The package holds a dataset of every veto cast in the UN Security
Council from 1946 to October 2026. This vignette shows what is in it,
how to count vetoes, how to use the roll calls, and how to join the data
to country-year and conflict-year panels.
The data
There are three tables. vetoes has one row per vetoed
draft. A draft vetoed by two members, such as China and Russia, is one
row; this follows the UN’s own count.
nrow(vetoes)
#> [1] 274
vetoes[vetoes$year == 2026, c("veto_id", "date", "vetoed_by", "topic")]
#> veto_id date vetoed_by
#> 272 272 2026-04-07 China; Russia
#> 273 273 2026-09-17 China; Russia
#> 274 274 2026-10-30 Russia
#> topic
#> 272 Iran: maritime security in the Strait of Hormuz
#> 273 Iran: sanctions Panel of Experts renewal
#> 274 Western Sahara: MINURSO renewal
Most vetoes fell on draft resolutions. In the early years the Council
also voted on parts of drafts, on amendments, on membership applications
and on whether a question was procedural. vetoed_item
records which.
table(vetoes$vetoed_item)
#>
#> amendment application draft oral proposal
#> 14 10 217 2
#> part of draft procedural question
#> 28 3
veto_votes holds the roll call: one row per Council
member per draft, 11 members until 1965 and 15 since. Here is the vote
on the 2024 draft on weapons in outer space.
subset(veto_votes, draft_symbol == "S/2024/302", c(member, permanent, vote, veto))
#> member permanent vote veto
#> 3488 China TRUE abstain FALSE
#> 3489 France TRUE yes FALSE
#> 3490 Russia TRUE no TRUE
#> 3491 United Kingdom TRUE yes FALSE
#> 3492 United States TRUE yes FALSE
#> 3493 Algeria FALSE yes FALSE
#> 3494 Ecuador FALSE yes FALSE
#> 3495 Guyana FALSE yes FALSE
#> 3496 Japan FALSE yes FALSE
#> 3497 Malta FALSE yes FALSE
#> 3498 Mozambique FALSE yes FALSE
#> 3499 Sierra Leone FALSE yes FALSE
#> 3500 Slovenia FALSE yes FALSE
#> 3501 South Korea FALSE yes FALSE
#> 3502 Switzerland FALSE yes FALSE
veto_countries has one row per country a draft concerns,
with ISO3, Correlates of War and Gleditsch–Ward codes as they stood on
the date of the vote.
head(veto_countries[c("date", "country", "iso3c", "cown", "gwno", "vetoed_by")])
#> date country iso3c cown gwno vetoed_by
#> 1 1946-02-16 France FRA 220 220 Russia
#> 2 1946-02-16 United Kingdom GBR 200 200 Russia
#> 3 1946-02-16 Lebanon LBN 660 660 Russia
#> 4 1946-02-16 Syria SYR 652 652 Russia
#> 5 1946-06-18 Spain ESP 230 230 Russia
#> 6 1946-06-18 Spain ESP 230 230 Russia
Each draft also carries a short account of what it would have done,
the reason the vetoing member gave, and what followed.
cat(vetoes$context[vetoes$veto_id == 254])
#> The United States and Albania tabled a draft that would have condemned Russia's referendums in four occupied Ukrainian regions; the vote came on the day Russia announced their annexation. Russia vetoed it, calling the draft a provocation and saying the votes reflected local will. The vote was 10 in favour, 1 against and 4 abstentions (Brazil, China, Gabon and India). On 12 October 2022 the General Assembly adopted resolution ES-11/4 at its emergency special session.
Counting vetoes
count_vetoes() counts drafts by year, or member vetoes
by member. A joint veto counts once in a yearly count and once for each
member in a count by member.
count_vetoes("member")
#> member n
#> 3 Russia 164
#> 5 US 95
#> 4 UK 32
#> 1 China 23
#> 2 France 18
count_vetoes("member", from = 1990)
#> member n
#> 3 Russia 44
#> 5 US 28
#> 1 China 20
#> 2 France 0
#> 4 UK 0
The plot has one square per member veto. It shows the Soviet vetoes
of the late 1940s and 1950s, the American vetoes of the 1970s and 1980s,
and the rise since 2011.
plot_vetoes(title = NULL)
By decade and member:
by_member <- count_vetoes(c("year", "member"))
by_member$decade <- by_member$year %/% 10 * 10
xtabs(n ~ decade + member, data = by_member)
#> member
#> decade China France Russia UK US
#> 1940 0 2 46 0 0
#> 1950 1 2 45 2 0
#> 1960 0 0 18 1 0
#> 1970 2 7 7 14 21
#> 1980 0 7 4 15 46
#> 1990 2 0 2 0 6
#> 2000 2 0 4 0 10
#> 2010 9 0 19 0 3
#> 2020 7 0 19 0 9
What the vetoes were about
Region and issue area are coded for every draft, one code each. The
issue area follows a fixed order of precedence set out in the
codebook.
sort(table(vetoes$issue_area), decreasing = TRUE)
#>
#> UN membership and appointments Use of force and ceasefires
#> 63 61
#> Occupation, territory and settlements Decolonization and self-determination
#> 26 22
#> Accountability and investigations Peace operations
#> 21 21
#> Political crises and human rights Sanctions
#> 21 16
#> Humanitarian access and protection Non-proliferation and disarmament
#> 12 9
#> Other thematic
#> 2
with(vetoes, table(issue_area, period = ifelse(year < 1990, "1946-89", "1990-2026")))
#> period
#> issue_area 1946-89 1990-2026
#> Accountability and investigations 13 8
#> Decolonization and self-determination 22 0
#> Humanitarian access and protection 2 10
#> Non-proliferation and disarmament 5 4
#> Occupation, territory and settlements 18 8
#> Other thematic 0 2
#> Peace operations 13 8
#> Political crises and human rights 15 6
#> Sanctions 11 5
#> UN membership and appointments 61 2
#> Use of force and ceasefires 40 21
Using the roll calls
A veto is cast by a permanent member, but elected members sometimes
vote with it. These are the elected members that most often voted
against a draft that a permanent member vetoed:
with_veto <- subset(veto_votes, vote == "no" & !permanent)
head(sort(table(with_veto$member), decreasing = TRUE), 10)
#>
#> Poland Ukrainian SSR Bolivia
#> 23 22 7
#> Canada West Germany Ceylon
#> 5 5 4
#> Czechoslovakia South Africa United Arab Republic
#> 4 3 3
#> East Germany
#> 2
The number of votes in favour of a vetoed draft gives a rough measure
of how isolated the vetoing member was. Since 1990:
recent <- subset(vetoes, year >= 1990)
table(recent$vote_yes)
#>
#> 9 10 11 12 13 14
#> 7 10 16 8 19 14
round(tapply(recent$vote_yes, recent$vetoed_by, mean), 1)
#> China China; Russia Russia US
#> 13.5 11.1 11.8 12.3
Joining country-year data
veto_country_year() collapses the country table to one
row per country and year. Join it to any panel keyed on COW or
Gleditsch–Ward codes, and set years without a veto to zero.
cy <- veto_country_year()
head(cy)
#> iso3c cown gwno country year n_vetoes veto_china veto_france
#> 175 AFG 700 700 Afghanistan 1980 1 0 0
#> 9 ALB 339 339 Albania 1946 1 0 0
#> 14 ALB 339 339 Albania 1947 6 0 0
#> 164 AGO 540 540 Angola 1976 1 0 0
#> 169 AGO 540 540 Angola 1977 1 0 1
#> 183 AGO 540 540 Angola 1981 1 0 0
#> veto_russia veto_uk veto_us
#> 175 1 0 0
#> 9 1 0 0
#> 14 6 0 0
#> 164 0 0 1
#> 169 0 1 1
#> 183 0 0 1
panel <- expand.grid(cown = c(652, 666, 369), year = 2010:2024)
panel <- merge(panel, cy[c("cown", "year", "n_vetoes")], by = c("cown", "year"), all.x = TRUE)
panel$n_vetoes[is.na(panel$n_vetoes)] <- 0
xtabs(n_vetoes ~ year + cown, data = panel)
#> cown
#> year 369 652 666
#> 2010 0 0 0
#> 2011 0 1 1
#> 2012 0 2 0
#> 2013 0 0 0
#> 2014 1 1 0
#> 2015 1 0 0
#> 2016 0 2 0
#> 2017 0 5 1
#> 2018 0 1 1
#> 2019 0 2 0
#> 2020 0 2 0
#> 2021 0 0 0
#> 2022 2 1 0
#> 2023 0 1 4
#> 2024 0 0 3
Palestine, Western Sahara and a few historical entities have no COW
or Gleditsch–Ward code. Join on iso3c if you need them.
Joining UCDP conflict-year data
Drafts about an armed conflict have a UCDP conflict identifier,
ucdp_conflict_id. The code below joins vetoes to the
UCDP/PRIO Armed Conflict Dataset, which you download from UCDP.
tmp <- tempfile(fileext = ".zip")
download.file("https://ucdp.uu.se/downloads/ucdpprio/ucdp-prio-acd-261-csv.zip", tmp)
acd <- read.csv(unz(tmp, "UcdpPrioConflict_v26_1.csv"))
v <- subset(vetoes, !is.na(ucdp_conflict_id))
v$n_vetoes <- 1
vy <- aggregate(n_vetoes ~ ucdp_conflict_id + year, data = v, FUN = sum)
names(vy)[1] <- "conflict_id"
acd_v <- merge(acd, vy, by = c("conflict_id", "year"), all.x = TRUE)
acd_v$n_vetoes[is.na(acd_v$n_vetoes)] <- 0
table(intensity = acd_v$intensity_level, veto = acd_v$n_vetoes > 0)
Two cautions. Drafts with ucdp_match == "past" concern a
mission or settlement after a conflict ended, such as the UN force in
Cyprus, so they have no active conflict-year to join. Vetoes after 2025
fall outside UCDP version 26.1.
The conflicts with the most vetoes:
head(sort(table(vetoes$ucdp_conflict_id), decreasing = TRUE))
#>
#> 234 299 298 235 336 203
#> 48 18 9 7 7 6
Joining UN documents
draft_symbol, meeting_record and
meeting_number follow UN style (S/2023/970,
S/PV.9499). They join to UN Digital Library voting records
and to resolution data that keep the symbol. For parts of drafts and
amendments, draft_symbol gives the document voted on and
draft_note says which part.
vetoes[vetoes$veto_id == 259, c("draft_symbol", "meeting_record", "meeting_number")]
#> draft_symbol meeting_record meeting_number
#> 259 S/2023/970 S/PV.9499 9499