The ten-year Treasury yield touched 5.35% last week, its highest since 2002. On September 16 the Fed raised rates for the first time since 2023, and yesterday’s hot inflation report put another hike in play for October.

We have been working through the conditions that tend to show up before a major market top: stretched valuations, record stock ownership by households, rosy earnings forecasts, growth running hot enough to stoke inflation. Each one is necessary. None is sufficient. This column adds one more piece of evidence, and it may be the most useful: what the dollar does while rates are rising.

The reason is liquidity, the supply of easy money and credit sloshing around the world. Liquidity is mother’s milk for markets. When it flows, investors pay up; when it dries up, everything gets repriced. The Fed controls the faucet at home, but the dollar tells you whether the drought is spreading worldwide.

First, a caution. One rate hike is not a tightening cycle. But the bond market isn’t betting on one and done. After yesterday’s inflation report, futures traders put better than 95% odds on another hike by year-end and 40% odds on two, double the day before. Once a single hike starts to look like a series, the dollar matters a great deal more.

Think of the dollar as a check-engine light. It doesn’t cause the problem or tell you when the car will stop. It warns that the system is under strain.

Which dollar? Not the one on the evening news. That index, DXY, is mostly the euro and the yen. The better gauge is the Fed’s broad index, which covers 26 currencies weighted by how much America trades with each country. Right now DXY is at an 18-month high while the broad index is below last November’s level, mostly because the euro and yen are falling for their own reasons.

“Fed hikes hurt stocks” is too simple, and often wrong. The table shows the last eight tightening cycles. What separates the bad ones from the mild ones is not how much the Fed raised. It is whether the broad dollar rose while stocks were falling.

The worst damage, in 2022, came with the biggest dollar surge. The mild pullbacks of 1994 and 2006 came with barely any dollar move. Hikes confirmed by a rising dollar are the dangerous ones, because American tightening has become a global squeeze.

Here is how that squeeze works. Foreign governments and companies owe trillions in dollars, and a stronger dollar makes those debts heavier. Oil, copper and wheat are priced in dollars, so they get pricier for everyone else. American companies’ foreign sales shrink when converted back home. And money leaves riskier markets for the safety of U.S. Treasuries.

The dollar isn’t a trading rule; it missed the mortgage risk building in 2006. But as a single reading of global liquidity, it is one of the best we have.

So where are we? In July, Fed Chair Kevin Warsh held rates and told the market to figure out the path for itself. It did. The ten-year yield rose more than 0.8% in the third quarter, the biggest quarterly jump this century. He turned hawkish at Jackson Hole and delivered the September hike.

The dollar firmed with him. The broad index is up more than 3% from its September low. That is a fraction of the 11% surge after Russia invaded Ukraine in 2022, but it is broad, and it is moving the wrong way for stocks.

The other warning lights are on, too. The S&P 500 ended September within 1% of a record, but a handful of giant companies are carrying it. Weight all 500 stocks equally and the index fell seven straight weeks, a streak seen only in 2002 and 2022. The typical stock is 16% off its high, and the extra yield investors demand for junk bonds is the widest since April.

Until last week the check-engine light was steady. Steady means get it looked at. Flashing means pull over. It has started to flash.

None of this calls for selling everything. It is the kind of moment disciplined investors recognize - when what has run furthest gets a second look, and portfolios drift back toward their targets. If the dollar settles down, stocks still face higher rates, but the case for a deep, worldwide decline weakens. If it keeps climbing while the Fed keeps hiking, we have one more necessary condition checked off.

The Fed’s rate decision is the headline. The dollar is the diagnostic.

Share

In 2021, when nobody wanted anything to do with offshore drilling, Noble bought a rival, Pacific Drilling, and sent two of its ships to the scrapyard: the Pacific Bora, built in 2010, and the Pacific Mistral, built in 2011. Rig-data firm Esgian puts the proceeds at $14.5 million and $10 million. Today a comparable ship rents for about $461,000 a day, on Transocean’s latest contracts. The Mistral went for roughly three weeks of rent.

That is what the bottom of a cycle looks like. Our view: a decade of neglect has turned offshore drilling into a genuine investment opportunity, and smart money is already buying.

The ships that matter are drillships, self-propelled vessels that work in the deepest water. S&P Global counted 119 in early 2017. Westwood Global Energy put the working fleet at 89 at the end of 2025 and expects about 93% of it to be busy this year. The only new supply is a few half-finished hulls stuck in Asian shipyards. A quarter of the fleet was cut up within a decade, the industry was cut to the bone, and nothing is being built to replace it.

So where did the money go? Into shale, the wells drilled on land in places like Texas. Shale wells come online in months and need no billion-dollar ships.

The catch is that they fade fast. The International Energy Agency (IEA) studied roughly 15,000 fields and found that if shale investment stopped, output would fall more than 35% within a year. Call it the shale treadmill: you run hard just to stay in place. And it keeps getting steeper. The Dallas Fed’s survey says producers now need $66 a barrel to profit from a new well, up from $49 in 2020. Rystad puts deepwater costs at about the same level, but a deepwater well produces for years.

The IEA says nearly 90% of supply spending since 2019 has gone to replacing what existing fields lose. If investment stopped today, the world would lose 5.5 million barrels a day each year, roughly what Brazil and Norway produce combined. Yet oil spending is falling for a third straight year.

Shale also no longer answers when prices call. After Russia invaded Ukraine in 2022, US output jumped by a million barrels a day within a year, according to the Energy Information Administration (EIA). This time, output was 13.9 million barrels a day before the Iran war began, and the EIA estimates it was still 13.9 million last quarter. Seven months of oil above $90 have produced no growth at all. The best land has been drilled, and the majors haven’t raised budgets. The quick barrels are gone.

So the industry is heading back out to sea. More than 70% of recently approved conventional oil projects are at sea, according to the IEA, with very deep water off Brazil, the US and Guyana doing the heavy lifting.

Reaching those barrels takes drillships, and nobody can simply order more. Transocean’s last two drillships ordered without a customer were priced at $540 million apiece in 2014; by 2020 the pair was expected to cost about $2.25 billion. Westwood now puts a new one above $1 billion and at least three years to build.

Westwood reckons a new ship needs $650,000 to $700,000 a day in rent to earn a decent return. Noble says the best ships fetch in the mid-$400,000s. Rates would have to rise by roughly half to justify an order, and shipyards, burned by last cycle’s cancellations and full of other work, aren’t eager anyway.

Meanwhile, the work is coming back. Noble’s CEO said the industry signed the most deepwater contracts in the first half of 2026 “in well over a decade,” and that 95% of the available fleet is booked. Transocean’s takeover of Valaris will leave fewer owners competing for it.

Rents stay flat only if oil demand falls. The EIA expects it to be higher in 2027 than before the war, even after a year in which the Strait of Hormuz was effectively closed. The only quick source of extra ships is the mothballed fleet referred to as “cold stacked” which Westwood says costs $85 million to $90 million a ship just to wake up.

The stock market hasn’t noticed. When Transocean and Valaris announced their deal, the pair was worth about $17 billion, debt included. That buys 33 drillships plus 40 other rigs. At $1 billion apiece, the drillships alone would cost twice that to build. Noble is valued at about $8 billion for 24 floating rigs, and Seadrill at about $3.4 billion for nine drillships plus a few others. By the industry’s own estimates, the public drillers trade at roughly a quarter to a third of what it would cost to rebuild their fleets. Picture buying a house for a third of its cost to build, in a town where the builders are booked for years.

Someone has noticed, though. We’ve been following Eldorado Drilling, a Norwegian company controlled by Darren Maupin’s Pilgrim Global. This year it paid $300 million for the drillship Deep Value Driller and $257.6 million for Vantage Drilling, buying scarce ships at a fraction of replacement cost.

It’s tempting to credit the war, but that gets it backward. The headlines didn’t cause this shortage; they exposed it. We see the same pattern from copper to rare earths: years of low prices starved investment, and lost capacity can’t be rebuilt quickly.

The odds favor whoever already owns the assets. The world spent a decade pouring money into wells that fade within a year, and the places it must turn to next can only be reached by a fleet it spent that same decade scrapping. Noble collected $10 million for the Pacific Mistral in 2021, and at the time it probably seemed sensible. Real contrarian investing means taking the other side of the trades that seem sensible.

Share

Subscribe now

The loudest claim about artificial intelligence is that it will take jobs quickly. The more useful historical lesson is that a technology can make work less necessary long before it can make workers disappear.

Consider the caboose.

For more than a century, the small car at the rear of an American freight train was part office, part dormitory, part lookout post, and part safety system. It carried the conductor and rear-end crew, often including a brakeman and flagman. It was also, eventually, a rolling symbol of the fight over whether technology’s gains belonged solely to employers, or had to be shared with workers.

That is a story worth remembering as companies announce ever more ambitious plans to use AI.

Before George Westinghouse invented his automatic air brake in 1869, stopping a freight train was a brutal labor problem. Brakemen clambered across moving freight cars, often at night and in bad weather, responding to whistle signals from the locomotive. They turned the handwheel on each car to apply its brakes. The system was slow, dangerous, and prone to catastrophe.

Westinghouse’s invention changed the economics and safety of freight railroading. Compressed air could apply brakes across a train rapidly and, crucially, automatically if the train broke apart. The technology removed the original and most perilous reason for large numbers of brakemen: men were no longer needed to sprint across car roofs and manually set brakes.

Yet the caboose did not vanish in 1870, 1880, or even 1920. It remained standard on American freight trains well into the 1980s.

That was partly because a railroad crew did more than brake a train. Crews switched cars, coupled and uncoupled equipment, inspected trains, protected stopped movements with flags and flares, kept records and handled waybills. The caboose offered a place to work, eat, and sleep on long runs. It gave a crew a vantage point over the train at a time when overheated bearings, shifted loads and broken equipment could be detected only by people looking, smelling and listening.

But technology kept narrowing the role. Automatic signals, radios, centralized dispatching, roller bearings, wayside detectors, and computerized paperwork weakened the old case for a crew at the rear of every train. By the late twentieth century, an electronic end-of-train device could report brake-line pressure and confirm that the rear car was intact and moving. The device did in a few pounds what a 25-ton caboose once did imperfectly.

The caboose then became a rolling exhibit in the debate over “featherbedding”, the pejorative term for union rules that force an employer to retain more workers than management believes necessary. Railroad executives saw a costly relic carrying employees whose central tasks had been replaced by machinery. Unions saw a safety platform, a workplace, and a negotiated claim on the productivity gains that machinery created.

Both views had elements of truth.

“Featherbedding” was never a neutral description. It was management’s word for work it did not want to pay for. Railroad workers could reasonably respond that someone still had to deal with a disabled train, a shifted load, a broken coupler, or a signal failure. The fact that a machine could perform a routine task did not settle who bore the risk when routine failed.

Railroads could not simply declare the caboose obsolete. They had to bargain. They had to revise work rules, overcome state laws requiring cabooses or minimum train crews, and compensate workers whose careers and seniority systems had been built around a different operating model.

The end came only when electronics, economics, labor agreements and regulation aligned. End-of-train devices appeared in the late 1960s; Florida East Coast had eliminated cabooses by 1972. But major railroads did not broadly shed them until the 1980s, after protracted crew-consist negotiations. Virginia, the last state with a caboose requirement, repealed it only in 1988.

That history should temper the current AI panic.

Through September’s employment report, the evidence for an economy-wide wave of direct AI job destruction remains thin. The Bureau of Labor Statistics reported payroll growth of 29,000 in September, unemployment of 4.2 percent and little monthly change across most major industries. Those are not especially strong numbers, but neither are they evidence of a sudden AI layoff shock.

What the data show instead is a low-hire, low-fire economy. Information employment fell by 10,000 in September, financial activities lost 7,000, and professional and business services lost 9,000. Those declines may reflect cost cutting, slower demand and restructuring. AI may be contributing at the margin, but national data cannot yet cleanly attribute the losses to it.

AI may nevertheless be changing the labor market at the edges. It can make managers more cautious about entry-level hiring, permit firms to leave vacancies unfilled and reinforce pressure to economize on back-office work. Yet Stanford researchers recently found little evidence that AI is causing broad job losses or materially reducing employment and postings in the occupations most exposed to it.

In other words, hiring may have slowed in certain white-collar functions. But many workers are still comfortably in the caboose.

The caboose lesson is not that unions can stop technology forever. They cannot. It is that invention is not adoption, and adoption is not instant displacement. Between a capable technology and a vanished job stand contracts, regulations, safety rules, liability and organizational inertia. Above all, workers are not merely inputs to be deleted from a spreadsheet. They are voters and citizens, with a say at the job site, in the union hall and at the ballot box. If AI’s gains are concentrated among corporate owners while the costs fall broadly on workers and communities, the political response may prove as consequential as the software itself.

Westinghouse’s air brake made the caboose less necessary. It did not make the caboose disappear. That took a century of further innovation and decades of bargaining. AI may move faster. But history suggests the “jobs are gone tomorrow” forecast is more sales pitch than timetable.

Subscribe now

Share

Anyone who has followed our work knows we write about gold a lot. That is not accidental. Gold has been a large part of how we think about protecting purchasing power, at times as much as a quarter of a portfolio over the last five years. We treat it as insurance against the slow leak of currency debasement - and, when the setup is right, as a position worth managing actively.

We have written about Bitcoin too. We have owned it, reduced it, and generally refused to turn it into a religion. Bitcoin inspires unusually strong opinions. We would rather let the data decide.

Right now, the data are sending an unusually simple message.

We tend to look at an investment through three lenses: the economic story, the models, and the charts. Today, all three point in the same direction.

What Bitcoin Is Acting Like

For years, the biggest question around Bitcoin has been what it actually is. Digital gold? A volatile technology stock? A currency? A speculative toy that moves with easy money?

We still do not know. We do not think investors need to know.

What matters is how it has been trading. Over the past six months, Bitcoin has moved more in line with gold than with the broad stock market. In plain English, when gold has zipped or zagged, Bitcoin has more often followed gold than the S&P 500.

That does not make Bitcoin “digital gold.” Correlations change, sometimes violently. And one difference remains enormous: central banks keep buying physical gold. They have not, in any serious way, adopted Bitcoin as a reserve asset.

Markets can change before the story does. For the moment, we think Bitcoin is behaving more like a monetary asset and less like another high-octane tech trade. The phrase “for the moment” matters. If Bitcoin starts trading again like a software stock on steroids, that conclusion goes back on the shelf.

The Models And The Charts

The second lens is quantitative. We use models that combine liquidity, sentiment, momentum, and a few fundamental inputs. They are not crystal balls. No model is. Their job is simpler: tilt the odds and keep emotion out of the room.

Our models are in a buy.

Then there are the charts, which may be the cleanest part of the setup. Bitcoin recently pushed through about $82,500, a level we had been watching. It is now around $86,000 and trading above its 50-, 100-, and 200-day moving averages - the simple trend lines many investors use to ask whether price is still heading the right way.

Over the past few years, Bitcoin has largely been driven by headlines around regulatory developments. We do not try to trade these headlines, but it is hard to ignore that this breakout happened following what was seemingly bad news with the Clarity Act failure.

Even more interesting is Bitcoin versus gold. After a long stretch of lagging, the Bitcoin-to-gold ratio looks as if it is breaking higher from a base. Bitcoin is not merely rising. It is starting to outperform the asset that has been one of our highest-conviction holdings.

That matters.

How Much Is Enough?

Our approach is to use market size as a useful starting point, not a law of nature.

Gold is roughly a $30 trillion market. Bitcoin is roughly $2 trillion. If a 25% gold allocation is comfortable in a hard-asset sleeve, scaling by relative size lands near 1.7%. Call it 2%.

That is not a commandment. Inside a broader “hard currency” bucket, we think it is a sensible first cut: gold remains the main position; Bitcoin gets a meaningful but much smaller weight. Conviction can then push the allocation up or down.

Could this be wrong? Absolutely. Bitcoin has humbled plenty of investors. A breakout can fail in a hurry. That is why the exit matters as much as the entry.

Our line in the sand is about $82,500. Above it, we own some Bitcoin. If price breaks decisively back below that level, we trim or get out and reassess. We do the same if Bitcoin starts acting like a high-beta tech stock again.

From about $86,000, our line is only 4% to 5% lower. The upside is not nearly as well defined. The setup is not certainty. It is asymmetry: a relatively small, clear invalidation versus an open-ended reward if the trend holds.

Maybe Bitcoin is becoming digital gold. Maybe it is not. We do not need to settle the philosophy seminar today.

We may let conviction move the weight from 2% but will watch our line of $82,500 and the correlations to tell us when the thesis may have changed.

Share

On the morning of September 7, 2010, a Chinese fishing trawler rammed two Japanese coast guard ships near a cluster of disputed islands, and the next day Japan arrested its captain. Beijing demanded his release. Tokyo refused. Within about two weeks, shipments of an obscure family of metals called rare earths were reported stalling at Chinese customs, with no ban announced and none needed. Japanese industry, which bought nearly 90% of these metals from China, went into a panic. Japan freed the captain on day seventeen. The price of those metals rose roughly tenfold over the following year.

That is what it looks like when one country controls a material the rest of the world cannot function without. In 2010 the lesson was Japan’s to learn. In 2026 it is ours, and this time the stakes are larger than car factories.

Start with what these materials are. Critical minerals are the raw ingredients of modern hardware: rare earths for the powerful magnets inside every electric motor, gallium for radar chips, graphite for batteries, tungsten for armor and drill bits. The list runs to 60 commodities on the US government’s official register, and they hide inside everything from an iPhone’s speaker to a fighter jet’s targeting system. Per Pentagon figures, a single F-35 contains more than 900 pounds of rare earth materials. A Virginia-class submarine contains 9,200 pounds. There is no substitute at any price; the physics of magnetism does not negotiate.

Pause on the economics of that dependence, because it is the engine of the whole investment case. Critical minerals are a vanishingly small share of the cost of the products they make possible, a point the IEA makes explicitly in its 2026 outlook. The rare earth content of an F-35 is a rounding error on the program that builds it. But without it, the jet does not fly. An input that is both irreplaceable and cheap relative to the finished product creates the least price-sensitive buyer in capitalism: a defense contractor choosing between paying ten times more for magnets and delivering no aircraft at all will pay ten times more every time, because the alternative is not a thinner margin but no product. Demand for these materials does not fall when prices rise. It simply pays. That is the structure in which prices move in multiples rather than percentages.

Here is the part most investors get wrong. The problem is not who owns the mines. Rare earth ore comes out of the ground in California, Australia, and a dozen other places. The problem is what happens next. Raw ore is useless until it is chemically separated and refined into pure material, an industrial process China spent thirty years mastering while the West outsourced it as dirty, low-margin work, the kind of acid-soaked, tailings-pond business the environmentally conscious West long ago decided it would rather not live next to. Think of wheat and flour: it does not matter how many wheat fields you own if one country owns nearly every flour mill on earth. China mines about 61% of the world’s rare earths but refines roughly 90% of them, per the International Energy Agency, and the agency’s 2025 outlook found China was the top refiner for 19 of the 20 strategic minerals it tracks. When new US export controls hit in April 2025, America had no refined production of any of the seven restricted rare earths.

Why does this matter now, after decades of the same dependence? Because Beijing stopped treating the monopoly as a business and started using it as a weapon, in writing. In December 2024, China barred exports of gallium, germanium, and antimony to the United States. In April 2025 it imposed licensing on seven rare earths and their magnets; exports of those magnets halved within a month, and Ford shut its Explorer plant in Chicago for a week because it could not buy a component that costs a few dollars per car. In October 2025, China claimed the right to license any product made anywhere containing even 0.1% Chinese-origin rare earth content. Then came a truce. After Trump and Xi met in Busan that month, Beijing suspended the October rules and the US-specific mineral ban for a year, and Washington paused a rule of its own in return. But the April licensing regime never stopped, and the suspensions lapse in November 2026 unless both sides extend them. China repealed nothing; it pressed pause. Embargoes fade. Laws and licensing bureaucracies wait.

Now the money. When supply of an irreplaceable material gets restricted, the price does not rise politely. Tungsten rose sixfold between January 2025 and April 2026, per the IEA. Antimony jumped 144% in 2025, per the US Geological Survey. Gallium and heavy rare earths already trade in Europe at roughly five times Chinese domestic prices, because two markets now exist for the same atoms and Western buyers can only shop in the expensive one. The 2010 episode, a tenfold move triggered by one fishing boat, marks how far prices can travel. And governments have put a floor under the downside: the Pentagon guaranteed the leading US producer a minimum price of $110 per kilogram for its key product for ten years, roughly $50 above the Chinese market price at signing, and Japan wrote the identical floor into its Australian supply contract in March 2026. Washington and Tokyo are paying producers to exist. That is not a subsidy; that is an admission.

Do not mistake the headlines about American projects for a solution. They are the beginning of one, which is a different thing. MP Materials in California can supply about 11.5% of the world’s rare earths and already separates the lighter ones, but it is only now building the heavy rare earth separation and magnet-making steps that Beijing’s April controls target. Lynas became the first company outside China to separate heavy rare earths in May 2025, one facility against an industry. The honest benchmark is Japan, which started this race in 2011 with state money and total urgency and, fifteen years later, has cut its dependence on China from 90% to somewhere between 60% and 70%. Refineries take years to permit, years to build, and years to master. Every one of those years belongs to the incumbent.

The counterargument is a lasting Washington-Beijing settlement that unwinds the export controls and merges the two price tiers. But Beijing has tightened these rules every year since 2020, and it kept tightening during the Busan truce, adding compounds to the control list in January 2026 and passing a supply-chain security law in March. The pause is a delay, not a reversal. And even a real settlement would not undo what the past two years taught Western governments, which are now stockpiling and paying for supply that does not route through China. While there is always the risk that China will temporarily loosen its restrictions in an attempt to inhibit the West’s drive to build refining capacity, a buyer who has been cut off once does not go back to buying just in time.

That is the investment case in one sentence: until the West builds enough refining capacity to meet its own needs, China sets the price and controls the supply, and everything the West builds in the meantime, every jet, every drone, every data center, bids for material at whatever price the monopolist will accept. In 2010 a fishing boat proved the flour mills could be closed. They have barely changed hands since.

Share Bob Haber

For the better part of two decades, commodities were the asset class investors visited the way one visits an elderly relative: briefly, reluctantly, and only when circumstances demanded it. Inflation flares, oil spikes, a war breaks out somewhere inconvenient, and suddenly everyone remembers that barrels and bushels exist. The crisis passes, the money rotates back into whatever software company is promising to disrupt the concept of profit, and the physical world goes back to being ignored.

Why own copper when you can own a stock that goes up every day?

That logic worked until it stopped working, which is how all such logic ends. We believe the ground has shifted beneath the market’s feet, and that commodities now deserve something they have not been granted in a generation: a permanent seat at the portfolio table. Not a trade. Not a hedge to be slapped on when CNBC gets nervous. A structural allocation - for most investors, somewhere between 10% and 20%.

This is not a call to sell your stocks, buy a bunker, or convert your net worth into Krugerrands. It is a call to recognize what commodities actually are.

The Stuff Beneath the Stuff

Stocks are claims on businesses. Bonds are promises to pay you back later, hopefully in money that still buys something. Real estate is dirt with paperwork. Commodities are more elemental than any of these; they are the raw inputs without which nothing else functions. Oil and gas to move people and power industry. Copper, aluminum, and steel to build the grids, homes, and data centers everyone suddenly wants. Uranium to keep the lights on. Grain to keep 8 billion people fed and, more importantly, calm.

Here is the part the market keeps missing: a faster computer does not reduce the need for electricity. Artificial intelligence does not run on vibes. It runs on chips, wire, cooling water, concrete, and staggering quantities of baseload power. A nation cannot rearm or reshore its industrial base with an app. The more digital the world becomes, the more physical infrastructure it demands underneath - a truth so obvious it has been almost entirely ignored.

Gold Got the Memo. Nothing Else Did.

Let us be clear: this is a different story than precious metals (in fact, we view it as a completely separate asset allocation). Gold made (what we believe is) a first move, and a powerful one, as investors slowly absorbed the implications of runaway deficits, currency debasement, and a world order held together with duct tape. Silver came along for the ride.

The rest of the complex? Left at the station.

Industrial metals remain remarkably cheap relative to gold. Energy is cheaper still. An ounce of gold currently buys roughly 44 to 46 barrels of crude, depending on your preferred flavor of benchmark. That is down from the absurd 79-barrel print seen at the extreme, but it remains historically stretched. In 2005, that same ounce bought about seven barrels.

Read that ratio carefully, because it is telling you something. The market has figured out that money is losing its purchasing power. It has not yet figured out the value of the things money must ultimately buy.

A Century of Neglect, Priced In

Compare the broad commodity complex to the Dow Jones Industrial Average and you find the ratio sitting near 100-year lows. One hundred years. Commodities have rarely been this despised relative to financial assets in the modern era.

Charts guarantee nothing, and anyone who tells you otherwise is selling a newsletter (ahem). But starting valuations matter, and they matter most at extremes. The last few times physical assets were this unloved - the 1930s, the 1970s stagflation, the early 2000s - the cycles that followed were violent and long. The pattern repeats for a simple reason: commodities can be ignored for years, but eventually the world rediscovers that it cannot live without them.

Meanwhile, look at what the market has chosen to cherish. Enormous valuations for the companies that consume raw materials. Rounding-error valuations for the companies that produce them. Technology, communications, and finance dominate the indices while energy and materials - the sectors supplying the fuel and metal that make everything else possible - have been squeezed to a sliver.

That arrangement is unstable. Unstable arrangements resolve.

The Scarcity Machine

Cheap valuations alone don’t make a bull market. Collisions do, and one is coming between constrained supply and relentless demand.

Start with supply. The extractive industries spent the past decade in penance - punished by shareholders for past sins of overexpansion and ordered to return capital rather than deploy it. A new copper mine takes a decade or more from discovery to production. Pipelines, refineries, and power plants require permits, skilled labor, and patience, three things in structurally short supply. The capacity that should have been built during the era of free money simply wasn’t. When demand arrives, supply cannot be summoned by next quarter’s earnings call. Price does the summoning.

Now the demand side, where three freight trains share one track:

Rearmament. Governments are spending on defense at levels not seen in decades, and war machines are made of steel, aluminum, explosives, and fuel - not PowerPoint.

The AI buildout. Strip away the breathless narrative and the AI boom is a hardware and power boom wearing a software costume. Data centers demand electricity around the clock and consume copper, concrete, and cooling capacity in quantities the grid was never designed to provide.

Emerging Markets. The IMF has the developing world growing at more than twice the developed world’s pace, 3.8% vs. 1.7% this year. Economies like India, with 1.4 billion people, are entering the resource-intensive phase of development, where even modest per-capita increases in steel, copper, and energy compound into enormous demand. And many are commodity producers themselves – higher prices make them richer, which funds the infrastructure that consumes more of what made them rich. We have seen this movie before. It was called China, and it drove the greatest commodity bull market of our lifetimes.

More demand. Less supply. And a world newly willing to pay for security of supply rather than merely the lowest sticker price. Do the arithmetic.

Long Paper, Short Steel

There is a harder edge to this story, and it is geopolitical. Step back and look at the West’s aggregate position as if it were a single trading book. We are long financial assets - equities, bonds, private credit, claims stacked on claims - and structurally short the physical inputs required to grow, build, and defend ourselves. Our adversaries run the opposite book.

While Western capital spent two decades financializing everything in sight, China spent those same decades cornering the physical world. It dominates the refining and processing of critical minerals, controls vast swaths of rare earth supply, stockpiles crude and copper with strategic intent, and has locked up mines and ports across Africa and Latin America. Russia, for its part, sits on energy, fertilizer, and grain - and has demonstrated a willingness to use all three as weapons. The West outsourced its supply chains to the very nations most likely to squeeze them, then declared victory because the stock market went up.

Anyone short a commodity eventually meets the person who owns it, and the terms of that meeting are rarely pleasant. Every export restriction on rare earths, every gas cutoff, every embargo threat is a margin call on the West’s short position. Covering that short - rebuilding mines, smelters, refineries, and stockpiles at home and among allies - will take a decade and trillions of dollars, and every bit of it is demand for the commodities themselves. Owning physical assets is no longer just an investment thesis. It is a hedge against being on the wrong side of the most crowded trade in geopolitics.

The Dollar Does The Rest

There is one more tailwind, and it blows from Washington. The United States - along with most of the developed world - is running wartime deficits during peacetime while carrying debt loads that would make a Habsburg blush. The menu of escapes is short: grow out of it, cut spending, raise taxes, default, inflate, or quietly rob savers through financial repression.

We invite readers to guess which options politicians will choose. Growth, inflation, and suppressed real rates are the path of least resistance, and that combination is poison for cash and long bonds but rocket fuel for real assets. A softer dollar sweetens the deal further: since most commodities are priced in dollars, a weaker greenback makes them cheaper for the rest of the world to buy, and dollar prices tend to rise accordingly.

What We Own and Why

We prefer broad, diversified exposure over hero trades. Nobody rings a bell at the bottom of the uranium market, and we are not in the ticker-guessing business. Our preferences:

• Industrial metals, copper and the electrification complex above all

• Energy, spanning oil, natural gas, and uranium

• Select producers - well-run miners and energy companies that learned capital discipline the hard way

• Diversified commodity strategies that can roam across energy, metals, and agriculture

Let us also be honest about the ride. Commodities are volatile, cyclical, political, and periodically terrifying. That is precisely why they belong in a portfolio as a permanent, diversified allocation rather than a headline-chasing trade you’ll abandon at the first 15% drawdown.

Most portfolios today are a monoculture of financial assets - stocks, bonds, private equity, credit - paper claims stacked on paper claims. Commodities offer something the rest of the portfolio cannot: direct ownership of the physical economy at the very moment the world is moving from abundance to scarcity, from efficiency to resilience, and from disinflation to structurally higher costs.

The market spent twenty years pricing atoms as if only bits mattered. That era is ending.

We like commodities. A lot. And the time is now.

Subscribe now

AI is starting to deliver real productivity growth, but it is early days - not yet a discernible mega-trend. The only honest way to judge AI’s success on a macro level is productivity: output per hour that becomes GDP, split among labor, capital, and government - three claimants, each in an unusually tight spot of its own. So far, the gain shows up as record corporate margins: the S&P 500’s net margin hit 16.9% in Q2 2026, the highest on record in FactSet data, which begins in 2009. In the same quarter, Labor’s share of non-farm business sector output fell to 52.9%, the lowest since 1947. Whatever productivity AI is generating so far is flowing to shareholders and AI-stack owners, not workers.

That combination - a promising, still-early technology, record margins, and most of the productivity gain flowing straight to capital’s bottom line - is exactly the setup for bubble-type investing: a technology can be real, and the market’s pricing of who captures its value can still be wrong. It is also not sustainable; labor will not stay quiet indefinitely while its share shrinks. The live question is not whether AI is real, but whether the gains spread widely enough that nobody has to lose for someone else to win.

As investors, our job is to believe the technology is transformational, exactly as America’s railroads were in the mid-to-late 1800s, and figure out which horse is still standing when the dust settles. There is one path where Nvidia, the hyperscalers, corporate America, employees, and the government’s balance sheet all win together. History shows how narrow it is: we have threaded it only once.

The one time every piece lined up was the second half of the 1990s. Productivity accelerated from roughly 1.4%-1.5% a year (1973-95) to 2.5%-2.9% (1995-2000), while unemployment fell the whole time, from 6.1% in 1994 to 4.0% in 2000. The dividend went into output and wages, not layoffs, and debt healed too, falling from 48% of GDP in 1993 to 32% by 2001. Acceleration, full employment, and shrinking debt, all at once: the template, achieved exactly once in fifty years.

The more common pattern for a transformative build-out is uglier, and it comes from the railroads. The rail network delivered the era’s productivity miracle, but its financiers fared worse than the economy that used it. Over 100 railroads failed within a year in the Panic of 1893; by 1894, nearly 200 were in receivership. Consumers captured the gains; equity holders did not. Only after that shakeout did Morgan’s “Morganization” consolidate survivors into pricing power, horizontal mergers first, then vertical, culminating in 1901 when Morgan bought Carnegie’s steel empire to form U.S. Steel. Andrew Carnegie, unlike Morgan the financier, was the operator: his mills, upstream of the railroads, earned double-digit returns throughout the shakeout.

Map today’s AI chain onto that pattern and the fit is close. Nvidia looks like Carnegie Steel: roughly 75% gross margins and about 81% share of AI accelerators, selling into every lab. Nvidia’s August 26 earnings underscored that moat: revenue up 106% year-over-year to $96.2 billion - evidence the boom is intact and, arguably, turning bubblicious. The frontier labs are the railroads: laying the track everyone will run on with economics that don’t clear. OpenAI booked $13.1 billion of revenue in 2025 against a $20.9 billion operating loss and has since roughly doubled its run rate past $40 billion. Demand was never the question in 1872 either. The hyperscalers play Morgan and Vanderbilt at once – financing the buildout and running their own lines. Roughly $725 billion of AI capex in 2026, some of it buying Nvidia silicon, some of it buying custom chips to route around Nvidia’s toll booth.

Morganization never preceded the pain it followed it by a quarter-century. We may be living in 1872: 35,000 miles of track had gone down between 1866 and 1873, the technology never in doubt, but almost nobody could yet tell which lines carried sound capital structures and which were rotten underneath. The panic was still more than a year away. Every dollar of fresh AI capex makes that eventual clearing event larger, not smaller.

There is a second 1872 parallel citizens, not just investors, live with. The Civil War left roughly $2.7 billion in debt neither party wanted to honestly fix. The 1873 Coinage Act pushed the other way: ending free coinage of silver put the country on a de facto gold standard, and the deflation that followed helped turn that year’s panic into a 65-month depression. Substitute today’s $40 trillion gross debt headed toward 136% of GDP by 2036: a debt problem of our own making, ignored by both parties.

The employment leg is narrower still. The 1990s had a fast-growing labor force, room to reallocate displaced workers into new roles without touching unemployment. Today’s labor force is barely growing, and boomers are retiring - no cushion this time. Aggregate unemployment has not risen yet, but there is a real entry-level squeeze: employment for 22-to-25-year-olds has declined 16% in AI-exposed occupations relative to other occupations.

The debt leg cuts against a shortcut some hope for: buying enough productivity from job losses to grow out of the deficit. We tried that already. Between 2001 and 2003, productivity surged - 4.1% in 2002 and 3.7% in 2003 - even as unemployment climbed from 4.8% in 2001 to 6.0% in 2003. Debt held by the public rose from 31% to 36% of GDP, as automatic stabilizers outweighed the growth benefit. Productivity purchased with job losses makes the debt problem worse.

Put the legs together and the perfect path is three needles, threaded at once. Productivity must accelerate on something like a 1990s scale, without the labor-force slack that made it painless last time. The vendor layer must consolidate into real pricing power without a panic that touches aggregate employment, because the moment unemployment turns up, the stabilizer math flips debt healing into deterioration. And corporate America must convert AI spending into actual profit well beyond the 5% of pilots managing that today.

The 1990s got by roughly doubling trend productivity, 1.4%-1.5% to 2.5%-2.9%, with only two claimants: labor, near its normal income share, and a government starting at 47.8% debt-to-GDP. This time there are three hands out: labor needs to reverse a record-low 52.9% share, capital needs its record margins to hold or valuations come apart, and Washington needs growth to arrest a ratio more than double 1993’s level. Satisfying all three plausibly takes something closer to tripling trend productivity, not doubling it; even Goldman Sachs’ AI estimate produces a 1.5% annual uplift to productivity growth over a decade. A sustained tripling is rarer still: it has almost never, if ever, been achieved in a reasonable timeframe.

That is a narrow corridor; narrower than either the “productivity will show up” bulls or the “1873 all over again” bears assume. Measured the only honest way, on productivity big enough to fund labor, capital, and government all at once, it means threading a very tiny needle, or, if football is the more honest metaphor, throwing a 60-yard Hail Mary and having it land in the right hands. It has happened before, exactly once. Betting on Nvidia’s moat, the hyperscalers absorbing the buildout, and corporate America converting AI capex into profit before a credit event forces the issue is a bet on history’s rarest outcome, not its most common one. Expect an extended bubble-type run until one of the three needy hands comes up short - labor turns political, capital pulls back spending, or Washington hits a debt pothole that outruns growth. Is AI the savior that finally lets us grow out of a debt problem both parties ignore, or the bringer of the next investment bust? On the evidence of 1872 through 1898: perhaps both, in sequence, and we do not yet know how long the sequence runs.

Subscribe now

Last week we wrote about the dominoes falling: Hamiltonian statecraft returning to Washington, Beijing quietly building a hard-asset monetary architecture, and a Treasury market that had started sending an uncomfortable message about long-term rates. Our conclusion was simple: keep the core gold position, add on weakness, and let the market - not emotion - tell us when to get more aggressive.

Days later, the market is talking louder than we expected, and today’s volatile trading is proof.

On Wednesday, Treasury Secretary Scott Bessent doubled the department’s liquidity-support buybacks of 10-to-30-year government debt, from $2 billion to at least $4 billion per operation. Long yields, which had just hit multi-decade highs, fell hard. The 30-year Treasury yield dropped nearly 10 basis points to roughly 5.19%, down from a 19-year high above 5.33% set a day earlier. The 10-year slid to about 4.64% from a 20-month high near 4.75%. Stocks jumped, the dollar sank toward a three-month low, and gold surged above $4,500 an ounce, up more than 12% for the month.

One day of relief is not a policy regime.

By Thursday, most of that surprise drop in yields had already been unwound. The market is testing the Fed yet again, and it is a powerful test, because the underlying setup argues for higher term compensation, not lower.

Start with the term premium - the extra yield investors demand for holding long-dated debt rather than rolling short paper. The New York Fed’s own model puts the 10-year term premium at roughly 0.65% to 0.75% today, comfortably below the 65-year median near 1.4% and well under the Great Moderation-era average near 1.65%. Investors are being paid historically little to hold long bonds at a moment overflowing with fiscal risk. Then add nominal GDP, a strong long-run magnet for long yields: current-dollar output grew at a 7.9% annualized pace last quarter and is running well north of 5% year over year. Layer in unemployment at 4.1%, near the tightest levels of this cycle. And Brent crude is a whisper from $95 a barrel, with obvious implications for headline inflation if it breaks through. Put those four forces together, and the long end has every incentive to keep testing Washington’s tolerance.

That is the pattern we flagged last week. Treasury’s move is not quantitative easing. Bessent reshuffled the government’s own debt profile; he did not create new money the way Federal Reserve QE does. Useful, but limited against forces this strong. The federal government still carries more than $40 trillion in gross debt, still faces upward of $1 trillion in net interest expense this year, and still needs buyers for a growing pile of new issuance.

Which brings us to the domino we said would matter most: the Federal Reserve.

Next week we get our first real look at how the new Fed intends to handle this. From August 27 through 29, the Kansas City Fed hosts its annual policy symposium at Jackson Lake Lodge in Wyoming; the same forum where past chairs have moved markets with a single paragraph. This year the speaker is new. Kevin Warsh, sworn in as the Fed’s 17th chair in May after the closest confirmation vote in the position’s history, delivers his first Jackson Hole address as chair. The official theme is financial innovation in payments. The unofficial one is whether a Fed under new leadership - facing a president who wants lower rates, an underpriced term premium, and an economy running hot on nominal growth - is preparing to lean toward easier policy anyway.

We are not predicting what Warsh will say. We are telling clients to watch closely, because the answer will move gold more than almost anything else left on this year’s calendar.

Our framework rests on two confirmations we have not yet received: the Fed shifting from restraint toward accommodation, and gold’s own volatility settling down enough to confirm this move has staying power rather than a spike. Neither is in place yet. But the forces pushing on the long end are real - which is exactly why we are telling clients to watch markets more closely right now.

Our conclusion has not changed; instead, it has sharpened. Keep the strategic gold position. Add on weakness with more confidence, because every domino that falls - the deficit, the yield spike, the Treasury intervention - makes the debasement case harder to dismiss. And watch Jackson Hole closely. Powell’s successor is about to tell us, in his own words, how comfortable he is running a Fed that has just watched the bond market need rescuing twice in a single week, against a backdrop screaming for higher, not lower, long-term rates.

Subscribe now

Some rivalries make sense on paper. The Fed versus inflation. Gold versus real rates. And then there are the ones nobody saw coming - like America’s first Treasury Secretary finding his most enthusiastic modern disciple not in Washington, but in Beijing.

Alexander Hamilton built the young United States on one non-negotiable idea: a currency is only as good as the hard asset behind it. The Coinage Act of 1792, his handiwork, pegged the dollar to a fixed weight of gold and silver. When America traded abroad, there was no reserve currency to hide behind: imbalances were settled the only way they could be, in metal.

Stability at home, settlement abroad, both resting on the same foundation. That’s the part of “Hamiltonian economics” nobody in Washington talks about this year.

And Washington is talking about Hamilton… a lot. Treasury Secretary Scott Bessent wrote a whole Wall Street Journal op-ed crediting Hamilton with inspiring the administration’s “economic statecraft.” Vice President Vance told a podcast audience that “American economic policy on the Right is now much more Alexander Hamilton than it is Milton Friedman.” Trade Representative Jamieson Greer used his Davos keynote to resurrect what he called “the Hamiltonian economic system that too many have forgotten.” Even Fed Chair Kevin Warsh, without ever saying Hamilton’s name, spent his pre-confirmation speeches arguing the Fed drifted from the boundaries the founders intended and needs to answer to the Treasury again - an idea that traces straight back to the bank Hamilton chartered in 1791.

The funny thing is, while Washington was rediscovering Hamilton’s rhetoric, Beijing was quietly building his plumbing.

China has spent the past two years constructing, brick by brick, the infrastructure for a world that settles trade in gold instead of dollars: a Shanghai Gold Exchange vault in Hong Kong, yuan-denominated gold contracts, a prospective vault in Saudi Arabia built for an oil-for-gold pipeline that never touches a dollar. That’s the overt part. The covert part is bigger: Beijing officially reported buying a trickle of gold in 2025, while analysts at Goldman Sachs and Société Générale, working the trade data, put the real number at ten times that - and independent researchers now peg China’s true reserves at 5,000 to 8,000 tons, more than triple what it discloses to the IMF. Russia, locked out of dollar clearing since 2022, is feeding that appetite and helping pilot a gold-backed settlement currency for the BRICS bloc.

Two governments, two entirely different reasons, arriving at the same conclusion: gold is reclaiming its old job as the asset that settles accounts when nobody fully trusts anyone else’s paper. Washington got there through nostalgia. Beijing got there through strategy. I’ll leave it to the historians to decide who understood Hamilton better.

What I care about is what it means for your portfolio, and that’s where I’m headed next. For the first time in about six months, gold is finally showing some pulse.

That doesn’t mean the correction is over, and I don’t want to get ahead of myself here, and neither should you. But after months of lower highs, souring sentiment, and the kind of volatility I flagged back in April, the tape has finally handed investors something worth reacting to.

Start with the Fed. At its July meeting, the committee held rates at 3.50%–3.75%, brushing off a few hawks itching to hike again. Then the jobs numbers came in soft, taking even more air out of the “one more hike” crowd. Gold noticed, climbing back to its best levels in almost two months.

Then there’s the currency story, which doesn’t get enough attention. The Treasury joined Japan in buying yen to defend it near 40-year lows, with the New York Fed doing the actual buying on Treasury’s behalf - the first joint yen intervention in over a decade. I don’t care much about the yen specifically. What I care about is the signal: even the world’s most powerful finance ministries admit they can’t just sit on their hands when fiat currencies get unruly in a system this leveraged. That’s not a one-off. That’s the whole debasement thesis playing out in real time, just wearing a different jersey.

Meanwhile, the pillar I keep coming back to - central bank buying - hasn’t budged an inch. I covered China’s numbers above, but China isn’t the only one still at the table. Poland added another 82 tons in the first half of 2026, pushing its reserves to roughly 632 tons, and it’s not slowing down. Korea just joined the party too: the Bank of Korea rolled out a new framework in August that lets it buy gold directly from domestic miners - another central bank quietly diversifying away from paper.

Stack that demand backdrop on top of sentiment that’s about as washed-out as I’ve seen it, and the ingredients for something more constructive are assembling.

Just don’t confuse “constructive” with “safe.” The GVZ, gold’s version of the VIX, closed last week around 25–26 and has since ticked back toward the high 20s, still well above the sub-20 reading that’s historically told us the market has settled down. My read: we’ve likely gone from downtrend to chop, not from downtrend straight into the next bull leg.

None of this changes the long-term thesis I’ve been pounding on for years: fiscal deterioration, exploding sovereign debt, a fracturing geopolitical order, and central banks buying gold like it’s going out of style (it isn’t). We remain comfortable at the high end of our historical range in precious metals, and I’d rather own too much gold in this environment than too little. What has changed is the tactic: with sentiment this beaten down and the fundamentals intact, I’m comfortable adding into weakness rather than waiting for a perfect entry that chop-driven markets rarely deliver.

Which brings me back to Hamilton. He didn’t build the credit of a brand-new country by timing the tape. He built it slowly, coin by coin, on a foundation he trusted more than the mood of the moment. Gold investors would do well to borrow his patience. Keep the core position. Add when the market hands you weakness. And let the discipline, not the daily chop, decide when this turns into the next leg higher.

Subscribe now

The bullish case is real

Start with the obvious: investors are not hallucinating the good news.

Large U.S. companies are producing extraordinary profits. Earnings for large-cap stocks are projected to grow by about 20% year over year this quarter, extending the long run of profit expansion. Profit margins and returns on equity for many of the market’s dominant businesses - especially in technology, semiconductors, and software - remain near historical highs according to Bloomberg. Small-business survey data from the NFIB also point to healthy sales and earnings trends, which supports the idea that productivity gains are broadening beyond a handful of mega-cap companies.

Then there is AI. If artificial intelligence pushes productivity into a meaningfully higher gear, it could support profits for much longer than many skeptics think. That would help justify premium valuations for the companies building the chips, software, cloud systems, and platforms that make the technology work. It could also help the country more broadly: faster productivity growth makes wage gains, tax receipts, and fiscal management easier over time.

So yes, investors have reasons to be optimistic. This may be one of the strongest collections of companies ever assembled in the S&P 500.

But so is the crowding

Still, the positive story is not the whole story.

Federal Reserve survey data show that U.S. households now hold a record share of their financial wealth in equities. In other words, people are not cautiously leaning into stocks. They are already there. And they are paying historically rich prices to stay there.

Leverage reinforces that message: FINRA data reported by Advisor Perspectives show U.S. margin debt climbed to roughly $1.42 trillion in May, a record high and more than 50% above year-earlier levels. On an inflation-adjusted basis, margin debt is also at an all-time high. That does not tell anyone when the market will peak, but it does tell you that speculation is running hot.

The same goes for the growing use of double- and triple-levered ETFs, many of them tied to the market’s most volatile stocks and sectors. Bloomberg-cited reporting shows assets in leveraged and inverse U.S. equity ETFs approaching $200 billion, with massive rebalancing flows on volatile days. Zero-days-to-expiration options tell a similar story: since 2022, same-day options volume has exploded, turning more of the market into a vehicle for intraday speculation rather than long-term investing.

Put it together and the picture becomes harder to ignore. Households are heavily committed, leverage is high, speculation is visible, and prices are rich. Those are not reasons to panic, but they are reasons to stop pretending this is a normal market.

The hedge may not hedge

There is another issue that deserves more attention than it gets: the traditional hedge is under pressure.

For much of the last 30 or 40 years, investors could reasonably assume that long-term government bonds would help offset equity losses. That assumption has become much less reliable. The past five years have shown repeated episodes in which stocks and long-duration bonds fell together, most notably in 2022, when the classic 60/40 portfolio suffered one of its worst drawdowns since the 1930s. That should not be dismissed as a one-off.

Part of the problem is valuation. Term-premium models from the New York Fed and San Francisco Fed show that the extra compensation investors receive for holding long-term Treasurys remains low relative to long-run history. Trading Economics, using Fed data, puts the 10-year term premium around 0.73% in July 2026, well below the levels seen in the early 1990s. In plain English, investors are taking a lot of duration risk without being paid especially well for it.

Part of the problem is supply. U.S. Treasury and Bloomberg data indicate that the federal government must finance nearly $12 trillion of marketable debt over the next year, more than double its needs just before COVID. At the same time, the U.S. is already absorbing about 38% of global savings, while the trailing 10-year growth rate of global savings is among the lowest on record. If you own long-term government bonds because you think they are portfolio insurance, you need to consider the possibility that they may, at key moments, add to portfolio stress instead of reducing it.

Policy still matters

This is where policy becomes central.

The U.S. sovereign balance sheet is under real strain. Data from the U.S. Treasury, the Social Security Administration, and Bloomberg show widening deficits, rising net interest costs, and borrowing needs that continue to climb. Treasury projections also point to ongoing financing support from the Federal Reserve in the hundreds of billions of dollars per year. That is another way of saying the system is leaning more heavily on monetary support to carry a growing fiscal burden.

Gold matters in that kind of world. Over long stretches of time, gold has often been a useful diversifier against both financial stress and the erosion of fiat purchasing power. But even gold is not a simple answer here. It moved sharply earlier this year, and it remains sensitive to the path of real interest rates and to the possibility that the Fed might still tighten in the near term.

And then there are the people. Democracies can change rules faster than markets expect. Tax policy, spending choices, regulation, and redistribution can all shift materially over the next two and a half years if voters decide the current arrangement has gone far enough. Investors near all-time highs should keep one eye not only on profits and policy, but on the electorate that can reshape both.

The profits are real. The AI story may be real. But so are the historically stretched valuations, the crowding, the speculative leverage, the questionable bond hedge, and the fiscal strain in the background. Investors do not need a dramatic forecast. They need perspective - and a clear understanding that near all-time highs, risk often hides in the relationships people assume will keep working.

Subscribe now

America has always been good at celebrating spectacle while ignoring the strain building underneath it. In the same news cycle that hands a trillionaire fantasy the spotlight, tens of millions of retirees are being told to brace for a future in which a Social Security check of roughly $1,600 a month is no longer something they can count on the way they once did.

Put that contrast into numbers and it stops being a question of how anyone feels. At roughly $1,600 a month, it would take a Social Security recipient roughly 52 million years to accumulate a single trillion dollars. That is the scale of the wealth fantasy now pulling in the headlines, the applause, and most of the political oxygen.

That gap is what makes the present moment so fragile. For a narrow slice of the country, the economy looks spectacular; for a much larger slice, it feels impossible, and the political system is starting to register the difference. Whoever takes the oath of office in January 2029 will inherit something close to $39 trillion of debt held by the public, with interest expense on track to be the single largest line in the federal budget. The job will not be optional: fix Social Security and Medicare in the first term, before the trust funds run dry and the arithmetic turns even less forgiving.

The same tension runs straight through the markets. Credit-card stress is climbing, auto-loan delinquencies are elevated, and student loans are turning into one more pressure point for households that were already living too close to the edge. At the top, wealth compounds on itself; at the bottom, the cushion is wearing through. Nobody should confuse that with a healthy, broad-based economy; it is a split-screen economy, and the two screens drift a little further apart every quarter.

Darius Dale of 42 Macro has been one of the clearest voices putting hard figures on the gap. The top 10% of households own roughly 87% of the stock market, while the bottom 50% own barely 1%. Median household income now runs at about 10% of the income of the top 5%, down from 26% in 1981. And nearly 60% of adults still cannot cover a $1,000 emergency out of savings. Taken together, those figures point to something other than broad-based prosperity: a system that has quietly grown brittle.

The policy response carries its own risk, and it is every bit as concrete. A wealth tax levied on illiquid securities would function as a forced-selling regime for private-company stock, restricted shares, partnership interests, and concentrated single-name holdings. Taxing paper wealth as though it were cash would not reform much of anything; it would simply pull the next crash forward, because when the tax comes due on an asset that cannot be sold in size without moving the market, something must give. Because one side of the political spectrum is actively promoting versions of this idea, in a growing number of states and increasingly out in the open among leading DSA-aligned politicians, investors should price it as a real probability rather than file it away as a fringe slogan.

I do not reach for the third rail of politics often, and I reach for it warily even now. But as an investor, I cannot pretend the large object in the road is not there. The choice in front of us is no longer abstract: it is detour or crash.

We have lived through a version of this before. The original Gilded Age ran on anarchist violence, repeated market crashes, and a political backlash that eventually reined in concentrated business power through trust-busting and broader reform. When institutions stop producing legitimacy, the system tends to resolve the tension one of two ways, through reform or through rupture. The late 1800s and early 1900s did not drift gently into a healthier balance; they were dragged there, and the dragging was not always peaceful.

That is why the next few years matter more than they appear to. They are not somewhere off in the distance; they sit right in the windshield. A great many of the purchases, financings, and lockups being struck today have to work inside that window. If both parties in Washington keep dodging the truth about debt, taxes, entitlement reform, and the widening split between the asset-rich and everyone else, the adjustment, when it finally comes, will be less orderly and a good deal more painful.

The headlines will keep chasing the trillion-dollar fantasy; spectacle always sells. The politics, and before long the markets, will be set by the $1,600 check and by the tens of millions of people who are quietly concluding that the math no longer works for them.

Subscribe now

Summer reading in the Berkshires, the rolling hills of western Massachusetts, can make economic history feel unusually tangible. Drive through Lenox or Stockbridge and you pass grand houses set behind stone walls and old trees, reminders of a time when some of the richest families in America built elaborate summer “cottages” far from the financial centers that made them rich. Many of these estates were used only briefly each season, and their owners often maintained several other homes in New York, Newport, Bar Harbor or elsewhere. The Berkshires were not where the money was made. They were where financial and industrial wealth went to display itself.

That is part of what makes the place so suggestive. The Berkshire estates are the physical afterlife of a capital cycle. Some of the names are still recognizable, but many are not. Wheatleigh was built by Henry H. Cook, a financier, banker, railroad director and real estate operator whose fortune was large enough to create one of the great houses of Lenox, even if his name now barely registers outside local history. Eastover was the estate of Harris Fahnestock Jr., heir to a banking family deeply connected to railroad finance in the Jay Cooke era. The houses remain. The names fade. The infrastructure boom that produced them changed the country anyway.

That thought sat in the background while reading Liaquat Ahamed’s 1873 and Richard White’s Railroaded, two books that together capture both the grandeur and the corruption of the railroad age. What they suggest is not simply that AI resembles railroads because both were exciting technologies. It is that railroads were one of the largest and most transformative investment booms in American history, a true general-purpose system that reordered settlement, finance, politics and industrial production all at once.

Contemporaries understood the scale in almost biblical terms. The Library of Congress notes that between 1871 and 1900 the United States added roughly 170,000 miles of new railroad track on top of the 45,000 miles already built, helping bind the country into a single national market. Ralph Waldo Emerson called railroad iron a “magician’s rod, in its power to evoke the sleeping energies of land and water,” while one Rocky Mountain newspaper in 1866 described the Pacific Railroad as “the one remedy for every evil, social, political, financial, and industrial.” Those claims sound extravagant now, but they capture how central rail had become. In economic history, railroads are not a side note. They are one of the key organizing facts of 19th-century American growth.

That is one reason the analogy to AI is so interesting. The rhetoric around AI has the same totalizing quality. Sam Altman has described AI in world-changing terms, and leaders at Davos have called it one of the defining technologies of the age, one that will reshape society, economies and politics. The comparison is not absurd. The mistake is assuming that because a technology is truly transformative, its labor effects, capital cycle and market returns will all be simple.

Take labor. One might imagine that as railroads spread across the country, many people assumed the horse was doomed. That is roughly how some current discussions of AI sound now: the machine is here, therefore the worker must be on the way out. Yet the historical pattern was almost the opposite. Yale’s energy history project shows that the U.S. draft-animal population, mostly horses and mules, rose from roughly four million in 1840 to about twenty-four million by 1900, even as railroads remade long-distance transportation.

The reason is straightforward. Rail displaced horses in one narrow function, long-haul transport, but it increased demand for them almost everywhere else. Rail depots needed drayage. Growing railroad towns needed more cartage. Farms connected to wider markets needed more local hauling. Rail automated one visible task while enlarging the surrounding economic system in ways that created more complementary work for horses and for the humans who depended on them.

That may prove to be the better framework for thinking about AI. A new technology can destroy specific tasks without reducing total labor demand in the near term. It can shift the bottleneck, enlarge the market or create new layers of complementary work that were not previously visible. The true collapse of the horse economy came later, with internal combustion. Once cars, trucks and tractors became broadly substitutive across transport and agriculture, the horse population fell sharply, down roughly 70% from peak by 1950 and about 85% by 1960. Rail was not the horse killer. The broader platform shift came later.

The second lesson is about duration. Today’s AI capex boom is already large enough to matter for macro analysis. Recent work places AI-related infrastructure spending at roughly 1.2% of U.S. GDP in 2025, an extraordinary figure for what still feels like an early-stage buildout. But the railroad precedent still dwarfs it in one crucial respect: persistence. Historical reconstructions suggest rail capital expenditures remained elevated for decades, with estimates around 2.4% of GDP in the 1870s and nearer 6% at the 1880s peak when upstream sectors are included.

Rail was not a brief investment burst. It was a regime. Even major busts in the middle of the process did not really stop it. The Panic of 1873 was severe, and overbuilding was already evident, yet railroad expansion resumed and in some respects intensified afterward. A future AI bust, whenever it arrives, may tell us less about the long-run importance of the infrastructure than about the soundness of the first financing structure.

The third lesson is that the shenanigans likely start early. They certainly did in rail. The subsidy-and-self-dealing dynamic appeared in the 1860s, not after the boom had fully matured. The Pacific Railroad Acts created incentives for overbuilding and abuse, while the Crédit Mobilier scandal showed how quickly politically supported infrastructure could become a vehicle for insider enrichment. By the time rail capex was running above 2% of GDP in the 1870s, the system was already fragile enough to help trigger a major panic.

That sequence feels relevant now. If AI is becoming an infrastructure complex rather than just a software category, then the distortions have probably already begun. The likely modern forms are different, but the resemblance is plain: subsidy competition for power and land, heroic utilization assumptions, accounting opacity around massive fixed investments and a market tendency to confuse technological inevitability with security-level inevitability.

Seen from a Berkshire porch, with 1873 and Railroaded on the table and a lineup of Gilded Age houses a short drive away, the conclusion is not that AI is a bubble and therefore a mirage. It is that the most important technologies often arrive through precisely this mix of genuine utility, extravagant promotion, political entanglement and periodic financial embarrassment. Railroads changed America even though many railroad securities proved disastrous. AI may do the same. The busts and the excesses may already be here, but history suggests they are more likely to mark the early chapters of the story than the end.

Subscribe now

The University of Michigan’s consumer sentiment index just hit an all‑time low, even as the stock market trades not far from record highs. That’s not a rounding error: a national survey that has been tracking how Americans feel about their finances and the economy since the 1950s just printed its worst reading on record in what is still widely described as a “strong” economy. If markets are supposed to discount the future and surveys are supposed to capture the present, that gap should make all of us uncomfortable.

It helps to know what the Michigan survey actually is. For roughly 70 years, the University of Michigan’s Survey Research Center has called households across the country and asked a simple set of questions: How do you feel about your current financial situation? Do you expect business conditions to be better or worse over the next year? Is this a good time to buy major household items? The answers are combined into an index that economists use as a real‑time gauge of the national mood. It is not a Wall Street poll or a social‑media sentiment score; it is Main Street, talking about Main Street, in a way that has been consistent across generations.

Against that history, a new all‑time low is telling us something profound. Households are effectively saying that today feels worse than the oil shocks of the 1970s, the early‑1980s double‑dip recession, the dot‑com bust, the global financial crisis, and the COVID shock. At the same time, equity investors, staring at earnings and discount rates, see enough strength in corporate profits to keep major indices near their highs. Squaring that circle requires looking beyond the unemployment rate and into the plumbing of the labor market and the national income accounts.

On the surface, the labor market still looks solid. Weekly unemployment insurance claims from the Department of Labor remain low by historical standards, and the insured unemployment rate is only modestly above pre‑pandemic levels. If all you watched were those two series, you might reasonably conclude that we have a tight labor market and that talk of “hidden weakness” is overblown. The reality is more complicated. Other official data from the Bureau of Labor Statistics show that the private‑sector hiring rate has fallen back to levels last seen in 2009, while the layoff rate remains below pre‑COVID norms and the quits rate has retreated to 2015–2016 territory. Companies are not doing large‑scale firings, but they are not doing much hiring either, and workers do not feel confident enough to jump to better jobs.

The household survey adds another layer. Since 2022, both the employment‑population ratio and the labor‑force participation rate have rolled over. Since November 2025, the headline unemployment rate has edged lower even as total employment fell by roughly 808,000, simply because about 1.17 million people left the labor force and therefore no longer count as unemployed. That is the world of Uber drivers, delivery gig workers, and part‑timers: you stay technically “employed,” you do not file for unemployment insurance, but your income can fall sharply. The official data are doing their job; they are just not capturing the erosion in the quality and trajectory of work that people feel in their daily lives.

To understand why the Michigan survey is so depressed, you have to follow the money. The Bureau of Economic Analysis publishes a series called real personal income excluding transfer payments, which strips out government benefits and adjusts for inflation to show the purchasing power generated by the private economy. From the end of the financial crisis through 2020, that series grew at about a 2.9% annual rate, a remarkably steady trend. Households built a mental model of “normal” around that line: each year, on average, their real incomes crept higher, giving them room to plan, save, and spend.

When inflation picked up in 2022, that line broke. Today, real personal income ex‑transfers sits more than 1.1 trillion dollars below where it would be if that pre‑pandemic trend had continued. In per‑capita terms, the gap is about 3,187 dollars relative to the 2009–2020 trend. That missing several thousand dollars of purchasing power shows up in rent checks, grocery bills, car payments, and medical deductibles. You may not lose your job. You may not show up in jobless claims. But relative to what you thought your life would look like - the implicit promise of that old trend line - you are behind, and every month you fall a little farther behind. The Michigan survey is simply that math, rendered in human emotion.

So why isn’t Wall Street panicking? Because the stock market lives in the world of profits and interest rates, not survey data. On that scorecard, corporate America has rarely looked better. BEA’s national income accounts show after‑tax corporate profit margins rising again in the latest Q4 data to 16.2%, near the highest level recorded since the 1950s and well above the roughly 13% level on the eve of the pandemic. If you plot those margins back to 1954, the post‑COVID period stands out as a regime of margin supremacy.

A decomposition of those margins reveals where the gains came from. As a share of gross value added, the compensation of employees has fallen by about 2.6 percentage points since Q4 2018, while net interest expense has also declined as a share of revenues. In plain language, businesses are paying workers a smaller slice of the pie and paying less to creditors, keeping more for shareholders. Many large firms locked in long‑term, low‑rate financing in 2021–2022, something that was never available to Main Street households rolling mortgages or financing used cars. Layer onto this the promise of artificial intelligence, which offers more output with fewer workers, and you can see why boards feel little urgency to change course.

Could this imbalance resolve through a classic downturn? It could. Leading indicators like housing permits and manufacturing new orders, tracked by the Federal Reserve and the Census Bureau, have already weakened in response to tighter monetary policy. If that weakness spills over into broader production, unemployment will eventually rise in ways that even the headline numbers cannot mask. But with after‑tax margins north of 16% and AI‑driven cost cuts still in their early innings, it is not obvious why corporate America would voluntarily compress profitability to ease the pressure on household balance sheets. From the vantage point of the C‑suite, the current strategy is working exactly as designed.

From the vantage point of a household taking the Michigan survey, it is a different story. Real incomes are a trillion dollars below trend, the job ladder is missing rungs, and the lion’s share of the expansion’s gains has gone to capital rather than labor. That combination produces a record‑low sentiment reading at the same time it supports near‑record equity prices. It is a divergence that cannot persist forever. If corporate America continues to “crush labor” in the national accounts - paying less out in compensation as a share of value added while keeping margins elevated - the mobs with the pitchforks will only grow louder, and eventually they will head for the ballot box.

None of this analysis would be possible without the extremely careful work of parsing official data from the Federal Reserve, the Bureau of Economic Analysis, the Bureau of Labor Statistics, the Census Bureau, and the Department of Labor into a coherent macro picture. I am indebted to EPB Research for assembling and visualizing these series so clearly; any errors in interpretation here are mine, not theirs. The Michigan survey may be telling us how people feel, but the underlying numbers from the major statistical agencies tell us why - and for now, the story they are telling is one of record profits, stalled real incomes, and a political fuse that grows shorter by the month.

Subscribe now

Fog of war is the old phrase for the uncertainty that descends when events outrun understanding. In war, commanders must act before they have a full picture of the battlefield. In markets, the same thing happens when prices, policy, and geopolitics collide all at once. So far in 2026, that is exactly where gold investors have been living.

I remain convinced about the long‑run case for gold, but the day‑to‑day ride has become so violent that staying with the position feels like an emotional stress test. Gold’s first quarter was a reminder that even strong secular bull markets do not move in straight lines. The metal pushed higher in January, then suffered two separate corrections of more than 20% from peak to trough in just a few weeks - leaving many investors wondering whether they owned a store of value or a theme‑park ride.

When I say “gold vol,” I am talking about how wild the short‑term price swings are, as seen through the options market. One common gauge is an index called GVZ, sometimes nicknamed the “Gold VIX,” which estimates how much gold is likely to move over the next month, annualized. Low readings suggest a calm tape; high readings tell you to expect big swings, both up and down.

You don’t need to memorize the exact formula to understand the message. When GVZ lives in the teens, gold tends to trend more smoothly. When it jumps into the high 20s or 30s, the market is bracing for turbulence. Recently that “vol thermometer” has been running hot, which matches what any gold holder has felt in their stomach, even if they’ve never heard of GVZ.

High gold vol does two things at once.

First, it changes price behavior. In the first quarter, gold twice dropped more than 20% in a matter of days, then bounced back almost as quickly. On paper, the full quarter still ended positive - roughly high single‑digit returns - but almost nobody experiences it that way. What people remember are the air pockets, not the net result.

Second, and more important, high gold vol changes investor behavior. It tempts people to overreact: to sell after a big drop, to hedge after the risk has already materialized, or to stay on the sidelines waiting for a “better entry” that never feels comfortable. In calmer tapes, investors can let the fundamentals work. In a tape like this, they’re being asked to make allocation decisions under conditions of partial visibility. That is what makes this such a classic fog‑of‑war moment for gold.

We’ve seen versions of this movie before. In earlier work I’ve argued that gold is the “currency of war,” a strategic asset that governments and investors turn to when paper promises look less reliable. I’ve also written that in a world of ballooning debts and financial repression, “you need gold” long before the inflation shows up in the official statistics. None of that logic has changed. What has changed is that, in the short run, gold has been pulled into violent cross‑currents of leverage, liquidity, and geopolitics.

The war, the Fed, and the next chapter

The current war has tightened energy markets and added to an already unstable fiscal backdrop. That combination points toward higher inflation over time, even if the monthly data wobbles. The crucial question for gold is how the Federal Reserve responds if we have persistent inflation.

There are two broad paths:

1. The “tough” path: The Fed decides to fight higher inflation with higher interest rates, or at least by keeping rates elevated for much longer. In that world, gold can suffer, at least for a while. Higher real yields and a stronger dollar are usually headwinds. If we get a genuine tightening campaign into rising inflation, I would expect gold to struggle or trade sideways, even if the long‑term story remains intact.

2. The “drift” path: The Fed chooses nothing or lower rates, allowing inflation to run hotter than its official target. That may sound unlikely at first but look at the labor market. We are drifting through what feels like a jobless recovery - an employment picture that is squishy beneath the surface, with weak hiring, soft real wage gains, and a growing gap between headline job numbers and how people actually feel about work. In that environment, I think it’s more realistic to expect hesitation, gradualism, and a bias toward easing rather than a brave Volcker‑style stand.

My base case is that the Fed ultimately chooses the second path. Political and social pressures around employment, combined with a heavy debt load, make a full‑on war against inflation difficult to sustain. That doesn’t mean rates crash tomorrow. It does mean that, over time, the central bank is more likely to err on the side of protecting growth and jobs, even if that lets inflation simmer.

For gold, that is the more supportive scenario. It suggests a world in which the current war‑driven chaos and high volatility are part of a larger transition toward weaker money, higher inflation, and greater demand for hard assets.

How I’m navigating this

So where does this leave someone trying to live with gold in a war economy?

• I would not disturb core strategic gold positions just because the recent path has been uncomfortable.

• If you are underweight relative to your long‑term target, periods of fear and high vol are, historically, moments to “pick away,” not to wait for a perfectly calm entry that may never arrive.

• I would be cautious about leverage and high‑beta gold plays until volatility calm - this is a time to think in terms of ballast, not bravado.

Gold today is noisy, volatile, and tied up with war and politics. It is also under‑owned in many portfolios, still being accumulated by central banks, and supported by the same long‑term forces - fiscal strain, monetary experimentation, and geopolitical tension - that have been building for years. For me, that combination is enough to stay constructive: hold your core, add carefully on weakness, and remember that the ultimate tell is not the latest headline, but when gold volatility finally breaks lower and lets the secular bull market show itself more clearly.

Subscribe now

In the late 1990s, investment grade telecom companies issued bonds by the hundreds of billions to wire the world with fiber optic cable. The infrastructure was real, the technology was transformative, and the debt was priced as though the revenues to service it were equally certain. When demand fell short of projections, the gap between what was owed and what was earned proved insurmountable, and investment grade carriers of the boom became some of the largest bankruptcies of the early 2000s. The fiber proved essential to the modern economy and the digital age, but bondholders took the pain.

A quarter century later, the borrowers are a different breed: cash‑generative businesses with genuine earnings, not the leveraged carriers of the 1990s. But the basic setup rhymes. Investment grade credit conditions are far more challenging than they look in a fact sheet or glossy pitch book. Issuance has surged to near‑record levels, with projections for 2026 pointing to new all‑time highs in total supply. As Matt Tracy notes, “the five major AI hyperscalers issued $121 billion in U.S. corporate bonds last year, versus an average $28 billion per year between 2020 and 2024.”1 Alphabet felt confident enough to sell a 100‑year bond, and Amazon set the record for the largest single corporate bond deal in history. When even your cloud provider is terming out to 2124, maybe it is not just “prudent balance sheet management.”

And that is just the visible tip of the iceberg. Private credit has become a multi‑trillion‑dollar asset class operating alongside public investment grade debt, with valuations set periodically by managers rather than continuously by markets. At this scale, it is still a relatively new creature, and the stress scenarios are largely untested. What history does tell us is that when credit stress surfaces in less liquid markets, it rarely stays contained; it seeps into public investment grade through downgrades, forced selling, and repricing that hits an already saturated market. By the time you see it show up in your “core bond” portfolio, the adjustment is not gentle.

These issuance trends are landing on top of already troubling internals. Roughly half of the investment grade bonds outstanding is BBB‑rated, meaning there have never been more bonds sitting one notch above junk, waiting for a cyclical downgrade2. Should conditions worsen, many of those bonds could get pushed into high yield. Investment grade funds that are mandated to own only investment grade paper then become forced sellers, regardless of price. We saw a version of this movie during the 2020 COVID downturn, when ratings pressure met illiquidity and forced selling cascaded through the system.

All of this issuance needs buyers. The buyers who were once eager are starting to sound uneasy. Vanguard, one of the largest players in the IG primary market, has said publicly that investors are “full” and warned that hiding in quality “may not play out well this time.”3 BlackRock has been similarly blunt, writing that investment grade “does not look especially cheap.”4 When two of the largest fixed income managers on the planet are effectively telling you they are stuffed and not being paid enough, that is not a backdrop where you want to reach for extra credit risk. When the bouncers at the club say, “We’re over capacity,” you do not push harder to get inside.

This is where our investment philosophy matters. Valuation is rarely a reliable short-term determinant of return, but that does not mean you can ignore it, because when conditions get murky—war, recession, inflation, policy mistakes—the bottom can be a lot deeper than you expected. You cannot time when cheap gets rewarded or expensive gets punished, but you can choose whether to be the last one holding richly‑priced credit when the lights go out.

We also need to be precise about definitions. We have long viewed non‑investment‑grade debt—“high yield,” “leveraged loans,” and much of private credit—as equity by another name and classify it as such in our framework. We think it behaves like equity, it sells off like equity, and its returns are driven by the same growth, margin, and liquidity conditions that drive the stock market. If you think junk and non‑investment‑grade debt belong in your “bond pile,” you have to answer a simple question: why do those instruments correlate so tightly with equity indexes, especially in stress periods? When the equity market falls sharply, high yield and leveraged loans do not cushion the blow; they amplify it. Calling them “income” does not magically make them ballast.

What is different, and more concerning, today is that even investment grade credit is not worth the risk in our view. At current valuations, you are being asked to take credit, downgrade, and liquidity risk for a modest pickup over Treasury bills, in an environment where fundamentals are slipping and relative performance is already rolling over. To own IG now, you need a long list of things to go right simultaneously: rates need to fall, AI needs to monetize on the optimistic schedule now baked into capital plans and balance sheets, record supply has to be absorbed smoothly by investors who already tell you they are “full,” and private‑credit stress must remain neatly contained in an opaque, untested market. That is a low‑odds package deal.

Meanwhile, the macro environment is anything but benign. The structural case against long duration has not weakened; it has strengthened. Inflation dynamics remain unsettled; the probability of recession is rising; and we are living through multiple wars whose first‑ and second‑order economic effects are still unfolding. A deteriorating macro backdrop hurts corporate earnings, widens credit spreads, and makes refinancing difficult—on top of record forecast issuance—considerably more treacherous. In this world, we believe a portfolio of tight‑spread corporate bonds is not the safe harbor many investors think they own. It is a leveraged expression of “nothing really bad happens.”

Investors were right about the long‑term importance of fiber in the 1990s, but the timing mismatch destroyed bondholder capital. Today, investors are probably right about the long‑term importance of AI. The problem is the bond math: rates are higher, investment grade spreads—while off their very tightest levels—remain historically thin, and this issuance cycle is measured in trillions, not billions. When spreads are this tight and the macro tape looks like this, the juice simply is not worth the squeeze.

We are avoiding credit risk in this environment and recommend investors do the same. For now, we prefer the simplicity and safety of short‑term U.S. Treasury bills until the wars and their knock‑on effects are better understood and credit spreads truly compensate for the risks on the table. Sometimes the hardest move in portfolio management is the easiest to explain: you do not have to reach.

Subscribe now

The Odyssey is about to hit theaters, but investors are already living through their own preview. In Homer’s tale, Odysseus must thread his ship between two monsters - Scylla, the rock‑dwelling devourer, and Charybdis, the whirlpool that can swallow the entire vessel - with no safe route, only a narrow strait where the best you can do is choose your damage and keep moving. That is where I believe markets now find themselves: squeezed between the inflation monster on one side and the recession vortex on the other, with the Federal Reserve - the supposed captain of the ship - effectively below deck and making no visible course corrections.

For years, investors could pretend these monsters were distant myths. Inflation was “transitory,” recessions were mild and quickly papered over with liquidity, and every dip felt like a buying opportunity. War and geopolitics have changed that script. The war in Iran is choking global energy supply and, as Niall Ferguson reminds us in The Free Press1, history shows shocks like this “rarely end without a recession.” Energy supply disruptions, higher defense spending, and renewed fiscal looseness have resurrected the inflation threat just as the cumulative impact of past tightening and higher real rates is starting to weigh on growth. The result is not a clean “inflation or recession” choice, but a messy, overlapping risk of seeing both monsters in the same passage.

In this updated Odyssey, inflation plays the part of Scylla. It is no longer the theoretical beast policymakers hoped they had slain in 2023. Hedgeye2, an independent research firm whose macro work now explicitly points to a renewed inflation up‑leg, with a reasonable path to 4%‑plus headline CPI by the fourth quarter as energy and other commodities re‑accelerate. War‑related energy shocks, higher transport and insurance costs, and the second‑round effects of elevated wages and government spending all support that view. It doesn’t take much: a spike in crude, shipping reroutes through longer and less secure routes, and another round of fiscal stimulus or military outlays, and 4% headline stops being a “wild” scenario and starts to look like the central tendency of this regime rather than the tail.

Scylla is dangerous not just because prices rise, but because she lives in a narrow channel. The closer the ship sails to her, the more likely she is to tear away men - but hugging the other side is not safe either. In market terms, the inflation monster threatens to eat away at real returns, compress multiples for long‑duration assets, and keep policy tighter than markets would like even if central bankers are reluctant to say so aloud. With 4% inflation and “sickening” fiscal arithmetic, 4.5% nominal yields on longer‑term bonds are not a savior; after inflation, the real pickup is marginal, and duration risk remains significant. At those levels, investors are not being overpaid for stepping out the curve; they are being offered a modest real return to underwrite fiscal behavior that looks increasingly unsustainable.

On the opposite side swirls Charybdis: the recession vortex. Ferguson’s warning is straightforward: the war in Iran is choking global energy supply, and shocks like this rarely end without a recession. Higher energy costs act like a tax on households and businesses; margins get squeezed, real incomes weaken, and at some point, demand simply buckles. If policy has already been tight and balance sheets are not pristine, that buckling turns into a full‑fledged downturn as credit spreads widen and earnings expectations reset lower. Historically, that has not meant a polite 5% dip and then “all clear.” Across post‑war U.S. recessions and associated bear markets, the stock market has often fallen 20%–30% or more from peak to trough, with several episodes in the 30%–50% range. A “standard” recession is entirely consistent with 20%–30% drawdowns from highs; investors conditioned to expect a quick, shallow correction are sailing with the wrong chart.

Complicating all this is the behavior - and now the composition - of the captain. The Federal Reserve is formally “data‑dependent,” but in practice the reaction function looks backward‑looking and slow in a regime that is changing rapidly. Policy settings still reflect yesterday’s soft‑landing narrative, not today’s war‑driven supply shock combined with persistent fiscal excess. On top of that, we are dealing with a changing captain and an undercurrent of personal animosities among key players that feels unprecedented for a modern central bank. It is hard enough to steer between Scylla and Charybdis with a unified bridge crew; it is far more dangerous when the officers are divided, some are about to leave the ship, and others are more focused on internal score‑settling than on the rocks dead ahead. In the Odyssey metaphor, the instruments are blinking, the crew is shouting about changing currents and rising winds, and the captain is down below, distracted by politics and personnel instead of the storm forming off the bow.

What should investors do in this strait? I would resist the temptation to declare victory over either monster or to anticipate an early, “clean” recession that quickly resets valuations to bargain levels. With inflation likely to push back toward 4% and fiscal policy deteriorating, I do not expect a garden‑variety downturn where a mild growth scare instantly makes 4.5% bonds and 5% pullbacks irresistible gifts. Recession risks are real, but the path is likely to be choppy, and the policy response slower and more constrained than the market’s muscle memory assumes.

Against that backdrop, “sell the rallies” is not a slogan; it is a discipline. When the market surges on a dovish interpretation of a speech, a temporary reprieve in oil, or a better‑than‑feared data print, I would treat that as an opportunity to reduce equity risk rather than chase it. My bias is to move methodically toward a target where only about one‑quarter to one‑third of total assets sit in equities, recognizing this as a range, not a precise point. In a regime of overlapping inflation and recession risks, with geopolitical instability and a distracted, divided captain, that means less sail up and more ballast in the hull.

What does that ballast look like - and where does gold fit? I continue to believe the long‑term case for gold is strong: structurally higher inflation risk, strained fiscal arithmetic, and eroding confidence in fiat promises all support owning real monetary assets over time. But gold also just came off a massive year‑end surge that left it technically overbought, with volatility uncomfortably high. After that kind of move, it is unrealistic to expect instant gratification from fresh allocations. I would not disturb core strategic gold holdings, but I would not assume that gold will provide a smooth or immediate hedge against every equity drawdown while it works off that overbought condition. Patience is essential: the destination still looks attractive, but the path will be bumpier than many would like.

The Odyssey did not end because Odysseus found a way around the monsters; it ended because he acknowledged their power, accepted some losses, and kept his focus on reaching shore. Investors don’t need to forecast the precise path of inflation and growth, or the exact date of the next policy move. They do need to recognize that the easy, one‑way era is over, the war‑shadowed regime is more treacherous, and the captain - changing, distracted, and riven by personal animosities - is not guiding them out of danger in real time. This is the moment to shorten sail, secure the rigging, keep your gold ballast in place, and be ready for the kind of 20%–30% equity waves that a real recession can still deliver.

Subscribe now

The investing world looks like it is quietly pivoting from a decade of financial engineering and pricey, illiquid alternatives toward a far older source of returns: real assets, with oil now leading the way. The Federal Reserve has not yet blessed this regime shift with a full easing cycle, and long-term inflation expectations remain anchored, with the 5-year, 5-year forward inflation rate - the market’s best guess of average inflation in years 6-10 from now - still sitting just a bit above 2%. But if most investors remain concentrated in S&P 500 index funds and U.S. government bonds in classic 60/40 portfolios, we think they could find themselves badly underexposed to the asset mix this new paradigm demands. We believe you should consider moving toward a 5-10% allocation to commodities if you haven’t already; we began this process late last year - likely more once we see how the Fed chooses to navigate the current crisis.

Why 5y5y matters for this story

The 5-year, 5-year forward inflation rate sounds arcane, but the idea is simple. It answers the question: “Starting five years from now, what inflation rate does the bond market expect, on average, over the following five years?” Technically, it is derived from nominal Treasury and TIPS yields at the 5- and 10-year points, backing out the implied inflation over years 6-10.

Policymakers use this number as a reality check on their credibility. If the 5y5y stays near target - roughly 2% in the U.S. - the Fed can argue that long term expectations are “anchored,” even if current inflation is noisy. That is essentially where we are now: the 5y5y has drifted up only modestly since the pandemic and recent inflation spikes and remains near its pre‑COVID range. In other words, the bond market has not yet fully priced a commodity driven inflation regime; that tension between muted expectations and tightening real asset fundamentals is precisely where opportunity lives.

Why bonds no longer carry the load

For years, bonds were the all-purpose answer to portfolio design: ballast, income, and risk reduction, all for the low price of tracking error versus a policy benchmark. Look under the hood and the math is less generous. The Bloomberg Municipal Bond Index, for example, recently offered a yield to worst in the mid 3% range; dress that up for tax benefits and you can tell yourself you are “earning” around 6%, but in real, after inflation terms you are barely treading water - and that assumes you believe official inflation statistics.

In many cases you are effectively paying highly indebted borrowers for the privilege of lending them money. In an absolute return framework, longer duration bonds simply do not carry their historical weight. The one exception we make is low duration, high grade paper and T bills, held not as return engines but as volatility dampeners. That is a harsh statement about an asset class that still occupies 40% of the iconic 60/40 portfolio and often more in institutional lineups. When the presumed risk-free asset no longer reliably preserves purchasing power, investors are forced to look elsewhere for both income and inflation insurance.

The commodity rotation, with oil at the core

We began pushing in that direction years ago by telling investors to fund large gold positions directly out of their bond sleeves - at times as high as one third of portfolio assets, and 25% as a long-term core allocation. Over the last year, the thesis has broadened from a “gold versus bonds” trade into a full-fledged commodity rotation. We have urged clients to sell remaining bond positions and redeploy into commodities and related equities, with a particular emphasis on energy. Recent flare ups in the Middle East have accelerated that move, but they are not the argument; they are simply catalysts within a much larger structural story.

The key change is that the energy transition narrative has finally had to confront arithmetic. Late in 2025, even the IEA conceded that oil demand will not decline meaningfully in the coming decade - and in fact will increase. That admission arrived after years of chronic under investment in traditional energy infrastructure. The industry is now discovering less than half as many barrels per year as it did last decade while spending roughly 300 billion dollars less annually than at the last capex peak, and nearly 90% of current upstream spending is being consumed just to offset natural decline in existing fields.

The issue is not that the planet is “running out” of oil in a geologic sense; it is that the cost - financial, technical, political - of extracting the marginal barrel is rising. Since 2015, the marketed offshore rig fleet has shrunk by roughly one third, with some segments down nearly 60%. That is not a spreadsheet curiosity; it shows up in the contracting calendar, where most awarded rig days are now booked years into the future. Building a new ultra deepwater ship takes four to five years and close to a billion dollars, and at current day rates contractors are understandably reluctant to commit fresh capital. Overlay mounting evidence that U.S. shale - especially the Permian, which provided the overwhelming bulk of global supply growth last decade - is hitting peak growth, and the outline of a higher for longer oil price regime comes into focus.

We’ve seen this movie before

If this sounds like the setup to a commodity “supercycle,” that is because we have lived through similar episodes. In the 2000s, a powerful boom in China and the broader emerging world drove a decade long surge in commodity prices, with crude oil, copper and a host of metals and agricultural products vastly outperforming developed market equities. Brent crude and base metals such as copper hit record highs into 2011, powered by strong demand and constrained supply growth. Investors who treated commodities as a niche sideshow missed a major macro trade.

Go back further and the 1966-1980 period offers an even starker template. A mix of war spending, social programs, easy money, and repeated oil shocks drove a long, grinding rise in inflation and a powerful run in real assets. Gold rose roughly five-fold in the 1970s and then spiked about 400% from the late 1970s to its 1980 peak, while oil prices jumped more than 200% around the time of the Iranian revolution and Iran–Iraq war. The result was a long era in which commodities and real assets dramatically outperformed financial assets, and traditional stock-bond mix struggled to preserve real wealth. The common denominator: underappreciated inflation pressure colliding with constrained supply.

Starting near market weight - and willing to go higher

Today, long-term inflation expectations as seen through the 5y5y are still calm, hovering only a bit over 2% and not far from pre pandemic norms. The Fed has not yet chosen between easing into commodity-driven inflation or staying tight into a slowdown. We think that is precisely where the opportunity lies. Commodities, broadly defined, represent on the order of 10% of the global investable universe - much smaller than global equities and bonds but far from a trivial rounding error. We therefore think of a 5-10% allocation not as a wild overweight but as starting roughly at or slightly below market weight, with room to increase if the regime shift we expect continues to unfold.

In practical terms, we think investors should target that 5-10% commodities and commodity-linked allocation as a standing portfolio position, funded primarily out of long duration bonds. That is a starting point, not a ceiling. Once it becomes clearer how the Fed intends to navigate this crisis - and whether it is willing to risk higher realized inflation to cushion markets - we can argue for larger weights. Within that sleeve, we favor a mix: direct commodity exposure, energy producers, services, and select metals and minerals that benefit both from chronic under investment in legacy energy and from the enormous physical demands of any realistic energy transition.

Bonds and broad equity indexes will not disappear from portfolios, nor should they. But they should stop monopolizing portfolios. The earlier investors begin reallocating toward real assets, the less they will have to do it under duress if and when this new regime becomes obvious to everyone else.

Subscribe now

Diversification started as a survival trick, not a math exercise. People who lived off ships and trade routes understood one hard truth: if one voyage went bad, it could ruin you. That’s the world Shakespeare captured in The Merchant of Venice. Antonio isn’t bragging when he says, “My ventures are not in one bottom trusted, nor to one place; nor is my whole estate upon the fortune of this present year.” He’s explaining risk management. His cargo isn’t on one ship, it’s not all headed to one port, and his future doesn’t depend on a single season. For a 16th‑century merchant, that was the difference between a bad year and bankruptcy.

Translate that into today’s terms and you get the basic investment idea everyone’s heard - don’t put all your eggs in one basket. Return is the easy part. If half of what you own earns 10% and the other half earns 8%, you end up around 9%. You don’t need a finance degree for that. The subtle part isn’t how each investment behaves on its own - it’s how they behave together when the world gets punched in the face. If everything you own tends to get hit at the same time, you’re not really diversified, no matter how many line items show up on your statement.

The “behave together” idea is what professionals call correlation. You don’t need the formula; you just need the picture. If two investments almost always move in the same direction at the same time, owning both is like owning a lemonade stand and a surfboard rental at the same beach. When it’s cold and gloomy, they both suffer. If they often move differently, one zigs when the other zags, they’re more like a lemonade stand at the beach and a hot chocolate stand at the bottom of a ski mountain: a bad day for one isn’t automatically a bad day for the other. Over time, this pattern of moving together or not ends up driving your overall risk more than how “jumpy” each piece is on its own.

For most investors, the 60/40 portfolio - roughly 60% stocks and 40% bonds - became the modern version of Antonio’s fleet of ships. From about 2000 through 2020, that mix didn’t just work; it looked brilliant. Stocks delivered growth while bonds threw off income and, just as important, often acted as a shock absorber when stocks sold off. In those two decades, stocks and bonds frequently moved in opposite directions, so the bumps from one were smoothed by the other. It felt like a beautifully engineered machine: one part sped up just as the other slowed down.

Look at a chart of the rolling stock–bond correlation over the last half‑century and you can see the regime shift. The 1970s, 1980s and much of the 1990s show mostly positive or unstable correlation - stocks and bonds often rising and falling together in a world where inflation shocks and rate volatility dominated. Post‑2000, the line spends most of its time below zero, marking the longest sustained stretch of negative correlation on record in the U.S. That unusual 2000–2020 window is what trained investors to believe “when stocks fall, bonds will save me” was almost a law of nature.

Historic Stock-Bond Correlation

Source: Created by Proficio Capital Partners. Data from Robert J. Schiller (schillerdata.com). Stocks are S&P 500 nominal total returns. Bonds are nominal 10-Year US Treasuries. The series being analyzed is the rolling 10-year stock-bond monthly correlation.

Here’s the psychological trap: when a relationship holds for twenty years, our brains quietly promote it from “pattern” to “physics.” Two full decades of bonds reliably helping when stocks hurt made the 60/40 portfolio feel less like a strategy that fits a particular era and more like a permanent rule of how markets work. But if we zoom out, the longer history tells a different story. From roughly 1965 to 2000, inflation fears ran supreme as the government tried to deliver both “guns and butter” - war and social programs at the same time, with persistent pressure on prices. In that world, inflation shocks hurt both stocks and bonds: bonds suffered as yields rose to keep up, and stocks suffered as higher rates and uncertainty compressed valuations. For long stretches, they moved together instead of offsetting each other.

If today’s governments also want “guns and butter” - large defense budgets, ambitious social spending, and little appetite for restraint - why are we so sure the 2000–2020 chapter is the template, and not the earlier 1965–2000 experience? High debt levels, repeated fiscal packages, and renewed geopolitical tension rhyme far more with that inflation‑sensitive era than with the calm, disinflationary 1990s. After 2020, those echoes turned into data. Inflation woke up. Supply chains frayed. Stocks and bonds suddenly shared the same core problem: both disliked rising inflation and rising interest rates. Instead of one zigging while the other zagged, they started to slump together when inflation or rate shocks hit.

In 2022, this stopped being theory and became a line on your statement. Stocks had a bad year, and, for many investors, their “safe” bond allocation dropped sharply at the same time. A traditional 60/40 portfolio suffered one of its worst years in modern history. The seat belt that was supposed to save you in a crash failed at the exact moment you needed it most. What had felt like a law of physics turned out to be one favorable twenty‑year chapter sitting on top of a much messier history.

This is where gold enters the story, not as a relic, but as a different kind of “ship” in the fleet. Gold cares less about quarterly earnings or next year’s GDP print and more about one thing: sustained purchasing power. Over long stretches, its role has been to keep pace with the erosion of paper money, not to compete with stocks for growth or bonds for income. That makes its behavior under stress very different. Because gold’s main job is to track the loss of fiat purchasing power, it tends to come into its own in the exact environments that hurt both stocks and bonds: persistent or surprising inflation, doubts about fiscal discipline, concern about currency debasement. When inflation rises and eats into bonds’ fixed payments, and higher rates compress stock valuations, gold is not mechanically wired into that punishment. It can, and often does, go its own way - rising as investors look for something that isn’t just another claim on future cash flows in a weakening currency.

That means gold often passes both the qualitative and quantitative tests of a true diversifier. Qualitatively, you can explain why it matters without mentioning GDP, profit margins or the next Fed meeting: its story is about purchasing power. Quantitatively, over full cycles - especially inflationary or confidence‑shaking ones - gold has frequently shown low or even negative correlation with traditional stock and bond portfolios, particularly when those portfolios are under the most stress.

Gold isn’t the only way to step outside the stock‑and‑bond weather system, but true examples are rare. Consider riparian water rights in the U.S. Southwest. Access to water in a dry region is needed whether GDP is up 3% or down 2%. Value there is driven by scarcity, local law and physical necessity, not the credit cycle or earnings season. The same goes for some kinds of mineral royalties or infrastructure access where cash flows depend more on usage and scarcity than on broad market risk. These are the sorts of assets that can, in principle, pass the uncorrelated test: you can describe why they make money without talking about “the economy,” and their prices don’t reliably rise and fall with stocks or bond yields.

The catch is that these investments are hard. You don’t buy serious water rights with a three‑letter ticker. You need to understand law, local politics, supply and demand, contract terms, and illiquidity. They demand research and discipline. And very few things marketed as “alternatives” actually behave in alternative way. Most so‑called alternatives quietly smuggle in equity risk, credit risk, or rate sensitivity under a different label. That’s the key differentiator: a true uncorrelated alternative has a different engine; a pseudo‑alternative just paints the same engine a different color.

If you’re not a math person and don’t want to become one, the practical question is simple and very old: how do you arrange your ships so that one storm can’t sink you? The answer isn’t to worship 60/40, or gold, or any single formula. It’s to be brutally honest about what really drives the things you own - and to deliberately add assets whose engines are not all wired to the same macro story. In a world that increasingly rhymes with the inflation‑sensitive decades of 1965–2000, diversification must go back to what it was for Shakespeare’s merchant: a survival skill, not a spreadsheet trick. The goal isn’t to squeeze out one more decimal place of return in calm seas; it’s to make sure no single storm - no inflation surprise, no policy mistake, no loss of faith in money itself - can decide your entire financial fate.

Subscribe now

James Carville said that if there were reincarnation, he wanted to come back as the bond market because “you can intimidate everybody1.” Back in the mid‑1990s, after a brutal backup in yields during Bill Clinton’s first term, Carville - the “Ragin’ Cajun” political strategist who engineered Clinton’s 1992 victory and later served as a senior presidential adviser - learned the hard way that bond markets can veto bad ideas faster than voters ever will. In 2026, that lesson feels almost quaint; the real power now sits where nominal growth, term premium, and quiet money printing collide at the long end of the U.S. Treasury curve.​

A 7% Economy Priced Like 3%

Start with the most basic mismatch: the economy versus the yield curve. With real growth running in the 3–4% range and inflation still comfortably above the Fed’s 2% target, nominal GDP has been printing close to 7%. That is not a “Goldilocks” environment; that is a hot economy, especially when you consider what is coming next: one big fiscal bill already in the pipeline, a likely pivot to lower short‑term policy rates, and an incremental push toward deregulation that tends to free up credit creation and risk‑taking.​

Historically, long‑term government bond yields have had a loose but persistent tendency to track nominal GDP over time2. They may sit modestly below nominal growth when investors believe in a durable disinflation story, or above it when inflation and fiscal policy look unhinged, but they rarely stay miles apart for long. Today’s long‑bond yields, still anchored far below recent nominal growth, are effectively a bet that this 7% world is an illusion - that both growth and inflation will glide back to the pre‑pandemic “2 and 2” regime without leaving a scar on the price level or the term structure. That is a bold assumption in a world of trillion‑dollar deficits, rising geopolitical risk, and a political class on both sides of the aisle that treats fiscal restraint as optional.​

Term Premium: Still Not Scared Enough

The term premium is supposed to be the market’s fear gauge for lending long. It is the extra yield investors demand to hold a 10‑ or 30‑year bond instead of rolling a series of short‑term bills, compensating them for interest‑rate risk, inflation uncertainty, and fiscal or political surprises. The Federal Reserve, the New York Fed, and a host of private models all try to infer this premium from the shape of the curve and survey expectations for future short rates.​

What those models are telling us now is unsettling: despite higher headline yields, the inferred term premium remains below its long‑run average and well under the peaks associated with prior episodes of bond‑vigilante revolt. In plain English, investors are still not being paid much extra to take on long‑dated Treasury risk versus sitting in short bills, even as nominal growth runs hot and fiscal arithmetic deteriorates. When you see a low term premium in a high‑nominal‑growth, high‑deficit world, one of two things is wrong: either the macro backdrop is about to cool dramatically, or long bonds are mispriced.​

This is where Carville’s line meets reality. A subdued term premium is the bond market whispering, not intimidating. But history suggests that when macro and policy conditions look like this, whispers can turn into screams quickly, with a sharp repricing at the long end as investors suddenly insist on being paid for duration and fiscal risk.​

RMP: New Acronym, Old Habit

Now layer in the Fed. Late in 2025, the Federal Reserve introduced another three‑letter program - Reserve Management Purchases, or RMP - under which it is buying Treasury bills again at a pace currently around 40–50 billion dollars per month. Officially, this is “not QE,” just a technical step to maintain “ample” reserves, smooth TGA swings, and ease year‑end funding noise. Practically, it means the Fed has stopped shrinking and is once again quietly expanding its balance sheet via outright purchases of government paper.​

The names do not matter; the balance sheet does. Markets have seen this movie before: whenever the Treasury’s funding needs run ahead of the private sector’s appetite, some new acronym appears - QE, RMP, SRF, and some future variant - to sanitize the reality of renewed money printing. If you own long‑duration bonds, what you are long is a promise that the central bank can keep inflation expectations anchored while repeatedly expanding its balance sheet to absorb public debt. That is not a risk‑free asset; it is a leveraged macro bet.​

The Only Balance Sheet Big Enough

Meanwhile, the supply side of the equation is exploding. Darius Dale at 42 Macro3 estimates that, over the next twelve months, the U.S. Treasury will need to market on the order of 12 trillion dollars of gross issuance - refunding existing debt plus new borrowing. He characterizes that flow as roughly 230% of all U.S. domestic savings and about 40% of total global savings, a claim on the world’s balance sheets by a single sovereign that has no modern precedent.​

At the same time, traditional foreign anchors are wobbling. Japanese investors, long stalwarts of the Treasury market, are increasingly staying home as rising domestic yields and a tentative exit from ultra‑easy policy make yen assets more attractive. Other foreign official buyers are either diversifying away from Treasuries for geopolitical reasons or simply unable to keep pace with the scale of issuance. Put starkly: if foreigners buy less, and domestic savings are already claimed, the only balance sheet large and flexible enough to bridge the gap - if current trends persist - is the Fed’s.​​

That is where term premium and nominal growth intersect with institutional reality. The more the Fed is forced into RMP‑style purchases to keep the Treasury market functioning, the more investors will question its independence and the more they will demand compensation at the long end for the risk of de facto fiscal dominance. Once that psychology shifts, term premium can move hundreds of basis points, not tens, and long‑duration bonds can deliver equity‑like drawdowns - with none of the upside.​

How To Position When Vigilantes Ride

In a world of 7% nominal GDP, suppressed term premium, RMP‑fueled balance‑sheet expansion, and a 12‑trillion‑dollar Treasury funding calendar, the most vulnerable asset class is the one investors still call “safe”: long‑duration, investment‑grade bonds. That is where you are most exposed to three converging risks - higher real rates, a normalization of term premium, and a loss of faith in the Fed’s ability (or willingness) to resist fiscal pressure.​

I believe the portfolio implications are straightforward:

• Sell or materially underweight long‑duration investment‑grade bonds, especially at the 10‑ to 30‑year maturities where an adjustment in term premium will hit hardest.

• Maintain a healthy allocation to short‑duration paper - T‑bills and very short notes - so you earn an acceptable yield today while preserving the optionality to buy future bargains if and when the long end blows out.​

• Hold meaningful exposure to commodities and related real‑asset plays, which historically benefit when the U.S. runs “hot” on nominal growth, fiscal expansion, and repeated rounds of balance‑sheet support.​

• Hedge the entire structure with hard currency in the broad sense: physical gold and silver, plus carefully selected miners that provide leveraged upside to an environment of ongoing financial repression and skepticism about fiat.​

Do not bother learning the acronyms; they will keep changing. Do pay close attention to the combination of nominal growth, term premium, and quiet money printing if you own long‑duration bonds. That trio is exactly what empowers Carville’s bond market to “intimidate everybody” - and it is back in the saddle.

Subscribe now

That is the most important risk rule in asset allocation.​

“I’m sorry” is what you say when a perfectly sensible holding lagged for a year. “I’m sorry we didn’t keep up with the Mag 7.” “I’m sorry that hedge cost a bit in a melt‑up.” “I’m sorry we owned ballast when momentum was the only religion.” Those are the ordinary regrets of a grown‑up portfolio. They are the cost of not driving 95 miles an hour in a rainstorm just because the road looks dry.​

“I’m really sorry” is different. That is what you say when you take the wrong kind of risk - one you did not fully understand, could not explain, or did not size properly - and the result is a major, possibly permanent loss of capital. That is when trust breaks. That is when the client stops asking questions because they have already decided you did not protect them when it mattered.​ A rational asset allocation accepts that you will almost always have something to be a little sorry about each year. One line item will look foolish next to whatever is working now. That is not a bug; that is the price of designing a portfolio where it is very hard to blow up the whole ship at once. The math behind this simply reminds you that what matters is not how each asset behaves in isolation, but how they behave together - how they co‑vary. Assets that do not move in lockstep, that zig when the others zag, can make the entire portfolio far more resilient, even when each component is risky on its own.

That is where hard currency assets come in. Gold, silver, platinum, palladium and the companies that mine them sit in a different place in the financial ecosystem than traditional stocks and bonds. They are not promises to deliver cash flows inside the fiat system; they are claims on real, finite tangible assets that do not care who wins the next election or where the policy rate settles next quarter. In a regime of fiscal profligacy and monetary improvisation – a debasement regime - those differences begin to matter more.

Bitcoin increasingly belongs in that conversation as well. It is not the same as gold, and it behaves differently in stress, but it shares important traits: it is outside the central bank balance sheet and features a hard‑coded supply path. For some allocators, that makes it a small but meaningful member of the hard‑currency family; a digital cousin to the metals that can add another imperfectly correlated leg to the stool.​

Let’s look back at 2025. It was not a “really sorry” year for most balanced portfolios. A plain‑vanilla 60/40 did well enough that many clients shrugged and moved on. Yet inside that comfort, something important happened: gold and hard currency assets quietly earned an “A”; not because they were the top performers every day, but because they did exactly what they were supposed to do in a policy backdrop biased toward dilution. Gold was up dramatically for the year, and it did so while the fiscal and external ledgers were flashing red and the central bank was quietly admitting that 3% might be the new 2%.​

That is covariance at work in the real world. When policymakers choose liquidity over discipline, and the currency does more of the adjusting, you want a dance partner in the portfolio that does not follow equities tick‑for‑tick. You want assets whose risk is different - not lower every day, not painless, just different. Hard currency assets have historically provided that: they tend to struggle in long disinflationary, credibility‑rich regimes, but they tend to shine when governments lean on the printing press, explicitly or implicitly, to fund themselves.​

The mistake is to judge those assets on a one‑year, stand‑alone scoreboard. In any given year, you can almost guarantee you will be “sorry” about your gold, or your miners, or your silver, or your bitcoin. They will lag the hottest sector at exactly the moment cocktail‑party conversation makes you doubt your discipline. But that discomfort is what you are paid to endure. The question is not “did gold beat the S&P this year?” The question is “if the currency itself is doing more of the adjusting over the next decade, what happens to my clients if I own no hard currency at all?”​

In a debasement regime, the real risk is not that you owned gold, and it went sideways for a year. The real risk is that you woke up one cycle too late and realized all your wealth was tied to claims inside the system - claims that only work if policymakers choose discipline over convenience. That is when “I’m really sorry” enters the chat: “I’m really sorry we didn’t own insurance when it was still affordable. I’m really sorry we let a decade of good returns convince us we did not need a dance partner for our stocks.”​

A better goal is simpler and more uncomfortable: accept that every year you will be a little sorry about something in the portfolio, including your hard currency sleeve. Size it so you can live with the drawdowns. Rebalance into it when it is out of favor instead of chasing whatever just had its moment. Treat it not as a trade, but as a permanent allocation to assets that are not someone else’s promise.​

If you do that, the odds of a catastrophic, “really sorry” outcome drop sharply. You will still have to explain why gold did nothing one year, why miners lagged another, why bitcoin scared the committee in a drawdown. But you won’t have to explain why a decade of quiet debasement, funded deficits, and financial repression left your clients with far less real wealth than the statements implied. In a world where policy keeps nudging us toward that path, gold and its hard‑currency cousins are not a speculation at the edge of the portfolio. They are the essential dance partners that make sure, when the music changes, you are not left dancing alone.

Subscribe now