add 38

.gitattributes

@@ -75,4 +75,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.mat filter=lfs diff=lfs merge=lfs -text
 *.tif filter=lfs diff=lfs merge=lfs -text
 *.adf filter=lfs diff=lfs merge=lfs -text
-*.pptx filter=lfs diff=lfs merge=lfs -text
+*.pptx filter=lfs diff=lfs merge=lfs -text
+*.Rda filter=lfs diff=lfs merge=lfs -text

38/paper.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:58d468259a9bf5ded115c28f209ba7173a4d7c2ac03ff324284d623664fbb8d4
+size 343008

38/replication_package/Codebooks - Indecent Disclosures.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ea55e4222c62a6b9bee0251c1d5173c7b941fe69f871464b457900040c628293
+size 102232

38/replication_package/ReadMe.txt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:13a9aef4f88cd64d2e10a862a643a4a6246fbcd9282f32bd2d24f3faa41455a2
+size 1107

38/replication_package/Replication.R

@@ -0,0 +1,1075 @@
+### Replication
+
+require("pacman")
+
+pacman::p_load( stargazer, foreign, stringr, data.table, ggplot2,lfe, xtable, openxlsx, zoo, lme4, stringi) 
+rm(list=ls())
+my_log <- file("my_log.txt")
+
+specify_decimal <- function(x, k) format(as.numeric(round(x, k), nsmall=k))
+ihs <- function(x) log(x + sqrt(x^2+1))
+
+mod_stargazer <- function(est) {
+  capture.output(est)
+}
+multiplot <- function(..., plotlist=NULL, file, cols=1, layout=NULL) {
+  require(grid)
+  
+  plots <- c(list(...), plotlist)
+  
+  numPlots = length(plots)
+  
+  if (is.null(layout)) {
+    layout <- matrix(seq(1, cols * ceiling(numPlots/cols)),
+                     ncol = cols, nrow = ceiling(numPlots/cols))
+  }
+  
+  if (numPlots==1) {
+    print(plots[[1]])
+    
+  } else {
+    grid.newpage()
+    pushViewport(viewport(layout = grid.layout(nrow(layout), ncol(layout))))
+    
+    for (i in 1:numPlots) {
+      matchidx <- as.data.frame(which(layout == i, arr.ind = TRUE))
+      
+      print(plots[[i]], vp = viewport(layout.pos.row = matchidx$row,
+                                      layout.pos.col = matchidx$col))
+    }
+  }
+}
+
+load("cands.Rda")
+load("els.Rda")
+
+
+##################################################################################################
+#################            MAIN TEXT                                          ##################
+##################################################################################################
+
+### TABLE 1
+
+### Models (in order)
+
+est1<-felm(perc_elected_partial~interactedtreatment + treatment + after+numberelected_log + electionyear|factor(regionid)+ factor(unit_type)|0 | regionid + electionyear,data=els, psdef=FALSE)
+
+est2<-felm(perc_elected_partial~interactedtreatment + treatment + after+numberelected_log+population_log+territory_log+income_log + electionyear| factor(regionid)+factor(unit_type) |0 | regionid + electionyear,data=els, psdef=FALSE)
+
+est3<-felm(perc_elected_partial~after + interactedtreatment + electionyear| factor(oktmo) |0 | regionid + electionyear,data=els, psdef=FALSE)
+
+est4<-felm(perc_elected_full~interactedtreatment + treatment + after+numberelected_log + electionyear|factor(regionid)+ factor(unit_type) |0 | regionid + electionyear,data=els, psdef=FALSE)
+
+est5<-felm(perc_elected_full~interactedtreatment + treatment + after+numberelected_log+population_log+territory_log+income_log + electionyear| factor(regionid)+factor(unit_type) |0 | regionid + electionyear,data=els, psdef=FALSE)
+
+est6<-felm(perc_elected_full~after + interactedtreatment + electionyear| factor(oktmo) |0 | regionid + electionyear,data=els, psdef=FALSE)
+
+est7<-felm(cands_per_seat~interactedtreatment + treatment + after+numberelected_log + electionyear|factor(regionid)+ factor(unit_type) |0 | regionid + electionyear,data=els, psdef=FALSE)
+
+est8<-felm(cands_per_seat~interactedtreatment + treatment + after+numberelected_log+population_log+territory_log+income_log + electionyear| factor(regionid)+factor(unit_type) |0 | regionid + electionyear,data=els, psdef=FALSE)
+
+est9<-felm(cands_per_seat~after + interactedtreatment + electionyear| factor(oktmo) |0 | regionid + electionyear,data=els, psdef=FALSE)
+
+### Layout 
+
+Region <- list(c("Unit Type, Region Fixed Effects","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{No}}"))
+
+UnitType <- list(c("Unit Type Fixed Effects","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{No}}"))
+
+MuniType <- list(c("Municipality Fixed Effects","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{Yes}}"))
+
+LinearTrend <- list(c("Linear Time Trend","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}"))
+
+originallayout="D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3}"
+newlayout="D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3}| D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3}| D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3}"
+
+t_ro<-mod_stargazer(stargazer(est1,est2,est3,est4,est5,est6,est7,est8,est9, omit="electionyear",keep.stat=c("n","rsq"),dep.var.caption="",covariate.labels=c("Treatment Group * Second Period Election","Treatment Group","Second Period Election","No. Seats (log)","Mun. Population (log)","Mun. Territory (log)","Mun. Revenue (log)"),ci=FALSE, ci.level=0.95, digits=3, column.sep.width="-18pt", align=TRUE, dep.var.labels.include=TRUE, dep.var.labels=c("Part-Time Incumbents (\\%)","Full-Time Incumbents (\\%)","Candidates per Seat"),font.size="small",model.names=FALSE,header=FALSE,add.lines=c(Region,MuniType,LinearTrend),out.header=FALSE,notes.append=FALSE, notes.label="",notes=c(""),star.cutoffs=NA))
+t_ro[11]<-"\\hline \\bigstrut "
+t_ro <-gsub(originallayout, newlayout, t_ro, fixed =TRUE)
+
+t_ro <-gsub("\\multicolumn{10}{r}{} \\\\ ","", t_ro, fixed =TRUE)
+
+t_ro[6]<-paste(" \\resizebox{.99\\textwidth}{!}{",t_ro[6],sep="")
+t_ro[42]<-paste(t_ro[42],"}",sep="")
+
+sink(file="Main_AvgCandsNoHeader.tex")
+cat(t_ro)
+sink()
+
+
+### TABLE 2
+
+est1<-felm(cands_perc_bus~interactedtreatment + treatment + after+numberelected_log + electionyear|factor(regionid)+ factor(unit_type) |0 | regionid + electionyear,data=els, psdef=FALSE)
+
+est2<-felm(cands_perc_bus~interactedtreatment + treatment + after+numberelected_log+population_log+territory_log+income_log + electionyear| factor(regionid)+factor(unit_type) |0 | regionid + electionyear,data=els, psdef=FALSE)
+
+est3<-felm(cands_perc_bus~after + interactedtreatment + electionyear| factor(oktmo) |0 | regionid + electionyear,data=els, psdef=FALSE)
+
+est4<-felm(cands_perc_directors~interactedtreatment + treatment + after+numberelected_log + electionyear|factor(regionid)+ factor(unit_type) |0 | regionid + electionyear,data=els, psdef=FALSE)
+
+est5<-felm(cands_perc_directors~interactedtreatment + treatment + after+numberelected_log+population_log+territory_log+income_log + electionyear| factor(regionid)+factor(unit_type) |0 | regionid + electionyear,data=els, psdef=FALSE)
+
+est6<-felm(cands_perc_directors~after + interactedtreatment + electionyear| factor(oktmo)|0 | regionid + electionyear,data=els, psdef=FALSE)
+
+est7<-felm(cands_perc_entre~interactedtreatment + treatment + after+numberelected_log + electionyear|factor(regionid)+ factor(unit_type) |0 | regionid + electionyear,data=els, psdef=FALSE)
+
+est8<-felm(cands_perc_entre~interactedtreatment + treatment + after+numberelected_log+population_log+territory_log+income_log + electionyear| factor(regionid)+factor(unit_type) |0 | regionid + electionyear,data=els, psdef=FALSE)
+
+est9<-felm(cands_perc_entre~after + interactedtreatment + electionyear| factor(oktmo) |0 | regionid + electionyear,data=els, psdef=FALSE)
+
+t_ro<-mod_stargazer(stargazer(est1,est2,est3,est4,est5,est6,est7,est8,est9, omit="electionyear", keep.stat=c("n","rsq"),dep.var.caption="",covariate.labels=c("Treatment Group * Second Period Election","Treatment Group","Second Period Election","No. Seats (log)","Mun. Population (log)","Mun. Territory (log)","Mun. Revenue (log)"),ci=FALSE, ci.level=0.95, digits=3, column.sep.width="-18pt", align=TRUE, dep.var.labels.include=TRUE, dep.var.labels=c("All Businesspeople (\\%)","Firm Directors (\\%)","Individual Entrepreneurs (\\%)"),font.size="small",model.names=FALSE,header=FALSE,add.lines=c(Region,MuniType,LinearTrend),out.header=FALSE,notes.append=FALSE, notes.label="",notes=c(""),star.cutoffs=NA))
+
+t_ro[11]<-"\\hline \\bigstrut "
+t_ro <-gsub(originallayout, newlayout, t_ro, fixed =TRUE)
+t_ro <-gsub("\\multicolumn{10}{r}{} \\\\ ","", t_ro, fixed =TRUE)
+
+t_ro[6]<-paste(" \\resizebox{.99\\textwidth}{!}{",t_ro[6],sep="")
+t_ro[42]<-paste(t_ro[42],"}",sep="")
+
+
+sink(file="Main_Business.tex")
+cat(t_ro)
+sink()
+
+
+
+### TABLE 3
+### A warning message appears from the felm command because the fixed effects absorb several of the non-time-varying variables within. This is to be expected and can be ignored.
+
+est1<-felm(perc_elected_partial~treatment*after*reg_pressfreedom+electionyear+ lngdp + log_pop  + resource_grppct + reg_urbanshare  + log_mincome + reg_sharepensm | factor(oktmo) |0 | regionid + electionyear,data=els,psdef=FALSE)
+
+est2<-felm(perc_elected_partial~treatment*after*reg_dem_media+electionyear++ lngdp + log_pop  + resource_grppct + reg_urbanshare  + log_mincome + reg_sharepensm | factor(oktmo) |0 | regionid + electionyear,data=els,psdef=FALSE)
+
+est3<-felm(perc_elected_partial~treatment*after*fn_budget_log+electionyear+ lngdp + log_pop  + resource_grppct + reg_urbanshare  + log_mincome + reg_sharepensm | factor(oktmo) |0 | regionid + electionyear,data=els,psdef=FALSE)
+
+est4<-felm(perc_elected_partial~treatment*after*log_justice+electionyear+ lngdp + log_pop  + resource_grppct + reg_urbanshare  + log_mincome + reg_sharepensm | factor(oktmo) |0 | regionid + electionyear,data=els,psdef=FALSE)
+
+est5<-felm(cands_perc_entre~treatment*after*reg_pressfreedom+electionyear+ lngdp + log_pop  + resource_grppct + reg_urbanshare  + log_mincome + reg_sharepensm | factor(oktmo) |0 | regionid + electionyear,data=els,psdef=FALSE)
+
+est6<-felm(cands_perc_entre~treatment*after*reg_dem_media+electionyear+ lngdp + log_pop  + resource_grppct + reg_urbanshare  + log_mincome + reg_sharepensm | factor(oktmo) |0 | regionid + electionyear,data=els,psdef=FALSE)
+
+est7<-felm(cands_perc_entre~treatment*after*fn_budget_log+electionyear+ lngdp + log_pop  + resource_grppct + reg_urbanshare  + log_mincome + reg_sharepensm | factor(oktmo) |0 | regionid + electionyear,data=els,psdef=FALSE)
+
+est8<-felm(cands_perc_entre~treatment*after*log_justice+electionyear+ lngdp + log_pop  + resource_grppct + reg_urbanshare  + log_mincome + reg_sharepensm| factor(oktmo) |0 | regionid + electionyear,data=els,psdef=FALSE)
+
+Muni <- list(c("Regional Covariates","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}"))
+
+Region <- list(c("Municipality FE; Linear Time Trend","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}"))
+
+t_ro<-mod_stargazer(stargazer(est1,est2,est3,est4,est5,est6,est7,est8, keep.stat=c("n","rsq"),dep.var.caption="",keep=c("after","treatment:after","treatment:after:reg_pressfreedom","treatment:after:reg_dem_media"   ,"treatment:after:audits_allpeople","treatment:after:ENFORCE"),
+   covariate.labels=c("Second Election","Treatment Group * Second Election",
+    "Second Election * GDF Press Freedom",     
+    "\\textbf{Treatment Group * Second Election * GDF Press Freedom}",
+    "Second Election * TP Press Freedom",     
+     "\\textbf{Treatment Group * Second Election * TP Press Freedom}",
+    "Second Election * Regional Tax Agency Budget",
+     "\\textbf{Treatment Group * Second Election * Regional Tax Agency Budget}",
+     "Second Election * Law Enforcement Personnel",
+      "\\textbf{Treatment Group * Second Election * Law Enforcement Personnel}"),ci=FALSE, ci.level=0.95, digits=3, column.sep.width="-20pt", align=TRUE, dep.var.labels.include=TRUE, dep.var.labels=c("Part-Time Incumbents (\\%)","Independent Entrepreneurs (\\%)"),font.size="small",model.names=FALSE,header=FALSE,add.lines=c(Muni,Region),out.header=FALSE,notes.append=FALSE, notes.label="",notes=c(""),star.cutoffs=NA))
+
+t_ro <-gsub("\\multicolumn{9}{r}{} \\\\ ","", t_ro, fixed =TRUE)
+
+
+t_ro[6]<-paste(" \\resizebox{.99\\textwidth}{!}{",t_ro[6],sep="")
+t_ro[50]<-paste(t_ro[50],"}",sep="")
+
+
+sink(file="Heterogeneity_MainInteractions.tex")
+cat(t_ro)
+sink()
+
+
+
+
+
+### FIGURE 1
+
+vrns_chart<-els[,list(vrns=uniqueN(vrn)),by=c("electionyear","treatment")]
+vrns_chart$treatment<-as.character(vrns_chart$treatment)
+vrns_chart$electionyear<-as.character(vrns_chart$electionyear)
+
+ggplot(vrns_chart, aes(x = electionyear, y = vrns, fill = treatment)) +
+  geom_bar(stat = "identity")+scale_fill_grey(start = 0.6, end = 0.3,name="",breaks=c("0", "1"),labels=c("Control     ", "Treatment     "))+ylim(0,7000)+xlab("\nElection Year")+ylab("Number of Elections\n")+theme_bw()+
+  theme(legend.key = element_rect(size = 5),
+        legend.key.size = unit(1.5, 'lines'),
+        plot.title=element_text(size=14,hjust = 0.65),
+        axis.text=element_text(size=14),axis.title=element_text(size=16),
+        legend.text=element_text(size=16)) +geom_vline(aes(xintercept=7.5),colour="darkgrey", linetype="dashed")+ annotate("text", label = "Amendment\n In Effect", x = 8.5, y = 5000, size = 4, colour = "black", angle=0)+geom_vline(aes(xintercept=5.5),colour="black")+ ggtitle("First Period Election                          Second Period Election")+
+  annotate("rect", xmin = 7.5, xmax = 9.5, ymin = 0, ymax = Inf,
+           alpha = .15)
+ggsave(filename = "ElectionsByYear.pdf", height=6, width=10)
+
+
+
+
+##################################################################################################
+#################            APPENDIX                                           ##################
+##################################################################################################
+
+
+
+### FIGURE A1
+
+cands$coded_profession<-0
+cands$coded_profession[cands$entrepreneur==1]<-1
+cands$coded_profession[cands$director==1]<-2
+cands$coded_profession[cands$teacher==1]<-3
+cands$coded_profession[cands$accountant==1]<-4
+cands$coded_profession[cands$doctor==1]<-5
+cands$coded_profession[cands$lowerclass==1]<-6
+cands$coded_profession[cands$nowork==1]<-7
+cands$coded_profession[cands$force==1]<-8
+cands$coded_profession[cands$official==1]<-9
+cands$coded_profession[cands$ngo==1]<-10
+
+prop<-as.data.frame(prop.table(table(cands$coded_profession)))
+
+prop$Freq<-prop$Freq*100
+prop$label <-paste0(specify_decimal(prop$Freq,1),"%",sep="")
+prop$Var1<-as.factor(prop$Var1)
+prop_breaks<-unique(as.factor(prop$Var1))
+
+prop <- transform(prop, Var1=reorder(Var1, Freq) ) 
+
+ggplot(prop, aes(x=Var1,y=Freq))+geom_bar(stat = "identity")+xlab("")+ylab("\nPercentage of Candidates (%)")+ geom_text(aes(label=label), position=position_dodge(width=0.9), hjust=-0.2,size=5)+ylim(0,25)+ coord_flip() + theme_bw()+ scale_x_discrete(breaks=prop_breaks,labels=c("Other","Entrepreneur","Firm Director","Education","Private Sector Professional","Health Care","Blue Collar","Unemployed / Pensioner","Law Enforcement","Government / SOE","Civil Society"))+theme(axis.text=element_text(size=18),axis.title=element_text(size=18)) + guides(fill=guide_legend(title=" "))
+
+ggsave(filename = "Professions.pdf", height=3.5, width=10)
+
+
+### TABLE A2
+
+els_first<-subset(els, sequence==1)
+els_second<-subset(els, sequence==2)
+
+SummaryTables <- data.frame(title=numeric(0),treatment= numeric(0),control= numeric(0),difference= numeric(0),pvalue=numeric(0))
+
+SummaryTables[1 ,] <- c("(1) Population (log)", 
+                        mean(els_first$population_log[els_first$treatment==1],na.rm=TRUE), 
+                        mean(els_first$population_log[els_first$treatment==0],na.rm=TRUE),
+                        summary(lm(population_log~treatment, data=els_first))[[4]][2],
+                        summary(lm(population_log~treatment, data=els_first))[[4]][8])
+
+SummaryTables[2 ,] <- c("(2) Territory (log)", 
+                        mean(els_first$territory_log[els_first$treatment==1],na.rm=TRUE), 
+                        mean(els_first$territory_log[els_first$treatment==0],na.rm=TRUE),
+                        summary(lm(territory_log~treatment, data=els_first))[[4]][2],
+                        summary(lm(territory_log~treatment, data=els_first))[[4]][8]
+)
+SummaryTables[3 ,] <- c("(3) Revenue (log)", 
+                        mean(els_first$income_log[els_first$treatment==1],na.rm=TRUE), 
+                        mean(els_first$income_log[els_first$treatment==0],na.rm=TRUE),
+                        summary(lm(income_log~treatment, data=els_first))[[4]][2],
+                        summary(lm(income_log~treatment, data=els_first))[[4]][8])
+SummaryTables[4 ,] <- c("(4) City Settlement", 
+                        mean(els_first$gorpos[els_first$treatment==1],na.rm=TRUE), 
+                        mean(els_first$gorpos[els_first$treatment==0],na.rm=TRUE),
+                        summary(lm(gorpos~treatment, data=els_first))[[4]][2],
+                        summary(lm(gorpos~treatment, data=els_first))[[4]][8] )
+
+SummaryTables[5 ,] <- c("(5) Rural Settlement", 
+                        mean(els_first$selpos[els_first$treatment==1],na.rm=TRUE), 
+                        mean(els_first$selpos[els_first$treatment==0],na.rm=TRUE),
+                        summary(lm(selpos~treatment, data=els_first))[[4]][2],
+                        summary(lm(selpos~treatment, data=els_first))[[4]][8])
+
+SummaryTables[6 ,] <- c("(6) City District", 
+                        mean(els_first$gorokrug[els_first$treatment==1],na.rm=TRUE), 
+                        mean(els_first$gorokrug[els_first$treatment==0],na.rm=TRUE),
+                        summary(lm(gorokrug~treatment, data=els_first))[[4]][2],
+                        summary(lm(gorokrug~treatment, data=els_first))[[4]][8])
+
+SummaryTables[7 ,] <- c("(7) Municipal Rayon", 
+                        mean(els_first$munrayon[els_first$treatment==1],na.rm=TRUE), 
+                        mean(els_first$munrayon[els_first$treatment==0],na.rm=TRUE),
+                        summary(lm(munrayon~treatment, data=els_first))[[4]][2],
+                        summary(lm(munrayon~treatment, data=els_first))[[4]][8])
+
+SummaryTables[8 ,] <- c("(8) Number Seats", 
+                        mean(els_first$numberelected[els_first$treatment==1],na.rm=TRUE), 
+                        mean(els_first$numberelected[els_first$treatment==0],na.rm=TRUE),
+                        summary(lm(numberelected~treatment, data=els_first))[[4]][2],
+                        summary(lm(numberelected~treatment, data=els_first))[[4]][8])
+
+SummaryTables[9 ,] <- c("(9) Number Candidates per Seat", 
+                        mean(els_first$cands_per_seat[els_first$treatment==1],na.rm=TRUE), 
+                        mean(els_first$cands_per_seat[els_first$treatment==0],na.rm=TRUE),
+                        summary(lm(cands_per_seat~treatment, data=els_first))[[4]][2],
+                        summary(lm(cands_per_seat~treatment, data=els_first))[[4]][8])
+
+SummaryTables[10 ,] <- c("(10) Part-time Deputy Candidates (%)", 
+                         mean(els_first$perc_elected_partial[els_first$treatment==1],na.rm=TRUE), 
+                         mean(els_first$perc_elected_partial[els_first$treatment==0],na.rm=TRUE),
+                         summary(lm(perc_elected_partial~treatment, data=els_first))[[4]][2],
+                         summary(lm(perc_elected_partial~treatment, data=els_first))[[4]][8])
+
+SummaryTables[11 ,] <- c("(11) Full-time Deputy Candidates (%)", 
+                         mean(els_first$perc_elected_full[els_first$treatment==1],na.rm=TRUE), 
+                         mean(els_first$perc_elected_full[els_first$treatment==0],na.rm=TRUE),
+                         summary(lm(perc_elected_full~treatment, data=els_first))[[4]][2],
+                         summary(lm(perc_elected_full~treatment, data=els_first))[[4]][8])
+
+SummaryTables[12 ,] <- c("(12) Businessperson Candidates (%)", 
+                         mean(els_first$cands_perc_bus[els_first$treatment==1],na.rm=TRUE), 
+                         mean(els_first$cands_perc_bus[els_first$treatment==0],na.rm=TRUE),
+                         summary(lm(cands_perc_bus~treatment, data=els_first))[[4]][2],
+                         summary(lm(cands_perc_bus~treatment, data=els_first))[[4]][8])
+
+SummaryTables[13 ,] <- c("(13) Candidate Age", 
+                         mean(els_first$age[els_first$treatment==1],na.rm=TRUE), 
+                         mean(els_first$age[els_first$treatment==0],na.rm=TRUE),
+                         summary(lm(age~treatment, data=els_first))[[4]][2],
+                         summary(lm(age~treatment, data=els_first))[[4]][8])
+
+SummaryTables[14 ,] <- c("(14) Female Candidates (%)", 
+                         mean(els_first$female[els_first$treatment==1],na.rm=TRUE), 
+                         mean(els_first$female[els_first$treatment==0],na.rm=TRUE),
+                         summary(lm(female~treatment, data=els_first))[[4]][2],
+                         summary(lm(female~treatment, data=els_first))[[4]][8])
+
+SummaryTables$treatment<-prettyNum(specify_decimal(as.numeric(SummaryTables$treatment),3),big.mark=",")
+SummaryTables$control<-prettyNum(specify_decimal(as.numeric(SummaryTables$control),3),big.mark=",")
+SummaryTables$difference<-prettyNum(specify_decimal(as.numeric(SummaryTables$difference),3),big.mark=",")
+SummaryTables$pvalue<-as.numeric(SummaryTables$pvalue)
+
+SummaryTables[15 ,] <- c("(15) Number of Elections", 
+                         prettyNum(length(unique(els_first$vrn[els_first$treatment==1])),,big.mark=","), 
+                         prettyNum(length(unique(els_first$vrn[els_first$treatment==0])),,big.mark=","), 
+                         "",
+                         "")
+SummaryTables$pvalue=NULL
+
+colnames(SummaryTables) <- c(" ","Treated Elections","Control Elections","Difference")
+S1<- capture.output(print.xtable(xtable(SummaryTables, digits=3, align="llccc",caption.placement='top',floating=TRUE,tocharFun=prettyNum),include.rownames =FALSE,hline.after=c(0,3,7,14,15)))
+
+sink(file="PreTreatmentTable.tex")
+cat(S1)
+sink()
+
+
+#### TABLE A3
+
+elections_summary<-els[,list(numbercands, numberelected, cands_per_seat, perc_elected_partial, perc_elected_full, cands_perc_directors, cands_perc_entre, female, age, income_log, population_log,territory_log,lngdp,log_pop,resource_grppct,reg_urbanshare,log_mincome,reg_sharepensm,reg_pressfreedom,reg_dem_media,fn_budget_log,log_justice,audits_allpeople,ENFORCE)]
+
+electionstable<-mod_stargazer(stargazer(elections_summary,covariate.labels=c(
+  "No. Candidates",
+  "No. Seats",
+  "Candidates per Seat",
+  "Part-Time Incumbents (\\%)",
+  "Full-Time Incumbents (\\%)",
+  "Firm Directors (\\%)",
+  "Entrepreneurs (\\%)",
+  "Female (\\%)",
+  "Mean Age",
+  "Revenue (log)","Population (log)","Territory (log)","Regional GDP (log)","Regional Population (log)","GDP from Natural Resources (\\%)","Urbanization (\\%)","Average Income (log)","Share of Pensioners (\\%)","GDF Press Freedom","TP Press Freedom","Regional Tax Agency Budget (log)","Law Enforcement Personnel (log)","Audit Risk","Enforcement Expenditures"),summary.stat=c("n","min","max","mean","median"),header=FALSE,digits=3,star.cutoffs = NA))
+electionstable[19]<-paste("\\hline ",electionstable[19],sep="")
+electionstable[22]<-paste("\\hline ",electionstable[22],sep="")
+
+electionstable <-gsub("0.00000", "0", electionstable, fixed =TRUE)
+
+sink(file="ElectionsStats.tex")
+cat(electionstable)
+sink()
+
+
+### FIGURE C1
+
+els_first$treatment<-as.character(els_first$treatment)
+els_first_r<-subset(els_first, is.na(territory_log)==FALSE & is.na(income_log)==FALSE & is.na(population_log)==FALSE)
+
+est_r<-felm(cands_per_seat~numberelected_log+territory_log+income_log+population_log+electionyear|factor(regionid)+factor(unit_type) |0 | regionid,data=els_first_r)
+
+els_first_r$residuals_per_seat<-est_r$residuals
+
+diff<-felm(residuals_per_seat~treatment, data=els_first_r)
+coef<-specify_decimal(diff$coefficients[2],3)
+p<-specify_decimal(diff$pval[2],2)
+
+plot_per_seat<-ggplot(els_first_r,aes(x = residuals_per_seat,fill=treatment)) + geom_density(alpha=0.25)+scale_fill_brewer(palette = "Set1",name="",breaks=c("0", "1"),labels=c("Control     ", "Treatment     "))+xlab(paste0("\nResiduals\n\nDiff: ",coef,' p: ',p,"\n",sep=""))+ylab("Density\n")+
+  theme(
+    plot.title=element_text(size=14,hjust = 0.65),
+    axis.text=element_text(size=12),axis.title=element_text(size=12),
+    legend.text=element_text(size=12))+ggtitle("(a) Candidates Per Seat")
+
+est_r<-felm(perc_elected_partial~numberelected_log+territory_log+income_log+population_log+electionyear+cands_per_seat+electionyear| factor(electionyear)+factor(regionid)+factor(unit_type) |0 | regionid,data=els_first_r)
+
+els_first_r$residuals_partial<-est_r$residuals
+diff<-felm(residuals_partial~treatment, data=els_first_r)
+coef<-specify_decimal(diff$coefficients[2],3)
+p<-specify_decimal(diff$pval[2],2)
+
+plot_partial<-ggplot(els_first_r,aes(x = residuals_partial,fill=treatment)) + geom_density(alpha=0.25)+scale_fill_brewer(palette = "Set1",name="",breaks=c("0", "1"),labels=c("Control     ", "Treatment     "))+xlab(paste0("\nResiduals\n\nDiff: ",coef,' p: ',p,"\n",sep=""))+ylab("Density\n")+
+  theme(
+    plot.title=element_text(size=14,hjust = 0.65),
+    axis.text=element_text(size=12),axis.title=element_text(size=12),
+    legend.text=element_text(size=12))+ggtitle("(b) Part-Time Incumbents (%)")
+
+est_r<-felm(perc_elected_full~numberelected_log+territory_log+income_log+population_log+electionyear| factor(electionyear)+factor(regionid)+factor(unit_type) |0 | regionid,data=els_first_r)
+
+els_first_r$residuals_full<-est_r$residuals
+diff<-felm(residuals_full~treatment, data=els_first_r)
+coef<-specify_decimal(diff$coefficients[2],3)
+p<-specify_decimal(diff$pval[2],2)
+
+plot_full<-ggplot(els_first_r,aes(x = residuals_full,fill=treatment)) + geom_density(alpha=0.25)+scale_fill_brewer(palette = "Set1",name="",breaks=c("0", "1"),labels=c("Control     ", "Treatment     "))+xlab(paste0("\nResiduals\n\nDiff: ",coef,' p: ',p,"\n",sep=""))+ylab("Density\n")+
+  theme(
+    plot.title=element_text(size=14,hjust = 0.65),
+    axis.text=element_text(size=12),axis.title=element_text(size=12),
+    legend.text=element_text(size=12))+ggtitle("(c) Full-Time Incumbents (%)")+xlim(-.025,.025)
+
+est_r<-felm(cands_perc_bus~numberelected_log+territory_log+income_log+population_log+electionyear| factor(electionyear)+factor(regionid)+factor(unit_type) |0 | regionid,data=els_first_r)
+
+els_first_r$residuals_bus<-est_r$residuals
+diff<-felm(residuals_bus~treatment, data=els_first_r)
+coef<-specify_decimal(diff$coefficients[2],3)
+p<-specify_decimal(diff$pval[2],2)
+
+plot_bus<-ggplot(els_first_r,aes(x = residuals_bus,fill=treatment)) + geom_density(alpha=0.25)+scale_fill_brewer(palette = "Set1",name="",breaks=c("0", "1"),labels=c("Control     ", "Treatment     "))+xlab(paste0("\nResiduals\n\nDiff: ",coef,' p: ',p,"\n",sep=""))+ylab("Density\n")+
+  theme( plot.title=element_text(size=14,hjust = 0.65),
+         axis.text=element_text(size=12),axis.title=element_text(size=12),
+         legend.text=element_text(size=12))+ggtitle("(d) Businesspeople (%)")
+
+est_r<-felm(cands_perc_directors~numberelected_log+territory_log+income_log+population_log+electionyear| factor(electionyear)+factor(regionid)+factor(unit_type) |0 | regionid,data=els_first_r)
+
+els_first_r$residuals_directors<-est_r$residuals
+diff<-felm(residuals_directors~treatment, data=els_first_r)
+coef<-specify_decimal(diff$coefficients[2],3)
+p<-specify_decimal(diff$pval[2],2)
+
+plot_directors<-ggplot(els_first_r,aes(x = residuals_directors,fill=treatment)) + geom_density(alpha=0.25)+scale_fill_brewer(palette = "Set1",name="",breaks=c("0", "1"),labels=c("Control     ", "Treatment     "))+xlab(paste0("\nResiduals\n\nDiff: ",coef,' p: ',p,"\n",sep=""))+ylab("Density\n")+
+  theme( plot.title=element_text(size=14,hjust = 0.65),
+         axis.text=element_text(size=12),axis.title=element_text(size=12),
+         legend.text=element_text(size=12))+ggtitle("(e) Firm Directors (%)")
+
+est_r<-felm(cands_perc_entre~numberelected_log+territory_log+income_log+population_log+electionyear| factor(electionyear)+factor(regionid)+factor(unit_type) |0 | regionid,data=els_first_r)
+
+els_first_r$residuals_entre<-est_r$residuals
+diff<-felm(residuals_entre~treatment, data=els_first_r)
+coef<-specify_decimal(diff$coefficients[2],3)
+coef<-ifelse(coef=="0",'0.001',coef)
+p<-specify_decimal(diff$pval[2],2)
+
+plot_entre<-ggplot(els_first_r,aes(x = residuals_entre,fill=treatment)) + geom_density(alpha=0.25)+scale_fill_brewer(palette = "Set1",name="",breaks=c("0", "1"),labels=c("Control     ", "Treatment     "))+xlab(paste0("\nResiduals\n\nDiff: ",coef,' p: ',p,"\n",sep=""))+ylab("Density\n")+
+  theme( plot.title=element_text(size=14,hjust = 0.65),
+         axis.text=element_text(size=12),axis.title=element_text(size=12),
+         legend.text=element_text(size=12))+ggtitle("(f) Entrepreneurs (%)")
+
+pdf(file = "ResidualsHistograms.pdf", height = 12, width = 12)
+multiplot(plot_per_seat,plot_full,plot_directors,plot_partial,plot_bus,plot_entre,cols=2)
+dev.off()
+
+
+
+### TABLE D1
+
+cands_e<-subset(cands, elected_izbirkom==1)
+
+### Only take candidates that are in the first sequence and get their treatment status
+els_first_t<-els_first[,list(vrn,treatment,unit_type,population_log,territory_log,income_log)]
+
+cands_e<-merge(cands_e, els_first_t,by=c("vrn"))
+
+### Create indicator for whether candidate ran again
+g_ran_again<-cands[cands$vrn %in% els_second$vrn]
+
+g_ran_again<-g_ran_again[,list(fullname, birthyear,oktmo,elected_izbirkom,nextparty=party)]
+g_ran_again$reran<-1
+setnames(g_ran_again,"elected_izbirkom","elected_izbirkom_again")
+
+cands_e<-merge(cands_e, g_ran_again,by=c("fullname","oktmo","birthyear"),all.x=TRUE,all.y=FALSE)
+cands_e$reran[is.na(cands_e$reran)==TRUE]=0
+
+### Demographics
+
+cands_e$log_age<-log(cands_e$age)
+cands_e$numberelected_log<-log(cands_e$numberelected)
+cands_e$numbercands_log<-log(cands_e$numbercands)
+
+cands_e$systemic_opposition<-ifelse(cands_e$party=="kprf" | cands_e$party=="ldpr" | cands_e$party=="sr" | cands_e$party=="rod",1,0)
+cands_e$other_opposition<-ifelse(cands_e$party=="patriots" | cands_e$party=="oth" | cands_e$party=="yab" ,1,0)
+cands_e$ur<-ifelse(cands_e$party=="ur" ,1,0)
+cands_e$nextparty_ur<-ifelse(cands_e$nextparty=="ur" ,1,0)
+
+#### MODELS
+
+est1<-felm(reran~treatment | factor(regionid)+ factor(unit_type) + factor(electionyear)|0 | regionid,data=cands_e)
+
+est2<-felm(reran~treatment + female + log_age + businessperson + onlyincumbent + partial_deputy +ur + systemic_opposition + other_opposition| factor(regionid)+ factor(unit_type) + factor(electionyear)|0 | regionid,data=cands_e)
+
+est3<-felm(reran~treatment + female + log_age + businessperson + onlyincumbent + partial_deputy +ur + systemic_opposition + other_opposition+numberelected_log + numbercands_log + population_log + territory_log + income_log| factor(regionid)+ factor(unit_type) + factor(electionyear)|0 | regionid,data=cands_e)
+
+est4<-felm(elected_izbirkom_again~treatment | factor(regionid)+ factor(unit_type) + factor(electionyear)|0 | regionid,data=subset(cands_e,reran==1))
+
+est5<-felm(elected_izbirkom_again~treatment + female + log_age + businessperson + onlyincumbent + partial_deputy +ur + systemic_opposition + other_opposition| factor(regionid)+ factor(unit_type) + factor(electionyear)|0 | regionid,data=subset(cands_e,reran==1))
+
+est6<-felm(elected_izbirkom_again~treatment + female + log_age + businessperson + onlyincumbent + partial_deputy +ur + systemic_opposition + other_opposition+numberelected_log + numbercands_log + population_log + territory_log + income_log| factor(regionid)+ factor(unit_type) + factor(electionyear)|0 | regionid,data=subset(cands_e,reran==1))
+
+t_ro<-mod_stargazer(stargazer(est1,est2,est3,est4,est5,est6, omit="electionyear",keep.stat=c("n","rsq"),dep.var.caption="",covariate.labels=c("Treatment Group","Female","Age (log)","Businessperson","Full-time Incumbent (previous term)","Part-time Incumbent (previous term)","Ruling Party","Systemic Opposition","Other Opposition","Council Size","No. Cands (first election)","Population (log)","Territory (log)","Revenue (log)"),ci=FALSE, ci.level=0.95, digits=3, column.sep.width="-10pt", align=TRUE, dep.var.labels.include=TRUE, dep.var.labels=c("Incumbent Re-ran in Second Election","Incumbent Won in Second Election"),font.size="small",model.names=FALSE,header=FALSE,add.lines=c(UnitType,Region),out.header=FALSE,notes.append=FALSE, notes.label="",notes=c(""),star.cutoffs=NA))
+
+### Layout
+
+Region <- list(c("Region, Year Fixed Effects","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}"))
+
+UnitType <- list(c("Unit Type Fixed Effects","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}"))
+
+originallayout="D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3}"
+newlayout=" D{.}{.}{-3} D{.}{.}{-3}| D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3}| D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3}"
+
+t_ro[9]<-paste(t_ro[9],"\\cmidrule(l{15pt}r{15pt}){2-4}\\cmidrule(l{15pt}r{15pt}){5-7}\\\\",sep="")
+
+t_ro[11]<-"\\hline \\bigstrut "
+t_ro <-gsub("\\multicolumn{7}{r}{} \\\\ ","", t_ro, fixed =TRUE)
+
+sink(file="Appendix_Reran_andWon.tex")
+cat(t_ro)
+sink()
+
+
+
+
+
+
+
+##### TABLE D2
+
+est1<-felm(reran~treatment | factor(regionid)+ factor(unit_type) + factor(electionyear)|0 | regionid,data=subset(cands_e, ur==1))
+
+est2<-felm(reran~treatment + female + log_age + businessperson + onlyincumbent + partial_deputy| factor(regionid)+ factor(unit_type) + factor(electionyear)|0 | regionid,data=subset(cands_e, ur==1))
+
+est3<-felm(reran~treatment + female + log_age + businessperson + onlyincumbent + partial_deputy+numberelected_log + numbercands_log + population_log + territory_log + income_log| factor(regionid)+ factor(unit_type) + factor(electionyear)|0 | regionid,data=subset(cands_e, ur==1))
+
+est4<-felm(elected_izbirkom_again~treatment | factor(regionid)+ factor(unit_type) + factor(electionyear)|0 | regionid,data=subset(cands_e, ur==1 & reran==1))
+
+est5<-felm(elected_izbirkom_again~treatment + female + log_age + businessperson + onlyincumbent + partial_deputy| factor(regionid)+ factor(unit_type) + factor(electionyear)|0 | regionid,data=subset(cands_e, ur==1 & reran==1))
+
+est6<-felm(elected_izbirkom_again~treatment + female + log_age + businessperson + onlyincumbent + partial_deputy+numberelected_log + numbercands_log + population_log + territory_log + income_log| factor(regionid)+ factor(unit_type) + factor(electionyear)|0 | regionid,data=subset(cands_e, ur==1 & reran==1))
+
+t_ro<-mod_stargazer(stargazer(est1,est2,est3,est4,est5,est6, omit="electionyear",keep.stat=c("n","rsq"),dep.var.caption="",covariate.labels=c("Treatment Group","Female","Age (log)","Businessperson","Full-time Incumbent (previous term)","Part-time Incumbent (previous term)","Council Size","No. Cands (first election)","Population (log)","Territory (log)","Revenue (log)"),ci=FALSE, ci.level=0.95, digits=3, column.sep.width="-10pt", align=TRUE, dep.var.labels.include=TRUE, dep.var.labels=c("Incumbent Re-ran in Second Election","Incumbent Won in Second Election"),font.size="small",model.names=FALSE,header=FALSE,add.lines=c(UnitType,Region),out.header=FALSE,notes.append=FALSE, notes.label="",notes=c(""),star.cutoffs=NA))
+
+### Layout
+Region <- list(c("Region, Year Fixed Effects","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}"))
+
+UnitType <- list(c("Unit Type Fixed Effects","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}"))
+
+originallayout="D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3}"
+newlayout=" D{.}{.}{-3} D{.}{.}{-3}| D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3}| D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3}"
+
+t_ro[9]<-paste(t_ro[9],"\\cmidrule(l{15pt}r{15pt}){2-4}\\cmidrule(l{15pt}r{15pt}){5-7}\\\\",sep="")
+
+t_ro[11]<-"\\hline \\bigstrut "
+t_ro <-gsub("\\multicolumn{7}{r}{} \\\\ ","", t_ro, fixed =TRUE)
+
+sink(file="Appendix_Reran_onlyUR.tex")
+cat(t_ro)
+sink()
+
+
+
+
+##### TABLE D3
+
+est1<-felm(nextparty_ur~treatment | factor(regionid)+ factor(unit_type) + factor(electionyear)|0 | regionid,data=subset(cands_e, ur==1 & reran==1))
+
+est2<-felm(nextparty_ur~treatment + female + log_age + businessperson + onlyincumbent + partial_deputy| factor(regionid)+ factor(unit_type) + factor(electionyear)|0 | regionid,data=subset(cands_e, ur==1& reran==1))
+
+est3<-felm(nextparty_ur~treatment + female + log_age + businessperson + onlyincumbent + partial_deputy+numberelected_log + numbercands_log + population_log + territory_log + income_log| factor(regionid)+ factor(unit_type) + factor(electionyear)|0 | regionid,data=subset(cands_e, ur==1& reran==1))
+
+t_ro<-mod_stargazer(stargazer(est1,est2,est3,omit="electionyear",keep.stat=c("n","rsq"),dep.var.caption="",covariate.labels=c("Treatment Group","Female","Age (log)","Businessperson","Full-time Incumbent (previous term)","Part-time Incumbent (previous term)","Council Size","No. Cands (first election)","Population (log)","Territory (log)","Revenue (log)"),ci=FALSE, ci.level=0.95, digits=3, column.sep.width="-10pt", align=TRUE, dep.var.labels.include=TRUE, dep.var.labels=c("UR Incumbent Re-Ran with UR"),font.size="small",model.names=FALSE,header=FALSE,add.lines=c(UnitType,Region),out.header=FALSE,notes.append=FALSE, notes.label="",notes=c(""),star.cutoffs=NA))
+
+
+t_ro <-gsub("\\multicolumn{4}{r}{} \\\\ ","", t_ro, fixed =TRUE)
+
+sink(file="Appendix_Reran_PartySwitching.tex")
+cat(t_ro)
+sink()
+
+
+###### TABLE D4
+
+est1<-felm(perc_elected_partial~interactedtreatment + treatment + after+numberelected_log + electionyear|factor(regionid)+ factor(unit_type)|0 | oktmo+ electionyear,data=els, psdef=FALSE)
+
+est2<-felm(perc_elected_partial~interactedtreatment + treatment + after+numberelected_log+population_log+territory_log+income_log + electionyear| factor(regionid)+factor(unit_type) |0 | oktmo+ electionyear,data=els, psdef=FALSE)
+
+est3<-felm(perc_elected_partial~after + interactedtreatment + electionyear| factor(oktmo) |0 | oktmo+ electionyear,data=els, psdef=FALSE)
+
+est4<-felm(perc_elected_full~interactedtreatment + treatment + after+numberelected_log + electionyear|factor(regionid)+ factor(unit_type) |0 | oktmo+ electionyear,data=els, psdef=FALSE)
+
+est5<-felm(perc_elected_full~interactedtreatment + treatment + after+numberelected_log+population_log+territory_log+income_log + electionyear| factor(regionid)+factor(unit_type) |0 | oktmo+ electionyear,data=els, psdef=FALSE)
+
+est6<-felm(perc_elected_full~after + interactedtreatment + electionyear| factor(oktmo) |0 | oktmo+ electionyear,data=els, psdef=FALSE)
+
+est7<-felm(cands_per_seat~interactedtreatment + treatment + after+numberelected_log + electionyear|factor(regionid)+ factor(unit_type) |0 | oktmo+ electionyear,data=els, psdef=FALSE)
+
+est8<-felm(cands_per_seat~interactedtreatment + treatment + after+numberelected_log+population_log+territory_log+income_log + electionyear| factor(regionid)+factor(unit_type) |0 | oktmo+ electionyear,data=els, psdef=FALSE)
+
+est9<-felm(cands_per_seat~after + interactedtreatment + electionyear| factor(oktmo) |0 | oktmo+ electionyear,data=els, psdef=FALSE)
+
+t_ro<-mod_stargazer(stargazer(est1,est2,est3,est4,est5,est6,est7,est8,est9, omit="electionyear",keep.stat=c("n","rsq"),dep.var.caption="",covariate.labels=c("Treatment Group * Second Period Election","Treatment Group","Second Period Election","No. Seats (log)","Mun. Population (log)","Mun. Territory (log)","Mun. Revenue (log)"),ci=FALSE, ci.level=0.95, digits=3, column.sep.width="-18pt", align=TRUE, dep.var.labels.include=TRUE, dep.var.labels=c("Part-Time Incumbents (\\%)","Full-Time Incumbents (\\%)","Candidates per Seat"),font.size="small",model.names=FALSE,header=FALSE,add.lines=c(Region,MuniType,LinearTrend),out.header=FALSE,notes.append=FALSE, notes.label="",notes=c(""),star.cutoffs=NA))
+
+### Layout
+
+Region <- list(c("Unit Type, Region Fixed Effects","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{No}}"))
+
+UnitType <- list(c("Unit Type Fixed Effects","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{No}}"))
+
+MuniType <- list(c("Municipality Fixed Effects","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{Yes}}"))
+
+LinearTrend <- list(c("Linear Time Trend","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}"))
+
+originallayout="D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3}"
+newlayout="D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3}| D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3}| D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3}"
+
+t_ro[11]<-"\\hline \\bigstrut "
+t_ro <-gsub(originallayout, newlayout, t_ro, fixed =TRUE)
+
+t_ro <-gsub("\\multicolumn{10}{r}{} \\\\ ","", t_ro, fixed =TRUE)
+
+sink(file="Main_AvgCands_MuniCluster.tex")
+cat(t_ro)
+sink()
+
+
+##### TABLE D5
+
+est1<-felm(cands_perc_bus~interactedtreatment + treatment + after+numberelected_log + electionyear|factor(regionid)+ factor(unit_type) |0 | oktmo+ electionyear,data=els, psdef=FALSE)
+
+est2<-felm(cands_perc_bus~interactedtreatment + treatment + after+numberelected_log+population_log+territory_log+income_log + electionyear| factor(regionid)+factor(unit_type) |0 | oktmo+ electionyear,data=els, psdef=FALSE)
+
+est3<-felm(cands_perc_bus~after + interactedtreatment + electionyear| factor(oktmo) |0 | oktmo+ electionyear,data=els, psdef=FALSE)
+
+est4<-felm(cands_perc_directors~interactedtreatment + treatment + after+numberelected_log + electionyear|factor(regionid)+ factor(unit_type) |0 | oktmo+ electionyear,data=els, psdef=FALSE)
+
+est5<-felm(cands_perc_directors~interactedtreatment + treatment + after+numberelected_log+population_log+territory_log+income_log + electionyear| factor(regionid)+factor(unit_type) |0 | oktmo+ electionyear,data=els, psdef=FALSE)
+
+est6<-felm(cands_perc_directors~after + interactedtreatment + electionyear| factor(oktmo)|0 | oktmo+ electionyear,data=els, psdef=FALSE)
+
+est7<-felm(cands_perc_entre~interactedtreatment + treatment + after+numberelected_log + electionyear|factor(regionid)+ factor(unit_type) |0 | oktmo+ electionyear,data=els, psdef=FALSE)
+
+est8<-felm(cands_perc_entre~interactedtreatment + treatment + after+numberelected_log+population_log+territory_log+income_log + electionyear| factor(regionid)+factor(unit_type) |0 | oktmo+ electionyear,data=els, psdef=FALSE)
+
+est9<-felm(cands_perc_entre~after + interactedtreatment + electionyear| factor(oktmo) |0 | oktmo+ electionyear,data=els, psdef=FALSE)
+
+t_ro<-mod_stargazer(stargazer(est7,est8,est9,est4,est5,est6,est1,est2,est3, omit="electionyear", keep.stat=c("n","rsq"),dep.var.caption="",covariate.labels=c("Treatment Group * Second Period Election","Treatment Group","Second Period Election","No. Seats (log)","Mun. Population (log)","Mun. Territory (log)","Mun. Revenue (log)"),ci=FALSE, ci.level=0.95, digits=3, column.sep.width="-18pt", align=TRUE, dep.var.labels.include=TRUE, dep.var.labels=c("All Businesspeople (\\%)","Firm Directors (\\%)","All Businesspeople (\\%)","Individual Entrepreneurs (\\%)"),font.size="small",model.names=FALSE,header=FALSE,add.lines=c(Region,MuniType,LinearTrend),out.header=FALSE,notes.append=FALSE, notes.label="",notes=c(""),star.cutoffs=NA))
+
+t_ro[11]<-"\\hline \\bigstrut "
+t_ro <-gsub(originallayout, newlayout, t_ro, fixed =TRUE)
+t_ro <-gsub("\\multicolumn{10}{r}{} \\\\ ","", t_ro, fixed =TRUE)
+
+sink(file="Main_Business_MuniCluster.tex")
+cat(t_ro)
+sink()
+
+
+
+##### TABLE D6
+
+est1<-felm(perc_elected_partial~interactedtreatment + treatment + after+ electionyear| factor(oktmo) |0 | regionid + electionyear,data=subset(els, reg_pressfreedom == 2))
+
+est2<-felm(perc_elected_partial~interactedtreatment + treatment + after + electionyear| factor(oktmo) |0 | regionid + electionyear,data=subset(els, reg_pressfreedom == 3))
+
+est3<-felm(perc_elected_partial~interactedtreatment + treatment + after + electionyear| factor(oktmo) |0 | regionid + electionyear,data=subset(els, reg_pressfreedom == 4))
+
+est4<-felm(perc_elected_full~interactedtreatment + treatment + after + electionyear| factor(oktmo) |0 | regionid + electionyear,data=subset(els, reg_pressfreedom == 2))
+
+est5<-felm(perc_elected_full~interactedtreatment + treatment + after + electionyear| factor(oktmo) |0 | regionid + electionyear,data=subset(els, reg_pressfreedom == 3))
+
+est6<-felm(perc_elected_full~interactedtreatment + treatment + after + electionyear| factor(oktmo) |0 | regionid + electionyear,data=subset(els, reg_pressfreedom == 4))
+
+est7<-felm(cands_perc_entre~interactedtreatment + treatment + after+population_log + electionyear| factor(oktmo) |0 | regionid + electionyear,data=subset(els, reg_pressfreedom == 2))
+
+est8<-felm(cands_perc_entre~interactedtreatment + treatment + after + electionyear| factor(oktmo) |0 | regionid + electionyear,data=subset(els, reg_pressfreedom == 3))
+
+est9<-felm(cands_perc_entre~interactedtreatment + treatment + after + electionyear| factor(oktmo) |0 | regionid + electionyear,data=subset(els, reg_pressfreedom == 4))
+
+est10<-felm(cands_perc_directors~interactedtreatment + treatment + after + electionyear| factor(oktmo) |0 | regionid + electionyear,data=subset(els, reg_pressfreedom == 2))
+
+est11<-felm(cands_perc_directors~interactedtreatment + treatment + after + electionyear| factor(oktmo) |0 | regionid + electionyear,data=subset(els, reg_pressfreedom == 3))
+
+est12<-felm(cands_perc_directors~interactedtreatment + treatment + after + electionyear| factor(oktmo) |0 | regionid + electionyear,data=subset(els, reg_pressfreedom == 4))
+
+t_ro<-mod_stargazer(stargazer(est1,est2,est3,est4,est5,est6,est7,est8,est9,est10,est11,est12, keep=c("interactedtreatment","after","population_log"),keep.stat=c("n","rsq"),dep.var.caption="",covariate.labels=c("Treatment Group * Second Period Election","Second Period Election","Mun. Population (log)"),ci=FALSE, ci.level=0.95, digits=3, column.sep.width="-15pt", align=TRUE,dep.var.labels.include = FALSE, column.labels = c("Low","Medium","High","Low","Medium","High","Low","Medium","High","Low","Medium","High"),font.size="small",model.names=FALSE,header=FALSE,add.lines=c(Region,UnitType),out.header=FALSE,notes.append=FALSE, notes.label="",notes=c(""),star.cutoffs=NA))
+
+originallayout="D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3}"
+newlayout="D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3}| D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3}|| D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3}| D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3}"
+
+
+Region <- list(c("Unit Type, Region Fixed Effects","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}"))
+
+UnitType <- list(c("Linear Time Trends","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}"))
+
+t_ro[11]<-"\\hline \\bigstrut "
+t_ro <-gsub(originallayout, newlayout, t_ro, fixed =TRUE)
+t_ro <-gsub("\\multicolumn{7}{r}{} \\\\ ","", t_ro, fixed =TRUE)
+
+t_ro[7]<-" \\textbf{Outcome:}& \\multicolumn{3}{c}{Part-Time Incumbents (\\%)} & \\multicolumn{3}{c}{Full-Time Incumbents (\\%)} & \\multicolumn{3}{c}{Independent Entrepreneurs (\\%)} & \\multicolumn{3}{c}{Firm Directors (\\%)} \\\\ \\cmidrule(l{15pt}r{15pt}){2-4}\\cmidrule(l{15pt}r{15pt}){5-7}\\cmidrule(l{15pt}r{15pt}){8-10}\\cmidrule(l{15pt}r{15pt}){11-13}\\\\"
+t_ro[9]<-paste("\\textbf{Level of Press Freedom:}",t_ro[9],sep="")
+sink(file="Terciles_PressFreedom.tex")
+cat(t_ro)
+sink()
+
+
+
+
+###### TABLE D7
+
+est1<-felm(perc_elected_partial~interactedtreatment + after + electionyear| factor(oktmo) |0 | regionid + electionyear,data=subset(els, fn_budget_log <= 13.64463))
+
+est2<-felm(perc_elected_partial~interactedtreatment + after + electionyear| factor(oktmo) |0 | regionid + electionyear,data=subset(els, fn_budget_log > 13.64463 & fn_budget_log<13.95877))
+
+est3<-felm(perc_elected_partial~interactedtreatment + after + electionyear| factor(oktmo) |0 | regionid + electionyear,data=subset(els, fn_budget_log>=13.95877))
+
+est4<-felm(perc_elected_full~interactedtreatment + after + electionyear| factor(oktmo) |0 | regionid + electionyear,data=subset(els, fn_budget_log <= 13.64463))
+
+est5<-felm(perc_elected_full~interactedtreatment + after + electionyear| factor(oktmo) |0 | regionid + electionyear,data=subset(els, fn_budget_log > 13.64463 & fn_budget_log<13.95877))
+
+est6<-felm(perc_elected_full~interactedtreatment + after + electionyear| factor(oktmo) |0 | regionid + electionyear,data=subset(els, fn_budget_log>=13.95877))
+
+est7<-felm(cands_perc_entre~interactedtreatment + after + electionyear| factor(oktmo) |0 | regionid + electionyear,data=subset(els, fn_budget_log <= 13.64463))
+
+est8<-felm(cands_perc_entre~interactedtreatment + after + electionyear| factor(oktmo) |0 | regionid + electionyear,data=subset(els, fn_budget_log > 13.64463 & fn_budget_log<13.95877))
+
+est9<-felm(cands_perc_entre~interactedtreatment + after + electionyear| factor(oktmo) |0 | regionid + electionyear,data=subset(els, fn_budget_log>=13.95877))
+
+est10<-felm(cands_perc_directors~interactedtreatment + after + electionyear| factor(oktmo) |0 | regionid + electionyear,data=subset(els, fn_budget_log <= 13.64463))
+
+est11<-felm(cands_perc_directors~interactedtreatment + after + electionyear| factor(oktmo) |0 | regionid + electionyear,data=subset(els, fn_budget_log > 13.64463 & fn_budget_log<13.95877))
+
+est12<-felm(cands_perc_directors~interactedtreatment + after + electionyear| factor(oktmo) |0 | regionid + electionyear,data=subset(els, fn_budget_log>=13.95877))
+
+Region <- list(c("Unit Type, Region Fixed Effects","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}"))
+
+UnitType <- list(c("Linear Time Trends","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}"))
+
+originallayout="D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3}"
+newlayout="D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3}| D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3}|| D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3}| D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3}"
+
+t_ro<-mod_stargazer(stargazer(est1,est2,est3,est4,est5,est6,est7,est8,est9,est10,est11,est12, keep=c("interactedtreatment","after","population_log"),keep.stat=c("n","rsq"),dep.var.caption="",covariate.labels=c("Treatment Group * Second Period Election","Second Period Election","Mun. Population (log)"),ci=FALSE, ci.level=0.95, digits=3, column.sep.width="-15pt", align=TRUE,dep.var.labels.include = FALSE, column.labels = c("Low","Medium","High","Low","Medium","High","Low","Medium","High","Low","Medium","High"),font.size="small",model.names=FALSE,header=FALSE,add.lines=c(Region,UnitType),out.header=FALSE,notes.append=FALSE, notes.label="",notes=c(""),star.cutoffs=NA))
+
+t_ro[11]<-"\\hline \\bigstrut "
+t_ro <-gsub(originallayout, newlayout, t_ro, fixed =TRUE)
+t_ro <-gsub("\\multicolumn{7}{r}{} \\\\ ","", t_ro, fixed =TRUE)
+
+t_ro[7]<-" \\textbf{Outcome:}& \\multicolumn{3}{c}{Part-Time Incumbents (\\%)} & \\multicolumn{3}{c}{Full-Time Incumbents (\\%)} & \\multicolumn{3}{c}{Independent Entrepreneurs (\\%)} & \\multicolumn{3}{c}{Firm Directors (\\%)} \\\\ \\cmidrule(l{15pt}r{15pt}){2-4}\\cmidrule(l{15pt}r{15pt}){5-7}\\cmidrule(l{15pt}r{15pt}){8-10}\\cmidrule(l{15pt}r{15pt}){11-13}\\\\"
+t_ro[9]<-paste("\\textbf{Law Enforcement Capacity:}",t_ro[9],sep="")
+sink(file="Terciles_AuditRisk.tex")
+cat(t_ro)
+sink()
+
+
+
+
+##### TABLE D8
+
+est1<-felm(perc_elected_partial~treatment*after*kgiscore+electionyear+ lngdp + log_pop  + resource_grppct + reg_urbanshare  + log_mincome + reg_sharepensm | factor(oktmo) |0 | regionid + electionyear,data=els,psdef=FALSE)
+
+est2<-lmer(perc_elected_partial~treatment*after*kgiscore+electionyear+ lngdp + log_pop  + resource_grppct + reg_urbanshare  + log_mincome + reg_sharepensm+factor(unit_type) + (1|regionid),data=els)
+
+est3<-felm(perc_elected_partial~treatment*after*audits_allpeople+electionyear+ lngdp + log_pop  + resource_grppct + reg_urbanshare  + log_mincome + reg_sharepensm | factor(oktmo) |0 | regionid + electionyear,data=els,psdef=FALSE)
+
+est4<-lmer(perc_elected_partial~treatment*after*audits_allpeople+electionyear+ lngdp + log_pop  + resource_grppct + reg_urbanshare  + log_mincome + reg_sharepensm+factor(unit_type) + (1|regionid),data=els)
+
+est5<-felm(perc_elected_partial~treatment*after*ENFORCE+electionyear+ lngdp + log_pop  + resource_grppct + reg_urbanshare  + log_mincome + reg_sharepensm | factor(oktmo) |0 | regionid + electionyear,data=els,psdef=FALSE)
+
+est6<-lmer(perc_elected_partial~treatment*after*ENFORCE+electionyear+ lngdp + log_pop  + resource_grppct + reg_urbanshare  + log_mincome + reg_sharepensm+factor(unit_type) + (1|regionid),data=els)
+
+est7<-felm(cands_perc_entre~treatment*after*kgiscore+electionyear+ lngdp + log_pop  + resource_grppct + reg_urbanshare  + log_mincome + reg_sharepensm | factor(oktmo) |0 | regionid + electionyear,data=els,psdef=FALSE)
+
+est8<-lmer(cands_perc_entre~treatment*after*kgiscore+electionyear+ lngdp + log_pop  + resource_grppct + reg_urbanshare  + log_mincome + reg_sharepensm+factor(unit_type) + (1|regionid),data=els)
+
+est9<-felm(cands_perc_entre~treatment*after*audits_allpeople+electionyear+ lngdp + log_pop  + resource_grppct + reg_urbanshare  + log_mincome + reg_sharepensm | factor(oktmo) |0 | regionid + electionyear,data=els,psdef=FALSE)
+
+est10<-lmer(cands_perc_entre~treatment*after*audits_allpeople+electionyear+ lngdp + log_pop  + resource_grppct + reg_urbanshare  + log_mincome + reg_sharepensm+factor(unit_type) + (1|regionid),data=els)
+
+est11<-felm(cands_perc_entre~treatment*after*ENFORCE+electionyear+ lngdp + log_pop  + resource_grppct + reg_urbanshare  + log_mincome + reg_sharepensm | factor(oktmo) |0 | regionid + electionyear,data=els,psdef=FALSE)
+
+est12<-lmer(cands_perc_entre~treatment*after*ENFORCE+electionyear+ lngdp + log_pop  + resource_grppct + reg_urbanshare  + log_mincome + reg_sharepensm+factor(unit_type) + (1|regionid),data=els)
+
+Muni <- list(c("Regional Covariates","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}"))
+
+Region <- list(c("Municipality FE; Linear Time Trend","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{No}}"))
+
+MLM <- list(c("Unit Type FE; Region RE","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{Yes}}"))
+
+
+t_ro<-mod_stargazer(stargazer(est1,est2,est3,est4,est5,est6,est7,est8,est9,est10,est11,est12, keep.stat=c("n","rsq"),dep.var.caption="",keep=c("treatment","after","treatment:after:kgiscore","treatment:after:fn_budget_log" ,"treatment:after:log_justice_salary"),
+                              covariate.labels=c("Treatment Group","Second Election","
+                                                 Treatment Group * Second Election",
+                                                 "Treatment Group * KGI Score",
+                                                 "Second Election * KGI Score",
+                                                 "\\textbf{Treatment Group * Second Election * KGI Score}",
+                                                 "Treatment Group * Audit Risk",
+                                                 "Second Election * Audit Risk",
+                                                 "\\textbf{Treatment Group * Second Election * Audit Risk}",
+                                                 
+                                                 "Treatment Group * Enforcement Exp.",
+                                                 "Second Election *Enforcement Exp. (log)",
+                                                 "\\textbf{Treatment Group * Second Election * Enforcement Exp.}" ),ci=FALSE, ci.level=0.95, digits=3, column.sep.width="-20pt", align=TRUE, dep.var.labels.include=TRUE, dep.var.labels=c("Part-Time Incumbents (\\%)","Independent Entrepreneurs (\\%)"),font.size="small",model.names=FALSE,header=FALSE,add.lines=c(Muni,Region,MLM),out.header=FALSE,notes.append=FALSE, notes.label="",notes=c(""),star.cutoffs=NA))
+
+t_ro[9]<-paste(t_ro[9],"\\cmidrule(l{15pt}r{15pt}){2-7}\\cmidrule(l{15pt}r{15pt}){8-13}\\\\",sep="")
+
+t_ro <-gsub("\\multicolumn{13}{r}{} \\\\ ","", t_ro, fixed =TRUE)
+
+t_ro <-gsub("(0.000)","", t_ro, fixed =TRUE)
+
+sink(file="Heterogeneity_EnforcementRobustness.tex")
+cat(t_ro)
+sink()
+
+
+
+
+##### TABLE D9
+
+est1<-lmer(perc_elected_partial~treatment*after*reg_pressfreedom+electionyear+ lngdp + log_pop  + resource_grppct + reg_urbanshare  + log_mincome + reg_sharepensm+factor(unit_type) + (1|regionid),data=els)
+
+est2<-lmer(perc_elected_partial~treatment*after*reg_dem_media+electionyear+ lngdp + log_pop  + resource_grppct + reg_urbanshare  + log_mincome + reg_sharepensm+factor(unit_type) + (1|regionid),data=els)
+
+est3<-lmer(perc_elected_partial~treatment*after*fn_budget_log+electionyear+ lngdp + log_pop  + resource_grppct + reg_urbanshare  + log_mincome + reg_sharepensm+factor(unit_type) + (1|regionid),data=els)
+
+est4<-lmer(perc_elected_partial~treatment*after*log_justice+electionyear+ lngdp + log_pop  + resource_grppct + reg_urbanshare  + log_mincome + reg_sharepensm+factor(unit_type) + (1|regionid),data=els)
+
+est5<-lmer(cands_perc_entre~treatment*after*reg_pressfreedom+electionyear+ lngdp + log_pop  + resource_grppct + reg_urbanshare  + log_mincome + reg_sharepensm+factor(unit_type) + (1|regionid),data=els)
+
+est6<-lmer(cands_perc_entre~treatment*after*reg_dem_media+electionyear+ lngdp + log_pop  + resource_grppct + reg_urbanshare  + log_mincome + reg_sharepensm+factor(unit_type) + (1|regionid),data=els)
+
+est7<-lmer(cands_perc_entre~treatment*after*fn_budget_log+electionyear+ lngdp + log_pop  + resource_grppct + reg_urbanshare  + log_mincome + reg_sharepensm+factor(unit_type) + (1|regionid),data=els)
+
+est8<-lmer(cands_perc_entre~treatment*after*log_justice+electionyear+ lngdp + log_pop  + resource_grppct + reg_urbanshare  + log_mincome + reg_sharepensm+factor(unit_type) + (1|regionid),data=els)
+
+
+Muni <- list(c("Regional Covariates, Linear Time Trend","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}"))
+
+Region <- list(c("Unit Type FE; Region RE","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}"))
+
+t_ro<-mod_stargazer(stargazer(est1,est2,est3,est4,est5,est6,est7,est8, keep.stat=c("n","rsq"),dep.var.caption="",keep=c("treatment","after","treatment:after:reg_pressfreedom","treatment:after:reg_dem_media" ,"treatment:after:fn_budget_log","treatment:after:log_justice"),
+                              covariate.labels=c("Treatment Group","Second Election","Treatment Group * Second Election","Treatment Group * GDF Press Freedom",
+       "Second Election * GDF Press Freedom",     
+       "\\textbf{Treatment Group * Second Election * GDF Press Freedom}",
+        "Treatment Group * TP Press Freedom",
+        "Second Election * TP Press Freedom",     
+        "\\textbf{Treatment Group * Second Election * TP Press Freedom}",
+        "Treatment Group * Regional Tax Agency Budget",
+        "Second Election * Regional Tax Agency Budget",
+        "\\textbf{Treatment Group * Second Election * Regional Tax Agency Budget}",
+         "Treatment Group * Law Enforcement Personnel",
+         "Second Election * Law Enforcement Personnel",
+          "\\textbf{Treatment Group * Second Election * Law Enforcement Personnel}"),ci=FALSE, ci.level=0.95, digits=3, column.sep.width="-20pt", align=TRUE, dep.var.labels.include=TRUE, dep.var.labels=c("Part-Time Incumbents (\\%)","Independent Entrepreneurs (\\%)"),font.size="small",model.names=FALSE,header=FALSE,add.lines=c(Muni,Region),out.header=FALSE,notes.append=FALSE, notes.label="",notes=c(""),star.cutoffs=NA))
+
+t_ro[9]<-paste(t_ro[9],"\\cmidrule(l{15pt}r{15pt}){2-5}\\cmidrule(l{15pt}r{15pt}){6-9}\\\\",sep="")
+
+t_ro <-gsub("\\multicolumn{9}{r}{} \\\\ ","", t_ro, fixed =TRUE)
+
+
+sink(file="MultilevelModels.tex")
+cat(t_ro)
+sink()
+
+
+
+
+###### TABLE D10
+
+els[,minlevelincumbency:=min(perc_elected_partial),by="oktmo"]
+
+est1<-felm(perc_elected_partial~interactedtreatment + after + electionyear | factor(oktmo) |0 | regionid + electionyear,data=subset(els), psdef=FALSE)
+
+est2<-felm(perc_elected_partial~interactedtreatment + after + electionyear | factor(oktmo) |0 | regionid + electionyear,data=subset(els,minlevelincumbency>0), psdef=FALSE)
+
+est3<-felm(perc_elected_partial~interactedtreatment + after + electionyear | factor(oktmo) |0 | regionid + electionyear,data=subset(els,minlevelincumbency>0.1), psdef=FALSE)
+
+est4<-felm(perc_elected_partial~interactedtreatment + after + electionyear | factor(oktmo) |0 | regionid + electionyear,data=subset(els,minlevelincumbency>0.2), psdef=FALSE)
+
+est5<-felm(perc_elected_full~interactedtreatment + after + electionyear | factor(oktmo) |0 | regionid + electionyear,data=subset(els), psdef=FALSE)
+
+est6<-felm(perc_elected_full~interactedtreatment + after + electionyear | factor(oktmo) |0 | regionid + electionyear,data=subset(els,minlevelincumbency>0), psdef=FALSE)
+
+est7<-felm(perc_elected_full~interactedtreatment + after + electionyear | factor(oktmo) |0 | regionid + electionyear,data=subset(els,minlevelincumbency>0.1), psdef=FALSE)
+
+est8<-felm(perc_elected_full~interactedtreatment + after + electionyear | factor(oktmo) |0 | regionid + electionyear,data=subset(els,minlevelincumbency>0.2), psdef=FALSE)
+
+MuniType <- list(c("Municipality Fixed Effects","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}"))
+
+LinearTrend <- list(c("Linear Time Trend","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}"))
+
+MinIncumbency <- list(c("Minimum Incumbent Constraint","\\multicolumn{1}{c}{\\text{None}}","\\multicolumn{1}{c}{\\text{0\\%}}","\\multicolumn{1}{c}{\\text{10\\%}}","\\multicolumn{1}{c}{\\text{20\\%}}","\\multicolumn{1}{c}{\\text{None}}","\\multicolumn{1}{c}{\\text{0\\%}}","\\multicolumn{1}{c}{\\text{10\\%}}","\\multicolumn{1}{c}{\\text{20\\%}}"))
+
+t_ro<-mod_stargazer(stargazer(est1,est2,est3,est4,est5,est6,est7,est8, omit="electionyear",keep.stat=c("n","rsq"),dep.var.caption="",covariate.labels=c("Treatment Group * Second Period Election","Treatment Group","Second Period Election"),ci=FALSE, ci.level=0.95, digits=3, column.sep.width="-5pt", align=TRUE, dep.var.labels.include=TRUE, dep.var.labels=c("Part-Time Incumbents (\\%)","Full-Time Incumbents (\\%)"),font.size="small",model.names=FALSE,header=FALSE,add.lines=c(MinIncumbency,MuniType,LinearTrend),out.header=FALSE,notes.append=FALSE, notes.label="",notes=c(""),star.cutoffs = NA))
+
+t_ro[11]<-"\\hline \\bigstrut "
+
+originallayout="D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3}"
+newlayout="D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3}| D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3} D{.}{.}{-3}"
+t_ro <-gsub(originallayout, newlayout, t_ro, fixed =TRUE)
+t_ro <-gsub("\\multicolumn{9}{r}{} \\\\ ","", t_ro, fixed =TRUE)
+
+sink(file="Appendix_IncumbencyMinimum.tex")
+cat(t_ro)
+sink()
+
+
+
+###### TABLE D11
+
+#### Do this because Stargazer isn't great at removing certain variables
+els[,treatment2:=treatment]
+
+MuniType <- list(c("Municipality Fixed Effects","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}"))
+
+LinearTrend <- list(c("Linear Time Trend","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}"))
+
+est1<-felm(perc_elected_partial~interactedtreatment + after+electionyear | factor(oktmo) |0 | regionid + electionyear,data=els,psdef=FALSE)
+
+est2<-felm(perc_elected_partial~interactedtreatment + after+electionyear | factor(oktmo) |0 | regionid + electionyear,data=subset(els, LowURSeats==0),psdef=FALSE)
+
+est3<-felm(perc_elected_partial~interactedtreatment + after+electionyear | factor(oktmo) |0 | regionid + electionyear,data=subset(els, LowURSeats==1),psdef=FALSE)
+
+est4<-felm(perc_elected_partial~interactedtreatment + after+electionyear | factor(oktmo) |0 | regionid + electionyear,data=subset(els, republic==1),psdef=FALSE)
+
+est5<-felm(perc_elected_partial~interactedtreatment + after+electionyear | factor(oktmo) |0 | regionid + electionyear,data=subset(els, republic==0),psdef=FALSE)
+
+est6<-felm(perc_elected_partial~interactedtreatment + after+electionyear | factor(oktmo) |0 | regionid + electionyear,data=subset(els, FarCrimea == 1),psdef=FALSE)
+
+t_ro<-mod_stargazer(stargazer(est1,est2,est3,est4,est5,est6, omit=c("electionyear","treatment2"), keep.stat=c("n","rsq"),dep.var.caption="",covariate.labels=c("Treatment Group * Second Period Election","Second Period Election"),ci=FALSE, ci.level=0.95, digits=3, column.sep.width="5pt", align=TRUE, dep.var.labels.include=TRUE,column.labels = c("Full Sample","Low UR Seats","High UR Seats","Ethnic Republic","Not Ethnic Republic","No 2017"), column.separate = c(1,1,1,1,1), dep.var.labels=c("Part-Time Incumbents (\\%)"),font.size="small",model.names=FALSE,header=FALSE,add.lines=c(MuniType,LinearTrend),out.header=FALSE,notes.append=FALSE, notes.label="",notes=c(""),star.cutoffs=NA))
+
+t_ro <-gsub("\\multicolumn{6}{r}{} \\\\ ","", t_ro, fixed =TRUE)
+
+sink(file="Party_Partial.tex")
+cat(t_ro)
+sink()
+
+
+####### TABLE D12
+
+
+est1<-felm(cands_perc_entre~interactedtreatment + after+electionyear | factor(oktmo) |0 | regionid + electionyear,data=els,psdef=FALSE)
+
+est2<-felm(cands_perc_entre~interactedtreatment + after+electionyear | factor(oktmo) |0 | regionid + electionyear,data=subset(els, LowURSeats==0),psdef=FALSE)
+
+est3<-felm(cands_perc_entre~interactedtreatment + after+electionyear | factor(oktmo) |0 | regionid + electionyear,data=subset(els, LowURSeats==1),psdef=FALSE)
+
+est4<-felm(cands_perc_entre~interactedtreatment + after+electionyear | factor(oktmo) |0 | regionid + electionyear,data=subset(els, republic==1),psdef=FALSE)
+
+est5<-felm(cands_perc_entre~interactedtreatment + after+electionyear | factor(oktmo) |0 | regionid + electionyear,data=subset(els, republic==0),psdef=FALSE)
+
+est6<-felm(cands_perc_entre~interactedtreatment + after+electionyear | factor(oktmo) |0 | regionid + electionyear,data=subset(els, FarCrimea == 1),psdef=FALSE)
+
+
+t_ro<-mod_stargazer(stargazer(est1,est2,est3,est4,est5,est6, omit=c("electionyear","treatment2"), keep.stat=c("n","rsq"),dep.var.caption="",covariate.labels=c("Treatment Group * Second Period Election","Second Period Election"),ci=FALSE, ci.level=0.95, digits=3, column.sep.width="5pt", align=TRUE, dep.var.labels.include=TRUE,column.labels = c("Full Sample","Low UR Seats","High UR Seats","Ethnic Republic","Not Ethnic Republic","No 2017"), column.separate = c(1,1,1,1,1), dep.var.labels=c("Individual Entrepreneurs (\\%)"),font.size="small",model.names=FALSE,header=FALSE,add.lines=c(MuniType,LinearTrend),out.header=FALSE,notes.append=FALSE, notes.label="",notes=c(""),star.cutoffs=NA))
+
+t_ro <-gsub("\\multicolumn{6}{r}{} \\\\ ","", t_ro, fixed =TRUE)
+
+sink(file="Party_Entrepreneur.tex")
+cat(t_ro)
+sink()
+
+
+
+
+
+####### TABLE E1
+
+est1<-felm(partialname_per_seat~interactedtreatment + treatment + after+numberelected_log + electionyear|factor(regionid)+ factor(unit_type) |0 | regionid + electionyear,data=els, psdef=FALSE)
+
+est2<-felm(partialname_per_seat~interactedtreatment + treatment + after+numberelected_log+population_log+territory_log+income_log + electionyear| factor(regionid)+factor(unit_type) |0 | regionid + electionyear,data=els, psdef=FALSE)
+
+est3<-felm(partialname_per_seat~after + interactedtreatment + electionyear| factor(oktmo) |0 | regionid + electionyear,data=els, psdef=FALSE)
+
+est4<-felm(business_per_seat~interactedtreatment + treatment + after+numberelected_log + electionyear|factor(regionid)+ factor(unit_type) |0 | regionid + electionyear,data=els, psdef=FALSE)
+
+est5<-felm(business_per_seat~interactedtreatment + treatment + after+numberelected_log+population_log+territory_log+income_log + electionyear| factor(regionid)+factor(unit_type) |0 | regionid + electionyear,data=els, psdef=FALSE)
+
+est6<-felm(business_per_seat~after + interactedtreatment + electionyear| factor(oktmo) |0 | regionid + electionyear,data=els, psdef=FALSE)
+
+est7<-felm(urwins_per_seat~interactedtreatment + treatment + after+numberelected_log + electionyear|factor(regionid)+ factor(unit_type) |0 | regionid + electionyear,data=els, psdef=FALSE)
+
+est8<-felm(urwins_per_seat~interactedtreatment + treatment + after+numberelected_log+population_log+territory_log+income_log + electionyear| factor(regionid)+factor(unit_type) |0 | regionid + electionyear,data=els, psdef=FALSE)
+
+est9<-felm(urwins_per_seat~after + interactedtreatment + electionyear| factor(oktmo) |0 | regionid + electionyear,data=els, psdef=FALSE)
+
+est10<-felm(incumbent_success_rate~interactedtreatment + treatment + after+numberelected_log + electionyear|factor(regionid)+ factor(unit_type) |0 | regionid + electionyear,data=els, psdef=FALSE)
+
+est11<-felm(incumbent_success_rate~interactedtreatment + treatment + after+numberelected_log+population_log+territory_log+income_log + electionyear| factor(regionid)+factor(unit_type) |0 | regionid + electionyear,data=els, psdef=FALSE)
+
+est12<-felm(incumbent_success_rate~after + interactedtreatment + electionyear| factor(oktmo) |0 | regionid + electionyear,data=els, psdef=FALSE)
+
+t_ro<-mod_stargazer(stargazer(est1,est2,est3,est4,est5,est6,est7,est8,est9,est10,est11,est12, omit="electionyear", keep.stat=c("n","rsq"),dep.var.caption="",covariate.labels=c("Treatment Group * Second Period Election","Treatment Group","Second Period Election","No. Seats (log)","Mun. Population (log)","Mun. Territory (log)","Mun. Revenue (log)"),ci=FALSE, ci.level=0.95, digits=3, column.sep.width="-18pt", align=TRUE, dep.var.labels.include=TRUE, dep.var.labels=c("Part-time Inc. Winners (\\%)","Bus. Winners (\\%)","UR Winners (\\%)","Inc. Success (\\%)"),font.size="small",model.names=FALSE,header=FALSE,add.lines=c(UnitType,Region,MuniType,LinearTrend),out.header=FALSE,notes.append=FALSE, notes.label="",notes=c(""),star.cutoffs=NA))
+
+t_ro[11]<-"\\hline \\bigstrut "
+t_ro <-gsub(originallayout, newlayout, t_ro, fixed =TRUE)
+t_ro <-gsub("\\multicolumn{13}{r}{} \\\\ ","", t_ro, fixed =TRUE)
+
+sink(file="Main_Winners.tex")
+cat(t_ro)
+sink()
+
+
+####### TABLE E2
+
+est1<-felm(age~interactedtreatment + treatment + after+numberelected_log + electionyear|factor(regionid)+ factor(unit_type) |0 | regionid + electionyear,data=els, psdef=FALSE)
+
+est2<-felm(age~interactedtreatment + treatment + after+numberelected_log+population_log+territory_log+income_log + electionyear| factor(regionid)+factor(unit_type) |0 | regionid + electionyear,data=els, psdef=FALSE)
+
+est3<-felm(age~after + interactedtreatment + electionyear| factor(oktmo) |0 | regionid + electionyear,data=els, psdef=FALSE)
+
+est4<-felm(female~interactedtreatment + treatment + after+numberelected_log + electionyear|factor(regionid)+ factor(unit_type) |0 | regionid + electionyear,data=els, psdef=FALSE)
+
+est5<-felm(female~interactedtreatment + treatment + after+numberelected_log+population_log+territory_log+income_log + electionyear| factor(regionid)+factor(unit_type) |0 | regionid + electionyear,data=els, psdef=FALSE)
+
+est6<-felm(female~after + interactedtreatment + electionyear| factor(oktmo) |0 | regionid + electionyear,data=els, psdef=FALSE)
+
+t_ro<-mod_stargazer(stargazer(est1,est2,est3,est4,est5,est6, omit="electionyear", keep.stat=c("n","rsq"),dep.var.caption="",covariate.labels=c("Treatment Group * Second Period Election","Treatment Group","Second Period Election","No. Seats (log)","Mun. Population (log)","Mun. Territory (log)","Mun. Revenue (log)"),ci=FALSE, ci.level=0.95, digits=3, column.sep.width="-18pt", align=TRUE, dep.var.labels.include=TRUE, dep.var.labels=c("Mean Age","Female (\\%)","Education Level"),font.size="small",model.names=FALSE,header=FALSE,add.lines=c(UnitType,Region,MuniType,LinearTrend),out.header=FALSE,notes.append=FALSE, notes.label="",notes=c(""),star.cutoffs=NA))
+
+t_ro[11]<-"\\hline \\bigstrut "
+t_ro <-gsub(originallayout, newlayout, t_ro, fixed =TRUE)
+t_ro <-gsub("\\multicolumn{7}{r}{} \\\\ ","", t_ro, fixed =TRUE)
+
+sink(file="Main_Demo.tex")
+cat(t_ro)
+sink()
+
+
+
+###### TABLE E3
+
+est1<-felm(cands_perc_official~after + interactedtreatment+electionyear| factor(oktmo) |0 | regionid,data=els)
+
+est2<-felm(cands_perc_doctor~after + interactedtreatment+electionyear| factor(oktmo) |0 | regionid,data=els)
+
+est3<-felm(cands_perc_teacher~after + interactedtreatment+electionyear| factor(oktmo) |0 | regionid,data=els)
+
+est4<-felm(cands_perc_force~after + interactedtreatment+electionyear| factor(oktmo) |0 | regionid,data=els)
+
+est5<-felm(cands_perc_accountant~after + interactedtreatment+electionyear| factor(oktmo) |0 | regionid,data=els)
+
+est6<-felm(cands_perc_ngo~after + interactedtreatment+electionyear| factor(oktmo) |0 | regionid,data=els)
+
+est7<-felm(cands_perc_lowerclass~after + interactedtreatment+electionyear| factor(oktmo) |0 | regionid,data=els)
+
+est8<-felm(cands_perc_notwork~after + interactedtreatment+electionyear| factor(oktmo) |0 | regionid,data=els)
+
+Region <- list(c("Region Fixed Effects","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{No}}"))
+
+UnitType <- list(c("Unit Type Fixed Effects","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{No}}","\\multicolumn{1}{c}{\\text{No}}"))
+
+MuniType <- list(c("Municipality Fixed Effects","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}","\\multicolumn{1}{c}{\\text{Yes}}"))
+
+t_ro<-mod_stargazer(stargazer(est1,est2,est3,est4,est5,est6,est7,est8,omit="electionyear",keep.stat=c("n","rsq"),dep.var.caption="",covariate.labels=c("Treatment Group * Second Period Election","Second Period Election"),order=c("interactedtreatment","after"),ci=FALSE, ci.level=0.95, digits=3, column.sep.width="-5pt", align=TRUE, dep.var.labels.include=TRUE, dep.var.labels=c(
+  "Government (\\%)",
+  "Health Care (\\%)",
+  "Education (\\%)",
+  "Law Enforcement (\\%)",
+  "Professional (\\%)",
+  "Civil Society (\\%)",
+  "Blue Collar (\\%)",
+  "Unemployed (\\%)"),font.size="small",model.names=FALSE,header=FALSE,add.lines=c(UnitType,Region,MuniType,LinearTrend),out.header=FALSE,notes.append=FALSE, notes.label="",notes=c(""),star.cutoffs = NA))
+
+t_ro <-gsub("\\multicolumn{9}{r}{} \\\\ ","", t_ro, fixed =TRUE)
+
+sink(file="ProfessionsDiD.tex")
+cat(t_ro)
+sink()
+my_log <- file("my_log.txt")
+sink(my_log, append = TRUE, type = "output")
+sink(my_log, append=TRUE, type="message")
+
+con <- file("test.log")
+sink(con, append=TRUE)

38/replication_package/cands.Rda

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:db4b934d9b6acb79b192ab9aab0cc4da8e546ebb9bc855b30791e6907d59b16e
+size 23356019

38/replication_package/els.Rda

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a7b5972081d468fe0e5cec1e64e0bb18eabc313715f7c56c5d488b4ee22855f0
+size 1466102

38/should_reproduce.txt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bdd6a0d3fa3c58213acec4b2949638f45635114bb4a10cecec2ecb3b63853c84
+size 15