The Long Wave

Nikolai Kondratieff was shot by Stalin's firing squad in 1938, which is a poor way to reward a man for being right. The Soviet economist had identified something the regime found ideologically intolerable: capitalism, rather than collapsing under its own contradictions, appeared to move in long waves of approximately 40-60 years. Expansion, stagnation, contraction, recovery — and then the cycle began again. The pattern was not a circle but a spiral, each revolution at a higher level of technological and institutional complexity.

The idea was simple and, for the Soviets, dangerous: capitalism adapts. It does not merely repeat; it reinvents. Each long wave is driven by a cluster of transformative technologies — canals and textiles in the first wave, railroads and steel in the second, electrification and chemicals in the third, automobiles and petrochemicals in the fourth. The fifth wave, beginning in the 1970s and 1980s, was driven by information technology and telecommunications. If the pattern holds, we are now in the early expansion phase of a sixth wave, driven by artificial intelligence, biotechnology, and renewable energy.

But it is the shape of these cycles that deserves attention, not merely their existence. Each wave follows a rough pattern: a technological breakthrough enables a new mode of production. Capital floods in. A speculative boom inflates asset prices beyond any rational valuation. The boom collapses. A period of institutional adjustment follows — new regulations, new social contracts, sometimes new political orders. And then the next expansion begins, built on the wreckage and the lessons of the last.

History does not repeat, but it rhymes. The rhyme scheme, however, is more regular than most people are comfortable admitting.

The Rhyme We Are Living

Consider the period from 1920 to 1955. A technological revolution (electrification, mass production, radio, aviation) fuels an extraordinary economic expansion. Asset prices surge. Inequality widens to levels not seen in generations. Credit expands recklessly. The culture is exuberant, experimental, unmoored from Victorian constraints. Then: collapse. A financial crisis (1929) triggers a depression. The depression breeds political extremism. Political extremism leads to war. War destroys and, paradoxically, rebuilds — creating the conditions for a new, more broadly shared prosperity.

Now consider the period from 2010 to the present. A technological revolution (mobile computing, cloud infrastructure, machine learning, generative AI) fuels an extraordinary economic expansion. Asset prices surge. Inequality widens to levels not seen in generations. Credit expands — this time through venture capital, crypto speculation, and government deficit spending. The culture is exuberant, experimental, and algorithmically unmoored. The parallels are not subtle.

The compound annual growth rate of the S&P 500 from 2010 to 2024 was approximately 11.4%, compared to 24.8% for the Dow Jones from 1924 to 1929. The numbers differ in magnitude but share a shape: sustained growth rates that outstrip underlying economic fundamentals. The ratio of market capitalization to GDP — Warren Buffett's preferred valuation metric — reached 195% in late 2024, exceeding the 2000 dot-com peak of 148% and approaching the 2021 high of 200%.

The debt picture is similarly rhyming. US federal debt-to-GDP reached approximately 123% in 2024. The formula is straightforward:

where $D_t$ is debt, $Y_t$ is GDP, $r$ is the average interest rate on debt, $G_t$ is government spending, $T_t$ is tax revenue, and $g$ is real GDP growth. When $r > g$ — when the interest rate on debt exceeds economic growth — the ratio compounds upward without bound unless primary surpluses ($T_t - G_t > 0$) offset it. The US has not run a primary surplus since 2001.

The Mechanism of the Turn

What makes cycles turn? Not just the mathematics of overvaluation, though that matters. The deeper mechanism is institutional exhaustion. The institutions built to manage one technological paradigm become increasingly misfit for the next. The regulatory frameworks, educational pipelines, labor market structures, and political coalitions that served the previous expansion become rigidities in the new one.

The 1920s economy ran on mass production, but its institutional infrastructure — weak labor protections, minimal financial regulation, gold-standard monetary policy — could not absorb the shocks that mass production generated. The concentration of income meant that consumption could not keep pace with production. The lack of financial regulation meant that speculative leverage amplified the downturn. The gold standard meant that monetary policy could not respond flexibly.

Every boom builds the institutions that will fail in the next crisis. This is not a bug. It is the mechanism.

Today's institutional mismatches are different in substance but similar in structure. Our educational system produces knowledge workers for an economy that is rapidly devaluing knowledge work. Our regulatory frameworks were designed for an era of scarcity — scarce information, scarce communication channels, scarce production capacity — and are confronting an era of abundance. Our tax systems are optimized for income from labor and physical capital, while an increasing share of value accrues to intellectual property and data.

The growth rate of AI capability, measured by various benchmarks, follows an approximately exponential curve. If we model capability $C$ as a function of compute investment $I$:

where $\alpha$ has been estimated at roughly 0.5-0.7 for recent large language models (the so-called "scaling laws"), then doubling compute produces a 40-60% increase in capability. This is not exponential in time — it is exponential in investment. The distinction matters because investment can accelerate or decelerate based on economic conditions, while exponential-in-time models imply inevitability.

The War Question

The darkest rhyme in the Kondratieff pattern is the association between long-wave downturns and major wars. The Napoleonic Wars coincided with the trough of the first wave. The American Civil War with the second. World War I and II bracketed the trough of the third. The Cold War's hottest phases — Korea, Vietnam — fell during the fourth wave's transition.

Correlation is not causation, but the mechanism is not mysterious. Economic contractions produce unemployed young men, resource competition, nationalist politics, and leaders who find foreign conflict useful for domestic legitimacy. The question is not whether the pattern exists but whether we can break it.

The optimistic case: nuclear deterrence, economic interdependence, and institutional learning (the UN, the EU, the WTO, however imperfect) have changed the game. The pessimistic case: those same institutions are eroding, great-power competition is intensifying, and the technologies of the new wave (AI, drones, cyber weapons) lower the threshold for conflict while raising the ceiling of destruction.

I do not know which case is correct. But I note that the people most confident in the optimistic case tend to be those with the least historical awareness, while the people most confident in the pessimistic case tend to be those with the least imagination about how institutions can adapt.

The long wave is turning. The question is not whether it will turn — the mathematics of debt accumulation, institutional rigidity, and technological disruption make some form of correction inevitable. The question is whether we manage the turn with intelligence and foresight, or whether we stumble through it the way our great-grandparents did: with a decade of suffering followed by a generation of rebuilding. The rhyme is there. The ending is not yet written.