Agent-based risk management : a regulatory approach to financial markets

This paper provides market risk calculation for an equity-based trading portfolio. Instead of relying on the purely stochastic internal model method, which banks currently apply in line with the Basel regulatory requirements, we propose to include also alternative price mechanisms from the financial literature into the regulatory framework. For this purpose a modified version of the model by Lux and Marchesi (2000) is developed, capturing the realistic feature that parts of the investors do not follow the assumption of no arbitrage, but are motivated by rules of thumb and market psychology instead. Although both the standard stochastic as well as the behavioral model are restricted to 250 trading days, the latter is able to capitalize possible turbulences on financial markets and likewise the well-known phenomenon of excess volatility - even if the last 250 days reflect a calm market for which the efficient-market hypothesis could hold. Thus it is argued that a value-at-risk-based maximum approach in the regulatory framework would create better capital requirements with respect to their level and counter-cyclicality. This in turn could reduce the extent to which (irrational) bubbles arise since market participants would have to anticipate comprehensively the costs of such bubbles bursting. Furthermore a key ratio is deduced from the agent-based construction to lower the influence of speculative derivatives.