Not logged in.
Quick Search - Contribution
Contribution Details
| Type | Bachelor's Thesis |
| Scope | Discipline-based scholarship |
| Title | Investors’ Diversification Preferences in the Gain and Loss Domain |
| Organization Unit | |
| Authors |
|
| Supervisors |
|
| Language |
|
| Institution | University of Zurich |
| Faculty | Faculty of Business, Economics and Informatics |
| Number of Pages | 49 |
| Date | 2021 |
| Zusammenfassung | This paper examines the investment behaviour of lay investors and focuses on decision preferences in different return scenarios. Through use of a survey, how the participants behave in both positive and negative return environments is examined. Participants were provided with graphs of three stocks (A, B and C) with returns of the last five years. Their task was to choose either asset B or C to add to their portfolio, which already consisted of stock A. The choice (B or C) consisted of one positively correlated asset and one negatively correlated asset. The focus of this work is on the cognitive factors that impact decision preferences of participants, depending on whether they are in a gain, loss or neutral domain. For the study, various effects from psychological science approaches that could lead to distortions in the participants’ decision preferences were applied. In particular, aspects of Kahnemann and Tversky’s (1979) work regarding framing were applied, whereby the influence of representation differences (which are responsible for irrational decision observations) were examined more closely. In investigating these and other bias effects from the behavioural finance sphere, several graphs, which differed in their presentation, were presented to survey participants. That is, identical graphs were presented in a different order and each with disparate information about the average value of the assets to be compared. The survey was supplemented with basic financial questions based on Mitchell and Lusardi’s (2011) "big three" questions to sort the respondents into two groups. Participants who answered all six economic comprehension questions correctly were assigned to the group of highly literate people. On the other hand, participants who answered one or more questions incorrectly were placed in a second group of low literate people. This resulted in two groups, each with participants carrying relatively similar levels of financial literacy. The groups were compared in terms of their decision-making preferences. In a second step, a statistical evaluation using linear regression tested the influence of variable parameters and their level of significance. Evaluation of the results revealed that individuals demonstrated a slight tendency to behave in a more risk-averse manner in a gain domain than in a loss domain. This behaviour was expected according to the prospect theory of Kahnemann and Tversky (1979), due to the risk aversion of individuals observed in a positive gain environment and the willingness to take risks in a loss environment. This behaviour is irrational, and it led to biases, especially in the loss domains. The greatest bias was measured in the two loss domain graphs: PL (portfolio with loss outcomes) and PLS (portfolio with loss shifted outcomes). Here, 45% of the participants surveyed chose the riskier variant of the positively correlated asset in each case. The comparison of the two groups (of high and low financial literacy) demonstrated that participants with certain prior financial knowledge are more consistent regarding their decision preferences than those with low literacy. A linear regression was created, whereby the sum of the number of rational decisions was regressed on the number of correct answers to the comprehension questions. The model produced a P-value of 0.036 with a significance level of alpha = 0.1% and alpha = 5%. The result thus reveals that financial literacy has a significant influence on decision-making performance. In a second step of the statistical evaluation, further parameters (e.g., age, gender and personal risk tolerance) were considered by means of multiple regression, but no significant influence could be proven. |
| Export |
BibTeX
|