Economist Writing Every Day

There is much to admire in the bundle of Canadian policy, but housing and construction regulation remains a largely unmitigated disaster.

The Canadian housing market… 😬 pic.twitter.com/pdr2oeLa5B

— Alec Stapp (@AlecStapp) July 9, 2023

I note this only as a reminder that

1. It’s safer to admire individual policies rather than national bundles. There isn’t a nation on earth that gets anywhere close to everything right.
2. What you gain from an optimal policy is often just slack that softens the impact of getting something else completely wrong.
3. Often you only really feel the true cost of bad policy when political tides undermine what was previously buttressing your entire system. Case in point: the NHS and Brexit.

What happens to the Canadian economy when the housing market is still strangling disposable income and an anti-immigrant political movement rises to power on the false but persuasive accusation that immigrants, and not bad housing policy, is to blame? Leveraging all my gifts of analysis and foresight, I predict bad things. Bad things will happen.

[Author’s note: I almost used the word “radical” in the title, but stepped back from the abyss. Being wrong is always more forgivable than being sweaty and clichéd.]

The principal objective of disinformation campaigns is to lower the value of publicly available knowledge so the purveyor’s preferred narrative can compete in the market for ideas. This is particularly attractive to authortarian regimes who have control of large media outlets. If the integrity of public information is sufficiently low, then wholesale fabrications can dominate based on reach and volume.

But what if you are a benevolent democratic regime without a state media outlet trying to compete with targeted disinformation campaigns? Your enemies have flooded the zone with so many conspiracy theories and falsehoods that your ability to steer the public discourse is significantly hindered. How do you recover control of the narratives driving elections and, in turn, policy?

The classic power solution would be to eliminate the disinformation. Constraints on production and dissemination of information, bans on outlets, criminal prosecution for promoting foreign propaganda. Standard command and control governance. But what if the genie is already out of the bottle? Maybe you can limit access to foreign-owned outlets (i.e. TikTok), but eventually everything is going to leak through via other outlets. Half of Instagram Reels is essentially TikTok on a two week delay after all. And that doesn’t solve the problem of domestic disinformation/misinformation. If disinformation has reduced the price of lying to zero, then we should expect news and campaigns to indulge whenever it serves their bottom lines, which means lies will find every crack in the media regulatory firewall, like water on concrete.

(Brief aside: maybe you don’t believe in the horseshoe theory of political politics as it relates to authoritarianism and identity, but it sure does seem to accurately describe affinities for conspiracy theories.)

If you are a benevolent democratic regime seeking to retain office for yourself or your political party, how do you communicate with a public unable to distinguish truth from opposition deception? How do you produce something with signal value when the world is being purposefully and strategically filled with so much noise?

---

What if you didn’t focus on communicating with the public, per se, at least in the short run? What if you gave up on communicating broadly, for a moment, and focused entirely on the subset who could independently extract signal from noise? You’d lose elections, right? There’s not enough “signal extracters” to compete with “noise voters”, are there?

This is going to sound mathy and, at first, elitist, but hear me out. Maybe there are enough signal extractors simply because noise voters cancel each other out. This is not a new theory. This is classic statistics and political economy. If we assume that noise voters are purely random in who they vote for, then the Law of Large Numbers kicks in and you essentially get an even split of noise voters across all candidates, allowing the election-within-the-election “signal extractors” to determine the final winner.

If that all sounds just a little too cute and too convenient, its because it probably is. Assuming that noise voters are randomly distributed across parties and platform is a pipe dream. At this moment in the US and abroad, authoritarians and social conservatives are far more invested in pursuing noise voters, to varying degrees of success, by serving them up bespoke misinformation at every turn. Not that we should expect this to stay constant. As we speak a Kennedy (!) is running for the Democratic nomination on what is essentially a platform of disinformation, conspiracy theories, and pure hokum.

At the end of the day, we have to increase the value of signal campaigns relative to noise. How do we do that? Education! Public service! A recommitment to civic duty! A recommitment to God! A blogging revival! Ha. You wish. Sorry, those are certainly aspirational, if not inspirational, solutions. But I think those whither and die in the face of unrepent bullshit and lies. I have a different answer.

What if the solution to disinformation is more misinformation? But first, an aside.

---

There is a classic story in game theory that professors still put in front of their students to this day: if you had to meet someone in New York City tomorrow, but couldn’t communicate with them, where would you go and when would you go there? When posited to New Yorkers, specifically, and East Coasters broadly, it’s amazing how many people give the same answer: the clock in Grand Central Station at noon.

This is known as a focal point (or Schelling Point). It allows for coordination without communication. Focal points show up in culture and social norms on fairly regular basis simply because they are so useful. They emerge, over time, from thousands of repeated interactions, with certain norms taking hold when they create advantages for their adherents. The seeds of these focal points are when enough people find something useful that it becomes duplicated. Like meeting a friend arriving in town at the train station.

The truth can be a natural focal point, not because it is necessarily pretty or inviting, but because it is actually there.

---

So, again, what if the solution to disinformation is more misinformation? Not debunking the lies and bullshit, but heaping more out the window until it covers every surface? The reason that targeted disinformation works is that it reduces the advantage of telling the truth, allowing your preferred narrative to compete. The weakness of disinformation and lies, though, is that they are nearly costless to supply. Noise voters aren’t shopping for the best answer, they’re shopping for the answer that they would prefer to be true. So give it to them! Give them exactly the answer they want. Give everyone the exact answer they want. Flood the zone to the point of total saturation.

If everyone can find their own truth, then the Law of Large Numbers can actually dominate the outcome. If everyone can be fed exactly the story they need to hear to vote for Candidate A and exactly the story they need to hear to vote for Candidate B, then their vote will be effectively a coin flip. They only votes remaining to be determined outside of our probabilistic system? The signal voters. But it gets better, because the truth has an advantage in this landscape: it’s a superior focal point. If beliefs are blades of grass in a lawn fertilized with pure and utter bullshit, the truth will look like all the others, but it will be just a little taller. As people observe signal voters collecting around it, it will grow and grow until people decide, absent communication, to meet at the tall blade of grass.

---

How do you create such an infinite system of bespoke false narratives for the tiniest slices of the electorate? Targeted large language models. Artificial intelligence. The exact thing that some people fear will destroy democracy and enable authoritarians everywhere. If everyone is receiving their perfect cocktail of flattering, angering, entertaining disinformation, the only people that will determine elections will be those with an abnormal resistance to bullshit. Narratives flooding the internet, produced by a million AIs at a million typewriters, will ensure that each of us will stumble upon the exact sonnet you most want to hear, telling you which aliens caused which problems, which conspiracy cost people jobs, and which reason the world is worse than when you were sixteen.

No one will be fully, purely resistant, but we, each of us, have dimensions on which we actually know what we are talking about. Our own experiences, tacit knowledge, and expertise what will dominate our decision making process and tilt the balance of our vote towards the best outcome. A lot, if not most, of us, have a signal voter within. If our lesser proclivities are nullified in the aggregate by the power of statisics and perfectly curated bullshit, then the political carnival might just leave us governed by the better angels of our nature. A curious, counter-intuitive distillate, to be sure. But maybe also a functioning, more resilient democracy.

Just don’t read the comments ever again.