Do sunk costs affect prices in the housing market?
Dimuthu Ratnadiwakara and Vijay Yerramilli
30 Dec 2020
rm(list=ls())
library(stargazer)
library(plyr)
library(lfe)
library(zoo)
library(scales)
library(rdrobust)
library(data.table)
library(ggplot2)
library(dplyr)
library(fst)
library(fixest)zdata[,Selling_premium:=zdata$sales_price/zdata$adj_purch_price]
zdata[,Renter_fraction:=zdata$renters/zdata$totalpopulation]
zdata[,Purchase_price:=SalesPriceAmount_prev]
zdata[,Listing_price:=listing_amount]
zdata[,Selling_price:=sales_price]
zdata[,Listing_premium:=list_premium]
zdata[,Property_tax_last_year:=prop_tax_prev_year]
zdata[,Property_tax_rate_last_year:=prop_tax_rate]
zdata_reg <- zdata[PurchaseYear<2008 & ListingYear>2013 & house_age>0 ]
zdata_reg[,Time_on_market:=as.numeric(sale_date)-as.numeric(listed_date)+1]
zdata_reg[,zip_list_month:=paste(zip,ListingMonth)]
zdata_reg[,tract_list_month:=paste(GEOID,ListingMonth)]
zdata_reg[,zip3:=floor(zip/100)]
setorder(zdata_reg,listed_date)zdata_reg[,predprice:= 0]
zd <- NULL
listmonths <- unique(zdata_reg$ListingMonth)
# pb <- txtProgressBar(min=1, max=length(listmonths), style=3)
for(i in 1:length(listmonths)) {
# setTxtProgressBar(pb, i)
temp <- CA_all[Sale_month < listmonths[i] & Sale_month> as.Date(paste0((year(listmonths[i])-1),"-",month(listmonths[i]),"-01")) & hpi_inflation>0 & sqft>0 & house_age>0 & !is.na(Purchase_price) & !is.na(hpi_inflation) & !is.na(beds) & !is.na(baths) & !is.na(sqft) & !is.na(house_age) & !is.na(yr) & !is.na(sales_price) & sales_price>0 & LotSizeSquareFeet>1000 & LotSizeSquareFeet<500000 & sales_price<2000000 & sales_price>100000 & adj_purch_price<2000000 & adj_purch_price>100000 & abs(sales_price-adj_purch_price) < 100000 ]
temp[,adjbin:=ntile(adj_purch_price,3)]
adjbinmaxmin <- temp[,.(minap=min(adj_purch_price),maxap=max(adj_purch_price)),by=adjbin]
t <- zdata_reg[ListingMonth==listmonths[i]]
t[,adjbin:=ifelse(adj_purch_price < adjbinmaxmin[adjbin==1]$maxap,1,
ifelse(adj_purch_price < adjbinmaxmin[adjbin==2]$maxap,2,3))]
# ifelse(adj_purch_price < adjbinmaxmin[adjbin==3]$maxap,3,
# ifelse(adj_purch_price < adjbinmaxmin[adjbin==4]$maxap,4,5))))]
if(nrow(temp)<30) {
t[,predprice:=adj_purch_price]
} else {
hedonicmodel <- feols(data=temp,sales_price~adj_purch_price+(beds)*factor(adjbin)+(baths)*factor(adjbin)+log(sqft)*factor(adjbin)+log(house_age)*factor(adjbin)+log(LotSizeSquareFeet)*factor(adjbin)|zip3)
t[,predprice:=(predict(hedonicmodel,newdata = t))]
}
zd <- rbind(zd,t)
}
zdata_reg <- zd
rm(zd)# zdata_reg[,predprice:=adj_purch_price]#exp(predict(hedonicmodel,newdata = zd))]
# zdata_reg[,predprice:=(predict(hedonicmodel,newdata = zdata_reg))]#
# zdata_reg[,predprice:=exp(predict(hedonicmodel,newdata = zdata_reg))]#
zdata_reg[,nominalloss:=Purchase_price-predprice]
zdata_reg[,nominalloss:=ifelse(nominalloss<=0,1,nominalloss)]
zdata_reg[,lppp:=listing_amount/predprice]
zdata_reg[,sppp:=sales_price/predprice]
zdata_reg <- zdata_reg[abs(predprice-sales_price)<400000]
zdata_reg[,lp:=floor(listing_amount/1000)*1000]
temp <- zdata_reg[,.(predprice=mean(predprice,na.rm=T),adj_purc=mean(adj_purch_price,na.rm=T),.N),by=lp]Figure 2
# ggplot(temp[N>10 & lp<500000],aes(x=lp,y=adj_purc))+geom_point()+geom_smooth(method="lm")+theme_minimal()+labs(x="Listing price",y="Adjusted purchase price")+ scale_x_continuous(labels = scales::comma)+ scale_y_continuous(labels = scales::comma)+labs(title="Adjusted purchase price")
ggplot(temp[N>10 & lp<1000000 & lp>=300000],aes(x=lp,y=predprice))+geom_point()+theme_minimal()+labs(x="Listing price ($)",y="Predicted price ($)")+ scale_x_continuous(labels = scales::comma)+ scale_y_continuous(labels = scales::comma)#+labs(title="Predicted price")
ggplot(temp[N>10 & lp<1000000 & lp>=300000],aes(x=lp,y=adj_purc))+geom_point()+theme_minimal()+labs(x="Listing price ($)",y="Adjusted purchase price ($)")+ scale_x_continuous(labels = scales::comma)+ scale_y_continuous(labels = scales::comma)#+labs(title="Predicted price")
Table 1
stargazer(zdata_reg[,c("Purchase_price","predprice","Listing_price","lppp","sppp","Property_tax_last_year","Property_tax_rate_last_year","purchase_hpi","hpi_inflation","ownership_years","beds","baths","sqft","LotSizeSquareFeet","house_age","LTV_prev","avg_school_rating","avg_school_distance","totalpopulation","medianage","medianhouseholdincome","Renter_fraction")],summary.stat = c("mean","sd","p25","median","p75","n"),digits = 4,
covariate.labels =c("Purchased price ($)","Predicted price ($)","Listing price ($)","Listing price/Predicted price","Selling price/Predicted price","Effective tax rate","Property taxes paid ($)","HPI_{Purchase}","HPI_{List}/HPI_{Purchase}","Years of ownership","Number of bedrooms","Number of bathrooms","House area (sq. ft)","Lot area (sq. ft)","Age of the house (years)","Loan-to-value_{Purchase}","GreatSchools rating","Distance to schools (miles)","Census tract population","Census tract median age","Census tract median income","Census traction fraction of renters") ,type="text" ) ###
## ===========================================================================================================
## Statistic Mean St. Dev. Pctl(25) Median Pctl(75) N
## -----------------------------------------------------------------------------------------------------------
## Purchased price () 449,528.1000 354,288.2000 225,000 361,500 575,000 25,026
## Predicted price () 603,514.1000 430,144.7000 338,198.8000 489,983.0000 729,357.8000 25,026
## Listing price () 654,899.4000 459,747.0000 386,082.5000 550,837.5000 779,900.0000 25,026
## Listing price/Predicted price 1.1292 0.3546 0.9272 1.0321 1.1968 25,026
## Selling price/Predicted price 1.0962 0.3397 0.9041 1.0061 1.1593 25,026
## Effective tax rate 5,643.8920 4,055.4290 3,117 4,710.0 7,001.1 25,026
## Property taxes paid () 0.0104 0.0037 0.0081 0.0099 0.0119 25,026
## HPIPurchase 421,687.7000 246,589.6000 247,200 382,200 543,800 25,026
## HPIList/HPIPurchase 1.4367 0.6136 0.9895 1.2437 1.7092 25,026
## Years of ownership 12.5718 2.7665 11 12 14 25,026
## Number of bedrooms 3.4188 0.8610 3 3 4 25,026
## Number of bathrooms 2.4713 0.9838 2.0000 2.0000 3.0000 25,026
## House area (sq. ft) 1,980.0340 817.6019 1,407 1,805 2,382 25,026
## Lot area (sq. ft) 16,963.6400 93,945.1700 5,662 7,200 10,018 25,026
## Age of the house (years) 39.0676 22.7821 19 35 56 25,026
## Loan-to-valuePurchase 0.6400 0.4096 0.3941 0.7841 0.8000 25,002
## GreatSchools rating 6.8593 2.0251 5.3333 7.0000 8.6667 25,000
## Distance to schools (miles) 1.4142 1.1362 0.8000 1.1333 1.6667 25,007
## Census tract population 5,882.0820 2,931.9990 4,132.0000 5,377.0000 6,813.0000 25,021
## Census tract median age 40.0760 7.8132 34.6000 39.2000 44.4000 25,019
## Census tract median income 81,087.6000 30,982.7500 58,000.0000 76,263.0000 99,708.0000 25,018
## Census traction fraction of renters 0.3226 0.1708 0.1923 0.2937 0.4298 25,019
## -----------------------------------------------------------------------------------------------------------Table 2
# stargazer(hedonicmodelfelm,no.space = T,align = T,omit.stat=c("ser","f", "rsq"),style = "qje",omit.table.layout = "n", dep.var.labels.include = FALSE,covariate.labels = c("log(Purchase price)","log(HPI inflation)","Number of bedrooms","Number of bathrooms","House area","log(Age of the house)","log(Lot area)"), add.lines = list(c("Year", "Yes")),type="text")#Table 3
controls1 = "log(predprice)+beds+baths+sqft+avg_school_rating+avg_school_distance+walk_score+log(house_age)+log(LotSizeSquareFeet)+LTV_prev+medianage+log(medianhouseholdincome)+Renter_fraction|zip_list_month|"
controls2 = "log(predprice)+beds+baths+sqft+avg_school_rating+avg_school_distance+walk_score+log(house_age)+log(LotSizeSquareFeet)+LTV_prev+medianage+log(medianhouseholdincome)+Renter_fraction|tract_list_month|"
controls3 = "log(predprice)+beds+baths+sqft+avg_school_rating+avg_school_distance+walk_score+log(house_age)+log(LotSizeSquareFeet)+LTV_prev+medianage+log(medianhouseholdincome)+Renter_fraction+log(nominalloss)|tract_list_month|"
controls4 = "log(adj_purch_price)+beds+baths+sqft+avg_school_rating+avg_school_distance+walk_score+log(house_age)+log(LotSizeSquareFeet)+LTV_prev|zip_list_month|"
cluster = "|zip"
instruments = "ownership_years" #
endo_var = "Property_tax_rate_last_year"regs <- list()
regs[[1]] <- felm(as.formula(paste("I(prop_tax_rate*100)~ownership_years+",controls1,"0",cluster,sep="")),data=zdata_reg)
regs[[2]] <- felm(as.formula(paste("I(prop_tax_rate*100)~ownership_years+",controls2,"0","|GEOID",sep="")),data=zdata_reg)
# .printtable(regs)
stargazer(regs,no.space = T,align = T,omit.stat=c("ser","f", "rsq"),style = "qje",omit.table.layout = "n", dep.var.labels.include = F,covariate.labels = c("Ownership years","log(Predicted price)","Number of bedrooms","Number of bathrooms","House area","GreatSchools rating","Distance to schools","Distance to amenities","log(Age of the house)","log(Lot area)","Loan-to-value_{Purchase}","Census tract median age","log(Census tract median income)","Census tract fraction of renters"),type="text")###
## ====================================================
## (1) (2)
## ----------------------------------------------------
## Ownership years -0.053*** -0.050***
## (0.001) (0.001)
## log(Predicted price) -0.757*** -0.817***
## (0.017) (0.023)
## Number of bedrooms 0.015*** 0.011*
## (0.004) (0.007)
## Number of bathrooms 0.006** 0.006
## (0.003) (0.008)
## House area 0.0002*** 0.0002***
## (0.00001) (0.00002)
## GreatSchools rating 0.014*** 0.014
## (0.003) (0.009)
## Distance to schools 0.002 -0.001
## (0.004) (0.011)
## Distance to amenities -0.0003** -0.001**
## (0.0001) (0.0003)
## log(Age of the house) -0.084*** -0.109***
## (0.008) (0.012)
## log(Lot area) 0.036*** 0.035***
## (0.006) (0.010)
## Loan-to-valuePurchase 0.037*** 0.032***
## (0.007) (0.010)
## Census tract median age 0.002**
## (0.001) (0.000)
## log(Census tract median income) 0.119***
## (0.020) (0.000)
## Census tract fraction of renters 0.087**
## (0.035) (0.000)
## N 23,966 23,966
## Adjusted R2 0.648 0.680
## ====================================================Table 5
olsformula <- as.formula(paste("log(Listing_price)~Property_tax_rate_last_year+",controls1,"0",cluster,sep=""))
ivformula <- as.formula(paste("log(Listing_price)~",controls1,"(",endo_var,"~",instruments,")",cluster,sep=""))
ivformula_tract <-as.formula(paste("log(Listing_price)~",controls2,"(",endo_var,"~",instruments,")","|GEOID",sep=""))
ivformula_tract_loss <-as.formula(paste("log(Listing_price)~",controls3,"(",endo_var,"~",instruments,")","|GEOID",sep=""))
regs <- list()
regs[[1]] <- felm(olsformula,data=zdata_reg)
regs[[2]] <- felm(ivformula,data=zdata_reg)
regs[[3]] <- felm(ivformula_tract,data=zdata_reg)
regs[[4]] <- felm(ivformula_tract_loss,data=zdata_reg)
stargazer(regs,no.space = T,align = T,omit.stat=c("ser","f", "rsq"),style = "qje",omit.table.layout = "n", dep.var.labels.include = FALSE,
covariate.labels = c("Effective tax rate","log(Predicted price)","Number of bedrooms","Number of bathrooms","House area","GreatSchools rating","Distance to schools","Distance to amenities","log(Age of the house)","log(Lot area)","Loan-to-value_{Purchase}","Census tract median age","log(Census tract median income)","Census tract fraction of renters","log(Nominal loss)"),
column.labels=c("OLS","IV"),column.separate=c(1,3),
add.lines = list(c("zipmonth", "Yes", "Yes", "No","No"),
c("tractmonth", "Yes", "Yes", "No","No"),
c("Cond. F. Stat","",round(condfstat(regs[[2]])[[1]],2),round(condfstat(regs[[3]])[[1]],2),round(condfstat(regs[[4]])[[1]],2))),type="text")###
## ============================================================================
## OLS IV
## (1) (2) (3) (4)
## ----------------------------------------------------------------------------
## Effective tax rate 15.997***
## (0.825)
## log(Predicted price) 0.366*** 0.320*** 0.314*** 0.298***
## (0.017) (0.017) (0.022) (0.023)
## Number of bedrooms 0.009** 0.010** 0.009* 0.009*
## (0.004) (0.004) (0.005) (0.005)
## Number of bathrooms 0.007 0.008 0.012* 0.012*
## (0.005) (0.005) (0.006) (0.006)
## House area 0.0002*** 0.0002*** 0.0002*** 0.0002***
## (0.00001) (0.00001) (0.00002) (0.00002)
## GreatSchools rating 0.013*** 0.014*** 0.011* 0.011*
## (0.003) (0.003) (0.006) (0.006)
## Distance to schools -0.004 -0.004 -0.003 -0.003
## (0.003) (0.003) (0.005) (0.005)
## Distance to amenities -0.0003*** -0.0003*** -0.001*** -0.001***
## (0.0001) (0.0001) (0.0002) (0.0002)
## log(Age of the house) -0.001 -0.008 -0.021** -0.021**
## (0.005) (0.005) (0.008) (0.008)
## log(Lot area) 0.060*** 0.061*** 0.056*** 0.057***
## (0.006) (0.006) (0.008) (0.008)
## Loan-to-valuePurchase 0.003 0.004 -0.001 -0.0004
## (0.003) (0.003) (0.005) (0.005)
## Census tract median age 0.003*** 0.003***
## (0.0005) (0.0005) (0.000) (0.000)
## log(Census tract median income) 0.076*** 0.083***
## (0.011) (0.012) (0.000) (0.000)
## Census tract fraction of renters 0.055*** 0.059***
## (0.020) (0.020) (0.000) (0.000)
## log(Nominal loss) 0.001**
## (0.0005)
## `Property_tax_rate_last_year(fit)` 10.680*** 11.499*** 9.942***
## (0.960) (1.447) (1.525)
## zipmonth Yes Yes No No
## tractmonth Yes Yes No No
## Cond. F. Stat 217.57 115.17 94.83
## N 23,966 23,966 23,966 23,966
## Adjusted R2 0.947 0.947 0.953 0.953
## ============================================================================Table 6
zhvi <- readRDS("C:/Users/dratnadiwakara2/Documents/OneDrive - Louisiana State University/Raw Data/Zillow Research Data/Zip_Zhvi_SingleFamilyResidence.rds")
zhvi <- data.table(zhvi)
zhvi <- zhvi[state=="CA"]
zhvi <- zhvi[month %in% as.Date(c("2006-01-01","2014-01-01"))]
zhvi <- zhvi[,c("zipcode","month","zhvi")]
zhvi <- dcast(zhvi,zipcode~month,value.var = "zhvi")
names(zhvi) <- c("zipcode","val2006","val2014")
zhvi <- zhvi[complete.cases(zhvi),]
zhvi <- data.table(zhvi)
zhvi[,pricechange:=log(val2014/val2006)]
hist <- ggplot()+geom_histogram(data=zhvi,aes(x=pricechange),fill="dodgerblue4",color="dodgerblue4",alpha=0.75)+labs(x="Price Change 2006 to 2014",y="Number of Zipcodes")+ theme_minimal()+scale_y_continuous(labels = comma)
zhvi[,pricechangeqt:=ntile(pricechange,16)]
zhvi <- zhvi[,c("zipcode","pricechange","pricechangeqt")]
zhvi[,zipcode:=as.numeric(zipcode)]
zdata_reg <- merge(zdata_reg, zhvi, by.x="zip",by.y="zipcode",all.x=T)
# print(hist)regs <- list()
regs[[1]] <- felm(olsformula,data=zdata_reg[nominalloss>1])
regs[[2]] <- felm(olsformula,data=zdata_reg[nominalloss<=1])
regs[[3]] <- felm(olsformula,data=zdata_reg[pricechangeqt<=4])
regs[[4]] <- felm(olsformula,data=zdata_reg[pricechangeqt %in% 5:16])
regs[[5]] <- felm(ivformula,data=zdata_reg[pricechangeqt<=4])
regs[[6]] <- felm(ivformula,data=zdata_reg[pricechangeqt %in% 5:16])
stargazer(regs,no.space = T,align = T,omit.stat=c("ser","f", "rsq"),style = "qje",omit.table.layout = "n", dep.var.labels.include = FALSE,
covariate.labels = c("Effective tax rate","log(Predicted price)"),
column.labels=c("OLS","IV"),column.separate=c(4,2),
add.lines = list(c("Controls", "Yes", "Yes", "Yes","Yes","Yes","Yes"),c("zipmonth", "Yes", "Yes", "Yes","Yes","Yes","Yes"),c("Cond. F. Stat","","","","",round(condfstat(regs[[5]])[[1]],2),round(condfstat(regs[[6]])[[1]],2))),
keep=c("Property_tax_rate_last_year","predprice"),type="text")###
## =============================================================================================
## OLS IV
## (1) (2) (3) (4) (5) (6)
## ---------------------------------------------------------------------------------------------
## Effective tax rate 21.626*** 15.654*** 15.370*** 16.180***
## (2.589) (0.885) (2.025) (0.882)
## log(Predicted price) 0.632*** 0.338*** 0.462*** 0.353*** 0.469*** 0.302***
## (0.049) (0.017) (0.049) (0.017) (0.075) (0.017)
## `Property_tax_rate_last_year(fit)` 15.873*** 9.944***
## (4.022) (0.980)
## Controls Yes Yes Yes Yes Yes Yes
## zipmonth Yes Yes Yes Yes Yes Yes
## Cond. F. Stat 10.73 222.29
## N 4,598 19,368 3,776 20,170 3,776 20,170
## Adjusted R2 0.944 0.951 0.861 0.934 0.861 0.933
## =============================================================================================Table 7
regs <- list()
regs[[1]] <- felm(ivformula,data=zdata_reg[Listing_price<listing_q2])
regs[[2]] <- felm(ivformula,data=zdata_reg[Listing_price>listing_q2])
CA <- read_fst("C:/Users/dratnadiwakara2/Documents/OneDrive - Louisiana State University/Projects/Sunk Cost/Processed Data/all_homes_prev/CA_all_homes_prev.fst",columns = c("BuildingAreaSqFt","zip","Bedrooms"),as.data.table = T)
CA_summary <- CA[,.(
sqft_lb=quantile(BuildingAreaSqFt,0.25, na.rm=T),
sqft_hb=quantile(BuildingAreaSqFt,0.75, na.rm=T)),
by=zip]
temp <- CA[,.(
beds_lb=quantile(Bedrooms,0.25, na.rm=T),
beds_hb=floor(quantile(Bedrooms,0.75, na.rm=T)+0.99)),
by=zip]
CA_summary <- merge(CA_summary,temp,by="zip")
zdata_reg <- merge(zdata_reg,CA_summary,by="zip",all.x=T)
zdata_reg[,outlider_sqft:=ifelse(sqft>sqft_hb ,1,0)]
zdata_reg[,outlider_beds:=ifelse(beds>beds_hb,1,0)]
regs[[3]] <- felm(ivformula,data=zdata_reg[outlider_sqft==0 & outlider_beds==0])
regs[[4]] <- felm(ivformula,data=zdata_reg[outlider_sqft==1 | outlider_beds==1])
redfin <- read.csv("C:/Users/dratnadiwakara2/Documents/Research/UH computer27/sunkcost_2019/home_sales_redfin.csv")
redfin$month <- as.Date(redfin$month,origin="1900-01-01")
redfin <- data.table(redfin)
redfin <- redfin[redfin$month<="2016-01-01"]
redfin <- redfin[,.(median_home_sales=median(home_sales)),by=list(zip)]
redfin[,low_activiy:=ifelse(redfin$median_home_sales<quantile(redfin$median_home_sales,0.75),1,0)]
names(redfin) <- c("zip","median_home_sales_2","low_activity_2")
zdata_reg <- merge(zdata_reg,redfin,by="zip",all.x=TRUE)
regs[[5]] <- felm(ivformula,data=zdata_reg[low_activiy==1])
regs[[6]] <- felm(ivformula,data=zdata_reg[low_activiy==0])
temp <- zdata_reg[!duplicated(zdata_reg$GEOID)]
median_age <- median(temp$medianage,na.rm = TRUE)
regs[[7]] <- felm(ivformula,data=zdata_reg[medianage<=median_age])
regs[[8]] <- felm(ivformula,data=zdata_reg[medianage>median_age])
medincome <- median(temp$medianhouseholdincome,na.rm = TRUE)
regs[[9]] <- felm(ivformula,data=zdata_reg[medianhouseholdincome<=medincome])
regs[[10]] <- felm(ivformula,data=zdata_reg[medianhouseholdincome>medincome])
median_renterfrac <- median(temp$Renter_fraction,na.rm = TRUE)
regs[[11]] <- felm(ivformula,data=zdata_reg[Renter_fraction<=median_renterfrac])
regs[[12]] <- felm(ivformula,data=zdata_reg[Renter_fraction>median_renterfrac])
stargazer(regs,no.space = T,align = T,omit.stat=c("ser","f", "rsq"),style = "qje",omit.table.layout = "n", dep.var.labels.include = FALSE,
covariate.labels = c("Effective tax rate"),
column.labels=c("L Pr.","H Pr","L Sz","H Sz","L Act","H Act","L Ag","H Ag","L Inc","H Inc","L Ren","H Ren"),
add.lines = list(
c("Controls", rep("Yes",12)),
c("zipmonth", rep("Yes",12)),
c("Cond. F. Stat",round(c(condfstat(regs[[1]])[[1]],condfstat(regs[[2]])[[1]],condfstat(regs[[3]])[[1]],condfstat(regs[[4]])[[1]],condfstat(regs[[5]])[[1]],condfstat(regs[[6]])[[1]],condfstat(regs[[7]])[[1]],condfstat(regs[[8]])[[1]],condfstat(regs[[9]])[[1]],condfstat(regs[[10]])[[1]],condfstat(regs[[11]])[[1]],condfstat(regs[[12]])[[1]]),2))),
keep=c("Property_tax_rate_last_year"), flip=T,
type="text")###
## ====================================================================================================================================
## L Pr. H Pr L Sz H Sz L Act H Act L Ag H Ag L Inc H Inc L Ren H Ren
## (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12)
## ------------------------------------------------------------------------------------------------------------------------------------
## Effective tax rate 4.350*** 13.203*** 8.302*** 16.169*** 10.839*** 9.778*** 6.247*** 12.880*** 16.183*** 8.641*** 9.966*** 13.141***
## (1.063) (1.470) (1.075) (2.094) (1.485) (1.251) (1.478) (1.313) (1.998) (1.030) (1.269) (1.589)
## Controls Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
## zipmonth Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
## Cond. F. Stat 118.11 89.79 159.48 71.65 107.39 131.47 57.36 152.94 43.24 200.69 135.55 71.05
## N 12,297 11,556 15,556 8,410 10,351 12,347 9,336 14,630 8,464 15,502 15,420 8,546
## Adjusted R2 0.959 0.958 0.945 0.949 0.940 0.943 0.941 0.943 0.922 0.941 0.949 0.941
## ====================================================================================================================================Table 8 (Panel A)
regs <- list()
regs[[1]] <- felm(as.formula(paste("log(sales_price)~Property_tax_rate_last_year+",controls1,"0",cluster,sep="")),data=zdata_reg)
regs[[2]] <- felm(as.formula(paste("log(sales_price)~",controls1,"(",endo_var,"~",instruments,")",cluster,sep="")),data=zdata_reg)
regs[[3]] <- felm(as.formula(paste("log(Time_on_market)~Property_tax_rate_last_year+",controls1,"0",cluster,sep="")),data=zdata_reg)
regs[[4]] <- felm(as.formula(paste("log(Time_on_market)~",controls1,"(",endo_var,"~",instruments,")",cluster,sep="")),data=zdata_reg)
stargazer(regs,no.space = T,align = T,omit.stat=c("ser","f", "rsq"),style = "qje",omit.table.layout = "n", dep.var.labels.include = FALSE,
covariate.labels = c("Effective tax rate","log(Predicted price)"),
column.labels=c("log(Sales price)","log(Days on market)"),column.separate=c(2,2),
add.lines = list(c("Controls", "Yes", "Yes", "Yes","Yes","Yes","Yes"),c("zipmonth", "Yes", "Yes", "Yes","Yes","Yes","Yes"),c("Cond. F. Stat","",round(condfstat(regs[[2]])[[1]],2),"",round(condfstat(regs[[4]])[[1]],2))),
keep=c("Property_tax_rate_last_year","predprice"), type="text")# ##
## =========================================================================
## log(Sales price) log(Days on market)
## (1) (2) (3) (4)
## -------------------------------------------------------------------------
## Effective tax rate 16.208*** -1.495
## (0.834) (1.764)
## log(Predicted price) 0.367*** 0.321*** -0.062** 0.003
## (0.017) (0.017) (0.025) (0.034)
## `Property_tax_rate_last_year(fit)` 10.863*** 6.060*
## (1.037) (3.171)
## Controls Yes Yes Yes Yes
## zipmonth Yes Yes Yes Yes
## Cond. F. Stat 217.57 217.57
## N 23,966 23,966 23,966 23,966
## Adjusted R2 0.923 0.922 0.342 0.341
## =========================================================================temp <- zdata_reg
temp[,purch_month_no:=as.numeric(format(temp$PurchaseMonth,"%m"))]
temp[,sale_month_no:=as.numeric(format(temp$sale_date,"%m"))]
temp <- temp[!is.na(temp$predprice) & !is.na(temp$GEOID)]
temp[,zip_purch_list:=paste(temp$zip,temp$purchase_list_year)]
adj_proptax <- felm(Property_tax_rate_last_year~log(predprice)|zip_purch_list,data=temp)
adj_proptax_resid <- data.frame(adj_proptax$residuals)
names(adj_proptax_resid) <- c("adj_proptax_resid")
adj_list <- felm(log(Listing_price)~log(predprice)|zip_purch_list,data=temp)
adj_list_resid <- data.frame(adj_list$residuals)
names(adj_list_resid) <- c("adj_list_resid")
adj_sale <- felm(log(Selling_price)~log(predprice)|zip_purch_list,data=temp)
adj_sale_resid <- data.frame(adj_sale$residuals)
names(adj_sale_resid) <- c("adj_sale_resid")
adj_purch <- felm(log(Purchase_price)~0|zip_purch_list,data=temp)
adj_purch_resid <- data.frame(adj_purch$residuals)
names(adj_purch_resid) <- c("adj_purch_resid")
temp <- cbind(temp,adj_proptax_resid)
temp <- cbind(temp,adj_list_resid)
temp <- cbind(temp,adj_sale_resid)
temp <- cbind(temp,adj_purch_resid)
temp[,DocumentDate_prev:=as.Date(temp$DocumentDate_prev,origin = "1970-01-01")]
temp[,purch_week:=as.numeric(strftime(temp$DocumentDate_prev,format="%V"))-20]
temp[,postJune:=ifelse(purch_week>-2,1,0)]
temp[,zip_purch_year:=paste(zip,PurchaseYear)]Table 9
dscregs <- list()
dscregs[[1]] <- felm(log(Purchase_price)~postJune+purch_week|zip_purch_year|0|zip,data=temp)
dscregs[[2]] <- felm(beds~postJune+purch_week|zip_purch_year|0|zip,data=temp)
dscregs[[3]] <- felm(baths~postJune+purch_week|zip_purch_year|0|zip,data=temp)
dscregs[[4]] <- felm(log(sqft)~postJune+purch_week|zip_purch_year|0|zip,data=temp)
dscregs[[5]] <- felm(log(house_age)~postJune+purch_week|zip_purch_year|0|zip,data=temp)
# dscregs[[6]] <- felm(log(0.0001+LTV_prev)~postJune|zip_purch_year|0|zip,data=temp)
stargazer(dscregs,no.space = T,align = T,omit.stat=c("ser","f", "rsq"),style = "qje",omit.table.layout = "n", dep.var.labels.include = F,
covariate.labels = c("Purchased after June 01"),
add.lines = list(c("zippurchmonth", "Yes", "Yes", "Yes","Yes","Yes","Yes")),
column.labels=c("log(Purchase price)","No beds","No baths","log(House area)","log(Age of house)","log(Loan-to-value)"),type = "text")##
## ===============================================================================================
## log(Purchase price) No beds No baths log(House area) log(Age of house)
## (1) (2) (3) (4) (5)
## -----------------------------------------------------------------------------------------------
## Purchased after June 01 0.037*** -0.042** -0.034* -0.025** 0.023*
## (0.010) (0.019) (0.020) (0.010) (0.013)
## purch_week 0.0004 0.001 -0.00001 0.0003 -0.001
## (0.0003) (0.001) (0.001) (0.0004) (0.0004)
## zippurchmonth Yes Yes Yes Yes Yes
## N 25,021 25,021 25,021 25,021 25,021
## Adjusted R2 0.645 0.173 0.145 0.182 0.485
## =============================================================================================== # keep=c("Property_tax_rate_last_year","predprice"),
# type="text")#Table 10 (Panel A)
regs = list()
regs[[1]] = felm(Property_tax_rate_last_year~postJune+purch_week+log(predprice)+beds+baths+sqft+avg_school_rating+avg_school_distance+walk_score+log(house_age)+log(LotSizeSquareFeet)+LTV_prev+medianage+log(medianhouseholdincome)+Renter_fraction|zip_purch_list|0|zip,data=temp)
regs[[2]] = felm(log(Listing_price)~postJune+purch_week+log(predprice)+beds+baths+sqft+avg_school_rating+avg_school_distance+walk_score+log(house_age)+log(LotSizeSquareFeet)+LTV_prev+medianage+log(medianhouseholdincome)+Renter_fraction|zip_purch_list|0|zip,data=temp)
regs[[3]] = felm(log(Selling_price)~postJune+purch_week+log(predprice)+beds+baths+sqft+avg_school_rating+avg_school_distance+walk_score+log(house_age)+log(LotSizeSquareFeet)+LTV_prev+medianage+log(medianhouseholdincome)+Renter_fraction|zip_purch_list|0|zip,data=temp)
stargazer(regs,no.space = T,align = T,omit.stat=c("ser","f", "rsq"),style = "qje",omit.table.layout = "n", dep.var.labels.include = FALSE,
covariate.labels = c("Purchased after June 01","Week of purchase","log(Predicted price)"),
column.labels=c("Effective tax rate","log(Listing price)","log(Selling price)"),
add.lines = list(c("Controls", "Yes", "Yes", "Yes","Yes","Yes","Yes"),c("zippurchmonth", "Yes", "Yes", "Yes","Yes","Yes","Yes")),
keep=c("postJune","predprice","purch_week"),type="text")###
## ================================================================================
## Effective tax rate log(Listing price) log(Selling price)
## (1) (2) (3)
## --------------------------------------------------------------------------------
## Purchased after June 01 -0.0002*** -0.012*** -0.011**
## (0.0001) (0.004) (0.005)
## Week of purchase 0.00001*** 0.0004*** 0.0004**
## (0.00000) (0.0001) (0.0002)
## log(Predicted price) -0.007*** 0.249*** 0.256***
## (0.0002) (0.014) (0.014)
## Controls Yes Yes Yes
## zippurchmonth Yes Yes Yes
## N 23,966 23,966 23,966
## Adjusted R2 0.670 0.942 0.936
## ================================================================================Table 10 (Panel B)
regs = list()
regs[[1]] = felm(Property_tax_rate_last_year~postJune+purch_week+I(purch_week^2)+log(predprice)+beds+baths+sqft+avg_school_rating+avg_school_distance+walk_score+log(house_age)+log(LotSizeSquareFeet)+LTV_prev+medianage+log(medianhouseholdincome)+Renter_fraction|zip_purch_list|0|zip,data=temp)
regs[[2]] = felm(log(Listing_price)~postJune+purch_week+I(purch_week^2)+log(predprice)+beds+baths+sqft+avg_school_rating+avg_school_distance+walk_score+log(house_age)+log(LotSizeSquareFeet)+LTV_prev+medianage+log(medianhouseholdincome)+Renter_fraction|zip_purch_list|0|zip,data=temp)
regs[[3]] = felm(log(Selling_price)~postJune+purch_week+I(purch_week^2)+log(predprice)+beds+baths+sqft+avg_school_rating+avg_school_distance+walk_score+log(house_age)+log(LotSizeSquareFeet)+LTV_prev+medianage+log(medianhouseholdincome)+Renter_fraction|zip_purch_list|0|zip,data=temp)
stargazer(regs,no.space = T,align = T,omit.stat=c("ser","f", "rsq"),style = "qje",omit.table.layout = "n", dep.var.labels.include = FALSE,
covariate.labels = c("Purchased after June 01","Week of purchase","Week of purchase^{2}","log(Predicted price)"),
column.labels=c("Effective tax rate","log(Listing price)","log(Selling price)"),
add.lines = list(c("Controls", "Yes", "Yes", "Yes","Yes","Yes","Yes"),c("zippurchmonth", "Yes", "Yes", "Yes","Yes","Yes","Yes")),
keep=c("postJune","predprice","purch_week"),type="text")#, type="text"##
## ================================================================================
## Effective tax rate log(Listing price) log(Selling price)
## (1) (2) (3)
## --------------------------------------------------------------------------------
## Purchased after June 01 -0.0002*** -0.011** -0.010*
## (0.0001) (0.005) (0.006)
## Week of purchase 0.00002*** 0.0004 0.0003
## (0.00000) (0.0002) (0.0002)
## Week of purchase2 -0.00000 0.00000 0.00000
## (0.00000) (0.00001) (0.00001)
## log(Predicted price) -0.007*** 0.249*** 0.256***
## (0.0002) (0.014) (0.014)
## Controls Yes Yes Yes
## zippurchmonth Yes Yes Yes
## N 23,966 23,966 23,966
## Adjusted R2 0.670 0.942 0.936
## ================================================================================Table 10 (Panel C)
Local linear running variable (3-months)
regs = list()
regs[[1]] = felm(Property_tax_rate_last_year~postJune+purch_week+log(predprice)+beds+baths+sqft+avg_school_rating+avg_school_distance+walk_score+log(house_age)+log(LotSizeSquareFeet)+LTV_prev+medianage+log(medianhouseholdincome)+Renter_fraction|zip_purch_list|0|zip,data=temp[abs(purch_week)<=12])
regs[[2]] = felm(log(Listing_price)~postJune+purch_week+log(predprice)+beds+baths+sqft+avg_school_rating+avg_school_distance+walk_score+log(house_age)+log(LotSizeSquareFeet)+LTV_prev+medianage+log(medianhouseholdincome)+Renter_fraction|zip_purch_list|0|zip,data=temp[abs(purch_week)<=12])
regs[[3]] = felm(log(Selling_price)~postJune+purch_week+log(predprice)+beds+baths+sqft+avg_school_rating+avg_school_distance+walk_score+log(house_age)+log(LotSizeSquareFeet)+LTV_prev+medianage+log(medianhouseholdincome)+Renter_fraction|zip_purch_list|0|zip,data=temp[abs(purch_week)<=12])
stargazer(regs,no.space = T,align = T,omit.stat=c("ser","f", "rsq"),style = "qje",omit.table.layout = "n", dep.var.labels.include = FALSE,
covariate.labels = c("Purchased after June 01","Week of purchase","log(Predicted price)"),
column.labels=c("Effective tax rate","log(Listing price)","log(Selling price)"),
add.lines = list(c("Controls", "Yes", "Yes", "Yes","Yes","Yes","Yes"),c("zippurchmonth", "Yes", "Yes", "Yes","Yes","Yes","Yes")),
keep=c("postJune","predprice","purch_week"),type="text")#,type="text"##
## ================================================================================
## Effective tax rate log(Listing price) log(Selling price)
## (1) (2) (3)
## --------------------------------------------------------------------------------
## Purchased after June 01 -0.0002** -0.009 -0.013*
## (0.0001) (0.007) (0.007)
## Week of purchase 0.00002** 0.0002 0.0004
## (0.00001) (0.001) (0.001)
## log(Predicted price) -0.007*** 0.233*** 0.241***
## (0.0002) (0.014) (0.016)
## Controls Yes Yes Yes
## zippurchmonth Yes Yes Yes
## N 12,662 12,662 12,662
## Adjusted R2 0.675 0.942 0.940
## ================================================================================