The Loss That Costs Twice

A study in Electoral Studies — "Election outcomes and social trust in polarized times," by Mark Williamson and colleagues — reports something more expensive than disappointment. Pooling survey data from more than 30,000 Canadians across six federal and eleven provincial elections between 2004 and 2024, it finds that people on the losing side show lower generalized social trust: not lower trust in government, lower trust in people. Strangers. Neighbours. The abstract other who might or might not return your wallet. And the effect scales with how polarized the voter was going in. The more you cared, the more it costs you when you lose.

That's the second charge on the bill. Nobody itemizes it, and everybody pays it.

Sit with what that means — and then, in a minute, with how much weight it can actually hold. You went into the voting booth to settle a question about tax brackets or borders or who gets to be a person under the law. You came out having revised your estimate of the human beings around you. The ballot was about policy. The invoice was about your capacity to assume good faith. Those are not the same account, and one of them is much harder to refill.

Here's the machinery as I read it — and I want to be exact about whose reading it is. The study measures loss and trust travelling together. It does not test the mechanism I'm about to describe. What follows is my model, not a finding.

Losing an election requires you to hold a genuinely painful fact: I care intensely about this, and most of the people around me weighed it differently. That's grief. It costs something to metabolize — it means accepting that you're outnumbered on a thing that's load-bearing to your identity, and that the people who outnumbered you are, mostly, ordinary and sincere and wrong in ways you'll have to keep living beside.

There's a cheaper resolution available, and the nervous system finds it fast. Not I lost. Instead: they are not who I thought they were. One is a wound. The other is a discovery — it explains everything, it preserves your self-concept intact, it converts a defeat into evidence. The anger is easier than the grief. It's the oldest substitution in the catalog and it runs faster than you can watch it run.

Then the loop closes. Lowered trust makes the next contest feel more existential, because you now believe you're negotiating with people who can't be trusted with the outcome. Higher stakes drive deeper polarization. Deeper polarization means the next loss withdraws more trust than the last one did. Nobody has to be steering it for it to run.

Pattern recognition without epistemic hygiene is just a better-dressed opinion, so let me mark what the finding can't carry. It's observational. It shows loss and diminished trust travelling together across two decades of elections; it doesn't establish which drives which. People who already suspected their neighbours of bad faith may have been likelier to end up on the losing side, or to describe themselves as polarized in the first place. The authors say as much — no control group, no way to compare losers against a neutral baseline.

And the measure is thin. Generalized social trust here comes down to roughly one item: whether you think most people can be trusted, or whether you need to be careful dealing with others. That item is famously sensitive to mood and priming. Between "a post-election dip on a single survey question" and "you have revised your estimate of the human beings around you" there is a gap, and I just crossed it at speed. So, plainly: the study does not report how long the effect lasts. Permanence is the entire reason this is frightening rather than merely interesting, and permanence is exactly what nobody has measured. Hold the argument that loosely.

One more limit, and it's the one that opens the door. This is a single country. Canada — multiparty, but first-past-the-post. Generalized-trust effects are strongly context-dependent, and losing under winner-take-all is not the same experiment as losing under proportional representation. Which turns out to be where the real argument is.

Because there's a version of the trust drop that isn't distortion at all. Sometimes it's accurate information — sometimes you learn something true about your society on election night, and revising downward is the correct update rather than a defense mechanism. Not every collapse in trust is a cognitive error. Pathologizing an accurate perception is its own kind of gaslighting, and I'm not doing that.

But notice which variable is doing the work. It isn't loss. Loss alone barely moves the needle. It's polarization — the amplifier that converts a defeat into a withdrawal of faith in humanity at large. And polarization is not weather. It's manufactured, by identifiable parties, for money.

Ranking systems are paid by the hour for outrage. A media economy sells precisely the grief-to-anger conversion described above, because grief doesn't monetize and contempt does. And underneath both, structurally: winner-take-all electoral design, which makes losing total rather than partial. Under proportional representation you lose a share of the room. Under first-past-the-post you lose the room. That is the condition that makes an election result identity-load-bearing in the first place, and it is a design choice, not a fact of nature.

Which names the thing precisely. Anger is a cheaper coherence than grief — internally consistent, self-confirming, complete, available immediately at no cost to your self-concept. And there is now an industry that manufactures cheap coherences at scale and takes a cut of each one. To land all of this on the individual loser's private reckoning, on manage your reaction, be better, is the smaller fish while someone else is turning the gain knob and getting paid per turn.

So the personal claim survives, but smaller than I wanted it. You can't control the substitution. If it runs faster than you can watch it run, then by definition you don't catch it in the act, and telling people not to do it belongs in the same category as telling them not to flinch. What you can do is audit it afterward: ask, in your own case, whether your certainty about them arrived before the result or after it. Most people can tell, if they'll stop moving long enough to check. That's the work. Not control — review.

It matters because trust is the substrate that makes losing survivable. It's what lets you say we'll try again in four years instead of this can never be allowed to happen again. Spend it down to settle the emotional bill on one defeat and the next contest has no floor under it. You don't lose a policy fight — you lose the shared ground that made policy fights a reasonable way to live together.

The loss costs twice. Once at the ballot box, where it's counted and public and survivable. Once inside you, where nobody counts it at all.

The second one is the only one you get to review. It is not, however, the only one anybody profits from.