China’s local government debt is the accumulated borrowing of provinces, cities, and counties, most of it built up off the books through state-owned companies known as local government financing vehicles. By 2024, debt held by those vehicles alone reached an estimated 66 trillion yuan, roughly $9.5 trillion, or about half of China’s annual economic output.1 Stack the on-budget bonds on top and many provinces now spend more servicing debt than they do on schools or hospitals. Beijing approved a 12 trillion yuan cleanup program in late 2024, but the fiscal mismatch that created the problem is still there.
Where the Debt Came From
The story starts in 1994. Premier Zhu Rongji rewrote the country’s fiscal rules through what became known as the tax-assignment reform. Before it, local governments negotiated their contributions to Beijing case by case, and the central share of national revenue had slipped to about 22 percent. The new rules sent the most lucrative taxes upward automatically. Central revenue more than doubled in the first year, and Beijing’s share of total fiscal revenue jumped to 56 percent.
Spending responsibilities didn’t move. Provinces and cities were still on the hook for education, healthcare, infrastructure, social security, and day-to-day administration. Beijing collected the money; localities paid the bills. And local governments were legally barred from running deficits or borrowing directly, so there was no clean way to close the gap. The prohibition didn’t remove the need for capital. It pushed the borrowing underground.
How Local Governments Actually Borrowed
The workaround was the local government financing vehicle, or LGFV. Starting in the late 1990s and accelerating after the 2008 global financial crisis, cities set up state-owned companies whose only job was to borrow for public works. Because these were legally separate corporations, their debts stayed off government balance sheets and outside any borrowing limits. Roads, bridges, subway lines, industrial parks, and whole new districts got built this way.
The vehicles raise money through a mix of bank loans and shadow banking products: trust loans, wealth management products, and other instruments issued outside the traditional banking channel. Shadow financing typically carries shorter repayment windows and higher interest rates than a government could get borrowing in its own name. What made lenders willing was an unspoken assumption that if a vehicle couldn’t pay, the local government behind it would. That implicit guarantee let vehicles with weak cash flow keep raising money well beyond what their own books could support.
The assets pledged behind the loans are often infrastructure that produces little or no revenue: a highway to a district that never developed, an industrial park half empty. Lenders didn’t scrutinize project economics when they believed the state stood behind the loan. Cheap credit funded ambitious projects, which justified more borrowing, which created more implicit government liabilities. Total LGFV debt hit 66 trillion yuan in 2024, up from 57 trillion the year before.
Why Land Sales Made the System Work
All land in China is state-owned, and local governments control the right to lease it for development. Selling long-term use rights to property developers became the financial engine of the whole arrangement. At the 2021 peak, land transfer revenue brought in 8.7 trillion yuan nationwide, about 38 percent of total local government income that year.
The money flowed in two ways. Cities auctioned land rights and used the proceeds to fund operations and service debt. They also transferred land to their financing vehicles at appraised values, which inflated those vehicles’ balance sheets and let them borrow more. As long as developers kept bidding and prices kept rising, the cycle sustained itself.
The property downturn that began in 2021 broke it. By 2025, land transfer revenue had fallen to roughly 4.15 trillion yuan, less than half the 2021 peak. Land-related income dropped from 38 percent of local revenue in 2021 to 27 percent by 2023. When land values fall, the damage compounds. Collateral behind existing LGFV loans loses value, which can trigger demands for more security or early repayment. Developers stop bidding, so future revenue dries up too. Municipalities in that squeeze face a hard choice: cut services, find new revenue, or default. Most are cutting services while waiting for Beijing to help with the rest.
What Beijing Has Tried Before
In 2014, Beijing tried to bring local borrowing into the open. The revised Budget Law, effective January 1, 2015, let provincial-level governments issue bonds directly for the first time, under quotas approved by the State Council. Prefectures and lower levels still can’t borrow on their own and have to go through their province. Bonds come in two flavors: general bonds for projects without their own revenue, repaid from tax income, and special purpose bonds for projects expected to earn income, repaid from that income.
The State Council’s rules that same year declared that financing vehicles had no authority to borrow on behalf of governments and ordered vehicles whose only business was government borrowing to be shut down. In practice, LGFV borrowing kept growing. Officials found workarounds, and enforcing the ban across thousands of vehicles proved far harder than writing it.
The Ministry of Finance now manages local borrowing through an annual quota system. The national legislature sets a ceiling; the Ministry allocates it by province based on financial health and project pipeline. A risk classification sorts regions by debt-to-revenue ratios, and high-risk provinces face restrictions or a full freeze on new borrowing. Officials who exceed their quotas can lose their positions, and Beijing can withhold fiscal transfers from provinces that don’t comply.
Underpinning all of this is an official no-bailout principle. The 2014 rules stated Beijing would not rescue local governments that borrowed themselves into insolvency, the point being to curb moral hazard. In practice the principle is more aspirational than absolute. Beijing hasn’t directly bailed out a province, but debt-swap programs and special bond allocations function as indirect rescues. The central government keeps enough ambiguity in the arrangement that local officials can’t treat borrowing as risk-free.
The 2024–2028 Resolution Program
In November 2024, the Standing Committee of the National People’s Congress approved a 12 trillion yuan program to convert hidden local debt into official, lower-cost obligations. The stated goal is to shrink hidden debt from 14.3 trillion yuan to 2.3 trillion yuan by the end of 2028.
The program has three pieces. The ceiling on special local government debt was raised by 6 trillion yuan in one move, from 29.52 trillion to 35.52 trillion, giving provinces headroom to issue new bonds and retire hidden obligations. Starting in 2024, 800 billion yuan a year from new special purpose bond issuances is earmarked for replacing hidden debt, totaling 4 trillion yuan over five years. And Beijing committed to guaranteeing repayment of 2 trillion yuan in hidden debt originating from shantytown redevelopment projects.
Mechanically it works like earlier swaps but at a larger scale. High-interest, short-term hidden debt gets replaced with official bonds at lower rates and longer maturities. Commercial banks holding the original LGFV loans accept the new bonds; they earn less interest but carry less default risk because the debt now sits on the government’s formal balance sheet. Finance Minister Lan Fo’an estimated the program would save local governments about 600 billion yuan in interest costs over five years.
The swap buys time. It does not erase the obligation. Converting hidden debt to official debt makes it cheaper and more transparent, but unless local governments grow revenue or cut spending, the burden shrinks only in servicing cost, not in size. The program also requires localities to stop creating new hidden debt while they work through the old, a condition that has proven difficult to enforce in past cleanup rounds.
What the Strain Looks Like on the Ground
The pressure is no longer abstract. Some cities have cut or halted bus and subway service. Civil servant salary delays have been reported across multiple provinces. In Shandong, officials described on social media being paid only one month’s wages per quarter under what local administrators called a “guarantee four, strive for six” policy. In Shenzhen, one of China’s wealthiest cities, some civil servants saw salaries cut by 25 percent and bonuses eliminated. In rural Guangdong, monthly bonuses of 1,000 yuan simply stopped.
Researchers have described wage arrears as systematic and widespread, unlikely to resolve quickly. The knock-on effects matter. When local officials go unpaid, the incentive to supplement income through corruption rises, and administrative fines on citizens and businesses tend to climb as governments hunt for revenue.
Longer-term obligations are also being pulled into the squeeze. Local governments in some provinces have diverted funds from individual social security accounts, money meant for workers’ future pensions, to cover payments to current retirees. One study estimated the resulting social security debt could exceed 8 percent of GDP if the practice continues unchecked. The problem is worst in older industrial provinces with shrinking workforces and growing retiree rolls. Wealthier coastal provinces with younger populations have stayed solvent.
For a system where government employment and state-funded services carry much of the weight of social stability, cracks of this kind matter beyond the balance sheet. The resolution program is as much about preserving public trust as it is about cleaning up the books.