributions to ensure that the pension fund will meet future payment obligations. The money in the pension fund can grow by means of investment. On an individual basis, workers can build up additional third pillar investments, which comprise tax-supported schemes, such as life insurance.
This paper focuses on the second pillar of the pension system. At the end of 2005, 695 pension funds served the Dutch market. In total they managed more than EUR 630 billion in pension capital (125 percent of GDP) for around 6.3 million active members, 2.4 million pensioners and 8 million inactive participants3, on a population of 16 million. As mentioned before, under Dutch law, employees must participate in the pension scheme offered by their employer. Generally, both the employer and the employee pay pension contributions, the employer bearing most of the
2 For more about the Dutch pension fund system, see Bikker, J and Dreu, De J, DNB working paper, no. 109, august 2006
cost. Also, most of the pension schemes in the Netherlands are defined benefit (DB) schemes, meaning that the benefits are based on a formula indicating the exact benefit that one can expect upon retiring. Indeed investment risk and portfolio management are entirely under the control of the company. The payouts made to retiring employees participating in this defined benefit plan are determined by factors such as salary history and the duration of employment. This contrasts with defined contribution (DC) schemes in which the financial contribution is fixed and the eventual benefit depends on the return on the funds invested. There is no way to know how much the plan will ultimately give the employee upon retiring. The amount contributed is fixed, but the benefit is not.Most pension contracts in the Netherlands guarantee only a nominal benefit, but pension funds do aim to link pensions to the wage or price index, provided there is sufficient pension capital. The Dutch situation can therefore best be described as a nominal DB pension combined with a ‘target benefit’ index-linked pension4.
Internationally, the Dutch pension fund system is unusual. The size of the second pillar is enormous and based entirely on funding. Not many countries have similarly high savings for their private pensions (OECD, 2004). Moreover, under the Dutch system, most pension schemes are DB schemes. In the US and in the UK, DC schemes dominate the market and their proportion is increasing. In the Netherlands a big portion of the third pillar provision, falls into this category. The fully funded feature of the second and third pillars encourages sufficient savings as demographic pressures rise. This is an important development as the number of citizens over 65 years, as a percentage of citizens between 20–64 years, is expected to double over the next 25 years.
Dutch pension funds can be broadly divided into three categories: industry-wide, corporate and occupational5. Industry-wide pension funds are mostly organized by sector, such as the steel or health care industry. ABP, the pension fund for all government employees, is also classified as an industry-wide fund. They are mostly initiated by unions in cooperation with employers’ organizations. Corporate pension funds, on the other hand, are typically related to one single company or some companies, e.g. Akzo Nobel, Shell. They are separate legal entities, run
4
Bikker, J and Dreu, De J, DNB working paper, no. 109, august 2006
5
The original Dutch terms are respectively: bedrijfstakpensioenfondsen, ondernemingspensioenfondsen and
directly by the sponsor company and, often, the union of the employee. Occupational pension funds offer pension plans to a specific group of professionals, for example to general practitioners. Participants are natural persons who practice a certain profession. In contrast to industry and company funds, occupational pension funds deal directly with workers and not with employers. Other types of pension funds include saving funds, but they constitute a very small share of the industry.
In terms of total assets and number of active participants, the Dutch pension system is dominated by industry-wide funds (see Table 2 and 3). However, the number of company pension funds is much higher (see Table 1), which reflects the on average smaller size of these institutions.
Table 1 Number of pension funds by type of pension fund 2001-2005 (source: DNB, 2006)
Pension fund type 2001 2002 2003 2004 2005 Industry-wide pension funds 96 100 99 102 103 Company pension funds 754 694 667 631 581 Occupational pension funds 11 11 11 12 11
Total 861 805 777 745 695
Table 2 Active member by type of pension fund 2001-2005 (source: DNB, 2006)
x 1.000 2001 2002 2003 2004 2005
Industry-wide pension funds 4901 5225 5077 5290 5284 Company pension funds 963 954 933 911 890 Occupational pension funds 40 41 42 45 46
Table 3 Total assets by type of pension fund 2001-2005 (source: DNB, 2006)
EUR million 2001 2002 2003 2004 2005
Table 4 indicates The Netherlands as having the biggest pension fund market relative to the
Dutch economy in comparison to the other G10 countries in 2005. The total asset value of the Dutch pension funds relative to the Dutch GDP is 124.9 percent, suggesting that Dutch pension savings are at relatively high level. This is partly due to the relatively small size of the Dutch economy alongside the developed Dutch financial and pension fund markets.
Table 4 Total asset value of pension funds of G10 countries as a percentage of GDP 2005 (source: OECD)
Graph 1 indicates that following a period of heavy decline, the total asset value of Dutch pension
funds has been increasing since 2003. Indeed, the sudden decrease in the world equity markets between 2000 and 2003 and the exposure to equity and equity-related products in the pension portfolios adequately explains the magnitude of the decline in pension fund assets in the Netherlands in that period. With the decline in (real) interest rates, a bond dominated portfolio no longer guaranteed sufficient returns to safeguard indexation, given the contribution level. One of the responses of pension funds was to invest more in the stock market, which increased average returns, but also mounted risk (Bikker and Vlaar, 2006).
Graph 1 Total asset value of Dutch Pension Funds as a percentage GDP 2001-2005(source: OECD) 0 20 40 60 80 100 120 140 2001 2002 2003 2004 2005
Given the large equity holding in pension portfolios in the Netherlands (see Graph 2), the recession in the equity market, inevitably, put downward pressure on pension asset growth. In the past few years Dutch pension funds have steadily recovered from the equity market down-turn of the early 2000s and shown robust asset growth. Total pension fund assets in the Netherlands amounted to EURO 636, 8 million in 2005, up from EURO 474, 6 million in 2001.
Graph 2 Structure of assets of pension funds in the Netherlands, 2005 (source: OECD)
Cash and Deposits Bills and bonds issued Loans
Shares
Land and Buildings Other investments
rates towards historically low levels (implying lower expected future returns, raising the discounted costs of future pension benefits), together with the ageing of the population in the Netherlands challenged the solvency of pension funds (Blikker and Vlaar 2006). This focused the attention of policy makers on the cost of supplementary retirement provisions. As a result, the sustainability of the Dutch pension systems has been called into question. In response, contribution premiums were increased6, pension schemes were made less generous, inflation and wage indexation were lowered and part of the risks was shifted from companies to participants (Bikker and Dreu, 2006).
Graph 3 Contributions, pension payment and direct investment return from Dutch pension funds under supervision in million euros (source: Statistical bulletin sept.2006)
0 5000 10000 15000 20000 25000 30000 2000 2001 2002 2003 2004 2005 Gross contribution Gross pension benefit Direct investment return
Graph 3 shows this process of decline in investment return, and increase of contribution premium
to recover for the losses for the Dutch pension funds under supervision. Dutch pension funds are under the supervision of De Nederlandsche Bank (DNB, The Dutch Central Bank). It is the official duty of the DNB to work towards a reliable financial system in which financial institutions meet their obligations. Good supervision encourages pension funds to meet their obligations and work soundly. The supervision concerns particularly their financial health, the
6
soundness of their management, their administrative organization and their internal control. Moreover the expertise and integrity of policy makers are reviewed. In short, the DNB checks whether a pension fund handles the money which has been entrusted to it in a justified and professional manner. The supervision is aimed at the protection of the interests of the customer, be this an individual or a company.
The above data illustrate clearly the increased importance of the accumulated institutional capital, its investment and its return. In the next sub-chapter the management of the invested capital of Dutch pension funds will be dealt with.
1.3 Dutch Pension fund investment management
As mentioned before in a defined benefit pension, which applies to almost all Dutch pension schemes, investment risk is typically assumed by the plan sponsor/employer and not by the individual. Because of this, investment of pension assets is one of the key functions performed by pension funds. For pension funds, three principles tend to drive investment decision-making: matching assets and liabilities, year-to-year and over the long term; risk management through portfolio diversification; and cost management through the market for financial services (Clark, 2000). Pension funds invest billions of euros annually in the stock and bond markets, and are therefore major players in the supply-demand balance of the markets. In order to promote both the performance and the financial security of pension plan benefits, it is critical that this function is implemented and managed responsibly. Policymakers have therefore a crucial role to ensure that regulations encourage prudent management of pension fund assets so as to meet the retirement income objectives of the pension plan7.
The investment function varies depending on the type of pension plan. In the case of the Dutch defined benefit plans, the goal of the investment function is to generate the highest possible returns consistent with the liabilities and liquidity needs of the pension plan, while taking into account the risk aversion of affected parties. Pension funds face considerable uncertainty about expected returns on invested assets compared to the relative stability of expected liabilities (Clark,
7
2000). The regulation of pension fund asset management should be based on the basic retirement income objective of a pension fund. Furthermore, it should assure that the investment management function is undertaken in accordance with the prudential principles of security, profitability, and liquidity using risk management concepts such as diversification and asset-liability matching8. The Dutch pension system has developed hybrid financial service conglomerates, intimately linking pension fund sponsors, while boards of directors overlap one another. As a result, many of the largest funds act both as the consumers and the suppliers of financial services. Custodial services, insurance, and investment management services can be found in Dutch pension fund related companies. Nevertheless, the Dutch pensions have sought to purchase expert advice and advanced financial products from London and Wall Street firms (Clark, 2000).
Despite the big investment management business, there is considerable evidence that pension plans tend to under-perform against the relevant standards of excellence. Lakonishok, Schleifer and Vishny (1992) report that the pension fund industry have consistently underperformed the market. The authors report that pension fund managers may trade too much, incurring large execution and transaction costs, and may be unlucky with their timing. In chapter 2 the advantages and limitations of a series of performance measures applicable to mutual funds and pension funds will be reviewed and compared.
8
Chapter 2 Literature review
2.1 Models of selectivity and timing and mutual fund performance
As mentioned earlier the investment management process can be separated into the activities of stock selection and market timing. Stock selection is based on the forecast of company specific events, and thus on the prices of individual securities. Market timing, on the other hand, refers to the forecast of price movements of the general security market as a whole (Fama, 1972).
The study of investment performance has long been searching for ways to distinguish between the ability to time the market and the ability to select stocks. This distinction can be seen as a good tool for attribution of an investment manager’s performance. Nowadays, it has even become standard practice to model selectivity and timing simultaneously. Mutual fund performance has received considerable attention in the US market, in contrast to pension fund performance. Moreover, there are numerous studies that cover a variety of periods. In this sub-chapter empirical evidence on investment performance of mutual funds will be discussed, the following sub-chapter will deal with pension funds.
Jensen (1968) proposed a risk adjusted performance measure that is sometimes referred to as the Differential Return Method. The model is
(1)
where Rpt is the excess return (net of risk-free rate) on the portfolio p. Rmt is the excess (net of
risk-free rate) return on the market portfolio, αp is a measure of security selection ability, βp
measures the sensitivity of the portfolio to the market return, upt is a random error, which has expected value of zero, and t denotes time. The model assumes that the risk level of the portfolio under consideration is stationary through time and ignores the market timing skill of the managers. Higher alpha represents superior performance of the fund and vice versa. Jensen applied his measure to examine annual rates of return net of management expenses on the portfolios of 115 open-end mutual funds for the period 1945-1964. He not only concluded that
the mutual funds were not able to predict security prices well enough to outperform a buy-the-market-and-hold policy, but also that there is very little evidence that any individual mutual fund was able to do significantly better than that which was expected from mere random chance. His conclusions held even when the fund returns were measured gross of management expenses. His findings of inferior performance by actively managed mutual fund’s portfolios are typical of many subsequent studies. Limitation of this model is that it considers only systematic risk not the entire risk associated with the fund and an ordinary investor can not mitigate unsystematic risk, as his knowledge of market is primitive (Mutualfundsindia Research Team).
Before Jensen, Treynor and Mazuy (1966) formulated a model for the assessment of both selectivity and market timing performance. This model, see equation (2), is an extension of equation (1), adding a quadratic term to test for market timing skill. Treynor and Mazuy (1966) wrote the excess return as
(2)
Whereγ , is the estimated indicator of market-timing performance, ∈ is the residual excess pt
return on the portfolio. Treynor and Mazuy argued that while a positive value of αp suggest
selectivity ability, a positive for γ is indicative of market-timing ability. A negative value for γ is interpreted as a lack of ability of fund managers to time the market correctly. An insignificant value for γ can be interpreted either as a lack of timing ability or as an omission by an investment fund manager to time the market. Treynor and Mazuy argued that if managers could time the market they would hold a larger share of volatile securities in a bull market and a larger share of less volatile securities in a bear market. They report evidence of timing ability for only one of 57 funds. Other studies using their method produce similar evidence of no timing ability.
There have been a number of extensions to the Jensen approach since the late 1960s. Fama (1972) and Jensen (1972) for example addressed the issue that the Jensen model of 1968 assumes that the risk level of the portfolio in consideration is stationary through times and ignores the market timing skill of the managers and suggested a somewhat finer breakdown of performance. They identified two dimensions of investment performance, where portfolio managers differentiate
between selection decisions and forecasting decisions. There proposal was empirically tested by Kon (1983). Kon concluded that individual mutual funds did exhibit significant positive timing ability. However, their study showed that mutual fund managers, as a group, have no special information regarding the unanticipated market portfolio returns. Grinblatt and Titman (1989b, 1991) addressed the potential bias that occurs when market timing ability is present, while performance measurement models exclude its empirical testing. They showed that successful market timers cause the estimate of systematic risk
β
to be biased upwards and the intercept term alphaα,
to be biased downwards.Henriksson and Merton (1981), also extending Jensen’s (1968) model, decomposed performance into selectivity and timing as follows:
(3)
where βp2 is the coefficient estimating timing ability and Yt is max [0,−xt]. The term βp2 is used by Henriksson and Merton to capture the market timing component of investment performance following Jensen (1972) demonstration of potential bias in the estimation. Jensen suggested that funds attempting to time the market will bias βp upward and the αp downward if
market timing (βp2) is ignored. The Henriksson- Merton model assumes that fund managers
target two systematic risk levels: one where the manager forecasts the risk-free asset to outperform the market portfolioβp , and the other where the market return is expected to
outperform the risk-free rateβp2. Successful market timing exists where the estimate of βp2 in
the model is significantly positive. The model does not predict the size of the return differential between a risky assets and the risk-free asset, but rather considers the direction of the forecast that a portfolio manager uses to re-weight the portfolio between a risky asset and the risk-free asset.
Bhattacharya and Pfleiderer (1983) offered a market timing approach that extends the theoretical approach of Jensen (1972). By correcting an error made in Jensen (1972) they showed that one can use a simple regression technique to obtain measures of timing and selection ability. Unlike Jensen, who assumed that the investment manager uses the unadjusted forecast of the market
return in the timing decision, Bhattacharya and Pfleiderer assumed that the investment manager used the adjusted forecast of the market return to minimize the variance of the forecast error. Breen et al. (1986) proposed that mutual fund’s performance measurement studies involving the tests of market timing ability, employing Treynor and Mazuy (1966), Henriksson and Merton (1981) or Bhattacharya and Pfleiderer (1983) approaches should also control for the influence of heteroskedasticity9. Breen et al. showed that ignoring heteroskedasticity results in rejection of the null-hypothesis of no timing ability more frequently than should be the case.
There are other models in the literature that permit identification and separation of selectivity and timing skills of mutual fund investment managers. Their results depend heavily on both the methodology used and the observation period considered. In addition, most of these studies were performed on the US or UK market and the situation may differ elsewhere. The focus of this paper is on the performance of Dutch pension funds, so in the next paragraph the theoretical literature on pension fund performance measurement will be reviewed.
2.2 Models of selectivity and timing and pension fund performance
As mentioned earlier, little research has been done on the investment performance of pension fund managers due to the absence of publicly available information about pension fund portfolios. But the existing evidence on the average performance of pension funds relative to external
benchmarks has been disappointing. From these studies, only a few modeled the attribution of security selection and market timing skill to the performance. In this subchapter
As mentioned in chapter 1, in 1986, BHB published a study about the investment performance of 91 large pension funds measured from 1973 to 1985. They compared the returns that would have been produced by the investment policy asset allocation of each pension fund with the actual return generated by actively managing the pension fund investments. The framework allowed them to measure the impact of security selection and market timing on performance. The results showed that on average the return generated by the investment policy portfolios dominated the
9
return attributed by the actively managed portfolios. BHB concluded that, on average, security selection and market timing added no significant value. A 1991 follow-up study by Brinson, Singer and Beebower showed the same results. The methodology of Brinson et al. will be discussed further in the next chapter as it is the method of research used in this study.
Other researchers like, Coggin, Fabozzi and Rahman (1993), used the Treynor and Mazuy (1966) and the Bhattacharya and Pfleiderer (1983) models, to investigate the investment performance of a random sample of 71 US equity pension fund managers for the period January 1983 through December 1990, and found that pension fund investment managers are on average better stock pickers than market timers. They argued that both selectivity and timing are sensitive to the choice of benchmark when management style is taken into consideration. For example they found that funds that target value strategies yielded out-performance of 2.1 per cent per annum, but funds that adopted growth strategies under-performed by -0.96 per cent.
In the UK Blake, Lehmann, & Timmermann, (1999) examined the asset allocations of a sample of 364 UK pension funds that retained the same fund manager over the period 1986-1994. They found that the total return is dominated by asset allocation. Average return from stock selection is negative and average return to market timing very negative.
Ibbotson and Kaplan’s (2000) report is build on the 1986 work of BHB. Their study found that only about 40% of the variation between funds is derived from differences in asset allocation. The remaining 60% of fund performance variation results from such other factors as the timing of moves between asset classes, security style (e.g. value or growth stock) within asset classes, security selection, and expenses. On average, asset allocation ultimately accounted for all of the absolute level of performance of the portfolios they studied.
Almost all the studies reviewed in this subchapter concluded that, on average, the asset allocation (investment policy) dominates the total performance of the pension fund’s investments.
Chapter 3 Methodology
The following chapter explains the methodology used to examine the investment results. There are a lot of techniques available to identify the causes for possible under- or over- performance. Yet, this thesis uses the method developed in the study that initially confirmed the importance of investment policy: “Determinants of Portfolio Performance,” a study by Gary Brinson, Randolph Hood and Gilbert Beebower (BHB), published in the July/August 1986 issue of the Financial
Analyst Journal. Based on the BHB study it has been a commonly held belief for over 10 years that, on average, investment policy explains more than 90 percent of portfolio return variability.
BHB describe the investment process used to manage a portfolio as a hierarchy of three decisions taken one after the other. Decision 1 and 2 are the choice of asset classes in which to invest and the choice of fixed normal asset class weights, these two decisions together form the investment policy. Decision 3 is the security selection and market timing decision and is referred to as investment strategy. In their research, BHB analyzed 91 large corporate pension plans that each had at least 40 quarters of performance history over the ten-year period 1974–1983. Their goal was to determine, from historical investment data on the 91 plans, which investment decisions, investment policy or investment strategy, had the biggest impact on total return and on the variability of the total return.
Each portfolio was assigned a composite benchmark of common stocks/bonds/cash based upon the plan’s average asset allocation over the 10-year period. BHB used the R2 statistic to describe that the investment policy decision explained, on average, over 93 percent of the variability of the plans’ returns. The R2 statistic measures the strength of a relationship between variables. The 91
plans had an average R2 of 0.93, indicating a strong relationship between the variability of a
portfolio’s actual returns and the return of portfolio’s composite benchmark. However, BHB also emphasized that security selection and market timing activities are still important
3.1 The framework of Brinson, Hood and Beebower
For their analysis four portfolios have been distinguished. Table 5 illustrates the framework and the computational requirements for analyzing portfolio returns.
Table 5 a Simplified Framework and Computational Requirements for Return Accountability
SELECTION Actual Passive Actual TIMING Passive In this framework, = p i
w policy (passive weight for asset class i =
a i
w actual weight for asset class i
= p i
R passive return for asset class i
= a i
R active return for asset class i
The first portfolio referred to as Portfolio I and represented in Quadrant I is the benchmark. The return on this portfolio is a consequence of the investment policy adopted by the pension fund, which is the long-term asset allocation plan to control overall risk and meet fund objectives. To
IV Actual Portfolio Return ) ( ) ( ia i a i R w IV R =
∑
IIPolicy and Timing Return ) ( ) ( ip i a i R w II R =
∑
IIIPolicy and Security Selection Return ) ( ) ( ia i p i R w III R =
∑
Icalculate the benchmark return, BHB used the weights of all asset classes, specified in advance and the passive return assign to each asset class. Portfolio II, shown in Quadrant II shows the return effects of investment policy and market timing. To calculate this return BHB used the actual weight for all asset classes and the passive return for all asset classes.
Quadrant III represents returns due to investment policy and security selection. Calculation of this return requires the passive weight for all asset classes and the active return for all asset classes. Portfolio IV, in Quadrant IV, shows the actual return to the total fund for the period and is the result of the actual portfolio weights and returns. The framework separates the effects of investment policy, market timing and security selection and the interaction of the latter, on total performance. Table 6 shows these contributors to total performance and their calculation method.
Table 6 Calculation of contributors to Total Performance
Return due to Calculation method
Timing R(II)- R(I)
Selection R(III)- (RI)
Interaction R(IV)- R(III)-R(II) +R(I) Total performance R(IV)- R(I)
The difference between the return of portfolio II and I gives the market timing effect on total return, and the difference between the return on portfolio III and I give the security selection effect. The joint effect of the interaction of both market timing and security selection on performance can be computed by summating the returns of Portfolio IV and I and subtracting the returns of portfolio III and II.
Chapter 4 Data
4.1 Characteristics of sample
This study covers 3510 Dutch pension funds from 2001 through 2005. The pension funds participating in this study were among those included in the World Market Company (WM Company) Universe of Dutch Pension Funds in 2004 and gave permission to mention them by name. 104 pension funds met the criterion above and of these 35 of them had sufficient return and investment weight information to be included in this analysis. The WM Company is the largest pension funds measurer in the UK. The company provides a variety of measurement services to pension funds both in the UK and The Netherlands. One of their most important services is the measurement of pension fund returns and the provision of appropriate league tables and benchmarks clients. In all cases the data are self-reported by the fund manager to the WM Company.
The main form of performance measurement among the WM universe of pension funds is the comparison of the return of the pension fund with the median fund in its league table. Pension fund trustees are concerned that their fund should outperform or at least approximately equal the median fund in the league table most of the time. The WM Company typically does not make the data available to the general public. The only information resource left is the annual reports, which depending on the pension fund or the investment managers differs in their composition and their information. For this reason all the return and weight data used for this research are collected from the annual reports of 2001 through 2005 of the different pension funds.
The internal benchmark weight each pension fund applied for the different asset classes was used as the policy weight for each asset class, in the cases where the benchmark weight were not reported the median weight of the WM universe of Dutch pension funds11. The same applied to the passive return for each asset class, the pension fund’s annual internal benchmark return was used. This was assumed to be the best alternative considering that a benchmark should meet the objectives of the underlying investment strategies and be a fair representation of the universal
10
See Appendix for a list of the included pension funds.
11
characteristics of the market sectors included. No other benchmark can meet these requirements better than the internal benchmark. Again in case the internal benchmark return is not available, the median return of WM universe of Dutch pension funds will be used instead. The WM Universe of Dutch pension funds is an aggregation of Dutch pension funds with similar investment briefs. It is based on the most representative cross-section of the Dutch pension fund industry, including both private and public sector funds, a variety of fund structures and a full range of asset management and is good for peer group comparisons. For both the actual weight and the active return the funds actual annual return and the actual weight at the end of each year were used.
Since this sample is small (35), the question arises as to whether the sample is unusual relative to the larger universe of Dutch pension portfolios. Table 7 shows the total assets of the sample relative to the total assets of the pension funds under supervision of the DNB. At year-end 2005, total assets of the participating institutions ranged from less than EURO13 million to over EURO190 billion.
Table 7 Total Assets of sample plus percentage of total assets Dutch pension funds under supervision
year Total assets sample % total assets pension funds under supervision of DNB12 2001 303895470800 64% 2002 291308661000 66% 2003 328081497900 66% 2004 367816460400 67% 2005 426607962200 67%
From the asset perspective it can be concluded that our sample is reliable in the sense that the total asset is far above halve of the total assets of the Dutch pension funds under supervision. This is the case because the sample includes Algemeen Burgerlijk Pensioenfonds (ABP) and
Pensioenfonds voor Gezondheid, Geestelijke en Maatschappelijke belangen (PGGM). ABP is by
12
far the largest pension fund in the Netherlands and the second largest in the world. With total assets exceeding 168 billion, ABP manages almost one third of all pension fund assets in the Netherlands. PGGM is in size surely a second, although it is about a third in size compared to ABP, PGGM is roughly three times larger than the number three on the biggest (in terms of assets).
The asset allocation of this capital is shown in the Graph 3. Graph 4 accentuates once more the ability of the sample to represent the Dutch pension market. The asset allocation of the sample does a good job in imitating the asset structure of the pension funds under the supervision of the
DNB, except for the share of mortgages in the total investment, which is higher for the pension funds under supervision. However the percentage is so low that its importance can be ignored. Most of the capital is being invested in equity and bonds (investments in each one of them is more than 40 percent). These holdings high light the importance of the developments in the equity and bonds market on the investment return of these pension funds and need for good investment management.
Graph 3 Asset allocation of the sample of 35 pension funds, 2001-2005
Graph 4 Asset allocation of the Dutch pension funds under supervision of DNB, 2001-2005 (source: DNB) 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 2001 2002 2003 2004 2005 mortgages loans bonds equity property other investment
Graph 5 shows that the contribution received by the pension funds in the sample of 35 pension
funds together with the direct income from investment has been enough to pay the pension fund’s obligations over the period 2001-2005. However, the indirect income from investment over this period indicates the losses endured by the pension funds during the years 2001 and 2002. The indirect income from investment has been left out of the graph, because of its illiquidity.
Graph 5 Contributions, pension payment & income from investment for the sample of 35 pension funds in million euros, 2001-2005
0 2000 4000 6000 8000 10000 12000 14000 16000 18000 2001 2002 2003 2004 2005 Contribution Pension paid
Graph 6 Equity allocations by Region for 200513 Europe United States Japan Pacific Emerging Markets Other
Investments are distributed among regions according to the objectives set by the different pension funds. Graph 6 shows the allocation of the equity capital by region for the sample of Dutch pension funds. Most of the equity capital is invested in Europe and the United States. The economic and market developments of the different regions were also of influence on the investments result of the pension funds. In the next paragraph some key economic and market development s for the years 2001-2005 will be discussed.
4.2 Important economic and stock market developments during sample period
The period 2001-2005 was an eventful period influencing the development of the global economy and the stock markets. The worldwide recession that started in 2000 continued in 2001. With the recession a long period of economic growth came to an end. The recession was caused by the more than tripling of the oil prices in the years 1999-2000, the tightening of the monetary policy, undertaken by a number of industrial countries in 1999, and the decline in the technology sector. Not only the scope of the economic recession was striking, but also the nature of it; the recession this time was not caused by a drop in demand, but by problems on the supply side of the economy. The strong economic expansion that took place in the second half of nineties led to considerable imbalances in the form of overcapacity and financial shortages. The industry sector was hit the hardest. Over-investments in particularly the technology and telecom sector,
13
stimulated by both high growth and high profit expectations, were the cause of overcapacity.
The terrorist attacks on the 11th of September 2001 took place at a moment when the American economy was already in a sluggish state. The American consumer’s confidence took a terrible blow. This is because, on one hand, their sense of security had been damaged, and on the other hand, they feared more attacks and possibly a huge increase of unemployment. Because of this the economic recession in the US continued to deepen.
In Europe, the economic growth also showed signs of decline in the middle of 2000. But the recession in Europe was less powerful than in the US, just as the economic expansion in the years 1995-2000 was less powerful than in the US. Across Europe the growth rate diverged considerably. Japan had both large structural and cyclical problems and has been in a recession since the middle of 2001, the fourth recession of the previous decade.
With the economic recession in the industrial countries the situation in the developing countries had also deteriorated.
The year 2001 was a bad year for the stock markets, share prices dropped, particularly caused by disappointing profit developments. Also the slow economic development was detrimental for the financial climate. The terrorist attacks of 11 September led to a panic response on the stock exchanges, where the rates dropped in a few days time with approximately 10 percent.
Meanwhile, the Japanese economy recovered from the recession but was still weak. The economic development in the Emerging markets differed strongly by region. The development in Asia was relatively favorable. This region, traditionally strongly dependent on exports, showed an increase of the domestic demand in 2002.
The year 2002 was characterized by the depreciation of the dollar. The depreciation was strongest with respect to the euro. 2002 was a disastrous year for the stock market. All over the world the share prices decreased on average with 25%. With that the bear market reached its heaviest phase so far. The year 2002 topped a serial of three consecutive years of losses on the stock markets. It was for the first time since the aftermath of the stock market crash of 1929 that the rates decreased three years consecutively. The Amsterdam stock exchange in 2002, however, was the worst performing stock market with a loss of 36 percent. For the European international investors 2002 was tougher.
The global economic recovery continued in 2003. Initially the economic growth was obstructed by the war in Iraq and the SARS-epidemic. This led to a fall in the consumer’s and producer’s confidence and had a hindering effect on the economic activity. The United States was the engine behind worldwide economic recovery, but the Asian countries also showed a strong growth. Europe was slow, but was also going to profit from the worldwide expansion.
The recovery of the European economy, however, continued with difficulty. In the first half of 2003 there was growth stagnation in the euro zone; a number of countries were going through a recession. In the second half of the year 2003 the economic growth increased, but was still lagging behind the other regions.
The Japanese economy also experienced economic recovery during 2003. Particularly two factors underpin that. First, in the previous years the improvement of the business climate had carried out strong reforms. Japan continued to profit from the growing international trade. Although several emerging markets were struggling with structural, financial and political problems, these countries did profit from the worldwide economic recovery. Asia was the fastest growing region.
Because of the appreciation of the euro, especially with respect to the dollar, but also with respect to the Japanese yen and the British pound, the return measured in euro lay significantly lower than the return measured in local currency.
The economic growth cycle reached its peak in 2004. The two major driving forces behind the recovery of the world economy were the strong cyclical recovery of the US and Chinese economies. Together, these two constituted an economic fly wheel for the rest of the world. The Emerging Markets, Japan and Europe, though in more limited degree, profited of the increase of the international trade and the enormous demand of raw materials and capital goods from China. Even though, political tensions (Iraq, terrorism, etc.) were regularly in the forefront.
The year 2004 was all in all a good year for the stock markets all over the world. The most important factors behind the favourable performances of the stock markets in 2004 were the favourable macroeconomic developments, the further improvement of the businesses profitability of, a favourable valuation both in historical perspective and an animation of companies, raised against other placement categories. Even though it took considerable time, before the favourable conditions translated themselves in price increases. Investors were insecure about the stock market, because of political developments and the impact of the strong increase of the oil price. Just in the last months this risk aversion decreased. Measured In local currency the Emerging markets performed the best, followed by Europe, whereas the US stayed behind. Emerging markets, mostly represented as raw material producers, profited of the strong growth of the Chinese economy, whereas Europe profited of an improving profitability. Converted to euro the return difference was in the advantage of Europe with respect to the US and Japan.
The economy of Europe has recovered after a number of difficult years, even though the recovery was considered fragile. The large overcapacity within Europe and the cheap imports from Asia had an oppressive impact on the prices, causing the high oil price to have only a restricted inflationary function. The Japanese economic recession was coming to an end after 15 years recession and deflation.
Despite the influence of the increasing oil price, the stock market performed well worldwide in the 2005.
Chapter 5 Results
One of the reasons why investors, in this case pension fund investment managers, practice active investment management is because there is a possibility to outperform the market, or so they believe. To what extent does the data in this research confirm this believes?
Table 8 shows the annual average relative return (against the benchmark) of the 35 pension funds
in the sample for each year in the sample period, expressed as a percentage.
Table 8 Average Annual Relative Return (against benchmark) for the years 2001-2005
year Average relative return
standard deviation
total pension funds under benchmark % of funds under benchmark 2001 -0.5% 3.1% 19 54.3% 2002 -7.0% 5.2% 32 91.4% 2003 0.5% 2.1% 14 40.0% 2004 0.1% 1.4% 19 54.3% 2005 0.2% 3.0% 12 34.3%
Over the 5-year period, the pension funds (on an average basis) underperformed the market the first two consecutive years. 2002 was the worst year, with an average relative underperformance of -7 percent and 91. 4 percent of the pension fund in the sample underperforming. As mentioned in chapter 4 the year 2002 was a disastrous year for shares. All over the world the share prices decreased on average with 25 percent. From the figures it can be concluded that the 2003-2005 period showed a small out performance. The results emphasize the financial markets developments over the period 2001-2005.
Table 9 and Table 10 show the worst and best performing pension funds, respectively, from the
Table 9 Average Annual Relative Return (against benchmark) of 35 Dutch pension funds for the years 2001-2005
10 worst-performing pension funds from the sample rank by average relative performance
Fund 5-year average size of investments Rank by size of investments 1.Grontmij 409398000 30
2.Tandartsen and Tandarts-Specialisten 1181321800 23
3.Woningcorporatie (ASW) 2751140800 15
4.Randstad 255639000 34
5.Rabobank 6019920000 9
6.Koopvaardij 2924597000 14
7.CSM Suiker 276155200 33
8.Meubelindustrie and Meubileringbedrijven 1029067400 24
9.Grafische Bedrijven 7283600000 7
10.Gasunie 588264400 27
Table 10 Average Annual Relative Return (against benchmark) for the years 2001-2005
10 best-performing pension funds from the sample Rank by average relative performance
It not possible to conclude by studying Table 9 and Table 10, that a pension fund needs to have a certain size to be able to beat the market. The average relative performance can not be related to size. Nonetheless, it is worthwhile to address the fact that the smallest pension fund, the pension fund for the Verloskundigen got the highest average annual relative performance.
Following the calculation of the relative performance the next question that arises is as to how this performance was achieved. To what extend can a manager’s performance be attributed? Is it investment policy, stock selection, or market timing?Performance attribution analysis attempts to explain portfolio performance in terms of these activities that compose the investment management process.
To analyze the relative importance of investment policy versus the investment strategy, the annual returns for each of the 35 pension funds were calculated for the years 2001 up to 2005.
Table 11 repeats the framework outlined in Table 5 and shows the quantification of the effects of
Table 11 Average Returns by Activity, 35 Pension Funds, 2001-2005 SELECTION Actual Passive Actual TIMING Passive
Active Returns due to
Timing 1.23%
Security Selection -0.60% Interaction 1.13% Total Active Return 1.76%
Table 12 shows the different effects of investment strategy and investment policy in more
detail. The effect of market timing on the average continuously compounded annual return of individual pension funds ranged from + 9.53 to -12.29 percent per year over the 2001-2005 period. The effect of security selection ranged from + 9.44 to -12.90 percent. When you add it all together, the average pension fund gained 1.76 percent per year. However, its effect on the individual pension plan ranged, from a low of -12.37 percent per year to a high of +14.96 percent, which is a range of 27.33 percent.
IV Actual Portfolio Return % 39 . 24 ) (IV = R II Policy and Timing
Return % 86 . 23 ) (II = R III
Policy and Security Selection Return % 03 . 22 ) (III = R I
Table 12 Average Return of sample of 35 Pension Funds, 2001-2005
Total Portfolio Returns Average Return Minimum Return Maximum Return Standard Deviation Policy 22.63% 8.48% 39.75% 5.32%
Policy and Timing 23.86% 6.27% 39.40% 5.40% Policy and Selection 22.03% 13.31% 35.37% 5.13% Actual Portfolio 24.39% 6.19% 38.59% 6.18%
Active Returns
Timing only 1.23% -12.29% 9.53% 4.81%
Security Selection Only -0.60% -12.90% 9.44% 4.98%
Other 1.13% -2.56% 19.36% 3.81%
Total Active Return 1.76% -12.37% 14.96% 6.22%
The study shows that active management is very important, but that the average value added to the return by the active management is small. Its importance with respect to investment policy is indicated by the relative size of the returns. Table 12 points out that the investment policy provides the larger portion of the return 22.63 percent, with active management accounting for an average 1.76 percent of the average annual actual return.
Further analysis shows the ability of investment policy to determine actual pension fund return.
Table 13a shows the relative amount of variance contributed by each quadrant to the return of the
portfolio. The percentages in each quadrant represent the average amount of variance of the actual total portfolio return explained by each of the quadrants. The figures were calculated by the regression of each pension fund’s actual total return against, in turn, its investment policy return, policy and timing return and policy and selection return and are the average of 35 R-squares of the regressions.
Table13a Percentage of Total Return Variation explained by Investment Activity, Average of 35 Pension Funds, 2001-2005 SELECTION Actual Passive Actual TIMING Passive
Table 13b shows that for individual pension funds investment policy explained no less than 39.18
percent and up to 99.57 percent of the total actual return variation.
Table 13b Percentage of Total Return Variation explained by Investment Activity, Average of 35 Pension Funds, 2001-2005
Average Minimum Maximum Standard Deviation
Policy 91.57% 39.18% 99.57% 13.15%
Policy and Timing 95.10% 56.59% 99.95% 7.85% Policy and Selection 97.24% 79.85% 99.97% 4.85%
The results in Table 13a and 13b clearly show that total return of a Dutch pension fund is dominated on average by investment policy decisions.
IV Actual Portfolio
Return 100.0%
II Policy and Timing
Return 95.1%
III
Policy and Security Selection Return
97.2%
I
Policy Return (Passive Portfolio Benchmark)
Chapter 6: Conclusion
Appendix 1
Participating pension funds and their investment managers
CORPORATE PENSION FUNDS:
• Ahold • Akzo Nobel • Gasunie • Grontmij • ING • Internatio-Müller • Koninklijke Nedlloyd • Rabobank • Randstad • Shell • Stork • TDV • Thales • Unilever “progress” • Vopak
INDUSTRY WIDE PENSION FUNDS:
• ABP • Architectenbureaus • Bibliotheekpersoneel • Bouwnijverheid • Detailhandel • Grafische Bedrijven • Horecabedrijf • Koopvaardij • Metalelektro • Meubelindustrie en Meubileringbedrijven
• Metaal en Technische Bedrijfstakken • PGGM • PNO Media • Schilders-, Afwerkings- en Glaszetbedrijf • Waterbouw • Woningcorporaties(ASW)
PENSION FUNDS MANAGED BY:
• Huisartsen
• Medisch Specialisten
References
Beasley, J.E. Meade, N., and Chang, T.J., 2003, ‘An evolutionary heuristic for the index tracking problem’, European Journal of Operational Research, Vol.148, pp. 621--643.
Bhattacharya, S., Pfleiderer, P., 1983, ‘A Note on performance evaluation’, Technical Report 714, Stanford University.
Bikker, J.A., and Dreu, J. de, 2006, ‘Pension fund efficiency: the impact of scale, governance and pension plan design’, DNB working paper, No. 109.
Bikker, J.A. and Vlaar,P.J.G., 2006, ‘Conditional Indexation in Defined Benefit Pension Plans’,
DNB Working paper, No.086.
Blake, D., Lehmann, B. N., and Timmermann, A., 1999, Asset allocation dynamics and pension fund performance, Journal of Business, Vol.72, No.4, pp. 429–461.
Brinson,G.P.,Hood,L.R. and Beebower,G.L.,1986, ‘Determinants of portfolio performance’,
Financial Analysts Journal, Vol. 42, No. 4, pp. 39-44.
Brinson, G. P., Singer, B. D., and Beebower, G. L., 1991, “Determinants of portfolio performance II’: an update, Financial Analysts Journal, Vol. 47. No.3, pp. 40–48.
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.
Fama, E.F., 1972, ‘Components of investment performance’, The Journal of Finance, Vol.27, pp. 551-567.
Good, W.R., 1984, ‘Accountability for Pension Performance’, Financial Analysts Journal, Vol.40, no. 1, pp. 39-42
Grinblatt, M. and Titman, S., 1989, ‘Mutual fund performance: Analysis of quarterly portfolio holdings’, Journal of Business, 3: pp.393-416.
Grinblatt, M and Titman, S., 1992, ‘The persistence of mutual fund performance’, The
Journal of Finance, Vol. 47, pp.1977-1984.
Grinold, R.C. and Kahn, R.N., 2000, ‘Active Portfolio Management’, Irwin Library of
Investmeny & Finance, New York, McGraw-Hill
Henriksson, R.D., and Merton, R.C., 1981, ‘On market timing and mutual fund performance II: Statistical procedures for evaluating forecasting skills’, Journal of Business, Vol. 54, No. 4, pp. 513-533.
Hodgson, T.M., Breban, S., Ford, C.L., Streatfield, M.P. and Urwin, R.C., 2000, "The concept of investment efficiency and its application to investment management structures", British Actuarial Journal, Vol.6, Part III, pp. 462-469.
Ibbotson, R. G. and Kaplan, P. D., 2000, ‘Does asset allocation policy explain 40, 90 or 100 percent of performance?’, Financial Analysts Journal, Vol. 56, No.1, pp.26–33.
Ippolito, R.A. and. Turner, J.A., 1987, “Turnover, fees and pension plan performance”, Financial
Analysts Journal, Vol. 43, No. 6, pp.16-26.
Jencen, M.C., 1968, ‘The performance of mutual funds in the period 1945-1964’, Journal of
Finance, Vol. 23, pp. 389-416
Jencen, M.C., 1972, ‘The Foundations and Current State of Capital Market Theory’, in: Studies in
Jensen, M. C., 1968, ‘The performance of mutual funds in the period 1945-1964’, The
Journal of Finance, 23: pp.3-43.
Kon, S.J., 1983, ‘The Market-Timing Performance of Mutual Fund Managers’, The Journal of
Business, Vol. 56, No. 3, pp. 323-347
Lakonishok, J.A., Shleifer, A. and Vishny, R.W., 1992, ‘The structure and performance of the money management industry’, Brookings Papers on Economic Activity, pp. 339-391.
Lintner, J. (1965), ‘The valuation of risk assets and the selection of risky investments in stock portfolios and capital budgets, Review of Economics and Statistics’, Vol. 47 No.1, pp. 13-37.
OECD, Guidelines on pension fund asset management, January 2006
Sharpe W.F, 1964, ‘Capital Asset Prices: A Theory of Market Equilibrium under Conditions of Risk’. The Journal of Finance, Vol. 19, No. 3, pp. 425-442.
Treynor, J. and Mazuy M., 1966, ‘Can mutual funds outguess the market?’, Havard
Business Review, Vol. 44, No. 4, pp. 131-136.
Treynor, J.L., 1961, ‘Market value, Time and Risk’, Unpublished Manuscript, no. 95-209
William F. Sharpe, ‘The Arithmetic of Active Management’, The Financial Analysts
