Asset Allocation
What you must know
Capital market expectations (build-up, DCF, risk-premium and econometric approaches), the economic balance sheet (assets, liabilities, net worth including human capital and pension claims), strategic versus tactical allocation, asset-only versus liability-relative optimisation, the surplus efficient frontier and surplus Sharpe ratio, risk budgeting, Monte Carlo simulation, rebalancing policy, and goals-based investing.
Traps
- In liability-relative work the risk-free asset is the asset that hedges the liability, not T-bills or cash. The tangency portfolio and the resulting mix therefore differ from the asset-only case.
- Surplus optimisation maximises surplus return per unit of surplus risk, not asset return. A large deficit can make a defensive allocation optimal, and the best hedging asset may have a poor standalone Sharpe ratio.
- Monte Carlo: adding simulation paths reduces sampling error only. It does not change the expected return or make the forecast more accurate.
- Asset-only frontiers ignore the liability entirely, so an allocation that is efficient in asset space can be badly inefficient in surplus space.
- Tactical tilts must be justified by a forecasted deviation from strategic weights, not by recent performance.
Worked example: Roy's safety-first ratio
Threshold return R_L = 3%. Portfolio A: E(R) = 9%, sigma = 12%, so SFR = (9 - 3)/12 = 0.50. Portfolio B: E(R) = 11%, sigma = 18%, so SFR = (11 - 3)/18 = 0.44. Choose A, because it maximises the probability of clearing the threshold. Rule of thumb: SFR = (E(R) - R_L)/sigma; the higher the ratio, the lower the probability of a shortfall.