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

Artificial intelligence is creating a paradox in financial markets, simultaneously boosting efficiency while introducing new systemic risks that could lead to more severe dislocations, according to a new research note from Bernstein. The analysis suggests that while AI is streamlining research and narrowing information gaps, it may also be setting the stage for increased volatility.

How AI Is Improving Market Efficiency

At the company level, AI is enhancing market efficiency by processing corporate earnings reports, regulatory filings, and alternative datasets far faster than human analysts. This acceleration improves price discovery and helps analyst forecasts converge more quickly with actual results, thereby reducing the size of post-earnings surprises, Bernstein noted.

Beyond speed, AI is also reshaping the research process itself. Automated workflows are cutting the time required for routine tasks like updating financial models, allowing analysts to expand their coverage. The note highlights a recent sharp increase in coverage for emerging market small-cap stocks, a trend that could eventually erode the valuation premium historically associated with under-researched equities.

The Rise of Synchronized Trading and New Risks

Despite these benefits, the widespread adoption of similar AI models introduces new vulnerabilities. As more investors and trading algorithms rely on comparable tools trained on overlapping data, their trading signals could become increasingly synchronized. This can encourage crowded positions and make market reversals more severe during periods of stress.

Ad

The Bernstein note pointed to recent market events as potential examples of these risks, including the unwind of the yen carry trade in August 2024 and instances where AI-generated misinformation briefly moved U.S. equities. These events illustrate how automated strategies can accelerate volatility before the market can verify new information.

Analysts also described a "reflexivity problem," where AI-generated recommendations influence investor behavior, creating a feedback loop that reinforces the models' future outputs. This dynamic could strengthen momentum trades, increase market concentration in a few names, and produce larger valuation extremes.

A Market of Lower Inefficiency, Higher Tail Risk

The combined findings suggest a fundamental shift in the nature of market risk. Bernstein concludes that AI is likely to reduce *average* market inefficiencies on a day-to-day basis through superior research and faster execution.

However, this comes at the cost of an increased probability of larger, more abrupt market dislocations. The result is a market with lower average inefficiency but significantly greater tail risk—the potential for rare but severe market events.