Sunday, 30 September 1984. About nine in the evening, a dormitory at Peking University.
Undergraduates in the Biology Department’s 1981 intake, cytogenetics majors, are arguing about a banner. In the morning their column marches past Tiananmen in the parade for the People’s Republic’s thirty-fifth birthday. They want to say that they support the reform policy. The parade’s organisers have ordered that students carry nothing but paper flowers.
Someone proposes “Education needs reform.” Someone else proposes “Reform should be accelerated.” “Long live Deng Xiaoping” is turned down: many of them find the wording inappropriate, and lacking in cordiality. They settle on “Hello, Comrade Deng Xiaoping!” Then they cut it to “Hello, Xiaoping!” Peking University’s own account says a long line would not catch the eye; another account says the longer form felt stiff and not warm.
The green paper the students’ union has given them is too thin and fragile. One student lays his own plastic sheet underneath it as a backing. They roll the banner up, wind ribbon round it and top it with paper flowers, so that it passes for a tall bouquet. By four in the morning they are in line.
Monday, 1 October. The air over the square is cool and hazy, thick enough to hide most of the air force fly-past. It is the first military parade since 1959: more than ten thousand troops, mechanised columns and missiles, half a million people standing in assigned places. Deng Xiaoping reviews the troops from an open car and gives the day’s only speech. “We have placed socialist modernization above everything else in our work,” he says.
As the Peking University column passes below the rostrum, the students open the banner. By the Golden Water Bridge a China Youth Daily photographer, He Yanguang, takes the picture. The Christian Science Monitor’s correspondent, in his report on the parade, does not mention it.
The banner greeted one man by his given name. Much of the policy it welcomed had started somewhere else, and lower down: in a machine-tool plant in Sichuan, in the villages of Anhui, on a peninsula across the water from Hong Kong. This chapter goes back six years to find them.
1978SichuanIn the years after the Cultural Revolution, at least a third of China’s state enterprises were not producing normally. Nearly a quarter of the state’s independently accounted industrial firms, 24.3 percent, were making losses, which came to 3.75 billion yuan. On 30 June 1978, the planner Fang Weizhong recalled, Hua Guofeng heard a report on a vice-premier’s tour of five European countries and said that foreign firms’ management had lessons worth borrowing. In October a delegation from the State Economic Commission went to study management in Japan.
In October 1978 Sichuan, whose Party chief was Zhao Ziyang, chose six state enterprises for a trial. They were a machine-tool plant, the Chongqing steelworks, a seamless-tube mill in Chengdu, a chemical works, a nitrogen-fertiliser plant at Xindu and a silk mill at Nanchong. If an enterprise met its plan for the year, it could keep a small share of its profit and pay its workers a small bonus. The economist Zhang Jun calls the change trivial by later standards, and at the time an ice-breaking voyage.
The machine-tool plant was Ningjiang. It had been split off from a Nanjing plant and moved in 1965 to Guan county, now the city of Dujiangyan, and it made small precision machine tools. The state supplied its materials, took its output, set its wages and absorbed its losses. For years it turned out five or six hundred machines a year. It could make more than a thousand.
Around September 1978 a secretary of the provincial Party committee, Du Xingyuan, came to the plant and asked what it needed. He spoke of loosening the ropes that bound the enterprise. The director, Liu Lunbao, and the Party secretary, Zhao Chuanren, asked that the makers of machines be allowed to deal directly with the people who used them, and listed twelve demands. The Party’s own account sums them up in one line: give us some power. In January 1979 the province issued fourteen articles on enterprise powers. Ningjiang gained the right to sell its own machines. Once it met its targets, it could keep 7 percent of its profit and pay bonuses.
The chief engineer, Chen Shanyang, then proposed an advertisement. Opponents in the plant asked a question: the state had never declared machinery to be a commodity, so would advertising it be a political error? After several nights of argument the plant’s Party committee decided to advertise, and to do it in the People’s Daily. The paper’s first answer, Zhao Chuanren recalled, was that a Party newspaper did not carry advertisements. The First Ministry of Machine-Building took the copy to Beijing, and a vice-minister, Sun Youyu, signed it off. The whole effort took more than three months. The fee was 8,000 yuan, which Zhao said was the price of one of the plant’s machines.
It ran on Monday 25 June 1979, across nearly a quarter of page four, and its largest type read, in translation:
Accepts direct orders from customers at home and abroad.
The notice was signed with the plant’s name, its address in Guan county and a telegraph code, 3068. The Party’s history site calls it the first advertisement for capital goods in the People’s Republic. On 5 August the People’s Daily ran a front-page lead backing the plant’s practice of setting output by sales. Within two months Ningjiang had signed contracts for 1,300 machines, more than twice its old annual output. For years afterwards, Liu Lunbao said, it needed only one salesman. By 1982 it was making 1,710 machines a year. The best-paid workers, Liu said, took home several hundred yuan a month in wages and bonus, more than the director. Bo Yibo, a veteran of the Party’s economic leadership, later said the advertisement had earned merit in China’s economic reform.
The trial spread by stages. In May 1979 eight large enterprises, among them the Capital Iron and Steel works, the Tianjin Bicycle Factory and the Shanghai Diesel Engine Plant, began trials of their own. In July the State Council issued five documents, on enterprise powers, profit retention, a fixed-asset tax, depreciation and working-capital loans, after an argument at a conference in Chengdu between the Ministry of Finance and local governments that wanted the trials. By the end of 1979, by Zhang Jun’s count, about 4,200 enterprises were in them. A Party history counts 6,600 by June 1980. In 1982 more than 36,000 state industrial firms were moved onto a rule of fixed payments to the state, with the excess kept by the firm.
The political scientist Sebastian Heilmann has a name for this method: experimentation under hierarchy, a way of making policy that legitimises local initiative while the centre keeps final control. Its two working parts are the model experiment, or trial point, and the spread “from point to surface” once a trial is judged to work.
Judging was the hard part. The economists Theodore Groves, Yongmiao Hong, John McMillan and Barry Naughton studied state enterprises in four provinces, with yearly data from 1980 to 1989. When output decisions moved to the firm and firms kept more of their profit, managers paid more in bonuses and hired more workers on fixed-term contracts, and productivity rose with both. Greater autonomy, though, tended not to raise what the firms paid to the state. Zhang Jun records the same pattern: handing firms power and profit made them prone to over-expand, and fed inflation, and as the state’s revenue weakened it halted the contracts and from 1983 began to replace remitted profit with tax. A trial point could work for its workers and still not work for the treasury.
In the countryside, the scholar Zhao Shukai writes, policymakers did not plan or design the change: farmers pressed ahead step by step, and the grassroots kept quiet about it, so those at the top did not know.
Less than a month after the meeting in Xiaogang, the Third Plenum’s draft decision on agriculture still carried two bans: no dividing the land to farm alone, and no contracting output to households. In April 1979 Document 31 added a third: no splitting accounting units into smaller ones. In 1979 the state raised the quota price for grain by 20 percent. Grain sold above the quota fetched 50 percent more than the new quota price, in effect 80 percent over the old one. In September 1979 the Fourth Plenum softened one ban. “No household contracting” became “also do not contract output to households”, with exceptions for remote households in the hills.
On 31 May 1980 Deng spoke to senior officials about rural policy. He named Feixi county first: most of its production teams had adopted household quotas, and production had risen sharply. Then he named Fengyang county, where nearly all the teams worked an all-round contract system and production had risen enough to change the county’s prospects. Some comrades feared for the collective economy. “I think their fears are unwarranted,” he said.
Fengyang was Xiaogang’s county.
Replay of an earlier scene: Xiaogang, 24 November 1978The prologue told that meeting. Look again at one clause. If the scheme failed and the team’s cadres went to prison or were executed, the others would raise their children to eighteen. The farmers feared that what they were doing could get them killed. In Anhui the idea had been tried before, openly, and with permission.
Fengyang county had been starving. In one of its communes, Liyuan, the population fell from 5,730 in 1957 to 2,870 in 1961. More than half had starved to death or fled, the county’s own Party researchers wrote later. Grain output there fell 77 percent between 1955 and 1961.
On 6 March 1961 Anhui’s provincial committee, answering peasants’ demands, decided to try a system that fixed output to the field and responsibility to the person. By the end of autumn 85.4 percent of the province’s production teams used it, and more than 20 percent of areas elsewhere in China had something similar. Anhui reported to the Centre and to Mao in writing three times, on 20 March, 27 April and 24 July, and had Mao’s consent. In December 1961 Mao told Anhui’s provincial Party secretary, Zeng Xisheng, that he had changed his mind: now that the production team was the basic accounting unit, the fields were no longer needed. After that, the Party’s chronology records, the “responsibility fields” were wrongly attacked as the “individual-farming wind”.
On 30 January 1962, within two months of turning against Anhui’s fields, Mao told a conference of more than seven thousand cadres that any mistakes the Centre had made ought to be his direct responsibility. In July Vice-Premier Deng Zihui told the Central Party School that field management could be contracted to households, and many leaders agreed. In September, at the Tenth Plenum, Mao made class struggle the theme and condemned the individual-farming wind, which the chronology glosses as contracting output to households. Deng Zihui’s proposals were condemned with it.
In February 1964 Mao told two visiting foreign Party leaders that people in the Party’s Rural Work Department wanted “three freedoms and one contract”: private plots, free markets, responsibility for profit and loss, and contracting output to households. This, he said, was the domestic programme of revisionists, and they included Central Committee members, secretaries of the Secretariat and vice-premiers. The chronology calls his estimate completely unrealistic.
1965ShanghaiOn 10 November 1965 Shanghai’s Wenhui Bao printed Yao Wenyuan’s attack on a play, Hai Rui Dismissed from Office. It tied the play’s theme of “returning land” to the individual-farming wind and called the play a poisonous weed. Mao approved its publication. The Party’s chronology calls the article the fuse of the Cultural Revolution.
Read against 1961, the clause about the children carries more weight. What the Xiaogang farmers signed had been done openly in their own province, welcomed by farmers by the Party’s own account, and then condemned as part of a revisionist programme.
After Deng’s talk the bans gave way in stages. In September 1980 Document 75 allowed household contracting in poor areas and told ordinary areas not to adopt it. In Anhui, teams farming by household rose from about 8,700, or 23 percent, in May 1980 to about 20,000, or 53.9 percent, by the end of the year. In 1981 planners drew a design that limited household contracting to about a fifth of localities. By the central rural work conference that October, it was expected to pass half within the year. Document No. 1 of January 1982 called household contracting one of the responsibility systems of the socialist collective economy and left the choice of form to the farmers. Zhao, who dates the high-level dispute from December 1978 to January 1982, sums up its three years in a line: the farmers pressed forward, and the government retreated step by step. In Shaanxi, which Greenhalgh describes as somewhat late to the farm reforms, 99.7 percent of the former production teams farmed by household contract by the end of 1983.
The harvests came. By 1984 China’s rice output was 30 percent above 1978, and its wheat output 63 percent above. Farm families began to buy what the communes had owned. Households held 1 percent of China’s tractors in 1980 and 43 percent at the start of 1984.
Who made them is still argued over. The economist Justin Yifu Lin, measuring the growth in farm output from 1978 to 1984, finds that higher prices worked mainly by drawing more inputs onto the fields. John McMillan, John Whalley and Lijing Zhu, measuring productivity rather than output, give most of its rise to the change in how farmers were paid. Both put the household system first. Neither gives it all of the credit.
Counted by what went into the fields Lin's Table 6
The whole bar is the 42.2 percent rise in crop output, 1978–84.
- The household system46.9%
- Fertiliser32.2%
- Land, labour and capitalland −1.8%, labour +4.5%, capital +10.8%13.6%
- Crop mix and cropping intensitycropping intensity −1.9%, nongrain ratio +3.7%1.8%
- Unexplained5.6%
1973 · the 43 Plan buys 26 plant sets abroad, fertiliser among them; their share of the harvest is unmeasured
Counted by what changed in prices Lin's Table 7
The whole bar is the 42.2 percent rise in crop output, 1978–84.
- The household system42.2%
- Prices, net of both kindsstate procurement prices +16.0%, market prices −0.8%15.2%
- A steady trend, which includes better supply of inputs29.8%
- Unexplained12.7%
Two of Lin's printed shares are slips, and the bars use the shares his points give: Multiple cropping (MCI) is printed 1.94 where its points give −1.94; Trend (T) is printed 29.74 where its points give 29.84.
The harvest had several fathers
Crop output rose 42.2 percent between 1978 and 1984. Justin Yifu Lin split that growth in two ways. Counting inputs, the household responsibility system accounts for 46.9 percent of it and chemical fertiliser for 32.2. Counting prices, the system accounts for 42.2 and prices for 15.2, with a steady trend and what is unexplained making up the rest.
Source: Justin Yifu Lin, “Rural Reforms and Agricultural Growth in China,” American Economic Review 82(1), 1992, pages 34–51 (Tables 6 and 7). The 1973 marker is from Zhou Hong (2019), not from Lin, and is tied to no bar.
See the numbers
| Table | Row | Points of growth | Share, from the points | Share, as printed |
|---|---|---|---|---|
| Table 6 | Land | −0.74 | −1.75 | −1.75 |
| Table 6 | Labor | 1.91 | 4.52 | 4.52 |
| Table 6 | Capital | 4.57 | 10.82 | 10.82 |
| Table 6 | Fertilizer | 13.60 | 32.20 | 32.20 |
| Table 6 | Household-farming reform (HRS) | 19.80 | 46.89 | 46.89 |
| Table 6 | Multiple cropping (MCI) | −0.82 | −1.94 | 1.94 |
| Table 6 | Ratio of nongrain crops (NGCA) | 1.56 | 3.69 | 3.69 |
| Table 6 | Residual | 2.35 | 5.56 | 5.57 |
| Table 7 | Household-farming reform (HRS) | 17.82 | 42.20 | 42.20 |
| Table 7 | [Market price]/[input price] (MP) | −0.32 | −0.76 | −0.76 |
| Table 7 | [State procurement price]/[input price] (GP) | 6.75 | 15.98 | 15.98 |
| Table 7 | Trend (T) | 12.60 | 29.84 | 29.74 |
| Table 7 | Residual | 5.38 | 12.74 | 12.74 |
One possible input had been ordered under Mao.
After a poor harvest in 1972, Beijing concluded that its farms needed far more fertiliser. In 1972–73 it contracted for thirteen large ammonia-urea complexes from American, Dutch, French and Japanese firms. American analysts expected them to supply about 3.5 million tonnes of nitrogen a year once they were running.
1973In March 1973, to strengthen agriculture, light industry and basic industry, the Centre approved a wider programme, the 43 Plan, named for its size: $4.3 billion to import twenty-six complete sets of equipment. By the account Wikipedia cites to a 1989 Chinese study of foreign economic cooperation, China paid from its foreign-exchange reserves and on deferred terms: Mao-era China was buying foreign process plants partly on credit five years before the reform began.
Lin’s price effect runs through inputs in general, not through these plants in particular.
1979ZhongnanhaiBack on the coast, five years earlier, the trial points were drawn on a map. On 6 January 1979 the Ministry of Transport and Guangdong province asked the State Council to let China Merchants, a Hong Kong company under the ministry, build an industrial zone in Shekou commune, in Bao’an county. On 31 January the company’s executive vice-chairman, Yuan Geng, went to Zhongnanhai with a vice-minister of transport, Peng Deqing, and a map of Hong Kong. Vice-Premier Li Xiannian studied it, marked out the whole Nantou peninsula and offered it to him. By the Shenzhen government’s account, Yuan judged that too large: the money at stake was huge, and a failure would shake everything around it. With a red pen he marked 2.14 square kilometres at the peninsula’s southern tip. That became the Shekou industrial zone. In July 1979 the blasting to clear the site began. In April 1979 Guangdong’s first secretary, Xi Zhongxun, proposed that his province take the lead by carving out parts of Shenzhen, Zhuhai and Shantou, and Deng agreed. On 26 August 1980 the Standing Committee of the National People’s Congress approved regulations for special economic zones in Guangdong.
Beside the road to the ferry for Hong Kong stood a billboard. It carried two lines from a set of slogans Yuan had written, chosen by the zone’s management committee: Time is money, efficiency is life. It was meant to show people from Hong Kong that Shekou valued efficiency. The Shenzhen government dates the lines to 1982, written on Yuan’s way back from a business trip to Hong Kong.
A slogan that would have been heresy a decade earlier
Yuan Geng's billboard faced the water across from Hong Kong. When a Sanyo delegation arrived in 1983, its paperwork was done in three days.
Source: Shenzhen Government, the Shekou story; Wikipedia, Yuan Geng
In 1983 Sanyo Electric sent seven people to Shekou to negotiate a factory. They expected the talks to take about a week. The paperwork was done in three days, and the factory was running about six weeks later.
In 1984 Deng visited Shekou, and Yuan raised the slogan with him. On 24 February, back in Beijing after a tour of Guangdong, Fujian and Shanghai, Deng told a few leading members of the Central Committee that the guideline was to open and not to close. He praised Shenzhen’s rapid construction, said Xiamen’s zone was too small, and quoted Shekou: “Their slogan is ‘Time is money, efficiency is life.’” By 30 June he could tell a Japanese delegation that China had opened fourteen large and medium-sized coastal cities.
The billboard spoke to investors. Foreign direct investment into China, which had trailed India’s until 1981, reached $1.26 billion in 1984. Over the five years 1980–84 China took in about $2.65 billion and India about $0.27 billion. China’s merchandise exports rose from $9.96 billion in 1978 to $26.1 billion in 1984.
In 1992, the year of the journey told in A Special Train, foreign investment into China rose from $4.4 billion to $11.2 billion. In 1993 it reached $27.5 billion. In the Shenzhen zone, foreign-owned factories multiplied.
Picture one of the young women who came from the villages to work in them. She is a composite, assembled from the scholarship rather than from one life. She is in her late teens, from a village inland, on a train to the factory towns of the Pearl River delta. The word for her is dagongmei, working girl. The sociologist Pun Ngai, who spent eight months on the floor and in the dormitories of an electronics factory in Shenzhen, describes women like her: twelve-hour shifts, four or five years in the factories on average, and then home. A household-registration rule dating from 1958 keeps a rural-born worker from settling in the city, and her family expects her to marry by her late twenties.
Chain C061958 → 2020Planted: the registration rule returns in A Job Card194720262020In 2020 the census counted 71.09 million migrant children, living away from the place where they were registered. It counted 66.93 million children left behind: living at their registered home without one or both parents, who had been away for more than six months. In the countryside the left-behind numbered 41.77 million, nearly four rural children in ten.
The census does not trace these children to the zone. It counts every child with a parent, mother or father, away for more than six months, wherever that parent went, and no source used here splits the 66.93 million by destination.
In the parade that carried the Peking University banner, one float bore the words “Contracting linked to output is good”. Shekou’s slogan rode on another, and from then on it was known across the country.
In October Deng told the visiting West German chancellor, Helmut Kohl, that China regarded reform as a revolution, though not, of course, a cultural revolution, and that the coming plenum would turn reform on the cities. On 20 October the Central Committee adopted that decision. Deng remarked at the adoption, and repeated two days later, that it read like “the draft of a textbook on political economy”. Six years of trial points had come first: a newspaper notice from a machine-tool plant, a contract in a mud house, a billboard on a ferry road. How much of the draft came from them, the sources read for this book do not say.
In Delhi, over the same six years, the state revived a failed car company and lowered a line in its labour law.