In 1985, Mehra & Prescott introduced the so-called equity premium puzzle, which captured that the historic equity premium in the US for the period 1889-1978, could by no means be explained with a traditional asset pricing model, based on expected utility theory. The authors concluded that the discrepancy had to be explained, by investors being much more risk averse than had previously been assumed, or by concluding that the traditional model did not capture actual investor behavior. This thesis investigates an alternative to the standard investor preferences assumed in the model tested by Mehra & Prescott (1985), namely myopic loss aversion as proposed by Benartzi & Thaler (1995). The model is based on prospect theory of Tversky & Kahneman (1979; 1992), as well as established concepts from the field of behavioral finance. It is descriptive in nature, as it is based on experimental studies of human decision making under risk, rather than relying on assumptions of a purely rational individual. First we expand the study of the equity premium puzzle to the small open economy of Sweden, using the original method of Mehra & Prescott (1985), and the statistical reformulation of the puzzle by Kocherlakota (1996). We start by investigating the period 1919-2010, and reach the surprising conclusion that the equity premium puzzle is not present in Sweden for that period. Based on the previous findings of the puzzle, we re-evaluate our period of analysis, and find that the years before 1925 are highly atypical. We thus eliminate the years from our dataset, and for the revised period from 1925-2010 we find an equity premium puzzle similar to the one found in the US. The result is found to be robust against variations in data input, as well as further changes of the sample period. Next, we investigate whether myopic loss aversion is able to serve as a potential solution to the equity premium puzzle in Sweden. We find that the puzzle can be explained by assuming that investors are loss averse, meaning that a loss hurts them twice as much as a gain pleases them, and that they are myopic in their investment strategy, meaning that they on average evaluate their portfolios too often. Our conclusion is supported by observations of optimal asset allocation, and we find that the implied equity premium would fall, if investors were to evaluate the return on their equity investments, in a more aggregated manner. Finally, we test the sensitivity of our method, against different changes in input variables, and over different sample periods, and again we find that our conclusion is robust.
|Educations||MSc in Finance and Strategic Management, (Graduate Programme) Final Thesis|
|Number of pages||136|