Why Chinese banks now act like local governments

China’s banking sector is undergoing a profound transformation. At the 2023 Central Financial Work Conference, it was explicitly stated for the first time that financial work must uphold its “political and people-oriented nature” (政治性、人民性). This redefinition significantly reshapes the role of banks: they are no longer regarded as purely profit-driven commercial entities, but as an integral part of the state governance system.

Higher-quality customers and reduced risks

“Political nature” refers to aligning financial activity with national development strategies and directing credit resources toward priority areas identified by the state. The “people-oriented” dimension focuses on advancing public welfare objectives, including safeguarding employment and promoting livelihood initiatives such as the guaranteed completion and delivery of housing projects.

Together, these principles effectively establish a “positive list” for banking operations, fundamentally reshaping how banks are run. Credit extension is no longer a matter of doing whatever is commercially viable; instead, it prioritises projects on the approved list that must be supported, while exercising caution toward — or even excluding — activities outside it.

To fulfil national financial objectives, regulatory authorities have established a range of binding performance indicators, setting specific targets for the growth and allocation of credit in areas such as technological innovation, manufacturing, green finance and inclusive finance. For instance, inclusive lending is subject to a “two-increases” requirement: its growth rate must not fall below the average growth rate of total lending, while the number of borrowing accounts must continue to expand.

To support the implementation of these objectives, the government also provides extensive backing for bank lending activities. Banks receive “whitelists” of qualified high-quality borrowers to facilitate priority approval. They are granted access to a wide range of administrative data — including business registration, tax, social security, utilities, real estate, and intellectual property records — which enables more precise credit risk modelling and customer profiling.

In addition, risk-sharing funds and guarantee institutions are established to reduce the credit risk borne by banks. In some localities, Housing and Urban-Rural Development authorities even share real-time construction site camera feeds with banks, allowing joint monitoring of project progress to ensure timely housing delivery. Taken together, these performance targets and state-bank coordination mechanisms increasingly make banks resemble quasi-governmental institutions.

Homogeneous competition in credit sector

As the allocation of credit is increasingly guided by state priorities and policy-driven whitelists, banks are competing within an increasingly narrow pool of eligible borrowers. Financing needs outside these priority areas receive far less attention, leaving banks to vie for many of the same clients. The result is a highly homogeneous competitive landscape, marked by several distinct features.

First, competition on lending rates has become increasingly intense. Although the Loan Prime Rate (LPR) has remained broadly stable, actual borrowing costs have continued to fall as banks offer ever larger discounts below the benchmark rate. In some cases, branches under pressure to meet technology-lending targets have extended loans at rates below their own funding costs, effectively subordinating commercial considerations to policy objectives.

Second, leading banks are utilising data shared by the government to build intelligent risk control and marketing systems, and to draw up their own whitelists. As soon as an enterprise recognised as “specialised, sophisticated, distinctive and innovative” meets pre-set criteria — in terms of tax payments, electricity consumption and so on — the system triggers an alert, prompting account managers to contact the client quickly with a pre-approved credit facility proposal. Some banks also monitor corporate cash flow movements via their back-office systems and leverage this information to poach high-quality clients from other banks.

Third, there is a shift from single-loan transactions to “full lifecycle” management. Some banks are willing to secure a client even if an individual loan is unprofitable. The rationale is that once a partnership is established, it can be extended to engage a range of high-value services, such as IPO underwriting, wealth management for high-level executives, and payroll services for employees. Some banks have also introduced “equity option loans”, which offer extremely low interest rates and acquire options to purchase future equity in a company. The idea is to offset short-term losses with long-term returns.

Fourth, there is also a shift from simply selling products to “building ecosystems” to retain customers through integrated systems. Seeking to get ahead in the fintech racetrack, banks have begun to put digital tools in place for businesses free of charge. These include systems for enterprise resource planning (ERP), financial management, human resource management and supply chain management. Once a company starts using these tools, its day-to-day operational data, personnel information and transaction records get accumulated within the bank’s system. This effectively makes the company deeply bound to the bank since the cost of migrating everything to a different bank would be forbiddingly high.

Notably, the core driver of the whole competition is no longer profit growth, but rather meeting performance targets — whether it be the growth rate of inclusive loans or the proportion of technology-related lending. These are essentially mission metrics. When business objectives shift from “profit maximisation” to “accomplishing specific missions”, interest rates lose their market function of regulating the supply and demand of capital.

This non-market-based competitive model is giving rise to a series of adverse consequences. The first among them is the misallocation of resources. Enterprises with genuine innovative potential but not on the approved lists are unable to access low-cost funding. They instead find themselves in a financing crisis amid market competition. 

The second consequence is signal distortion. As interest rates fail to accurately reflect asset risk levels, it becomes difficult for investors and policymakers to identify market bubbles and potential crises through price signals. The third consequence is a decline in capabilities. With banks focusing more on securing policy resources, their ability to independently assess corporate operational competence and technological prospects is gradually weakening.

Banks as ‘touters’ targeting businesses

The business model currently employed by banks for inclusive finance and fintech bears a striking resemblance to how China’s local governments strove to attract investments many years ago.

How do they compete for clients? In the past, mayors would lead delegations to more developed regions to attract investment, offering tax incentives such as “three exemptions and two reductions”. Today, bank presidents lead teams into industrial parks, offering incentives such as deep discounts below the LPR, loans without collateral or guarantees, and interest subsidies for first-time borrowers. Some banks even provide ERP systems and financial software free of charge. By effectively paying companies to become customers, banks are increasingly adopting the same approach that local governments once used when offering land and factory buildings to attract investment.

The same logic is evident in how targets are selected. In the past, local governments designated industrial parks and development zones, with policy incentives largely reserved for firms operating within them. Today, banks focus on whitelists of “specialised, sophisticated, distinctive and innovative” enterprises and “high-tech enterprises”. In effect, this is a financial version of the industrial park model. Companies on these lists enjoy abundant access to financing, while those outside them are often filtered out by banks’ screening mechanisms, regardless of their growth potential.

Even more striking is how a tactic once associated with local governments — “shut the door and beat the dog” — has re-emerged in the banking sector in a new guise: so-called “customer-base harvesting”. This typically takes place through three main channels:

First, the bank provides systems free of charge, so that it gets the entirety of a company’s operational data and management processes under its thumb. For the company to switch to another bank, it would have to rebuild the whole nexus of systems from scratch and migrate all the data. The cost of doing so would be so prohibitive that the company would have to stay on and keep working with the original bank. 

Second, banks may require clients to take up additional services, thereby increasing the effective cost of financing. For example, after obtaining a loan at an exceptionally low interest rate of 3.0%, a company may be encouraged to purchase a package of related products, such as payroll insurance or employer’s liability insurance. Banks may also charge various fees under different headings, including assessment, consultancy, or advisory fees.

An even more common practice is to link loans with deposit requirements. For instance, for a loan of 10 million RMB (US$1.48 million), a bank may require the borrower to keep 3 million RMB on deposit. Given the low returns on deposits, this arrangement can significantly raise the company’s effective financing cost.

Third, preferential terms are withdrawn upon loan renewal. Many low-interest loans only offer preferential rates for the first year. When the loan is renewed the following year, banks may cite changes in risk assessments to withdraw the preferential rate and reprice the loan at a spread above the LPR.

Downside of expansion of scale

When banks, much like local governments once did in their pursuit of GDP growth, focus narrowly on expanding inclusive and technology-related lending, they may be meeting today's targets at the expense of tomorrow's risks. Over time, this approach could lead to a substantial build-up of non-performing loans.

One consequence is that banks’ ability to price risk is being undermined, weakening their resilience to future losses. Under normal market conditions, higher-risk borrowers are charged higher interest rates, with the additional margin helping to absorb potential losses from future defaults. Today, however, banks are driving lending rates ever lower in order to meet policy targets. As interest margins are squeezed, profits available for loss provisioning are correspondingly reduced. If non-performing loans rise sharply, banks may find it difficult to recover their principal. In such a scenario, even the modest profits accumulated in recent years could be quickly wiped out.

Another concern is that the “whitelist” model leads to a high degree of risk homogeneity, which translates to a powder keg of systemic risks. When all banks grant loans based on the same list, the scale of financing for companies on that list swells rapidly. A company that originally required only RMB 10 million in funding may be courted by multiple banks and ultimately obtain RMB 50 million or even more.

Over-financing leads either to reckless corporate expansion and overcapacity, or to idle circulation of funds, disconnected from the real economy. Once the industry’s development cycle takes a downturn, risks will manifest in high concentration among the whitelisted enterprises, and all the banks involved in lending will be impacted. This follows exactly the same logic as seen in the intense explosion of local government debts in the past: high-quality entities ultimately became a source of risk due to over-financing.

Finally, headline growth can mask underlying weaknesses, leaving the credit system exposed to future stress. Pressure to meet administrative targets has fuelled the rise of intermediaries that help unqualified firms qualify for whitelists by manipulating data and credentials. The result is that banks increasingly fill their balance sheets with assets that appear sound on paper but are of questionable quality in practice.

As medium- and long-term loans with maturities of three to five years begin to enter their peak repayment period, the economic environment may prove less supportive than expected. If conditions deteriorate, firms that have relied on intermediaries and sustained themselves through continuous refinancing will see their weaknesses quickly exposed. Any resulting stress could then spread through the broader credit system.

It should be noted that there is a fundamental difference between banks’ non-performing loans and local government debts. Local governments have nifty tools to fall back on, such as taxing power and land reserves. They can also resolve debt issues through measures like loan extensions and debt swaps. For banks, however, the only risk buffer is their capital. Should non-performing loans erupt on a large scale, their capacity for credit expansion would be directly exhausted, resulting in financial shocks and economic costs that would be far more difficult to bear.

Fortunately, positive signs have recently emerged. As this article was being written, the National Financial Regulatory Administration issued a notice abolishing the mandatory growth targets for inclusive loans to micro and small businesses that had been in place for many years. This adjustment may suggest that regulators have begun to recognise the risks associated with the increasing administrative control over banking operations and are taking steps to correct course.

This article was first published in Lianhe Zaobao as “中国银行业为何越来越像地方政府？”.

Related: Why desperate local governments fund corporate fantasies | Shadow banking under fire: AVIC Trust’s collapse