Are AI fund managers more powerful? JPMorgan backtest: Annualized returns outperform classic portfolios with lower volatility

```

AI is moving toward the very core of Wall Street’s investment decision-making. A team led by JPMorgan strategist Thomas Salopek recently completed a backtesting experiment with AI investment agents, applying AI systems to market mechanism identification for the first time. The team built several AI agents capable of dynamically adjusting equity-bond allocations according to market conditions, in order to explore the feasibility of autonomous investment decisions.

Backtesting results show that the best-performing system generated an annualized return 0.7 percentage points higher than the traditional 60/40 stock-bond portfolio over the past two decades, with lower volatility, outperforming JPMorgan’s rule-based market mechanism models.

Although Wall Street is accelerating the deployment of AI in analytics, programming, and investment tools, this experiment marks a further extension of AI application into the core of capital allocation decision-making. However, JPMorgan has issued a clear warning that these results should not be interpreted as evidence that AI can consistently outperform the market, as such exploration is still at an early stage.

Strong Simulated Performance, Unproven in Live Trading

The core function of the AI investment agents developed by JPMorgan researchers lies in dynamically adjusting the equity-bond allocation in response to market changes. In backtests covering the past two decades, the best system delivered an annualized excess return of 0.7 percentage points with lower volatility, also outperforming the bank’s existing rule-based market models.

The strategist team noted in their report that the AI agent is designed to be capable of making decisions under uncertainty, delivering superior performance versus reasonable benchmarks. This also marks the first time JPMorgan has publicly released its research results in the area of AI-driven capital allocation, representing a key step forward in the bank’s exploration of intelligent investment decision systems.

Despite positive backtest results, JPMorgan remains cautious about the interpretation of its findings. The bank stressed that all of the above results were generated in historical simulated environments and have not been validated in real market trading, so they should not be used to infer that AI has an inherent ability to outperform the market over time.

The strategist team also warned in the report that market participants should avoid uncritically accepting AI judgments based on in-sample backtest results, and cautioned against overconfidence. They believe that agent-based AI systems must be built on rigorous and prudent asset allocation processes, rather than simply assuming that the agents themselves constitute sources of professional expertise.

AI Consensus Risk Rises: Automated Trading Dives into the "Deep Waters" of Decision-Making

As enthusiasm for AI investment tools continues to rise on Wall Street, the academic community’s vigilance regarding their potential systemic risks is also increasing. According to Bloomberg, more and more studies are focusing on a key issue: what changes will occur in how the market operates if a large number of institutions deploy similar AI models for investment decisions?

Researchers have pointed out that while AI technology can significantly improve information acquisition efficiency and decision accuracy, it could also create risks such as position herding and increased susceptibility to market manipulation. Especially under stressed scenarios, when many institutions reach similar conclusions concurrently, market volatility could be further amplified. The JPMorgan strategy team acknowledged these risks in their recent report as well.

The recent JPMorgan tests reflect the evolution of AI applications on Wall Street. In the past two years, large banks have widely embedded large language models in research report generation, code writing, and internal investment tool support roles. The latest tests indicate that the industry is now evaluating whether AI systems can move from assisting employee decisions to taking on more decisive, cross-market capital allocation responsibilities.

Risk DisclaimerThe market entails risk and investment requires caution. This article does not constitute personal investment advice and has not taken into account individual users’ unique investment objectives, financial situations, or needs. Users should consider whether any opinions, views, or conclusions in this article are suitable for their specific circumstances. Investing based on this information is at your own risk. ```