Services, Jobs and Economic Development in Africa - Online appendix

Next content

Uncategorized

Southeastern Europe: Transitions, Prospects, Crossroads

Recent publications Klodiana Beshku, The EU as a geopolitical power: the case of the Western Balkans region as the periphery to the core, EUI Robert Schuman Centre Working Paper, 2021 Klodiana Beshku, Orjana Mullisi, The European Union as a...

Social Media

This page hosts the online appendix for the article ‘Services, Jobs and Economic Development in Africa’ by Leonardo Baccini, Matteo Fiorini, Bernard Hoekman and Marco Sanfilippo. This online appendix showcases graphs and maps on sectoral and occupational dynamics in Africa over available census waves in the IPUMS International database. It also includes a note for replicating the data cleaning procedure. This appendix was originally developed by Lorenzo Sileci, a Research Associate at the EUI Robert Schuman Centre for Advanced Studies (Global Governance Programme), and PhD Student in Environmental Economics at the London School of Economics and Political Science (LSE). If you use the graphs, please cite us as ‘Baccini, L., Fiorini, M., Hoekman, B. and Sanfilippo, M. (2021) Services, Jobs and Economic Development in Africa, EUI RSC Working Paper 2021/82’.

Aggregate Occupational Dynamics

Scatterplots showing aggregate occupational dynamics across the last two available waves in the IPUMS dataset.

Go to the data

Aggregate Sectoral Dynamics

Scatterplots showing aggregate sectoral dynamics across the last two available waves in the IPUMS dataset.

Go to the data

Disaggregated Occupational Dynamics

Scatterplots showing disaggregated occupational dynamics across the last two available waves in the IPUMS dataset.

Go to the data

Disaggregated Sectoral Dynamics

Scatterplots showing disaggregated sectoral dynamics across the last two available waves in the IPUMS dataset.

Go to the data

Sectoral Dynamics Maps

Series of maps showing the dynamics of aggregate sectors across countries included in the dataset. Maps show admin-level sectoral shares for the last two available waves and the differential between waves.

Go to the data

Within-sector Labour Force Composition

Bar plots showing within-sector labour force composition across the last two available waves in the IPUMS dataset.

Go to the data

Aggregate Occupational Dynamics

Here we provide a series of scatterplots which describe the dynamics of surveyed individuals' sectoral allocation in the last two available waves of the IPUMS dataset. We first examine the whole dataset and then refine the exercise by looking at individual countries

As we can see in the scatterplot below, the general sectoral dynamics for the African countries included in the dataset underpin a movement from the primary to the tertiary sector over the years.

Click on the image to view it fullscreen or to download it

Back to menu

Aggregate Sectoral Dynamics

As we can see in the scatterplot below, the general sectoral dynamics for the African countries included in the dataset underpin a movement from the primary to the tertiary sector over the years.

Click on the image to view it fullscreen or to download it

Back to menu

Disaggregated Occupational Dynamics

Here we provide a series of faceted scatterplots which describe the dynamics of surveyed individuals' occupational allocation in the last two available waves of the IPUMS dataset. We first examine the whole dataset and then refine the exercise by looking at individual countries. Here, each occupational category is reported in a different facet.

Click on the image to view it fullscreen or to download it

Back to menu

Disaggregated Sectoral Dynamics

Here we provide a series of faceted scatterplots which describe the dynamics of surveyed individuals' sectoral allocation in the last two available waves of the IPUMS dataset. We first examine the whole dataset and then refine the exercise by looking at individual countries. Here, each sector is reported in a different facet.

Click on the image to view it fullscreen or to download it

Back to menu

Aggregate Sectoral Dynamics Maps

Click on the image to view it fullscreen or to download it

Back to menu

IPUMS Data Cleaning Replication Code

Back to menu

This replication note has been written by Lorenzo Sileci (Research Associate EUI RSCAS, GGP and LSE PhD candidate) in July 2021

Collapsing IPUMS Data at the lowest available geographic scale

In the following, I illustrate how I have aggregated the IPUMS survey data at the finest available geographic scale (administrative level 2 for most countries, except Botswana - where only administrative levek 1 is available - and Rwanda - where only the last wave is available at the administrative 2 level, and hence data from said wave has been resampled to the administrative level 1).

I use the Benin's last wave as an example of how the aggregation has been performed.

First of all, data must be read into R:

### BENIN ###

library(tidyverse)

## ── Attaching packages ─────────────────────────────────────── tidyverse 1.3.1 ──

## ✓ ggplot2 3.3.3     ✓ purrr   0.3.4
## ✓ tibble  3.1.2     ✓ dplyr   1.0.6
## ✓ tidyr   1.1.3     ✓ stringr 1.4.0
## ✓ readr   1.4.0     ✓ forcats 0.5.1

## ── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
## x dplyr::filter() masks stats::filter()
## x dplyr::lag()    masks stats::lag()

# Import data from IPUMS

if (!require("ipumsr")) stop("Reading IPUMS data into R requires the ipumsr package. It can be installed using the following command: install.packages('ipumsr')")

## Loading required package: ipumsr

ddi <- read_ipums_ddi("~/Dropbox/Work/RA/services micro ssa/Lorenzo_RA_Work/IPUMS_data/ipumsi_00030.xml")
benin <- read_ipums_micro(ddi)

## Use of data from IPUMS-International is subject to conditions including that
## users should cite the data appropriately. Use command `ipums_conditions()` for
## more details.

# Examine variable names

colnames(benin)

##  [1] "COUNTRY"   "YEAR"      "SAMPLE"    "SERIAL"    "HHWT"      "URBAN"    
##  [7] "GEOLEV1"   "GEOLEV2"   "PERNUM"    "PERWT"     "RESIDENT"  "NCHILD"   
## [13] "NCHLT5"    "AGE"       "SEX"       "MARST"     "MARSTD"    "CHBORN"   
## [19] "CHSURV"    "BIRTHSLYR" "BIRTHSURV" "MORTMOT"   "RELIGION"  "RELIGIOND"
## [25] "ETHNICBJ"  "LANGBJ"    "SCHOOL"    "LIT"       "EDATTAIN"  "EDATTAIND"
## [31] "YRSCHOOL"  "EMPSTAT"   "EMPSTATD"  "LABFORCE"  "OCCISCO"   "INDGEN"   
## [37] "IND"       "MIGRATEP"

# Inspect hh and person weights: in this instance they are all equal

summary(benin$HHWT)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##      10      10      10      10      10      10

summary(benin$PERWT)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##      10      10      10      10      10      10

# Import IPUMS_GIS world level adm2 shapefile and restrict to Benin, renaming geolev2 variable for merging purposes

benin.map <- sf::read_sf(dsn = "~/Dropbox/Work/RA/services micro ssa/Lorenzo_RA_Work/world_geolev2_2019", layer = "world_geolev2_2019") %>%
  filter(CNTRY_NAME == "Benin") %>%
  rename(GEOLEV2 = "GEOLEVEL2") %>%
  dplyr::select(-CNTRY_CODE, -BPL_CODE)

Optionally, in the _wa datasets, the waves get restricted to working age population only:

# benin <- benin %>%
#   filter(!AGE==999 & AGE > 14 & AGE < 65)

Population Shares

I first calculate population counts for each administrative area by counting observations within each precinct. I then merge the dataset restricted to population with the map of Benin's administrative units, in order to be able to plot them. Here, I offer a spatial representation of the population distribution via a chloropleth map.

# 1. Population counts

benin_pop_2013 <- benin %>%
  filter(YEAR == 2013) %>%
  group_by(GEOLEV2) %>%
  tally(name = "population")

# Merge with calculated shares

benin.pop.2013.spatial <- merge(benin.map, benin_pop_2013, by = "GEOLEV2")
rm(benin_pop_2013)

# Plot Population Shares:

ggplot() +
  geom_sf(data = benin.pop.2013.spatial$geometry) +
  geom_sf(data = benin.pop.2013.spatial, aes(fill=population)) +
  scale_fill_viridis_c(option = "plasma") + theme_void()

Urban/Rural Shares

As for in all categorical variables, in order to calculate geographical shares for each level of the variable, I first need to "dummify" each category. Here, I exclude observations labelled as "unknown" (the same will be done for all "not in universe" observations), and calculate 1/0 dummies for each level of the URBAN variable. Then, I aggregate the two dummies at the administrative level 2, and calculate the shares by multiplying them with the individual survey weights provided by IPUMS:

# 1. URBAN/RURAL SHARES

benin_ruralurb <- benin %>%
  filter(!URBAN == 9) %>%
  mutate(urban = ifelse(URBAN == 2, 1, 0),
         rural = ifelse(URBAN == 1, 1, 0))

# summarise for each district 

benin_ru_2013 <- benin_ruralurb %>%
  filter(YEAR == 2013) %>%
  group_by(GEOLEV2) %>%
  summarise(urban = weighted.mean(urban, PERWT),
            rural = weighted.mean(rural, PERWT))


# Merge with calculated shares

benin.ru.2013.spatial <- merge(benin.map, benin_ru_2013, by = "GEOLEV2")
rm(benin_ruralurb, benin_ru_2013)

# Plot Urban Shares:

ggplot() +
  geom_sf(data = benin.ru.2013.spatial$geometry) +
  geom_sf(data = benin.ru.2013.spatial, aes(fill=urban)) +
  scale_fill_viridis_c(option = "plasma")  + theme_void()

Residency Shares

Same thing for the share of resident surveyed individuals:

# 3. RESIDENCY SHARES

benin_resident <- benin %>%
  mutate(resident_present = ifelse(RESIDENT == 1, 1, 0),
         resident_absent = ifelse(RESIDENT == 2, 1, 0),
         resident_visitor = ifelse(RESIDENT == 3, 1, 0),
         resident_defacto = ifelse(RESIDENT == 4, 1, 0),
         resident_unknown = ifelse(RESIDENT == 9, 1, 0))

# summarise for each district 

benin_re_2013 <- benin_resident %>%
  filter(YEAR == 2013) %>%
  group_by(GEOLEV2) %>%
  summarise(resident_present = weighted.mean(resident_present, PERWT),
            resident_absent  = weighted.mean(resident_absent,  PERWT),
            resident_visitor = weighted.mean(resident_visitor, PERWT),
            resident_defacto = weighted.mean(resident_defacto, PERWT),
            resident_unknown = weighted.mean(resident_unknown, PERWT))

# Merge with calculated shares

benin.re.2013.spatial <- merge(benin.map, benin_re_2013, by = "GEOLEV2")
rm(benin_resident, benin_re_2013)

Number of Children and Children under 5

These two variables need not be dummified as the weighted mean per district already yields the average number of children and children under 5 years of age for each administrative area:

# 4. NUMBER OF CHILDREN DISTRICT SHARES

# summarise for each district 

benin_nchild_2013 <- benin %>%
  filter(YEAR == 2013) %>%
  group_by(GEOLEV2) %>%
  summarise(n_children = weighted.mean(NCHILD, PERWT))

# Merge with calculated shares

benin.nchild.2013.spatial <- merge(benin.map, benin_nchild_2013, by = "GEOLEV2")
rm(benin_nchild_2013)


# 5. NUMBER OF CHILDREN UNDER 5 DISTRICT SHARES

# summarise for each district 

benin_nchlt5_2013 <- benin %>%
  filter(YEAR == 2013) %>%
  group_by(GEOLEV2) %>%
  summarise(n_children_u5 = weighted.mean(NCHLT5, PERWT))

# Merge with calculated shares

benin.nchlt5.2013.spatial <- merge(benin.map, benin_nchlt5_2013, by = "GEOLEV2")
rm(benin_nchlt5_2013)

Age

Here I calculate mean age for each district, and exclude "not in universe" observations (AGE == 999):

# 6. AGE DISTRICT SHARES

# summarise for each district 

benin_age_2013 <- benin %>%
  filter(!AGE==999 & YEAR == 2013) %>%
  group_by(GEOLEV2) %>%
  summarise(age = weighted.mean(AGE, PERWT))

# Merge with calculated shares

benin.age.2013.spatial <- merge(benin.map, benin_age_2013, by = "GEOLEV2")
rm(benin_age_2013)

Gender

Gender is a binary dummy without unknowns and NIU:

# 7. GENDER SHARES

benin_sex <- benin %>%
  mutate(female = ifelse(SEX == 2, 1, 0),
         male = ifelse(SEX == 1, 1, 0))

# summarise for each district 

benin_sex_2013 <- benin_sex %>%
  filter(YEAR == 2013) %>%
  group_by(GEOLEV2) %>%
  summarise(female = weighted.mean(female, PERWT),
            male = weighted.mean(male, PERWT))

# Merge with calculated shares

benin.sex.2013.spatial <- merge(benin.map, benin_sex_2013, by = "GEOLEV2")
rm(benin_sex, benin_sex_2013)

Marital Status

Here, I need to exclude both unknowns and NIU and to dummify the 4 categories of the marital status variable in order to calculate the incidence of each category for each administrative unit.

benin_marst <- benin %>%
  filter(!MARST == 0 & !MARST==9) %>%
  mutate(single = ifelse(MARST == 1,1,0),
         married = ifelse(MARST == 2, 1, 0),
         divorced = ifelse(MARST == 3, 1, 0),
         widowed = ifelse(MARST == 4, 1, 0))

benin_marst_2013 <- benin_marst %>%
  filter(YEAR == 2013) %>%
  group_by(GEOLEV2) %>%
  summarise(single = weighted.mean(single, PERWT),
            married = weighted.mean(married, PERWT),
            divorced = weighted.mean(divorced, PERWT),
            widowed = weighted.mean(widowed, PERWT))

# Merge with calculated shares

benin.marst.2013.spatial <- merge(benin.map, benin_marst_2013, by = "GEOLEV2")
rm(benin_marst, benin_marst_2013)

Newborn Children and Surviving Newborn Children

As before, I exclude unknown and NIU observations, dummify each category and calculate shares for each category:

# 9. NEWBORN CHILDREN SHARES

benin_chborn <- benin %>%
  filter(!CHBORN == 98 & !CHBORN==99) %>%
  mutate(nochildren = ifelse(CHBORN == 0, 1, 0),
         onechild = ifelse(CHBORN == 1, 1, 0),
         twochildren = ifelse(CHBORN == 2, 1, 0),
         threechildren = ifelse(CHBORN == 30, 1, 0))


benin_chborn_2013 <- benin_chborn %>%
  filter(YEAR == 2013) %>%
  group_by(GEOLEV2) %>%
  summarise(nochildren = weighted.mean(nochildren, PERWT),
            onechild = weighted.mean(onechild, PERWT),
            twochildren = weighted.mean(twochildren, PERWT),
            threechildren = weighted.mean(threechildren, PERWT))

# Merge with calculated shares

benin.chborn.2013.spatial <- merge(benin.map, benin_chborn_2013, by = "GEOLEV2")
rm(benin_chborn, benin_chborn_2013)


# 10. SURVIVING CHILDREN SHARES

benin_chsurv <- benin %>%
  filter(!CHSURV == 98 & !CHSURV==99) %>%
  mutate(nosurv_children = ifelse(CHSURV == 0, 1, 0),
         onesurv_children = ifelse(CHSURV ==1,1,0),
         twosurv_children = ifelse(CHSURV == 2, 1, 0),
         threesurv_children = ifelse(CHSURV == 30, 1, 0))

benin_chsurv_2013 <- benin_chsurv %>%
  filter(YEAR == 2013) %>%
  group_by(GEOLEV2) %>%
  summarise(nosurv_children = weighted.mean(nosurv_children, PERWT),
            onesurv_children = weighted.mean(onesurv_children, PERWT),
            twosurv_children = weighted.mean(twosurv_children, PERWT),
            threesurv_children = weighted.mean(threesurv_children, PERWT))

# Merge with calculated shares

benin.chsurv.2013.spatial <- merge(benin.map, benin_chsurv_2013, by = "GEOLEV2")
rm(benin_chsurv, benin_chsurv_2013)

Last Year Births and Surviving Births

Exactly as before, I exclude unknown and NIU observations, dummify each category and calculate shares for each category:

# 11. LAST YEAR BIRTHS SHARES

benin_births <- benin %>%
  filter(!BIRTHSLYR == 9 & !BIRTHSLYR == 8) %>%
  mutate(no_births = ifelse(BIRTHSLYR == 0, 1, 0),
         one_three_births = ifelse(BIRTHSLYR ==1,1,0),
         four_more_births = ifelse(BIRTHSLYR == 4, 1, 0))

benin_births_2013 <- benin_births %>%
  filter(YEAR == 2013) %>%
  group_by(GEOLEV2) %>%
  summarise(no_births = weighted.mean(no_births, PERWT),
            one_three_births = weighted.mean(one_three_births, PERWT),
            four_more_births = weighted.mean(four_more_births, PERWT))

# Merge with calculated shares

benin.births.2013.spatial <- merge(benin.map, benin_births_2013, by = "GEOLEV2")
rm(benin_births, benin_births_2013)


# 12. LAST YEAR SURVIVING BIRTHS SHARES

benin_birthsurv <- benin %>%
  filter(!BIRTHSURV == 9 & !BIRTHSURV == 8) %>%
  mutate(one_three_birthsurv = ifelse(BIRTHSURV ==1,1,0),
         four_more_birthsurv = ifelse(BIRTHSURV == 4, 1, 0))

benin_birthsurv_2013 <- benin_birthsurv %>%
  filter(YEAR == 2013) %>%
  group_by(GEOLEV2) %>%
  summarise(one_three_birthsurv = weighted.mean(one_three_birthsurv, PERWT),
            four_more_birthsurv = weighted.mean(four_more_birthsurv, PERWT))

# Merge with calculated shares

benin.birthsurv.2013.spatial <- merge(benin.map, benin_birthsurv_2013, by = "GEOLEV2")
rm(benin_birthsurv, benin_birthsurv_2013)

Mother Mortality

# 12. MOTHER MORTALITY SHARES

benin_mortmot <- benin %>%
  filter(!MORTMOT == 9 & !MORTMOT == 8) %>%
  mutate(mother_alive = ifelse(MORTMOT == 1, 1, 0),
         mother_dead = ifelse(MORTMOT == 2, 1, 0))

benin_mortmot_2013 <- benin_mortmot %>%
  filter(YEAR == 2013) %>%
  group_by(GEOLEV2) %>%
  summarise(mother_alive = weighted.mean(mother_alive, PERWT),
            mother_dead = weighted.mean(mother_dead, PERWT))

# Merge with calculated shares

benin.mortmot.2013.spatial <- merge(benin.map, benin_mortmot_2013, by = "GEOLEV2")
rm(benin_mortmot, benin_mortmot_2013)

Religion

# 13. RELIGION SHARES

benin_religion <- benin %>%
  filter(!RELIGION == 0 & !RELIGION == 9) %>%
  mutate(religion_atheist = ifelse(RELIGION == 1,1,0),
         religion_buddhist = ifelse(RELIGION == 2,1,0),
         religion_hindu = ifelse(RELIGION == 3, 1, 0),
         religion_jewish = ifelse(RELIGION == 4, 1, 0),
         religion_muslim = ifelse(RELIGION == 5, 1, 0),
         religion_christian = ifelse(RELIGION == 6, 1, 0),
         religion_other = ifelse(RELIGION == 7, 1, 0))

benin_religion_2013 <- benin_religion %>%
  filter(YEAR == 2013) %>%
  group_by(GEOLEV2) %>%
  summarise(religion_atheist =weighted.mean(religion_atheist, PERWT),
            religion_buddhist =weighted.mean(religion_buddhist, PERWT),
            religion_hindu = weighted.mean(religion_hindu, PERWT),
            religion_jewish = weighted.mean(religion_jewish, PERWT),
            religion_muslim = weighted.mean(religion_muslim, PERWT),
            religion_christian =weighted.mean(religion_christian, PERWT),
            religion_other = weighted.mean(religion_other, PERWT))

# Merge with calculated shares

benin.religion.2013.spatial <- merge(benin.map, benin_religion_2013, by = "GEOLEV2")
rm(benin_religion, benin_religion_2013)

Ethnicity and Language

Here, there is a slight degree of added complexity. Ethnicity and language variables are indeed not uniform across IPUMS samples (i.e. different countries) due of course to the different ethnic and linguistic composition of each country. I directly compute the degree of ethnic and linguisitic fractionalisation at the lowest administrative level by adapting Alesina et al. (2003), "Fractionalization" and calculating: \[ FRACT_j = 1 - \sum_{i=1}^{N}{s_{ij}^{2}} \] where \(FRACT_j\) is the fractionalisation index for administrative unit \(j\), and \(s_{ij}\) is the share of ethnic/linguistic group \(i \in (1,..., N)\) in administrative unit \(j\). My code first counts the number of individuals for each ethnic/linguistic group in each administrative unit, then calculates the share \(s_{ij}\) by dividing it for the administrative unit's population, and finally applies Alesina et al. (2003)'s formula in order to calculate \(FRACT_j\). I also produce a map of ethnic fractionalisation indices to illustrate the concept.

# 14. ETHNICITY SHARES

benin_ethnic <- benin %>%
  filter(YEAR == 2013) %>%
  count(GEOLEV2, ETHNICBJ, name = "frac") 

benin_ethnic_2013 = merge(benin_ethnic, benin.pop.2013.spatial, by = "GEOLEV2") %>%
  dplyr::select(-one_of(c("CNTRY_NAME", "ADMIN_NAME", "geometry"))) %>%
  mutate(share_i2 = (frac/population)^2) %>%
  group_by(GEOLEV2) %>%
  summarise(ethnic_fract = 1-sum(share_i2))

# merge with spatial units

benin.ethnic.2013.spatial <- merge(benin.map, benin_ethnic_2013, by = "GEOLEV2")
rm(benin_ethnic, benin_ethnic_2013)

# Plot the fractionalisation index

ggplot() +
  geom_sf(data = benin.ethnic.2013.spatial$geometry) +
  geom_sf(data = benin.ethnic.2013.spatial, aes(fill=ethnic_fract)) +
  scale_fill_viridis_c(option = "plasma") + theme_void()

# 15. LANGUAGE SHARES

benin_lang <- benin %>%
  filter(YEAR == 2013) %>%
  count(GEOLEV2, LANGBJ, name = "frac") 

benin_lang_2013 = merge(benin_lang, benin.pop.2013.spatial, by = "GEOLEV2") %>%
  dplyr::select(-one_of(c("CNTRY_NAME", "ADMIN_NAME", "geometry"))) %>%
  mutate(share_i2 = (frac/population)^2) %>%
  group_by(GEOLEV2) %>%
  summarise(lang_fract = 1-sum(share_i2))

# merge with spatial units

benin.lang.2013.spatial <- merge(benin.map, benin_lang_2013, by = "GEOLEV2")
rm(benin_lang, benin_lang_2013)

School Attendance

# 16. SCHOOL ATTENDANCE SHARES

benin_school <- benin %>%
  filter(!SCHOOL == 0 & !SCHOOL == 9) %>%
  mutate(school_attend = ifelse(SCHOOL == 1, 1, 0),
         school_no =  ifelse(SCHOOL == 2, 1, 0),
         school_past = ifelse(SCHOOL == 3, 1, 0),
         school_never = ifelse(SCHOOL == 4, 1, 0))

benin_school_2013 <- benin_school %>%
  filter(YEAR == 2013) %>%
  group_by(GEOLEV2) %>%
  summarise(school_attend =weighted.mean(school_attend, PERWT),
            school_no =weighted.mean(school_no, PERWT),
            school_past = weighted.mean(school_past, PERWT),
            school_never = weighted.mean(school_never, PERWT))

# Merge with calculated shares

benin.school.2013.spatial <- merge(benin.map, benin_school_2013, by = "GEOLEV2")
rm(benin_school, benin_school_2013)

Literacy

# 17. LITERACY SHARES

benin_lit <- benin %>%
  filter(!LIT == 0 & !LIT==9) %>%
  mutate(literacy_no = ifelse(LIT == 1, 1, 0),
         literacy_yes = ifelse(LIT == 2, 1, 0))


benin_lit_2013 <- benin_lit %>%
  filter(YEAR == 2013) %>%
  group_by(GEOLEV2) %>%
  summarise(literacy_no =weighted.mean(literacy_no, PERWT),
            literacy_yes =weighted.mean(literacy_yes, PERWT))

# Merge with calculated shares

benin.lit.2013.spatial <- merge(benin.map, benin_lit_2013, by = "GEOLEV2")
rm(benin_lit, benin_lit_2013)

Educational Attainment

# 18. EDUCATIONAL ATTAINMENT SHARES

benin_edattain <- benin %>%
  filter(!EDATTAIN == 0 & !EDATTAIN == 9) %>%
  mutate(edattain_noprimary = ifelse(EDATTAIN == 1, 1, 0),
         edattain_primary = ifelse(EDATTAIN == 2, 1, 0),
         edattain_secondary = ifelse(EDATTAIN == 3, 1, 0),
         edattain_university = ifelse(EDATTAIN == 4, 1, 0))


benin_edattain_2013 <- benin_edattain %>%
  filter(YEAR == 2013) %>%
  group_by(GEOLEV2) %>%
  summarise(edattain_noprimary =weighted.mean(edattain_noprimary, PERWT),
            edattain_primary =weighted.mean(edattain_primary, PERWT),
            edattain_secondary = weighted.mean(edattain_secondary,PERWT),
            edattain_university = weighted.mean(edattain_university,PERWT))

# Merge with calculated shares

benin.edattain.2013.spatial <- merge(benin.map, benin_edattain_2013, by = "GEOLEV2")
rm(benin_edattain, benin_edattain_2013)

Years of Schooling

# 19. YEARS OF SCHOOLING DISTRICT SHARES

benin_yrschool <- benin %>%
  filter(!YRSCHOOL == 99 & !YRSCHOOL == 90 & !YRSCHOOL == 91 & 
           !YRSCHOOL == 92 & !YRSCHOOL == 93 & !YRSCHOOL == 94 & 
           !YRSCHOOL == 95 & !YRSCHOOL == 96 & !YRSCHOOL == 98)

# summarise for each district 

benin_yrschool_2013 <- benin_yrschool %>%
  filter(YEAR == 2013) %>%
  group_by(GEOLEV2) %>%
  summarise(yrschool = weighted.mean(YRSCHOOL, PERWT))

# Merge with calculated shares

benin.yrschool.2013.spatial <- merge(benin.map, benin_yrschool_2013, by = "GEOLEV2")
rm(benin_yrschool, benin_yrschool_2013)

Employment Status

# 20. EMPLOYMENT SHARES

benin_empstat <- benin %>%
  filter(!EMPSTAT == 0 & !EMPSTAT==9) %>%
  mutate(employment_yes = ifelse(EMPSTAT == 1, 1, 0),
         employment_no  = ifelse(EMPSTAT == 2, 1, 0),
         employment_inactive = ifelse(EMPSTAT == 3, 1, 0))


benin_empstat_2013 <- benin_empstat %>%
  filter(YEAR == 2013) %>%
  group_by(GEOLEV2) %>%
  summarise(employment_yes =weighted.mean(employment_yes, PERWT),
            employment_no = weighted.mean(employment_no,PERWT),
            employment_inactive = weighted.mean(employment_inactive,PERWT))

# Merge with calculated shares

benin.empstat.2013.spatial <- merge(benin.map, benin_empstat_2013, by = "GEOLEV2")
rm(benin_empstat, benin_empstat_2013)

Labour Force Participation

# 21. LABOUR FORCE PARTICIPATION SHARES

benin_labforce <- benin %>%
  filter(!LABFORCE == 9 & !LABFORCE==8) %>%
  mutate(labforce_no = ifelse(LABFORCE == 1, 1, 0),
         labforce_yes = ifelse(LABFORCE == 2, 1, 0))

benin_labforce_2013 <- benin_labforce %>%
  filter(YEAR == 2013) %>%
  group_by(GEOLEV2) %>%
  summarise(labforce_no =weighted.mean(labforce_no, PERWT),
            labforce_yes = weighted.mean(labforce_yes,PERWT))

# Merge with calculated shares

benin.labforce.2013.spatial <- merge(benin.map, benin_labforce_2013, by = "GEOLEV2")
rm(benin_labforce, benin_labforce_2013)

Occupation

# 22. OCCUPATION SHARES

benin_occisco <- benin %>%
  filter(!OCCISCO == 99 & !OCCISCO==98 & !OCCISCO == 97) %>%
  mutate(occupation_legislators = ifelse(OCCISCO == 1, 1, 0),
         occupation_professionals = ifelse(OCCISCO == 2, 1, 0),
         occupation_technicians = ifelse(OCCISCO == 3, 1, 0),
         occupation_clerks = ifelse(OCCISCO == 4, 1, 0),
         occupation_services = ifelse(OCCISCO == 5, 1, 0),
         occupation_agrifish = ifelse(OCCISCO == 6, 1, 0),
         occupation_crafts = ifelse(OCCISCO == 7, 1, 0),
         occupation_plant = ifelse(OCCISCO == 8, 1, 0),
         occupation_elementary = ifelse(OCCISCO == 9, 1, 0),
         occupation_army = ifelse(OCCISCO == 10, 1, 0),
         occupation_other = ifelse(OCCISCO == 11, 1, 0))

benin_occisco_2013 <- benin_occisco %>%
  filter(YEAR == 2013) %>%
  group_by(GEOLEV2) %>%
  summarise(occupation_legislators = weighted.mean(occupation_legislators, PERWT),
            occupation_professionals = weighted.mean(occupation_professionals, PERWT),
            occupation_technicians = weighted.mean(occupation_technicians, PERWT),
            occupation_clerks = weighted.mean(occupation_clerks, PERWT),
            occupation_services =weighted.mean(occupation_services, PERWT),
            occupation_agrifish =weighted.mean(occupation_agrifish, PERWT),
            occupation_crafts = weighted.mean(occupation_crafts, PERWT),
            occupation_plant = weighted.mean(occupation_plant, PERWT),
            occupation_elementary = weighted.mean(occupation_elementary, PERWT),
            occupation_army = weighted.mean(occupation_army, PERWT),
            occupation_other = weighted.mean(occupation_other, PERWT))

# Merge with calculated shares

benin.occisco.2013.spatial <- merge(benin.map, benin_occisco_2013, by = "GEOLEV2")
rm(benin_occisco, benin_occisco_2013)

Aggregated Sectors

Here I calculate sectoral shares with three target categories: primary, secondary and tertiary sector. Therefore, I group agriculture and mining into the primary sector, manufacturing, construction and "other industry" into the secondary sector and all other categories into the tertiary.

# 23. 3-WAY SECTORAL SHARES

benin_sectors <- benin %>%
  filter(!INDGEN == 0 & !INDGEN == 998 & !INDGEN== 999) %>%
  mutate(sector_primary = ifelse(INDGEN == 10 | INDGEN == 20, 1, 0),
         sector_secondary = ifelse(INDGEN == 30 | INDGEN == 50 | INDGEN == 130, 1, 0),
         sector_tertiary = ifelse(INDGEN == 40 | INDGEN == 60 | INDGEN == 70|
                                    INDGEN == 80 | INDGEN == 90 | INDGEN == 100|
                                    INDGEN == 110| INDGEN == 111| INDGEN == 112|
                                    INDGEN == 113| INDGEN == 114| INDGEN == 120, 1, 0))

benin_sectors_2013 <- benin_sectors %>%
  filter(YEAR == 2013) %>%
  group_by(GEOLEV2) %>%
  summarise(sector_primary   =   weighted.mean(sector_primary, PERWT),
            sector_secondary = weighted.mean(sector_secondary, PERWT),
            sector_tertiary  =  weighted.mean(sector_tertiary, PERWT))

# Merge with calculated shares

benin.sectors.2013.spatial <- merge(benin.map, benin_sectors_2013, by = "GEOLEV2")
rm(benin_sectors, benin_sectors_2013)

I then plot primary, secondary and tertiary shares by district:

ggplot() +
  geom_sf(data = benin.sectors.2013.spatial$geometry) +
  geom_sf(data = benin.sectors.2013.spatial, aes(fill=sector_primary)) +
  scale_fill_viridis_c(option = "viridis")  + theme_void()

ggplot() +
  geom_sf(data = benin.sectors.2013.spatial$geometry) +
  geom_sf(data = benin.sectors.2013.spatial, aes(fill=sector_secondary)) +
  scale_fill_viridis_c(option = "viridis")  + theme_void()

ggplot() +
  geom_sf(data = benin.sectors.2013.spatial$geometry) +
  geom_sf(data = benin.sectors.2013.spatial, aes(fill=sector_tertiary)) +
  scale_fill_viridis_c(option = "viridis")  + theme_void()

The uneven spatial distribution of sectors is apparent, with the secondary and tertiary sectors concentrating in the southern, coastal part of the country (where the capital Porto-Novo is located) and the rest of the country devoted mainly to agriculture (with the exception of the district of Parakou, the most important city in the central part of Benin).

Disaggregated Sectoral Shares

Here, sector are decomposed fully, with each category of the INDGEN variable accounting for one specific subsector.

# 24. SECTORAL SHARES

benin_indgen <- benin %>%
  filter(!INDGEN == 0 & !INDGEN == 998 & !INDGEN== 999) %>%
  mutate(sector_agriculture = ifelse(INDGEN == 10, 1, 0),
         sector_mining = ifelse(INDGEN == 20, 1, 0),
         sector_manufacturing = ifelse(INDGEN == 30, 1, 0),
         sector_utilities = ifelse(INDGEN == 40, 1, 0),
         sector_construction = ifelse(INDGEN == 50, 1, 0),
         sector_trade = ifelse(INDGEN == 60, 1, 0),
         sector_accommodation = ifelse(INDGEN == 70, 1, 0),
         sector_transport = ifelse(INDGEN == 80, 1, 0),
         sector_finance = ifelse(INDGEN == 90, 1, 0),
         sector_public = ifelse(INDGEN == 100, 1, 0),
         sector_unspec_services = ifelse(INDGEN == 110, 1, 0),
         sector_business = ifelse(INDGEN == 111, 1, 0),
         sector_education = ifelse(INDGEN == 112, 1, 0),
         sector_health = ifelse(INDGEN == 113, 1, 0),
         sector_other_services = ifelse(INDGEN == 114, 1, 0),
         sector_private_hh = ifelse(INDGEN == 120, 1, 0),
         sector_other_industry = ifelse(INDGEN == 130, 1, 0))


benin_indgen_2013 <- benin_indgen %>%
  filter(YEAR == 2013) %>%
  group_by(GEOLEV2) %>%
  summarise(sector_agriculture =weighted.mean(sector_agriculture, PERWT),
            sector_mining =weighted.mean(sector_mining, PERWT),
            sector_manufacturing =weighted.mean(sector_manufacturing, PERWT),
            sector_utilities =weighted.mean(sector_utilities, PERWT),
            sector_construction =weighted.mean(sector_construction, PERWT),
            sector_trade =weighted.mean(sector_trade, PERWT),
            sector_accommodation =weighted.mean(sector_accommodation, PERWT),
            sector_transport =weighted.mean(sector_transport, PERWT),
            sector_finance =weighted.mean(sector_finance, PERWT),
            sector_public = weighted.mean(sector_public, PERWT),
            sector_unspec_services = weighted.mean(sector_unspec_services, PERWT),
            sector_business = weighted.mean(sector_business, PERWT),
            sector_education = weighted.mean(sector_education, PERWT),
            sector_health = weighted.mean(sector_health, PERWT),
            sector_other_services = weighted.mean(sector_other_services, PERWT),
            sector_private_hh = weighted.mean(sector_private_hh, PERWT),
            sector_other_industry = weighted.mean(sector_other_industry, PERWT))

# Merge with calculated shares

benin.indgen.2013.spatial <- merge(benin.map, benin_indgen_2013, by = "GEOLEV2")
rm(benin_indgen, benin_indgen_2013)

Migration Shares

As in Hohmann (2018), I encounter the problem of the different harmonisation of the migration variable. Indeed, different samples (i.e. different countries) have categorised internal migration differently. There are three main versions of the migration variable: MIGRATE1, MIGRATE5 and MIGRATEP. The suffixes reflect how migration status is indexed for each individual in each wave. In particular, an individual is considered to be a migrant if they lived in a different administrative unit or a different country altogether at a particular point in the past: MIGRATE1 registers this information 1 year in the past, MIGRATE5 5 years in the past, and MIGRATEP indicates whether the individual lived in a different administrative unit "at some point in the past".

# 25. MIGRATION SHARES

benin_migratep <- benin %>%
  filter(!MIGRATEP == 0 & !MIGRATEP==99 & !MIGRATEP == 98) %>%
  mutate(migrant_no = ifelse(MIGRATEP == 10 | MIGRATEP == 11, 1, 0),
         migrant_yes = ifelse(MIGRATEP == 12 | MIGRATEP == 20 | MIGRATEP == 30, 1, 0))


benin_migratep_2013 <- benin_migratep %>%
  filter(YEAR == 2013) %>%
  group_by(GEOLEV2) %>%
  summarise(migrant_no =weighted.mean(migrant_no, PERWT),
            migrant_yes =weighted.mean(migrant_yes, PERWT))

# Merge with calculated shares

benin.migratep.2013.spatial <- merge(benin.map, benin_migratep_2013, by = "GEOLEV2")
rm(benin_migratep, benin_migratep_2013)

Variables disaggregated by gender

I then turn to the cross-products of the gender variable with employment status, occupation and disaggregated sectors. For each level of employment status, occupation and sector, I calculate the female and male share.

# 26. Gender EMPLOYMENT SHARES
# only select the variable "EMPSTAT" in order to aggregate its shares at adm2 level

benin_gender_empstat <- benin %>%
  filter(!EMPSTAT == 0 & !EMPSTAT==9) %>%
  mutate(female_employment_yes =      ifelse(EMPSTAT == 1 & SEX == 2, 1, 0),
         female_employment_no =       ifelse(EMPSTAT == 2 & SEX == 2, 1, 0),
         female_employment_inactive = ifelse(EMPSTAT == 3 & SEX == 2, 1, 0),
         male_employment_yes =      ifelse(EMPSTAT == 1 & SEX == 1, 1, 0),
         male_employment_no =       ifelse(EMPSTAT == 2 & SEX == 1, 1, 0),
         male_employment_inactive = ifelse(EMPSTAT == 3 & SEX == 1, 1, 0))

# summarise for each district 

benin_gender_empstat_2013 <- benin_gender_empstat %>%
  filter(YEAR == 2013) %>%
  group_by(GEOLEV2) %>%
  summarise(female_employment_yes =weighted.mean(female_employment_yes, PERWT),
            female_employment_no = weighted.mean(female_employment_no,PERWT),
            female_employment_inactive = weighted.mean(female_employment_inactive,PERWT),
            male_employment_yes =weighted.mean(male_employment_yes, PERWT),
            male_employment_no = weighted.mean(male_employment_no,PERWT),
            male_employment_inactive = weighted.mean(male_employment_inactive,PERWT))

# Merge with calculated shares

benin.gender.empstat.2013.spatial <- merge(benin.map, benin_gender_empstat_2013, by = "GEOLEV2")
rm(benin_gender_empstat, benin_gender_empstat_2013)


# 27. GENDER OCCUPATION SHARES

benin_gender_occisco <- benin %>%
  filter(!OCCISCO == 99 & !OCCISCO==98 & !OCCISCO == 97) %>%
  mutate(female_occupation_legislators    = ifelse(OCCISCO == 1 & SEX == 2, 1, 0),
         female_occupation_professionals  = ifelse(OCCISCO == 2 & SEX == 2, 1, 0),
         female_occupation_technicians    = ifelse(OCCISCO == 3 & SEX == 2, 1, 0),
         female_occupation_clerks         = ifelse(OCCISCO == 4 & SEX == 2, 1, 0),
         female_occupation_services       = ifelse(OCCISCO == 5 & SEX == 2, 1, 0),
         female_occupation_agrifish       = ifelse(OCCISCO == 6 & SEX == 2, 1, 0),
         female_occupation_crafts         = ifelse(OCCISCO == 7 & SEX == 2, 1, 0),
         female_occupation_plant          = ifelse(OCCISCO == 8 & SEX == 2, 1, 0),
         female_occupation_elementary     = ifelse(OCCISCO == 9 & SEX == 2, 1, 0),
         female_occupation_army          = ifelse(OCCISCO == 10 & SEX == 2, 1, 0),
         female_occupation_other         = ifelse(OCCISCO == 11 & SEX == 2, 1, 0),
         male_occupation_legislators    = ifelse(OCCISCO == 1 & SEX == 1, 1, 0),
         male_occupation_professionals  = ifelse(OCCISCO == 2 & SEX == 1, 1, 0),
         male_occupation_technicians    = ifelse(OCCISCO == 3 & SEX == 1, 1, 0),
         male_occupation_clerks         = ifelse(OCCISCO == 4 & SEX == 1, 1, 0),
         male_occupation_services       = ifelse(OCCISCO == 5 & SEX == 1, 1, 0),
         male_occupation_agrifish       = ifelse(OCCISCO == 6 & SEX == 1, 1, 0),
         male_occupation_crafts         = ifelse(OCCISCO == 7 & SEX == 1, 1, 0),
         male_occupation_plant          = ifelse(OCCISCO == 8 & SEX == 1, 1, 0),
         male_occupation_elementary     = ifelse(OCCISCO == 9 & SEX == 1, 1, 0),
         male_occupation_army          = ifelse(OCCISCO == 10 & SEX == 1, 1, 0),
         male_occupation_other         = ifelse(OCCISCO == 11 & SEX == 1, 1, 0))

# summarise for each district 

benin_gender_occisco_2013 <- benin_gender_occisco %>%
  filter(YEAR == 2013) %>%
  group_by(GEOLEV2) %>%
  summarise(female_occupation_legislators = weighted.mean(female_occupation_legislators, PERWT),
            female_occupation_professionals = weighted.mean(female_occupation_professionals, PERWT),
            female_occupation_technicians = weighted.mean(female_occupation_technicians, PERWT),
            female_occupation_clerks = weighted.mean(female_occupation_clerks, PERWT),
            female_occupation_services =weighted.mean(female_occupation_services, PERWT),
            female_occupation_agrifish =weighted.mean(female_occupation_agrifish, PERWT),
            female_occupation_crafts = weighted.mean(female_occupation_crafts, PERWT),
            female_occupation_plant = weighted.mean(female_occupation_plant, PERWT),
            female_occupation_elementary = weighted.mean(female_occupation_elementary, PERWT),
            female_occupation_army = weighted.mean(female_occupation_army, PERWT),
            female_occupation_other = weighted.mean(female_occupation_other, PERWT),
            male_occupation_legislators = weighted.mean(male_occupation_legislators, PERWT),
            male_occupation_professionals = weighted.mean(male_occupation_professionals, PERWT),
            male_occupation_technicians = weighted.mean(male_occupation_technicians, PERWT),
            male_occupation_clerks = weighted.mean(male_occupation_clerks, PERWT),
            male_occupation_services =weighted.mean(male_occupation_services, PERWT),
            male_occupation_agrifish =weighted.mean(male_occupation_agrifish, PERWT),
            male_occupation_crafts = weighted.mean(male_occupation_crafts, PERWT),
            male_occupation_plant = weighted.mean(male_occupation_plant, PERWT),
            male_occupation_elementary = weighted.mean(male_occupation_elementary, PERWT),
            male_occupation_army = weighted.mean(male_occupation_army, PERWT),
            male_occupation_other = weighted.mean(male_occupation_other, PERWT))

# Merge with calculated shares

benin.gender.occisco.2013.spatial <- merge(benin.map, benin_gender_occisco_2013, by = "GEOLEV2")
rm(benin_gender_occisco, benin_gender_occisco_2013)


# 28. GENDER SECTORAL SHARES

benin_gender_indgen <- benin %>%
  filter(!INDGEN == 0 & !INDGEN == 998 & !INDGEN== 999) %>%
  mutate(sector_female_agriculture      = ifelse(INDGEN == 10 & SEX == 2, 1, 0),
         sector_female_mining           = ifelse(INDGEN == 20 & SEX == 2, 1, 0),
         sector_female_manufacturing    = ifelse(INDGEN == 30 & SEX == 2, 1, 0),
         sector_female_utilities        = ifelse(INDGEN == 40 & SEX == 2, 1, 0),
         sector_female_construction     = ifelse(INDGEN == 50 & SEX == 2, 1, 0),
         sector_female_trade            = ifelse(INDGEN == 60 & SEX == 2, 1, 0),
         sector_female_accommodation    = ifelse(INDGEN == 70 & SEX == 2, 1, 0),
         sector_female_transport        = ifelse(INDGEN == 80 & SEX == 2, 1, 0),
         sector_female_finance          = ifelse(INDGEN == 90 & SEX == 2, 1, 0),
         sector_female_public          = ifelse(INDGEN == 100 & SEX == 2, 1, 0),
         sector_female_unspec_services = ifelse(INDGEN == 110 & SEX == 2, 1, 0),
         sector_female_business        = ifelse(INDGEN == 111 & SEX == 2, 1, 0),
         sector_female_education       = ifelse(INDGEN == 112 & SEX == 2, 1, 0),
         sector_female_health          = ifelse(INDGEN == 113 & SEX == 2, 1, 0),
         sector_female_other_services  = ifelse(INDGEN == 114 & SEX == 2, 1, 0),
         sector_female_private_hh      = ifelse(INDGEN == 120 & SEX == 2, 1, 0),
         sector_female_other_industry  = ifelse(INDGEN == 130 & SEX == 2, 1, 0),
         sector_male_agriculture      = ifelse(INDGEN == 10 & SEX == 1, 1, 0),
         sector_male_mining           = ifelse(INDGEN == 20 & SEX == 1, 1, 0),
         sector_male_manufacturing    = ifelse(INDGEN == 30 & SEX == 1, 1, 0),
         sector_male_utilities        = ifelse(INDGEN == 40 & SEX == 1, 1, 0),
         sector_male_construction     = ifelse(INDGEN == 50 & SEX == 1, 1, 0),
         sector_male_trade            = ifelse(INDGEN == 60 & SEX == 1, 1, 0),
         sector_male_accommodation    = ifelse(INDGEN == 70 & SEX == 1, 1, 0),
         sector_male_transport        = ifelse(INDGEN == 80 & SEX == 1, 1, 0),
         sector_male_finance          = ifelse(INDGEN == 90 & SEX == 1, 1, 0),
         sector_male_public          = ifelse(INDGEN == 100 & SEX == 1, 1, 0),
         sector_male_unspec_services = ifelse(INDGEN == 110 & SEX == 1, 1, 0),
         sector_male_business        = ifelse(INDGEN == 111 & SEX == 1, 1, 0),
         sector_male_education       = ifelse(INDGEN == 112 & SEX == 1, 1, 0),
         sector_male_health          = ifelse(INDGEN == 113 & SEX == 1, 1, 0),
         sector_male_other_services  = ifelse(INDGEN == 114 & SEX == 1, 1, 0),
         sector_male_private_hh      = ifelse(INDGEN == 120 & SEX == 1, 1, 0),
         sector_male_other_industry  = ifelse(INDGEN == 130 & SEX == 1, 1, 0))


# summarise for each district 

benin_gender_indgen_2013 <- benin_gender_indgen %>%
  filter(YEAR == 2013) %>%
  group_by(GEOLEV2) %>%
  summarise(sector_female_agriculture =weighted.mean(sector_female_agriculture, PERWT),
            sector_female_mining =weighted.mean(sector_female_mining, PERWT),
            sector_female_manufacturing =weighted.mean(sector_female_manufacturing, PERWT),
            sector_female_utilities =weighted.mean(sector_female_utilities, PERWT),
            sector_female_construction =weighted.mean(sector_female_construction, PERWT),
            sector_female_trade =weighted.mean(sector_female_trade, PERWT),
            sector_female_accommodation =weighted.mean(sector_female_accommodation, PERWT),
            sector_female_transport =weighted.mean(sector_female_transport, PERWT),
            sector_female_finance =weighted.mean(sector_female_finance, PERWT),
            sector_female_public = weighted.mean(sector_female_public, PERWT),
            sector_female_unspec_services = weighted.mean(sector_female_unspec_services, PERWT),
            sector_female_business = weighted.mean(sector_female_business, PERWT),
            sector_female_education = weighted.mean(sector_female_education, PERWT),
            sector_female_health = weighted.mean(sector_female_health, PERWT),
            sector_female_other_services = weighted.mean(sector_female_other_services, PERWT),
            sector_female_private_hh = weighted.mean(sector_female_private_hh, PERWT),
            sector_female_other_industry = weighted.mean(sector_female_other_industry, PERWT),
            sector_male_agriculture =weighted.mean(sector_male_agriculture, PERWT),
            sector_male_mining =weighted.mean(sector_male_mining, PERWT),
            sector_male_manufacturing =weighted.mean(sector_male_manufacturing, PERWT),
            sector_male_utilities =weighted.mean(sector_male_utilities, PERWT),
            sector_male_construction =weighted.mean(sector_male_construction, PERWT),
            sector_male_trade =weighted.mean(sector_male_trade, PERWT),
            sector_male_accommodation =weighted.mean(sector_male_accommodation, PERWT),
            sector_male_transport =weighted.mean(sector_male_transport, PERWT),
            sector_male_finance =weighted.mean(sector_male_finance, PERWT),
            sector_male_public = weighted.mean(sector_male_public, PERWT),
            sector_male_unspec_services = weighted.mean(sector_male_unspec_services, PERWT),
            sector_male_business = weighted.mean(sector_male_business, PERWT),
            sector_male_education = weighted.mean(sector_male_education, PERWT),
            sector_male_health = weighted.mean(sector_male_health, PERWT),
            sector_male_other_services = weighted.mean(sector_male_other_services, PERWT),
            sector_male_private_hh = weighted.mean(sector_male_private_hh, PERWT),
            sector_male_other_industry = weighted.mean(sector_male_other_industry, PERWT))


# Merge with calculated shares

benin.gender.indgen.2013.spatial <- merge(benin.map, benin_gender_indgen_2013, by = "GEOLEV2")
rm(benin_gender_indgen, benin_gender_indgen_2013)

Variables disaggregated by age

Similarly I calculate the cross-products of the age variable with employment status, occupation and disaggregated sectors. I divide the sample into three age groups: youth (15-24), adults (25-54) and seniors (55-64), covering the entire span of working age population. For each of these three levels, I calculate the respective employment, occupation and sectoral shares.

# 31. AGE-specific EMPLOYMENT SHARES
# only select the variable "EMPSTAT" in order to aggregate its shares at adm2 level

benin_age_empstat <- benin %>%
  filter(!EMPSTAT == 0 & !EMPSTAT==9) %>%
  mutate(youth_employment_yes       = ifelse(EMPSTAT == 1 & AGE < 25 & AGE > 14, 1, 0),
         youth_employment_no        = ifelse(EMPSTAT == 2 & AGE < 25 & AGE > 14, 1, 0),
         youth_employment_inactive  = ifelse(EMPSTAT == 3 & AGE < 25 & AGE > 14, 1, 0),
         adult_employment_yes       = ifelse(EMPSTAT == 1 & AGE < 55 & AGE > 24, 1, 0),
         adult_employment_no        = ifelse(EMPSTAT == 2 & AGE < 55 & AGE > 24, 1, 0),
         adult_employment_inactive  = ifelse(EMPSTAT == 3 & AGE < 55 & AGE > 24, 1, 0),
         senior_employment_yes      = ifelse(EMPSTAT == 1 & AGE > 54 & AGE < 65, 1, 0),
         senior_employment_no       = ifelse(EMPSTAT == 2 & AGE > 54 & AGE < 65, 1, 0),
         senior_employment_inactive = ifelse(EMPSTAT == 3 & AGE > 54 & AGE < 65, 1, 0))

# summarise for each district 

benin_age_empstat_2013 <- benin_age_empstat %>%
  filter(YEAR == 2013) %>%
  group_by(GEOLEV2) %>%
  summarise(youth_employment_yes =weighted.mean(youth_employment_yes, PERWT),
            youth_employment_no = weighted.mean(youth_employment_no,PERWT),
            youth_employment_inactive = weighted.mean(youth_employment_inactive,PERWT),
            adult_employment_yes =weighted.mean(adult_employment_yes, PERWT),
            adult_employment_no = weighted.mean(adult_employment_no,PERWT),
            adult_employment_inactive = weighted.mean(adult_employment_inactive,PERWT),
            senior_employment_yes =weighted.mean(senior_employment_yes, PERWT),
            senior_employment_no = weighted.mean(senior_employment_no,PERWT),
            senior_employment_inactive = weighted.mean(senior_employment_inactive,PERWT))


# Merge with calculated shares

benin.age.empstat.2013.spatial <- merge(benin.map, benin_age_empstat_2013, by = "GEOLEV2")
rm(benin_age_empstat, benin_age_empstat_2013)


# 34. AGE-specific OCCUPATION SHARES
# only select the variable "OCCISCO" in order to aggregate its shares at adm2 level

benin_age_occisco <- benin %>%
  filter(!OCCISCO == 99 & !OCCISCO==98 & !OCCISCO == 97) %>%
  mutate(youth_occupation_legislators   = ifelse(OCCISCO == 1 & AGE < 25 & AGE > 14, 1, 0),
         youth_occupation_professionals = ifelse(OCCISCO == 2 & AGE < 25 & AGE > 14, 1, 0),
         youth_occupation_technicians   = ifelse(OCCISCO == 3 & AGE < 25 & AGE > 14, 1, 0),
         youth_occupation_clerks        = ifelse(OCCISCO == 4 & AGE < 25 & AGE > 14, 1, 0),
         youth_occupation_services      = ifelse(OCCISCO == 5 & AGE < 25 & AGE > 14, 1, 0),
         youth_occupation_agrifish      = ifelse(OCCISCO == 6 & AGE < 25 & AGE > 14, 1, 0),
         youth_occupation_crafts        = ifelse(OCCISCO == 7 & AGE < 25 & AGE > 14, 1, 0),
         youth_occupation_plant         = ifelse(OCCISCO == 8 & AGE < 25 & AGE > 14, 1, 0),
         youth_occupation_elementary    = ifelse(OCCISCO == 9 & AGE < 25 & AGE > 14, 1, 0),
         youth_occupation_army         = ifelse(OCCISCO == 10 & AGE < 25 & AGE > 14, 1, 0),
         youth_occupation_other        = ifelse(OCCISCO == 11 & AGE < 25 & AGE > 14, 1, 0),
         adult_occupation_legislators   = ifelse(OCCISCO == 1 & AGE < 55 & AGE > 24, 1, 0),
         adult_occupation_professionals = ifelse(OCCISCO == 2 & AGE < 55 & AGE > 24, 1, 0),
         adult_occupation_technicians   = ifelse(OCCISCO == 3 & AGE < 55 & AGE > 24, 1, 0),
         adult_occupation_clerks        = ifelse(OCCISCO == 4 & AGE < 55 & AGE > 24, 1, 0),
         adult_occupation_services      = ifelse(OCCISCO == 5 & AGE < 55 & AGE > 24, 1, 0),
         adult_occupation_agrifish      = ifelse(OCCISCO == 6 & AGE < 55 & AGE > 24, 1, 0),
         adult_occupation_crafts        = ifelse(OCCISCO == 7 & AGE < 55 & AGE > 24, 1, 0),
         adult_occupation_plant         = ifelse(OCCISCO == 8 & AGE < 55 & AGE > 24, 1, 0),
         adult_occupation_elementary    = ifelse(OCCISCO == 9 & AGE < 55 & AGE > 24, 1, 0),
         adult_occupation_army         = ifelse(OCCISCO == 10 & AGE < 55 & AGE > 24, 1, 0),
         adult_occupation_other        = ifelse(OCCISCO == 11 & AGE < 55 & AGE > 24, 1, 0),
         senior_occupation_legislators   = ifelse(OCCISCO == 1 & AGE > 54 & AGE < 65, 1, 0),
         senior_occupation_professionals = ifelse(OCCISCO == 2 & AGE > 54 & AGE < 65, 1, 0),
         senior_occupation_technicians   = ifelse(OCCISCO == 3 & AGE > 54 & AGE < 65, 1, 0),
         senior_occupation_clerks        = ifelse(OCCISCO == 4 & AGE > 54 & AGE < 65, 1, 0),
         senior_occupation_services      = ifelse(OCCISCO == 5 & AGE > 54 & AGE < 65, 1, 0),
         senior_occupation_agrifish      = ifelse(OCCISCO == 6 & AGE > 54 & AGE < 65, 1, 0),
         senior_occupation_crafts        = ifelse(OCCISCO == 7 & AGE > 54 & AGE < 65, 1, 0),
         senior_occupation_plant         = ifelse(OCCISCO == 8 & AGE > 54 & AGE < 65, 1, 0),
         senior_occupation_elementary    = ifelse(OCCISCO == 9 & AGE > 54 & AGE < 65, 1, 0),
         senior_occupation_army         = ifelse(OCCISCO == 10 & AGE > 54 & AGE < 65, 1, 0),
         senior_occupation_other        = ifelse(OCCISCO == 11 & AGE > 54 & AGE < 65, 1, 0))

# summarise for each district 

benin_age_occisco_2013 <- benin_age_occisco %>%
  filter(YEAR == 2013) %>%
  group_by(GEOLEV2) %>%
  summarise(youth_occupation_legislators = weighted.mean(youth_occupation_legislators, PERWT),
            youth_occupation_professionals = weighted.mean(youth_occupation_professionals, PERWT),
            youth_occupation_technicians = weighted.mean(youth_occupation_technicians, PERWT),
            youth_occupation_clerks = weighted.mean(youth_occupation_clerks, PERWT),
            youth_occupation_services =weighted.mean(youth_occupation_services, PERWT),
            youth_occupation_agrifish =weighted.mean(youth_occupation_agrifish, PERWT),
            youth_occupation_crafts = weighted.mean(youth_occupation_crafts, PERWT),
            youth_occupation_plant = weighted.mean(youth_occupation_plant, PERWT),
            youth_occupation_elementary = weighted.mean(youth_occupation_elementary, PERWT),
            youth_occupation_army = weighted.mean(youth_occupation_army, PERWT),
            youth_occupation_other = weighted.mean(youth_occupation_other, PERWT),
            adult_occupation_legislators = weighted.mean(adult_occupation_legislators, PERWT),
            adult_occupation_professionals = weighted.mean(adult_occupation_professionals, PERWT),
            adult_occupation_technicians = weighted.mean(adult_occupation_technicians, PERWT),
            adult_occupation_clerks = weighted.mean(adult_occupation_clerks, PERWT),
            adult_occupation_services =weighted.mean(adult_occupation_services, PERWT),
            adult_occupation_agrifish =weighted.mean(adult_occupation_agrifish, PERWT),
            adult_occupation_crafts = weighted.mean(adult_occupation_crafts, PERWT),
            adult_occupation_plant = weighted.mean(adult_occupation_plant, PERWT),
            adult_occupation_elementary = weighted.mean(adult_occupation_elementary, PERWT),
            adult_occupation_army = weighted.mean(adult_occupation_army, PERWT),
            adult_occupation_other = weighted.mean(adult_occupation_other, PERWT),
            senior_occupation_legislators = weighted.mean(senior_occupation_legislators, PERWT),
            senior_occupation_professionals = weighted.mean(senior_occupation_professionals, PERWT),
            senior_occupation_technicians = weighted.mean(senior_occupation_technicians, PERWT),
            senior_occupation_clerks = weighted.mean(senior_occupation_clerks, PERWT),
            senior_occupation_services =weighted.mean(senior_occupation_services, PERWT),
            senior_occupation_agrifish =weighted.mean(senior_occupation_agrifish, PERWT),
            senior_occupation_crafts = weighted.mean(senior_occupation_crafts, PERWT),
            senior_occupation_plant = weighted.mean(senior_occupation_plant, PERWT),
            senior_occupation_elementary = weighted.mean(senior_occupation_elementary, PERWT),
            senior_occupation_army = weighted.mean(senior_occupation_army, PERWT),
            senior_occupation_other = weighted.mean(senior_occupation_other, PERWT))

# Merge with calculated shares

benin.age.occisco.2013.spatial <- merge(benin.map, benin_age_occisco_2013, by = "GEOLEV2")
rm(benin_age_occisco, benin_age_occisco_2013)


# 37. AGE-specific SECTORAL SHARES

benin_age_indgen <- benin %>%
  filter(!INDGEN == 0 & !INDGEN == 998 & !INDGEN== 999) %>%
  mutate(sector_youth_agriculture      = ifelse(INDGEN == 10 & AGE > 14 & AGE < 25, 1, 0),
         sector_youth_mining           = ifelse(INDGEN == 20 & AGE > 14 & AGE < 25, 1, 0),
         sector_youth_manufacturing    = ifelse(INDGEN == 30 & AGE > 14 & AGE < 25, 1, 0),
         sector_youth_utilities        = ifelse(INDGEN == 40 & AGE > 14 & AGE < 25, 1, 0),
         sector_youth_construction     = ifelse(INDGEN == 50 & AGE > 14 & AGE < 25, 1, 0),
         sector_youth_trade            = ifelse(INDGEN == 60 & AGE > 14 & AGE < 25, 1, 0),
         sector_youth_accommodation    = ifelse(INDGEN == 70 & AGE > 14 & AGE < 25, 1, 0),
         sector_youth_transport        = ifelse(INDGEN == 80 & AGE > 14 & AGE < 25, 1, 0),
         sector_youth_finance          = ifelse(INDGEN == 90 & AGE > 14 & AGE < 25, 1, 0),
         sector_youth_public          = ifelse(INDGEN == 100 & AGE > 14 & AGE < 25, 1, 0),
         sector_youth_unspec_services = ifelse(INDGEN == 110 & AGE > 14 & AGE < 25, 1, 0),
         sector_youth_business        = ifelse(INDGEN == 111 & AGE > 14 & AGE < 25, 1, 0),
         sector_youth_education       = ifelse(INDGEN == 112 & AGE > 14 & AGE < 25, 1, 0),
         sector_youth_health          = ifelse(INDGEN == 113 & AGE > 14 & AGE < 25, 1, 0),
         sector_youth_other_services  = ifelse(INDGEN == 114 & AGE > 14 & AGE < 25, 1, 0),
         sector_youth_private_hh      = ifelse(INDGEN == 120 & AGE > 14 & AGE < 25, 1, 0),
         sector_youth_other_industry  = ifelse(INDGEN == 130 & AGE > 14 & AGE < 25, 1, 0),
         sector_adult_agriculture      = ifelse(INDGEN == 10 & AGE > 24 & AGE < 55, 1, 0),
         sector_adult_mining           = ifelse(INDGEN == 20 & AGE > 24 & AGE < 55, 1, 0),
         sector_adult_manufacturing    = ifelse(INDGEN == 30 & AGE > 24 & AGE < 55, 1, 0),
         sector_adult_utilities        = ifelse(INDGEN == 40 & AGE > 24 & AGE < 55, 1, 0),
         sector_adult_construction     = ifelse(INDGEN == 50 & AGE > 24 & AGE < 55, 1, 0),
         sector_adult_trade            = ifelse(INDGEN == 60 & AGE > 24 & AGE < 55, 1, 0),
         sector_adult_accommodation    = ifelse(INDGEN == 70 & AGE > 24 & AGE < 55, 1, 0),
         sector_adult_transport        = ifelse(INDGEN == 80 & AGE > 24 & AGE < 55, 1, 0),
         sector_adult_finance          = ifelse(INDGEN == 90 & AGE > 24 & AGE < 55, 1, 0),
         sector_adult_public          = ifelse(INDGEN == 100 & AGE > 24 & AGE < 55, 1, 0),
         sector_adult_unspec_services = ifelse(INDGEN == 110 & AGE > 24 & AGE < 55, 1, 0),
         sector_adult_business        = ifelse(INDGEN == 111 & AGE > 24 & AGE < 55, 1, 0),
         sector_adult_education       = ifelse(INDGEN == 112 & AGE > 24 & AGE < 55, 1, 0),
         sector_adult_health          = ifelse(INDGEN == 113 & AGE > 24 & AGE < 55, 1, 0),
         sector_adult_other_services  = ifelse(INDGEN == 114 & AGE > 24 & AGE < 55, 1, 0),
         sector_adult_private_hh      = ifelse(INDGEN == 120 & AGE > 24 & AGE < 55, 1, 0),
         sector_adult_other_industry  = ifelse(INDGEN == 130 & AGE > 24 & AGE < 55, 1, 0),
         sector_senior_agriculture      = ifelse(INDGEN == 10 & AGE > 54 & AGE < 65, 1, 0),
         sector_senior_mining           = ifelse(INDGEN == 20 & AGE > 54 & AGE < 65, 1, 0),
         sector_senior_manufacturing    = ifelse(INDGEN == 30 & AGE > 54 & AGE < 65, 1, 0),
         sector_senior_utilities        = ifelse(INDGEN == 40 & AGE > 54 & AGE < 65, 1, 0),
         sector_senior_construction     = ifelse(INDGEN == 50 & AGE > 54 & AGE < 65, 1, 0),
         sector_senior_trade            = ifelse(INDGEN == 60 & AGE > 54 & AGE < 65, 1, 0),
         sector_senior_accommodation    = ifelse(INDGEN == 70 & AGE > 54 & AGE < 65, 1, 0),
         sector_senior_transport        = ifelse(INDGEN == 80 & AGE > 54 & AGE < 65, 1, 0),
         sector_senior_finance          = ifelse(INDGEN == 90 & AGE > 54 & AGE < 65, 1, 0),
         sector_senior_public          = ifelse(INDGEN == 100 & AGE > 54 & AGE < 65, 1, 0),
         sector_senior_unspec_services = ifelse(INDGEN == 110 & AGE > 54 & AGE < 65, 1, 0),
         sector_senior_business        = ifelse(INDGEN == 111 & AGE > 54 & AGE < 65, 1, 0),
         sector_senior_education       = ifelse(INDGEN == 112 & AGE > 54 & AGE < 65, 1, 0),
         sector_senior_health          = ifelse(INDGEN == 113 & AGE > 54 & AGE < 65, 1, 0),
         sector_senior_other_services  = ifelse(INDGEN == 114 & AGE > 54 & AGE < 65, 1, 0),
         sector_senior_private_hh      = ifelse(INDGEN == 120 & AGE > 54 & AGE < 65, 1, 0),
         sector_senior_other_industry  = ifelse(INDGEN == 130 & AGE > 54 & AGE < 65, 1, 0))

# summarise for each district 

benin_age_indgen_2013 <- benin_age_indgen %>%
  filter(YEAR == 2013) %>%
  group_by(GEOLEV2) %>%
  summarise(sector_youth_agriculture =weighted.mean(sector_youth_agriculture, PERWT),
            sector_youth_mining =weighted.mean(sector_youth_mining, PERWT),
            sector_youth_manufacturing =weighted.mean(sector_youth_manufacturing, PERWT),
            sector_youth_utilities =weighted.mean(sector_youth_utilities, PERWT),
            sector_youth_construction =weighted.mean(sector_youth_construction, PERWT),
            sector_youth_trade =weighted.mean(sector_youth_trade, PERWT),
            sector_youth_accommodation =weighted.mean(sector_youth_accommodation, PERWT),
            sector_youth_transport =weighted.mean(sector_youth_transport, PERWT),
            sector_youth_finance =weighted.mean(sector_youth_finance, PERWT),
            sector_youth_public = weighted.mean(sector_youth_public, PERWT),
            sector_youth_unspec_services = weighted.mean(sector_youth_unspec_services, PERWT),
            sector_youth_business = weighted.mean(sector_youth_business, PERWT),
            sector_youth_education = weighted.mean(sector_youth_education, PERWT),
            sector_youth_health = weighted.mean(sector_youth_health, PERWT),
            sector_youth_other_services = weighted.mean(sector_youth_other_services, PERWT),
            sector_youth_private_hh = weighted.mean(sector_youth_private_hh, PERWT),
            sector_youth_other_industry = weighted.mean(sector_youth_other_industry, PERWT),
            sector_adult_agriculture =weighted.mean(sector_adult_agriculture, PERWT),
            sector_adult_mining =weighted.mean(sector_adult_mining, PERWT),
            sector_adult_manufacturing =weighted.mean(sector_adult_manufacturing, PERWT),
            sector_adult_utilities =weighted.mean(sector_adult_utilities, PERWT),
            sector_adult_construction =weighted.mean(sector_adult_construction, PERWT),
            sector_adult_trade =weighted.mean(sector_adult_trade, PERWT),
            sector_adult_accommodation =weighted.mean(sector_adult_accommodation, PERWT),
            sector_adult_transport =weighted.mean(sector_adult_transport, PERWT),
            sector_adult_finance =weighted.mean(sector_adult_finance, PERWT),
            sector_adult_public = weighted.mean(sector_adult_public, PERWT),
            sector_adult_unspec_services = weighted.mean(sector_adult_unspec_services, PERWT),
            sector_adult_business = weighted.mean(sector_adult_business, PERWT),
            sector_adult_education = weighted.mean(sector_adult_education, PERWT),
            sector_adult_health = weighted.mean(sector_adult_health, PERWT),
            sector_adult_other_services = weighted.mean(sector_adult_other_services, PERWT),
            sector_adult_private_hh = weighted.mean(sector_adult_private_hh, PERWT),
            sector_adult_other_industry = weighted.mean(sector_adult_other_industry, PERWT),
            sector_senior_agriculture =weighted.mean(sector_senior_agriculture, PERWT),
            sector_senior_mining =weighted.mean(sector_senior_mining, PERWT),
            sector_senior_manufacturing =weighted.mean(sector_senior_manufacturing, PERWT),
            sector_senior_utilities =weighted.mean(sector_senior_utilities, PERWT),
            sector_senior_construction =weighted.mean(sector_senior_construction, PERWT),
            sector_senior_trade =weighted.mean(sector_senior_trade, PERWT),
            sector_senior_accommodation =weighted.mean(sector_senior_accommodation, PERWT),
            sector_senior_transport =weighted.mean(sector_senior_transport, PERWT),
            sector_senior_finance =weighted.mean(sector_senior_finance, PERWT),
            sector_senior_public = weighted.mean(sector_senior_public, PERWT),
            sector_senior_unspec_services = weighted.mean(sector_senior_unspec_services, PERWT),
            sector_senior_business = weighted.mean(sector_senior_business, PERWT),
            sector_senior_education = weighted.mean(sector_senior_education, PERWT),
            sector_senior_health = weighted.mean(sector_senior_health, PERWT),
            sector_senior_other_services = weighted.mean(sector_senior_other_services, PERWT),
            sector_senior_private_hh = weighted.mean(sector_senior_private_hh, PERWT),
            sector_senior_other_industry = weighted.mean(sector_senior_other_industry, PERWT))

# Merge with calculated shares

benin.age.indgen.2013.spatial <- merge(benin.map, benin_age_indgen_2013, by = "GEOLEV2")
rm(benin_age_indgen, benin_age_indgen_2013)

Variables disaggregated by migration status

I repeat the same exercise by interacting employment, occupation and sector with two categories of migration: migrant and native (i.e. non-migrant according to the definition of migration given above).

# 40. MIGRANT EMPLOYMENT SHARES
# only select the variable "EMPSTAT" in order to aggregate its shares at adm2 level

benin_migrant_empstat <- benin %>%
  filter(!EMPSTAT == 0 & !EMPSTAT==9) %>%
  mutate(migrant_employment_yes      = ifelse(EMPSTAT == 1 & MIGRATEP == 12 | MIGRATEP == 20 | MIGRATEP == 30, 1, 0),
         migrant_employment_no       = ifelse(EMPSTAT == 2 & MIGRATEP == 12 | MIGRATEP == 20 | MIGRATEP == 30, 1, 0),
         migrant_employment_inactive = ifelse(EMPSTAT == 3 & MIGRATEP == 12 | MIGRATEP == 20 | MIGRATEP == 30, 1, 0),
         native_employment_yes       = ifelse(EMPSTAT == 1 & MIGRATEP == 10 | MIGRATEP == 11, 1, 0),
         native_employment_no        = ifelse(EMPSTAT == 2 & MIGRATEP == 10 | MIGRATEP == 11, 1, 0),
         native_employment_inactive  = ifelse(EMPSTAT == 3 & MIGRATEP == 10 | MIGRATEP == 11, 1, 0))

# summarise for each district 

benin_migrant_empstat_2013 <- benin_migrant_empstat %>%
  filter(YEAR == 2013) %>%
  group_by(GEOLEV2) %>%
  summarise(migrant_employment_yes =weighted.mean(migrant_employment_yes, PERWT),
            migrant_employment_no = weighted.mean(migrant_employment_no,PERWT),
            migrant_employment_inactive = weighted.mean(migrant_employment_inactive,PERWT),
            native_employment_yes =weighted.mean(native_employment_yes, PERWT),
            native_employment_no = weighted.mean(native_employment_no,PERWT),
            native_employment_inactive = weighted.mean(native_employment_inactive,PERWT))

# Merge with calculated shares

benin.migrant.empstat.2013.spatial <- merge(benin.map, benin_migrant_empstat_2013, by = "GEOLEV2")
rm(benin_migrant_empstat, benin_migrant_empstat_2013)


# 42. MIGRANT OCCUPATION SHARES

benin_migrant_occisco <- benin %>%
  filter(!OCCISCO == 99 & !OCCISCO==98 & !OCCISCO == 97) %>%
  mutate(migrant_occupation_legislators   = ifelse(OCCISCO == 1  & MIGRATEP == 12 | MIGRATEP == 20 | MIGRATEP == 30, 1, 0),
         migrant_occupation_professionals = ifelse(OCCISCO == 2  & MIGRATEP == 12 | MIGRATEP == 20 | MIGRATEP == 30, 1, 0),
         migrant_occupation_technicians   = ifelse(OCCISCO == 3  & MIGRATEP == 12 | MIGRATEP == 20 | MIGRATEP == 30, 1, 0),
         migrant_occupation_clerks        = ifelse(OCCISCO == 4  & MIGRATEP == 12 | MIGRATEP == 20 | MIGRATEP == 30, 1, 0),
         migrant_occupation_services      = ifelse(OCCISCO == 5  & MIGRATEP == 12 | MIGRATEP == 20 | MIGRATEP == 30, 1, 0),
         migrant_occupation_agrifish      = ifelse(OCCISCO == 6  & MIGRATEP == 12 | MIGRATEP == 20 | MIGRATEP == 30, 1, 0),
         migrant_occupation_crafts        = ifelse(OCCISCO == 7  & MIGRATEP == 12 | MIGRATEP == 20 | MIGRATEP == 30, 1, 0),
         migrant_occupation_plant         = ifelse(OCCISCO == 8  & MIGRATEP == 12 | MIGRATEP == 20 | MIGRATEP == 30, 1, 0),
         migrant_occupation_elementary    = ifelse(OCCISCO == 9  & MIGRATEP == 12 | MIGRATEP == 20 | MIGRATEP == 30, 1, 0),
         migrant_occupation_army         = ifelse(OCCISCO == 10  & MIGRATEP == 12 | MIGRATEP == 20 | MIGRATEP == 30, 1, 0),
         migrant_occupation_other        = ifelse(OCCISCO == 11  & MIGRATEP == 12 | MIGRATEP == 20 | MIGRATEP == 30, 1, 0),
         native_occupation_legislators   = ifelse(OCCISCO == 1 & MIGRATEP == 10 | MIGRATEP == 11, 1, 0),
         native_occupation_professionals = ifelse(OCCISCO == 2 & MIGRATEP == 10 | MIGRATEP == 11, 1, 0),
         native_occupation_technicians   = ifelse(OCCISCO == 3 & MIGRATEP == 10 | MIGRATEP == 11, 1, 0),
         native_occupation_clerks        = ifelse(OCCISCO == 4 & MIGRATEP == 10 | MIGRATEP == 11, 1, 0),
         native_occupation_services      = ifelse(OCCISCO == 5 & MIGRATEP == 10 | MIGRATEP == 11, 1, 0),
         native_occupation_agrifish      = ifelse(OCCISCO == 6 & MIGRATEP == 10 | MIGRATEP == 11, 1, 0),
         native_occupation_crafts        = ifelse(OCCISCO == 7 & MIGRATEP == 10 | MIGRATEP == 11, 1, 0),
         native_occupation_plant         = ifelse(OCCISCO == 8 & MIGRATEP == 10 | MIGRATEP == 11, 1, 0),
         native_occupation_elementary    = ifelse(OCCISCO == 9 & MIGRATEP == 10 | MIGRATEP == 11, 1, 0),
         native_occupation_army         = ifelse(OCCISCO == 10 & MIGRATEP == 10 | MIGRATEP == 11, 1, 0),
         native_occupation_other        = ifelse(OCCISCO == 11 & MIGRATEP == 10 | MIGRATEP == 11, 1, 0))


# summarise for each district 

benin_migrant_occisco_2013 <- benin_migrant_occisco %>%
  filter(YEAR == 2013) %>%
  group_by(GEOLEV2) %>%
  summarise(migrant_occupation_legislators = weighted.mean(migrant_occupation_legislators, PERWT),
            migrant_occupation_professionals = weighted.mean(migrant_occupation_professionals, PERWT),
            migrant_occupation_technicians = weighted.mean(migrant_occupation_technicians, PERWT),
            migrant_occupation_clerks = weighted.mean(migrant_occupation_clerks, PERWT),
            migrant_occupation_services =weighted.mean(migrant_occupation_services, PERWT),
            migrant_occupation_agrifish =weighted.mean(migrant_occupation_agrifish, PERWT),
            migrant_occupation_crafts = weighted.mean(migrant_occupation_crafts, PERWT),
            migrant_occupation_plant = weighted.mean(migrant_occupation_plant, PERWT),
            migrant_occupation_elementary = weighted.mean(migrant_occupation_elementary, PERWT),
            migrant_occupation_army = weighted.mean(migrant_occupation_army, PERWT),
            migrant_occupation_other = weighted.mean(migrant_occupation_other, PERWT),
            native_occupation_legislators = weighted.mean(native_occupation_legislators, PERWT),
            native_occupation_professionals = weighted.mean(native_occupation_professionals, PERWT),
            native_occupation_technicians = weighted.mean(native_occupation_technicians, PERWT),
            native_occupation_clerks = weighted.mean(native_occupation_clerks, PERWT),
            native_occupation_services =weighted.mean(native_occupation_services, PERWT),
            native_occupation_agrifish =weighted.mean(native_occupation_agrifish, PERWT),
            native_occupation_crafts = weighted.mean(native_occupation_crafts, PERWT),
            native_occupation_plant = weighted.mean(native_occupation_plant, PERWT),
            native_occupation_elementary = weighted.mean(native_occupation_elementary, PERWT),
            native_occupation_army = weighted.mean(native_occupation_army, PERWT),
            native_occupation_other = weighted.mean(native_occupation_other, PERWT))

# Merge with calculated shares

benin.migrant.occisco.2013.spatial <- merge(benin.map, benin_migrant_occisco_2013, by = "GEOLEV2")
rm(benin_migrant_occisco, benin_migrant_occisco_2013)


# 44. MIGRANT SECTORAL SHARES

benin_migrant_indgen <- benin %>%
  filter(!INDGEN == 0 & !INDGEN == 998 & !INDGEN== 999) %>%
  mutate(sector_migrant_agriculture      = ifelse(INDGEN == 10 & MIGRATEP == 12 | MIGRATEP == 20 | MIGRATEP == 30, 1, 0),
         sector_migrant_mining           = ifelse(INDGEN == 20 & MIGRATEP == 12 | MIGRATEP == 20 | MIGRATEP == 30, 1, 0),
         sector_migrant_manufacturing    = ifelse(INDGEN == 30 & MIGRATEP == 12 | MIGRATEP == 20 | MIGRATEP == 30, 1, 0),
         sector_migrant_utilities        = ifelse(INDGEN == 40 & MIGRATEP == 12 | MIGRATEP == 20 | MIGRATEP == 30, 1, 0),
         sector_migrant_construction     = ifelse(INDGEN == 50 & MIGRATEP == 12 | MIGRATEP == 20 | MIGRATEP == 30, 1, 0),
         sector_migrant_trade            = ifelse(INDGEN == 60 & MIGRATEP == 12 | MIGRATEP == 20 | MIGRATEP == 30, 1, 0),
         sector_migrant_accommodation    = ifelse(INDGEN == 70 & MIGRATEP == 12 | MIGRATEP == 20 | MIGRATEP == 30, 1, 0),
         sector_migrant_transport        = ifelse(INDGEN == 80 & MIGRATEP == 12 | MIGRATEP == 20 | MIGRATEP == 30, 1, 0),
         sector_migrant_finance          = ifelse(INDGEN == 90 & MIGRATEP == 12 | MIGRATEP == 20 | MIGRATEP == 30, 1, 0),
         sector_migrant_public          = ifelse(INDGEN == 100 & MIGRATEP == 12 | MIGRATEP == 20 | MIGRATEP == 30, 1, 0),
         sector_migrant_unspec_services = ifelse(INDGEN == 110 & MIGRATEP == 12 | MIGRATEP == 20 | MIGRATEP == 30, 1, 0),
         sector_migrant_business        = ifelse(INDGEN == 111 & MIGRATEP == 12 | MIGRATEP == 20 | MIGRATEP == 30, 1, 0),
         sector_migrant_education       = ifelse(INDGEN == 112 & MIGRATEP == 12 | MIGRATEP == 20 | MIGRATEP == 30, 1, 0),
         sector_migrant_health          = ifelse(INDGEN == 113 & MIGRATEP == 12 | MIGRATEP == 20 | MIGRATEP == 30, 1, 0),
         sector_migrant_other_services  = ifelse(INDGEN == 114 & MIGRATEP == 12 | MIGRATEP == 20 | MIGRATEP == 30, 1, 0),
         sector_migrant_private_hh      = ifelse(INDGEN == 120 & MIGRATEP == 12 | MIGRATEP == 20 | MIGRATEP == 30, 1, 0),
         sector_migrant_other_industry  = ifelse(INDGEN == 130 & MIGRATEP == 12 | MIGRATEP == 20 | MIGRATEP == 30, 1, 0),
         sector_native_agriculture      = ifelse(INDGEN == 10 & MIGRATEP == 10 | MIGRATEP == 11, 1, 0),
         sector_native_mining           = ifelse(INDGEN == 20 & MIGRATEP == 10 | MIGRATEP == 11, 1, 0),
         sector_native_manufacturing    = ifelse(INDGEN == 30 & MIGRATEP == 10 | MIGRATEP == 11, 1, 0),
         sector_native_utilities        = ifelse(INDGEN == 40 & MIGRATEP == 10 | MIGRATEP == 11, 1, 0),
         sector_native_construction     = ifelse(INDGEN == 50 & MIGRATEP == 10 | MIGRATEP == 11, 1, 0),
         sector_native_trade            = ifelse(INDGEN == 60 & MIGRATEP == 10 | MIGRATEP == 11, 1, 0),
         sector_native_accommodation    = ifelse(INDGEN == 70 & MIGRATEP == 10 | MIGRATEP == 11, 1, 0),
         sector_native_transport        = ifelse(INDGEN == 80 & MIGRATEP == 10 | MIGRATEP == 11, 1, 0),
         sector_native_finance          = ifelse(INDGEN == 90 & MIGRATEP == 10 | MIGRATEP == 11, 1, 0),
         sector_native_public          = ifelse(INDGEN == 100 & MIGRATEP == 10 | MIGRATEP == 11, 1, 0),
         sector_native_unspec_services = ifelse(INDGEN == 110 & MIGRATEP == 10 | MIGRATEP == 11, 1, 0),
         sector_native_business        = ifelse(INDGEN == 111 & MIGRATEP == 10 | MIGRATEP == 11, 1, 0),
         sector_native_education       = ifelse(INDGEN == 112 & MIGRATEP == 10 | MIGRATEP == 11, 1, 0),
         sector_native_health          = ifelse(INDGEN == 113 & MIGRATEP == 10 | MIGRATEP == 11, 1, 0),
         sector_native_other_services  = ifelse(INDGEN == 114 & MIGRATEP == 10 | MIGRATEP == 11, 1, 0),
         sector_native_private_hh      = ifelse(INDGEN == 120 & MIGRATEP == 10 | MIGRATEP == 11, 1, 0),
         sector_native_other_industry  = ifelse(INDGEN == 130 & MIGRATEP == 10 | MIGRATEP == 11, 1, 0))


# summarise for each district 

benin_migrant_indgen_2013 <- benin_migrant_indgen %>%
  filter(YEAR == 2013) %>%
  group_by(GEOLEV2) %>%
  summarise(sector_migrant_agriculture =weighted.mean(sector_migrant_agriculture, PERWT),
            sector_migrant_mining =weighted.mean(sector_migrant_mining, PERWT),
            sector_migrant_manufacturing =weighted.mean(sector_migrant_manufacturing, PERWT),
            sector_migrant_utilities =weighted.mean(sector_migrant_utilities, PERWT),
            sector_migrant_construction =weighted.mean(sector_migrant_construction, PERWT),
            sector_migrant_trade =weighted.mean(sector_migrant_trade, PERWT),
            sector_migrant_accommodation =weighted.mean(sector_migrant_accommodation, PERWT),
            sector_migrant_transport =weighted.mean(sector_migrant_transport, PERWT),
            sector_migrant_finance =weighted.mean(sector_migrant_finance, PERWT),
            sector_migrant_public = weighted.mean(sector_migrant_public, PERWT),
            sector_migrant_unspec_services = weighted.mean(sector_migrant_unspec_services, PERWT),
            sector_migrant_business = weighted.mean(sector_migrant_business, PERWT),
            sector_migrant_education = weighted.mean(sector_migrant_education, PERWT),
            sector_migrant_health = weighted.mean(sector_migrant_health, PERWT),
            sector_migrant_other_services = weighted.mean(sector_migrant_other_services, PERWT),
            sector_migrant_private_hh = weighted.mean(sector_migrant_private_hh, PERWT),
            sector_migrant_other_industry = weighted.mean(sector_migrant_other_industry, PERWT),
            sector_native_agriculture =weighted.mean(sector_native_agriculture, PERWT),
            sector_native_mining =weighted.mean(sector_native_mining, PERWT),
            sector_native_manufacturing =weighted.mean(sector_native_manufacturing, PERWT),
            sector_native_utilities =weighted.mean(sector_native_utilities, PERWT),
            sector_native_construction =weighted.mean(sector_native_construction, PERWT),
            sector_native_trade =weighted.mean(sector_native_trade, PERWT),
            sector_native_accommodation =weighted.mean(sector_native_accommodation, PERWT),
            sector_native_transport =weighted.mean(sector_native_transport, PERWT),
            sector_native_finance =weighted.mean(sector_native_finance, PERWT),
            sector_native_public = weighted.mean(sector_native_public, PERWT),
            sector_native_unspec_services = weighted.mean(sector_native_unspec_services, PERWT),
            sector_native_business = weighted.mean(sector_native_business, PERWT),
            sector_native_education = weighted.mean(sector_native_education, PERWT),
            sector_native_health = weighted.mean(sector_native_health, PERWT),
            sector_native_other_services = weighted.mean(sector_native_other_services, PERWT),
            sector_native_private_hh = weighted.mean(sector_native_private_hh, PERWT),
            sector_native_other_industry = weighted.mean(sector_native_other_industry, PERWT))

# Merge with calculated shares

benin.migrant.indgen.2013.spatial <- merge(benin.map, benin_migrant_indgen_2013, by = "GEOLEV2")
rm(benin_migrant_indgen, benin_migrant_indgen_2013)

Cross-products between sector and educational attainment

I compute similar cross-products for sector and educational attainment, in order to be able to examine the distribution of sectors across different levels of education.

benin_edu_indgen <- benin %>%
  filter(!INDGEN == 0 & !INDGEN == 998 & !INDGEN== 999) %>%
  mutate(edu_noprimary_agriculture      = ifelse(INDGEN == 10 & EDATTAIN == 1, 1, 0),
         edu_noprimary_mining           = ifelse(INDGEN == 20 & EDATTAIN == 1, 1, 0),
         edu_noprimary_manufacturing    = ifelse(INDGEN == 30 & EDATTAIN == 1, 1, 0),
         edu_noprimary_utilities        = ifelse(INDGEN == 40 & EDATTAIN == 1, 1, 0),
         edu_noprimary_construction     = ifelse(INDGEN == 50 & EDATTAIN == 1, 1, 0),
         edu_noprimary_trade            = ifelse(INDGEN == 60 & EDATTAIN == 1, 1, 0),
         edu_noprimary_accommodation    = ifelse(INDGEN == 70 & EDATTAIN == 1, 1, 0),
         edu_noprimary_transport        = ifelse(INDGEN == 80 & EDATTAIN == 1, 1, 0),
         edu_noprimary_finance          = ifelse(INDGEN == 90 & EDATTAIN == 1, 1, 0),
         edu_noprimary_public          = ifelse(INDGEN == 100 & EDATTAIN == 1, 1, 0),
         edu_noprimary_unspec_services = ifelse(INDGEN == 110 & EDATTAIN == 1, 1, 0),
         edu_noprimary_business        = ifelse(INDGEN == 111 & EDATTAIN == 1, 1, 0),
         edu_noprimary_education       = ifelse(INDGEN == 112 & EDATTAIN == 1, 1, 0),
         edu_noprimary_health          = ifelse(INDGEN == 113 & EDATTAIN == 1, 1, 0),
         edu_noprimary_other_services  = ifelse(INDGEN == 114 & EDATTAIN == 1, 1, 0),
         edu_noprimary_private_hh      = ifelse(INDGEN == 120 & EDATTAIN == 1, 1, 0),
         edu_noprimary_other_industry  = ifelse(INDGEN == 130 & EDATTAIN == 1, 1, 0),
         edu_primary_agriculture       = ifelse(INDGEN == 10 & EDATTAIN == 2, 1, 0),
         edu_primary_mining            = ifelse(INDGEN == 20 & EDATTAIN == 2, 1, 0),
         edu_primary_manufacturing     = ifelse(INDGEN == 30 & EDATTAIN == 2, 1, 0),
         edu_primary_utilities         = ifelse(INDGEN == 40 & EDATTAIN == 2, 1, 0),
         edu_primary_construction      = ifelse(INDGEN == 50 & EDATTAIN == 2, 1, 0),
         edu_primary_trade             = ifelse(INDGEN == 60 & EDATTAIN == 2, 1, 0),
         edu_primary_accommodation     = ifelse(INDGEN == 70 & EDATTAIN == 2, 1, 0),
         edu_primary_transport         = ifelse(INDGEN == 80 & EDATTAIN == 2, 1, 0),
         edu_primary_finance           = ifelse(INDGEN == 90 & EDATTAIN == 2, 1, 0),
         edu_primary_public           = ifelse(INDGEN == 100 & EDATTAIN == 2, 1, 0),
         edu_primary_unspec_services  = ifelse(INDGEN == 110 & EDATTAIN == 2, 1, 0),
         edu_primary_business         = ifelse(INDGEN == 111 & EDATTAIN == 2, 1, 0),
         edu_primary_education        = ifelse(INDGEN == 112 & EDATTAIN == 2, 1, 0),
         edu_primary_health           = ifelse(INDGEN == 113 & EDATTAIN == 2, 1, 0),
         edu_primary_other_services   = ifelse(INDGEN == 114 & EDATTAIN == 2, 1, 0),
         edu_primary_private_hh       = ifelse(INDGEN == 120 & EDATTAIN == 2, 1, 0),
         edu_primary_other_industry   = ifelse(INDGEN == 130 & EDATTAIN == 2, 1, 0),
         edu_secondary_agriculture     = ifelse(INDGEN == 10 & EDATTAIN == 3, 1, 0),
         edu_secondary_mining          = ifelse(INDGEN == 20 & EDATTAIN == 3, 1, 0),
         edu_secondary_manufacturing   = ifelse(INDGEN == 30 & EDATTAIN == 3, 1, 0),
         edu_secondary_utilities       = ifelse(INDGEN == 40 & EDATTAIN == 3, 1, 0),
         edu_secondary_construction    = ifelse(INDGEN == 50 & EDATTAIN == 3, 1, 0),
         edu_secondary_trade           = ifelse(INDGEN == 60 & EDATTAIN == 3, 1, 0),
         edu_secondary_accommodation   = ifelse(INDGEN == 70 & EDATTAIN == 3, 1, 0),
         edu_secondary_transport       = ifelse(INDGEN == 80 & EDATTAIN == 3, 1, 0),
         edu_secondary_finance         = ifelse(INDGEN == 90 & EDATTAIN == 3, 1, 0),
         edu_secondary_public         = ifelse(INDGEN == 100 & EDATTAIN == 3, 1, 0),
         edu_secondary_unspec_services= ifelse(INDGEN == 110 & EDATTAIN == 3, 1, 0),
         edu_secondary_business       = ifelse(INDGEN == 111 & EDATTAIN == 3, 1, 0),
         edu_secondary_education      = ifelse(INDGEN == 112 & EDATTAIN == 3, 1, 0),
         edu_secondary_health         = ifelse(INDGEN == 113 & EDATTAIN == 3, 1, 0),
         edu_secondary_other_services = ifelse(INDGEN == 114 & EDATTAIN == 3, 1, 0),
         edu_secondary_private_hh     = ifelse(INDGEN == 120 & EDATTAIN == 3, 1, 0),
         edu_secondary_other_industry = ifelse(INDGEN == 130 & EDATTAIN == 3, 1, 0),
         edu_university_agriculture     = ifelse(INDGEN == 10 & EDATTAIN == 4, 1, 0),
         edu_university_mining          = ifelse(INDGEN == 20 & EDATTAIN == 4, 1, 0),
         edu_university_manufacturing   = ifelse(INDGEN == 30 & EDATTAIN == 4, 1, 0),
         edu_university_utilities       = ifelse(INDGEN == 40 & EDATTAIN == 4, 1, 0),
         edu_university_construction    = ifelse(INDGEN == 50 & EDATTAIN == 4, 1, 0),
         edu_university_trade           = ifelse(INDGEN == 60 & EDATTAIN == 4, 1, 0),
         edu_university_accommodation   = ifelse(INDGEN == 70 & EDATTAIN == 4, 1, 0),
         edu_university_transport       = ifelse(INDGEN == 80 & EDATTAIN == 4, 1, 0),
         edu_university_finance         = ifelse(INDGEN == 90 & EDATTAIN == 4, 1, 0),
         edu_university_public         = ifelse(INDGEN == 100 & EDATTAIN == 4, 1, 0),
         edu_university_unspec_services= ifelse(INDGEN == 110 & EDATTAIN == 4, 1, 0),
         edu_university_business       = ifelse(INDGEN == 111 & EDATTAIN == 4, 1, 0),
         edu_university_education      = ifelse(INDGEN == 112 & EDATTAIN == 4, 1, 0),
         edu_university_health         = ifelse(INDGEN == 113 & EDATTAIN == 4, 1, 0),
         edu_university_other_services = ifelse(INDGEN == 114 & EDATTAIN == 4, 1, 0),
         edu_university_private_hh     = ifelse(INDGEN == 120 & EDATTAIN == 4, 1, 0),
         edu_university_other_industry = ifelse(INDGEN == 130 & EDATTAIN == 4, 1, 0))

# summarise for each district 

benin_edu_indgen_2013 <- benin_edu_indgen %>%
  filter(YEAR == 2013) %>%
  group_by(GEOLEV2) %>%
  summarise(edu_noprimary_agriculture =weighted.mean(edu_noprimary_agriculture, PERWT),
            edu_noprimary_mining =weighted.mean(edu_noprimary_mining, PERWT),
            edu_noprimary_manufacturing =weighted.mean(edu_noprimary_manufacturing, PERWT),
            edu_noprimary_utilities =weighted.mean(edu_noprimary_utilities, PERWT),
            edu_noprimary_construction =weighted.mean(edu_noprimary_construction, PERWT),
            edu_noprimary_trade =weighted.mean(edu_noprimary_trade, PERWT),
            edu_noprimary_accommodation =weighted.mean(edu_noprimary_accommodation, PERWT),
            edu_noprimary_transport =weighted.mean(edu_noprimary_transport, PERWT),
            edu_noprimary_finance =weighted.mean(edu_noprimary_finance, PERWT),
            edu_noprimary_public = weighted.mean(edu_noprimary_public, PERWT),
            edu_noprimary_unspec_services = weighted.mean(edu_noprimary_unspec_services, PERWT),
            edu_noprimary_business = weighted.mean(edu_noprimary_business, PERWT),
            edu_noprimary_education = weighted.mean(edu_noprimary_education, PERWT),
            edu_noprimary_health = weighted.mean(edu_noprimary_health, PERWT),
            edu_noprimary_other_services = weighted.mean(edu_noprimary_other_services, PERWT),
            edu_noprimary_private_hh = weighted.mean(edu_noprimary_private_hh, PERWT),
            edu_noprimary_other_industry = weighted.mean(edu_noprimary_other_industry, PERWT),
            edu_primary_agriculture =weighted.mean(edu_primary_agriculture, PERWT),
            edu_primary_mining =weighted.mean(edu_primary_mining, PERWT),
            edu_primary_manufacturing =weighted.mean(edu_primary_manufacturing, PERWT),
            edu_primary_utilities =weighted.mean(edu_primary_utilities, PERWT),
            edu_primary_construction =weighted.mean(edu_primary_construction, PERWT),
            edu_primary_trade =weighted.mean(edu_primary_trade, PERWT),
            edu_primary_accommodation =weighted.mean(edu_primary_accommodation, PERWT),
            edu_primary_transport =weighted.mean(edu_primary_transport, PERWT),
            edu_primary_finance =weighted.mean(edu_primary_finance, PERWT),
            edu_primary_public = weighted.mean(edu_primary_public, PERWT),
            edu_primary_unspec_services = weighted.mean(edu_primary_unspec_services, PERWT),
            edu_primary_business = weighted.mean(edu_primary_business, PERWT),
            edu_primary_education = weighted.mean(edu_primary_education, PERWT),
            edu_primary_health = weighted.mean(edu_primary_health, PERWT),
            edu_primary_other_services = weighted.mean(edu_primary_other_services, PERWT),
            edu_primary_private_hh = weighted.mean(edu_primary_private_hh, PERWT),
            edu_primary_other_industry = weighted.mean(edu_primary_other_industry, PERWT),
            edu_secondary_agriculture =weighted.mean(edu_secondary_agriculture, PERWT),
            edu_secondary_mining =weighted.mean(edu_secondary_mining, PERWT),
            edu_secondary_manufacturing =weighted.mean(edu_secondary_manufacturing, PERWT),
            edu_secondary_utilities =weighted.mean(edu_secondary_utilities, PERWT),
            edu_secondary_construction =weighted.mean(edu_secondary_construction, PERWT),
            edu_secondary_trade =weighted.mean(edu_secondary_trade, PERWT),
            edu_secondary_accommodation =weighted.mean(edu_secondary_accommodation, PERWT),
            edu_secondary_transport =weighted.mean(edu_secondary_transport, PERWT),
            edu_secondary_finance =weighted.mean(edu_secondary_finance, PERWT),
            edu_secondary_public = weighted.mean(edu_secondary_public, PERWT),
            edu_secondary_unspec_services = weighted.mean(edu_secondary_unspec_services, PERWT),
            edu_secondary_business = weighted.mean(edu_secondary_business, PERWT),
            edu_secondary_education = weighted.mean(edu_secondary_education, PERWT),
            edu_secondary_health = weighted.mean(edu_secondary_health, PERWT),
            edu_secondary_other_services = weighted.mean(edu_secondary_other_services, PERWT),
            edu_secondary_private_hh = weighted.mean(edu_secondary_private_hh, PERWT),
            edu_secondary_other_industry = weighted.mean(edu_secondary_other_industry, PERWT),
            edu_university_agriculture =weighted.mean(edu_university_agriculture, PERWT),
            edu_university_mining =weighted.mean(edu_university_mining, PERWT),
            edu_university_manufacturing =weighted.mean(edu_university_manufacturing, PERWT),
            edu_university_utilities =weighted.mean(edu_university_utilities, PERWT),
            edu_university_construction =weighted.mean(edu_university_construction, PERWT),
            edu_university_trade =weighted.mean(edu_university_trade, PERWT),
            edu_university_accommodation =weighted.mean(edu_university_accommodation, PERWT),
            edu_university_transport =weighted.mean(edu_university_transport, PERWT),
            edu_university_finance =weighted.mean(edu_university_finance, PERWT),
            edu_university_public = weighted.mean(edu_university_public, PERWT),
            edu_university_unspec_services = weighted.mean(edu_university_unspec_services, PERWT),
            edu_university_business = weighted.mean(edu_university_business, PERWT),
            edu_university_education = weighted.mean(edu_university_education, PERWT),
            edu_university_health = weighted.mean(edu_university_health, PERWT),
            edu_university_other_services = weighted.mean(edu_university_other_services, PERWT),
            edu_university_private_hh = weighted.mean(edu_university_private_hh, PERWT),
            edu_university_other_industry = weighted.mean(edu_university_other_industry, PERWT))

# Merge with calculated shares

benin.edu.indgen.2013.spatial <- merge(benin.map, benin_edu_indgen_2013, by = "GEOLEV2")
rm(benin_edu_indgen, benin_edu_indgen_2013)

Cross-products between sector and occupation

I compute similar cross-products for sector and occupation, in order to be able to analyse the distribution of occupational roles across sectors.

benin_occisco_indgen <- benin %>%
  filter(!INDGEN == 0 & !INDGEN == 998 & !INDGEN== 999) %>%
  mutate(legislators_agriculture      = ifelse(INDGEN == 10 & OCCISCO == 1, 1, 0),
         legislators_mining           = ifelse(INDGEN == 20 & OCCISCO == 1, 1, 0),
         legislators_manufacturing    = ifelse(INDGEN == 30 & OCCISCO == 1, 1, 0),
         legislators_utilities        = ifelse(INDGEN == 40 & OCCISCO == 1, 1, 0),
         legislators_construction     = ifelse(INDGEN == 50 & OCCISCO == 1, 1, 0),
         legislators_trade            = ifelse(INDGEN == 60 & OCCISCO == 1, 1, 0),
         legislators_accommodation    = ifelse(INDGEN == 70 & OCCISCO == 1, 1, 0),
         legislators_transport        = ifelse(INDGEN == 80 & OCCISCO == 1, 1, 0),
         legislators_finance          = ifelse(INDGEN == 90 & OCCISCO == 1, 1, 0),
         legislators_public          = ifelse(INDGEN == 100 & OCCISCO == 1, 1, 0),
         legislators_unspec_services = ifelse(INDGEN == 110 & OCCISCO == 1, 1, 0),
         legislators_business        = ifelse(INDGEN == 111 & OCCISCO == 1, 1, 0),
         legislators_education       = ifelse(INDGEN == 112 & OCCISCO == 1, 1, 0),
         legislators_health          = ifelse(INDGEN == 113 & OCCISCO == 1, 1, 0),
         legislators_other_services  = ifelse(INDGEN == 114 & OCCISCO == 1, 1, 0),
         legislators_private_hh      = ifelse(INDGEN == 120 & OCCISCO == 1, 1, 0),
         legislators_other_industry  = ifelse(INDGEN == 130 & OCCISCO == 1, 1, 0),
         professionals_agriculture       = ifelse(INDGEN == 10 & OCCISCO == 2, 1, 0),
         professionals_mining            = ifelse(INDGEN == 20 & OCCISCO == 2, 1, 0),
         professionals_manufacturing     = ifelse(INDGEN == 30 & OCCISCO == 2, 1, 0),
         professionals_utilities         = ifelse(INDGEN == 40 & OCCISCO == 2, 1, 0),
         professionals_construction      = ifelse(INDGEN == 50 & OCCISCO == 2, 1, 0),
         professionals_trade             = ifelse(INDGEN == 60 & OCCISCO == 2, 1, 0),
         professionals_accommodation     = ifelse(INDGEN == 70 & OCCISCO == 2, 1, 0),
         professionals_transport         = ifelse(INDGEN == 80 & OCCISCO == 2, 1, 0),
         professionals_finance           = ifelse(INDGEN == 90 & OCCISCO == 2, 1, 0),
         professionals_public           = ifelse(INDGEN == 100 & OCCISCO == 2, 1, 0),
         professionals_unspec_services  = ifelse(INDGEN == 110 & OCCISCO == 2, 1, 0),
         professionals_business         = ifelse(INDGEN == 111 & OCCISCO == 2, 1, 0),
         professionals_education        = ifelse(INDGEN == 112 & OCCISCO == 2, 1, 0),
         professionals_health           = ifelse(INDGEN == 113 & OCCISCO == 2, 1, 0),
         professionals_other_services   = ifelse(INDGEN == 114 & OCCISCO == 2, 1, 0),
         professionals_private_hh       = ifelse(INDGEN == 120 & OCCISCO == 2, 1, 0),
         professionals_other_industry   = ifelse(INDGEN == 130 & OCCISCO == 2, 1, 0),
         technicians_agriculture     = ifelse(INDGEN == 10 & OCCISCO == 3, 1, 0),
         technicians_mining          = ifelse(INDGEN == 20 & OCCISCO == 3, 1, 0),
         technicians_manufacturing   = ifelse(INDGEN == 30 & OCCISCO == 3, 1, 0),
         technicians_utilities       = ifelse(INDGEN == 40 & OCCISCO == 3, 1, 0),
         technicians_construction    = ifelse(INDGEN == 50 & OCCISCO == 3, 1, 0),
         technicians_trade           = ifelse(INDGEN == 60 & OCCISCO == 3, 1, 0),
         technicians_accommodation   = ifelse(INDGEN == 70 & OCCISCO == 3, 1, 0),
         technicians_transport       = ifelse(INDGEN == 80 & OCCISCO == 3, 1, 0),
         technicians_finance         = ifelse(INDGEN == 90 & OCCISCO == 3, 1, 0),
         technicians_public         = ifelse(INDGEN == 100 & OCCISCO == 3, 1, 0),
         technicians_unspec_services= ifelse(INDGEN == 110 & OCCISCO == 3, 1, 0),
         technicians_business       = ifelse(INDGEN == 111 & OCCISCO == 3, 1, 0),
         technicians_education      = ifelse(INDGEN == 112 & OCCISCO == 3, 1, 0),
         technicians_health         = ifelse(INDGEN == 113 & OCCISCO == 3, 1, 0),
         technicians_other_services = ifelse(INDGEN == 114 & OCCISCO == 3, 1, 0),
         technicians_private_hh     = ifelse(INDGEN == 120 & OCCISCO == 3, 1, 0),
         technicians_other_industry = ifelse(INDGEN == 130 & OCCISCO == 3, 1, 0),
         clerks_agriculture     = ifelse(INDGEN == 10 & OCCISCO == 4, 1, 0),
         clerks_mining          = ifelse(INDGEN == 20 & OCCISCO == 4, 1, 0),
         clerks_manufacturing   = ifelse(INDGEN == 30 & OCCISCO == 4, 1, 0),
         clerks_utilities       = ifelse(INDGEN == 40 & OCCISCO == 4, 1, 0),
         clerks_construction    = ifelse(INDGEN == 50 & OCCISCO == 4, 1, 0),
         clerks_trade           = ifelse(INDGEN == 60 & OCCISCO == 4, 1, 0),
         clerks_accommodation   = ifelse(INDGEN == 70 & OCCISCO == 4, 1, 0),
         clerks_transport       = ifelse(INDGEN == 80 & OCCISCO == 4, 1, 0),
         clerks_finance         = ifelse(INDGEN == 90 & OCCISCO == 4, 1, 0),
         clerks_public         = ifelse(INDGEN == 100 & OCCISCO == 4, 1, 0),
         clerks_unspec_services= ifelse(INDGEN == 110 & OCCISCO == 4, 1, 0),
         clerks_business       = ifelse(INDGEN == 111 & OCCISCO == 4, 1, 0),
         clerks_education      = ifelse(INDGEN == 112 & OCCISCO == 4, 1, 0),
         clerks_health         = ifelse(INDGEN == 113 & OCCISCO == 4, 1, 0),
         clerks_other_services = ifelse(INDGEN == 114 & OCCISCO == 4, 1, 0),
         clerks_private_hh     = ifelse(INDGEN == 120 & OCCISCO == 4, 1, 0),
         clerks_other_industry = ifelse(INDGEN == 130 & OCCISCO == 4, 1, 0),
         services_agriculture      = ifelse(INDGEN == 10 & OCCISCO == 5, 1, 0),
         services_mining           = ifelse(INDGEN == 20 & OCCISCO == 5, 1, 0),
         services_manufacturing    = ifelse(INDGEN == 30 & OCCISCO == 5, 1, 0),
         services_utilities        = ifelse(INDGEN == 40 & OCCISCO == 5, 1, 0),
         services_construction     = ifelse(INDGEN == 50 & OCCISCO == 5, 1, 0),
         services_trade            = ifelse(INDGEN == 60 & OCCISCO == 5, 1, 0),
         services_accommodation    = ifelse(INDGEN == 70 & OCCISCO == 5, 1, 0),
         services_transport        = ifelse(INDGEN == 80 & OCCISCO == 5, 1, 0),
         services_finance          = ifelse(INDGEN == 90 & OCCISCO == 5, 1, 0),
         services_public          = ifelse(INDGEN == 100 & OCCISCO == 5, 1, 0),
         services_unspec_services = ifelse(INDGEN == 110 & OCCISCO == 5, 1, 0),
         services_business        = ifelse(INDGEN == 111 & OCCISCO == 5, 1, 0),
         services_education       = ifelse(INDGEN == 112 & OCCISCO == 5, 1, 0),
         services_health          = ifelse(INDGEN == 113 & OCCISCO == 5, 1, 0),
         services_other_services  = ifelse(INDGEN == 114 & OCCISCO == 5, 1, 0),
         services_private_hh      = ifelse(INDGEN == 120 & OCCISCO == 5, 1, 0),
         services_other_industry  = ifelse(INDGEN == 130 & OCCISCO == 5, 1, 0),
         agrifish_agriculture       = ifelse(INDGEN == 10 & OCCISCO == 6, 1, 0),
         agrifish_mining            = ifelse(INDGEN == 20 & OCCISCO == 6, 1, 0),
         agrifish_manufacturing     = ifelse(INDGEN == 30 & OCCISCO == 6, 1, 0),
         agrifish_utilities         = ifelse(INDGEN == 40 & OCCISCO == 6, 1, 0),
         agrifish_construction      = ifelse(INDGEN == 50 & OCCISCO == 6, 1, 0),
         agrifish_trade             = ifelse(INDGEN == 60 & OCCISCO == 6, 1, 0),
         agrifish_accommodation     = ifelse(INDGEN == 70 & OCCISCO == 6, 1, 0),
         agrifish_transport         = ifelse(INDGEN == 80 & OCCISCO == 6, 1, 0),
         agrifish_finance           = ifelse(INDGEN == 90 & OCCISCO == 6, 1, 0),
         agrifish_public           = ifelse(INDGEN == 100 & OCCISCO == 6, 1, 0),
         agrifish_unspec_services  = ifelse(INDGEN == 110 & OCCISCO == 6, 1, 0),
         agrifish_business         = ifelse(INDGEN == 111 & OCCISCO == 6, 1, 0),
         agrifish_education        = ifelse(INDGEN == 112 & OCCISCO == 6, 1, 0),
         agrifish_health           = ifelse(INDGEN == 113 & OCCISCO == 6, 1, 0),
         agrifish_other_services   = ifelse(INDGEN == 114 & OCCISCO == 6, 1, 0),
         agrifish_private_hh       = ifelse(INDGEN == 120 & OCCISCO == 6, 1, 0),
         agrifish_other_industry   = ifelse(INDGEN == 130 & OCCISCO == 6, 1, 0),
         crafts_agriculture     = ifelse(INDGEN == 10 & OCCISCO == 7, 1, 0),
         crafts_mining          = ifelse(INDGEN == 20 & OCCISCO == 7, 1, 0),
         crafts_manufacturing   = ifelse(INDGEN == 30 & OCCISCO == 7, 1, 0),
         crafts_utilities       = ifelse(INDGEN == 40 & OCCISCO == 7, 1, 0),
         crafts_construction    = ifelse(INDGEN == 50 & OCCISCO == 7, 1, 0),
         crafts_trade           = ifelse(INDGEN == 60 & OCCISCO == 7, 1, 0),
         crafts_accommodation   = ifelse(INDGEN == 70 & OCCISCO == 7, 1, 0),
         crafts_transport       = ifelse(INDGEN == 80 & OCCISCO == 7, 1, 0),
         crafts_finance         = ifelse(INDGEN == 90 & OCCISCO == 7, 1, 0),
         crafts_public         = ifelse(INDGEN == 100 & OCCISCO == 7, 1, 0),
         crafts_unspec_services= ifelse(INDGEN == 110 & OCCISCO == 7, 1, 0),
         crafts_business       = ifelse(INDGEN == 111 & OCCISCO == 7, 1, 0),
         crafts_education      = ifelse(INDGEN == 112 & OCCISCO == 7, 1, 0),
         crafts_health         = ifelse(INDGEN == 113 & OCCISCO == 7, 1, 0),
         crafts_other_services = ifelse(INDGEN == 114 & OCCISCO == 7, 1, 0),
         crafts_private_hh     = ifelse(INDGEN == 120 & OCCISCO == 7, 1, 0),
         crafts_other_industry = ifelse(INDGEN == 130 & OCCISCO == 7, 1, 0),
         plant_agriculture     = ifelse(INDGEN == 10 & OCCISCO == 8, 1, 0),
         plant_mining          = ifelse(INDGEN == 20 & OCCISCO == 8, 1, 0),
         plant_manufacturing   = ifelse(INDGEN == 30 & OCCISCO == 8, 1, 0),
         plant_utilities       = ifelse(INDGEN == 40 & OCCISCO == 8, 1, 0),
         plant_construction    = ifelse(INDGEN == 50 & OCCISCO == 8, 1, 0),
         plant_trade           = ifelse(INDGEN == 60 & OCCISCO == 8, 1, 0),
         plant_accommodation   = ifelse(INDGEN == 70 & OCCISCO == 8, 1, 0),
         plant_transport       = ifelse(INDGEN == 80 & OCCISCO == 8, 1, 0),
         plant_finance         = ifelse(INDGEN == 90 & OCCISCO == 8, 1, 0),
         plant_public         = ifelse(INDGEN == 100 & OCCISCO == 8, 1, 0),
         plant_unspec_services= ifelse(INDGEN == 110 & OCCISCO == 8, 1, 0),
         plant_business       = ifelse(INDGEN == 111 & OCCISCO == 8, 1, 0),
         plant_education      = ifelse(INDGEN == 112 & OCCISCO == 8, 1, 0),
         plant_health         = ifelse(INDGEN == 113 & OCCISCO == 8, 1, 0),
         plant_other_services = ifelse(INDGEN == 114 & OCCISCO == 8, 1, 0),
         plant_private_hh     = ifelse(INDGEN == 120 & OCCISCO == 8, 1, 0),
         plant_other_industry = ifelse(INDGEN == 130 & OCCISCO == 8, 1, 0),
         elementary_agriculture      = ifelse(INDGEN == 10 & OCCISCO == 9, 1, 0),
         elementary_mining           = ifelse(INDGEN == 20 & OCCISCO == 9, 1, 0),
         elementary_manufacturing    = ifelse(INDGEN == 30 & OCCISCO == 9, 1, 0),
         elementary_utilities        = ifelse(INDGEN == 40 & OCCISCO == 9, 1, 0),
         elementary_construction     = ifelse(INDGEN == 50 & OCCISCO == 9, 1, 0),
         elementary_trade            = ifelse(INDGEN == 60 & OCCISCO == 9, 1, 0),
         elementary_accommodation    = ifelse(INDGEN == 70 & OCCISCO == 9, 1, 0),
         elementary_transport        = ifelse(INDGEN == 80 & OCCISCO == 9, 1, 0),
         elementary_finance          = ifelse(INDGEN == 90 & OCCISCO == 9, 1, 0),
         elementary_public          = ifelse(INDGEN == 100 & OCCISCO == 9, 1, 0),
         elementary_unspec_services = ifelse(INDGEN == 110 & OCCISCO == 9, 1, 0),
         elementary_business        = ifelse(INDGEN == 111 & OCCISCO == 9, 1, 0),
         elementary_education       = ifelse(INDGEN == 112 & OCCISCO == 9, 1, 0),
         elementary_health          = ifelse(INDGEN == 113 & OCCISCO == 9, 1, 0),
         elementary_other_services  = ifelse(INDGEN == 114 & OCCISCO == 9, 1, 0),
         elementary_private_hh      = ifelse(INDGEN == 120 & OCCISCO == 9, 1, 0),
         elementary_other_industry  = ifelse(INDGEN == 130 & OCCISCO == 9, 1, 0),
         army_agriculture       = ifelse(INDGEN == 10 & OCCISCO == 10, 1, 0),
         army_mining            = ifelse(INDGEN == 20 & OCCISCO == 10, 1, 0),
         army_manufacturing     = ifelse(INDGEN == 30 & OCCISCO == 10, 1, 0),
         army_utilities         = ifelse(INDGEN == 40 & OCCISCO == 10, 1, 0),
         army_construction      = ifelse(INDGEN == 50 & OCCISCO == 10, 1, 0),
         army_trade             = ifelse(INDGEN == 60 & OCCISCO == 10, 1, 0),
         army_accommodation     = ifelse(INDGEN == 70 & OCCISCO == 10, 1, 0),
         army_transport         = ifelse(INDGEN == 80 & OCCISCO == 10, 1, 0),
         army_finance           = ifelse(INDGEN == 90 & OCCISCO == 10, 1, 0),
         army_public           = ifelse(INDGEN == 100 & OCCISCO == 10, 1, 0),
         army_unspec_services  = ifelse(INDGEN == 110 & OCCISCO == 10, 1, 0),
         army_business         = ifelse(INDGEN == 111 & OCCISCO == 10, 1, 0),
         army_education        = ifelse(INDGEN == 112 & OCCISCO == 10, 1, 0),
         army_health           = ifelse(INDGEN == 113 & OCCISCO == 10, 1, 0),
         army_other_services   = ifelse(INDGEN == 114 & OCCISCO == 10, 1, 0),
         army_private_hh       = ifelse(INDGEN == 120 & OCCISCO == 10, 1, 0),
         army_other_industry   = ifelse(INDGEN == 130 & OCCISCO == 10, 1, 0),
         other_agriculture     = ifelse(INDGEN == 10 & OCCISCO == 11, 1, 0),
         other_mining          = ifelse(INDGEN == 20 & OCCISCO == 11, 1, 0),
         other_manufacturing   = ifelse(INDGEN == 30 & OCCISCO == 11, 1, 0),
         other_utilities       = ifelse(INDGEN == 40 & OCCISCO == 11, 1, 0),
         other_construction    = ifelse(INDGEN == 50 & OCCISCO == 11, 1, 0),
         other_trade           = ifelse(INDGEN == 60 & OCCISCO == 11, 1, 0),
         other_accommodation   = ifelse(INDGEN == 70 & OCCISCO == 11, 1, 0),
         other_transport       = ifelse(INDGEN == 80 & OCCISCO == 11, 1, 0),
         other_finance         = ifelse(INDGEN == 90 & OCCISCO == 11, 1, 0),
         other_public         = ifelse(INDGEN == 100 & OCCISCO == 11, 1, 0),
         other_unspec_services= ifelse(INDGEN == 110 & OCCISCO == 11, 1, 0),
         other_business       = ifelse(INDGEN == 111 & OCCISCO == 11, 1, 0),
         other_education      = ifelse(INDGEN == 112 & OCCISCO == 11, 1, 0),
         other_health         = ifelse(INDGEN == 113 & OCCISCO == 11, 1, 0),
         other_other_services = ifelse(INDGEN == 114 & OCCISCO == 11, 1, 0),
         other_private_hh     = ifelse(INDGEN == 120 & OCCISCO == 11, 1, 0),
         other_other_industry = ifelse(INDGEN == 130 & OCCISCO == 11, 1, 0))



# summarise for each district 

benin_occisco_indgen_2013 <- benin_occisco_indgen %>%
  filter(YEAR == 2013) %>%
  group_by(GEOLEV2) %>%
  summarise(legislators_agriculture =weighted.mean(legislators_agriculture, PERWT),
            legislators_mining =weighted.mean(legislators_mining, PERWT),
            legislators_manufacturing =weighted.mean(legislators_manufacturing, PERWT),
            legislators_utilities =weighted.mean(legislators_utilities, PERWT),
            legislators_construction =weighted.mean(legislators_construction, PERWT),
            legislators_trade =weighted.mean(legislators_trade, PERWT),
            legislators_accommodation =weighted.mean(legislators_accommodation, PERWT),
            legislators_transport =weighted.mean(legislators_transport, PERWT),
            legislators_finance =weighted.mean(legislators_finance, PERWT),
            legislators_public = weighted.mean(legislators_public, PERWT),
            legislators_unspec_services = weighted.mean(legislators_unspec_services, PERWT),
            legislators_business = weighted.mean(legislators_business, PERWT),
            legislators_education = weighted.mean(legislators_education, PERWT),
            legislators_health = weighted.mean(legislators_health, PERWT),
            legislators_other_services = weighted.mean(legislators_other_services, PERWT),
            legislators_private_hh = weighted.mean(legislators_private_hh, PERWT),
            legislators_other_industry = weighted.mean(legislators_other_industry, PERWT),
            professionals_agriculture =weighted.mean(professionals_agriculture, PERWT),
            professionals_mining =weighted.mean(professionals_mining, PERWT),
            professionals_manufacturing =weighted.mean(professionals_manufacturing, PERWT),
            professionals_utilities =weighted.mean(professionals_utilities, PERWT),
            professionals_construction =weighted.mean(professionals_construction, PERWT),
            professionals_trade =weighted.mean(professionals_trade, PERWT),
            professionals_accommodation =weighted.mean(professionals_accommodation, PERWT),
            professionals_transport =weighted.mean(professionals_transport, PERWT),
            professionals_finance =weighted.mean(professionals_finance, PERWT),
            professionals_public = weighted.mean(professionals_public, PERWT),
            professionals_unspec_services = weighted.mean(professionals_unspec_services, PERWT),
            professionals_business = weighted.mean(professionals_business, PERWT),
            professionals_education = weighted.mean(professionals_education, PERWT),
            professionals_health = weighted.mean(professionals_health, PERWT),
            professionals_other_services = weighted.mean(professionals_other_services, PERWT),
            professionals_private_hh = weighted.mean(professionals_private_hh, PERWT),
            professionals_other_industry = weighted.mean(professionals_other_industry, PERWT),
            technicians_agriculture =weighted.mean(technicians_agriculture, PERWT),
            technicians_mining =weighted.mean(technicians_mining, PERWT),
            technicians_manufacturing =weighted.mean(technicians_manufacturing, PERWT),
            technicians_utilities =weighted.mean(technicians_utilities, PERWT),
            technicians_construction =weighted.mean(technicians_construction, PERWT),
            technicians_trade =weighted.mean(technicians_trade, PERWT),
            technicians_accommodation =weighted.mean(technicians_accommodation, PERWT),
            technicians_transport =weighted.mean(technicians_transport, PERWT),
            technicians_finance =weighted.mean(technicians_finance, PERWT),
            technicians_public = weighted.mean(technicians_public, PERWT),
            technicians_unspec_services = weighted.mean(technicians_unspec_services, PERWT),
            technicians_business = weighted.mean(technicians_business, PERWT),
            technicians_education = weighted.mean(technicians_education, PERWT),
            technicians_health = weighted.mean(technicians_health, PERWT),
            technicians_other_services = weighted.mean(technicians_other_services, PERWT),
            technicians_private_hh = weighted.mean(technicians_private_hh, PERWT),
            technicians_other_industry = weighted.mean(technicians_other_industry, PERWT),
            clerks_agriculture =weighted.mean(clerks_agriculture, PERWT),
            clerks_mining =weighted.mean(clerks_mining, PERWT),
            clerks_manufacturing =weighted.mean(clerks_manufacturing, PERWT),
            clerks_utilities =weighted.mean(clerks_utilities, PERWT),
            clerks_construction =weighted.mean(clerks_construction, PERWT),
            clerks_trade =weighted.mean(clerks_trade, PERWT),
            clerks_accommodation =weighted.mean(clerks_accommodation, PERWT),
            clerks_transport =weighted.mean(clerks_transport, PERWT),
            clerks_finance =weighted.mean(clerks_finance, PERWT),
            clerks_public = weighted.mean(clerks_public, PERWT),
            clerks_unspec_services = weighted.mean(clerks_unspec_services, PERWT),
            clerks_business = weighted.mean(clerks_business, PERWT),
            clerks_education = weighted.mean(clerks_education, PERWT),
            clerks_health = weighted.mean(clerks_health, PERWT),
            clerks_other_services = weighted.mean(clerks_other_services, PERWT),
            clerks_private_hh = weighted.mean(clerks_private_hh, PERWT),
            clerks_other_industry = weighted.mean(clerks_other_industry, PERWT),
            services_agriculture =weighted.mean(services_agriculture, PERWT),
            services_mining =weighted.mean(services_mining, PERWT),
            services_manufacturing =weighted.mean(services_manufacturing, PERWT),
            services_utilities =weighted.mean(services_utilities, PERWT),
            services_construction =weighted.mean(services_construction, PERWT),
            services_trade =weighted.mean(services_trade, PERWT),
            services_accommodation =weighted.mean(services_accommodation, PERWT),
            services_transport =weighted.mean(services_transport, PERWT),
            services_finance =weighted.mean(services_finance, PERWT),
            services_public = weighted.mean(services_public, PERWT),
            services_unspec_services = weighted.mean(services_unspec_services, PERWT),
            services_business = weighted.mean(services_business, PERWT),
            services_education = weighted.mean(services_education, PERWT),
            services_health = weighted.mean(services_health, PERWT),
            services_other_services = weighted.mean(services_other_services, PERWT),
            services_private_hh = weighted.mean(services_private_hh, PERWT),
            services_other_industry = weighted.mean(services_other_industry, PERWT),
            agrifish_agriculture =weighted.mean(agrifish_agriculture, PERWT),
            agrifish_mining =weighted.mean(agrifish_mining, PERWT),
            agrifish_manufacturing =weighted.mean(agrifish_manufacturing, PERWT),
            agrifish_utilities =weighted.mean(agrifish_utilities, PERWT),
            agrifish_construction =weighted.mean(agrifish_construction, PERWT),
            agrifish_trade =weighted.mean(agrifish_trade, PERWT),
            agrifish_accommodation =weighted.mean(agrifish_accommodation, PERWT),
            agrifish_transport =weighted.mean(agrifish_transport, PERWT),
            agrifish_finance =weighted.mean(agrifish_finance, PERWT),
            agrifish_public = weighted.mean(agrifish_public, PERWT),
            agrifish_unspec_services = weighted.mean(agrifish_unspec_services, PERWT),
            agrifish_business = weighted.mean(agrifish_business, PERWT),
            agrifish_education = weighted.mean(agrifish_education, PERWT),
            agrifish_health = weighted.mean(agrifish_health, PERWT),
            agrifish_other_services = weighted.mean(agrifish_other_services, PERWT),
            agrifish_private_hh = weighted.mean(agrifish_private_hh, PERWT),
            agrifish_other_industry = weighted.mean(agrifish_other_industry, PERWT),
            crafts_agriculture =weighted.mean(crafts_agriculture, PERWT),
            crafts_mining =weighted.mean(crafts_mining, PERWT),
            crafts_manufacturing =weighted.mean(crafts_manufacturing, PERWT),
            crafts_utilities =weighted.mean(crafts_utilities, PERWT),
            crafts_construction =weighted.mean(crafts_construction, PERWT),
            crafts_trade =weighted.mean(crafts_trade, PERWT),
            crafts_accommodation =weighted.mean(crafts_accommodation, PERWT),
            crafts_transport =weighted.mean(crafts_transport, PERWT),
            crafts_finance =weighted.mean(crafts_finance, PERWT),
            crafts_public = weighted.mean(crafts_public, PERWT),
            crafts_unspec_services = weighted.mean(crafts_unspec_services, PERWT),
            crafts_business = weighted.mean(crafts_business, PERWT),
            crafts_education = weighted.mean(crafts_education, PERWT),
            crafts_health = weighted.mean(crafts_health, PERWT),
            crafts_other_services = weighted.mean(crafts_other_services, PERWT),
            crafts_private_hh = weighted.mean(crafts_private_hh, PERWT),
            crafts_other_industry = weighted.mean(crafts_other_industry, PERWT),
            plant_agriculture =weighted.mean(plant_agriculture, PERWT),
            plant_mining =weighted.mean(plant_mining, PERWT),
            plant_manufacturing =weighted.mean(plant_manufacturing, PERWT),
            plant_utilities =weighted.mean(plant_utilities, PERWT),
            plant_construction =weighted.mean(plant_construction, PERWT),
            plant_trade =weighted.mean(plant_trade, PERWT),
            plant_accommodation =weighted.mean(plant_accommodation, PERWT),
            plant_transport =weighted.mean(plant_transport, PERWT),
            plant_finance =weighted.mean(plant_finance, PERWT),
            plant_public = weighted.mean(plant_public, PERWT),
            plant_unspec_services = weighted.mean(plant_unspec_services, PERWT),
            plant_business = weighted.mean(plant_business, PERWT),
            plant_education = weighted.mean(plant_education, PERWT),
            plant_health = weighted.mean(plant_health, PERWT),
            plant_other_services = weighted.mean(plant_other_services, PERWT),
            plant_private_hh = weighted.mean(plant_private_hh, PERWT),
            plant_other_industry = weighted.mean(plant_other_industry, PERWT),
            elementary_agriculture =weighted.mean(elementary_agriculture, PERWT),
            elementary_mining =weighted.mean(elementary_mining, PERWT),
            elementary_manufacturing =weighted.mean(elementary_manufacturing, PERWT),
            elementary_utilities =weighted.mean(elementary_utilities, PERWT),
            elementary_construction =weighted.mean(elementary_construction, PERWT),
            elementary_trade =weighted.mean(elementary_trade, PERWT),
            elementary_accommodation =weighted.mean(elementary_accommodation, PERWT),
            elementary_transport =weighted.mean(elementary_transport, PERWT),
            elementary_finance =weighted.mean(elementary_finance, PERWT),
            elementary_public = weighted.mean(elementary_public, PERWT),
            elementary_unspec_services = weighted.mean(elementary_unspec_services, PERWT),
            elementary_business = weighted.mean(elementary_business, PERWT),
            elementary_education = weighted.mean(elementary_education, PERWT),
            elementary_health = weighted.mean(elementary_health, PERWT),
            elementary_other_services = weighted.mean(elementary_other_services, PERWT),
            elementary_private_hh = weighted.mean(elementary_private_hh, PERWT),
            elementary_other_industry = weighted.mean(elementary_other_industry, PERWT),
            army_agriculture =weighted.mean(army_agriculture, PERWT),
            army_mining =weighted.mean(army_mining, PERWT),
            army_manufacturing =weighted.mean(army_manufacturing, PERWT),
            army_utilities =weighted.mean(army_utilities, PERWT),
            army_construction =weighted.mean(army_construction, PERWT),
            army_trade =weighted.mean(army_trade, PERWT),
            army_accommodation =weighted.mean(army_accommodation, PERWT),
            army_transport =weighted.mean(army_transport, PERWT),
            army_finance =weighted.mean(army_finance, PERWT),
            army_public = weighted.mean(army_public, PERWT),
            army_unspec_services = weighted.mean(army_unspec_services, PERWT),
            army_business = weighted.mean(army_business, PERWT),
            army_education = weighted.mean(army_education, PERWT),
            army_health = weighted.mean(army_health, PERWT),
            army_other_services = weighted.mean(army_other_services, PERWT),
            army_private_hh = weighted.mean(army_private_hh, PERWT),
            army_other_industry = weighted.mean(army_other_industry, PERWT),
            other_agriculture =weighted.mean(other_agriculture, PERWT),
            other_mining =weighted.mean(other_mining, PERWT),
            other_manufacturing =weighted.mean(other_manufacturing, PERWT),
            other_utilities =weighted.mean(other_utilities, PERWT),
            other_construction =weighted.mean(other_construction, PERWT),
            other_trade =weighted.mean(other_trade, PERWT),
            other_accommodation =weighted.mean(other_accommodation, PERWT),
            other_transport =weighted.mean(other_transport, PERWT),
            other_finance =weighted.mean(other_finance, PERWT),
            other_public = weighted.mean(other_public, PERWT),
            other_unspec_services = weighted.mean(other_unspec_services, PERWT),
            other_business = weighted.mean(other_business, PERWT),
            other_education = weighted.mean(other_education, PERWT),
            other_health = weighted.mean(other_health, PERWT),
            other_other_services = weighted.mean(other_other_services, PERWT),
            other_private_hh = weighted.mean(other_private_hh, PERWT),
            other_other_industry = weighted.mean(other_other_industry, PERWT))

# Merge with calculated shares

benin.occisco.indgen.2013.spatial <- merge(benin.map, benin_occisco_indgen_2013, by = "GEOLEV2")
rm(benin_occisco_indgen, benin_occisco_indgen_2013)

Merging all together

Here, I merge all spatially explicit datasets together using their geography as the ID variable, in order to obtain a dataset with administrative units as observational units and the calculated shares as variables. Datasets constructed this way will have a large horizontal dimension, with each column corresponding to a variable.

benin.spatial.2013 <- merge(benin.pop.2013.spatial, sf::st_drop_geometry(benin.ru.2013.spatial), by = c("CNTRY_NAME", "ADMIN_NAME", "GEOLEV2"))
benin.spatial.2013 <- merge(benin.spatial.2013, sf::st_drop_geometry(benin.re.2013.spatial), by = c("CNTRY_NAME", "ADMIN_NAME", "GEOLEV2"))
benin.spatial.2013 <- merge(benin.spatial.2013, sf::st_drop_geometry(benin.nchild.2013.spatial),  by = c("CNTRY_NAME", "ADMIN_NAME", "GEOLEV2"))
benin.spatial.2013 <- merge(benin.spatial.2013, sf::st_drop_geometry(benin.nchlt5.2013.spatial),by = c("CNTRY_NAME", "ADMIN_NAME", "GEOLEV2"))
benin.spatial.2013 <- merge(benin.spatial.2013, sf::st_drop_geometry(benin.age.2013.spatial),by = c("CNTRY_NAME", "ADMIN_NAME", "GEOLEV2"))
benin.spatial.2013 <- merge(benin.spatial.2013, sf::st_drop_geometry(benin.sex.2013.spatial),by = c("CNTRY_NAME", "ADMIN_NAME", "GEOLEV2"))
benin.spatial.2013 <- merge(benin.spatial.2013, sf::st_drop_geometry(benin.marst.2013.spatial),by = c("CNTRY_NAME", "ADMIN_NAME", "GEOLEV2"))
benin.spatial.2013 <- merge(benin.spatial.2013, sf::st_drop_geometry(benin.chborn.2013.spatial),by = c("CNTRY_NAME", "ADMIN_NAME", "GEOLEV2"))
benin.spatial.2013 <- merge(benin.spatial.2013, sf::st_drop_geometry(benin.chsurv.2013.spatial),by = c("CNTRY_NAME", "ADMIN_NAME", "GEOLEV2"))
benin.spatial.2013 <- merge(benin.spatial.2013, sf::st_drop_geometry(benin.births.2013.spatial),by = c("CNTRY_NAME", "ADMIN_NAME", "GEOLEV2"))
benin.spatial.2013 <- merge(benin.spatial.2013, sf::st_drop_geometry(benin.birthsurv.2013.spatial),by = c("CNTRY_NAME", "ADMIN_NAME", "GEOLEV2"))
benin.spatial.2013 <- merge(benin.spatial.2013, sf::st_drop_geometry(benin.mortmot.2013.spatial),by = c("CNTRY_NAME", "ADMIN_NAME", "GEOLEV2"))
benin.spatial.2013 <- merge(benin.spatial.2013, sf::st_drop_geometry(benin.religion.2013.spatial),by = c("CNTRY_NAME", "ADMIN_NAME", "GEOLEV2"))
benin.spatial.2013 <- merge(benin.spatial.2013, sf::st_drop_geometry(benin.ethnic.2013.spatial),by = c("CNTRY_NAME", "ADMIN_NAME", "GEOLEV2"))
benin.spatial.2013 <- merge(benin.spatial.2013, sf::st_drop_geometry(benin.lang.2013.spatial),by = c("CNTRY_NAME", "ADMIN_NAME", "GEOLEV2"))
benin.spatial.2013 <- merge(benin.spatial.2013, sf::st_drop_geometry(benin.school.2013.spatial),by = c("CNTRY_NAME", "ADMIN_NAME", "GEOLEV2"))
benin.spatial.2013 <- merge(benin.spatial.2013, sf::st_drop_geometry(benin.lit.2013.spatial),by = c("CNTRY_NAME", "ADMIN_NAME", "GEOLEV2"))
benin.spatial.2013 <- merge(benin.spatial.2013, sf::st_drop_geometry(benin.edattain.2013.spatial),by = c("CNTRY_NAME", "ADMIN_NAME", "GEOLEV2"))
benin.spatial.2013 <- merge(benin.spatial.2013, sf::st_drop_geometry(benin.yrschool.2013.spatial),by = c("CNTRY_NAME", "ADMIN_NAME", "GEOLEV2"))
benin.spatial.2013 <- merge(benin.spatial.2013, sf::st_drop_geometry(benin.empstat.2013.spatial),by = c("CNTRY_NAME", "ADMIN_NAME", "GEOLEV2"))
benin.spatial.2013 <- merge(benin.spatial.2013, sf::st_drop_geometry(benin.labforce.2013.spatial),by = c("CNTRY_NAME", "ADMIN_NAME", "GEOLEV2"))
benin.spatial.2013 <- merge(benin.spatial.2013, sf::st_drop_geometry(benin.occisco.2013.spatial),by = c("CNTRY_NAME", "ADMIN_NAME", "GEOLEV2"))
benin.spatial.2013 <- merge(benin.spatial.2013, sf::st_drop_geometry(benin.sectors.2013.spatial), by = c("CNTRY_NAME", "ADMIN_NAME", "GEOLEV2"))
benin.spatial.2013 <- merge(benin.spatial.2013, sf::st_drop_geometry(benin.indgen.2013.spatial), by = c("CNTRY_NAME", "ADMIN_NAME", "GEOLEV2"))
benin.spatial.2013 <- merge(benin.spatial.2013, sf::st_drop_geometry(benin.migratep.2013.spatial),by = c("CNTRY_NAME", "ADMIN_NAME", "GEOLEV2"))
benin.spatial.2013 <- merge(benin.spatial.2013, sf::st_drop_geometry(benin.gender.empstat.2013.spatial),by = c("CNTRY_NAME", "ADMIN_NAME", "GEOLEV2"))
benin.spatial.2013 <- merge(benin.spatial.2013, sf::st_drop_geometry(benin.gender.occisco.2013.spatial),by = c("CNTRY_NAME", "ADMIN_NAME", "GEOLEV2"))
benin.spatial.2013 <- merge(benin.spatial.2013, sf::st_drop_geometry(benin.gender.indgen.2013.spatial),by = c("CNTRY_NAME", "ADMIN_NAME", "GEOLEV2"))
benin.spatial.2013 <- merge(benin.spatial.2013, sf::st_drop_geometry(benin.age.empstat.2013.spatial),by = c("CNTRY_NAME", "ADMIN_NAME", "GEOLEV2"))
benin.spatial.2013 <- merge(benin.spatial.2013, sf::st_drop_geometry(benin.age.occisco.2013.spatial),by = c("CNTRY_NAME", "ADMIN_NAME", "GEOLEV2"))
benin.spatial.2013 <- merge(benin.spatial.2013, sf::st_drop_geometry(benin.age.indgen.2013.spatial),by = c("CNTRY_NAME", "ADMIN_NAME", "GEOLEV2"))
benin.spatial.2013 <- merge(benin.spatial.2013, sf::st_drop_geometry(benin.migrant.empstat.2013.spatial),by = c("CNTRY_NAME", "ADMIN_NAME", "GEOLEV2"))
benin.spatial.2013 <- merge(benin.spatial.2013, sf::st_drop_geometry(benin.migrant.occisco.2013.spatial),by = c("CNTRY_NAME", "ADMIN_NAME", "GEOLEV2"))
benin.spatial.2013 <- merge(benin.spatial.2013, sf::st_drop_geometry(benin.migrant.indgen.2013.spatial),by = c("CNTRY_NAME", "ADMIN_NAME", "GEOLEV2"))
benin.spatial.2013 <- merge(benin.spatial.2013, sf::st_drop_geometry(benin.edu.indgen.2013.spatial),by = c("CNTRY_NAME", "ADMIN_NAME", "GEOLEV2"))
benin.spatial.2013 <- merge(benin.spatial.2013, sf::st_drop_geometry(benin.occisco.indgen.2013.spatial),by = c("CNTRY_NAME", "ADMIN_NAME", "GEOLEV2"))

Night Lights Aggregation

Once the various spatial datasets have been merged together, raster datasets can be aggregated to the whole matrix. Here, I do not need to dummify categories nor to collapse data from the individual level to the geographic level of choice. It suffices to extract the raster values to the desired administrative units by choosing the needed mathematical operator. I compute both the mean level and the Sum of Lights for each district.

NTL2013 <- raster::raster("~/Dropbox/Work/RA/services micro ssa/Lorenzo_RA_Work/NTL/F182013.v4/F182013.v4c_web.stable_lights.avg_vis.tif")

benin.spatial.2013 <- benin.spatial.2013 %>%
  mutate(ntl_mean=exactextractr::exact_extract(NTL2013, benin.spatial.2013, "mean"),
         ntl_sum=exactextractr::exact_extract(NTL2013, benin.spatial.2013, "sum"))

I plot the SoL for Benin in 2013, highlighting how the capital city and in general the southernmost (comparatively service-rich) regions are at the right tail of the nightlights distribution. This fact seems to confirm the correlation between developed economic activity and nightlights.

ggplot() +
  geom_sf(data = benin.spatial.2013$geometry) +
  geom_sf(data = benin.spatial.2013, aes(fill=ntl_sum)) +
  scale_fill_viridis_c(option = "viridis")  + theme_void()

LSE, Department of Geography and Environment; Grantham Research Institute for Climate Change and the Environment; l.sileci@lse.ac.uk

Back to menu

Within-sector Labour Force Composition

Click on the image to view it fullscreen or to download it

Back to menu