PIMS/Shell Lunchbox Lecture: Long Term Asset Allocation for the Patient Investor

  • Date: 11/04/2014
Peter Forsyth, Cheriton School of Computer Science, University of Waterloo

After graduating in 1979, Peter Forsyth was a Senior Simulation Scientist at the Computer Modelling Group (CMG) in Calgary, where he developed petroleum reservoir simulation software. After leaving CMG, Peter was the founding President of software startup Dynamic Reservoir Systems (DRS), also in Calgary.

In 1987, Peter joined the University of Waterloo, where he is now a Professor in the Cheriton School of Computer Science. Peter's current research focuses on Computational Finance, with particular focus on numerical methods for Hamilton Jacobi Bellman partial differential equations. He is a member of the Editorial Board of Applied Mathematical Finance and the Journal of Computational Finance.  During 2008-2013, he was the Editor-in-Chief of JCF.

Peter is currently a director of Aquanty, a software startup specializing in integrated modelling of three dimensional surface/subsurface water flows. Aquanty specializes in simulating the impact of industrial activity and climate change on water resources.


Calgary Place Tower (Shell)


Long Term Asset Allocation for the Patient Investor

Many studies have shown that the most important factor in investment success is asset allocation, as opposed to stock picking.  At its most basic level, this amounts to the choice of how much to allocate to bonds and stocks.

In this talk, I will formulate the portfolio allocation problem as the solution to a problem in optimal stochastic control.

I will discuss semi-self-financing mean-variance (MV) dynamic asset allocation strategies which are superior to self-financing MV portfolio strategies. Our strategies are built upon a Hamilton-Jacobi-Bellman (HJB) equation approach for the solution of the portfolio allocation problem.

MV strategies are often criticized for penalizing the upside as well as the downside. However, under our strategies, the MV portfolio optimization problem can be shown to be equivalent to maximizing the expectation of a well-behaved utility function of the portfolio wealth.

We show that, for long term investors, the use of dynamic MV strategies can achieve the same expected value with a much smaller standard deviation and probability of shortfall compared to a constant mix strategy.
Other Information: 

Location: Calgary Place Tower 1 (330 5th Avenue SW), Room 1106


Time: 12:00-1:00 pm





PIMS is grateful for the support of Shell Canada Limited, Alberta Enterprise and Advanced Education, and the University of Calgary for their support of this series of lectures.