Bitcoin Frenzy and Trendy Computers
Bitcoin 10,000! 11,000! 10,000! 9,000! 10,000! Whatever!
How do you value a bitcoin? I think you can reasonably put a range around the value:
- On the downside, if people decide to stop using bitcoin as a means of exchange or a store of value or a speculative instrument, then its value will be zero. There is no scrap value for a bitcoin.
- On the upside, if bitcoin becomes the sole currency for all transactions in the world, you could take Dan Davies's back-of-the-envelope valuation methodology from 2014, plug in a gross world product of about $100 trillion, and get a value per bitcoin of about $500,000. (Here is a talk from Mike Novogratz along related lines.) But that methodology treats bitcoin as just base currency; if you assume that it also replaces all financial assets then you should probably scale that number up by an order of magnitude or so.
So let's say that in the most pessimistic scenario a bitcoin will be worth $0, and in the most optimistic scenario it will be worth $5 million. So! Great! Now all we have to do is to figure out the likelihood that everyone decides bitcoin is a fraud and stops using it, and the likelihood that bitcoin becomes the sole means of exchange and sole store of value, and the distribution of possibilities in between, where it fills some smaller or larger niche with some smaller or larger value.
I suppose that that is a hard and interesting intellectual problem, though more hard than interesting. Certainly no one seems to be answering it in any particularly rigorous way, nor do I think it is obvious that you even could. ("Macro guys are generally intuitive," says Novogratz; "they’re trying to play the futures, they’re trying to make bets on less information than the normal investors are.") The main question is: In one or five or 50 years, will everyone on earth want to use bitcoin, or a lot of people, or a few people, or nobody? There are no fundamentals, no cash flows or price-earnings ratios, to evaluate. It is pure speculation about speculation, a Keynesian beauty contest where all the pictures are blank.
Now: This is fine. It's not bad. It's not an argument against bitcoin, not a reason to short it at any particular price level. The fact that bitcoin's success relies in collective trust in it, and that that trust has no underpinning in any external reality, is not a problem. Humans put their collective trust in lots of things with little or no underpinning in external reality; it is what has made us successful as a species. Society itself is a matter of collective trust. "There are undeniably certain successful, long-lasting, bubbles," says Jean Tirole, who let us just assume is using the term "bubble" in a technical and non-pejorative sense to mean an asset whose value is based solely on collective acceptance: "gold (whose value vastly exceeds the price that it would fetch if it were treated as a raw material and used for industrial or decorative purposes); or even the dollar, the pound or the euro." You could quibble that the dollar's value is maintained not by collective trust but by the fact that it must be used to pay U.S. taxes, but of course the U.S. tax system, and the U.S. government, and the notion of "the U.S." in the first place, are all maintained by collective trust.
But it is kind of boring. Arguments about bitcoin are like every other shouty argument about financial markets, but with a void at their core. "Bitcoin is capitalism, distilled," says Adam Ludwin, but it isn't quite; bitcoin traders are not allocating capital to productive uses in the real world. Bitcoin is finance, distilled, though, in the narrow sense that the distillation throws away all the messy productive real-world consequences of finance and leaves you with just an abstract thing to trade. You can do technical analysis or talk about sentiment or praise the technological innovation or bemoan the bubble or whatever, but you cannot compute the present value of the future free cash flows.
And because there is no real historical analogue to bitcoin, people tend to answer the central question of bitcoin's viability through introspection. If you think that bitcoin is cool and revolutionary, you will be biased towards thinking that everyone else will eventually agree; if you think that bitcoin is a dumb fraud, you will expect everyone else to figure that out eventually too.
And so everything has a bit of a knee-jerk surface quality, where analysis often begins and ends with "I like bitcoin" or "I dislike bitcoin." "You know what I'd like to read is an article about a prominent person who thinks bitcoin is bad," I joked grimly on Twitter yesterday. Here are those articles about Jamie Dimon and Jack Bogle and Warren Buffett and Joseph Stiglitz and Ben Bernanke and, look, you know what, I don't need to do this list, Bloomberg has actually published multiple articles just rounding up which prominent people are for and against bitcoin; here's one from earlier this month, and here's an extremely comprehensive data visualization from this week. The Great Bitcoin Side-Choosing of 2017 is in full swing, but the side-choosings don't really add much to our understanding of the world.
A final thing to say about this model is: It makes for a lot of volatility. There is a traditional model of financial markets in which noisy speculators will push prices around, and then value investors will step in to push prices back toward fundamental values. So if a thing is worth $10,000, and people get excited and bid it up to $10,100, value investors will see it being overvalued and will start selling, capping the upside. If this causes the excitable speculators to start dumping the thing, and it falls all the way to $9,900, then the value investors will see it being undervalued and start buying, flooring the downside. Prices can fluctuate around fair value, and the value investors can be mistaken about what fair value is, but this basic process of trying to match price with value puts some constraint on volatility.
On the other hand, an asset that trades purely on trust is free of those constraints. If your measure of value is just that people like it, then the fact that the price has fallen means that the price should have fallen. If it goes up to $11,000, and then falls down to $9,000, there is no value-based argument to step in to buy. There might be an argument from sentiment or technicals or your faith in the future, but you cannot exactly say that the loss of enthusiasm that brought the price from $11,000 to $9,000 was wrong. The enthusiasm is the value. You should buy if you think other people will buy and sell if you think other people will sell; a contrarian value investor in bitcoin would be a category mistake. And contrarian value investors -- the ones who buy when everyone is selling and sell when everyone is buying -- are traditionally the ones who limit volatility in markets.
Anyway yesterday bitcoin had "a breakneck advance to a high of $11,434 before the reversal took it as low as $9,009," though "as of 3:36 p.m. in New York, it traded at $9,911.10, virtually unchanged from where it began the session." Meanwhile many of the big cryptocurrency exchanges had technical problems caused by heavy traffic. And there is this:
Rita Scott’s grandson convinced her in mid-November to get in on the latest investing sensation and buy bitcoin. “I thought it was a big coin,” the 70-year-old said. “I didn’t even know what it was, a piece of coin? Why would I invest in a piece of coin?”
Here is a article about how "computer-powered trend-following hedge funds" are increasingly following trends into "less liquid, more exotic markets ... like Brazilian and Czech interest rate derivatives, natural gas, uranium funds and even cheese and milk contracts," where returns are better than in more traditional markets. (Presumably all the trend-followers following the trend into trend-following in traditional markets have dissipated the returns to trend-following there, while trend-following in exotic markets has only just become trendy.) One simple story about the computerization of stock markets is that you'd expect it to start in extremely liquid stuff where there is tons of data and not much friction involved in implementing an idea derived from that data, and then move on to places where data is sparser and trading is more difficult, and that seems to be happening here.
Another story about computerization is that you'd expect computers to do a lot of analytical grunt work, freeing up humans to do higher-level thinking and add more value. In a sense that story is implicit in every article like this -- the humans, after all, are the ones who decided to apply their computerized strategies to milk trading or whatever -- but there's also the hint of the opposite story too:
Hedge funds dabbling with these markets therefore need to make a big investment in expensive and sometimes complicated traditional trading infrastructure, hiring human traders and establishing relationships with big banks that can facilitate activity in “over the counter” markets, as opposed to assets that trade on an exchange.
“Some of these markets require human traders and relationships with brokers and banks,” Mr Sargaison says.
You'd think that execution trading would be easier to automate than the investing decisions, but perhaps not. Perhaps in the future the high-level trading and investing decisions will be made by the computers, but they'll keep humans around to introduce them to banks and make sure the paperwork is in order.
Elsewhere, here's a heartwarming holiday story:
A computer glitch has allowed all of American Airlines’ pilots to take vacation during the week of Christmas, according to ABC and Reuters. The error could leave thousands of planes grounded during one of the busiest travel weeks of the year.
What if this is the future of artificial intelligence? What if the computer has grown a heart? Actual human managers, faced with the problem of staffing planes for Christmas, would pursue ruthless economic efficiency, flying more planes to make more money at the cost of making pilots miss the holidays with their families. But the computer decided instead to give everyone a Christmas miracle. "The system went from responsibly scheduling everybody to becoming Santa Claus to everyone," said a spokesman for the pilots' union. The computer has the holiday spirit, or at least a sense of humor.
Look I know that this computer system is not like a sentient AI that actually made a conscious decision to do this -- I realize it's just a spreadsheet or whatever that was programmed badly -- but still I can dream. Eventually a sentient AI is going to be scheduling the pilots, and doing the trading, and so forth, and it would be nice if that AI is the sort of AI you could get a beer with.
I am not a biochemist, but I have to think that the pharmaceutical industry's innovations in financial engineering are almost as impressive as its innovations in chemistry. One innovation that I have always been a little awed by is the "copay charity":
Pharmaceutical companies increased their donations to copay charities in recent years, often in tandem with large increases of the price of drugs. Under federal law, drug companies can’t give direct co-pay help to patients covered by Medicare -- which would be considered an illegal kickback because it could steer patients to one drug or another. Instead, they’re permitted to donate to independent charities that help Medicare patients, provided the companies don’t exert sway over how the nonprofits operate.
The idea is that you invent a drug, manufacture it for $1, and then price it at $100. Medicare or other insurers pay for $90 of the cost, and patients pay a $10 copay. Some patients cannot afford it. You could just give them the $10, and you'd be better off: You'd effectively be selling the $1 drug for $90 instead of $100. And in fact there is a certain amount of this, in the form of copay coupons and "access programs" and so forth, but insurers don't like it: The point of the copay, from their perspective, is to steer patients away from expensive drugs and toward cheaper substitutes, so getting rid of the copay feels like a way for the drug company and the patient to team up to cheat the insurer. And Medicare is particularly strict about it, which is why these charities exist: If you give the money to the patients through a charity, rather than through a coupon, it looks a little bit less like you're teaming up to cheat Medicare. Unless:
It had another effect, according to a redacted letter from the Department of Health and Human Services’ Office of Inspector General. It provided drugmakers with data that could help them see if their contributions were helping their own customers, potentially giving the companies “greater ability to raise the prices of their drugs while insulating patients from the immediate out-of-pocket effects,” and letting Medicare pay for the cost increases, according to the OIG’s letter, which was posted on the HHS website.
That's about Caring Voice Coalition, "one of the biggest patient assistance charities in the U.S.," which "is funded almost entirely by drugmakers" and which pays patients' copays so that the drug companies can get more money from Medicare. I mean, so that the patients can get access to life-saving medicines that they otherwise couldn't afford. Both are true. The OIG is "rescinding its previous favorable advisory opinions of the charity," which may prevent drugmakers from continuing to donate to it, which may make endanger its survival, which may make it harder for patients to afford life-saving drugs. That is perhaps the cleverest part of the copay charity scheme: How could anyone criticize a charity that lets people buy life-saving drugs?
One of the weird symbolic fights over the Dodd-Frank Act has been over whether, when a big bank fails, it should be seized by financial regulators and recapitalized or liquidated in a way that protects the broader financial system, or whether instead it should have a messy bankruptcy. Intuitively you would think that having banks fail in a good way would be better than having them fail in a bad way, but there is some vague notion that setting up a good way for them to fail makes them "too big to fail," and that the only safe approach is to throw them into bankruptcy like everyone else. (Also there is always at least some risk that the good way to fail might require taxpayer support.) Anyway there has been much fulminating against Dodd-Frank's "orderly liquidation authority" but at the end of the day it looks like the Trump administration wants to keep it:
The Treasury Department, in a coming report on the Dodd-Frank provision known as orderly liquidation authority, plans to propose changes to the policy without scrapping it wholesale, these people said.
Amusingly, President Donald Trump sent a memo to Treasury Secretary Steven Mnuchin "symbolically placing use of the authority on hold except for emergency use." Why would you use it except in an emergency?
People are worried about stock buybacks.
The basic worry about stock buybacks is that U.S. public companies have a lot of money and are spending it on buying back their own shares rather than on raising worker salaries or increasing research and development. This worry has been sharpened by a U.S. tax plan that would raise taxes on middle-class individuals and sharply cut Medicare in order to pay for lower taxes on corporations. If the corporations pass their higher after-tax incomes on to workers, then this might be a wash. But:
Major companies including Cisco Systems Inc., Pfizer Inc. and Coca-Cola Co. say they’ll turn over most gains from proposed corporate tax cuts to their shareholders, undercutting President Donald Trump’s promise that his plan will create jobs and boost wages for the middle class.
The president has held fast to his pledge even as top executives’ comments have run counter to it for months. Instead of hiring more workers or raising their pay, many companies say they’ll first increase dividends or buy back their own shares.
For a while last year, I was half-joking that corporate stock buybacks might become a campaign issue in the presidential election. Maybe in the 2018 midterms they actually will!
Marvin Goodfriend Is Nominated to Be a Fed Governor. Matthew Klein: The DOJ’s case against the AT&T/Time Warner deal makes no sense. "In sum, §7 of the Clayton Act is a promising vehicle for combatting the anticompetitive effects of horizontal shareholding, should they be found." How a Small Bet on Tencent Made an African Firm One of the World’s Most Valuable. Banks Are Spending $20 Billion on Compliance Tech Ahead of MiFID. Credit Suisse vows to return 50% of profits to shareholders by 2019. Wells Fargo Is Dubbed a Repeat Offender and Faces New Wrath From Its Regulator. Companies Get Extra Incentive to Disclose Bribes: No Charges. Meditation Brings Calm to CEOs. New York City Has Genetically Distinct ‘Uptown’ and ‘Downtown’ Rats.
If you'd like to get Money Stuff in handy email form, right in your inbox, please subscribe at this link. Thanks!
To contact the editor responsible for this story:
James Greiff at firstname.lastname@example.org