The impact of biases on simulation-based risk aggregation: modeling cognitive influences on risk assessment

This paper develops a systematic approach to quantifying the effect of judgmental biases on aggregate risk measures. Starting with the standard risk management process, we derive the areas that require expert judgment as input in order to aggregate risk into risk measures such as Earnings at Risk. We specify three possible gateways for biases and identify several psychological theories to quantify deviations of expert judgments from objective probabilities. The impact of these cognitive biases on the aggregate risk measure is investigated via Monte Carlo simulation experiments. Through experimental design, we can determine the size of both the average and the possible interaction effects of the different biases. The results show that aggregate risk is systematically underestimated if it is based on biased subjective judgment. Moreover, the existence of interaction effects indicates potential problems of simple debiasing strategies. Copyright Springer Verlag 2011