In this thesis we develop an asset allocation model. The framework for this model involves using past returns to produce forecasts of expected returns for 16 asset classes, representing the investment universe for Danish pension funds. The turmoil of 2007/2008 is just one among many financial crises over the past 30 years. These events with low probability but high impact can, and should, challenge conventional ideas about portfolio construction. We will never be able to predict when such events will occur and how bad they will be. However, we do know, that we empirically observe such events with much greater frequency than current models allow for. Our concern is not the risk of assets going down 5% but the sort of risk that can wipe you out. In financial literature it is called tail risk. Modern portfolio theory teaches us that, for a rational investor, the goal is to hold a portfolio that achieves the highest risk-adjusted performance. Although our aim to maximize risk-return efficiency is fully consistent with financial theory, successful implementation of the theory depends not only on its conceptual grounds, but also on the reliability of the assumptions to the model. In part II of the thesis we investigate and find three specific weaknesses in conventional risk assessment, which all leads to underestimation of portfolio risk. These findings support the need for a new asset allocation framework, because investors will gain a practical benefit: the potential to improve portfolio efficiency and resilience, in light of a clearer understanding of portfolio risk. The model we have built is non-parametric, meaning that it does not rely on data belonging to any particular distribution, and that it does not assume that the structure of a model is fixed. Instead the model provides a method of simulating future returns scenario for each asset in the portfolio. This is done by using the empirical returns to forecast expected returns. This model should be a tool, to be used by all kinds of investors with their personal investment universe, in the aim of understanding the risk taken. In this thesis we have used the Danish pension funds to construct the investmentuniverse. Other empirical asset allocations framework has been made before ours. What is special in this thesis is that we use the model for, not only finding the optimal portfolio, consisting stocks and bonds, but we also add a second group of investments. In this thesis we call them alternative investments. This group of investments is especially hard to model, due to the lack of good indexes representing their true characteristics, and thereby the risk involved in these kind of investment. With our model we find that adding or increasing the exposure to some of the alternative investments can benefit the overall portfolio efficiency for Danish pension funds. We found that, especially exposure to commodities should be very attractive for the Danish pension funds given their current allocation.
|Uddannelser||Cand.merc.fir Finansiering og Regnskab, (Kandidatuddannelse) Afsluttende afhandling|