Shifting the Curve: Why Climate Matters More

Plus! Online-to-Offline; Consolidation; Unintended Consequences; Supply Chains as Banking Systems; International Software Nationalism

Shifting the Curve: Why Climate Matters More

Simon Kuznets once observed an inverted-U-shaped relationship between GDP growth and economic inequality: when a country gets rich, it tends to benefit some people more than others—the ones who worked in whichever industry produces competitive exports, levered land-owners who bought in the right places, company founders who start the first company of a given kind in the country and sustain a monopoly, etc.—but as the economy matures, it gets more equal. This may have been an artifact of when the measurements were taken, since inequality has gone back up in many places since Kuznets first wrote, but it's a compelling general model.

And some of his followers have applied it elsewhere: an environmental Kuznets curve. A very poor country is, on average, a low-emissions country: the natural resources stay in the ground, farming happens by hand, and when people need to travel, they walk. A country that's getting richer is, more often than not, a country that's getting dirtier: more cars on the road means more gasoline being burned, coal is a cheap source of power, new houses and offices get thrown together with poor insulation and that's offset with air conditioning rather than retrofitting buildings to be more efficient. In a country growing GDP fast, many of the buildings were built by a significantly poorer country than the one they're in, and there are energy-inefficient substitutes for what could otherwise be architectural features.

As indicators of wealth shift from questions like "How many meals have you had to skip this month?" or "How many of your children wear shoes?" to more pleasant ones like "What color is the message bubble when you send a text?" or "How many vacations do you take each year?" concerns shift, too. So, over time, the carbon-intensity of GDP growth declines. Some of this happens naturally; richer countries have more service-sector jobs, and those jobs tend to need less energy and fewer materials. But there are policy reasons, too. Once people aren't worried about survival, they start worrying more about living conditions, and especially about what their children and grandchildren will inherit. A country that's rich on paper, but where the air is dangerous to breath, is not a great inheritance to pass down.

This describes an earlier generation of environmental concerns, dealing with direct, tangible consequences of industrialization: smog in the air, waste in rivers, etc. In the US and most of Europe, these problems have gotten a lot better over time. A four-day bout of smog in London killed 4,000 people in 1952, but by one measure air pollution there has been declining steadily since the late 19th century, and is now at pre-industrial levels. In other countries that got wealthier more recently, this is still the most salient environmental concern: catch-up growth happens faster when there are light regulations, and sometimes a business that's untenable for environmental reasons in one country can shift to another place where the rules are more loose. Before Wuhan was famous for other reasons, it made international news because of anti-smog protests (Wuhan is an important auto manufacturing center).

As the world economy grows, more countries pass the point where subsistence is likely and environmental concerns get more important, so we should generally expect more places to worry about pollution. If the normal developmental path is followed, India's situation will continue to worsen for a while—at a GDP per capita of about $2,100, they're still at the rising point of the curve, though this is somewhat offset by a model that's done better exporting services than goods. If African countries grow faster than the global average—something I consider more likely given recent capital flows and rising wages in China ($)—that will also lead to a temporary rise in emissions as they move up the curve. Meanwhile, emissions in richer countries are not rising along with economic growth. In the US, for example, they peaked in 2005 and have been declining since then. Some of the worst pollutants have gotten banned, and some of the more egregious manufacturing practices, the sort that dyed rivers weird colors or set them on fire were banned.

But there's another force that's shifting the Kuznets curve, and will affect the development path of poor countries in the future. It feels very dated to talk about environmental issues in terms of tangible things like plastic six-pack rings that choke dolphins, or chemical spills. The primary environmental concern is a) more abstract, and b) more apocalyptic. The old environmental movement was a global approach to local concerns: pollution was bad, but there was a limit to how far it could travel. But concerns about climate change are naturally global in scope. Individual countries can cut emissions, but those cuts are easily offset by other countries changing their behavior—especially if emissions regulations in one country merely shift an industry to another country that cares less. In the old model, this might be a fair trade: let someone else have smog as long as we have blue skies! But if the concern is changes in average temperature, and its variance, then local solutions don't work at all.

There's a growing channel for transmitting that policy view around the world: US multinationals are an important force for integrating poorer countries into the world economy, identifying their potential exports, and tapping in to their consumer markets. And those multinationals tend to be run by people who share rich-country concerns about climate change. At their scale, they often interact with national governments in the countries in which they operate, and many of those governments are full of people who have been educated in the US or Western Europe and have also imbibed these norms. Robin Hanson calls them the world forager elite. Sometimes, this happens in a direct way: In one of the great acts of passive-aggressive diplomacy, the US embassy in Beijing installed an air quality monitor on its roof and started tweeting about air quality levels. After a brief effort to stop this, the Chinese government concluded that it would actually be easier to reduce pollution. Global media and a semi-shared global outlook make it hard for countries to ignore their effect on a global issue, even if they'd prefer to have other priorities.

Thus the list of countries that have, or are considering, net-zero carbon emissions targets includes some powers that try to have a global leadership role (the US, China, the UK, France), but many other smaller countries that would otherwise be surprising contributors to solving a global problem.

Most of these climate targets are far in the future, and given the age of world leaders, most of the people who are making these promises, or will make them, expect the promise to come due after they're dead. But with enough countries making these commitments, it will get hard for any country not to have some stated carbon-neutrality target. And that affects the path forward for developing countries. There are, roughly, four options:

  1. Give up on growth: if you don't climb the Kuznets curve, you don't have to cut emissions because you don't have any. This is politically untenable, and it's rather unfair for rich countries to ask poor countries to stay poor.

  2. Grow fast now, make up for it later: this is basically a bet that future technologies will be able to cost-effectively reduce emissions dramatically, and that the best way to save up for them is to profit from emissions-heavy industries for as long as possible and then switch. Alternatively, it might be a bet on future institutional capability: perhaps the government is not effective enough to pursue clean growth today, but a bigger country with a stronger government might be able to fix those problems in the next generation. It's the St. Augustine approach: Lord make me carbon-neutral, but not yet.

  3. Grow, with an eye to efficiency rather than maximum speed. This will trim a bit off of GDP growth—it is simply much easier to get things done without paperwork. But it might actually lead to a higher peak GDP. Waste can be worth eliminating even without externalities. It's often a sign of low efficiency. Even within emissions sources, there's wide variation in efficiency. This study looks at oil fields in terms of their energy efficiency—how many barrels of oil (or their energy equivalent) need to be burned to extract one barrel? A few are close to 1:1, like Canadian oil sands, while some oil fields get energy returns of 100:1. EVs have a lower maintenance cost per mile than ICs, because they have fewer moving parts. Part of the Toyota Production System is to identify and eliminate waste, on the grounds that a perfectly efficient process will be more cost-effective, and waste is a visible sign of that. (That is, of course, easier with a car company that’s using man-made inputs than it is for, say, a mining company that’s constrained by what form the raw resources come in.) Aiming for low waste means forgoing some opportunities, but it also means selecting the ones that have a longer runway.

  4. Invest in, or hope for, geoengineering. Geoengineering is a broad category, which ranges from proven-but-costly options to unproven-but-very-compelling ones.1 And it doesn't "solve" climate change so much as it addresses specific consequences thereof. But in historical terms, it's something to bet on, because so many tradeoffs between the things people like and the consequences they don't get solved by technology that adds new possibilities, rather than by accepting hard tradeoffs.

As climate change gets more salient, and as the global growth model increasingly leans on American-run companies and American-educated functionaries, the old model will cease to work because people in power won't want to try it for fear of the consequences. But given the scale of investment that would be required to reverse climate change, and ideally to implement more deliberate climate control—once we know we can make temperatures go up or down, why not figure out what the ideal is and set them there rather than reifying "natural" as whatever conditions happened to prevail during some tiny and arbitrarily selected bit of earth's history? For some developing countries, this will mean that the classic path to growth is cut off; they won't be able to do South Korea-style heavy industry, for example, because South Korean steel and chemical companies will be more competitive at complying with complex environmental rules and detailed audits, even if their cost structure is worse. But there will be other growth opportunities that lean in to zero emissions as a target; it's easier to build many things if you know the goals from the start, and an economy suitable for 2050 is one of those things.


A busy week off-site: I did a podcast interview with Nate Meyvis of Branch Prediction, which was lots of fun. I contributed this piece on bubbles to the new A16Z pro-tech tech news site. And Packy McCormick interviewed me for Stripe Sessions. (His newsletter is worth reading.)


The WSJ profiles online retailers that have started using offline methods like print catalogues to attract an audience ($). It's a rare retail business that only works in a single channel. These do show up sometimes, but they're usually more of a way to arbitrage some strange trait of the channel than a real business—for most companies, the ideal is to find one channel that works, scale up until it's saturated, and then find a new way to reach similar customers. Many direct-to-consumer brands reach the same conclusion: e-commerce is a great way to scale a business, but it inevitably means scaling it on a platform owned by somebody else, and that somebody else is usually good at price discrimination. In direct mail, physical retail, and other offline channels, the results may be harder to measure, but profitable approaches are more sustainable.


Another story that's in some ways a reframing of the story above: Buzzfeed's CEO is trying to consolidate other digital publishers ($, The Information)—slightly ironic, because he'd been advocating that for years but hadn't done it until recently, while his competitors had done some M&A. The Buzzfeed thesis is that distribution channels like search and social media have power, and that media companies can only counteract that at scale. Meanwhile, bigger media operations can experiment with more revenue options. This seems like a natural evolution in online media: there was rapid speciation in media outlets, but many of them turned out to be unsustainable, or at least to be not especially compelling businesses. A superior operator in a bad industry can produce a decent business from rapidly acquiring and fixing the worst-performing competitors in that business, and Buzzfeed actually looks like it could play this role well.

Unintended Consequences

Colorado passed a law mandating that companies post salary expectations for jobs, and as a result some companies are posting job ads that specifically exclude Coloradans from applying ($, WSJ). This is partly something they can do because so many people are either location-agnostic or at least more flexible in where they end up—some of us haven't finished unpacking from the Great Covid Exodus from certain cities. But it does illustrate how hard it is to pass a rule that's effective and doesn't immediately lead other economic actors to route around it. Part of the reason employers don't post salaries is, naturally, that they want to be in a better negotiating position. And one reason they don't is that they may get a great applicant for a different position, and be in the awkward position of saying "We're offering you more money than most people get for doing this job, but less than you would have gotten if you'd qualified for that job."

Supply Chains as Banking Systems

Last week I wrote ($) about how the cost of working capital can put financial pressure on some links in supply chains:

The two natural models for this are 1) a model with a separate layer of bank-like institutions that provide capital, so the central companies in any supply chain can stay capital-efficient, or 2) a model where those central companies take advantage of their low cost of capital and their information advantage to act as merchant banks for suppliers, giving them liquidity on terms they couldn't otherwise get.

A new example of this: Stanley Black & Decker is increasing capital expenditures in order to make strategic investments in its suppliers ($, WSJ). While some companies aren't doing this, it's a hard trend to ignore: in a world where most big companies use a just-in-time approach to inventory, the ones with strategic stakes and good working relationships with their suppliers will be much better at working around shortages. And that collides with the trend towards shrinkflation and fixed prices over supply- and demand-responsive pricing for consumer goods: companies seem to prefer shortages to price hikes for some goods. The net result is that it will be more expensive to be the key company in a given supply chain, but that the outcome will be steadier sales, and an opportunity to dominate the market when other suppliers get disrupted. This is an important trend to keep an eye on.

International Software Nationalism

Nigeria's President has opened an account on India-based Twitter competitor Koo following Nigeria's Twitter ban. Koo has been more willing than Twitter to follow the dictates of the Indian government. This highlights an interesting paradox in the rise of nationalist politics around the world: in theory, world leaders with an internationalist outlook should be more willing to cooperate, because they have shared goals. In practice, though, it's a form of coopetition, because each leader's idea of the world order is a little different, and international cooperation means ceding some nation-state power to a higher authority. So within that cooperation is competition over who will be #1. Meanwhile, nationalists in India can cooperate with nationalists in Nigeria, because there's no single prize they're competing for—a world where India can determine its own destiny is not incompatible with one where Nigeria can do the same. This doesn't work if two states share a disputed border, or compete with each other economically, but it does point to a loose coalition of states united by what they don't have in common, and using tools like Koo over ones like Twitter.

  1. That’s “costly” in the sense that it’s a big-ticket expense, not in the sense that it’s not worth doing. Building global hydrocarbon infrastructure—all those rigs, ships, pipelines, refineries, and gas stations, not to mention the road system and the capital cost of manufacturing cars to use all that gasoline, required many trillions of dollars.