| An Evaluation of Asset Pricing Models using the Minimum Distance Test |
|
Bong Jun Kim, Jae Ho Cho |
| 최소거리검정을 이용한 자산가격결정모형의 평가 |
|
김봉준, 조재호 |
|
|
|
|
| Abstract |
|
Hansen and Jagannathan (1997) suggest that the minimum distance between the arbitrage-free stochastic discount factor set and the implied discount factor set of reference assets can be interpreted as the maximum pricing error of a theoretical model. Using this metric, this paper attempts to compare the performances of several asset pricing models in explaining returns of reference portfolios constructed with KOSPI stocks. Specifically we test the standard CAPM of Sharpe (1964) et al., three-factor model of Fama and French (1993), the consumption CAPM of Lucas (1978), the production CAPM of Cochrane (1996), the conditional CAPM of Jagannathan and Wang (1996), the fivefactor model of Campbell (1996), and their conditional versions. For instrumental variables in conditional models, we use such variables as cyclical components of industrial production index, the amount of export and the dividend yield of the market portfolio. For each of unconditional and conditional models, we carry out unconditional and conditional tests. In the conditional test, the reference portfolios are rebalanced as the information set is time-varying. The results of the paper are summarized as follows. All of unconditional models are rejected at 1% significance level while conditional versions of the three-factor model of Fama and French and the five-factor model of Campbell are not at 10% significance level. Annualized maximum pricing errors of these models are smaller than those of unconditional models by 10~14%. In particular, as the conditional model of Campbell survives even in the conditional test at 10% significance level, it performs best in the light of the least maximum pricing error. |
| Key Words:
시간가변적 모형,오일러등식,조건부 검정,최소거리 검정,확률할인요소,Generalized Method of Moment Test,Minimum Distance Test,Stochastic Discount Factor,Unconditional and Conditional Models,Unconditional and Conditional Tests |
|
PDF Links
Full text via DOI
Download Citation
