The Changing Concept of Financial Risk

The recent rapid accumulation of anomalous empirical research results has made clear that the classical definition of financial risk based on asset classes only is ready for a epistemological change. Currently, the definition of financial risk suffers from three major deficiencies: (1) financial risk is insufficiently measured by the conventional second - order moments; (2) financial risk is assumed to be stable and all distribution moments are assumed to be time-invariant; and (3) pricing observations are assumed to exhibit only serial dependencies, which can be simply removed by simple inverse transformations, like the geometric Brownian motion, Markov, ARIMA, or (G)ARCH models. But (1) higher - order moments are acknowledged by experienced traders to be influential in asset and derivative valuations; (2) distributions of returns are observed to be nonstationary; and (3) difficult to observe long term dependencies have surprised many risk managers. Based on accumulated empirical evidence, a new funtional definition of financial risk that takes account of asset classes, time and time horizons is required to fully capture the concept as required in the empirical financial markets.