by Aswath Damodaran.
In valuation practice, it is common to use point estimates for input variables to arrive at a point estimate of value. This practice has its roots in the twentieth century, where data was difficult or expensive to access, and practitioners had weak tools. As data becomes more accessible and models more powerful, it is worth examining whether we can make our valuations of projects and businesses richer and more informative, by drawing on established statistical tools and we introduce three in this paper.
The first is scenario analysis, where you lay out a broad range of scenarios, and estimate a value under each one. If the scenarios represent the complete spectrum of outcomes, you could even derive an expected value. The second is decision trees, suited for projects or businesses that have to deal with sequential risk, where you have to make it through one step to get to the next. The third and most comprehensive approach is simulations, where you replace point estimates of one or more input variables with probability distributions, and estimate a value distribution for an asset/business, rather than a single value. We look at the why you may pick one approach over the other and pitfalls to avoid along the way.
Download article here.