Providing top-level academic expertise
Examples of the problems on which JB Consultants have worked include:
-
Absolute return portfolios:
Where the objective is to produce a stock (equity) portfolio that (as best as can be achieved) generates returns independent of the underlying market. In this area a factor based regression model has been developed.
-
Consistent portfolios:
Where the objective is to choose a portfolio that produces returns out-of-sample that are consistent with the returns seen in-sample, (when the portfolio was chosen). It is well know that often "optimised" portfolios fail to live up to their in-sample promise when held out-of-sample. Here density forecasting can be used to highlight regions of portfolio risk-return space that lead to consistent portfolios.
-
Enhanced indexation:
Where a portfolio that out-performs an index must be chosen. Here the degree of out-performance may be specified, or may be left unspecified. For such problems a number of nonlinear models, with varying objective functions, that are solved using an evolutionary algorithm have been developed.
-
Index tracking:
Where the objective is to choose a portfolio that tracks a specified market index. Here, rather than engage in full replication, a subset of assets must be chosen, and the decision problem relates to the subset of assets to hold and the investment in the chosen assets. Such problems often involve rebalancing (incurring transaction cost) over time. In these areas a number of different models have been developed.
-
Markowitz mean-variance portfolio optimisation:
Where the objective is to develop a tradeoff curve of risk (as measured by volatility, variance in portfolio return), against mean portfolio return. Such problems often need the basic Markowitz model enhanced by special purpose constraints, for example cardinality constraints restricting the number of assets in the portfolio chosen, and these present algorithmic challenges in terms of solving large scale problems. In this area a number of different algorithms have been developed.
-
Momentum portfolios:
These seek to exploit asset return momentum (e.g. in equity markets). In this area a model that uses a Sortino ratio objective to produce a portfolio that itself exhibits momentum (rather than a more simplistic portfolio in which each individual asset exhibits momentum, and the assets are typically equally weighted) has been developed.