TABLE OF CONTENTS Stephen Wright and Stephen Satchell
(The views and opinions expressed in this article are those of the authors and are not necessarily those of UBS. UBS accepts no liability over the content of the article. It is published solely for informational purposes and is not to be construed as a solicitation or an offer to buy or sell any securities or related financial instruments.)
Realistic forecasts merge information from many sources. This information may be subjective or objective; absolute or relative; strategic or tactical; incomplete or contradictory; macroeconomic, factor, style or stock specific. No matter how complex, we need a coherent framework to allow the insight from these multiple inputs to be merged in a way that is both statistically rigorous and intuitively reasonable.
In reality, forecasts of return are probability distributions that can become arbitrarily complex (multivariate and/or multimodal) if they are to fully reflect all the above insight. Unfortunately there is no single correct way of constructing such a distribution; the most effective approach depends on the characteristics of the system being forecast, as well as the information quality and skill available to the forecaster.
This chapter builds on our earlier paper A robust cross-sectional approach to equity forecast construction (Satchell and Wright, 2005), which introduced the concept of a mixture of normals approach in the context of rank scorecards. Here we extend this approach to show how insight from equilibrium models, relative value views, expected factor return, and expected asset return as well as stochastic scenario...
