Gold pullback risks mount

To put this into perspective, since 1974 the legal limit in stock markets has been 2x. Near 24x, each dollar deployed in gold futures has 24x the price impact on gold as a dollar invested outright! And the risks of being wrong are dreadful, as a mere 4.2% gold price move against specs’ bets would wipe out 100% of their capital risked. That forces these guys to maintain myopic ultra-short-term trading time horizons.

Rather than months or weeks, such extreme leverage demands focusing on days or hours. No one can afford to be wrong for long at 24x, which is exceedingly-unforgiving. Gold-futures specs have to react to gold-moving news very quickly, shifting capital fast. And their most-watched catalysts have long been major US-economic-data releases, including both consumer and wholesale inflation and monthly jobs reports.

The latest iteration of the latter was released a week ago on Jobs Friday, proving terrible. The Bureau of Labor Statistics claimed the US economy only created 142k jobs in August, under the +161k expected. On top of that prior-two-month revisions were really negative, slashing away 86k previously-claimed jobs! I’ve long believed revisions should be added to the current headline, which would’ve crushed it to just +56k.

The internals proved ugly too. Last month crappier part-time jobs soared 527k, while superior full-time jobs collapsed 438k. In addition 1,325k native-born Americans left the payrolls, while 635k foreign-born people took jobs. That rare headline nonfarm-payrolls miss was also the second in a row after the BLS originally reported July at +114k versus +185k expected. All that was certainly dovish for Fed rate cuts.

Speculators have long aggressively bought gold futures on Fed-dovish jobs reports, implying labor-market weakness. Their leveraged buying drives big gold gains on those Jobs Fridays, from 1% to 2%+. Yet this past Friday, gold’s initial post-data surge was constrained at just +0.5%. Then it reversed to plunge 1.5% in intraday trading before rebounding some to a -0.7% close. Such abnormal weakness spawned some worries.

I heard from plenty of our newsletter subscribers wondering why the heck gold sold off after such ugly monthly jobs data. There was some Fedspeak interpreted as more hawkish than traders expected that day, which contributed. But the main reason gold didn’t surge to another record close was specs’ gold-futures positioning is overextended. They don’t have much available capital firepower left to do big buying.

While these guys wield outsized influence over gold prices, their ranks are small. Only a tiny fraction of all traders are bold enough or foolish enough to run 10x, 20x, and even sometimes 30x+ leverage on volatile gold. And the amount of capital they collectively control is also really little compared to broader markets. Gold-futures speculators can only do so much buying before they exhaust their capital firepower.

The best way I’ve found to infer how much of that is likely available or has been expended is looking at specs’ total longs and shorts compared to their trading ranges in recent years. This chart superimposes gold over both, showing whether spec longs and shorts are relatively-high or -low. Gold’s pullback risks are really mounting given their overall positioning today. That implies a buy-lower opportunity is nearing.

Gold’s current upleg has proven a remarkable record breakout surge, soaring a mighty 38.7% at best in 10.8 months as of late August! Gold’s first new nominal record in 3.3 years was achieved back in early December, and gold has powered to 28 more since then. A year ago $2,050 would’ve seemed really high, yet now $2,500 is starting to feel normal! Some unique drivers contributed to gold’s powerful upleg.

Its centerpiece was blasting up 20.0% between mid-February to mid-April. Gold-futures buying certainly played a role in that, with total spec longs soaring 67.3k contracts higher. Yet all that came during the first half of gold’s remarkable breakout surge, then gold-futures buying vanished. As analyzed extensively in our subscription newsletters, Chinese investors and world central banks took the gold-buying baton.

Due to gold-futures speculators’ finite capital, uplegs driven solely by them typically enjoy 20%-to-25% gains at best. This 39% one challenging 40%+ monster status has grown much larger thanks to atypical outsized buying from Chinese and central bankers. But considered across this entire gold upleg, specs’ gold-futures buying has still been massive. They continued to dominate gold over much of this upleg’s life.

Measured from gold’s early-October bottoming to late August’s latest interim high, they added a colossal 124.4k long contracts while buying to cover another 86.0k short ones. Combining for 210.3k, that was the equivalent of 654.2 metric tons of gold! But specs’ gold-futures positioning is only reported weekly in the Commitments of Traders reports, current to Tuesday closes that don’t perfectly match gold’s troughs and peaks.

If total spec buying is instead considered from the CoT Tuesdays closest to gold’s upleg-to-date span, their buying grows to 143.7k longs and 82.4k shorts. That combines for 703.2t in gold-equivalent terms! This is enormous by any standard, specs have been huge gold-futures buyers. They have done so much buying that their probable capital firepower available for fueling this gold upleg has mostly been expended.

Major gold uplegs are birthed by frenzied gold-futures short covering. Back in early October spec shorts had soared to a lofty 174.4k contracts, which wasn’t far under late September 2022’s 3.8-year high of 185.3k! In recent years spec shorts have found strong secular support around 95k contracts. They were already covered down to that point by late January 2024, stalling before oddly falling much lower into late June.

That was a deep 4.1-year secular low, implying specs have fully exhausted their short covering. Indeed those contracts have rebounded near their support since. While spec shorts can go lower, precedent argues they aren’t likely to and if they do such extremes won’t last long. Specs have vastly more room to short sell aggressively from here than keep buying, which is short-term-bearish for gold and somewhat worrying.

Spec longs are way more important for gold’s fortunes. Since this gold upleg was born, on average spec longs have outnumbered shorts by a whopping 3.3x! So longs have proportionally-more influence on gold’s short-term price action. Back in early October when gold bottomed, total spec longs were really low at just 264.8k contracts. But since then they soared as high as 408.5k in late August, challenging resistance.

That has run around 415k contracts in recent years. While total spec longs can briefly surge higher on excessive greed and euphoria, such spikes reverse sharply within weeks. Once total spec longs hit and exceed their secular resistance, probabilities heavily favor an imminent reversal. Largely out of capital firepower to keep buying gold futures, speculators have far more room to sell their excessively-bullish bets.

Triggering big mean-reversion long dumping and serious shorting usually requires some notable catalyst, like Fed-hawkish inflation-data or jobs upside surprises. A recent case in point was Jobs Friday in early June, when the BLS reported the US added a huge 272k jobs in May compared to +190k expected. Gold plummeted 3.6% on close that day, its worst daily loss in 3.6 years! But some China-central-bank data contributed.

With the Fed’s long-awaited and hyper-anticipated next rate-cutting cycle recently increasingly targeted for starting at next week’s FOMC meeting, specs haven’t felt compelled to dump gold futures. So gold has remained more resilient than it ought to given such lopsided gold-futures positioning. But downside risks abound while spec shorts remain super-low and spec longs really-high, big selling could erupt anytime.

Despite the Fed guaranteeing a maiden cut next week, that FOMC meeting could still spark major gold-futures selling. With every-other FOMC decision including this imminent one, the Fed releases its top officials’ projections for its federal-funds rate in coming years. The last dot plot published in mid-June implied just one 25-basis-point rate cut in 2024 followed by another four in 2025, for 125bp of projected cuts.

The new dot plot will almost certainly include more than five cuts by the end of next year, but maybe not enough for traders. Their futures-implied rate cuts exiting August totaled 100bp in 2024 before another 108bp in 2025, or over eight total cuts! If top Fed officials only see six or seven, the gold-futures guys might not like that. Especially if currency traders consider it Fed-hawkish-enough to really bid up the US dollar.

Periodic mid-upleg pullbacks are healthy and essential to rebalance sentiment, bleeding off greed before it soars to upleg-slaying extremes. They typically force gold at least back to its 50-day moving average, which is $2,445 midweek. But in both late June and late July, gold fell well under its 50dma. Gold could easily plunge all the way to lower support of this past summer’s high-consolidation trading range, or $2,300!

That would make for a large 8.9% total pullback since late August, nearing 10%+ correction levels. I’m not forecasting that, but it wouldn’t be unusual at all. Gold and gold-stock speculators and investors need to be ready for a bigger selloff given specs’ excessively-bullish gold-futures positioning. That includes having trailing stop losses on trades and preparing shopping lists for new trades as gold starts bottoming.

Yet despite specs’ mostly-expended gold-futures buying, they don’t necessarily have to sell soon. If gold keeps powering higher on balance on Chinese-investor and central-bank buying, specs might not be spooked into dumping. If top Fed officials signal more, bigger, or faster rate cuts than traders expect, the gold-futures guys’ worries that could spawn selling will wane. One more factor is gold’s biggest wildcard of all.

Extraordinarily American stock investors who usually play a major role in fueling monster gold uplegs have been totally missing in action during this one! I wrote a whole essay last week analyzing that and why gold’s demand comeback will be super-bullish for it. Sooner or later gold will have powered high enough for long enough to finally attract back American stock investors to gold-ETF shares, accelerating gold’s upside.

The major gold ETFs act as conduits for the vast pools of stock-market capital to slosh into gold, and that can easily overpower whatever the way-smaller gold-futures speculators are doing. So this gold upleg could still surge considerably higher before inevitable mean-reversion gold-futures selling kicks in. Yet still gold-selloff risks are high when total spec longs are excessively-high and total spec shorts are super-low.

Instead of being feared, healthy pullbacks create the best mid-upleg buying opportunities. Especially in gold stocks, which leverage gold’s material price swings around 2x to 3x. Thanks to these high prevailing gold prices, gold-mining profits have soared to spectacular record heights! Gold-stock prices still need to double to quadruple to reflect $2,500 gold, let alone where it is heading as this powerful secular bull grows.

The bottom line is gold pullback risks are mounting. Speculators’ super-leveraged gold-futures trading is the dominant driver of gold’s short-term price action. These traders have done massive buying during this upleg, leaving their shorts well under secular support and longs challenging secular resistance. Such positioning implies their likely capital firepower for buying has largely been exhausted, which is gold-bearish.

Any day some catalyst could spark major gold-futures selling which could quickly snowball, slamming gold into a larger pullback. Especially if the US dollar surges big on some Fed-hawkish news. But this doesn’t guarantee an imminent gold selloff. If gold continues rallying on balance with Chinese investors, central banks, or even American stock investors soon buying, specs shouldn’t be in any hurry to dump gold futures.

(By Adam Hamilton)