Abstract
Strategic asset allocation is discussed in a discrete-time economy, where the rates of return from asset classes are explained in terms of some observable and hidden factors. We extend the existing models by incorporating long-term memory in the rates of return and observable economic factors, which have been documented in the empirical literature. Hidden factors are described by a discrete-time, finite-state, hidden Markov chain noisily observed in a fractional Gaussian process. The strategic asset allocation problem is discussed in a mean-variance utility framework. Filtering and parameter estimation are also considered in the hybrid model.
Original language | English |
---|---|
Pages (from-to) | 609-626 |
Number of pages | 18 |
Journal | Methodology and Computing in Applied Probability |
Volume | 16 |
Issue number | 3 |
DOIs | |
Publication status | Published - Sept 2014 |
Keywords
- Fractional Gaussian VAR process
- Hidden Markov models
- Long memory
- Mean-variance utility
- Strategic asset allocation