There’s a question nobody in polite economic circles likes to answer directly… who benefits from stagflation?
The standard answer is “nobody.” Stagflation is the economic equivalent of a house fire in a flood, everything bad happens at once. Growth stalls, prices rise, and the policy toolkit that works for one problem makes the other worse.
But nobody benefits is a cop-out. Somebody always benefits. The question is who, and whether any of them might be quietly comfortable with the arrangement.
The Winners’ Table
Governments carrying massive debt. Inflation is the oldest sovereign debt trick in the book. You borrow in today’s dollars, you repay in tomorrow’s weaker ones. Take $30 trillion in national debt, run 6% inflation for several years, and the real burden shrinks meaningfully; no restructuring required, no default headlines. Governments still suffer politically. Voters feel the price increases at the grocery store. But the balance sheet quietly improves while everyone argues about eggs.
Commodity producers. When stagflation is supply-shock driven, and it usually is, the firms sitting on the supply side of the shock print money. Oil companies. Mining firms. Agriculture producers. Revenue climbs while demand stays surprisingly sticky because people still need to eat and drive. The 1970s minted energy fortunes. The pattern repeats.
Companies with pricing power. The firms that can raise prices faster than their costs rise don’t just survive stagflation; they widen margins. Energy. Defense. Monopolistic tech platforms. Some infrastructure plays. The common thread isn’t sector, it’s leverage… if your customers can’t leave, inflation is a feature, not a bug.
Owners of real assets. Commodities, real estate, gold, resource equities, anything denominated in stuff rather than promises tends to hold or gain value. Cash and bonds, the traditional “safe” assets, quietly bleed purchasing power. The retiree with a fixed pension and a savings account gets hollowed out. The landlord with a paid-off building does fine.
Highly leveraged borrowers. This one’s counterintuitive. If you borrowed a million dollars at a fixed rate before inflation spiked, inflation is doing you a favor. Your debt is denominated in nominal dollars. Every year of inflation shrinks the real weight of that repayment. The catch is you have to survive long enough to benefit, leverage kills if cash flow doesn’t hold, but for those who do, inflation is an accelerant on the path to real-terms debt relief.
A caveat on the winners
This [Winners’] table assumes a level of control that might not exist. The landlord with a paid-off building “does fine”, until rent controls arrive or social unrest makes ownership a liability instead of an asset. The highly leveraged borrower benefits from inflating away debt, unless they have to refinance at 10% in the middle of it. The commodity exporter prints money, until demand destruction finally catches up or governments impose windfall taxes. Sometimes the fire gets so big it burns the people who started it to stay warm.
And beneath all of these tactical risks sits the structural one… social cohesion. That’s the ultimate limit on the winners’ table. Every asset on it, the real estate, the commodity revenue, the inflated-away debt, depends on a functioning society underneath it.
Courts that enforce contracts. Infrastructure that moves goods. Streets that are safe enough to do business on. Even a billionaire needs a working post office.
When the losers’ list gets long enough, when real wages fall far enough, when enough people feel the system is rigged for someone else’s benefit, the winners’ assets don’t just lose value. They become stranded by instability.
Capital requires stability to function as capital. Without social cohesion, it’s just numbers on a screen in a country that doesn’t work anymore.
The Losers’ Table
The losers are easier to identify because there are more of them… wage earners whose pay doesn’t keep pace, savers watching their purchasing power evaporate, fixed-income retirees on pensions designed for a different price level, and small businesses without the market power to pass costs through.
Real wages fall. That’s the defining feature. The economy stagnates, but prices don’t.
The AI Accelerant
Here’s where the picture gets uncomfortable.
A weaker labor market doesn’t just cause pain, it opens a door. Specifically, the door marked “automation.” When labor is strong and expensive, companies still want to automate, but the political and social friction is high. Layoffs generate backlash. Unions push back. Executives hesitate.
When labor weakens, automation becomes politically easier. Headcount reductions get filed under efficiency instead of greed.
And AI slots directly into the playbook: reduce headcount, invest in productivity tools, automate processes. This isn’t theoretical. It’s the historical pattern.
Major automation waves follow economic stress with striking regularity:
- The 1970s and 80s brought industrial automation.
- The early 2000s brought outsourcing and enterprise software.
- Post-2008 brought cloud infrastructure and SaaS.
- The 2020s are bringing AI.
Economic pressure forces efficiency upgrades. It always has. What’s different this time is the breadth of the tool. Previous automation waves hit blue-collar work hardest. AI hits white-collar work, management layers, knowledge work, the professional class that previously considered itself automation-proof.
Reading the Dashboard: ISM Services
If you want to see stagflation forming in real time, the ISM Services report is one of the better dashboards available. But you have to read it correctly, and most people don’t.
The ISM Prices Paid index is a diffusion index, not a price level. It doesn’t measure how much prices rose. It measures how many firms reported that their input costs went up compared to last month.
- A reading of 50 means no net change, equal numbers of firms seeing increases and decreases.
- Above 50 means more firms are still seeing costs rise. The higher above 50, the broader and faster the increases.
So when Prices Paid fell from 66.6 in January 2026 to 63.0 in February, that doesn’t mean prices dropped. It means prices were still rising, just for fewer firms, at a less intense pace.
The rate of price acceleration cooled, but the level of prices remained high.
That distinction matters enormously for interpreting the data. A reading of 63 still means a clear majority of services firms are reporting higher costs month over month.
Now look at the other side of the chart: Employment. It crossed back above 50 in January 2026, and the narrative immediately softened. The “stagflation jaws”, the gap between Prices Paid and Employment, narrowed from nearly 22 points in October 2025 to about 11 points in February.
That narrowing is real. But don’t confuse it with health.
We are not looking at “everything is fine because employment ticked above 50 once.” We are looking at services firms still reporting rising input costs while the broader labor market weakens and white-collar jobs get cut in the name of efficiency and AI.1
That is not healthy expansion. That is a stagflationary setup with a different mask on.
Employment crossing 50 in a services diffusion survey doesn’t mean the labor market is strong. It means the rate of deterioration paused. Meanwhile, the structural forces, AI-driven headcount reduction, margin optimization through automation, real wage compression, continue to build underneath the surface.
The jaws narrowed. The teeth didn’t go away.
The Rhyme Nobody Wants to Hear
For anyone inclined to read the cooling ISM numbers as an all-clear, here’s the historical overlay that should give you pause.
The gold line is year-over-year CPI from 1966 to 1982, the full stagflation era. The white line is 2014 to present. The two are time-shifted to align the cycles, and the first hump rhymes almost perfectly. The post-COVID inflation spike tracks the early 1970s spike in both magnitude and shape. The disinflation coming down follows a strikingly similar path. And right now, at 2.4% YoY, we’re sitting at roughly the point in the 1970s analog where the first wave’s relief gave way to the second wave’s acceleration; the one that took inflation above 12%.
The chart doesn’t predict. But it contextualizes.
The 1970s didn’t deliver one inflation shock. They delivered two. The first one scared everyone. The pullback calmed them down. Then the second wave hit harder, because the structural conditions that caused the first wave were never resolved, they were just temporarily masked by a brief disinflationary window.
Look at the conditions today through that lens. Supply chains are still fragile. Energy markets are being reshaped by war.2 Fiscal deficits remain enormous.3 And the Fed, having declared a degree of victory on inflation, has limited room to respond aggressively if the second wave arrives, especially with a labor market that’s already weakening under AI-driven restructuring.
The 1970s didn’t deliver one inflation shock. They delivered two. The first one scared everyone. The pullback calmed them down. Then the second wave hit harder.
The ISM data shows the inflationary pressure never fully cleared. The overlay shows what happened last time the pressure didn’t fully clear. Neither chart tells you what will happen. Together, they tell you what the setup looks like, and the setup looks familiar.
Does Anyone Actually Want This?
Rarely does anyone design stagflation. But certain actors tolerate inflation longer than they otherwise might when it’s solving a problem for them. Governments watching their debt burden erode in real terms aren’t in a rush to crush inflation. Commodity exporters benefiting from elevated prices aren’t lobbying for demand destruction. Firms using the cover of a weak labor market to restructure around AI aren’t asking for tighter conditions.
It’s a side-effect of conflicting objectives, not a conspiracy. But the incentives are real, and they point in a consistent direction.
The Instability Equation
The market is currently pricing a specific scenario… AI productivity gains combined with weaker labor bargaining power equals higher corporate margins. And on paper, that math works.
For a while.
The tension arrives when AI reduces wages faster than it reduces prices. If the productivity gains flow to capital and the cost reductions don’t flow to consumers at the same pace, consumer demand weakens. The very workforce being “optimized” is also the customer base. You can’t squeeze the labor piston indefinitely without eventually starving the demand engine.
And here's what makes this cycle different from every previous one… speed.4
Historically, labor shifted from the fields to the factories over decades. Generations had time to retrain, relocate, adapt. The social fabric stretched but didn't tear. If AI shifts labor out of the knowledge office over months, not decades, not even years, the demand engine doesn't just starve. It seizes.
There's no graceful adjustment period when the displacement curve goes vertical. The consumer base collapses before the productivity gains have time to flow back as lower prices or new industries.
The economy doesn’t gently rebalance. It stalls at speed.
In piston terms: the resource piston is running hot, the currency piston is unstable, the labor piston is compressing, and the capital piston is expanding. That configuration generates profits for capital owners and crushing pressure on labor income. It’s the classic stagflationary profile, and it’s why stagflation always becomes politically explosive.
The winners can do the math. The losers can feel it. And the gap between those two experiences is where the real instability lives.
If you found this analysis useful, consider sharing it with someone navigating these crosscurrents. The economic weather is shifting, understanding who holds the umbrella matters.