Risk aversion defines our attitudes and preferences regarding risk taking. While risk aversion is commonly associated with the finance industry, risk aversion is everywhere: our daily activities, the major decisions we make in life, our respect for the rules and laws of society and even the choices we make in our personal relationships. In this article I focus on financial risk aversion.
Finance industry and academic research are polarised
Financial risk aversion defines our attitudes to taking financial risk. It is an area where industry and academia are poles apart, and in my opinion not well understood by industry in particular. The academic definition of risk aversion is specific, namely the reluctance of a person to accept a bargain with an uncertain payoff rather than another bargain with a more certain, but possibly lower, expected payoff.
Research into risk aversion requires many different skillsets: economists, finance specialists, psychologists, actuaries and behavioural science experts, just to name a few specialties which have a role to play in better understanding risk aversion.
Relative versus absolute risk aversion
Many in the industry summarise the risk aversion of an individual as a single number based on the answers to a questionnaire; I believe this is a dangerous approach. Risk aversion is much more than a numeric measure of a risk attitude: the form of our risk preferences is also crucial to understand.
In the academic literature risk preferences can take many forms, including absolute risk aversion and relative risk aversion. Absolute risk aversion means that we have an aversion to losing a certain dollar amount. Relative risk aversion means that we have an aversion to losing a certain percentage of our portfolio. So as a portfolio grows in size, all else equal, an individual whose risk preferences take the form of absolute risk aversion would ideally roll into a less aggressive portfolio while an individual whose risk preferences take the form of relative risk aversion would not change their investment strategy.
Many academics and industry practitioners claim that relative risk aversion (that is, percentage-based) is intuitively more appropriate and there is some support for this from research experiments. However there are also experiments and case studies that reveal that people express absolute risk aversion preferences. Bringing this back to industry, a relative risk aversion assumption is evident in the design of many basic default super funds (constant risk exposure regardless of account balance), whereas lifecycle funds have some sympathy to the concepts of absolute risk aversion (the risk of the fund decreases with age which is broadly related to account balance).
The theory of measuring risk aversion has been a highly detailed area of research for many years. Examples include lottery style tests and qualitative questionnaires. The better quality questionnaire approaches have been calibrated alongside lottery style tests for consistency. A lottery style test offers an individual the choice of two lotteries with varying payouts and risks. By running a number of permutations across the same individual, an estimate of risk aversion can be derived.
Research findings on risk aversion
The academic literature has revealed some novel findings on risk aversion. For instance:
- Men are generally less risk averse (more risk seeking) than women
- We tend to become more risk averse as we age
- Tall people are on average less risk averse than short people
- Financial risk aversion has some consistency with our overall ‘life’ risk aversion
- An individual’s self-assessment of their own risk aversion is more accurate than the informed opinion of their financial planner.
In theory, particularly in academia, risk aversion is often assumed to be exactly known and is used as a direct input into determining how an individual should optimally invest their savings. This is done to illustrate an important research question, rather than in ignorance of the uncertainties in measuring risk aversion.
Industry uses inadequate risk tolerance tests
In industry, attempts are made to measure risk aversion. These generally take the form of a risk tolerance questionnaire. Answers are aggregated and mapped into a recommended portfolio allocation. Many of the approaches I have seen are too simplistic. In particular no attempt is taken to measure the form of risk aversion. Many financial planners account for this by using the risk aversion estimate alongside a consideration of an individual’s capacity to take financial risk (the status of their personal balance sheet and income profile) and their financial objectives, to determine an appropriate investment strategy.
The industry seems to make some simplifications which could have a great cost to individuals. Specifically, many portfolios appear to be constructed using a mean variance framework, whereby one designs a portfolio which maximises expected return for a determined tolerance to investment volatility. Portfolios designed using this approach are only optimal for an individual whose preferences take the special form of relative risk aversion where the pain of a loss is equal to the pleasure of an equivalent sized gain. Is this really an appropriate assumption? Many behavioural science experiments would suggest not: the pain of a loss is often greater than the pleasure of an equivalent sized gain.
Understanding the risk aversion of individuals is an important issue for all industry participants. The obvious area is financial planning where, as stated previously, my opinion is that risk profiling practices can be improved. The other area is the design of the investment strategy for a super fund where effectively there are embedded assumptions built into the design of default funds. For both groups there is a lot of powerful science about how risk aversion can be measured and how it should be translated into portfolio construction that is largely untouched by industry.
David Bell is Chief Investment Officer at AUSCOAL Super. He is working towards a PhD at University of New South Wales.