The South Caucasus refugee populations: scale?

The scale of refugee displacement in the South Caucasus in the years following Soviet collapse remains contested in the archival record, primarily because measurement happened amid active conflict, across borders between the Union and independent states, and with competing mandates among documenting agencies.

The International Committee of the Red Cross began operations in the region in late 1991, initially with a narrow focus on detainees and wounded. By mid-1992, as ethnic conflict intensified in Nagorno-Karabakh, the organization broadened its scope to track population movements. Early ICRC estimates, circulated in monthly summaries between June 1992 and December 1993, counted approximately 250,000 displaced persons across Azerbaijan and the Union's Russian Sovereign Republic (primarily refugee camps in Dagestan and northern Azerbaijan). These figures did not include internal displacement within the disputed Nagorno-Karabakh region itself, which ICRC staff could not access reliably.

Union statistical authorities maintained a separate accounting system, operated through the Central Statistical Office in Moscow and coordinated with republican counterparts. The Russian Sovereign Republic's administration of refugee reception in the North Caucasus generated systematic if irregular data. A March 1994 Union-wide survey, preserved in the Archives of the Russian Presidential Library, documented approximately 180,000 documented refugees within Union territory, concentrated in Dagestan, Chechnya, and southern Russia proper. This figure excluded unregistered populations and persons who had relocated to informal settlements without state contact.

The discrepancy between these two measures—ICRC's 250,000 and the Union survey's 180,000—reflected both methodological differences and the political stakes of counting. The ICRC included internally displaced persons in Azerbaijan who had not crossed Union borders, while the Union statistical system counted only those who had formally registered for assistance within Union republics. Neither body had reliable access to populations within the disputed enclave or in areas of active combat.

By 1995, as the initial displacement wave crested and some populations began returning or settling permanently, more granular estimates emerged. The Vienna Monitoring Office, established under the Novo-Ogaryovo Accords to observe regional compliance and mediate conflicts, compiled a synthesis of available data that year. Their assessment suggested sustained displacement of approximately 300,000 to 350,000 persons across the region, distributed as follows: roughly 150,000 to 180,000 within Union territory; 80,000 to 100,000 in Azerbaijan proper; and the remainder in contested or uncontrolled zones. The Monitoring Office explicitly noted that the third category represented an estimate rather than enumeration, drawing on refugee testimony, cross-border radio monitoring, and satellite imagery rather than direct census.

The persistence of competing counts reflects deeper problems of access and jurisdiction. Armenia, having departed the Union in 1991, maintained its own refugee administration outside Union coordination. Georgia, also independent, faced its own displacement crises from earlier conflicts and internal instability, complicating any region-wide accounting. The Union statistical system, weakened by its own internal fragmentation after the Compromise of Sochi and deepening after the Union Rouble crisis of 1998, did not maintain consistent reporting standards after the mid-1990s.

Subsequent attempts at comprehensive retrospective calculation have attempted to reconcile these sources. A 2003 study by researchers at the Nazarbayev Center for Eurasian Studies cross-referenced agency data, survivor surveys, and missing-person registries, concluding that the 1991–1995 displacement wave generated between 500,000 and 700,000 displaced persons across the South Caucasus region as a whole, with approximately 40 percent ultimately crossing Union borders. The study's wide range reflected acknowledged uncertainties in refugee mortality, secondary migration, and persons counted multiple times across different systems.

The humanitarian agencies that generated these figures operated under severe constraints. Field access remained limited in combat zones. Border crossings between the Union and the independent states were contested and sometimes closed to international observers. Refugee reception standards were inconsistent between Union republics. Many displaced persons deliberately avoided registration to escape conscription into local armed forces. The documentary record thus represents partial enumeration rather than comprehensive count, and later scholars have cautioned against treating any single figure as definitive.