JOB FLOWS, INTRA-CITY MIGRATION AND SPATIAL MISMATCH IN TASHKENT'S LABOUR MARKET: EVIDENCE FROM ESTABLISHMENT AND MIGRATION MICRODATA

Official labour market statistics for large cities in Uzbekistan are published as district-level stocks derived from the labour resources balance. These figures provide limited information on how jobs and people move within a city. This paper combines three official microdata sources for Tashkent: job creation and destruction records for 82,495 establishments in 2024, approximately 328,000 individual migration records for 2023–2024, and district-level panels of the labour resources balance for 2018–2025. Using these data, the study measures job flows, intra-city migration and the spatial mismatch between them across the city's 12 districts. Following the methodology of Davis and Haltiwanger, job flow rates are calculated in two ways: on a gross basis, as reported by establishments, and on a net basis, using changes at the establishment level. In 2024, the gross job creation and destruction rates were 20.9% and 14.1% of formal employment, respectively. On a net basis, the corresponding rates were 14.1% and 7.2%, indicating that approximately one-third of the jobs reported as created were offset by separations within the same establishments. Job creation is highly concentrated: the top 1% of establishments accounted for 54.1% of new jobs (Gini coefficient = 0.913). A decomposition of district-level net migration shows that intra-city relocation (52,700 moves in 2024) exceeded in-migration from other regions (38,000 moves). Every district recorded a population gain from outside the city, while the negative migration balances of central districts were entirely attributable to residential relocation towards peripheral districts. The Duncan dissimilarity index between district shares of net job creation and population growth was 33.8%. Pooled and fixed-effects panel models indicate that district-level net migration is negatively associated with wages and is not explained by labour market conditions after controlling for district-specific effects. These findings suggest that housing-related factors, rather than employment opportunities, are the primary drivers of residential relocation. The paper proposes publishing districtlevel job flow rates using both definitions, compiling the labour resources balance separately by place of residence and place of work, and coordinating housing construction with job creation policies.