Hi all,
I'm currently working on a project using panel data trying to assess contractual language and the effect on firm outcomes. My supervisor has also asked me to test for firm status effect on contractual language. I wanting to show both of these on the same graph for the contractual terms as outcome and "cause" (using in quotes to make clear that I'm using the term "cause" loosely). Here are the regressions I'm running:
Within the macro
are the variables "franchisee_noncompete", "right_to_purchase", "right_1st_refusal", and "excl_dealing".
This is my code for my coefficient plot:
When I run the code for my coefficient plot, I receive this image, attached:
The top and the bottom three confidence intervals are where PE takeover is the outcome. The four confidence intervals where it says "treat_by_post" are where PE takeover is the "cause". The first confidence interval line in the "treat_by_post" section corresponds to "franchisee_noncompete", the second corresponds to "right_to_purchase", etc.
Is it possible to make the confidence interval lines in the "treat_by_post" section show up under a different color with the restraint that they correspond to? That is, there should be two lines (one for cause, the other for outcome) in the area marked by "Franchisee Bound by Noncompete", two lines in the area marked "Right to Purchase Assets", so on and so forth? Thanks for your help in advance.
I'm currently working on a project using panel data trying to assess contractual language and the effect on firm outcomes. My supervisor has also asked me to test for firm status effect on contractual language. I wanting to show both of these on the same graph for the contractual terms as outcome and "cause" (using in quotes to make clear that I'm using the term "cause" loosely). Here are the regressions I'm running:
Code:
*** Regression for Takeover as Cause
local i = 1
foreach restraint of local restraints {
reghdfe `restraint' treat_by_post, abs(fruns year) vce(cluster fruns)
estimates store est_restraint_outcome_`i'
local i = `i' + 1
}
*** Regression for Takeover as Outcome
local i = 1
foreach restraint of local restraints {
keep if year >= 2013
reghdfe treat_by_post `restraint', abs(fruns year) vce(cluster fruns)
estimates store est_PE_outcome_`i'
local i = `i' + 1
}
Code:
`restraints'
This is my code for my coefficient plot:
Code:
* Coefplot for Non-Competition Restraints
coefplot ///
(est_PE_outcome_1, label("`: variable label franchisee_noncompete'")) ///
(est_restraint_outcome_1, label("`: variable label franchisee_noncompete'")) ///
(est_PE_outcome_2, label("`: variable label right_to_purchase'")) ///
(est_restraint_outcome_2, label("`: variable label right_to_purchase'")) ///
(est_PE_outcome_3, label("`: variable label right_1st_refusal'")) ///
(est_restraint_outcome_3, label("`: variable label right_1st_refusal'")) ///
(est_PE_outcome_4, label("`: variable label excl_dealing'")) ///
(est_restraint_outcome_4, label("`: variable label excl_dealing'")), ///
pstyle(p1) ///
xline(0) drop(_cons) legend(off) xtitle(Coefficient) ///
note("Starting in 2013 for PE Takeover as Outcome" "Starting in 2009 for PE Takeover as Cause")
The top and the bottom three confidence intervals are where PE takeover is the outcome. The four confidence intervals where it says "treat_by_post" are where PE takeover is the "cause". The first confidence interval line in the "treat_by_post" section corresponds to "franchisee_noncompete", the second corresponds to "right_to_purchase", etc.
Is it possible to make the confidence interval lines in the "treat_by_post" section show up under a different color with the restraint that they correspond to? That is, there should be two lines (one for cause, the other for outcome) in the area marked by "Franchisee Bound by Noncompete", two lines in the area marked "Right to Purchase Assets", so on and so forth? Thanks for your help in advance.
Comment