How have Afghan government and allied security force casualties been counted in war death estimates?
Executive summary
Estimates of Afghan government and allied security-force deaths have been produced through a mix of official tallies, UN and NGO documentation, and academic research; these different streams yield divergent totals because they use different definitions (military vs. police vs. contractors), data sources, and adjustment methods [1] [2] [3]. Researchers such as the Brown University Costs of War project and independent scholars have combined reported figures, surveys and modeling to reach higher aggregate counts—while others have applied “reduction factors” to official Afghan claims or treated Afghan force deaths separately from coalition tallies, producing lower or more conservative estimates [4] [3] [5].
1. Official registers and Afghan government reporting: raw counts with political context
Afghan ministries and security institutions maintained casualty lists—military, police and sometimes paramilitary units—that were periodically published or leaked and subsequently incorporated into aggregate tallies, but these reports were neither standardized nor independently verifiable; some researchers explicitly down‑weighted Afghan government numbers, applying reduction factors as large as 75% when integrating them into wider estimates because of suspected double‑counting, incomplete records, or political inflation [4]. The coalition-era tracker “Coalition casualties in Afghanistan” notes large totals attributed to Afghan forces—reporting as many as 45,000 fatalities between 2014–2019 and an estimated ~69,095 total by August 2021—showing that official and secondary public tallies often treat Afghan security forces as a distinct allied casualty category [1].
2. UN, NGOs and systematic documentation: event‑based attribution
UNAMA and other UN actors documented civilian harm and attributed a portion of civilian deaths to pro‑government forces (PGFs) and Afghan Air Force operations, producing annual counts of civilian casualties caused by government operations and noting spikes when Afghan forces increased aerial strikes (e.g., UNAMA finding PGFs responsible for 25% of mid‑2021 civilian casualties and large numbers in 2020) —these products focus on civilian harm but also reveal patterns in government force lethality and risk that indirectly inform force‑death accounting [2]. NGOs and casualty‑recording networks seek “every casualty” databases and use field-based verification, but security constraints mean personnel inside the country often must withhold personal details and cannot maintain comprehensive force‑death registers [6].
3. Academic syntheses and the Costs of War approach: aggregated and modeled totals
Academic projects aggregate official counts, media reports, contractor lists and survey data to estimate broader totals; the Costs of War research and related academic papers explicitly include Afghan military and police deaths as allied losses, citing conservative totals “over 21,000” in some summaries and project‑level estimates of 66,000–69,000 Afghan military and police deaths in others—illustrating the gap between piecemeal official registers and modeled aggregations that attempt to capture the full allied toll [3] [7] [5]. These projects also fold in contractor fatalities and allied troop deaths, producing comprehensive human‑cost figures that treat Afghan security forces as a central component of allied casualties [8] [7].
4. Methodological problems: definitions, access, and double counting
Scholarly reviews stress that modern casualty counting is hampered by definitional inconsistency (who counts as “security forces”?), limited access for neutral monitors, and political incentives to under‑ or over‑report; passive observation, survey extrapolation, and administrative rosters each introduce distinct biases, and earlier attempts sometimes excluded or heavily discounted Afghan reports for lack of corroboration, producing wide uncertainty around both lower‑ and upper‑bound estimates [9] [4]. Independent compendia warned that contractor deaths and local police/militias were frequently omitted from official allied totals, and that survey‑based extrapolations cannot simply be ported between conflicts without adjusting methods [10] [9].
5. Political incentives and contested narratives
Governments and coalitions had incentives to emphasize or minimize certain categories: Western media and official trackers focused on coalition troop deaths, Afghan authorities had reasons to highlight losses to attract support or obscure desertions, and independent researchers at times discounted Afghan numbers as part of methodological conservatism—each stance reflects implicit agendas that shaped which deaths entered headline tallies versus buried footnotes [4] [3] [10]. Analysts therefore often present multiple totals rather than a single definitive figure, with dissenting estimates reflecting different tradeoffs between inclusiveness and verifiability [5] [11].
6. Bottom line — what can be said with confidence
Afghan government and allied security‑force casualties are counted through a patchwork: direct official lists, UN/NGO event verification, media and contractor records, and modeled academic aggregations; this patchwork produces wide but overlapping ranges (tens of thousands to perhaps ~70,000 Afghan security deaths in prominent estimates) because of differing inclusion rules and methodological adjustments, and any single published number must be read against those caveats and the source’s approach to Afghan‑origin data [1] [7] [3] [4]. Where sources diverge, the gap typically reflects whether researchers accepted Afghan government tallies at face value, adjusted them downward, or supplemented them with survey/model estimates to account for unrecorded deaths [4] [9] [5].