Please use this identifier to cite or link to this item: http://scholarbank.nus.edu.sg/handle/10635/147498
Title: AN EMPIRICAL STUDY OF THE FAMA FRENCH FIVE-FACTOR MODEL AND THE HOU XUE AND ZHANG Q-FACTOR MODEL
Authors: HUANG CAN
Issue Date: 2016
Citation: HUANG CAN (2016). AN EMPIRICAL STUDY OF THE FAMA FRENCH FIVE-FACTOR MODEL AND THE HOU XUE AND ZHANG Q-FACTOR MODEL. ScholarBank@NUS Repository.
Abstract: Recent study by Hou, Xue and Zhang (HXZ) suggests that the HXZ q-factor model performs better empirically than the five-factor model by Fama and French (FF) in capturing anomalies that are not captured by the classic FF three-factor model (HXZ 2015). However, results from tests using the discovered anomalies can be subject to the choice of test assets. As such, these tests may not be the optimal method of comparing the empirical performance of the HXZ q-factor model and that of the FF five-factor model. It remains possible that there exist undiscovered anomalies that the five-factor model explains better. Moreover, another possibility is that there exist assets priced by neither model and using these as test assets is not useful in telling which model performs better. This paper attempts to take a more general approach in comparing the empirical performance of the five-factor model and that of the q-factor model. From statistical tests that are independent of the choice of test assets, more general results are obtained to determine whether one model would better price any asset, including potential undiscovered anomalies. Three types of tests are applied, namely the Gibbons Ross Shanken (GRS) tests, the modified GRS tests with White’s heteroscedasticity-corrected standard errors, and the Generalised Method of Moments (GMM) estimations. All tests show consistent results that the HXZ q-factor model performs better empirically than the FF five-factor model. There is significant evidence, at the 95% confidence level, that the HXZ model explains what the FF model explains, and more. Specifically, the GRS test results show that there is no significant evidence that the FF model provides any additional explanatory power. The Method of White tests give consistent results: with correction for heteroscedasticity, there is strong evidence that the q-factor model can explain something the FF model cannot. This result is further concluded from the more efficient GMM estimation with J-tests. So the conclusion drawn is that there is no significant evidence that there exist assets that are priced by the FF five-factor model but not the HXZ q-factor model. The set of assets that are priced correctly by the HXZ factors is a super set of the set of assets that are priced correctly by the FF factors. In summary, there is significant evidence that the HXZ factors explain more than what the FF factors do, but there is no significant evidence that the FF factors explain more than what the FF factors do.
URI: http://scholarbank.nus.edu.sg/handle/10635/147498
Appears in Collections:Bachelor's Theses (Restricted)

Show full item record
Files in This Item:
File Description SizeFormatAccess SettingsVersion 
Huang Can_A0084641E.pdf693.34 kBAdobe PDF

RESTRICTED

NoneLog In

Page view(s)

3
checked on Sep 27, 2018

Download(s)

1
checked on Sep 27, 2018

Google ScholarTM

Check


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.