Thinking like an economist 22: Risky decision making
Most economic decisions involves risk, in the sense that the decision maker doesn't know what the exact consequences of different choices will be. This article discusses risk aversion, the importance of expectations, and the value of taking a flexible approach to decisions as risks are resolved over time.
Last year I invested in some real estate in Perth. I'd been considering doing this for a while. The previous year I had suspected that the real estate market in Perth had peaked and was about to fall, so I held off, waiting for a bargain. Unfortunately, prices boomed. In buying last year, I was worried that my expectation for the previous year would come true a year later. I decided, on balance, to buy anyway – clearly, a risky decision. Happily, prices boomed again, this time to my benefit.
Most decisions involve risk, but in such big decisions, the importance of risk really stands out. For the purpose of this discussion, risk means that you don't know what the actual outcome from a decision will be, but you can assign probabilities to a number of potential outcomes. (Another time I will distinguish between risk and uncertainty. This initial definition is relevant to both.) In the real estate example, the possible outcomes are different rates of increase or decrease in house prices after the purchase. Variation in your earning power, and therefore your capacity to service an investment load, might be another source of risk.
As described in Bernstein's (1996) wonderfully interesting and readable history covering 'the remarkable story of risk', the idea of using probabilities to analyse risky decisions is surprisingly recent, dating from the 17th century.
Today, the body of literature on risk is vast and very diverse, but here are a few essential features that apply to pretty much all of it.
• Risk is essentially about decision making when the outcomes are probabilities rather than certainties.
• Risk is intertwined with time. If you are dealing with risk, the outcomes you are considering are in the future (or at least your knowledge about them is).
• Risk involves both up-side and down-side. In general conversation, people often use the word 'risk' to mean the chance of a bad outcome. However, in a thorough decision process you have to weigh up both bad and good possible outcomes.
• A decision option can be 'more risky' in the sense that it has a bigger range of possible outcomes. Other things being equal, most people prefer 'less risky' options, but a minority like to take risks.
Someone who prefers less risky options is said to be 'risk-averse' (not 'risk-adverse'!). For a risk-averse decision maker, there is a trade-off between what she expects to gain through a decision (a weighted average of the possible outcomes) and the riskiness of the decision (which economists often represent by the standard deviation of the possible outcomes). She would be prepared to take a more risky course of action, provided that the expected outcome is sufficiently higher. Conversely, she would be prepared to pay something (a 'risk premium') in order to avoid riskiness.
Many books and academic articles about risk make a big fuss about risk aversion, as if it was the most important issue related to risk. Actually, its effects on decision making are fairly modest compared to some other things, especially compared to the decision maker's expectations about the likely outcomes. People vary in both their risk attitudes (i.e. their degree of risk aversion or risk preference) and their expectations about the future. Variation in expectations is by far the more important factor in explaining differences between the decisions of different people.
The other thing that is more important than risk aversion is the ability to respond to risk as it unfolds. Often you can soften the blow following a bad outcome, or enhance the benefits following a good outcome, by making sensible refinements to your original decision after you learn more about its consequences. For example, a business might gear up to meet expected higher demand for Australian flags around the time of a big international sporting event. If the relevant Australian sporting team is doing worse than expected in the lead up to the event, the business might be wise to scale down their production of flags, and the converse applies as well. For most decision makers, being responsive and flexible in the face of different circumstances like this matters much more than accounting accurately for their risk aversion (Pannell et al., 2000). By 'matters more' I mean that it makes a bigger difference to their welfare.
David Pannell, The University of Western Australia, David.Pannell@uwa.edu.au
Bernstein, P.L. (1996). Against the Gods: The Remarkable Story of Risk, Wiley, New York.
Pannell, D.J., Malcolm, L.R. and Kingwell, R.S. (2000). Are we risking too much? Perspectives on risk in farm modelling. Agricultural Economics 23(1): 69-78. Full paper (65 K)
p.s. 3 April 2006
I enjoyed your discussion of risky decision making as we've just been studying it in my Microeconomics class.
My lecturer gave a really good example by way of trying to explain risk from the perspective of an economist that might be useful: Consider two events - falling out of an aeroplane with a parachute and falling out of an aeroplane without a parachute. Which is more risky? Falling out of the aeroplane with a parachute is clearly more risky as the outcome is less certain. The parachute may or may not open; whereas you know, with certainty, the outcome of falling out of an aeroplane without a parachute.
It was a good example because even at the Master's level, some of my fellow students thought falling out of the plane with a parachute was less risky.
DP: This example neatly highlights that the more risky alternative is not
necessarily the worse option.
Pannell Discussions are brief pieces on issues and ideas in economics, science, the environment, natural resource management, politics, agriculture and whatever else.
|69. Climate change 6 Feb 2006||71. Agonising decisions 10 Apr 2006|
URL for this page: http://dpannell.fnas.uwa.edu.au/pd/pd0070.htm