Dear Statalisters,
I'm using Stata 15.1.
I have a time series of 24 months, and would like to identify a point of structural break, by assuming there are no "jumps" (so, in a segmented regression scenario). I can't see a way to enforce such no-discontinuity assumption. I know I can use the "breakvars" option in the "estat sbsingle" command and not include the constant in the list (e.g.: breakvars (month)) but, in that case, for each possible point, I should build a new variable (e.g., generate month_centered=month-12.5, for month 13) in order to have the "y" axis precisely where the structural break is expected, and then use the "estat sbknown" command.
Is there a way to compare all points of structural break simultaneously, by including the condition of continuity (so, in a segmented regression setting)?
I'm using Stata 15.1.
I have a time series of 24 months, and would like to identify a point of structural break, by assuming there are no "jumps" (so, in a segmented regression scenario). I can't see a way to enforce such no-discontinuity assumption. I know I can use the "breakvars" option in the "estat sbsingle" command and not include the constant in the list (e.g.: breakvars (month)) but, in that case, for each possible point, I should build a new variable (e.g., generate month_centered=month-12.5, for month 13) in order to have the "y" axis precisely where the structural break is expected, and then use the "estat sbknown" command.
Is there a way to compare all points of structural break simultaneously, by including the condition of continuity (so, in a segmented regression setting)?
