Assessing Risk in Graphically Presented Financial Series

摘要

It has been argued that traders use their natural sensitivity to the fractal properties of price graphs to assess risk and that they are better able to do this when given price change as well as price level information. This approach implies that risk assessments should be higher when the Hurst exponents are lower, that this relationship should be stronger in the presence of price change information and that risk assessment should depend more strongly on the Hurst exponent than on the standard deviation of the series. Participants in Experiment 1 decided which of two assets was riskier by inspecting graphs of their price series. Graphs with lower Hurst exponents were selected only by those who were less emotionally stable and hence more sensitive to risk. However, when both price series and price change series were presented, the assets with lower Hurst exponents were selected by all participants. In a second experiment, participants were given both price level and price change series for a number of assets and rated the risk of trading in each one. Ratings depended more strongly on Hurst exponents than on other measures of volatility. They also depended on indicators of potential loss. Human risk assessment deviates from the way that risk is measured in modern finance theory: it requires integration of information relevant to both uncertainty and loss aversion, thereby imposing high attentional demands on traders. These demands may impair risk assessment but they can be eased by adding displays of price change information.