Pandemics and Hyperinflations

I wrote an article a few years ago about hyperinflation in ancient Rome (and blogged about it here), arguing that the social trust in issuing bodies has been a foundation for monetary value long before modern institutions.

I got a random notification that someone had actually read and cited my work in a recent article “The US Money Explosion of 2020, Monetarism and Inflation: Plagued by History?” I really liked the author’s concept: inflation during pandemic periods is staved off for years because of saving rates, but then the post-crisis period is actually when the most inflation occurs.

This passed my ‘gut check’: during a crisis, who blows their entire budget? It also passed my historical-precedent check, and not only because he researched the Spanish flu and medieval precedent; in the Roman hyperinflation, the inflation lagged decades behind the expanded monetary volume, and in fact came right as the civil wars that nearly brought the Empire to its knees came to an end.

So, in short, inflation-hawks, you are probably right to fear the dramatic expansion of the money supply; however, you won’t feel vindicated for potentially years to come. In an age where people look for causes today to become results tomorrow (EVERY DAY, the WSJ tells me “stocks moved up/down because MAJOR EVENT TODAY”), we need to lengthen our time horizons of analysis and recognize that, just maybe, the ramifications of today’s policies will not really be felt for years. Or, put in a more dire light, by the time we realize who is right, it will be too late to reassert social trust in monetary value, and the dollar will follow the denarius into histories of hyperinflations.

Casual Empiricism: USPS

It looks to me (as I refresh tracking numbers) that the post office is still reeling after several months of attempted voter suppression. It also looks to me like even though Trump is on his way out, there is no reason to believe that someone just as terrible couldn’t come along at any point in the next 50 years and outdo him.

As far as the USPS goes I think there’s a fairly simple solution that should make most people happy: split the USPS in two: a private for-profit firm that delivers junk mail and competes with UPS and Amazon, and a government agency that handles government business including things like distributing ballots and census surveys.

But the USPS is just one small part of a much larger problem. When the Trump II comes along, he’ll have more powers, including (very likely) a lot more power to mess with the health care sector. There are a lot of reasons I don’t like the idea of more government in health care, but this one should be terrifying to everyone.

“A criticism of Indian Americans by an Indian national in the US”

[This is from my inbox, by Vishnu Modur. I reproduce it here with his permission – BC]

This Atlantic article got me thinking. As an Indian national in the U.S., I would like to make a limited point about some (definitely not all) Indian Americans. In my interactions with some Indian Americans, the topic of India induces, if you will, a conflicting worldview. India —the developing political state—is often belittled in some very crude ways, using some out-of-context recent western parallels by mostly uninformed but emboldened Indian Americans.

Just mention Indian current affairs, and some of these well-assimilated Indian Americans quickly toss out their culturally informed, empathetic, anti-racist, historically contingent-privilege rhetoric to conveniently take on a sophisticated “self-made” persona, implying a person who ticked all the right boxes in life by making it in the U.S. This reflexive attitude reversal comes in handy to patronize Indians living in India. They often stereotype us as somehow lower in status or at least less competent owing to the lack of an advanced political state or an ”American” experience—therefore deficient in better ways of living and a higher form of ”humanistic” thinking.

This possibly unintentional but ultimately patronizing competence-downshift by a section of Indian Americans results in pejorative language to sketch generalizations about Indian society even as they recognize the same language as racist when attributed to American colored minorities.

In the last decade, I have learned that one must always take those who openly profess to be do-gooders, culturally conscious, anti-racist, and aware of their privileged Indian American status as a contingency of history with a bucket load of salt. Never take these self-congratulatory labels at face value. Discuss the topic of India with them to check if Indian contexts are easily overlooked. If they do, then obviously, these spectacular self-congratulatory labels are just that — skin-deep tags to fit into the dominant cultural narrative in the U.S.

Words of the economist Pranab Bardhan are worth highlighting: “Whenever you find yourself thinking that some behavior you observe in a developing country is stupid, think again. People behave the way they do because they are rational. And if you think they are stupid, it’s because you have failed to recognize a fundamental feature of their current economic environment.”

Disruption arises from Antifragility

One of my favorite classics about why big businesses can’t always innovate is Clayton Christiansen’s The Innovator’s Dilemma. It is one of the most misunderstood business books, since its central concept–disruption–has been misquoted, and then popularized. Take the recent post on Investopedia that says in the second sentence that “Disruptive technology sweeps away the systems or habits it replaces because it has attributes that are recognizably superior.” This is the ‘hype’ definition used by non-innovators.

I think part of the misconception comes from thinking of disruption as major, public, technological marvels that are recognizable for their complexity or for even creating entire new industries. Disruptive innovations tend instead to be marginal, demonstrably simpler, worse on conventional scales, and start out by slowly taking over adjacent, small markets.

It recently hit me that you can identify disruption via Nassim Nicholas Taleb’s simple heuristics of recognizing when industry players are fragile. Taleb is my favorite modern philosopher, because he actually brought a new, universally applicable concept to the table, that puts into words what people have been practicing implicitly–but without a term to use. Anti-fragility is the inverse of fragile and actually helps you understand it better. Anti-fragile does not mean ‘resists breaking,’ which is more like ‘robust;’ instead, it means gains from chaos. Ford Pintos are fragile, Nokia phones are robust, but mechanical things are almost never anti-fragile. Bacteria species are anti-fragile to anti-biotics, as trying to kill them makes them stronger. Anti-fragile things are usually organic, and usually made up of fragile things–the death of one bacterium makes the species more resistant.

Taleb has a simple heuristic for finding anti-fragility. I recommend you read his book to get the full picture, but the secret to this concept is a simple thought experiment. Take any concept (or thing), and identify how it works (or fails to work). Now ask, if you subject it to chaos–by that, I mean, if you try to break it–and slowly escalate how hard you try, what happens?

  • If it gets disproportionately harmed, it is fragile. E.g., traffic: as you add cars, time-to-destination gets worse slowly at first, then all of the sudden increases rapidly, and if you do it enough, cars literally stop.
  • If it gets proportionately harmed or there is no effect, it is robust. Examples are easy, since most functional mechanical and electric systems are either fragile (such as Ford Pintos) or robust (Honda engines, Nokia phones, the Great Pyramids).
  • If it gets better, it is anti-fragile. Examples are harder here, since it is easier to destroy than build (and anti-fragility usually occurs based on fragile elements, which gets confusing); bacterial resistance to anti-biotics (or really, the function of evolution itself) is a great one.

The only real way to get anti-fragility outside of evolution is through optionality. Debt (obligation without a choice) is fragile to any extraneous shock, so a ‘free option’–choice without obligation, the opposite, is pure anti-fragility. Not just literal ‘options’ in the market; anti-fragile takes a different form in every case, and though the face is different, the structure is the same. OK, get it? Maybe you do. I recommend coming up with your own example–if you are just free riding on mine, you don’t get it.

Anyway, back to Christiansen. Taleb likes theorizing and leaves example-finding to you, while Christiansen scrupulously documented what happened to hundreds of companies and his concepts arose from his data; think about it like Christiansen is Darwin, carefully measuring beaks, and recognizing natural selection, where Taleb is Wallace, theorizing from his experience and the underlying math of reality. Except in this case, Taleb is not just talking about natural selection, he is also showing how mutation works, and giving a theory of evolution that is not restricted to just biology.

I realized that you can actually figure out whether an innovation is disruptive using this heuristic. It takes some care, because people often look at the technology and ask if it is anti-fragile–which is a mistake. Technologies are inorganic, so usually robust or fragile. Industries are organic, strategies are organic, companies are organic. Many new strategies build on companies’ competencies or existing customer bases, and though they may meet the ‘hype’ definition above, they give upside to incumbents, and are thus not fragilizing. Disruption happens when a company has an exposure to a strategy that it has little to gain from, but that could cannibalize its market if it grows, as anti-fragile things are wont to do.

The questions is: is a given incumbent company fragile with respect to a given strategy? Let’s start with some examples–first Christiansen’s, then my own:

  • Were 3″ drive makers fragile with respect to using smaller drives in cars?
    • In my favorite Christiansen anecdote, a 3″ drive-making-CEO, whose company designed a smaller 1.8″ drive but couldn’t sell it to their PC or mainframe customers, complained that he did exactly what Christiansen said, and built smaller drives, and there was no market. Meanwhile, startups were selling 1.8″ drives like crazy–to car companies, for onboard computers.
    • Christiansen notes that this was a tiny market, which would be an 0.01% change on a big-company income statement, and a low-profit one at that. So, since these companies were big, they were fragile to low-margin, low-volume, fast-growing submarkets. Meanwhile, startups were unbelievably excited about selling small drives at a loss, just so that Honda would buy from them.
    • So, 3″ drive makers had everything to lose (the general drive market) and a blip to gain, where startups had everything to gain and nothing to lose. Note that disruptive technologies are not those that are hard to invent or that immediately revolutionize the industry. Big companies (as Christiansen proved) are actually better at big changes and at invention. They are worse at recognizing value of small changes and jumps between industries.
  • Were book retailers fragile with respect to online book sales?
    • Yes, Amazon is my Christiansen follow-on. Jeff Bezos, as documented in The Everything Store, gets disruption: he invented the ‘two-pizza meeting’, so he ‘gets’ smallness; he intentionally isolates his innovation teams, so he ‘gets’ the excitement of tiny gains and allows cannibalism; he started in a proof-of-concept, narrow, feasible discipline (books) with the knowledge that it would grow into the Everything Store if successful, so he ‘gets’ going from simple beginnings to large-scale, well, disruption.
    • The Everything Store reads like a manual on how to be disrupted. Barnes & Noble first said “We can do that whenever we want.” Then when Bezos got some traction, B&N said “We can try this out but we need to figure out how to do it using our existing infrastructure.” Then when Bezos started eating their lunch, B&N said “We need to get into online book sales,” but sold the way they did in stores, by telling customers what they want, not by using Bezos’ anti-fragile review system. Then B&N said “We need to start doing whatever Bezos does, and beat him by out-spending,” by which time he was past that and selling CDs and then (eventually) everything.
    • Book sellers were fragile because they had existing assets that had running costs; they were catering to customers with not just a book, but with an experience; they were in the business of selecting books for customers, not using customers for recommendations; they treasured partnerships with publishers rather than thinking of how to eliminate them.
  • Now, some rapid-fire. Think carefully, since it is easy to fall into the trap of thinking industry titans were stupid, not fragile, and it is easy to have false positives unless you use Taleb’s heuristic.
    • Car companies were fragile to electric sports cars, and Elon Musk was anti-fragile. Sure, he was up-market, which doesn’t follow Christiansen’s down-market paradigm, but he found the small market that the Nissan Leaf missed.
    • NASA was fragile to modern, cheap, off-the-shelf space solutions, and…yet again…Elon Musk was anti-fragile.
    • Taxis were fragile to app-based rides.
    • Hotels were fragile to app-based rentals.
    • Cable was fragile to sticks you put in your TV.
    • Hedge funds were fragile to index funds, currently are fragile to copy trading, and I hope to god they break.
  • Lastly, some counter-examples, since it is always better to use the via negativa, and assuming you have additive knowledge is dangerous. If you disagree, prove me wrong, found a startup, and make a bajillion dollars by disrupting the big guys who won’t be able to find a market:
    • There is nothing disruptive about 5G.
    • Solar and wind are fragile and fragilizing.
    • What was wrong with WeWork’s business model? Double fragility–fixed contracts with building owners, flexible contracts with customers.
    • On a more optimistic note, cool tech can still be sustaining (as opposed to disruptive), like RoboAdvisors or induction stoves or 3D printed shoes.
    • Artificial intelligence or blockchain any use you have heard of (but not in any that you don’t know yet).

So, to summarize, if a company is fragile to a new strategy, the best it can do is try to robustify itself, since it has little upside. Many innovations give upside to incumbents at the marginal cost of R&D, and thus sustain them; disruption happens when the incumbents have little to gain from adopting a strategy, but startups have a high exposure to positive impact from possible adoption of a strategy due to the potential growth from small-market, incremental/simplifying opportunities, which is definitionally anti-fragility to the strategy.

Now, I hope you have a tool for judging whether industrial incumbents are fragile. Rather than trying to predict success or failure of any, you should just use Taleb’s heuristic–that will help you sort things into ‘hyped as disruptive’ vs. ‘actually probably disruptive.’ A last thought: if you found this wildly confusing, just remember, disruptive innovations tend to steal the jobs of incumbents. So, if an incumbent (say, a Goldman Sachs/Morgan Stanley veteran writing the definition of “disruptive” for Investopedia) is talking about a banking or trading technology, it is almost certainly not disruptive, since he would hardly tell you how to render him extraneous. You will find out what is disruptive when he makes an apology video while wearing a nice watch and French cuffs.

Prediction market update

The market for who wins the presidency closed this morning! But the Electoral College margin of victory market was still open and at 98 cents for the already certain outcome. Maxing out my position there would mean $17 for free! So I did, and the market dipped to 97 cents.

This truly is the dumbest jack in the box. We all know exactly what’s going to happen, and yet…

Offensive advantage and the vanity of ethics

I have recently shifted my “person I am obsessed with listening to”: my new guy is George Hotz, who is an eccentric innovator who built a cell phone that can drive your car. His best conversations come from Lex Fridman’s podcasts (in 2019 and 2020).

Hotz’s ideas bring into question the efficacy of any ethical strategy to address ‘scary’ innovations. For instance, based on his experience playing “Capture the Flag” in hacking challenges, he noted that he never plays defense: a defender must cover all vulnerabilities, and loses if he fails once. An attacker only needs to find one vulnerability to win. Basically, in CTF, attacking is anti-fragile, and defense is fragile.

Hotz’s work centers around reinforcement learning systems, which learn from AI errors in automated driving to iterate toward a model that mimics ‘good’ drivers. Along the way, he has been bombarded with questions about ethics and safety, and I was startled by the frankness of his answer: there is no way to guarantee safety, and still depends on human drivers to intervene to protect themselves. Hotz basically dismisses any system that claims to take an approach to “Level 5 automation” that is not learning-based and iterative, as driving in any condition, on any road, is an ‘infinite’ problem. Infinite problems have natural vulnerabilities to errors and are usually closer to impossible where finite problems often have effective and world-changing solutions. Here are some of his ideas, and some of mine that spawned from his:

The Seldon fallacy: In short, 1) It is possible to model complex, chaotic systems with simplified, non-chaotic models; 2) Combining chaotic elements makes the whole more predictable. See my other post for more details!

Finite solutions to infinite problems: In Hotz’s words regarding how autonomous vehicles take in their environments, “If your perception system can be written as a spec, you have a problem”. When faced with any potential obstacle in the world, a set of plans–no matter how extensive–will never be exhaustive.

Trolling the trolley problem: Every ethicist looks at autonomous vehicles and almost immediately sees a rarity–a chance for an actual direct application of a philosophical riddle! What if a car has to choose between running into several people or alter path to hit only one? I love Hotz’s answer: we give the driver the choice. It is hard to solve the trolley problem, but not hard to notice it, so the software alerts the driver whenever one may occur–just like any other disengagement. To me, this takes the hot air out of the question, since it shows that, as with many ethical worries about robots, the problem is not unique to autonomous AIs, but inherent in driving–and if you really are concerned, you can choose yourself which people to run over.

Vehicle-to-vehicle insanity: While some autonomous vehicle innovators promise “V2V” connections, through which all cars ‘tell’ each other where they are and where they are going and thus gain tremendously from shared information. Hotz cautions (OK, he straight up said ‘this is insane’) that any V2V system depends, for the safety of each vehicle and rider, on 1) no communication errors and 2) no liars. V2V is just a gigantic target waiting for a black hat, and by connecting the vehicles, the potential damage inflicted is magnified thousands-fold. That is not to say the cars should not connect to the internet (e.g. having Google Maps to inform on static obstacles is useful), just that safety of passengers should never depend on a single system evading any errors or malfeasance.

Permissioned innovation is a contradiction in terms: As Hotz says, the only way forward in autonomous driving is incremental innovation. Trial and error. Now, there are less ethically worrisome ways to err–such as requiring a human driver who can correct the system. However, there is no future for innovations that must emerge fully formed before they are tried out. And, unfortunately, ethicists–whose only skin in the game is getting their voice heard over the other loud protesters–have an incentive to promote the precautionary principle, loudly chastise any innovator who causes any harm (like Uber’s first-pedestrian-killed), and demand that ethical frameworks precede new ideas. I would argue back that ‘permissionless innovation‘ leads to more inventions and long-term benefits, but others have done so quite persuasively. So I will just say, even the idea of ethics-before-inventions contradicts itself. If the ethicist could make such a framework effectively, the framework would include the invention itself–making the ethicist the inventor! Since instead, what we get is ethicists hypothesizing as to what the invention will be, and then restricting those hypotheses, we end up with two potential outcomes: one, the ethicist hypothesizes correctly, bringing the invention within the realm of regulatory control, and thus kills it. Two, the ethicist has a blind spot, and someone invents something in it.

“The Attention”: I shamelessly stole this one from video games. Gamers are very focused on optimal strategies, and rather than just focusing on cost-benefit analysis, gamers have another axis of consideration: “the attention.” Whoever forces their opponent to focus on responding to their own actions ‘has the attention,’ which is the gamer equivalent of the weather gauge. The lesson? Advantage is not just about outscoring your opponent, it is about occupying his mind. While he is occupied with lower-level micromanaging, you can build winning macro-strategies. How does this apply to innovation? See “permissioned innovation” above–and imagine if all ethicists were busy fighting internally, or reacting to a topic that was not related to your invention…

The Maginot ideology: All military historians shake their heads in disappointment at the Maginot Line, which Hitler easily circumvented. To me, the Maginot planners suffered from two fallacies: one, they prepared for the war of the past, solving a problem that was no longer extant. Second, they defended all known paths, and thus forgot that, on defense, you fail if you fail once, and that attackers tend to exploit vulnerabilities, not prepared positions. As Hotz puts it, it is far easier to invent a new weapon–say, a new ICBM that splits into 100 tiny AI-controlled warheads–than to defend against it, such as by inventing a tracking-and-elimination “Star Wars” defense system that can shoot down all 100 warheads. If you are the defender, don’t even try to shoot down nukes.

The Pharsalus counter: What, then, can a defender do? Hotz says he never plays defense in CTF–but what if that is your job? The answer is never easy, but should include some level of shifting the vulnerability to uncertainty onto the attacker (as with “the Attention”). As I outlined in my previous overview of Paradoxical genius, one way to do so is to intentionally limit your own options, but double down on the one strategy that remains. Thomas Schelling won the “Sveriges Riksbank Prize in Economic Sciences in Memory of Alfred Nobel” for outlining this idea in The Strategy of Conflict, but more importantly, Julius Caesar himself pioneered it by deliberately backing his troops into a corner. As remembered in HBO’s Rome, at the seminal engagement of Pharsalus, Caesar said: “Our men must fight or die. Pompey’s men have other options.” However, he also made another underappreciated innovation, the idea of ‘floating’ reserves. He held back several cohorts of his best men to be deployed wherever vulnerabilities cropped up–thus enabling him to be reactive, and forcing his opponent to react to his counter. Lastly, Caesar knew that Pompey’s ace-in-the-hole, his cavalry, was made up of vain higher-class nobles, so he told his troops, instead of inflicting maximum damage indiscriminately, to focus on stabbing their faces and thus disfigure them. Indeed, Pompey’s cavalry did not flee from death, but did from facial scars. To summarize, the Pharsalus counter is: 1) create a commitment asymmetry, 2) keep reserves to fill vulnerabilities, and 3) deface your opponents.

Offensive privacy and the leprechaun flag: Another way to shift the vulnerability is to give false signals meant to deceive black hats. In Hotz’s parable, imagine that you capture a leprechaun. You know his gold is buried in a field, and you force the leprechaun to plant a flag where he buried it. However, when you show up to the field, you find it planted with thousands of flags over its whole surface. The leprechaun gave you a nugget of information–but it became meaningless in the storm of falsehood. This is a way that privacy may need to evolve in the realm of security; we will never stop all quests for information, but planting false (leprechaun) flags could deter black hats regardless of their information retrieval abilities.

The best ethics is innovation: When asked what his goal in life is, Hotz says ‘winning.’ What does winning mean? It means constantly improving one’s skills and information, while also seeking to find a purpose that changes the world in a way you are willing to dedicate yourself to. I think the important part of this that Hotz does not say “create a good ethical framework, then innovate.” Instead, he is effectively saying do the opposite–learn and innovate to build abilities, and figure out how to apply them later. The insight underlying this is that the ethics are irrelevant until the innovation is there, and once the innovation is there, the ethics are actually easier to nail down. Rather than discussing ‘will AIs drive cars morally,’ he is building the AIs and anticipating that new tech will mean new solutions to the ethical questions, not just the practical considerations. So, in summary, if you care about innovation, focus on building skills and knowledge bases. If you care about ethics, innovate.

The Seldon Fallacy

Like some of my role models, I am inspired by Isaac Asimov’s vision. However, for years, the central ability at the heart of the Foundation series–‘psychohistory,’ which enables Hari Seldon, the protagonist, to predict broad social trends across thousands of galaxies over thousands of years–has bothered me. Not so much because of its impact in the fictional universe of Foundation, but for how closely it matches the real-life ideas of predictive modeling. I truly fear that the Seldon Fallacy is spreading, building up society’s exposure to negative, unpredictable shocks.

The Seldon Fallacy: 1) It is possible to model complex, chaotic systems with simplified, non-chaotic models; 2) Combining chaotic elements makes the whole more predictable.

The first part of the Seldon Fallacy is the mistake of assuming reducibility, or more poetically, of NNT’s Procustean Bed. As F.A. Hayek asserted, no predictive model can be less complex than the model it predicts, because of second-order effects and accumulation of errors of approximation. Isaac Asimov’s central character, Hari Seldon, fictionally ‘proves’ the ludicrous fallacy that chaotic systems can be reduced to ‘psychohistorical’ mathematics. I hope you, reader, don’t believe that…so you don’t blow up the economy by betting a fortune on an economic prediction. Two famous thought experiments disprove this: the three-body problem and the damped, driven oscillator. If we can’t even model a system with three ‘movers’, because of second-order effects, how can we model interactions between millions of people? Basically, with no way to know which reductions in complexity are meaningful, Seldon cannot know whether, in laying his living system into a Procustean bed, he has accidentally decapitated it. Using this special ability, while unable to predict individuals’ actions precisely, Seldon can map out social forces with such clarity that he correctly predicts the fall of a 10,000-year empire. Now, to turn to the ‘we can predict social, though not individual futures’ portion of the fallacy: that big things are predictable even if their consituent elements are not.

The second part of the Seldon Fallacy is the mistake of ‘the marble jar.’ Not all randomnesses are equal: drawing white and black marbles from a jar (with replacement) is fundamentally predictable, and the more marbles drawn, the more predictable the mix of marbles in the jar. Many models depend on this assumption or similar ones–that random events distribute normally (in the Gaussian sense) in a way that increases the certainty of the model as the number of samples increases. But what if we are not observing independent events? What if they are not Gaussian? What if someone tricked you, and tied some marbles together so you can’t take out only one? What if one of them is attached to the jar, and by picking it up, you inadvertently break the jar, spilling the marbles? Effectively, what if you are not working with a finite, reducible, Gaussian random system, but an infinite, Mandelbrotian, real-world random system? What if the jar contains not marbles, but living things?

I apologize if I lean too heavily on fiction to make my points, but another amazing author answers this question much more poetically than I could. Just in the ‘quotes’ from wise leaders in the introductions to his historical-fantasy series, Jim Butcher tells stories of the rise and fall of civilizations. First, on cumulative meaning:

“If the beginning of wisdom is in realizing that one knows nothing, then the beginning of understanding is in realizing that all things exist in accord with a single truth: Large things are made of smaller things.

Drops of ink are shaped into letters, letters form words, words form sentences, and sentences combine to express thought. So it is with the growth of plants that spring from seeds, as well as with walls built from many stones. So it is with mankind, as the customs and traditions of our progenitors blend together to form the foundation for our own cities, history, and way of life.

Be they dead stone, living flesh, or rolling sea; be they idle times or events of world-shattering proportion, market days or desperate battles, to this law, all things hold: Large things are made from small things. Significance is cumulative–but not always obvious.”

–Gaius Secundus, Academ’s Fury

Second, on the importance of individuals as causes:

“The course of history is determined not by battles, by sieges, or usurpations, but by the actions of the individual. The strongest city, the largest army is, at its most basic level, a collection of individuals. Their decisions, their passions, their foolishness, and their dreams shape the years to come. If there is any lesson to be learned from history, it is that all too often the fate of armies, of cities, of entire realms rests upon the actions of one person. In that dire moment of uncertainty, that person’s decision, good or bad, right or wrong, big or small, can unwittingly change the world.

But history can be quite the slattern. One never knows who that person is, where he might be, or what decision he might make.

It is almost enough to make me believe in Destiny.”

–Gaius Primus, Furies of Calderon

If you are not convinced by the wisdom of fiction, put down your marble jar, and do a real-world experiment. Take 100 people from your community, and measure their heights. Then, predict the mean and distribution of height. While doing so, ask each of the 100 people for their net worth. Predict a mean and distribution from that as well. Then, take a gun, and shoot the tallest person and the richest person. Run your model again. Before you look at the results, tell me: which one do you expect shifted more?

I seriously hope you bet on the wealth model. Height, like marble-jar samples, is normally distributed. Wealth follows a power law, meaning that individual datapoints at the extremes have outsized impact. If you happen to live in Seattle and shot a tech CEO, you may have lowered the mean income in the group by more than the average income of the other 99 people!

So, unlike the Procustean Bed (part 1 of the Seldon Fallacy), the Marble Jar (part 2 of the Seldon Fallacy) is not always a fallacy. There are systems that follow the Gaussian distribution, and thus the Marble Jar is not a fallacy. However, many consequential systems–including earnings, wars, governmental spending, economic crashes, bacterial resistance, inventions’ impacts, species survival, and climate shocks–are non-Gaussian, and thus the impact of a single individual action could blow up the model.

The crazy thing is, Asimov himself contradicts his own protagonist in his magnum opus (in my opinion). While the Foundation Series keeps alive the myth of the predictive simulation, my favorite of his books–The End of Eternity (spoilers)–is a magnificent destruction of the concept of a ‘controlled’ world. For large systems, this book is also a death knell even of predictability itself. The Seldon Fallacy–that a simplified, non-chaotic model can predict a complex, chaotic reality, and that size enhances predictability–is shown, through the adventures of Andrew Harlan, to be riddled with hubris and catastrophic risk. I cannot reduce his complex ideas into a simple summary, for I may decapitate his central model. Please read the book yourself. I will say, I hope that as part of your reading, I hope you take to heart the larger lesson of Asimov on predictability: it is not only impossible, but undesirable. And please, let’s avoid staking any of our futures on today’s false prophets of predictable randomness.

Triple-blinded trials in political economy

In medicine, randomized controlled trials are the most highly regarded type of primary study, as they separately track treatment and control groups to determine whether an observed effect is actually caused by the intervention.

Bias, the constant bane of statisticians, can be minimized further by completing a blinded trial. In a single-blinded trial, the patient population is not informed which group they are in, to prevent knowledge of therapy from impacting results. Placebos are powerful, so blinding has helped identify dozens of therapies that are no better than sugar pills!

However, knowledge can contaminate studies in another way–through the physicians administering the therapies. Bias can be further reduced by double blinding, in which the physicians are also kept in the dark about which therapy was administered, so that their knowledge does not contaminate their reporting of results. In a double-blind trial, only the study administrators know which therapy is applied to each patient, and sometimes an independent lab is tasked with analysis to further limit bias.

Overall, these blinding mechanisms are meant to make us more certain that the results of a study are reflective of an intervention’s actual efficacy. However, medicine is not the only field where the efficacy of many interventions is impactful, highly debated, and worthy of study. Why, then, do we not have blinded studies in political economy?

We all know that randomized controlled trials are pretty much impossible in political economy. North/South Korea and West/East Germany were amazing accidental trials, but we can still hope that politicians and economists make policies that can at least be tracked to determine their ‘change from baseline’ even if we have no control group. Because of how easy it is to harm socioeconomic systems and sweep the ruinous results under the rug, I personally consider it unethical to intervene in a complex system without careful prior consideration, and straight up evil to do so without plans to track the impact of that intervention. So, how can politicians take an ‘evidence-based approach’ to their interventions?

I think that, in recent years, politicians–especially in the US and especially liberals and COVID-reactionaries–have come up with an amazing new experimental method: the triple blinded study. Examples include the ACA, the ARRA, and the recent $3 trillion stimulus package. In a triple blinded study, politicians carefully draft bills so that they are (1) too long for anyone, especially the politicians themselves, to read; (2) filled with a mish-mash of dozens of strategies implemented simultaneously or that are delegated vaguely to administrative agencies; and (3) have no pre-specified metrics by which the policy will be judged, thus blinding everyone to any useful study of signal and response.

I am reminded of one of the most painful West Wing episodes ever made, in which “President Bartlett” is addressing an economic crisis, and is fielding dozens of suggestions from experts–without being able to choose among the candidate interventions. Donna, assistant to his Deputy Chief of Staff, tells a parable about how her grandmother would use ‘a little bit of this, a little bit of that’ to cure minor illnesses. Inspired, Bartlett adopts a policy of ALL suggested economic interventions, thus ensuring that we try everything–and learn nothing. I shudder to think that this strategy was ever broached publicly…and copied from fiction into reality.

In this way, politicians have cleverly enabled us to reduce the bias caused by any knowledge of the intervention or its impact. The patients (citizens), physicians (politicians), and study administrators (economists?) are all kept carefully in the dark so that none of them can know how a policy impacted the economy. Thus, anyone debating any of these topics is given the full freedom to invent whatever argument they want, cherry-pick any data they want, and continue peddling their politics without ever being called to task by the data.

Even more insanely, doctors are held not only to the standard of evidence-based medicine, but also to that of of the precautionary principle–where passivity is preferred to action and novel methods are treated with special scrutiny. “Evidence-based policy”, on the other hand, is a buzzword and not an actual practice to align with RCTs, and any politician who actually followed the precautionary principle would be considered ‘do-nothing’. Thus, we carefully keep both evidence and principles of ‘do no harm’ far from the realm of political action, and continue a general practice across politics of the blind making sure that they lead the blind.

In sum, political leaders, please ignore Donna. Stop intentionally blinding us to policy impacts. Stop doing triple-blinded studies with the future of our country. Sincerely, all data-hounds, ever.

Are we over profit maximization?

Friedman: A business is obligated to maximize shareholder value, nothing more.

Everyone else: That’s crazy! Profit maximizing businesses roll over all sorts of other stakeholders and fail to live up to basic ethical standards.

This relates to a complaint I’ve made before. Markets are good at generating prices that reflect aggregate views on the relative scarcity/importance of various goods. Markets aren’t good at charity. To roll other things in there means a good old fashioned price is now a price plus an obligation to do some moral calculus in how we each interact with the complex adaptive system that is the world economy. It’s a recipe for disaster.

So what do we do? We recognize the gap between a world where Friedman’s advice is reasonable and the world we live in, then we figure out how to close that gap. That Friedman’s doesn’t match our world says more about our world than it does about Friedman’s argument.

Rather than move Friedman’s starting point by trying to juggle competing demands of various stakeholders without markets, we should think about the legal framework these stakeholders are acting in.

If we refine our understanding of who has what rights to make what decisions we’ll see that the reason profit maximizers (and vote maximizers) sometimes do bad things is because it’s the best choice available to them. The answer isn’t to say “businesses lobby business therefore they shouldn’t respond to incentives!” it’s to say “therefore we should restrict opportunities to seek rents!”

Coase wasn’t trying to tell us that spillovers don’t matter. He was trying to tell us that transaction costs do matter and whenever they’re present, we need to be careful in allocating rights that have spillover effects. By the same token, we should think of Friedman’s advice as saying “in a perfect world, corporations should maximize profits, but the world needs work.”

Pop Epistemology

I believe in gravity. I don’t believe in the flat earth conspiracy. But I haven’t done the work to verify either. Instead, I trust that some social process of “science” has done a reasonably good job of assembling and verifying the knowledge that keeps my house from collapsing or my car from exploding.

There are some areas where I’m qualified to hold an opinion. But honestly, it’s a pretty small set of things and subject to an infinity of caveats. The things I “know” are really things I believe because they were taught to me by sources I trust. It’s an imperfect system, but it works tolerably well and it frees up my time to do things like working, and having a life. I’m not going to “do my research” because that would mean not doing something with higher marginal benefit.

What Trumpians realize is that sowing distrust in sources of knowledge gives them an advantage in the marketplace of ideas. What’s worse is that they’re not wrong about the fundamental ambiguity of knowledge. I haven’t got enough time, energy, or inclination to verify that the sun will in fact rise again tomorrow. I can’t scientifically test the veracity of claims of what sorts of noodley appendages touch us all.

Do I know that Joe Biden is a better candidate than Trump? If I’m being honest, the answer is no. I’m not terribly comfortable with that, so I might decide against being honest. I know enough to verify that at least one of the candidates is a turd sandwich of a human being.

What I know for sure about this mess is that the problems are complex. Even a well funded team of experts with broad powers would have infinite problems sorting things out. And the sorts of people we try to put in power are less capable than well funded teams of experts with broad powers.

As always, I hope we learn a valuable lesson here. Complex systems are always going to confound our simple human sensibilities. Given the complexity of society, we should avoid aggregating so much power into the hands of politicians–especially when “the other guy” sometimes gets hold of that power.

Interpretation is Everything

I’ve got a thing for models. And COVID has meant a lot of cool little models of disease transmission have been coming across my desk. This has been fun for me. But it’s also an intellectual minefield. Models help us tell stories and think through versions of the world that haven’t happened, but could. And they leave us feeling confident that we understand the world we’re operating in.

But it’s worth remembering a key inescapable fact: you always have to use your best judgment. There are no straightforward conclusions you can get for free without taking a risk of being wrong. A model showing that masks are worth it misses knock on effects. That doesn’t mean the model is useless, just that it only captures one part of the world.

Take the humble supply and demand model. We take a couple of lines, add in some other conditions (e.g. taxes, transaction costs, price controls, etc.), do a little algebra, and voila! You’ve got yourself a conclusion: subsidizing a good will result in people buying more (despite the private benefits of those extra units being less than the private costs). If you find some reasonable estimates of the elasticity of supply and demand for a product you can figure out how much impact a subsidy would have. Ceteris paribus.

All models rely on the ceteris paribus assumption in some form. If a model didn’t hold something constant it wouldn’t be a model anymore, it would just be a copy of reality.

In the case of supply and demand we’re rolling pretty much all the interesting things into that all-else-held-equal assumption. Language, history, legal structure, current events, politics, technology, and all the infinite possible interactions between things. Subsidizing face masks in 2019 would have seemed like a mistake, holding constant the state of affairs in 2019. Sure, we could have figured out that there was some sort of positive spill-over for masks even without a pandemic. But we could have also identified any number of other threats competing for scarce resources.

My advice to students: maintain humility. (My advice to non-students: maintain a student mindset.) Economics provides an incredibly powerful set of tools, but it doesn’t make you a god. There’s no getting around the fact that you’ve got to simplify reality to understand it and there’s no fool-proof formula for identifying things that make sense to hold constant in a constantly changing world.

Why the US is behind in FinTech, in two charts

The US is frankly terrible at innovation in banking. When Kenya (and its neighbors) has faster adoption of mobile banking–as they have since at least 2012–it is time to reconsider our approach.

Here is the problem: we made new ideas in banking de facto illegal. Especially since the 2008 financial crisis, regulatory bodies (especially the CFPB) has piled on a huge amount of potential liability that scares away any new entrant. Don’t believe me? Let’s look at the data:

bank creation

Notice anything about new bank creation in the US after 2008?

A possible explanation, in a “helpful resource” provided to banking regulators and lawyers for banks:

regulatory complexity

This shows: 8 federal agencies reporting to the FSOC, plus another independent regulatory body for fintech (OFAC/FinCEN). Also, the “helpful” chart notes state regulations just as an addendum in a circle…probably because it would take 50 more, possibly complex and contradictory charts.

So, my fellow citizens, don’t innovate in banking. No one else is, but they are probably right.

The Blind Invisible Hand

Kevin recently wrote a post that really tickled my brain. It touches on the computational aspect of entrepreneurship. There are a couple points I’d like to follow up on.

First I’d argue that the uncertain entrepreneur is not the analog of the blind watchmaker. This is a minor quibble, but I think it’s good to keep our language tidy and that includes clarifying our metaphors. The Blind Watchmaker is a perfect metaphor for the emergent order in markets. But the watch is the market as a whole. Any one entrepreneur is just a tiny component of the system–potentially an ingenious component, but always dwarfed by the genius of the system as a whole. The watch maker in biology is the process of evolution. In markets, the closest idea we have is the invisible hand–also an evolutionary process.

Second and more importantly, I’d like to poke at the genetic component of the metaphor to show how much harder social evolution is than biological evolution.

Evolution is a process that acts on the substrate of “replicators”. DNA replicates (in genes) and so do ideas/jokes/norms/etc. (in memes). I guess we could just say “a business model is a type of meme!” and be done with it. But even thinking about what Internet jokes spread means stepping away from the abstract genetic alphabet of strings of A’s, T’s, C’s, and G’s.

The replicators of entrepreneurial evolution occur at more than one level (as I understand it, the idea of multi-level selection is controversial in biology, but inevitable here): little patterns of behavior make up larger patterns. A burger restaurant is sort of like a buffalo. And the business model (e.g. McDonald’s franchise) is sort of like the species as a whole or perhaps something even broader. All the various ways to market burgers compete across a range of niches, but we don’t have a literal genetic code to analyze. We might, hypothetically, be able to isolate the appropriate atomic unit of economic life, but I’m skeptical it would be terribly useful (at least for human understanding).

Still, what entrepreneurial and biological evolution have in common is that they are, fundamentally, complex sets of computations (in out-of-equilbrium systems) on a non-silicon medium. Entrepreneurs indeed face a different situation than genes, but that’s only because they’re dealing with multiple (tangled) layers of evolution spanning large scale things like:

  • human culture,
  • legal systems,
  • economic patterns and business models,

through medium-scale things like the particular landscape of a particular market at a given time and place, down to micro things like the particular ISO specifications of some particular size of bolt.

It’s true that “unlike evolution, you…are trying to achieve something beyond replication…” as an entrepreneur. But at the end of the day a) your apparently high minded goals are really just their own evolving and replicating memes, and b) your apparently high minded goals are really just setting the stage for the atomic unit of evolution that really matters: the proper size and shape of a paperclip. It’s like Dawkins wrote in The Selfish Gene: It’s not really the organism (entrepreneur) that matters, it’s the gene (atomic unit of whatever sort of evolution).

The Blind Entrepreneur

Entrepreneurs usually make decisions with incomplete information, in disciplines where we lack expertise, and where time is vital. How, then, can we be expected to make decisions that lead to our success, and how can other people judge our startups on our potential value? And even if there are heuristics for startup value, how can they cross fields?

The answer, to me, comes from a generalizable system for improvement and growth that has proven itself– the blind watchmaker of evolution. In this, the crucial method by which genes promulgate themselves is not by predicting their environments, but by promiscuity and opportunism in a random, dog-eat-dog-world. By this, I mean that successful genes free-ride on or resonate with other genes that promote reproductive success (promiscuity) and select winning strategies by experimenting in the environment and letting reality be the determinant of what gene-pairings to try more often (opportunism). Strategies that are either robust or anti-fragile usually outperform fragile and deleterious strategies, and strategies that exist within an evolutionary framework that enables rapid testing, learning, mixing, and sharing (such as sexual reproduction or lateral gene transfer paired with fast generations) outperform those that do not (such as cloning), as shown by the Red Queen hypothesis.

OK, so startups are survival/reproductive vehicles and startup traits/methods are genes (or memes, in the Selfish Gene paradigm). With analogies, we should throw out what is different and keep what is useful, so what do we need from evolution?

First, one quick note: we can’t borrow the payout calculator exactly. Reproductive success is where a gene makes more of itself, but startups dont make more of themselves. For startups the best metric is probably money. Other than that, what adaptations are best to adopt? Or, in the evolutionary frame, what memes should we imbue in our survival vehicles?

Traits to borrow:

  • Short lives: long generations mean the time between trial and error is too long. Short projects, short-term goals, and concrete exits.
  • Laziness: energy efficiency is far more important than #5 on your priority list.
  • Optionality: when all things are equal, more choices = more chances at success.
  • Evolutionarily Stable Strategies: also called “don’t be a sucker.”
  • React, don’t plan: prediction is difficult or even impossible, but being quick to jump into the breach has the same outcome. Could also be called “prepare, but don’t predict.”
  • Small and many: big investments take a lot of energy and effectively become walking targets. Make small and many bets on try-outs and then feed those that get traction. Note– this is also how to run a military!
  • Auftragstaktik: should be obvious, central planning never works. Entrepreneurs should probably not make any more decisions than they have to.
  • Resonance: I used to call this “endogenous positive feedback loops,” but that doesn’t roll off the tongue. In short, pick traits that make your other traits more powerful–and even better if all of your central traits magnify your other actions.
  • Taking is better than inventing: Its not a better startup if its all yours. Its a better startup if you ruthlessly pick the best idea.
  • Pareto distributions (or really, power laws): Most things don’t really matter. Things that matter, matter a lot.
  • Finite downside, infinite upside: Taleb calls this “convexity”. Whenever presented with a choice that has one finite and one infinite potential, forget about predicting what will happen– focus on the impact’s upper bound in both directions. It goes without saying– avoid infinite downsides!
  • Don’t fall behind (debt): The economy is a Red Queen, anyone carrying anything heavy will continually fall behind. Debt is also the most likely way companies die.
  • Pay it forward to your future self: squirrels bury nuts; you should build generic resources as well.
  • Don’t change things: Intervening takes energy and hurts diversity.
  • Survive: You can’t win if you’re not in the game. More important than being successful is being not-dead.

When following these guidelines, there are two other differences between entrepreneurs and genes: One, genes largely exist in an amoral state, whereas your business is vital to your own life, and if you picked a worthwhile idea, society. Two, unlike evolution, you actually have goals and are trying to achieve something beyond replication, beyond even money. Therefore, you do not need to take your values from evolution. However, if you ignore its lessons, you close your eyes to reality and are truly blind.

Our “blind” entrepreneur, then, can still pick goals and construct what she sees as her utility. But to achieve the highest utility, once defined, she will create unknowable and unpredictable risk of her idea’s demise if she does not learn to grow the way that the blind watchmaker does.

Wats On My Mind: City Management Games

I’ve been playing a city management game called Sim Empire. It’s a lot like the old classics of Pharaoh, Caesar, or Anno Domini. You are building a town out of nothing, lay out the streets, houses, businesses, and municipal buildings – even houses of worship. The more of your citizens’ needs you can satisfy, the more lavish their homes become – and therefore the more you can collect from them in taxes.

The game has made me aware once again of the sheer beauty of the invisible hand of the market. Here then are some random thoughts on the economy of these types of games.

My citizens don’t have enough grain. If I don’t build enough grain farms, they could starve (in some games, yes). It’s a wonder they don’t revolt and throw me out of office! Oh, but I built enough police stations to cover every square pixel, so they daresn’t, and enough military that no outsider feels safe ‘liberating’ them. If only I allowed free markets, though, some entrepreneurial bitizen would notice the price of grain was high, farm some land, and provide for everyone. No tyrant needed!

The one chief advantage my underlings have is a powerful one: if I don’t provide for their every whim, they will refuse to pay taxes. Apparently my military apparatus is not sufficient to take their money by force, despite being strong enough to remain in power. If I’ve neglected the game for a while due to the pressures of real life, I see 75% of the country simply refusing to pay taxes and nothing to do about it.

Actually, there is one thing I can do about it: go to the free market. What? I thought there wasn’t a free market in my empire. Well, there isn’t, but there is a free international market with no tariffs, quotas, or other restrictions. Well, there is one restriction: no trading outside of 6am-6pm. The one chief advantage Sim Empire has over its older cousins is that I can work with other tyrants. If one has too much wood or grain, there is a marketplace where they can sell their excess to me. I can also sell my excess stone or porcelain.

I’ve noticed, though, that this free market is rather odd. The price of raw materials is higher than the price of finished products. Clay, for instance, right now costs 40-45 gold and wood costs 50, while porcelain – made from clay and wood! – costs 35-40. And you get less porcelain than you put in clay and wood! It’s a real money loser. It occurs to me that I should stop my porcelain factories altogether, sell the clay and wood I used to be using on the market, buy porcelain, and pocket the difference. If enough of us do that, the prices ought to revert. … But why are they doing that in the first place?

I am pleased to announce our Empire runs on hard metal money: gold. No fiat currency here! So no inflation, right? I’m actually dubious. There is no actual limit on the amount of gold I personally can amass, nor on the amount other players can create. The developers never come in to take gold out of the system, so I actually predict as the number of players increase and the amount of gold increases faster than the number of goods being traded, the prices of goods ought to go up over time as well. For an example of real life silver and gold-based currencies, economies, and countries being destroyed by inflation, head on over to Crash Course History for Spain and China.

A reminder that being on the gold standard won’t solve all your problems.