This thesis studies the Black-Litterman approach implemented on a set of domestic equity indexes with an econometric model for the specification of the views and their confidence levels. We selected global and local macroeconomic factors as external regressors in the time series analysis to provide a better forecast of the conditional mean and variance. We find that the views and omegas derived with this technique represents an effective input, especially in periods of high volatility. The Black-Litterman portfolio built with these specifications is able to consistently outperform the market-weighted portfolio and a comparable benchmark index. The sensitivity of the model to its parameters has also been investigated. The stability of the asset allocations, the relatively low portfolio variance, the opportunity to update the views monthly and its rebalancing frequency lead to the conclusion that this approach represents a valuable strategy for active portfolio management.
|Uddannelser||Cand.merc. Erhvervsøkonomi, (Kandidatuddannelse) Afsluttende afhandling|