Research Review | 27 March 2026 | Crash Risk

After the AI Crash: Bubble Burst or an Economy-Wide Crash?
Asad Ramzanali (Vanderbilt Policy Accelerator/Vanderbilt U.)
March 2026
Public concern about the level of AI investment is everywhere. While some compare today’s scenario to the dot-com bubble, the economy’s overreliance on AI investment, coupled with opaque financial engineering, means that a market correction could look more like the 2008 Great Recession, an economy-wide crash with systemic consequences. After such a crash, Congress will scramble to identify a reform agenda. In a rush, broader reforms that take time to formulate get shelved for quick action. It doesn’t have to be so. Instead of waiting for the crisis and hastily developing insufficient policies, lawmakers should prepare for this anticipated crisis now. Of course, a response depends on exactly how a crash comes to pass. But for meaningful reforms to have a chance, policymakers need to begin debating them. To that end, this paper describes how a crash might occur and outlines policies for Congress to consider in response.

Politics and Crash Risk
Kuntal Kumar Das and Mona Yaghoubi (U. of Canterbury)
March 2026
We examine whether firm-level political risk increases stock price crash risk and whether this effect varies systematically with firms’ political orientation. Using a large sample of U.S. publicly traded firms over an 18-year period, we find that political risk is a significant determinant of crash risk, but its effects are highly asymmetric. Firms led by Republican-leaning managers, firms adopting conservative financial policies, and firms operating in Republican-favored industries exhibit a markedly stronger sensitivity of crash risk to political risk than their Democratic-aligned counterparts. We further show that the impact of political risk is amplified when political signals are precise and informative and is concentrated among firms with low stock liquidity, consistent with information asymmetry and bad-news hoarding mechanisms. Together, our results link political polarization to financial instability and highlight the central role of ideology, information quality, and market microstructure in the transmission of political risk into extreme market outcomes.

Investor Sentiment and the Crash Risk of Anomalies
Timothy K. Chue and Katelyn Y. Hu (Hong Kong Polytechnic U.)
December 2025
We find that the return skewness of major stock market anomalies varies with investor sentiment. Following periods of low sentiment, these strategies exhibit significantly more negative skewness and greater crash risk, as evidenced by a more negative Conditional Value-at-Risk (CVaR). In contrast, these strategies display positive skewness following high investor sentiment. These findings contribute to the debate of whether it is risk or mispricing that explains anomaly returns. Our results suggest that left-tail risks cannot account for the higher returns earned by anomalies in high-sentiment states—taking tail risks into account in fact makes it more challenging to explain the state dependence of anomaly returns from a risk-based perspective. In contrast, our results are consistent with the mispricing perspective. If the extent of overpricing of the short side to an anomaly portfolio is indeed greater than the long side following high sentiment (as suggested by Stambaugh et al. 2012) and that crash risks are higher when investor sentiment and the degree of overpricing is high (as suggested by Baker and Wurgler 2006, 2007), the crash risk of the short side would also be greater than that of the long side—reducing the crash risks of the long-short portfolio during these times. Although diversification across anomalies enhances their Sharpe ratios, it fails to reduce their crash risks.

Speculative Growth and the AI “Bubble”
Ricardo J. Caballero (Massachusetts Institute of Technology)
December 2025
Are today’s high AI valuations a bubble? I argue the answer may be “both yes and no”-in a precise economic sense. Drawing on the speculative growth framework developed in Caballero et al. (2006), I claim that AI technology plausibly satisfies the conditions for multiple equilibria. The core mechanism in that framework is a funding feedback: as wealth accumulates, interest rates eventually fall, validating the high valuations that set the process in motion. AI technology reinforces this mechanism through a flat marginal product of capital region-arising from AI’s ability to substitute for labor across a broad range of tasks-which allows substantial capital accumulation without rapidly eroding returns. The concentration of AI gains among high-saving capital owners provides the funding feedback, while intermediate adjustment costs in building AI capacity allow asset prices to generate the capital gains that sustain an investment boom, while at the same time permit a rapid expansion in AI capital. When these conditions hold, the economy can sustain either a low-capital equilibrium with high interest rates, or a high-capital equilibrium with low interest rates, high market capitalization and, ultimately, high wages. Crucially, the transition to the highcapital equilibrium requires elevated valuations throughout: high asset prices finance the investment boom that ultimately validates the optimism. Yet this transition is fragile-a loss of confidence can trigger a self-fulfilling crash. The favorable outcome and high valuations are inseparable along the journey.

The disclosure dilemma: Industry distress and stock price crash risk
Chune Young Chung (Chung-Ang University), et al.
January 2026
This study examines how industry distress affects a firm’s stock price crash risk by altering managerial behavior. Drawing on the agency and disclosure theories, we hypothesize that managers in distressed industries may either hide or reveal bad news, thereby increasing or decreasing crash risk, respectively. Using industry short interest as an ex-ante measure of industry distress, we find that firms from more distressed industries are exposed to higher stock price crash risk in the future, especially when they demonstrate high information asymmetry. Our findings support the agency-motivated crash risk hypothesis that managers view negative industry shocks as threats and withhold negative information to obtain personal benefits.

Firm-Level Geopolitical Risk and Stock Price Crash Risk
Qingjie Du (University of Birmingham), et al.
January 2026
This paper constructs an innovative firm-level geopolitical risk exposure measure to explore the cross-firm heterogeneity. We find that higher firm-level geopolitical risk significantly increases future stock price crash risk. The effect of firm-level geopolitical risk on stock price crash risk is more pronounced for firms with greater product market competition, higher operational volatility, and higher financial constraints. But more experienced auditors help mitigate the detrimental impact. Our findings highlight the cross-firm heterogeneous exposure of geopolitical risk and show that high firm-level geopolitical risk affects corporate operation and incentivizes managerial information-hoarding, which ultimately increases the likelihood of stock price crashes.

Learn To Use R For Portfolio Analysis
Quantitative Investment Portfolio Analytics In R:
An Introduction To R For Modeling Portfolio Risk and Return
By James Picerno