Today is the weekly free edition of The Diff, going out to 22,814 readers, up 34 from last week.

In this issue:

• A Normal Country?

• Negative-Marginal-Cost Billboards

• Spillovers

• Piecemeal Privatization

• Financial Protectionism

A Normal Country?

Now, land is being divided up again just the way it used to be in the bad old days... What's going to happen after I'm dead?

—Mao Zedong to Chen Yun, arguing against letting peasant families sell surplus crops, 1962

Well, you can just stop and think of what could happen if anybody with a decent system of government got control of that mainland. Good God... There'd be no power in the world that could even—I mean, you put 800 million Chinese to work under a decent system... and they will be the leaders of the world.

—Richard Nixon, 1971

Is China special, or is China following the same path as other countries that catch up economically and militarily to the previously dominant power? And, if the latter is true, is that good news or very bad news?

This is an important question for everyone: it matters to people in China, naturally, and it's very concerning to the world's current superpower. But it also matters to the countries that deal with the United States (i.e. everyone not currently under sanctions), countries that sell to or buy from China (pretty much any place with either raw materials or any kind of consumer market, i.e. almost everyone). And, naturally, it matters to China’s neighbors, especially the ones with close diplomatic ties to the US.

There is, of course, no definitive answer. Historical precedents are helpful, but only take you so far. At a high level, the contest between China and the US looks a whole lot like the one that took place between the US and UK in the late 19th and early 20th centuries, culminating in the First World War:

• Germany started from a lower base; it was politically fragmented and under-developed.

• Germany was united both through military conquest (Prussia's various wars) and through militarism; the PRC was formed after a military victory, and the post-Mao period of reform began when Deng Xiaoping sidelined Hua Goufeng in 1978, in part because of Deng's much stronger support network in the army. (This was, as the saying goes, very legal & very cool, but it also looks suspiciously like a gentle sort of coup.)

• Germany wanted to catch up to Britain's colonial empire, but because it started late it ended up with some of the least desirable colonies; China, which has been similarly late to geopolitical influence, has ended up working with countries that are pariahs or near-pariahs, with weaker relationships with the rest of the world.

• The arms race between Britain and Germany centered on enormous battleships that required colossal amounts of steel and machine tools; economic competition directly enabled military competition. As AI's military applications get more obvious, China's rapid growth in that area looks more and more like an attempt to leapfrog the US's military. The US today, like Britain then, a) defends open markets politically, or at least advocates for them in other places, and b) defends them practically, by having a big navy. Being unable to do this, or at least being one of two powers that can plausibly decide who gets to do business with whom, would be a significant change.

On the other hand, the US/China and UK/Germany comparison has its issues. For one thing, a very important driver of their naval competition was the fact that Emperor Wilhelm II was a committed Anglophile, and the grandson of Queen Victoria. He liked wearing a British naval uniform. He hosted a regatta to compete with the Royal Family's, which ended up being so popular that all the hotels in Kiel were booked and a cruise ship had to handle the overflow. (No, Dreamforce was not first.) When Germany started expanding its fleet in the 1860s, the first ships they bought were made in England, and it took a quarter-century to make bigger ships at home. Germany's Grand Admiral was also a Great Britain fan, who read English novels and sent his daughters to college there. The First World War is straight out of a Girardian literary analysis: Germany's leaders were obsessed with the UK, wanted to be the UK, and, when they concluded they couldn't, tried to kill it instead.¹

This comparison does not apply to China. Sure, the Politburo liked House of Cards, and they've mostly traded in their Mao Suits for more Alan Flusser-approved attire. China's leaders are definitely not trying to be the United States of America. They're happy to learn from the US, both from successes and mistakes, but the goal is not convergence.

The Baseline

To look at where we are, it's useful to look at where we started. If you happen to enjoy Chinese history, you're in luck given that there's a whole lot of it. China's story more or less requires a cyclical view of history, given that the country has been, off and on, the most powerful in the world, but has reverted to poverty and chaos in the meantime.

One way to look at China's path to the present is that the country has gone through an extended period of catch-up growth. It grew real GDP per capita at an 8.4% annual rate in the forty years after the beginning of reforming and opening up in 1978. But that's actually pretty close to what South Korea did in the forty years through the year 2000 (+7.3%). Japan's turnaround started earlier, and the data are messier, but the country was close to subsistence in the immediate postwar period, and by 1980 ranked among the world's rich countries. All of these countries started their growth in a period of deep deprivation—in the 1950s, people were bartering hair for food ($, Economist). In 1947, a Japanese judge, Yamaguchi Yoshitada, refused to buy black-market rice and chose to subsist only on government-mandated rations. He starved to death. Those circumstances a) mean that future prosperity is compared to a ridiculously low number, a point where poverty is measured in calories rather than dollars per day, and b) those crises lead to reforms, and tend to breed a cadre of ruthlessly effective leaders.

Why did China's growth start as late as it did? In one sense, China started early: in the 1930s, the country was developing an industrial sector, and had switched from importing textile to importing cotton and exporting clothes. Shanghai had a thriving movie industry (one source says it was second only to Hollywood), and the second-highest foreign-born population of any city in the world, behind New York.

Looking back at where China was in 1978 is an iterative process of looking at horrific episodes in history, zooming back a few years, finding more horror, and repeating the process: before the opening up process began there was the Cultural Revolution, in which every institution in China except the military got turned inside out by the party. Before that was the Great Leap Forward, in which the country ran a rather accelerated development plan (a goal of outproducing the UK in steel was accelerated from a planned fifteen years to three) leading to brutal famines. Before that, the CCP regime started off with a series of purges—Russia had to adapt Marxism to a mostly agrarian country by declaring rural landlords class enemies; China, at least according to Frank Dikötter's The Tragedy of Liberation, didn't really have a landlord class, but party cadres did have a landlord-executing quota, so they had to invent the class in order to purge it.² Before that was a civil war, before that a war with Japan, before that a period during which significant parts of China were ruled by warlords, and before that the slow decline of the Qing, which included many abusively unequal treaties with European powers.

China's early reforms under communists led to cascading poverty: the state shut down private businesses, and often subjected their owners to abuse or execution, which destroyed much of the societal function that economies need as a substrate. One business owner who survived noticed that the workers he was most friendly to were the cruelest during this process, and realized it was because of the recursive nature of these purges: the workers most friendly to known capitalists had to be the most enthusiastic in their denunciations, or they'd be next. That dynamic breaks a lot of the social bonds in a country; people work together better when they trust one another, and waves of political action broke those ties. One of the paradoxes of the communist revolution was that its ostensible beneficiaries, peasants and workers, were not necessarily all that enthusiastic about the revolution. Some elites felt a bit better about it—when communist forces besieged Beijing, the nationalist general defending the city negotiated a surrender through his daughter, who was a communist party member.

Meanwhile, the government collectivized agriculture, aggregating family-owned and community-run plots into collective farms. The earlier communities were quite sophisticated, with a web of norms and social obligations that governed sharing land, tools, and labor; some extended families were large enough and powerful enough that they actually issued formal shares to define membership. Collectivization was in part an attempt to impose, top-down, an idealized version of a system that already existed.

In urban areas the state issued ration cards that were required for households to get access to food. This had two effects: it made urban workers entirely dependent on the party-state, which had by this point more or less merged, and it gave the government detailed knowledge of who lived where, and the ability to retaliate against individuals by cutting off their access to meals. A centralized food distribution system is, among other things, a rolling census where the punishment for failure to fill out the form is starvation.

It's interesting to compare this to the agricultural situation in other miracle economies: Taiwan, South Korea, and Japan also shook up land ownership around this time, but mostly by breaking up large estates and giving individual farmers ownership of their plots. As it turns out, these smaller farms were far more productive than huge ones; many of the big efficiency gains from modern agriculture simply weren't available, whereas the efficiency gain from aligned incentives was. China, weirdly enough, launched an agricultural reform in reverse, consolidating small plots into giant state-owned farms, with the attendant drop in labor efficiency. While policy was reducing food supply, it was creating potential energy for later reforms.

Meanwhile, outside of agriculture, the country's industrial sector went through a very different kind of reform. Poor countries have a notoriously difficult time collecting taxes, but poor communist countries have a way to do it: nationalize every major industry, and fund the government's budget out of their P&L. Since production and prices are set by state planners, not market forces, this essentially means that the government can charge whatever tax rate it wants, as long as the money's there. State-run power companies can set their prices high, state-owned tractor companies can compel collectives to buy their wares even if they're overpriced, the resource extraction sector can mandate low wages and have the government keep the difference.³ This system works, in a way, if every company produces according to plan (which they didn't), and if the data are collected accurately (they weren't—one of the darkest examples of this is that grain production quotas were set by weight, not calories, so communes that were having trouble hitting their numbers could just decide to stop drying out grain before storage, or even add water to it. This led to tons and tons of rotting grain). In a completely state-controlled system, all of the economics of intermediate steps in the economy—anything other than selling goods and services to end users—is an accounting fiction that can drift far from reality. And as harvests got worse and production problems piled up, exacerbated by political chaos and famines, the fictions got very elaborate indeed.

So by 1978, extended periods of peace and prosperity in China were basically mythical; they simply had not happened within living memory, and to know about them you'd have to read books (assuming those books hadn't been burned during the Cultural Revolution).

Reform and Opening Up

The economist Steven Radelet once wrote: "In 1976, Mao single-handedly and dramatically changed the direction of global poverty with one simple act: he died." After a brief period of uncertainty, Deng Xiaoping took power in 1978. Anyone who had been hoping that the post-Mao period would involve loosening up controls would have been disappointed to see Deng in charge; he'd been a committed communist since the 1920s, and fought in the civil war, had been admonished by Mao during the postwar agricultural reforms to execute fewer landlords, had gotten purged and then returned to power thanks to his military connections. He was not exactly a bright-eyed reformer. But he was a pragmatist: despite developing misgivings about Mao, he was fairly cautious in his criticisms, and developed a practice of tactically apologizing for his ideological deviations in order to move on to the next order of business. (You can formally relinquish power and informally increase it if the people you rely on know that you recognize how the game is played. It's actually quite a bit like Twitter trolling, where the goal is to make someone else look dumb by acting dumb yourself and getting earnestly corrected by them.)

For a country with a centrally-controlled economy, the post-1978 reforms are surprisingly unplanned. At the level of the central government, early reforms mostly consisted of seeing things happen and not stopping them. For example, China's agricultural liberalization went like this: in 1977, a few collective farms started experimenting with a "family contract" system, in which farmers who grew more than their food quota were allowed to keep or sell the surplus. Deng allowed this to expand, and food supplies rose 8% in the next year.

The government did have a plan for export-driven growth. They were going to sell oil. To anyone looking at oil consumption stats now, that's ludicrous; China is at the top of the list of the world's biggest oil importers. But in 1978, oil production had been growing by 15% annualized for a decade, and, in a place where almost no one could afford a car, that growth led directly to exports. They even had a Stakhanovite folk hero oilfield worker, whose story had been featured in two movies and a stage play. Unfortunately, production at that oil field peaked soon after, and China needed a new source of hard currency to satisfy their import plans.

What they settled on was the export-processing zone. This is a sort of economic airlock that allows a country to indirectly interact with the world market, without enabling the free flow of goods, services, and people. EPZs were entities that could import raw materials without tariffs, and then sell finished goods back tariff-free. They were approved gradually and carefully, so the country limited which ways it would industrialize, and kept its manufacturing focused on fairly simple supply chains. But it gave other exporters a way to tap into China's effectively unlimited labor pool, and gave the country some exports. Hong Kong, for example, was a major exporter of toys at the time (and, in fact, exported as much as the entire mainland in 1978). As Hong Kong's wages rose, it made sense for some of their manufacturing work to move to China instead. (For more on Hong Kong, toys, and globalization, see here ($).)

This pattern persisted, at many different levels: eventually, state planners started keeping the planned part of the economy roughly static, while letting the rest of the economy grow, in effect slowly privatizing by default. But they kept control over some key industries, and maintained the ability to intervene in the private sector financially and politically. The liberalization process was surprisingly gradual, with many goods subject to price controls through the 90s.

And in some parts of the economy, loosening control in a relative sense meant increasing it in an absolute sense. As the banking sector evolved from one state-run bank to multiple state banks, policy banks, and more independent local ones, the central bank started acting more like a central bank, using reserve requirements to keep the economy on cruise control and keep inflation in check. A government banking monopoly can’t intervene in a fine-grained way, but a more diversified banking sector, all tied to a powerful central bank, can make high-level decisions about credit availability and outsource the details to individual banks.

The ration card system went away, but its informational legacy lived on in the hukou system, which ties social services access to legal residence, and which is hard to change. China's urbanization was the largest peacetime migration in human history, but it was tightly controlled (the state still looks at urban/rural population flows as a key economic indicator). Other countries don't have this economic lever; internal population movements are exogenous to policy in most places, but endogenous in China.

Other post-communist countries often privatized in a "big-bang" where state-run companies were suddenly private, government assets were sold, and rules quickly legalized market behaviors and formalized property rights. This usually resulted in chaos, inflation, and a shrinking economy. Some parts of a state-run economy may work okay, and some may be disaster areas, but when the worst companies get privatized and fail, their workers are unemployed—a demand shock—and their customers can't get their products—a supply shock. China's aggressively piecemeal approach was not without its own bouts of economic chaos, with inflation hitting 25% in 1989 and again in 1995, but it was relatively well-controlled.

Any look at recent headlines around Chinese tech companies shows that this is still the broad strategy. It's not an embrace of markets as an end goal, but appreciation of them as a mechanism that helps in some cases but doesn't achieve the state's goals in others. And while the government certainly prefers a higher GDP to a lower one, it's aware of side benefits from large companies: mandatory facial recognition and high market-share platforms are an even better tracking tool than ration cards (although this is complicated by the fact that companies don't see it as being in their interest to share data ($, FT)). The model is not convergence, but adoption: applying workable ideas in a CCP-controlled context, but being willing to jettison the ones that don't apply, or that carry risks.

Survivorship Bias

Why is the party so cautious? Their record speaks for itself economically, given that China has grown fast while ignoring the Washington Consensus. (They've also been able to ignore other parts of that consensus: China was granted most-favored nation status by the US, conditional on making changes that would reduce human rights abuses. When that status came up for renewal, China hadn't met the terms, but got it anyway.) The party's political success is also notable, since it's the world's longest-surviving communist regime.

At another level, though, this bias towards caution reflects how tenuous China's political situation was, and how close individual leaders came to dying during political convulsions. Deng, as mentioned above, was purged for being a "capitalist roader"—his experience was mostly boring (he got a part-time job at a tractor company for a few years), but other family members suffered more. His son was thrown out of a window by Red Guards, and then couldn't get admission to a hospital; he ended up paralyzed. Xi Jinping's father got in trouble for publishing the wrong book, and Xi himself was exiled for a while, during which he lived in cave houses and memorized Faust.

China is the rare country to have ordered its soldiers to shoot at the armies of multiple nuclear powers. Mao ordered a large-scale intervention during the Korean War (Douglas MacArthur suggested using nuclear weapons to retaliate, but was relieved of his command). And Chinese soldiers were shooting at Russians near the Ussuri river border in 1969. Nothing makes the stakes quite as clear as Mao's line: "What if they killed 300 million of us? We would still have many people left."

China had a diplomatic opening-up process that started a bit earlier than the economic one, and ran roughly in parallel. This was not driven by some abstract desire to be a participant on the world stage, but by conflicts with the USSR. And those conflicts were caused by Soviet Russia being insufficiently communist: after Stalin died, Mao saw the denunciation of Stalin as a personal threat to his own rule. This led to a great deal of friction between Russia and China; at one conference, a Russian diplomat lost his temper with a Chinese counterpart said "If you want Stalin, you can have him in a coffin! We'll send him to you in a special railway car!"

It also led to the very real threat of war, which prompted China to reach out to the US and begin establishing more formal ties. This process started with high-level diplomatic talks, and progressed to increasingly close economic ties. After the fall of the Soviet Union, the original justification for the relationship no longer applied, but it was useful enough to both sides that it persisted.

A very good model for how the CCP behaves is to ask what maximizes the odds that they'll celebrate the 100-year anniversary of the founding of the PRC in 2049. The country has clearly come close to the brink many times, and its leaders have personal experience with what that can mean. Lucky countries like the US and Australia, who largely avoided national crises except for the self-inflicted kind, simply don't have the same level of risk aversion. In most real-world domains, risk aversion is not expressed by passivity, since that just means letting someone else set the agenda. Instead, it's displayed through strategic defensive action—always being in a position to improve one's position, and ensuring that action opens up more possibilities than it closes so there's always a new option in response.

Hamburger University and Xi Jinping Thought

One fascinating broad question about the world is: how do enormous organizations actually act at a local level? The Pope has to supervise 1.2 billion Catholics, and yet any Catholic, nearly anywhere, can go to mass and confession. The US military has almost 1.4 million active-duty personnel, and any high-level strategy ultimately boils down to specific people pulling triggers. And McDonald's can serve almost the same burger at more than 39,000 locations around the globe. How do large groups achieve consistent individual outcomes? The answer, in McDonalds' case, is Hamburger University, which trains the best of the best (of McDonald's managers and franchisees) in the fine art of a consistent burger-and-fries experience. At some point, any organization with lots of personnel will need some standardized way to coordinate their behavior, and standardization usually involves writing things down.

A very useful feature of China is that it's big. This is nice for commentators; as a country of 1.4 billion people improves its relative standard of living, it's going to continuously set records. (More cement in three years than the US used in a century is a classic of the genre.) But China's size also affects its governance: no other country has to operate on that kind of scale. If Guangdong province were an independent nation, for example, it would be tied with Mexico in 10th place on the list of most populous nations, and would have the 13th largest GDP, just after Australia. If the CCP has some policy goal, it faces the Catholicism/US Military/McDonald's Problem at unprecedented scale: how does it get implemented?

China's answer is a surprisingly decentralized system, where 85% of spending decisions happen at the provincial level or below. Data flows up the org chart, and then targets flow down. Best practices flow up a bit, and then sideways: if something works in one city, the province will find out about it, and pretty soon it will be tested in more cities in the province. At that point, if it still works, it's something that works at the province level, and the central government will find out and spread it to other provinces.

The mechanism for this is The Party. The Chinese Communist Party is a vast organization. Only consumer-facing tech companies can credibly claim that they've gotten more signups. It has over 90 million members. Considering the usual information decay in a game of telephone, it's not useful for a directive to be transmitted down the chain of command; some things have to be broadcast. This makes China a surprisingly open country: if the party line is that denouncing Stalin was the beginning of the end of the USSR, that will find its way into an official speech.

A one-party state does not make state and party completely synonymous. The state is subordinate to the party, and the party is bigger than the state. One survey showed that one third of business owners were members of the communist party, compared to overall party membership of just under 10% of adults. Party members earn about 10% more than average, although one study used twins as controls and argued that this is because of who joins the party, not what the party does for them. Either way, the party is very central, and its behavior is coordinated in public, because there's no way to send a private message to ninety million people without it falling into the hands of person #90,000,001 who wasn't supposed to get it. Better to make the speeches public and wrap the details in jargon than to miscommunicate.

Normal, Weirdly Enough

Different growth models can produce great results. Venice and England each had restrictive trade laws and death penalties for industrial espionage at points in their history when they were fairly rich places. Hong Kong grew fast with a highly libertarian model, the USSR grew fast with the opposite. In that very meta sense, it's not exceptional to see a country where the economy liberalizes but the political situation, if anything, gets more tightly controlled over time.

A general economic development principle is that markets usually work best—the USSR’s growth push is an exception, but a brief and bloody one—and the parts of the economy that are regulated need to be so clear and ironclad that they’re basically equivalent to a feature of the natural world. China fits into this model nicely. The country does tolerate a surprisingly free market, at least for a self-styled communist state, and has used the market to provide full employment and a much higher standard of living. But a market economy was never the end goal. It was something between an accident and an emergency measure, driven by the need to fund defense against the USSR and by the failure of the collectivized agricultural economy.

Even if the numbers line up with other countries, the comparison falls flat: China has taken a strange path to being a more normal economy, but "normal" is a waypoint, not a destination. The country's political model is flexible enough to shy away from market reforms in the event that they cause more disruption than the party thinks they're worth, and the technology enabled by new wealth provides a more authoritarian safety valve.

Deng Xiaoping once remarked that his government would be viewed as legitimate as long as real GDP per capita kept growing fast. Deng worried that if the economy only grew at four or five percent annually, the government's legitimacy would be questioned. This is true, but insufficient: Deng's predecessor did not achieve 5%+ GDP growth over any sustained period, and left China quite a bit poorer than he'd found it. But he did have other ways to maintain legitimacy. As growth slows, the natural question becomes whether a gradual rate of improvement is enough to keep the government popular, or if other approaches to legitimacy turn out to be their best option.

Further reading: Henry Kissinger has been chatting with China's leaders for decades, and is thus spectacularly well-informed. On the other hand, he's a diplomat, so he has to be diplomatic. On China is a great little book. Frank Dikötter's Tragedy of Liberation fills in some blanks between the end of the civil war and the start of the Great Leap Forward. What happened in between was not pleasant. The Ezra Vogel biography of Deng is extremely detailed, and a good look at how giant bureaucracies function. The Pantsov/Levine bio is very much in dialogue with it, and fills in some important gaps. Graham Allison's Destined for War is much more thoughtful than it sounds—it's not an argument that US/China military conflict is inevitable so much as it's an argument that the historical record is grim. Robert K. Massie's Dreadnought is a very thorough analysis of the British/German arms race, and is the source for all the amusing details about Wilhelm. Barry Naughton's book on China's economy is quite helpful for understanding all the China-specific economic arrangements out there, and putting them in historical context, but Arthur Kroeber's book is much more terse.

Share

Elsewhere

Negative-Marginal-Cost Billboards

Netflix is a business that's great at creating content, delivering it, and marketing it. The content creation part of the business is very visible, since a decent fraction of the most talked-about shows and movies are exclusive to the platform. Their content delivery is also pretty visible, because they use writeups of meaty problems they've solved as a recruiting tool. Their marketing gets a bit less attention, but is naturally an important part of the model: Netflix's shows give it a regular hook for ads, and increasingly the company has to solve the targeting problem that most of its addressable audience in the US already subscribes. The new Netflix merch store is a nice cost-effective way to keep promoting their shows: a branded t-shirt is a walking billboard, but it's also revenue-producing. The economics of apparel retail are a lot less favorable than those of streaming, at least on the margin, but streaming economics work better when the cost of user acquisition is low, and this particular marketing channel pushes that cost below zero.

Spillovers

A few months ago I wrote about positive economic spillovers from the pandemic ($), and better treatments for other diseases was at the top of the list. The Economist profiles South Korea's biotech industry ($), highlighting the case of Seegene's 9x increase in annual sales in 2020. Biotech may join the cohort of industries that are viewed, not just as a source of exports and growth, but as a strategic necessity.

Piecemeal Privatization

South Africa's state-controlled electricity company has struggled to keep the lights on, and the government is allowing private companies to generate their own power for industrial purposes ($, FT). This is a close echo of the Chinese model: they're not privatizing the existing power company, but they are allowing mines and metals companies to build their own power generation. If more of the country's institutional effectiveness is embedded in these private firms than in the state, that's a useful kind of outsourcing: it leads to less demand for the state power company, and ensures that mistakes there don't impede South Africa's ability to produce the raw materials exports that ultimately fund the government.

Financial Protectionism

An independent investigation shows that Toshiba's management worked with the Japanese government to thwart a shareholder activist campaign ($, FT), which ultimately succeeded. I've previously written some optimistic things about Japanese equities ($): valuations are low, and corporate governance reforms make shareholder activism more viable. But this story indicates that, at least for the largest companies, it's still difficult for outsiders to have influence.