The Coming Carbon Short-Squeeze
Plus! The Streaming Race; Captive Banks; Construction Credit; Illiquidity Doesn't Always Pay; Instacart Shifts Further Towards Ads
Welcome to the Friday edition of The Diff! This newsletter goes out to 25,055 subscribers, up 127 from last week.In this issue:
The Coming Carbon Short-Squeeze
The Streaming Race
Illiquidity Doesn't Always Pay
Instacart Shifts Further Towards Ads
The Coming Carbon Short-Squeeze
Of the world's twenty largest carbon emitters, 19 have an explicit target for when their net emissions will reach zero. For most countries, that will require some combination of subsidizing lower-emissions or negative-emissions activities, taxing higher-emissions activities, and outright banning some things (clearing rainforests, flaring natural gas). And, right now, it's hard to see how the world hits its collective targets.
You can think of three broad models for how we take emissions from N to 0:
The classic new technology rollout cadence, where things start small and compound. Solar used to only make sense for satellites, then as an occasional high-priced gimmick, then only with subsidies. Now it's cost-competitive with coal in the aggregate, which makes it a better choice for power uses where intermittent access is fine and cheap energy matters above all (e.g. making aluminum, mining bitcoin, desalination, charging EVs whose batteries can last all day) but not quite good enough for the full grid (coal gets a lot more competitive once the sun goes down). In that model, we might see net emissions per dollar of GDP drop 0.1% one year, 0.2% the next, 0.4% the year after that, etc., such that at first it looks like zero progress is being made, until suddenly emissions are plummeting at an accelerating pace. This looks nice on a graph, but the late stages of this deployment involve some combination of massive improvements in both solar efficiency and batteries (especially batteries) coupled with trillions of dollars of capital expenditures, and probably radical improvements in carbon capture and sequestration, too.
The more pessimistic path is that taxes and subsidies let us pluck a lot of the low-hanging fruit, so the global economy experiences a series of step-function improvements in carbon efficiency, followed by a lull. A global $30/ton carbon tax would instantly render some coal power plants uneconomical, and would make old internal combustion cars more expensive than new EVs and more fuel-efficient models. It would mean that re-insulating older houses would pay for itself, that energy conservation would be unavoidable, that coal-inefficient steel mills would run at a loss even with cheap labor and subsidized capital, etc. So the most emissions-heavy swathe of the global economy would promptly shut down, and newly price-competitive options would replace it. But then we'd run out of low-hanging fruit and have to make harder tradeoffs: replacing clunkers with fuel-efficient cars is one thing, but are owners of the world's ~1.3 billion cars all going to wait in line for a new all-electric vehicle? And will we have enough lithium to make them? And what about fertilizer? Are we going to spin up natural gas alternatives fast enough, or just make do with either less food or a lot more agricultural labor? There's actually a precedent for these kinds of questions, driven by cost rather than climate: the US economy was fairly freewheeling with energy from the 20s through the early 70s, when oil was cheap and (in real terms) getting cheaper. When the oil crisis hit, some of the waste went away. But it was difficult to keep cutting.
Average the cadence of a favorable logistic growth curve for emissions reduction from new technologies and an unfavorable one from eliminating the least efficient old ones., and you get a linear path: some things start out easy—shutting down a coal plant earlier, or deciding not to decommission a low-emission power source. Some parts get easier over time (the billionth car battery will be a lot cheaper and more efficient than the millionth was). This looks like a sloppy method, with assumptions that will irritate both the Panglossians and the more climate apocalypse-minded. But that's fine: whatever actually happens with climate change will irritate both sides, and will probably include some technology-driven solutions with their smooth exponential curves, and some social and institutional changes with their frustratingly human-driven gradualism, wild swings back and forth, measurement issues, conflicts of interest, etc.
And, taking a simple linear model and applying it to current emissions, we need to eliminate around 1.5 billion metric tons of emissions globally each year, which amounts to some mix between the avoided emissions of decommissioning 490 coal power plants every year, or the sequestered carbon from planting 7.6 billion trees.
What pushes people to do this now? One underrated but important driver of human behavior is accounting. Accounting makes the difference between something being a theoretical concern and being a target that can be optimized for and traded off against different goals. One of the reasons companies seem to display more agency than other organizations is that they are required to keep accurate books, and structured to maximize particular variables. A nonprofit or a government agency can do many things that feel like they're making a difference, but nobody in the mining business ever got rich by "raising awareness" of copper instead of digging it up and selling it for a return that meets or exceeds the cost of capital.1 Accountants have played a significant role in determining which side of the worth-it line companies fall on for a long time.
Specifically, the interesting parts of accounting are not just keeping track of when cash flows come in and out, but of how to account for intangibles. Using depreciation instead of treating capital expenditures as an operating expense or just putting everything on the books at cost will affect what companies buy and how they use it. Writing down impaired assets helps firms avoid the sunk cost fallacy (or helps discourage them from tricking their investors into ignoring sunk costs). Taking reserves based on likely future actions like refunds and credit losses is a way for a company to see how profitable a given transaction will be, rather than the most profitable it can look.
And, soon enough, accounting standards used in 140 jurisdictions will incorporate standardized metrics for measuring emissions. Accounting creates the kind of legibility that governments love and rely on; rules about taxes, imports and exports, IP usage, and treatment of employees are infeasible to enforce without consistent standards, and rules about emissions fall into the same category. That's particularly relevant because so many emissions are in parts of the supply chain the more regulated companies can't touch ($, WSJ). Right now, an iron miner selling to China can quite honestly say that while they'd prefer that Chinese steel mills didn't use energy-inefficient processes, their hands are tied. But if they're required to report Scope 3 emissions—the ones from their customers using their product, not just from their own operations—and if they're going to get taxed based on them, those inefficiencies start to hit their bottom line. Oil companies have also been more eager to set targets for scope 1 and scope 2 emissions than the much larger scope 3. But accounting standards are designed to solve exactly this sort of inconsistency. And since the total cost needs to be accounted for somewhere, it will probably end up at the level where it's easiest to measure, which means tracking the downstream effects of big companies rather than trying to estimate the carbon impact of every single driver and homeowner.
There's an interesting lobbying dynamic with respect to emissions, too. The most valuable and profitable companies in the world are not in the business of digging up raw materials or bashing them into useful shapes; they're doing some combination of selling bits and selling devices with a high value-to-weight ratio, and thus a high value-to-emissions ratio. These companies are not especially worried about the impact of more expensive emissions; if it hurts them, it's through the broad economic impact, not through direct costs. Meanwhile, many of these same companies are in regulatory crosshairs for different reasons; if the narrative that big companies are good is unlikely to win, the next best alternative for them is a narrative that big companies are bad because of climate change, not because of their impact on elections or society. It's unclear how much of this will filter down to accounting as such, versus how much will affect how that accounting gets translated into taxes and regulations, but directionally we should expect the weight of corporate lobbying and PR to be on the side of expensive emissions.2
So that's the longer-term perspective. But there's a short-term reason to be interested in the carbon credits market, too. Klima DAO is a DAO designed to encourage the purchase of tokenized carbon credits, and it's built to incentivize customers to lock them up in order to maximize their stake in the system. In one sense it's a financial version of carbon sequestration that subsidizes the real-world sequestration through the market mechanism of higher prices. But it's also an effort to craft a game theory-aware smart contract designed to corner the market. Market corners have a long and dubious history, and one of the things that usually kills them is defectors; if you and the Hunt Brothers are buying up silver, every time the price ticks up you have a stronger incentive to sell, and once enough people sell the corner can't be sustained. (Especially if, as in the case of the silver corner, the rules get changed.) Corners don't come out of nowhere; they work best when there's a reason to think the price of an asset could go parabolic for fundamental reasons, and that price changes could be self-fulfilling; the Hunt Brothers were nervous about a collapse of the dollar, and a dollar collapse is a much more plausible scenario when precious metals are soaring.
A DAO won't necessarily be able to fully corner the market (Klima's market cap is around $800m, and it's volatile), so that alone is not enough to cause a sudden spike in emissions credit prices. But the fact that corners can be organized through DAOs, and that smart contracts can be built to disincentivize defection, means that it's possible for prices to suddenly reflect the long-term view that emissions credits will get more valuable as emissions themselves move from being a perceived social cost to a material expense on companies' P&L statements. If there are fundamental reasons to think that the private sector is accidentally net short carbon offsets, and if there’s a way for speculators to systematize the kind of behavior WallStreetBets did in a more organic and self-organizing fashion, short squeezes in emissions credits get more and more likely.
The Streaming Race
Netflix's Squid Game was a massive hit in terms of viewers, but it wasn't able to reverse slow growth in subscribers in mature markets last quarter, and slower subscriber growth has been a common theme among streaming video providers ($, FT). (In fairness to Netflix, US search interest in Squid Game peaked in the first week of Q4, and searches for Netflix rose then, too, so its biggest contribution may be delayed.) In one sense, the fact that streaming growth has started to slow demonstrates that streaming companies were behaving optimally in the last two years: Covid was a behavioral reset, and created much larger demand for at-home entertainment, so any company that had planned to pivot to a streaming model would have wanted to accelerate those plans. Some of them, of course, would fail to make that pivot—the world doesn't need that many general-interest streaming services—but anyone who didn't pull growth forward in 2020 and 2021 would find themselves competing with more formidable alternatives, with far more data, in 2022 and beyond.
The Economist highlights another risk in China's financial system: banks controlled by property developers and manufacturers ($). Companies that need capital, or whose customers need funding to make purchases from them, can benefit from creating a financing arm, but if they take control of an existing bank and redirect its loans towards their own interests, it creates systemic risks. If a company's in-house financing arm collapses, that's very bad news for the company, but if their in-house financing arm is a bank that has lots of other business on its books, then the damage is more contagious.
Techcrunch highlights a new startup, Flexbase,that is providing financing for construction companies, which often have trouble collecting payments from customers. The explosion in niche working capital providers for small businesses continues apace: if there's a sector whose biggest constraint is access to cash, there's room for a specialty provider who can underwrite their particular risks. In the case of construction, one of the drivers of slow payments is the complexity of invoices, so Flexbase automates that as well. Financial problems often turn out to be information problems, and another way to phrase that is to note that the biggest addressable market for some business-focused software products is to make the software free and monetize through financing instead.
Illiquidity Doesn't Always Pay
All else being equal, a liquid asset is worth more than an equivalent-but-illiquid asset, which naturally means that illiquid companies should produce higher returns to compensate investors. While this makes sense in theory, the excess returns from illiquid companies have been declining over time, and now the premium mostly exists for very tiny US-listed companies and slightly larger companies overseas. One way to understand this is through portfolio theory: if a given company is illiquid, investors need a higher return, but one way to get synthetic liquidity is to have a diversified portfolio with as many illiquid positions as possible—take twice as many positions and you've doubled your per-position liquidity. (At least as long as buyers trying to earn returns from owning illiquid assets are not the main price-setters for those assets, in which case they'll all correlate and the liquidity advantage will disappear.)
Instacart Shifts Further Towards Ads
Grocery stores have evolved into a model where selling food is a business with poor returns but selling in-store ad space makes enough money to offset this, and Instacart is moving in the same direction. The company is offering cheaper or free delivery for orders scheduled in advance, and highlighting more special offers. Both of these are efforts to make advertising a bigger driver of revenue: the offers do this in an obvious and direct way, but cheaper delivery also means that more of Instacart's economics are determined by what people order (and thus what products Instacart recommends to them) and less by treating delivery as a product. The price discrimination options in delivery are weaker, since customers are conscious of the cost; in advertising they're more widely available, because advertisers don't experience sticker shock as long as the return on their ads is good enough.
The latest roles from companies in the Diff network:
A startup is building a tool to help small businesses untangle their vendor relationships and make better choices about what products they buy. They are looking for experienced full-stack and backend engineers to join the team. (SF, remote)
A startup working in the fraud detection space is looking for people with data engineering and/or ML experience based in European time zones.
We’re on the lookout for fundamental equity researchers who have experience with alternative data, or people who have extensive data analysis experience, with an interest in equities.
If you're interested in one of these roles, or want to find out what else is available in the Diff network, please reach out. Diff Jobs is free for candidates.
This overstates how much the corporate form is effective, relative to how much we entrust corporations with the particular set of jobs whose output can be so measured. And naturally it means there's a need to measure and mitigate the side effects of the pursuit of profit-above-all-else.
It helps that big tech companies are in the fairly unique category where ambitious capital expenditures plans often make their share prices go up. The market has developed a Pavlovian sense that when someone smart enough to create a centi-billion dollar company decides to spend tens of billions of dollars on datacenters or chip fabs, they probably have a good reason to.