Biases and Heuristics in Portfolio Management – Determinants for non-optimal Portfolio Diversification
von Ibrahim Filiz
Datum der mündl. Prüfung:2019-01-23
Erschienen:2019-02-22
Betreuer:Prof. Dr. Kilian Bizer
Gutachter:Prof. Dr. Markus Spiwoks
Gutachter:Prof. Dr. Waldemar Toporowski
Dateien
Name:Dissertation von Ibrahim Filiz_18022019.pdf
Size:1.94Mb
Format:PDF
Zusammenfassung
Englisch
Markowitz (1952) shows that it is useful for risk-averse investors to split capital among different investment instruments. Practice shows, however, that investors often have non-optimal diversified portfolios. There are many possible reasons for securities portfolios not being optimally diversified. In the last two decades, experimental economic research in particular has identified some heuristics and cognitive distortions (such as the 1/N heuristic, home bias, mental accounting or the illusion of control) which can contribute towards non-optimal portfolio diversification. Additional research efforts in this field can be found in this dissertation. The first contribution examines the influence of herding, status-quo bias and gambler’s fallacy on diversification behavior. However, neither herding nor status-quo biases significantly contribute to non-optimal portfolio choices. Gambler’s fallacy, however, plays an important role in these decisions. Many participants are zealous to notice patterns in a history of random events and to infer from these pattern the sequence of future events. The second study addresses the question of whether positive and negative emotions have an influence on diversification behavior, and it reveals that only a small part of subjects take rational decisions and always choose the optimal portfolio. In addition, the study shows that the mood of subjects has an influence on their portfolio decisions and thus also on their exposure to risk. Experimental research on diversification behavior requires a clear differentiation between risk-averse, risk-neutral and risk-loving subjects, because decisions which can be absolutely meaningful for a risk-loving subject are completely inconceivable for a risk-averse subject and vice versa. The procedures previously used to determine risk preference exhibit a number of weaknesses. Therefore in the third contribution a new procedure for the determination of risk preference is proposed. Future interest rate trends are of key significance in portfolio management, because bond and share prices are significantly influenced by interest rates. Therefore, the fourth contribution takes an in-depth look at the forecasting competence of analysts in the Asia-Pacific region. Overall it can be stated that - at least in some countries and for some forecast horizons - forecasts of future interest rate trends in the Asia-Pacific region are more successful than those made in other parts of the world. The influence of the overconfidence bias on portfolio decisions by economic subjects can be tremendous. There are hardly any studies on whether positive or negative emotions have an influence on self-assessment. Whether emotions can impair possible learning effects in relation to self-assessment has not been investigated at all up to now, but the fifth study deals with this issue. Overall, the results of the study show that emotions have no significant impact on self-assessment. In respect of the occurrence of learning effects, however, there is a difference between the moods.
Keywords: behavioral finance; biases and heuristics; experimental economics; investor rationality; portfolio choice; non-optimal diversification; risk exposure; risk preference; herding; status-quo bias; gambler’s fallacy; emotions; overconfidence bias; learning effect; forecast accuracy