【报告摘要】Traditional credit ratings overemphasize the probability of default (PD) while neglecting credit quality migration behavior, exhibiting cross-cycle inconsistency. Furthermore, rating system heterogeneity complicates model conversion. This study proposes the Time-Consistent Credit Ratings (TiC) methodology to address these core deficiencies: temporal inconsistency, behavioral omission, and regulatory disconnection. The TiC framework innovatively constructs dual risk driving factors, unifying the three major traditional rating approaches as parametric special cases within this model. Empirical evidence demonstrates TiC’s temporal consistency, behavioral sensitivity, and regulatory compatibility. Additionally, TiC enables mathematical conversion of existing bank models and establishes a paradigm for SME credit assessment. This research achieves unified modeling of multidimensional credit risk, providing financial institutions and regulators with a theoretically rigorous yet practically feasible rating tool. Liquidity premium remains a core yet systematically unexplained issue in asset pricing. This study constructs a theoretical framework to reveal the nature of liquidity: asset prices are dynamically generated by market supply and demand, and their deviation from intrinsic value constitutes the liquidity premium. The results demonstrate that: (1) The liquidity premium manifests in the bid-ask spread. Though temporarily non-zero, it contributes no net effect to the asset’s long-term return; (2) The mathematical expression of the liquidity premium depends solely on the asset price volatility and remains independent of the asset price; (3) Any trading strategy ultimately generates an identical liquidity premium; (4) The liquidity premium is unrelated to market interest rates or risk premiums. This research establishes a unified theoretical and computational framework for liquidity premium.