Value at risk (VaR) is an approach used in risk management to measure downs
ide risk. Not all VaRs, however, are created equal. Defining and accurately
measuring market risk is a considerable task. VaR estimates depend on a nu
mber of inputs, including assumptions, data parameters, and methodology Acc
ordingly, comprehending the optimal use of these various inputs and how the
y might impact the VaR forecast is necessary to avoid biasing portfolio ris
k estimates. The authors examine three equity portfolios of varying degrees
of diversification, using twenty common VaR models developed through four
VaR techniques to clearly demonstrate the ramifications of using inappropri
ate models. They find that the particular characteristics of a portfolio mu
st guide and determine which VaR model may be applied in order to extract a
ccurate VaR estimates. Using the wrong VaR model will, bias the behavior of
portfolio managers and cause them to be exposed to much more risk than the
y desire.