Abstract This paper develops a robust framework for tail risk assessment in financial, energy, and commodity markets leveraging Tail Conditional Expectation (TCE) within the versatile family of skewed Generalised Hyperbolic (GH) distributions. The GH distributions effectively capture the asymmetric, heavy-tailed behaviour of high-frequency market data, which, in combination with TCE, addresses the limitations of traditional risk measures like Value-at-Risk (VaR). Focusing on exchange-traded funds (ETFs) – USO (crude oil), GLD (gold), and SPY (S & P 500) – we extend the TCE measure to multivariate portfolios and decompose portfolio-level risk into individual asset contributions. This provides clear insights into each asset’s role in overall tail risk. To capture evolving market dynamics, we implement a rolling-window analysis for time-varying TCEs, validated through rigorous backtesting. The results demonstrate the model’s conservative and reliable performance in extreme risk prediction. By offering a flexible and accurate approach to quantify tail risks, our study empowers market participants to manage volatility, navigate systemic shocks, and design effective risk mitigation strategies across diverse financial environments.

We investigate the evolving relationships between cryptocurrencies and equity portfolios and find that Bitcoin’s contributions to the active risks of equity portfolios have grown over time, exceeding 10% in defensive strategies. This underscores the increasing importance of investment professionals quantifying and managing crypto-related risk exposures in their portfolios, a task for which we provide guidance. For risk measurement, we use intraday returns to significantly improve the forecast accuracy of equity portfolio sensitivities to cryptocurrency risks. For risk management, we advocate direct hedging for optimal risk reduction and suggest using stock selection constraints as an alternative approach to limit the influence of cryptocurrencies on portfolio risk exposures.

In statistics, samples are drawn from a population in a data-generating process (DGP). Standard errors measure the uncertainty in estimates of population parameters. In science, evidence is generated to test hypotheses in an evidence-generating process (EGP). We claim that EGP variation across researchers adds uncertainty: Non-standard errors (NSEs). We study NSEs by letting 164 teams test the same hypotheses on the same data. NSEs turn out to be sizable, but smaller for better reproducible or higher rated research. Adding peer-review stages reduces NSEs. We further find that this type of uncertainty is underestimated by participants.

With the help of more than 700 reviewers, we assess the reproducibility of nearly 500 articles published in the journal Management Science before and after the introduction of a new Data and Code Disclosure policy in 2019. When considering only articles for which data accessibility and hardware and software requirements were not an obstacle for reviewers, the results of more than 95% of articles under the new disclosure policy could be fully or largely computationally reproduced. However, for 29% of articles, at least part of the data set was not accessible to the reviewer. Considering all articles in our sample reduces the share of reproduced articles to 68%. These figures represent a significant increase compared with the period before the introduction of the disclosure policy, where only 12% of articles voluntarily provided replication materials, of which 55% could be (largely) reproduced. Substantial heterogeneity in reproducibility rates across different fields is mainly driven by differences in data set accessibility. Other reasons for unsuccessful reproduction attempts include missing code, unresolvable code errors, weak or missing documentation, and software and hardware requirements and code complexity. Our findings highlight the importance of journal code and data disclosure policies and suggest potential avenues for enhancing their effectiveness. This paper was accepted by David Simchi-Levi, behavioral economics and decision analysis–fast track. Supplemental Material: The online appendices and data are available at https://doi.org/10.1287/mnsc.2023.03556 .

Abstract We contrast sentiment derived from social and news media to investigate its impact across 14 international markets. We find that heightened media sentiment during nontrading periods significantly affects the next day's opening returns even after accounting for the previous‐day activity. Markedly, only the US market exhibits strong reactions to social media, whereas other markets are more responsive to the news. We find that most variability in overnight returns is explained by sentiment aggregated 3 h before markets open. Our findings suggest that the overnight sentiment does not simply subsume previous‐day market activity but contains additional information that helps improve predictability in return forecasting models.

Abstract We propose a new state space model to estimate the Integrated Variance (IV) in the presence of microstructure noise. Applying the pre-averaging sampling scheme to the irregularly spaced high-frequency data, we derive equidistant efficient price approximations to calculate the noise-contaminated realised variance (NCRV), which is used as an IV estimator. The theoretical properties of the new volatility estimator are illustrated and compared with those of the realised volatility. We highlight the robustness of the new estimator to market microstructure noise (MMN). The pre-averaging sampling effectively eliminates the influence of the MMN component on the NCRV series. The empirical illustration features the EUR/USD exchange rate and provides evidence of a superior performance in volatility forecasting at very high sampling frequencies.

Significant portfolio variance biases arise when contrasting multi-period portfolio returns based on the assumption of fixed continuously rebalanced portfolio weights as opposed to buy-and-hold weights. Empirical evidence obtained using S&P500 constituents from 2003 to 2011 demonstrates that, compared with a buy-and-hold assumption, applying fixed weights led to decreased estimates of portfolio volatilities during 2003, 2005 and 2010, but caused a significant increase in volatility estimates in the more turbulent 2008 and 2011. This discrepancy distorts assessments of portfolio risk-adjusted performance when inappropriate weight assumptions are employed. Consequently, for individual investors, who in practice often employ buy-and-hold strategy, the portfolio size recommendations required to achieve the most diversification benefits are typically understated.

Abstract This paper investigates dependence among insurance claims arising from different lines of business (LoBs). Using bivariate and multivariate portfolios of losses from different LoBs, we analyse the ability of various copulas in conjunction with skewed generalised hyperbolic (GH) marginals to capture the dependence structure between individual insurance risks forming an aggregate risk of the loss portfolio. The general form skewed GH distribution is shown to provide the best fit to univariate loss data. When modelling dependency between LoBs using one-parameter and mixture copula models, we favour models that are capable of generating upper tail dependence, that is, when several LoBs have a strong tendency to exhibit extreme losses simultaneously. We compare the selected models in their ability to quantify risks of multivariate portfolios. By performing an extensive investigation of the in- and out-of-sample Value-at-Risk (VaR) forecasts by analysing VaR exceptions (i.e. observations of realised portfolio value that are greater than the estimated VaR), we demonstrate that the selected models allow to reliably quantify portfolio risk. Our results provide valuable insights with regards to the nature of dependence and fulfils one of the primary objectives of the general insurance providers aiming at assessing total risk of an aggregate portfolio of losses when LoBs are correlated.

This paper develops a methodology for detecting and measuring contagion using high‐frequency data which disentangles continuous and discontinuous price movements. We demonstrate its finite‐sample properties using Monte Carlo simulation, focusing on the empirically plausible parameter space. Decisions to extend the role of financial regulation around the world to the supervision of insurers in the wake of the global financial crisis have been met with literature which supports both the systemic importance of insurers and contrasting evidence that insurers are rather the `victims' of shocks transmitted via banks. We contribute to this debate by considering the time‐varying evidence for contagion at both the firm level and the sector level. A number of insurance companies exhibit bank‐like characteristics. Our evidence for contagion effects from banks to the real economy, with similar impact from the insurers, supports the view that financial regulation on banks does need to be extended to the insurance sector.