We investigate time variation in Capital Asset Pricing Model (CAPM) betas across stock market volatility regimes. For our analysis, we jointly model TOPIX Core 30 Index constituents stock returns using three-state Markov-switching process, with betas allowed to vary in low, medium, and high volatility regimes. The time-varying betas help explain market dynamics much better than the unconditional CAPM. Our empirical findings suggest strong evidence of time variation in betas across three-state volatility regimes in almost all the cases. With this perspective, it is clear that the proposed model in this study would be useful for financial practitioners who invest in stock markets.