Contract Awards In Investment Project Financing: An Analysis On Supplier Locality Influence on WB Financing


Sourced from https://datacatalog.worldbank.org/search/dataset/0066219/Contract-Awards-in-Investment-Project-Financing “The data available here now includes all contract awards financed by The World Bank under Investment Project Financing (IPF) operations. The data source is STEP (Systematic Tracking of Exchanges in Procurement), which is required to be used by Borrowers in all IPF operations subject to the World Bank’s Procurement Regulations. Data is entered by Borrowers.”Supplier Country / Economy” represents the place of supplier registration, which may or may not be the supplier’s actual country / economy of origin. Information does not include awards to subcontractors, nor does it account for cofinancing.

Please note that for contracts awarded to joint-ventures of multiple companies, the total contract value was split equally amongst the members of the joint-venture.” 

library(tidyverse)
## ── Attaching core tidyverse packages ──────────────────────── tidyverse 2.0.0 ──
## ✔ dplyr     1.1.4     ✔ readr     2.1.5
## ✔ forcats   1.0.0     ✔ stringr   1.5.1
## ✔ ggplot2   3.5.1     ✔ tibble    3.2.1
## ✔ lubridate 1.9.3     ✔ tidyr     1.3.1
## ✔ purrr     1.0.2     
## ── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
## ✖ dplyr::filter() masks stats::filter()
## ✖ dplyr::lag()    masks stats::lag()
## ℹ Use the conflicted package (<http://conflicted.r-lib.org/>) to force all conflicts to become errors
library(Hmisc)
## 
## Attaching package: 'Hmisc'
## 
## The following objects are masked from 'package:dplyr':
## 
##     src, summarize
## 
## The following objects are masked from 'package:base':
## 
##     format.pval, units
library(readr)
library(ggplot2)

contract_awards <- readr::read_csv("Contract_Awards_in_Investment_Project_Financing.csv", show_col_types = FALSE)
#Hmisc::describe(contract_awards)
#detach("package:psych", unload = TRUE)

# Simplify column names
contract_awards <- contract_awards |> rename(review_type = `Review type`, project_id = `Project ID`, borrower_country = `Borrower Country / Economy`, supplier_country = `Supplier Country / Economy`)




Same Country Suppliers and Borrowers Influence Financing?


Using the World Bank Contract Awards data set, lets determine whether a contract in which the supplier and borrower are from the same country positively correlates with a higher financing approval, negatively correlates, or has no correlation.

Assumptions: Limiting the data set where Review type = Post only to narrow the scope to decisively awarded and financed projects. 

post_contracts <- contract_awards |>
  dplyr::filter(contract_awards$review_type == "Post")

nrow(contract_awards)
## [1] 247762
nrow(post_contracts)
## [1] 224654


Reduce the data set size to just the subset of columns necessary to make the determination: Borrower country, Supplier country, Project ID

post_contracts <- post_contracts |>
  dplyr::select(project_id, borrower_country, supplier_country, Region)

#head(bor_sup_cnty, n=20)


# Count unique project Ids grouping by Region, Borrower Country, and Supplier Country
result <- post_contracts |>
  dplyr::group_by(Region, borrower_country, supplier_country) |>
  dplyr::summarize(proj_count = dplyr::n_distinct(project_id))
## `summarise()` has grouped output by 'Region', 'borrower_country'. You can
## override using the `.groups` argument.
# Add a column that indicates that a borrower and supplier country are the same
result <- result |>
  mutate(local = (borrower_country == supplier_country))

result
## # A tibble: 3,522 × 5
## # Groups:   Region, borrower_country [149]
##    Region                borrower_country supplier_country      proj_count local
##    <chr>                 <chr>            <chr>                      <int> <lgl>
##  1 AFRICA                Africa           Burundi                        1 FALSE
##  2 AFRICA                Africa           Congo, Democratic Re…          1 FALSE
##  3 AFRICA                Africa           France                         1 FALSE
##  4 AFRICA                Africa           Kenya                          1 FALSE
##  5 AFRICA                Africa           Uganda                         1 FALSE
##  6 AFRICA                Africa           United Kingdom                 1 FALSE
##  7 EAST ASIA AND PACIFIC Cambodia         Australia                      3 FALSE
##  8 EAST ASIA AND PACIFIC Cambodia         Bangladesh                     2 FALSE
##  9 EAST ASIA AND PACIFIC Cambodia         Cambodia                      25 TRUE 
## 10 EAST ASIA AND PACIFIC Cambodia         Cameroon                       1 FALSE
## # ℹ 3,512 more rows



What would the distribution look like on a weighted plot. The expectation is that a positive correlation would suggest a tight grouping around the 45 degree line. 

AFRICA <- result |> dplyr::filter(Region == "AFRICA")
E_ASIA <- result |> dplyr::filter(Region == "EAST ASIA AND PACIFIC")
ES_AFRICA <- result |> dplyr::filter(Region == "Eastern and Southern Africa")
EC_ASIA <- result |> dplyr::filter(Region == "EUROPE AND CENTRAL ASIA")
LATIN <- result |> dplyr::filter(Region == "LATIN AMERICA AND CARIBBEAN")

#lowest : AFRICA                       EAST ASIA AND PACIFIC        Eastern and Southern Africa  EUROPE AND CENTRAL ASIA      LATIN AMERICA AND CARIBBEAN 
#highest: LATIN AMERICA AND CARIBBEAN  MIDDLE EAST AND NORTH AFRICA OTHER                        SOUTH ASIA                   Western and Central Africa  

ggplot(AFRICA, aes(x = borrower_country, y = supplier_country, size = proj_count, color = local)) + geom_point()

ggplot(E_ASIA, aes(x = borrower_country, y = supplier_country, size = proj_count, color = local)) + geom_point()

ggplot(ES_AFRICA, aes(x = borrower_country, y = supplier_country, size = proj_count, color = local)) + geom_point()

ggplot(EC_ASIA, aes(x = borrower_country, y = supplier_country, size = proj_count, color = local)) + geom_point()

ggplot(LATIN, aes(x = borrower_country, y = supplier_country, size = proj_count, color = local)) + geom_point()




Conclusion


From the scatter plots there is a loose but not convincing correlation suggesting the relationship is nothing more than coincidence. I would conclude there is no correlation between local borrower and supplier financing.