Abstract
In the market model the return on an asset is modeled as a linear function of the return on a market index with slope parameter beta. The coefficient beta is often used as a measure of the sensitivity of the asset’s return to the market and to measure the component of the variance of the return that is explained by the market. However, both of these interpretations require the additional assumption that the error term in the market model has mean 0 conditional on the return on the market index, an assumption that is often difficult to verify in practice. In this paper, a nonparametric version of the market model is proposed that does not require such an assumption. This nonparametric model replaces the beta coefficient of the market model with a “beta curve” describing the relationship between the asset’s return and that of the market locally near a given value of the market return. The proposed model is applied to stock returns, as well as to returns on mutual funds. Corresponding tests of the market model are given and it is shown that the nonparametric model often provides an improvement over the standard parametric market model.
Original language | English (US) |
---|---|
Pages (from-to) | 179-199 |
Number of pages | 21 |
Journal | Annals of Finance |
Volume | 12 |
Issue number | 2 |
DOIs | |
State | Published - May 1 2016 |
Funding
This work was supported by National Science Foundation Grant DMS-1308009.
Keywords
- Beta
- Correlation curve
- Market model
- Nonparametric function estimation
ASJC Scopus subject areas
- Finance
- Economics, Econometrics and Finance(all)