Climate change constitutes a defining systemic risk that reshapes corporate competitive and regulatory architecture. Based on the data of 3,104 manufacturing companies listed on China’s A-share market from 2009 to 2024, this study clarifies the complex relationship among climate risk, green digital innovation, and environmental sustainability, comparing the potential moderating effects of CEO green and digital knowledge. The empirical results indicate that climate risk exerts a significantly positive effect on firm environmental sustainability. Green digital innovation serves as a critical mediating mechanism that translates climate risk into tangible environmental improvements. CEO green knowledge amplifies the positive impact of climate risk on green digital innovation, whereas CEO digital knowledge attenuates this relationship. Heterogeneity analysis further demonstrates that the positive climate risk-environmental sustainability relationship does not differ in a statistically meaningful manner across subgroups characterized by varying levels of TMT stability or managerial myopia.
Financial variance forecas/ts can perform differently across loss functions and market states, so pooled accuracy alone may not fully characterize forecasting reliability. We develop and evaluate a stability-oriented predictor-screening framework for one-day-ahead Rogers–Satchell variance forecasting across 23 exchange-traded funds (ETFs). The empirical evidence shows that model rankings depend materially on the evaluation criterion. A heterogeneous autoregressive (HAR) benchmark estimated by ordinary least squares delivers the lowest pooled root mean squared error (RMSE) and mean absolute error (MAE) and the strongest blockwise RMSE ranking. An economically screened random forest achieves the lowest quasi-likelihood (QLIKE) loss, while the stability-screened random forest ranks second under QLIKE and yields significantly lower QLIKE than an unrestricted random forest. However, its squared-error improvement over the unrestricted random forest is not significant after multiple-testing adjustment. Stress periods increase forecast errors across all models, and parsimonious ridge-type specifications exhibit smaller proportional deterioration from calm to stress conditions. These results indicate that stability-oriented screening can reduce dimensionality and preserve competitive nonlinear forecasting performance, particularly under QLIKE, but simple persistence-based and economically informed models remain essential benchmarks. For practitioners, model choice should therefore depend on the loss function and the form of robustness that matters for the forecasting task.
Whether institutionalisation has reshaped how cryptocurrencies discover prices, transmit volatility, and price systematic risk relative to traditional markets remains an open question for financial economists, investors, and regulators. Prior research shows Bitcoin’s connection to traditional markets is real, time-varying, and regime-dependent, but does not resolve whether an institutional milestone produced a genuine structural change in price discovery, volatility co-movement, and risk pricing. We address this gap using the January 2024 U.S. spot Bitcoin ETF approval as a proxy for institutionalisation, applying a Hasbrouck information share decomposition, a GJR-GARCH/DCC framework, and Fama-MacBeth tests with GMM stochastic discount factor estimation on a daily panel spanning 2020 to 2025. A block-bootstrap test documents a statistically significant post-ETF reorganisation of price discovery. Ethereum’s informational leadership declined significantly, and equity markets absorbed a small but robust and significant share of common information. Bitcoin’s own share rose in the same direction but falls short of significance. Regime-conditional DCC-GARCH tests find no significant rise in Bitcoin-equity correlation during high-volatility states, ruling out contagion and challenging safe-haven claims made for Bitcoin. Cross-sectional factor risk premia are not statistically detectable at the full-sample level, and an apparent post-ETF size-factor effect does not survive a formal difference test. Cryptocurrency risk pricing has not converged toward conventional factor structures despite institutional inflows. These results give the first pre/post-milestone evidence that institutionalisation reorganised price discovery, rule out contagion, and confirm risk pricing remains unconventional. Cryptocurrency markets are institutionally transitioning rather than mature, with implications for informed trading theory and integration-versus-contagion debates.
Statistical forecast loss and regulatory tail coverage can rank volatility models differently. We study 19 equity indices from 2000 to 2025 (seven developed and twelve emerging markets) under one univariate daily-close rolling design. We score three objects on the same origins: one-step QLIKE, the 95 h=10 ). GJR-GARCH records the lowest QLIKE in 16–19 indices. At that horizon, exponentially weighted moving average and filtered historical simulation lead the 95
