The composition of that growth, however, tells a very different story. A disproportionate share of the improvement came from the infrastructure being built around artificial intelligence: software, servers, computer equipment, data centers, and the electricity infrastructure needed to support them. Once that unusually concentrated investment cycle is separated from the rest of the economy, the underlying growth rate appears much closer to 1% than to a genuine economic boom.

This distinction matters because the AI buildout is not evidence that demand is accelerating across the entire economy. It is evidence that a small number of extraordinarily profitable technology companies are spending enormous amounts of money on a highly capital-intensive investment cycle. That spending counts as GDP today, regardless of whether the data centers, chips, and models eventually produce returns sufficient to justify their cost.

At the same time, the companies funding this boom are beginning to experience a sharp reduction in free cash flow. The labor market is creating very few jobs, hundreds of thousands of people are leaving the labor force, and full-time employment is declining. These are not the conditions normally associated with a broad economic expansion.

The AI buildout is not hiding a booming economy. It is preventing a slowing economy from looking as weak as it really is.

Subscribe now

The Headline GDP Number Hides an Increasingly Concentrated Economy

Real GDP increased at an annualized rate of 2.1% during the first quarter of 2026. The figure was revised upward from the second estimate, largely because imports were lower than previously calculated. Investment, exports, government spending, and consumer spending all contributed positively to growth. However, the underlying measures were considerably weaker than the headline GDP figure.

There is no official category called “AI investment” in the national accounts. The Bureau of Economic Analysis does not separate an AI server from a computer purchased for traditional business activity, nor does it determine what portion of software investment is specifically connected to generative AI. Any estimate therefore requires assumptions about which spending categories should be included.

The Federal Reserve constructed one useful proxy by combining software investment, data-center construction, investment in power facilities, computer and peripheral equipment, and the net-export effect of computer equipment. The Fed explicitly notes that the calculation is imperfect because many of these categories include non-AI activity and because the buildout relies heavily on imported equipment.

During the first quarter of 2026, software contributed approximately 0.53 percentage points to real GDP growth, while computer and peripheral equipment contributed another 0.59 percentage points. Data centers and power facilities each added around 0.03 percentage points. Together, those positive categories contributed approximately 1.18 percentage points to the economy’s total growth rate of 2.1%.

Looked at on a gross basis, these categories therefore represented approximately 56% of total GDP growth. This is the origin of the claim that AI-related spending explained roughly half of economic growth during the quarter.

Share

Outside the AI Buildout, the Economy Is Growing Near Stall Speed

If the net AI-buildout contribution of approximately 50% is removed from the 2.1% GDP growth rate, the remaining economy grew at an annualized rate of approximately 1.08% (2.14-1.06).

Neither calculation is a perfect counterfactual because different components of GDP interact with each other. The companies constructing data centers pay workers, purchase services, consume electricity, and generate income that circulates through the rest of the economy. It would therefore be incorrect to assume that the economy could simply remove AI investment without affecting any other category.

Nevertheless, the exercise reveals the degree of concentration. Outside the direct AI-related spending categories, the economy appears to be growing somewhere around 1% to 1.2%. It describes an economy operating close to stall speed, not one experiencing a broad demand-driven boom.

The weakness also explains why the apparent strength in GDP is not being confirmed by employment. Capital-intensive investment in servers, software, chips, and data centers can add substantially to GDP without creating the same number of jobs that would accompany a housing boom, a manufacturing expansion, or strong growth in consumer-facing services.

Share Signals Over Noise

The Buildout Is Beginning to Consume the Hyperscalers’ Cash Flow

Alphabet provides one of the clearest examples of the financial cost of the buildout. During the second quarter of 2026, the company generated $39.1 billion in operating cash flow but spent $44.9 billion on capital expenditures, primarily related to technical infrastructure. On a simple operating-cash-flow-minus-capex basis, Alphabet produced quarterly free cash flow of approximately negative $5.8 billion.

This does not mean that Alphabet has become an unprofitable or structurally cash-burning company. Its advertising, subscription, and cloud businesses continue to generate enormous operating income, and its trailing free cash flow remains positive. The second-quarter result does show, however, that the current pace of infrastructure investment exceeded the cash generated by the business during that quarter.

Alphabet expects its 2026 capital expenditures to reach between $175 billion and $185 billion, illustrating the extraordinary scale of the spending required to remain competitive in AI infrastructure.

Oracle represents an even more extreme case. During fiscal year 2026, Oracle generated a record $32 billion of operating cash flow, but reported free cash flow of negative $23.7 billion as it expanded Oracle Cloud Infrastructure. The difference implies capital expenditures of approximately $55.7 billion during the year.

Oracle also announced plans to raise between $45 billion and $50 billion through a combination of debt and equity financing during calendar year 2026. The company has an enormous contracted backlog and may eventually generate substantial revenue from that infrastructure, but it must finance the construction long before most of the associated revenue is recognized.

This is the financial reality behind the GDP contribution. The AI investment boom supports economic growth today because the hyperscalers are spending aggressively. But that same spending is consuming their cash flow and increasing their dependence on continued revenue growth, capital-market access, and future returns.

Share

Why the Headline P/E Ratio Can Give the Wrong Impression

The valuation problem becomes more complicated because the earnings of some hyperscalers are also being increased by gains on their investments in private companies.

During the second quarter of 2026, Alphabet reported $98 billion in other income and expense, including approximately $99 billion in net gains on equity securities. The company stated that these gains were primarily related to unrealized increases in the value of SpaceX and another private company. These gains increased pre-tax income and net income, but they did not come from advertising sales, Google Cloud revenue, subscriptions, or operating efficiencies.

A gain can be legitimate without being recurring operating income. Alphabet owns assets whose estimated market values increased, and accounting rules required those changes to be recognized. However, the gain expanded the earnings denominator used in the price-to-earnings ratio without producing a corresponding increase in revenue or operating cash flow.

As of July 29, 2026, Alphabet’s market capitalization was approximately $4.1 trillion and its headline P/E ratio was close to 16.9. On the surface, that multiple can appear relatively inexpensive for a company growing rapidly in cloud computing and AI.

But the earnings figure includes enormous unrealized gains that may not recur. A valuation based only on headline net income can therefore make the stock appear cheaper than it would on the recurring operating earnings produced by Search, YouTube, Cloud, and the company’s other core businesses.

The P/E is not mathematically incorrect. The problem is the quality and repeatability of the earnings being used in the calculation.

Subscribe now

Price-to-Sales Reveals What the Earnings Multiple Is Hiding

Price-to-sales is not a perfect valuation metric. It ignores profit margins, capital intensity, taxes, and differences in business quality. A company with high margins and recurring revenue should reasonably trade at a higher sales multiple than a low-margin cyclical company.

However, price-to-sales is useful when reported earnings contain unusually large non-operating items. Revenue cannot be increased by revaluing a private investment. Comparing market capitalization with sales therefore provides a cleaner indication of how much investors are paying for the company’s underlying commercial activity.

The contrast is revealing. The earnings multiple looks relatively moderate, while the sales multiple remains at a level normally associated with exceptional growth, high margins, and durable competitive advantages. Part of that difference comes from the extraordinary unrealized gains included in net income.

This does not automatically mean that Alphabet is overvalued. It means that investors should not interpret the headline P/E as if every dollar of current earnings came from repeatable operations.

What Happens When AI Investment Stops Accelerating?

The greatest risk is not that hyperscalers suddenly stop spending. The risk is that their spending simply stops accelerating.

GDP growth measures changes in economic activity. A company can continue investing $150 billion per year in infrastructure and still provide little additional contribution to GDP growth if it spent approximately the same amount during the previous year. A high level of capex supports the level of GDP, but it is the growth rate of that capex that determines its contribution to GDP growth.

The Federal Reserve Bank of St. Louis demonstrated this dynamic during 2025. AI-related investment categories contributed approximately 1.3 percentage points to GDP growth during the first quarter and 1.16 percentage points during the second quarter. Their combined contribution then declined to approximately 0.48 percentage points in the third quarter, even though the level of investment remained elevated. The contribution declined because the rate of investment growth slowed.

This is the most important vulnerability in the current economic outlook. The economy does not need hyperscalers to cut their capital budgets for the AI boost to weaken. Capex merely needs to grow more slowly than it did during the previous period.

If Alphabet, Microsoft, Amazon, Meta, Oracle, and the other major infrastructure investors eventually stabilize their spending, the economy will lose one of its largest marginal sources of growth. The data centers will remain, the servers will continue operating, and investment will still be historically high, but the contribution to growth will fade.

At that point, the underlying economy will become much more visible. If consumption, housing, manufacturing, and employment have not recovered by then, headline GDP could fall rapidly toward the weak growth rates currently visible beneath the AI investment cycle.

Leave a comment

The Buildout Is Supporting GDP Before It Has Proven Its Return

There is nothing imaginary about the AI infrastructure being constructed. The data centers are real, the servers are real, electricity demand is real, and the models are becoming increasingly capable. The investment is creating economic activity and may ultimately produce major productivity gains.

The uncertainty lies in the return.

Alphabet’s negative quarterly free cash flow and Oracle’s deeply negative annual free cash flow show that the infrastructure must be financed today, while the associated profits may arrive years later. In the meantime, competitive pressure is increasing, the price of model access is falling, and businesses are still learning how to reorganize themselves around the technology.

The risk is that the industry builds capacity based on revenue and demand assumptions that prove too optimistic. If returns disappoint, companies may slow investment, depreciate assets more aggressively, cancel projects, or become more selective about future data-center commitments.

Because AI-related capex currently explains such a large share of economic growth, a reduction in investment momentum would no longer be only a technology-sector story. It would become a macroeconomic shock.

Share

Final Takeaway

The U.S. economy grew by 2.1% during the first quarter of 2026, but that headline materially overstates the breadth of the expansion. Depending on the methodology, AI-related investment explained approximately one-third to more than one-half of the quarter’s growth.

Outside that buildout, the economy appears to be expanding at roughly 1% to 1.2%. That conclusion is consistent with weak gross domestic income, modest private domestic demand, and an increasingly fragile labor market.

At the same time, the companies funding the AI boom are consuming enormous amounts of cash. Alphabet’s quarterly capital expenditures exceeded operating cash flow, while Oracle generated negative free cash flow of $23.7 billion during its fiscal year. Reported earnings and P/E ratios are also being affected by large unrealized gains on private investments, making some valuations appear more attractive than the underlying operating figures suggest.

None of this means that AI will fail. Artificial intelligence may become one of the most important technologies in economic history. It may eventually transform productivity, reduce costs, and generate entirely new industries.

But the current economic contribution comes primarily from building the infrastructure, not from a completed productivity revolution. GDP is receiving the spending today, while investors are still waiting for proof that the returns will justify the capital committed.

The larger macroeconomic conclusion is therefore difficult to ignore. The U.S. economy is not experiencing a broad boom. It is experiencing an extraordinarily concentrated investment cycle that is temporarily offsetting weakness everywhere else.

If AI investment continues accelerating, that cycle can support growth for longer. If it merely stabilizes, the contribution to GDP will fade. And if it slows while employment and consumption remain weak, the underlying economic contraction will become much harder to hide.

The AI buildout is not evidence that the economy has escaped the slowdown. It may be the only reason the slowdown has not yet appeared more clearly in the headline GDP number.

This publication is for educational and informational purposes only. It does not constitute investment advice.

Artificial intelligence will almost certainly change the global economy. Over the next twenty or thirty years, it may transform how companies produce goods, design products, write software, conduct research, manage supply chains, and organize their workforces. Denying that potential would be a mistake. However, accepting that AI will eventually increase productivity does not mean that the productivity gains assumed in today’s valuations will arrive immediately.

The history of electricity provides one of the clearest examples of this distinction. Electric power and electric motors existed decades before they produced a visible acceleration in economy-wide productivity. Factories had originally been designed around steam engines, central drive shafts, and belts that transmitted mechanical power throughout the building. Simply replacing the steam engine with an electric motor did not fundamentally improve the production process because the rest of the factory remained organized around the old technology.

The real productivity gains appeared when manufacturers redesigned the factory itself. Smaller electric motors allowed individual machines to operate independently, production floors could be reorganized, workflows could be improved, and workers could be placed around the production process rather than around a central source of mechanical power. The San Francisco Federal Reserve notes that electrification required new infrastructure, redesigned production, reengineered factories, and retrained workers before it generated sustained productivity gains. Research on the period similarly finds that electricity’s most significant contribution to American productivity became visible during the early twentieth century and particularly the 1920s.

Artificial intelligence will probably follow a similar path. Giving employees access to an AI assistant may improve certain tasks, but it does not automatically reorganize an entire company. The larger gains will emerge when businesses redesign their operations around the technology, change how information moves through the organization, retrain employees, rewrite internal processes, and create products that could not have existed before AI. That transformation is much more complicated than purchasing a software subscription.

This distinction is essential for investors. A technology can be genuinely revolutionary while the financial market still overestimates how quickly it will produce profits. The long-term productivity story may be correct, while the current valuation story may still be dangerously premature.

Share Signals Over Noise

The Technology and the Investment Are Not the Same Thing

Financial bubbles are often built around real innovations. The internet was real in 1999. Railroads were real in the nineteenth century. Telecommunications transformed the world, and electricity became one of the most important technologies in economic history. The fact that an innovation eventually succeeds does not mean that every company associated with it deserves any valuation at any price.

This is where much of the current AI debate goes wrong. Investors are usually forced into one of two camps. Either artificial intelligence is a revolutionary technology and therefore current valuations are justified, or AI is a bubble and therefore the technology must be overhyped. Those are not the only possibilities.

AI can become one of the most important general-purpose technologies in history while the companies developing today’s models still struggle to produce returns that justify their valuations. Productivity gains may eventually flow primarily to the businesses using AI rather than the companies building the foundational models. Competition may reduce the price of intelligence faster than demand increases. The technology may create enormous value for society while destroying value for investors who paid too much for the wrong layer of the ecosystem.

The correct question is therefore not whether AI works. The correct question is whether the companies receiving today’s extraordinary valuations will eventually capture enough of the value created by AI to justify those valuations.

Share

The Financing Machine Behind Frontier AI

The leading AI laboratories are extraordinarily capital-intensive businesses. Training frontier models requires vast amounts of computing power, while operating those models creates continuing inference expenses every time a customer sends a request. Revenue is growing rapidly, but so are research, infrastructure, customer acquisition, and computing costs.

According to shareholder documents reported by The Information and summarized by Reuters, OpenAI generated approximately $5.7 billion in revenue during the first quarter of 2026 but burned through about $3.7 billion in cash during the same period.

Anthropic has also experienced extraordinary growth and attracted enormous amounts of capital. In May 2026, the company raised $65 billion at a reported post-money valuation of $965 billion. The company’s revenue was growing rapidly, particularly through enterprise and coding products, but it was still operating at a loss and depended on outside capital to finance its expansion.

This does not mean that either company is destined to fail. Young technology companies frequently lose money while building infrastructure, developing products, and acquiring customers. The important point is that these companies are not currently funding their entire expansion from internally generated profits. Their business models continue to depend on the willingness of investors and strategic partners to provide additional capital.

As long as capital remains abundant, the system can continue expanding. New investors provide cash, the AI company increases spending, revenue grows, and the next funding round is completed at a higher valuation. The rising valuation then affects not only the private company and its employees, but also the financial statements of the public companies that invested in it.

That is where the private AI boom begins to enter the earnings of the S&P 500.

Share Signals Over Noise

How a Funding Round Becomes Corporate Earnings

Consider the relationship between Amazon and Anthropic. Amazon initially provided billions of dollars to Anthropic through convertible investments and later held nonvoting preferred shares in the company. That investment gave Amazon exposure to any increase in Anthropic’s private valuation.

When Anthropic completes a new funding round, the price paid by new investors can provide an observable reference point for the value of Amazon’s existing shares. If the new round takes place at a significantly higher valuation, accounting rules may require Amazon to mark the value of its investment upward. The increase is then recognized through other income, even though Amazon has not sold its shares and Anthropic has not distributed profits to Amazon.

Amazon’s first-quarter 2026 filing provides a clear example. The company reported that upward adjustments to private-company equity investments reflected observable price changes, primarily involving its nonvoting preferred shares in Anthropic. Amazon recorded approximately $12.3 billion of upward adjustment from the observable price change and reported total pre-tax gains of $16.8 billion from its Anthropic investments during the quarter. Its balance-sheet carrying value for the preferred shares rose from approximately $14.8 billion at the end of 2025 to $32 billion by March 31, 2026.

The economic sequence is important. Amazon invests in Anthropic. Anthropic spends the capital while attempting to expand its revenue and technology. A later group of investors purchases shares at a higher valuation. That higher price allows Amazon to increase the accounting value of the shares it already owns. The increase appears as non-operating income and raises Amazon’s reported net income and earnings per share.

No Anthropic shares needed to be sold for the gain to appear. Amazon did not receive the entire gain in cash. Anthropic did not need to become profitable. The new funding round itself provided the observable price that supported the higher accounting value.

This is permitted under accounting rules, and the gain is economically meaningful because Amazon owns an asset that investors were willing to value more highly. However, it is fundamentally different from profit generated by selling products, delivering cloud services, or improving operating efficiency.

Share Signals Over Noise

The Q1 Earnings Surge Was Not Entirely What It Appeared to Be

The effect became visible at the index level during the first-quarter 2026 earnings season. FactSet reported that the blended year-over-year earnings growth rate for the S&P 500 jumped from 15% to 27.1% over the week in which Alphabet, Amazon, and Meta released their results. Those three companies accounted for 71% of the net dollar-level increase in S&P 500 earnings during that reporting period.

The reported figures were remarkable. Alphabet’s GAAP earnings included a $37.7 billion net gain, primarily resulting from unrealized gains on non-marketable equity securities. Amazon’s results included the $16.8 billion of pre-tax gains from its Anthropic investments. Meta’s earnings were also boosted by an $8.03 billion income-tax benefit. These items were legitimate components of GAAP earnings, but they were not ordinary operating profits produced by the companies’ day-to-day businesses.

This does not mean that the underlying businesses were weak. Alphabet reported 22% revenue growth, a 30% increase in operating income, and a two-percentage-point expansion in operating margin. Amazon’s North American, international, and AWS operating businesses also improved. The companies generated real revenue and real operating profits.

The problem is not that the operating businesses were imaginary. The problem is that headline EPS growth combined two very different sources of earnings: profits generated by core operations and gains created by revaluing private investments. When those categories are presented together, the index can appear to be experiencing a much broader and more sustainable earnings acceleration than the underlying operating data alone would suggest.

The exact claim that Anthropic-related revaluations (or similar investments) represented 40% of total S&P 500 earnings growth. A small group of companies caused a dramatic increase in reported index earnings, and a significant portion of their EPS surprises came from non-operating gains and tax effects rather than recurring operating income.

Subscribe now

These Are Real Accounting Gains, but They Are Not Recurring Sales

There is an important difference between calling an accounting gain non-operating and calling it fake. The gains are not fabricated. Amazon owns a stake in Anthropic, and investors were willing to pay a much higher price for comparable shares. Marking that asset upward reflects the observable market value available at the time of the funding round.

However, an unrealized gain should not be treated as if it had the same quality as recurring operating income. A dollar earned by selling cloud services or advertising can potentially be repeated quarter after quarter if the business continues to perform. A dollar created by revaluing a private investment requires another increase in the investment’s value before it can contribute again.

This produces a difficult comparison problem. If Anthropic’s valuation rises from one funding round to the next, Amazon may record another gain. If the valuation remains unchanged, there may be no new gain. If a later financing takes place at a lower price, or if the investment becomes impaired, part of the previous increase could be reversed.

Amazon’s own filing states that future observable price changes or impairments will continue to be recognized through other income. This means the mechanism works in both directions. A rising private valuation can increase public-company earnings, but a falling valuation can eventually reduce them.

The gain therefore represents a claim on potential future value, not proof that the future profits required to support that value have already been generated.

Share

The Valuations Assume Something Close to Monopoly Economics

A valuation approaching one trillion dollars (like OpenAI or Anthropic) requires much more than rapid revenue growth. It requires the market to believe that the company will eventually produce extraordinary free cash flow, maintain high profit margins, and retain a substantial portion of the economic value created by its technology.

That assumption would be easier to defend if the leading American AI laboratories were building durable monopolies. A company with unique technology, limited competition, high switching costs, and strong pricing power can convert rapid adoption into extraordinary long-term profitability. Under those conditions, investors may be willing to pay an enormous valuation today because the company could dominate a highly profitable industry tomorrow.

The problem is that the competitive structure of artificial intelligence increasingly points in the opposite direction. OpenAI and Anthropic do not operate in an isolated market. They compete with Google, Meta, xAI, DeepSeek, Alibaba’s Qwen, Moonshot’s Kimi, and a growing number of open-weight and lower-cost models. Customers can already use several models simultaneously, select different providers for different tasks, and move workloads toward whichever model offers the best combination of performance, speed, and price.

The price differences are particularly important. DeepSeek V4 Flash therefore costs approximately 1.9% of Claude Sonnet’s standard price and about 1.1% of Claude Opus’s price. Depending on the specific model and workload being compared, the Chinese alternative can therefore operate at approximately 1% to 3% of the price of leading American frontier models. A figure close to 1.5% is not unreasonable as an illustration of this enormous pricing gap, although the exact percentage changes depending on the model, the mix between input and output tokens, and the use of caching.

This does not prove that every DeepSeek response is equivalent to the output of Claude or ChatGPT. Differences in reliability, reasoning quality, coding ability, security, latency, enterprise support, and regulatory compliance remain important. Nevertheless, the economic signal is difficult to ignore. Capable Chinese models are being offered at only a small fraction of the prices charged by leading American laboratories.

The competitive threat is no longer theoretical. Google Cloud has made DeepSeek R1 available through Vertex AI Model Garden, allowing developers to access the model through Google’s own cloud infrastructure and APIs. Google’s current documentation also lists several DeepSeek models as managed or self-deployed options within Gemini Enterprise Agent Platform. In other words, one of the largest American developers of frontier AI is simultaneously providing enterprise customers with access to a much cheaper Chinese competitor.

This is exactly what should concern investors paying monopoly-like valuations for OpenAI and Anthropic. A monopoly can preserve prices because customers have nowhere else to go. A competitive commodity market behaves differently. When several providers offer similar capabilities, customers compare costs, providers lower prices, and a larger share of the economic value flows to the users of the technology rather than to the companies producing it.

The leading American laboratories are already responding by offering smaller models, cheaper tiers, batch discounts, caching discounts, and promotional pricing. Those reductions may accelerate adoption, but they also make it more difficult to achieve the extraordinary margins required to justify current private valuations. Chinese models are gaining market share.

The core contradiction is that Anthropic and OpenAI are being valued as if they will eventually possess monopoly-like pricing power, while the market for artificial intelligence is rapidly developing commodity-like characteristics. Model performance is converging, alternatives are multiplying, and the price of intelligence continues to fall.

If Anthropic or OpenAI were the only companies capable of delivering advanced intelligence, their current valuations might eventually be supported by enormous margins and durable pricing power. But if customers can obtain sufficiently similar results from Chinese or open-weight models at approximately 1% to 3% of the cost, the long-term economics become much more difficult.

Subscribe now

What Could Break the Loop

The valuation loop does not require the AI industry to collapse. It only requires future outcomes to be less extraordinary than the outcomes already embedded in current prices. A funding round completed at the same valuation as the previous round would eliminate the next upward revaluation. A down round could reduce the value of earlier investments. An IPO priced below the latest private valuation could force investors to confront the difference between a negotiated private funding price and the price available in a liquid public market.

Price competition represents another risk. If model performance continues converging while inference prices fall, the laboratories may generate more usage but less profit per unit. Revenue could continue rising while the path to sustainable margins moves further into the future. That would weaken the fundamental justification for the valuations supporting the accounting gains.

Finally, the gains themselves can reverse. Amazon explicitly accounts for future observable changes in the price of its private investments and for potential impairments. The same accounting mechanism that increases earnings when valuations rise can reduce earnings when those valuations fall.

The fragility is therefore not only in the distant future. It is already present in today’s financial statements.

The Bubble Is Not AI—It Is the Timing of the Expectations

Calling AI a bubble does not mean that the technology is useless. It means that the financial claims placed on the technology may have moved much faster than the cash flows capable of supporting them.

Electricity changed the world, but the transformation required decades of infrastructure, experimentation, organizational redesign, and complementary investment. The same may be true for artificial intelligence. Companies will need time to restructure themselves around AI before the largest productivity gains become visible.

The market is not waiting for that transformation. It is already valuing private laboratories as if they will become some of the largest and most profitable companies in history. It is already recognizing part of those valuation increases in the earnings of major public companies. It is already using those earnings to support the valuation of an increasingly concentrated stock market.

Meanwhile, the price of model access is falling, competitors are multiplying, capital requirements remain enormous, and the laboratories continue to depend on outside financing. The technology is improving, but the economic moat is not yet proven.

That is the contradiction at the center of the AI bubble. The market is capitalizing decades of potential productivity before it knows which companies will capture that productivity, what prices they will be able to charge, or whether their competitive advantages will survive.

The most important part of the current bubble is not simply that AI companies are expensive. It is the financial mechanism connecting private funding rounds, public-company investment gains, headline EPS growth, and market valuations. When new capital enters Anthropic at a higher valuation, Amazon can recognize a gain. When large gains appear in Amazon’s and Alphabet’s earnings, S&P 500 earnings growth looks stronger. When index earnings look stronger, investors become more comfortable paying elevated prices for the same companies financing the AI ecosystem.

This mechanism can continue while private valuations rise and capital remains abundant. But it does not create permanent profitability by itself. Eventually, the valuations must be supported by operating cash flows, pricing power, and durable competitive advantages.

If AI laboratories develop monopoly-like businesses, current valuations may eventually appear reasonable. If models become increasingly interchangeable and prices continue falling, the productivity benefits may flow mainly to users while the laboratories struggle to earn adequate returns on the capital invested.

At this stage of the cycle, the level of job creation matters less than the direction. What we are seeing is not just a single weak print, but a gradual deceleration in hiring momentum. When job growth slows to these levels, it becomes increasingly difficult for the labor market to absorb new entrants and sustain income growth across the economy.

Subscribe now

Negative Revisions Confirm the Trend

The revisions also moved in the wrong direction. April and May were revised down by a combined 74,000 jobs, which means the labor market was weaker than previously reported. This matters because one weak month can always be dismissed as noise, but weak payroll growth combined with negative revisions tells a different story.

Revisions are often overlooked, but they are critical for understanding the true trajectory of the labor market. When past data is consistently revised lower, it suggests that the initial strength was overstated. In this case, the revisions reinforce the idea that the labour market is not stabilizing; it is gradually losing momentum.

What made the report confusing was that the unemployment rate actually fell from 4.3% to 4.2%. At first glance, that looks like good news. But this is exactly why looking only at the unemployment rate can be misleading late in the business cycle.

The unemployment rate did not fall because the economy created a large number of jobs. It fell because a large number of people left the labor force. This distinction is crucial. A declining unemployment rate driven by job creation signals strength. A declining unemployment rate driven by falling participation signals weakness.

Share

The Collapse in Labor Force Participation

In June, the labor force declined by 720,000 people, while household employment fell by 507,000. The labor force participation rate dropped from 61.8% to 61.5%, one of the sharpest monthly declines outside of the Covid period.

This is the most important part of the report. If people stop participating in the labor force, they are no longer counted as unemployed, even if the underlying labor market is weakening. This creates a statistical illusion where the unemployment rate improves while the real economy deteriorates.

Participation is one of the most important indicators of labor market health because it reflects both opportunity and confidence. When participation falls sharply, it often signals that workers are becoming discouraged or that job opportunities are not sufficient to keep them engaged.

This is why the unemployment rate loses relevance when participation is falling. To be counted as unemployed, a person must not have a job and must be actively looking for work. If that person stops looking, they disappear from the unemployment calculation. As a result, the labor market can appear stronger than it actually is. This dynamic becomes especially important late in the cycle, when structural weaknesses begin to emerge beneath stable or even improving headline indicators.

For investors and policymakers, relying too heavily on the unemployment rate in this environment can lead to incorrect conclusions about the strength of the economy.

The Real Driver: Labor Income and Consumption

For macro analysis, the key question is not only whether the unemployment rate is rising or falling. The key question is whether labor income is expanding enough to support consumption.

The U.S. economy is driven by consumption, and consumption depends heavily on employment, wages, hours worked, and participation. If fewer people are working, or if fewer people are even participating in the labor market, that becomes a direct signal about future demand.

A weakening labor market does not immediately translate into a recession, but it gradually erodes the foundation of consumption. Over time, this can lead to slower economic growth and increased vulnerability to shocks.

The full-time employment data reinforces the same message. Full-time workers declined by 514,000 in June. This is important because full-time employment is usually a better measure of stable income generation than headline payrolls alone.

A person losing a full-time job and replacing it with part-time work may still appear employed, but from an economic perspective, that household is likely weaker. Income becomes less predictable, spending becomes more cautious, and financial resilience declines.

This shift from full-time to part-time employment is often an early sign of labor market deterioration that is not immediately visible in headline figures. Full-time jobs support credit creation, housing demand, durable goods consumption, and overall spending confidence. They are the backbone of economic stability.

When full-time employment begins to deteriorate, the economy becomes more fragile, even if the headline unemployment rate does not immediately reflect it. This is because full-time employment is closely tied to long-term financial commitments such as mortgages, car loans, and consumer credit.

A decline in full-time employment can therefore have a disproportionate impact on economic activity, amplifying the effects of a broader slowdown.

A Labor Market Losing Momentum

The June report fits into a broader pattern that has been developing for months. Hiring is slowing, participation is falling, full-time employment is weakening, and previous payroll numbers are being revised lower.

None of these signals points to a strong expansion. Instead, they point to a labor market that is gradually losing its ability to support consumption and sustain economic growth. This type of slowdown is often subtle at first, but it can become more pronounced over time if the underlying trends are not reversed.

This matters because the economy does not consume the unemployment rate. The economy consumes income. If income growth weakens because fewer people are working full-time or because more people are leaving the labor force, then consumption eventually weakens as well. And if consumption weakens, the broader economy follows.

This is why focusing on income-related indicators provides a more accurate picture of economic health than relying solely on the unemployment rate.

The main takeaway from June is not that payrolls were only 57,000. The main takeaway is that the unemployment rate fell for the wrong reason. A falling unemployment rate caused by lower participation is not a strength. It is a warning sign.

If this trend continues, the labor market may continue to look better on the surface than it really is underneath. That is exactly why investors should stop focusing only on the unemployment rate and start paying closer attention to participation, full-time employment, and household employment.

Those are the numbers that truly tell us whether the economy is still expanding or whether the labor market is quietly weakening beneath the headline.

Whether the cycle lasts exactly 18.6 years is ultimately beside the point. Markets do not operate according to calendars. What makes this framework valuable is not the precise timing, but the economic forces that repeatedly produce these long expansions and painful corrections.

Housing is unlike almost every other asset because it is overwhelmingly financed with debt. Every home purchase creates a loan, every new mortgage expands the banking system’s balance sheet, and every increase in lending creates additional purchasing power. Unlike stocks, where one investor simply buys shares from another, housing transactions often generate entirely new credit. That credit becomes new spending power, which supports construction, employment, consumption, and ultimately economic growth.

The cycle, therefore, begins with credit rather than with house prices themselves.

As banks become more willing to lend, financing becomes easier, more buyers enter the market, and prices begin to rise. Those higher prices increase the value of the collateral backing existing homes, allowing banks to extend even more credit. Rising collateral values encourage larger mortgages, larger mortgages generate stronger demand, and stronger demand pushes prices even higher.

This feedback loop can continue for years. Eventually, however, something changes.

Share Signals Over Noise

Debt Spiral

Eventually, debt begins growing faster than incomes. Affordability deteriorates. Monthly mortgage payments become increasingly difficult to sustain. Credit standards tighten. Demand slows. The same mechanism that fueled the expansion now begins working in reverse, reducing credit creation and removing one of the largest sources of demand in the economy.

This is why housing should not simply be viewed as another sector. It is one of the primary engines of the entire economic cycle.

Another important consequence of rising home prices is the wealth effect. As homeowners watch the value of their properties appreciate year after year, they naturally become more confident about their financial position. They spend more, borrow more against home equity, renovate their properties, purchase durable goods, and generally consume more than they otherwise would have. Rising housing wealth, therefore, stimulates aggregate demand far beyond the housing market itself.

When the cycle reverses, that process reverses as well. Falling home prices reduce household wealth, borrowing slows, consumption weakens, and credit creation contracts. What initially appears to be a housing correction gradually spreads throughout the broader economy.

This is precisely why housing is also one of the most important drivers of inflation. Much of today’s inflation debate continues to revolve around oil prices, tariffs, geopolitical conflicts, or supply chain disruptions. These factors certainly matter in the short run, but they rarely create lasting inflation by themselves. Oil prices can temporarily push headline inflation higher, just as wars or tariffs can increase the cost of specific goods. However, these shocks generally fade once supply conditions normalize.

Housing is fundamentally different.

Shelter represents roughly one-third of the Consumer Price Index, making it by far the single largest component of measured inflation. More importantly, housing inflation tends to move slowly and persistently. Rental contracts adjust gradually, home prices feed into rents with a lag, and mortgage markets evolve over years rather than months. When housing enters a sustained boom, inflation becomes much more difficult to bring back down because its largest component remains elevated for an extended period.

This is one of the reasons why I continue to believe that structural inflation almost always requires a strong housing market.

Without expanding mortgage credit, rising home prices, and accelerating rents, it becomes extremely difficult for inflation to remain persistently high. Temporary oil shocks can create headlines, but they rarely change the long term direction of inflation unless housing and credit are reinforcing the move.

Share

How the Cycle Works: 14 Up, 4 Down

Once you understand the structure of the cycle, you can’t unsee it. Every recession, every housing boom, every “unexpected” financial crisis snaps into place. The cycle has five distinct phases, and each one has a characteristic feel — a psychology, a set of signals — that repeats with remarkable consistency.

1. Recovery (years 1–7): Land prices begin recovering from the trough. Credit loosens. Construction picks up. Confidence returns slowly. This is the phase where the best long-term investment opportunities typically appear — though most people are still too traumatized by the prior bust to act.

2. Mid-cycle peak and recession (around year 7): A brief slowdown — a mini-recession — occurs roughly halfway through the upswing. Stock markets correct. Then the recovery resumes. This is often mistaken for the end of the cycle, leading investors to sell too early.

3. Winners’ Curse phase (years 12–14): The final speculative surge. Everyone is “all-in.” Land prices rise fastest. Leverage is highest. Volatility increases. This is the most dangerous phase for investors who mistake momentum for safety. The psychology of this phase is powerful — FOMO is at its peak, and the very people who should be reducing exposure are often increasing it.

4. Cycle peak: Land prices top out. Rising interest rates become unsustainable — but they are the effect, not the cause. The cause is land values too high relative to productive capacity. The peak is often invisible in real-time because confidence is at its maximum.

5. 4-year downturn: Credit contraction. Land values fall. Bankruptcies cascade through the banking system. Stock markets typically crash 1–2 years after the land peak, not simultaneously. The severity of the downturn depends on how much leverage has built up during the boom.

Historical Cycle Record

What convinced me that this wasn’t theory was the historical record. Nearly 200 years of data across wildly different political environments, monetary systems, and technological eras — and the same pattern keeps showing up. The cycle is not a product of capitalism, democracy, gold standards, or fiat money. It is a product of how land is owned and taxed.

The Stock Market and the Land Cycle: Two Distinct Events

Interestingly, housing cycles have also tended to lead broader financial markets. In previous cycles, residential real estate frequently began slowing before equities reached their final peak. Credit conditions weakened first, housing activity softened, and only later did stock markets experience their most significant corrections. This sequence reflects the central role housing plays within the broader credit cycle.

That observation fits remarkably well with the environment we see today.

Housing has already weakened considerably, yet U.S. equities continue trading near historically extreme valuations. Investor optimism surrounding artificial intelligence, future productivity gains, and corporate earnings remains exceptionally high, even as several macroeconomic indicators continue pointing toward slower growth.

This divergence does not necessarily mean equities must decline immediately. Late-cycle environments often include a final period of optimism before financial conditions fully tighten. History shows that speculative enthusiasm can persist long after the underlying economy has begun slowing.

This is precisely why I do not view the 18.6 year cycle as a timing indicator. It does not tell us the exact month or year when markets will reverse. Instead, it provides context. It reminds us that housing, credit, inflation, and financial markets have always been deeply interconnected. Every major housing boom eventually becomes a credit story. Every credit story eventually becomes an inflation story. And every inflation story ultimately influences interest rates, monetary policy, and equity valuations.

The cycle-by-cycle evidence for this lag

It's also worth distinguishing end-of-cycle crashes from mid-cycle corrections. End-of-cycle crashes — where the full land cycle terminates — produce the most severe and sustained damage: 1929–32 (−89% Dow), 2007–09 (−57% S&P 500). But the cycle shows that there is usually a mid-cycle pullback in real estate markets, which in a similar fashion leads to an unwind in other speculative bubbles occurring at the time, such as the dot-com bust of 2001–2002 (−49% S&P 500).

Whether the next turning point arrives exactly on schedule is impossible to know. What is becoming increasingly clear, however, is that many of the underlying conditions that have historically marked the later stages of previous cycles are already visible today. Housing is slowing, shelter inflation is easing, credit growth has moderated, and financial assets remain priced for extraordinary optimism.

The calendar may never repeat perfectly. The mechanisms behind these cycles, however, have proven remarkably consistent.

However, the reality is much less dramatic.

The inflation measures that investors, economists, and policymakers have relied on for decades are not disappearing. Headline CPI will continue to be published. Core CPI will continue to exist. The Personal Consumption Expenditures index, commonly known as PCE, will remain available. Americans will still be able to see exactly how much prices are changing across the economy.

What may change is something much more subtle. The Federal Reserve may choose to place greater emphasis on a different inflation metric when making policy decisions.

That distinction is critical because changing the lens through which inflation is interpreted is not the same as changing inflation itself.

Share

Understanding the Different Inflation Measures

The most familiar measure is Headline CPI. This is the number most people recognize because it includes everything consumers actually purchase. Food, energy, housing, transportation, medical services, recreation, and dozens of other categories all contribute to the final figure.

The strength of Headline CPI is that it reflects the inflation consumers experience in everyday life. The weakness is that it can be extremely noisy. A drought can temporarily push food prices higher. A war in the Middle East can send oil prices soaring. Floods, tariffs, trade disruptions, or geopolitical tensions can all produce short-term spikes that may have little to do with the underlying direction of the economy.

This is precisely why economists developed Core CPI. Core inflation excludes food and energy, not because those categories are unimportant, but because they tend to be heavily influenced by temporary supply shocks. The objective is to identify whether inflationary pressures are becoming embedded across the broader economy rather than reacting to events that may reverse quickly.

The Federal Reserve has historically preferred another measure entirely, the Personal Consumption Expenditures index. PCE differs from CPI because it adjusts the weights assigned to different categories more frequently and better captures the tendency of consumers to substitute toward cheaper alternatives when relative prices change. It also places less emphasis on housing than CPI, which is why it often reports slightly lower inflation readings.

Now another metric has entered the public debate: Trimmed Mean Inflation.

Unlike Core CPI, which excludes the same categories every month, Trimmed Mean inflation removes the largest monthly price increases and the largest monthly declines regardless of where they occur. If airline tickets suddenly jump forty percent, those observations may be excluded. If gasoline prices collapse sharply, they may also be excluded. If eggs surge because of a temporary supply shortage, those extreme movements may not influence the final number.

The objective is straightforward. Reduce noise and identify the underlying trend. Importantly, this is not some newly invented measure designed to lower inflation. Variations of Trimmed Mean inflation have existed for years and have been studied extensively by regional Federal Reserve banks and academic economists.

The metric itself is not new. The only question is whether policymakers decide it deserves greater weight.

Why Is The Fed Looking At It Now?

The answer may simply be that the economy is slowing.

Over the last several years, inflation has repeatedly been influenced by supply-side disturbances. We have experienced post-pandemic disruptions, wars, tariffs, commodity volatility, and periodic energy shocks that have distorted short-term inflation readings.

At the same time, many of the cyclical indicators that tend to lead inflation have continued to weaken. Credit growth has slowed. Housing inflation is drastically coming down. The labor market has gradually softened.

Several leading indicators continue pointing toward slower economic activity.

If policymakers believe these external shocks are temporary while the broader economy is naturally disinflating, it becomes understandable why they would pay closer attention to measures designed to smooth volatility.

This does not mean they are saying inflation is lower than it really is. It means they may be saying that the inflation relevant for monetary policy is lower than what temporary shocks imply.

There is a significant difference between the two.

What If The Fed Is Right?

The more interesting question is not whether the Fed is manipulating inflation. The more interesting question is whether the cycle itself is delivering the disinflation that many investors still refuse to acknowledge.

For months, my thesis has remained remarkably consistent. Housing inflation, which accounts for roughly one-third of CPI, continues to decline. The housing market itself remains under pressure from elevated mortgage rates, rising inventories, and deteriorating affordability.

Credit creation has slowed significantly compared to the post-pandemic period. Consumers are becoming more sensitive to price increases. Businesses face greater pressure on margins.

These are not the conditions typically associated with the beginning of a new structural inflation regime; they are the conditions often observed during economic slowdowns. If this interpretation is correct, then measures such as Trimmed Mean inflation may simply be capturing what the broader cycle has been signalling all along.

Inflation is slowing because demand is slowing. Not because the Fed has changed the formula.

Share Signals Over Noise

What The Numbers Are Actually Saying

Headline PCE inflation, the Fed’s official target measure, has continued to move lower from its post-pandemic peaks and has recently been running around 3.8% due to the oil shock on a year-over-year basis. Core PCE, which excludes food and energy, is also trending in the range 3%-3.5% over the last 12 months.

Trimmed Mean PCE, meanwhile, has remained slightly above the Federal Reserve’s 2% target, generally fluctuating in the low-to-mid 2% area.

For critics of the measure, this persistence is often presented as evidence that Trimmed Mean inflation understates the problem. In reality, many economists interpret it in exactly the opposite way.

Because Trimmed Mean inflation systematically removes the most extreme monthly price movements, it tends to capture the broad-based trend affecting the majority of goods and services in the economy. It does not react as aggressively to temporary spikes in gasoline prices, sudden collapses in airfare, or isolated shortages affecting a handful of products.

The fact that Trimmed Mean inflation has remained modestly above 2% suggests that underlying inflation pressures have not completely disappeared, but neither are they accelerating in a way consistent with a new inflationary spiral. Instead, it points toward an economy that is gradually normalizing after an extraordinary period of disruption.

Perhaps most importantly, if the United States were truly entering a new structural inflation regime, all of these measures would eventually move higher together. The chart below shows that during the post-pandemic period, inflation rose to 7% year over year under the Trimmed Mean PCE metric. This means that in a new inflationary regime, Headline inflation, Core measures, PCE, and Trimmed Mean inflation would all rise.

No statistical technique can permanently hide a broad-based inflation problem. If inflation becomes embedded throughout the economy, the signal eventually overwhelms the noise regardless of how the data are filtered.

That is precisely why using measures such as Trimmed Mean inflation to guide monetary policy is not inherently problematic. Policymakers are not attempting to redefine inflation. They are attempting to distinguish between temporary disturbances and persistent trends. If they are wrong, the data will reveal it soon enough.

The challenge is that not all valuation metrics are created equal.

Over the years, investors have developed dozens of ways to measure whether a market is expensive or cheap. Some focus on earnings, others on assets, sales, cash flows, or economic output. Each attempts to answer the same question: how much are investors paying relative to what they are actually receiving?

The answer, however, depends heavily on which metric is being used.

Buffett Indicator

One of the most frequently cited valuation measures is the Buffett Indicator, which compares the total market capitalization of U.S. equities to GDP. The idea is intuitive. If the stock market becomes significantly larger than the economy itself, valuations may be stretched.

The problem is that the economy today looks very different from the economy that existed when this indicator became popular decades ago.

The United States has increasingly evolved into a service-driven and globally integrated economy. Many of the largest companies in the world generate substantial portions of their revenues and profits outside the United States. These earnings contribute directly to their market valuations, but they do not necessarily contribute to U.S. GDP. As a result, the relationship between domestic economic output and market capitalization has become less straightforward than it once was.

This does not mean the Buffett Indicator is useless. It simply means that its limitations are larger today than they were thirty or forty years ago. The signal still matters, but the interpretation requires more caution.

Share Signals Over Noise

Other Metrics

Other metrics, such as Price to Book and Price to Sales, attempt to solve different problems. Price to Book focuses on the relationship between market value and net assets, while Price to Sales looks at valuations relative to revenue generation. Both can provide useful information, particularly when comparing companies within the same industry. However, neither fully captures the economic cycle or the sustainability of profits.

The metric that receives perhaps the most attention on Wall Street today is the forward price-to-earnings ratio.

In my view, this is also the least useful valuation measure. The reason is simple. Forward earnings are not earnings. They are forecasts.

The forward P/E ratio compares today’s stock price with expected earnings over the next twelve months. The denominator is not based on what companies have earned, but on what analysts expect them to earn.

This distinction is critical.

When economic conditions are strong and investor optimism is high, analysts tend to increase their earnings forecasts. As those projections rise, the forward P/E ratio appears lower and more attractive, even as prices continue moving higher. In effect, the metric often reflects optimism rather than valuation.

The reverse occurs during bear markets. Analysts become pessimistic, earnings expectations fall, and the forward P/E ratio suddenly appears expensive despite the fact that stock prices may have already declined significantly.

In this sense, the forward P/E ratio functions less as a valuation tool and more as a sentiment indicator. It tends to look most attractive near major peaks and least attractive near major bottoms. That is precisely the opposite of what investors should want from a valuation framework.

Share

The Preferred One

This is why I continue to prefer the Shiller CAPE ratio as the most useful long-term valuation metric.

Unlike forward P/E, CAPE relies on actual earnings rather than forecasts. More importantly, it uses inflation-adjusted earnings averaged over the previous ten years. This solves one of the biggest weaknesses of the traditional price-to-earnings ratio, namely that earnings can fluctuate dramatically from one quarter to the next due to recessions, temporary shocks, accounting adjustments, or unusually strong periods of profitability.

By averaging earnings over an entire decade, CAPE smooths these distortions and provides a more stable picture of valuation.

No valuation metric is perfect, and CAPE is certainly not a timing tool. Markets can remain overvalued for years. Bubbles can become even larger before they burst. But if the objective is to understand where valuations stand relative to history, CAPE has consistently proven to be one of the most informative measures available.

And today, it is sending a very clear message. Current CAPE readings sit remarkably close to the levels reached during the peak of the dot-com bubble in 2000. While we have not yet exceeded those extremes, we are much closer than most investors appreciate.

In fact, a move of roughly 10 percent in the S&P 500 would likely be enough to push valuations back toward the same territory that defined one of the most speculative periods in financial history.

Markets are currently exhibiting many of the characteristics typically associated with late-cycle behaviour. Optimism remains elevated, investors continue extrapolating strong earnings growth into the future, and every decline is viewed as a buying opportunity. The combination of artificial intelligence enthusiasm, improving geopolitical sentiment, and expectations for easier monetary policy has created an environment where risk appetite remains extraordinarily strong.

Under these conditions, a final euphoric phase is entirely possible.

This is why valuation should never be used as a timing mechanism. The market can become significantly more expensive before reality eventually catches up. Valuation tells us very little about what will happen next week or next month.

What it does tell us is how unusual the current environment has become. Years from now, when investors look back at this period, they may debate the exact catalyst that eventually ended the cycle. They may discuss interest rates, earnings, liquidity, geopolitics, or any number of other factors. But one thing is unlikely to be debated. The evidence was visible.

Valuations were stretched. Expectations were elevated. Risk premiums were compressed. Investors simply chose to focus on the narrative rather than the numbers.

Whether the market peaks tomorrow or years from now is impossible to know. But history suggests that when valuations approach the extremes of 2000, future long-term returns tend to become increasingly constrained.

That is not a prediction. It is simply what the data has shown, repeatedly, across more than a century of market history.

This decline matters because oil sits at the center of the global macroeconomic system. It is not simply another commodity whose price fluctuates independently. Oil feeds directly and indirectly into transportation, logistics, industrial production, and consumer prices. When oil prices spike due to geopolitical tensions, markets immediately begin pricing in higher inflation expectations. But when those tensions ease, and oil begins to fall, the process works in reverse.

Understanding this mechanism is critical because it helps investors see how geopolitical events ultimately filter through to bonds, equities, and broader financial conditions.

If an agreement is eventually confirmed, one of the most important consequences would be the normalization of energy flows through the Strait of Hormuz. That matters because the Strait represents one of the most important oil transit routes in the world. If markets become confident that oil supply disruptions are no longer a risk, the geopolitical premium embedded in crude prices begins to disappear.

As oil prices decline, inflation expectations also begin to ease.

This is where the bond market comes into play.

Why Falling Oil Prices Push Bond Prices Higher

Oil has a powerful effect on inflation expectations because higher energy costs are passed through to the rest of the economy. Transportation costs rise, production costs increase, and businesses begin passing some of those costs on to consumers. Even categories that appear unrelated to energy are indirectly affected because the modern economy depends heavily on transportation and logistics.

As inflation expectations rise, investors demand higher yields from bonds in order to compensate for the expected erosion in purchasing power. This is why periods of rising oil prices are often associated with rising interest rates and falling bond prices.

But if oil prices begin to decline because geopolitical tensions ease, the opposite happens. Lower oil prices reduce inflation pressure, which reduces the need for higher yields. As yields fall, bond prices rise. This effect becomes especially pronounced in long duration bonds, which are significantly more sensitive to changes in interest rates.

That is why instruments such as TLT (an ETF that buys 20+ year duration bonds) tend to rally when long-term yields decline. The longer the duration of the bond, the more sensitive it becomes to changes in rates, making these assets particularly reactive to shifts in inflation expectations.

This is one of the most important relationships in macroeconomics. Oil influences inflation expectations. Inflation expectations influence interest rates. Interest rates influence bond prices.

And once interest rates begin to move lower, the effects do not stop with bonds.

Share Signals Over Noise

How Lower Rates Support Equities

The equity market is closely tied to interest rates, even though many investors focus only on earnings, headlines, or technical indicators. At the core of equity valuation is the process of discounting future cash flows.

When investors value a company, they estimate the future stream of profits or cash flows the business is expected to generate over time. Those future cash flows are then discounted back to the present using an interest rate, typically anchored to the so-called risk-free rate, which is generally based on Treasury yields.

This is where the concept of the equity risk premium becomes important.

Equities are inherently riskier than government bonds. Because of that, investors demand additional expected returns in exchange for holding stocks rather than simply buying risk free Treasuries. If Treasury yields rise sharply, bonds become increasingly attractive relative to equities because investors can earn higher returns without taking equity risk.

As a result, higher interest rates tend to compress equity valuations.

But if oil prices decline, inflation expectations ease, and interest rates begin to fall, bonds become less attractive from a yield perspective. This lowers the discount rate used in equity valuation models and simultaneously pushes investors toward riskier assets in search of higher returns.

In other words, falling rates tend to support higher equity valuations. This is why markets often react positively to falling yields, even when those lower yields are initially caused by macroeconomic weakness. The valuation effect can dominate in the short term, especially in sectors with long-duration cash flows such as technology.

However, this dynamic must be understood within the broader context of current valuations. The market is already extremely expensive by historical standards. The Shiller CAPE ratio remains near 42, one of the highest readings ever recorded and second only to the peak of the dot com bubble.

At the same time, measures such as the Buffett Indicator continue to suggest that U.S. equities remain significantly detached from underlying economic output.

That context matters because lower rates can justify higher valuations only up to a point. If valuations are already stretched, falling rates may fuel another euphoric phase, but they also increase the fragility of the system.

Subscribe now

Understanding Markets Through Macroeconomics

One of the most important lessons for investors is that markets are rarely driven by isolated headlines. To understand where bonds or equities may move next, it is often necessary to step back and analyze the broader macroeconomic chain of events underlying the market.

Many investors believe that predicting the stock market requires focusing directly on stocks themselves. In reality, understanding equities often begins far away from the equity market. It begins with understanding the economy, inflation, interest rates, liquidity, and, in this case, the geopolitical forces shaping oil prices.

By analyzing the relationship between oil, inflation expectations, interest rates, bonds, and equities, investors can better understand why markets react the way they do and why seemingly unrelated events can produce powerful financial consequences.

The market is not reacting only to headlines about a potential agreement between the United States and Iran. It is reacting to what that agreement could mean for oil, inflation, interest rates, and ultimately the entire financial system.

For centuries, gold functioned as money. It was used directly as a medium of exchange, a store of value, and a unit of account. Transactions were settled in gold or in currencies directly backed by it. Under various forms of the gold standard, governments and central banks held gold reserves while issuing paper currency redeemable for a fixed amount of gold. In that sense, gold was not simply another commodity. It was the monetary foundation of the system itself.

This historical role is important because it explains why many investors still instinctively view gold as “real money.” If gold was once the ultimate monetary asset, then comparing financial assets against gold can appear, at least conceptually, like a way of measuring real wealth rather than nominal wealth.

But the monetary system no longer functions this way.

For more than a century, gold has not been used as a practical medium of exchange in everyday economic activity. People do not buy homes, pay salaries, settle invoices, or conduct business transactions in gold. Modern economies operate through fiat currencies, primarily the U.S. dollar and other sovereign currencies. These are the units in which contracts are written, taxes are paid, debts are denominated, and investments are measured.

That distinction changes the logic behind how investment performance should actually be evaluated.

The Problem With the S&P to Gold Ratio

One of the most common arguments made today is that the S&P 500 has not truly appreciated in “real terms” because, when measured against gold, it has significantly underperformed since the peak in 2000. According to this framework, the S&P to gold ratio peaked during the dot-com bubble and has been in a long-term decline ever since.

The implication is that investors have not actually created wealth over the last two decades, but have merely experienced nominal gains caused by currency debasement. At first glance, the argument sounds sophisticated. But the underlying logic is deeply flawed.

The only coherent way to measure the real return of an investment is by adjusting it for the inflation of the currency used to purchase that investment in the first place. If an investor buys the S&P 500 using U.S. dollars, then the relevant question is whether the purchasing power of those dollars increased after accounting for inflation. That is the actual economic reality faced by the investor.

The investor is not purchasing the S&P with gold.

This point matters because investment returns should be evaluated relative to the monetary unit through which economic transactions occur. What ultimately matters is whether the investment increased the investor’s purchasing power within the economy where they consume goods and services. Since those transactions occur in dollars, not gold, the relevant comparison is against dollar inflation.

Using gold as the denominator introduces an arbitrary benchmark that has no direct relationship with how modern economic activity functions. And once that door is opened, the framework quickly becomes inconsistent.

Why use gold instead of silver? Why not Bitcoin? Why not palladium, coffee, copper, or any other asset that can also fluctuate relative to equities? None of these assets is used as the primary medium of exchange in the economy. None is the actual monetary unit used to buy the S&P 500 in the first place.

The selection of gold is not based on economic necessity, but on historical nostalgia.

Share

The Contradiction Inside the Ratio

The deeper issue with the S&P to gold ratio is that it often produces conclusions that are completely disconnected from observable economic reality.

Since 2022, for example, the ratio has suggested that the S&P 500 has effectively been in decline relative to gold. But during this same period, the U.S. economy has continued to expand. Corporate earnings have grown, productivity has improved, and many sectors have experienced significant increases in profitability. Margins have remained elevated, technological investment has accelerated, and economic activity has clearly not resembled a stagnant or collapsing environment.

Yet according to the ratio, almost none of that growth “counts.”

Taking the longer term perspective makes the contradiction even larger. If one looks at the S&P relative to gold since the year 2000, the ratio suggests that there has been little to no real wealth creation for over two decades. That conclusion is difficult to reconcile with the reality of what actually occurred during that period.

Over the last twenty-five years, the global economy has experienced enormous technological progress, the rise of entirely new industries, major increases in productivity, massive growth in corporate earnings, and the creation of some of the largest and most profitable companies in history. To argue that none of this generated meaningful real growth simply because gold outperformed equities stretches the framework beyond credibility.

The ratio effectively assumes that gold is the ultimate and permanent measure of value, even though gold itself is a volatile financial asset whose price fluctuates based on speculation, liquidity conditions, central bank demand, and investor psychology.

Share

When the Ratio Becomes Absurd

The final inconsistency appears during recessions and financial crises. In periods of severe economic stress, both equities and gold can decline simultaneously due to liquidity needs and forced selling. Investors often sell whatever they can in order to raise cash. But if the S&P 500 falls more sharply than gold, the ratio itself rises.

Under this framework, the investor would appear to be experiencing “real gains” precisely during a recessionary collapse. That creates a completely illogical result.

Imagine an environment where unemployment is rising, consumption is collapsing, credit markets are freezing, and economic activity is deteriorating sharply. Under the S&P to gold framework, investors could still appear richer in “real terms” simply because gold declined less than equities.

But how can real wealth be increasing in an environment where the economy itself is weakening dramatically?

This exposes the central flaw of the ratio. It confuses relative price movements between two volatile financial assets with actual purchasing power and economic prosperity.

Real wealth is not determined by whether one speculative asset outperforms another. It is determined by productive capacity, income generation, corporate profitability, innovation, and the purchasing power investors retain within the economy where they actually live and transact.

Gold can absolutely be a useful hedge, a reserve asset, or a store of value under certain conditions. But using it as the sole denominator for measuring all investment performance yields conclusions that are often economically incoherent.

The fact that the S&P-to-gold ratio has declined since 2000 does not mean the economy failed to grow. It means gold outperformed equities over that particular period. Those are not the same thing.

When oil prices rise sharply, they tend to push headline inflation higher. This increase feeds directly into inflation expectations, which in turn affects bond markets. Investors demand higher yields to compensate for the expected loss of purchasing power, particularly at the front end of the curve. As yields rise, bond prices fall, and financial conditions begin to tighten. This is the initial phase of an oil shock, where inflation fears dominate and interest rates move higher.

However, this is only the first part of the cycle.

The second phase, which tends to be overlooked, is the economic impact of higher energy costs. Oil is a largely inelastic necessity. When prices rise, households and businesses cannot simply stop consuming it. Instead, they reallocate spending away from other areas of the economy in order to absorb the higher cost. This shift reduces discretionary consumption, compresses margins, and ultimately weakens economic activity. Over time, the same shock that initially pushed inflation higher begins to act as a drag on growth.

And this is where the direction of interest rates reverses.

Cycles in Practice: Three Historical Episodes

Looking at past cycles, the pattern is remarkably consistent. Oil shocks tend to produce an initial spike in interest rates, followed by a much larger decline as the economic damage becomes evident.

During the Gulf War, oil prices surged approximately 155% in 1990. This sharp increase led to a rise in inflation expectations and a corresponding short-term rise in interest rates. However, as the higher energy costs began to weigh on economic activity, the effect reversed. Growth slowed, demand weakened, and interest rates entered a sustained downward trend.

A similar dynamic unfolded during the dot-com cycle, culminating in the Dot-com Bubble. Oil prices increased by roughly 247% leading into the peak, an even larger move than during the Gulf War. Once again, the immediate impact was higher inflation expectations and rising yields. But the longer-term outcome was the same. From 2000 to 2003, interest rates declined significantly as the economy weakened and the equity bubble unwound.

The Global Financial Crisis provides perhaps the clearest example. Between 2007 and mid 2008, oil prices surged by approximately 176%. This drove a temporary increase in inflation and pushed interest rates higher in the short term. However, the economic consequences were severe. Higher energy costs compounded an already fragile system, accelerating the downturn. As demand collapsed, interest rates followed, falling sharply as the crisis unfolded.

Across all three episodes, the sequence is identical. Oil shocks push rates higher initially through inflation expectations, but ultimately lead to lower rates as they damage the underlying economy.

Share

Today: The Same Setup, Different Cycle Position

The current environment (See Image above for 2026) appears to be following the same script. Since the beginning of the latest geopolitical tensions, oil prices have increased by approximately 117%. This has already translated into higher inflation expectations and a rise in interest rates, particularly at the long end of the curve. The 10-year Treasury yield, for example, has risen to around 4.41 percent after being below 4 percent prior to the shock.

At first glance, this may seem like the beginning of a structurally higher rate environment. But history suggests otherwise.

What matters is not just the magnitude of the oil shock, but the condition of the economy when it occurs. In 2022, when oil prices also surged following the conflict between Russia and Ukraine, the economic backdrop was very different. Households still held elevated savings from the post-pandemic period, and the credit cycle was in a much earlier stage. This allowed the economy to absorb higher energy costs more effectively, at least temporarily.

Today, the situation is far more fragile. Savings rates are lower, credit conditions are tighter, and the economy is much closer to the end of the cycle. In this context, the ability to absorb higher oil prices is significantly reduced. The same mechanism that played out in previous cycles becomes more powerful, as the burden of higher energy costs forces a sharper adjustment in consumption and investment.

If the historical pattern holds, the current rise in interest rates should not be interpreted as a lasting shift upward, but rather as the first phase of the same cycle we have seen before. The initial inflation shock pushes yields higher, but the subsequent economic slowdown drives them lower.

In that sense, oil shocks do not create sustained inflationary regimes. They accelerate the cycle. And at this stage of the cycle, that acceleration tends to end the same way, with weaker growth, tighter conditions, and ultimately, lower interest rates.

This level of attention would make sense if the Fed directly controlled the most important interest rates in the economy. But it does not.

The Federal Reserve controls only one part of the interest rate structure, and it is not the part that ultimately drives real economic activity. What the Fed sets is the short-term price of money, specifically the federal funds rate, which influences borrowing costs at very short maturities, typically ranging from overnight lending to roughly one year. These are the rates at which banks lend to each other, manage reserves, and interact with the Fed itself.

This segment of the curve matters for the plumbing of the financial system. It affects liquidity conditions, short-term funding markets, and the cost of leverage for financial institutions. But it is not the rate that determines whether a household decides to buy a car, whether a family can afford a home, or whether a small business chooses to invest and expand.

Those decisions depend on long-term interest rates. If someone wants to finance a car, the relevant rate is closer to five years. If they want to take out a mortgage, the relevant rate is twenty or thirty years. If a business wants to borrow to fund operations or invest in growth, the cost of capital is tied to longer duration yields, credit spreads, and overall financial conditions, not the overnight rate set by the Fed.

This distinction is critical, yet often ignored. There is a persistent tendency to treat the Fed’s policy rate as if it were the central driver of credit creation in the economy. In reality, it is only one piece of a much larger system. Credit expands when long-term borrowing costs decline, when financial conditions ease, and when lenders are willing to extend capital. None of these are directly controlled by the Fed in a mechanical way.

Share

The 2-Year Yield

Even more importantly, the Fed itself is not acting independently of the market in the way many assume. Its decisions are not made in a vacuum. In practice, the Fed tends to follow the signal coming from the bond market, particularly the two-year Treasury yield.

The two-year yield is one of the most informative indicators of where monetary policy is headed because it reflects expectations about growth, inflation, and future policy. When the two-year yield begins to decline sharply, it signals that the market expects weaker economic conditions and eventual rate cuts. When it rises, it reflects expectations of stronger growth or higher inflation, and therefore tighter policy.

Historically, the relationship between the Fed and the two-year yield has been remarkably consistent.

In 2000, the two-year yield began to fall well before the Fed started cutting rates. As it moved decisively below the Fed’s policy rate, it signaled a shift in economic conditions that the Fed eventually had to acknowledge. Rate cuts followed shortly after.

A similar pattern occurred in 2007. By the second half of that year, the two-year yield had already dropped significantly below the Fed’s policy rate, reflecting growing stress in the system. The Fed responded with aggressive cuts, but only after the market had already moved.

The same dynamic repeated in 2019, when the two-year yield rolled over and declined below the Fed’s rate, leading to a new easing cycle. More recently, toward the end of 2024 and into 2025, the two-year yield again moved sharply below the policy rate, and the Fed followed with rate cuts.

The pattern also works in reverse. In late 2021, as inflation accelerated, the two-year yield began to rise rapidly, anticipating tighter monetary policy. The Fed did not lead this move. It followed it, raising rates after the market had already repriced expectations.

This sequence highlights a key point. The Fed does not lead the market. It reacts to it.

Share

Expected Outcome

This is why the outcome of each Fed meeting should rarely be surprising. The direction of policy is already embedded in the bond market, particularly in the two-year yield, long before the official decision is announced. The Fed is effectively validating what the market has already priced in.

At the same time, even this dynamic should not distract from the broader issue.

The policy rate itself is not the most important variable for the real economy. What ultimately matters is whether financial conditions are tightening or easing, whether long-term yields are rising or falling, and whether credit is expanding or contracting.

Focusing exclusively on the Fed risks missing what actually drives economic activity. The market’s obsession with every Fed decision creates the illusion that monetary policy is the central force shaping the cycle. In reality, the Fed is responding to the same underlying conditions as everyone else, and the rate it controls is only one piece of a much larger and more complex system.

Understanding that distinction is essential. Because once you see it, the question shifts from what the Fed will do next to what the market is already telling you.

The argument is straightforward and, at first glance, compelling. U.S. equities, measured through indicators such as the Shiller CAPE ratio, are trading near levels last seen during the dot-com bubble (see image 2), while emerging markets, energy, and metals appear deeply discounted by comparison.

From a long-term perspective, this argument is not entirely wrong. There is a strong case to be made that over the next decade, assets tied to commodities, emerging economies, and regions like Latin America could outperform the United States. Cycles exist, and periods of extreme overvaluation in one region often coincide with undervaluation in another. Capital eventually rotates, and when it does, the moves can be powerful and sustained.

The problem is not the direction of the idea. The problem is the timing.

The Rotation

What is often missing from this narrative is the assumption that this rotation happens smoothly, or even immediately. The implicit belief is that investors can simply reduce exposure to U.S. equities and reallocate toward emerging markets, benefiting from both the relative valuation gap and the expected outperformance. History suggests that this is not how these transitions occur.

To understand this dynamic, it is useful to revisit the last time we saw a similar setup. In early 2000, at the peak of the dot com bubble, valuation spreads between the United States and emerging markets were also extreme. U.S. equities were trading at historically elevated multiples, while many emerging markets, including Latin America, appeared cheap. On paper, the opportunity looked obvious.

Share

But what followed was very different from what many would expect today. When the S&P 500 and the Nasdaq peaked in March 2000, capital did not immediately rotate into emerging markets. Instead, the first phase of the adjustment was driven by risk aversion. The U.S. dollar strengthened (See chart 2 in the image above), U.S. Treasuries performed well, and global investors moved toward safety and liquidity.

Emerging markets, far from benefiting, were hit harder. Brazil, as one example, declined roughly 70 percent in dollar terms between 2000 and late 2002 (See chart 3 in the image above). This was not an isolated case, but part of a broader pattern in which higher-risk assets underperformed significantly during the initial phase of the downturn. The valuation gap did not protect them. In fact, it became irrelevant in the face of tightening financial conditions and a global flight to safety.

Only after that adjustment phase, once the cycle had reset, did the opportunity in emerging markets truly begin to materialize. From the early 2000s onward, Latin America and commodity-driven economies entered a strong bull market, benefiting from a weaker dollar, improving global growth, and a renewed appetite for risk.

Share

The sequence is critical.

There is a structural difference between recognizing a long-term opportunity and capturing it at the right time. The transition from an overvalued U.S. market to an outperforming emerging market cycle does not happen in a single step. It happens in phases, and the first phase is almost always characterized by deleveraging, rising demand for liquidity, and a preference for safety.

This is why the idea of being simultaneously bearish on the United States and bullish on emerging markets in the short term is problematic.

Global markets are deeply interconnected. The United States is not just another market. It is the central node of the global financial system, the primary reserve currency issuer, and the deepest source of liquidity. When U.S. equities decline sharply, it is rarely an isolated event. It typically reflects tightening financial conditions, reduced risk appetite, and a broader contraction in global liquidity.

In that environment, capital does not rotate into higher-risk regions. It retreats. Investors tend to reduce exposure to the most volatile and least liquid assets first. That often includes emerging market equities, Latin American assets, and corporate debt from riskier jurisdictions. These markets, which may look cheap in relative terms, become even cheaper as capital exits in search of safety.

If the United States were to experience a significant correction, whether 30, 40, or even 50 percent, it is highly unlikely that emerging markets would remain stable. History suggests the opposite. They would likely experience deeper drawdowns, not because their fundamentals are worse, but because they carry higher perceived risk and lower liquidity.

Only after that process unfolds, once valuations reset and the global cycle stabilizes, does the opportunity begin to emerge.

This is where the longer-term bullish case for commodities, Latin America, and emerging markets becomes more relevant. A weaker dollar, improving global demand, and a renewed expansion in credit can create the conditions for sustained outperformance. But those conditions typically appear after the downturn, not before it.

In the current environment, it is entirely possible to be structurally bearish on U.S. equities over the long term while also being cautious on emerging markets in the short term. These views are not contradictory. They reflect an understanding that markets move through phases, and that correlations tend to rise during periods of stress.

At the peak of a cycle, everything is connected. And when the adjustment begins, it is rarely the cheapest assets that are spared.

Valuation metrics are supposed to serve a simple purpose: to anchor expectations. At their core, they exist to compare what investors are paying today with what they are actually receiving in return, whether that means earnings, sales, book value, cash flows, or economic output. The best valuation tools are not designed to predict the future with precision, but to impose discipline by showing when prices have moved too far away from what is observable and sustainable.

Historically, the most useful valuation measures have one thing in common: they rely on realized data rather than forecasts. That distinction matters because realized data is tangible, while forecasts are assumptions. The more a valuation metric depends on future growth projections, future margins, or future economic conditions, the more it stops being a measure of value and starts becoming a reflection of sentiment.

This is precisely the problem with the forward P/E ratio.

The Numbers

The forward P/E ratio compares today’s stock price with expected earnings over the next twelve months. On the surface, the logic appears reasonable. If the S&P 500 is trading at around 21 to 22 times forward earnings, then valuations may not seem particularly extreme, especially compared with the most euphoric periods of recent years. Today, the forward P/E ratio for the S&P 500 sits around 22, depending on the measure used.

For many investors, that relative stability is used as proof that the market is expensive for a reason. Prices are high, but earnings are expected to grow fast enough to justify them. That argument sounds convincing until one looks at what sits behind the denominator.

Forward earnings are not facts. They are estimates built on assumptions about future growth, future demand, labor costs, margins, financing conditions, and consumer behavior. They are projections, not reality. And history shows that those projections are almost always revised lower when the cycle begins to deteriorate.

Analysts tend to extrapolate recent strength into the future. When markets are rising, margins are elevated, and economic conditions still appear stable, earnings forecasts move higher alongside stock prices. That optimism makes valuations appear more reasonable than they actually are. As a result, forward P/E ratios often look the cheapest precisely when risk is the highest and right before earnings begin to disappoint.

There is also a deeper flaw in the concept itself. The forward P/E compares today’s price with tomorrow’s earnings, mixing two completely different points in time. Price is a present variable. Forward earnings belong to the future. If those earnings actually materialize, the price is unlikely to remain where it is today. Investors would bid the shares even higher as those stronger earnings become more visible.

In other words, higher future earnings do not justify today’s price. They justify a higher future price.

Share

Today

This problem becomes even more dangerous late in the cycle, which is exactly where we are today. The market is currently being supported by extremely optimistic assumptions around artificial intelligence, productivity gains, and continued earnings growth. Large financial institutions continue to raise their year end targets for the S&P 500 based on stronger expected profits.

Those same optimistic assumptions are what feed directly into forward earnings estimates. The result is that the forward P/E ratio remains stable, not because prices are disciplined, but because analysts keep raising the denominator alongside the numerator. Prices rise, earnings expectations rise with them, and the ratio creates the illusion that valuations are not expanding.

That is why the forward P/E ratio often functions less as a valuation tool and more as a sentiment indicator. When confidence is high, earnings forecasts rise and valuations appear reasonable. When confidence fades, those same forecasts are revised lower, often abruptly, making the market suddenly look expensive after prices have already fallen.

We saw this dynamic clearly in 2022. Before the bear market began, the forward P/E ratio remained close to historical averages, suggesting that valuations were manageable. Yet the S&P 500 still declined roughly 20 percent as interest rates moved higher, liquidity tightened, and earnings expectations began to weaken. The forward P/E ratio offered little warning because it was anchored to earnings estimates that had not yet adjusted to reality.

Subscribe now

What to Actually Pay Attention To

Looking at other valuation measures, the answer becomes much clearer.

The Buffett Indicator, which compares total U.S. stock market capitalization with GDP, is currently estimated near 220 to 230 percent, a level that exceeds even the dot-com bubble.

At the same time, the Shiller CAPE ratio remains around 40, close to the highest valuation levels ever recorded and only slightly below the peak reached in 2000.

Unlike forward P/E, these measures do not rely on hypothetical future earnings. They compare prices with realized profits, realized economic output, and long-term averages. That is why they tend to be far more useful when trying to understand whether markets are truly expensive.

And right now, they are all telling the same story. The market is not cheap. It is historically expensive.

That does not mean the bubble has to end immediately. Overvalued markets can always become even more overvalued, especially when optimism is high and liquidity remains abundant. In fact, another rally could easily push the market toward the same valuation extremes seen during the dot-com bubble.

But the stability of the forward P/E ratio should not be interpreted as evidence that valuations are under control. It should be interpreted as evidence that earnings expectations remain extraordinarily optimistic.

Historically, that has never been a reason for comfort.

There are many different ways to measure whether the market is expensive. Price to Sales, Price to Book, market capitalization to GDP, often referred to as the Warren Buffett Indicator, and several other metrics all point to the same conclusion. By many of these measures, the current U.S. stock market is the most overvalued in history, above 1929, above 1968, above 2000, and above 2021.

The one major valuation metric that still remains slightly below the extremes of 2000 is the CAPE ratio, also known as the Shiller Cyclically Adjusted Price to Earnings Ratio.

The CAPE ratio measures the market price relative to the average of the last 10 years of inflation-adjusted earnings. Instead of using only the latest twelve months of profits, which can be heavily distorted by recessions, profit booms, or temporary shocks, CAPE smooths earnings across a full decade. This makes it one of the most useful long-term valuation tools because it removes much of the noise that can distort traditional price-to-earnings ratios.

A standard price-to-earnings ratio can look artificially cheap when profits are temporarily inflated, or artificially expensive when profits collapse during a recession. CAPE helps avoid that problem. It is not perfect for timing short-term market moves, but historically it has been one of the strongest indicators of long-term returns. The higher the starting CAPE ratio, the lower the future returns have tended to be. The lower the starting CAPE ratio, the better the future returns have tended to be.

Analisis

Today, the CAPE ratio for the S&P 500 stands near 40. That is already an extremely elevated level by historical standards. The long-term median CAPE ratio is only around 16, meaning the market today trades at more than double its historical norm.

However, despite how stretched valuations already are, we are still below the all-time peak reached during the dot-com bubble in 2000, when the CAPE ratio briefly climbed above 44. That means we are surprisingly close.

To move from a CAPE ratio of roughly 40 today to the 44 level reached in 2000, the market would only need to rise approximately another 13 to 14 percent, assuming earnings remain relatively stable. That is not a particularly large move, especially in the context of a late-cycle rally driven by optimism around artificial intelligence, easier monetary policy expectations, and the possibility of geopolitical tensions easing.

This is what makes the current setup so interesting. We are already in one of the most overvalued markets in history, yet we are still close enough to the 2000 peak that another relatively modest rally could push valuations back to the highest level ever recorded in more than 150 years of market history. That would place this cycle firmly above 1929, above the late 1960s, above 2021, and second only to the dot-com bubble, or perhaps even above it.

And if that happens, we would once again be living through history. The comparison to prior valuation peaks is important because valuations matter most when viewed through the lens of future returns.

Share

Time to Recover

In 1929, the market reached one of the highest valuation levels in history, reaching a CAPE of 31.48, before collapsing into the Great Depression. In nominal terms, without inflation adjustment, investors had to wait roughly 25 years to fully recover their investment.

The late 1960s produced a different type of outcome. Valuations were elevated, though not as extreme as 1929 or 2000, but the inflationary environment of the 1970s made real returns extremely poor. Even though the market did not collapse in the same dramatic way as it did after 1929, investors still spent more than two decades with little to no real progress. In inflation-adjusted terms, it took roughly 27 years for the market to recover from the 1966 peak.

Then came 2000. The dot-com bubble remains the clearest modern example of what happens when investors overpay for future growth. The CAPE ratio reached an all-time high of 44.19 in March 2000. After that peak, the S&P 500 went through two major bear markets, first with the collapse of technology stocks and then with the Global Financial Crisis. In real terms, investors had to wait approximately 17 years before fully recovering.

This is one of the most important lessons in investing. Valuations do not tell you what the market will do next month, next quarter, or even next year. Overvalued markets can always become even more overvalued. Bubbles can continue much longer than most people expect. That is exactly what may be happening now.

But valuations do tell you a lot about what future long-term returns are likely to look like. When investors buy into periods of extreme optimism and elevated multiples, future returns tend to disappoint. Not necessarily because earnings collapse immediately, but because so much future growth has already been priced in.

Share

Today

Today, the CAPE ratio is once again telling a very similar story. The market may still have room for one final euphoric push, especially if the geopolitical backdrop improves and investors continue chasing the same handful of large technology names that have driven most of the gains in recent years. Another 10 to 15 percent rally is enough to put us back at the most expensive valuations in modern history.

But history also suggests that when investors start from these levels, the following decade tends to be much less exciting. That is not a prediction. It is simply what the data has shown, over and over again, across every major bubble of the last century.

That concern was not irrational. Whenever oil rises sharply, especially due to geopolitical tensions, the impact tends to appear quickly in the headline inflation data. Energy becomes more expensive, transportation costs increase, and companies begin passing those higher costs through to consumers. Food prices can also be affected indirectly, since production, logistics, and distribution all depend heavily on fuel.

This is precisely why markets pay attention not only to headline inflation, but also to core inflation.

Headline inflation includes everything, including volatile categories such as food and energy. Core inflation, by contrast, excludes those categories in order to provide a cleaner view of the underlying inflation trend. The reason this distinction exists is that energy prices can be heavily influenced by geopolitical events, while food prices can be distorted by droughts, floods, or supply shortages. These are often temporary shocks that may lift inflation in the short term without necessarily changing the broader inflation cycle. That is exactly what happened in March.

The Data

Headline CPI increased 0.9 percent month over month and 3.3 percent year over year, a notable jump from the 2.4 percent annual inflation rate recorded in February. On the surface, that looks like a meaningful reacceleration in inflation. However, once we look beneath the surface, it becomes clear that the vast majority of the increase came from energy. The energy index surged 10.9 percent in March alone, while gasoline prices rose 21.2 percent during the month and were up 18.9 percent compared with a year earlier. Fuel oil increased even more sharply, rising 30.7 percent month over month and 44.2 percent year over year. Gasoline alone accounted for nearly three-quarters of the monthly increase in headline CPI.

The impact of higher energy prices was also visible across transportation and related categories. Transportation inflation rose 5 percent year over year, with gasoline being the single largest contributor. Airline fares also moved higher and contributed to the increase in inflation.

At the same time, food inflation remained relatively contained. The overall food index was unchanged during the month. Food consumed at home actually declined by 0.2 percent, while food away from home increased by only 0.2 percent. This suggests that, at least for now, the energy shock has not yet translated into a broad-based acceleration in food prices.

Share

The Core Inflation

The more important part of the report, however, was core inflation.

Core CPI rose only 0.2 percent in March, up from 2.5 percent year over year in February. That is barely a change, especially considering the magnitude of the oil shock that occurred during the month. More importantly, the result came in below expectations, as economists had been looking for a larger increase.

One of the categories that appears to have contributed to the slight increase in core inflation was airline fares. The Bureau of Labor Statistics specifically highlighted airline tickets as one of the categories that moved higher during the month, which makes sense given the sharp rise in jet fuel prices. Transportation data shows that airline fares accounted for roughly 4 percent of annual transportation inflation in March.

This is important because it reinforces the idea that the inflation pressure we saw in March was concentrated in categories directly exposed to higher energy costs, rather than reflecting a broad based increase across the economy. In other words, the report looked much more like a temporary external shock than the beginning of a new structural inflation cycle.

Share

Looking Ahead

Looking ahead, the focus should increasingly shift toward the month over month data rather than the year over year figures. The reason is simple. Oil prices today are significantly higher than they were a year ago, which means that annual comparisons are likely to remain elevated for some time even if oil prices stabilize. In other words, year-over-year inflation may continue looking high simply because the comparison period was much lower.

What matters now is whether oil continues to rise from here. If oil stabilizes, or if the geopolitical premium begins to fade, then the monthly inflation data should begin to soften again. If oil continues to move higher, we may continue to see temporary upward pressure on headline CPI. But even in that scenario, the key variable remains core inflation, and that part of the report still looks remarkably contained.

The market reaction reflected that interpretation exactly. Stocks did not collapse, Treasury yields did not surge, and there was no major repricing of interest rate expectations. Investors understood that the inflation report was driven primarily by energy and that core inflation remained largely under control.

The market knows that these types of shocks are temporary. And unless they begin feeding into broader categories, especially housing, wages, and core services, inflation is likely to continue following the same downward trend that was already in place before the geopolitical tensions began.

This has become the dominant storyline. The assumption is straightforward. If oil moves higher, inflation follows. The logic feels intuitive and, at least on the surface, historically grounded. Energy costs feed directly into transportation, production, and ultimately consumer prices. A shock in oil, especially one driven by geopolitical risk, tends to show up quickly in headline inflation, reinforcing the perception that inflation itself is once again accelerating.

But this interpretation, while not entirely incorrect in the short term, misses the broader structure of how inflation actually works.

To understand inflation, it is necessary to understand its composition. The Consumer Price Index in the United States is not driven primarily by volatile components like energy, but rather by more persistent categories. The most important of these is housing, specifically rent and shelter costs, which account for roughly 35 percent of the entire index. This is not a marginal component. It is the core of the core.

What this implies is critical. For inflation to move structurally higher, housing must move higher. Without an acceleration in rent inflation, it becomes extremely difficult for the remaining 65 percent of the index to offset that weight and generate sustained upward pressure on the overall number.

Share

The history…

The data support this clearly. As shown in the chart below, whenever rent inflation has been around 3 percent, headline inflation has historically settled near 2.4 percent. This relationship has repeated itself across decades, appearing consistently in multiple periods, including the late 1950s, the late 1960s, the early 1970s, the early 1980s, the early 2000s, the mid 2000s, the late 2010s, and most recently in the latest April reading. Today, we are once again in that same configuration, with rent inflation near 3 percent and headline inflation printing at 2.4 percent. This is not a coincidence. It is structured.

And yet, much of the current inflation debate is being framed through the lens of the 1970s, a period often cited as the archetype of inflation driven by oil shocks. The narrative is well known. The oil embargo, supply disruptions, and surging energy prices are seen as the primary catalysts behind the multiple inflation spikes that defined that decade. This comparison has resurfaced repeatedly in recent years, particularly whenever energy markets experience volatility. However, this narrative is incomplete.

Share Signals Over Noise

The seventies

While oil shocks did play a role in amplifying inflation during the 1970s, they were not the underlying driver of the cycle. The structural force behind those inflationary waves was housing. The demographic surge driven by the baby boom generation created a significant increase in demand for housing as millions of individuals entered the labor force and began forming households. This demand translated directly into higher rent inflation, which, given its weight in the CPI, pushed overall inflation higher.

The chart above makes this dynamic evident. The three major inflation peaks of the 1970s align closely with corresponding peaks in rent inflation. This is a critical point: oil may have exacerbated volatility, but the persistence and magnitude of inflation during that period were rooted in housing. Without that structural pressure, the impact of oil alone would not have been sufficient to sustain multiple inflation cycles.

Subscribe now

Today

If the current inflation concerns are being driven primarily by oil, then by definition, they are more likely to be temporary. Energy shocks can create short-term spikes in inflation, but they do not, on their own, generate sustained inflationary regimes. For that to happen, the housing component must accelerate. And that is not what the data is showing.

Rent inflation today is running at approximately 3 percent, a level historically consistent with moderate, stable inflation rather than an accelerating cycle. More importantly, the underlying conditions in the housing market suggest downward pressure, not upward acceleration. Mortgage rates remain elevated, significantly constraining affordability. At the same time, inventory levels are rising, with more sellers entering the market relative to buyers. This imbalance points toward continued price adjustments rather than renewed expansion.

In this environment, it is difficult to construct a scenario in which rent inflation meaningfully accelerates. Without that acceleration, the probability of a structurally higher inflation regime remains low. This does not mean inflation cannot rise in the short term. A spike in oil prices, particularly if geopolitical tensions escalate further, can push headline inflation higher over the coming months. These effects are real and tend to be quickly reflected in the data. However, they are also self-limiting. Higher energy costs act as a constraint on demand, weakening consumption and ultimately reducing the very pressures that caused the increase.

The market, once again, is focusing on the most visible variable, oil, while underestimating the most important one, housing.

As shown in the chart, the relationship between rent inflation and headline inflation has been remarkably stable over time. If rent inflation begins to decline from current levels, which is consistent with the direction implied by housing market fundamentals, then headline inflation is likely to follow. The path may include short-term volatility, particularly in response to energy markets, but the underlying trend would remain downward.

There is a growing narrative that a second wave of inflation is imminent, driven by geopolitics and supply shocks. History suggests otherwise. Without a corresponding surge in housing, such a scenario lacks the structural foundation required to sustain it.

Oil can move inflation in the short term. It cannot redefine the cycle. Housing can. And right now, housing is not pointing up.

Since that point, the global economy has been progressively weakening. Credit conditions have tightened, growth has slowed, and labor markets have begun to show early signs of stress. Oil, as it often does, reflected this shift early. The move lower toward the mid-$50 range over the past 6 months was not an anomaly but a signal, confirming that demand was softening and that the underlying engine of the global economy was losing momentum.

Understanding oil requires understanding what it represents. Oil is not simply a commodity traded on supply and demand imbalances in isolation. It is deeply embedded in the structure of economic activity. It fuels transportation, supports industrial production, and sits at the core of global trade and logistics. When oil rises due to endogenous forces, it typically reflects expansion: more consumption, more production, and more movement across the economy. But when oil declines in the absence of a supply shock, it signals the opposite. It reflects fewer transactions, declining industrial output, and weakening demand across sectors.

This is why oil acts as a real-time check on economic conditions. A slowdown in consumption translates into fewer shipments, fewer deliveries, and reduced fuel usage across transportation networks. Even large-scale logistics systems, such as those driven by e-commerce demand, become more sensitive to declining activity. At the same time, industrial sectors reduce output, further compressing energy demand. Oil, in this sense, is neither leading nor lagging the economy; it is moving with it, confirming the direction of the cycle.

Share

Possible Paths

Against this backdrop, the current geopolitical shock introduces uncertainty, but it does not fundamentally alter the underlying trajectory. Instead, it creates a set of possible paths that, despite appearing different on the surface, are likely to converge toward a similar outcome.

Scenario 1

In a scenario where tensions between the United States and Iran de-escalate and the Strait of Hormuz remains open, the risk premium currently embedded in oil prices would likely unwind quickly. Markets would shift their focus back to fundamentals, which remain weak (the current state of the US economy). The pre-existing trend in Oil, which was already pointing downward, would reassert itself, potentially pushing oil prices into the $40-$50 range. Such a move would not represent a new development, but rather a continuation of the same macro story that has been unfolding since 2022: slowing growth, weakening demand, and tightening financial conditions.

Scenario 2

Alternatively, if the conflict escalates further, oil prices could move significantly higher in the short term due to supply disruptions. However, interpreting such a move as bullish would be a mistake. A spike driven by exogenous factors does not signal economic strength; it acts as a constraint on the system. Higher oil prices increase costs across the board, functioning effectively as a tax on both consumers and producers. In an already fragile economic environment, this added pressure becomes difficult to absorb.

The mechanism through which this pressure manifests is relatively straightforward. If higher energy costs are passed on to consumers, headline inflation rises in the short term, reducing real purchasing power. But unlike earlier stages of the cycle, consumers today are far less capable of absorbing these increases. Credit is no longer expanding at the same pace, savings buffers have diminished, and labor market conditions are gradually softening. As a result, higher prices translate directly into reduced demand, further weakening economic activity.

If, on the other hand, companies are unable to pass these costs through, the pressure shifts to margins. Businesses that rely heavily on energy inputs, particularly in transportation, logistics, and manufacturing, face rising costs without a corresponding increase in revenue. This compresses profitability, weakens balance sheets, and increases the likelihood of cutbacks, layoffs, or even defaults. In both cases, the system adjusts, but the adjustment is contractionary.

This dynamic is not theoretical. It closely mirrors what occurred in 2008, when oil prices surged sharply (54%) in the first half of the year despite an already deteriorating economic backdrop. That surge did not reflect strength; it exacerbated existing weaknesses, accelerating the downturn. Once demand ultimately collapsed, oil followed, falling 77% from its peak. See image 1 below.

Scenario 3 (Less likely)

There is, in theory, a third scenario in which oil prices remain elevated, in the range of $100 to $110, without triggering a significant economic response. However, this outcome assumes a level of stability that is inconsistent with the nature of the current shock. Geopolitical disruptions are inherently binary; they either escalate or resolve. They do not typically sustain a prolonged equilibrium at elevated price levels without inducing second-order effects on demand or supply. More importantly, such a scenario would require a degree of resilience in the global economy that is difficult to justify given the current stage of the cycle.

Share

Ultimately, while the short-term paths may differ, the longer-term outcome appears increasingly constrained. If the conflict resolves, oil prices are likely to decline as weak fundamentals reassert themselves, reconnecting with the same downward trend that was already in place before the geopolitical shock interrupted it. If the conflict escalates, any resulting price spike is unlikely to be sustained, as it would place additional pressure on an already fragile system, accelerating the deterioration in economic activity, compressing demand, and ultimately triggering the same type of reversal historically observed in late-cycle environments.

What matters here is not the path, but the mechanism. Oil, in this context, is not driving a new cycle, it is reacting within the constraints of an existing one. A higher price driven by exogenous disruption does not create demand; it destroys it. And in a system where credit is tightening, labor markets are softening, and consumption is already under pressure, that destruction tends to happen faster and more aggressively.

This is why the range of possible outcomes is narrower than it appears. The market may attempt to price different narratives in the short term, oscillating between geopolitical risk and supply concerns, but the underlying macro backdrop imposes limits on how far those narratives can extend. At this stage of the cycle, elevated prices become self-correcting, not because supply improves, but because demand weakens to the point where prices can no longer be sustained.

The global economy is not in a position to absorb prolonged energy shocks without consequence. Whether the adjustment comes through the consumer, through corporate margins, or through a broader contraction in activity, the end result converges toward the same direction: lower demand and, eventually, lower prices.

Higher interest rates play a central role in this transition. Debt that was once cheap to roll over becomes more costly. Companies facing higher refinancing costs are forced to cut expenses, delay investments, and in many cases reduce their workforce. Households feel the same pressure, as credit card rates, mortgages, and loans become more expensive. The result is a gradual slowdown in economic activity, which eventually leads to a contraction.

The Contraction is Normal

This process is not an anomaly. It is the mechanism through which the system resets itself. Debt is restructured, inefficient capital is removed, and the foundation is laid for the next cycle. Recessions are not accidents, they are necessary phases that follow periods of expansion. Without them, imbalances would continue to build indefinitely.

Within this framework, the labor market becomes one of the most important indicators to monitor. Employment is the backbone of consumption, and consumption represents the majority of economic activity. When the labor market weakens, the entire system begins to adjust.

One of the most revealing indicators today is the average duration of unemployment, which currently stands at 25 weeks. At first glance, this may not seem alarming. In fact, in the context of the cycle, it can be interpreted as a normal development. As we approach the end of an expansion, it is expected that it becomes harder for unemployed individuals to find new jobs. Hiring slows, job openings decline, and the time it takes to reenter the workforce increases.

However, when we place this number in a long term context, the picture becomes more concerning. As shown in Image 1, the average number of weeks unemployed prior to each recession has been trending higher over time. In 1970, before entering recession, the average duration was around eight weeks. By the 1974 downturn, it had increased to roughly nine to ten weeks. In 1979, it moved above ten weeks. By the 2000 recession, it was closer to twelve weeks. And before the 2008 financial crisis, it had risen to approximately seventeen weeks.

The pattern is clear. Each cycle begins from a weaker labor market starting point than the previous one.

Historically, outside of the 2008 crisis, the average duration of unemployment during recessions tended to remain below roughly twenty-two weeks. That range held relatively consistently for decades. But the 2008 crisis marked a structural break. As shown in Image 1, the average duration surged dramatically, eventually peaking above forty weeks around 2012. That was not just a cyclical move, it was a shift in the behavior of the labor market.

Share

Today

Today, we are entering the late stage of the cycle with the average already at 25 weeks, before a recession has even begun. This is the key point. If the previous cycle entered recession at around seventeen weeks (2008) and then surged above forty, what should we expect now that we are starting from a significantly higher base?

It is reasonable to assume that, in the next downturn, this metric could rise to forty five or even fifty weeks. As shown in Image 1, the increases in unemployment duration during recessions have been substantial, and there is little evidence to suggest that the next cycle will behave differently. If anything, the structural trend points toward longer periods of unemployment.

This is why the current situation is more concerning than it appears on the surface. The level itself may seem consistent with a late cycle environment, but the trend is what matters. We are not entering the next recession from a position of strength. We are entering it from a position that is already weaker than in previous cycles.

And this is not an isolated signal. Other labor market indicators are pointing in the same direction. The overall number of unemployed individuals is rising. Job creation is slowing. Revisions to employment data continue to trend downward. Each of these elements reinforces the same conclusion. The labor market is deteriorating before the recession has even officially begun.

This matters because employment drives consumption, and consumption drives GDP. If individuals remain unemployed for longer periods, their spending declines. As spending declines, businesses generate less revenue, leading to further layoffs. This feedback loop is what transforms a slowdown into a full economic contraction.

The implication is straightforward. If the labor market is already showing signs of structural weakness before entering a recession, the downturn itself is likely to be more severe. Starting from a higher baseline of unemployment duration means that the adjustment process will be longer and more painful.

This cycle is not starting from the same conditions as previous ones. It is starting from a weaker position. And if history is any guide, that difference will matter once the recession fully unfolds.

So what is private credit? In plain terms, it’s when investment funds, hedge funds, and private equity firms lend money directly to mid-sized companies, the kind of businesses that your local bank stopped wanting to deal with after the 2008 financial crisis. When the crisis hit, regulators tightened the rules on banks, making it harder and more expensive for them to lend to riskier borrowers. That left a huge gap in the market, and a new industry rushed in to fill it. Over the past fifteen years, private credit has grown from a niche corner of finance into a nearly $2 trillion global market, with the United States at its center. It’s fast, it’s flexible, and it offers returns of around 8 to 10 percent a year, which, in a world of low interest rates, made it look like pure gold to investors.

The reason it grew so fast is the same reason it’s now starting to worry serious people. Private credit isn’t traded on any public exchange. There are no daily price updates, no ticker symbols, and no transparency. When a bank makes a loan, regulators can see it. When a private credit fund makes a loan, it’s largely invisible to the broader market. That opacity was seen as a feature, it kept nervous investors from panicking during market swings. But opacity in good times becomes a very dangerous thing when the tide starts to turn.

And the tide is turning. As of early 2026, the signs of stress in the private credit market are hard to ignore. Default rates, meaning the percentage of borrowers who stop paying back their loans, have climbed to 5.8%, the highest level ever recorded for this market. High-profile bankruptcies, like those of First Brands and Tricolor, have left private lenders absorbing real, painful losses. Some funds have quietly begun restricting investors from withdrawing their money, a classic sign that things aren’t as liquid as they appeared. Banks, meanwhile, are tightening the credit they extend to these funds themselves, squeezing the whole system from the top down.

Share

PIK

There’s another warning signal worth understanding, even if it sounds technical at first: something called PIK, or “payment-in-kind.” Normally, when a company borrows money, it pays back interest in cash. When a company is in trouble and can’t afford even those interest payments, lenders sometimes agree to let the company pay its interest with more debt instead of real money (Ponzi Scheme?). It’s like paying your credit card bill by taking out another credit card. When PIK usage starts rising across a credit market, and it is rising right now, it means a growing number of borrowers are already drowning, just slowly enough that nobody’s officially declared a crisis yet.

None of this means the economy is in freefall tomorrow. Many analysts are quick to point out that sectors like artificial intelligence and certain parts of technology remain genuinely strong, and that some private credit funds are performing just fine. But that’s exactly what makes the current situation so tricky: there’s an enormous gap between the winners and the losers, and because this market is so opaque, most investors can’t easily tell which funds are healthy and which are quietly falling apart. That kind of hidden fragility is what makes financial crises so sudden when they finally arrive.

Share Signals Over Noise

The Credit Cycle

Here’s the bigger picture, and this is the part that really matters: credit is not just one piece of the economy. It is the engine of the economy. When credit flows freely and cheaply, businesses borrow to invest, hire workers, launch new products, and expand. Workers earn wages, consumers spend, and the cycle feeds itself. The whole boom of the last decade was fueled, in large part, by easy credit, and private credit was among the biggest sources of that fuel. On the flip side, almost every serious recession in modern economic history has been preceded not by a war or an oil shock, but by a credit contraction. When lenders get scared and stop lending, businesses can’t refinance their debts, payrolls get cut, spending collapses, and the whole machine grinds to a halt. It’s not a theory, it’s a pattern that has repeated itself with remarkable consistency.

Wars and oil prices are real economic forces, but they are external shocks. They raise costs, they create uncertainty, they can absolutely cause pain. But an economy with healthy, flowing credit can absorb those shocks. It adjusts, it adapts, it pushes through. What it cannot absorb is a simultaneous freeze of credit. If lenders stop lending at the same moment that a war is raising prices and uncertainty is spreading, there is no cushion left.

What makes private credit particularly dangerous as a fault line is precisely what made it attractive in the first place: its size, its speed, and its secrecy. A $2 trillion market that operates largely in the dark, with no public prices and limited regulatory oversight, is not a market that corrects slowly and visibly. It’s a market that can look perfectly stable right up until the moment it isn’t. The current cycle is being described by insiders as the first real stress test this industry has ever faced, because the last time credit markets truly seized up, in 2008, private credit barely existed. Nobody actually knows how bad it could get when this machine hits a real wall.

The honest answer is that we are in genuinely uncertain territory. The defaults are rising. The stress signals are multiplying. The market is opaque enough that even the experts are working with incomplete information. What is clear is that the people watching the economy most carefully are not losing sleep over oil prices or the latest geopolitical standoff. They’re watching the private credit market, because they know that when credit stops flowing, everything else stops too. The next time someone tells you the big risk is whatever’s dominating the headlines, it’s worth remembering: the biggest risks are almost always the ones that don’t make the front page until it’s too late.

But this conclusion oversimplifies the mechanics behind currencies. It blends together two different ideas that should not be treated the same. On one side, there is the long-term debasement argument tied to money supply growth. On the other, there is the relative value of the dollar against other currencies, which is what the DXY actually measures.

When we talk about gold, silver, or other scarce assets, the debasement argument holds over long horizons. More dollars in the system tend to push the price of finite assets higher. But the DXY is not measuring scarcity. It is measuring relative strength, primarily against the euro, which dominates the index composition.

This distinction is critical because currencies do not move in isolation. The dollar does not weaken simply because more of it exists. It weakens or strengthens based on global liquidity conditions, relative growth, and, most importantly, credit dynamics. This is where the current narrative begins to break down.

The dollar today appears to be stabilizing around the 98 to 100 range. This level is not random. In a macro environment that is shifting toward disinflation, and potentially deflation, the dollar tends to find support. Deflationary pressures increase the demand for liquidity, and the dollar remains the core of that global liquidity system. When liquidity tightens, the system does not look for alternatives. It moves toward dollars. This is why periods of stress, not expansion, tend to coincide with dollar strength. And increasingly, the macro data is pointing in that direction rather than toward continued expansion.

Is the dollar dying?

The first chart highlights something that directly contradicts the “dollar is dying” narrative. Since 2008, the dollar has been in a long-term appreciation trend. This does not mean it moves in a straight line, but the direction over time has been higher, not lower. The idea of structural decline simply does not align with this data.

DXY (Indice Dolar) vs EEM (EFTs Emerging Economies)

However, the more important insight comes from within that trend. The dollar does not just drift higher. It moves in cycles. Roughly every three to four years, we see strong upward moves in the dollar, followed by periods of consolidation or decline. These cycles are not random. They are deeply tied to global liquidity and credit conditions.

DXY (Indice Dolar) vs ILF (EFTs Latin American Economies)

Share

This is where the relationship with emerging markets becomes clear. In the first chart, every major bottom in the dollar since 2008 has coincided with a peak in emerging markets. The same pattern appears in the second chart for Latin American markets. The timing is not perfect, but the relationship is consistent.

The reason behind this is structural. Companies and economies in emerging and Latin American markets generate most of their revenues in local currency. However, a significant portion of their liabilities, financing, and even some operating costs are denominated in dollars. This creates an inherent currency mismatch. When the dollar strengthens, this mismatch becomes a problem. Servicing dollar-denominated debt becomes more expensive in local currency terms. Profitability declines, balance sheets weaken, and financial conditions tighten. Even if local economies remain stable, the currency effect alone creates pressure.

This is why a stronger dollar tends to act as a tightening force globally. It is not just a reflection of strength in the United States. It is a mechanism that drains liquidity from the rest of the world. And this is exactly the type of dynamic that tends to hurt emerging and Latin American markets the most.

Share

The current macro backdrop reinforces this view. We are beginning to see signs of tightening liquidity across the system. The increase in repo usage is one of those signals. Repo markets (short-term funding from the Federal Reserve to commercial banks to meet their liquidity needs) reflect short-term funding issues, and when usage rises, it often indicates that liquidity is becoming scarce. This type of behavior typically appears late in the cycle, not at the beginning. In early-cycle environments, liquidity is abundant, credit is expanding, and funding is easily available. What we are seeing now is closer to the opposite. The system is beginning to show signs of stress beneath the surface.

At the same time, credit conditions are deteriorating. Credit is the core driver of the economic cycle. It fuels consumption, investment, and expansion. When credit growth slows or contracts, the entire system begins to adjust. Demand weakens, growth slows, and financial conditions tighten.

In that environment, the demand for dollars increases. Not because of strength, but because of necessity. The global system is built on dollar funding, and when that funding becomes scarce, the dollar rises. This is not a theory. It is a pattern that has repeated across multiple cycles.

So the implication is not subtle. The current weakness in the dollar may not be the beginning of a structural decline, but rather a pause within a broader cycle. If the macro environment continues to shift toward tighter liquidity and weaker credit, the next move in the dollar is likely higher. And if that happens, the consequences for emerging and Latin American markets follow the same script we have seen before. What is currently perceived as strength in those markets may, in reality, be the late stage of a cycle that is about to turn.

The inflation of the 1970s was not just about oil. It was fundamentally driven by demographics. The entry of the baby boomer generation into their prime consumption years created a powerful and sustained demand shock, particularly in housing. This was not a temporary disruption. It was a structural expansion of demand across the economy, pushing prices higher over an extended period of time.

If we break down that decade, we can identify three distinct inflationary waves. The first, around 1969–1970, emerged as the economy overheated after years of expansion in the 1960s. Fiscal spending tied to the Vietnam War and Great Society programs had already stretched capacity, while the labor market tightened significantly. Housing demand accelerated as younger households entered the market, and inflation began to build even before the major oil shocks materialized.

The second wave, peaking in 1974–1975, is often reduced to the oil embargo, but that is only part of the story. By the time energy prices surged, inflationary pressures were already embedded in the system. Wage growth had accelerated, credit expansion remained strong, and housing costs continued to push higher. The oil shock acted as an amplifier rather than the origin. Importantly, this cycle ended not because oil prices normalized, but because the economy rolled over and unemployment began to rise sharply, compressing demand.

The third and final surge, in 1979–1980, followed a similar pattern but with even stronger momentum. Inflation expectations had become more entrenched, and the economy experienced another push from energy prices after the Iranian Revolution. However, beneath the surface, the same dynamics were at play. Demand had been running hot for years, supported by demographics and credit. But once again, as monetary conditions tightened and the labor market deteriorated, the system could not sustain elevated inflation. The downturn that followed ultimately broke the cycle.

What ties these three episodes together is not just the presence of supply shocks, but the persistence of demand. Each wave required a strong labor market, rising incomes, and continuous consumption. And in each case, the turning point was the same. When unemployment bottomed and began to rise, inflation followed it down.

The mechanism is straightforward. Inflation, at its core, requires sustained demand. Even in an environment of high commodity prices, if individuals begin to lose their jobs, income falls, consumption weakens, and pricing power disappears. Without income, there is no capacity to absorb higher prices. The demand side breaks, and inflation inevitably declines.

This dynamic held consistently throughout the 1970s. Despite multiple oil shocks and persistent supply-side pressures, inflation could not sustain itself once the labor market deteriorated. The idea that inflation can remain elevated in the face of rising unemployment is not supported by the historical record.

Share

Today

Fast forward to today, and the contrast becomes clear. While markets are once again focused on energy prices and geopolitical risks, the labor market is telling a very different story. Recent non-farm payroll data has shown clear signs of weakness, with disappointing prints over the past few months and significant downward revisions to prior estimates. What initially appeared as a resilient labor market in 2025 now looks far more fragile.

This matters more than any short-term spike in oil. A weakening labor market undermines the very foundation required to sustain inflation. Without job growth, without income expansion, and with increasing uncertainty around employment, demand cannot hold. And without demand, inflation cannot persist.

So while the narrative of a “new 1970s” is gaining traction, the underlying conditions simply do not align. The inflation of that decade was built on a structural demand boom driven by demographics. Today, we are facing the opposite. Even if we experience short-term inflationary shocks, the broader trajectory is constrained by a deteriorating labor market. And in that environment, sustained inflation becomes not just unlikely, but fundamentally unsustainable.