David Friedman is an anarcho-capitalist economist and legal scholar, author of works including Machinery of Freedom and Legal Systems Very Different from Ours. This conversation covers dating markets, the future of reputation and contract enforcement, Bitcoin and cryptocurrency, prediction markets, the surprising uniformity of modern legal systems, Friedman’s theory of property rights and commitment strategies, automation and inequality, and advice for young people interested in technology.
Dating markets and marriage
The simple economic model of marriage treats everyone as identical and imagines an implicit “price” in terms of the terms of marriage (how much the husband does, how much decision-making power the wife gets, etc.). In this model, legalizing polygamy actually benefits women and harms men, because men competing for multiple wives bid up the “price” of a wife — meaning monogamous men must offer better terms. The reverse holds for polyandry.
The harder problem is sorting: people differ not only in how desirable they are as spouses, but in how desirable they are to specific others. Friedman’s own wife was, from his perspective, a “one in ten thousand catch,” but other men didn’t recognize it because tastes differ. This makes the dating problem fundamentally about search costs, which economics handles poorly.
Online dating should have helped — in principle, computers can screen millions of profiles for objective characteristics and narrow the field before people interact. In practice, it hasn’t clearly improved outcomes. Friedman notes that even successful platforms like OKCupid reportedly got worse after being taken over by businesspeople rather than enthusiasts.
Arranged marriages vs. courtship: Friedman recounts a conversation with a woman from southern India whose parents chose her husband (she had a veto but didn’t do the searching). She was still happily married, while his own first marriage had ended. He is genuinely uncertain whether the modern Western system of courtship works better or worse than the traditional system where parents search and children approve. A clean test would compare Indian immigrants in the US who marry partners chosen by their parents versus those who choose freely — this would separate quality of choice from cultural support for marriage.
On “efficiency”: Friedman’s benchmark is a “bureaucrat god” — an omniscient, omnipotent, benevolent actor who could reallocate everything optimally. No real system achieves this, but competitive markets come closer than central planning because they aggregate decentralized information and give each person control over themselves. The dating market is certainly inefficient compared to the ideal, but the interesting question is whether it beats the main historical alternative (parental choice), and Friedman doesn’t claim to know the answer.
The future of reputation and contract enforcement
In a world of strong privacy through encryption, where two parties can transact without any third party observing, contracts could still be enforced through reputational mechanisms — but this world has downsides too, such as enabling cyber-ransomware and anonymous kidnapping payments via cryptocurrency.
Friedman is not predicting this world will happen; he is exploring what it would look like if it did. He notes that the NSA has spent decades trying to prevent such a world, and that governments doing less controlling would have both advantages and disadvantages.
Reputation already matters enormously in commerce — Friedman’s negative online review of a solar battery installer got an immediate response and a $500 discount. Most store behavior is driven more by reputational concerns than legal threats.
“Woke” signaling and reputation: Friedman is skeptical that political agreement is a reliable reputational signal, since people within ideological groups frequently attack each other. He expects reputation to become less tied to real-space identity over time, not more, because online interactions with strangers (e.g., dealers in India he’s never met) already work on reputation without real-space links. Platforms like eBay and Amazon already provide reputational ratings for transactions.
Honor culture is, in Friedman’s view, the real-space version of what online reputation systems do more effectively. The internet’s effect on reputation is analogous to the automobile’s effect on courtship — a technological change with obvious effects and less obvious ones (closed cars gave couples private space away from community surveillance).
How Friedman predicted (and missed) Bitcoin
Friedman wrote a chapter on e-cash in Future Imperfect (2008, the same year as the Bitcoin white paper). He imagined a centralized issuer — a private bank or institution — creating anonymous digital cash backed by dollars or commodity bundles, similar to Chaum’s digital cash from decades earlier.
What he missed: Bitcoin’s decentralized architecture with no issuer at all. The reason Chaum-style digital cash never happened is that it requires a reliable issuer in a country with functioning legal institutions, and no such country would tolerate money laundering becoming unenforceable. Bitcoin solved this by having no issuer — it just exists online.
Bitcoin’s problem: its price is too unstable for use as a transactional currency. If a price is set in Bitcoin, the seller doesn’t know what value he’ll receive, and buyers can’t easily compare prices over time.
Stablecoins are the missing piece: Friedman has sketched a mechanism on his blog for a cryptocurrency whose units automatically subdivide or consolidate to maintain a stable value (e.g., pegged to a basket of dollars, euros, and Swiss francs). The technical challenge is the oracle problem — how the software learns real-world exchange rates. He notes that Ethereum and other platforms are working on smart contracts that require exactly this kind of real-world information.
Friedman’s thinking on cryptocurrency was partly shaped by early interactions with cypherpunks, where ideas were exchanged back and forth — he borrowed from them and they borrowed from him.
Prediction markets
Friedman has known about prediction markets longer than almost anyone because he corresponded with Robin Hanson before Hanson became an economist. Hanson’s insight was simple: a stock market’s nominal function is allocating capital, but as a side effect it generates information (prices reveal what informed people believe). Prediction markets do this deliberately.
Futures markets for agricultural products are already prediction markets — if the future price of wheat is much higher than the spot price, the “smart money” expects a bad harvest.
Current prediction markets seem to have limits on bet sizes and house cuts. For them to work well, they need to allow large bets, so that informed money (people who care about profit, not ideology) dominates. If only small bets are possible, prices get driven by ideologues. Friedman cites a case where a US presidential election market briefly surged for one candidate — likely someone trying to manipulate perceptions — but smart money quickly corrected it.
Regulation vs. technology as barriers to progress
Friedman previously thought regulation couldn’t stop progress globally because there are many independent countries. He now thinks he may have been too optimistic.
COVID-19 challenge trials illustrate the problem: the vaccine was developed in less than a week; the rest of the time was testing. Challenge trials (vaccinate healthy young adults, deliberately expose them to the virus) could have had a usable vaccine available in perhaps two months. No Western country was willing to do them. Friedman suspects this reflects a shared elite culture where regulators across countries imitate each other and seek each other’s approval — Robin Hanson’s “global elite” theory.
Nuclear power is another case where regulation may have globally curtailed technical progress through the same mechanism.
Friedman suspects that even something as desirable as ending aging would face regulatory barriers in some countries (e.g., the FDA demanding lengthy testing, or governments worried about Social Security bankruptcy), but probably not all of them — so progress likely can’t be fully stopped.
The surprising uniformity of modern legal systems
Friedman’s book Legal Systems Very Different from Ours deliberately avoids comparing modern state legal systems (Japanese vs. European vs. US) because they are too similar. He looks at systems like medieval Icelandic law, pirate law, prison gangs, and traditional Somali law.
Modern legal systems are remarkably uniform: no country uses marketable tort claims, or requires plaintiffs to bear risk for filing unsuccessful suits, or has genuinely non-state legal systems. Even countries that talk about following Sharia mostly have statutory systems borrowing some rules from Islamic law rather than the traditional scholar-based system.
Friedman speculates that the uniformity comes from elite imitation — regulators and law professors worldwide know each other, talk to each other, and copy what seems to work elsewhere. There’s also a practical incentive: if you’re doing business across borders, similar legal systems are easier to navigate.
Non-state legal systems (university rules, corporate rules, prison gangs, pirate codes) show much more diversity. Friedman suspects that if you looked at all non-state legal systems, you’d find far more variety than among state systems.
Even seemingly independent domains like university rulebooks may be more similar than expected, though Friedman notes it’s unclear whether they resemble US criminal or tort law.
Friedman’s theory of property rights and commitment strategies
The core idea (from his article “A Positive Account of Property Rights” and Machinery of Freedom, 3rd edition): property rights arise from commitment strategies. In nature, animals mark territories and fight more desperately the farther into their territory a trespasser comes. A fight to the death is a loss for both sides, so if one side can credibly commit to fighting at any cost, the other backs off.
The human equivalent is not primarily geographic — it’s boundaries in “rights space.” People draw lines such that if someone crosses them, they will bear unreasonably large costs to fight back. The boundary must be at a Schelling point — a natural division both sides recognize (like 50-50 splits between bank robbers who don’t want to keep arguing until police arrive).
Civil order exists because people have reasonably consistent pictures of each other’s commitment strategies. Your neighbor won’t accept you dumping trash on his property, even if fighting you legally costs more than tolerating it, because surrendering would signal that you’re a wimp others can push around.
Government, in Friedman’s definition, is “that organization against which we drop our commitment strategies.” You resist an individual extorting you at high cost; you don’t resist a government the same way because (a) it’s much more powerful, and (b) other people don’t interpret your surrender to government as weakness the way they would surrender to a private individual. The UK sent its only aircraft carrier to fight Argentina over the Falklands because if they didn’t, it might be Gibraltar next — they were maintaining a commitment strategy.
The poor and property rights: those with fewer resources relevant to conflict are less able to maintain commitment strategies. But the relevant resource isn’t always money — in medieval Iceland, it was how many competent warriors your kin and friends were. People can also form coalitions by contract (as in Somalia, where “piles of shields” are contractual defense groups, not kinship-based) to pool their capacity for retaliation.
Automation, AI, and inequality
Friedman is skeptical of the “Moore’s Law for Everything” thesis (associated with Sam Altman) that automation will make a few people enormously wealthy while leaving large segments of the population worse off.
His first approximation: why can’t the non-automated population keep doing what they were doing before, with the same technology, trading among themselves? They’d be about as well off as a productive but low-wealth society (e.g., Hong Kong in 1950). It’s hard to see how they’d be worse off.
Comparative advantage still applies: if techies become 100x more productive at some things and only 2x more productive at others, non-techies can focus on the things where the gap is smallest and trade. The more different your trading partners are, the greater the gains from trade for both sides.
The pattern Friedman does* expect is that some people get a lot better off and others get a little better off — not that the poor get absolutely worse off. This may already be happening: technological progress increases the value of intelligence relative to strength, benefiting techies, doctors, and lawyers.
The horse-and-car analogy (raised by the interviewer): a horse retained some comparative advantage over a car, but horses don’t own themselves. Friedman acknowledges this is a real concern — if AI can do almost everything humans can do, the analysis changes — but he still thinks there will be things humans can do that computers can’t (personal servants, childcare, live entertainment), and those activities can be traded for the things computers produce more cheaply.
Economics of medieval reenactment
Friedman’s participation in the Society for Creative Anachronism (SCA) gave him a more sympathetic understanding of gift cultures. In the SCA, it feels more natural to do favors and donate materials (like gemstones for medieval jewelry projects) than to charge money, even though he does sell a medieval cookbook with his wife.
He recognizes that gift cultures exist within modern life too — when friends invite you to dinner, they don’t expect you to pay them $20. He finds it economically interesting that people sometimes forgo the convenience of money in exchange for the accounting burden of tracking favors, but he doesn’t fully understand why.
Advice for young futurists
Friedman would be tempted by computer programming and cryptography, but notes that the most interesting fields attract the most smart people, making it harder to stand out. He left theoretical physics for economics partly because physics was “overpopulated with high IQ people” and he could say more original things in economics.
What he’d most enjoy: independent programming projects — inventing things equivalent to Tetris, or building tools to help teach economics (he has ideas for programs to accompany his price theory textbook). He finds making machines that work deeply satisfying, and programming makes this much easier than physical engineering.
He also writes novels for fun but recognizes he lacks the talent of truly good fiction writers (including his younger son, who thinks about narrative structure more carefully).
