As investors, we are always looking for the best asset managers to manage our money at the lowest possible cost. But in order to determine who the best managers are and what their skills are worth to us, we need to measure those managers’ performance against some reference point or benchmark.
According to online encyclopedia Wikipedia, the term “benchmark” originated from the chiselled marks that surveyors made in stone structures and into which they inserted a bracket that held a levelling rod, or bench. The benchmark ensured that the rod was placed in the same position in future.
But when it comes to measuring managers’ performance, it seems that the issue of which benchmarks to use is no longer cast in stone.
For many years, it has been common practice in the investment industry to use as a benchmark an index that represents the market or markets in which a portfolio invests.
The unit trust industry, for example, publishes the performance of equity, bond and real estate unit trust funds compared with their respective indices, which represent the universe of securities in which those funds can invest, Leon Campher, the chief executive officer of the Association for Savings & Investment SA (Asisa), says.
The FTSE/JSE All Share index (Alsi) is used as the benchmark of the universe of shares in which a domestic equity general fund can invest.
A share is included in an index, such as the Alsi or the Morgan Stanley Capital index (MSCI), based on its market capitalisation. The market capitalisation of a share is the total value of all the company’s shares in the market – the number of shares in issue multiplied by the price of the share.
The unit trust industry publishes the performance of asset allocation funds without any standard sub-category benchmark. Asset allocation funds are typically benchmarked against composite indices. These indices are made up of market capitalisation indices that represent the different asset classes in which an asset allocation fund can invest.
A common composite asset class benchmark for domestic balanced funds, for example, is 60 percent of the Alsi, 10 percent of the FTSE/JSE Listed Property index, 20 percent of the SA Bond index and 10 percent of the Short Term Fixed Interest index.
When, based on his or her outlook for the markets, a fund manager deviates from a strategic composite benchmark, it is referred to as tactical asset allocation.
Problems with traditional benchmarks
The market capitalisation indices that are typically used as benchmarks can be problematic, particularly in an equity market that is dominated by a few large shares, as is the case in South Africa. The
Alsi includes some large shares. The weighting of these shares in the index may be so large that managers may not want to hold similarly high concentrations in their portfolios. A large share may be listed on other stock exchanges and have large holdings in the hands of foreign shareholders, making it less accessible to South African investors and fund managers. Managers may be prevented by regulation 28 under the Pension Funds Act from holding such a high weighting of a share in their portfolio if they are managing a retirement fund investment.
Unit trust fund managers who manage retirement fund portfolios are restricted by regulation 28 from investing more than five percent of the portfolio in a share with a market capitalisation of less than R2 billion and from investing more than 10 percent of the portfolio in a share with a market capitalisation of R2 billion or more.
Unit trust fund managers are also required in terms of the Collective Investment Schemes Control Act to have enough liquid investments (principally cash) that they can use to pay you out for your units should you decide to redeem them. This cash holding will, purely by chance, benefit a manager when the equity market declines and will place a drag on performance when the market does well.
To compensate for the problems of foreign ownership and share ownership limits, the Shareholder Weighted index (Swix) and the Capped index (Capi) were developed for both the Alsi and the Top 40 indices. The Swix downweights shares that have a large proportion of foreign shareholders, while the Capi caps at 10 percent the weighting of any single share in the relevant index.
Many managers of domestic equity funds have adopted either the Swix or the Capi as a benchmark, despite the fact that the benchmark for funds in this sub-category remains the Alsi.
But domestic equity managers have struggled to out-perform the Swix, as well as the Alsi.
Recent research by Daniel Wessels, of Cape Town-based financial planning practice Martin Eksteen Jordaan Wessels, shows that less than 50 percent of domestic equity general, value and growth funds have out-performed either the Alsi or the Swix over the one-, three-, five-, seven- and 10-year periods to the end of December 2009.
Roland Rousseau, a quantitative analyst who worked for Deutsche Bank before starting A-Dex, consulting on portfolio construction and the design of more “intelligent” indices, says that on average at least two-thirds of actively managed funds worldwide under-perform a common, broadly diversified market index.
Campher says the fact that the Alsi is the benchmark for the domestic equity general sub-category does not necessarily mean that your fund can or should beat this benchmark.
Each fund will set its own benchmark that is appropriate to the performance it is likely to deliver given its investment mandate and style, and this benchmark must be disclosed on the fund’s fact sheet, he says.
Sunette Mulder, a senior policy adviser for Asisa, says you, as an investor, or your financial adviser, must take note of the fund’s benchmark and its performance relative to its chosen benchmark.
Market indices aren’t good
South African fund managers choose the market indices against which their funds’ performance should be benchmarked, and the investing public is left with the difficult task of determining if that benchmark is a fair one.
Rousseau says that, internationally, institutional investors, such as “smart” pension funds and multi-managers, have realised that indices do not make good benchmarks. The problem with indices, he says, is that the choice of which index to use is largely subjective and arbitrary; it tells you nothing about the skill of the manager.
If you look at the performance of Allan Gray Equity Fund relative to the Alsi over the past two, three and four years to the end of October last year, it appears to have out-performed this index, but if you compare its performance with the Swix, it has under-performed, Rousseau says. It is difficult to tell whether or not the manager has skill, he says.
If you then consider the performance of the Marriott Dividend Growth Fund, it appears that this fund’s manager has greater skill than the Allan Gray manager, Rousseau says. Finally, if you add the performance of the FTSE/JSE Dividend Plus index, it appears as if no active manager has skill relative to this passive index, he says.
As a result, Rousseau says, a manager can appear to be skilled relative to one index but not to another, and “it makes no sense for a manager to be both skilled and unskilled at the same time”.
Know the risks
When you are presented with a return of 10 percent a year by Manager A and a return of 15 percent a year by Manager B, you automatically assume that Manager B must be more skilled, Rousseau says. But you do not know if Manager B was just lucky or if the higher return was as a result of the manager taking additional risk.
In the same way, if you feel unwell at night, take a pill and feel better the next morning, you do not know if you feel the way you do because the pill cured you or if your immune system fought off the illness, Rousseau says.
If you hire a manager and it out-performs its benchmark index after one year, does this mean you found the right manager or did you just choose an inappropriate benchmark? Should you keep this manager? Should you instead invest with a manager with a five-year track record of out-performance, even though research shows that past performance is no indicator of future performance, Rousseau asks.
The missing link in identifying managers with true skill is knowledge of the level of risk they take, he says.
The return of a manager’s fund on its own, or relative to a benchmark index that does not reflect the same risks to which the actively managed fund is exposed, is completely meaningless, Rousseau says.
“Only once we compare the active manager’s returns to the risks taken, can we judge if somebody is skilled or not. Indices don’t do this.” He says that Thomas Schneeweis explains this well in his 1999 paper “Alpha, alpha, who’s got alpha?” in the Journal of Alternative Investments.
“Comparing your portfolio return to the Standard & Poor’s 500 or any other benchmark is inappropriate unless your strategy responds only to the same return drivers that drive the S&P500 or the cited benchmark. Similarly, it is not appropriate to say that you have a positive net risk-adjusted return simply because your return is greater than the risk-free rate, unless your portfolio is risk-free,” Schneeweis wrote.
Anne Cabot-Alletzhauser, the head of Alexander Forbes’s research institute who has been in the multi-management industry for 16 years, says that performance alone tells you nothing about a manager’s skill.
She agrees with Rousseau that the way to measure skill is by measuring the risks a manager takes. People are quick to assume – wrongly – that out-performance of a market index shows how skilled a manager is, Cabot-Alletzhauser says.
Rousseau says that 80 to 90 percent of a fund’s returns come from the risks to which it is exposed, but these risk factors are out of a manager's control. The manager does not have any control over how the rand moves or what return the equity market delivers, for example. Often, investors are willing to pay very high fees for a higher return, when all they are getting is higher risk, not necessarily any higher skill, he says.
If you want to beat an asset allocation or a balanced fund with a benchmark based on a 60-percent allocation to equities, 30 percent to bonds and 10 percent to cash, you simply need to go overweight in equities, Rousseau says. It is a well-known fact that, irrespective of a fund manager’s skill, equities deliver a higher return over time than do bonds and cash, because equities are more risky, he says.
Equities’ return above cash is positive over nearly every three- to five-year period historically, Rousseau says.
“You therefore don’t need an active manager to out-perform this asset allocation benchmark over time; you just need more equity risk and a little patience.”
Balanced fund managers therefore charge fees (even very high performance-based fees) for taking additional risk, irrespective of their level of skill, Rousseau says.
In the long run, fund managers very seldom, or only for very short periods, go underweight in equities, clearly indicating that their returns are driven by excess equity risk, not tactical skill, Rousseau says.
Similarly, he says, to beat a general equity market index, such as the FTSE 100, you just need to go overweight on small-cap shares, and, to beat the MSCI, you just need to go overweight on emerging market indices. Taking these positions increases your risk, but they also increase your returns.
Globally, small caps deliver higher returns above cash than do traditional equity indices, and emerging markets, which have higher risks, deliver more than developed markets, Rousseau says.
Cabot-Alletzhauser says the manager of a balanced fund that boasted that it had over a 10-year period been able to out-perform the common strategic benchmark of 60 percent domestic equities, 20 percent bonds, five percent cash and 15 percent global equities missed the point that every manager in the Alexander Forbes Large Manager Watch had achieved the same “miracle”. More careful analysis showed that this out-performance had nothing to do with getting the asset allocation right: as the bull market from 2003 to 2007 progressed, all the managers in the Large Manager Watch allowed their allocations to equities to drift higher than 60 percent, to a maximum of 75 percent of the fund (the maximum equity exposure for retirement funds), she says.
Over the past 10 years, good stock-picking by managers has massively outweighed the contribution to their performance from asset allocation – both good or bad, Cabot-Alletzhauser says. Yet, she says, institutional investors, such as retirement funds, are using surveys that rank managers based on their performance relative to their peers to determine which managers are most skilled at asset allocation.
Customised benchmarks
Rousseau says there is no single ready-made index that can match all the risk drivers or exposures of an active fund. Instead, appropriate benchmarks that take into account the specific risks an active fund manager takes must be tailor-made to suit each fund. If the manager’s returns out-perform this risk-adjusted benchmark, you can assume that the manager has skill, Rousseau says.
Another popular, more robust test of the manager’s skill would be to calculate all the possible active portfolios with the same exposure to market sectors as the active manager’s fund, he says. Only if the active manager out-performed the average return of all of these possible portfolios with the same risk or sector exposures, could you conclude that the manager was skilled, Rousseau says.
Many managers have been fired by South African investors, despite delivering a good risk-adjusted return, because their return was lower than an arbitrary index such as the Swix, Rousseau says.
Managers are often wrongly penalised for being less risky (that is, more conservative) than the index, he says.
Cabot-Alletzhauser says one approach that can help to get to the heart of the skill in managing South African equities is a methodology known as the portfolio opportunity distributions approach.
By comparing a manager’s performance against an outcome that has been derived from Monte Carlo simulations of thousands of different portfolios constructed from the same universe of shares in which the manager would have been able to invest, you can at least know you are comparing like with like.
In the absence of that approach, she argues that investors who use conventional indices must appreciate that there are times in the market’s cycle when a manager that earns a mere one percent more than the market as measured by a market capitalisation index should be deemed highly skilled, and there are times when it is difficult to say whether a manager that achieved six-percent out-performance of the market is skilled. This is because the out- or under-performance is more often a function of market structure changes than of manager skill, Cabot-Alletzhauser says.
If the markets become highly concentrated in terms of performance being driven by just a handful of shares, asset managers that, in terms of regulations, are required to hold diversified portfolios of shares, with no more than 10 percent of the portfolio in a single share, cannot possibly out-perform the market, regardless of skill, she says.
Similarly, if the market does not provide some measure of differential performance between its various segments, it is extremely difficult for an asset manager to demonstrate differential performance from the market, Cabot-Alletzhauser says.
Finally, manager skill is as much about portfolio construction as it is about picking the right shares or asset classes, she says.
When determining the skill of a manager, it is necessary to get an insight into the efficiency of the manager’s portfolio construction because this determines how efficiently a manager can use its stock selection skills to out-perform a market benchmark. This means the manager needs to understand thoroughly how each asset or share in the portfolio behaves in relation to the others and to the chosen benchmark to determine the optimal weighting that will translate into out-performance, Cabot-Alletzhauser says.
Simplistically, she says, if equities are more volatile than bonds but we have no idea which of the two asset classes will out-perform, a manager that favours bonds over equities at a given point in time would have to have a significantly higher overweighting in bonds than in equities to translate that view into out-performance. But a manager that favours equities over bonds would require a significantly lower overweighting in equities for the same magnitude of out-performance. This is often a counter-intuitive “ask” for fund managers, she says.
The bottom line, Cabot-Alletzhauser says, is not to chase managers who have been performing well relative to an index if you do not clearly understand both the changing structure of the market over the time period, the appropriateness of the benchmark and the nature of the mandate constraints that might significantly influence the skill-versus-luck debate.
Common way to measure risks
Rousseau says the most common way to measure the risks a fund manager takes is to use what has become known as the Fama-French three-factor model.
Two American economists, Eugene Fama and Kenneth French, demonstrated that variations in fund managers’ returns can be explained largely by how exposed they were to three common risks that drive returns: a value risk, a small-cap risk and a general market risk.
Fama and French’s research showed that returns above those delivered by the equity market can be earned by taking the risks of holding small-cap and value companies. Holding these shares earns you a higher return than the market but at higher risk, they found. The higher returns from these risks are not due to manager skill, Rousseau says. Anybody, even an unskilled investor, can harvest these higher long-term returns by exposing their fund to these additional risks, he says.
Rousseau says you can build a benchmark made up of the three market forces in the same proportions as the manager is exposed to these risks in its portfolio.
Research by the University of Cape Town’s Professor Paul van Rensburg, now Rousseau’s business partner, has shown that a fund’s exposure to the resources and financial indices can explain 80 percent of the return variation between funds, Rousseau says.
Cabot-Alletzhauser agrees with Rousseau that you need to know whether a manager’s performance was a function of an active strategy or a systemic bias embedded in the manager’s style or philosophy.
She says the remarkable performance of large local balanced fund managers from the start of the decade to the market crash of 2008 was a distinct function of the fact that the persistent equity exposure in these portfolios simply rose as the market rose.
Not everyone agrees
Pieter Koekemoer, the head of personal investments at Coronation, says most active fund managers change their exposure to different segments of the market based on their research into market conditions. Determining when to go overweight or underweight in a particular asset class or segment of the market is often an active decision, rather than a reflection of a style bias, he says.
For example, if a manager is overweight in small-cap shares at a particular time, that may be an active outcome based on current valuation measurements rather than a permanent feature of the portfolio, Koekemoer says. If a benchmark changes continually to capture those changes, it will capture the manager’s style after the fact, he says.
However, if a manager has a permanent exposure to a particular style, that should be taken into account when setting its benchmark, he says.
Rousseau says the correct way to capture a manager’s risk exposure is not to use hindsight but rather a “forward-rolling window” as described by William Sharpe, the economist who developed the Sharpe ratio to measure risk-adjusted performance.
Rousseau says similar research that he and Van Rensburg conducted demonstrates that in South Africa over 90 percent of the variability in the returns of actively managed equity funds can be explained by common factors, such as value and momentum risks, without resorting to hindsight.
Koekemoer says many asset managers all over the world are now trying to analyse winning investment styles and to build a model to replicate them. But this does not mean you should create a new benchmark each time a manager decides it is time to shift its focus in the market based on its analysis of the drivers of the performance of the market.
He says to use simulations of all the possible portfolios a manager could hold and consider an average of those returns will not help investors much, because it is likely that a manager will out-perform some of those simulations and under-perform others. However, these portfolios are not available to invest in, and it is not possible to know upfront which ones will out-perform.
However, Rousseau says that investors and multi-managers often mislead themselves by looking at a manager’s historical track record, which, studies have shown, has no relationship whatsoever to future performance.
What we do know upfront from research, he says, is that certain higher-risk exposures, such as value or momentum investing, do consistently deliver higher returns on average, going forward.
Piet Viljoen, the founder of independent asset manager Regarding Capital Management (RE:CM), says an active manager should take risks but should change the exposure to those risks depending on how they are priced in the market.
You should not use benchmarks or performance relative to benchmarks over short, arbitrary periods to determine which is the best manager to manage your money, Viljoen says. Instead, you should take the following three steps:
* Look for a manager that has a sound investment philosophy and a process that you can understand.
* Make sure you are comfortable with the investment philosophy and process, and make sure the manager can describe the profile of returns you can expect to earn. The manager should be able to tell you what returns you should expect above inflation over the longer term.
* Check the manager’s past return profile against its philosophy to see if it makes sense.
These steps are a qualitative process rather than a quantitative one, Viljoen says.
Simple benchmarks that represent the investment opportunities available to the manager should then be used to measure returns over at least a full market cycle, from peak to peak or trough to trough, he says.
RE:CM uses the Alsi as the benchmark for its domestic equity fund. Viljoen says the returns of the Capi, Swix and Alsi will converge over time. Over a full business cycle – from the peak of the market to its next peak or from trough to trough – they will deliver the same results.
From the local equity market trough in October 2003 to the next trough in April 2009 – some five-and-a-half years – the Alsi returned 147 percent, the Top 40 index returned 140 percent, the Alsi Capi returned 156 percent, the Alsi Swix returned 164 percent and the FTSE/JSE Equally Weighted Top 40 index, which equally weights the top 40 shares, returned 165 percent. These were differences of less than two percent a year, Viljoen says.
The economy grows at between two and five percent a year above inflation and that is the return to which you can be exposed, he says. A manager must look from there how it can find additional value.
Trying to identify the exposure to risk in a port-folio is just too complicated, and it is easier to choose a manager that can make a good investment, Viljoen says. If you decide to track various indices rather than invest with an active manager, you will still need to pay an expert to decide which index to track.
He says being able to understand the profile of the returns that a manager's philosophy will deliver is the same as building a customised benchmark to see what returns a manager can deliver, as Rousseau suggests.
RE:CM’s philosophy is to buy good-quality companies the earnings of which are more robust than those of the average companies and to buy these companies cheaply to benefit from their potential return to fair value, Viljoen says.
The returns earned by RE:CM’s funds will depend on the opportunities in the market. When the market is cheap, the prospects of returns will be high, and when it is expensive, the prospects of returns will be lower, he says. Over a full business cycle, however, RE:CM expects a return higher than that of the Alsi.
Smart beta is raising the bar
The development of products that can capture returns obtained through a particular investment style or tilt to the market – known as smart beta products – will raise the benchmark for active managers and give ordinary investors a better idea of the returns that can be delivered by tilting a portfolio to certain market segments.
Smart beta products track a style index or use computer-based investment strategies to deliver returns, without the costs of active management.
One smart beta product that has reached the South African market is the Research Affiliates Fundamental index (Rafi). The Rafis are known as fundamental indices, and they seek to avoid what is commonly cited as a disadvantage of market-weighted indices: market capitalisation indices are overweight in overvalued shares and underweight in undervalued shares.
The FTSE/JSE Rafi Alsi and the FTSE/JSE Rafi 40 index use fundamentals – such as an analysis of dividends, cash flow, quality of earnings, book value and sales – to weight shares in these indices.
The enhanced Rafi methodology has other accounting metrics, such as quality of earnings and financial distress, to determine the shares in the index, along with more frequent rebalancing of the shares.
Paul Stewart, the managing director of Plexus, says Plexus has decided to track the Rafi indices and the enhanced Rafi indices because market capitalisation indices are not good benchmarks.
Rafi indices elsewhere in the world now have a five-year track record, and have out-performed the relevant indices by two to four percent a year in developed markets and in excess of five percent in emerging markets, he says.
The Rafi index-tracking investments do not deliver returns entirely without skill, because someone still has to do the work to get the shares in the correct weightings, Stewart says.
More and more fund managers are asking to use the Rafi indices as a benchmark for equity portfolios, and similar methodologies are being researched for use in the bond market, he says.
Market capitalisation bond indices allocate the most money to countries with the highest debt levels. Fundamental bond indices will determine in which countries’ bond markets to invest in accordance with their gross domestic product, land mass, population and electricity usage, Stewart says.
Benchmarking against the drivers of risk, as recommended by Rousseau, is a very technical exercise, he says, and it is more important for you, as an investor, to identify the risk factors you want to avoid.
Investment consultants do take account of the risks a manager takes when investing, because they measure performance in risk-adjusted returns such as Sharpe or information ratios, Stewart says.
Rousseau believes the Dividend Plus index is a much better benchmark against which to measure active fund managers than the Swix or Alsi. The FTSE/JSE Dividend Plus index, which weights stocks with a higher dividend yield, captures quite well the excess returns from value investing, he says.
“We can show scientifically that most fund managers in South Africa have a diluted value bias in their funds’ risk profiles. So a value index like the Dividend Plus index is a much better benchmark to measure skill against than the Swix or Alsi, because it more accurately reflects the investment style and risks of active managers than any other index,” he says.
Anyone, even those without any skill, can buy the Satrix Dividend Plus Exchange Traded Fund (ETF). This ETF performed better than all actively managed domestic equity general funds over the three years to the end of October last year, confirming that most managers are diluted value managers, Rousseau says. According to etfSA, the Satrix Dividend Plus ETF returned 11.91 percent a year over this period.
Koekemoer says it is not necessarily appropriate to use enhanced indices such as the Rafi and the Dividend Plus as benchmarks, because these represent additional investment options that compete with actively managed funds, albeit at a cheaper price.
But investors need to decide whether they want to track an index – and, if so, which index – or if they want to back an active manager to deliver out-performance, he says.
An index-tracking investment, such as an ETF, will never change the index it is tracking. Instead a new fund that tracks a new index will be launched when these indices become available. Active fund management, however, is a continuous process, and an active fund manager can adapt to market conditions.
Investment funds that track fundamental indices may, however, raise the bar for active managers, Koekemoer says, because it is more difficult for managers to claim superior skill for the out-performance of a chosen index, such as the Alsi or the Swix, when an investment that tracks the Rafi or the Dividend Plus index can also reliably deliver more than the chosen index.
George Herman, an investment strategist at financial planning specialists Citadel, says the development of non-market capitalisation indices is very positive, but investors need to realise that the choice to use a fundamental index is an important one and can involve greater risks. The additional complexity will not necessarily ensure better results, he says.
While research shows that fundamental indices do deliver better results, they may not always do so, Herman says.
Pin-pointing alpha
Benchmarks are used to distinguish the returns you can earn from the market – commonly known as beta – and those that are a result of a manager out-performing the market – known as alpha.
Active fund managers argue that the compounding effect of the out-performance they can earn for you is significant. But before you agree to pay for this, you need to know what you are getting and hence the importance of an appropriate benchmark.
Rousseau argues that as the construction of benchmarks becomes more refined, the distinction between alpha and beta becomes more exact, and investors should pay appropriately only for the returns that are indeed better than those they would have earned from a passive investment.
Currently, managers earn a fee on the assets they have under management, but in Australia consultants say that asset managers do not deserve fees based on assets under management that are a result of market movements, Rousseau says.
They say managers are like the flea on the back of an elephant (the market). If the elephant moves, the flea (the manager) cannot take credit for it. Australian consultants are proposing that if a manager earns a fee of two percent on, for example, a R100-million investment and the market moves up by 10 percent, the manager should continue to earn two percent of R100 million and not two percent of R110 million. The manager therefore earns fees only on the portion due to out-performance and not on those due to general market movements that are completely out of the manager’s control, Rousseau says.
The same rule applies when markets decline. In this case, the manager will not earn lower fees simply because the market declined. The manager’s overall fees will, however, decline whenever it under-performs its benchmark. While this may be a discussion for your retirement fund trustees to have with your fund’s asset manager, as passive investments that deliver more than the market become more widely accessible, there is also likely to be pressure on the fees that asset managers charge on your investments. As the gap between the return of these investments and that of active managers narrows, more investors will question the need to pay the level of fees they are being charged.
A rethink on some performance fees may also be called for. Performance fees set hurdle rates, or minimum returns, above which the performance fees will apply. These hurdle rates may be the same as or higher than the benchmark for your investment.
Koekemoer says that when it comes to performance fees, hurdle rates should be fair. He says the two key questions you need to ask are:
* Given the expected return and out-performance, is the expected fee reasonable relative to the returns you will earn?
* Are you giving the manager a free option to earn a fee for good returns without having to experience pain for under-performance, or does the fee reward good returns and punish the manager for under-performing?
BENCHMARKS AND THE TARGET FOR YOUR INVESTMENTS
You need a financial plan to ensure that you will reach your goals despite potential mishaps, such as illness or disability. A good financial plan will distinguish between short-, medium- and long-term goals, and will establish a means for you to achieve those goals.
A long-term goal, such as saving enough for your retirement, may involve your saving 10 percent of your income for the rest of your working life and earning a particular return on those savings. Usually, that return is expressed as a return above inflation – for example, inflation as measured by the consumer price index (CPI) plus five percent or CPI plus three percent. That is your goal, but it should not necessarily be the benchmark that is used to determine if your fund manager is performing how it should, Anne Cabot-Alletzhauser, the head of the Alexander Forbes research institute, says.
Some managers offer funds that target returns above CPI as a packaged solution for you, and it may seem logical to choose one of these funds that matches your targeted return.
Roland Rousseau, the managing director of A-Dex, says “the harsh reality of using a cash or CPI benchmark is that there is a zero-percent chance of identifying any manager with skill”, because these benchmarks do not represent what the fund invests in and do not provide relevant comparisons if your goal is to identify whether or not a manager is skilled.
There are many portfolios that will beat CPI or cash without any skill, Rousseau says.
For example, a naive “zero-skill” portfolio that is equally spread among equities, as represented by the FTSE/JSE All Share index (Alsi), bonds, as represented by the JSE All Bond index, and cash would have delivered about CPI + 10 percent over the past decade, he says.
Rousseau says cash and CPI benchmarks are typically used by hedge funds and absolute return funds. But cash and CPI usually have no relevance to the manager's investment process, and the portfolio should be benchmarked against something that involves the same level of risk.
“Cash and CPI just don’t cut it as benchmarks”; they are merely target returns, he says.
Another way to look at this is to check if your benchmark’s returns have a high correlation with the active manager’s returns. If the benchmark is not very highly correlated to the active fund's returns, the benchmark is inappropriate and useless. Why bother using a benchmark such as CPI + X percent that has no statistical relationship with the active manager’s returns, Rousseau asks.
Absolute return funds and hedge funds are volatile (that is, risky), while a cash or CPI benchmark has no volatility, indicating that there is no meaningful correlation between the active fund and the cash benchmark, he says.
If an absolute return fund has no correlation to the cash return benchmark, we might as well use the water levels of the Vaal River or birth rates in the Western Cape as a benchmark, because these also will not have a correlation to the active fund’s returns, Rousseau says. But on this point there are different views in the investment industry.
Pieter Koekemoer, the head of personal investments at Coronation, says asset managers use cash and inflation-linked targets, because consumers need good returns over time after costs – this is what defines whether or not they will retire comfortably.
An inflation-linked benchmark can be a good option if you invest in an asset allocation fund that has a risk constraint, such as to invest no more than 60 percent of the fund in growth assets and not to lose your capital over 12 months, Koekemoer says. Such a fund could not just take unconstrained risks to achieve a return that out-performs its benchmark, because it needs to remain within a clearly defined risk budget, he says.
However, when an asset allocation fund does not have a short-term risk constraint and aims to maximise returns over the long term, it should rather have a composite benchmark made up of indices that represent the likely asset classes in which the fund will invest.
Paul Stewart, the managing director of Plexus, also does not agree that using cash and inflation as benchmarks is a bad thing.
While Stewart agrees that you cannot invest in an asset that will provide you with inflation plus three percent, he does not believe it is wrong to measure returns against such a benchmark.
He says that multi-managers such as Plexus will measure the performance of the managers of the underlying assets against benchmarks that can determine whether or not those managers are skilled, but the combined return of the asset classes is measured in terms of inflation, and the risks are constrained by setting a tracking error (a limit on the deviation from a benchmark).
Plexus believes there will be an evolution of indices, but Stewart says somebody still has to choose in which indices you should invest and in what proportions, and the way in which the indices are put together will give good multi-managers a competitive edge.
The manager must decide how to combine the asset classes in order to achieve a particular return above inflation, and this becomes the minimum return. If a manager earns a performance fee for beating that minimum return, it means the underlying asset classes have beaten their benchmarks, Stewart says.
Piet Viljoen, the founder of independent asset manager Regarding Capital Management (RE:CM), says it does not matter if you use a benchmark such as the Alsi or one based on inflation, because the share market as represented by the Alsi will deliver an above-inflation return. The Alsi, he says, has historically returned inflation plus six percent.
He says that using an inflation-related benchmark means that a manager cannot deliver a negative return and then claim that this was because the benchmark returned a negative return.
Many people believe you should measure the risks a manager takes by its under- or out-performance of a benchmark over an arbitrary period of time, but Viljoen believes risk should be defined as the chance of permanently losing money. You can evaluate that from the manager’s performance and processes, he says.
George Herman, an investment strategist at Citadel, says investors need to determine their investment objectives and to select appropriate benchmarks to measure whether they are achieving those objectives.
Younger investors who have a high-growth objective should not choose an inflation-linked benchmark, Herman says, but rather a market index.
Although you can get benchmarks that aim to out-perform inflation plus a certain number of percentage points, the risk of not achieving CPI over the short term increases. A pensioner who is not able to take much risk can consider a CPI benchmark, Herman says.
However, he acknowledges that CPI is an easy benchmark to beat, and this can be achieved simply by investing in inflation-linked bonds.
HALLMARKS OF A GOOD BENCHMARK
A good investment benchmark should meet a number of criteria.
In a paper entitled “Dynamic benchmarking: are you using the right yardstick”, Roland Rousseau, the managing director of A-Dex, has drawn up a list of criteria for a good benchmark based on international research that he conducted while working for Deutsche Bank.
The criteria identified in international research are:
* The benchmark should be unambiguous. It must be specified, and a quantifiable way of measuring it must be stipulated, Rousseau says.
* A benchmark should be investable. The benchmark (which can consist of various indices) should be replicable or trackable, Rousseau says. An example of a commonly used benchmark that is not replicable is a peer benchmark – for example, when a manager is expected to out-perform the average of its peers. You cannot guarantee the performance of the peer group, Rousseau says, and you cannot replicate it.
CPI + X percent is not replicable or trackable and therefore is inappropriate, he says. A benchmark that contains an illiquid stock is also not replicable, because the share may not be liquid enough or the fund may be too large to hold a liquid amount of that share.
Funds of differing sizes, with different risks and with different liquidity, therefore all need different benchmarks precisely because their benchmarks must be investable, Rousseau says.
Pieter Koekemoer, the head of retail at Coronation, says comparing performance to peer groups can be relevant, as long as the funds in the peer group have similar constraints on the risks they can take. This will show the range of available returns, he says.
Peer group performance should not, however, be used as a hurdle for performance fees, he says.
George Herman, an investment strategist at Citadel, says a peer benchmark can be problematic in determining the top manager, because you may be identifying a manager who is the best of a bunch of fools.
As a benchmark, peer groups are not anchored in reality, he says, and you should use peer group comparisons only to discuss with your manager why its performance differs from that of the group.
* A benchmark should be appropriate. A benchmark must be appropriate for the risk and performance characteristics of the fund, Rousseau says.
It would be inappropriate to use a growth benchmark for a value fund manager or a large-cap benchmark for a small-cap manager, he says. This is exactly why peer group benchmarks that do not take risk properly into account are so dangerous, he says.
Every manager has a unique and constantly changing blend of common risks, and comparing them on a like-for-like basis without adjusting for this risk is inappropriate.
* A benchmark must be specified in advance. If you cannot say in advance what the benchmark is, it makes it difficult for a fund manager to take a position against it, Rousseau says.
A benchmark that stipulates that a manager should beat the median return of its peer group is not possible to define in advance and inevitably leads to conflict between the manager and the investor, he says.
* A benchmark should include only as many shares on which a manager can have an informed opinion. Managers try to beat the benchmark by being informed about shares, Rousseau says.
Large fund managers have large research teams and can be informed about a larger number of shares. Small fund managers cannot, for example, know all 500 shares in the United States's Standard & Poor’s 500 (S&P500).
* A benchmark should have a low turnover of securities when it is rebalanced. Rousseau says if there is too much churning of securities in the index, it will become costly to track. An active manager that attempts to out-perform the benchmark should have a higher turnover, however.
* A good benchmark should not force a fund manager to take a negative position – that is, go underweight on a stock the manager likes. For example, the size of Anglo American in the All Share index (Alsi) is so big that no fund manager will dare go overweight in this share, Rousseau says.
* A benchmark’s performance should be measurable. If you cannot frequently measure the performance of a benchmark or study its behaviour, it is difficult to beat it, Rousseau says.
A peer group benchmark (for example, to out-perform the average of the manager’s peers) is not measurable and not much use to a fund manager.
Rousseau has a few criteria of his own for benchmarks, based on his research:
* A benchmark should be fair. This means you should have a 50-percent chance of under- or out-performing the benchmark.
Rousseau says there are times when the probability of beating the Alsi is not more than 20 percent and there are times when it is extremely easy to beat the Alsi, even without any skill.
Indices, such as the Alsi, with a high concentration of stocks are not fair benchmarks for a retirement fund that is prevented from holding a high weighting in a single share.
In a strong bull market, Rousseau says, it is difficult for a portfolio manager with a large fund to out-perform a market index such as the Alsi. The performance of large stocks will often have high correlations with each other in a bull market, and the Alsi will be close to being optimal or unbeatable, he says.
If there is no chance of beating a benchmark, you may as well track it passively, Rousseau says.
He says the above suggests that we should have different benchmarks for active and passive investing. Most pension funds in South Africa have the same benchmark for active as for passive investing.
Rousseau says that having different benchmarks for active and passive investing does not mean that the two ways of investing are mutually exclusive – you can still allocate your investments across both methods, but you will need different benchmarks for each part of the portfolio if you want to identify whether or not a manager is skilled.
* A benchmark should be stable. The S&P500 and the Alsi have high concentrations in only a few stocks and are not very stable (that is, they are more volatile than actively managed funds), Rousseau says.
* A benchmark should be dynamic. It should change continuously with the portfolio for which it was designed, Rousseau says, especially if your benchmark needs to track appropriately the changing risks of active managers.
If you do not use dynamic benchmarks, the active fund’s risks will quickly deviate from the benchmark’s risk profile, making the benchmark inappropriate.
This article was first published in the 1st quarter 2011 edition of Personal Finance magazine.