Dr. Vijay Jog
December 2, 2009
Scope and Context
The Investment Landscape
Existing Research-Based Evidence: Investment Returns and Performance
Canadian Investment Landscape
Investment Landscape: Costs and Returns
Implications of the Investment Cost Structure on Retirement Income Adequacy
Overall Key Observations and Conclusions
Appendix A: Glossary of Terms – Mutual Funds
This paper presents an analysis of the potential impact of investment returns on the retirement income adequacy of Canadians.
Rather than solely looking at whether Canadians are saving enough, Jog argues that there should be equal emphasis (if not more) placed on how as Canadians we can make smarter investment choices. Accordingly, the paper attempts to document empirically the interaction between investment choices, cost of investments, investment returns and retirement income and wealth.
The first section of this paper looks at existing research-based evidence on investment returns and performance through the performance of individual investors, mutual funds and other professionally managed savings. The key message from this section is that individual investors typically make poor decisions and while professionally managed funds outperform passive strategies, the investor incurs costs that are not compensated by returns.
In the second section, Jog looks at the Canadian investment landscape by focussing on the experience of a hypothetical Canadian individual using data over the last twenty five years. The date shows that: 1) maximising savings through an RRSP makes a significant difference depending on the investment vehicle chosen and the tax rate faced by the individual at retirement; 2) the choice of investment matters; 3) even after saving 18% of salary each year for 25 years, the end period wealth can only sustain the individual for twenty years or less; and 4) being a regular and an early saver matters.
The third section looks at the costs and returns in the investment landscape. The conclusions drawn from this section are that a "do-it-yourself" investor can invest at a relatively low cost though it is still somewhat higher than the cost of investing through a private sector defined benefit plan with the costs of investing with a financial advisor being the highest.
The forth section examines the implications of the investment cost structure on Retirement Income Adequacy. Table 15 looks at assumptions on cost and the impact on retirement income and notes that using a portfolio with both advice and active management that earns the same rate of return as the underlying index would result in a loss of 4 years of wealth. Table 16 estimates overall costs of the entire pension system.
The paper concludes by highlighting 7 overall key observations: 1) retirement income adequacy depends on tax assistance for savings (RRSP); 2) skipping years of saving or starting late has a significant implication on wealth accumulation; 3) investments of pension assets were worth $2.1 trillion as of 2007; 4) these pension assets are invested in a variety of securities (including 30% in foreign securities); 5) a high proportion (55%) is invested through employer sponsored plans (90% are defined benefit) ; 6) active management, over the long-run, does not add incremental value over a passive index and finally; 7) using 2007 data, the estimated overall cost of investment is 78 bps per retirement assets totalling $17.7 billion for $2.1 trillion of total assets. Of that $17.7 billion, the costs associated with investment advice, administration and active management account for $9.3 billion, almost 50% of the total costs.
Acknowledgements: This paper is written for the Research Working Group on Retirement Income Adequacy. I would like to thank many people for providing input, data and comments including: Kim Duxbury of Sunlife Financials, Blake Hill of Manulife Financial; Louis-Georges Mongeau of Standard Life of Canada; Terrie Miller of CEM Benchmarking Inc.; Keith Ambachtsheer of KPA Advisory Services Ltd.; Dennis Yanchus of Investment Funds Institute of Canada; Ronald Sanderson of Canadian Life and Health Insurance Association of Canada; Alfred LeB lanc of Department of Finance, Government of Canada; Christopher Donnelly and Paul Bean of RBC Asset Management Inc.; Jack Mintz, University of Calgary; Suzanne Paquette, Université Laval; and many others who also commented on the paper. I would also like to thank Mariela Wong, Ruben Palencia and Patricia Robertson for providing expert research assistance under very tight deadlines.
Retirement income adequacy of Canadians depends upon a variety of sources of income. However, five of these seem be the most important.
Each of these five sources/components is associated with different levels of risks and costs. The first source is probably the least risky while the risk of the second component is employer bankruptcy along with a high unfunded pension plan (e.g., Nortel or GM Canada). The risk of the other three sources comes from investment choices, rates of returns and costs associated with the investment vehicles being used.
Thus, retirement income adequacy depends not only on the various types of savings vehicles and the timing and amount of savings, but also on the investment choices and returns at least on those investments which are not guaranteed by the government or the employer. It is also important to note that, except for public sector plans, the risk of employer bankruptcy with an underfunded DB plan is non-trivial since, in many cases, this is the time when DB plans also become severely underfunded due to a decline in the market value of the pension fund.
The intent of this paper is to focus on the investment component (as opposed to the savings component) of the retirement income debate and its implication on retirement income adequacy. More specifically, the paper focuses on expanding our overall understanding of the investment choices available to a Canadian resident and investment performance of various savings and investment vehicles and intermediaries who assist in channelling individual savings so that they can provide adequate post retirement income to individuals.
The importance of investment returns on the retirement adequacy of Canadians should not be underestimated even though the focus of traditional debate is whether Canadians are saving enough from their current income to prepare them for their post-retirement life, especially in a world where life expectancy is expected to increase almost monotonically. A simple example about the importance of investment returns would suffice to emphasise their importance. If rather than earning, say, an 8% return annually, an individual earns 6.5% per year, then $1,000 savings would result in $4,828 (at 6.5%) as compared to $6,848 (at 8%) at the end of 25 years, a reduction of 30%. In other words, the same individual would have to save almost twice the amount to account for a poor investment choice/lower investment return. This simple example illustrates the potential impact of investment returns on retirement income adequacy and why we must place equal (if not more) emphasis on how Canadians can invest better as we do on how Canadians can be encouraged to save more. The former (making smarter investment choices) may be much easier for most Canadians than the latter (increasing savings rates).
Understanding the investment landscape in Canada is not that easy, since like any other developed country Canada has a gamut of investment vehicles and intermediaries available to individual Canadians. In this paper, we focus only on traded assets (e.g., stocks and bonds) and not on assets that are either held for consumption (home) or for leisure (cottage) and thus are considered to be somewhat illiquid and non tradable. We investigate costs and returns of both the direct vehicles for investment (e.g., buying shares of a company by placing a direct order) as well as investing through the various intermediaries (e.g., DB or DC plans and managed mutual funds).
In essence, there are three ways to accumulate wealth for post-retirement years: personal decisions (RRSP), through DB, DC or group RRSP plans, or through an intermediary that provides both advisory and investment options. In each case, the ultimate investment can be in an individual security (e.g., common stock or bond) or a passive index (either an Exchange Traded Fund (ETF) or an Index fund) or an actively managed (and recommended by a financial advisor) mutual fund or a segregated fund with or without an insurance guarantee. In all cases, the ultimate choice can be based on a "do-it-yourself" process or through a financial advisor.1
In addition, the assets underlying the DB and DC plans are being invested by pension fund managers and insurance companies on behalf of investors and employers. Moreover, a portion of CPP assets are now being invested by the CPP Investment Board.
It is also clear that these different choices and vehicles would be associated with differences in costs and rates of returns. The typical assumption would be that, on average, an individual expects to be better off with his investment choice when the costs incurred to undertake an investment are more than offset by a higher risk-adjusted return.
Accordingly, the main objective of the paper is to document empirically this interaction between the investment choices, cost of investments and investment returns. Our review includes a broad range of choices and investment vehicles to provide a better understanding of costs (historical and current) and risk adjusted returns. We also provide a perspective on existing research evidence on various issues associated with this trade-off. With respect to empirical evidence on historical rates of return and costs, we investigate the following vehicles/assets:
In addition, wherever the data permit, we investigate and document gross and net returns with appropriate benchmarks, differences between public and private sector plans, possible rationale for costs and differences between gross and net returns, economies of scale and scope – aspects that have direct implications on our understanding of the influence of investment choices on retirement income adequacy.
Our data comes from a variety of diverse sources including CANSIM, Statistics Canada, Sedar, CEM Benchmarking Inc. (for pension funds), IFIC (for mutual funds), CLHIA (for segregated funds) and other publicly available sources. We are aware that the diversity of sources may pose a challenge in cross comparisons; we do our best to ensure that a clearer picture emerges from our analysis with appropriate caveats.
There exists a large body of evidence on the investment returns earned on different vehicles and through different choices by individuals. While most of the research is from the United States, some of it is specific to the Canadian environment.
Some of the most interesting evidence comes from a series of papers by Barber and Odeon (1999, 2000) and Odeon (1999) about the performance of individual investors. They investigate investments and trading patterns of a large number of active individual investors using the data set from a large discount brokerage firm in the United States. While the commissions for buying and selling stocks were much higher in those days3, they show that, on average, individuals typically make poor sell and purchase decisions. Odeon (1999) states that in his sample, the investors tend to buy securities that have risen or fallen more over the previous six months than the securities they sell. They sell securities that have, on average, risen rapidly in recent weeks and they sell far more previous winners than losers.
The evidence presented by Barber and Odeon (1999) leads them to conclude that the gross returns even before accounting for transaction costs earned by these households are quite ordinary, on average. The net returns after accounting for the bid-ask spread and commissions were even poorer. They found that the average household underperformed a value-weighted market index 1.1% annually and, after accounting for the fact that the average household tilts its common stock investments toward small value stocks with high market risk, the underperformance averages 3.7% annually. The average household turned over approximately 75% of its common stock portfolio annually and the poor performance of the average household can be traced to the costs associated with this high level of trading.
They conclude by saying "our main point is simple: Trading is hazardous to your wealth. We believe that these high levels of trading can be at least partly explained by a simple behavioural bias: People are overconfident, and overconfidence leads to too much trading." (Italics added)
Similar conclusions are presented in a recent paper by Frazzini and Lamont (2008). Using the flow of funds data into U.S. mutual funds, they conclude that fund flows from retail investors into mutual funds are "dumb money". By reallocating across different mutual funds, retail investors reduce their wealth in the long run. This "dumb money" effect is strongly related to the value effect. High sentiment also is associated high corporate issuance, interpretable as companies increasing the supply of shares in response to investor demand.
As far as we know, there is no direct evidence (à la Barber and Odeon) on the investment performance of Canadian retail investors. The research analysing flow of funds into Canadian mutual funds (Deaves, 2004; Sinha and Jog, 2005) concludes that, unlike the U.S. investor, Canadian investors do not invest disproportionately in winning funds, and they do seem to punish losing funds. They claim that past performance and past asset allocations, as well as fund size and the size of the fund family are significant determinants of current fund flows and Canadians are more forgiving of losing funds.
In addition to investing money on their own, investors have increased their dependence on investing through mutual funds. As per the Ontario Securities Act, a "Mutual Fund" means an issuer whose primary purpose is to invest money provided by its security holders and whose securities entitle the holder to receive on demand, or within a specified period after demand, an amount computed by reference to the value of a proportionate interest in the whole or in part of the net assets, including a separate fund or trust account, of the issuer. "Mutual fund in Ontario" means a mutual fund that is a reporting issuer or that is organized under the laws of Ontario, but does not include a private mutual fund4. Worldwide, the mutual fund industry net assets reached $26.2 trillion peak by year-end 2007, declining to $19 trillion by year-end 2008; the corresponding numbers for Canada are Cdn$636 billion and Cdn$506 billion. As can be seen, the net asset values declined by 21% in 2008 – mostly due to the decline in valuation of the equities.
Given this tremendous growth in mutual funds around the world, it may come as no surprise that the performance of mutual funds is probably the most researched topic in the investment field. Our analysis shows that 166 mutual fund papers were published in a 22-year period (1987 to 2008) in the top 12 finance journals.5 The interest of this research is focussed in three broad areas: 1) fund performance; 2) persistence in performance; and 3) relation between flow of funds and performance (see Exhibit 1). Given this breadth of research, it is impossible to review individual papers, so we provide some broad conclusions without referring to a specific paper; a list of key papers is provided in the reference section.
Since the research is still predominantly U.S.-based, we provide some key observations from this large body of literature. The first observation is that the U.S. mutual funds underperform on a net (net of management fees) return basis, fund managers do not have significantly positive market timing ability and fund style migration does not create value for investors. Second, there is no persistence in mutual fund performance; the performance is mean reverting. Third, the results on the relation between the flow of funds and fund performance are mixed. Some studies show that investors do chase past performance (and then are disappointed with actual performance) and are reluctant to abandon bad performing funds due to the costs of switching; others show that investors (at least in Canada) do not chase performance to a great extent. The same conclusions are valid for Canadian mutual funds (Sinha and Jog (2005) and Deaves (2004)). Overall, the conclusion from this voluminous literature is that actively managed mutual funds, net of costs, underperform passive investment strategies and that there is no performance persistence.
Then the legitimate question to ask is that, in light of this evidence, why do investors invest in actively managed mutual funds? One could speculate on at least five reasons. First and perhaps the most dominant reason may be investor ignorance about investment choices due to lack of knowledge or lack of time. Many may simply not know or have time to find out how easy or difficult it is to invest in passively managed index funds or ETFs and how to choose amongst them and when. And thus, they are more than willing to pay for advisory fees and for the associated active management. Second, many investors rely on financial advisors for choosing investment vehicles; these advisors may have financial incentives to steer investors towards actively managed funds rather than passively managed lower fee index funds or ETFs.6 Third, the media continues to interview fund managers that outperform the passive strategies (even though the evidence shows non-persistence) and talks and writes about undervalued stocks, indicating that there is a way to find undervalued stocks by relying on professional advice. Fourth, as noted by Barber and Odean (1999, 2000) and recently by Statman, Thorley, and Vorkink (2006), overconfidence is probably the other major reason investors are willing to incur the extra fees, expenses, and transaction costs of active strategies either by themselves or by searching for professionally managed mutual funds. Lastly, it is also possible that an individual investor may think that she may actually be worse off in making the right timing decisions with respect to when and how much to invest, when to rebalance the asset mix (equity versus bonds), what type of passive vehicle to choose – and she is actually better off letting the professionals handle it.
In addition to mutual funds, a significant fraction of savings is channelled through DB or DC plans. However, many of these, in turn, choose mutual fund managers to invest their funds but due to their scale they pay lower fees than those paid by retail investors. Many also invest in passive index funds or manage the funds by themselves or give them to institutional asset managers who act as external managers of pension funds. Compared to the voluminous research on the mutual fund performance, however, there is scant evidence in the academic literature on the performance of these funds or the institutional asset managers (IAMs) they hire.7 This is in spite of the fact that in the United States alone these funds represent a total asset value of more than $6 trillion, of which 40% to 50% is invested in equities. In Canada, of the $2.2 trillion pension assets, trusteed DB plans constituted $1.0 trillion as of 2007. The two recent papers partially fill this research gap.
In a study of institutional asset managers, Busse et al. (2006), using a large sample of IAMs in the United States came to three conclusions. First, even after accounting for fees and expenses, IAMs, on average, outperform the benchmark which is significant both statistically and economically. Second, a significant majority of the IAMs use performance-based fee structures, quite unlike the mutual fund fee structure where MERs are independent of performance which may lead to reduced agency costs. Third, and more significantly, there is a significant inflow of new funds into high performing (first quartile) asset managers whose performance declines due to this high influx of funds. As a result, they state that "however, the persistence that is the source of potential gains for plan sponsors is its very own death knell: we find that portfolios in the winner deciles draw an inﬂux of capital from plan sponsors, and in the year following this capital inﬂow, the excess returns disappear." These results of outperformance found by Busse et al. are, however, in contrast to some earlier papers such as Lakonishok et al. (1992) and Coggin, et al. (1993) who find that performance is poor on average, investment managers have limited skill in selecting stocks, and Survival bias and a short time series does not allow for a robust conclusion about persistence.8
The Bauer and Frehen (2008) paper provides evidence on the U.S. DB and DC pension funds using benchmarks self-selected by pension funds. They report that DB and DC pension funds slightly underperform benchmarks and a passively managed large cap investment is, in relative terms, most attractive.9 In addition, they conclude that persistence in yearly pension fund equity performance is weak or nonexistent.
In summary, this extensive evidence on investment performance of individuals either investing directly or through intermediaries leads to some key conclusions. First, left to himself (or herself), an individual investor performs poorly as he makes wrong market timing, asset selection and security selection decisions . Second, there is very little evidence that professionally managed funds outperform the passive strategy, especially after considering costs. Third, there is very little evidence that there is persistence in performance. Fourth, performance based fees may be a partial solution to reducing agency costs, but while this mechanism may be available to larger pension funds, the individual investor does not have access to it as mutual funds do not provide such performance based fee contracts.10 Fifth, to gain access to professional management, the investor incurs costs which are not compensated by excess returns. Sixth, this leaves the question as to why an individual investor would invest on her own or search for actively managed mutual funds.
With this backdrop, in the following section we provide direct empirical evidence on the investment landscape faced by a typical Canadian investor and its implications on the retirement income of Canadians.
One of the basic tenets of finance is that high returns are associated with high risk, and since equities typically have high risk (higher variations in periodic returns) overall they should provide investors with a higher return. To set the stage, Exhibit 2 provides three different snapshots of historical evidence on returns on three types of securities: t-bills, long-term bonds and TSX index. Exhibit 2 shows 82 years of history (beginning in 1925 but excluding 2008), Exhibit 3 shows 53 years of history beginning in 1956 and Exhibit 4 starts in 1984. In each of these exhibits, we track the cumulative value of $1 invested in the beginning of the relevant period assuming reinvestment and excluding costs of investment and taxes. As can be seen, over the long period, stocks provide a higher end-period wealth. However, and most surprisingly, in the last 25 years long term bonds have outperformed stocks. In addition, three other observations are important. First, the longer the time period, the higher is the impact of compounding returns and higher the end-period wealth. Second, stock returns are more volatile than bond returns. Third, investment choices do matter in end-period wealth, focussing only on savings and not worrying about investment may not be prudent.
Next, we focus on the experience of a hypothetical Canadian individual using the data over the last 25 years. Our main purpose is to show how choices made by an individual with respect to savings percent, RSP versus non-RRSP and asset selection have implications on end period wealth.
We start at the end of December 31, 1983 and assume that, in that year, this individual earned $40,000 taxable income and decided to save 18% of her salary. She has two choices: to invest 18% ($7,200) in an RRSP (maximum allowed) or the same amount in a taxable account. Since the RRSP contribution gets a tax deduction at a 35% tax rate, she invests the tax refund of $2,520 in a parallel taxable account. At the end of 25 years, the RRSP account faces taxes at withdrawal, but the taxable account does not face such penalty as these amounts faced taxes on their annual investment returns depending on the type of asset classes chosen (see below). We further assume that her salary grows by CPI + 1% and she continues to invest the same percentage of her salary each year at the end of the year and reinvests any interest and dividend receipts after paying the associated taxes. We make some simplified assumption about tax rates. To keep this illustration simple and manageable, we further assume that this individual faces a marginal 35% tax rate throughout her career and a 20% tax rate at the end of 25 years.11 This tax rate of 35% applies to the investment income from T-Bills, the corresponding tax rates would be: 20% effective tax rate on bonds (meaning 60% of the total return on the bond index is interest) and 0.48% of equity assuming a 3% dividend payout with a 16% effective tax rate on dividends; we assume that the investor holds a buy-and-hold portfolio. Note that this is an illustrative example and these rates and assumptions can be changed.
We then chose four asset classes: t-bills, long term bonds, equity and a balanced portfolio of 60% equity, 30% long-term bonds and 10% t-bills. Table 1 shows the results of the various investment choices and compares the end-period wealth as a multiple of 60% of the 2008 salary which has grown to $76,554. We use 60% of salary as it is typically said that this percentage allows for the same standard of living in the post-retirement period. We then show the impact on end period wealth if this individual does not invest for five years during that period and if he starts investing in 1998 (15 years later than our base case).12
Table 1 leads to the following observations. First, even after noting that the RRSP account will have tax implications at withdrawal, maximising savings through an RRSP (tax deferred vehicle) makes a significant difference depending on the investment vehicle chosen and the tax rate faced by the individual at retirement. More specifically, since interest is taxable, the table illustrates that holding debt in a RRSP account benefits more than holding equity, all else being the same. Second, the choice of investment matters. A very risk-averse investor who invests exclusively in T-Bills would have significantly less wealth than if they chose either equity or a blended portfolio. In addition, since interest is taxable, the table also illustrates that holding debt in a RRSP account benefits more than holding equity, all else being the same. Third, even after saving 18% of salary each year for 25 years, the end-period wealth can only sustain the individual for 20 years or less.13 Fourth, being a regular and an early saver matters. There is a significant difference in the end of period wealth between a regular saver and a saver who skips some years or starts late.
Note that the above example assumes that the cost of investment is zero. However, we know that is not the case and thus it is important to understand the cost structure that a typical Canadian individual investor may face.
As noted earlier, there are many ways to invest one’s savings: the investment vehicles range from direct investment in stocks and bonds, Exchange Traded Funds (ETFs), index funds mostly marketed by the banks, bank managed non index funds (similar to other mutual funds), mutual funds, DB plans, DC plans, group RRSPs (administered mostly by insurance companies), insurance company funds with insurance and embedded guarantees, company sponsored DB plans and lastly the CPP Investment Board for CPP contributions. In some cases, financial advisors may be involved whereas some individuals may make their own decisions (outside of defined benefit plans). Table 2 shows the aggregate value of the pension-related assets as of 2007. As can be seen, of the $2.2 trillion in assets (note, this is prior to 33% decline in equity values in 2008), trusteed pension plans constitute $921 million followed by RRSP assets of $739 million.
Also, in each case, the ultimate investment securities can vary: for example, money market, bonds and equities including both domestic and foreign. Thus, the range of investment vehicles and intermediaries and choices is wide. Table 3 shows the asset mix of the Pension assets as of 2006. As can be seen, pension funds invest 41% of their assets through mutual funds or IAMS; the total equity component is 44% and bonds constitute 41%; foreign assets (equities and bonds) represent 24%. Next we turn our attention to costs and, wherever available, investment returns.
As noted earlier, our data comes from various sources: public and private. Since DB plans constitute the most significant fraction of retirement assets, we started with them; the data for this part of the analysis comes from both Statistics Canada (Table 4A) which is aggregated for all funds from 1993 to 2008, and from CEM Benchmarking Inc. (Table 4B) beginning from 1990 (53 funds) to 2008 (81 funds).14 As shown, the costs of managing DB plans vary across private versus public sector funds but on average a private sector DB plan’s costs are anywhere between 30 to 45 bps and corresponding figures for a public sector plan are between 25 to 35 bps. Using the CEM Data, we find that the total costs do not decline monotonically with the total asset size; for both 2007 and 2008, the costs seem to stabilise at 35 to 40 bps using data on all funds. Thus, there do not seem to be economies of scale after a particular size and expenses seem to be somewhat independent of fund size at these levels.15 Similarly, we find very little evidence that larger funds consistently outperform smaller funds at the total fund level, not withstanding considerable differences in asset mix of these funds. We also find that the larger the fund, the larger are their investments in non-conventional and potentially non-benchmarkable and illiquid assets (e.g., private equity and hedge funds).
It should also be noted that one would expect a DB plan to have a lower cost because of the nature of the plan. It is designed to provide for a stream of payments over time that are somewhat predictable, it lowers the cost of portfolio management because both inflows and outflows are lower and predictable and, the portfolio manager can invest without worrying about changes in plan members’ opinions about asset mix and withdrawals and having to respond to member questions.
Next we focus our attention on the investment returns of pension funds in the CEM database. While we have data on every investment category, we restrict our focus to two major investment classes, Canadian equity and Canadian bonds, as these typically constitute a significant portion of pension assets. The database consists of benchmarks self-selected by pension fund respondents, but to avoid self selection bias, we use the TSE index and CDN bond Index as the respective benchmarks. The results of this comparison are shown in Table 5A and 5B.
In both tables, the first column shows the year, the corresponding benchmark returns used for comparisons, # Funds, Weighted (by assets) Average Return – All Funds, Weighted Average Return – Funds Quartile 1, # Funds whose returns for that year are higher than the corresponding benchmark, and % Funds that exceeded the benchmark return. As can be seen, in equity portfolios, 60% of the funds exceed the benchmark returns and the performance is better more in "down" years than in "up years". This may suggest a more conservative investment policy of stock selection and changes in cash- stock mix.
The results for bond portfolios are not good; except for recent years and one or two years in between, the bond portfolios have significantly underperformed the TSX DEX long-term bond index. However, one can argue that the bond portfolios of the sample funds also include short-term bonds and thus the index chosen may not be representative of the benchmark. Since it is almost impossible to find the right benchmark, we consider another approach to compare the performance of sample funds.
The CEM database provides data for 19 years beginning in 1990; however, the data is not available for every year or for every fund since some funds started providing data only later in the period and some funds stopped providing data in the latter part of the period. So, we decided to focus on those funds where we have at least five years of data. Next, we categorised these funds in four quartiles based on their performance (this is a standard practice). We categorised the funds based on four types of returns: total fund return, Canadian equity returns, Canadian bond returns and U.S. equity returns since the last three categories constitute a significant percentage of total assets. Table 6 shows the results of that analysis.
As can be seen, we have data on gross returns for 134 funds for at least five out of the 19 years. If there was (perfect) persistence in superior performance, we would expect that the fund in quartile one (Q1) – the high return quartile – remains in quartile one in all the years it is in the database and, if so, this will correspond to a value of 100% in the "Q1 persistence" column. Similarly, if the fund remains in quartile 4 (Q4) – low return quartile – all the years, it would show a value of 100% in that column. Thus, the value of 24% in the "Q1 persistence" column indicates that, on average, a typical fund remained in first quartile for only five of the 19 years; in other years, it was in quartiles 2, 3 or 4. Similarly a value of 25% in the "Q4 persistence" column indicates that a typical fund was in Q4 for four out of 19 years.
As seen from table 6, what is remarkable is the lack of any persistency over the 19 year period; while there is some year-to-year Q1 persistency displayed by some funds, they are unable to maintain their membership in Q1 over a longer period. The results are almost identical for all four categories. Although not shown here, we also see very little relationship between costs of the fund and weighted average return performance; the correlation coefficients between costs and asset weighted total returns, Canadian stock returns, Canadian bond returns and U.S. equity returns are -0.037, -0.028, -0.006 and -0.050, respectively. The overall conclusion is that pension funds in this sample do not show any consistency and do not outperform the benchmark even without accounting for costs.
Investments in these plans are typically administered by insurance companies on behalf of corporate clients, but in general managed by third party money managers. Exhibits 5A and 5B provide aggregate statistics on assets managed by insurance companies under different types of plans and the corresponding member coverage. We were fortunate to get confidential data from large insurance companies to get a representative fraction of managed assets.
This data allows us to calculate the costs of investing that are borne by investors and their net rate of returns.16 To maintain confidentiality, we have decided to not report the size of the assets under management (AUMs) and made minor changes to the data without impacting it in any fundamental way. We believe that the results provided in Table 7 can thus be viewed as a highly representative examination of over $120 billion retirement assets managed by Insurance companies.
As can be seen, the overall cost structure indicates that the cost of investing for all types of plans is approximately 70 BPS; the costs of managing RPP Plans is approximately 60 BPS with somewhat higher costs for managing RRSP assets (92 bps). The overall costs can also be broken down (table 7B) by "commissioned" and non-commissioned assets and there is a considerable difference (almost 60 basis points); the cost structure of "non-commissioned" plans is just a few points over the costs of managing DB plans.17 As shown in Table 7C, the costs also vary by asset class: managing domestic and North American equities (90 bps) costs less than global equities (124 bps).
It should be noted that the higher costs of managing "non-commissioned" assets compared to the DB plans may arise for three reasons. First, these plans require a considerable administrative infrastructure that the captive DB plans do not have. Second, the captive DB plan itself does not have a requirement to earn a return for the shareholders, whereas that is not the case for shareholder-owned insurance companies. Third, it is not clear what cost structure will be incurred by DB plans if they were to now compete for assets outside their own organisation and have to incur all the administrative costs associated with dealing with individuals. We should also note that the asset mix of this group is significantly different than the DB plan sample and, therefore, we cannot make any conclusions on whether there is a difference in the relative performance of these groups, although it seems highly unlikely that a difference exists.
Next, Table 7D shows the inherent economies of scale that depend both on the number of plan members and the average asset value. These costs are based on representative typical plans based on sample pricing scenarios by insurance companies. As can be seen, the larger the plan (based on # of members) and higher the average asset value per member, the smaller the corresponding costs. For a plan with 25 members and average asset value per member of $25,000, the costs are 1.25%, whereas for a plan with 7,500 members and with average asset value of $50,000, the costs decline to 0.44%. Similarly, if the plan decides to invest in a passive index, the costs are naturally lower. It should also be noted that these assets are in turn invested in mutual funds. The costs are lower due to lower regulatory, administrative and reporting requirements and, for the most part, a lack of individual advice provided except in some special situations.
In Canada, insurance companies also provide investment vehicles that provide guarantees to savers on their investments – guaranteed investment funds (GIFs). These come in a variety of shapes and forms and include: Guaranteed Withdrawal Balance (GWB), Annual Lifetime Withdrawal Amount (LWA), Annual Guaranteed Withdrawal Amount (GWA) and a GWB Bonus, to name a few.18 In essence, these guarantees reduce the risk of investments and are affected by the timing and the amounts of deposits and withdrawals. Insurance companies, in turn, invest these amounts in mutual funds and charge an additional amount for providing these guarantees. The costs of these "guarantees" can be inferred from differences between the total cost of these investment products with guarantees and the costs of the corresponding underlying mutual fund. Table 7E provides a sample of the representative costs associated with these products. As can be seen, typical additional cost of insurance is approximately 27 bps and the cost of Guaranteed Withdrawal Balance (GWB) is an additional 62 bps. As noted earlier, these are representative numbers and would vary across insurance companies and by the underlying mutual fund category.
Next, we turn our attention to the CPP Investment Board (CPPIB) which was formed in 1997 when the Canadian government decided to invest CPP assets in capital markets and established the CPP Investment Board as a federal Crown Corporation by an Act of Parliament with a mandate to invest the assets of the Canada Pension Plan in a way that maximizes returns without undue risk of loss. Its first investment took place in 1999; the funds invested through the auspices of CPPIB are now in excess of $100 billion. Table 8 outlines the growth in CPPIB’s portfolio, reported rates of return and costs.
A review of Table 8 leads to some interesting results. Prior to 2007, the CPPIB’s cost structure represented 7 basis points of assets. It invested (according to its annual reports) through two external investment management firms (TD and Barclays Global) and, except for 2001, did not do much better than the benchmark. However, there seemed to be a clear policy change in 2006 with an increased investment in non Canadian private assets and what is labelled as "inflation sensitive assets (real estate, inflation-linked bonds and infrastructure)". During the same period, consulting fees went up 265%, salaries went up by 425% and the analysis shows that CPPIB reported that it earned 2.5% higher than the benchmark with the inflation sensitive assets and non Canadian private assets earning excess returns. The costs of running CPPIB now have reached 20 bps; still lower than DB plans but the growth of the cost base has been significant. .
Mutual funds are a big part of the Canadian investment landscape as they constitute approximately $500 billion of investment by Canadians at the end of 2008. Naturally, just as the investment performance of managed mutual funds has received considerable attention, even more attention has been given to the fees (management expense Ratios – MERs) charged by mutual funds especially since they do not seem to generate consistently higher returns than passive indexing. A recent paper by Khorana et al (2007) provides a comprehensive analysis of mutual fund fees using data on 46,580 mutual funds offered for sale in eighteen countries with assets in excess of $10 trillion. To account for cross sectional variations across countries, they define costs as total shareholder cost (TSC):
Total Shareholder Cost = TER + initial load/5 + back-end load at five years/5
where TER, in effect, is defined as "all annual operating costs (including administration/share registration, trustee/custody, audit and legal fees), not just the basic annual management charge." Their results for year 2002 show that the TSC for Canada for bonds, equity and full sample are 1.84%, 3.00% and 2.41%, respectively. These results also show that Canada has one of the highest costs for investing through mutual funds using three different measures of management fees. Obviously there are many challenges with such cross country comparisons.19 This Morningstar study (2007) gives Canada an overall grade of "B+". More specifically, the study gives Canada an "A" grade in Investor protection, transparency in prospectus and report, transparency in sales and media; a "B+" in distribution/Choice; a grade of "C" in taxation; and an "F" in Fees and Expenses. So if the costs are higher in Canada, it is also clear that they are well reported and with full transparency. It is also possible that other factors such as economies of scale, costly regulatory requirements and differential levels of individual advice may be at play here.
In this section we provide more recent evidence on these fees and also ensure that the comparisons are made in the right context. Accordingly, we discuss mutual funds in general including Exchange Traded funds (ETFs), index funds and managed mutual funds.
Table 9 shows the most recent data on 103 ETFs traded on the TSX constituting $28 billion in market value; the sub groupings show that they comprise 11 categories and are marketed by five organisations. Equity ETFs represent over 40% of the market value, followed by Canadian bond ETFs, leveraged and commodity ETFs. These ETFs can be bought and sold as stocks and represent the lowest cost of investment in broadly diversified portfolios of underlying securities. The costs vary by type with equity bond and balanced funds costing 40 bps and specialty funds (which do have some active management) costing approximately 80 bps. It should be noted that the managed funds by insurance companies analysed in Table 7 have a similar cost structure. However, for an individual investor, ETFs may constitute the lowest cost of investment in the capital market and most are eligible for RRSP investments.
Next in the cost ladder are funds marketed and sold by Canada’s five large banks. These banks provide opportunities to invest in indexed funds (similar to standard ETFs) as well as their actively managed funds. Table 10 shows the results of our analysis of these RRSP-eligible funds. As can be seen, the cost of investing in index funds is approximately 100 bps and for non-indexed funds is 200 bps. This indicates that the cost of active management charged by the banks is 100 bps. It should be noted that there are some advantages in investing in index funds over ETFs. An individual can make a regular deposit and does not have to incur costs of buying ETFs regularly ($10 per trade for an active high income investor but $30 for a regular investor). For an investor who wants to invest $1000 per month through ETFs, these costs would be significant. On the other hand, once ETFs are bought, there are no annual fees.
It should be also noted that if individual wants advice as to the asset mix decisions (bond versus equity) from an advisor and then invests the funds in ETFs, the advisor may increase advisory fee since he would now get no part of trailer fees. The question really is, what is the adequate level of compensation for pure advisory services and why does one not find a large number of advisors who simply provide advisory services and do not automatically steer investors to the most cost effective funds?
Table 11 shows the variety of domiciled mutual funds currently available to a Canadian investor along with their average and median expense ratios and table 12 shows it for the segregated funds.21 A few observations are in order. As can be seen a Canadian investor has a choice of 46 types of mutual funds and forty types of segregated funds. The expense ratios (MERs and MAFs) vary considerably across the various types of funds. Overall money market funds exhibit MERs of less than 100 bps, bond funds and North American equity funds just below 200 bps and specialty and global funds approximately 250 to 300 bps. In the segregated fund categories, the numbers are approximately 30 to 40 bps higher perhaps reflecting costs of insurance guarantees. Overall, an investor is subject to 200 bps costs which however are significantly lower than reported by Khorana et al (2007).
We now provide a more detailed analysis of the mutual fund cost structure by separating various components of mutual fund fees.22 We focus on various components of the cost structure: passive investment through ETFs or index funds, management advisory fees (MAFs) and management Expense ratios (MERs) of similar funds, MFs that are sold to individual retail investors (series A) compared to those sold targeted to advisors offering fee-based accounts (series F). Appendix A provides a glossary of the necessary terms. Data in these two tables are based on top 10 funds by assets in each respective CIFSC categories and thus exclude many smaller funds.23 Also note that these numbers are significantly lower than reported by Khorana et al (2007) as they focus on assets that are of most relevance to Canadian investors and those that are marketed by large well-known funds.
These two tables allow us to analyse the cost structure and services embedded in a typical mutual fund sold through the retail advice channel from the point of view of the retail investor. Based on Tables 13 and 14, we separate out the various cost components for a typical Canadian Equity fund and a typical Canadian Fixed Income fund sold through the retail advice channel from the point of view of the retail investor in the chart below. (Numbers are average but highly representative).
As noted earlier, from the point of view of the retail mutual fund investor, they could buy an ETF or an index fund through a discount brokerage, foregoing the cost of advice and services provided by the portfolio manager and individual advisor or alternatively, they could purchase a mutual fund with portfolio management and individual advice. The primary service provided by an index fund is passive money management – the buying and selling of securities to match the particular index tracked. The costs associated with this portfolio management service are represented here by the management advisory fee (MAF) of the lowest cost index fund in each CIFSC category.24 The costs associated with the second choice are made up of the additional cost of the services provided by the portfolio manager25 and the additional cost of the services provided by individual advisor.26 In addition to these costs, there are also administrative costs and GST that are included in the total management expense ratio (MER) of both passively and actively managed mutual funds.
We are able to draw some broad conclusions from this analysis. First, a "do it yourself" investor can invest relatively inexpensively either through ETF or through an index fund and it would cost approximately 40 to 70 bps – somewhat higher than the cost of investing through a private sector DB plan. Second, it costs approximately 70 to 80 bps to get advice from a financial advisor whose natural inclination may be to recommend managed mutual funds; this is not to say that advisors do not explain the options of ETFs and Index funds. Third, the cost of pure active management is about 60 bps. Fourth, although not shown here, there is an additional cost incurred by these funds which represents the cost of trading, termed as the Trading Expense Ratio (TER), that is almost 23 bps for equity funds and very small for bonds and approximately10 bps across all funds.
However, as noted earlier, there is a considerable difference between a mutual fund and a DB plan. A mutual fund, whether passively or actively managed, must price daily, must provide a prospectus and information form annually, must provide financial statements and a report of fund performance semi-annually as well as an annual report of its independent review committee among other documents. The unit holder must also receive account statements at least annually and is provided access online or by phone to his/her account balance through either the fund company or the advisor or both, 24/7. These do add to the cost structure of a mutual fund compared to a DB plan. Thus, one could say even though active management may not result in consistent excess returns; it is simply a cost of avoiding bad investment decisions that may be made by an uninformed investor.
We estimate the implications of this cost structure on retirement income adequacy in two ways. First, we turn our attention to the representative investor and estimate the impact of costs of investing on his retirement income using estimates from Table 14. Table 15 shows our assumptions about costs and its impact on the retirement income of this representative individual (see Table 1 for base case). As can be seen, for a blended portfolio (with blended costs of 1.6%), using ETFs alone would result in reduction of eight months (0.7 years) worth of wealth; advice only (but investment in ETF) costs 2.8 years of wealth and using both advice and active management that earns the same rate of return as the underlying index would result in a loss of four years of wealth. These differences are somewhat higher for an all equity portfolio than an all bond portfolio since the cost of advice and active management is relatively higher for equity portfolio relative to the bond portfolio. Note that a crucial assumption in this table is that active management does not, on average, earn excess returns than the underlying passive index. If advisor-recommended active management earns 1.85% per year higher than the underlying equity index and earns 1% higher than the underlying bond index, an investor would be indifferent between investing through ETFs and active management with advice.
Table 16 estimates overall costs of the entire pension system; given the complexity, these are rough estimates. Here we take the Canadian retirement asset values as of 2007 from Table 2 and associate estimated costs of investing to each of the asset classes based on our analysis so far. For social security plans, and trusteed public sector plans, we use costs of public sector plans as a proxy (0.35%), and the corresponding costs for trusteed private sector plans of 0.44%; we estimate these costs based on Table 4A and 4B. For insurance company contracts, we assume that 90% of these are non-guaranteed and the rest are guaranteed; accordingly, we use a blended cost of 0.7 % using estimates from table 7A and 7E. For government annuities, we use an arbitrary 0.1% (using no-advisory T-bill costs from Table 15 as a proxy). For Deferred profit sharing plans and RRSP deposits, we use 0.50% and 0.90% estimated from Table 7A. For mutual funds and segregated funds, cost estimates are more slightly complicated. We know from Table 3 that the overall asset mix resembles our blended portfolio (used to provide estimates of wealth in Table 15) but with a 28% investment in either foreign funds or in foreign assets. Their blended costs are higher than domestic equity; these can be estimated as approximately 2.25% (see Table 14). We also know that of the approximately $911 billion of pension assets (table 3), insurance companies have $35 billion under their management (CLHIA data) which, on average, cost 0.65%. So, using the 1.60% annual costs of a blended portfolio (table 15) and assuming that this represent 60% of these assets; knowing that foreign assets actually have a 2.25% costs and represent 30% of these assets and a small portion (the remaining 10%) of these assets are under insurance company management (cost 0.65%); we arrive at an estimate of the overall costs of 1.7%. Since we have no detailed knowledge of the asset mix of "other individual RRSP accounts" and neither do we know how these are invested, we use the same estimate (1.7%) as above. One may argue that these assets include foreign assets, index funds, bank managed non-index funds and mutual funds and may not have any economies of scale, so their costs would be actually higher and 1.7% is actually a conservative number.
Table 16 shows that these costs, 78 bps for a representative retirement dollar, were worth $16.5 billion. If we assume that 80% of assets classified under mutual funds and other RRSP accounts are invested through advisory services and active management (the rest invested through index funds and ETFs or without any advisory services), these costs (1.7% times 80% of assets under the last two categories in Table 16) can be estimated to be $9.4 billion. These costs represent investments being made by Canadians in financial advisory sector of the economy with the expectation that it would provide them with higher than benchmark investment returns or to prevent them from making bad investment mistakes if they are left on their own.
The main intent of the paper was to focus on the investment component (as opposed to the savings component) of the retirement income debate and its implication on retirement income adequacy. The paper focussed on analysing various investment choices available to a Canadian resident, and investment performance and associated costs of various savings and investment vehicles and intermediaries who assist in channelling individual savings so that they can provide adequate post retirement income to individuals. As can be seen, this is a complex area as the investment field comprises of many investment vehicles. We offer following observations.
First, retirement income adequacy critically depends on the tax assistance for savings (RRSP versus non–RRSP), timing of investments and the type of investment. An individual investing 18% of her saving since 1983 in blended portfolio at zero cost and invested in corresponding passive equity, debt and t-bills would have saved 18 years worth of 2008 replacement income (60% of 2008 salary). The same amount invested outside RRSP would have resulted in 14 years worth replacement income. Both these numbers assume a 35% tax rate during work years and a 20% tax rate at the retirement year.
Second, these numbers are sensitive to investment choices and savings patterns; skipping a few years or starting late has significant impact on wealth accumulation at retirement. Third, as of 2007, $2.1 trillion worth of pension assets were invested through a variety of vehicles ranging from assets under social security type programs (CPP and QPP) to assets invested by individuals through individual RRSP accounts. Fourth, the composition of these assets shows that these assets are invested in many different types of securities; Canadians have invested approximately 30% of their assets in foreign securities. Fifth, a significant fraction (55%) of these assets is invested through employer sponsored pension plans, and 90% of these are in DB plans. Thus, the returns and costs of assets under these plans have a considerable impact on the retirement adequacy of Canadians. Our analysis indicates that the costs incurred by these plans are in the range of 35 to 45 bps and there is no evidence that these plans show any persistence in performance or that they consistently outperform the passive investment options. It should be noted that these funds do not have any distribution or advisory or administration or reporting or regulatory costs associated with administering a large number of individual accounts. In addition, we also review the costs and performance of CPPIB and find that the costs have increased rapidly and the focus of investment has shifted from marketable assets to non-marketable assets; the latter are hard to benchmark.
Sixth, we spent considerable efforts documenting and detailing costs of other financial intermediaries, namely insurance companies and mutual funds. Our analysis of the overall cost structure of insurance companies that manage defined contribution plans (RPPs and Group RRSPs) indicates that the overall cost is approximately 70 BPS; the cost of managing RPP Plans is approximately 60 BPS with somewhat higher cost for managing RRSP assets (92 bps). With respect to mutual funds, we find that the combined costs of advice and active management approximately 160 bps – split almost equally amongst the two components and varying considerably depending on the type of asset (Canadian equity, foreign equity, bonds, money market, specialty funds, etc.). It should be noted that costs associated with mutual funds cannot be compared with a DB plan; one would expect a higher cost associated with a mutual fund. We also show that the costs of bank-managed mutual funds are no different than a comparable non-bank managed mutual fund. Given the fact that banks with their existing retail infrastructure and continuous relationship with clients may have lower incremental costs, this observation is somewhat surprising. Although the debate on the benefits of active management would continue, we document that the existing academic evidence shows that active management, on the long run, does not add incremental value over a passive index. This does not mean that, from time to time, active managers do not outperform the benchmarks or that investors would do better on their own. Actually, the available U.S. based evidence indicates that, when investors are left on their own, they make bad investment decisions . One may conclude that many investors are not sophisticated and do need advisory services from financial advisors. The question is whether financial advisors should channel the savings in actively managed funds or in ETFs or index funds; if they do the latter, then what cost structure would be appropriate for such a purely advisory service? We also note that a recent survey of mutual fund investors indicated that 83% use financial advisors and that the industry receives high marks from analysts for investor protection, transparency in prospectus and report, transparency in sales and Media and in distribution/Choice. So if the costs are higher in Canada in comparison to other countries, it is also clear that these costs are well reported and with full transparency. We also note that costs noted by Khorana et al seem to be much higher than those we report here. It is also possible that other factors such as economies of scale, costly regulatory requirements and differential levels of individual advice may be at play.
Lastly, we provide rough estimates of costs associated with various components of pension assets. Using the 2007 data from Statistics Canada, we estimate the overall cost of investment to be 78 bps per retirement assets totalling $17.7 billion for $2.1 trillion of total assets. Of the $17.7 billion costs, the cost associated with investment advice, administration and active management account for $9.3 billion, almost 50% of the total costs.
Aggarwal, Rajesh K., Galin Georgiev, and Jake Pinato, Detecting Performance Persistence in Fund Managers, Journal of Portfolio Management 2007, pp. 110-119.
Alessie, Rob, Stefan Hochguertel, and Arthur van Soest, Ownership of Stocks and Mutual Funds: A Panel Data Analysis," Review of Economics and Statistics, 2004, pp.783-796.
Ang James S., Gregory L. Nagel and Jun Yang, "A Critical Long View of Capital Markets and Institutions: Realized Returns from Corporate Assets, 1950-2003", Working paper 2006.
Antolin, P., "Pension Fund Performance", OECD Working Papers on Insurance and Private Pensions, #20, OECD Publishing, doi: 1787/240402404057, 2008.
Baks, Klaas P., Andrew Metrick, and Jessica Wachter, Should Investors avoid all Actively Managed Mutual Funds? A Study in Bayesian Performance Evaluation, Journal of Finance, 2001, pp. 45-85.
Barber, Brad M., and Terrence Odean, Boys will be boys: Gender, overconfidence, and common stock investment, Quarterly Journal of Economics 116, 2001, pp .261-292.
Barber, Brad M., Terrence Odean, and Lu Zheng, Out of sight, out of mind: The effects of expenses on mutual fund flows, Journal of Business, 2005, pp. 2095-2119.
Bogle, John C., The Mutual Fund Industry 60 years Later: For Better Or Worse? Financial Analysts Journal, 2005, pp. 15-24.
Bogle, John C., A question so important that it should be hard to think about anything else, Journal of Portfolio Management, 2008, pp. 95-102.
Brown, S. J., and Goetzmann, W. N., Performance Persistence, Journal of Finance, 1995, pp. 679-698.
Busse Jeffrey A., Amit Goyal and Sunil Wahal, Performance Persistence in Institutional Investment Management, http://ssrn.com/abstract=890319, June 2006,
Chan, Louis K. C., Hsiu-Lang Chen, and Josef Lakonishok, On Mutual Fund Investment Styles, Review of Financial Studies, 2002, pp. 1407-1437.
Carhart, Mark, On persistence in mutual fund performance, Journal of Finance, 1997, pp. 57-82.
Chen, Joseph, Harrison Hong, Ming Huang, and Jeffrey Kubik, Does fund size erode mutual fund performance? The role of liquidity and organization, American Economic Review, 2004, pp.1276-1302.
Cici, Gjergji, Gibson, Scott and Moussawi, Rabih, For Better or Worse? Mutual Funds in Side-By-Side Management Relationships with Hedge Funds SSRN: http://ssrn.com/abstract=905600, (December 14, 2006).
Coggin, T.D., F. J. Fabozzi, and S. Rahman, The investment performance of U.S. equity Pension fund managers: An empirical Investigation, 1993, pp.1039-1055.
Deaves, Richard, The Comparative Performance of Load and no-Load Mutual Funds in Canada, Canadian Journal of Administrative Sciences, 2004, pp. 326-333.
Del Guercio Diane and Paula A. Tkac, The Determinants of the Flow of Funds of Managed Portfolios: Mutual Funds vs. Pension Funds, Journal of Financial and Quantitative Analysis, 2002, pp. 523-557
Elton, Edwin, Martin Gruber, and Christopher Blake, Incentive fees and mutual Funds, Journal of Finance 2003, pp. 779-804.
Fama, Eugene F., and Kenneth R. French, Mutual fund performance, working paper, Dartmouth College, 2008
Frazzini Andrea and Owen A. Lamont, Dumb Money: Mutual Fund Flows and the Cross-Section of Stock Returns, Journal of Financial Economics, 2008, pp. 299-322.
French, Kenneth, R., Cost of Active Investing, 2008: http://ssrn.com/abstract=1105775
Golec, Joseph and Laura Starks, Performance fee contract change and mutual fund risk, Journal of Financial Economics 2004, pp. 93-118.
Gruber, M., Another puzzle: the growth in actively managed mutual funds, Journal of Finance, 1996, pp. 783-807
Jensen, Michael C. The Performance of Mutual Funds in the Period 1945–1964, Journal of Finance, 1968, pp. 389–416.
Ajay Khorana, Ajay, Henri Servaes and Peter Tufano, "Mutual Fund Fees Around the World", 2007 Forthcoming, Review of Financial Studies.
Lakonishok, J., A. Shleifer, and R. W. Vishny, The structure and performance of the Money management industry, Brookings Papers on Economic Activity: Macroeconomics, 1992, pp. 339 - 391.
McGuigan, Thomas P., The Difficulty of Selecting Superior Mutual Fund Performance, Journal of Financial Planning 2006, pp. 50-56.
Malkiel, Burton G., Returns from Investing in Equity Mutual Funds 1971 to 1991, Journal of Finance, June 1995, pp. 549 –72.
Sinha Rajeeva and Vijay M. Jog, "Fund Flows and Performance: A study of Canadian Equity funds" Canadian Investment Review, summer 2007, pp.28-34.
Sinha Rajeeva and Vijay M. Jog, "Performance of Canadian Mutual Funds and Investors", Advances in Investment Analysis and Portfolio Management, 2007, pp. 227-259.
Statman, Meir, What do investors want? Journal of Portfolio Management August 2004, 153-161.
Statman, Meir, Steven Thorley, and Keith Vorkink, Investor overconfidence and trading volume, Review of Financial Studies, 2006, pp. 1531-1565.
Stoll, Hans R., 1993, Equity trading costs in-the-large, Journal of Portfolio Management Summer, 41-50.
Sun Chengye, Review of Studies on Mutual Funds Performance, unpublished document, Sprott School of Business, Carleton University, Canada, 2009.
Tonks, I., Performance persistence of pension-fund managers, Journal of Business, 2005, pp. 1917-1942.
Investment Choices and End period Wealth
|Type of plan||1990||1995||2000||2001||2002||2003||2004||2005||2006||2007|
|millions of dollars|
|Canada Pension Plan||40,577||40,611||41,660||46,004||52,119||65,716||78,760||94,471||113,581||122,729|
|Quebec Pension Plan||14,323||13,880||18,350||16,452||18,125||18,754||22,421||26,973||32,239||34,703|
|Employer-sponsored pension plans||322,263||523,093||818,781||792,713||778,691||850,547||940,752||1,048,170||1,162,236||1,222,789|
|Government consolidated revenue arrangements2||89,170||131,909||180,082||165,576||177,055||180,197||184,194||188,241||193,738||199,089|
|Insurance company contracts3||30,968||38,411||53,389||49,538||51,552||57,802||61,441||66,667||70,986||73,118|
|Government of Canada annuities4||829||628||495||505||427||395||369||344||331||319|
|Deferred profit sharing plans5||1,887||2,162||7,661||10,179||6,316||8,924||15,326||18,995||23,542||29,177|
|Individual registered saving plans6||135,263||273,281||411,464||409,829||378,932||428,348||500,475||570,776||641,003||739,295|
|Deposits in RRSP accounts7||97,588||134,760||102,504||104,531||106,346||110,305||110,634||110,243||111,836||114,472|
|Mutual funds and segregated funds in RRSP accounts8||12,748||74,364||188,581||190,056||175,348||193,697||211,255||231,458||250,301||251,454|
|Other individual registered saving plans9||24,927||64,157||120,379||115,242||97,238||124,346||178,586||229,075||278,866||373,369|
Source: Statistics Canada, Latest Developments in the Canadian Economic Accounts, catalogue number 13-605-XIE.
|Assets||Total||Defined benefit||Defined contribution||Combination||Other|
|Pooled, mutual and investment funds:|
|- Equity fund (Canadian)||74,452,423||8.2||65,912,150||7.9||2,661,545||12.8||5,436,723||10.1||442,007||7.9|
|- Bond fund (fixed income)||94,907,920||10.4||83,853,332||10.1||2,707,465||13.0||7,050,188||13.1||1,296,937||23.1|
|- Mortgage fund||9,066,703||1.0||8,745,553||1.1||45,233||0.2||95,144||0.2||180,776||3.2|
|- Real estate fund||16,579,726||1.8||16,326,281||2.0||19,879||0.1||229,709||0.4||3,859||0.1|
|- Money market fund||7,584,145||0.8||6,566,259||0.8||352,836||1.7||654,935||1.2||10,117||0.2|
|- Foreign fund||103,148,747||11.3||92,991,348||11.2||2,770,348||13.3||6,970,066||13.0||416,986||7.4|
|- Canadian common and preferred||131,725,751||14.4||118,550,129||14.3||3,043,975||14.6||9,008,671||16.8||1,122,978||20.0|
|- Foreign common and preferred||156,864,923||17.2||145,272,190||17.5||2,416,752||11.6||8,278,548||15.4||897,435||16.0|
|- Other Canadian (corporate)||40,439,759||4.4||36,260,391||4.4||1,153,915||5.6||2,746,053||5.1||279,402||5.0|
|Cash, deposits, short-term:|
|- Cash, deposits, GICs||7,425,101||0.8||6,653,954||0.8||284,247||1.4||427,624||0.8||59,277||1.1|
|- Government of Canada t-bills||4,246,326||0.5||3,845,439||0.5||96,239||0.5||265,471||0.5||39,178||0.7|
|- Foreign short-term investments||909,241||0.1||809,234||0.1||4,823||0.0||91,384||0.2||3,802||0.1|
|- Other short-term paper||9,997,425||1.1||9,029,684||1.1||83,353||0.4||841,880||1.6||42,510||0.8|
|- Accrued interest and dividends receivable||2,537,685||0.3||2,360,022||0.3||46,188||0.2||121,008||0.2||10,470||0.2|
|- Accounts receivable||3,393,121||0.4||3,162,655||0.4||24,408||0.1||200,680||0.4||5,380||0.1|
|- Other assets||17,896,859||2.0||17,602,133||2.1||89,550||0.4||204,898||0.4||279||0.0|
|Source: Statistics Canada|
Canadian Defined benefit Plans
Assets and Costs
Source: Statistics Canada
Canadian Defined benefit Plans
Assets and Costs
Source: CEM Benchmarking Inc.
Comparison of Equity portfolio returns with Benchmark
Comparison of Bond portfolio returns with Benchmark
Total fund returns (131 funds)
Canadian Equity Returns (131 funds)
Canadian Bond Returns (131 funds)
U.S. Equity Returns (119 funds, post 1994 only)
Assets Managed by Insurance Companies
Management fees and net ROR by Asset type
Management fees and net ROR by commissioned and Non-Commissioned
Management fees and net ROR by Asset type
The total assets of sub categories do not add to 100% as come sub categories are not comparable across the respondents and to respect confidentiality are excluded in this table.
Economies of Scale: Plan Members and Asset Size
Cost of Guarantees
CPP Investment Board
Source: various annual reports
Exchange Traded Funds
Costs and market Value
Bank managed Mutual funds
Indexed and managed (non-Indexed)
|Bank||Canadian - Equity||Global - Equity||Money Market - Canadian||US - Equity||Canadian - Bond||Global - Bond||Candian Balanced||Total/ Average|
|Bank||Canadian - Equity||Global - Equity||Money Market - Canadian||US - Equity||Canadian - Bond||Global - Bond||Candian Balanced||Total/ Average|
|Source: Various Bank Web sites|
Mutual Funds Choices Available to Canadian Investors
Mutual Funds: Type and Expense Ratios
Source: Morningstar PALTrak at September 30, 2009
Segregated Funds: Type and Expense Ratios
Source: Morningstar PALTrak at September 30, 2009
Costs of Active Management and Mutual funds
Source: Investment Funds Institute of Canada
Cost of Active Management and MERs
Source: Investment Funds Institute of Canada
Impact of Costs on Portfolio value
Aggregate Costs of Investing on Pension Assets
Studies in Performance of Mutual funds categorised by key attributes.
Source: Chengye Sun (2009)
Historical Rates of Return – Value of $1 invested in 1925
Note: in 2008 TSX had a -33 percent return and as consequence the 2008 value would have been $1,763.43. The corresponding value for LT bonds and T-bills would have been $134.59 and $41.55, respectively.
Source: Fundamentals of Corporate Finance – Richard Brealey, Stewart Myers, Alan Marcus, Elizabeth Maynes and Devashis Mitra, McGraw – Hill Ryerson, Third Canadian Edition.
Value of $1 invested in 1956
Compiled by the author. Source: IFIC.
Investment returns from 1984
Compiled by the author. Source: IFIC.
Insurance Company Managed Assets
Insurance Company Managed Plan Members
Source: Canadian Life and Health Insurance Association
Series A – This can also be known by other names but, in general, denotes a series sold to retail investors and whose Management Expense Ratio (MER) will include compensation to the distributor (dealer and advisor). The series usually allows multiple purchase options – sales charge option (FE) (commission can be charged directly by the advisor on the fund purchase), deferred sales charge (DSC) option (the advisor is compensated on the sale by the fund company and the investor’s purchase is subject to fees if the investor redeems within a certain time frame) or a no load option.
Series F – Series targeted to advisors offering fee-based accounts. This series will have the distribution costs stripped out leaving only the advisory fees charged for portfolio management as well as administration fees, legal fees, filing fees, auditor fees and GST.
Series E: This series is sold only via the discount brokerage channel.
Series I: This is the series sold to institutions - large pensions, insurance companies, fund-of-funds etc. The MER and MAF on the I series are often negotiated individually with each institution so the posted MERs or MAFs for this series will typically understate the true costs.
Management Expense Ratio (MER): For each series, the management expense ratio is calculated based on the total expenses of that series of the Fund (including Goods and Services Tax and interest, and management advisory fees (MAF) but excluding brokerage commissions and other portfolio transaction costs), and is expressed as an annualized percentage of daily average Net Asset Value of that series of the Fund during the period.
Management Expense Ratio (MER) before waivers and fees absorbed by the fund manager: At its sole discretion, the Manager may waive a portion of the management and advisory fees or absorb a portion of the operating expenses of certain Funds. Such waivers and absorptions can be terminated at any time, but can be expected to continue for certain series of the Funds until such time as these series of the Funds are of sufficient size to reasonably absorb all management and advisory fees and expenses incurred in their operation.
Trading Expense Ratio (TER): The trading expense ratio represents total commissions and other portfolio transaction costs expressed as an annualized percentage of daily average Net Asset Value of the Fund during the period. In some cases, TERs are charged at different rates for each series in other cases there is one rate charged to the Fund as a whole. It does not include trailer commissions, which are paid by the fund management company out of their management advisory fees (MAF) and is not included in the MER. Therefore, the total on-going cost of ownership would be the MER plus the TER.
Gross versus net Returns: If the reported net return is 5% for the year and the MER was 2% and TER was 0.4% then the gross return is roughly 7.4%.
2 It should be noted that segregated funds managed by insurance companies for workplace plans are not the same as the segregated funds in the retail world. More specifically, retail segregated funds include an insurance component that guarantees a specific return on the investor's capital at maturity. However, segregated funds that cater to group RRSPs or defined contribution plan simply mean that the market based assets are kept separate from the general assets of the company. These segregated funds, in turn, buy units in the underlying pooled or mutual funds. This has implications on the corresponding costs as we note later in the paper.
3 They state that the average commission paid when a security is purchased is 2.23% of the purchase price. The average commission on a sale is 2.76% of the sale price. Thus, if one security is sold and the sale proceeds are used to buy another security, the total commissions for the sale and purchase averaged about 5%. However, in the last 10 years, due to the growth in discount brokerage houses, costs of trading a security have declined significantly.
4 Ontario Securities Act Sec. 1(1). The other two types are closed-end fund and unit investment trust.
5 The top finance 10 journals are ranked by Oltheten et al (2005); they are (listed by ranking): Journal of Finance (JF), Journal of Financial Economics (JFE), Review of Financial Studies (RFS), Journal of Financial and Quantitative Analysis (JFQA), Journal of Business (JB), American Economic Review (AER), Journal of Political Economy (JPE), Econometrica (ECO), Journal of Banking and Finance (JBF), and Financial Management (FM).We also include Journal of Portfolio Management (JPM) and Financial Analyst Journal (FAJ). The number of published mutual fund studies has been growing rapidly over years. From 1991 to 1995, 11 papers were published, while from 2001 to 2005, 64 papers were published, the number is five times larger. From 2006 to 2008, 32 studies were published.
6 This does not mean that an advisor steers the investor where he (the advisor) receives the highest fees, actually the fees seem to vary little across various mutual funds. For example, for most equity funds the standard is a 0.5% trailer fee and 1% when it is sold front-end. Fixed income funds will tend to have lower trailers followed by money market funds. Clearly, advisors do have motivation beyond just earning high fees; these include building and retaining client assets by providing them the best advice from a risk-adjusted return point of view, However, it is also natural that the motivation to steer a client into a low fee, no trailer-no front end loaded mutual fund with similar risk-return characteristics may not be there.
7 A recent paper by Antolini (2008) provides a comparison of aggregate pension fund performance by country on a risk adjusted basis using relatively standard investment performance measures. He concludes that, in spite of the data limitations, pension funds have generally outperformed a hypothetical but realistic benchmark portfolio; his results for Canada shows a 2% excess returns by pension funds during the 1990-2005 period.
8 However, Christopherson et al. (1998) do find some evidence of persistence but among poorly performing investment managers.
9 They also focus on comparing the pension fund performance with mutual funds matched for size and type and conclude (self evident) that pension funds do better than mutual funds net of costs. However, it is not clear that such comparison is a valid one. First, the benchmarks were self selected by the reporting pension funds in their data set and second, mutual funds are for profit and provide a multitude of other services to individual retail investors and thus naturally have a very different cost structure.
10 For some mixed evidence, please see Elton et al (2003) and Cici et al (2006).
11 These are important assumptions since the benefit of RRSP-type vehicles is directly dependent on the tax rates faced by the individual during her working life and during the post-retirement period.
12 We have arbitrarily selected years 1986, 1991, 1997, 2001 and 2005 as years where investor did not save.
13 This is an illustrative example and we have excluded the implications of taxes post withdrawal period, expected returns on the portfolio and inflation in the post- 2008 period. All these can be easily modeled. Also note that OAS/CPP combine to provide an indexed life annuity of about $18,000/yr as a base.
14 During the 19 years, a total of 239 funds appear in the database.
15 CEM analysis using global funds as well as Canadian funds shows economies of scale after accounting for differences in asset mix of funds. It also claims that larger funds perform relatively better than smaller funds meaning return to cost ratio improves as size improves. However, our somewhat simpler analysis using all Canadian funds and only at the aggregate level (using correlation coefficients and sorting by asset size) indicates that these correlations between assets size, cost and returns vary considerably across years.
16 These costs often referred to as fund management fees include, but are not limited to, operating expenses for both the segregated fund or mutual fund and any underlying fund, and investment management fees. Investment management fees pay for professional investment managers to select the fund's investments. These fees also pay for keeping records of accounts, GST and other member servicing costs. Operating expenses for both the segregated fund or mutual fund and any underlying fund are generally made up of brokerage commissions and other expenses of buying and selling securities for a fund and any expenses relating to the operation of a fund, including legal, audit, trustee, custodial and safekeeping fees, interest, operating and administrative costs (other than advertising, distribution and promotional expenses), member servicing costs and costs of financial and other reports used by the fund.
17 The "commissions" in these plans are paid to an advisor as compensation for the plan-related services they provide to the employer as well as plan member services such as personalised investment advice and investment education.
18 A detailed explanation of each of these guaranteed products is beyond the scope of this paper. However, the information is readily available from any insurance company that sells these products.
19 This study is also not without its challenges, the most notable being that it is based simply on the perceptions of Morningstar analysts.
20 An ETF is an investment vehicle that combines key features of traditional mutual funds and individual stocks. ETFs are open-ended funds which, like index mutual funds, represent portfolios of securities that track specific indexes. A distinct difference is that ETFs trade like stocks and can be bought and sold (long or short) on an exchange and can employ the same trading strategies used with stocks. ETF units can essentially provide unlimited liquidity.
21 Although not shown in detail here, we also analysed 108 funds comprising of $48 billion that are classified as 5 Star funds by Morningstar. The methodology followed by Morningstar for its ratings is as follows. For each fund with at least a three-year history, Morningstar calculates a Morningstar Rating™ based on a Morningstar Risk-Adjusted Return measure that accounts for variation in a fund's monthly performance (including the effects of sales charges, loads and redemption fees), placing more emphasis on downward variations and rewarding consistent performance. (Each share class is counted as a fraction of one fund within this scale and rated separately, which may cause slight variations in the distribution percentages.) The top 10% of the funds in an investment category receive 5 stars, 22.5% receive 4 stars, 35% receive 3 stars, the next 22.5% receive 2 stars, and the bottom 10% receive 1 star. The Overall Morningstar Rating is a weighted average of the funds' three-, five-, and 10-year (if applicable). We found that the weighted MERs are significantly lower than average MERs, meaning larger MFs have a lower MER and that the MERs range from 234 bps for Canadian equity funds, 160 bps for equity and approximately 40 bps for bonds and money markets.. We also calculated various correlations between asset size, MERs and performance and found a negative correlation between asset size and MERs meaning that larger funds have lower MERs; however even in the 5 star funds, the correlation between MERs and performance is negative for both 1-year and 3 –year performance; higher MERs do not seem to result in higher returns.
22 We thank Dennis Yanchus of Investment Fund Institute of Canada for the detailed data and analysis used in these tables.
23 CIFSC stands for Canadian Investment Funds Standards Committee which classifies each fund in 20 categories.
24 Only CIFSC categories which have an index fund option available are represented in table 13 and 14.
25 This is represented by the difference between the MAF of the index fund and asset-weighted MAF of a sample of actively managed funds within the CIFSC category.
26 This is represented by the difference between the MAF charged by active fund managers on their retail versus fee-based mutual fund series. See the glossary for more details.