Byrne's Eye View

Most of the comments I get are spam, and most of the spam I get is generic, but in the last few days I've noticed something strange. Before they get down to the business of selling cut-rate pharmaceuticals and porn, spam comments introduce themselves with "Great site! Very informative! Click here for..."

Since spammers are in the marketing business, and everything they do is automated, their market research gets done automatically. So this spam is only common because it works. Someone out there thought "Well. This guy is turning my otherwise decent blog into a Discount Xanax Emporium. But he's so nice about it."

Prediction markets can't have bubbles or crashes, because both phenomena require indefinite extrapolation in one direction -- and betting on specific yes/no questions precludes this. That's convenient. In a conventional financial market, in which most products can swing indefinitely in any direction, this would look like a bubble:

And there's an echo effect from Paul's recent performance that feels like a bubble -- supporters aren't saying "I like Ron Paul," as often as they're saying "Look at how many people like Ron Paul!"

But it's not a bubble yet. Someone with enough financial resources can still bet against this Sancho-playing-Quixote character, and they'll be able to do so knowing exactly how much they can lose.

This is why prediction markets are immune to 1999 (and 1987): in 1999, the story wasn't "Internet stocks are undervalued by X%," but "Too many people shorted Internet stocks Y% ag, and as long as you buy to flip before they buy to cover, you'll come out ahead." Short selling and infinite potential losses can easily create trading opportunities that require you to buy what's expensive and sell what's already cheap (to "run the models in reverse"). I can't see that persisting in a yes/no market.

So one other thing: could other financial markets operate in the same bubble-free space? Sure, but the paperwork would be brutal. Right now, stocks represent an indefinite stream of future income. If you want to avoid speculation, split each stock into a series of bets on how much they'll make each quarter ("Less than $10 million," "$10-20 million," "$20-30 million,") etc. Then, cut off the 'tail' ("They'll earn some amount more than $50 million,"), give it to a charity, and stipulate that it must be spent and can't be sold. This turns equities into a fine-grained fixed-income/preferred product, demonstrating a fundamental law of finance: you can get rid of speculative behavior, but only if you're willing to forgo fun.

It isn't hard to instantly win any argument: It is undeniable that someone who believes something either a) lives according to their principles, or b) does not. This makes everyone either a) self-serving, or b) a hypocrite.

As long as arguments are about people and not truth, a) or b) offers a guaranteed one-sentence riposte to any conceivable position.

Did pulling out of Vietnam represent the opposite of the sunk cost fallacy?

This is a line I've heard a lot lately (not couched in economic terms, of course), so I can't tell if it's opportunism or a real change.[1] But what's more interesting is the general concept: are there situations where the marginal gain is positive, but the total return is lower than expected, in which people avoid making a marginal investment because they're disgusted with their earlier losses?

[1] The way I heard it, the only question about the Vietnam war was how much it would cost the US to lose.

You might consider reading Michael Lewis's piece on pricing natural disaster risk. Or you might not: Lewis knows who he's writing for, so he never misses an opportunity to veer towards friendly metaphors ("In nature’s casino, they had set themselves up as the house, and yet they didn’t know the odds..." not exactly. Given enough consecutive good years in the insurance business, the average insurer will act like a moron because in the short run, it pays -- the average intelligence of insurers is an average from good years and bad, so it would be implausible if they weren't disproportionately stupid in advance of herd-culling catastrophes).

I hadn't thought of this before, but isn't there a conflict of interest when private equity companies 1) issue lots of high-yield debt, and 2) make money trading distressed debt? Even if it doesn't cause anything sinister, it makes the company more willing to chase edgier deals: at best, it's a clever way to diversify.

It feels like 1999 (or, uh, 1987[1] or early 2007) all over again: investors have earned a 200%+ annualized return on a ten-figure investment.

Goldman Sachs Group Inc.'s Global Equity Opportunities hedge fund rose 12 percent last week after the securities firm shored up the money-losing pool with $3 billion of cash, investors said.

...

New York-based Goldman put $2 billion into the fund and raised $1 billion from investors including Maurice ``Hank'' Greenberg, the former chairman of American International Group Inc., and billionaire Eli Broad.

[1] This is conjecture. I've read a lot about 1987, but at the time I wasn't tracking the market. Also, I hadn't learned to walk yet.

Bernanke should raise rates a hundred basis points. The Dow would drop 8% or so, but we'd all be better off -- we'd finally be rid of the Sawtooth of Stupidity. The Sawtooth is the pattern that goes like this:

1) There is bad news, so stocks drop.

2) The drop in stock prices is interpreted, by the Fed, as further bad news.

3) The Fed cuts rates, so stocks return to their previous level.

That's almost fine, but the problem is that between 2 and 3 is a huge amount of pointless speculation -- Fedwatching, instead of actual investing or even old-fashioned speculating. Traders end up thinking something like "The subprime crisis could be slightly worse, in which case the Dow drops slightly, or really worse, in which case we get a rate cut and rally 200 points tomorrow." There's an eerie space where expected returns go up as fundamentals deteriorate, and that space is entirely due to the Fed reacting to reactions instead of reacting to fundamentals.

It's not hard to put a stop to it. Today, the market went up because traders thought the economy might go down enough for the Fed to wake up and push interest rates down. Sawtooth. All the Fed has to do is invert the pattern once, and let the market swing faster than the economy rather than the other way around, to make the Sawtooth vanish for good.

Second Life now features deposit insurance. Like everything else in Second Life, it's largely pointless except in the context of a blizzard of pointless activity that adds up to lots of fun.

What I have to wonder about is this: are there any Second Life exchanges? Has anyone standardized contracts? Do any of the Second Life banks discuss how they lend their money, and do any of them lend their money to people speculating in other banks' deposits?

Today's Wall Street Journal reports on a rumor of a rumor: perhaps someone might think that someone else might claim that Warren Buffett is going to bail out the mortgage industry. It's cargo cult finance: what contrarians do is unexpected, thus what is unexpected is what contrarians will do.

Not really.

Buffett isn't (yet) betting on subprimes, but lots of people are betting on Buffett: Berkshire Hathaway is up 9% this month, even though the market is down by about 5%. And why? Berkshire hasn't changed extraordinarily since then -- it's just a flight to quality. Or rather, it should be -- but equally solid insurance companies like White Mountain are down as much as the broader market, which makes me suspect that investors aren't just investing.

They're praying: buying Berkshire is evidence that they expect everything to go down, and something to go down more than it should -- and they're outsourcing the details to Warren Buffett. But Berkshire Hathaway has been the same crypto-put for about two decades, so revaluing it on the basis of data older than I am is hardly rational. I hate to say this, and I hate to even think that way, but: if I were a shorter-term trader, I'd be shorting Berkshire Hathaway right now

The other day, I mentioned a currency crisis in Second Life: a few SL bankers have discovered that fraud pays better than business, and now that they've been found out, it's hurting the entire economy.

In what's probably unrelated news, Eve Online has hired an economist. So far, Eve's inflationary crises have been mild and quickly averted, and their economy has achieved something notable: you can ignore it. The obvious reaction is to note that this illustrates the importance of regulation rather than unfettered market craziness. But take a look at the details:

"As a real economist I had to spend months trying to find data to test an economic theory but if I was wrong, I wasn't sure if the theory was wrong or the data was wrong. At least here I know the data is right," Guodmundsson said.

As new players join, [the game's parent company] adds new planets and asteroids that can be exploited, one of several "faucets" that serve to inject funds into the universe and keep the economy ticking.

So what it really illustrates is that an economy is more stable when it's run by all-seeing, all-knowing supply-siders. Unfortunately, it's the nature of information to reproduce itself and the nature of technology to cheapen itself, so centralization is an increasingly silly anachronism. Eve Online has an economy that runs like broken clockwork -- as long as you're constantly retuning and resetting it, it sporadically resembles reality. But they could offer something closer to a real economy -- one in which natural resources are scarce and most wealth takes the form of skills and knowledge. This, incidentally, is a great description of Second Life.

Second Life is a mess, but it's a realistic one.

In case you're wondering why I haven't blogged much lately: my girlfriend and I have consolidated book collections, and given each other to-read lists. Top shelf, hers. Bottom shelf, mine. Shelf flipped for easy reading (click for full size):

These are actually the emergency to-read shelves. The actual to-read lists are about three times as big.

I'm glad Roger keeps us up-to-date on this stuff: there are now two exchanges for trading large-cap unregistered stocks.This was inevitable; the problem with regulation isn't that the SEC is too strict or too lenient, but that, no matter what, they're not precisely as strict as everyone wants them to be. The only way to achieve that is to have multiple regulation schemes, or one scheme flexible enough to handle every preference.

For now, it's a problem for the ibanks. But in the future, I think it's going to be solved by startups wielding smart contracts.

Patent trolls at SCO finally lost their lawsuit against Novell, in a painfully tedious dispute over who owns Unix -- the operating system and oral tradition used by 90% of the top 10% of developers, dot.commers, and computer scientists. I won't go into the details of the case, because I'm not an expert (I'm a Unix user, but a ruling in favor of SCO is about as enforceable as the happy birthday copyright, so it doesn't affect me).

What interested me about this news was how many coders profited from it. We all benefit as consumers, of course,[1] but a few people profited directly by betting against SCO's stock. Shorting SCO (a good idea, by the way) was a popular maneuver among open-source developers. But that sounds like a terrible strategy to me: the lawsuit's effect on them would be the opposite of it's effect on SCO, so shorting SCO meant doubling down on a bet they didn't want to make in the first place.

This argument didn't appeal to anyone, because nobody in the open-source world cheered for SCO. But left-ish political leanings are common (not universal) in the open-source world. So to balance out the positive connotations of SCO's legal loss, I'd ask: would you feel the same way about a pro-war Republican who made a killing on Halliburton stock? It's the same maneuver -- aligning financial interests with moral beliefs -- but nobody I know would clap and call it clever.

[1] Google would probably have to start charging for search queries of Linux wasn't free.

There's a bank run in Second Life.

I've read enough to wonder if someone could make money in Second Life finance. As far as I know, you can't create an effective bank in Second Life because 1) in a post-scarcity economy, nobody needs to borrow anything, 2) if someone wants to borrow something, their assets won't produce income, and 3) in a game, entertainment is more important than value, so it's better to turn your bank into a casino than to run it as a bank.

If you really wanted to succeed as a Second Life banker, you'd have to do what American banks used to: since there wasn't an FDIC, they'd spend ridiculous sums on their offices so they had a lot to lose by leaving. If a Second Life bank spent half of its assets hiring well-known SL designers to make the offices look as gaudy as possible, it might fool people into making further deposits.

Of course, this would give you lots of interest-bearing liabilities and no income-generating assets. How you'd resolve this problem is left as an exercise to the reader.

There's a simmer backlash against Facebook, Myspace, and other social networks because they turn their back on the open tradition of the web. As far as I know, they have to: it's hard to make money from a product if you don't control how they use it, and the best way to do that is to control how you view your data so you view it with their ads.

But it's not at all hard to import data from one of these networks, and there's no reason the network element can't survive when the information is hosted elsewhere. So there's an opportunity to create an open social network with the following features:

• Easy updating of Myspace, Facebook, LinkedIn from a central site: if you change jobs, you shouldn't have to edit every profile.
• Easy backup: if Myspace loses data or Friendster goes under, you shouldn't lose the profile or friends-list you spent so long crafting.
• Indexing of people with similar interests: as far as I know, this is what makes these companies so successful: if you're interested in, say, Mozart, you can click on the word 'Mozart' in a profile and get a list of everyone with the same interests. For the open social network to succeed, it would need the same scale as the rest -- which means making it easier to add an existing account to the new service than to join the network.
• Open access, when appropriate: scriptable privacy filters (showing your work-related blog entries to everyone -- except your boss) would be even better than Facebook's decent privacy protection or LinkedIn's trust-based model.

The best way to fix social networks is to transcend them: the 'walled garden' approach was fine when it was the best we could do, but users have outgrown it. A better social network is one startup away.

The Minneapolis bridge collapse was a disaster for the language, too:

investigating professionals will remain on scene. They will remain on scene until they understand and see what, in fact, they believe is the causal factor here or a probable series of causal factors.

Yes. The "causal factor". I bet this disaster recovery business is exhausting -- he can't wait to go to sleep in a nice, warm unit of bedular furnishing.

It reminds me of this: when you analyze the text for patterns, law doesn't look like English -- it looks like bytecode.

I first heard of this in a programming textbook. When you're dealing with human beings, you can pack information more densely by being less accurate: similes, analogies, and references all make it easy to say a lot with a little. When you're writing legal code or computer code, you have to make everything you say explicitly clear, because colliding interpretations make vague statements worse than useless.

I can think of an easy way to let government employees talk like humans, without worrying about two million-dollar commas and the like: port the legal code to Prolog, and treat commentary as comments.

How is Japan rich despite lacking natural resources, while Nigeria, with its abundant oil supplies, is poor? One surprising answer is monetary: Japan invented money more or less independently, while Nigeria borrowed the idea from elsewhere.

When someone invents money, they do it for a reason: they need an easy way to store and transfer value. This leads straight to the question of what form value takes, and that usually leads to an answer involving precious metals. [1] But a better answer might be "Value is what we're working to get," and thus "A currency we work for should be backed by what we'll sell it for." Indirectly, gold is a good answer, since it's also a store of value so it's an approximate representation of what you're working to get.

But a better example would be whatever you expect to spend it on. This is how it worked in Japan: their currency was the koku, a measure of weight that represented the amount of rice a person would consume in a year. This is simple, obvious, and brilliant: for people who spend most of their money on food, this means no inflation -- and for people who mortgage their farms, it means their debt is proportionate to their income.

The koku is obsolete, but the idea behind it isn't. If Nigeria launched an oil-backed currency tomorrow, they would 1) create a banking system out of their oil industry, 2) solve the inflation problem, and 3) become a better store of value than the euro, the dollar, or gold.

The implementation is easy: instead of paying people in Naira (or under the table in dollars), Nigerian oil companies offer them certificates redeemable for the value of 1/100th of a barrel of oil, payable on demand in dollars, Euros, pounds, or whatever else. Over the next few years, they refinance their debt in Petros, so the amount they owe is proportionate to the value of the oil they own.

This would radically recapitalizing the country: the pre-Petro economy has flucuating assets and constant debt; the post-Petro economy will match the value of debt to their income from oil. Nigeria doesn't need a speculative oil portfolio, and nobody needs to be beholden to a volatile world market: the Petro would let Nigeria sell oil short to balance its exposure, while keeping the value of any gains in productivity.

"A debt denominated in income" should be a familiar concept: it's another way to define a stock. In Nigeria's case, though, this isn't a typical stock: it's a 'tracking stock' that reflects the performance of the most valuable and least desirable part of their economy.

This plan isn't hard to criticize, so I'll address a few objections:

• It's confusing. It is. It shouldn't be. What's more intuitive -- a store of value backed by something of value, or a currency that derives its value from being valued as a currency? Circular logic works when it has to, but we shouldn't use it when there's a better alternative.


• It's hard to implement. Also true. Less true of Nigeria than anywhere else. I can't think of another country with a lower price/book value. If you're going to have a revolution, you should target whatever structure is most decayed. Nobody's going to miss the Naira, and nobody's going to fight too hard to keep it around in the face of a better alternative.


• It encourages fraud. Letting private companies issue currency does increase that chance that someone will be defrauded, but it decreases the magnitude of the fraud. Right now, Nigeria's central bank has a monopoly on currency, so the Naira's value is only stable as long as the people at the bank are utterly incorruptible. When the Petro is implemented, a few Dunn & Bradstreet types will get in the business of valuing Petros (or insuring them). In a market economy, insurance is synonymous with scale, so these guys will be bigger than the companies they insure -- they'll have more to gain from honesty, and more to lose from fraud.[2]

Nigerias has a vast amount of oil, and they have a lot to gain by getting money for it before they even have to extract it from the ground. Introducing an oil-backed currency is a way to forget about speculators and forget about inflation -- it gives them the escape velocity to leave the Third World.

What are they waiting for?

[1] That's how it worked recently. The earliest Greek currency was a sixty-pound slab of iron representing a cow. And Ithaca, NY, has a currency that represents an hour of labor -- good news if you're a janitor, not so great if you're a neurosurgeon.

[2] That's a subject for another essay, but a quick defense of the 'insurance = scale' statement would be that the Federal government is, roughly, a very large company that reinsures us against the risk of war, and insures us against the risk of being poor when we're old, losing our jobs, not being able to drive where we want to, getting mugged, getting hit by a hurricane when we live below sea level, etc., etc. etc.

By popular demand, I'm switching from markdown back to normal HTML, because people who subscribe to my RSS feed are getting the raw markdown, which is no fun at all.

Often the most unlikely of rumors caused panic in the markets. In two years, for example, Paul Volcker resigned from his post as chairman of the Federal Reserve seven times and died twice.

--Michael Lewis, Liar's Poker

Rumors survive because they reproduce faster than any other kind of thought. Any individual rumor dies out quickly, because on the whole they aren't selected for longevity.

If there's a neo-Darwinian view of rumors, that is it: they're like viral infections that take advantage of the normal process of sharing information. They hijack the mechanism because false-but-compelling notions spread faster than true-but-boring ones, even if the compelling twist is a random mutation or a transcription error ("Fed Chief resigned to lower interest rates," can be parsed in two very different ways). In terms of selection theory, they fit the r-side of the chart: rumors reproduce early, reproduce many at a time, and tend to be small (does anyone ever passed on a 5000-word exegesis when the kernel of exciting misconceptions would do?).

Recast in economic terms, I like to think of rumors as a currency that can be infinitely duplicated by whoever owns it: if I have a fact that could be worth $10,000 to me, I might share it with ten accomplices, each of whom expects about half the benefits. That way, I've collected about $50,000 of IOUs (which, in this context, I'd probably end up exchanging for more baseless rumors).

Like any hyperinflationary currency, rumors are a popular trade because they're fun: limited duration (no one cares once everyone knows) and highly variable outcomes make them about as volatile as short-term, out-of-the-money options, so they attract a similar breed of speculator.

Options, of course, have standard contracts and institutionalized markets. Truemors is a start -- but it's a bad one. What would be interesting is a rumor market that lets you assess the 'fundamentals' (the reputation of whoever shares the rumor -- or just a way for people to make public bets that their information is good).

Paul Graham asks: "Could you reproduce Silicon Valley elsewhere, or is there something unique about it?." His answer: it's unique, but if you want to replicate it a few of the right people are all you need.

How would you recreate Manhattan? Like Silicon Valley, it's an enormously productive area despite a dearth of natural resources (besides access to water, which is not as valuable as it used to be). Most of the industries here rely on skilled, but not ridiculously skilled, labor (you couldn't have Silicon Valley without a few professors in the top .1% of their fields, but the media, finance, and law industries here can be quite successful without that kind of elite academic backing -- as long as the city still has a healthy fraction of the top 5% in each field).

What gives Manhattan its competitive advantage? The cost of living is so high that time is too valuable to waste.

Wasting someone's time is an easy way to impose a massive externality: if I interrupt a bond analyst or a tax attorney for a minute, I've spent a few dollars of someone's time at a lower cost to myself -- and it's unlikely that my interruption will provide that much gross benefit to either of us. But if I'm already paying for my time, I'm much less likely to waste it pestering people with a small probability of success -- it's a flat tax on sales spiels, sob stories, product pitches, and every other trick for converting a certain loss of time into a possible transfer of money.

Time here is valuable because real estate is valuable, and we're all forced to consume some real estate just to stick around. It's a virtuous cycle: people who can't afford the rent are priced out of the city, so the people who do live here are wealthier on average, so anyone who wants to spend time with wealthy people (e.g. someone in finance, law, or the media) has an incentive to move here, pushing prices up again for another round of the same.[1]

Three hundred years ago, geography made New Amsterdam a place worth living by bringing ships in; today, geography makes Manhattan worth the steep cost by keeping poor people out -- since we're on an island, sprawl goes from a continuum to a binary equation, lopping off most of left half of the net-worth bell curve.

It doesn't take an island to pull this off, though: if you really wanted to create another Manhattan (or an even Manhattanier Manhattan), all you recreate is the steep cost of showing up: a $5,000 annual fee for living in the city would suffice. Of course, no one will pay $5,000 to be the first person there, so you might require a few hundred highly skilled 'founders' to start the first couple hedge funds, law firms, and cable TV stations. Thereafter, the incentive structure is easy to understand: if you want to deal exclusively with very rich people with very valuable time, borrow whatever you can and move in (keeping in mind that valuable time for rich people means high wages for limo drivers, janitors, and assistants of all sorts). A few provisions for expelling people who commit crimes would also be helpful.

The most compelling part of this thought experiment is that the cost of an actual government would be absurdly cheap. If you made a Venn diagram of "People with access to $5,000," in one circle and "People who commit lots of crimes," in the other, the sliver of intersection would be the entire criminal population -- a police force about 10% as large as that of a no-fee city wouldn't be out of the question.

It's hard to imagine creating this in an existing city (expelling all the poor people wouldn't be good PR, even if what's left is an economic Garden of Eden), but one trick might be to do what Social Security did, and change the legal but not economic incidence of the tax; instead of charging people to be there, you'd just charge their employer (or school) to have them there. [2]

I'm not sure who would want to live in this kind of city, but a good way to guess would be to ask: who wouldn't be able to live there, and who would be happier without those people? Someone with lots of ideas and no plan for implementing any of them is out until the first success; this gives venture capitalists more time to focus on genuinely interesting plans. Aspiring writers with movie scripts and no movie experience can't afford the entry fee; film directors can use the hassle-free time as a dividend to work on more projects. Resident's won't participate in high total, low average class-action lawsuits, so the lawyers who create those won't prosper, while more productive ones (who craft low-dispute merger contracts and effective tax strategies) will. Depending on whether your taste in books leans more to Upton Sinclair or Ayn Rand, this is either hell or paradise -- a condition that's usually a good sign of an underexploited niche.

Most people probably agree that it's better to have Manhattan than to not have it, so I have to wonder -- why not create Manhattan, only more so?

[1] Indirectly, we spend more on real estate than the cost of our rent, because the square footage taken up by laundromats and restaurants and power plants also costs that much. If you want to figure out the difference, the best way would to to sum up all real estate value in New York, and find out how much it would cost to buy the same amount of land in, say, the middle of Kansas. That would get a rough estimate of how much people pay to price poor people out of New York.

[2] The effect is the same (whether I have to pay someone $105,000 so he'll willingly pay the $5K, or $100,000 plus $5K to the government, our balance sheets look the same), but the PR is much better.