Hello everybody,
I am new to Stata, this forum and econometric models but I hope that you can help me or at least give some hints because I have searched for hours now and have not found any solution how to fix my problem with STATA.
My Problem:
To let you know what I am talk about I attached a reserach paper which I would like to rebuild as well as my STATA dataset (excel) with my own values.
My research is based on the attached resreach paper "Did the value premium survive the subprime credit crisis?" here my main problem is that I do not know how to rebuild the table on page 173 with the regression which is mentioned in footnote 12 in Stata.
In the resreach paper there is a dummy variable included in order to measure the impact of the crisis and non crisis period in order to research the impact of a crisis period on the factor loadings of a 4 factor Carhart regression model (extension of the Fama&French 3 factor model).
But unlike the paper I would like to extend the research (and include 2 crisis periods) within the observation period 31/12/1998 - 30/05/2014 (so that there are overall 5 dummy variables).
I collected a dataset which consists of the excess returns of 8 Indices as well as the Fama and French Factors (Mkt-Rf, SMB, HML, WML).
The whole observation period: 31/12/1998 – today (186 monthly observations)
And the dummy variables should divide the whole sample into different periods (which are basically the crisis / non-crisis periods) which are:
Pre-Crisis 1 31/12/1998 – 29/02/2000 (15 monthly observations) (variable per1)
Dot Com Bubble: 29/02/2000 – 31/10/2002 (32 monthly observations) (variable per2)
Pre-Crisis : 31/10/2002 – 30/03/2007 (53 monthly observations) (variable per3)
Financial Crisis: 30/03/2007 – 31/12/2008 (21 monthly observations) (variable per4)
Post crisis: 31/12/2008 – today (65 monthly observations) (variable per5)
I know that if I do not use dummy variables (and calculate the regression for the whole periode) then the model should look similar like this (please see variables in my excel sheet):
http://en.wikipedia.org/wiki/Fama-Fr...e-factor_model
regress INDEX1G MktRF SML HML WML
But here I would like to include the dummy variables in order to observe if the crisis periods have an impact on the beta loadings and on the alpha coefficient.
The model becomes much more difficult since there are 5 dummy variables which I would like to observe in pairs:
First Period:
Pre-Crisis 1 31/12/1998 – 29/02/2000 (15 monthly observations) (dummy variable per1)
Dot Com Bubble: 29/02/2000 – 31/10/2002 (32 monthly observations) (dummy variable per2)
+abnormal Change from Pre-Crisis to Crisis
Second Period:
Pre-Crisis : 31/10/2002 – 30/03/2007 (53 monthly observations) (dummy variable per3)
Financial Crisis: 30/03/2007 – 31/12/2008 (21 monthly observations) (dummy variable per4)
+abnormal Change from Pre-Crisis to Crisis
Third Period:
Post crisis: 31/12/2008 – today (65 monthly observations) (dummy variable per5)
How can I possibly implement that in Stata?
I am grateful for any help!
Cheers,
Dirk
I am new to Stata, this forum and econometric models but I hope that you can help me or at least give some hints because I have searched for hours now and have not found any solution how to fix my problem with STATA.
My Problem:
To let you know what I am talk about I attached a reserach paper which I would like to rebuild as well as my STATA dataset (excel) with my own values.
My research is based on the attached resreach paper "Did the value premium survive the subprime credit crisis?" here my main problem is that I do not know how to rebuild the table on page 173 with the regression which is mentioned in footnote 12 in Stata.
In the resreach paper there is a dummy variable included in order to measure the impact of the crisis and non crisis period in order to research the impact of a crisis period on the factor loadings of a 4 factor Carhart regression model (extension of the Fama&French 3 factor model).
But unlike the paper I would like to extend the research (and include 2 crisis periods) within the observation period 31/12/1998 - 30/05/2014 (so that there are overall 5 dummy variables).
I collected a dataset which consists of the excess returns of 8 Indices as well as the Fama and French Factors (Mkt-Rf, SMB, HML, WML).
The whole observation period: 31/12/1998 – today (186 monthly observations)
And the dummy variables should divide the whole sample into different periods (which are basically the crisis / non-crisis periods) which are:
Pre-Crisis 1 31/12/1998 – 29/02/2000 (15 monthly observations) (variable per1)
Dot Com Bubble: 29/02/2000 – 31/10/2002 (32 monthly observations) (variable per2)
Pre-Crisis : 31/10/2002 – 30/03/2007 (53 monthly observations) (variable per3)
Financial Crisis: 30/03/2007 – 31/12/2008 (21 monthly observations) (variable per4)
Post crisis: 31/12/2008 – today (65 monthly observations) (variable per5)
I know that if I do not use dummy variables (and calculate the regression for the whole periode) then the model should look similar like this (please see variables in my excel sheet):
http://en.wikipedia.org/wiki/Fama-Fr...e-factor_model
regress INDEX1G MktRF SML HML WML
But here I would like to include the dummy variables in order to observe if the crisis periods have an impact on the beta loadings and on the alpha coefficient.
The model becomes much more difficult since there are 5 dummy variables which I would like to observe in pairs:
First Period:
Pre-Crisis 1 31/12/1998 – 29/02/2000 (15 monthly observations) (dummy variable per1)
Dot Com Bubble: 29/02/2000 – 31/10/2002 (32 monthly observations) (dummy variable per2)
+abnormal Change from Pre-Crisis to Crisis
Second Period:
Pre-Crisis : 31/10/2002 – 30/03/2007 (53 monthly observations) (dummy variable per3)
Financial Crisis: 30/03/2007 – 31/12/2008 (21 monthly observations) (dummy variable per4)
+abnormal Change from Pre-Crisis to Crisis
Third Period:
Post crisis: 31/12/2008 – today (65 monthly observations) (dummy variable per5)
How can I possibly implement that in Stata?
I am grateful for any help!
Cheers,
Dirk
