Dean Ball, a fellow at the libertarian Mercatus Center and otherwise classical liberal type wrote a long tweet last week pointing out that Trump critics have not properly engaged with the intellectual foundations upon which Trump’s economic policies are (possibly) based. In particular, that there are positive externalities from a manufacturing sector in the US, and that government intervention is the only way to obtain the benefits of that externality. What particularly caught my eye was when Rohit Krishnan replied with the standard free market keyhole solution; that Trump’s tariffs are too much, and you should tailor your interventions into the market to focus on the smallest possible intervention with subsidies for particular products. That is exactly what I would have suggested. Yet:

https://x.com/deanwball/status/1907641739369943396

Alright, so I’ve been called out as not thinking about the problem at all, point taken, let’s try and engage with the cited people and see what they have to say.

The Intellectual Case

Dean mentions Stephen Miran (Trump’s chair of the Council of Economic Advisors) and Oren Cass. Cass I’m more familiar with, and I find quite unconvincing. While he discusses benefits from manufacturing related to innovation and national security, which we will definitely revisit, he spends a lot of time discussing the fact that loss of manufacturing impacts people via job losses and community. This seems very close to bemoaning dynamism itself. Economies adjust, technology advances, and there are winners and losers. We should not ban light bulbs to make sure candlemakers remain employed, just like banning automation at ports to keep longshoremen jobs is a bad idea. For this reason, I’ll be focusing on Miran since I find his explanation of manufacturing losses far more interesting, and his arguments about national security and innovation far more convincing.

Miran published two white papers last year “Brittle Versus Robust Reindustrialization” in February for the Manhattan Institute and “A User’s Guide to Restructuring the Global Trading System” in November for Hudson Bay Capital.

The earlier paper from February is actually somewhat contrary to Dean’s dismissal of targeted tariffs as “whack a mole”. Miran spends the majority of the paper arguing for fairly standard conservative supply side reforms— or what we might these days call “Abundance” policies: regulation slashing, NEPA reform, occupational licensing reform, reducing powers of unions, land use reform, and moving towards a Destination Based Cash Flow Tax instead of a standard corporate tax.

Miran then turns towards the demand side and envisions a more robust defense industrial policy with stronger buy-American provisions. He sees this as both economically and militarily beneficial. Even though he doesn’t specify expanding defense spending, he criticizes current buy-American rules as far too lenient and thus sees a radical shift in production as possible without much change in spending. Miran also discusses expanding these buy-American rules to strategic industries like semiconductors, pharmaceuticals, satellites, etc.

Next up: tariffs. Miran is a fan, particularly in defense supply chains to reduce Amerian trade linkages with China. He also suggests that in some cases, tariffs will be required not just on products with Chinese country of origin, but for all imports of that product if the industry is strategically important, but he doesn’t specify any details here. Nonetheless “Both tariffs and expanded Buy American requirements should be phased in gradually, with credible pre-commitments, to give firms the time and necessary planning horizon for reconfiguring their supply chains.”

Subscribe now

Restructuring Global Trade

Miran’s second paper, published in November is far more radical. He says this is not a policy prescription, but a discussion for what the next Trump administration might do. Since he was then appointed chair of CEA, it seems like it may in fact be quite prescriptive.

The foundational point, mostly absent in his previous paper, is that the US dollar’s role as the world reserve currency has an unrelenting upwards impact on the dollar’s value, which disadvantages exports and particularly manufacturing. While in a typical model of international trade, exchange rates will always adjust to balance trade flows (if your country buys more imports, your currency declines, which makes future imports more expensive), the reserve currency role creates a continuous demand for dollars unrelated to trade flows. Thus, as the world economy grows, the demand for dollars continues to expand, which results in a very strong dollar. This is known as the Triffen Dilemma, named after the economist who noted this phenomenon in the 60s. The dilemma makes US exports unattractive and thus harms the manufacturing sector.

Taking increased manufacturing as a given policy goal, Miran explores the policy options available for the Trump administration to reduce the negative impacts of the Triffen Dilemma, the first of which is tariffs.

Miran makes a sophisticated case, including some points about tariffs that I was either unaware of, or had forgotten since my international finance class in college. Particularly that large economies can maximize welfare by having positive tariffs because they can extract price reductions from foreign producers given the country’s large market power. Of course, Miran declines to note that this is only welfare enhancing if you disregard the welfare of non-Americans. Miran also points out that there can be currency adjustments, as there was between the US and China during Trump’s last administration, which meant that American consumers didn’t see price increases as a result of the tariffs because the Chinese yuan lost value. In practice, this means the tariffs were paid for by Chinese citizens broadly who dealt with a less valuable currency and pricier imports. Thus, depending on currency adjustments, tariffs can raise prices or they can adjust the currency values such that the US treasury can effectively tax foreigners.

The paper goes into a lot of detail of ideal tariff implementation (gradual) and how these can be used for leverage to help achieve US national security goals:

He also notes the administration needs to be careful with tariff implementation to avoid retaliation, as that would risk the US shooting past any theoretical “welfare enhancing” tariff level or even international trade broadly blowing up broadly.

What was completely absent from this section? Any mention that tariffs could improve the US manufacturing sector.

The Global Currency System

Finally, we arrive at the big guns. The Triffin Dilemma results in an overvalued US dollar, so the government could pursue a policy of direct devaluation. The problem, as Miran notes, is that if you devalue the dollar, people won’t want to hold dollar denominated assets.

Miran then details potential multi-lateral approaches Trump could take to devalue the currency while avoiding a run on US treasuries. He suggests a hypothetical “Mar-a-lago Accord” where the US leverages tariffs to bring allies to the negotiating table. Basically, as long as the US is the security provider to our allies, they should help pay for it via buying treasuries, and in particular “century bonds” to tamp down long term interest rates, and if they don’t, they will lose access to the US consumer market via large tariffs. In his own “Feasibility” section though, Miran admits this approach may be pretty tough to get working. Much of global US dollar foreign currency reserves are not in the hands of traditional allies and they may not be able to be convinced. Moreover “a large fraction of the U.S. debt is held by private sector investors, both institutional and retail. These investors will not be convinced to term out their Treasury holdings as part of some sort of accord. A run by these investors out of USD assets has potential to overwhelm the bid for duration coming from a term out from the foreign official sector.” Sounds dangerous to me.

Miran next turns to unilateral currency approaches. He acknowledges as much, but these are incredibly risky. First, the president can leverage the same law empowering the Treasury’s sanction powers to disincentivize the accumulation of foreign exchange reserves. For example:

…impose a user fee on foreign official holders of Treasury securities, for instance withholding a portion of interest payments on those holdings…Some bondholders may accuse the United States of defaulting on its debt, but the reality is that most governments tax interest income, and the U.S. already taxes domestic holders of UST securities on their interest payments…Of course, a user fee risks inducing volatility…

You betcha some bondholders will complain!

Another unilateral approach is to counteract foreign holders of USD reserves by having the government accumulate foreign currency reserves, thus pulling their currency out of the market, reducing supply, and appreciating their currencies. Miran notes that this introduces the mirror risk of the above; if the administration adds a user fee on Chinese owned US government debt, but simultaneously tries to appreciate the yuan by buying Chinese government debt, the Chinese government will just add a user fee onto the debt owned by the US Treasury, or worse just default on their obligations to block American attempts to appreciate the yuan and harm American taxpayers.

Finally, Miran points out that because these currency approaches are much riskier, tariffs are clearly the first choice before any radical policies are undertaken.

Free Markets Finished?

Having read these papers, I can see Dean’s point quite well. China has established a massive manufacturing juggernaut, while the US has mostly not kept up. And while classical liberals can certainly critique that China could have a much wealthier populace and less political oppression if they didn’t do all this central planning, the risk is still that they’ve done enough to now threaten others with poverty and oppression. Miran isn’t a leftist or a socialist or an idealist. He clearly buys into the fundamental economic reality that markets work and efficiently allocate capital, yet he has spotted a market distortion that places the global world order and American national security at risk. It frustrates my libertarian leanings, but perhaps abandonment of free market principles is required to save the free market system. And yes, we’ll get to the fact that Trump has only pretended to do the reading, but I also understand Dean’s annoyance with people claiming “there is no plan”. Because clearly there is a really detailed one here from Stephen Miran, who seems to be the only person taking geopolitical situation clearly.

But the more I thought about it, the more issues I see in the plan. As a precursor, I want to reemphasize that this plan to reformulate the entire global trade system is extremely ambitious, radical, and risky. My counterpoints and alternatives are not compared to a dull status quo, but a plan that risks our diplomatic ties with our allies, the dollar’s status as the reserve currency, a US debt crisis, and indeed our entire economy. A plan that risky should be certain there are not alternatives or issues. I think there are. First, the elephant in the room we’ve been ignoring.

We’re Not Doing the Plan

Miran is constantly discussing the trade-offs, the risks, and the need for clarity when plotting this new global trading order. If allies are spooked off, if you can’t bring people to the negotiating table, if you upset financial markets, then the plan isn’t going to work. Therefore, Miran argues that what is needed is supply side reforms and simultaneous gradual tariff ramping up. This is a core feature of the plan, not some text copy-pasted onto the end. If we’re trying to get other countries to do something, we have to make it clear what each player’s options are and what the consequences are. He repeats many of these points in his April 7 speech.

Trump didn’t bother with any of that. Hours before the April 2 tariffs were announced, it still wasn’t clear the administration even knew what they were going to do. The announced tariffs are extremely high, inviting retaliation, something else Miran says should be avoided. They are also not separated into buckets based on national security requirements. By not differentiating, many of our former allies are now joining with China in adding retaliatory tariffs. This is a nightmare scenario for the Miran plan.

In Miran’s April 7th speech, a major thing missing is the need for caution and gradual tariffs. Presumably Miran knows the ship has sailed there. Maybe there was a faction struggle inside the White House, and Steve Bessent and Peter Navarro had way more radical plans than even Miran’s very risky one. Or maybe Miran never cared about being cautious. Or perhaps Trump is just a mercantilist and thinks it’s 1700. Regardless, the plan is dead.

And that’s just the points about tariff gradualism and clear negotiating goals. If you were really trying to target manufacturing specifically, you’d need to be actively working to make immigration easier and more streamlined (more workers for US assembly lines). And of course, why would you tariff agricultural goods? We don’t want to push Americans into farming, that would harm manufacturing even more.

Nonetheless, unlike Trump and the whole White House, I did do the reading, so I might as well add some actual critiques of the plan. First, I’d like to attack the fundamental premise.

Do you actually need manufacturing?

Miran spends no time defending this policy goal, as he takes it as given that there is bipartisan interest in this. But I’d like to talk about how necessary this really is. I loved last year’s ACX review of How the War Was Won, and there’s no doubt that US productive capacity won the war for the allies, but are we planning to fight World War 2 again? We have nuclear deterrents now. Can’t you just give Taiwan nukes and avoid the 41 page white paper and the collapse of the US economy altogether? Who cares about made up accords, and Triffen Dilemmas and quasi-legal government debt vehicles if you just say “hey, we made a Pacific version of NATO, don’t cross us on this, Trump is nuts” and then we just keep letting the economic system chug along.

Moreover, it’s worth noting the US fought several wars while still being a manufacturing powerhouse and things didn’t turn out great: Korea and Vietnam are not really on the US military’s career highlight reels. Perhaps it’s possible that having a dominant manufacturing sector is necessary but not sufficient.

Dean Ball does argue from his perspective that manufacturing is less about national security production during a war, but rather about innovation and how AI will soon transform the economy. AI will change everything digitally, but will require manufacturing prowess to leverage those digital advances into the real world. If China owns all the manufacturing capability, then even if US labs get to advanced AI or AGI (artificial general intelligence) first, they won’t be able to capitalize and we will end up being stuck with CCP-value aligned AIs instead. This seems reasonable, but it’s also quite difficult to foresee how important manufacturing output today is related to what happens in a crazy AI futures. Still, let’s keep this in mind.

How important is the reserve currency to national security?

I won’t spend much on this, because I tend to agree with Miran that the national security benefits are good. But it’s not like US government dollar sanctions have really dealt any blow to Chinese ambitions. Russia is having a tough time, but again it’s not clear America is achieving our goals even if the sanctions on Russia are challenging for them. As long as we are considering upending the entire system, I’d like to at least know this isn’t the answer. Strangely, Samuel Hammond points out that Treasury Secretary Scott Bessent has said positive things about putting the US back on a gold standard. In essence this position would aim not just for some rebalancing to avoid the Triffen effects, but actually swing the balance far into the other direction, forcing the economy to become more like China, producing industrial goods and consuming far less. This seems like a bad idea too as we not only give up the national security benefits but also become significantly poorer to appease the desires of some elites.

The Special Case of Canada

Suppose we take the Triffen Dilemma seriously and that we must maintain reserve status while also requiring a manufacturing base for national security purposes. Why are we not leveraging our relationship with Canada to address all three? Canada has their own currency. They aren’t impacted by the Triffen Dilemma. If we put up tariffs or defense quotas around imports to the United States and also Canada, the Canadian dollar would not face headwinds from being a reserve currency, it would just depreciate, sparking a manufacturing boom. What’s more, we (used to) have very low trade barriers with Canada, meaning we could take advantage of the manufacturing there (and lower Canadian dollar) for cheap manufacturing inputs and could easily see spillover effects for some manufacturing within the US taking Canadian outputs and adding high tech value adds. I think this approach would also satisfy Dean’s concerns about AI and manufacturing; if US labs developed AGI and leveraged Canadian manufacturing to build robots, chip fabs, drones, or whatever crazy sci-fi things we are going to cook up, that’s totally fine and would benefit both countries immensely.

The problem with having a manufacturing base in China and low trade barriers is that China is a geopolitical adversary and they have an incentive if tensions get bad to cut us off from their manufacturing base. Taiwan is an ally, but vulnerable to Chinese invasion. There are similar if lesser concerns about South Korea or Japan. Canada, however, is ideally located deep within the American security shield. They are NATO members so we are already treaty bound to protect them, we have an integrated defense in NORAD, and our armed forces undertake joint operations routinely. Canada can be defended almost as if it is US soil. And of course, especially relevant to American conservatives, there are huge cultural overlaps between Canada and the US as well. They are also a liberal democracy, Canadian actors have huge roles in American pop culture, they primarily speak English, and they even participate in our national sports leagues.

Would Canada assent to have large tariff barriers raised against others besides the US, and have their currency depreciate against the dollar? There are costs. For this arrangement to work, there would likely need to be an additional binding treaty to make sure policy agreements between the countries could not be suddenly revoked. This means an explicit loss of flexibility in policy options going forward. However, that’s essentially what NAFTA, NORAD, and NATO are: binding agreements that make everyone better off by restricting everyone’s options. Having unique access to the US market when many other countries are cut off could be quite enticing, and the US could also sweeten the deal with other benefits.

This seems much simpler than any of Miran’s longshot policy prescriptions to reshape the entire global trading order and doesn’t require trying to coerce dozens of nations, instead just working with a single close ally. Of course, Trump has alienated Canada so any agreement would be pretty hopeless at this point, but I think this is an under-discussed error in our foreign policy.

One final point: Canada is pretty small. Its entire population is less than California and its economy is also much smaller than the US, so there may be limitations on how much manufacturing from Asia can be replaced solely by Canada. Mexico is obviously the second best option for this “just put the manufacturing nearby but in a country that doesn’t use the US dollar” plan. They have a much larger population and many manufacturing plants exporting to the US already given their generally lower wages. However their economy is even smaller and their present security arrangement with the US would be totally inadequate. Not to mention the political challenges of Trump trying to suddenly build a relationship with Mexico. At the very least though, this build-in-Canada approach could be one step of several undertaken.

We Tried Nothing and We’re All Out of Ideas

We have no empirical data on the Triffen Dilemma’s magnitude, especially in a floating currency regime. Full Stop. Triffen originally wrote about the reserve currency problem in relation to the Bretton Woods system. Miran is arguing for reconfiguring the entire global economy, risking total economic collapse, but we don’t even know if this is the key insight or how much its contributing to the problem. Miran himself notes that despite the Triffen Dilemma (which forces countries to buy US debt), US government borrowing costs are not particularly low. Miran is mostly puzzled by this and moves on, but it could be quite significant. Especially since a major plank of the plan is to bring trading partners to the table and force them to buy longer term treasuries; they already do this, and apparently it doesn’t accomplish what Miran wants already!

Moreover, the Triffen Dilemma impacts all exports, not just manufacturing. Taking steps to devalue the dollar could improve manufacturing, but it also would just make US services cheaper too. We already have network effects related to services like finance, technology, and healthcare. Why wouldn’t those sectors see surges in growth even larger than manufacturing? Other countries have seen declines in manufacturing as percent of GDP and they do not issue the reserve currency. It seems plausible that Triffin effects could be dominated by other policies.

Miran spends most of the February paper discussing supply side reforms. He mentions them offhandedly in the "User’s Guide”, but we should try these first. And there’s a huge menu to choose from: NEPA reform, zoning, occupational licensing, DBCFT, even consumption taxes. Maybe the magnitude of manufacturing expansion would be large enough to be considered a success, maybe not, but it would doubtless have a positive impact. Not to mention a huge economic boom in construction and energy production as well. Supply side reforms are also gaining popularity on the left with Ezra Klein and Derek Thompson’s new book. Additionally, they do not risk destroying the global economy and upsetting our allies. There’s also a good political angle for Trump; the construction industry is a much better place to improve the prospects for blue collar male workers than manufacturing; it’s far bigger and less prone to automation. Turning manufacturing into an inefficient jobs program would harm the national security goals anyway.

Also, just as throwaway, Miran didn’t mention it, but immigration is key for manufacturing. SpaceX and Tesla are American manufacturing companies because Elon was able to come to the US; TSMC dominates semiconductors because Morris Chang decided not to stay in the US. The administration is certainly not even considering this strategy.

It’s also not clear we need to make a binary call as to whether we have or have not addressed the Triffen Dilemma. We need to hit some level of manufacturing base that would address national security or innovation concerns. Once a given national security or AI inflection points happen, there will likely be a huge amount of investment thrown into the problem. Existing assembly lines would need to be retooled anyway. The goal is to have enough expertise to quickly build what is needed. This is likely a spectrum. Moreover, the US is not starting from zero; there is plenty of advanced manufacturing. Regulatory and tax reform, coupled with Operation Warp Speed style incentives could do a lot before needing to gamble the entire global trading system.

The Fatal Conceit

Dean’s original tweet that inspired this post argued that classical liberals were in some sense naive and had not fully grasped the magnitude of the problem. Simply sticking to laissez-faire principles with a few keyhole solutions won’t cut it. But I think in fact it’s planners like Stephen Miran and everyone else further off the deep end in the Trump administration who have not grokked the free market critique. I argued earlier that we we weren’t following Miran’s plan. That’s true, but central planning will always fail. If they had understood Hayek, they’d know planners do not have the information available and cannot beat the market. If they had understood James Buchanan, they’d know they wouldn’t be able to pursue a singular goal of domestic manufacturing, but have to deal with a political coalition of interest groups all wanting different goals.

And the consequences really are dire. I don’t believe the markets have fully embraced how strongly the White House wants to keep these tariffs. Miran might want to keep them only as a negotiating tool, but virtually no one else does. And they may try to go much further, like remove the dollar as the reserve currency entirely. The impact on the dollar and treasury rates could be quite negative, with really catastrophic tail risks. The chaos we see isn’t just because the plan fell apart; it’s because the planners thought they could control the incentives of billions of actors in the market and come out unscathed. Their hubris doomed them, and perhaps will doom the US more broadly.

Share

Jason Crawford is the Founder of the Roots of Progress Institute, and the related blog (also substack). I’ve been reading through some of the essays in his latest large project that’s he calling the “Techno-Humanist Manifesto”. It’s really good. If there’s one takeaway from this post, it’s that you should check out the essays he has written so far.

Crawford’s long term thesis is that technology and industry have transformed humanity, creating beauty and flourishing where there used to be only poverty and struggle. We can pursue this dream and lift our society to future heights that today we cannot even fathom. This did not used to be a radical position, but we’ve turned away from this idea and taken negative views of technology and even progress. This is particularly surprising because human welfare really has continued to improve across a robust array of dimensions. Jason Crawford (along with some others like Tyler Cowen and Patrick Collison) has helped to establish a broader “Progress Studies” movement, and Scott Alexander has a nice summary post on a big conference they had this year.

The call to arms to defend progress resonates with me. Crawford’s writings are a full-throated celebration of the society humanity has built and the titanic achievements that have become overlooked because they are so common now. From the intro:

To our ancient ancestors, our mundane routines would seem like wizardry: soaring through the air at hundreds of miles an hour; making night bright as day with the flick of a finger; commanding giant metal servants to weave our clothes or forge our tools; mixing chemicals in vast cauldrons to make a fertilizing elixir that grants vigor to crops; viewing events or even holding conversations from thousands of miles away; warding off the diseases that once sent half of children to an early grave. We build our homes in towers that rise above the hills; we build our ships larger and stronger than the ocean waves; we build our bridges with skeletons of steel, to withstand wind and storm. Our sages gaze deep into the universe, viewing colors the eye cannot see, and they have discovered other worlds circling other Suns; they have found the atoms of Democritus; they can tell us the system of the heavens and the mechanism of life; they can at long last turn base metals into gold.1 Once, these accomplishments, and their benefits to humanity, were referred to simply as “progress.”

The most powerful Kings of 500 years ago had no access to things we take for granted: cheap food, central heating, instantaneous communication with our relatives wherever they are on the planet. Delivery pizza could not have been purchased at any price. From Chapter 1:

We forget that everything had to be invented, even the most mundane features of daily life. Basic fasteners such as the zipper, the paper clip, the safety pin, and the rubber band were not invented until the 1800s.6 Simple food containers such as soda cans, plastic bottles, and milk cartons required a remarkable amount of engineering: to choose the right materials, to make them easily openable, to manufacture them cheaply.

But today progress is not seen as an obvious good. People regularly talk as if we are living in the worst times of the world and that they would choose to be born in the past rather than today. Crawford argues that not only is this despair depressing for us psychologically, but it is harming our ability to push progress further.

If society believes that scientific, technological and industrial progress is harmful or dangerous, people will work to slow it down or stop it. Activists have obstructed all forms of energy—nuclear power, oil and gas, even solar and wind14—and the average American benefits from no more energy today than fifty years ago.

Crawford wants to make the case that progress is good and thus we should actively work to achieve it, instead of thwarting it as we often do. Chapter Two dives into the sheer powerlessness of early humans, and yet how completely we have overcome the state of nature. We barely consider it day-to-day, but humans are essentially the only animals that do not find their own food. Civilization allowed for the invention of professions outside of farming or gathering. But even these new jobs were extremely difficult. Forges didn’t have thermometers, and ships didn’t have satellite positioning systems. If the winter lasted longer than expected or you slipped and broke your leg at the wrong time, you could easily starve. The extent to which literally none of these gargantuan challenges to our ancestors bother us in the slightest today is amazing.

Chapters 3 and 4 focus on human achievement and the importance of fulfilling human desires as an end in itself. Only humans can cooperate and work together across generations to build society. Natural beauty is great, but it is human enterprises that allow the discovery of how the world works and appreciate it. Evolution can only learn through generational guessing; we can learn by experimentation and communication. Crawford also takes on the belief that modern capitalism causes meaninglessness. Growth doesn’t just mean you can buy more crap you don’t need:

And more of that life can be devoted to the things we enjoy most—recreation, loved ones, personal enrichment, self-actualization—because less and less of it is devoted to basic necessities. At the start of the 20th century, the average American household devoted the vast majority of its expenditure, about 80%, just to food, clothing, and housing; by the end of the century, that was down to 50%.16 More wealth doesn’t just mean more of the same stuff, and it doesn’t even just mean more luxurious versions of the same stuff: it means that once you have fed, clothed, and housed yourself, you still have money left for dancing lessons, camping gear, a family photo session, or a trip to Taipei.

The Future

Crawford hasn’t gotten to the part of the manifesto thinking about the future yet. I figured I might take a brief crack at it.

Let’s consider a society that leans into progress, rather than shuns it. In this world, we allow technological innovation to flourish, we celebrate it as our destiny as a species, and we construct policy around getting more of it.

I think we would firstly see a diminishing of many problems that plague our society. Climate change would basically be solved. We already know enhanced rock weathering can absorb carbon from the atmosphere. With abundant clean energy from next-generation nuclear power, solar, and fusion, we could power atmospheric carbon capture at scale while providing cheap electricity to everyone.

Staying with energy, for a moment, with practically unlimited cheap electricity, we could desalinate seawater at massive scale, turning deserts into gardens. Cities like Phoenix and Las Vegas could become lush oases, their streets lined with shade trees and flowing fountains. Water-intensive agriculture could expand into previously arid regions. Energy-hungry industries like aluminum smelting and chemical manufacturing could become dramatically cheaper. We could run massive indoor farming operations anywhere, freeing agriculture from the constraints of climate and weather. Data centers could expand without environmental concern, supporting ever more powerful AI systems and computational tools. In a world of energy abundance, many of our current resource constraints simply disappear.

Energy also contributes to transportation improvements. With cheap enough energy, synthetic fuels could make carbon-neutral aviation commonplace and dramatically reduce ticket prices, since fuel costs are typically a quarter of airlines' operating expenses. Moreover, regulations have kept back supersonic travel for decades. We could have both cheaper and significantly faster air travel.

Housing abundance would transform our cities and quality of life. Advanced materials and automated construction techniques could slash building costs. Today we spend huge portions of our income and price out people who want to move to cities with the best job prospects just to live in century-old buildings. Instead, every family could start with a brand new house by default, always having the latest amenities, building materials, and spaces adapted for today’s living rather than 100 years ago. With reformed zoning laws and embracing of new construction methods, housing could become so plentiful that the concept of a "housing crisis" becomes as foreign to our grandchildren as food scarcity is to us.

Space colonization would open up entirely new frontiers for humanity. Regular shuttles to orbital habitats could become as common as international flights are today. Mining asteroids could provide abundant raw materials while preserving Earth's environment. Permanent bases on the Moon and Mars would not just be scientific outposts, but the beginning of new branches of human civilization.

Perhaps most transformative would be advances in health and longevity. We're already seeing promising developments in gene therapy, regenerative medicine, and understanding of aging processes. We know we could create platform vaccines to eliminate entire families of viruses, but we just need policy that rewards their creation. A society embracing progress could see dramatic extensions of healthy lifespan. Diseases that terrify us today could become as manageable as infections in the antibiotic era.

Beyond these specific domains, we might see innovations we can barely imagine today. Just as someone from 1924 could hardly conceive of the internet, our future technologies might seem like magic to us. Artificial intelligence could help us solve previously intractable problems in science and engineering. New forms of education could help everyone reach their full potential. Advances in psychology and neuroscience could help us better understand ourselves and each other, leading to more effective institutions and stronger communities.

The key insight is that progress compounds upon itself. Each breakthrough opens new possibilities, and a society oriented toward progress would be better positioned to seize these opportunities rather than fear them. As Crawford suggests, our ancestors would view our current lives as wizardry - perhaps it's time we embraced our role as wizards of the future.

I Do Have Some Concerns

Finally, we turn to today, and political events which got me thinking about this area. I was not a supporter of Donald Trump during the 2024 election. However, there’s no arguing that Democrats and left-leaning groups have largely opposed the positive and triumphant view of progress I have outlined above. Degrowthers, climate activists, unions, and more have taken the stance that economic growth and progress are bad, zero sum, and should be opposed. Not everyone in the Democratic party believes this; there is a growing number of advocates of the “Abundance Agenda” including Ezra Klein, Matthew Yglesias, Noah Smith, and others.

The Biden administration itself did not embrace this worldview. The CHIPS Act was beleaguered by government regulations preventing much of the investment money from getting to projects quickly. The anti big business stance of much of administration’s regulatory arms, especially the FTC, seemed specifically opposed to growth or innovation, even if it benefited consumers. Kamala Harris, as Biden’s VP, deserves reasonable suspicion when it comes to issues of progress and advancing humanity, and I don’t really begrudge anyone who opposed her on those grounds.

What was most surprising to me was that many who I would consider to be on the pro-progress side enthusiastically embraced Donald Trump. I’m not talking about overall frustration with the political system yet ultimately finding Trump more palatable, but rather actively campaigning for the former commander-in-chief. Elon Musk and Marc Andreeson were probably the biggest advocates going to bat for the once and future president. While the first Trump administration wasn’t terrible when it came to innovation policy, it was fairly mixed. A deregulatory environment is good, but generally attacks on free trade are unhelpful for trying to build supply chains for complex goods. Operation Warp Speed was a spectacular demonstration of how government policy can incentivize powerful scientific breakthroughs.

The problem is that Trump didn’t run on Operation Warp Speed, or on bringing that kind of policy success to other parts of the government. He ran away from his record here. He’s appointed RFK Jr., a techno-skeptic conspiracy theorist to be HHS Secretary. RFK has said that the polio vaccine killed more people than polio via vaccine induced soft tissue cancers. And Trump has no concept of economic growth, positive sum trade, or technological progress. He believes bilateral trade deficits are direct money transfers, and that automation at U.S. ports would harm jobs. These are positions of a luddite.

Of course, an administration is made up of many people. Jim O’Neill is a great pick for HHS Deputy Secretary, and Jared Isaacman has potential to really accelerate space exploration at NASA. Both seem likely to be much more aligned towards a progress agenda than anyone who could have conceivably been appointed by Kamala Harris. And, so far at least, Elon Musk continues to hover around government and be involved in everything. When the intra-coalition argument about legal skilled immigration blew up, I expected Musk and any legal immigration advocates to quickly lose. After all, Trump’s new Deputy Chief of Staff and previous Trump administration veteran Stephen Miller has repeatedly sought to curtail legal immigration to the U.S. But I was fairly surprised that Trump sided with Musk, and indeed Kalshi shows a 60% chance of expanded H1-Bs under Trump as of the time of writing.

It remains to be seen what will actually happen during the next four years. Steve Bannon has already called Elon “truly evil” in response to the whole coalition fight, and Trump has a history of very public fall outs with his most important advisors. I will take whatever victories I can get, but I would caution anyone who shares my viewpoint to avoid identifying too strongly with any given political party. Even if you judge Trump to be better on this issue, he’s not a true believer that the world can be improved by technology or development. More than anything, Trump is simply on Trump’s side, and he has little consistent ideology. The real work of advancing human capabilities will continue regardless of political allegiances, driven by those who genuinely believe in technology's power to improve the world.

Tyler Cowen asserts that if you are concerned about existential risk, you should be short the market, and he has a list of 10 bad responses he says EAs give if asked why they aren’t short.

If you pose the “have you thought through being short the market?” question, one hears a variety of answers that are what I call “first-order wrong.”  That is, there may well be more sophisticated defenses of those points of view, but you just hear the first-order response, designed to dispose of the question without much further thought.  A few of those responses are:

1. “Why should I have to gamble?” (Given your other views, it is hedging not gambling)

2. “There is already evidence I am right.  My friends and I made a lot of money buying Nvidia stock.”

3. “I don’t know how to short the market.”  Or “Amateur investors shouldn’t short the market!”

4. “Did the stock market predict Hitler and WWII?”

5. “How possibly can I cash in if the world ends very suddenly?  After all, the AGI has an incentive to deceive us.”

6. “But I don’t know when the world is going to end!”

7. “Why should I short the market when I can earn so much more going long on Nvidia!?”

8. “Well, I am not buying stocks!”

9. “If the world is ending soon, what do I need money for?”

10. “But if the world doesn’t end, things will be really great.”

And more.  (I’ve even heard “Are you short the market?”)  I will leave it as an exercise to the reader to work out what is wrong with these responses.  In most cases o1 and Claude can come to your aid, if needed.

Some of these sounded pretty reasonable to me, notably 5, 6, and 9, but as Tyler didn’t find it necessary to tell me exactly how I was a moron, I turned to Claude for those:

(5) "Can't cash in if world ends suddenly" The flaw: This assumes binary outcomes (total extinction vs business as usual). One could profit from market drops during early signs of AI problems, before any theoretical point of no return.

(6) "Don't know when world will end" The flaw: Markets offer various instruments with different time horizons. If someone has strong views about AI risk within a particular timeframe, they could choose corresponding investment vehicles.

(9) "What do I need money for if world ends?" The flaw: Again assumes binary outcomes. Money would be valuable in many scenarios short of complete extinction, and could fund efforts to prevent worst outcomes.

Claude makes a good point that outcomes are not necessarily binary; there’s a wide range of possible AI outcomes besides “my Nvidia stock goes to the moon” and “all humans are dead”. Additionally, there’s a question of whether the market will actually drop in the cases where bad things happen; if AIs have largely taken over the management of companies, it’s possible we will see skyrocketing valuations as AIs trade with each other even as the world gets worse for humans as the AIs battle it out amongst themselves. Thus, I have created this color-coded chart (Claude helped):

You can see here why a lot of EAs might not initially think about hedging. If you’re guessing the two most likely scenarios are in the right or left column, hedging isn’t going to do much.

Since initially encountering fast takeoff scenarios (right-most column) when reading Superintelligence, generative AI has taken off and we’ve seen how real models work. They have training bottlenecks related to physical chips and energy consumption; maybe there are untapped software improvements that will allow future models to scale without any hardware improvements, but no frontier AI companies are acting that way. The takeaway for me is that the future is likely multiple companies or models struggling with limited physical resources even as their abilities and potentially monetary resources grow. This makes the middle scenario more likely.

I think right now I’m skeptical of the slow takeoff / market-keeps-going-up scenario, as multiple agents coming into conflict sounds like it has to be bad for markets, but these futures are very strange already and I’m not confident about anything here. Since the slow takeoff / market crash scenario is the only scenario where we need to hedge though, it seems we should broadly take long positions except for the hedge needed in slow takeoff / market crash.

Does that mean you should short the market? Potentially, or at least there should be some hedging for AI induced geopolitical destabilization. Now that we’ve finally bought Tyler’s premise, how would you actually go about doing that?

Timing

Claude’s critique of (6) sounded reasonable (“Markets offer various instruments with different time horizons”) but I couldn’t really think of a good way to short the market if you had a big range of possible AI timelines. So I asked Claude to elaborate. I gave it parameters of a 10% chance of “really bad outcomes” but I said the timeline was anywhere from 4-20 years from now.

It considered and dismissed a few hedging strategies: long-dated put options, VIX/volatility products, gold/bitcoin, and deep out of the money options.

The core problem is that most hedging instruments are designed for shorter timeframes. Traditional portfolio insurance gets very expensive over a 16-year horizon.

Let me flip this around: What would an ideal hedging instrument look like for your scenario?

• Long dated (4-20 years)

• Small regular premium

• Big payoff in crisis

• Not tied to specific timing

• Doesn't drag much on portfolio returns

Actually... I realize I'm not seeing anything in traditional markets that really fits these requirements well. This suggests Cowen's call to "learn more finance" might be insufficient - we might need new financial instruments specifically designed for long-term catastrophic risk scenarios.

This is a genuinely hard problem that perhaps explains some of the EA reluctance better than Cowen acknowledges.

Well there you have it. Claude even specifically mentioned the VIX contango issue with investing long term volatility products. You can read more about it here, but basically it means you are paying a premium to hold VIX futures if the market doesn’t think the VIX is about to spike. You can’t trade on the VIX directly, but instead you can buy VIX Futures ETFs. The contango issue cashes out in these ETFs as a continually dropping valueover time. Even the medium term VIXM dropped 90% of its value in about 5 years.

And had you been holding it during the last big volatility event at the beginning of Covid, you would have only recovered to the 2017 level about 3.5 years prior. In other words, if you think AI volatility is more than 3.5 years away (I’d assign a big chunk of probability to that), anything you put in now will be wiped out, assuming the market crash is similar to Covid. On the other hand, perhaps the crash will be much worse than Covid, and thus the VIX will see a much bigger increase, and perhaps other EAs think we are much closer to a bad-but-not-megadeath event. If so, here’s something to buy.

Strategy Implementation

Of course, how much should you buy? What percent of your portfolio should it be? What is the actual expected value in the situation given timing uncertainty? These are not simple questions.

Even the most straightforward cases of portfolio theory are much more complicated than it might seem. Financial services companies like Vanguard, have put together low expense ratio, diversified, easy to purchase funds that will provide solid diversification and exposure with minimal overhead and cost. They even have target date funds that you can purchase based on your expected retirement and they will automatically rebalance towards more conservative assets as you near the target dates. But if you want to do anything more complicated than that, you have to make some assumptions on your own and it’s not particularly clear which way to go! These include additional diversification, like exposure to real estate, commodities, cryptocurrencies, or anything else. If you want to take into account more lumpy income or concern that your job is as risk, now you have to calculate what percentage of your assets you should keep liquid. How about the decision to add to a Traditional vs Roth IRA? Here’s a cool 10,000 word article on Bogleheads to work through in your spare time.

Given that even relatively common and straightforward retirement portfolios have still unanswered questions, asking EAs why they haven’t devised a formal financial portfolio strategy for tail risk AI scenarios seems a tad unfair, even if technically Tyler was just prodding at the specific takes he listed.

I’m about to get snarky, so let me make sure I concede at least a small point: looking at some sort of insurance product isn’t a ridiculous idea. Given the complexity and expense of holding hedging positions, I’m not sure hedging is worth the portfolio cost, although clearly there’s a wide range of possible probabilities to consider. Moreover, just slightly increasing your consumption now is not an unreasonable response either and is a form of hedging across time.

Of course, all this is assuming the need to hedge against a market drop is correct. Eliezer is skeptical:

And actually in Tyler’s original post, while Tyler dismisses point (2) (“There is already evidence I am right.  My friends and I made a lot of money buying Nvidia stock”), after taking a quick look at the magnitude of the challenge in front of us, I think EAs deserve some credit here for calling the Nvidia run up. Apparently Tyler himself has never actually achieved any impressive trades based on his views despite haranguing EAs about not doing it:

Perhaps Tyler ought to study finance a bit to get more comfortable with it!

It’s 2007. I realize there’s a gaping film pop culture void in my general knowledge. I’m constantly bombarded by friends referencing “royales with cheese” or Agent Smith and I have no idea what they are talking about. I decide it’s time to rectify this problem. It’s still early days on the internet, and after some investigating, I find several forums dedicated to posting links for illegally uploaded movies, sometimes even in 480p. I create a list, eventually stored on my brand new iPhone (it’s an iPod and a phone, can you imagine?). Every time someone mentions a movie, it goes on the list. Eventually I supplement with some torrenting, but the big break through is figuring out how to rip DVDs onto my computer. I get a Netflix account with the DVDs mailed to you (no late fees!). When the mail comes in, I rip it that day and send it back.

After several years I’ve seen hundreds of movies. Some classics like The Shining and Predator, but also a bunch of middling stuff that made it onto my list because it was the late 2000s and that’s what I heard was good (Mr. and Mrs. Smith, Æon Flux). Tokyo Drift will always be a guilty pleasure, but I know there have been better films. Yet despite the mountain of cinema I’ve seen, my list of movies to watch has actually gotten longer somehow…

It’s 2013. Actually, I need to read some more books. Sure, I had my libertarian awakening in high school after reading The Fountainhead and then tearing through Atlas Shrugged, Economics in One Lesson, and Capitalism and Freedom. But there’s so much more! Classic science fiction like Heinlein, Asimov, and Orson Scott Card that I never read. Modern fantasy like A Song of Ice and Fire which is a TV show now. And there’s a whole bunch of nonfiction I should catch up on. You ever realize how much useful stuff is actually *in* a textbook? It’s crazy. I wonder how I’ll read all this. Someone just sent me a link to something called “The Sequences”, I hope they’re short…

It’s 2017. They tell me the golden age of television is upon us. I was behind on Breaking Bad but was able to finish it a year or so after the actual show stopped airing. I’m caught up with Game of Thrones, which is the greatest television phenomenon ever created. Actually HBO is a behemoth right now. This first season of Westworld is breathtaking. So is True Detective. I need to get around to watching The Wire, The Sopranos, and The Americans eventually. I’ve also heard it’s worth going back to some network TV comedies like The Office, Parks and Recreation, and Community. There’s some cool animation happening too: I’ve been watching Rick and Morty and Archer but I want to watch Avatar: the Last Airbender. And Star Wars is now at Disney so I guess I need to watch The Clone Wars? Plus there’s this new series on SyFy called The Expanse which looks dope.

When will I have the time for this? I can’t believe I’d ever thought I’d finish my movie or reading list, I gotta put that on hold for these TV shows. At least I’m caught up on Stranger Things…

It’s 2020. There’s nothing to do but sit at home and consume more content. Maybe I should get back into videogames. Hmm I’ve got some 200 games in my Steam library, including, let me see….74 that I’ve apparently never played at all. If those aren’t good enough, I could always buy some more from the apparently 100,000 titles that are on the Steam store. Some of these are story based and will take a few to dozens of hours to complete. Some are open world games and can take hundreds. Some are different every time, encouraging replayability. Some are not story based at all, and instead focus on strategy or tactics. People sink thousands of hours into Crusader Kings or Stellaris. I recommended a friend Factorio and he disappeared for months.

If I don’t even want to purchase a game, there’s a plethora of esports titles where you can compete against other players. There isn’t really an “end” to the game, you just try to master them as a skill, like chess or bridge. There’s also crafting games like Minecraft, where not only is the game itself open-ended, you can create almost anything you want within it. Or if you wanted both crafting and esports, you’ve got Fortnite.

What about MMOs? World of Warcraft has a back catalog of 20 years of content. Final Fantasy XIV is big now, Runescape is still going strong. EVE Online if you want a second job. Each of these games could probably take over your life if you let it. And we haven’t even gotten to consoles; the PlayStation 5 and new Xboxes comes out later this year, but you can buy the previous generation consoles right now. There’s also a huge community working on making old games playable. It’s a legal gray area, but you can go and play hundreds of games from old consoles with emulation software, including some of the greatest games every created.

So I’ve got probably 5-10 thousand hours of content off the top of my head here and we haven’t even gotten to gaming in real life! Last year, before the pandemic I was playing board games with friends every other weekend. There are tons of amazing board games to check out, not just Catan. Of course, right now my friends are actually starting a new D&D campaign on Discord. I’ve never done that before but at a few hours a week, that should be pretty manageable…

It’s 2022. So it turns out at that first D&D session, the DM asked if we had seen an episode of Critical Role. I didn’t know what that was, so I looked it up. They had just posted to their YouTube channel episode 110 of “Campaign 2”. It was 4 hours long. Dear God. Now, I certainly haven’t consumed the thousands of hours of Critical Role content available (they’re now in the middle of Campaign 3 of course), but the dozen episodes I’ve seen are quite entertaining. It does however, have to compete with tons of other great content on YouTube. The YouTube fountain of content is itself probably bigger than any of the last content entries combined now that I think of it. I’m subscribed to some 50 channels, probably 20 of which I often watch with a few I watch every upload. I suspect I’m a YouTube outlier, but the quality of content you can find is incredible and generally better than anything on cable TV.

There’s just so much breadth. I follow PC hardware content, science and math channels, rocketry and space, gaming, that primitive technology guy and so much more that I never thought I’d be interested in. There’s this delightful Canadian guy who goes camping in weird places and his channel is just called “Camping with Steve”. I have zero interests in camping or going outdoors in general, and yet I watch every single one of his videos.

YouTube also pairs well with other content. There’s great breakdowns and additional background videos for many popular TV shows and movies. The Corridor Crew channels breaks down VFX and stunts in films, and I now often use their videos as inspiration for my ever growing movie list. If you play almost any game it’s worth checking YouTube for tips or setup tweaks that improve the gaming experience.

To me, podcasts also belong in the same space as YouTube videos; the barrier to entry is very low, the content breadth and magnitude is enormous, but it optimizes for much longer than typical videos. Checking recently and it appears I have some 100 hours of podcasts in my backlog. I love the 80000 Hours podcast, but the jokes about the episode lengths write themselves. Robert Wiblin must be stopped.

It’s 2024, there’s this new thing called Substack…

II. Freedom?

Am I honestly complaining about having too much content to consume? Is this actually a problem? I don’t always self-identify as a libertarian these days, but I certainly lean towards that tradition. Isn’t this the promise of a free society? People can create things for each other! Stories, games, true crime podcasts! You didn’t have this freedom if you lived in the Soviet Union or Nazi Germany, and you still don’t if you live in China, North Korea, or Cuba. Not only does liberal democracy allow for free enterprise and flourishing marketplace, but also a vast ecosystem of creativity and content generation. And many, in fact, make their living through selling their content.

Most weeks I listen to the Reason Roundtable podcast where the editors of the libertarian magazine Reason talk about the news and how much they hate the government, etc. They end each podcast with a full throated embrace of capitalism and discuss what they, as consumers, have been consuming this week. Each editor has their own quirks, but it’s typically much of what I’ve catalogued in the previous section and even more: travel, sports, music, museums, and whatever else is happening. I think this is actually underrated among libertarians who too often gripe and commiserate on what government ought not to do or where it ought not to interfere. Libertarians and classical liberals often don’t put forth a positive vision of the world because they are arguing for the unplanned and the unseen, which cannot be predicted beforehand.

But my story of overwhelmed inability to catch up is really just a story of abundance and pure creation. This must be celebrated! These are the fruits of a free society. To denigrate the content tsunami is to spit in the face of freedom.

Still, some part of me is uneasy. Perhaps if there’s too much to watch or listen or play, then I’ll just have to miss some things, and that’s ok. But I do have two objections. One is that there is a small externality: the more stuff there is, the less shared experiences I have with others. A large shared pop culture is rewarding. You can discuss content and stories with others who participate. But if everyone is listening to their own podcast or YouTube channels, we end up with less in common to discuss. Perhaps we stop hanging out entirely with other people if the content is so good and addictive.

And two, the future looks weird. Ben Thompson from Stratechery traces content platforms’ changing focus over the years. First, Facebook compiled user generated content from your friends and acquaintances in real life. Content was specific and clustered. Then we moved towards a larger broadcast model where followers would accumulate to big accounts. Instagram and Twitter got big on this model, but you saw YouTube and Facebook move towards this too. Eventually, there was so much content from the accounts you followed, social platforms started recommending and ranking all the content into a newsfeed so you didn’t necessarily see everything. The ranking algorithm started simply and became quite complex with machine learning models involved in larger and larger amounts of content to compile and rank.

YouTube was probably the first to start moving towards the next level, where content recommendations weren’t just coming from your subscriptions, but actually content across the whole site. TikTok really blew this open with an entirely algorithmically driven feed of short videos, and YouTube and Instagram quickly followed suit. You used to see big content aggregation Instagram or Facebook pages where you could view content from many people in one place. Now the platforms themselves find the content scattered on their site and deliver it directly to consumers (with the ad revenue) and skip the aggregation pages entirely. You no longer need to request to YouTube or TikTok to deliver the content from a given creator. The machine learning algorithms quickly figure out what you like and give you what you need without needing to find creators and subscribe or follow. It’s still there as an option, but it’s just another signal in the giant inference machine.

And now Ben Thompson argues the next step is going to be not just algorithmically selected content, but algorithmically generated content. Today, each person’s feed is different and specifically tailored to their interests as determined by a big ML model. But if the perfect content for your interests doesn’t exist (I cannot find a fictional documentary of alternate history where NBA has no salary cap and is run by the Yakuza), the algorithm can’t show it to you. But in the future, we could generate the content they want without needing someone to actually make it.

In some sense, this has always been what was happening: NBC paid for a pilot to make Star Trek in the 60s because they thought it would draw audiences. In the 2010s when Netflix saw that their short term cheap content licensing deals were not going to be renewed, they knew they had to start developing their own original content. Instead of just guessing as NBC execs had done in the 60s, they used their extensive understanding of their own customers to target specific genres of shows that they felt could do well on their platform. And thus we got big hits like Squid Game and Stranger Things, but also smaller things like Beasts of No Nation and Narcos. But a big enough difference in degree may really be a difference in kind.

For example, I’m curious about what is happening at Spotify. It is not against their terms of service to create AI music, and it seems that this is becoming cheap enough that I’m getting actual ads urging me to stream specific AI-generated music playlists on Spotify. Artists have long decried Spotify’s low royalties for streaming, but if generative AI is much cheaper, perhaps the future of the platform will be decidedly weighted in favor of generated music. Does the future look like a bunch of music enthusiasts who don’t even have a favorite artist?

In 10 years when your Kindle recs are all books generated directly for you that didn’t exist five minutes ago, and when your TikTok feed is custom made documentaries directed in the style of Jon Bois on exactly the thing you wanted to learn more about, are we going to be totally socially isolated in ways that we long for the 2020s? I want to say that the desire for real human interaction will push us to demand from our algorithms shared experiences, but the trend has consistently been in the other direction. As I said above, the content tsunami is a result of human liberalism and the free market; the default is that the tsunami gets bigger.

Of course, there is nothing older than people decrying the decline of the civilization within which they live, so who knows? Maybe people in the future will laugh at how small minded I’m being.

It would seem the perfect first post for a blog titled Calibrations. So are prediction markets any good at predicting things, and should we care what they say?

Prediction markets allow participants to buy shares of an event occurring or not, analogous to a bet or wager. For example, the website Polymarket has political event markets. A "share" pays out if the election occurs as you predict and pays nothing if you predict incorrectly. A market is created by people buying and selling shares on the outcomes of the events, and the price of the share represents the market’s current probability estimate for the outcome. For example, you can buy a share of Kamala Harris winning the presidential election which will pay out if she wins, but will be worth nothing if she loses. A price of 60 cents would indicate a 60% chance of her winning.

Data on Accuracy

So are prediction markets any good? Well, we can actually look at the history of prediction markets and see how often a market that predicts any given outcome actually ends up resolving in that way. Maxim Lott at electionbettingodds.com has a track record page for the prediction markets he tracks.

This is the calibration of the markets. If something that is predicted to happen 60% of the time actually happens that often, we say it’s well calibrated (and thus the name of this blog). We can also see the calibration for some other prediction sites here from calibration.city which has some excellent visualizations:

I suspect some of the systematic bias you see in the chart above and below the 50% mark is from the fact that many of these markets default to 50% and require betters to provide liquidity to move the market away from the start. To make an enticing market to bet in, the market makers have to incorrectly price a market so that bettors have an incentive to wager and earn their winnings, and we see this in the data as a the market makers systematically “betting” the wrong way to start the market. In fact, if you go to the calibration.city page and weight the y-axis for resolutions by market volume, this bias is reduced (since larger volume markets would have less weight on the initial price) although it does not disappear.

Alright, prediction markets are well-calibrated, but does that mean they are accurate? Are they actually predicting anything? Not necessarily!

Suppose we know that Republicans win the presidency about half the time and Democrats win the other half. We check the prediction markets and they always say democrats have a 50% chance to win. This is a well calibrated prediction, but it doesn’t actually tell us much information at any given point in time. We want to know about this particular upcoming election. What we can do then is use a Brier score to mathematically measure accuracy, and not just calibration. The Brier score is the means squared error for a set of predictions and their outcomes, and thus we can use it to measure the accuracy of probabilistic predictions. Here are the Brier scores from calibration.city:

I arranged by market volume and did a time weighted average, but you can also try it yourself. Overall it seems market volume didn’t have much effect, perhaps because the existing Brier scores are already pretty low around 0.15, and also perhaps because higher volume markets could reflect exciting changes in the underlying event leading to both more uncertainty and more market trading.

We can actually decompose the Brier score into calibration, resolution, and uncertainty components, although exactly why this works is beyond my current statistical understanding. This stack exchange post was helpful, but unfortunately I don’t have the time to put together a full script to grab the data from Manifold and run the decomposition.

There’s also real limitations to comparing Brier scores. You’re really only supposed to compare them across the same events; if the underlying event is more uncertain, you would expect higher (worse) Brier scores. But Kalshi, Manifold, and Polymarket are all predicting slightly different things. The legal difficulties prediction market platforms face exacerbates the accuracy measurement problem; PredictIt has to limit the number and size of bets, Polymarket and Betfair are banned in the United States, and Manifold only uses play money. Still, you’d expect much of this to contribute to worse Brier scores. The fact that these are pretty solid is encouraging. If anyone wants to run an actual Murphy decomposition on prediction market data on a narrow topic, I would be interested to see the results.

I think we can still draw some conclusions though:

• Prediction markets are well calibrated: if they say something will happen with a given percentage chance, that’s actually a good estimate of how likely it will happen.

• Prediction markets have pretty good (low) Brier scores, and that puts an lower limit on how practically bad their resolution decomposition component could be even if they were perfectly calibrated.

• Prediction markets do all this despite major regulatory barriers.

The Bigger Picture

Taking a step back though, I think Matthew Green’s critique gets to something deeper: sure prediction markets are well calibrated in the face of uncertainty, but they don’t have any magical insight besides pushing the best guesses to the top. In other words, they’re not actually reducing the fundamental uncertainty in the world. Why is everyone so excited about this?

I think the answer is price discovery.

First let’s take the demand side. Perhaps you’ve looked up a prediction market price for an event and found the price to be…about what you expected.

The Fed has been talking about rate cuts this year, they indicated in their last meeting that it’s likely coming and the market took a sharp tumble last week before recovering. It seems pretty obvious to anyone paying attention that the likelihood of a Fed rate cut by the end of the year is high. But not everyone was paying attention. A prediction market's existence creates a publicly known "price" of the best estimate of whether an event will occur according to the market participants. Being able to know a well calibrated estimate of an uncertain event is a huge positive externality that prediction markets provide the whole world.

This is a big deal! If you don’t have a price to check, you’re left doing all the research yourself and you only have so much time. Worse, you might have to rely on very non-quantified pundit opinions with no track record. A price cuts through everything to give you a single best guess.

Next let’s take the supply side. Think about the amount of time, effort, resources, money, labor, algorithm design, etcetera that trading firms undertake to get an edge in financial trading. There are potentially billions of dollars on the line if you can correctly predict whether an equity or future will rise or fall in price. Prediction markets can harness these incentives to uncover truth not just about company financial data, but about anything we are curious about. They tie reality and truth about the world to financial incentives. They push us to discover information about what is happening and what will happen.

Unfortunately, they are also legally limited. Event based betting is for all practical purposes banned in the United States even though political decision making is very high stakes. Our lives are worse because of this! We should want our political decision-making processes scrutinized and better understood; prediction markets provide financial incentives to achieve those ends. Prediction markets could be providing us with real time information on whether Congress will pass legislation or how an administration will respond to crises.

Of course, the actual application of prediction markets could be much broader than politics. Nominal GDP futures markets could do a better job informing the Fed of forward looking expectations than stock markets or lagging economic data. Or imagine if we had an ongoing prediction market about possible pandemics on the horizon. That would have been very handy to refer to in January or February 2020 and may have alerted us sooner than simply relying on broad media reporting, which didn't pick up the potential danger of the pandemic until early March.

Perhaps today prediction markets don’t tell us much besides aggregating some polling data, but if we legalized them, maybe they could tell us a lot more!

Uncharted Territory

The finance industry may be driven purely by profit, but the development of new financial instruments has actually brought all sorts of benefits, some of them very widely distributed. Index funds and ETFs allow regular people with savings access to market returns without needing to expose themselves to specific stock risk and at a fraction of the cost of mutual funds. Foreign exchange and commodity derivatives can help companies reduce their risk by locking in prices now; airlines are trying to run an airline business, not predict oil prices in six months.

Prediction markets could do even more if we let them.

A conditional prediction market is a single trading market with two events allowing investors to trade on the relationship between the two outcomes. For example, you could make a prediction market for a political parties’ election prospects conditional on whether they replace their candidate or not. Democrats took months to come around to the idea that their current presidential candidate might do poorly. If they had been able to compare conditional bets on who would win, they might have been able to select a different candidate much earlier!

Conditional markets could have all sorts of interesting applications. We could answer questions like “what will different policies’ impact be on unemployment next year?” or “which scientific research approach is most likely to improve 5 year survival rates of a major cancer?”. But I think the most important point to underline is that we don’t know because the markets aren’t allowed. It took a while to develop ETFs that regular investors could use for their retirement accounts. If we allowed people to experiment and try things, we could probably create some pretty cool things.

Even today, with all the legal challenges prediction markets face, they offer great value! People on Twitter will continue to claim that prediction markets are systemically biased or simply reflecting quirky beliefs of people overrepresented in the betting pool. I think this is too quick of a dismissal, but to some extent, they’re right! And I think it’s because you have to jump through a bunch of hoops to correct the markets and collect your winnings; as long as the CFTC wages a crusade against PMs, it will be challenging to make much money on them and thus they may often have incorrect prices. But we don’t have to live this way! We could look into the future if we allowed ourselves to.