Time-varying Beta Fama-MacBeth regressions
Posted: Tue Dec 09, 2014 1:23 am
Good day,
Please assist.
I am attempting to run fama-macbeth regressions with time varying betas. I installed the fmb add-in but upon reading the documentation, I realised that risk premia is estimated using constant OLS beta estimates. Hence the need for manual fama-macbeth estimation.
Subsequently, I regressed monthly returns (returns_m) for 16 time series on 5 factors for t-1 to t-60, to set-up time-varying OLS betas (beta1-5) for 144 periods.I then converted the data into a dated panel format, please see FMB data.WF1 attached.
I need to estimate the following equation for all cross sections to determine coefficients (gamma):
returns_m c beta1 beta2 beta3 beta4 beta5
The means of gammas are then the fama-macbeth coefficients. Please could someone explain how to compile these fama-macbeth coefficients, theirs tstats and pvalues as well as the R-squared of the regression.
Thank you,
Johan
Please assist.
I am attempting to run fama-macbeth regressions with time varying betas. I installed the fmb add-in but upon reading the documentation, I realised that risk premia is estimated using constant OLS beta estimates. Hence the need for manual fama-macbeth estimation.
Subsequently, I regressed monthly returns (returns_m) for 16 time series on 5 factors for t-1 to t-60, to set-up time-varying OLS betas (beta1-5) for 144 periods.I then converted the data into a dated panel format, please see FMB data.WF1 attached.
I need to estimate the following equation for all cross sections to determine coefficients (gamma):
returns_m c beta1 beta2 beta3 beta4 beta5
The means of gammas are then the fama-macbeth coefficients. Please could someone explain how to compile these fama-macbeth coefficients, theirs tstats and pvalues as well as the R-squared of the regression.
Thank you,
Johan