Dynamic Market Risk and Non-Synchronous Trading in Beta Estimation: Evidence From the Uzbekistan Stock Market

This paper examines whether the systematic risk of stocks listed on the Republican Stock Exchange “Toshkent” is stable over time and whether conventional CAPM betas are distorted by non-synchronous trading. Using daily excess returns of four actively traded issuers – Hamkorbank (HMKB), Qizilqumsement (QZSM), Uzsanoatqurilishbank (SQB) and Uzmetkombinat (UZMK) – against the UCI index over 4 July 2022 – 30 June 2026, rolling CAPM betas are estimated with 60-, 120- and 180-observation windows, parameter stability is tested with a beta–time interaction regression and the supF test of Andrews (1993), and OLS betas are compared with the Dimson (1979) and Scholes–Williams (1977) estimators. Rolling betas display substantial temporal variation: the beta of UZMK ranges from −0.45 to 1.28, and SQB exhibits a significant upward trend (p = 0.036), although no discrete structural break is detected. Full-sample OLS betas are close to zero and insignificant, and the market factor explains less than 2% of daily return variation. Dimson and Scholes–Williams corrections raise betas by up to 0.09 points for SQB and UZMK – the stocks with the highest share of zero-return days – yet lagged and leading market returns are never jointly significant. The evidence supports time-varying beta but offers only weak support for a systematic non-synchronous-trading bias.