The Long Wave

So far we’ve watched creative destruction happen one collision at a time: an incumbent here, a challenger there, a single moat breached. That’s the close-up. Now step back until individual firms blur and a rhythm comes into view. Destruction is not a one-off event that a lucky industry survives and is then safe. It arrives in waves — long swells of building followed by brutal seasons of clearing — and the tide never stops coming in. This lesson zooms out from the single crossover to the whole ocean, and asks the systems question: not whether the gale will blow again, but where in its cycle you are standing right now.

Boom and bust are one cycle, not a boom plus an accident

Picture a forest fire — the ecologist’s kind, not the disaster-movie kind. The years of lush overgrowth and the wildfire that razes it are not two separate events, one good and one bad. They are one cycle. The overgrowth creates the fuel; the fire clears the deadwood and returns nutrients to the soil so the next generation can grow. Suppress the fire forever and you don’t get eternal forest — you get a tinderbox that eventually burns catastrophically. Creative destruction has the same two-stroke engine.

The build-out phase is the boom. A promising new technology appears, the scent of profit spreads, and capital stampedes in. Firms proliferate — dozens, hundreds, chasing the same opportunity. Nobody yet knows the right scale, the winning business model, or how much capacity the market can absorb, so the system finds out the only way it can: by over-building. Too much track gets laid, too much fibre gets buried, too many near-identical startups get funded. Over-investment isn’t a bug here; it’s how a decentralised economy gropes toward the answer.

The clearing phase is the bust — the shakeout. Demand can’t support all that capacity, prices collapse, and the weak firms die. This looks like pure carnage, and for the people inside it, it is painful. But structurally it is the destruction half of creative destruction doing exactly its job: writing off the excess, killing the business models that didn’t work, and — crucially — freeing the resources. The capital, the skilled workers, the cheap leftover infrastructure all get reallocated to the survivors and to the next wave. The bust is not the failure of the boom. It is the boom’s necessary second half.

Worked example: the railway manias and the dot-com wave

Britain’s railway mania of the 1840s is the archetype. Investors poured money into railway companies at a frenzy; Parliament authorised thousands of miles of line, much of it duplicating existing routes or serving nowhere in particular. Share prices soared, then crashed. Most of the manic-era companies were ruined or absorbed. But when the smoke cleared, Britain was left with a dense national rail network — physical infrastructure, built with investors’ losses, that powered decades of growth. The speculators got destroyed; the rails stayed.

The dot-com boom and bust of 1995–2002 ran the identical script one wave later. Capital flooded into anything with a “.com”; telecom firms buried enormous quantities of fibre-optic cable on the bet that internet traffic would explode. In 2000–2001 it broke: the NASDAQ lost around three-quarters of its value, thousands of companies vanished, and vast fibre capacity sat “dark,” unused. Yet look at what the clearing phase left behind. The genuinely strong survivors — Amazon, Google — emerged into a field cleared of frivolous competitors. And that “wasted” dark fibre? It became the absurdly cheap backbone that made the next wave — streaming video, cloud computing, the mobile internet — economically possible. The mania over-built the road; the crash cleared the traffic; the next wave drove on it.

Kondratiev long waves and the clustering of innovation

Zoom out even further — past the single boom-bust to the scale of human lifetimes — and a more contested, more spectacular pattern is claimed. In the 1920s the Soviet economist Nikolai Kondratiev argued that capitalist economies move in long waves of roughly 45–60 years: a long expansion, then a long contraction, repeating. (Stalin’s regime found the implication — that capitalism self-renews rather than collapsing — inconvenient, and Kondratiev was executed in 1938. The idea outlived him.)

Schumpeter picked up the baton and supplied the engine. His insight: innovations don’t arrive smoothly, one per year like clockwork. They cluster. Every so often a general-purpose technology — a foundational advance useful almost everywhere — appears, and a swarm of related innovations clusters around it, powering one long wave of growth before its potential is exhausted and the economy waits for the next. The usual stylised sequence:

Why innovation clusters: a reinforcing loop with a delay

Why would innovation bunch up like this instead of dribbling out evenly? The answer is a feedback loop — and if you took the feedback-loops course, this is a textbook reinforcing loop with a delay, which is precisely the structure that produces waves rather than a smooth ramp.

Here’s the loop. A general-purpose technology — steam, electricity, the transistor, the internet — does something special: it lowers costs across many industries at once. Cheap motive power, cheap light and machinery, cheap computation, cheap communication. That broad cost drop makes a whole swarm of complementary innovations suddenly worth pursuing. Electricity didn’t just replace steam engines; it made possible the assembly line, the home appliance, the elevator (and therefore the skyscraper), refrigeration, radio. Each of those is a business, and each successful business throws off profits and frees up resources — which then fund the build-out of the next set of innovations. Success feeds success: more investment → more complementary breakthroughs → more profit → more investment.

The delay is what turns that reinforcing loop into a wave instead of an instant explosion. It takes years — often decades — to build the infrastructure, retrain the workers, redesign the factories, and figure out the business models that exploit a general-purpose technology. (Factories didn’t reap electricity’s full productivity gains until managers stopped copying the steam-era layout and reorganised the whole plant around small electric motors — that took a generation.) So the surge builds slowly, accelerates as the loop compounds, over-shoots, and then exhausts the technology’s potential — at which point growth stalls and the economy waits, restless, for the next general-purpose technology to restart the loop. Reinforcing loop + long delay = boom, saturation, bust, pause. A wave.

No moat is permanent — the Red Queen never lets you rest

Now bring the long view crashing back down to the single firm, because it changes the whole meaning of a “competitive advantage.” If destruction comes in waves that never stop, then no moat is permanent. Every defensive advantage an incumbent builds — scale, brand, network effects, patents, a lock on distribution — is eventually breached by the next wave. Not might be. Will be, given enough time.

This is the direct tie to the moats and Red Queen courses. In Lewis Carroll’s Through the Looking-Glass, the Red Queen tells Alice, “it takes all the running you can do, to keep in the same place.” That is the incumbent’s true condition. A moat is not a wall you build once and then relax behind — it is a treadmill you can never step off. The environment (rivals, technologies, customer expectations) keeps advancing, so an advantage that stands still is an advantage decaying in relative terms. To merely hold your position, you must keep reinvesting, keep improving, and — hardest of all — keep self-disrupting: cannibalising your own profitable products before a challenger does it for you. Standing still is falling behind. That is not a motivational slogan; it is the logical consequence of a world where the waves keep coming.

Reading where an industry sits on its wave

Here’s where the long-wave view stops being history and becomes a practical instrument. Point it at any industry and ask one question: is its general-purpose technology early or late in its wave? The answer tells a strategist or investor almost everything about where the risk and the rents are.

An early industry is in the land-grab. The technology is fresh, no one has won yet, and many entrants are pouring in. This is where Schumpeterian rents are richest — get there first with a genuine innovation and you collect above-normal profit before imitators arrive. It’s also where uncertainty is highest and most entrants will die. Early = high rents, high mortality, everything to play for.

A late industry is consolidated. The shakeout has happened, a few large incumbents dominate, the technology is mature, margins are fat and stable, and life is quiet. It feels safe. It is the most dangerous kind of safe. Because a mature, sleepy, high-margin industry is exactly the profile a challenger targets — juicy profits to steal, complacent incumbents anchored by the incumbent’s curse, and a general-purpose technology ripe to be replaced by the next wave. The quiet is not permanence; it is the calm at the top of the wave, right before it breaks.

Two practical readings fall out of this:

• The biggest rents are early — in the build-out, before imitation competes the innovator’s profit away.
• The biggest destruction is in the clearing phase — the shakeout is where the most firms die, so it’s where the model’s carnage concentrates. A late, fat industry has the furthest to fall.

Where the long-wave lens goes wrong (three pitfalls)

The long wave is a powerful lens and, exactly for that reason, easy to over-trust. Three traps catch people:

1. Treating the wave as a precise clock. The single most common abuse is grabbing the “45–60 year” figure and forecasting the next crash to the year, or insisting a downturn is “due” because the calendar says so. Remember the caveat: strict periodicity is empirically weak. The clustering of innovation is the defensible part; the timing is not. Use the wave to understand structure — build-out, saturation, clearing — not to date the future. Over-fitting dates to a stylised pattern is astrology with charts.

2. Assuming every boom is a healthy wave. Not every frenzy is creative destruction laying down useful infrastructure. Some booms are pure bubbles — speculative manias with no lasting productivity gain, where the “clearing phase” leaves behind nothing valuable, just losses and a hangover. The railway mania left a rail network; plenty of manias have left only wreckage. Telling a genuine wave (real general-purpose technology, real complementary innovation, real leftover infrastructure) from a hollow bubble is hard in the moment and is exactly the discipline the next lesson is about: not all destruction is creative.

3. Mistaking a temporary lull for permanent safety. The flip side of pitfall 1. Between waves, or during a mature industry’s long profitable plateau, it’s tempting to conclude the gale has finally, permanently stopped — that this moat really is forever. It never is. A quiet stretch is an interval, not an end. The wave that isn’t visible yet is still forming. Read calm as “between waves,” never as “the waves are over.”

Putting it together

The whole lesson collapses into a single shift of altitude. Zoom in and creative destruction is a duel: incumbent versus challenger, one moat, one crossover. Zoom out and it’s a tide — long waves of building and clearing, powered by innovation that clusters around general-purpose technologies, each wave laying the infrastructure and freeing the resources for the next. The firm-level lessons taught you how an incumbent falls. The long wave tells you the fall is not a one-time hazard to be survived, but a permanent condition to be run inside — forever, on the Red Queen’s treadmill.

Where this goes next

You now hold the temporal view: destruction has a rhythm, it comes in waves powered by clustered innovation, and it never stops — so no moat is permanent and the only stable strategy is to keep running. But notice how much of this lesson kept flagging a caveat. Not every boom is a healthy wave. Strict periodicity is contested. Some destruction leaves only wreckage. The model is powerful precisely because it’s a strong tendency — and a strong tendency is not an iron law.

The final teaching lesson turns that honesty into the main event: where the model lies. We’ll separate genuine creative destruction from mere churn and rent-seeking, face the survivorship bias that makes the winners look inevitable, reckon squarely with the human cost the word “creative” so easily glosses over, and give the incumbents who actually adapt their due. The gale is real — but knowing exactly where the map stops matching the territory is what turns a slogan into a tool.