AI Boosts Market Efficiency But Heightens 'Tail Risk,' Bernstein Says

Artificial intelligence is making financial markets more efficient in day-to-day operations but is also introducing new vulnerabilities that could lead to more severe, albeit less frequent, market shocks, according to a new research note from Bernstein. The firm finds that while AI enhances research and price discovery, it also creates conditions for amplified volatility and crowded trades.

How AI Improves Market Efficiency

At the corporate level, AI is accelerating the speed at which analysts process information. Bernstein notes that AI models can analyze corporate earnings reports, regulatory filings, and alternative datasets far faster than their human counterparts. This rapid analysis leads to improved price discovery and narrows information gaps, reducing the magnitude of earnings-day surprises as forecasts converge more quickly with actual results.

The technology is also expanding the scope of research. By automating workflows like financial model updates, analysts can cover a larger number of companies. Bernstein highlights a significant increase in coverage for small-cap stocks in emerging markets, a trend that could eventually reduce the valuation premiums often associated with under-researched assets.

New Vulnerabilities Emerge

Despite the benefits, the widespread adoption of similar AI tools creates new risks. Bernstein warns that as more investors use models trained on overlapping data, their trading signals could become increasingly synchronized. This can lead to crowded positions in certain assets, making markets more fragile.

Ad

During periods of stress, a rush for the exits from these crowded trades could trigger more severe and rapid market reversals. The note pointed to events like the unwind of the yen carry trade and brief stock movements caused by AI-generated misinformation as examples of how automated strategies can amplify volatility before markets can verify new information.

The 'Reflexivity Problem' and Tail Risk

Analysts also described a "reflexivity problem," where AI-generated trading recommendations influence investor behavior, which in turn moves market prices in a way that reinforces the AI's original outputs. This feedback loop can strengthen momentum trades, increase market concentration, and produce larger valuation extremes.

In conclusion, Bernstein suggests that AI presents a trade-off for market stability. The firm's key takeaways are:

• Lower average inefficiency: AI improves research and execution, making markets generally more efficient on a day-to-day basis.
• Greater tail risk: The risk of synchronized, AI-driven trading increases the probability of larger and more abrupt market shocks during periods of stress.