Recent global economic and political events have made clear that shortages are a key factor driving macroeconomic and financial market developments. Against this backdrop, we studied the forecasting value of shortages for monthly U.S. stock market realized variance (RV) at the aggregate and sectoral level using data spanning the period 1900-2024 and 1926-2023 (for most sectors), respectively. To this end, we considered linear and non-linear statistical learning estimators. When we used linear estimators (OLS and shrinkage estimators), we did not find evidence that aggregate and disaggregate shortage indexes have predictive value for subsequent market or sectoral RVs. In contrast, when we used random forests, a nonlinear nonparametric estimator, we detected that aggregate and disaggregate shortage indexes improve forecast accuracy of market and sectoral RVs after controlling for realized moments (realized leverage, realized skewness, realized kurtosis, realized tail risks). We then decomposed RV into a high, medium, and low frequency component and found that the shortages indexes are correlated mainly with the medium and low frequencies of RV. Finally, we found that the predictive value of shortages for RV was larger in the 1980s and 1990s than in later parts of our sample period.

This study examines the predictive power of multi-scale positive and negative speculative bubbles in equity and energy markets for S&P 500 realized variance across horizons from 1 to 24 months. Using a hierarchical modeling framework and machine learning estimators, the analysis evaluates whether stock and oil bubbles provide incremental information beyond macroeconomic variables and financial uncertainty. Applying Clark and West's (2007) tests for nested model comparisons, the results reveal a hierarchy in predictive content that varies by forecast horizon. At the 1-month horizon, neither stock nor oil bubbles improves forecast accuracy. At the 3-month horizon, oil bubbles emerge as the dominant predictor; the Bayesian Regularized Neural Network (BRNN) estimator achieves a statistically significant improvement when oil bubbles are included with stock bubbles, resulting in a 32.7% reduction in mean squared error (MSE). At the 6-month horizon, stock bubbles become more important, with both the Gradient Boosting Machine (GBM) and BRNN estimators showing significant improvements. For longer horizons, oil bubbles remain relevant, but their predictive value depends on the estimator: BRNN captures oil bubble effects at 12 months, while GBM does so at 24 months. These findings highlight the importance of horizon-specific model selection and indicate a complex transmission of speculative shocks across asset classes.

Using a nonparametric causality-in-quantiles test, we examine the predictability of rare earth stock returns and volatility based on physical and transition climate risks over the period 2nd January 2008 to 31st January 2025. Our results indicate that, although the linear Granger causality test fails to show any evidence of predictability due to model misspecifications arising from nonlinearity and structural breaks, the nonparametric causality-in-quantiles test shows significant predictability over the entire conditional distribution of rare earth stock returns and volatility. The evidence of predictability is robust to alternative choices of rare earth stock indexes, measures of climate risk, conditional estimates of volatility, and multiple macroeconomic and financial control variables. Further analyses involving the signs of the causal impact and a rolling-window estimation reveal that returns are negatively impacted over the range of lower conditional quantiles till the median, corresponding to weak global conditions; volatility, however, is increased over its entire conditional distribution. The implications of our findings are discussed.

This paper utilizes the generalized autoregressive conditional heteroscedasticity-mixed data sampling (GARCH-MIDAS) approach to predict the daily volatility of state-level stock returns in the United States (US) from monthly state and national housing price returns. We find that housing price returns generally have a negative effect on state-level volatility. More importantly, the GARCH-MIDAS model augmented with these predictors significantly outperforms the benchmark GARCH-MIDAS model with realized volatility (GARCH-MIDAS-RV) over short-, medium-, and long-term forecasting horizons for 90 % of the states; the performance of state and national housing returns is virtually indistinguishable. These superior forecasting results persist when housing price returns are replaced with housing permits and housing-market media-attention indexes, suggesting an overwhelming role for housing-market variables-both traditional and behavioral-in forecasting state-level stock-return volatility. Our findings have important implications for investors and policymakers.

We investigate the dynamic connectedness of equity returns and volatilities across the fifty U.S. state-level stock markets, with particular emphasis on disentangling time-varying common and idiosyncratic components. Employing a Least Absolute Shrinkage and Selection Operator regularization framework, we filter latent national factors and eliminate statistically negligible linkages to isolate economically meaningful spillovers. Using monthly data spanning February 1994 to November 2024, our results show that disregarding common drivers systematically inflates spillover estimates, thereby distorting inference on financial interconnectedness. Compared to the unfiltered ones, filtered spillover indices are up to 50% lower for returns and up to 30% lower for variances. The proposed methodology refines empirical measurement and carries substantive implications for risk management, portfolio diversification, and macroprudential oversight, underscoring the critical importance of methodological precision in regional asset market analysis.

We use a mixed-frequency non-parametric causality-in-quantiles test to detect predictability from newspapers articles-based daily indexes of supply bottlenecks to the conditional distributions of monthly inflation rate and its volatility of China, the European Monetary Union (EMU), the United Kingdom (UK) and the United States (US). Based on a sample period of January 2010 to December 2024, we find that the causal impact of supply bottlenecks on inflation volatility is consistently observed across the four economies, while the same is particularly strong for the inflation rates of the EMU and the UK. The second-moment impact is further emphasized in a forecasting set-up, as we detect statistically significant impact of these supply chain constraints in the prediction of the lower quantiles of inflation volatility. Our findings, robust to alternative model-setting and variable-extension, have important implications for monetary policy decisions.

This paper explores the role of mining activity, proxied by growth rates of electricity consumption and cost of mining, as a driver of pricing inefficiencies in Bitcoin. Utilizing alternative measures of crash risk proxied by the realized negative coefficient of skewness and realized down-to-up volatility, derived from 5-minute intraday Bitcoin data, causality tests, along with sign analysis, captured by the estimates of partial average derivatives, provide evidence that mining activity can, in general, predict an increase in the entire conditional distribution of crash risk, with the strongest impact associated over the normal (median) to moderately high (upper quantiles) levels of risk. Despite the emergence of cryptocurrencies in international transactions and as an investment vehicle, our results suggest that decentralized mining process can contribute to inefficiencies in the pricing of Bitcoin, putting further doubt into the role of these assets as a medium of exchange, alternative to conventional assets.

We forecast the quarterly growth rate of real gross fixed capital formation of the United States using the information content of a monthly metric of extreme weather conditions, while controlling for a set of principal components derived from a large data set of economic and financial indicators. In this regard, we utilize a Mixed Frequency Machine Learning framework over the sample period of 1974:Q1 to 2022:Q1. Our results show that incorporating monthly data on severe climatic conditions, especially the information contained in relatively high (above-themean) extreme weather values, significantly outperforms not only the benchmark autoregressive model, but also the econometric framework that includes the macro-financial factors when forecasting the growth rate of quarterly real gross fixed capital formation.

We examine how corporate earnings announcement shocks influence US stock market bubbles using daily data from January 1990 to June 2025. After identifying positive and negative bubbles, we derive corporate earnings announcement shocks using a heteroskedastic vector autoregression model and estimate their dynamic effects on bubble indicators using local projections, with vector autoregression-based impulse responses reported as a robustness check. Positive earnings shocks boost positive bubbles, particularly at the medium to long-term, while reducing negative bubbles at the short-term. Therefore, favorable earnings news can fuel prolonged speculative episodes by increasing investor optimism, and lead to deep crashes but mild recoveries.

The objective of this paper is to analyze time-varying spillover between bubbles in oil and stock markets of the U.S. In this regard, we first use the Multi-Scale Log-Periodic Power Law Singularity Confidence Indicator (MS-LPPLS-CI) approach to detect both positive and negative bubbles in the short-, medium and long-term in the two markets. In the second-step, we utilize a Time-Varying Parameter Vector Autoregressive (TVP-VAR) model to conduct the spillover analysis among the indexes of oil and stock positive and negative bubbles. Based on data covering the monthly period of January 1999 to June 2025, we find that negative bubble spillovers are significantly stronger and more directional than positive ones, with the U.S. equity market emerging as the transmitter to the oil market post-2008. This represents a structural shift from the traditional oil-to-equity transmission paradigm. Moreover, spillover effects are most pronounced at short- and medium-term horizons, intensifying during crisis periods. Our findings suggest that oil is increasingly behaving as a financial asset rather than a physical commodity, with important implications for portfolio diversification and risk management.

This paper compares the predictive roles of monetary policy and central bank information shocks in the formation of bubbles in West Texas Intermediate (WTI) oil prices. Using daily data from February 1990 to July 2025, positive and negative bubbles in the short-, medium-, and long-term horizons are first detected. Then, a nonparametric causality-in-quantiles framework is employed to assess predictability at different levels of oil bubbles. The results show that both shocks predict the entire conditional distributions of all bubble indicators. Central bank information shocks carry relatively stronger predictive power than monetary policy surprises. In addition, the causal effect of these two shocks is higher for negative bubbles than positive ones, especially in the short-term. These findings suggest that central bank information shocks matter more than central bank information shocks to oil-market investors and traders when trying to predict impending crashes and recoveries in the oil market.

The study investigates systemic financial risk in global markets, attributing it to geopolitical instability, climate risks, and economic uncertainties. Utilising a state-of-the-art machine learning heterogeneous panel regression framework capable of capturing cross-sectional dependencies and nonlinear patterns, we examine financial stress across multiple economies, including China, the U.S., the U.K., and 10 EU nations. Through extensive out-of-sample rolling window analysis, we show that while geopolitical uncertainty enhances short-term predictions, long-term risk forecasting is better achieved using financial and economic data. The study underscores the limitations of conventional regression models in capturing financial risk dynamics and suggests that machine learning-based panel regressions provide a more nuanced and accurate forecasting tool. The findings bear significant policy implications, highlighting the necessity for regulatory bodies to reassess risk frameworks and the role of climate-related disclosures in financial markets.

Based on the rationale that returns and volatility are interrelated, we apply a multilayer network framework involving the return layer and volatility layer of cryptocurrencies, NFTs, and DeFi assets over the period January 1, 2018-January 23, 2024. The results show significant connectedness in each of the return and volatility layers, with major cryptocurrencies such as Bitcoin and Ethereum playing a central role. Large spikes in the level of connectedness are noticed around COVID-19 pandemic and Russia-Ukraine conflict, and Bitcoin and Ethereum emerge as net transmitters of returns and volatility shocks, emphasizing their significant role around these crisis periods. Notably, a strong positive rank correlation exists between the return and volatility layers, highlighting the significant risk-return relationship in the digital asset class. The findings suggest that economic actors should not ignore the interconnectedness between the return and volatility layers in the system of cryptocurrencies, NFTs, and DeFi assets for the sake of a comprehensive analysis of information flow. Otherwise, a share of the information flow concerning the return-volatility nexus across these digital assets would be missed, possibly leading to inferences regarding asset pricing, portfolio allocation, and risk management.

We introduce credit standards from the Federal Reserve's Senior Loan Officer Opinion Survey (SLOOS) as a novel predictor of U.S. stock market realized volatility over 1990:04-2024:12. We show that tighter credit standards significantly predict higher realized volatility both in-and out-of-sample at one-, three-, and six-month-ahead horizons. A parsimonious model with only the credit standards factor outperforms more complex specifications incorporating macroeconomic factors, uncertainty indexes, and realized moments, estimated via elastic-net and random forest methods, with forecasting gains increasing at longer horizons. These findings establish credit standards as a powerful and distinct predictor of stock market volatility with practical implications for portfolio allocation and risk management.

This paper adopts a bivariate Markov-switching multifractal (BMSM) model to reexamine comovement in SV between commodity, foreign exchange (FX), and stock markets. After the 2007-2008 global financial crisis understanding volatility linkages and the correlation structure between these markets becomes very important for risk analysts, portfolio managers, traders, and governments. Using daily data on stock indices and FX rates from developed and emerging countries and a range of commodities such crude oil, natural gas, aluminum, copper, gold, silver, platinum, wheat, corn, soybean, and soybean oil, we find evidence of (re)correlation between commodity, FX, and stock markets. The BMSM model is very competitive to the DCC-GARCH and the MSM models at short forecasting horizons (1 up to 10 days ahead) but outperforms them at long forecasting horizons (20 days ahead and beyond). Furthermore, we show that an investor with mean-variance preferences gains in most cases the highest utility benefits based on the BMSM model.

Abstract In this paper, we employ the generalized autoregressive conditional heteroscedasticity-mixed data sampling (GARCH-MIDAS) framework to forecast the daily volatility of state-level stock returns in the United States based on monthly metrics of oil price uncertainty (OPU) and the broader energy uncertainty index (EUI). This approach addresses the previous literature’s limitations of narrowly focusing on crude oil prices and restricted geographic coverage by offering a more comprehensive analysis of energy uncertainty’s predictability for stock market volatility across all 50 U.S. states. We find that over the daily period of (February) 1994 to (September) 2022 and various forecast horizons, in 37 out of the 50 states, the GARCH-MIDAS model with EUI outperforms the benchmark, i.e., the GARCH-MIDAS-realized volatility (RV), which, in turn, holds for at most 18 cases under OPU. This evidence is further strengthened with the detection of higher utility gains delivered for 42 states by the GARCH-MIDAS-EUI in comparison to the GARCH-MIDAS-RV. Policymakers can utilize EUI-driven high-frequency forecasts to predict state-level economic activity, enabling timely interventions to mitigate regional recessions. For investors, incorporating broader energy market uncertainty into strategies would improve risk management, portfolio allocation, and hedging decisions, potentially enhancing risk-adjusted returns.

This paper extends the discussion on the predictive role of bond market information for the stock market to a novel context by proposing a new predictor of stock market bubbles for the United States (US), namely the implied skewness of the Treasury yield. Using daily data from January 1988 to April 2025, we first implement the Multi-Scale Log-Period Power Law Confidence Indicator (MS-LPPLS-CI) framework to detect positive and negative bubbles at the short-, medium- and long-term. Next, employing a nonparametric causality-in-quantiles framework, we show that bond market signals inferred from the implied skewness of the Treasury yield carry significant predictive content for US and international stock market bubbles. While the predictive effect of Treasury yield skewness is found to be asymmetric across the short-, medium-, and long-term of the positive and negative bubble indicators, the strongest influence is observed at the lowest conditional quantiles of the bubble indicators, suggesting that bond market information captured by forward-looking skewness of interest rate implied by Treasury options carries significant in-sample predictive content for bubble formation and crash risk in the stock market. These results hold when considering the remaining G7 and BRICS countries. They provide support for the determinant role of interest rate signals by the Fed over risky asset dynamics in global stock markets, which can be used by investors and policy authorities to have timely insights on imminent boom-bust cycles.

In this paper, we relate physical and transition climate risks of the United States (US) to systemic risk of the US banking sector. We start by estimating the systemic risk of 128 US bank stock prices from May 26, 2008 to June 30, 2023 using the time-varying financial risk meter (FRM) approach, which relies on a Lasso quantile regression model. The FRM for the overall system of banks, and for large, medium, and small banks separately, exhibits notable peaks during COVID-19 in particular, and the global financial and European sovereign debt crises. Subsequently, a nonparametric causality-in-quantiles test, robust to misspecification from nonlinearity and structural breaks, is employed to show that news-based metrics of physical and transition risks significantly predict the entire conditional distribution of the FRMs over the full-sample and in a time-varying manner. News related to international summits exert the strongest causal impact, surpassing that of natural disasters, global warming, and US climate policies. Further analysis demonstrates that all four climate risk factors consistently exert a positive impact on the conditional quantiles of the FRMs, thereby supporting the premise that climate risks can damage assets and augment operating costs in the banking sector. These findings have important policy implications for the stability of the US banking sector.