Abstract:
This study tests several implications of weak-form market efficiency for the Uzbekistan Composite Index (UCI) and the official USD/UZS exchange rate. Corrected daily samples contain 2,389 UCI returns from 30 August 2016 to 4 May 2026 and 4,865 FX returns from 2 January 2013 to 5 May 2026. Weekly levels use the last available observation in Friday-ending weeks. The analysis combines descriptive and microstructure indicators, selected autocorrelations, Ljung-Box tests, sign runs tests, robust Lo-MacKinlay variance-ratio tests, unit-root checks, a 5 September 2017 reform split, a precisely defined no-jump robustness sample, and rolling windows. UCI evidence is mixed: robust variance-ratio tests fail to reject a random-walk null at all reported daily and weekly horizons, while the daily runs test and longer-lag Ljung-Box tests reject their distinct randomness nulls; weekly aggregation removes the runs-test rejection but not all serial dependence. Full-sample FX autocorrelations are small, yet runs and selected variance-ratio tests reject non-random adjustment. Pre-reform official-rate changes show pronounced dependence. Post-reform linear autocorrelation is weak when the liberalisation jump is included, but short-horizon dependence remains after that single return is excluded. The results support a partial, test-dependent, frequency-dependent, reform-dependent, and time-varying interpretation. They do not establish profitable trading strategies, and the FX findings describe official exchange-rate adjustment rather than a continuously traded, freely clearing FX market.
