Chow BreakPoint Test Questions
Posted: Thu May 07, 2009 8:36 pm
1. Should I complete the Chow Break Point Test with all Independent Variables in a time series regression, or just the IV that I am concerned with? My F statistic is higher when I include all the variables.
2. Should I select the date by logic? One date, two dates?
a. If I select two dates, is it checking for 3 parallel equations in 3 quartiles?
3. Checking F Statistic: If over 5% then we can accept the null hypothesis of no break?
a. Is the following correct?
i. If I have 33 observation periods, and 8 IVs, then I have 9 coefficients, so: 33-9 = 24 Degrees of Freedom.
b. If I have a lagged variable that is a duplicate, such as oil price and oil price (t-1), do I still count the lagged oil price as an
IV for calculating the degrees of freedom?
i . Do I have to do anything different for the Chow test if I include a lagged variable?
c. Using Table 4 (Critical Values for the Fm,∞ Distrbution), in “Introduction to Econometrics” (James H. Stock and Mark W. Watson) on page 759, the 5% level is 1.52 at 24 degrees of freedom, so based on the results below can we accept the null hypothesis that there are no breaks?
i. Did I calculate the significance level properly?
Chow Breakpoint Test: 1991
Null Hypothesis: No breaks at specified breakpoints
Varying regressors: All equation variables
Equation Sample: 1975 2007
F-statistic 1.553979 Prob. F(9,15) 0.2164
Log likelihood ratio 21.73895 Prob. Chi-Square(9) 0.0097
Wald Statistic 13.98581 Prob. Chi-Square(9) 0.1228
Original Equation:
2. Should I select the date by logic? One date, two dates?
a. If I select two dates, is it checking for 3 parallel equations in 3 quartiles?
3. Checking F Statistic: If over 5% then we can accept the null hypothesis of no break?
a. Is the following correct?
i. If I have 33 observation periods, and 8 IVs, then I have 9 coefficients, so: 33-9 = 24 Degrees of Freedom.
b. If I have a lagged variable that is a duplicate, such as oil price and oil price (t-1), do I still count the lagged oil price as an
IV for calculating the degrees of freedom?
i . Do I have to do anything different for the Chow test if I include a lagged variable?
c. Using Table 4 (Critical Values for the Fm,∞ Distrbution), in “Introduction to Econometrics” (James H. Stock and Mark W. Watson) on page 759, the 5% level is 1.52 at 24 degrees of freedom, so based on the results below can we accept the null hypothesis that there are no breaks?
i. Did I calculate the significance level properly?
Chow Breakpoint Test: 1991
Null Hypothesis: No breaks at specified breakpoints
Varying regressors: All equation variables
Equation Sample: 1975 2007
F-statistic 1.553979 Prob. F(9,15) 0.2164
Log likelihood ratio 21.73895 Prob. Chi-Square(9) 0.0097
Wald Statistic 13.98581 Prob. Chi-Square(9) 0.1228
Original Equation: