Journal of Law and Financial Management
|
|
Home
| Databases
| WorldLII
| Search
| Feedback
Journal of Law and Financial Management |
|
By Tom Valentine[*]
This article provides an overview of some of the more important theoretical and practical issues relating to risk profile in the context of the financial planning process. It is argued that current industry best practice in risk profiling does not match well with theoretical literature on financial risk and risk taking. Alternatives to current practices are briefly reviewed.
An important aspect of providing advice on investment is incorporating appropriate information on risk into that advice. Appropriate information has two components. First, the adviser must understand, and where relevant provide the client with information on, the risks inherent in each type of investment. Second, the adviser should establish a risk profile for each client (sometimes this process is described as determining the client’s risk tolerance, although these two objectives are slightly different) and to make recommendations in sympathy with this risk profile. This approach is strongly supported by the Financial Planning Association (FPA) and the Australian Securities and Investments Commission (ASIC) and it is seen as compliant with the “know your client” and “know your product” rules originating from Section 851 of the Corporations Law[1].
A number of attempts have been made within the Investment Advice Industry to develop operational procedures to meet these requirements. It is the purpose of this paper to evaluate these approaches and this will be done in two steps. The initial step will be to clarify the theoretical basis of the process under consideration. That is, what are we attempting to achieve and what would be the ideal although not necessarily practical way of achieving it? The second step is to look at the techniques currently used to see how closely they approximate the ideal.
In undertaking this analysis, this paper considers first, how the risk of assets should be measured and how this is currently done within the industry. Secondly, the paper discusses the meaning of a “risk profile” and the various ways such a profile can be constructed for an investor. The final part of this section considers whether such a profile actually exists.
The third section of the paper raises the question - is it desirable to force advisers to make judgements in this area? An alternative is a prescriptive approach in which advisers make recommendations in the best interests of their clients.
The risk of any single investment is that its return will differ from its expected value. It is often measured by the volatility of historical returns, although this approach is always subject to the caveat that the past is not always a good predictor of the future. One can attempt to predict future volatility of return on the basis of a priori analysis, but there are few examples of success from this approach.
An important element in considering the volatility of returns is that the volatility can be calculated over a range of time periods and the risk of a given asset class will depend on the investment horizon chosen. For example, share returns are highly volatile on a monthly basis but not on a ten-year basis. When the risk of share investment is described to an investor, it should be the risk appropriate to the investor’s investment horizon.
This risk is often divided into a number of sources[2]. Some examples are:
Establishing the relationship between these sources of risk and the volatility of the returns on specific assets requires complex analysis. There would be little point in attempting to explain these relationships to individual investors. However, they are useful for some purposes. For example, investors need to be told that high-yield bonds pay that high yield because they have a low credit rating which refl ects the greater likelihood that their issuers will default. This example indicates that examining the volatility of asset returns does not provide comprehensive information on the risk involved in a particular asset.
A major concern about current industry practice (as encouraged by the FPA and ASIC) is that the investors’ time horizon is treated in a very cavalier fashion. The risk of assets is measured on the basis of the volatility of very short-term returns. Typically, the volatility of monthly returns is used for this purpose. This approach leads to assets being classified in the following order, going from least risky to most risky:
Figure 1 is an example of this type of ordering taken from a fund manager’s brochure.
However, this ordering could be misleading for an investor with a long time horizon. Investors in superannuation should have a long time horizon. Given that termination benefits can be rolled over into an allocated pension, even investors close to retirement have a long time horizon. In this case, shares cannot be regarded as a particularly risky investment. Over long periods, shares have yielded 6-7% over fixed-interest investments and with little risk. This is not to deny that, for investors with a short investment horizon (such as a couple who want to invest money which is to be a deposit on a home to be purchased within the next year), shares are a risky investment. But, so is every other asset but cash. Fixedinterest investments can suffer significant capital losses if they cannot be held to maturity.
Another problem with the current approach is that it is based on a “building blocks” process. We add the risks to which each asset is subject and those which are subject to the greatest number of sources of risk are judged to be the riskiest assets. This problem can be illustrated by the comparison between investment in Australian and overseas shares. The building block approach is illustrated in figure 2 below.
Therefore, overseas shares are regarded as riskier than domestic shares. However, there are reasons for doubting this conclusion. First, there is no reason to assume that the risks of share investment are the same for Australian shares and overseas shares. It is possible that offshore share markets are more or less volatile than the domestic share market. Certainly, Australian investors would have a wider range of possible investments overseas than is available domestically. Secondly, there is no reason why foreign exchange risk should be regarded as additive to the other risks. It is possible that the value of the Australian dollar is negatively correlated with the performance of overseas share markets so that foreign exchange risk actually reduces the overall risk of investment in offshore shares.
This last issue points to a very important problem with the current approach - it takes very little account of the possible use of diversification to reduce risk. In the context of a diversified portfolio, the risk of a specific asset is the amount by which it increases the variance of the portfolio return when it is added to a portfolio. The risk of a particular asset in this sense might be considerably less than its risk measured on a stand-alone basis, i.e. as the standard deviation of returns. This means that an asset which appears to be very risky in its own right might be very useful to include in a portfolio if:
(i) it yields a high return; and
(ii) its return is not correlated with the returns on assets already included in the portfolio or is negatively correlated with them.
Ignoring this possibility will prevent advisers from offering clients portfolios which could give them a higher return with lower risk, i.e. it is possible that they are offering them inefficient portfolios.
Mainstream Finance theory has developed an elaborate theory of individuals’ risk profiles. Recent work by Danthine and Donaldson[3] provide a clear exposition of this theory. Its starting point is the utility of money function U(Y) where Y is the investor’s wealth. We offer the investor a choice of Y with certainty or a gamble in which the outcomes (Y+h) and (Y-h) each have a probability of ½. The expected utility of the gamble (E(U)) is:
If U(Y) > ½U(Y+h) + ½U(Y-h), the investor is risk averse. This will happen if the utility function is strictly concave, i.e. its second derivative is negative so that the marginal utility of wealth declines as wealth increases. Measures of risk aversion can be calculated from this utility function[4]. For example, the measure of absolute risk aversion is:
where U' (Y) and U''(Y) are, respectively, the first and second derivatives of U(Y).
Danthine and Donaldson suggest[5] that the reciprocal of this measure can be described as “risk tolerance”. This concept gives a precise meaning to the term risk tolerance which is used rather loosely in the financial planning industry. It is generally assumed that investors are risk averse, i.e. that the second derivative of the utility function is negative.
If the utility function is quadratic or if the distribution of returns is normal, an investor’s expected utility of a portfolio’s rate of return is a function only of the expected return on the portfolio and the standard deviation of the return[6]. In this case, the investor’s risk profile is fully represented by indifference curves between the expected return and standard deviation of the return. Along each indifference curve, the level of expected utility is the same. Indifference curves are convex to the origin if either one of the assumptions mentioned at the beginning of this paragraph are valid. This indifference map is the closest theoretical equivalent to the vague industry concept of a “risk profile”.
Mainstream theory indicates that an investor’s portfolio should be created in two independent steps. First, the investor chooses how much of the portfolio will be allocated to a riskfree asset and how much to a portfolio of shares. Secondly, the portfolio of shares is on the efficient frontier which is the set of portfolios that have the highest return for a given standard deviation. This frontier is represented by the curve in Figure 3. The line in Figure 3 shows portfolios that are a mixture of the share portfolio at A and the risk-free asset. The line starts at the point where the whole portfolio is invested in the risk-free asset, earning the risk-free return (Rf). At A, the full portfolio is invested in shares. Portfolios on line B above A involve borrowing to invest in the share portfolio represented by A.
Investors can access the portfolios along the line B. They will choose the one that is just touched by the highest possible indifference curve (the highest expected utility).
How can we use this theory to develop processes to measure the risk profiles of investors? A number of approaches are suggested by the theory.
(A) We can attempt to determine the utility curve U(Y) by offering investors choices between various gambles and certain outcomes. This would allow us to determine a measure of risk tolerance (aversion).
(B) We can attempt to trace out the relevant part of the indifference map by offering investors choices amongst investments with different expected returns and standard deviations of returns. This approach is similar to the “revealed preference” method of estimating the indifference curves of consumers.
(C) It is likely that the expected utility function depends on some exogenous variables such as the age of the investor, the marital status of the investor, the investment experience of the investor, etc. Also, investors’ past decisions (e.g. job history, insurance decisions) may throw light on their risk tolerance. Questions can be asked about these exogenous variables to assist in tracing out the indifference map.
(D) The theoretical analysis discussed above suggests that investors could be presented with choices of the portfolios along the line in Figure 3. All of these portfolios would include:
- The share portfolio A; and
- A holding of the risk-free asset or a borrowing.
What is the share portfolio A? Theory suggests that it is the market portfolio (i.e. an indexed portfolio or a portfolio with a beta of unity). This means that we would offer investors a choice amongst portfolios including an indexed share portfolio and a holding of or short position in the riskfree security. In other words, the theory suggests that investors should hold an indexed portfolio of shares. This conclusion is supported by the recently reported empirical results of Drew and Stanford who conclude that “low-cost (33 basis points per annum, nil entry load) passive asset selection provides superior risk-adjusted returns to fund members than the returns achieved through a high-cost (186 basis points per annum, 4% entry load) active asset selection strategy”[7]. Similar conclusions are also reached in other recent research[8].
This approach is also consistent with the insight that a portfolio with a given beta can be created by:
High share betas arise from high leverage in the company concerned. This dependence is shown by the Hamada relationship[9]:
This formulation refers to the case where the corporate taxation system is a classical one. Davey[10] argues that relatively few investors would accept the risk of gearing . However, investors with high-growth portfolios are already exposed to a geared position.
One problem with this approach is that the “market portfolio” is not restricted to shares. We would actually need to use the full market portfolio including property, fixed-interest investments and offshore investments. Effort would need to be put into developing an efficient frontier, e.g. can the risk/return trade-off be improved by replacing some of the risk-free asset with high yield bonds? However, the argument in this paper suggests that this effort is more useful than the time devoted to risk profiling.
The portfolio presented to clients should incorporate aftertax returns. That is, the choices should be adjusted to take account of the taxation position of the investor, e.g. whether the investor is on the top marginal tax rate of 48.5% or a superannuation fund on a tax rate of 15%.
This theory indicates the nature of the questions which need to be asked to establish the risk profiles or risk tolerance of investors. They should incorporate a numeric risk-return trade-off so that investors have a choice between objective alternatives. The questions should relate to the portion of the indifference map which is attainable by the investor and to the time horizon appropriate to the investor. That is, they should refer to some portfolios in which investors might possibly invest. It would also be useful to provide respondents with examples of the type of variations in returns which result from standard deviations of a different size.
The theoretical approaches set out above are not free from fl aws, and many problems have been identified with the theory outlined in the previous subsection. These are summarised below, but an excellent overview is provided by Nofsinger (2001)[11].
A number of authors have argued that there is a convex section in investors’ utility functions, i.e. that the second derivative of the utility function is positive. This type of utility function produces “risk loving” behaviour. Levy and Levy (2001) survey the literature in this area and carry out an experiment to test this possibility[12]. They conclude that most individuals are not risk averse.
One example of this literature is the work of Kahneman and Tversky[13] who argue that preferences are not defined over actual returns but over gains and losses relative to a benchmark[14]. This approach is called “prospect theory”. In this theory, investors may have an aversion to losses relative to the benchmark.
There is also an extensive literature (much of it based on the experimental approach) which indicates that investors do not make decisions under risk in a rational or consistent way. The following list is largely taken from Nofsinger (2001)[15]:
HHHTTT or HTHTTH
Many people perceive a causal regularity in the first outcome. One example of this illusion is the “hot hands” fallacy in which it is believed that fund managers who have been successful will continue to be successful;
These examples raise significant concerns about the process of risk profiling. They show that individuals’ reactions to risk are very complex and it may not be possible to identify their characteristics by asking a small number of questions. They also indicate that investors’ own views on their risk profiles and risk tolerance can be seriously misleading. Most importantly, this discussion (especially evidence of myopia, inconsistent preferences and cognitive dissonance) raises serious doubts as to whether such a thing as a risk profile actually exists.
Current practice in the financial planning industry has little relationship with the theory outlined in above. One of the highly regarded risk profiling tools used in the industry is ProQuest (2002). It satisfies the “best practice” requirements set out by authors such as Callan and Johnson[16] and Davey[17]. However, this questionnaire does not consistently establish a risk profile in any of the ways indicated by theory for the following reasons:
As outlined by Davey (2002)[18], the current approach requires advisers to develop standard portfolios with different risk/return characteristics and to choose for each client the portfolio which most suits the client’s risk tolerance. It is unlikely that these portfolios are efficient or represent appropriate positions for the clients given their time horizon.
Advisers have been urged to use a risk profiling tool and to provide clients with advice on how recommendations relate to their measured profile[19]. Chambers warns of the possible legal consequences of failing to take account of clients’ risk tolerance:
“regardless of your approach, the law (both Corporations Act and common law) requires that the adviser must have an understanding of the client’s risk tolerance and must make an informed judgement as to whether that risk tolerance is suited to the strategy or strategies necessary to achieve the client’s goals.” [20]
Similar arguments are advanced by Bobbin[21]. In spite of these comments, little is actually known about the legal situation because very few cases have been decided in this area. The most significant example of a similar situation is the foreign currency loan cases decided in the early nineties, which are surveyed by Weerasooria[22]. A major claim of the borrowers in these cases was that the bank concerned had not adequately explained the risks of foreign currency loans to them (and how they could be managed). In the main, the Courts regarded a failure to do this as negligence. Banks which could demonstrate that they had explained the risks were not found to be liable.
These cases were decided on a variety of small points. This situation is typical of the law and underlines the futility of having such questions decided by Courts staffed by lawyers with no commercial experience and little, if any, relevant knowledge. In this environment, people are happy to jump on small legal points because they do not understand the underlying transactions. However, one common point is that banks were not required to measure their clients’ risk tolerance. The only conclusion that comes out of this experience is that investment advisers must clearly explain the risk of their recommended investments to clients.
Nevertheless, many advisers make use of standard risk profiling instruments in spite of considerable cynicism about their outputs. In an era of litigious clients encouraged by ambulance-chasing lawyers, it is useful to be able to say that you have performed a risk profiling exercise using an instrument accepted industry-wide.
An alternative to risk profiling would be to adopt a prescriptive approach. Regulations could be introduced to require advisers to explain the risks of recommended investments and to choose ones which are in their clients’ best interests. The latter requirement implies that advisers need to understand their clients’ financial positions and needs. However, they would not need to measure risk tolerance. For example, assume that a client is thirty years old and is saving for retirement. Then that client should be put into shares and property because history suggests that these assets will give the best ultimate outcome. This would be the case even if the client appears to be a very timid investor. The adviser would also need to pay attention to the construction of a properly structured portfolio which provides a favourable risk-return trade-off.
Current practice in the financial planning industry requires advisers to measure the risk profiles or risk tolerance of their clients. However, the meaning attached to these concepts is quite vague. This lack of clarity is strange because their meaning is very clear in Finance theory. Risk tolerance can be equated to the reciprocal of the coefficient of risk aversion. The risk profile is the map of the investor’s indifference curves between the expected return and the variance of the return. Investors will then choose the portfolio (amongst those available) which puts them on the highest possible indifference curve.
The theory also suggests that investors actually choose amongst portfolios combining a fixed-interest investment (or borrowing) and the market portfolio. This result suggests that we can trace out the relevant portion of the risk profile by asking investors to choose amongst portfolios consisting of the risk-free asset and the market portfolio in different proportions.
Many concerns have been raised about this standard model. There is considerable evidence that investors’ attitudes to risk are inconsistent, myopic and unstable. This evidence raises the question - does a meaningful risk profile exist? Existing risk profiling instruments do not provide information which throws light on risk profiles as defined in the theoretical model. Indeed, it is not clear what they are measuring.
Many advisers use risk profiling instruments simply because this provides them with protection in the event of litigation arising from their advice. The actual legal parameters are unclear. Nevertheless, the points made in the previous two paragraphs suggest that the whole process is wasteful and pointless. The law and regulations should be changed so that any requirement to construct a risk profile is removed.
Instead, advisers should make recommendations in the best long-term interests of their clients. The time and effort that currently goes into risk profiling would be much better used in improving the structuring of portfolios so that they provide clients with the most advantageous risk/return tradeoff. Of course, the risks of recommended investments must be clearly explained to clients. Also, the risks of gearing need to be explained and the plan should include measures (such as access to cash) for mitigating these risks.
[1]. As it then was.
[2]. Macquarie/Financial Planning Association, (2001), Understanding Risk to Achieve Your Financial Goals, FPA, Sydney.
[3]. Danthine, J., & Donaldson, J., (2002), Intermediate Financial Theory, Pearson, New Jersey.
[4]. Danthine, J., & Donaldson, J., (2002), Intermediate Financial Theory, Pearson, New Jersey, P. 44.
[5]. Danthine, J., & Donaldson, J., (2002), Intermediate Financial Theory, Pearson, New Jersey, P. 44.
[6]. Danthine, J., & Donaldson, J., (2002), Intermediate Financial Theory, Pearson, New Jersey, pp. 93 - 97.
[7]. Drew, M.E and Stanford, J., (2001), “Asset Selection and Superannuation Fund Performance: A Note for Trustees”, Economic Papers, Vol. 20, No. 1, pp. 57 - 66. See in particular, p.64.
[8]. Drew, M.E., Stanford, J., and Taraneko, P., (2001), “Hot Hands and Superannuation Fund Performance: A Note for Second Trustees”, Economic Papers, Vol 20, No. 4, pp. 18 - 25.
[9]. Brigham, E., and Houston, J., (1999), Fundamentals of Financial Management, (9th ed), Harcourt College Publishers, Florida. See pp. 617 - 619.
[10]. Davey, G., (2002), “Scientific Risk Profiling”, ProQuest, Sydney.
[11]. Nofsinger, J., (2001), Investment Madness, Prentice Hall, New Jersey.
[12]. Levy, M. & Levy, H., (2001), “Testing for Risk Aversion: A Stochastic Dominance Approach”, Economic Letters, Vol 71, pp. 233 - 240.
[13]. Kahneman, D., & Tverskey, A., (1979), “Prospect Theory: An Analysis of Decision Under Risk”, Econometrica, Vol. 47, pp. 263 - 291.
[14]. See on this point; Danthine, J., & Donaldson, J., (2002), Intermediate Financial Theory, Pearson, New Jersey, at pp. 39 - 40.
[15]. Nofsinger, J., (2001), Investment Madness, Prentice Hall, New Jersey.
[16]. Callan, V., and Johnson, M (2001), “Some Guidelines for Financial Planners in Measuring and Advising Clients About Their Levels of Risk Tolerance”, Journal of Personal Finance, Vol. 1, No.1, pp. 31 - 44.
[17]. Davey, G., (2002), “Risk Profiling and Gearing”, Resnik Comminications Gearing and Investment Conference, July 31 - August 1, Sydney.
[18]. Davey, G., (2002), “Risk Profiling and Gearing”, Resnik Comminications Gearing and Investment Conference, July 31 - August 1, Sydney.
[19]. Callan, V., and Johnson, M (2001), “Some Guidelines for Financial Planners in Measuring and Advising Clients About Their Levels of Risk Tolerance”, Journal of Personal Finance, Vol. 1, No.1, pp. 31 - 44.
[20]. Chambers, L., (2002), “Risk Tolerance: Its All About Returns”, Money Management, August 29, p.20.
[21]. Bobbin, P. (undated), “An Adviser’s Obligation In Establishing A Client’s Risk Tolerance”, The Argyle Partnership, Sydney.
[22]. Weerasooria, W., (1993), Banking Law and the Financial System in Australia, 3rd Ed, Butterworths, Sydney. See in particular, Chapter 17.
[*] Professor of Banking & Finance, University of Western Sydney.
AustLII:
Copyright Policy
|
Disclaimers
|
Privacy Policy
|
Feedback
URL: http://www.austlii.edu.au/au/journals/JlLawFinMgmt/2003/5.html
