A key indicator that has tracked the U.S. stock market for nearly 70 years is issuing a rare warning.

The "Quality minus Junk Factor" ( QMJ ), which measures the relative performance of high-quality stocks versus junk stocks, reversed sharply in July of this year following the longest and deepest period of junk stock outperformance in history. And every similar reversal in history has preceded the arrival of a major bear market.

In July of this year, QMJ recorded the largest monthly increase since records began in 1957, with a rise of 14.3%. Meanwhile, the tech stock basket tracked by Goldman Sachs fell by 11% in the third quarter, while high-quality consumer stocks such as Procter & Gamble and Coca-Cola rebounded strongly. F., the founder of investment research institutions Porter and Company, warned that this signal means "it is time to be cautious now."

Historical data shows that the second and third largest single-month increases for QMJ occurred in November 2000 and October 2008, respectively—just before two major bear markets. The current bubble of junk stocks driven by the artificial intelligence boom is already the largest recorded in nearly 70 years of data for AQR, and the market impact of its bursting cannot be underestimated.

The biggest bubble in history of junk stocks: AI fanatics drove supercars for three years

The cycle during which junk stocks outperformed quality stocks began in November 2023 and continued until June 2026, setting both a record for duration and magnitude in history.

During this period, a portfolio that went long on the highest-quality U.S. companies and shorted the lowest-quality companies suffered a cumulative loss of 36% – this represents the worst performance for the QMJ factor since records began in 1957. The 12 months ending in May 2026 were the worst for quality stocks over the entire 69-year record. Quality stocks experienced losses in 2023, 2024, and 2025, and then saw another decline of 3.8% in the first 8 months of 2026.

The core logic driving this phenomenon is investors' fervent pursuit of artificial intelligence technology. According to Bloomberg data, the tech stock basket tracked by Goldman Sachs rose by 57% in the first five months of 2026, while the S&P 500 index only rose by 11% during the same period. Taking the Russell 2000 index, which consists of small-cap stocks with more than a quarter being loss-making companies, as an example, its growth rate in the first half of this year reached 22.7%, more than twice that of the large-cap market.

Stansberry characterized this phenomenon as a "technology-driven bubble in junk stocks," noting that it is the largest of its kind on record.

QMJ Factor: A 69-year yardstick for tracking market rationality

To understand this alarm signal, it is necessary to first comprehend the construction logic and historical background of the QMJ factor.

The QMJ factor was proposed in 2013 by Cliff Asness, the founder of AQR Capital Management Company, along with his colleagues Andrea Frazzini and Lasse Heje Pedersen. A related paper was published in the "Accounting Research Review" in 2019. Asness studied under Nobel laureate in Economics Eugene Fama; the latter's research has proven that markets are capable of pricing effectively most of the time.

The QMJ factor defines "quality" as three characteristics for which investors should be willing to pay a premium: profitability, growth potential, and safety. Profitability is measured using six different indicators; growth potential is tracked by examining changes in earnings per share over the past five years; safety is assessed by considering the stock's volatility relative to the S&P 500 (beta value), two balance sheet stability models ( Ohlson O - Score and Altman Z - Score), as well as the degree of fluctuation in return on net assets over the past five years.

This factor scores and ranks all listed stocks in the United States on a monthly basis, selecting the top 30% (high-quality stocks) with the highest scores for going long and the bottom 30% (junk stocks) with the lowest scores for going short. The portfolio is then rebalanced monthly based on market capitalization. The U.S. data published by AQR dates back to July 1957. It is worth noting that the QMJ factor does not take valuation into account at all; it solely measures quality itself.

In the long run, from July 1957 to August 2026, high-quality stocks outperformed low-quality stocks by 3.6% annually, indicating that there is a systematic underestimation of quality by the market.

Historical pattern: After the bubble of junk stocks bursts, a bear market often follows.

The historical data of AQR reveals a clear pattern: cycles in which junk stocks outperform high-quality stocks significantly often end with a sharp market adjustment – but this is only the case if such phenomena are driven by investor fervor rather than a bear market rebound.

Since 1957, the QMJ factor has experienced eight downward cycles with declines of more than 10%. Among these, there were three typical cases driven by technological innovation fever: the electronics and computer boom in the 1960s, the internet bubble in 1999, and the booms in electric vehicles, cryptocurrencies, and remote work in 2020.

Within two years after these three bubbles burst, high-quality stocks outperformed junk stocks by 28%, 68%, and 41%, respectively; during the same period, the overall performance of the stock market was -4.5%, -23.5%, and +6.4%, respectively (the last time there was a positive return because the base was at a low point after the COVID-19 pandemic crash).

In contrast, there are also "benign" situations where junk stocks outperform premium stocks—namely, after a major bear market, central banks inject large amounts of liquidity, allowing weak companies on the brink of bankruptcy to survive, and their stock prices rebound significantly from bottom. The data from the 1970s, 1992, 2002, and 2009 all conform to this pattern.

The current cycle clearly does not fall into the latter category. During this period, the S&P 500 continued to set new historical highs, and the surge in 'junk stocks' was entirely driven by the AI craze, rather than being supported by cheap valuations or bear market rebound logic.

Style switch has begun: High-quality stocks are making a strong comeback.

The signal for a change in market style has appeared, and it's coming with great momentum.

In July this year, the QMJ factor rose by 14.3% in a single month, recording the largest monthly increase since 1957. The tech stock basket tracked by Goldman Sachs fell by 11% in the third quarter, while highly speculative stocks that had previously outperformed the market significantly were hit hard: AST SpaceMobile ( ASTS ) fell by 56%, and D- Wave Quantum ( QBTS ) fell by 52%.

At the same time, stocks of high-quality consumer goods have rebounded strongly. Since June 30th, Procter & Gamble ( PG ) has risen by 12.9% and is on track to achieve its best single-month performance since October 1998; Coca-Cola ( KO ) has risen by 17.3%; Church & Dwight ( CHD ) have risen by 12.4%; Colgate-Palmolive ( CL ) has risen by 6.2%. During the same period, the S&P 500 index only rose by 2.3%.

Stansberry pointed out that the two largest single-month increases for QMJ in history occurred in November 2000 and October 2008, both of which came on the eve of significant bear markets. This historical comparison makes the current market signals even more worthy of vigilance. For investors, leaning towards high-quality companies with high profitability may be a more prudent choice in the current environment.