What this covers

Matt Levine discusses how financial markets work and where his understanding breaks down. The conversation ranges across derivatives pricing, cryptocurrency, mergers and acquisitions, market structure, and regulatory priorities, but returns repeatedly to a central tension: whether the Efficient Markets Hypothesis can account for what actually happens in finance. Levine argues that attention to specific risk exposures—measuring an interest rate swap by its actual delta rather than notional size—and recognition of human incentive structures make sense of most phenomena. Yet cryptoassets present a genuine puzzle. Bitcoin and its peers behave like stores of value untethered to fundamentals, their prices apparently decoupled from incorporation of real information in ways that weaken his prior confidence in market efficiency.

The conversation pulls in adjacent terrain: how indexing creates a feedback loop toward fewer, larger public companies; why mergers persist despite empirical evidence they destroy value; how moral hazard might worsen if clearinghouses replace banks in derivatives settlement; and whether regulators should focus enforcement on institutional bond traders or retail fraud. Levine contends that the IPO pop is consistent with market efficiency, that Regulation FD contains a striking tension with actual practice, and that perceived asymmetries in shorting costs and exchange withdrawal friction may explain why some cryptoassets trade above fundamental value. He also defends jargon as efficient shorthand, argues that the SEC should police egregious ICO fraud to preserve market confidence, and suggests that finance has shed some of its cultural mystique as tech companies have absorbed the mantle of the mysterious priesthood.

Sharpest takeaway

Levine argues that financial markets are best understood through an efficient-markets lens combined with attention to specific risk exposures and human dynamics, but that genuinely weird phenomena like cryptoassets challenge that framework because their prices appear decoupled from fundamentals.

  • Derivative size by notional is misleading; what matters is the specific risk exposure
  • Crypto behaves like a store of value untethered to fundamentals, weakening pure efficient-markets confidence
  • Markets may have decoupled financial asset prices from human emotional reactions to news

The argument · threads9 threads · 34 claims
0.75

Cryptocurrency has weakened efficient-markets belief by showing Bitcoin's price ignores fundamentals and disruption can come from outside banking's structured web.

6 pointscentrality 5/5
  • Crypto has weakened Levine's efficient-markets fundamentalism because Bitcoin's price behavior does not reflect incorporation of fundamental information, and Bitcoin demonstrates that something big can disrupt the financial system from entirely outside its structured sociological web of banks, prime brokers, hedge funds, and retail investors—contrary to his prior belief that entrenched relationships make such disruption very hard.

    “Whatever Bitcoin is doing, it's not that. It is just weird on that basis.”

  • Bitcoin's first-mover advantage and fame can entrench it, and if it stays above a threshold price for a year, that increases its likelihood of staying above that price for another year, so by force of repetition it eventually becomes a store of value—contrary to the crypto community's overrated belief that people will use a hundred different currencies for a hundred different activities.

    “if Bitcoin stays above X thousand dollars for a year, then that increases its likelihood of staying above X thousand dollars for another year. Eventually it, by force of repetition, becomes a store of value.”

  • High Bitcoin prices, negative real (and sometimes nominal) yields on government securities, and high blue-chip stock prices may all be part of one phenomenon—a scarcity of good stores of value relative to growing global wealth—a transmuted version of the financial crisis story where insufficient supply of low-risk assets led people to build supposedly low-risk assets that turned out to be risky.

    “Sure, it's a transmuted story of the financial crisis. The supply of low-risk assets was insufficient to meet the demand, and so people built things that were low-risk financial assets that turned out to be risky.”

  • The best available (though bad) model for valuing Bitcoin is to treat it as a store of value not tethered to anything else, and estimate that some percentage of world financial wealth currently held in stores like gold would shift into Bitcoin; the older currency/transaction-velocity model is no longer persuasive because proponents now claim Bitcoin is a store of value rather than a transactional currency.

    “Some percentage of people's wealth is in gold. If they shifted half of that to Bitcoin, Bitcoin would be worth X. That is the bad but best available model.”

  • Some cryptoassets are probably too highly valued due to technical limits to arbitrage: a perception (real or not) that shorting them is hard, the inherent extra danger of shorting something super volatile versus going long, and exchange withdrawal limits where it is easy to buy bitcoins but slow and complicated to sell them and withdraw cash—so the risks of buying and selling are asymmetric.

    “I do think that there are some obvious limits to arbitrage. One is that there's a perception that it's hard to short a lot of cryptoassets.”

  • Uber's long-term bet relies on dominating self-driving cars, but this is not super compelling because building a self-driving car is hard while building the routing app to dispatch one is relatively easy, so if Tesla or Apple builds the best self-driving car, consumers will simply download that company's app, making the routing app a weak moat despite its global network effect.

    “it's hard to build a self-driving car. It's relatively easy to build a routing app to send the self-driving car to you.”

0.48

Risk in derivatives should be measured by actual exposures like delta and DV01, not by notional figures that misrepresent true economic danger.

2 pointscentrality 4/5
  • The right way to think about derivatives markets is by their actual risk exposures (DV01 for interest rate swaps, delta for equity swaps), not by quadrillion-dollar notional figures, because the probability of losing the full notional on short-term interest rate swaps is extremely low and notional itself is not a meaningful risk factor.

    “The right way to think about it, in the way that you would do it if you're actually working in the derivatives market, is to think about the risk exposures of it.”

  • The IPO pop is consistent with efficient markets because an IPO is a huge price discontinuity rather than a continuous price series; the banks do non-efficient-market M&A-style work negotiating one-off with big investors to find a price, then deliberately sell about 15 percent below it because everyone—issuers keeping most of the stock and IPO investors—benefits when the stock rises and is harmed when it falls, so no one has an incentive to overprice it.

    “The IPO is a huge discontinuity. You don't have a price and then you have a price.”

0.48

Crypto lacks the mature infrastructure derivatives have; no one understands what Ether is, making crypto instruments weirder rather than better at hedging known risks.

2 pointscentrality 4/5
  • Mature derivatives markets reflect the ability to understand and fine-tune exposures to a well-understood underlying, but cryptocurrency is the opposite: no one really knows what Ether is, so instruments like CryptoKitties make exposures weirder and more complicated rather than hedging or precisely magnifying known risks.

    “I think of derivatives markets as being a mark of maturity and of the ability to understand what exposures exist in some instrument or in some economic reality.”

  • The crypto frenzy is not analogous to businesses issuing gift certificates, because no store quintuples its market value by issuing gift certificates; if the outcome were merely that each business has its own low-velocity token like Starbucks accepting Starbucks tokens, that would be a sad and uninteresting outcome for the thousands of cryptocurrencies.

    “No store quintuples its market value by issuing gift certificates. The crypto frenzy is not . . . If this were a story of gift certificates, it would be a lot less interesting.”

0.44

Writing succeeds by focusing on comparative advantage and matching one's personal working style rather than covering whatever topic others ignore.

9 pointscentrality 2/5
  • It is easy to avoid writing what everyone else writes by focusing on comparative advantage: even a really good Trump take would only be the 50th best of 12,000 written that morning, whereas writing about something esoteric like KodakCoin or gaming of FIRC regulations makes you one of very few or the only voice, and readers welcome the relief from political news.

    “even if I write a really good take on Trump, it will only be the 50th best take on Trump of the 12,000 written that morning, whereas if I write about KodakCoin, I'll be only 1 of 20 people writing about Kodak.”

  • Levine has no formal framework for deciding whether something is weird; he uses 'weird' as a near-synonym for interesting, writing about whatever matches his own biases toward finding things interesting, funny, and quirky, which has resonated with an audience.

    “I like weird, so when I use weird it's often just a synonym for interesting.”

  • Levine's writing productivity rests on ruthlessly focusing on places where he has an advantage or can add value—never writing about something just because someone needs to—and on matching his personal metabolism (a daily panic deadline producing reliably, whereas a weekly deadline produces only every three weeks).

    “I try to be pretty ruthlessly focused on places where I have an advantage or where I can add value.”

  • The valuable and weird part of being a classics major is the sense of direct engagement with humans who lived 2,000 years ago—experiencing their consciousness through their own language, as in Cicero's personal letters—and discovering how many of their concerns are both similar to and strikingly different from our own.

    “this sense of direct engagement with humans who lived 2,000 years ago.”

  • Studying law is easier than studying classics because the dense texts analyzed in law are mostly in English and embedded in everyday culture (through shows like Law and Order), whereas reading Greek poetry like Thucydides requires far more investment for someone not raised speaking ancient Greek.

    “If you're not raised speaking ancient Greek, it is much easier to read even a 19th century legal decision than to read Thucydides.”

  • US public companies are now fewer, older, bigger, and more profitable, which makes indexing more rational because the spread between companies and the returns to stock picking are lower; conversely, the rise of indexing and focus on scale makes a $100 million IPO or a weird non-index company less attractive to large funds, creating a feedback loop where indexing drives the trend toward fewer, larger, more profitable public companies.

    “in a world where all companies are the same — they're established, they're profitable, the spread between companies is narrower — the returns to the stock picking are going to be a little lower, and it's more rational to index.”

  • We are almost certainly not sufficiently prepared for hacks; experts warn the world will end horribly if you don't change your password and they are probably telling the truth, and as attacks progress from stealing email addresses to jackpot ATM attacks to compromising the central computers our financial lives and identities depend on, the consequences become catastrophic.

    “It seems overwhelmingly likely to me that we're not worried enough about hacks.”

  • Footnotes have become more appropriately valued because the internet, through hyperlinks, has valorized intertextuality—having a main narrative line plus asides—diminishing the older notion that a single distraction-free narrative is the highest achievement of writing.

    “The notion of an intertextuality, like having multiple lines of the text, having a main line and having asides, is a valuable notion and one that, I think, has been valorized by internet writing in a way. You have links.”

  • Book 9 of the Iliad is a footnote to all Western culture because the canonically most famous heroic warrior, offered the choice between immortal fame through death in battle or a long forgotten life on his farm, initially chooses to be forgotten—forcing an existential examination within a culture that valorizes deathless fame, prefiguring later existentialist thought.

    “The forcing of that choice is the central point of the highest work of Greek art, sort of prefigures a lot of existentialist thought in the future, I think.”

0.38

Mergers persist despite empirical evidence they underperform because empire-building bias and the human desire to act let people persuade themselves they will beat the odds.

3 pointscentrality 3/5
  • Empirical data showing acquisitions are bad on average is trustworthy but doesn't mean your merger will be bad; mergers persist because of empire-building bias and because everyone can cite reasons their deal is better and roughly 49 percent will be right, so charismatic bankers and the human desire to do stuff let people persuade themselves they are in the winning minority—closely analogous to active management.

    “The data is like, on average, they're a little bad. So you say, "Here are the reasons why we are better." Everyone can say that, and 49 percent of them will be right. You throw in some empire building”

  • An efficient-markets believer should not worry that people aren't worried enough, because one assumes a giant pool of people smarter than oneself has reached a reasonable conclusion; if a price like the VIX were too low it would be higher, so apparent mispricings should be deferred to rather than second-guessed.

    “I assume that there's a giant pool of people who are smarter than I am, and they're coming to a reasonable conclusion.”

  • Low volatility may reflect that financial asset prices have decoupled from human emotion about the non-financial world, because as people build more rational investing tools markets react less to noisy news; the strongest counterargument is that investing is not a solvable technological problem but an interpersonal fight where as you get better someone else gets better, leaving human biases as the residue.

    “The explanation that I like is that the prices of financial assets have somewhat decoupled from human emotion about the world that is not financial assets.”

0.38

SEC enforcement should prioritize protecting retail investors from clear fraud like pump-and-dump schemes and ICO scams over policing informed institutional markets.

3 pointscentrality 3/5
  • An SEC enforcement division should prioritize protecting uninformed retail investors from fraud, including egregious ICO frauds that make the market look bad, because allowing people to run pump-and-dump groups and high-five each other publicly undermines confidence in markets even when the victims are sophisticated.

    “You don't really want to be the country where people are committing fraud, running pump-and-dump groups, talking about it, having articles written about it, and high-fiving each other. It doesn't seem to encourage confidence in the markets.”

  • Regulators and prosecutors are drawn to gray-area institutional bond market cases because they involve bigger trades, bigger penalties, and are sexier than retail fraud, but this is less bang for the buck since it polices a market of informed investors who can take care of themselves, whereas clear harms like a boiler room defrauding a grandma of her pension are unambiguously wrong.

    “There's a lot of gray-area behavior in institutional bond markets that has gotten a lot of focus because it is gray area, so it's more interesting to bring a case if you're a prosecutor or an enforcer.”

  • Levine states he has an aesthetic objection to ICO frauds.

    “I have an aesthetic objection to ICO frauds.”

0.34

Valuing and interpreting financial deals requires recognizing that banks and clients may hold different economic intuitions from the same trade, extracting different payoffs.

2 pointscentrality 3/5
  • Regulation FD bars companies from disclosing material nonpublic information to one investor without disclosing to all, yet companies constantly meet with investors who then trade, supposedly learning nothing material—revealing a striking tension between a rule that on its face would prohibit substantive meetings and the lived practice of finance where companies have fiduciary duties to and routinely meet with their owners.

    “companies meet constantly with their investors... and yet they're never disclosing material nonpublic information. Then the investors leave those meetings and go trade.”

  • In complicated financial trades, the economic intuitions the bank explains to a corporate client (like a CFO) are not necessarily the intuitions the bank itself holds, and the two sides can have overlapping rather than opposing interests, each extracting a different and complicated set of payoffs not described in the public documents—an awareness valuable for interpreting financial deals.

    “the economic intuitions I was explaining to the CFO were not necessarily the economic intuitions that we had.”

0.27

Finance is becoming less mystified and more seen as obfuscation; tech now occupies the mystique position finance once held, and regulatory fragmentation especially in insurance is excessive.

4 pointscentrality 2/5
  • Post-financial-crisis regulatory reform has largely had its intended effects, so systemically important financial institutions probably aren't making huge, levered, directional bets on non-traditional banking products; while you never fully know, Levine would be surprised if such bets exist at regulated institutions today.

    “the reform of financial regulations since the financial crisis has had a lot of its intended effects.”

  • Geographic clustering of finance in Midtown will probably decay, because to the extent trading is computers trading with each other co-location is unnecessary, but to the extent finance is M&A and very high-dollar sales—where people strongly believe in in-person meetings—it is not as amenable to fragmentation as one might think.

    “to the extent that trading consists of computers trading with each other, it does seem somewhat unnecessary that we all sit within 10 blocks of each other.”

  • American financial regulation is almost certainly too federalistic and fragmented, with insurance regulation being the clearest example—too state-based, fragmented, and gameable—while corporate law is effectively unified anyway because it is a creature of Delaware, so there's no strong case for federalizing it.

    “It's very clearly too state-based and fragmented and gameable”

  • Centralizing derivatives risk into clearinghouses may create more moral hazard than leaving it with banks, because a clearinghouse is a member association relying on its members for capitalization, so its concern about losing money is more attenuated than a bank trading desk where individuals directly do not want to lose money.

    “Seems to me that a clearinghouse has more moral hazard than a bank. A clearinghouse is often a member association, where you're ultimately relying on the members for your capitalization and for your protection against blowing up.”

0.24

Technical jargon and legalese carry precise meanings and are more efficient than plain language, making aesthetic criticism of them overblown.

3 pointscentrality 2/5
  • Jargon and legalese are underrated and people are too quick to criticize them, because a field's technical terms often carry a set of precise meanings that are easier and more efficient to use than plain language, making aesthetic criticisms of legalese often overblown.

    “in general, people are too quick to criticize jargon and legalese and all these things.”

  • In a specialized economy every sector angles to seem complicated, mysterious, and important to defer to a priesthood, and finance sits high in that sweet spot—but it has come down somewhat, as the public now finds finance confusing in a non-awe-inspiring way and wants to ban or radically simplify it rather than defer to its mystique, with tech (e.g. Facebook) now mystified far more.

    “we live in a specialized economy, and everyone is angling for that notion that what they do is complicated and mysterious and important.”

  • An activity is nerdy due to two clusters of factors: an academic barrier requiring intelligence or learning, and unpopularity or uncoolness; finance became nerdier academically over the last decade (a trading floor that once needed a fun person now needs a physics PhD) while computer programming shed its uncoolness as Zuckerberg-style stories made it a path to power.

    “What makes an activity nerdy, there are two clusters of factors. One is having some sort of academic barrier, where you have to be smart to do it, or you have to learn something to do it. The other, traditionally, is unpopularity, right? You're a nerd because you're not cool.”