[This article was first published on Macroeconomic Observatory - R, and kindly contributed to R-bloggers]. (You can report issue about the content on this page here)
Want to share your content on R-bloggers? click here if you have a blog, or here if you don't.
Want to share your content on R-bloggers? click here if you have a blog, or here if you don't.
We gather macroeconomic data for France, Germany, Italy, Spain & the Euro area, which can be used for various purposes, especially for the estimation and the calibration of general equilibrium models. This database is built and automatically updated through DBnomics using the rdbnomics package. All the code is written in R, thanks to the @Rct16 and @Rstu16.
The Euro Area
First of all, we gather the following databases:
-
The database based on @Faga01 and used in the bayesian estimation of the DSGE model developped in @Smet03 for the Euro area : https://macro.cepremap.fr/article/2015-10/sw03-data/
-
The financial database similar to @Chri14a, but for the Euro area : https://macro.cepremap.fr/article/2016-06/cmr14-EA-data/
-
The fiscal database for the Euro Area, based on the paper of @ppp09 : https://macro.cepremap.fr/article/2019-11/fipu-EA-data/
-
The foreign block database for the Eurozone : https://macro.cepremap.fr/article/2019-12/open-EA-data/
sw03 <- read_csv("https://shiny.cepremap.fr/data/EA_SW_rawdata.csv") %>%
filter(period >="1980-01-01")
fipu <- read_csv("https://shiny.cepremap.fr/data/EA_Fipu_rawdata.csv")
finance <- read_csv("https://shiny.cepremap.fr/data/EA_Finance_rawdata.csv")
open <- read_csv("https://shiny.cepremap.fr/data/EA_Open_rawdata.csv")
EA_rawdata <-
sw03 %>%
inner_join(fipu,by="period") %>%
inner_join(finance,by="period") %>%
inner_join(open,by="period") %>%
rename(unempbenef=unemp) %>%
mutate(pop=1000*pop) %>%
add_column(country="EA")
France, Germany, Italy & Spain
Additionnally to the Euro area data, country data for France, Germany, Italy and Spain is needed. Different sources have been used: Eurostat, IMF (WEO and IFS), BIS, OECD and ECB.
Data Retrieval & Seasonal Adjustment
In this section we get country data on :
- Gross Domestic Product
- consumption
- investment
- GDP deflator
- compensation of employees
- hours worked
- investment deflator
- loans to non-financial corporations
- entrepreneurial net worth
- short-term interest rate
- lending rate
- total government expenditure
- government consumption
- government investment
- government social transfers
- government interest payments
- government debt
- world demand
- total government revenue
- unemployment benefits
- nominal effective exchange rate
- imports
- exports
- population
As in the Euro area database, except oil prices which are the same by assumption.
Some of these data may not be seasonally adjusted, therefore, we use the seasonal package developed by Sax (2016) to remove the seasonal component.
Compensation of Employees
We use the namq_10_a10 database from Eurostat to obtain these data.
wage_de_fr <-
rdb("Eurostat","namq_10_a10",mask = "Q.CP_MEUR.SA.TOTAL.D1.DE+FR") %>%
add_column(var="wage")
wage_es_it <-
rdb("Eurostat","namq_10_a10", mask = "Q.CP_MEUR.SCA.TOTAL.D1.ES+IT") %>%
add_column(var="wage")
Hours Worked
We use the namq_10_a10_e database from Eurostat to retrieve these data.
hours <-
rdb("Eurostat","namq_10_a10_e",mask = "Q.THS_HW.TOTAL.SCA.EMP_DC.IT+DE+FR+ES") %>%
add_column(var="hours")
Gross Domestic Product
We use the namq_10_gdp database from Eurostat to obtain these data.
gdp <-
rdb("Eurostat","namq_10_gdp",mask = "Q.CLV10_MEUR.SCA.B1GQ.IT+DE+FR+ES") %>%
add_column(var="gdp")
Consumption
We use the namq_10_gdp database from Eurostat to retrieve these data.
conso <-
rdb("Eurostat","namq_10_gdp",mask = "Q.CLV10_MEUR.SCA.P31_S14_S15.IT+DE+FR+ES") %>%
add_column(var="conso")
Investment
We use the namq_10_gdp database from Eurostat to obtain these data.
inves <-
rdb("Eurostat","namq_10_gdp",mask = "Q.CLV10_MEUR.SCA.P51G.IT+DE+FR+ES") %>%
add_column(var="inves")
GDP Deflator
We use the namq_10_gdp database from Eurostat to retrieve these data.
defgdp <-
rdb("Eurostat","namq_10_gdp",mask = "Q.PD10_EUR.SCA.B1GQ.IT+DE+FR+ES") %>%
add_column(var="defgdp")
Investment Deflator
We use the namq_10_gdp database from Eurostat to obtain these data.
definves <-
rdb("Eurostat","namq_10_gdp",mask = "Q.PD10_EUR.SCA.P51G.IT+DE+FR+ES") %>%
add_column(var="definves")
Population
Given the scarcity of data to build long quarterly series for this variable, we use 2 databases from Eurostat that we will chain and interpolate by country in the next section: lfsq_pganws & demo_pjanbroad.
pop_recent <-
rdb("Eurostat","lfsq_pganws",mask = "Q.THS.T.TOTAL.Y15-64.POP.IT+DE+FR+ES") %>%
add_column(var="pop_recent")
pop_old <-
rdb("Eurostat","demo_pjanbroad",mask = "A.NR.Y15-64.T.IT+DE+FR+ES") %>%
add_column(var="pop_old")
Government Consumption
Given the scarcity of data to build long quarterly series for Italy and Germany, we use 2 databases from Eurostat that we will chain and interpolate by country in the next section: gov_10q_ggnfa & gov_10a_main.
pubcons_recent_fr <-
rdb("Eurostat","gov_10q_ggnfa",mask = "Q.MIO_EUR.SCA.S13.P3.FR") %>%
add_column(var="pubcons_recent")
pubcons_recent_it_de_es_nsa <-
rdb("Eurostat","gov_10q_ggnfa",mask = "Q.MIO_EUR.NSA.S13.P3.IT+DE+ES") %>%
add_column(var="pubcons_recent")
pubcons_old_it_de <-
rdb("Eurostat","gov_10a_main",mask = "A.MIO_EUR.S13.P3.IT+DE") %>%
add_column(var="pubcons_old")
Since the data retrieved from the gov_10q_ggnfa database is not seasonally adjusted for Italy and Germany, we use the seasonal package developed by Sax (2016):
df_nsa_q <-
pubcons_recent_it_de_es_nsa %>%
select(period,country=geo, value)
to_deseason <-
df_nsa_q %>%
spread(country, value)
deseasoned_q <-
bind_rows(lapply(unique(df_nsa_q$country),
function(var) deseason(var_arrange = var,
source_df = to_deseason)))%>%
mutate(Origin = "Adjusted Series",country=var)%>%
select(-var)
df_nsa_q %<>% mutate(Origin = "Unadjusted Series")
plot_df <-
bind_rows(df_nsa_q,deseasoned_q) %>%
na.omit()
ggplot(plot_df,aes(period,value,colour=Origin))+
geom_line()+
facet_wrap(~country ,scales ="free_y",ncol = 2)+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL) +
theme(legend.title=element_blank()) +
theme(strip.text = element_text(size=12)) +
theme(plot.title = element_text(size=16)) +
ggtitle("Government Consumption")
pubcons_recent_it_de_es <- deseasoned_q %>% filter(Origin=="Adjusted Series") %>% select(country,-Origin,value,period) %>% mutate(var="pubcons_recent")
Government Investment
Given the scarcity of data to build long quarterly series for Italy and Germany, we use 2 databases from Eurostat that we will chain and interpolate by country in the next section: gov_10q_ggnfa & gov_10a_main.
pubinves_recent_fr <-
rdb("Eurostat","gov_10q_ggnfa",mask = "Q.MIO_EUR.SCA.S13.P51G.FR") %>%
add_column(var="pubinves_recent")
pubinves_recent_it_de_es_nsa <-
rdb("Eurostat","gov_10q_ggnfa",mask = "Q.MIO_EUR.NSA.S13.P51G.IT+DE+ES") %>%
add_column(var="pubinves_recent")
pubinves_old_it_de <-
rdb("Eurostat","gov_10a_main",mask = "A.MIO_EUR.S13.P51G.IT+DE") %>%
add_column(var="pubinves_old")
Since the data retrieved from the gov_10q_ggnfa database is not seasonally adjusted for Italy and Germany, we use the seasonal package developed by Sax (2016):
df_nsa_q <-
pubinves_recent_it_de_es_nsa %>%
select(period,country=geo, value)
to_deseason <-
df_nsa_q %>%
spread(country, value)
deseasoned_q <-
bind_rows(lapply(unique(df_nsa_q$country),
function(var) deseason(var_arrange = var,
source_df = to_deseason)))%>%
mutate(Origin = "Adjusted Series",country=var)%>%
select(-var)
df_nsa_q %<>% mutate(Origin = "Unadjusted Series")
plot_df <-
bind_rows(df_nsa_q,deseasoned_q) %>%
na.omit()
ggplot(plot_df,aes(period,value,colour=Origin))+
geom_line()+
facet_wrap(~country ,scales ="free_y",ncol = 2)+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL) +
theme(legend.title=element_blank()) +
theme(strip.text = element_text(size=12)) +
theme(plot.title = element_text(size=16)) +
ggtitle("Government Investment")
pubinves_recent_it_de_es <- deseasoned_q %>% filter(Origin=="Adjusted Series") %>% select(country,-Origin,value,period) %>% mutate(var="pubinves_recent")
Government Social Transfers
Given the scarcity of data to build long quarterly series for Italy and Germany, we use 2 databases from Eurostat that we will chain and interpolate by country in the next section: gov_10q_ggnfa & gov_10a_main.
tfs_recent_fr <-
rdb("Eurostat","gov_10q_ggnfa",mask = "Q.MIO_EUR.SCA.S13.D62PAY.FR") %>%
add_column(var="tfs_recent")
tfs_recent_it_de_es_nsa <-
rdb("Eurostat","gov_10q_ggnfa",mask = "Q.MIO_EUR.NSA.S13.D62PAY.IT+DE+ES") %>%
add_column(var="tfs_recent")
tfs_old_it_de <-
rdb("Eurostat","gov_10a_main",mask = "A.MIO_EUR.S13.D62PAY.IT+DE") %>%
add_column(var="tfs_old")
Since the data retrieved from the gov_10q_ggnfa database is not seasonally adjusted for Italy and Germany, we use the seasonal package developed by Sax (2016):
df_nsa_q <-
tfs_recent_it_de_es_nsa %>%
select(period,country=geo, value)
to_deseason <-
df_nsa_q %>%
spread(country, value)
deseasoned_q <-
bind_rows(lapply(unique(df_nsa_q$country),
function(var) deseason(var_arrange = var,
source_df = to_deseason)))%>%
mutate(Origin = "Adjusted Series",country=var)%>%
select(-var)
df_nsa_q %<>% mutate(Origin = "Unadjusted Series")
plot_df <-
bind_rows(df_nsa_q,deseasoned_q) %>%
na.omit()
ggplot(plot_df,aes(period,value,colour=Origin))+
geom_line()+
facet_wrap(~country ,scales ="free_y",ncol = 2)+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL) +
theme(legend.title=element_blank()) +
theme(strip.text = element_text(size=12)) +
theme(plot.title = element_text(size=16)) +
ggtitle("Government Social Transfers")
tfs_recent_it_de_es <- deseasoned_q %>% filter(Origin=="Adjusted Series") %>% select(country,-Origin,value,period) %>% mutate(var="tfs_recent")
Government Interest Payments
Given the scarcity of data to build long quarterly series for Italy and Germany, we use 2 databases from Eurostat that we will chain and interpolate by country in the next section: gov_10q_ggnfa & gov_10a_main.
intpay_recent_fr <-
rdb("Eurostat","gov_10q_ggnfa",mask = "Q.MIO_EUR.SCA.S13.D41PAY.FR") %>%
add_column(var="intpay_recent")
intpay_recent_it_de_es_nsa <-
rdb("Eurostat","gov_10q_ggnfa",mask = "Q.MIO_EUR.NSA.S13.D41PAY.IT+DE+ES") %>%
add_column(var="intpay_recent")
intpay_old_it_de <-
rdb("Eurostat","gov_10a_main",mask = "A.MIO_EUR.S13.D41PAY.DE+IT") %>%
add_column(var="intpay_old")
Since the data retrieved from the gov_10q_ggnfa database is not seasonally adjusted for Italy and Germany, we use the seasonal package developed by Sax (2016):
df_nsa_q <-
intpay_recent_it_de_es_nsa %>%
select(period,country=geo, value)
to_deseason <-
df_nsa_q %>%
spread(country, value)
deseasoned_q <-
bind_rows(lapply(unique(df_nsa_q$country),
function(var) deseason(var_arrange = var,
source_df = to_deseason)))%>%
mutate(Origin = "Adjusted Series",country=var)%>%
select(-var)
df_nsa_q %<>% mutate(Origin = "Unadjusted Series")
plot_df <-
bind_rows(df_nsa_q,deseasoned_q) %>%
na.omit()
ggplot(plot_df,aes(period,value,colour=Origin))+
geom_line()+
facet_wrap(~country ,scales ="free_y",ncol = 2)+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL) +
theme(legend.title=element_blank()) +
theme(strip.text = element_text(size=12)) +
theme(plot.title = element_text(size=16)) +
ggtitle("Government Interest Payments")
intpay_recent_it_de_es <- deseasoned_q %>% filter(Origin=="Adjusted Series") %>% select(country,-Origin,value,period) %>% mutate(var="intpay_recent")
Total Government Expenditure
Given the scarcity of data to build long quarterly series for Italy and Germany, we use 2 databases from Eurostat that we will chain and interpolate by country in the next section: gov_10q_ggnfa & gov_10a_main.
totexp_recent_fr <-
rdb("Eurostat","gov_10q_ggnfa",mask = "Q.MIO_EUR.SCA.S13.TE.FR") %>%
add_column(var="totexp_recent")
totexp_recent_it_de_es_nsa <-
rdb("Eurostat","gov_10q_ggnfa",mask = "Q.MIO_EUR.NSA.S13.TE.IT+DE+ES") %>%
add_column(var="totexp_recent")
totexp_old_it_de <-
rdb("Eurostat","gov_10a_main",mask = "A.MIO_EUR.S13.TE.DE+IT") %>%
add_column(var="totexp_old")
Since the data retrieved from the gov_10q_ggnfa database is not seasonally adjusted for Italy, Germany, and Spain, we use the seasonal package developed by Sax (2016):
df_nsa_q <-
totexp_recent_it_de_es_nsa %>%
select(period,country=geo, value)
to_deseason <-
df_nsa_q %>%
spread(country, value)
deseasoned_q <-
bind_rows(lapply(unique(df_nsa_q$country),
function(var) deseason(var_arrange = var,
source_df = to_deseason)))%>%
mutate(Origin = "Adjusted Series",country=var)%>%
select(-var)
df_nsa_q %<>% mutate(Origin = "Unadjusted Series")
plot_df <-
bind_rows(df_nsa_q,deseasoned_q) %>%
na.omit()
ggplot(plot_df,aes(period,value,colour=Origin))+
geom_line()+
facet_wrap(~country ,scales ="free_y",ncol = 2)+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL) +
theme(legend.title=element_blank()) +
theme(strip.text = element_text(size=12)) +
theme(plot.title = element_text(size=16)) +
ggtitle("Total Government Expenditure")
totexp_recent_it_de_es <- deseasoned_q %>% filter(Origin=="Adjusted Series") %>% select(country,-Origin,value,period) %>% mutate(var="totexp_recent")
Total Government Revenue
Given the scarcity of data to build long quarterly series for Italy and Germany, we use 2 databases from Eurostat that we will chain and interpolate by country in the next section: gov_10q_ggnfa & gov_10a_main.
totrev_recent_fr <-
rdb("Eurostat","gov_10q_ggnfa",mask = "Q.MIO_EUR.SCA.S13.TR.FR") %>%
add_column(var="totrev_recent")
totrev_recent_it_de_es_nsa <-
rdb("Eurostat","gov_10q_ggnfa",mask = "Q.MIO_EUR.NSA.S13.TR.IT+DE+ES") %>%
add_column(var="totrev_recent")
totrev_old_it_de <-
rdb("Eurostat","gov_10a_main",mask = "A.MIO_EUR.S13.TR.DE+IT") %>%
add_column(var="totrev_old")
Since the data retrieved from the gov_10q_ggnfa database is not seasonally adjusted for Italy, Germany, and Spain, we use the seasonal package developed by Sax (2016):
df_nsa_q <-
totrev_recent_it_de_es_nsa %>%
select(period,country=geo, value)
to_deseason <-
df_nsa_q %>%
spread(country, value)
deseasoned_q <-
bind_rows(lapply(unique(df_nsa_q$country),
function(var) deseason(var_arrange = var,
source_df = to_deseason)))%>%
mutate(Origin = "Adjusted Series",country=var)%>%
select(-var)
df_nsa_q %<>% mutate(Origin = "Unadjusted Series")
plot_df <-
bind_rows(df_nsa_q,deseasoned_q) %>%
na.omit()
ggplot(plot_df,aes(period,value,colour=Origin))+
geom_line()+
facet_wrap(~country ,scales ="free_y",ncol = 2)+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL) +
theme(legend.title=element_blank()) +
theme(strip.text = element_text(size=12)) +
theme(plot.title = element_text(size=16)) +
ggtitle("Total Government Revenue")
totrev_recent_it_de_es <- deseasoned_q %>% filter(Origin=="Adjusted Series") %>% select(country,-Origin,value,period) %>% mutate(var="totrev_recent")
Government Debt
Given the scarcity of data to build long quarterly series for this variable, we use 2 databases, one from Eurostat and the other from the IMF, that we will chain and interpolate by country in the next section: gov_10q_ggdebt & WEO, respectively.
debt_recent <-
rdb("Eurostat","gov_10q_ggdebt",mask = "Q.GD.S13.MIO_EUR.IT+DE+FR+ES") %>%
add_column(var="debt_recent")
debt_old <-
rdb(provider_code = "IMF", dataset_code = "WEO:latest", mask = "DEU+ESP+FRA+ITA.GGXWDG") %>%
add_column(var="debt_old") %>%
select(geo='weo-country',period,value,var) %>%
mutate(geo=str_sub(geo,1,2)) %>%
filter(year(period)<=max(debt_recent$period))
df_nsa_q <-
debt_recent %>%
select(period,country=geo, value)
to_deseason <-
df_nsa_q %>%
spread(country, value)
deseasoned_q <-
bind_rows(lapply(unique(df_nsa_q$country),
function(var) deseason(var_arrange = var,
source_df = to_deseason)))%>%
mutate(Origin = "Adjusted Series",country=var)%>%
select(-var)
df_nsa_q %<>% mutate(Origin = "Unadjusted Series")
plot_df <-
bind_rows(df_nsa_q,deseasoned_q) %>%
na.omit()
ggplot(plot_df,aes(period,value,colour=Origin))+
geom_line()+
facet_wrap(~country ,scales ="free_y",ncol = 2)+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL) +
theme(legend.title=element_blank()) +
theme(strip.text = element_text(size=12)) +
theme(plot.title = element_text(size=16)) +
ggtitle("Debt")
debt_recent <- deseasoned_q %>% filter(Origin=="Adjusted Series") %>% select(country,-Origin,value,period) %>% mutate(var="debt_recent")
Loans to Non-Financial Corporations
We use the CNFS database from the Bank for International Settlements to obtain these data.
loans_nfc <-
rdb("BIS","total_credit",mask = "Q.IT+DE+FR+ES.N.A.M.XDC.A") %>%
add_column(var="loans_nfc") %>%
select(geo=BORROWERS_CTY,period,value,var)
Entrepreneurial net worth
networth <-
rdb("OECD","MEI", mask = "FRA+DEU+ITA+ESP.SPASTT01.IXOB.Q") %>%
select(period,value,geo=LOCATION) %>%
add_column(var="networth") %>%
mutate(geo=case_when(
geo=="FRA" ~ "FR",
geo=="DEU" ~ "DE",
geo=="ESP" ~ "ES",
geo=="ITA" ~ "IT"))
Short-term interest rate
shortrate <-
rdb("OECD","MEI", mask = "FRA+DEU+ITA+ESP.IR3TIB01.ST.Q") %>%
select(period,value,geo=LOCATION) %>%
add_column(var="shortrate") %>%
mutate(geo=case_when(
geo=="FRA" ~ "FR",
geo=="DEU" ~ "DE",
geo=="ESP" ~ "ES",
geo=="ITA" ~ "IT"))
Lending Rate
Given the scarcity of data to build long quarterly series for this variable, we use 2 databases, one from the European Central Bank and the other from the IMF, that we will chain and interpolate by country in the next section: MIR & IFS, respectively.
lendingrate_recent <-
rdb("ECB","MIR",mask = "M.IT+DE+FR+ES.B.A2A.A.R.A.2240.EUR.N") %>%
select(geo=REF_AREA,period,value) %>%
mutate(period=paste(year(period),quarter(period))) %>%
group_by(geo,period) %>%
summarize(value=mean(value,na.rm = T)) %>%
mutate(var="lendingrate_recent",
period=yq(period))
lendingrate_old <-
rdb("IMF","IFS",mask = "Q.IT+DE+FR+ES.FILR_PA") %>%
add_column(var="lendingrate_old") %>%
select(geo=REF_AREA,period,value,var)
World Demand
We use the foreign demand specific to the four countries, that we also built for this project : https://macro.cepremap.fr/article/2020-02/foreign-demand-euro-countries-data/
world_demand <-
read_csv("https://shiny.cepremap.fr/data/Foreign_demand.csv") %>%
rename(geo=country)
Unemployment benefits
Given the scarcity of data to build long quarterly series for this variable, we proceed in this way: we first determine quarterly series using the ratio of quarterly social expenditures on annual unemployment benefits, and then we use the seasonal package developed by Sax to remove the seasonal componenet from the series. Then we retrieve annual data on unemployment benefits using the spr_exp_sum database from Eurostat.
In the first place, we retrieve government social expenditures and compute their quaterly share for each year:
filter <- "Q.MIO_EUR.NSA.S13.D62PAY.IT+DE+FR+ES"
df <- rdb("Eurostat","gov_10q_ggnfa",mask = filter)
socialexp <-
df %>%
mutate(year=year(period),country=geo) %>%
select(period,value,year,country) %>%
group_by(year,country) %>%
mutate(sum=sum(value),
ratio=value/sum) %>%
ungroup() %>%
select(-value,-year,-sum)
Then, we retrieve the latest annual data on unemployment benefits, put them in a quarterly table and use the previous ratio of quarterly social expenditures to compute quarterly unemployment benefits:
filter <- "A.MIO_EUR.S13.GF1005.TE.IT+DE+FR+ES"
df <- rdb("Eurostat","gov_10a_exp",mask = filter)
recent_unempbenef <-
df %>%
mutate(year=year(period),country=geo) %>%
select(period,value,year,country) %>%
spread(country, value)
recent_unempbenef_q <-
tibble(period=seq(min(recent_unempbenef$period),
length.out=nrow(recent_unempbenef)*4,
by = "quarter"),
year=year(period)) %>%
left_join(recent_unempbenef,by="year") %>%
select(-period.y,-year) %>%
rename(period=period.x) %>%
gather(country,value,-period)
unempbenef_q <-
recent_unempbenef_q %>%
inner_join(socialexp,by=c("period"="period","country"="country")) %>%
mutate(value=value*ratio) %>%
select(-ratio)%>%
na.omit()
Since we have unadjusted data, we use the seasonal package developed by Sax (2016) on these data:
to_deseason <-
unempbenef_q %>%
spread(country, value)
unempbenef_q_deseasoned <-
bind_rows(lapply(unique(unempbenef_q$country),
function(var) deseason(var_arrange = var,
source_df = to_deseason)))%>%
mutate(Origin = "Adjusted Series",country=var)%>%
select(-var)
unempbenef_q %<>% mutate(Origin = "Unadjusted Series")
plot_unempbenef <-
bind_rows(unempbenef_q,unempbenef_q_deseasoned) %>%
na.omit()
ggplot(plot_unempbenef,aes(period,value,colour=Origin))+
geom_line()+
facet_wrap(~country ,scales ="free_y",ncol = 2)+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL) +
theme(legend.title=element_blank()) +
theme(strip.text = element_text(size=12)) +
theme(plot.title = element_text(size=16)) +
ggtitle("Unemployment benefits")
unempbenef_recent <- unempbenef_q_deseasoned %>% filter(Origin=="Adjusted Series") %>% select(geo=country,-Origin,value,period) %>% mutate(var="unempbenef_recent")
Now, using the spr_exp_sum database from Eurostat we retrieve the annual series, that will be interpolated and chained to the quarterly series in the next section.
filter <- "A.UNEMPLOY.MIO_EUR.IT+DE+FR+ES"
df <- rdb("Eurostat","spr_exp_sum",mask=filter)
unempbenef_old <-
df %>%
add_column(var="unempbenef_old") %>%
select(period,value,geo,var)
Nominal effective exchange rate
df <- rdb("BIS","eer",mask="M.N.B.IT+DE+FR+ES")
neer <-
df %>%
select(period,value,geo=REF_AREA) %>%
mutate(period=paste(year(period),quarter(period))) %>%
group_by(geo,period) %>%
summarize(value=mean(value)) %>%
mutate(period=yq(period),
var="neer")
Imports and exports
df <- rdb("OECD","EO",mask = "FRA+DEU+ITA+ESP.MGSV+XGSV.Q")
imports_exports_volume <-
df %>%
select(period,value,geo=LOCATION,value,var=VARIABLE) %>%
mutate(var=case_when(
var=="MGSV" ~ "imports",
var=="XGSV" ~ "exports")) %>%
mutate(geo=case_when(
geo=="FRA" ~ "FR",
geo=="DEU" ~ "DE",
geo=="ESP" ~ "ES",
geo=="ITA" ~ "IT"))
Merging data frames
We merge all the data frames with all the series, to chain and interpolate the special cases in the following section. The figure below shows the unchained series for France, Germany, Italy and Spain.
df <-
bind_rows(wage_de_fr,wage_es_it,
hours,
gdp,
conso,
inves,
defgdp,
definves,
pop_recent,pop_old,
debt_old,
lendingrate_recent,lendingrate_old,
networth,
shortrate,
loans_nfc,
totexp_recent_fr, totexp_old_it_de,
totrev_recent_fr, totrev_old_it_de,
intpay_recent_fr, intpay_old_it_de,
pubcons_recent_fr, pubcons_old_it_de,
pubinves_recent_fr, pubinves_old_it_de,
tfs_recent_fr,tfs_old_it_de,
world_demand,
unempbenef_recent,unempbenef_old,
neer,
imports_exports_volume) %>%
select(period,value,country=geo,var) %>%
bind_rows(pubcons_recent_it_de_es, pubinves_recent_it_de_es, tfs_recent_it_de_es, debt_recent,
totexp_recent_it_de_es, intpay_recent_it_de_es, totrev_recent_it_de_es) %>%
na.omit() %>%
filter(year(period)>=1991) %>%
arrange(var, period)
ggplot(df,aes(period,value,colour=country))+
geom_line()+
facet_wrap(~var,ncol=3,scales = "free_y")+
scale_x_date(expand = c(0.01,0.01)) +
theme + theme(strip.text=element_text(size=11)) +
xlab(NULL) + ylab(NULL)+
ggtitle("Unchained Series")
France: Chaining & Interpolating Data
Before chaining and interpolating the special cases, we first gather all data for France, and then we proceed case by case.
df_fr <- df %>% filter(country=="FR") %>% select(-country)
Government Debt
We first interpolate the annual series in order to obtain quarterly series, and then we chain the quarterly series from the two different databases.
debt_old_a <-
df_fr %>%
filter(var=="debt_old") %>%
mutate(value=1000*value) %>%
select(-var)
debt_old_q <-
tibble(period=seq(as.Date("1991-01-01"),
length.out = (nrow(debt_old_a)-1)*4+1,
by = "quarter"),
value=NA) %>%
left_join(debt_old_a, by="period") %>%
select(-value.x) %>%
rename(value=value.y) %>%
mutate(value=na.spline(value),
var="debt")
debt_recent <-
df_fr %>%
filter(var=="debt_recent") %>%
mutate(var="debt")
minDateDebtRecent <- min(debt_recent$period)
debt <-
chain(basis = debt_recent,
to_rebase= debt_old_q,
date_chain=minDateDebtRecent) %>%
mutate(var="debt")
plot_df <- bind_rows(add_column(debt_old_a,var="debt_old_a"),
mutate(debt_old_q,var="debt_old_q"),
mutate(debt,var="debt_chained"))
ggplot(plot_df,aes(period,value,col=var))+
geom_line()+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL)+
ggtitle("Government Debt")
Population
We first interpolate the annual series in order to obtain quarterly series, and then we chain the quarterly series from the two different databases.
pop_old_a <-
df_fr %>%
filter(var=="pop_old")
pop_old_q <-
tibble(period=seq(as.Date("1991-01-01"),
length.out = (nrow(pop_old_a)-1)*4+1,
by = "quarter"),
value=NA) %>%
left_join(pop_old_a, by="period") %>%
select(-value.x) %>%
rename(value=value.y) %>%
mutate(value=na.spline(value),
var="pop")
pop_recent <-
df_fr %>%
filter(var=="pop_recent") %>%
mutate(var="pop",value=value*1000)
pop <- chain(basis = pop_recent,
to_rebase= pop_old_q,
date_chain="2015-01-01")
plot_df <- bind_rows(mutate(pop_old_a,var="pop_old_a"),
mutate(pop_old_q,var="pop_old_q"),
mutate(pop,var="pop_chained"))
ggplot(plot_df,aes(period,value,col=var))+
geom_line()+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL)+
ggtitle("Population")
Lending Rate
We chain the quarterly series from the two different databases.
lendingrate_old <-
df_fr %>%
filter(var=="lendingrate_old") %>%
mutate(var="lendingrate")
lendingrate_recent <-
df_fr %>%
filter(var=="lendingrate_recent") %>%
mutate(var="lendingrate")
minDateLendingRateRecent <- min(lendingrate_recent$period)
lendingrate <-
chain(basis = lendingrate_recent,
to_rebase= lendingrate_old,
date_chain=minDateLendingRateRecent) %>%
mutate(var="lendingrate")
plot_df <- bind_rows(mutate(lendingrate_old,var="lendingrate_old"),
mutate(lendingrate_recent,var="lendingrate_recent"),
mutate(lendingrate,var="lendingrate_chained"))
ggplot(plot_df,aes(period,value,col=var))+
geom_line()+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL)+
ggtitle("Lending Rate")
Unemployment Benefits
We first interpolate the annual series in order to obtain quarterly series, and then we chain the two quarterly series.
unempbenef_old_a <-
df_fr %>%
filter(var=="unempbenef_old") %>%
mutate(value=value/4)
unempbenef_old_q <-
tibble(period=seq(as.Date("1991-01-01"),
length.out = (nrow(unempbenef_old_a)-1)*4+1,
by = "quarter"),
value=NA) %>%
left_join(unempbenef_old_a, by="period") %>%
select(-value.x,value=value.y) %>%
mutate(value=na.spline(value),
var="unempbenef")
unempbenef_recent <-
df_fr %>%
filter(var=="unempbenef_recent") %>%
mutate(var="unempbenef")
unempbenef <- chain(basis = unempbenef_recent,
to_rebase= unempbenef_old_q,
date_chain="2015-01-01")
plot_df <- bind_rows(mutate(unempbenef_old_a,var="unempbenef_old_a"),
mutate(unempbenef_old_q,var="unempbenef_old_q"),
mutate(unempbenef,var="unempbenef_chained"))
ggplot(plot_df,aes(period,value,col=var))+
geom_line()+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL)+
ggtitle("Unemployment Benefits")
Merging French Data
We gather all the final series in a dataframe.
FR_rawdata <-
df_fr %>%
filter(! var %in% c("lendingrate_old","lendingrate_recent",
"pop_old","pop_recent",
"debt_old","debt_recent",
"unempbenef_old","unempbenef_recent")) %>%
bind_rows(lendingrate,pop,debt,unempbenef) %>%
mutate(var=
case_when(var=="pubcons_recent" ~ "pubcons",
var=="pubinves_recent" ~ "pubinves",
var=="tfs_recent" ~ "tfs",
var=="totexp_recent" ~ "totexp",
var=="intpay_recent" ~ "intpay",
var=="totrev_recent" ~ "totrev",
TRUE ~ var)) %>%
spread(var,value) %>%
add_column(country="FR")
Spain: Chaining & Interpolating Data
Before chaining and interpolating the special cases, we first gather all data for France, and then we proceed case by case.
df_es <- df %>% filter(country=="ES") %>% select(-country)
Government Debt
We first interpolate the annual series in order to obtain quarterly series, and then we chain the quarterly series from the two different databases.
debt_old_a <-
df_es %>%
filter(var=="debt_old") %>%
mutate(value=1000*value) %>%
select(-var)
debt_old_q <-
tibble(period=seq(as.Date("1991-01-01"),
length.out = (nrow(debt_old_a)-1)*4+1,
by = "quarter"),
value=NA) %>%
left_join(debt_old_a, by="period") %>%
select(-value.x) %>%
rename(value=value.y) %>%
mutate(value=na.spline(value),
var="debt")
debt_recent <-
df_es %>%
filter(var=="debt_recent") %>%
mutate(var="debt")
minDateDebtRecent <- min(debt_recent$period)
debt <-
chain(basis = debt_recent,
to_rebase= debt_old_q,
date_chain=minDateDebtRecent) %>%
mutate(var="debt")
plot_df <- bind_rows(add_column(debt_old_a,var="debt_old_a"),
mutate(debt_old_q,var="debt_old_q"),
mutate(debt,var="debt_chained"))
ggplot(plot_df,aes(period,value,col=var))+
geom_line()+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL)+
ggtitle("Government Debt")
Population
We first interpolate the annual series in order to obtain quarterly series, and then we chain the quarterly series from the two different databases.
pop_old_a <-
df_es %>%
filter(var=="pop_old")
pop_old_q <-
tibble(period=seq(as.Date("1991-01-01"),
length.out = (nrow(pop_old_a)-1)*4+1,
by = "quarter"),
value=NA) %>%
left_join(pop_old_a, by="period") %>%
select(-value.x) %>%
rename(value=value.y) %>%
mutate(value=na.spline(value),
var="pop")
pop_recent <-
df_es %>%
filter(var=="pop_recent") %>%
mutate(var="pop",value=value*1000)
pop <- chain(basis = pop_recent,
to_rebase= pop_old_q,
date_chain="1998-01-01")
plot_df <- bind_rows(mutate(pop_old_a,var="pop_old_a"),
mutate(pop_old_q,var="pop_old_q"),
mutate(pop,var="pop_chained"))
ggplot(plot_df,aes(period,value,col=var))+
geom_line()+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL)
Lending Rate
We chain the quarterly series from the two different databases.
lendingrate_old <-
df_es %>%
filter(var=="lendingrate_old") %>%
mutate(var="lendingrate")
lendingrate_recent <-
df_es %>%
filter(var=="lendingrate_recent") %>%
mutate(var="lendingrate")
minDateLendingRateRecent <- min(lendingrate_recent$period)
lendingrate <-
chain(basis = lendingrate_recent,
to_rebase= lendingrate_old,
date_chain=minDateLendingRateRecent) %>%
mutate(var="lendingrate")
plot_df <- bind_rows(mutate(lendingrate_old,var="lendingrate_old"),
mutate(lendingrate_recent,var="lendingrate_recent"),
mutate(lendingrate,var="lendingrate_chained"))
ggplot(plot_df,aes(period,value,col=var))+
geom_line()+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL)+
ggtitle("Lending Rate")
Unemployment Benefits
We first interpolate the annual series in order to obtain quarterly series, and then we chain the two quarterly series.
unempbenef_old_a <-
df_es %>%
filter(var=="unempbenef_old") %>%
mutate(value=value/4)
unempbenef_old_q <-
tibble(period=seq(as.Date("1991-01-01"),
length.out = (nrow(unempbenef_old_a)-1)*4+1,
by = "quarter"),
value=NA) %>%
left_join(unempbenef_old_a, by="period") %>%
select(-value.x,value=value.y) %>%
mutate(value=na.spline(value),
var="unempbenef")
unempbenef_recent <-
df_es %>%
filter(var=="unempbenef_recent") %>%
mutate(var="unempbenef")
unempbenef <- chain(basis = unempbenef_recent,
to_rebase= unempbenef_old_q,
date_chain="2015-01-01")
plot_df <- bind_rows(mutate(unempbenef_old_a,var="unempbenef_old_a"),
mutate(unempbenef_old_q,var="unempbenef_old_q"),
mutate(unempbenef,var="unempbenef_chained"))
ggplot(plot_df,aes(period,value,col=var))+
geom_line()+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL)+
ggtitle("Unemployment Benefits")
Merging Spanish Data
We gather all the final series in a dataframe.
ES_rawdata <-
df_es %>%
filter(! var %in% c("lendingrate_old","lendingrate_recent",
"pop_old","pop_recent",
"debt_old","debt_recent",
"unempbenef_old","unempbenef_recent")) %>%
bind_rows(lendingrate,pop,debt,unempbenef) %>%
mutate(var=
case_when(var=="pubcons_recent" ~ "pubcons",
var=="pubinves_recent" ~ "pubinves",
var=="tfs_recent" ~ "tfs",
var=="totexp_recent" ~ "totexp",
var=="totrev_recent" ~ "totrev",
var=="intpay_recent" ~ "intpay",
TRUE ~ var)) %>%
spread(var,value) %>%
add_column(country="ES")
Germany: Chaining & Interpolating Data
Before chaining and interpolating the special cases, we first gather all data for France, and then we proceed case by case.
df_de <- df %>% filter(country=="DE") %>% select(-country)
Government Debt
We first interpolate the annual series in order to obtain quarterly series, and then we chain the quarterly series from the two different databases.
debt_old_a <-
df_de %>%
filter(var=="debt_old") %>%
mutate(value=1000*value) %>%
select(-var)
debt_old_q <-
tibble(period=seq(as.Date("1991-01-01"),
length.out = (nrow(debt_old_a)-1)*4+1,
by = "quarter"),
value=NA) %>%
left_join(debt_old_a, by="period") %>%
select(-value.x) %>%
rename(value=value.y) %>%
mutate(value=na.spline(value),
var="debt")
debt_recent <-
df_de %>%
filter(var=="debt_recent") %>%
mutate(var="debt")
minDateDebtRecent <- min(debt_recent$period)
debt <-
chain(basis = debt_recent,
to_rebase= debt_old_q,
date_chain=minDateDebtRecent) %>%
mutate(var="debt")
plot_df <- bind_rows(add_column(debt_old_a,var="debt_old_a"),
mutate(debt_old_q,var="debt_old_q"),
mutate(debt,var="debt_chained"))
ggplot(plot_df,aes(period,value,col=var))+
geom_line()+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL)+
ggtitle("Government Debt")
Population
We first interpolate the annual series in order to obtain quarterly series, and then we chain the quarterly series from the two different databases.
pop_old_a <-
df_de %>%
filter(var=="pop_old")
pop_old_q <-
tibble(period=seq(as.Date("1991-01-01"),
length.out = (nrow(pop_old_a)-1)*4+1,
by = "quarter"),
value=NA) %>%
left_join(pop_old_a, by="period") %>%
select(-value.x) %>%
rename(value=value.y) %>%
mutate(value=na.spline(value),
var="pop")
pop_recent <-
df_de %>%
filter(var=="pop_recent") %>%
mutate(var="pop",value=value*1000)
pop <- chain(basis = pop_recent,
to_rebase= pop_old_q,
date_chain="2015-01-01")
plot_df <- bind_rows(mutate(pop_old_a,var="pop_old_a"),
mutate(pop_old_q,var="pop_old_q"),
mutate(pop,var="pop_chained"))
ggplot(plot_df,aes(period,value,col=var))+
geom_line()+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL)+
ggtitle("Population")
Lending Rate
We chain the quarterly series from the two different databases.
lendingrate_old <-
df_de %>%
filter(var=="lendingrate_old") %>%
mutate(var="lendingrate")
lendingrate_recent <-
df_de %>%
filter(var=="lendingrate_recent") %>%
mutate(var="lendingrate")
minDateLendingRateRecent <- min(lendingrate_recent$period)
lendingrate <-
chain(basis = lendingrate_recent,
to_rebase= lendingrate_old,
date_chain=minDateLendingRateRecent) %>%
mutate(var="lendingrate")
plot_df <- bind_rows(mutate(lendingrate_old,var="lendingrate_old"),
mutate(lendingrate_recent,var="lendingrate_recent"),
mutate(lendingrate,var="lendingrate_chained"))
ggplot(plot_df,aes(period,value,col=var))+
geom_line()+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL)+
ggtitle("Lending Rate")
Unemployment Benefits
We first interpolate the annual series in order to obtain quarterly series, and then we chain the two quarterly series.
unempbenef_old_a <-
df_de %>%
filter(var=="unempbenef_old") %>%
mutate(value=value/4)
unempbenef_old_q <-
tibble(period=seq(as.Date("1991-01-01"),
length.out = (nrow(unempbenef_old_a)-1)*4+1,
by = "quarter"),
value=NA) %>%
left_join(unempbenef_old_a, by="period") %>%
select(-value.x,value=value.y) %>%
mutate(value=na.spline(value),
var="unempbenef")
unempbenef_recent <-
df_de %>%
filter(var=="unempbenef_recent") %>%
mutate(var="unempbenef")
unempbenef <- chain(basis = unempbenef_recent,
to_rebase= unempbenef_old_q,
date_chain="2015-01-01")
plot_df <- bind_rows(mutate(unempbenef_old_a,var="unempbenef_old_a"),
mutate(unempbenef_old_q,var="unempbenef_old_q"),
mutate(unempbenef,var="unempbenef_chained"))
ggplot(plot_df,aes(period,value,col=var))+
geom_line()+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL)+
ggtitle("Unemployment Benefits")
Government Consumption
We first interpolate the annual series in order to obtain quarterly series, and then we chain the quarterly series from the two different databases.
pubcons_old_a <-
df_de %>%
filter(var=="pubcons_old") %>%
mutate(value=value/4)
pubcons_old_q <-
tibble(period=seq(as.Date("1995-01-01"),
length.out = (nrow(pubcons_old_a)-1)*4+1,
by = "quarter"),
value=NA) %>%
left_join(pubcons_old_a, by="period") %>%
select(-value.x,value=value.y) %>%
mutate(value=na.spline(value),
var="pubcons")
pubcons_recent <-
df_de %>%
filter(var=="pubcons_recent") %>%
mutate(var="pubcons")
minDatePubConsRecent <- min(pubcons_recent$period)
pubcons <- chain(basis = pubcons_recent,
to_rebase= pubcons_old_q,
date_chain=minDatePubConsRecent)
plot_df <- bind_rows(mutate(pubcons_old_a,var="pubcons_old_a"),
mutate(pubcons_old_q,var="pubcons_old_q"),
mutate(pubcons,var="pubcons_chained"))
ggplot(plot_df,aes(period,value,col=var))+
geom_line()+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL)+
ggtitle("Government Consumption")
Government Investment
We first interpolate the annual series in order to obtain quarterly series, and then we chain the quarterly series from the two different databases.
pubinves_old_a <-
df_de %>%
filter(var=="pubinves_old") %>%
mutate(value=value/4)
pubinves_old_q <-
tibble(period=seq(as.Date("1995-01-01"),
length.out = (nrow(pubinves_old_a)-1)*4+1,
by = "quarter"),
value=NA) %>%
left_join(pubinves_old_a, by="period") %>%
select(-value.x,value=value.y) %>%
mutate(value=na.spline(value),
var="pubinves")
pubinves_recent <-
df_de %>%
filter(var=="pubinves_recent") %>%
mutate(var="pubinves")
minDatePubInvesRecent <- min(pubinves_recent$period)
pubinves <- chain(basis = pubinves_recent,
to_rebase= pubinves_old_q,
date_chain=minDatePubInvesRecent)
plot_df <- bind_rows(mutate(pubinves_old_a,var="pubinves_old_a"),
mutate(pubinves_old_q,var="pubinves_old_q"),
mutate(pubinves,var="pubinves_chained"))
ggplot(plot_df,aes(period,value,col=var))+
geom_line()+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL)+
ggtitle("Government Investment")
Government Social Transfers
We first interpolate the annual series in order to obtain quarterly series, and then we chain the quarterly series from the two different databases.
tfs_old_a <-
df_de %>%
filter(var=="tfs_old") %>%
mutate(value=value/4)
tfs_old_q <-
tibble(period=seq(as.Date("1995-01-01"),
length.out = (nrow(tfs_old_a)-1)*4+1,
by = "quarter"),
value=NA) %>%
left_join(tfs_old_a, by="period") %>%
select(-value.x,value=value.y) %>%
mutate(value=na.spline(value),
var="tfs")
tfs_recent <-
df_de %>%
filter(var=="tfs_recent") %>%
mutate(var="tfs")
minDateTfsRecent <- min(tfs_recent$period)
tfs <- chain(basis = tfs_recent,
to_rebase= tfs_old_q,
date_chain=minDateTfsRecent)
plot_df <- bind_rows(mutate(tfs_old_a,var="tfs_old_a"),
mutate(tfs_old_q,var="tfs_old_q"),
mutate(tfs,var="tfs_chained"))
ggplot(plot_df,aes(period,value,col=var))+
geom_line()+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL)+
ggtitle("Government Social Transfers")
Total Government Expenditure
We first interpolate the annual series in order to obtain quarterly series, and then we chain the quarterly series from the two different databases.
totexp_old_a <-
df_de %>%
filter(var=="totexp_old") %>%
mutate(value=value/4)
totexp_old_q <-
tibble(period=seq(as.Date("1995-01-01"),
length.out = (nrow(totexp_old_a)-1)*4+1,
by = "quarter"),
value=NA) %>%
left_join(totexp_old_a, by="period") %>%
select(-value.x,value=value.y) %>%
mutate(value=na.spline(value),
var="totexp")
totexp_recent <-
df_de %>%
filter(var=="totexp_recent") %>%
mutate(var="totexp")
minDateTotExpRecent <- min(totexp_recent$period)
totexp <- chain(basis = totexp_recent,
to_rebase= totexp_old_q,
date_chain=minDateTotExpRecent)
plot_df <- bind_rows(mutate(totexp_old_a,var="totexp_old_a"),
mutate(totexp_old_q,var="totexp_old_q"),
mutate(totexp,var="totexp_chained"))
ggplot(plot_df,aes(period,value,col=var))+
geom_line()+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL)+
ggtitle("Total Government Expenditure")
Total Government Revenue
We first interpolate the annual series in order to obtain quarterly series, and then we chain the quarterly series from the two different databases.
totrev_old_a <-
df_de %>%
filter(var=="totrev_old") %>%
mutate(value=value/4)
totrev_old_q <-
tibble(period=seq(as.Date("1995-01-01"),
length.out = (nrow(totrev_old_a)-1)*4+1,
by = "quarter"),
value=NA) %>%
left_join(totrev_old_a, by="period") %>%
select(-value.x,value=value.y) %>%
mutate(value=na.spline(value),
var="totrev")
totrev_recent <-
df_de %>%
filter(var=="totrev_recent") %>%
mutate(var="totrev")
minDateTotRevRecent <- min(totrev_recent$period)
totrev <- chain(basis = totrev_recent,
to_rebase= totrev_old_q,
date_chain=minDateTotRevRecent)
plot_df <- bind_rows(mutate(totrev_old_a,var="totrev_old_a"),
mutate(totrev_old_q,var="totrev_old_q"),
mutate(totrev,var="totrev_chained"))
ggplot(plot_df,aes(period,value,col=var))+
geom_line()+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL)+
ggtitle("Total Government Revenue")
Government Interest Payments
We first interpolate the annual series in order to obtain quarterly series, and then we chain the quarterly series from the two different databases.
intpay_old_a <-
df_de %>%
filter(var=="intpay_old") %>%
mutate(value=value/4)
intpay_old_q <-
tibble(period=seq(as.Date("1995-01-01"),
length.out = (nrow(intpay_old_a)-1)*4+1,
by = "quarter"),
value=NA) %>%
left_join(intpay_old_a, by="period") %>%
select(-value.x,value=value.y) %>%
mutate(value=na.spline(value),
var="intpay")
intpay_recent <-
df_de %>%
filter(var=="intpay_recent") %>%
mutate(var="intpay")
minDateIntPayRecent <- min(intpay_recent$period)
intpay <- chain(basis = intpay_recent,
to_rebase= intpay_old_q,
date_chain=minDateIntPayRecent)
plot_df <- bind_rows(mutate(intpay_old_a,var="intpay_old_a"),
mutate(intpay_old_q,var="intpay_old_q"),
mutate(intpay,var="intpay_chained"))
ggplot(plot_df,aes(period,value,col=var))+
geom_line()+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL)+
ggtitle("Government Interest Payments")
Merging German Data
We gather all the final series in a dataframe.
DE_rawdata <-
df_de %>%
filter(! var %in% c("lendingrate_old","lendingrate_recent",
"pop_old","pop_recent",
"debt_old","debt_recent",
"unempbenef_old","unempbenef_recent",
"totexp_recent", "totexp_old",
"intpay_recent", "intpay_old",
"totrev_recent", "totrev_old",
"pubcons_recent", "pubcons_old",
"pubinves_recent","pubinves_old",
"tfs_recent","tfs_old")) %>%
bind_rows(lendingrate,pop,debt,unempbenef,totexp,totrev,intpay,pubcons,pubinves,tfs) %>%
spread(var,value) %>%
add_column(country="DE")
Italy: Chaining & Interpolating Data
Before chaining and interpolating the special cases, we first gather all data for France, and then we proceed case by case.
df_it <- df %>% filter(country=="IT") %>% select(-country)
Consumption
Quarterly data on consumption is not available for Italy before 1996, but there is annual data. We interpolate the annual series in order to obtain quarterly series, and then we chain the two quarterly series in 1996-01-01.
conso_old_a <-
rdb("Eurostat","nama_10_gdp",mask = "A.CLV10_MEUR.P31_S14_S15.IT") %>%
select(period, value) %>%
add_column(var="conso_old") %>%
mutate(value=value/4)
conso_old_q <-
tibble(period=seq(as.Date("1995-01-01"),
length.out = (nrow(conso_old_a)-1)*4+1,
by = "quarter"),
value=NA) %>%
left_join(conso_old_a, by="period") %>%
select(-value.x,value=value.y) %>%
mutate(value=na.spline(value),
var="conso")
conso_recent <-
df_it %>%
filter(var=="conso")
minDateConsoRecent <- min(conso_recent$period)
conso<- chain(basis = conso_recent,
to_rebase= conso_old_q,
date_chain=minDateConsoRecent)
plot_df <- bind_rows(mutate(conso_old_a,var="conso_old_a"),
mutate(conso_old_q,var="conso_old_q"),
mutate(conso,var="conso_chained"))
ggplot(plot_df,aes(period,value,col=var))+
geom_line()+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL)+
ggtitle("Consumption")
Investment
Quarterly data on investment is not available for Italy before 1996, but there is annual data. We interpolate the annual series in order to obtain quarterly series, and then we chain the two quarterly series in 1996-01-01.
inves_old_a <-
rdb("Eurostat","nama_10_gdp",mask = "A.CLV10_MEUR.P51G.IT") %>%
select(period, value) %>%
add_column(var="inves_old") %>%
mutate(value=value/4)
inves_old_q <-
tibble(period=seq(as.Date("1995-01-01"),
length.out = (nrow(inves_old_a)-1)*4+1,
by = "quarter"),
value=NA) %>%
left_join(inves_old_a, by="period") %>%
select(-value.x,value=value.y) %>%
mutate(value=na.spline(value),
var="inves")
inves_recent <-
df_it %>%
filter(var=="inves")
minDateInvesRecent <- min(inves_recent$period)
inves<- chain(basis = inves_recent,
to_rebase= inves_old_q,
date_chain=minDateInvesRecent)
plot_df <- bind_rows(mutate(inves_old_a,var="inves_old_a"),
mutate(inves_old_q,var="inves_old_q"),
mutate(inves,var="inves_chained"))
ggplot(plot_df,aes(period,value,col=var))+
geom_line()+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL)+
ggtitle("Investment")
Investment Deflator
Quarterly data on investment deflator is not available for Italy before 1996, but there is annual data. We interpolate the annual series in order to obtain quarterly series, and then we chain the two quarterly series in 1996-01-01.
definves_old_a <-
rdb("Eurostat","nama_10_gdp",mask = "A.PD10_EUR.P51G.IT")%>%
select(period, value) %>%
add_column(var="definves_old")
definves_old_q <-
tibble(period=seq(as.Date("1995-01-01"),
length.out = (nrow(definves_old_a)-1)*4+1,
by = "quarter"),
value=NA) %>%
left_join(definves_old_a, by="period") %>%
select(-value.x,value=value.y) %>%
mutate(value=na.spline(value),
var="definves")
definves_recent <-
df_it %>%
filter(var=="definves")
minDateDefInvesRecent <- min(definves_recent$period)
definves<- chain(basis = definves_recent,
to_rebase= definves_old_q,
date_chain=minDateDefInvesRecent)
plot_df <- bind_rows(mutate(definves_old_a,var="definves_old_a"),
mutate(definves_old_q,var="definves_old_q"),
mutate(definves,var="definves_chained"))
ggplot(plot_df,aes(period,value,col=var))+
geom_line()+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL)+
ggtitle("Investment Deflator")
Government Debt
We first interpolate the annual series in order to obtain quarterly series, and then we chain the quarterly series from the two different databases.
debt_old_a <-
df_it %>%
filter(var=="debt_old") %>%
mutate(value=1000*value) %>%
select(-var)
debt_old_q <-
tibble(period=seq(as.Date("1991-01-01"),
length.out = (nrow(debt_old_a)-1)*4+1,
by = "quarter"),
value=NA) %>%
left_join(debt_old_a, by="period") %>%
select(-value.x) %>%
rename(value=value.y) %>%
mutate(value=na.spline(value),
var="debt")
debt_recent <-
df_it %>%
filter(var=="debt_recent") %>%
mutate(var="debt")
minDateDebtRecent <- min(debt_recent$period)
debt <-
chain(basis = debt_recent,
to_rebase= debt_old_q,
date_chain=minDateDebtRecent) %>%
mutate(var="debt")
plot_df <- bind_rows(add_column(debt_old_a,var="debt_old_a"),
mutate(debt_old_q,var="debt_old_q"),
mutate(debt,var="debt_chained"))
ggplot(plot_df,aes(period,value,col=var))+
geom_line()+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL)+
ggtitle("Government Debt")
Population
We first interpolate the annual series in order to obtain quarterly series, and then we chain the quarterly series from the two different databases.
pop_old_a <-
df_it %>%
filter(var=="pop_old")
pop_old_q <-
tibble(period=seq(as.Date("1991-01-01"),
length.out = (nrow(pop_old_a)-1)*4+1,
by = "quarter"),
value=NA) %>%
left_join(pop_old_a, by="period") %>%
select(-value.x) %>%
rename(value=value.y) %>%
mutate(value=na.spline(value),
var="pop")
pop_recent <-
df_it %>%
filter(var=="pop_recent") %>%
mutate(var="pop",value=value*1000)
pop <- chain(basis = pop_recent,
to_rebase= pop_old_q,
date_chain="2015-01-01")
plot_df <- bind_rows(mutate(pop_old_a,var="pop_old_a"),
mutate(pop_old_q,var="pop_old_q"),
mutate(pop,var="pop_chained"),
mutate(pop_recent,var="pop_recent"))
ggplot(plot_df,aes(period,value,col=var))+
geom_line()+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL)+
ggtitle("Population")
Lending Rate
We chain the quarterly series from the two different databases.
lendingrate_old <-
df_it %>%
filter(var=="lendingrate_old") %>%
mutate(var="lendingrate")
lendingrate_recent <-
df_it %>%
filter(var=="lendingrate_recent") %>%
mutate(var="lendingrate")
minDateLendingRateRecent <- min(lendingrate_recent$period)
lendingrate <-
chain(basis = lendingrate_recent,
to_rebase= lendingrate_old,
date_chain=minDateLendingRateRecent) %>%
mutate(var="lendingrate")
plot_df <- bind_rows(mutate(lendingrate_old,var="lendingrate_old"),
mutate(lendingrate_recent,var="lendingrate_recent"),
mutate(lendingrate,var="lendingrate_chained"))
ggplot(plot_df,aes(period,value,col=var))+
geom_line()+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL)+
ggtitle("Lending Rate")
Unemployment Benefits
We first interpolate the annual series in order to obtain quarterly series, and then we chain the two quarterly series.
unempbenef_old_a <-
df_it %>%
filter(var=="unempbenef_old") %>%
mutate(value=value/4)
unempbenef_old_q <-
tibble(period=seq(as.Date("1991-01-01"),
length.out = (nrow(unempbenef_old_a)-1)*4+1,
by = "quarter"),
value=NA) %>%
left_join(unempbenef_old_a, by="period") %>%
select(-value.x,value=value.y) %>%
mutate(value=na.spline(value),
var="unempbenef")
unempbenef_recent <-
df_it %>%
filter(var=="unempbenef_recent") %>%
mutate(var="unempbenef")
unempbenef <- chain(basis = unempbenef_recent,
to_rebase= unempbenef_old_q,
date_chain="2015-01-01")
plot_df <- bind_rows(mutate(unempbenef_old_a,var="unempbenef_old_a"),
mutate(unempbenef_old_q,var="unempbenef_old_q"),
mutate(unempbenef,var="unempbenef_chained"))
ggplot(plot_df,aes(period,value,col=var))+
geom_line()+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL)+
ggtitle("Unemployment Benefits")
Government Consumption
We first interpolate the annual series in order to obtain quarterly series, and then we chain the quarterly series from the two different databases.
pubcons_old_a <-
df_it %>%
filter(var=="pubcons_old") %>%
mutate(value=value/4)
pubcons_old_q <-
tibble(period=seq(as.Date("1995-01-01"),
length.out = (nrow(pubcons_old_a)-1)*4+1,
by = "quarter"),
value=NA) %>%
left_join(pubcons_old_a, by="period") %>%
select(-value.x,value=value.y) %>%
mutate(value=na.spline(value),
var="pubcons")
pubcons_recent <-
df_it %>%
filter(var=="pubcons_recent") %>%
mutate(var="pubcons")
minDatePubConsRecent <- min(pubcons_recent$period)
pubcons <- chain(basis = pubcons_recent,
to_rebase= pubcons_old_q,
date_chain=minDatePubConsRecent)
plot_df <- bind_rows(mutate(pubcons_old_a,var="pubcons_old_a"),
mutate(pubcons_old_q,var="pubcons_old_q"),
mutate(pubcons,var="pubcons_chained"))
ggplot(plot_df,aes(period,value,col=var))+
geom_line()+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL)+
ggtitle("Government Consumption")
Government Investment
We first interpolate the annual series in order to obtain quarterly series, and then we chain the quarterly series from the two different databases.
pubinves_old_a <-
df_it %>%
filter(var=="pubinves_old") %>%
mutate(value=value/4)
pubinves_old_q <-
tibble(period=seq(as.Date("1995-01-01"),
length.out = (nrow(pubinves_old_a)-1)*4+1,
by = "quarter"),
value=NA) %>%
left_join(pubinves_old_a, by="period") %>%
select(-value.x,value=value.y) %>%
mutate(value=na.spline(value),
var="pubinves")
pubinves_recent <-
df_it %>%
filter(var=="pubinves_recent") %>%
mutate(var="pubinves")
pubinves <- chain(basis = pubinves_recent,
to_rebase= pubinves_old_q,
date_chain="2003-01-01")
plot_df <- bind_rows(mutate(pubinves_old_a,var="pubinves_old_a"),
mutate(pubinves_old_q,var="pubinves_old_q"),
mutate(pubinves,var="pubinves_chained"))
ggplot(plot_df,aes(period,value,col=var))+
geom_line()+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL)+
ggtitle("Government Investment")
Government Social Transfers
We first interpolate the annual series in order to obtain quarterly series, and then we chain the quarterly series from the two different databases.
tfs_old_a <-
df_it %>%
filter(var=="tfs_old") %>%
mutate(value=value/4)
tfs_old_q <-
tibble(period=seq(as.Date("1995-01-01"),
length.out = (nrow(tfs_old_a)-1)*4+1,
by = "quarter"),
value=NA) %>%
left_join(tfs_old_a, by="period") %>%
select(-value.x,value=value.y) %>%
mutate(value=na.spline(value),
var="tfs")
tfs_recent <-
df_it %>%
filter(var=="tfs_recent") %>%
mutate(var="tfs")
minDateTfsRecent <- min(tfs_recent$period)
tfs <- chain(basis = tfs_recent,
to_rebase= tfs_old_q,
date_chain=minDateTfsRecent)
plot_df <- bind_rows(mutate(tfs_old_a,var="tfs_old_a"),
mutate(tfs_old_q,var="tfs_old_q"),
mutate(tfs,var="tfs_chained"))
ggplot(plot_df,aes(period,value,col=var))+
geom_line()+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL)+
ggtitle("Government Social Transfers")
Total Government Expenditure
We first interpolate the annual series in order to obtain quarterly series, and then we chain the quarterly series from the two different databases.
totexp_old_a <-
df_it %>%
filter(var=="totexp_old") %>%
mutate(value=value/4)
totexp_old_q <-
tibble(period=seq(as.Date("1995-01-01"),
length.out = (nrow(totexp_old_a)-1)*4+1,
by = "quarter"),
value=NA) %>%
left_join(totexp_old_a, by="period") %>%
select(-value.x,value=value.y) %>%
mutate(value=na.spline(value),
var="totexp")
totexp_recent <-
df_it %>%
filter(var=="totexp_recent") %>%
mutate(var="totexp")
minDateTotExpRecent <- min(totexp_recent$period)
totexp <- chain(basis = totexp_recent,
to_rebase= totexp_old_q,
date_chain=minDateTotExpRecent)
plot_df <- bind_rows(mutate(totexp_old_a,var="totexp_old_a"),
mutate(totexp_old_q,var="totexp_old_q"),
mutate(totexp,var="totexp_chained"))
ggplot(plot_df,aes(period,value,col=var))+
geom_line()+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL)+
ggtitle("Total Government Expenditure")
Total Government Revenue
We first interpolate the annual series in order to obtain quarterly series, and then we chain the quarterly series from the two different databases.
totrev_old_a <-
df_it %>%
filter(var=="totrev_old") %>%
mutate(value=value/4)
totrev_old_q <-
tibble(period=seq(as.Date("1995-01-01"),
length.out = (nrow(totrev_old_a)-1)*4+1,
by = "quarter"),
value=NA) %>%
left_join(totrev_old_a, by="period") %>%
select(-value.x,value=value.y) %>%
mutate(value=na.spline(value),
var="totrev")
totrev_recent <-
df_it %>%
filter(var=="totrev_recent") %>%
mutate(var="totrev")
minDateTotRevRecent <- min(totrev_recent$period)
totrev <- chain(basis = totrev_recent,
to_rebase= totrev_old_q,
date_chain=minDateTotRevRecent)
plot_df <- bind_rows(mutate(totrev_old_a,var="totrev_old_a"),
mutate(totrev_old_q,var="totrev_old_q"),
mutate(totrev,var="totrev_chained"))
ggplot(plot_df,aes(period,value,col=var))+
geom_line()+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL)+
ggtitle("Total Government Revenue")
Government Interest Payments
We first interpolate the annual series in order to obtain quarterly series, and then we chain the quarterly series from the two different databases.
intpay_old_a <-
df_it %>%
filter(var=="intpay_old") %>%
mutate(value=value/4)
intpay_old_q <-
tibble(period=seq(as.Date("1995-01-01"),
length.out = (nrow(intpay_old_a)-1)*4+1,
by = "quarter"),
value=NA) %>%
left_join(intpay_old_a, by="period") %>%
select(-value.x,value=value.y) %>%
mutate(value=na.spline(value),
var="intpay")
intpay_recent <-
df_it %>%
filter(var=="intpay_recent") %>%
mutate(var="intpay")
minDateIntPayRecent <- min(intpay_recent$period)
intpay <- chain(basis = intpay_recent,
to_rebase= intpay_old_q,
date_chain=minDateIntPayRecent)
plot_df <- bind_rows(mutate(intpay_old_a,var="intpay_old_a"),
mutate(intpay_old_q,var="intpay_old_q"),
mutate(intpay,var="intpay_chained"))
ggplot(plot_df,aes(period,value,col=var))+
geom_line()+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL)+
ggtitle("Government Interest Payments")
Merging Italian Data
We gather all the final series in a dataframe.
IT_rawdata <-
df_it %>%
filter(! var %in% c("conso","inves","definves","lendingrate_old","lendingrate_recent",
"pop_old","pop_recent",
"debt_old","debt_recent",
"unempbenef_old","unempbenef_recent",
"totexp_recent", "totexp_old",
"intpay_recent", "intpay_old",
"totrev_recent", "totrev_old",
"pubcons_recent", "pubcons_old",
"pubinves_recent","pubinves_old",
"tfs_recent","tfs_old")) %>%
bind_rows(conso,inves,definves,lendingrate,pop,debt,unempbenef,totexp,totrev,intpay,pubcons,pubinves,tfs) %>%
spread(var,value) %>%
add_column(country="IT")
Final database for the estimation
Implicit tax rates
We retrieve the data on the implicit tax rates (ITR) on consumption, labour and corporate incomes, that we built specifically for this project : https://macro.cepremap.fr/article/2019-11/implicit_tax_rates/
itr <- read_csv("https://shiny.cepremap.fr/data/ITR_eurodata.csv") %>%
rename(year=period)
itrq <-
tibble(period=EA_rawdata$period) %>%
mutate(year=year(period)) %>%
left_join(itr,by="year") %>%
na.omit() %>%
select(-year) %>%
gather(var,value,-period) %>%
separate(var,c("country","var")) %>%
spread(var,value)
Merge raw data
rawdata_var <- colnames(DE_rawdata) EA_rawdata_short <- EA_rawdata %>% select(rawdata_var,oil_prices)
Then, we gather the datasets for France, Germany, Italy, Spain and the Euro area in a unique data frame.
rawdata_df <-
bind_rows(EA_rawdata_short,DE_rawdata,ES_rawdata,FR_rawdata,IT_rawdata) %>%
left_join(itrq,
by = c("period", "country")) %>%
arrange(country,period) %>%
filter(year(period)>=1995,
period <= last_date)
save(rawdata_df, file = "rawdata.RData")
Normalize data
Then we select and normalize the data by population and prices.
data_df <-
rawdata_df %>%
transmute(period,
country,
gdp_rpc=1e+6*gdp/pop,
conso_rpc=1e+6*conso/pop,
inves_rpc=1e+6*inves/pop,
defgdp = defgdp,
wage_rph=1e+6*wage/defgdp/(hours*1000),
hours_pc=(hours*1000)/pop,
pinves_defl=definves/defgdp,
loans_nfc_rpc=1e+9*loans_nfc/pop/defgdp,
networth_rpc=1e+6*networth/pop/defgdp,
re=shortrate/100,
creditspread=(lendingrate-shortrate)/100,
pubcons_rpc=100*1e+6*pubcons/(defgdp*pop),
pubinves_rpc=100*1e+6*pubinves/(defgdp*pop),
tfs_rpc=100*1e+6*tfs/(defgdp*pop),
othgov_rpc=100*1e+6*(totexp-pubcons-pubinves-tfs-intpay)/(defgdp*pop),
debt_gdp=100*debt/(defgdp*gdp),
taun,tauwf,tauwh,tauc,
world_demand,
oil_prices,
neer,
imports_rpc=imports/pop,
exports_rpc=exports/pop) %>%
gather(var,value,-period,-country)
The figure below shows the final series for all the listed countries.
plot_data_df <-
data_df %>%
mutate(varname=
case_when(
var=="gdp_rpc" ~ "Real GDP per capita",
var=="conso_rpc" ~ "Real consumption \n per capita",
var=="inves_rpc" ~ "Real investment \n per capita",
var=="defgdp" ~ "GDP deflator",
var=="wage_rph" ~ "Real wage per hour" ,
var=="hours_pc" ~ "Hours worked per capita",
var=="pinves_defl" ~ "Real price of investment",
var=="loans_nfc_rpc" ~ "Real credit to \n NFC per capita",
var=="networth_rpc" ~ "Real net worth \n per capita",
var=="re" ~ "Short-term \n interest rate (APR)",
var=="creditspread" ~ "Credit spread (APP)",
var=="pubcons_rpc" ~ "Real public consumption\n per capita",
var=="pubinves_rpc" ~ "Real public investment\n per capita",
var=="tfs_rpc" ~ "Real social transfers\n per capita",
var=="othgov_rpc" ~ "Real other public\n expenditure per capita",
var=="debt_gdp" ~ "Debt-to-GDP ratio",
var=="taun" ~ "Implicit Tax Rate \n on labour income ",
var=="tauwh" ~ "Implicit Tax Rate \n on employees' SSC",
var=="tauwf" ~ "Implicit Tax Rate \n on employers' SSC" ,
var=="tauc" ~ "Implicit Tax Rate \n on consumption" ,
var=="world_demand" ~ "Foreign demand",
var=="oil_prices" ~ "Crude oil prices",
var=="neer" ~ "Nominal effective exchange rate",
var=="imports_rpc" ~ "Real imports per capita",
var=="exports_rpc" ~ "Real exports per capita"),
country_name=
case_when(
country=="FR" ~ "France",
country=="DE" ~ "Germany",
country=="IT" ~ "Italy",
country=="ES" ~ "Spain" ,
country=="EA" ~ "Euro Area")) %>%
na.omit()
tikz('estimated.tex', width=5.2,height=8.4, sanitize=TRUE)
ggplot(plot_data_df,aes(period,value,col=country_name))+
geom_line()+
facet_wrap(~varname,ncol=3,scales="free_y")+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL)+
theme(legend.title=element_blank()) +
theme(strip.text=element_text(size=8),
axis.text=element_text(size=7))+
ggtitle("Series for the estimation")+
theme(plot.title=element_text(size=12))
dev.off()
## png
## 2
ggplot(plot_data_df,aes(period,value,col=country_name))+
geom_line()+
facet_wrap(~varname,ncol=3,scales="free_y")+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL)+
theme(legend.title=element_blank()) +
theme(strip.text=element_text(size=10),
axis.text=element_text(size=9))+
ggtitle("Series for the estimation")+
theme(plot.title=element_text(size=15))
data <-
data_df %>%
unite("var",c("country","var")) %>%
mutate(period=gsub(" ","",as.yearqtr(period))) %>%
spread(var,value) %>%
select(-DE_oil_prices,-FR_oil_prices,-ES_oil_prices,-IT_oil_prices)
colnames(data)[1] <- ""
write.csv(data,"data_DE_EA_ES_FR_IT.csv",row.names = FALSE)
The data can be downloaded directly here.
Series for the calibration
For the calibration, we need additional series. The data needed for this purpose is retrieved below by variable.
Leverage of non-financial corporations
Using the nasq_10_f_bs database from Eurostat, we retrieve these data. The figure below shows the series for France, Germany, Italy, Spain and the Euro area 19.
debt <- rdb("Eurostat","nasq_10_f_bs",mask = "Q.MIO_EUR.S11.LIAB.F+F3+F4+F6.EA19+IT+DE+FR+ES")
leverage <-
debt %>%
select(value,period,country=geo,var=na_item) %>%
mutate(var=
case_when(
var=="F" ~ "total",
var=="F3" ~ "debt_securities",
var=="F4" ~ "loans",
var=="F6" ~ "pensions_reserves")) %>%
spread(var,value) %>%
arrange(country,period) %>%
transmute(period,
country,
value=(debt_securities+loans+pensions_reserves)/total,
var="leverage") %>%
na.omit()
ggplot(leverage,aes(period,value,colour=country))+
geom_line() +
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL) +
theme(legend.title=element_blank()) +
ggtitle("Leverage")
The depreciation rate of the capital stock
We compute the depreciation rate of the capital stock for France, Italy, Spain, Germany and the Euro Area since 1995. We use the data from the Penn World Table Feenstra et al. (2015) available here.
listcountry <- list("France" = "FRA",
"Germany" = "DEU",
"Italy" = "ITA",
"Spain" = "ESP",
"Euro Area" = "EA")
df <- read_dta("https://www.rug.nl/ggdc/docs/pwt100.dta") %>%
mutate(country=countrycode, period=as.Date(as.yearqtr(year))) %>%
filter(year(period)>=1995 & currency_unit=="Euro" & !grepl('MNE', country))
Step1: data for Euro Area countries
delta <- df %>% select(country, period, value=delta) %>% add_column(var="delta")
Step2: Euro Area GDP-weighted average
After retrieving the data on the average depreciation rate of the capital stock for the Euro Area countries, it is possible to build the GDP-weighted average for the Eurozone. First, it is necessary to establish the weights that will be used for this purpose, using the output-side real GDP at chained PPPs (in mil. 2011US$) of each country.
gdp <- df %>% select(country, period, value=rgdpo) EA_gdp <- gdp %>% group_by(period) %>% summarize(value=sum(value)) %>% ungroup() weights <- gdp %>% left_join(EA_gdp,by="period") %>% transmute(country, period, weight=value.x/value.y)
Now we apply these weights to our country data in order to build the Euro Area GDP-weighted average. The figure below shows the final series for France, Germany, Italy, Spain and the Euro Area.
delta_EA <-
delta %>%
left_join(weights,by=c("country","period"))
delta_EA <-
delta_EA %>%
transmute(period,value=value*weight) %>%
group_by(period) %>%
summarize(value =sum(value)) %>%
add_column(country="EA",var="delta")
delta_countries <-
delta %>%
filter(grepl('FRA|DEU|ITA|ESP', country))
delta_FIN <-
bind_rows(delta_countries,delta_EA)
ggplot(delta_FIN,aes(period,value,colour=country))+
geom_line()+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL) +
theme(legend.title=element_blank()) +
ggtitle("Depreciation rate of the capital stock")
The share of capital revenues in GDP
We proceed in two steps.
Step1: data for Euro Area countries
We obtain the share of labour compensation in GDP also from the Penn World Table, for the countries that compose the Euro Area. We deduce then the share of capital revenues in GDP
alpha <- df %>% select(country, period, value=labsh) %>% mutate(value=1-value) %>% add_column(var="alpha")
Step2: Euro Area GDP-weighted average
Now we apply the GDP-weights to our country data in order to build the Euro Area GDP-weighted average. The figure below shows the final series for France, Germany, Italy, Spain and the Euro Area.
alpha_EA <-
alpha %>%
left_join(weights,by=c("country","period"))
alpha_EA <-
alpha_EA %>%
transmute(period,value=value*weight) %>%
group_by(period) %>%
summarize(value =sum(value)) %>%
add_column(country="EA",var="alpha")
alpha_countries <-
alpha %>%
filter(grepl('FRA|DEU|ITA|ESP', country))
alpha_FIN <-
bind_rows(alpha_countries,alpha_EA)
ggplot(alpha_FIN,aes(period,value,colour=country)) +
geom_line() +
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL) +
theme(legend.title=element_blank()) +
ggtitle("Share of capital revenues in GDP")
The share of capital in GDP
We proceed in two steps.
Step1: data for Euro Area countries
We obtain the stock of capital in GDP also from the Penn World Table, for the countries that compose the Euro Area.
capital <- df %>% select(country,period,gdp=rgdpna,capital=rnna) %>% mutate(value=capital/gdp) %>% select(-capital,-gdp) %>% add_column(var="capital_gdp")
Step2: Euro Area GDP-weighted average
Now we apply the GDP-weights to our country data in order to build the Euro Area GDP-weighted average. The figure below shows the final series for France, Germany, Italy, Spain and the Euro Area.
capital_EA <-
capital %>%
left_join(weights,by=c("country","period"))
capital_EA <-
capital_EA %>%
transmute(period,value=value*weight) %>%
group_by(period) %>%
summarize(value =sum(value)) %>%
add_column(country="EA",var="capital_gdp")
capital_countries <-
capital %>%
filter(grepl('FRA|DEU|ITA|ESP', country))
capital_FIN <-
bind_rows(capital_countries,capital_EA)
ggplot(capital_FIN,aes(period,value,colour=country))+
geom_line()+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL) +
theme(legend.title=element_blank()) +
ggtitle("Stock of capital in GDP")
The share of crude oil imports in GDP
oil_import_value <-
rdb("Eurostat","nrg_ti_coifpm",mask="M.TOTAL.VAL_THS_USD.DE+FR+IT+ES+EU_V") %>%
select(period,oil_import=value,country=geo) %>%
mutate(year=year(period)) %>%
group_by(country,year) %>%
summarise(oil_import=sum(oil_import)) %>%
mutate(period=as.Date(paste0(year,"-01-01"))) %>%
select(-year) %>%
mutate(country=case_when(
country=="EU_V" ~ "EA",
TRUE ~ country))
# in US $
ea_gdp_usd <-
rdb("IMF","WEOAGG:latest",mask = "998.NGDPD.us_dollars") %>%
select(period,gdp=value,country=`weo-countries-group`)
gdp_usd <-
rdb("IMF","WEO:latest",mask = "FRA+DEU+ITA+ESP.NGDPD.us_dollars") %>%
select(period,gdp=value,country=`weo-country`) %>%
bind_rows(ea_gdp_usd) %>%
mutate(country=case_when(
country=="FRA" ~ "FR",
country=="DEU" ~ "DE",
country=="ITA" ~ "IT",
country=="ESP" ~ "ES",
country=="998" ~ "EA",
TRUE ~ country))
oil <-
oil_import_value %>%
left_join(gdp_usd) %>%
transmute(period,
value=oil_import/(gdp*1000000),
var="oil_imports_gdp")
ggplot(oil,aes(period,value,color=country))+
geom_line() +
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL) +
theme(legend.title=element_blank()) +
ggtitle("Share of oil imports in GDP")
The share of petrol in private consumption
petrol_weight <-
rdb("Eurostat","prc_hicp_inw",mask="A.CP07222.FR+DE+IT+ES+EA") %>%
select(value,country=geo,period) %>%
mutate(value=value/1000) %>%
rename(petrol_weight=value) %>%
filter(year(period)>=2015)
hhconso <-
rdb("Eurostat","nama_10_gdp",mask="A.PC_GDP.P31_S14.FR+DE+IT+ES+EA") %>%
select(period,conso_gdp=value,country=geo) %>%
filter(year(period)>=2015)
petrol_conso <-
hhconso %>%
left_join(petrol_weight,by=c("country","period")) %>%
na.omit() %>%
transmute(period,
country,
value=conso_gdp*petrol_weight/100,
var="petrol_conso")
ggplot(petrol_conso,aes(period,value,color=country))+
geom_line() +
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL) +
theme(legend.title=element_blank()) +
ggtitle("Petrol consumption to GDP ratio")
Share of final consumption in imports
imported_conso <-
rdb("Eurostat","naio_10_cp1700",mask="A.MIO_EUR.IMP.P3+P51G.TOTAL.DE+FR+IT+ES+EA19") %>%
select(period,value,country=geo,var=induse) %>%
spread(var,value) %>%
mutate(value=P3/(P3+P51G)) %>%
select(-P3,-P51G) %>%
filter(year(period)>=2010) %>%
mutate(country=case_when(country=="EA19" ~ "EA",
TRUE ~ country),
var="imported_conso")
ggplot(imported_conso,aes(period,value,color=country))+
geom_line() +
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL) +
theme(legend.title=element_blank()) +
ggtitle("Share of final consumption in total imports")
Miscellaneous
ea_gdp <-
rdb("Eurostat/namq_10_gdp/Q.CP_MEUR.SCA.B1GQ.EA19") %>%
select(period,gdp=value)
df <- rdb("ECB","TRD",mask = 'M.I8.Y.M+X.TTT.J8.4.VAL')
ea_trade <-
df %>%
transmute(period=paste(year(period),quarter(period),sep="-"),
value,
var=ifelse(grepl("Import",series_name),"imports","exports")) %>%
group_by(var,period) %>%
summarize(value = sum(value)) %>%
ungroup() %>%
mutate(period=yq(period)) %>%
spread(var,value) %>%
left_join(ea_gdp) %>%
transmute(period,
imports_gdp=imports/(gdp*1000),
exports_gdp=exports/(gdp*1000)) %>%
gather(var,value,-period) %>%
add_column(country="EA")
df <- rdb("OECD","EO",mask = "FRA+DEU+ITA+ESP.XGS+MGS+GDP.A")
imports_exports <-
df %>%
select(var=VARIABLE,period,value, country=LOCATION) %>%
spread(var,value) %>%
transmute(period, country,
imports_gdp=MGS/GDP,
exports_gdp=XGS/GDP) %>%
gather(var,value,-period,-country) %>%
mutate(country=case_when(
country=="FRA" ~ "FR",
country=="DEU" ~ "DE",
country=="ITA" ~ "IT",
country=="ESP" ~ "ES",
TRUE ~ country)) %>%
bind_rows(ea_trade)
alpha_delta_capital <-
bind_rows(alpha_FIN,delta_FIN,capital_FIN) %>%
mutate(country=case_when(
country=="FRA" ~ "FR",
country=="DEU" ~ "DE",
country=="ITA" ~ "IT",
country=="ESP" ~ "ES",
TRUE ~ country))
leverage2 <-
leverage %>%
mutate(country=case_when(
country=="EA19" ~ "EA",
TRUE ~ country))
share <- rdb("Eurostat","nama_10_gdp",mask="A.CP_MPPS.B1GQ.DE+FR+IT+ES+EA19")
share2 <-
share %>%
select(value,country=geo,period) %>%
spread(country,value) %>%
transmute(period,
FR=FR/EA19,
DE=DE/EA19,
IT=IT/EA19,
ES=ES/EA19) %>%
gather(country,value,-period) %>%
mutate(var="share")
Final series for the calibration, and steady state values by country
rawdata <-
bind_rows(EA_rawdata_short,
FR_rawdata,
ES_rawdata,
IT_rawdata,
DE_rawdata) %>%
left_join(itrq,by=c("country","period")) %>%
select(period,country,pubcons,pubinves,tfs,totexp,totrev,intpay,gdp,inves,tauk) %>%
filter(period<=max(rawdata_df$period)) %>%
gather(var,value,-country,-period) %>%
bind_rows(plot_data_df) %>%
select(-varname,-country_name)
hours_pc_meanEA <-
rawdata %>%
filter(var=="hours_pc",
country=="EA") %>%
summarise(value=mean(value,na.rm = T)) %>%
first()
rawdata_growth_ratio <-
rawdata %>%
spread(var,value) %>%
arrange(country) %>%
transmute(period,
country,
defgdp_growth=defgdp/lag(defgdp,4)-1,
gdp_rpc_growth=gdp_rpc/lag(gdp_rpc,4)-1,
definves_growth=pinves_defl/lag(pinves_defl,4)-1,
hours_pc_index=hours_pc/hours_pc_meanEA,
tfs_gdp=tfs/(defgdp/100*gdp),
pubcons_gdp=pubcons/(defgdp/100*gdp),
pubinves_gdp=pubinves/(defgdp/100*gdp),
totexp_gdp=totexp/(defgdp/100*gdp),
otherexp_gdp=(totexp-tfs-pubcons-pubinves-intpay)/(defgdp/100*gdp),
intpay_gdp=intpay/(defgdp/100*gdp),
totrev_gdp=totrev/(defgdp/100*gdp),
inves_gdp=inves/gdp,
shortrate=re,
tauk,taun,tauwh,tauwf,tauc) %>%
gather(var,value,-period,-country) %>%
bind_rows(alpha_delta_capital,
leverage2,
imports_exports,
imported_conso,
share2,
oil,
petrol_conso) %>%
filter(year(period) >= 1995,
year(period) <= 2019) %>%
mutate(varname=
case_when(
var=="alpha" ~ "Share of capital \n revenue in GDP",
var=="defgdp_growth" ~ "GDP deflator \n growth rate",
var=="definves_growth" ~ "Price of investment \n growth rate",
var=="delta" ~ "Depreciation rate \n of the capital stock" ,
var=="capital_gdp" ~ "Capital stock in GDP",
var=="exports_gdp" ~ "Exports-to-GDP ratio" ,
var=="gdp_rpc_growth" ~ "Real GDP per capita \n growth rate" ,
var=="hours_pc_index" ~ "Hours worked per \n capita index",
var=="imports_gdp" ~ "Imports-to-GDP ratio" ,
var=="imported_conso" ~ "Share of final consumption \n in total imports",
var=="oil_imports_gdp" ~ "Oil imports to GDP ratio",
var=="petrol_conso" ~ "Petrol consumption \n to GDP ratio ",
var=="intpay_gdp" ~ "Government interest \n payments to GDP ratio",
var=="inves_gdp" ~ "Investment-to-GDP ratio",
var=="leverage" ~ "Leverage of non \n financial corporations",
var=="share" ~ "Share of PPP GDP \n in Euro area PPP GDP",
var=="otherexp_gdp" ~ "Other government \n expenditures to GDP ratio",
var=="pubcons_gdp" ~ "Government consumption \n to GDP ratio",
var=="pubinves_gdp" ~ "Government investment \n to GDP ratio",
var=="shortrate" ~ "Short-term \n interest rate (APR)",
var=="taun" ~ "Implicit Tax Rate \n on labour income " ,
var=="tauwh" ~ "Implicit Tax Rate \n on employees' SSC",
var=="tauwf" ~ "Implicit Tax Rate \n on employers' SSC",
var=="tauc" ~ "Implicit Tax Rate \n on consumption",
var=="tauk" ~ "Implicit Tax Rate \n on corporate income",
var=="tfs_gdp" ~ "Government social \n transfers to GDP ratio",
var=="totexp_gdp" ~ "Total government \n expenditure to GDP ratio",
var=="totrev_gdp" ~ "Total government \n revenue to GDP ratio"),
country=
case_when(
country=="FR" ~ "France",
country=="DE" ~ "Germany",
country=="IT" ~ "Italy",
country=="ES" ~ "Spain",
country=="EA" ~ "Euro Area"))
## Error in melt_dataframe(data, id_idx - 1L, gather_idx - 1L, as.character(key_var), : All columns must be atomic vectors or lists. Problem with column 6.
tikz('calibrated.tex', width=5.2,height=8.4, sanitize=TRUE)
ggplot(rawdata_growth_ratio,aes(period,value,color=country))+
geom_line()+
facet_wrap(~varname,ncol=3,scales = "free_y")+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL)+
theme(legend.title=element_blank()) +
theme(strip.text=element_text(size=8),
axis.text=element_text(size=7))+
ggtitle("Series for the calibration")+
theme(plot.title=element_text(size=12))
dev.off()
## png
## 2
ggplot(rawdata_growth_ratio,aes(period,value,color=country))+
geom_line()+
facet_wrap(~varname,ncol=3,scales = "free_y")+
scale_x_date(expand = c(0.01,0.01)) +
theme + xlab(NULL) + ylab(NULL)+
theme(legend.title=element_blank()) +
theme(strip.text=element_text(size=11),
axis.text=element_text(size=9))+
ggtitle("Series for the calibration")+
theme(plot.title=element_text(size=15))
calibration <- rawdata_growth_ratio %>% mutate(Parameter=varname) %>...
To leave a comment for the author, please follow the link and comment on their blog: Macroeconomic Observatory - R.
R-bloggers.com offers daily e-mail updates about R news and tutorials about learning R and many other topics. Click here if you're looking to post or find an R/data-science job.
Want to share your content on R-bloggers? click here if you have a blog, or here if you don't.