General analysis

library(tidyverse)
library(scales)
library(patchwork)
library(pander)
library(kableExtra)
library(countrycode)
library(sf)
library(here)

source(here("lib", "graphics.R"))
source(here("lib", "pandoc.R"))

my_seed <- 1234
set.seed(my_seed)

df_donor <- readRDS(here("data", "derived_data", "df_donor.rds"))
df_donor_us <- readRDS(here("data", "derived_data", "df_donor_usaid.rds"))

df_country_aid <- readRDS(here("data", "derived_data", "df_country_aid.rds"))
df_country_aid_laws <- filter(df_country_aid, laws)

df_autocracies <- readRDS(here("data", "derived_data", "df_autocracies.rds"))

Overall data summary

num_countries <- df_country_aid_laws %>% distinct(gwcode) %>% nrow()
num_years <- df_country_aid_laws %>% distinct(year) %>% nrow()
year_first <- df_country_aid_laws %>% distinct(year) %>% min()
year_last <- df_country_aid_laws %>% distinct(year) %>% max()

Our data includes information about 142 countries across 25 years (from 1990–2014)

Summary of variables in model

(From our AJPS submission): The values here are slightly different from what we had at ISA and MPSA (and our ISQ submission) because we’re now using V-Dem 8.0 and AidData 3.1.

And now the values are even different-er, since we’re using V-Dem 10.0 and have really clean data now.

vars_to_summarize <- read_csv(here("data", "manual_data", "coefs.csv")) %>% 
  mutate(summary_order = 1:n())

vars_summarized <- df_country_aid_laws %>%
  select(one_of(vars_to_summarize$term)) %>%
  mutate(total_oda = total_oda / 1000000) %>%
  mutate(gdpcap_log = exp(gdpcap_log)) %>% 
  pivot_longer(names_to = "term", values_to = "value", everything()) %>%
  filter(!is.na(value)) %>% 
  group_by(term) %>%
  summarize(N = n(),
            Mean = mean(value),
            Median = median(value),
            `Std. Dev.` = sd(value),
            Min = min(value),
            Max = max(value)) %>% 
  ungroup() %>% 
  left_join(vars_to_summarize, by = "term") %>%
  arrange(summary_order) %>%
  mutate(across(c(Mean, `Std. Dev.`, Median, Min, Max),
                ~case_when(
                  format == "percent" ~ percent_format(accuracy = 0.1)(.),
                  format == "dollar" ~ dollar_format(accuracy = 1)(.),
                  format == "number" ~ comma_format(accuracy = 0.1)(.)))) %>% 
  mutate(N = comma(N)) %>% 
  select(category, subcategory, Variable = term_clean_table, Source = source, 
         Mean, `Std. Dev.`, Median, Min, Max, N)

vars_summarized %>% 
  select(-category, ` ` = subcategory) %>% 
  kbl() %>% 
  kable_styling() %>% 
  pack_rows(index = table(fct_inorder(vars_summarized$category))) %>% 
  collapse_rows(columns = 1, valign = "top")

	Variable	Source	Mean	Std. Dev.	Median	Min	Max	N
Outcome
	Total aid (constant 2011 USD, millions)	OECD and AidData	$1,261	$2,918	$445	$0	$63,233	3,435
	Proportion of contentious aid	OECD and AidData	6.8%	10.4%	2.9%	0.0%	100.0%	3,435
	Proportion of aid to domestic NGOs	USAID	4.2%	13.9%	0.2%	0.0%	100.0%	3,435
	Proportion of aid to foreign NGOs	USAID	13.1%	20.1%	0.9%	0.0%	100.0%	3,435
Treatment
	Total legal barriers	@christensen2013 and @Chaudhry:2016	1.6	1.9	1.0	0.0	9.5	3,435
	Barriers to advocacy	@christensen2013 and @Chaudhry:2016	0.3	0.5	0.0	0.0	2.0	3,435
	Barriers to entry	@christensen2013 and @Chaudhry:2016	0.8	0.9	1.0	0.0	3.0	3,435
	Barriers to funding	@christensen2013 and @Chaudhry:2016	0.4	0.9	0.0	0.0	4.5	3,435
	Core civil society index	@christensen2013 and @Chaudhry:2016	0.6	0.3	0.7	0.0	1.0	3,435
Confounders
Human rights and politics	Electoral democracy index (polyarchy)	V-Dem	0.4	0.2	0.4	0.0	0.9	3,435
	Political corruption index	V-Dem	0.6	0.3	0.7	0.0	1.0	3,435
	Rule of law index	V-Dem	0.5	0.3	0.4	0.0	1.0	3,435
	Civil liberties index	V-Dem	0.6	0.2	0.7	0.0	1.0	3,435
	Physical violence index	V-Dem	0.6	0.3	0.6	0.0	1.0	3,435
	Private civil liberties index	V-Dem	0.6	0.3	0.7	0.0	1.0	3,435
Economics and development	GDP per capita (constant 2011 USD)	UN	$6,418	$10,464	$2,758	$92	$82,410	3,435
	Trade as % of GDP	UN	78.4%	46.1%	69.6%	1.9%	441.6%	3,435
	Educational equality	V-Dem	0.1	1.3	-0.1	-3.2	3.6	3,435
	Health equality	V-Dem	0.1	1.4	-0.1	-3.3	3.5	3,435
	Infant mortality rate (deaths per 1,000 birhts)	V-Dem and Gapminder	44.7	34.3	35.3	2.2	171.0	3,435
Unexpected shocks	Internal conflict in last 5 years	UCDP/PRIO	0.3	0.5	0.0	0.0	1.0	3,435
	Natural disasters	EM-DAT	2.1	3.5	1.0	0.0	43.0	3,435

Aid stuff

Overall OECD aid

total_oecd_aid <- df_country_aid_laws %>%
  summarise(total = dollar(sum(total_oda))) %>% 
  pull(total)

OECD members donated $4,333,135,732,679 between 1990 and 2013.

plot_aid <- df_country_aid_laws %>%
  filter(year <= 2013) %>% 
  group_by(year) %>%
  summarise(total = sum(total_oda)) %>% 
  mutate(fake_facet_title = "ODA commitments (all OECD)")

fig_oecd_aid <- ggplot(plot_aid, aes(x = year, y = (total / 1000000000))) + 
  geom_line(size = 0.5) +
  labs(x = NULL, y = "Billions of USD") +
  scale_y_continuous(labels = dollar) +
  coord_cartesian(xlim = c(1990, 2015)) +
  theme_donors() + 
  facet_wrap(vars(fake_facet_title))
fig_oecd_aid

Proportion of contentious vs. noncontentious aid

df_donor %>%
  filter(year >= 1990) %>%
  group_by(purpose_contentiousness) %>%
  summarise(Total = sum(oda)) %>%
  ungroup() %>% 
  mutate(Percent = Total / sum(Total)) %>%
  mutate(Total = dollar(Total), 
         Percent = percent_format(accuracy = 0.1)(Percent)) %>% 
  rename(`Contentiousness` = purpose_contentiousness) %>% 
  kbl() %>% 
  kable_styling()

Contentiousness	Total	Percent
High	$267,373,936,986	6.2%
Low	$4,071,293,867,271	93.8%

plot_aid_contentiousness <- df_donor %>%
  filter(year >= 1990) %>%
  group_by(year, purpose_contentiousness) %>%
  summarise(total = sum(oda)) %>%
  mutate(perc = total / sum(total)) %>%
  filter(purpose_contentiousness == "High") %>% 
  mutate(fake_facet_title = "Contentious aid (all OECD)")

fig_oecd_contention <- ggplot(plot_aid_contentiousness, aes(x = year, y = perc)) + 
  geom_line(size = 0.5) +
  scale_y_continuous(labels = percent_format(accuracy = 1)) + 
  labs(x = NULL, y = "Percent") +
  coord_cartesian(xlim = c(1990, 2015)) +
  theme_donors() +
  facet_wrap(vars(fake_facet_title))
fig_oecd_contention

USAID aid

total_us_aid <- df_country_aid_laws %>%
  summarise(total = dollar(sum(oda_us))) %>% 
  pull(total)

total_us_aid_post_2000 <- df_country_aid_laws %>%
  filter(year > 1999) %>%
  summarise(total = dollar(sum(oda_us))) %>% 
  pull(total)

The US donated $389,773,308,222 between 1990 and 2014 and $266,150,720,203 between 2000 and 2014.

plot_aid_us <- df_country_aid_laws %>%
  filter(year >= 1990) %>% 
  group_by(year) %>%
  summarise(total = sum(oda_us)) %>% 
  mutate(fake_facet_title = "ODA commitments (USAID only)")

fig_us_aid <- ggplot(plot_aid_us, aes(x = year, y = (total / 1000000000))) + 
  geom_line(size = 0.5) +
  expand_limits(y = 0) +
  labs(x = NULL, y = "Billions of USD") +
  scale_y_continuous(labels = dollar) +
  coord_cartesian(xlim = c(1990, 2015)) +
  theme_donors() +
  facet_wrap(vars(fake_facet_title))
fig_us_aid

Proportion of US aid to types of NGOs

Total amounts over time:

usaid_total <- df_country_aid_laws %>% 
  summarise(total = sum(oda_us)) %>% pull(total)

channels_nice <- tribble(
  ~channel,         ~channel_clean,
  "oda_us_ngo_dom", "Domestic NGOs",
  "oda_us_ngo_int", "International NGOs",
  "oda_us_ngo_us",  "US-based NGOs"
)

df_country_aid_laws %>%
  pivot_longer(names_to = "channel", values_to = "total_oda_us",
               c(oda_us_ngo_dom, oda_us_ngo_int, oda_us_ngo_us)) %>% 
  group_by(channel) %>%
  summarise(total = sum(total_oda_us)) %>%
  ungroup() %>% 
  left_join(channels_nice, by = "channel") %>% 
  mutate(perc = total / usaid_total) %>%
  mutate(total = dollar(total), perc = percent(perc)) %>% 
  select(Channel = channel_clean, Total = total, Percent = perc) %>% 
  kbl() %>% 
  kable_styling()

Channel	Total	Percent
Domestic NGOs	$7,057,013,854	1.8%
International NGOs	$14,412,538,155	3.7%
US-based NGOs	$31,189,618,294	8.0%

The US clearly favors US-based NGOs or international NGOs over domestic NGOs.

usaid_total_yearly <- df_country_aid_laws %>%
  group_by(year) %>%
  summarise(annual_total = sum(oda_us)) %>% 
  mutate(fake_facet_title = "USAID ODA channeled through NGOs")

plot_usaid_channel <- df_country_aid_laws %>%
  filter(year >= 1990) %>% 
  pivot_longer(names_to = "channel", values_to = "total_oda_us", 
               c(oda_us_ngo_dom, oda_us_ngo_int, oda_us_ngo_us)) %>%
  group_by(year, channel) %>%
  summarise(total = sum(total_oda_us)) %>%
  left_join(usaid_total_yearly, by = "year") %>%
  mutate(perc = total / annual_total) %>%
  left_join(channels_nice, by = "channel")

fig_usaid_channel <- ggplot(plot_usaid_channel, 
                            aes(x = year, y = perc, colour = channel_clean)) + 
  geom_line(size = 0.5) +
  scale_y_continuous(labels = percent_format(accuracy = 1)) + 
  scale_colour_manual(values = channel_colors) +
  labs(x = NULL, y = "Percent") +
  guides(colour = guide_legend(title = NULL, reverse = TRUE, nrow = 2)) +
  coord_cartesian(xlim = c(1990, 2015)) +
  theme_donors() + 
  facet_wrap(vars(fake_facet_title))
fig_usaid_channel

USAID data categorizes all aid as government-channeled before 2000 because of some quirk in the data.

plot_usaid_channels_all <- df_donor_us %>%
  group_by(year, channel_subcategory_name) %>%
  summarise(total = sum(oda_us_2011)) %>%
  mutate(perc = total / sum(total)) %>%
  mutate(channel = ifelse(str_detect(channel_subcategory_name, "NGO"), "NGO", "Other"))

ggplot(plot_usaid_channels_all, 
       aes(x = year, y = perc, colour = channel_subcategory_name)) + 
  geom_line(size = 0.5) +
  scale_y_continuous(labels = percent_format(accuracy = 1)) + 
  labs(x = NULL, y = "Percent of US aid given through different channels") +
  guides(colour = guide_legend(title = NULL, ncol = 2)) +
  theme_donors()

So we just look at aid after 2000.

fig_usaid_channel_trimmed <- fig_usaid_channel +
  coord_cartesian(xlim = c(2000, 2015))
fig_usaid_channel_trimmed

All DV figures combined

fig_dvs <- (fig_oecd_aid + fig_oecd_contention) / 
  (fig_us_aid + fig_usaid_channel_trimmed) &
  theme_donors(base_size = 10) +
  theme(legend.text = element_text(size = rel(0.6)),
        axis.title.y = element_text(margin = margin(r = 3)),
        legend.box.margin = margin(t = -0.5, unit = "lines"))

fig_dvs

ggsave(here("analysis", "output", "fig-dvs.pdf"), fig_dvs,
       width = 6.5, height = 4.75, device = cairo_pdf)
ggsave(here("analysis", "output", "fig-dvs.png"), fig_dvs,
       width = 6.5, height = 4.75, dpi = 300, type = "cairo")

Legal restrictions on NGOs

DCJW / Chaudhry indexes

dcjw_indexes <- read_csv(here("data", "manual_data", "dcjw_index.csv")) %>%
  mutate(Laws = str_replace_all(Laws, "\n", "\n\n"))

caption <- "Description of indexes of NGO barriers {#tbl:ngo-barriers-index}"
dcjw_index_table <- pandoc.table.return(dcjw_indexes, keep.line.breaks = TRUE, 
                                        style = "grid", justify = "lll", caption = caption)

cat(dcjw_index_table)

Description of indexes of NGO barriers {#tbl:ngo-barriers-index}

Index	Maximum value	Laws
Barriers to entry	3	How burdensome is registration? (Not burdensome = 0; Burdensome = 1) In law, can an NGO appeal if denied registration? (Yes = 0; No = 1) Are barriers to entry different for NGOs receiving foreign funds? (Less burdensome = -1; Same = 0; More burdensome = 1)
Barriers to funding	5	Do NGOs need prior approval from the government to receive foreign funding? (Yes = 1; No = 0) Are NGOs required to channel foreign funding through state-owned banks or government ministries? (Yes = 1; No = 0) Are any additional restrictions on foreign support in place? (Yes = 1; No = 0) Are all NGOs prohibited from receiving foreign funds? (No = 0; Partially = 0.5; Yes = 1) Is a category of NGOs prohibited from receiving foreign funds? (No = 0; Partially = 0.5; Yes = 1)
Barriers to advocacy	2	Does the law restrict NGOs from engaging in political activities? (No = 0; Partially = 0.5; Yes = 1) Are restrictions on political activities different for NGOs receiving foreign funds? (Less restrictive = -1; Same = 0; More restrictive = 1)
Total barriers	10	—

cat(dcjw_index_table, file = here("analysis", "output", "tbl-ngo-barriers-index.md"))

NGO barriers over time

# barriers_total = advocacy + entry + funding
dcjw_questions <- read_csv(here("data", "manual_data", "dcjw_questions.csv")) %>% 
  filter(!ignore_in_index) %>% 
  select(barrier_group = barrier_display, barrier = question_clean, 
         barrier_display = question_display) %>% 
  mutate(barrier_group = paste0("Barriers to ", str_to_lower(barrier_group)))

df_barriers <- df_country_aid_laws %>% 
  group_by(gwcode, year) %>% 
  summarize(across(one_of(dcjw_questions$barrier), ~. > 0)) %>% 
  group_by(year) %>% 
  summarize(across(-gwcode, ~sum(.))) %>% 
  ungroup() %>% 
  pivot_longer(names_to = "barrier", values_to = "value", -year) %>% 
  left_join(dcjw_questions, by = "barrier") %>% 
  mutate(barrier_display = str_replace(barrier_display, "XXX", "\n")) %>% 
  arrange(desc(value)) %>% 
  mutate(barrier_display = fct_inorder(barrier_display, ordered = TRUE))

dcjw_entry_plot <- ggplot(filter(df_barriers, 
                                 barrier_group == "Barriers to entry"), 
                          aes(x = year, y = value, 
                              color = barrier_display,
                              linetype = barrier_display)) +
  geom_line(size = 0.5) +
  expand_limits(y = c(0, 90)) +
  scale_y_continuous(sec.axis = sec_axis(~ . / num_countries,
                                         labels = percent_format(accuracy = 1)),
                     expand = c(0, 0)) +
  scale_colour_manual(values = c("black", "grey80", "grey50"), name = NULL) +
  scale_linetype_manual(values = c("solid", "solid", "21"), name = NULL) +
  coord_cartesian(xlim = c(1990, 2015), ylim = c(0, 90)) +
  guides(color = guide_legend(nrow = 2)) +
  labs(x = NULL, y = "Number of countries") +
  theme_donors(10) + 
  theme(legend.justification = "left") +
  facet_wrap(vars(barrier_group))

dcjw_funding_plot <- ggplot(filter(df_barriers, 
                                   barrier_group == "Barriers to funding"), 
                            aes(x = year, y = value, 
                                color = barrier_display,
                                linetype = barrier_display)) +
  geom_line(size = 0.5) +
  scale_y_continuous(sec.axis = sec_axis(~ . / num_countries,
                                         labels = percent_format(accuracy = 1)),
                     expand = c(0, 0)) +
  scale_colour_manual(values = c("black", "grey80", "grey50", "black", "grey80"), name = NULL) +
  scale_linetype_manual(values = c("solid", "solid", "solid", "21", "21"), name = NULL) +
  coord_cartesian(xlim = c(1990, 2015), ylim = c(0, 40)) +
  guides(color = guide_legend(nrow = 3),
         linetype = guide_legend(nrow = 3)) +
  labs(x = NULL, y = "Number of countries") +
  theme_donors(10) + 
  theme(legend.justification = "left") +
  facet_wrap(vars(barrier_group))

dcjw_advocacy_plot <- ggplot(filter(df_barriers, 
                                    barrier_group == "Barriers to advocacy"), 
                            aes(x = year, y = value, 
                                color = barrier_display)) +
  geom_line(size = 0.5) +
  scale_y_continuous(sec.axis = sec_axis(~ . / num_countries,
                                         labels = percent_format(accuracy = 1)),
                     expand = c(0, 0)) +
  scale_colour_manual(values = c("black", "grey80"), name = NULL) +
  coord_cartesian(xlim = c(1990, 2015), ylim = c(0, 60)) +
  guides(color = guide_legend(nrow = 1)) +
  labs(x = NULL, y = "Number of countries") +
  theme_donors(10) + 
  theme(legend.justification = "left") +
  facet_wrap(vars(barrier_group))

df_ccsi_plot <- df_country_aid %>%
  left_join(df_autocracies, by = "gwcode") %>%
  group_by(year, autocracy) %>%
  nest() %>% 
  mutate(cis = data %>% map(~ mean_cl_normal(.$v2xcs_ccsi))) %>% 
  unnest(cis) %>% 
  ungroup() %>% 
  mutate(fake_facet_title = "Core civil society index",
         autocracy = factor(autocracy, 
                            labels = c("Democracies (Regimes of the World > 4)",
                                       "Non-democracies (Regimes of the World ≤ 4)"), 
                            ordered = TRUE))

fig_ccsi <- ggplot(df_ccsi_plot, aes(x = year, y = y)) +
  geom_ribbon(aes(ymin = ymin, ymax = ymax, fill = autocracy), alpha = 0.2) +
  geom_line(aes(color = autocracy), size = 0.5) +
  annotate(geom = "text", x = 2015, y = 0.1, hjust = "right", size = pts(7),
           label = "Higher values = more open civil society") +
  scale_colour_manual(values = c("black", "grey75"), name = NULL) +
  scale_fill_manual(values = c("black", "grey75"), name = NULL) +
  scale_linetype_manual(values = c("solid", "solid", "21")) +
  coord_cartesian(xlim = c(1990, 2015), ylim = c(0, 1)) +
  guides(color = guide_legend(nrow = 2)) +
  labs(y = "Average", x = NULL) +
  theme_donors(10) +
  theme(legend.justification = "left") +
  facet_wrap(vars(fake_facet_title))

barriers_summary <- 
  ((dcjw_entry_plot + dcjw_funding_plot) / 
     (dcjw_advocacy_plot + fig_ccsi)) &
  theme(legend.text = element_text(size = rel(0.6)),
        axis.title.y = element_text(margin = margin(r = 3)),
        legend.box.margin = margin(t = -0.5, unit = "lines"))

barriers_summary

ggsave(here("analysis", "output", "fig-barriers-summary.pdf"), barriers_summary,
       width = 6.5, height = 6, device = cairo_pdf)
ggsave(here("analysis", "output", "fig-barriers-summary.png"), barriers_summary,
       width = 6.5, height = 6, dpi = 300, type = "cairo")

Compulsory vs. burdensome registration

Laws requiring NGO registration aren’t necessarily a sign of oppression—even the US requires that nonprofits that earn above a certain threshold register as 501(c)(3) organizations. Though the figure below shows that compulsory regulation have increased over time, actual restriction has occurred too. Burdensome registration is not just another standard layer of bureaucracy.

df_regulation <- df_country_aid_laws %>%
  left_join(select(df_autocracies, gwcode, autocracy), by = "gwcode") %>%
  group_by(year, autocracy) %>%
  summarise(`Registration required` = sum(ngo_register) / n(),
            `Registration burdensome` = sum(ngo_register_burden) / n()) %>%
  ungroup() %>% 
  pivot_longer(names_to = "type_of_law", values_to = "value", -c(year, autocracy)) %>% 
  mutate(autocracy = 
           factor(autocracy, levels = c(TRUE, FALSE),
                  labels = c("Non-democracies", "Democracies")))

fig_regulation_burden <- ggplot(df_regulation, 
                                aes(x = year, y = value, color = type_of_law)) +
  geom_line(size = 0.5) +
  scale_y_continuous(labels = percent_format(accuracy = 1)) +
  scale_x_continuous(expand = c(0, 0)) +
  coord_cartesian(ylim = c(0, 0.7), xlim = c(1990, 2014)) +
  guides(colour = guide_legend(title = NULL)) +
  labs(x = NULL, y = "Proportion of countries\nwith regulation") +
  theme_donors(10) +
  facet_wrap(vars(autocracy))

fig_regulation_burden

ggsave(here("analysis", "output", "fig-regulation-burden.pdf"), fig_dvs,
       width = 6, height = 2.5, device = cairo_pdf)
ggsave(here("analysis", "output", "fig-regulation-burden.png"), fig_dvs,
       width = 6, height = 2.5, dpi = 300, type = "cairo")

Aid

Aid over time, by donor type

aid_donor_type_time <- df_donor %>%
  group_by(year, donor_type_collapsed) %>%
  summarise(total_aid = sum(oda, na.rm = TRUE))

ggplot(aid_donor_type_time, aes(x = year, y = total_aid / 1000000000,
                                colour = donor_type_collapsed)) +
  geom_line(size = 0.5) +
  labs(x = NULL, y = "Billions of USD",
       caption = "Source: OECD and AidData. 2011 dollars.") +
  guides(colour = guide_legend(title = NULL)) +
  scale_y_continuous(labels = dollar) + 
  theme_donors()

Aid over time, by contentiousness

aid_contention_time <- df_donor %>%
  group_by(year, purpose_contentiousness) %>%
  summarise(total_aid = sum(oda, na.rm = TRUE))

ggplot(aid_contention_time, aes(x = year, y = total_aid / 1000000000,
                                colour = purpose_contentiousness)) +
  geom_line(size = 0.5) +
  labs(x = NULL, y = "Billions of USD",
       caption = "Source: OECD and AidData. 2011 dollars.") +
  guides(colour = guide_legend(title = NULL)) +
  scale_y_continuous(labels = dollar) + 
  theme_donors()

Restrictions and aid

inv_logit <- function(f, a) {
  a <- (1 - 2 * a)
  (a * (1 + exp(f)) + (exp(f) - 1)) / (2 * a * (1 + exp(f)))
}

dvs_clean_names <- tribble(
  ~barrier, ~barrier_clean,
  "barriers_total", "All barriers",
  "advocacy", "Barriers to advocacy",
  "entry", "Barriers to entry",
  "funding", "Barriers to funding"
)

ivs_clean_names <- tribble(
  ~variable, ~variable_clean, ~hypothesis,
  "total_oda_lead1", "Total ODA", "H1",
  "prop_contentious_lead1", "Contentious aid", "H2",
  "prop_ngo_dom_lead1", "Aid to domestic NGOs", "H3",
  "prop_ngo_foreign_lead1", "Aid to foreign NGOs", "H3"
)

Restrictions and ODA (H₁)

df_plot_barriers_oda <- df_country_aid_laws %>% 
  select(year, gwcode, country, total_oda_lead1,
         one_of(dvs_clean_names$barrier)) %>% 
  pivot_longer(names_to = "barrier", values_to = "value", 
               one_of(dvs_clean_names$barrier)) %>% 
  left_join(dvs_clean_names, by = "barrier") %>% 
  mutate(barrier_clean = fct_inorder(barrier_clean, ordered = TRUE))

ggplot(df_plot_barriers_oda, 
       aes(x = value, y = total_oda_lead1, color = barrier_clean)) +
  geom_point(alpha = 0.5) +
  stat_smooth(method = "lm") +
  stat_smooth(data = filter(df_plot_barriers_oda, 
                            total_oda_lead1 > 10000000000), 
              method = "lm", linetype = "21") +
  scale_y_continuous(labels = dollar) +
  guides(color = FALSE) +
  labs(x = "Number of legal barriers", y = "Total ODA in next year",
       title = "Total ODA in next year",
       subtitle = "Dotted lines show trends when omitting observations\nwith less than $10,000,000,000 in ODA") +
  theme_donors(10) +
  theme(strip.text.x = element_text(margin = margin(t = 1, b = 1))) +
  facet_wrap(vars(barrier_clean), scales = "free_x", nrow = 2)

Restrictions and contentiousness (H₂)

df_plot_barriers_contention <- df_country_aid_laws %>% 
  select(year, gwcode, country, prop_contentious_lead1,
         one_of(dvs_clean_names$barrier))  %>% 
  pivot_longer(names_to = "barrier", values_to = "value", 
               one_of(dvs_clean_names$barrier)) %>% 
  left_join(dvs_clean_names, by = "barrier") %>% 
  mutate(barrier_clean = fct_inorder(barrier_clean, ordered = TRUE))

ggplot(df_plot_barriers_contention, 
       aes(x = value, y = prop_contentious_lead1, color = barrier_clean)) +
  geom_point(alpha = 0.5) +
  stat_smooth(method = "lm") +
  stat_smooth(data = filter(df_plot_barriers_contention, 
                            prop_contentious_lead1 > 0.05), 
              method = "lm", linetype = "21") +
  scale_y_continuous(labels = percent) +
  guides(color = FALSE) +
  labs(x = "Number of legal barriers", 
       y = "Proportion of contentious aid in next year",
       title = "Proportion of contentious aid in next year",
       subtitle = "Dotted lines show trends when omitting observations\nwith less than 5% contentious aid") +
  theme_donors(10) +
  theme(strip.text.x = element_text(margin = margin(t = 1, b = 1))) +
  facet_wrap(vars(barrier_clean), scales = "free_x", nrow = 2)

Restrictions and NGOs (H₃)

df_plot_barriers_ngos <- df_country_aid_laws %>% 
  select(year, gwcode, country, 
         prop_ngo_dom_lead1, prop_ngo_foreign_lead1,
         one_of(dvs_clean_names$barrier)) %>% 
  pivot_longer(names_to = "barrier", values_to = "value", 
               one_of(dvs_clean_names$barrier)) %>% 
  pivot_longer(names_to = "ngo_type", values_to = "prop_ngo", 
               c(prop_ngo_dom_lead1, prop_ngo_foreign_lead1)) %>% 
  left_join(dvs_clean_names, by = "barrier") %>% 
  left_join(ivs_clean_names, by = c("ngo_type" = "variable")) %>% 
  mutate(barrier_clean = fct_inorder(barrier_clean, ordered = TRUE))

ggplot(df_plot_barriers_ngos, 
       aes(x = value, y = prop_ngo, color = barrier_clean)) +
  geom_point(alpha = 0.5) +
  stat_smooth(method = "lm") +
  stat_smooth(data = filter(df_plot_barriers_ngos, 
                            prop_ngo > 0.05), 
              method = "lm", linetype = "21") +
  scale_y_continuous(labels = percent) +
  guides(color = FALSE) +
  labs(x = "Number of legal barriers", 
       y = "Proportion of aid to NGOs in next year",
       title = "Proportion of aid channeled to types of NGOs in next year",
       subtitle = "Dotted lines show trends when omitting observations\nwith less than 5% aid to NGOs") +
  coord_cartesian(ylim = c(0, 1)) +
  theme_donors(10) +
  theme(strip.text.x = element_text(margin = margin(t = 1, b = 1))) +
  facet_wrap(vars(variable_clean, barrier_clean), scales = "free_x", ncol = 4)

CCSI and all DVs (all hypotheses)

df_plot_ccsi_ngos <- df_country_aid_laws %>% 
  select(year, gwcode, country, 
         one_of(ivs_clean_names$variable), v2xcs_ccsi) %>% 
  pivot_longer(names_to = "variable", values_to = "value", 
               c(one_of(ivs_clean_names$variable))) %>% 
  left_join(ivs_clean_names, by = "variable") %>% 
  mutate(hypothesis_clean = paste0(hypothesis, ": ", variable_clean)) %>% 
  arrange(hypothesis_clean) %>% 
  mutate(hypothesis_clean = fct_inorder(hypothesis_clean, ordered = TRUE))
  
ggplot(df_plot_ccsi_ngos, 
       aes(x = v2xcs_ccsi, y = value, color = hypothesis)) +
  geom_point(alpha = 0.25) +
  scale_color_viridis_d(option = "plasma", end = 0.9) +
  guides(color = FALSE) +
  labs(x = "Civil society index", 
       y = "Variable value in next year",
       title = "Core civil society index") +
  theme_donors() +
  facet_wrap(vars(hypothesis_clean), scales = "free_y")

CIVICUS restrictions

civicus <- read_csv(here("data", "raw_data", "Civicus", "civicus_monitor_2017.csv"),
                    na = "Null") %>%
  mutate(Population = as.double(Population),  # Integers can't handle world population
         Rating = factor(Rating, levels = c("Open", "Narrowed", "Obstructed", 
                                            "Repressed", "Closed"), 
                         ordered = TRUE),
         iso3 = countrycode(Country, "country.name", "iso3c",
                            custom_match = c("Kosovo" = "XKK",
                                             "Micronesia" = "FSM")))

# Use the Robinson map projection
projection <- "ESRI:54030"

world_map <- read_sf(here("data", "raw_data", "ne_110m_admin_0_countries",
                          "ne_110m_admin_0_countries.shp")) %>% 
  filter(ISO_A3 != "ATA")

civicus %>% count(Rating) %>% pandoc.table()

Rating	n
Open	26
Narrowed	64
Obstructed	50
Repressed	35
Closed	20

map_with_civicus <- world_map %>% 
  # Fix some Natural Earth ISO weirdness
  mutate(ISO_A3 = ifelse(ISO_A3 == "-99", as.character(ISO_A3_EH), as.character(ISO_A3))) %>% 
  mutate(ISO_A3 = case_when(
    .$ISO_A3 == "GRL" ~ "DNK",
    .$NAME == "Norway" ~ "NOR",
    .$NAME == "Kosovo" ~ "XKK",
    TRUE ~ ISO_A3
  )) %>% 
  left_join(select(civicus, iso3, Rating), by = c("ISO_A3" = "iso3"))

plot_civicus_map <- ggplot() +
  geom_sf(data = map_with_civicus, aes(fill = Rating), size = 0.15, color = "black") +
  coord_sf(crs = st_crs(projection)) +
  scale_fill_manual(values = c("grey90", "grey70", "grey45",
                               "grey20", "black"),
                    na.translate = FALSE, name = "Civic space") +
  theme_donors_map() + 
  theme(legend.key.size = unit(0.7, "lines"))

plot_civicus_map

ggsave(here("analysis", "output", "fig-civicus-map.pdf"), plot_civicus_map,
       width = 5.5, height = 3, device = cairo_pdf)
ggsave(here("analysis", "output", "fig-civicus-map.png"), plot_civicus_map,
       width = 5.5, height = 3, dpi = 300, type = "cairo")

List of countries included in models

matrix_from_vector <- function(x, ncol) {
  n_balanced <- ceiling(length(x) / ncol) * ncol
  matrix(c(x, rep(NA, n_balanced - length(x))), ncol = ncol)
}

all_countries <- df_country_aid_laws %>% 
  distinct(country) %>% 
  arrange(country) %>% 
  pull(country) 

caption <- paste0("All countries included in models (N = ", 
                  length(all_countries),
                  ") {#tbl:countries}")

ncol_countries <- 4

tbl_countries <- all_countries %>% 
  matrix_from_vector(ncol = ncol_countries) %>% 
  pandoc.table.return(justify = paste0(rep("l", ncol_countries), collapse = ""), 
                      caption = caption, missing = "")

cat(tbl_countries)

All countries included in models (N = 142) {#tbl:countries}

Afghanistan	Ecuador	Liberia	Saudi Arabia
Albania	Egypt	Libya	Senegal
Algeria	El Salvador	Lithuania	Serbia
Angola	Equatorial Guinea	Madagascar	Sierra Leone
Argentina	Eritrea	Malawi	Singapore
Armenia	Estonia	Malaysia	Slovakia
Azerbaijan	Eswatini	Mali	Slovenia
Bahrain	Ethiopia	Mauritania	Solomon Islands
Bangladesh	Fiji	Mauritius	Somalia
Belarus	Gabon	Mexico	South Africa
Benin	Gambia	Moldova	South Korea
Bhutan	Georgia	Mongolia	South Sudan
Bolivia	Ghana	Montenegro	Sri Lanka
Bosnia & Herzegovina	Guatemala	Morocco	Sudan
Botswana	Guinea	Mozambique	Syria
Brazil	Guinea-Bissau	Myanmar (Burma)	Tajikistan
Bulgaria	Guyana	Namibia	Tanzania
Burkina Faso	Haiti	Nepal	Thailand
Burundi	Honduras	Nicaragua	Timor-Leste
Cambodia	Hungary	Niger	Togo
Cameroon	India	Nigeria	Trinidad & Tobago
Central African Republic	Indonesia	North Korea	Tunisia
Chile	Iran	North Macedonia	Turkey
China	Iraq	Oman	Turkmenistan
Colombia	Israel	Pakistan	Uganda
Comoros	Jamaica	Panama	Ukraine
Congo - Brazzaville	Jordan	Papua New Guinea	United Arab Emirates
Congo - Kinshasa	Kazakhstan	Paraguay	Uruguay
Costa Rica	Kenya	Peru	Uzbekistan
Côte d’Ivoire	Kosovo	Philippines	Venezuela
Croatia	Kuwait	Poland	Vietnam
Cuba	Kyrgyzstan	Portugal	Yemen
Cyprus	Laos	Qatar	Zambia
Czechia	Latvia	Romania	Zimbabwe
Djibouti	Lebanon	Russia
Dominican Republic	Lesotho	Rwanda

cat(tbl_countries, file = here("analysis", "output", "tbl-countries.md"))

Copyright © 2017–2021 Suparna Chaudhry and Andrew Heiss. Licensed under the MIT License. Code available on GitHub.