Modeling fund and portfolio risk: A bi-modal approach to analyzing risk in turbulent markets

Following the financial crisis of 2008, it has been argued that Value at Risk (VaR), and risk analysis in general, failed to alert risk managers of the turbulence on the horizon. This is a misguided view that should not have come as a surprise because many widely circulated academic papers and discussions suggested, well before the crisis, that simple VaR results could easily be misinterpreted if the circumstances for its proper use are not fully understood. This paper addresses some ways in which VaR concepts may be applied more effectively. Non-standard Monte Carlo simulations are utilized. Whereas standard mean-variance defined methodologies using Monte Carlo analysis may not capture how “fat” a lower tail may actually be, a bi-modal switching structure between assumed normal periods and possible turbulent economic periods may help resolve the problem. Lower boundaries (worst case paths) of the different (normal versus bi-modal) processes are mapped to illustrate implied riskiness of portfolios if turbulence occurs. The analysis implies that no mechanical risk analysis is sufficiently divorced from a judgment call about possible market disruptions; however, a bi-modal approach allows quantification of the said judgment in conjunction with empirical observations from history.