Sunday, 19 January 1992. Nine o’clock sharp, Shenzhen railway station.
A special train rolls in, and an 88-year-old man steps down. He is, officially, retired. He is, officially, on a family holiday. He wears a dark grey jacket and black Western trousers. The provincial and city leaders are waiting on the platform, and when someone suggests he rest first, Deng Xiaoping says he has come to see everything.
Eight years earlier this ground was paddy and fishponds. Now he is taken up to the revolving restaurant on the 53rd floor of the International Trade Centre, the building that stands for what the city calls Shenzhen speed, and he looks down at the plan of the city while the party secretary reports that income per head has risen from 600 yuan to 2,000. At a laser-disc factory that uses Philips technology he asks the chairman, Ye Huaming, whether he is General Ye Ting’s second son. “No, the fourth,” says Ye, holding up four fingers. Deng remarks that foreign capital is only about a quarter of Shenzhen’s economy.
Not one newspaper will print any of this. The first public report appears on 26 March.
He left Beijing on 17 January and would be back on 21 February. Between the two dates a freeze would break.
The freeze was two years old. China’s growth had dropped to 3.8 percent in 1990, and in December 1991 People’s Daily was still running articles that equated a market economy with abolishing public ownership. Deng’s answer, given in conversations and never in a speech, was a test. Judge a policy by whether it raises the productive forces, the nation’s strength and people’s living standards. Plan against market does not divide socialism from capitalism. Guard against the Left more than against the Right. In Zhuhai on 23 January he said the danger was mainly from the Left, and in Shanghai he said that Pudong and Shenzhen would have to answer with facts by the end of the century. The Politburo issued the talks as Document No. 2 on 28 February and endorsed them on 9 and 10 March.
Then the slogan had to be found. On 1 April Jiang Zemin asked for a formula for the coming Party Congress. Economists including Wu Jinglian proposed “socialist market economy”. On 25 May Jiang telephoned Zhu Rongji in Taiyuan to ask whether the words could be said, and on 12 June Deng agreed that Shenzhen already was one. In July and August the drafters tried to demote the phrase to a third-level heading, and Wu wrote again on 5 September. The Congress report of 12 October made it the goal: a socialist market economy “on the basis of public ownership as the mainstay”. A constitutional amendment followed in March 1993.
No law changed in 1992. What changed was what a provincial official could safely sign. Growth rose from 9.4 percent in 1991 to 14.3 percent in 1992. Foreign investment, as A Billboard showed, surged in the same two years. In Delhi, meanwhile, Manmohan Singh was keeping his promise to redeem the gold, as A Crate of Gold told it.
1992ShenzhenA freed market makes a boom, and a boom needs rules that did not yet exist. In May 1992 Shanghai released its daily price limits, and one share rose 105 percent in a day. In August Shenzhen, a city of about 800,000, announced five million lottery forms for new share issues. A buyer could take up to ten. People came from all over China. Estimates of the crowd run from half a million to 1.2 million. The forms, sold at 300 banks, cost about a third of a month’s pay, for a chance of under one in ten of receiving shares. By noon on 8 August more than 800,000 people were in the streets. On the 9th the forms sold out in two hours.
On the 10th thousands who thought officials had kept forms for themselves rioted. The riot police used tear gas and truncheons, trading was halted, and at least ten people were hurt. The Christian Science Monitor called it the worst unrest since Tiananmen. In the early hours of the 11th the city announced 500,000 more forms. On 12 October the State Council created a Securities Committee under Zhu Rongji, with the new securities regulator as its office. A liberalising signal made a boom, the boom made disorder, and the centre answered with a rule-maker. It is a pattern the chapter will see again.
India had its own version that same year, and it ended in a regulator too. From April 1991 to April 1992 the Sensex rose 274 percent, from 1,194 to 4,467. The State Bank of India then reported a shortfall in government securities, and the broker Harshad Mehta was found to have used bank receipts in the repo market to fund share buying, about Rs 3,500 crore in all. He was jailed in 1992. Control over capital issues ended that year and the securities regulator, SEBI, became the apex body, but the reaction to a repo scam was to ban repos in government securities, which dried up liquidity for years. Both countries met a market scandal with a rule-maker; the difference was what each had just freed.
1993DalianBy 1993 the economy was running hot. Investment in fixed assets rose about 40 percent in 1992 and 50 percent in 1993. Consumer prices rose 14.6 percent in 1993. Zhu Rongji, as vice-premier, planned a conference at Dalian from 10 to 13 June with World Bank guests, among them Franco Modigliani and Nicholas Stern. It urged positive real interest rates, credit ceilings and price reform, and on 24 June the State Council issued sixteen measures, thirteen of them economic tools rather than orders. Prices still rose 24.3 percent in 1994, but by 1996 inflation was 8.3 percent with growth near 10. AFP credits Zhu with the soft landing. India, for its part, had unified its exchange rates in its budget of 27 February 1993, about ten months before China did.
On 31 May 1993 a drafting group of 25 under Wen Jiabao had entered a guest compound at Yuquanshan, and after eight drafts the Party’s Third Plenum adopted a fifty-article Decision on 14 November. The market was to play the basic role in allocating resources; state firms were to become a “modern enterprise system”; distribution was to put efficiency first with fairness considered; social security was to be multi-tier. Opponents feared the words “labour market” would undermine the working class’s standing as masters of the enterprise.
199313:25Kuichong, ShenzhenFive days later, at 1.25 on the afternoon of Friday 19 November, a fire began in the ground-floor warehouse of the Zhili toy factory in Kuichong, in Shenzhen. It was a Hong Kong-run plant making toys for export. Of 418 registered workers, 404 were on site, most of them young women who had come from the countryside.
The account that Hong Kong’s Sing Tao Headline drew from the trade-union inquiry and the Workers’ Daily gives the layout. The windows were welded shut against theft and most were meshed. The east stairwell was locked behind a gate, and the west was piled with goods. Of the four ground-floor exits, three were sealed, and the fourth was a barred passage about eight metres long and one wide. A short circuit had started the fire. The fire crew, 21 men with two trucks, found no hydrant and drew water from a kilometre away.
Eighty-seven people died; accounts give 51 injured, though one of them counts 54 burn victims. The account names three of them. Xie Yunxia, 17, had quit four days earlier and been paid, but her foreman talked her into staying until the New Year. Wang Xiaofang, 21, had written home that she could send only 200 yuan because she was keeping back her fare: she would be home for the Spring Festival and marry after it. Chen Yuying, 19, was burned over three-quarters of her body and would have 37 operations and two amputations.
In December 1994 the factory director was sentenced to six years and the Hong Kong owner to two. Two fire officers received 17 and 10 years for dereliction of duty and bribery. The Labour Law had already been passed, on 5 July 1994, to take effect on 1 January 1995. The account says the fire helped bring the Labour Law into being.
This is the price of the light-factory step, paid in one afternoon. A Staircase told how the step works: hands before capital, a cheap product, a harsh export teacher. The step, as told there, says nothing about who inspects the exits.
India’s side of the toy step is a document, not a minute. The nearest one is a report from two years later. In August 1995 the Advisory Committee on Reservation asked a committee of officials to review the list of products reserved for small firms, and its report lists plastic toys and mechanical toys among 91 items it proposed to delete. So when the report was written the list still held them. The committee recorded the investment ceiling for a reserved unit as Rs 60 lakh or Rs 75 lakh. The officials in the small-industry office opposed removing any item, on the ground that reservation had created employment and spread entrepreneurship.
That is a parallel, not a jump. On one side of the border, 404 people worked in a toy plant behind sealed exits. On the other, the law capped what a toy maker could invest in a plant. A Staircase told the Indian half, and A Car Key drew the line on the workforce side.
Electronic toys came off the reserved list in 1998–99, as the next chapter records. The notifications this book read do not fix the day the rest followed, and so it gives none. What came after is easier to date. In February 2020 India raised its basic customs duty on toys from 20 percent to 60 percent, and in March 2023 to 70 percent. A quality control order for toys, issued in 2020, took effect on 1 January 2021, and every import consignment had to pass sample testing.
2023The Economic Survey of 2025–26 reports that between 2014–15 and 2022–23 India’s toy imports fell by 52 percent and its exports rose by 239 percent, and that India went from net importer to net exporter of toys from 2020–21. That is the government describing the result of its own policy. This book read no independent study of what the duty did to prices or to quality, and so it records the success as claimed. The shape matters more than the number: after decades of a wall that kept toy factories small, India put up a different wall, of duties and tests, to keep imports out.
On 1 January 1994 China changed its tax system.
Before it, taxes had been negotiated between the centre and each province. After it, they were sorted into central, local and shared. Value-added tax was shared 75:25. The top rate of enterprise income tax fell from 55 to 33 percent. A national tax service collected the central and shared taxes, and, after a two-year transition, localities lost the power to grant tax breaks. The centre’s share of revenue rose from 22.0 percent in 1993 to 55.7 percent in 1994.
It had to buy consent. The centre guaranteed that no province would receive less than its 1993 revenue, added refunds that would grow each year at 30 percent of the national growth rate of VAT and consumption tax, and promised transfers to poor provinces once its own share reached 60 percent. Revenue had fallen from 30.9 percent of GDP in 1978 to 10.7 percent in 1995, and the centre’s share of what was collected had fallen from two-thirds to a third by 1993. Wang Shaoguang and Hu Angang had argued in a 1993 report that this enfeebled macro-control, and that the centre had to recover its capacity. The centre needed it back, and the 1994 package was the way. Lou Jiwei, who took part in the reforms, called the whole 1993–94 package of price, tax, fiscal, financial, trade and investment reform the most important institutional milestone since reform began. That package also merged the official and swap-market exchange rates, and the yuan went from 5.76 to the dollar in 1993 to 8.62 in 1994, a loss of about a third of its dollar value. Merchandise exports rose from $91.7 billion to $121.0 billion.
Look at what the bargain taught the provinces. In Wang’s account, late in 1993 they learned that 1993, not 1992, would be the base year for the guarantee. Local collection ran 63.3 percent above the year before in October, 87.5 percent in November and 119.8 percent in December, and 28.3 percent over budget for the year. Some localities collected arrears, some leaned on banks to lend firms money to pay tax, and some took 1994 taxes in advance. The centre had lured the provinces into revealing what they could collect, at the price of very large refunds. After the refunds the share the centre could actually use was only about 20 to 30 percent, lower than before the reform. Wang counted the 55.7 percent as a gross figure and judged the centre’s rescue incomplete.
Spending duties stayed where they were. Localities kept pensions, the social burdens of state enterprises and local services, and extrabudgetary funds of 186 billion yuan, 36 percent of budgetary revenue, were left alone. Christine Wong, writing for the World Bank in 2000, found localities leaning on off-budget funds and on land as collateral, and one city, Hengyang, four months behind on paying its officials in 1999.
The economist Wu Jinglian, speaking in 2011 and quoted in China Reform magazine in 2012, described the market system built by the end of the century as half-controlled and half-market, with state monopoly strengthening, and warned that it could degenerate into crony capitalism. He had been one of those who proposed the phrase “socialist market economy” in 1992. One of those who proposed the label was describing, two decades on, a market in which the state kept the controls.
Door D3 · openedOn the imagined night that frames this book, the Finance Secretary’s folder, opened at eleven, ended on a question: where does a state’s money come from, and what does each source cost later? This is one answer. A centre short of revenue took it back and left the localities with the bills, and the localities came to depend on one asset they could sell.
In 1995 local governments’ income from selling land came to about 40 billion yuan.
2021In 2021 it was about 8.7 trillion, roughly 30 percent of local governments’ revenue, against 20 percent in 2012. In nominal yuan that is more than two hundred times the 1995 figure, over twenty-six years. A land-use right had first been auctioned in Shenzhen in 1987, and a 1988 constitutional amendment let the right be transferred, so the market was older than the tax split. What the split changed was who needed it.
Chain C081994 → 2021Planted: the cities that live on land19472026The arithmetic is the lesson, and it is also the trap, so the chapter stops here to say so.
Put the three pieces side by side. An old man’s train licensed what everyone below him already wanted. The factory it licensed was cheap to build and unsafe to work in. The tax split paid for the centre’s recovery by leaving the towns to find their own money. None of this is wrong in a single step. Each is what the next step stands on.
In Delhi, the same budget that unified the rupee had written a tax holiday for a different kind of climb, one that needed no factory at all. That lift is the next chapter.