SCM version 1(only KOR, LUX)
### load data from google sheet
combined_data <- read.csv("https://docs.google.com/spreadsheets/d/e/2PACX-1vRe67N_K9M-u4GgzLxD310RbZgaRTjk4wXYch-TQT2e29TXciZ2Pdz8vr50p8_LMfw_TrvOhpSNZ2yF/pub?gid=1175897076&single=true&output=csv")
##### SCM Version 1
######################
#install.packages("Synth", repos="http://cran.rstudio.com/")
library(Synth )
## ##
## ## Synth Package: Implements Synthetic Control Methods.
## ## See https://web.stanford.edu/~jhain/synthpage.html for additional information.
#data("synth.data") # Example dataset included in Synth package
country.list <- unique(combined_data$unit.num)
list <- setdiff(country.list, c(10,22,30,38,40,41,46,49,56, 72,76,90,100,107,110,112,116,117,144, 146,177,181,187,195,202,203,204,223,224,225,246,253,265))
#list <- setdiff(country.list, c(41,100,107,110,146,202,246,265))
dataprep.out <-
dataprep(
foo = combined_data,
predictors = c("alcohol","GDP","infant_mortality","cancer_death","doctors"),
predictors.op = "mean",
dependent = "life_expectancy_at_birth",
unit.variable = "unit.num",
unit.names.variable = "REF_AREA",
time.variable = "TIME_PERIOD",
treatment.identifier = 202,
controls.identifier = list,
time.predictors.prior = c(1986:1995),
time.optimize.ssr = c(1986:1995),
time.plot = 1986:2005
)
synth.out <- synth(dataprep.out)
##
## X1, X0, Z1, Z0 all come directly from dataprep object.
##
##
## ****************
## searching for synthetic control unit
##
##
## ****************
## ****************
## ****************
##
## MSPE (LOSS V): 0.3085476
##
## solution.v:
## 0.2521439 0.05408721 0.02503257 0.026605 0.6421313
##
## solution.w:
## 5.5e-09 6e-09 2.1e-09 2e-09 5.1e-09 4.1e-09 3.655e-07 5.55e-08 2.9e-09 1.12e-08 3e-10 5e-10 0.4671621 0.5328374 1.3e-09 4e-10 3.8e-09
synth.tables <- synth.tab(
dataprep.res = dataprep.out,
synth.res = synth.out)
synth.tables$tab.pred
## Treated Synthetic Sample Mean
## alcohol 35.300 12.156 10.345
## GDP 8797.800 21240.993 16985.929
## infant_mortality 4.368 9.373 10.136
## cancer_death 98.889 1292.199 1274.033
## doctors 1.123 1.485 2.304
synth.tables$tab.w
## w.weights unit.names unit.numbers
## 14 0.000 AUS 14
## 15 0.000 AUT 15
## 18 0.000 BEL 18
## 36 0.000 CAN 36
## 55 0.000 CZE 55
## 59 0.000 DNK 59
## 71 0.000 ESP 71
## 78 0.000 FRA 78
## 82 0.000 GBR 82
## 102 0.000 HUN 102
## 115 0.000 ISL 115
## 120 0.000 JPN 120
## 127 0.467 KOR 127
## 145 0.533 LUX 145
## 155 0.000 MEX 155
## 178 0.000 NOR 178
## 191 0.000 POL 191
synth.tables$tab.loss
## Loss W Loss V
## [1,] 3.548052 0.3085476
path.plot(dataprep.res = dataprep.out,
synth.res = synth.out,
Ylab="Life expectancy at birth",
Xlab="Year",
Ylim = c(70,82),
Legend=c("Taiwan","Synthetic Taiwan"),
Legend.position="bottomright",
Main = "SCM Version 1"
)
abline(v=1995, lty=2)
arrows(1995, 78, 1993, 78,
col = "black",
length = .1)
text(1991, 78,
"NHI Policy",
cex = 1)
gaps.plot(dataprep.res = dataprep.out,
synth.res = synth.out,
Ylab="Gap of life expectancy",
Xlab="Year"
)
abline(v=1995, lty=2)
arrows(1995, 1, 1993, 1,
col = "black",
length = .1)
text(1991, 1,
"NHI Policy",
cex = 1)
SCM version 2(KOR, LUX, ESP)
##### SCM Version 2
### load data from google sheet
combined_data <- read.csv("https://docs.google.com/spreadsheets/d/e/2PACX-1vRe67N_K9M-u4GgzLxD310RbZgaRTjk4wXYch-TQT2e29TXciZ2Pdz8vr50p8_LMfw_TrvOhpSNZ2yF/pub?gid=1175897076&single=true&output=csv")
#install.packages("Synth", repos="http://cran.rstudio.com/")
library(Synth )
#data("synth.data") # Example dataset included in Synth package
country.list <- unique(combined_data$unit.num)
list <- setdiff(country.list, c(10,22,30,38,40,41,46,49,90,100,107,110,112,116,117, 146,177,181,187,195,202,203,204,223,224,246,265))
#list <- setdiff(country.list, c(41,100,107,110,146,202,246,265))
dataprep.out <-
dataprep(
foo = combined_data,
predictors = c("alcohol","GDP","infant_mortality","cancer_death","doctors"),
predictors.op = "mean",
dependent = "life_expectancy_at_birth",
unit.variable = "unit.num",
unit.names.variable = "REF_AREA",
time.variable = "TIME_PERIOD",
treatment.identifier = 202,
controls.identifier = list,
time.predictors.prior = c(1986:1995),
time.optimize.ssr = c(1986:1995),
time.plot = 1986:2005
)
synth.out <- synth(dataprep.out)
##
## X1, X0, Z1, Z0 all come directly from dataprep object.
##
##
## ****************
## searching for synthetic control unit
##
##
## ****************
## ****************
## ****************
##
## MSPE (LOSS V): 0.2985003
##
## solution.v:
## 0.1701266 0.2136082 0.007266856 0.0001097161 0.6088886
##
## solution.w:
## 1.6676e-06 2.0092e-06 7.313e-07 8.193e-07 7.321e-07 4.5923e-06 2.4761e-06 0.2228325 3.131e-07 8.016e-07 2.52643e-05 1.7278e-06 2.51635e-05 2.008e-07 6.9e-09 0.5727735 2.052e-07 0.2043249 3.748e-07 2.254e-07 7.675e-07 2.512e-07 7.643e-07
synth.tables <- synth.tab(
dataprep.res = dataprep.out,
synth.res = synth.out)
synth.tables$tab.pred
## Treated Synthetic Sample Mean
## alcohol 35.300 11.160 10.024
## GDP 8797.800 13666.109 16897.810
## infant_mortality 4.368 9.663 9.851
## cancer_death 98.889 1255.513 1396.026
## doctors 1.123 1.373 2.445
synth.tables$tab.w
## w.weights unit.names unit.numbers
## 14 0.000 AUS 14
## 15 0.000 AUT 15
## 18 0.000 BEL 18
## 36 0.000 CAN 36
## 55 0.000 CZE 55
## 56 0.000 DEU 56
## 59 0.000 DNK 59
## 71 0.223 ESP 71
## 72 0.000 EST 72
## 76 0.000 FIN 76
## 78 0.000 FRA 78
## 82 0.000 GBR 82
## 102 0.000 HUN 102
## 115 0.000 ISL 115
## 120 0.000 JPN 120
## 127 0.573 KOR 127
## 144 0.000 LTU 144
## 145 0.204 LUX 145
## 155 0.000 MEX 155
## 178 0.000 NOR 178
## 191 0.000 POL 191
## 225 0.000 SWE 225
## 253 0.000 USA 253
synth.tables$tab.loss
## Loss W Loss V
## [1,] 2.870037 0.2985003
path.plot(dataprep.res = dataprep.out,
synth.res = synth.out,
Ylab="Life expectancy at birth",
Xlab="Year",
Ylim = c(70,82),
Legend=c("Taiwan","Synthetic Taiwan"),
Legend.position="bottomright",
Main = "SCM Version 2"
)
abline(v=1995, lty=2)
arrows(1995, 78, 1993, 78,
col = "black",
length = .1)
text(1991, 78,
"NHI Policy",
cex = 1)
gaps.plot(dataprep.res = dataprep.out,
synth.res = synth.out,
Ylab="Gap of life expectancy",
Xlab="Year"
)
abline(v=1995, lty=2)
arrows(1995, 1, 1993, 1,
col = "black",
length = .1)
text(1991, 1,
"NHI Policy",
cex = 1)
