Vanke vs Baoneng: The Takeover Defense of Last Resort Was the Government

The last move in China's most famous hostile takeover was made by an insurance regulator. On 24 February 2017 the China Insurance Regulatory Commission banned Yao Zhenhua from the insurance industry for ten years, fined Foresea Life Insurance 800,000 yuan and fined six of its senior executives 560,000 yuan between them. The findings included supplying false information about the source of a November 2015 capital increase and misusing insurance funds, including holding more than 30% of assets in stocks.³

Yao controlled Baoneng Group. Baoneng, through Foresea Life and an industrial company called Shenzhen Jushenghua, held 25.4% of China Vanke, the country's largest residential developer, and had been trying since 2015 to convert that into control of the board.²

Every founder who has read a hostile takeover guide knows the defense menu: a poison pill, a staggered board, a white knight, a litigation track, an investor relations campaign that reminds everyone how much they like the current plan. Vanke used two items from that menu and lost ground on both. The thing that ended the fight was not on the menu, because it is not a corporate instrument. It was a licence, and Vanke was not the party holding it.

That is the interesting part, and also the uncomfortable part. This case gets taught as the moment China discovered hostile takeovers. It is better read as the moment a company discovered that its defense budget had been sitting in a regulator's drawer the whole time.

Wang Shi founded Vanke in 1984 and gave away his claim on it. In the 1988 shareholding reform, 40% of net assets were allocated as employee collective shares. Wang relinquished his personal allocation. By his own account he chose reputation over ownership and expected to run the company on ability instead of a controlling stake.¹⁹

For twenty-seven years that worked, which is why nobody examined it. A founder with no shares and a dispersed register looks like the model of professional management until somebody decides to un-disperse the register.

Read Vanke's announcement of 6 December 2015 with that in mind. It reports that as of 4 December, Jushenghua and Foresea Life together held 2,211,038,918 A shares, 20.008% of the share capital, and had become the largest shareholder. It then explains, twice, that the company has no controlling shareholder and no de facto controller because of its scattered shareholding structure.¹ The paragraph was written to reassure the market that nothing had changed. To anyone with 20 billion yuan and a securities account, it reads as a product description.

The mechanical point is worth stating plainly, because it survives translation into any jurisdiction. Authority that rests on a coalition is priced. Authority that rests on shares is owned. Wang Shi had spent nearly three decades accumulating the first kind, and Baoneng spent five months buying enough of the second kind to make the difference visible.

Baoneng did not fund the raid. Its policyholders and four banks did, in that order.

Foresea Life sold universal life products, an instrument that pairs thin insurance coverage with an investment return, at average yields reported around 4% to 8%.³ Money arrives from the public with a promised return and a maturity attached, and until it is paid out it can buy things. That is float, and it is the cheapest acquisition currency ever legally assembled, with the single qualification that it belongs to someone else and has a date on it.

The second layer was leverage. Of Baoneng's 25.4%, Jushenghua held 18.73% and Foresea Life held 6.67%. Jushenghua acquired 10.34% of Vanke, about 1.14 billion shares, for 21.2 billion yuan: 6.5 billion of its own capital and 14.7 billion raised through nine asset management plans placed with Ping An, Guangfa, China Minsheng and China Construction Bank. The ratio was roughly one yuan of Baoneng money to two yuan of borrowed money, and the carrying cost worked out at 19.83 yuan a share.²

Look at where each participant sat in that structure. The banks took the senior tranche with a fixed return. The policyholders took a yield promise from an insurer. Jushenghua took the residual, which is the slice that captures the upside and also the slice that gets a phone call when the share price falls. Its carrying cost was 19.83 yuan a share, so the structure had a level below which the banks stop being passive investors and start being counterparties.²

The structural weakness is the one nobody notices while the stake is going up. Seven of the nine asset management plans matured at the end of 2017.² A control position assembled on instruments with maturity dates is a lease. Baoneng was the largest shareholder of China's biggest developer on a term that expired, and everyone opposing it eventually learned to read the calendar.

Wang Shi's public response to learning who now owned the largest block of his company was to say that Vanke did not welcome the move, arguing that Baoneng would harm the company's reputation and credibility.⁴

Strip the politeness and the argument is that the largest shareholder is the wrong sort of person to own Vanke. As a description of risk it aged well. Baoneng's money was borrowed against yield promises, its vehicle was an insurer with a supervisor, and within six years Jushenghua carried 87 billion yuan of debt with 40 billion falling due inside a year.¹⁷ Wang Shi read the balance sheet correctly.

As a defense it did nothing at all, because the share register does not have a column for whether the board approves of you. A founder telling his largest shareholder that it lacks the standing to own him is a claim about class, and the market clears on price.

The tell is what Vanke reached for next. Chinese listed companies cannot easily adopt the shareholder rights plan that a Delaware board would have had drafted and sitting in a drawer.¹⁴ So Vanke suspended its A shares on 18 December 2015 and kept them suspended for more than six months, resuming on 4 July 2016, when they immediately fell the 10% daily limit.¹⁴ Freezing the stock stops an accumulation and it also traps every shareholder who wanted out. It is the defense equivalent of pulling the fire alarm to end a meeting.

The real defense was dilution, dressed as an acquisition.

On 17 June 2016 Vanke's board approved buying assets held by SZMC Qianhai International Development, a wholly owned unit of Shenzhen Metro Group, by issuing 2.87 billion new shares at 15.88 yuan each. The consideration was 45.6 billion yuan and the issue price was a 35% discount to the average trading price over the preceding sixty days.⁵ The effect was to hand a state-owned subway operator more than 20% of Vanke and push Baoneng down the register.

A white knight transaction usually gets described as bringing in a partner who shares the board's view of long-term value. What it does arithmetically is issue cheap stock to a friendly holder so that an unfriendly holder owns less of the company. Everyone in the room understands this, which is why the fight over the plan did not come from Baoneng first.

It came from China Resources, the state conglomerate that had been Vanke's largest shareholder until December 2015 and held 15.31% with three board seats. China Resources voted against. Vanke counted the result as seven votes in favour out of ten, because independent director Zhang Liping did not vote, citing a conflict involving Blackstone. China Resources counted the board as eleven members, which puts a two-thirds threshold at eight, and publicly challenged whether the resolution had passed at all.⁶

Two state-linked shareholders and a founder-led board spent that weekend arguing about the denominator. The plan to save Vanke from an outside acquirer diluted the inside acquirer too, and China Resources discovered a principled objection to dilution at exactly the moment dilution reached it.

Baoneng did what a 24% holder does when the board issues stock to someone else. On 24 June 2016 Jushenghua and Foresea Life requisitioned an extraordinary general meeting to dismiss the board, including Wang Shi, and argued in the requisition that Wang had drawn more than 50 million yuan while studying abroad between 2011 and 2014.⁷ Vanke escalated in the other direction, reporting Baoneng's nine asset management plans to the CSRC, the Asset Management Association of China, the Shenzhen Stock Exchange and the CSRC's Shenzhen bureau, alleging that the structures breached disclosure rules.⁸

Neither move decided anything. The Shenzhen Metro share issuance was never completed; Vanke and Shenzhen Metro agreed to terminate the restructuring in December 2016, six months after failing to reach a consensus, with Baoneng and China Resources both opposed.²⁰ The board's own term expired on 27 March 2017 and the re-election simply did not happen; Yu Liang told the market "we do not have a time frame."¹⁵ For three months the directors Baoneng had asked shareholders to remove continued to sit, past the end of their mandate, while the register was rearranged around them.

Vanke did not out-vote Baoneng. Somebody bought the votes Baoneng did not have.

In January 2017 China Resources sold its entire 15.31% to Shenzhen Metro for about 37.2 billion yuan.¹¹ Evergrande, which had accumulated 14.07% during 2016, went next. On 9 June 2017 it disposed of its whole Vanke position for approximately 29,200 million yuan and booked a loss of 7,176 million yuan, which appears in its 2017 results under "Other Losses, Net."¹³

That figure deserves a moment. Evergrande paid roughly 7.18 billion yuan for the privilege of standing on the correct side of a governance dispute, and disclosed it to its own shareholders as a line item. Nobody sells a stake at that loss because the price was right.

By 21 June 2017, Vanke's circular to the Hong Kong exchange recorded the finished position: Shenzhen Metro held 1,689,599,817 A shares, 15.31% of the company, and ten Evergrande subsidiaries holding 1,553,210,974 A shares, 14.07%, had entrusted Shenzhen Metro with their voting rights, proposal rights and rights to attend general meetings.¹² Two days earlier Shenzhen Metro had used that block to propose the resolutions electing a new board.¹²

Wang Shi declined nomination. Yu Liang became chairman when the new board was elected on 30 June 2017.¹⁶ The founder who had spent eighteen months arguing that Vanke should not have a controlling shareholder left the company on the day it acquired one.

Run the counterfactual. Suppose China Resources and Evergrande had held. Baoneng would still have been the largest shareholder, still with a requisition on file, and still funded by an insurance subsidiary. What removed that possibility was upstream of the cap table entirely.

On 3 December 2016, CSRC chairman Liu Shiyu told a meeting of the Asset Management Association of China that asset managers using improperly sourced funds for leveraged buyouts "have been barbarians and ultimately become robbers in the industry," and added "evil monsters" and "poisonous pests" to the description.⁹ A securities regulator workshopping four separate insults in one speech is not a rhetorical failure. It was a price signal, and the market repriced accordingly.

Days later the CIRC suspended Foresea Life from selling universal life products, after a December report from the company showed it had failed to fix problems identified in a May inspection. The stated grounds included exceeding caps on medium and short-term investment products, misleading online promotion, and aggressive competition on settlement interest rates.¹⁰ That closed the tap. Then came 24 February 2017 and the ten-year ban on Yao himself.³

Set the penalties against the position. Foresea Life was fined 800,000 yuan.³ Jushenghua had put 21.2 billion yuan into a single tranche of the Vanke stake.² The fine was never the punishment. The punishment was that the person running the bid could no longer run an insurer, which left the funding engine with an operator disqualified from operating it.

There is no clause you can negotiate that produces this outcome. It is the defense of last resort in the literal sense: it is available only to companies the state has decided to keep, and it is delivered by an institution the target does not control and cannot instruct.

My reading is that Vanke is remembered as a governance victory and was closer to a transfer.

Wang Shi was right about Baoneng's fragility and wrong about what followed from it. Diagnosing that your acquirer is overleveraged tells you the bid may fail. It does not tell you that you are entitled to be the one who survives it, and Vanke's management spent a year arguing entitlement while the actual work was being done by two sellers and a regulator.

The outcome also failed the test that governance is supposed to apply, which is whether the people bearing risk were protected. Yao was barred from the insurance industry and Baoneng still cleared the trade. By December 2019 it had cut its Vanke holding to slightly under 5%, and Caixin calculated its net profit on the share sales at more than 13.7 billion yuan.^16b The raider was disqualified and paid. Evergrande, which had helped hold the line, wrote off 7.18 billion yuan.¹³ Minority holders got a six-month trading halt and a board whose term ran three months past its expiry. If the intervention was designed to protect anyone, the beneficiary was the target's stability rather than its shareholders.

Then there is the epilogue, which the 2017 coverage could not see. Shenzhen Metro won the register and inherited the obligation that comes with it. Vanke has since become a state rescue case: close to 138 billion yuan of cumulative net losses since 2024, a first-half 2026 net loss guided at 12 to 15 billion yuan against 11.95 billion a year earlier, and a July 2026 board slate on which all six non-independent nominees came from Shenzhen's government or its state-owned enterprises, ending the professional-manager system Wang Shi built the company around.¹⁸

So the defense worked and the thing being defended did not survive it. Vanke was kept out of Baoneng's hands and delivered to a shareholder that could genuinely control it, which is what the founder had spent his career arranging for the company not to have. That is the part I would want a founder to sit with. The question in a control fight is rarely whether you can stop this buyer. It is who is standing behind you, what they will want afterwards, and whether you would have accepted those terms if they had been written into a term sheet instead of a rescue.

Vanke's specifics are Chinese. The diagnostics are not.

• Read the raider's funding for its maturity dates, not its size. Baoneng's stake was 25.4% and its exposure was nine plans, four bank lenders and a 1:2 ratio with seven maturities landing at the end of 2017. A stake is a position. A funding stack is a schedule, and schedules are where a bid becomes negotiable.
• Find out which regulator can reach your acquirer. Insurance, banking, exchange listing, foreign investment screening, sector licensing. This is not a plan, because you do not control it. It is the single most important fact about the other side's staying power and most boards discover it late.
• Do the recusal arithmetic in your articles before you need it. Vanke's board split over whether two thirds of eleven directors means eight votes or whether a recusal moves the denominator to ten. Work out today what majority your articles require to issue shares, whether it runs on directors present or directors in office, and what a single conflicted director does to the count.
• Price the white knight in dilution, not gratitude. Vanke's rescue issued 2.87 billion shares at a 35% discount. Before you welcome a friendly holder, calculate the post-issue percentage for every existing shareholder, including the ones you expect to be on your side, and assume they will read the same table.
• Know the expiry date of your board's term. Vanke's expired on 27 March 2017 and the delay bought three months. That worked because nobody forced the point. Whether it works for you depends on your articles and your exchange, and it is worth knowing the answer before a requisition arrives.
• If you own nothing, name your coalition in writing. Wang Shi's authority rested on a register that held together through habit. List the holders whose support you rely on, size their blocks, and ask what each of them would take to sell. If you cannot answer for one of them, that is the one who sells.

One caveat on transplanting any of this. Takeover rules, disclosure thresholds, board procedure and regulatory reach vary sharply by jurisdiction. China's framework runs through the Company Law, the Measures for the Administration of Takeover of Listed Companies, exchange listing rules, and sector supervisors who exercise far more discretion than a Delaware court allows itself. Nothing above is advice on a specific transaction, and a real fight needs qualified counsel in the first week rather than the second month.

Not straightforwardly, which is why this fight took the shape it did. A China Daily news analysis published as trading resumed listed the absence of US-style shareholder rights plans among the market's regulatory gaps, alongside thin disclosure requirements for leveraged acquirers.¹⁴ That absence pushes defense toward three cruder tools: suspending trading, issuing discounted shares to a friendly buyer, and reporting the acquirer to regulators. Vanke used all three. The Japanese comparison is instructive here, because a defense that goes to a shareholder vote produces its own arithmetic problems, as the Kusuri no Aoki poison pill vote shows.

It lost the company and won the trade. Baoneng never got a board seat out of the stake, and Yao Zhenhua's ten-year ban removed the funding engine and his standing in the industry that supplied it.³ On the shares themselves, Caixin put Baoneng's net profit from selling down the Vanke position at more than 13.7 billion yuan by December 2019.^16b The wider group did not survive the decade in the same shape: by 2021 Jushenghua carried 87 billion yuan of debt, with 40 billion due within a year.¹⁷ Whether that counts as losing depends on which balance sheet you look at, which is generally true of raiders and almost never true of the boards that fight them.

Two filings in this story did more work than everything either side said in public.

The first is Vanke's disclosure of 6 December 2015, which told the market that the company had no controlling shareholder and no de facto controller.¹ The second is the circular of 21 June 2017, which recorded that one state shareholder held 15.31% and held the voting rights over another 14.07%.¹² Eighteen months and one sentence apart.

Open your articles of association this afternoon and find two numbers: the majority required to issue new shares, and the expiry date of your current board's term. If either answer takes more than ten minutes to establish, you have located the part of your defense that nobody has ever tested.