Only paid subscribers can submit a question. Any questions a paid subscriber submits over the next week will be fodder for my next regular Friday Q&A post on Economics for Investors. So ask away in the comments to this post..

Now with no further ado, let me go to your questions.

This week’s three questions and answers are Is the dollar losing its global primacy--and what does that mean to me? Does the U.S. economy feel the world’s pain? Is Trump’s 2027 budget a serious budget?

Question #1: You’ve been publishing a lot of data recently showing the dollar losing share in global currency markets. Is the dollar losing its global primacy--and what does that mean to me?—Amy L.

I don’t think there’s any doubt that central banks and global governments are diversifying their portfolios to reduce their exposure to the dollar and lower their risk if the dollar continues its decline.

In 2025 the U.S. dollar fell about 8% in broad trade-weighted terms in 2025, and the DXY, the dollar index against six other currencies, fell roughly 9-10%. The dollar lost about 15% vs. the euro, 9% vs. the pound, and 6-8% vs. the Canadian and Australian dollars, while staying roughly flat against the yen

The Iran war led to a stronger dollar, showing that the U.S. currency has’t lost its safe haven status in volatile times. Since the Iran war began on February 28, 2026, the U.S. dollar has strengthened by roughly 2.8% to 3% on the DXY.

Those short term gains for the dollar have disguised the long term trend for the duration of the Iran crisis, but I don’t think the long term trend has changed: the dollar looks to be in the midst of a continued retreat that’s driven by important long-term fundamentals. The U.S. gross national debt is about $39 trillion as of early April 2026, while debt held by the public is about $31 trillion. The likely trend is still up, with the Congressional Budget Office projecting debt held by the public to rise from 101% of GDP in 2026 to 120% by 2036. The debt is projected to keep rising steadily, driven by persistent deficits (including higher health care and retirement outlays as the U.S. population ages) higher interest costs (net interest is projected to exceed $1 trillion in fiscal 2026), the long-term reduction in tax receipts from recent tax cuts, and Congressional unwillingness (or inability if you prefer) to reduce the gap between receipts and outlays.

In other words, there is no near-term flattening in the official trend line—debt continues climbing through the rest of the decade.

You can see the effects of that now as global central banks moved to derisk their portfolios. The share of gold in foreign central bank reserves has surpassed the share of U.S. Treasuries for the first time since 1996.

By the second quarter of 2025 or early 2026, central banks held ~$4 trillion in gold versus ~$3.9 trillion in U.S. Treasuries, driven by aggressive gold buying from China, India, Turkey, and others.

Gold’s share of reserves rose from ~10% in 2021 to ~13% by late 2025, while U.S. debt holdings stagnated.

This marks a structural de-dollarization shift amid sanctions risk, geopolitical tensions, and gold’s ~70% price surge in 2025.

The crossover happened in 2025, with official gold reserves now exceeding $5 trillion in market value.

I expect this trend at central banks, mirrored in the actions of private asset portfolios, to continue once the crisis of the Iran war presses a little less urgently on money managers.

So what’ the effect of this de-dollarization—and why is it important to you?

I don’t look for the overnight collapse of the dollar—despite its problem the dollar and the dollar markets still play a critical role in the world financial system.

Rather I see a steady erosion in the dollar and an increasing sense that the dollar is increasingly risky.

This will have two important effects.

First, it will push up interest rates on Treasuries and everything pegged to them from corporate bonds to mortgages. Look for rising interest rates in the U.S. economy—despite anything that the Federal Reserve does to lower short-term interest rates. This means more expensive mortgages, car loans, and business financing. And higher inters payments by the Federal government in order keep selling U.S. Treasuries. This all means that the trend will be falling bond prices and slower economic growth. (Which last makes it harder to make those interest payments.)

Second, the decline in the dollar and in dollar-denominated assets will lead to an increased search for alternatives. That’s good for gold and silver prices. And for other assets still to be determined.

Read more

About 20% of global oil supply and 20% of global LNG supply have been disrupted by closing the Strait to energy shipping and by the destruction of terminal infrastructure in counties like Qatar.

Roughly 20% to 30% of global fertilizer trade has been affected by the Strait of Hormuz disruption.

About one-third of global helium supply has been affected by the Iran war and closing the Strait of Hormuz. Qatar accounts for roughly one-third of global helium production. (Helium is mainly used in the manufacture of AI and other semiconductors for cooling, purging, and process control.)

The economics of supply and demand says the prices for these commodities will rise because supplies are scarce.

Eventually, and if the war goes on long enough or the damage to infrastructure is severe enough, companies will find ways to create new supply. It won’t happen overnight as many of the facilities needed by this new supply take years to go from investment decision to first production—2 to 3 years in many cases for LNG facilities, for example. Restoring production at the petrochemical plant in Iran bombed over the weekend will taken estimated 3 to 5 years. The new production will eventually bring supply and demand back into alignment and the excess profits created by the supply shock will disappear. And because these new sources of supply take so long to go from decision to production, it’s quite likely that the volume of new supply will exceed future demand at some point and the market will go from “not enough” to “too much.” You’d prefer not to be an investor in one of the latecomers that winds up adding new supply into a glut.

All that is to say, however, that if you can find a company that looks likely to be one of the first to add new supply into a market with an excess of demand over supply—where that supply shortfall is likely to last for a couple of years (as seems likely in the case of the Iran war judging from estimates of three to five years to repair damage to infrastructure)—then you’ve probably found a good candidate for a profitable investment.

One sector I’d target for a dollar or two in my portfolio on supply shortfalls created by the Iran war is LNG. The consensus view before the war was that in 2026 LNG supply would grow faster than demand, creating likely downward pressure on prices with a meaningful surplus or “glut” beginning in 2026. Reuters cited forecasts of 460-484 million metric tons of global LNG supply in 2026, up as much as 10% year over year.

Read more

Unfortunately, what the global economy is experiencing now because of the Iran war is a supply side shock. Because Iran has closed the Strait of Hormuz to energy shipping and because its missiles or the threat of its missiles have led countries such as Qatar to close down energy facilities, global energy consumers can’t get the supplies of oil, natural gas, helium, naphtha, and more that they demand to heat their homess, cook their food, fertilize their crops, or make their plastics.

And Powell was pointing out, there is squat all that the Fed can do about this supply side problem.

The Fed can mainly influence demand, not the underlying supply of oil, labor, or goods. If inflation comes from a supply shock, the Fed may monitor it and keep policy steady unless it starts affecting inflation expectations or causes broader second-round effects.

Thanks to John Maynard Keynes central banks have a pretty good framework for understanding the demand side of the economy and pretty good tools for addressing demand-side slumps. Keynes thought the economy should be understood mainly through aggregate demand, not through the idea that supply automatically creates its own demand. In his view, an economy can settle at too low a level of output and employment because demand is insufficient, even if companies are able to produce more.

John Maynard Keynes: The portfolio he managed for King’s College outperformed the UK market by about 6 to 8 percentage points a year in the 1930s and 1940s.

(Before Keynes, economic orthodoxy, embodied in Say’s Law and the work of French economist Jean-Baptiste Say (January 5, 1767, to November 15, 1832) , held that production creates income and therefore creates demand for other goods and services. People often summarize this as “supply creates its own demand. When a good or service is produced and sold, the seller earns income that can be used to buy other goods and services. So, in the classical view, the act of production is what ultimately generates demand in the economy. Keynes pointed out that this is not literally true in all cases, especially because money can be held rather than spent and because recessions can involve weak demand.)

Our tools on the supply side are much less robust and as applied in supply side economics have a disputed track record. It’s clear that targeted government spending, tax breaks, and subsidies, for example, can influence supply, but it’s not clear what the lag might be nor how effective these measures might be in the face of market disincentives to increases in supply.

Powell wasn’t reminding the financial markets of the difficulties of designing policies to increase supply. His intention was, I think, to suggest that the financial markets remember that the Fed focuses on demand and that investors and traders should, therefore, project that the Fed will ignore the current supply shock in setting interest rates. A rate cut, I think he was implying, was more likely than the financial markets thought at the moment.

However, I believe, you and I should take a moment to think about the problems and difficulties of increasing supply. Because right now the current economic and financial market situation presents something close to a supply side crisis in some key areas of the global economy.

Read more

Only paid subscribers can submit a question. Any questions a paid subscriber submits over the next week will be fodder for my next regular Friday Q&A post on Economics for Investors. So ask away in the comments to this post..

Now with no further ado, let me go to your questions.

This week’s three questions and answers are Why has “safe asset” gold been anything but safe? Who’s right on stagflation, the Fed or the ECB? What about the great market rotation?

Question #1: Why has gold performed so poorly since the attack on Iran? I thought gold was supposed to be safe? Paul P.

A hedge that works in one market doesn’t necessarily work in another.

Gold is “normally” a great hedge on market volatility and geopolitical uncertainty.

But this time some “special” circumstances combined to make gold a spectacularly ineffective hedge.

Photo by Jingming Pan on Unsplash

To summarize:

• Gold gained about 110% from January 1, 2024 to December 31, 2025, rising from $2,062.90to $4,315.13 per ounce. That is an increase of about $2,252.23 per ounce.

• At the start of 2026 many traders and investors—and central banks—had essentially doubled their money in gold since the start of 2024.

• That meant there was a substantial reservoir of doubt in market sentiment—anyone who had made that much money in gold was inclined to stay on board for further gains but was worried about a “too-far/too-fast” correction.

• That combination meant that when financial market stress arrived with the start of the Iran war, many holders of gold had some doubts about holding onto that asset.

• When they looked around for what to sell to meet margin calls on other assets, they were inclined to think “Sell gold.”

• And that included central banks for energy importing countries that now faced increases in the cost of imported oil and natural gas—and drops in the value of their local currencies against the dollar. (Oil is denominated in dollars.) To support their local currencies central banks sold dollars. Central banks in countries that are energy importers have been taking profits from the run up and using those profits to support currencies that are under pressure as dollar-denominated oil soars in price.

• That squeeze seems to be especially severe for Asian energy importers. For example, Turkish gold reserves show a decline of 6 tons in the week of March 13 and another 52.4 tons in the week of March 20, marking a sharp drawdown in reserves.

• Gold sales and swap arrangements from Turkey alone total an estimated 58.4 tons. Which exceeds total outflows from gold-backed exchange-traded funds tracked by Bloomberg, which were about 43 tons over the same two-week period.

• The correction is a “buying opportunity” once tensions in the Middle East subside, Fidelity International money manager George Efstathopoulos told Bloomberg. “Inflation risks, fiscal pressures, and bond credibility are all still structural tailwinds for gold.”

• I’d stress the “once tensions in the Middle East subside” part of that quote. And add “once oil prices stabilize and maybe even retreat.” What central banks of oil importing countries fear most right now is a short-term run up in oil prices that leaves them short of funds to defend their currencies.

• In other words, yes, there will be a huge opportunity to put the gold trade back on--but not until some of the oil price risk comes out of the market.

Read more

Only paid subscribers can submit a question. Any questions a paid subscriber submits over the next week will be fodder for my next regular Friday Q&A post on Economics for Investors. So ask away.

Now with no further introduction, let me go to your questions.

This week’s three questions and answers are about what happens if there is no new chair at the Fed by the time Powell’s term is up, why 4.51% on the 10-year Treasury is important, and what the VIX is telling us.

Question #1: Jerome Powell’s term as chair of the Federal Reserve ends in May and Kevin Warsh’s nomination to replace him seems to be in trouble in the Senate. What happens if Warsh isn’t confirmed by May?—from Sara K.

Concretely nothing happens. The intangibles could be very important, however.

If North Carolina Republican Senator Thom Tillis continues to hold up President Donald Trump’s nomination of Kevin Walsh to the Federal Reserve, the Fed can simply choose a temporary chair. The rules are set up to allow this by either having the current chair continue in that role or by having the Board of Governors elect a temporary chair from among their numbers. The President does not have the power to pick a temporary chair, but it’s not completely impossible that President Trump could throw everything into chaos by insisting he can pick the temporary chair. Short of that possibility, though, Powell’s term ending before there’s a replacement confirmed by the Senate is a non-issue.

In concrete terms.

The intangibles could be very different. Without a confirmed replacement, the Fed is unlikely to feel comfortable with a big change in policy—like resuming interest rate cuts. But at this juncture not cutting rates in, say, October, would also be a big policy statement. Especially when the White House has made it so clear that it wants lower interest rates yesterday.

In this situation, we could be in for another episode in President Trump’s attempts to bring the Federal Reserve under the control of the White House.

That’s not exactly what a dollar facing serious questions from other governments and markets needs.

Read more

I think we’re in the early stages NOW of a credit crisis but that we’re not seeing the full danger because we think this crisis has to look like the last one.

It’s going to take me two parts to explain this idea. This is Part 1. Look for Part 2 tomorrow.

Those of us who were investors—or simply struggling with the larger economic meltdown—during the Minsky Moment event now labeled the Global Financial Crisis (mid‑2007 to early 2009, with the most acute phase in 2008) are excused for thinking that is how the credit crashes outlined by economist Hyman Minsky in the 1970s and 1980s play out. Credit problems in one part of the financial markets—the sub-prime mortgage market in 2007—caused by the collapse of lending standards ripples out into the larger market, causing the collapse of a major institution or two, Lehman Brothers and Bear Stearns, and becoming, finally a systemic crisis.

Read more

Only paid subscribers can submit a question. Any questions a paid subscriber submits over the next week will be fodder for my next regular Friday Q&A post on Economics for Investors. So ask away.

Now with no further introduction, let me go to your questions.

This week’s three questions and answers are about the possibility that the Iran war will cause a recession, the K-shaped economy, and the shocking state of disrepair in the unemployment safety net.

Question #1: Will the war in Iran push the U.S. economy into a recession?—from Erica N.

Probably not. On its own. But it may well be a close run.

The economy wasn’t exactly growing like a mushrooms after a September rain at the end of 2025. On Friday, the Commerce Department downwardly revised its estimate of economic growth in the fourth quarter of last year, saying that gross domestic product expanded at just 0.7% annual rate in the final three months of 2025. The initial report had put growth at a 1.4% rate, the department had reported in an “advance” estimate last month.

In most circumstances, the Federal Reserve would be looking to cut interest rates to stimulate age economy. But these aren’t most circumstances. Inflation, before the Iran war caused a big spike in oil prices, remained well above the Fed’s 2% target. On Friday a separate Commerce Department report showed that underlying inflation remained stubbornly elevated in January. The Federal Reserve’s preferred inflation gauge, the Personal Consumption Expenditures price index, rose 2.8% over the prior year, a slight decline from December but still well above the Fed’s 2 percent target.

And core PCE inflation--which strips out volatile food and energy prices and which is the inflation measure the Fe watches most closely--ticked up to 3.1% annual rate in January from 3.0% in December.

On Friday, as well, Brent crude, the European benchmark for oil, finished the day above $103 a barrel, up more than 40% since the U.S./Israeli war against Iran began almost two weeks ago.

That’s not good news for immediate inflation. And it’s not good news for medium term inflation when higher oil prices will work their way into price increases for everything from air travel to food to plastic goods of all kinds.

The big Iran-war question is how much higher gasoline prices will act as a drag on consumer spending. Patrick De Haan, head of gasoline market analysis at GasBuddy, estimates that “Americans today will spend roughly $250 million more [a day] on gasoline than they did 30 days ago.”

That’s an extra $250 million day spent on gasoline rather than on apparel or restaurant meals, or travel or pet supplies.

Which seems like a lot of money—hey, it is a lot of money—until you stop to think, first, that U.S. consumers buy 375 million to 377 million gallons of gas a day. Which at a pre-war price of $3.00 a gallon works out to spending of $1.1–1.2 billion per day on gasoline. (The AAA National Average was at $3.63 per gallon for regular gas as of March 13.)

And, second, that household‑budget surveys show gasoline as a 2–3% slice of total consumer spending, translating into around $2,000 to $2,500 dollars per household per year at recent prices.

Now tacking on a 21% increase to the household gasoline budget—which is what a jump from $3.00 to $3.63 a gallon works out to is indeed a nasty increase of more than $500 a year in spending for a household that is already struggling with creeping inflation and a big increase in the cost of health insurance, but it’s not enough to damp consumer spending in the economy as a whole to produce a recession.

For that we need a bigger shock.

How likely are we to get one?

I don’t think the U.S. economy is about to provide one. Yes, lots of families are stretched to near the end of their endurance and it’s hard to see how they ca keep up current levels of consumer spending. But, wage growth has so far stayed healthy. Unemployment is, perhaps, edging higher but it is still relatively low. Levels of consumer debt are rising—which is to be expected if so many families are struggling to make ends.

U.S. consumer debt is high in dollars but not yet at past “danger” levels. Total U.S. household debt is a record about $18.8 trillion (Q4 2025), up roughly $4.5–$4.6 trillion since the end of 2019. As a share of GDP and income, though, it is lower than in 2007–08: household debt is about 61–69% of GDP and about 81% of disposable income, versus far higher ratios before the global financial crisis The Fed’s debt‑service ratio (debt payments as a percentage of disposable income) was around 8% in mid‑2025, compared with 11–12% at the 2007 peak; the long‑run average is near 9%.

I think the most accurate way to look at the current high level of consumer debt is that it in itself is unlikely to cause a crisis, but the fact that consumers are so stretched now, before any crisis, means that when a downturn comes it’s likely to be especially nasty and long-lived.

I’m looking at three potential sources of a shock to the system that would be big enough to trigger a recession. My candidates are

1. An implosion in some part of the credit market big enough to create worry about systemic risk to the financial system, My current candidate is the growing mess in the private credit market. I’m trying t track that here for you. My current take is that, so far, the private credit market is able to find buyers for its riskier debt. As long as enough buyers feel they’re being paid enough to take on the extra risk, the sector will muddle through.

2. Some big, unexpected international crisis. Ukraine or Iran could certainly stress the system, but I don’t think either is big enough to provide the size of shock we need to create a recession or financial crisis. Which doesn’t mean I don’t think there’s nothing to worry about. An unexpected use of nuclear weapons by North Korea (on the Korean peninsula or elsewhere) would do the trick. So would an attack on Taiwan by the People’s Republic of China,

3. A demonstration of astounding financial fecklessness by the U.S. government. Given that current levels of financial insanity have been sufficient to endanger the long-term standing of the US. dollar, but not to provoke a panic among U.S. debtors, I’m not sure what kind of event would rise to the level of a big enough short-term shock to cause a recession. But I’m pretty sure that there is one and past behavior in Washington also guarantees that someone will attempt something monumentally stupid. The fact that I can’t think what that might be doesn’t give me much comfort.

Read more

With big implication for Amazon and AI.

Photo by Yender Gonzalez on Unsplash

In early March, OpenAI revised its Instant Checkout shopping feature. I don’t think it’s an overstatement to describe the move as a retreat from the company’s attack on Amazon and the rest of the e-commerce sector. (Remember that Open AI is hungry for any new revenue source to pay for all those data centers.) The company announced that its Instant Checkout direct “buy” button would no longer be inside ChatGPT responses —and the company would now send users looking to buy to merchant apps within ChatGPT instead.

Instant Checkout was a major effort to insert ChatGPT and OpenAI into the search/buy stream and to set up an alternative marketplace—eventually—to e-commerce sites like Amazon that collect substantial revenue from being the intermediary between the digital shopper and digital stores.

I won’t go so far as to say that Amazon breathed a sigh of relief—when you have a 40% share of all e-retail in the United States, I doubt you panic over anything.

But for investors trying to figure out how—and when—disruptive AI is going to be to existing customer decision flows I think this is another important data point for valuing stocks that include Amazon, Adobe, Salesforce, and Palo Alto Networks. And for assessing how fast AI companies will build the big revenue streams they need to justify current valuations.

In addition the OpenAI/Amazon case is a good introduction to the concept of monopsony, introduced into mainstream economics by Joan Robinson in her 1933 book, The Economics of Imperfect Competition. Monopsony, like its close relative monopoly, explains, in Robinson’s framework, why markets aren’t perfectly competitive and why some companies, like Amazon, are able to reap excess returns in their businesses. Robinson’s focus in 1933 was on the labor markets. There was this little thing called the Great Depression going on and Robinson, like many economists, wanted to explain how so many people could be without work and why demand had suddenly collapsed. Her research looked at the question of how the power of a big employer could change the dynamics of a nominally freely competitive job market.

Robinson never imagined a market-dominating beast like Amazon. But I think her concept of monopsony fits the current e-commerce economy very well.

Here’s what happened that led OpenAI to change course. OpenAI launched Instant Checkout in September 2025 with Etsy, then expanded it to include Shopify, Walmart, Target, and others. The idea was that users could complete purchases without leaving ChatGPT.

But according to The Information and Digital Commerce 360, users were mostly browsing without buying — treating ChatGPT as a research tool rather than a checkout terminal. And merchant customers decided that they weren’t comfortable with handing over so much control of their business to an AI agent. Shopify president Harley Finkelstein, for example, flagged concerns at a Morgan Stanley conference on March 3 about “the integrity of transaction activities” in agentic commerce — essentially, merchants worried about losing control of the customer relationship. And just as bad, the upside to a ChatGPT agent just didn’t look all that enticing. The in-chat buy buttons generated very low conversion rates from potential customers.

So OpenAI decided to refocus on what was actually working: product discovery and recommendations. The AI shopping agent isn’t dead—but it is being restructured from a direct checkout model to more of a discovery-and-handoff model.​ Instant Checkout is moving into merchant apps (like the Target and PC Express grocery apps that run inside ChatGPT), rather than appearing inline in chat responses.​ And the underlying Agentic Commerce Protocol (co-developed with Stripe) will live on as infrastructure, but it’ll power transactions through those apps rather than direct in-chat product listings.​ OpenAI said it’s “prioritizing making ChatGPT search and product discovery great” while letting merchants handle checkout in their own embedded apps.​

Two important lessons here, I think.

First, for AI as a whole, inserting an AI tool into an existing framework—whether it’s e-commerce or HR or network security—is much more complex than “drop-in and run.” More on what this means for the software companies and their stocks, which everybody now hates, on my next Applied Economics post coming this weekend.

Second, for Amazon in particular, it means that its very profitable monopsony will continue unchallenged for at least a little while longer.

Read more

The moves were totally reasonable. Investors and traders had feared that the war with Iran would spiral into a more destructive and wider regional war. And recent attacks on oil infrastructure in the Middle East and on water desalination plants on both sides of the Persian Gulf provided ample evidence that those fears were justified.

And then President Donald Trump gave a phone interview with CBS News in which he said that the war with Iran was progressing ahead of schedule. Trump told CBS News that the Iran war is “very complete, pretty much.” The President also reiterated that the conflict is “very far” ahead of his four-to-five-week timeline, adding that he was considering taking over the Strait of Hormuz.

And markets reversed.

Read more

Why is this exercise important?

Because it’s a real life example of the work on the role of monopolies in our economy by economists like Joan Robinson and Paul Sweezy. Their work begins with the extreme excess returns that companies with effective monopoly power generate--and points to the important role that monopolies play in the business cycle of boom and bust.

And because monopoly economics are critical to deciding if the current generation of AI stocks are really going to be worth what investors now say they are.

Importantly for the current moment, monopoly economics outlines why AI companies are willing to project capital spending budgets, most of it for data centers, compute, and networking, of $650 billion to $700 billion for 2026. If one of these AI companies can become an effective AI monopoly, then all this capital spending makes sense--for the victor. Citadel Securities puts AI capex at around 2% of global GDP in 2026.

Read more

Every week I’ll return with three more questions from readers of Economics for Investors about economics and the economy. I’ll put the first question and my answer in front of the paywall so everyone—free and paid subscribers alike can read it. The other two questions and my answers are for paid subscribers only.

Only paid subscribers can submit a question. Any questions a paid subscriber submits over the next week will be fodder for my next regular Friday Q&A post on Economics for Investors. So ask away.

Now with no further introduction, let me go to your questions.

This week’s three questions and answers are about forecasts of productivity gains from AI; the politics and economics of renewable energy, and why the March 18 Federal Reserve meeting is super important?

Question #1: Are projections of productivity gains and job losses from AI complete bullshit?--from Charles W.

Well, maybe not complete bullshit but what I’’d call the Klarna example certainly suggests that the numbers aren’t all they seem. And that we need to skeptically re-examine the very popular general story that assumes that huge job cuts from AI are just around the corner.

What is the Klarna example?

Klarna is a fintech company specializing in the buy-now/pay later market. Klarna lets people split purchases into installments and offers a broader “everyday money” app for shopping, payments, and personal finance. The company’s “buy now, pay later” service might, for example, allow a consumer to pay in 4 interest‑free installments, or to pay in 30 days, or to finance a purchase over several months, depending on country and merchant. Klarna also provides a shopping app and card so you can pay with Klarna online or in‑store, track orders, manage installments, and sometimes earn cashback or rewards.

As you might imagine, this business requires collecting and processing a ton or data, generating a decision on credit and credit terms, and then answering questions from consumers on issues like changing a payment date (routine) or an inability to pay on tie or ever (difficult and one-off in any cases.)

For Klarna getting more efficiency at handling data and serving customers would be a big deal—and for the company’s bottomline..

Starting in 2024, Klarna began claiming cost savings from AI. CEO Sebastian Siemiatkowski said its AI assistant was doing the work of roughly 700 customer‑service agents. A figure later updated to over 850. The result, the company claimed, was tens of millions of dollars in cost savings and big improvements in response times and resolution rates.

Klarna paused most hiring, cut large numbers of staff and outsourced roles, and framed the AI rollout as a major productivity and profitability win ahead of its IPO.

But…

By 2025–2026, reports began to describe big drops in customer satisfaction and in the quality of service. The drops were significant enough to prompt investor worry about damage to the brand..

That wasn’t exactly what a company that had only recent gone public wanted to tell investors. Klarna had gone public on 10 September 2025 at an IPO price of $40 per share. On its NYSE debut day it (ticker KLAR), closed at $45.8 dollars, up about 15 %. But by early 2026, KLAR had fallen to the low‑ to mid‑teens on continued losses and the danger that declining service quality posed to the long-term value of the brand.

So , Klarna started restoring human support: rehiring agents, piloting an “Uber‑style” flexible remote workforce, and moving internal staff into customer‑service roles. CEO Siemiatkowski publicly acknowledged the company had focused too much on efficiency and had to ensure customers could always reach a human.

I think it’’s worth diving a little deeper into the exact nature of the problems Klarna encountered with AI. The story seems to be that the company’s AI agents did a good job at maximizing efficiency where metrics like how long it took to end a customer query could be applied. The AI had been told what metrics to look at to measure its goals, and it did indeed maximize progress toward those measurable goals.

But where the goals were harder to quantify and measure—and where the company had done poor job of articulating those goals for the AI, and devising metrics to judge performance toward those goals—,the AI, not surprisingly had done a poor job. How was the poor AI supposed to improve customer satisfaction when the poor thing hadn’t even been trained on how to define the goal.

Reports note that the AI assistant could manage routine FAQs but struggled with nuanced tasks like disputes, refunds, and financial‑stress situations that require judgment, discretion, and emotional sensitivity. Customers described replies as generic, repetitive, or unhelpful, exposing “empathetic gaps” that damaged trust—especially problematic in a product dealing with debt and missed payments.

Klarna heavily promoted that its AI did the work of 700+ agents and handled millions of conversations in dozens of languages, creating expectations that it was a near‑drop‑in replacement for humans. When users then hit obvious limitations, the gap between the AI story and their experience amplified frustration and made the brand look like it had over‑promised and under‑delivered. There was visible public backlash about people feeling ignored or unable to reach a human.

The take-away story for AI and jobs is that replacing human beings is tougher than it looks—especially in a service company. In important ways, the AI revolution isn’t equivalent to replacing textile workers with spinning jennies and power looms in the nineteenth century. Many current jobs are very nuanced and very difficult to teach to a machine.

Bottom line, the AI industrial revolution is going to take longer than current stock market enthusiasm believes. And there’s going to be a huge role for intermediate software companies—exactly the Salesforces and ServiceNows that have been savaged recently—that can build systems to interface between AIs and the real world.

I think it’s going to take a while for the financial markets to grasp that.

Look at the recent market reaction to the news that Block (XYZ), is cutting more than 4,000 jobs, reducing headcount from a bit over 10,000 to just under 6,000—about a 40% workforce reduction. CEO Jack Dorsey, said “intelligence tools” and AI have “fundamentally changed” how you build and run a company, and that much smaller teams using these tools can do more, better. He framed it as a proactive, one‑time restructuring rather than a gradual series of cuts, and predicted that most companies will reach a similar conclusion and follow with AI‑driven restructurings over the next year.

Block explicitly presented this as an AI‑enabled efficiency move, not a distress layoff,. And the stock market treated the news as one of the first big, high‑profile “AI‑driven” workforce reductions in the S&P 500. After the layoff announcement and earnings, Block’s shares jumped about 24% in after‑hours trading.

Allow me to be cynical. From January until just before the AI layoff announcement in late February 2026, Block traded down hard, bottomed in early February, and was still sitting on a mid‑teens year-to-date loss before the post‑announcement spike. Investors had sold on worries about slowing growth, increasing competition, and margin pressure. Block had underperformed both the broader market and its fintech peers. (Block, formerly Square, is a financial‑technology and digital payments company that builds payment and financial services platforms: Square for in‑person and online commerce and Cash App for peer‑to‑peer payments, spending, saving, and investing. It also owns Afterpay (buy‑now‑pay‑later), TIDAL (music streaming), Bitkey (Bitcoin self‑custody wallet), and Proto (Bitcoin mining hardware and services.)

You don’t think Block could possibily be using AI window-dressing to make what is a huge downsizing look more positive than it is?

I expect to see more of these “AI revolutionizes work NOW” stories. At least as long as the stock market rewards them.

Read more

Read more

Using that same framework, it’s clear why silver, as spectacular as a run as it had in 2025—up nearly 160% on the year—and in the early days of 2026—up another 20% through January 20—isn’t as good a future bet for investors and traders as gold.

The big difference between gold and silver going forward is the role of the world’s central banks.

As I pointed out in that earlier post, global central banks are accumulating gold at a far faster pace than five years ago. Their buying is a major factor in increasing global demand for gold.

Silver, on the other hand, is not an official asset on the books of any of the world’s central banks, Modern reserve reporting (IMF and central‑bank balance sheets) is built around gold and foreign‑exchange assets, and silver is not recognized as an official reserve asset in that frameworks.​ There are persistent rumors that the central banks of Russia and India are buying silver, but at the moment these are just rumors. And the fact that international financial standards don’t recognize silver as a balance sheet asset puts a damper on central bank buying.

Which doesn’t mean that one central bank—through its power over all of China’s financial markets—isn’t playing a huge—and negative role—in the price of silver.

Let me construct a quick time line of recent silver prices.

The closing silver price on February 23, 2026 was about $87.70 per ounce in the spot market.

But not so long ago—January 29, 2026 to be precise—silver hit an all‑time high of $121.62 an ounce.

What happened to produce that crash?

Most explanations that I’ve read point to the “inevitable” unwinding of the huge speculative spike—a gain of 130% in 2025 into early 2026—on an upside surprise in the jobs report, a stronger than expected CPI inflation report, and the news that President Trump would nominate Kevin Warsh to head the Federal Reserve (instead of a candidate seen as more committed to lowering interest rates faster.) All these developments, the explanation goes, forced a re-evaluation of assumptions about how fast the Fed would cut interest rates. And a repricing of assets, like silver, that were see as inflation hedges. As silver prices broke key technical levels, stop‑loss orders and margin calls kicked in, creating cascading liquidations and “mechanical” selling that accounted for the bulk of the move, with one analysis estimating ~79% of the decline came from such mechanical factors rather than new information.

All that is true, but it leaves out one very important move by the government in Beijing with a very clear chronological connection to the timing of the collapse in silver.

The Shanghai Futures Exchange (SHFE) twice tightened trading rules on precious‑metals contracts in January 2026, raising margin requirements, widening daily price limits and adjusting fees for silver (and gold/copper/aluminum) “to rein in risk amid a sharp rally.”​From the January 22, 2026 settlement, SHFE raised margin requirements and widened daily price limits for silver (along with gold, copper, and aluminum. ) For silver contracts, the daily price fluctuation limit was lifted to 15%, with hedging margin at 16% and speculative/general positions at 17%, increasing the capital required to hold positions.​

A further tightening step came via a notice on January 26, 2026, effective from the night session of January 27, adjusting trading limits for silver (and tin) futures for certain client categories, including non‑member and overseas participants.

It’s not a coincidence, I believe, that this crack down on what Chinese authorities saw as excessive speculation took place at just the time when silver was peaking. The move in Shanghai took major money out of the silver trade.

The Chinese didn’t move on a whim—and the reasons for the tightening are important for the future price of silver.

All indications are that the Chinese government had become alarmed at what the speculative boom in silver was doing to costs in the Chinese solar industry—a key sector in the government’s planning for the growth in China’s economy.

Modern crystalline‑silicon solar cells typically use on the order of 75–135 milligrams of silver per cell, depending on the technology. As a rule of thumb, a standard panel with ~60–72 cells, at ~80–100 mg per cell, will contain roughly 5–10 grams of silver in total. Older cell technologies use move silver with about 20 grams per panel as a legacy average.

Historically silver has been a small slice of panel cost, but with the recent price spike it has become one of the largest cost components. Bloomberg data cited in January 2026 say as a result of the rapid increase in prices silver then accounted for about 29% of total solar panel production cost, up from 3.4% in 2023 and 14–15% in 2024–2025. Silver had overtaken poly silicon as the single largest input cost.

That wasn’t an acceptable development and the trend was deeply worrying to Chinese officials. So the government put the brakes on speculative trading that had pushed silver prices to historic highs and that looked likely to produce even higher prices in the immediate future.

The question now, for investors and traders in silver is whether China has now created an effective price cap on future silver prices.

Will China intervene again if silver approaches the $122 intraday high on January 29? Will it move even sooner the next time to stop speculation short of that level?

I believe the answer is very likely yess. “Thrifting,” the process of finding ways to reduce silver use in solar cells, will certainly contribute in the long-run to keeping the market price of solar cells to a level that doesn’t endanger China’s solar exports—or its dominant global market share. (China currently produces around 90% of the world’s solar cells, and controls more than 80% of manufacturing capacity across every major stage of the solar supply chain. A 2024 Statista breakdown for 2023 puts China’s share of global solar PV cell production at “almost 92%”, with Malaysia and Vietnam a very distant second and third.​)

It is notoriously difficult for central banking and market authorities to control speculative cash flows. So I don’t expect that the January tightening will be the last time Chinese authorities will have to intervene in the silver market.

But if I were a silver investor or trader right now I’d be treating $120 an ounce as a major risk target. At $85 an ounce now, silver seems a good speculative buy. If or when we approach $120 again, I’d think about taking profits and waiting for the next news from Shanghai.

As a subscriber, you get exclusive access to my 5 Economic Indicators Cheat Sheet -

a one-page guide to the key data every investor should track.

Inside you’ll find:

•…

Read more

Unlike the world’s supply of fiat currencies where supply will soar as the world’s indebted countries--just about everyone--print more money to pay their bills, the global supply of gold is increasing by just 1% a year or less.

Unlike demand for the U.S. dollar, the world’s fiat money flagship currency--where every day more investors want to hedge their risks by diversifying out of the dollar--demand for gold from global central banks is climbing and Wall Street strategists have modestly but noticeably increased their recommended allocation to gold for individual portfolios.

The problem with fiat currencies--you know the paper money you have in your wallet whether it’s dollars, or euros, or yen--is that this money has value only because governments say it does and because we are willing to accept it at something like its stated value. As such it depends on trust in the issuing state and its implicit promise to take reasonable care of the value of that money. (Fiat currencies aren’t backed--although they once were--kinda--by a physical commodity like gold or silver)

Currently a whole lot of people don’t trust the governments that issue these fiat currencies to take reasonable care of the value of their paper promises. They see run away discretionary spending, huge looming spending for non-discretionary items like retirement, health care, and interest payments. They increasingly see governments putting pressure on central banks to abandon rules of sound monetary management. And they see no credible plan for reducing spending or raising more revenue in order to reduce government debt.

They do fear that there is indeed a plan to address the problem of soaring debt. That plan, they suspect, relies on printing more money to cover the coming bills--and using the resulting increase in inflation to gradually “solve” the debt problem by making those fiat currencies worth less.

If you have that world view, you can understand the increasing demand for commodity currencies such as gold.

We can get a good read on the increase in demand for gold by looking at the numbers reported by central banks. Central bank gold holdings have risen strongly over the last five years, with cumulative net purchases well over 4,000 metric tons. From 2021 through 2025, central banks have bought roughly 800–1,100 metric tons per year, versus an average of about 400–500 metric tons in the preceding decade. International Monetary Fund and World Gold Council data indicate that official sector gold holdings (central banks plus IMF and some other official institutions) are now around 40,000 metric tons, roughly 20–25% of reported above‑ground gold stocks.

Photo by Jingming Pan on Unsplash

For the first time in decades, gold has overtaken U.S. Treasuries as the single largest reserve asset class by value for the official sector.

The consensus is that central banks will keep buying substantial amounts of gold in 2026 and for several years after, but at or slightly below the record pace of 2022–2024. For example, J.P. Morgan projects around 755–800 metric tons of central‑bank gold purchases in 2026, down from 1,000 metric tons in each of 2022–2024 but still well above the pre‑2022 norm of 400–500 metric tons. Goldman Sachs characterizes the post‑2022 surge as a multi‑year structural shift in reserve management (de‑dollarization, sanction risk, diversification), and assumes official‑sector accumulation continues for at least another three years at elevated levels (its base case is ~960 metric tons a year.) The World Gold Council’s 2025 Central Bank Gold Reserves Survey finds 95% of reserve managers expect global official gold holdings to rise over the next 12 months, and none expect a decline, with 43% planning to increase their own gold reserves.

That’s the demand side story in favor of gold.

But remember your basic economics--it’s supply and demand, right?

The supply side is equally important. And here gold has a decided competitive advantage over fiat currencies. Governments can increase the supply of any fiat currency by rolling the printing presses (actually having central banks buy government bonds with money created by increasing the size of their balance sheets) until financial markets revolt and investors refuse to buy more paper.

To increase the supply of gold, however, gold miners have to dig the yellow metal out of the ground. And here the projections are showing very modest--if any--increases in gold production and supply.

Global mined gold supply grew in 2021 by about 1–2% from 2020.

Global production in 2022 was up roughly 0%.

Global production in 2023 was up 2–3% from 2022.

Global production in 2024 was up about 1–2% from 2023.

And global production in 2025 was up about 0–1% from 2024.

Why? Analysts highlight structural problems facing mining companies: declining average reserve grades, long project lead times, and the need to replace declining production from mature assets.

For those reasons, the projections show that mined gold supply is likely to plateau over the next couple of years rather than surge, as the industry nears a production peak after years of only mild growth.

Here’s my take on all this: the long-term supply and demand trends--including a seemingly inexhaustible supply of bad policy decisions from global governments--are in favor of adding gold to your portfolio. Once upon time, Wall Street strategists recommended having 1% to 3% of a portfolio in gold. That recommendation is now closer to 3% to 5%. And we’re on our way to 10%. Increased portfolio allocations to gold by individual investors and central banks mean more gold demand when gold supply is growing very slowly. That adds up to higher gold prices.

That long-term trend won’t come without frightening short-term volatility. Even just in 2026 COMEX gold futures fell about 11% in a single day on January 30, 2026, described as the biggest one‑day loss in decades. On February 3, 2026, gold futures posted a ~6–7% daily gain. In early February gold had another nearly 10% one‑day drop. Which came just after a January 27–28, 2026 move characterized as gold’s biggest daily dollar gain ever.

As much as you can, ignore these swings. Stick to your plan for adding gold to your portfolio. Buy, to follow that plan, on the dips. Trade, if you have the experience, to take profits near the highs. And then rebuy.

This isn’t the easiest market for buying and holding gold. But that is the strategy you want to pursue now.

In my portfolios on my Internet sites JubakPicks.com (free and you can follow the portfolios if you register with your email) and JubakAMcom (subscription) you’ll find gold stocks, gold mining ETFs, and gold metal ETFs. I’ll be looking to find more gold plays to recommend in the coming weeks. Stay tuned.

But in my inbox, and over coffee with friends, the questions are different. They are deeper.

“If the Fed raises rates, why is my mortgage actually getting cheaper?”

“Is the U.S. Dollar actually losing its global status, or is that just a headline?”

“How does a demographic shift in Japan affect my tech portfolio in California?”

These aren’t stock market questions. They are Economic questions.

And I think there’s even a deeper set of questions that you need to be asking right now. And that I’d like to help answer for you.

You see as much as we want to pretend that economics is a neutral, unbiased science, the truth is that the information that it produces and that Wall Street uses in the strategies that drive financial markets is deeply influenced by assumptions and theories and readings on economic history that can be deeply misleading in the long run. Or even in the short run.

And that the success of your portfolio, the gains and the losses, depends on separating the questionable assumptions from what we really know about the economy.

Think of it this way. The stock market is the dashboard of a car. Economics is the engine. Most investors spend their whole lives staring at the speedometer. They cheer at that speedometer when asset prices are climbing. When things go sour, they yell at the instrument panel or even pound it.

When what they ought to be doing is finding better diagnostics to tell them what is going on under the hood. And maybe even pulling off to the side of the road, popping the hood, and taking a deep, in-person look at the engine.

I’m starting this Substack to open the hood. And to tell you how to find better diagnostics and to gain a better understanding of how economics and the economy work.

Economics for Investors is not an academic textbook. It is a field guide. I am going to take the complex, jargon-heavy “dismal science” and turn it into a tool that you can use to gain a competitive advantage in your investing. I will answer the questions that don’t fit into a daily market wrap, but are essential for long-term wealth.

If you’ve ever felt like the “experts” are speaking a different language, this is for you. We’re going to learn how to parse that language together.

So what can you expect here every week?

Every Tuesday, I’ll send you a Deep Dive post into how a core issue in economics is driving investment practice. Or at least how it should be, if you want to make the most money and avoid big losses.

Every Wednesday, I’ll send you an Applied Economics post that applies-–hence the name—an issue from economics to a decision you face in your portfolio about what to buy or sell.

Every Friday, I’ll post my answers to three questions I’ve received in recent days about economics and the economy. This is your chance to vent on things that are bugging you–or that just confuse you–about the current state of economics and our economy.

The occasional Deep Dive post and the first of my three answers to your questions will be available—and sent to email boxes-–of free and paid subscribers alike. So it’s worth becoming a free subscriber. Most Deep Dive posts, all Applied Economics posts, and the remaining last two answers in my Q&A will go out just to subscribers.

Hope you’ll come along on this adventure. I think it will be fun, educational, and profitable.

Jim Jubak

Yet another attempt to fix Kroger (KR). But this K-shaped economy may have left the 2nd largest grocer in the United State without a turnaround strategy.

Which leads to the musical question: What other retailers are caught in the same position?

On Monday, after a roughly year-long search, grocery giant Kroger announced its new top boss: veteran Walmart (WMT) executive Greg Koran, an alumnus of the only grocer in America larger than Kroger by sales.

But I have to wonder if anyone can right Kroger given the storms--from every direction--that Kroger faces.

Kroger sees competition from everywhere. Downmarket, the company is facing heat from discount grocers such as Aldi. Upmarket, it’s losing customers to quality-centric players such as Whole Foods and Sprouts.

A recent report from data analytics company Placer.AI found that foot traffic growth for “non-differentiated” grocery stores, such as those that make up the bulk of Kroger’s empire, lagged compared to the entire industry last year. For instance, visits to non-differentiated stores increased just 3.6% year-over-year in the fourth quarter of 2025, while visits to “quality-first” chains jumped 9.3% and visits to “savings-first” chains jumped 7.1%.

So-called “unicorn” chains--or those that “defy grocery’s traditional quality-price tradeoff” via extreme focus, (such as Trader Joe’s)--were up 9.2%, while the entire industry saw an increase of 5.3%. The trends were observed across all four quarters last year; in the first quarter, non-differentiated stores saw foot traffic decrease by more than 1%.

What’s more, Placer.AI found that the customer base for both up- and downstream competitors is eating into the middle:

The median household income of savings-focused chain customers has increased in the past three years, while the median household income of quality-focused chain customers has decreased.

In short, the middle class is splitting in two--trading up for quality or down for value--leaving traditional grocers like Kroger stranded in the ‘no-man’s-land’ of the K-shaped economy: “As price sensitivity rises and perceived quality differences narrow, the retailers winning today are those with the clearest answers to a simple question: Why shop here instead of anywhere else?” Placer.AI analyst Erich Kahner wrote.

Koran has his work cut out for him. Kroger has been focused on reducing costs for customers the past couple of years in a bid to compete with the Aldis of the world. The company has slashed roughly 1,000 corporate jobs in the past year, while closing underperforming locations, scaling up its (often-cheaper) store-brand labels, and, yes, adding more protein offerings.

What other retailers face a similar K-Shaped squeeze? Nike (NKE) comes to mind. Or names like J.Crew, Banana Republic, Coach, New Balance, Chipotle Mexican Grill, Michael Kors, and Tory Burch.

Any questions you submit over the next week will be fodder for my next regular Friday Q&A post on Economics for Investors. So ask away.

Now with no further introduction, let me go to your questions.

This week’s three questions and answers are about the neutral interest rate (and Fed interest rate policy), global economic volatility, and tariffs and productivity.

Question #1: Why does the Trump administration care so much about what economists call the neutral interest rate?--from Roxy197.

I love wonky questions and this is about as wonky as it gets. The neutral (or natural) real interest rate, r*, is the inflation‑adjusted interest rate consistent with the economy growing at full employment but where inflation remains within its target range. In other words, it’s the real interest rate that allows for maximum economic growth without accelerating inflation above the target rate. In nominal terms, a neutral fed funds rate is roughly “inflation plus r*,” so knowing r* helps judge if today’s policy rate is tight or easy.

That’s the easy part.

Deciding what the neutral rate is at any moment can be extremely contentious. Not surprising, actually, since the neural interest rate is, supposedly, the starting point for Federal Reserve interest rate policy.

If you want to juice or repress the economy, you need to set interest rates either below or above the neutral rate in order to get your desired economic result.

Right now the Federal Reserve claims that its Fed funds benchmark interest rate of 3.50% to 3.75% is neutral--that is it is neither low enough to stimulate the economy and increase inflation nor high enough to suppress economic growth and increase unemployment.

Economists in the Trump Administration argue that the Fed has got it wrong. The neutral interest rate is actually much lower than the Fed’s nominal rate at 3,5% to 3,75%. Hence, current Fed interest rate policy is needlessly restrictive and is suppressing growth.

In 2023, economists at the Brookings Institute summarized the median estimate by members of the Fed’s Open Market Committee of the long‑run real neutral rate at about 0.5%, with a long‑run nominal funds rate around 2.5–3%. So Fed policy was mildly restrictive.

But a 2025 Cleveland Fed model estimates the real r* for the U.S. is now around 1.5%, implying a medium‑run nominal neutral rate near 3.7% given their inflation assumptions. So Fed policy is effectively neutral.

Part of the debate is about whether or not r has permanently increased. With more increases on the way.

Higher public debt, re‑militarization, de‑globalization, AI‑driven investment, and stronger productivity could push r* above 1% and possibly higher if several forces coincide.

The other side of the debate argues that the aging of the global population, slower underlying growth, and global savings patterns that kept r* low pre‑COVID are likely to limit how far it can rise.

If true neutral is higher than many think (say nominal 3.5–4%), then current policy isn’t restrictive and interest rate policy would result in higher interest rates and higher long‑run Treasury yields.

And smaller gains in the stock market.

In practice, the debate is roughly: “Is the U.S. neutral nominal funds rate still near 3%, or has the post‑pandemic shift pushed it closer to the 3½–4% area—and for how long?”

My own conclusion is that the cited factors have pushed the nominal neutral rate higher and the world is looking at a move to higher rates and yields rather than a return to the lower neutral rate of the pre-Covid years.

Listen for the answer to this question from Kevin Warsh during hearings on his nomination to be the next chair of the Federal Reserve. I expect him to endorse the “neutral rate is lower” view. But it will be interesting to hear how directly he says this.

Question #2: In your 2014 letter explaining the decision to close your global mutual fund, you wrote that central bank cash flows and the resulting risk-on/risk-off dynamics had largely overwhelmed the fundamentals of individual companies, such that what mattered most was market exposure and timing rather than the specifics of what was owned. You also noted that while you believed market drivers would eventually change, you could not honestly say when. From today’s vantage point, do you think fundamentals have become more important again as drivers of returns, or do you still see structural and macro forces that continue to limit the effectiveness of traditional fundamental stock picking?--from Chance.

It would be hard for me to convincingly argue for a return to the importance of fundamentals in the financial markets after the Great Recession that followed on the sub-prime mortgage crisis or the Magnificent 7 market that has culminated, so far, in a company like OpenAI being awarded a market capitalization of $500 billion while still private and before the company has a convincing strategy for reaching profitability.

For me right now, though, the most important question is the future direction and force of global cash flows. In my opinion, we’re watching the end of the global loose money era as a temporary surplus of savings turns into a temporary scarcity. My reasons to think this include the end of China’s demographic bonus years from a young population, a rising tide of government debt around the world, the costs of aging populations, and the costs of global climate change. To me that all adds up to more expensive money in coming decades.

At the same time, I think we’re witnessing the end of a period of over-performance by U.S. assets. In 2025 and 2026 the world began slowly rebalancing its portfolios to move away from dollar assets. In 2026 so far, the U.S. stock market, even as it moved higher, has NOT been the best performing in the world.

One of the things I’ve learned in watching and participating in the equity markets for, my God, almost 40 years now is how tricky it is to get the call right on changes in the direction of long term trends. Those periods of trend change, like the current period, are characterized by increased volatility. And changes in the direction of long-term trends tend to take much longer to take effect than investors expect.

In short this isn’t an easy period for investors. And it’s not one likely to comfortably and consistently reward the application of the tools of fundamental analysis. One of the biggest challenges, I’d say, is likely to be sticking to any discipline of fundamental valuation when big, if potentially temporary, rewards go to investors and speculators who follow the popular trend of the moment.

Question #3: According to the historical record, do higher tariffs help or hurt the economy in the long run?--from BrentB

The research argues that higher tariffs--like those imposed by the Trump administration--reduce productivity growth. Which would be a shame since forecasts before the tariffs had been pointing to strong productivity growth in the last half of the 2020s.

Tariffs are likely to dampen U.S. productivity growth at the margin, mainly by raising input costs, increasing uncertainty, and allowing less‑efficient firms to survive.

Cross‑country economic research using five decades of data finds that higher import tariffs are followed by lower output and lower productivity over the medium term, alongside higher unemployment and inequality.

Historical U.S. evidence (from manufacturing during the nineteenth century U.S. Gilded Age) shows higher tariffs were associated with reduced labor productivity and lower output per establishment, consistent with more small, less‑productive firms entering markets under the umbrella of tariff protection.

Firm‑ and industry‑level studies (e.g., of South Africa, and US/EU manufacturing) document that tariffs worsen resource misallocation by keeping inefficient firms alive and shrinking the market share of more productive ones, lowering aggregate productivity by several percent when protection rises meaningfully.

These effects don’t show up overnight, but with tariffs already in place, tariffs could start weighing on productivity growth over the next year or two.

Which would be unfortunate timing since global economies need higher, non-inflationary growth to pay down current high levels of debt.

Before factoring in the effects of tariffs, a Richmond Fed study forecast trend U.S. labor productivity, after drifting down to about 1.4% in 2021, could be expected to rise gradually to about 2.0% by 2030, with the average for the 2020s around 1.7% to 2.0%. 2026, according to this forecast, would have been part of a decade in which underlying productivity runs noticeably stronger than the 1% to 1.25% norm of the 2010s, if not at the 3% plus rates of the late‑1990s tech boom.

Several banks and international forecasters tie their above‑consensus 2026 GDP calls (around 2½–3% real growth) to “above‑trend productivity growth” supported by AI‑linked investment.

Recent data show nonfarm business productivity growing close to 3% over the latest four quarters and nearly 5% annualized in the third quarter of 2025, which most officials view as cyclical and unlikely to persist.

If the current tariff regime is maintained or tightened, the balance of empirical evidence suggests it will trim U.S. productivity growth rather than boost it in 2026, partially offsetting gains from technology and reorganization and making the recent 3%‑type productivity prints harder to sustain.

Those are your three questions and my three answers for this week. Submit your questions for next week’s Q&A beginning next Tuesday on my Substack Economics for Investors. Any questions you submit over the next week using any vehicle will be fodder for my next regular Friday Q&A post on Economics for Investors.

But exactly how?

I think looking back to Adam Smith who wrote The Wealth of Nations (published in 1776) during the first Industrial Revolution in England can tell us a lot about what questions we should be asking--and maybe even give us an answer or two.

The following long piece is an example of the Big Picture posts--along with my weekly Micro Economics post and my weekly economics Q&A (yep, 3 posts a week)--that I will be putting up every Tuesday on my new Substack Economics for Investors. That Substack will be going live next week. And content like this will appear only on that Substack for subscribers to Economics for Investors. Look for the official announcement here SOON.

The big question, of course is

Will AI result in such massive productivity gains that work will become optional? (Yes. Elon Musk has said that advancing AI and robotics will make work effectively optional in 10 or 20 years. And, he added, money will become irrelevant.) AI had better produce big productivity gains in order to justify the trillions of dollars going into capital spending in the sector. And to justify those stock prices for META Platforms (META) and Nvidia (NVDA.)

Or will it result in an army of unemployed or underemployed workers as it generates massive layoffs? This is the fear that is producing such high levels of anxiety in so many American families. Even though surveys like the annual International AI Safety report say there is no evidence of significant job losses due to AI--YET--the news from corporate America just feeds those worries every day.

For example, in late January, Amazon confirmed cutting 16,000 corporate job cuts. That follows on cuts of 14,000 jobs in October. In its public memos Amazon describes the cuts as part of making Amazon “leaner,” reducing layers and bureaucracy, and reallocating resources into its biggest bets. CEO Andy Jassy and other company officials have repeatedly said that as Amazon embeds generative AI and automation, they “expect this will reduce our total corporate workforce as we get efficiency gains,” and that they will need fewer people in some existing roles and more in new AI‑related roles.

You can see how this might make some folks conclude that AI leads to job cuts.

The most famous “story” in Smith’s Wealth of Nations is about a revolution in productivity in a pin factory due to the introduction of specialization and the division of labor.

Smith describes a small pin factory where making a simple pin had been broken into about 18 distinct operations: drawing out the wire, straightening it, cutting it, pointing it, forming the head (itself several steps), attaching the head, whitening the pins, and finally putting them into paper for sale. If each worker had to make each pin by going through all those steps himself, Smith claims, each could perhaps make only 10–20 pins a day​.

But if workers divided the labor and specialized in just one or two parts of the process, this factory with 10 workers could produce around 48,000 pins per day or roughly 4,800 per worker, Smith claimed.

(The decision of the most famous classical economist in his most famous work to choose pins for such a key example of such a key explanatory principle isn’t as strange as it seems today. Pins were a big economic deal in 15th and 16th and later times. And a hot political issue. Pin imports into England in Tudor times were on the order of £40,000–£60,000 per year--a sum that is equal to £12 to £16 million in today’s pounds at a time when England’s entire GDP was something like £40 billion, a far cry from today’s £2.5–3 trillion. English monarchs restricted or banned pin imports mainly to protect and promote domestic pin‑making. In 1483, Richard III issued a statute prohibiting the import of foreign pins to support the London and English “pinners.” Elizabeth I later explicitly banned pin imports in an effort both to protect domestic manufacture and to ensure that pins met certain quality criteria. By Smith’s time, though, English pin makers had pretty much conquered the domestic market.)

The economic growth that Smith was seeing--and I’ll return in a minute to questions later economists have raised about Smith’s explanation for this growth--was explosively different than what England had seen. After a period of essentially stagnant economic activity--per capita GDP growth was close to zero in the first half of the 17th century--real per capita GDP growth jumps to about 0.8–0.9% per year in the second half of the century. With population growing roughly 0.3–0.4% per year, aggregate real GDP is rising around 1.1–1.3%. Not much by current standards of 2% to 3%, but radically different from 0%.

And that growth rate continued in Smith’s century. Aggregate real GDP grows around 0.8–1.0% per year on average before really taking off in the 19th century. From 1822 onward, real GDP per person steps up to about 1.0–1.1% per year, peaking near 1.9–2.0% per year in mid‑century. With population growing roughly 1.1–1.3% per year, aggregate real GDP growth is in the neighborhood of 2–3% per year, highest in the 1830s–1850s.

But what is shocking to me--and exceedingly important to answering the question of what AI will do to the job market--is the collapse of wages in many of the sub-industries most directly impacted by the industrial revolution Smith describes.

For many workers in the British textile sector, especially handloom weavers, real wages fell dramatically in the early 19th century, even as output and productivity surged.

By the late 1820s–1830s, many handloom weavers who had once earned around £1.5–£2 per week in the 1790s–1805 were down to 5–8 shillings per week. That’s roughly a 70–80% fall in nominal weekly earnings, with real income erosion compounded by prices and long hours.</strong>

And wages remained low even as productivity rose and economic growth accelerated. Factory spinners and weavers labor productivity and output per head rose sharply from roughly the 1810s to the mid‑1830s, while average real wages rose only slowly, so wage growth significantly lagged productivity.

This pattern even has a name. Economists call it the Engels pause after Karl Marx’s collaborator Friedrich Engels.

In a sensitive and detailed work of reporting, The Condition of the Working Class in England, Engels documented the poverty, despair, and exploitation of workers and their children created by the first phase of the Industrial Revolution. Often dated to about 1800–1850, it was a period when most gains from mechanization, factories, and capital deepening went to profits and owners, not to workers’ incomes.

Only later in the 19th century did real wages for broad segments of the working class begin to rise significantly, “ending” the pause as growth became more widely shared.

Modern economic historians use the “Engels pause” as a shorthand for a pattern of a decoupling between productivity growth and real wage growth during early industrialization in an economy as a whole or in an economic sector. Within an economy or sector during an Engels pause in an economy or a sector wage structures become polarized: a relatively small number of skilled spinners and weavers earned decent wages while a large pool of low‑paid piecers, children, and unskilled workers barely made enough to feed themselves. And often not that.

I have dragged you through all this history because I think the transition to an AI economy--whatever that will eventually look like--has all the characteristics of an Engels pause. (And if you have a sense of historical irony you may get a tickle from the thought that such Masters of Capital as Altman, Musk, and Zuckerberg have ushered in an economic stage--again--named after one of the founders of Marxism.)

What do I think that means for you as an investor--and as a citizen of this economy?

• First, we’re looking at a long period before the AI economy sorts itself out.

The original of the Engels pause lasted from 1800 to 1850. I think we’re in for a long period of economic turmoil.

• Second, we are looking at stagnant or falling wages for many--but not all–workers during this period.

• Third, how long this Engel’s pause lasts and how bad the drop in incomes for those workers caught up in the reorganization of the economy gets depends--on things like labor regulations, the willingness of government to put its heavy thumb on the scale to help workers, and the effectiveness of a new generation of union organizing. I’m not going to attempt to encapsulate the current political situation on those fronts. I’m pessimistic--but you undoubtedly have your own read. Just suffice it to say that the outcomes are not determined and will depend on the political events of the next decades.

• And fourth, we don’t yet know how the AI story will turn out.

Some very good economists working in what I’d call a post-Keynesian mode say Adam Smith got some crucial things wrong about what drives industrial revolutions. I’m not talking about the argument that Smith really didn’t understand the pin factory and how it made pins--and the possibility that despite his attestations to the contrary he never actually visited an English pin factory or even any pin factory.

To me, the most important critique of Smith’s observations comes from English economist Joan Robinson. (Robinson died in 1983 so while she doesn’t have the authority that comes from being a dead white male, she can at least claim the authority that comes with being dead.)

Robinson’s critique boils down to this; Smith’s whole understanding of the massive increase in productivity at the pin factory is wrong.

Smith ascribes that productivity to the specialization, the division of labor and the way that work is recognized in the factory. Wrong, Robinson says--with numbers to back her up. The Industrial Revolution in England was the result of the invention of new kinds of capital equipment--the spinning jenny, the flying shuttle, the water frame, and the spinning mule--and investment in new sources of power--Watt’s steam engine--to power this equipment. Giving the average worker more equipment to work with increased worker productivity.

Robinson emphasizes that productivity growth is endogenous: when profit‑seeking firms face expanding markets and labour scarcity, they actively search out inventions and speed up the diffusion of new techniques to raise output per worker. Technical progress is thus not “a natural phenomenon that falls like the gentle rain from heaven” but something accelerated by the pressure of accumulation and profit, with capitalists continually investing in machinery to fend off profit squeezes from rising wages.

In extending Keynesian economic cycles into the long run, Robinson stressed that inventions in such periods are often labour‑saving, so rising productivity can coexist with persistent unemployment if growth in demand and investment does not keep pace.

She distinguished this “Marxian unemployment” (from technical change and accumulation) from Keynesian unemployment (from deficient effective demand), and saw the Industrial Revolution as a textbook case where machines displaced labour even as output surged. In her analysis, a “golden age” of capitalism occurs when labour is relatively scarce and capital abundant (that is, interest rates are low), so wage rates rise roughly in line with productivity while profits are maintained by continual technical progress.

For much of the Industrial Revolution, however, she reads the historical evidence as closer to an Engels‑pause configuration: productivity and profits rose faster than real wages, with a growing share going to capital, reflecting the specific institutional “rules of the game” of early English capitalism, which gave capitalists a distinct edge in deciding how productivity gains would reshared out.

Robinson’s analysis points me to some very big unanswered questions about the AI economy.

How will work be organized between human workers and thinking machines?

Will workers actually get more capital equipment, which would lead to an increase in productivity?

Or will capital spending go into infrastructure like new electrical plants that have relatively little impact on how workers do their jobs?

And maybe most importantly, will the AI economy be dominated by relatively slow-moving big companies like Microsoft and OpenAI or by a decentralized army of AI empowered creators pushing ahead as fast as they can in all directions at once?

For investors, I think, this amounts to yet another of those instances where you have to figure out how to invest in a situation of high uncertainty. Where that uncertainty will last for a considerable length of time.

This long piece is an example of the Big Picture posts--along with my weekly Micro Economics post and my weekly economics Q&A (yep, 3 posts a week)--that I will be putting up every Tuesday on my new Substack, Economics for Investors. That Substack will be going live next week. And content like this will appear only on that Substack for subscribers to Economics for Investors.

Photo by Sándor Szabó on Unsplash

Some are important--but too much inside economics baseball for those site. Some just require too long an explanation to fit there.

Which is too bad. Because I believe, looking around me at the financial markets and the global economy that investors need to know more about economics right now because, right now, economics is critically important for investors who want to make money from stocks, bonds, and other assets.

Which is why, in the next ten days, I’ll be launching a new Substack, Economics for Investors.

And why one of the regular weekly posts of that Substack will be a Q&A where I answer readers’ questions about economics, the “dismal science” of understanding what makes the U.S. and global economies run.

Adam Smith—the father of the dismal science

And why I’m giving you a preview of the kinds of questions I’ll be answering in that Q&A. As well as asking you to submit your questions on economics here so I can answer some of them now and some when that Substack is up and running. </strong>

Today I’m going to answer three of the questions about economics and investing that I get asked most frequently.

Question #1: Why do we pay so much attention to GDP when it is such a flawed number?

GDP is a favorite whipping boy among my readers and it’s not hard to see why. If there’s a more flawed major economic statistic, I can’t name it.

To take one example of a problem, GDP--Gross Domestic Product--does not include unpaid domestic labor--largely performed by women--at all. Which leads to striking anomalies like this: If a wife died and the husband had to hire and pay a domestic worker to do exactly the same tasks, it would count as an increase in GDP.

Everyone knows this produces a huge distortion in national economic accounting. Since the 1993 and 2008 revisions of the UN System of National Accounts (SNA), the core GDP production boundary still excludes most unpaid household services (cooking, cleaning, childcare, elder care, etc.), so headline GDP and growth rates do not count that work.

In response to criticism, many countries and international bodies now regularly measure the value of unpaid household work using time‑use surveys and publish it as separate satellite accounts or “extended GDP” estimates, but they explicitly keep it outside official GDP to preserve comparability and the market‑activity focus. UN studies suggest unpaid household and care work is worth on the order of 9–15% of GDP globally, and as much as 30–40% of GDP in some countries.

The OECD estimates that including unpaid household services would boost measured output by roughly 15–33% of GDP in advanced economies.

The cynics among us are likely to say that unpaid household labor isn’t included in official GDP numbers because it is performed by women.

But this is only one of the many problems with GDP.

It counts “bads” as goods and ignores damage to the environment, communities, and individuals. GDP adds spending on pollution cleanup, crime, car accidents, and health costs of harmful products as positive output, even when they reflect social or environmental harm. At the same time, it does not subtract environmental degradation (resource depletion, carbon emissions, biodiversity loss), so economies can look “healthy” in GDP terms while running down natural capital.

GDP does not directly capture health, education quality, life expectancy, safety, leisure time, social cohesion, or subjective life satisfaction. You can raise GDP by pushing longer working hours or more overtime, even if work–life balance, mental health, and community life deteriorate. It ignores sustainability and future capacity. It tracks current production/consumption but not whether that is sustainable given natural resources, infrastructure, or human capital. Short‑term growth can come at the expense of over‑borrowing, over‑building, or over‑extracting, which GDP does not flag as a problem.

I find that the reasons we continue to use GDP, when everyone realizes that it is such a flawed measure, fascinating.

Let me flag two of the more “fascinating” aspects.

First, externalities and the costs of “bads.”

GDP emphasizes, by omission, one of the key battles of capitalism from the days of Adam Smith’s pin factory to now. Company profits and return on capital are, in many instances, hugely dependent on how much cost can be transferred from private to public accounts. This problem of “externalities” is at the core of the current attacks on environmental regulation by the Trump Administration. Think oil companies’ or automakers’ or food companies’ profit margins would look different--as in much lower--if they had to fully include the costs of “bads” such as air and water pollution? GDP accounting doesn’t include the costs of “bads” so corporate accounting is comfortably supported by the larger system

Second, GDP accounting is a continued reflection of the 19th century liberal defense of market capitalism against critics on the left such as Marx and Engels.

These leftist critiques argued that capitalism inevitably led to the immiseration of the working class. Not so, the defenders of market capitalism replied. If the economy overall grew, so would workers’ wages. Market capitalism would increase the size of the pie so everyone’s slice would be bigger.

Key to that argument was showing that the pie was indeed growing. And a measure such as GDP, the global standard for measuring growth, became a key “fact” in that argument. As long as GDP said the economy was growing, there was no need to question the system itself.

It is an irony, and testimony to the continued power of this growth ideology, that the most prominent example of the fixation on GDP growth is Communist China, where everything is okay, despite all evidence to the contrary, as long as officially-measured growth meets the official target of 5% each year.

Question #2: How can the stock market hit new record highs at the same time as consumer confidence hits a 10-year low?

This one is pretty simple--it all depends on who you ask.

The Standard & Poor’s 500 hit a new all‑time high on January 6, 2026, and then set record closing highs on January 26 and 27. At the same time, the Conference Board Consumer Confidence survey dropped sharply in the latest report, falling 9.7 points to 84.5 in January 2026 from a revised 94.2 in December-—that’s its lowest level since 2014. As with any survey, the results depend on whom you ask. Stock prices and stock market direction essentially represent a survey of the wealthiest 10% of the U.S. population. Those are the people who own stocks. The Top 10% by income have an average stock portfolio on the order of $200,000–$250,000 per household. The Bottom 50% by income average stock portfolio size is only around $3,000–$4,000 per household; many in this group own no stocks at all, so the median is near zero and the average is pulled up by a minority with 401(k)/IRA holdings.

Or to put it another way, the bottom 50% of Americans own roughly 1% of total U.S. stock‑market wealth, while the top 10% own about 89%.

So the current stock market records are, largely, based on the optimism of the top 10% of U.S. households by income. That optimism is, of course, heavily influenced by the fact that these folks have done very well in the stock market’s big bull run.

The Conference Board’s U.S. Consumer Confidence Survey is based on a national sample of about 5,000 U.S. households, with roughly 3,000 completed responses per month after non‑response. In other words, it is designed to be representative of U.S. households, with sampling and weighting by age, gender, income, and geography to match the national population.

So the difference in sentiment between the two ‘surveys” is a reflection of the current two track (or K-shaped) U.S. economy where the upper 10% is doing very well, and the bottom 50% is struggling with the rising costs of health insurance, food, home furnishings, appliances, and many other categories of goods and services.

Question #3. You frequently mention economist Hyman Minsky. You obviously think he was important. (He died in 1996.) Why?

Two reasons.

First, Minsky is part of what I would call the Behavioral Revolution in economics.

The economics that I was taught at Columbia in the 1980s--and what is still the dominant strain in economic thinking--assumes rational actors who do things designed to maximize utility. Confronted with a decision, these actors do a less or more complex calculation of the utility of each choice, and then follow that course. One investment produces a higher return on investment? That’s the one we’ll choose. Traditional economics assumes individuals have stable preferences, full information, and optimize rationally. Behavioral economics instead focuses on bounded rationality, limited attention, emotions, and social influences, and models how these factors systematically shape choices in markets, policy, and everyday life. Key figures include Daniel Kahneman, Amos Tversky, and RichardThaler.

Minsky’s 1986 book Stabilizing an Unstable Economy applied a similar approach to the credit cycle.

No one paid a whole lot of attention to Minsky’s “financial instability hypothesis” when his book came out.

His theory states that periods of stability encourage increasing leverage and risk-taking, making the system progressively more fragile and eventually triggering crisis. Minsky argued that periods of prolonged stability shift financing behavior in the entire economy from hedging to speculation to Ponzi schemes, as firms and lenders lever up based on past success. Once enough companies and lenders are in the speculative/Ponzi phases, a relatively small shock will force asset sales, trigger debt deflation, and produce a crisis.

Minsky explicitly emphasizes shifting expectations, herd-like optimism, and a move from caution to exuberance, but these are embedded in a structural story about Wall Street balance sheets and liability structures.

Minsky states that markets are inherently unstable even if agents are not wildly irrational, because success changes acceptable leverage and funding norms; finance endogenously transitions into fragile structures.

The 2007–2008 global financial crisis “suddenly” brought attention to Minsky’s hypothesis. The way that a global financial crisis could begin in a relatively out of the way corner of the financial system, the sub-prime mortgage market, and then lead to a systemic crisis, seemed like the perfect example of the credit cycle that Minsky had described. Economists began to speak of that crisis as a “Minsky moment.”

Which leads me to the second reason I think Minsky is important now. The current moment in the credit markets show exactly the worrying signs that Minsky’s postulated before a crisis.

The sectors under stress are different--it’s not subprime mortgages this time but the private credit market and cryptocurrencies and a loosening of Washington regulation of these markets--but it’s easy to see the loosening of credit standards and the increased risk-taking that Minsky described.

I think it’s really important right now for investors to see the risks in this market through the lens of Minsky’s credit cycle. To learn more, start with Minsky’s own short paper Hyman P. Minsky – “The Financial Instability Hypothesis” (Levy Institute Working Paper 74). You can find it online at https://www.levyinstitute.org/wp-content/uploads/2024/02/wp74.pdf. It’s about 20 pages long.

Okay, these are my first questions for my new project, Economics for Investors. Send me more questions in the “comment” to this post. I’ll answer the next batch here on my Economics for Investors Substack once it is formally launched in the next 10 days.