This is the weekly free edition of The Diff, the newsletter that tracks inflections in finance and tech. Last week’s subscriber-only posts:
- A Golden Age for Individual Investors, on how cheaper trades and changes in market structure make now a fun time to be individual investor.
- Vacations as the Linchpin of Eurozone Economic Stability, on the macro and political risk of a very bad summer travel season.
- Guidance: Too Little, Or Too Much? Companies are pulling guidance, and analysts are annoyed. But earnings guidance doesn’t make much sense at a time when one form of uncertainty dominates outcomes.
- Mexico’s Covid Crisis - Eventually: Mexico seems like the kind of country that should be economically hard-hit by Covid-19, but two of their biggest exports won’t face problems just yet.
To read these posts, and get access to future paywalled posts, sign up today!
VC and Public Choice Theory: Logo-Hunting and Optionality
Agency conflicts involving mispriced options are the root of all evil in finance. Incompetent people have little job security, so while they can make mistakes, those mistakes are usually small—the worse you are at your job, the lower the dollar threshold for a “career-ending” mistake. To really lose a lot of money, you need to hire competent, hardworking people, and then give them terrible incentives.
In capital markets, this usually takes the form of paying someone an annual bonus and letting them take once-every-N-years risks. In any one year, the payoff from underpricing ten-year-storm insurance is high. But every so often, the storm arrives. To a trader who earns an annual bonus, the payoff function looks like this: high odds of a good bonus, low odds of getting fired. Since the worst their comp can do is zero, it’s mathematically optimal for them to risk many times their annual profits on low-probability events. With an annual bonus and the absence of risk controls, a trader’s compensation is a call option on the proceeds of writing put options.
In venture capital, the internal structure of the typical fund creates an interesting variant on this: associates have a strong incentive to chase hot deals, and they’re well-positioned to help the company that’s raising money pitch itself to senior partners. The incentive comes from the fact that “Associate” is not typically a partner-track job, but it’s also a job that’s very appealing to people who want to become partners. A common career plan is to join one fund on a non-partner-track job, and parlay that job performance into a partner-track role somewhere else. How, though? The output from an associate isn’t easily measurable from the outside—you can look at aggregate performance and ask each person how much they contributed, but it’s a fact of life in financial firms that half the employees think they generated 100% of the alpha, and they other half thinks they did. There’s one good way to claim credit, and that’s to work on specific deals.
But the right deal for a fund is not necessarily the right deal for the resume. An investment that’s produced a 10x return and no media coverage is worse for one’s career than a 2x return from a famous company. And that selection effect is particularly powerful for weird companies. Everyone says they want to be a contrarian. That’s incorrect. Everyone wants to have been a contrarian, but most people aren’t wired for actual contrarianism. Other investors are their peers, so a contrarian is doing something that makes them look dumb to the people they compare themselves to. And given the skewed return distribution in venture, the median outcome of a contrarian bet is that your friends think it’s stupid and the returns agree.
This gives venture funds a tilt towards momentum. Which is not, on its own, a bad thing: most successful exits raise a series of overpriced-looking rounds until the very end. Strictly in annualized markup terms, the biggest momentum-chasing trade of all time was Facebook’s Instagram acquisition: in 2011, it was worth $25m; in early April 2012 it was raising at $500m (“Still, some top venture-capital firms have decided to pass on it because of the app’s high valuation and immaturity of its business…”), and a few days later it exited for $1bn. (A 2x return in 4 days yields a 2.94e29% annualized return. That’s pretty good.)
An investor bias towards momentum probably explains some of the distribution of returns: once a company is hot, its cost of capital drops, so it can grow faster and conquer more of its market. Whether that’s net good or bad depends on what happens next: if the company raises at an amazing valuation, it tends to mean competitors get funded; if it raises an astounding amount, it sucks all the oxygen out of the room. Hotness affects recruiting, too, and recruiting affects returns—improving Glassdoor ratings correlate with high returns. But, like many other signals, momentum has worse dollar-weighted than time-weighted returns: being late to an investment is the first step to being too late.
In an upside-focused business, downside risks may seem not to matter—but young companies are fragile enough that a single problem can bring them down. Ambitious startups have a specific vision of the future, and their near-term task is to avoid all the potential wrong turns they could make on the way. It’s like the Douglas Adams line: “There is an art, it says, or rather, a knack to flying. The knack lies in learning how to throw yourself at the ground and miss.” So an investment process that’s designed to downplay risks still hurts, even when the interesting variance is upside.
Do associates have the power to influence investments? If you look at org charts and job descriptions, the answer is no: they don’t make the decision, they just gather data and try to get introductions. But they are in a position to influence which datapoints get emphasized and which get ignored, or to articulate the company’s narrative about why a particular risk doesn’t really matter. It’s an echo of a great Steven Sinofsky piece on enterprise sales:
Every enterprise sales person I have worked with begins to build out the physical and logical org chart from the first engagements. You want to learn the management reporting structure as well as the power. You want to understand the budget and decision making processes. Great enterprise sales people also know that you invest in the full org chart and don’t just focus on the areas of most authority and power—you never know where an advocate or obstacle might appear from as a deal progresses.
The goal of talking to an associate is to convince a decisionmaker, and the associate can shape that decisionmaker’s view. And the sales job for a hot round is a lot easier for the associate than an odd company would be. If the deal has been validated by somebody else, the investment decision just hinges on making sure other funds didn’t miss anything glaring.
And making the decision easier for partners—finding a conventional way to sound contrarian—has non-monetary benefits for them, too. It takes a certain kind of grit to tolerate the lumpy returns-on-effort from investing in growth companies. At every level of the funnel, a huge amount of time is wasted: most time spent sourcing deals will be spent on deals that turn out not to be worth the investment, so everyone in venture is constantly writing off the sunk cost of research. You can spend hours, days, weeks, months exploring a thesis, and then find the one reason it’s totally uninvestable. And within the portfolio, the companies that need the most attention are the ones doing badly, not the ones doing well; the cost of a bad investment is not just the dollars lost, but how time-consuming it is to fail. (To make things even worse, the classic case where time spent on a failing company is a good investment is when the company is doing badly because it has the wrong CEO. So a partner’s returns are best when they are, in some sense, betraying someone they befriended.) The magic of compound interest and power-law return distribution means that the better the investing career, the more of the dollar returns come from just one investment, and money has declining marginal utility. So in retrospect the median return on time spent actually keeps dropping.
To be clear, I don’t think that every sub-partner VC is a rogue trader, and even if they were, it wouldn’t be an especially harmful kind of rogue trading. Buying overpriced options means you slowly bleed profits; it’s selling underpriced options that can blow up in your face. Venture funds aren’t big enough or levered enough to pose systemic risk in the short term, though malinvestment from them is bad for overall productivity growth in the very long term. The actual upshot is practically a koan: raising a hot round is an even better deal than it looks, because misaligned incentives encourage investors to overpay.
 The paper finds that this effect is stronger for early-stage companies, and that makes sense. When a mature company grinds out slow improvements in margins despite weak revenue growth, you can see it happen in the Glassdoor reviews. They go from “We love working here!” to “They started charging us for snacks.”
 Since money has declining marginal utility over time, activities that were a good use of time when you were merely well-off become a retrospectively worse use of time when you were super-rich. To everyone who invested in Uber’s seed round, the single biggest career mistake they’ve made was the failure to send a text that said “Hey Travis, could I actually put in another $5k?”
 All useful investing advice takes the form of a koan, because anything obvious gets priced out.
I’m doing a deeper dive on Amazon and cloud economics. Diff readers who either a) have been responsible for large cloud budgets, or b) have worked at AWS, Azure, GCP, Alibaba Cloud, IBM Cloud, or other such companies, are encouraged to reach out. We can default to off-the-record.
I’ve always felt that typos give a newsletter a sort of cozy, bohemian feel. Like chipped glassware and exposed brick. Reader feedback indicates that this is not the case, so I’m getting an editor. Readers are invited to wave a fond goodbye to “mroe,” “indocators,”(!?), sentences that were supposed to include the word “not” but did… not… etc.
Negative Oil Futures: Almost Precedented
Institutional Investor has a blow-by-blow story of WTI futures’ negative trading. One line that stood out: “While much remains in question, one fact cannot be contested: In all of market history, this had never happened in a standardized, exchange-traded contract…” While that’s true, it’s not a consequence of changes in market structure. It’s actually because of tighter environmental regulations: in the 1955 onion corner, prices nearly went negative. Instead:
The mesh bag that you put 50 pounds a bag of those onions in, the bag alone empty costs 20 cents. A 50-pound bag of onions, they went to 10 cents a bag… As long as there had been futures trading in onions, no one had ever seen a price that low. Traders who were stuck with onions literally couldn’t give them away, and they tried… They called orphanages. They called hospitals, schools, whatever. They tried to get rid of as many onions as they could, and the rest of them, they dumped in the Chicago River.
(Thanks to this incident, onions are no longer “a standardized, exchange-traded contract”; onion futures have been banned.)
In more general oil news, the US is pulling missiles out of Saudi Arabia. One of the long-run aftereffects of fracking is that the US can afford to be much less concerned with geopolitics in the Middle East. (I wrote about this a bit here and here, and you should also read Peter Zeihan’s book-length treatment. While fracking reduces US dependency on the Middle East, it also puts Middle Eastern countries in a very difficult position; fewer US troops in the region does not imply less military conflict.
Uber: Meet the New Boss
Uber is one of the few companies that can view 2020 as a roughly typical year as far as existential risks go. Under Dara Khosrowshahi, they’re picking fewer fights—but adopting the frenetic acquire-divest habit he perfected at Expedia. In tandem with their earnings announcement yesterday, they confirmed their deal to invest in bike-sharing company Lime, and transfer Uber’s ride-sharing subsidiary to them. I highlighted this deal when it was rumored yesterday, since the timing made it seem like it was driven by Lime’s capital needs (their spending should be weighted earlier in the year, and the revenue they’d use to pay it back is not forthcoming). But it also showcases the Expedia playbook of adroitly shifting assets into partially-owned affiliates. Expedia spun off both TripAdvisor and Trivago, and was itself a spinoff of IAC; this sort of restructuring is part of the model. Given the economics of ride-sharing and food delivery, with some markets that burn cash and some that are, on a standalone basis, quite profitable, Uber has room to do this again.
Uber also reported that Eats gross revenue was up 54% Y/Y for the quarter and accelerated to +89% in April. (Grubhub’s gross sales were +9% in the quarter, and orders were +20% in April—but this includes the headwind of lower corporate spending.) One factor that matters for PR: Uber’s take, which rose to 11.3% from 10.1% last quarter.
In other ridesharing news, The Information has a great interview with a senior vice president at Didi. One notable tactic:
Many [drivers are paying] auto mortgages; they have to pay on a monthly basis. So that puts a lot of pressure on the drivers. So what we did is we negotiated [with finance companies] so that the driver doesn’t have to pay the lease in March and a big part of April.
Historically, the Chinese government has used state-owned companies and banks as economic shock absorbers—it’s interesting to see a tech company use auto lenders the same way.
Zoom Makes Progress
Zoom announced the acquisition of hyper-secure chat service Keybase, and plans to offer end-to-end encryption. Encryption does not appear to be something customers were clamoring for in the past, but it’s certainly something Zoom said they offered, which heightened the negative press they got when that turned out to be untrue. Part of Zoom’s approach recently has been to pay for respected security brand names (they hired Alex Stamos, who had recently called on his previous boss to resign as CEO after he quit).
Separately, Zoom got New York’s attorney general to close an investigation into the company. The pivot from product development to hypergrowth to firefighting has been impressive, and it will be very interesting to see what they launch when their 90-day feature freeze is up.
California Approaches a 2009-Level Deficit
California’s state government now expects a $54bn deficit through July 1, 2021, approaching the $60bn shortfall they faced in 2009. California’s budget is unusually levered to income taxes and capital gains, making it more pro-cyclical than other states’. And the capital gains piece means that California also suffers when high-income earners are reluctant to cash out: a rally in tech stocks doesn’t help if firms are still delaying IPOs.
Last month, I argued that several decades of deflation in tradable goods had made it hard for retailers to respond to a shortage by raising prices. A new study partially confirms this, showing that new sellers of scarce goods are far more willing to mark them up than existing sellers. New merchants were willing to charge 8-9x more than 2019 prices for masks and hand sanitizer, while Amazon’s own price went up by less than 100%. In a testament to econ 101, the Amazon data is spotty because their more affordable products were largely out of stock by mid-February.
This Time is Different
From last month: a very interesting essay on 2008 as epistemological training for Covid-19:
[One] source of uncertainty stems from the feedback mechanism inherent in the models. What, for example, is the realised impact of policy intervention on behaviour? It seems that in the first three weeks of lockdown in the UK, compliance has been higher than anticipated – schools had been expected to operate at 20 percent capacity, but have been running at 2 percent; the furlough scheme had been expected to cost £10bn, but recent estimates are that it will cost £40bn.
This feedback effect could lead to a seesawing of outcomes: high compliance leads to fewer deaths, fewer deaths lead to relaxed rules and lower compliance, this leads to higher deaths, and the process repeats. It won’t oscillate smoothly, though, since every additional wave makes the average person trust models less.