Using Statistics and Mathematics to Model the way in which Interest Rates Evolve over Time

Andrew F. Siegel

Department of Statistics and Foster School of Business

University of Washington at Seattle


Mathematical modeling is all about choosing the simplest equations that act like reality. Models of the evolution of interest rates over time are surprisingly deep because simpler models are full of inconsistencies (called “arbitrage opportunities”) that do not reflect market efficiency. Sophisticated traders would exploit these opportunities, systematically changing interest rates, instantly rendering the model invalid. At any given moment there are many interest rates: one for each investment term (which can range from days to years), and a consistent model must prevent arbitrage opportunities within the vast linear space of combinations (portfolios) of investments with different terms. Many aspects of mathematics and statistics are involved in the creation of consistent arbitrage-free models. My research involves a change in focus: Building a linear system for market prices instead of for interest rates, which reduces model complexity while increasing model flexibility. The technical parts of this talk will be described conceptually in an intuitive way, and anyone with an interest in mathematics and its applications is welcome to attend!


Thursday, May 24

Steitz Hall 102