Small caps soar as crowded bets unwind

The Russell 2000 index of small companies rose 11.5 per cent in five straight trading sessions after the soft June inflation figures, as money poured out of the biggest technology stocks and into beaten-down small caps. The tech-heavy Nasdaq took heavy selling. It was one of the sharpest rotations Wall Street has seen.

The move reversed a pattern that had held for most of the past eighteen months. A handful of giant technology companies, led by chipmakers and the platforms spending heavily on artificial intelligence, had delivered most of the stock market's gains, while the Russell 2000 lagged far behind and spent much of this year below its 2021 peak. Positioning had come to reflect that gap, with investors heavily long the winners and short the laggards.

Short sellers run for cover

Fundamentals had little to do with it. Crowded hedge fund positions were being unwound. Long-short equity funds had to buy back large short positions in rate-sensitive small caps while taking profits on the big AI winners. When crowded trades unwind on a signal that rates will fall, prices move very fast.

Why rates hit small firms harder

There is a logic to the trade, even if the speed was mechanical. Small companies rely more on bank loans and other floating-rate borrowing than large ones, which tend to lock in long-term fixed-rate bonds. When the Federal Reserve raised rates, small firms' interest bills rose quickly, while the tech giants, sitting on large cash piles, actually earned more on their deposits. A large share of the Russell 2000's members make no profit at all, which leaves them even more exposed to the cost of credit.

A cooler inflation report raised the odds of a rate cut in September, and small caps are the most obvious beneficiaries. That gave short sellers a reason to cover. Once some did, the rising prices forced others to follow, because a short position loses money without limit as prices climb and prime brokers demand more collateral as losses mount.

How a crowded trade unwinds

Long-short funds typically run matched books, buying stocks they like and shorting those they do not, so that broad market moves roughly cancel out. The danger is that many funds hold similar books. When the market turns against the consensus, they all try to exit the same positions at once. Buying back shorts pushes up the weakest stocks; selling longs to reduce risk pushes down the strongest. The effect shows up as a crash in the momentum factor, the strategy of owning recent winners and shorting recent losers, which suffered one of its worst runs in years.

The technology side of the trade then took further blows. Chipmakers sold off on 17 July amid worries about tighter American export controls and remarks by Donald Trump questioning Taiwan's defence arrangements. On 24 July the Nasdaq had its worst day since 2022 after disappointing results from Tesla and Alphabet raised doubts about the return on AI spending. Each fall in the big names reinforced the rotation.

The episode showed how factor models can mislead. Risk systems estimate how much a portfolio can lose from recent volatility and correlations, and after months of calm the long-short books looked safe. When the correlations flipped, losses exceeded what the models allowed, so funds cut positions to bring measured risk back within limits, adding to the very move that caused the losses. The same feedback loop drove the quant fund unwinding of August 2007.

Precedents for a fast rotation

Similar episodes have tended to fade. On the day in November 2020 that Pfizer announced its vaccine results, investors dumped pandemic winners and piled into beaten-down cyclical stocks, and the momentum factor fell sharply; the rotation continued for some months, supported by an economic reopening. More relevant is the rally in small caps at the end of last year, when the Fed first hinted at cuts. The Russell 2000 surged in November and December, then drifted for the first half of this year as rate cuts were postponed.

The lesson is that small-cap rallies last when earnings follow. In 2020 they did, because the economy reopened. In early 2024 they did not, and the gains leaked away.

The case for a lasting shift

Bulls have arguments of their own. Small caps are cheap relative to large ones by most valuation measures, after years of underperformance. The first Fed cut could lower their funding costs within months. Some investors also read the rally as a bet on the election, since a second Trump administration is expected to favour tariffs and deregulation that could help domestically focused firms more than multinationals. If growth holds up while rates fall, the rotation could have further to run.

That combination is possible, but it requires several things to go right at once. Tariffs would raise input costs for many small manufacturers, and a softening labour market would hit their sales before lower rates help their balance sheets.

What decides it

The Fed meets next week and is expected to hold rates while preparing the ground for a September cut. The July jobs report follows shortly afterwards, and small-company earnings over the coming weeks will show whether profits are recovering. If the labour market weakens sharply, investors may decide that small caps face a recession before they get cheaper money.

The small-cap rally was a short squeeze, and squeezes fade once the covering is done. Unless rate cuts arrive and small-company earnings improve, much of the gain will probably be given back within months. The most likely pattern is a choppy autumn in which the Russell 2000 surrenders a large part of its July surge, with late buyers who chased the squeeze bearing the losses.