Exit Markets & the Redesign#
Divorce as an exit market, and the collected bill for every gap this semester left open
Three ideas, before anything else#
- exit market
- sunk cost
- bargaining power
Get these three straight now — the capstone asks you to use all three on a market of your own choosing.
Exit market#
Every market this semester assumed getting in was the hard part. But knowing you can leave — and what that actually costs — is just as much a market as finding someone in the first place: leaving means re-entering week 2’s search costs from scratch, at whatever age and circumstance you’re leaving from, against a pool that’s moved on without you.
A market for finding a replacement for what you’re about to give up, complete with its own search costs and its own uncertain outside options, is an exit market.
Sunk cost#
Four years of shared friends, a lease signed together, a routine that took a year to build — none of it makes staying the right call going forward. A purely forward-looking decision should only weigh what happens next, not what’s already been spent getting here. And yet what’s already spent is exactly what makes leaving feel unthinkable.
Money, time, or effort already spent — that a forward-looking decision should ignore, but that keeps pulling people to stay anyway — is a sunk cost.
Bargaining power#
Two people with identical facts — same outside options, same sunk costs, same years invested — can land on wildly different outcomes, because one of them can credibly threaten to walk and be believed, and the other can’t. Every model since week 9 has quietly assumed outside-option claims are just true, handed straight to a formula.
The capacity to make your threat to leave, or your claim about what you’re owed, actually count for something is bargaining power — and naming it is this week’s whole point.
Now, back to the point#
Alex and Sam, four years past the split in week 9, aren’t negotiating a lease anymore. They’re negotiating whether to stay together at all.
What the exit decision has to model#
| Cost | What it captures |
|---|---|
| Search cost | Re-entering the dating market from scratch, at a different age and life stage than the last time either of them searched |
| Sunk cost | Four years of shared friends, routines, and a signed lease — none of it relevant to what happens next, all of it hard to walk away from |
| Outside option | The value of life alone, or with someone else — exactly as self-reported and unverifiable as Alex’s and Sam’s outside-option claims back in week 9 |
Same facts, different outcome#
| Question | Alex has a real fallback (family nearby, no financial need for Sam) | Sam claims the identical fallback, falsely, and isn’t caught |
|---|---|---|
| Who makes concessions to stay together? | Sam | Sam, again |
| What actually decided that | A genuine difference in outside options | Nothing but whether the claim was believed |
Nash bargaining, applied honestly here, can’t tell these two rows apart. Only bargaining power can — and the model has no variable for it.
A market can tell you the price. It can’t tell you who gets to set it.
Tutorial 12#
Group post-mortem: for the running example you’ve carried since tutorial 1, name one thing the market model explained well this semester, and one thing it visibly failed to explain. The second question is the harder one.
The gaps, collected#
| Week | What the model couldn’t say |
|---|---|
| 1 | Why not being rushed has value the model never priced |
| 3 | Whether a costly signal is even about the thing it claims to signal |
| 4 | What’s actually behind an adverse-selection death spiral, only that one happens |
| 5 | Why the honest type still pays a cost, even once screening works |
| 6 & 8 | Who ends up playing which side of a one-sided algorithm, or gets labelled the proposer |
| 7 | Who breaks a standoff that neither side wants to lose first |
| 9 | Whose outside-option claim gets believed |
| 10 | Which of many sustainable equilibria two real people actually land in |
| 11 | Who gets to design the rule everyone else has to live under |
Nine different “the model can’t say X” moments. Nine sightings of the same missing variable, wearing a different costume each week.
What this framework never modeled#
Naming bargaining power doesn’t explain it. This model still has nothing to say about where it comes from — why one partner’s threat to leave is credible and the other’s isn’t, why one person’s claimed outside option gets believed and another’s doesn’t.
That’s not a gap next week fixes. There is no next week. It’s the question the capstone hands to you, for a market you choose yourself.
That’s the semester#
Twelve weeks modeled how relationships form, screen, bargain, cooperate, and end. The one thing they never modeled is why some people get to set the terms. Find that gap in a market of your own choosing, for the capstone.