With Nalan Baştürk, Stefano Grassi, Lennart Hoogerheide and Herman K. van Dijk.
Published in Journal of Econometrics, 2018.
A novel extended time-varying density combination scheme for dynamic asset-allocation based on mixing of alternative volatility models and alternative portfolio strategies. For computations: a new dynamic filter based on mixtures of Student’s t distributions (M-Filter).
PDF: Forecast Density Combinations of Dynamic Models and Data Driven Portfolio Strategies
Abstract
A dynamic asset-allocation model is specified in probabilistic terms as a combination of return distributions resulting from multiple pairs of dynamic models and portfolio strategies based on momentum patterns in US industry returns. The nonlinear state space representation of the model allows efficient and robust simulation-based Bayesian inference using a novel non-linear filter. Combination weights can be cross-correlated and correlated over time using feedback mechanisms. Diagnostic analysis gives insight into model and strategy misspecification. Empirical results show that a smaller flexible model-strategy combination performs better in terms of expected return and risk than a larger basic model-strategy combination. Dynamic patterns in combination weights and diagnostic learning provide useful signals for improved modelling and policy, in particular, from a risk-management perspective.