In this issue:
Bubbles and Easy Money
One theory of bubbles is that they are always and everywhere caused by an increase in the money supply. Defined broadly enough, this is a tautology: since a "bubble" is a temporary increase in the ratio of certain asset prices to a set amount of money, they require money to come from somewhere. And the inverse is certainly true; if the federal funds rate were raised to 10% overnight, asset prices would collapse, and the bubbliest assets would fall fastest. ("Why keep your money in constantly inflating Bitcoin, when deflationary fiat preserves your purchasing power?")
It is true that many recent bubbles have coincided with greater capital availability:
Housing price growth took off in the 2000s because the Fed cut rates to contain the aftermath of the dot-com bubble.
That bubble went through its manic phase in 1999, following the emergency rate cuts of the 1998 crisis. (That said, the Federal Funds rate started rising again in mid-1999, and by the market peak in March 2000, were at their highest level in five years.)
The monetary theory certainly gives people something to complain about. Since real rates have been declining for the last eight centuries, anyone who's lived a long time or read enough history will, on average, see lower interest rates than they're used to for most of their life. This is particularly true of anyone who started paying attention to the news in the 70s and 80s, and anchored to double-digit interest rates as a fact of life.
You can divide the monetary side of the bubble debate into three basic stories:
An increase in money and an unchanged quantity of investable assets implies, mathematically, that more money is required to buy the world's asset base.
The counterargument to this is that bubbles are endogenous to specific industries, that they're about the story and not the money. A money supply-driven bubble dynamic should affect everything, or every asset that's hard to make more of. Cannabis is not especially hard to grow, but there was a speculative mania for cannabis producers in 2017 and 2018. The dot-com boom was built on excitement about the Internet, not just the need to own stocks; it left plenty of dowdier companies behind.
If markets are efficient, there's no such thing as a bubble, just a judicious estimate of future payoffs that happens to be wrong some of the time.
None of these are entirely satisfactory, but there's an interesting meta theory: that bubbles are not caused by a change in monetary aggregates, but by a change in which money can (or must) access which opportunities. Take the recent Gamestop short squeeze. There are people who blame central banks, but it doesn't make much sense:
But there is a monetary story behind some bubbles. It's not a story of quantity, but of quality: a common signature of bubbles is that one form of money gets changed into another form that magnifies market impact.
The 1920s equity bubble was fueled by margin debt, but margin borrowing had existed long before then. In Reminiscences of a Stock Market Operator, the fictionalized Jesse Livermore talks about trading with 100:1 leverage in the early 1900s, so the 10:1 leverage some speculators used in the 20s doesn't sound that extraordinary. But what was new in the 20s was the source of those funds. In the late stages of the bull market, while banks were pulling back from margin lending, a new source of capital stepped in: large corporations with cash on hand.
This created a strange cycle: a company could invest in margin loans earning 10%. Its stock might trade at 20x earnings. So every dollar of cash the company kept on hand and didn't spend would produce $2 of market value. From the end of 1927 to the start of the crash in 1929, banks reduced their margin loan exposure from $2.6bn to $1.8bn. Non-banks, mostly corporations and wealthy families, increased theirs from $1.8bn to $6.6bn.
For any given lender, this was perfectly rational; the market was liquid, and rising, so the collateral was easy to seize and liquidate. Collectively, of course, it had the obvious problem that if everyone's getting margin calls, the margin calls will continue. (It had a secondary problem: it meant that companies were recycling their cash profits into the financial markets rather than into capital spending or wages—this was an early case of inflation in asset prices coupled with deflation everywhere else. Recycling profits into call money was a maximally bubbly behavior, since it increased stock prices while ultimately decreasing non-financial earnings.)
I've written before about the petrodollar recycling dynamic in the housing crash. Normally, high oil prices lead to slower economic growth, but the period from 2003-6 was a paradoxical one, where higher oil profits mostly accrued to companies that saved money in dollars, and the assets they chose to invest in were mortgage-backed securities, which kept housing prices elevated, allowing mortgage equity withdrawals to offset the implicit tax of higher gasoline prices.
The bubbliest-looking aspect of the market today is the performance of Gamestop. It's a small company, but there's definitely manic behavior on the part of investors (for more, see yesterday's issue). And Gamestop—along with other small-cap, heavily-shorted companies with brand names recognizable to young people—is the beneficiary of a qualitative change in money. Fiscal stimulus is usually targeted at people with a high propensity to spend, but that propensity has undergone major changes recently. In addition to the obvious ones (no restaurants and bars, much less travel), there are other large expenditures that are getting delayed: some people have moved back in with their parents, for example, cutting their rent to zero. Car sales have rebounded somewhat, although they remain below 2019 levels, because cars are more discretionary when more jobs can be done from home.
The usual trick with stimulus money is to balance between two forces: if someone is desperate to pay the bills, they'll spend their stimulus check right away, which gets the economy closer to baseline. On the other hand, most people won't adjust their long-term spending behaviors based on a one-time payment, so the default behavior for a windfall would be to save it. (Or, equivalently, in the best empirical study I could find, to pay down debt.)
If Robinhood, WeBull, and the like have made it easier for money to flow from savings accounts directly into equities, and WallStreetBets has ensured that those funds flow into the same small set of equities, this would explain why a handful of stocks have gotten extreme attention, while the rest of the market, while high, isn't experiencing the same kinds of gyrations. It's not a story about monetary aggregates at all; it's a story about a meaningful drop in the cost of living for people who use Robinhood and browse Reddit.
 The best analysis of this is "The Petrodollar Illusion," from Clarium Capital in 2006, which is sadly not available online.
For a while in March and April, this newsletter only covered Covid, because that was the only thing happening in the world at the time. Things quickly returned to normal. Thus with Gamestop.
Yesterday, Gamestop opened 50% above Friday's close, reached a price 50% higher than that, briefly went negative, and closed up a sedate 18%. (This morning, it’s up another 16%.) Occasionally, when one fund is known to have a money-losing trade, some of its other trades start performing poorly, either a) because the fund itself is reducing exposure, or b) because everyone knows what the fund owns and is betting that they'll have to close. It sounds unpleasant. Jim Cramer devotes an unusually manic-depressive chapter of his memoir to the same experience, and it sounds painful.
This happened to Melvin Capital yesterday. The WSJ reported on Friday that they were down 15% for the year ($) due to short bets, including Gamestop. As Gamestop rose this morning, Melvin's largest positions—Limited Brands, Expedia, Advance Auto Parts—all fell. And after Gamestop peaked, they mostly came back.
And Melvin itself got a $2.75bn rescue package from two larger funds ($, WSJ). The economics of this deal make sense: once a levered fund starts losing money because it's recently lost money, it can go straight zero. This is a familiar part of the story of Long-Term Capital Management, and also played a role in the collapse of Amaranth (which executed a more distressed version of the same deal in the end). Melvin's long-term track record is excellent, and Robinhood-risk is survivable given a big enough bankroll, so this ends up being retroactive black swan insurance. Expensive, but worth it as a way to stay alive.
Rudi Dutschke, a German Marxist borrowing some ideas from Antonio Gramschi, coined the term "Long march through the institutions" to describe his plans for a revolution: getfellow Marxists to join professions, to work well, rise in the ranks, and ultimately to stay loyal to the revolution. Bitcoin, as I've argued elsewhere, follows the same path: the owners get gradually more institutional over time, from anarchist hackers to esoteric private funds to major companies. Another group that's joined that roster: college endowment funds, including Harvard and Yale. College endowment funds are unique in that they may have the longest outlooks of any institution in the world: Harvard, Yale, and other elite US schools predate the founding of United States; Cambridge and Oxford were founded well before the British state reached its modern form; the Sorbonne has outlasted a monarchy, two empires, and at least four republics so far. Universities can tolerate losses in the short term, and can pay attention to extremely distant outcomes. For them, buying Bitcoin doesn't make sense as a short-term flyer, but it does make sense as a way to shift their reserve asset mix.
2 & 20 Media
Andreessen Horowitz arguably came into existence as a blog before it formalized its venture operation (A16Z was founded in 2009, while the Pmarchive dates back to 2007, although there were plenty of angel investments before the fund was launched). Now, they're scaling up their media operation. The tradeoff in tech journalism is that cynicism always gives writers something to write about—no company perfectly lives up to its public perception, every round looks overpriced, every product looks silly at first. But the returns from early-stage investing are distributed on a power law, and opinions vary, so cynicism about any given company is the consensus view. It's a bias analogous to shorting a basket of overpriced stocks; you can be right on average, but when you're wrong, you're really wrong. When VCs firms write about tech companies, they avoid that bias, although they naturally bring along the bias that they're talking their book. What the firm is trying to do now is retain the correct bias, without the conflict, by making the content operation more independent of the core business. The monetization model remains similar to what it was in the heyday of blog.pmarca.com: using high-signal content to attract the best deal flow.
In November, there were rumors that Twitter would acquire Substack, which were quickly denied by Substack. At the time, I wrote that this made sense, because Twitter was a free marketing channel for paid newsletters; Twitter wasn't capturing much of the value it was creating. They're taking steps to do so now, by acquiring newsletter platform Revue. Substack has already leaned in to Twitter as a discovery platform. Since the Twitter social graph is denser than the Substack social graph (most people have more Twitter followers than newsletter subscriptions), it's easier for Substack to find people who ought to be reading one another's long-form content by seeing who reads which short-form content. It's unclear if a captive newsletter platform will outweigh Substack's early lead. In the early 2000s, when eBay noticed that Paypal was profiting so much from their auctions, eBay launched its own payments solution. But distribution couldn't beat a faster pace of launches, so they ended up buying Paypal after all.
Airline Loyalty Programs as a DTC Brand
I'm perennially fascinated by the fact that running a hub-and-spoke airline in the US is a loss-leader designed to subsidize acquisition of credit card spend. That model was starting to work very nicely in the mid-2010s, since the airlines were well-capitalized and had preferential access to a high-spending demographic. Obviously, a 2/3 drop in flights, and a shift at the high end from first-class commercial to private planes, weakened that model. One adaptation: selling airline meals, or even starting a wine club, and tying it to loyalty points. For any loyalty-based business, the key dependency is to maximize the number of times each member interacts with the system. That produces more data, and more opportunities to auction off access to members. A wine club is far from American Airlines' core competency, but keeping AAdvantage members earning and burning points is at the center of their model.
Palantir, From The Inside
A former Palantir employee, who left in 2015, offers the long-term bull case: that Palantir is to large enterprises what Salesforce is to everyone else. The post is light on financials, but makes one very important point:
IMHO, the one thing that has not received enough attention in the IPO is Palantir’s ability to track the source and pedigree of every single piece of data in the system, empowering granular access control on a need-to-know basis, which enables organizations to share information across borders/boundaries.
Anyone who has tried to build this kind of control retroactively knows that it's incredibly painful. By design, this breaks all sorts of dependencies in a complicated data pipeline; often, the most convenient way to group user data, for example, is a way that implicitly deanonymizes users. As a rule, any new access control scheme that doesn't break large amounts of legacy code isn't preserving privacy. Since Palantir started with clients who are sensitive to data access and provenance, they had to spend early to build this, but now it's baked in. As companies get more sensitive to privacy—as "data is the new oil" makes more of them think of Deepwater Horizons rather than Spindletop—imposing easily-audited access control will be increasingly important.
Paul Graham once described Stripe's go-to-market strategy:
More diffident founders ask "Will you try our beta?" and if the answer is yes, they say "Great, we'll send you a link." But the Collison brothers weren't going to wait. When anyone agreed to try Stripe they'd say "Right then, give me your laptop" and set them up on the spot.
Perhaps Palantir's sales team will do something similar, by taking a look at a prospective client's current data infrastructure and showing a five-line SQL query that, if used, could generate tens of millions of dollars in GDPR fines.