Testing the Q-Factor Model: An Econometric Analysis of Stock Returns in EViews

Overview

An individual econometrics project testing whether Hou, Xue and Zhang's (2015) q-factor model - a four-factor extension of traditional asset pricing models incorporating market, size, investment, and return-on-equity factors - could explain the monthly excess returns of an assigned stock. The project combined EViews-based regression modelling with a full suite of diagnostic testing to assess the robustness of the results, supported by a recorded presentation of the key findings.

What I did

• •Imported and structured monthly stock price data and the four q-factor model variables into an EViews workfile
• •Calculated simple monthly returns for the assigned stock and converted them into excess returns using the monthly risk-free rate (3-month US T-bill)
• •Standardised all return series to consistent units ahead of estimation
• •Produced summary statistics, histograms, and time-series plots of the return series and each factor to describe their distributional properties and trends
• •Estimated a multiple linear regression of the stock's excess returns on the four q-factors, interpreting the coefficients, goodness of fit, and statistical significance of each beta, including a joint significance test
• •Reviewed the assumptions of the Classical Linear Regression Model and assessed the consequences of violating them
• •Plotted the regression residuals and ran White's test to formally test for heteroscedasticity, then re-estimated the model using Huber-White heteroscedasticity-adjusted standard errors
• •Assessed autocorrelation using the Durbin-Watson statistic and the Breusch-Godfrey test up to the tenth lag, then re-estimated the model using Newey-West standard errors to correct for both heteroscedasticity and autocorrelation
• •Tested the normality of the return series using the Jarque-Bera statistic and discussed the practical implications of non-normality
• •Checked for multicollinearity among the explanatory variables and evaluated its impact on the reliability of the regression inferences
• •Delivered a 5-minute recorded presentation summarising the methodology and key findings

Skills & tools