Portfolio & Risk Management: Asset Allocation and Risk Budgeting Optimization

In standard static Mean-Variance approach portfolio is presented by one allocation vector optimized in terms of expected returns & variance-covariance (VcV) matrix. Such one-dimensional approach is not suitable for Fixed Income: i) portfolio cannot be described by allocation vector only, and ii) returns VcV matrix is period dependent even if yield process is stationary. Multi-dimensional optimization problem is formulated in terms of risk-sensitivity matrix (RSM), allocation & yield vectors. Yield vector reflects term-structure, security & asset selection. Returns VcV matrix is expressed in terms of yield VcV matrix and RSM, which is specified by risk budgeting & duration management. So, optimal allocation is conditional on RSM, i.e. on risk & portfolio management strategies. Instantly efficient portfolio derived from static one-dimensional optimization will not be efficient after infinitesimal time transformation since RSM is not time-invariant. Multi-dimensional approach provides optimal allocation and duration management strategies for any risk budgeting constraint. An optimal RSM allows for a minimum Tracking Error portfolio to be more efficient than benchmark"s global efficient frontier. Infinite amount of optimal portfolio alteration decisions, subject of RSM choice, is benchmark dependent. General conclusion: integrated portfolio & risk management process is an important issue for asset management.