Andrew Smith (University College Dublin)

will speak on

Sampling-based Inference and Interest Rates

Time: 12:00PM
Date: Mon 14th October 2019
Location: Seminar Room SCN 1.25 [map]

Abstract: Methodology disagreements are a recurrent feature in financial modelling. Nowhere is this more striking than in models of interest rates. The orthodox economic view is that interest rates are stationary but cyclical and should generally be positive. However, the history of the past fifty years shows a more or less steady decrease in interest rates to current levels around zero, with the implication under extrapolation that rates will become more and more negative in future.
Computing constraints no longer force us to choose between conflicting models. Using modern techniques of sampling-based inference, multiple models can all be true at once, with weights reflecting how plausible they are. So is this the magic formula that allows protagonists to compromise and reach consensus about how to model interest rates? Or does another layer of complexity just give us more places to conceal judgements and assumptions?

(This talk is part of the Working Group on Statistical Learning series.)

PDF notice

Return to all seminars