About OpenKnowledge@NAU | For NAU Authors

Refining our understanding of beta through quantile regressions: Working paper series--10-07

Atkins, Allen B. and Ng, Pin (2010) Refining our understanding of beta through quantile regressions: Working paper series--10-07. Working Paper. NAU W.A. Franke College of Business.


Download (922kB) | Preview


The Capital Asset Pricing Model (CAPM) has been a key theory since the 1960's. One of its main contributions is to attempt to identify how the risk of a particular stock is related to the risk of the overall stock market using the risk measure, beta. If the relationship between an individual stock's returns and the returns of the market exhibit heteroskedasticity, then the estimates of beta for different quantiles of the relationship can be quite different. The behavioral ideas first proposed by Kahneman and Tversky (1979), which they called prospect theory, postulate that i) people exhibit "loss-aversion" in a gain frame, and, ii) people exhibit "risk-seeking" in a loss frame. If this is true people could prefer lower beta stocks after they have experienced a gain and higher beta stocks after they have experienced a loss. Stocks that exhibit converging heteroskedasticty (9.8% of our sample) should be preferred by investors and stocks that exhibit diverging heteroskedasticity (19.8% of our sample) should not be preferred. Investors may be able to benefit by choosing portfolios that are more closely aligned with their preferences.

Item Type: Monograph (Working Paper)
Publisher’s Statement: Copyright, where appropriate, is held by the author.
ID number or DOI: 10-07
Keywords: Working paper, Beta, risk preferences, portfolio management, quantile regression, heteroskedasticity
Subjects: H Social Sciences > HG Finance
NAU Depositing Author Academic Status: Faculty/Staff
Department/Unit: The W.A. Franke College of Business
Date Deposited: 17 Oct 2015 20:25
URI: http://openknowledge.nau.edu/id/eprint/1496

Actions (login required)

IR Staff Record View IR Staff Record View


Downloads per month over past year