Check out the episode:
You can find the shownotes through this link.
Are you interested in a governing system based on betting?
Our debate today works with the article titled Shall we vote on values, but bet on beliefs? from 2000, by Robin Hanson.
This is a great preparation to our next interview with Robin Hanson in episode 390 talking about the opportunities in the betting markets for governance.
Since we are investigating the future of cities, I thought it would be interesting to see how this new form of governance could work. This article presents the revolutionary governing system where the public votes on values but bets on beliefs, with design challenges and opportunities within the system.
[intro music]
Welcome to today’s What is The Future For Cities podcast and its Research episode; my name is Fanni, and today we will introduce a research by summarising it. The episode really is just a short summary of the original investigation, and, in case it is interesting enough, I would encourage everyone to check out the whole documentation. This conversation was produced and generated with Notebook LM as two hosts dissecting the whole research.
[music]
Speaker 1: Today, we’re tackling a really bold proposal for how we could run things, a concept known as Archy. The basic idea is it’s actually pretty simple on the surface, you separate what we value from what we believe. So the public gets to vote on our values. They defy what national wellbeing looks like, this idea of GDP plus, but then betting markets get to bet on beliefs. They decide which policies will actually get us to that goal. And that really brings us to our central question for today, to swapping out our traditional, political and frankly, academic systems for these decentralized markets. Making them the main engine for policy knowledge. Does that actually offer a real path to better government?
Speaker 2: You frame it as this necessary engineering solution, which I get the information. Problems in democracy are significant. You have low incentives for voters to be informed, a failure to aggregate knowledge. Those are absolutely real issues, right? I’d argue that we can’t just throw out our established if flawed institutions for a mechanism that hands over policy to what are at the end of the day financial instruments. We have to really weigh the promise of market efficiency against the huge institutional risks that come with turning governing into a high stakes bet.
Speaker 1: I understand the institutional caution, but let me offer a perspective that’s really rooted in how societies try to get reliable knowledge in the first place. Our current systems are demonstrably failing at this. Just look at the information gaps in democracy. You have widespread voter ignorance. I think it’s something like less than a third of us adults can even name their own representative. You see this reflected in these persistent public opinions that just fly in the face of scientific or economic consensus. How can we possibly expect good policy when the foundation is that shaky? And it’s not just democracy, even academia, which we’re supposed to trust for expert guidance, has its own structural flaws. Studies with selection biases, error rates are shockingly high. And peer review, there’s often very low agreement. We’re relying on systems where the incentive to actually be right is pretty
Speaker 2: weak. You’ve laid out a powerful case against the old ways, but my skepticism is really about assuming that prediction markets are some kind of silver bullet. You point to their success in these isolated tasks like beating government forecasts on orange juice futures, or getting racetrack odds, right? And sure, they work there because of a very specific incentive structure. The market concentrates influence among these super rational informed traders, the market makers, and yes, they’re unbiased on average. They put in more capital, they make fewer mistakes than everyone else.
Speaker 1: Precisely. That’s the key. That disproportionate influence of rational informed traders is the very mechanism we need to leverage. It aligns influence with actual competence.
Speaker 2: That’s a compelling argument for say, forecasting, but I frame it very differently when we’re talking about governance. Yes, voter knowledge has failures, but collective public opinion isn’t always chaos. It’s often quite stable and rational on big domestic issues. More importantly, we tend to just gloss over the known failures of speculative markets themselves. These markets have real pathologies, like excess volatility bubbles, especially on long-term predictions. And they also suffer from things like the long shot bias where people overpay for these low probability high payout bets. This Hayekian hypothesis that markets are graded aggregating information, it’s only roughly confirmed experiments show that if the information gets too complex, you can get these information traps where the market just settles on the wrong answer. You just can’t transplant a mechanism that’s good at predicting printer sales to the immense task of running an entire government. It’s a category error.
Speaker 1: Okay, but let’s ground this in incentives because that’s the real difference maker here. The problem with politics and expert committees is that experts, they simply declare their policy advice without much in the way of clear or direct incentives to be right. The incentives are for popularity, for career advantage, not for being accurate. The core advantage of Charry is that it solves this by making speculators pay for their mistakes. They literally have to put their money where their mouth is. If you consistently back policies that fail, you lose capital, you lose influence, and the market weeds you out. It’s this self-correcting monetary dynamic that forces policy to be guided by the most accurate predictions. Our democratic institutions just don’t have a mechanism like that.
Speaker 2: I’m sorry, but I just don’t buy that. The incentive is that clean, especially when you’re talking about national policy. Let me tell you why. When you attach that kind of massive financial prize to a policy decision, you open the door to huge risks of manipulation and corruption. The paper itself admits these problems first. If a market for a policy is thin and not enough traders, powerful groups could just buy policy via betting markets. They could pour in enough money to swing the prediction their way. It basically turns policy into an auction,
Speaker 1: but that risk is supposed to be mitigated by the rational speculators, right? If some rich fool tries to buy a bad policy, savvy traders should see that as free money and bet aggressively against them correcting the market.
Speaker 2: That’s a lovely theory, but in practice. We’re asking the system to rely on these perfectly rational, perfectly funded traders to always be there to offset someone with a malicious intent. And then there’s the much darker side. The idea that people could do harm to win bets. Imagine a huge corporation or even another country taking a massive bet against a certain policy. They could then credibly threaten to do something that would tank our national welfare unless the decision goes their way and suddenly speculators are forced to bet on the bad outcome. Not because the policy is flawed, but because they believe the threat is real. It’s no longer about finding truth. It’s about institutional blackmail.
Speaker 1: I do acknowledge the severity of those risks, coercion and policy buying. They are serious, but arch’s whole architecture is designed to mitigate that by cleanly separating values from beliefs. You are focusing on the corruption of the means, the market, but the entire system is anchored by the democratically chosen ends, that GDP plus function, that function. It could include discounted GDP, but also things like average lifespan, leisure hours, even measures of inequality. It dictates what we as a society want. The betting markets only figure out the causal path to get there, and the courts are the final backstop. If someone tries to rig the game by encoding a specific project into the welfare definition, say you get a welfare boost. Only if we build this bridge. The courts are supposed to invalidate it, that separation is key.
Speaker 2: That’s an interesting point about judicial review, but I think the definition of GDP plus is actually the fetal flaw here, not the safeguard. You are assuming we can create one function that encodes all our values. How do you even begin to account for complex things like population changes, births, immigration, or the risk of destruction, which is basically anything that prevents the bets from paying off. And when you try to define welfare so completely, you just magnify any measurement errors. It’s the classic problem. When a measure becomes a target, it ceases to be a good measure. Speculators won’t work to raise real welfare. They’ll work to raise measured welfare by finding and exploiting every little loophole. It’s just like that old story of the nail factory that got rewarded for the weight of nails that produced what happened? It started making a few giant useless nails. Future barky takes that exact incentive and applies it to everything.
Speaker 1: Actually has a way to deal with it recursion. So imagine a general policy gets approved at the national level, like increased green energy spending. That policy can then authorize smaller, more localized FUTA keys for the specific decisions like where to build a power plant. It delegates decisions down to the level where the knowledge actually is. And on top of that, there’s a veto mechanism built in any proposal can be immediately vetoed. If a secondary market estimates it would actually decrease welfare one year in the future. It’s a check and balance, a cooling off period to catch buggy decisions before they do real damage.
Speaker 2: I appreciate the thought behind recursion and the veto, but the complexity issue goes way beyond just local decisions or a one-year timeline. Think about the scope of government and externalities, for example. What about military strategy that needs secrecy? The proposal says a FUTA could approve a secret military regime. That relies on speculators agreeing it’s the best option. And if they agree, the market price itself becomes a public signal. You compromise the very secrecy you need. And what about external effects? The whole GDP plus ideas fundamentally about national welfare, our economy or lifespan. But what about international treaties, foreign aid, those require thinking about consequences for foreigners. Political debate for all its flaws forces us to at least grapple with those non-national externalities. U Turkey seems fundamentally myopic in a world that’s so interconnected.
Speaker 1: You’re highlighting critical design challenges, no question. But let’s step back and look at the cost of what we’re doing now. The status quo is failing to aggregate knowledge, and that leads to these massive differences in economic growth and wellbeing between nations. The policy choices we make are immensely important, and right now they’re often guided by very poor information. For tery offers a mechanism with an intense monetary incentive to align policy with accurate knowledge. Even if the design isn’t perfect, defining GDP plus managing secrecy, we have to see this as an engineering problem. The goal should be to test and refine these kinds of prototypes. If we’re really serious about solving this fundamental social problem of how to make good collective decisions, we can’t just ignore a tool that’s so consistently outperforms the experts.
Speaker 2: This has been a great discussion because it really highlights these inescapable weaknesses and how democracy processes information. I totally grant that the incentive problem in politics is acute, but the solution you’re proposing, I think it just swaps one set of systemic flaws for another. We’re trading the low incentives and messiness of democracy for the volatility, the moral hazards, and the perhaps unsolvable problem of defining what it is we even want. Until we can confidently define a corruption proof comprehensive GDP plus function and really eliminate the practical risk of manipulation, I think relying on democracy with all its acknowledged flaws is still the less perilous choice. The difficulty of defining what we want and then guarding that definition from being gamed makes the whole food tar machine just too fragile, I think, to bear the weight of a country.
[music]
What is the future for cities podcast?
Episode and transcript generated with Descript assistance (affiliate link).
What is the future for cities? podcast 
Leave a comment