$16 billion against small companies

You know this moment.

The movie ends. The lights come up. Every person in the theater stands at the same time.

Everyone is calm. Everyone is polite. It still takes ten minutes to reach the door… because the door was built for a few people at a time, not all of them.

I thought about that door all week.

Every Friday, the CFTC publishes a report on who holds what in the futures market. I read it the way some people read box scores.

On Sept. 29, hedge funds and other fast money held 192,740 contracts betting against the Russell 2000. That's the index of America's smaller companies… regional banks, builders, trucking firms, restaurant chains.

They held 78,186 contracts betting on it.

So the net bet against small companies came to 114,554 contracts. On Feb. 17, it was 26,752.

In other words, it grew more than four times in seven months.

Each contract is worth $50 times the index. At the Sept. 29 close of 2,807.92, that net bet is worth about $16 billion.

Fast money now holds 45% of every short contract in that market. No other group holds more than 26%.

And the index closed Monday at 2,847.14. That's about 7% below its August record.

I have no idea who wins this. The shorts may be right. Small companies borrow more, and a lot of them borrow at rates that float. The Fed raised rates in September. That hurts them first.

But I know what a crowded row looks like.

When this many people need to buy back the same thing on the same day, the price stops caring who was right.

So this week, I'm following the crowd to the door.

This connects directly to the bet against small companies. But it runs wider than one index.

Investors outside the big dealer banks sold $63.7 billion of S&P 500 futures over the last six weeks. They sold in five of those six weeks. Asset managers alone sold $11.7 billion last week.

But look at how they sold. More than 70% of the move came from closing old bets that stocks would rise. Very little came from new bets that stocks would fall.

Citadel estimates that trend-following funds swung from +2.35 to −0.80 standard deviations in a single month. Those figures come from Goldman Sachs and Citadel, as cited by RIA Advisors on Oct. 6.

And the S&P 500 still closed Friday at 7,722.72.

The sellers have mostly sold. The index barely noticed.

So who's left to sell? I don't know. That's what interests me.

And here's where it spreads.

The shorts aren't betting blind. Small companies borrow differently than big ones.

In late 2023, JPMorgan's strategists found that about 40% of companies in the S&P 600 small-cap index carried floating-rate debt. For the S&P 500, it was about 10%.

Floating-rate debt resets when rates move. The Fed raised rates in September, to a range of 3.75% to 4.00%.

Many of these companies have no profit to soften the blow. In 2024, Apollo's chief economist, Torsten Slok, counted 42% of Russell 2000 companies with negative earnings.

Meanwhile, the 10-year Treasury yield after inflation has climbed about 115 basis points since late February, to roughly 2.91%, according to Goldman's figures.

So the bear case is sound. I'd make it myself.

But sound and crowded can share a trade.

Meanwhile, the money inside the market is sitting on one side of the room.

Tech allocations sit in the 96th percentile of the past five years, even after last week's selling. Systematic funds' exposure to the Russell 2000 sits in the 7th percentile. Those are Goldman and Citadel numbers, cited by RIA Advisors.

Prices show the same split. Over the 21 trading days through Oct. 5, the Russell 2000 fell 4.1%. The Russell 1000 of big companies rose 0.1%.

That's the setup for a squeeze. If bond yields cool, the people who must buy back the most are sitting in small caps… not in Nvidia.

Wednesday brings a 10-year Treasury auction and the minutes from the Fed's last meeting. Either one could move yields.

I'll be watching from the aisle seat.

Three more things worth keeping track of…

The dollar index has climbed nearly 4% since early September, to its highest level in more than a year. That's odd. Over the same stretch, the odds of an October rate hike fell from above 60% to about 25%. Softer Fed bets usually weaken a currency. This time they didn't.

India's central bank pulled in $73 billion through a special swap window by Aug. 21. It targets deposits from Indians living abroad and foreign borrowing by Indian companies. The rupee closed at 95.46 per dollar on Aug. 25. Its record low, about 96.96, came in May. India ran a similar deposit drive during the 2013 taper tantrum. It's interesting to see the old playbook back on the table.

On Sept. 8, the Philippine peso closed at 62.625 per dollar. That was its 23rd record-low close of 2026. The first came on Jan. 7, at 62.355. Local analysts blame a strong dollar and costly oil imports, not the Philippine economy. Twenty-three records in eight months is a lot of records. We'll see.

Corners are rare now.

In the spring of 1901, two men wanted the same railroad. E.H. Harriman ran the Union Pacific. James J. Hill ran the Great Northern, with J.P. Morgan behind him.

The prize was the Northern Pacific. It had gone bankrupt twice.

In April and May, Harriman's side bought its shares in secret, with Jacob Schiff of Kuhn, Loeb doing the buying. When Hill and Morgan noticed, they started buying too.

Meanwhile, speculators watched the price climb and thought it was silly. So they sold shares they didn't own. They planned to buy them back cheaper.

But the two camps had bought up so much stock that almost none was left for sale.

The shorts needed shares. The shares were gone.

Over four days in May, Northern Pacific went from $110 a share to $1,000. It took about 17 hours of trading.

To raise cash, the shorts sold whatever else they owned. The rest of the market fell with them. May 9, 1901 became one of the first modern stock panics.

It ended only when the two camps made a deal and let the shorts off the hook.

The shorts weren't wrong about the railroad. It was a twice-broke line. They were wrong about the exit.

Today the crowd sits in index futures, not one railroad. Nobody can corner an index. But anybody can crowd one.

Being right isn't the same as getting out. We'll see.