8  Figures 4 and 7 & Table 7

This document was created on Monday May 29th and provides the output needed to replicate Corbi, Papaioannou, and Surico (2018) Namely, Figures 4 and 7 and table 7 from the published version of the paper.

We provide two sets of output: one for the full 2002-2019 sample and one reduced to the the common municipalities between the two data-sets.

This output was created using an adapted version of the author’s master.do file. It is split into three parts, one which does the cleaning and refers back to the master.do, and two others which create the figures and images. For more information, please see the README.txt file under ~/Dropbox/TransfersReplication/code/. All of the data is available to run their and our do files.

For an overview of the data cleaning code with some explanations behind each section, I have create an explainer available here.

8.1 Main Figures - Full - 2002-2019

Figure 4

Code

knitr::include_graphics("images/replication_images/figure_4_replication.png")

This figure replicates figure 4 from the oritinal Regional Transfers Multipliers paper using our 2002-2019 sample of interior municipalities. The FPM transfer data is deflated to 2000BRL. It is collected for all interior municipalities (as opposed to capitals or reserve municipalities).

FPM transfers around the cutoffs - 2000BRL Deflators

Figure 4 - Using Theoretical Transfer

Code

knitr::include_graphics("images/replication_images/figure_4_lfpmhat_replication.png")

This figure replicates figure 4 from the oritinal Regional Transfers Multipliers paper using our 2002-2019 sample of interior municipalities. The FPM transfer data is deflated to 2000BRL. It is collected for all interior municipalities (as opposed to capitals or reserve municipalities).

Theoretical FPM transfers around the cutoffs - 2000BRL Deflators

7 A) Residual Log(Total Private Sector Earnings)

Code

knitr::include_graphics("images/replication_images/figure_7a_replication.png")

This figure replicates figure 7a from the oritinal Regional Transfers Multipliers paper using our 2002-2019 sample of interior municipalities. It presents the residual log total earnings in the private sector. Earnings are deflated to 2000 BRL levels using the IPCA deflator. Here, we keep rows (mun-year) common to our data-set, which may include missing observatios for their and/or our data.

Residual log total earnings in the private sector - 2000BRL Deflators

B) Residual Log( Private Sector Employment)

Code

knitr::include_graphics("images/replication_images/figure_7b_replication.png")

This figure replicates figure 7b from the oritinal Regional Transfers Multipliers paperusing using our 2002-2019 sample of interior municipalities. It presents the residual log total employment in the private sector. Earnings in this data-set are deflated to 2000 BRL levels using the IPCA deflator.

Residual log total employment in the private sector - 2000BRL Deflators

C) Residual Log(Avg Private Sector Earnings)

Code

knitr::include_graphics("images/replication_images/figure_7c_replication.png")

This figure replicates figure 7c from the oritinal Regional Transfers Multipliers paper using our 2002-2019 sample of interior municipalities. It presents the residual log average municipality-year wages in the private sector. Earnings in this data-set are deflated to 2000 BRL levels using the IPCA deflator. Here, we keep rows (mun-year) common to our data-set, which may include missing observatios for their and/or our data. To be clear, we only look at firms that ought to be in the data, according to the rules specified in the paper.

Residual log average municipality-year wages in the private sector - 2000BRL Deflators

Table 7

Code

## Clean table 7 part A

# clean the lower part of the table --------------
table7A_lower <- paste0(dir_eemt, "subprojects/transfers_replication/ouput/regression_models/xlsx/table7A_LogPrivateEmployment_replication.xlsx") %>% 
  readxl::read_xlsx()  %>% as.data.table() %>% .[12:15] %>% .[, statistic := ""] %>% rename(., outcome=...1) %>% 
.[, Regression := ""]



# Clean Table A; numbers -----------------
table7A_LogPrivateEmployment_replication <- paste0(dir_eemt, "subprojects/transfers_replication/ouput/regression_models/xlsx/table7A_LogPrivateEmployment_replication.xlsx") %>% 
  readxl::read_xlsx()  %>% as.data.table() %>% .[3:6] %>% 
  .[, ...1 := zoo::na.locf(...1)] %>% melt.data.table(id.vars = c("...1")) %>% 
  .[!is.na(value)] %>% 
  .[, statistic := "coefficient"] %>% 
  .[stri_detect_fixed(value, "("), statistic := "se"]  %>% 
  .[, outcome := "Log Private Employment"]
  
table7A_LogPrivateWages <- paste0(dir_eemt, "subprojects/transfers_replication/ouput/regression_models/xlsx/table7A_LogPrivateWages_replication.xlsx") %>% 
  readxl::read_xlsx()   %>% as.data.table() %>% .[3:6] %>% 
  .[, ...1 := zoo::na.locf(...1)] %>% melt.data.table(id.vars = c("...1")) %>% 
  .[!is.na(value)] %>% 
  .[, statistic := "coefficient"] %>% 
  .[stri_detect_fixed(value, "("), statistic := "se"]     %>% 
  .[, outcome := "Log Private Avg. Wages"]


table7A_LogPrivateEarnings <- paste0(dir_eemt, "subprojects/transfers_replication/ouput/regression_models/xlsx/table7A_LogPrivateEarnings_replication.xlsx") %>% 
  readxl::read_xlsx()    %>% as.data.table() %>% .[3:6] %>% 
  .[, ...1 := zoo::na.locf(...1)] %>% melt.data.table(id.vars = c("...1")) %>% 
  .[!is.na(value)] %>% 
  .[, statistic := "coefficient"] %>% 
  .[stri_detect_fixed(value, "("), statistic := "se"]    %>% 
  .[, outcome := "Log Private Total Earnings"]
  
table7A_main <- rbind(table7A_LogPrivateEmployment_replication, table7A_LogPrivateWages) %>% 
  rbind(., table7A_LogPrivateEarnings) %>% 
  dcast.data.table(formula = outcome + statistic  ~ variable, value.var = "value" ) %>% 
  .[, Regression := "Reduced Form"]

# %>%    rbind(data.table(outcome=c("Reduced Form")), ., fill=T)


# Clean Table B; numbers -----------------
table7B_LogPrivateEmployment_replication <- paste0(dir_eemt, "subprojects/transfers_replication/ouput/regression_models/xlsx/table7B_LogPrivateEmployment_replication.xlsx") %>% 
  readxl::read_xlsx()  %>% as.data.table() %>% .[3:4] %>% 
  .[, ...1 := zoo::na.locf(...1)] %>% melt.data.table(id.vars = c("...1")) %>% 
  .[!is.na(value)] %>% 
  .[, statistic := "coefficient"] %>% 
  .[stri_detect_fixed(value, "("), statistic := "se"]  %>% 
  .[, outcome := "Log Private Employment"]
  
table7B_LogPrivateWages <- paste0(dir_eemt, "subprojects/transfers_replication/ouput/regression_models/xlsx/table7B_LogPrivateWages_replication.xlsx") %>% 
  readxl::read_xlsx()   %>% as.data.table() %>% .[3:4] %>% 
  .[, ...1 := zoo::na.locf(...1)] %>% melt.data.table(id.vars = c("...1")) %>% 
  .[!is.na(value)] %>% 
  .[, statistic := "coefficient"] %>% 
  .[stri_detect_fixed(value, "("), statistic := "se"]     %>% 
  .[, outcome := "Log Private Avg. Wages"]

table7B_LogPrivateEarnings <- paste0(dir_eemt, "subprojects/transfers_replication/ouput/regression_models/xlsx/table7B_LogPrivateEarnings_replication.xlsx") %>% 
  readxl::read_xlsx()    %>% as.data.table() %>% .[3:4] %>% 
  .[, ...1 := zoo::na.locf(...1)] %>% melt.data.table(id.vars = c("...1")) %>% 
  .[!is.na(value)] %>% 
  .[, statistic := "coefficient"] %>% 
  .[stri_detect_fixed(value, "("), statistic := "se"]    %>% 
  .[, outcome := "Log Private Total Earnings"]
  
table7B_main <- rbind(table7B_LogPrivateEmployment_replication, table7B_LogPrivateWages) %>% 
  rbind(., table7B_LogPrivateEarnings) %>% 
  dcast.data.table(formula = outcome + statistic  ~ variable, value.var = "value" ) %>% 
  # rbind(data.table(outcome=c("IV")), ., fill=T)
  .[, Regression := "Fuzzy IV"]


table7 <- rbind(table7A_main, table7B_main, fill=T) %>%
  rbind(., table7A_lower, fill=T)


table7_present <- table7 %>% copy() %>%
  .[, outcome_hidden := outcome] %>% copy() %>%
  .[statistic=="se", outcome := ""]  %>%
    .[, statistic := NULL] %>% 
    .[, outcome_hidden := NULL] 

Code

table7_gt<- table7_present  %>% 
  group_by(Regression)  %>% 
  gt()   %>% 
   tab_spanner(
    label = "Local estimates in differences",
    columns = c(`(5)`, `(6)`, `(7)`, `(8)` )
  ) %>%
   tab_spanner(
    label = "Local estimates in levels",
    columns = c(`(1)`, `(2)`, `(3)`, `(4)` )
  )   %>%
   tab_spanner(
    label = "<4%",
    columns = c(`(1)`, `(3)`, `(5)`, `(7)` )
  )    %>% 
   tab_spanner(
    label = "<2%",
    columns = c(`(2)`, `(4)`, `(6)`, `(8)` )
  )  %>%  tab_options(
    column_labels.font.size = "small",
    table.font.size = "small",
    row_group.font.size = "small",
    data_row.padding = px(3)
  ) %>%
  sub_missing(
  columns = everything(),
  rows = everything(),
  missing_text = "-"
)

# gtsave(table7_gt, paste0(dir_transfers_replication, "ouput/regression_models/gt_output/table7.pdf"))

table7_gt

Replicating Table 7: 2002-2019

	<4%	<2%	<4%	<2%	<4%	<2%	<4%	<2%
outcome	Local estimates in levels				Local estimates in differences
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
Reduced Form
Log Private Avg. Wages	0.010714	-0.015418	-0.013432	-0.002010	-0.046218	-0.065479	-0.058503	-0.054189
	(0.028)	(0.035)	(0.031)	(0.039)	(0.034)	(0.047)	(0.037)	(0.048)
Log Private Employment	-0.001516	-0.063229	-0.015277	-0.042993	-0.004728	0.030964	-0.021339	0.029839
	(0.055)	(0.058)	(0.063)	(0.073)	(0.044)	(0.054)	(0.053)	(0.064)
Log Private Total Earnings	0.007305	-0.081581	-0.029342	-0.047140	-0.053057	-0.037175	-0.084265	-0.029728
	(0.059)	(0.067)	(0.069)	(0.077)	(0.053)	(0.067)	(0.057)	(0.071)
Fuzzy IV
Log Private Avg. Wages	0.0115	-0.0164	-0.0148	-0.00219	-	-	-	-
	(0.0289)	(0.0361)	(0.0331)	(0.0404)	-	-	-	-
Log Private Employment	-0.00162	-0.0674	-0.0168	-0.0469	-	-	-	-
	(0.0571)	(0.0592)	(0.0677)	(0.0768)	-	-	-	-
Log Private Total Earnings	0.00781	-0.0869	-0.0322	-0.0515	-	-	-	-
	(0.0617)	(0.0685)	(0.0742)	(0.0803)	-	-	-	-
Municipality FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
State-year dummies	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Cutoff-year dummies	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
First-order polynomial	No	No	Yes	Yes	No	No	Yes	Yes

Table 7 - First Stage

Code

## Clean table 7 part A
# clean the lower part of the table --------------
table7A_first_stage <- paste0(dir_eemt, "subprojects/transfers_replication/ouput/regression_models/xlsx/table7A_firststage.xlsx") %>% 
  readxl::read_xlsx()  %>% as.data.table() %>% 
  rename(., outcome=...1) %>% .[1:15]


# lfpm <- table7A_first_stage %>%   copy() %>%   .[3:4] %>% .[, outcome := na.locf(outcome)] %>% 
#   .[, .(outcome, `(1)`, `(2)`, `(3)`, `(4)`)]
# Slfpm <- table7A_first_stage %>%   copy() %>%   .[5:6 ] %>% .[, outcome := na.locf(outcome)] %>% 
#     .[, .(outcome, `(5)`, `(6)`, `(7)`, `(8)`)]

Code

table7A_first_stage %>% 
  .[is.na(outcome), outcome := "" ] %>% 
  gt() %>%
  sub_missing(
  columns = everything(),
  rows = everything(),
  missing_text = "-"
)  %>% 
   tab_spanner(
    label = "Local estimates in differences",
    columns = c(`(5)`, `(6)`, `(7)`, `(8)` )
  ) %>%
   tab_spanner(
    label = "Local estimates in levels",
    columns = c(`(1)`, `(2)`, `(3)`, `(4)` )
  )   %>%
   tab_spanner(
    label = "<4%",
    columns = c(`(1)`, `(3)`, `(5)`, `(7)` )
  )    %>% 
   tab_spanner(
    label = "<2%",
    columns = c(`(2)`, `(4)`, `(6)`, `(8)` )
  )  %>%  tab_options(
    column_labels.font.size = "small",
    table.font.size = "small",
    row_group.font.size = "small",
    data_row.padding = px(3)
  ) 

Table 8.1: Table 7: First Stage

(a) The following table presents the first stage results from table 7. The first 4 columns present estimates of log(FPM) ~ log(Theoretical FPM) + XB. The latter four present estimates using the first differences of the same variables; log(S.FPM) ~ log(S.Theoretical FPM) + XB

	<4%	<2%	<4%	<2%	<4%	<2%	<4%	<2%
outcome	Local estimates in levels				Local estimates in differences
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
VARIABLES	lfpm	lfpm	lfpm	lfpm	Slfpm	Slfpm	Slfpm	Slfpm
	-	-	-	-	-	-	-	-
lfpmhat	0.979657***	0.985886***	0.951679***	0.961183***	-	-	-	-
	(0.005)	(0.006)	(0.006)	(0.007)	-	-	-	-
Slfpmhat	-	-	-	-	1.003015***	1.004928***	0.993876***	0.993824***
	-	-	-	-	(0.007)	(0.009)	(0.007)	(0.010)
Constant	0.304477***	0.210836**	0.726752***	0.583636***	0.000764**	0.002120***	0.002204***	0.003798***
	(0.072)	(0.089)	(0.087)	(0.107)	(0.000)	(0.001)	(0.001)	(0.001)
	-	-	-	-	-	-	-	-
Observations	12,969	6,171	12,969	6,171	9,301	3,617	9,301	3,617
R-squared	0.997	0.998	0.997	0.998	0.958	0.960	0.959	0.960
Municipality FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
State-year dummies	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Cutoff-year dummies	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
First-order polynomial	No	No	Yes	Yes	No	No	Yes	Yes

?(caption)

8.2 Main Figures - Common Sample - 2002-2014

Figure 4

Code

knitr::include_graphics("images/replication_images/figure_4_replication_common.png")

This figure replicates figure 4 from the oritinal Regional Transfers Multipliers paper using our 2002-2019 sample of interior municipalities. The FPM transfer data is deflated to 2000BRL. It is collected for all interior municipalities (as opposed to capitals or reserve municipalities).

FPM transfers around the cutoffs - 2000BRL Deflators

Figure 4 - Using Theoretical Transfer

Code

knitr::include_graphics("images/replication_images/figure_4_lfpmhat_replication_common.png")

This figure replicates figure 4 from the oritinal Regional Transfers Multipliers paper using our 2002-2019 sample of interior municipalities. The FPM transfer data is deflated to 2000BRL. It is collected for all interior municipalities (as opposed to capitals or reserve municipalities).

Theoretical FPM transfers around the cutoffs - 2000BRL Deflators

7 A) Residual Log(Total Private Sector Earnings)

Code

knitr::include_graphics("images/replication_images/figure_7a_replication_common.png")

This figure replicates figure 7a from the oritinal Regional Transfers Multipliers paper using our 2002-2019 sample of interior municipalities. It presents the residual log total earnings in the private sector. Earnings are deflated to 2000 BRL levels using the IPCA deflator. Here, we keep rows (mun-year) common to our data-set, which may include missing observatios for their and/or our data.

Residual log total earnings in the private sector - 2000BRL Deflators

B) Residual Log( Private Sector Employment)

Code

knitr::include_graphics("images/replication_images/figure_7b_replication_common.png")

This figure replicates figure 7b from the oritinal Regional Transfers Multipliers paperusing using our 2002-2019 sample of interior municipalities. It presents the residual log total employment in the private sector. Earnings in this data-set are deflated to 2000 BRL levels using the IPCA deflator.

Residual log total employment in the private sector - 2000BRL Deflators

C) Residual Log(Avg Private Sector Earnings)

Code

knitr::include_graphics("images/replication_images/figure_7c_replication_common.png")

This figure replicates figure 7c from the oritinal Regional Transfers Multipliers paper using our 2002-2019 sample of interior municipalities. It presents the residual log average municipality-year wages in the private sector. Earnings in this data-set are deflated to 2000 BRL levels using the IPCA deflator. Here, we keep rows (mun-year) common to our data-set, which may include missing observatios for their and/or our data. To be clear, we only look at firms that ought to be in the data, according to the rules specified in the paper.

Residual log average municipality-year wages in the private sector - 2000BRL Deflators

Table 7

Code

## Clean table 7 part A

# clean the lower part of the table --------------
table7A_lower_incommon <- paste0(dir_eemt, "subprojects/transfers_replication/ouput/regression_models/xlsx/table7A_LogPrivateEmployment-InCommon_replication.xlsx") %>% 
  readxl::read_xlsx()  %>% as.data.table() %>% .[10:13] %>% .[, statistic := ""] %>% rename(., outcome=...1) %>% 
.[, Regression := ""]


# Clean Table A; numbers -----------------
table7A_LogPrivateEmployment2 <- paste0(dir_eemt, "subprojects/transfers_replication/ouput/regression_models/xlsx/table7A_LogPrivateEmployment-InCommon_replication.xlsx") %>% 
  readxl::read_xlsx()  %>% as.data.table() %>% .[3:4] %>% 
  .[, ...1 := zoo::na.locf(...1)] %>% melt.data.table(id.vars = c("...1")) %>% 
  .[!is.na(value)] %>% 
  .[, statistic := "coefficient"] %>% 
  .[stri_detect_fixed(value, "("), statistic := "se"]  %>% 
  .[, outcome := "Log Private Employment"]
  
table7A_LogPrivateWages2 <- paste0(dir_eemt, "subprojects/transfers_replication/ouput/regression_models/xlsx/table7A_LogPrivateWages-InCommon_replication.xlsx") %>% 
  readxl::read_xlsx()   %>% as.data.table() %>% .[3:4] %>% 
  .[, ...1 := zoo::na.locf(...1)] %>% melt.data.table(id.vars = c("...1")) %>% 
  .[!is.na(value)] %>% 
  .[, statistic := "coefficient"] %>% 
  .[stri_detect_fixed(value, "("), statistic := "se"]     %>% 
  .[, outcome := "Log Private Avg. Wages"]


table7A_LogPrivateEarnings2 <- paste0(dir_eemt, "subprojects/transfers_replication/ouput/regression_models/xlsx/table7A_LogPrivateEarnings-InCommon_replication.xlsx") %>% 
  readxl::read_xlsx()    %>% as.data.table() %>% .[3:4] %>% 
  .[, ...1 := zoo::na.locf(...1)] %>% melt.data.table(id.vars = c("...1")) %>% 
  .[!is.na(value)] %>% 
  .[, statistic := "coefficient"] %>% 
  .[stri_detect_fixed(value, "("), statistic := "se"]    %>% 
  .[, outcome := "Log Private Total Earnings"]
  
table7A_main_incommon <- rbind(table7A_LogPrivateEmployment2, table7A_LogPrivateWages2) %>% 
  rbind(., table7A_LogPrivateEarnings2) %>% 
  dcast.data.table(formula = outcome + statistic  ~ variable, value.var = "value" ) %>% 
  .[, Regression := "Reduced Form"]

# %>%    rbind(data.table(outcome=c("Reduced Form")), ., fill=T)

# Clean Table B; numbers -----------------
table7B_LogPrivateEmployment2 <- paste0(dir_eemt, "subprojects/transfers_replication/ouput/regression_models/xlsx/table7B_LogPrivateEmployment-InCommon_replication.xlsx") %>% 
  readxl::read_xlsx()  %>% as.data.table() %>% .[3:4] %>% 
  .[, ...1 := zoo::na.locf(...1)] %>% melt.data.table(id.vars = c("...1")) %>% 
  .[!is.na(value)] %>% 
  .[, statistic := "coefficient"] %>% 
  .[stri_detect_fixed(value, "("), statistic := "se"]  %>% 
  .[, outcome := "Log Private Employment"]
  
table7B_LogPrivateWages2 <- paste0(dir_eemt, "subprojects/transfers_replication/ouput/regression_models/xlsx/table7B_LogPrivateWages-InCommon_replication.xlsx") %>% 
  readxl::read_xlsx()   %>% as.data.table() %>% .[3:4] %>% 
  .[, ...1 := zoo::na.locf(...1)] %>% melt.data.table(id.vars = c("...1")) %>% 
  .[!is.na(value)] %>% 
  .[, statistic := "coefficient"] %>% 
  .[stri_detect_fixed(value, "("), statistic := "se"]     %>% 
  .[, outcome := "Log Private Avg. Wages"]

table7B_LogPrivateEarnings2 <- paste0(dir_eemt, "subprojects/transfers_replication/ouput/regression_models/xlsx/table7B_LogPrivateEarnings-InCommon_replication.xlsx") %>% 
  readxl::read_xlsx()    %>% as.data.table() %>% .[3:4] %>% 
  .[, ...1 := zoo::na.locf(...1)] %>% melt.data.table(id.vars = c("...1")) %>% 
  .[!is.na(value)] %>% 
  .[, statistic := "coefficient"] %>% 
  .[stri_detect_fixed(value, "("), statistic := "se"]    %>% 
  .[, outcome := "Log Private Total Earnings"]
  
table7B_main_incommon <- rbind(table7B_LogPrivateEmployment2, table7B_LogPrivateWages2) %>% 
  rbind(., table7B_LogPrivateEarnings2) %>% 
  dcast.data.table(formula = outcome + statistic  ~ variable, value.var = "value" ) %>% 
  # rbind(data.table(outcome=c("IV")), ., fill=T)
  .[, Regression := "Fuzzy IV"]

table7_incommon <- rbind(table7A_main_incommon, table7B_main_incommon, fill=T) %>%
  rbind(., table7A_lower_incommon, fill=T) 

table7_incommon_present <- table7_incommon %>% copy() %>%
  .[, outcome_hidden := outcome] %>% copy() %>%
  .[statistic=="se", outcome := ""]  %>%
    .[, statistic := NULL] %>% 
    .[, outcome_hidden := NULL] 

Code

table7_incommon_gt <- table7_incommon_present %>% 
  group_by(Regression) %>% 
  gt()  %>% 
   tab_spanner(
    label = "Local estimates in levels",
    columns = c(`(1)`, `(2)`, `(3)`, `(4)` )
  )   %>%
   tab_spanner(
    label = "<4%",
    columns = c(`(1)`, `(3)` )
  )    %>% 
   tab_spanner(
    label = "<2%",
    columns = c(`(2)`, `(4)` )
  )  %>%  tab_options(
    column_labels.font.size = "small",
    table.font.size = "small",
    row_group.font.size = "small",
    data_row.padding = px(3)
  ) %>%
  sub_missing(
  columns = everything(),
  rows = everything(),
  missing_text = "-"
)

# gtsave(table7_incommon_gt, paste0(dir_transfers_replication, "ouput/regression_models/gt_output/table7_incommon.pdf"))

table7_incommon_gt

Table 8.2: Replicating Table 7: 2002-2014

	<4%	<2%	<4%	<2%
outcome	Local estimates in levels
	(1)	(2)	(3)	(4)
Reduced Form
Log Private Avg. Wages	0.061591**	0.050169	0.033099	0.044230
	(0.031)	(0.043)	(0.034)	(0.051)
Log Private Employment	0.003198	-0.026708	0.019555	0.010105
	(0.057)	(0.071)	(0.065)	(0.085)
Log Private Total Earnings	0.066876	0.023298	0.056161	0.054546
	(0.060)	(0.082)	(0.066)	(0.096)
Fuzzy IV
Log Private Avg. Wages	0.0656**	0.0529	0.0363	0.0479
	(0.0318)	(0.0428)	(0.0362)	(0.0523)
Log Private Employment	0.00341	-0.0281	0.0215	0.0109
	(0.0597)	(0.0708)	(0.0693)	(0.0871)
Log Private Total Earnings	0.0712	0.0245	0.0617	0.0590
	(0.0623)	(0.0820)	(0.0712)	(0.0990)
Municipality FE	Yes	Yes	Yes	Yes
State-year dummies	Yes	Yes	Yes	Yes
Cutoff-year dummies	Yes	Yes	Yes	Yes
First-order polynomial	No	No	Yes	Yes

This table replicates the output from table 7 using the intersection of the municipalities in our data and theirs, between 2002-2014.

Table 7 - First Stage

Code

## Clean table 7 part A

"C:/Users/alckm/Dropbox/EEMT/subprojects/transfers_replication/ouput/regression_models/xlsx/table7A_firststage_incommon.xlsx"

[1] "C:/Users/alckm/Dropbox/EEMT/subprojects/transfers_replication/ouput/regression_models/xlsx/table7A_firststage_incommon.xlsx"

Code

# clean the lower part of the table --------------
table7A_first_stage_common <- paste0(dir_eemt, "subprojects/transfers_replication/ouput/regression_models/xlsx/table7A_firststage_incommon.xlsx") %>% 
  readxl::read_xlsx()  %>% as.data.table() %>% 
  rename(., outcome=...1) %>% .[1:15]


# lfpm <- table7A_first_stage %>%   copy() %>%   .[3:4] %>% .[, outcome := na.locf(outcome)] %>% 
#   .[, .(outcome, `(1)`, `(2)`, `(3)`, `(4)`)]
# Slfpm <- table7A_first_stage %>%   copy() %>%   .[5:6 ] %>% .[, outcome := na.locf(outcome)] %>% 
#     .[, .(outcome, `(5)`, `(6)`, `(7)`, `(8)`)]

Code

table7A_first_stage_common %>% 
  .[is.na(outcome), outcome := "" ] %>% 
  gt() %>%
  sub_missing(
  columns = everything(),
  rows = everything(),
  missing_text = "-"
)  %>% 
   tab_spanner(
    label = "Local estimates in differences",
    columns = c(`(5)`, `(6)`, `(7)`, `(8)` )
  ) %>%
   tab_spanner(
    label = "Local estimates in levels",
    columns = c(`(1)`, `(2)`, `(3)`, `(4)` )
  )   %>%
   tab_spanner(
    label = "<4%",
    columns = c(`(1)`, `(3)`, `(5)`, `(7)` )
  )    %>% 
   tab_spanner(
    label = "<2%",
    columns = c(`(2)`, `(4)`, `(6)`, `(8)` )
  )  %>%  tab_options(
    column_labels.font.size = "small",
    table.font.size = "small",
    row_group.font.size = "small",
    data_row.padding = px(3)
  ) 

Table 8.3: Table 7: First Stage

(a) The following table presents the first stage results from table 7. The first 4 columns present estimates of log(FPM) ~ log(Theoretical FPM) + XB. The latter four present estimates using the first differences of the same variables; log(S.FPM) ~ log(S.Theoretical FPM) + XB

	<4%	<2%	<4%	<2%	<4%	<2%	<4%	<2%
outcome	Local estimates in levels				Local estimates in differences
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
VARIABLES	lfpm	lfpm	lfpm	lfpm	Slfpm	Slfpm	Slfpm	Slfpm
	-	-	-	-	-	-	-	-
lfpmhat	0.992804***	1.001072***	0.960541***	0.972053***	-	-	-	-
	(0.006)	(0.007)	(0.007)	(0.009)	-	-	-	-
Slfpmhat	-	-	-	-	1.018091***	1.032847***	1.008031***	1.018742***
	-	-	-	-	(0.008)	(0.010)	(0.008)	(0.012)
Constant	0.104443	-0.020119	0.589723***	0.416474***	-0.000339	0.000363	0.001393	0.002212
	(0.088)	(0.112)	(0.107)	(0.131)	(0.000)	(0.001)	(0.001)	(0.001)
	-	-	-	-	-	-	-	-
Observations	9,079	4,202	9,079	4,202	6,141	2,310	6,141	2,310
R-squared	0.997	0.998	0.997	0.998	0.961	0.963	0.961	0.964
Municipality FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
State-year dummies	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Cutoff-year dummies	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
First-order polynomial	No	No	Yes	Yes	No	No	Yes	Yes

?(caption)

Corbi, Raphael, Elias Papaioannou, and Paolo Surico. 2018. “Regional Transfer Multipliers.” The Review of Economic Studies 86 (5): 1901–34. https://doi.org/10.1093/restud/rdy069.