Institutions, Careers & the Paths We Can't Leave

So far this course has kept path dependence in a comfortable, arm’s-length costume: keyboards, cassette tapes, rail gauges. Easy to nod along to, because none of it is you. This lesson pulls the model off the workbench and points it at the things that actually run your life — the laws you live under, the company that signs your paycheck, the career you’ve spent a decade building, and the habits your brain runs on autopilot. Same engine. Different disguise. The uncomfortable news is that the reinforcing loop that froze QWERTY onto every keyboard on Earth is also the reason you can’t just quit the wrong job, and the reason a country stays poor a century after the accident that made it so.

The single idea to carry through the whole lesson: anything people organise their behaviour around develops increasing returns to staying put. Once everyone’s expectations, investments, and complementary arrangements are built on top of a path, the path is held in place not by being good but by being there — and dislodging it means moving everyone at once. That’s true of a legal code, a business model, and your professional identity in exactly the same way.

Institutional path dependence: why history keeps countries on their tracks

Analogy. Think of an old European city that grew up around medieval cart-paths. The streets are narrow, crooked, and maddening for cars — objectively a bad layout for modern traffic. Yet nobody bulldozes and re-grids it, because every building, sewer, property line, and address is wired to the existing streets. The layout persists not because it’s good but because everything else is built on top of it. Now replace “streets” with “laws, constitutions, property regimes, bureaucracies” and you have institutional path dependence.

Precise definition. The economic historian Douglass North (Nobel, 1993) built much of his life’s work on the claim that institutions — the formal and informal rules of a society — are profoundly path dependent. Once a rule is in place, a web of complementary organisations, expectations, and investments grows up around it: firms structure themselves to exploit it, careers specialise in navigating it, contracts assume it, and voters build their plans on it. Every one of those complements raises the cost of changing the rule, because changing it now strands all of them at once. So institutions persist far past the moment their original rationale died — and, crucially, “history matters” for why some economies stay rich and others stay poor. North’s point wasn’t just that the past influenced the present (trivially true of everything) but that small, early institutional choices got amplified into large, persistent, self-reinforcing divergences between nations that no amount of “just adopt better rules” easily reverses.

Worked example — a rule that outlived its reason. Consider administrative boundaries drawn by colonial powers, often with a ruler on a map, ignoring the geography and peoples underneath. The original rationale — the convenience of a distant empire — vanished with decolonisation. Yet the borders overwhelmingly persisted. Why? Because the moment independence arrived, an entire apparatus had already organised around each line: national governments, armies, currencies, capital cities, civil-service jobs, tax systems, and elite careers were all built on top of the inherited map. Redrawing a border now means dismantling all of that simultaneously and getting every neighbour to agree — a coordination problem so brutal that even manifestly dysfunctional boundaries survive for generations. The reason for the border is dead; the organisation around it keeps it alive. You can see the same accretion in any mature tax code (each loophole and carve-out has a constituency that reorganised around it) or in common-law quirks that persist purely because centuries of contracts and rulings assume them.

Pitfall. The tempting error is to read persistence as proof of efficiency — “if this institution is still here, it must be working, or it would have been reformed away.” That’s exactly backwards. Path dependence says persistence can be a symptom of lock-in, not fitness: the institution might survive because too many parties are organised around it to move at once, not because it’s the best available rule. Never let “it’s still standing” stand in for “it’s still the best design.”

Organisational lock-in: when a moat becomes a cage

Analogy. A championship team keeps running the exact playbook that won it three titles. It’s not stubbornness at first — the playbook works, the players are drilled in it, the coaches built careers on it. But the league adapts, the game speeds up, and the very mastery of the old playbook becomes the reason the team can’t switch to a new one. What made them win is now what makes them lose. Organisations do this constantly.

Precise definition. Firms accumulate routines, business models, and core competencies — “the way we do things here.” These are genuine assets: they’re why the firm succeeded, and they compound (better routines → more success → more investment in those routines). But that same compounding creates core rigidities (a term from the strategy scholar Dorothy Leonard-Barton): the competencies that made you win get so deeply embedded — in your hiring, your metrics, your incentives, your identity — that you can’t abandon them when the world changes. This is the competency trap: getting very good at the wrong thing, precisely because being good at it kept paying off right up until it didn’t. It’s the incumbent’s version of QWERTY — locked onto the layout that used to make sense.

Worked example — the cash cow that couldn’t be killed. The classic shape: a company invents or dominates a category, builds its entire cost structure, sales force, factories, and executive incentives around that cash cow, and then meets a discontinuity it is structurally incapable of embracing — even when it sees the discontinuity coming clearly. A film-and-camera giant that literally invented an early digital-imaging prototype in-house still couldn’t cannibalise its enormously profitable film business to chase digital, because every internal loop — sales commissions, factory investment, executive bonuses, the mental model of “we are a film company” — pulled toward protecting the cash cow. A dominant video-rental chain built on late-fee revenue and physical stores couldn’t reorganise around mail-and-streaming without gutting the very engine that funded it. In each case the firm wasn’t blind; it was locked in, held by the complements built on the old path exactly as North’s institutions are.

Pitfall. The seductive mistake is “the incumbent was just badly managed / not innovative enough.” Sometimes true — but the deeper, more useful reading is that success itself manufactures the rigidity. The routines were correct answers to the old environment; the firm optimised into them because they worked. Blaming stupidity lets you believe it won’t happen to you. Recognising it as structural lock-in — the predictable price of getting very good at one path — is the only version that helps you watch for it in your own organisation.

The sunk-cost trap: what path dependence feels like from the inside

Analogy. You’re 90 minutes into a three-hour film that turned out to be dreadful. Do you stay? The money and time are already spent whether you leave or not — a rational chooser weighs only the next 90 minutes against anything else you could do — yet the pull to stay “because I’ve already sat through so much” is almost physical. That pull is path dependence experienced from inside a single human skull.

Precise definition. A sunk cost is a cost you’ve already paid and cannot recover, no matter what you do next. The sunk-cost fallacy is letting those unrecoverable past costs drive a forward-looking decision — continuing down a path because you’ve invested in it, rather than because continuing is the best remaining option. This is the subjective face of path dependence: institutions lock in through external complements, but you lock yourself in through the felt sense that abandoning the path would “waste” everything you poured into it. Leaving feels like destroying value — even though the value is already gone either way.

The critical distinction — and it’s the whole toolkit. Not every reason to stay on a path is a fallacy. Draw a hard line:

The clean mental move: imagine someone dropped into your exact situation today with none of your past investment. If they’d still stay, your reason is a real switching cost. If they’d leave and only your history keeps you, you’re paying a sunk cost — and sunk costs are, decision-theoretically, none of your business.

Worked example. “I’m three years into a law degree I’ve come to hate, so I’ll finish it — I can’t waste three years.” Split it honestly. The forward-looking part: finishing might be cheap now (one year left, the credential opens doors) — that’s a legitimate reason and might well justify staying. The sunk-cost part: “I can’t waste the three years” is pure fallacy — those three years are spent identically whether you finish or quit, so they should carry zero weight in the decision. People routinely staple the two together and let the emotional sunk-cost half smuggle a bad decision past them. Un-staple them and the real question emerges: given where I am now, is the next year of this the best use of that year?

Pitfall. Over-correcting into “sunk costs never matter, so I should feel free to quit everything.” Two traps hide here. First, switching costs are real — resetting seniority, rebuilding a network, and relearning a field are genuine future costs that often do justify staying; dismissing them as “just sunk costs” is its own error. Second, serial quitting has its own compounding downside: bailing on every path the moment it gets hard means you never accumulate the increasing-returns assets that make any path valuable. The skill isn’t “always quit” or “always persist” — it’s cleanly separating the forward-looking ledger from the backward-looking one, then deciding on the forward one alone.

Career and skill lock-in: your résumé is an increasing-returns asset

Analogy. A specialist’s career is like planting a single enormous tree instead of a garden. Pour everything into one trunk and it grows tall, strong, and shade-giving faster than any diversified plot could — glorious as long as the climate suits that tree. But you can’t transplant a mature oak, and if the climate shifts, you’re standing under the one thing you can’t move. Depth and immovability are the same property viewed from two angles.

Precise definition. Your skills, professional network, reputation, and salary are classic increasing-returns assets, and they’re bonded to one path. Each year in a field, you get better at it (learning effects), you know more people in it (network effects), your reputation there deepens, and your pay tracks that accumulated specialisation. They compound — you already met compounding in the latticework — and compounding is spectacular while the path is good. But every one of those compounding assets is also a switching cost: change fields and you reset the learning curve, leave most of the network behind, start reputation near zero, and take a pay cut. Specialisation deepens the moat and the cage in the same motion. This is what “golden handcuffs” names precisely — the very rewards that make a path lucrative are what bind you to it.

Worked example. Two people, same starting point, at year twelve. Ada went deep: one industry, one specialty, now a top-paid recognised expert. Ben kept a portfolio: solid but less specialised skills, a broader network, more transferable tools. While the industry thrives, Ada is unambiguously ahead — her focused compounding won. Now the industry gets disrupted. Ada’s assets are worth a fraction elsewhere: her specialised skill doesn’t transfer, her network is inside the sinking field, and her senior salary makes her expensive to hire into a junior slot in a new domain. Ben takes a smaller hit — his transferable skills and broad network port to the next thing. The lesson isn’t “don’t specialise” — Ada’s depth was the right bet for over a decade and paid enormously. It’s that deep specialisation buys present returns by selling future option value, and you should make that trade knowingly, not by drifting into it.

How to think about it. Three moves keep the upside of compounding without walking blind into the cage:

1. Keep some option value. Deliberately preserve a little slack — a transferable skill, an outside relationship, a side project — that survives if your main path collapses. It costs a bit of focus and buys an exit.
2. Bias investment toward transferable skills. Between two ways to get equally good, prefer the one whose skill ports to other paths (general judgement, communication, tooling everyone uses) over the one that’s valuable only inside your current silo.
3. Watch the direction of the compounding. Compounding is only wonderful when the path is good. The dangerous case is a path that’s still compounding — you’re getting more senior, more expert, better paid — in a declining field. You feel successful right up to the moment the field goes, because the local signals (raise, title, respect) stay green while the ground erodes. Ask periodically: is my path compounding me toward more future options, or deeper into a single corner?

QWERTY of the mind: paradigms, habits, and cultural lock-in

Analogy. A shared language is the purest QWERTY there is. English spelling is a mess, but you can’t unilaterally reform it — the value of a spelling is entirely that everyone else uses the same one, so the “worse” convention holds because coordinating a switch is nearly impossible. Every adopter makes the standard more binding on the next.

Precise definition (kept brief). The same increasing-returns loop runs inside heads and cultures. A scientific paradigm, a professional’s mental model, a personal habit, and a cultural convention all lock in through a reinforcing loop of shared expectation and repeated use: the more a way of seeing or doing is practised and assumed by others, the more costly it is to think or act differently, so it entrenches — regardless of whether a better frame exists. A field can stay stuck on an outdated paradigm not because it’s correct but because every career, textbook, and grant is organised around it (institutional lock-in again, now made of ideas). A habit persists because the neural and situational complements built around it make the default cheaper than the alternative every single time you choose.

Pitfall. Don’t over-apply this into “all conventions are arbitrary lock-ins we should smash.” Many shared conventions are held in place by lock-in and also genuinely good — coordination itself has value, so a merely-adequate standard everyone shares can beat a superior one nobody else uses. The useful question is never “is this locked in?” (lots of good things are) but “is this locked in and worse than an available alternative I could actually coordinate a move to?”

The practical toolkit: choosing paths you can leave

Analogy. Some doors lock behind you; some you can walk back through. Before stepping through any door, the cheapest thing you can check is which kind it is — because a wrong step through a one-way door costs far more than a wrong step through a revolving one. Most people obsess over which room looks best and forget to check whether the door swings both ways.

Precise definition — the whole payoff in three questions. Path dependence’s practical lesson is not “avoid commitment” (commitment is where compounding returns come from) but “commit with your eyes open about reversibility.” Before stepping onto any significant path, run three checks:

1. Reversibility / option value first. Ask not only “how good is this path?” but “how easy is it to step off later?” A slightly-worse path you can exit cheaply often beats a slightly-better path that locks you in — because reversibility is itself worth a lot when you’re uncertain, and you are always more uncertain than you feel.
2. Prefer reversible moves when future switching costs are high. The higher the increasing returns a path will accumulate — the deeper the specialisation, the bigger the network, the more everything will get built on top — the more you should favour a move you can undo, because the lock-in you’re signing up for is correspondingly stronger.
3. Assume you’re under-weighting future switching costs. Here’s the systematic bias to correct for: when a path looks attractive now, humans reliably under-estimate how hard it will be to leave later. The complements that will bind you — the skills, the sunk investment, the organised expectations — don’t exist yet, so they’re invisible and feel free. They won’t stay invisible. Deliberately inflate your estimate of future switching costs before you commit.

Pitfall. The over-correction is commitment-phobia — treating all lock-in as danger and refusing to ever go deep, keeping every option open forever. But permanent optionality is its own trap: you never accumulate the increasing-returns assets (expertise, network, reputation, a moat) that make any path pay off, and you drift as a permanent generalist beaten by specialists everywhere. The model doesn’t say “never commit.” It says: commit deliberately, prefer reversible commitments while you’re uncertain, and price in the future switching costs your present enthusiasm is hiding from you.

Practice: sort the decisions

Try the two most confusable distinctions in this lesson. First, sort real decisions by how committing they are — whether stepping off later is cheap or expensive. Then we’ll separate the two forces that keep people on a path.

Recap quiz

Where this goes next

You now know that path dependence isn’t a museum piece about keyboards — it’s the physics of the institutions, companies, and careers you actually inhabit, and you have a toolkit for choosing paths you can still leave. Which raises the natural, hopeful question: if lock-in is this powerful, can it ever be broken?

Lesson 6 answers it honestly, from both sides. First, how locked-in systems do get overturned — the shocks, discontinuous leaps, coordinated switches, and deep-pocketed sponsors that occasionally pry a market or an institution off its frozen path. And then the intellectual honesty this course has promised throughout: the three ways the path-dependence model lies — the just-so-story trap (retrofitting “lock-in” onto any outcome you dislike), mistaking mere causation for genuine lock-in, and assuming lock-in is forever when much of it quietly isn’t. A model you only ever see winning is a superstition; next lesson, we stress-test this one.