Taking Strategy Seriously
Plus! Errata; Streaming as Instant Gratification; Batteries; Shrinking Analyst Teams; Diffident Voters; Oil, Recapitalized
Welcome to the Friday free edition of The Diff! This newsletter goes out to 23,311 subscribers, up 104 since last week. In this issue:
Indirect-to-Consumer: The E-commerce Rollup Business: there's been a profusion of companies that raise money to buy out merchants on Amazon and other platforms. Why does this model exist, and what are the risks?
Banking Wants to be Free, But Free Banking Wants to be Dangerous: (As seen in Money Stuff!) An important paper on stablecoins raises some interesting historical analogies—but takes them too far by assuming that the crypto ecosystem is tightly integrated with the rest of the financial world.
The Controlled Demolition of China Evergrande: China Evergrande is one of the most indebted companies in the world, and it looks like it's rapidly collapsing. But to understand what happens next, and whether this is the Lehman Moment for China's real estate market, it's important to look at what the company's real purpose is, and who wins and loses if it goes under.
Thanks for reading The Diff. For full access, and to join Q&A calls, subscribe today.
In this issue:
Taking Strategy Seriously
Streaming as Instant Gratification
Shrinking Analyst Teams
Taking Strategy Seriously
One of the guilty pleasures of equity investing is listening to the occasional brutal earnings call. An earnings call is usually a time to congratulate management on a great quarter ($, WSJ), hear updates on the company's strategy and explanations for nuances in quarterly results, and ask a few questions that help with financial modeling or reveal industry trends.
But sometimes it's an opportunity to call management out for disappointing performance. A few years ago, United Airlines had a disappointing quarter, with poor guidance; the company took credit for some new initiatives, but their impact was just enough to get the business to worse-than-expected. So they got questions like:
Coming and saying "we're on track for these initiatives but don't mind these numbers" doesn't feel right. It feels like you need to like reset expectations, whether it'd be not giving specific numbers for individual initiatives or somehow saying, hey, resetting a bar here for these initiatives because I don't think those numbers mean very much to anyone anymore.
[H]ow can we have any confidence in the 2018 [expense] story, particularly given the headwind that you guys went out of your way to lay out on this call?
And sometimes it was more of a comment than a question:
These numbers are not terribly useful. So what I'm looking for is kind of more useful numbers.
Since the key deliverable for a quarterly conference call is past numbers that are pleasant and future numbers that are both good and believable, this is roughly the equivalent of sending an entree back to the kitchen or walking out of a movie early, except that several billion dollars of market capitalization disappears in the process.
Normally, when a company reports a bad quarter, it's a chastening experience. They reconsider big plans and focus their attention on the basics. If they're doing so much that management's attention is divided, they need to scale back. UAL management made some ambiguous comments on the call about considering the marginal impact of changing their schedule and adding or pulling back capacity, but the general assumption was that they had learned an important lesson about biting off more than they could chew and would be more cautious in the future.
Their next earnings call included an in-person conference, and opened with a little comedy bit from the new head of investor relations:
Before we begin today's presentation, as safety is our highest priority, we like to provide a safety briefing. In the event of an emergency, please exit out the stairwell directly behind you and down the stairs... [much more in this vein] In an event of an active shooter, be prepared to run, hide, or if you're [United then-President and now CEO] Scott Kirby, fight.
There are two broadly useful stylized facts about airlines and scale, leading to a third corollary. They are:
Any given airline will get lower unit costs if it's bigger, since fixed costs are spread out over more activity. This is true at a micro level; with a 500-plane fleet, taking a plane out of service for repairs is easy, whereas with a five-plane fleet it requires the company to scramble. It's also true at a macro level; a bigger airline gets a better deal on new planes, can afford more brand marketing, and can build a more comprehensive network.
In the aggregate, when GDP grows faster than capacity, airlines have pricing power and the industry does well. When capacity grows faster than GDP, everyone is fighting for market share, fare wars break out, and profits suffer.
The corollary to this is that anyone who wants the industry to make a lot of money should complain when an airline decides to add more capacity, and celebrate when everyone grows slowly.
In the four years through 2017, UAL grew capacity at about 2.4% annualized, while nominal GDP grew 3.9% over the same period. They were playing along. At that presentation in January of 2018, they announced that they were doing something quite different, growing capacity 4-6% for the next few years. In some industries, this would not be a big deal; one company tries to expand, maybe others shrink a bit. But in airlines, it's significant, since it means the entire industry can tip from being slightly undersupplied (and thus profitable) to slightly oversupplied (and a low-margin, high-variance fight to the death). The list of US airlines that have gone bankrupt is long and inglorious, and includes Delta, United, and American (Southwest has avoided this, both through low costs and by hedging fuel aggressively in 2008, when oil rocketed1). This list, and the clusters of airlines that went under in the same difficult years, was the first place investors' minds went when United laid out its plans.
Over the next two days, United's stock dropped 15%, and airlines as a whole dropped 6%.
After that, something interesting happened: the strategy worked. In their presentation, United laid out a different model from the macro/micro analysis above. The key argument:
A hub and spoke airline is really a manufacturing company, and it is about manufacturing connections. The more connections you can drive at a hub, the higher profits you drive at that hub, the more options you have for customers to flow through that hub. And it's exponential. You add 1 flight into a hub that has 80 connections, you don't just add one market like a point-to-point carrier would be doing. You add 80 new markets. And that strengthens the whole network, and it makes the other 80 flights stronger at that one hub.
This is a crucial point. Airlines can be a commodity business, but they also have network effects, both at the hub level and through loyalty programs. Under the right circumstances, and with the right strategy, things flip: an airline can buy commoditized inputs like fuel, and long-lived assets like planes, and plug them into a model where they get above-average returns. For United, that meant making its hubs stronger, and taking back routes it had previously ceded to competitors; if two airlines are operating the same route, the one with more connections can get better economics, so it's suboptimal for airlines to cede share in markets where those economics apply. And, once the airline is sending more traffic to a given hub, that hub is both more profitable and harder to compete with; it's a competitive moat that expands with high margins, instead of the usual pattern where high margins attract competition.
This strategy worked quite well. From the day of that earnings call through mid-February 2020 (i.e. before Covid started affecting stock prices), UAL returned 16% while an airline ETF was down 6%. They took their model seriously, and found a way to drive better returns in a business that looked like it was squeezed between challenging competitive dynamics on the revenue side and volatile fuel prices and increasing labor prices on the cost side.
It's useful to look at their current strategies as an extension of this same kind of thinking. United has, notably, signed up for both Boom Supersonic's planes and electric short-haul aircraft from both Archer and Heart. One reason for the last two is that United is trying to cut emissions—they're targeting net zero emissions without carbon offsets by 2050, and while they're turn their fleet over a few times by then, it's a good idea for them to start pushing the cost curve down now so planes will be affordable later.
Normally, an equipment upgrade is a substitute for existing equipment. A plane gets old, and it gets replaced by something nicer and more fuel-efficient. But supersonic planes and four-passenger vertical-takeoff jets are not really substitutes for anything an airline currently flies. They're complements. A supersonic jet changes the cost/benefit tradeoff for different kinds of meetings; if you're in New York and someone in LA suggests a Zoom meeting later that day, it will be theoretically possible to counter with "Why don't I stop by your office instead?" (Will it be murderously expensive to schedule a last-minute trip on an already-expensive aircraft? Yes. Will that expense have signaling value over and above the usual benefits to in-person communication? Yes!)
The smaller electric aircraft may work for some city-to-city trips, but they're mostly a way to expand the effective area served by a specific airport. They're micro-spokes on the existing hubs. For an airline that already relies on hubs, this is disproportionately beneficial, since it effectively makes each hub bigger, and can also make any spoke more important, thus sending further traffic to the hubs.
What's especially interesting about this transition is that it's really an extension of United's earlier expansion strategy. Once you're convinced that adding capacity that strengthens hubs is strictly better than adding capacity anywhere else, and can generate above-average returns on capital, you have a good business plan—but it's a limited one, because at some point those hubs get saturated, and the opportunities for incremental growth diminish. At their recent investor day in June, United talked about flying larger planes to feed traffic to their hubs (this is both more cost-efficient if the hubs are getting traffic and a better flyer experience). And they're once again planning to grow capacity 4-6% for years, a number that used to be terrifying to investors but that now sounds achievable.
United is a case study in taking strategy seriously, looking at a tough industry and finding incremental ways to deploy capital at above-average returns. There’s a lot of energy devoted to figuring out these dynamics at an industry level, and identifying companies that can reliably employ capital at high returns. Aggregates can hide interesting variances within industries, and even within companies. It’s common to observe that excess returns and poor returns don’t always persist over long periods; there is mean reversion at the company and industry level. The United story is a case of how mean reversion actually happens: buried inside a challenged company in a tough industry, there was an opportunity for above-average growth.
Thanks to Jack Robling for suggesting this topic and sending some helpful observations and links.
In last week's issue, I mentioned some impressive VC performance stats from an FT article. Other recent sources have more modest VC performance numbers, much closer to equities, and the number the FT itself cites doesn't seem to appear in the papers of the researcher they mention. Apologies for the error, and thanks to Søren Fryland of IPQ Capital for the heads-up
Streaming as Instant Gratification
The Atlantic has a piece arguing that streaming a new release like Black Widow ends up weakening Disney's overall economics based on a press release from a theater lobbying group, but the math is incomplete. It is true that the streaming debut hurt box office sales—the theater industry group estimates that the movie would have done $92m to $160m in its first weekend at the box office, instead of $80 million, but this is offset by $60 million in streaming revenue. They correctly note that Disney has to share 15% of its streaming revenue with platforms (if subscribers sign up on them) but somehow omits the fact that theaters themselves take a larger cut.
One way to look at this analysis is that it's false in this instance but still useful. Disney understands customer lifetime value better than any other media company (in this piece ($) I look at Disney's M&A strategy in terms of filling in age gaps in its products). But other media companies might be tempted to copy Disney without copying the full model that makes it work. Disney benefits from streaming releases in two ways:
They get even more leverage when negotiating with theaters. (Not that they need it at this point), and
User acquisition and engagement increase their streaming user count and reduce their churn rate, and the gross profit from streaming is a lower bound on its benefits to Disney, since they can also use streaming to distribute future products and hype theme parks.
If other media companies can't do this, they should shy away from trying to copy Disney's model, since they don't have a complete copy of Disney's economics to back it up.
The marginal cost of deploying new solar and wind power is quite competitive with fossil fuels, if measured over the course of a day. Unfortunately, wind and the sun are not available on a steady 24/7 schedule in any specific location, so the available options are some combination of a) to continue to use other energy sources, b) wholesale social engineering designed to line up power consumption with peak solar and wind production, or c) batteries. Solar plus batteries has not been competitive with coal at providing 24-hour power, but Form, a battery startup, believes it can cut energy storage costs by 60-75%, making renewables competitive with fossil fuels ($, WSJ). It's unclear if this works; if it does, it's a huge development that solves an important problem and puts energy back on track to decline in cost over time. (That is a bigger deal than it sounds like, since accessing new and cheap sources of energy coincides with step-function changes in standards of living and even ways of living: domesticating animals created one new pattern of settling down, and cars and planes exert a lot of force on the size and structure of cities.)
Shrinking Analyst Teams
The number of full-time sell-side analysts, and the number of ratings on public companies, continues to decline. There are many factors behind this: when research and trading were a bundle, it was easier to overcharge (or to have trading subsidize research); as hedge funds have proliferated they've hired more sell-side analysts and made prices more efficient; and regulation has hurt some investor-unfriendly but company-friendly ways that analysts could produce revenue for their employers. I wrote last week about the fact that newsletters can move markets when they publish about individual stocks, and that's a sign that there is demand for original research on companies. The existing business model might be too wedded to a payment system that no longer applies, but there will always be demand for people who can find mispricings but don't have the capital to take full advantage of them.
A $24bn SPAC deal to take Lucid Motors public has failed because shareholders didn't bother to vote ($, FT). (Edit: they were later able to get the votes and the merger went through.) This is still a mostly theoretical risk with SPACs, since they can keep soliciting votes and the business they’re buying probably won't run out of cash while awaiting the merger. It illustrates a wide divergence between what SPAC traders are looking for and what the structure is meant for; if shareholders are day-trading, they may not care about corporate actions that much, and may get notified to vote after they've already sold the stock. Meanwhile, Robinhood doesn't have a strong interest in making it easy for people to vote their stock; while this is a nice feature for users, it's far down the list. The problem is actually fairly similar to the one the US encounters during midterm elections. Yes, the outcome matters to people and they care about it. No, they don't care that much, especially when the result is (usually) not in doubt and the signaling value of casting a vote is fairly weak.
As energy majors sell their most polluting assets to private companies, it means the average cyclical downturn will kill more companies, and fewer of them will be able to financially recover from accidents, or pay to clean them up ($, FT). This will have an interesting second-order effect on regulation: if there is a big oil spill caused by a small company that bought its assets from a larger one, there will be calls to restrict big oil companies from selling assets to smaller ones. (The alternative, which is economically equivalent, is to require companies to buy more comprehensive insurance, but if there are equivalent policies, one wonkish and one punitive, the punitive-sounding one will probably win.) This could leave major energy companies in the awkward position of being unable to dispose of assets in line with their long-term emissions plans.
It's a testament to oil price volatility that an article called "Southwest's fuel gamble: Hedges keeps fares in check" appeared in September 2008, and one with the headline "Southwest’s hedge on fuel backfires" came out two weeks later.