© 2019 by Gittins & Oreskovich Consulting

The GO Risk Measures

Gittins and Oreskovich created the Gittins-Oreskovich (GO) Risk Measures to address the gap in currently available tools for quantifying risk in pharmaceutical research.

In journal articles, as well as Dr. Oreskovich’s doctoral thesis, the GO measures are proffered as alternative risk measures to canonically used tools like Value-at-Risk (VaR), Variance of Net Present Value (VarNPV), and utility theory.

 

In Gittins and Oreskovich's opinion, none of the currently available approaches for risk quantification in drug development offers a wholly satisfactory procedure for quantifying risk.  A scale of valuation is required which improves on simply taking the expected net value with discounting at the risk free interest rate.  Taking the weighted average cost of capital as the discount rate takes account of the influence of uncertainty on interest rates.  However we still need some way of adding the influence of risk to a calculation of expected net value.

 

Using utilities rather than monetary values in principle does this job.  The downside is that there remains the non-trivial task of assigning a utility to a cash flow, and even if this could be done, the utilities assigned by different people would be different.

 

An obvious objective quantity to quote is the variance of NPV, to accompany its expected value, and this has often, and usefully, been done.  Often, however, what really matters in a risky situation is the probability of a very bad outcome.  This leads us to consider VaR, redefined so as to refer to a project of unknown length, and thus becoming simply a quantile of the NPV distribution.

 

In fact we can do better than VaR in three ways.  The first improvement is to consider the maximum indebtedness reached by the aggregate cash flow, rather than simply NPV, which is the aggregate cash flow at the end of the project.  We call this maximum the cash flow exposure, CFE.  It is the first of our proposed new measures of risk.  The second improvement is to consider the expected value ECFE (or even the entire distribution) rather than simply a quantile.  The third improvement is to add to the aggregate cash flow at each stage the expected value of future cash flows, and to consider the maximum indebtedness on this basis, taking into account future prospects.  This defines PCFE, the prospective cash flow exposure, and EPCFE, its expectation.  PCFE is the second of our proposed new measures of risk.

 

Both have probability distributions over possible future scenarios.  The best single-valued measure of risk is in each case the expectation of maximum financial exposure, ECFE and EPCFE.