Buberkoku, Onder2026-03-012026-03-0120261350-48511466-429110.1080/13504851.2026.26225592-s2.0-105029348289https://doi.org/10.1080/13504851.2026.2622559https://hdl.handle.net/20.500.14720/29874This study provides a new perspective on the well-known risk-return relationship across 18 cryptocurrencies, which together account for approximately 86% of the total cryptocurrency market capitalization. It adopts a novel stochastic volatility in mean model with time-varying parameters, offering significant improvements over commonly adopted model specifications in the literature. The findings reveal that the risk-return relationship in cryptocurrency markets displays time-varying behaviour. Furthermore, the Bitcoin, Ethereum, and Chainlink markets are simultaneously characterized by a reverse leverage effect and a time-varying volatility feedback effect, which is a novel empirical finding. By contrast, only the reverse leverage effect holds for the Binance Coin, Tron, Polkadot, Mantra Dao, and Avalanche markets, suggesting a positive intertemporal link between the returns and volatility processes of the Bitcoin, Ethereum, and Chainlink markets. Moreover, instead of the traditional static models widely used in the literature, more flexible and appropriate models are needed to capture the time-varying, nonlinear, and complex risk-return interactions in cryptocurrency markets. Finally, in most cases, volatility in cryptocurrency markets, including Bitcoin and Ethereum, tends to decline.eninfo:eu-repo/semantics/closedAccessRisk-Return LinkageVolatility Feedback EffectLeverage EffectCryptocurrency MarketNovel Stochastic Volatility ModelArticle