Welcome back to The Diff. Here are the subscribers-only posts you missed this week:
- A Painful Parable about Insurance and Incentives details the case of a fund that offered market-beating returns while hedging against the risk of a crash—and then promptly blew up during the crash. It’s a story about derivatives, backtests, speculation, and martingale bets.
- Inside the House Report on Big Tech and Competition: Part III: I wrapped up my series on the antitrust case against big tech companies with a look at Amazon and Apple, two companies whose core businesses are emphatically not monopolies, but that still have competitive advantages built in.
- Contrarian beliefs as a synthetic option applies the math of dynamic hedging to the pursuit of unconventional opportunities. The Black-Scholes model should inform how you think about niche obsessions.
- In How Fragile is China’s Financial System?, I review a report that measures how brittle China’s banking system is, in real-time. Some investors have been calling for a China collapse for a decade or more, and it’s interesting to evaluate why they either got it wrong or haven’t gotten it right just yet.
Subscribe now to read these pieces. At 2pm ET today, I’ll be joining subscribers on a call with former eBay CEO Devin Wenig to talk about competition and e-commerce.
This is the once-a-week free edition of The Diff, the newsletter about inflections in finance and technology. The free edition goes out to 15,708 subscribers, up 154 week-over-week.
In this issue:
- How the Pandemic is Fixing Some of the Real Estate Market’s Flaws
- Measuring the World’s Largest Economy
- Showrooming and Retail
- Google Continues to Disaggregate Pages
- Canada Policy Goes Post-Covid
- The Bus Factor
- Retro Trade Policy
- Payments and the Acquisition Option
How the Pandemic is Fixing Some of the Real Estate Market’s Flaws
There are two ways to look at any kind of investment: what it does and what it is. What a stock does, for example, is bounce around a lot in an unpredictable way, with a positive skew and a positive correlation to other risk assets. What it is, though, is a residual claim on a company’s assets: after paying off employees, suppliers, creditors, and the taxman (in roughly that order), the remaining money belongs to shareholders. These approaches aren’t contradictory, but they do get to an answer in different ways.
Housing is in an odd category as an investment. Unlike purely financial assets, it pays a return in consumption, not cash. And even as a financial asset, it’s a strange one. An American buying a home with a typical mortgage—fixed-rate with prepayment option—circa last year was getting:
- Continuous income in the form of shelter;
- An asset that rose in price with a) inflation, and b) conditions in the local labor market; and
- An option to reduce their interest burden and/or withdraw equity if interest rates dropped, without the obligation to pay more or invest more equity if rates rose.
This is a very strange package of securities. If you surveyed a cross-section of voters and asked them “Should Robinhood be allowed to sell highly-levered opaque interest rate derivatives?” a fair number of people would say no. But your bank will sell you those derivatives by default.
Historically, housing as an investment has a mixed track record.
- A mortgage gets paid down over time, so it’s a way to make people save money by default.
- Homes have appreciated over time. While their absolute returns aren’t great, the levered returns are fine, and a levered position that’s immune to margin calls is a good source of returns.
- For people who could get priced out of a neighborhood, owning their home is a good hedge. There was a time when teachers, police officers, factory workers, and other Richard Scarry characters could afford to live in places like San Francisco and Manhattan. As those cities got more dominated by superstar industries, it was less and less tenable for people who didn’t work in those industries to pay rent, much less own their own homes.
- A mortgage is leverage, and housing is consumption. Borrowing to fund consumption—even borrowing in order to buy future consumption in bulk—is generally not ideal.
- Homes have appreciated in the aggregate, but in specific cities and neighborhoods, they’ve performed very poorly as investments.
- Those neighborhoods are also where people work. The counterpoint to someone who can afford to stay in Park Slope because they bought their home decades ago is someone who can’t afford to leave Detroit for the same reason.
- The commission on a home sale transaction is generally 5-6%, and closing costs are another 2-4%. (Since these are costs associated with the transaction, the economic incidence is independent of legal incidence—they’re a cost of doing the deal, so they’re borne by buyer and seller. And since most current buyers are future sellers, it makes sense to assume an equal split on average.) Since the median down payment is 5-6%, this implies that most homebuyers are paying total transaction costs close to the amount of equity they start with. At the median home ownership duration of 13 years, someone who makes a round-trip transaction (buying one home, then selling it) pays a transaction cost that works out to around 0.65% of assets per year. Not terrible as far as asset management fees go, but not a steal, either.
High rates of home ownership are also a mixed bag socially. At one level, they’re the single most common way people experience capitalism—tapping financial markets to buy something, taking care of it, and then selling it at a higher price. And they make people more literally invested in local communities. But the more levered buyers are, the more widespread home ownership turns into widespread real estate speculation instead. And homeowners can vote in favor of their own interests, one of which is reducing the supply of homes to raise the value of their biggest asset. That option to refinance a mortgage has an important side effect on global interest rates: when rates drop, mortgage borrowers refinance, but mortgage lenders generally borrow at fixed interest rates. To stay hedged, they have to buy long-term debt in response—which pushes rates down further. All this affects the US treasury bond market, which is the closest thing the world has to a global benchmark for the price of money over long periods. So the entire global financial system is more volatile specifically so American homeowners can benefit from interest rate cuts.
All of these complexities are very much pre-Covid framings.
Today, home ownership is actually a much less economically dangerous choice, both for the marginal buyer and for the economy as a whole:
- The marginal homebuyer today is much more likely to be a remote worker leaving a high cost-of-living city to settle somewhere cheaper. They’re in a better position to pay a mortgage than the wave of new homeowners in the early 2000s.
- These workers' incomes don’t correlate with local housing markets. If big investment banks had laid off workers in 2019, it would have hit housing prices in Manhattan, Jersey City, and Greenwich. Today, it’s a smaller hit, spread out over low-tax states like Florida and Texas, places with lots of vacation homes like Hilton Head and the Poconos ($, WSJ).
- Some of these new homeowners plan to stay where they are for a while, even permanently, but others are already planning to return to big cities when it’s safe. The latter group is less likely to vote for restrictive zoning.
- A pre-Covid trend that accelerated soon after the pandemic hit is the rise of institutional investors who buy single-family homes. These investors make the housing market function more smoothly, because they’re able to bid for distressed assets when other investors don’t have access to capital. This makes housing prices overall less volatile. And the related rise of the iBuyers further dampens volatility (see my Opendoor writeup for more).
- Interest rates are exceptionally low right now. There’s no law of nature that says they can’t go lower, but as they get closer to zero, central banks try out less conventional forms of monetary policy. This makes refinancing smaller future risk, so mortgages don’t add extra volatility to long-term interest rates.
Every crisis causes some harm and obliterates some complacency. Usually, they’re harmful on balance, although there are happy exceptions (as it turns out, Sputnik was not a big national security risk, but the moon landing was a direct result of it). Covid is definitely on the net-harmful side of the ledger, barring some unlikely developments. But some Covid-induced shifts are unambiguously positive. Housing is still not a great investment, but a confluence of pandemic-related changes have indirectly conspired to raise it to the level of okay.
 There are exceptions to this. Chuying Agro-Pastoral Group paid a bond in ham, and there used to be an OTC-traded winery that paid a dividend partly in coupons for their wine. But in general, financial assets are cash-in, cash-out.
 Perhaps the most extreme example of this was in San Francisco, where a revolutionary terrorist group was able to afford San Francisco housing and bomb-making supplies in the 70s with a few of them working part-time jobs. This is not the most memorable detail in Days of Rage, but it’s certainly the one that best illustrates how much the world has changed since then.
 Here’s the mechanism: if you own a portfolio of mortgages, i.e. you’re the lender, you’re probably borrowing money to fund that portfolio. If you don’t want to take risk on interest rates, you match the duration of your borrowing to the duration of your portfolio. For example, if your mortgage portfolio’s average duration is 8 years, you’ll borrow with a similar duration. If rates decline, the duration of those mortgages goes down, because they get prepaid (which turns them into cash, with a duration of zero). But the duration of your debt goes up. So you buy duration to hedge, by purchasing treasury futures. Which means that whenever bonds go up, hedgers have to buy more bonds. Normally there are plenty of markets where hedgers have to adjust positions, and it nets out, because people on the other side of the trade are hedging in the opposite direction. In this specific market there’s a huge pool of borrowers who will never dynamically hedge, and a huge pool of lenders who have to, so the hedging trade is one-sided.
 Two that come to mind: a radical shift in public attitudes towards drug research, leading to more drugs developed faster and cheaper, which could potentially produce cures for unrelated diseases. And, a much more prosaic possibility: that for the next generation or so, people will mask up and stay home every flu season, and the R(t) of the seasonal flu will permanently decline.
A Word From Our Sponsors
Many of the Diff readers have investing or alternative data backgrounds, in some cases, both. One of the issues that people in the industry are familiar with is the usability of the data sets. Please join Sentieo, a fundamental research platform and a challenger to the legacy terminals, and alternative data provider Earnest Research as they go over several companies to demonstrate the combination of alternative data, NLP and ML applications. The webcast is on Thursday, October 22, at 2 pm ET. It is free and open to everyone: register here. (Can’t make it? Sign up, and they will email you the recording.)
I have a new piece on Palladium on the social status savings glut and the aging of American leaders. One point I didn’t realize until I read it: wars reset social status.
After the Revolutionary War, the first five presidents had served in the military in some capacity. Even John Adams, while not in uniform, had been chairman of the Continental Congress’ Board of War. Six of the seven presidents elected after the Civil War had served in the military, and every post-WWII president did the same until the election of Bill Clinton in 1992.
Piketty pointed out that war is one of the forces that resets economic inequality. It turns out to have that effect on other forms of inequality, too.
Measuring the World’s Largest Economy
China is, by one measure, now the largest economy in the world. China’s economy can purchase 16% more goods and services in China than America’s can in America. By dollar value of goods produced, not adjusted for cost of living, America is still well ahead.
Which metric matters?
Usually, measuring on a purchasing power parity basis is smarter, because it answers what is usually the right question to ask when comparing two countries: which country does the best job for its people? But increasingly, comparisons between the US and China are not comparing two abstract systems for allocating resources—capitalism with a safety net versus state-directed capitalism with minimal safety nets. Instead, it’s asking the harsher question of which country wins in a head-to-head struggle that’s denominated in money. In that case, the nominal numbers are more important. The US can afford to purchase more of the world’s goods and services, and can transfer more resources to favored groups than China can, even though one of the costs of that economic power is that domestically-produced goods and services are pricey for Americans.
Showrooming and Retail
Microsoft has agreed to pay Gamestop a percentage of digital sales derived from Xbox units purchased at their stores. This is a very interesting development: Amazon, Apple, and a handful of other companies can afford to tolerate “showrooming” at their retail stores, because they can do accurate sales attribution and measure the total purchase lift caused by those stores. E-commerce probably over-indexes on attribution, since checking prices online is so convenient, which means that some of the decline in physical retail is due to a market failure, not because the stores don’t serve a purpose in generating demand.
Back in 2011, in a previous incarnation of The Diff, sadly no longer online, I speculated about this:
Some bookstores are considering banning phones. Right now, that’s a good option. But Amazon could use their price checking app as a way to integrate the retail bookstore experience with Amazon’s technical backend. After all, they have enough information to know which bookstore gets credit for turning a customer on to a particular book—so why not explore the economics of paying that bookstore for the new customer?
For independent bookstores, the storefront is tough, but the real pain is in the back: ordering enough inventory, keeping boring titles in stock, dealing with the new releases you absolutely have to have in stock but that don’t necessarily sell all that well—the worst part of running an independent bookstore happens to be the part that Amazon excels at. If bookstores don’t need the capital for inventory, and can get a cut of Amazon’s rich [incremental - Ed.] margins, they can afford to maintain the bookstore experience while outsourcing the logistics to Amazon.
The future of the bookstore isn’t a “for lease” sign. It’s a combination of coffeeshop and showroom; the only physical thing customers will buy and take out of the store will be the Kindles on sale by the register.
Google Continues to Disaggregate Pages
I’ve referred before to Google dissolving the webpage by sending searchers directly to the part of the page with results they care about when they click. Today’s search announcements include another example of this, by displaying the relevant passages from pages in the search results themselves. This is part of a broader trend, where Google either a) heavily monetizes search clicks, or b) discourages the clicks from happening in the first place by surfacing results. The asymptote for Google is that every search result is either an answer or an ad.
(In other Google news, they now offer music search by humming, which is an amazing product—limited utility, but truly impressive, and it solves a problem that’s otherwise possible only through brute force.)
Canada Policy Goes Post-Covid
The Bank of Canada has discontinued some of its Covid-era interventions now that credit markets are calmer. Since 2008, central banks have unlearned some lessons from the inflationary 70s and learned new ones more appropriate to a deflationary world: it’s far safer to promise to potentially overreact, and then to do less, than to underpromise and see the situation escalate. Part of solving a financial crisis is repairing the underlying problems, but the best way to control the magnitude of the crisis is to credibly claim to backstop enough investors that nobody bothers with a bank run.
The Bus Factor
Applied Divinity Studies has a great piece on The Bus Factor, defined as “the number of people who would have to get hit by a bus for the project to fail.” Low bus factors are a form of operating leverage: a team that’s dependent on one person, or a handful of people, is also a team that’s disproportionately made up of high-impact people. It will get a lot more done for a given number of worker-days (as long as the bus doesn’t come). The Bus Factor is a good way to understand why companies seem to ship a lot less per employee as they grow. Once the core product works, the most important thing is to reduce the risk of it breaking, and that means layering on redundancy.
Retro Trade Policy
ProPublica has a profile of Robert Lighthizer, whose career follows a pattern I’ve seen before: extremely early success, a long fallow period, and then a late-career peak in influence. Lighthizer became Congress’s youngest staff director in 1978, but had to wait another four decades to make his biggest impact. (Two near-contemporaries with similar arcs: Donald Rumsfeld was the youngest Secretary of Defense in 1973, and one of the oldest ever under Bush; Cheney was White House Chief of Staff at age 34, and Vice President decades later.) What’s interesting about this arc is that it lets people solidify their views when they’re young, but enact what have become old-fashioned opinions much later. Lighthizer was trying to use trade policy to prevent deindustrialization in the 80s, when it was conventional and bipartisan; when the bipartisan consensus switched to free trade in the 90s, he was left behind. And now he’s back.
Payments and the Acquisition Option
Every platform business has an ongoing tradeoff between network effects (which are good for business) and surface area (which presents risk). As a payment network grows, for example, growth gets easier—but fraud gets bigger, and the PR risk of working with the wrong customers rises, too. Since the converse to easy growth with an established network effect is difficult growth early on, there can be many attempts, most of which fail. For the ones that succeed, there are a handful of options: be the top regional player (but always worry about competing with the global market leader), use home market dominance to become the global leader, or sell.
Paystack, which hoped to be the Stripe of Africa, will be literally that. They chose to sell.
The economics of this are repeatable in any business where building a viable network effect is expensive and requires some situational knowledge. Creating a payments network for American e-commerce businesses is very different from creating one in Nigeria, but a single payments provider that operates in both markets is more valuable than either of the two operated independently. This makes payments providers that operate in the developing world more fundable, because they have an exit opportunity: once they’ve outbuilt the biggest global companies in their home market, they’re worth the most as part of one of those companies. Selling a regionally successful company to a bigger competitor is the cheapest way for the seller to “buy” the buyer’s network effects.