Local Government Debt: Adding Pressure to China's Economic Slowdown

Local Governments in China have been hit hard by the real estate crisis that has unfolded since 2021: they are heavily reliant on land-related revenues to cover their expenses, which include the provision of basic social services as well as the funding of local investment projects. After three decades benefiting from buoyant property markets, the sudden decline of their land-related revenues is a major revenue shock.

The question of their heavy debt potentially threatens the financial and social stability of the country. The recent 13th plenum of the Party held in July 2024 announced a fiscal reform as one of its priorities, so as to " establish a fiscal relationship between the central and local governments that features well-defined powers and responsibilities and the appropriate allocation of resources, with an optimum balance between regions".

Local Government financial weakness: too dangerous to be ignored

In China, government finance is a study in contrast. On the one hand, the debt of the Central Government (CG) is very conservatively managed, and kept at a relatively low level compared to other major economies and to most developing countries. On the other hand Local Governments (LG) are structurally in deficit, and have to use all kinds of tricks and loopholes to make ends meet.

The official figures for China’s overall public debt put it at 71 tr RMB, or 69% of GDP at the end of 2023, consisting of CG debt for 30 Tr RMB (24% of GDP), and of LG debt for 41 Tr RMB (32% of GDP). The first category includes sovereign bonds issued by the Ministry of Finance of the PRC, while the second consists of bonds issued directly by provincial or municipal governments. The balance (13%) is owed by various Government funds. If things were that simple, the Chinese government would be one of the least indebted ones among major countries.

The real situation is more complex, however, as there are massive contingent (or off-balance sheet) liabilities at the local level that must be taken into account as well. These liabilities result from debts issued by entities called Local Government Financing Vehicles (LGFVs), which have been set up by LGs for the very purpose of working around various limitations imposed on their capacity to issue debt directly (see infra for more on the origin of this).

The IMF estimates the total amount of LGFV debts at 60 Tr RMB, or 48% of GDP in 2023, coming from 13% in 2014. A more accurate estimate of China’s government debt ratio (“augmented” ratio in IMF parlance) is therefore 117%, rather than the official 69%. Arguably, this still does take China’s government into dangerous territory: as of end of 2022, China’s ratio was higher than India’s (85%), close to the US’s (121%), but lower than Italy’s (144%) or Japan’s (267%) (see IMF). In addition, the government in China owns more assets than almost any other government in the world due to its substantial FX reserves and to its large public sector. There is still room for the Chinese government to borrow more if needed. Nevertheless, the rapid rise in government debt and the potential costs of bailing out of LGs were explicitly mentioned in the decisions by Moody’s and Fitch to put China’s sovereign rating on a negative watch in respectively December 2023 and April 2024, indicating that they are worthy of attention.

LGFVs however do not conduct any business by themselves, contrary to SOEs, and are mainly holding companies and funding vehicles set-up by government entities. China is not a federal state, moreover, and local authorities hold their mandates and authority from the central State: events of defaults by provinces or important cities would be immediately interpreted by financial markets and the public at large as a sign of financial weakness affecting the whole state apparatus. This is a risk Chinese authorities would certainly not want to take, as it could trigger a massive confidence crisis among the public, affecting not only LGs but also banks, which are State-owned in China, and -at least for the larger ones- benefit from the perception that the Government will not let them fail.

There are in addition several specific factors which make LG debt potentially toxic and expose the Chinese authorities to unacceptable risks

- LGs are responsible for the funding of the majority of local infrastructure and construction projects, and their difficulty or incapacity to do so would of course drag further down economic growth. But most importantly, LGs are also responsible for the provision of most basic public services, such as education, health, or retirement benefits, so that any situation affecting their capacity to meet their obligations could have very severe consequences, for the population as well as for the Government.
 
-The banking sector is heavily exposed to LGFV debt. Fitch estimates Chinese banks’ exposure to LGFVs  at around 15% of their balance sheet. To put things in perspective, this is more than banks’ direct loans to property developers, which is around 4% of their total loans. Furthermore, the weakest banks are also the most exposed ones, as small regional banks tend to be less well capitalized and more exposed to LGFVs than large national banks. They would be particularly vulnerable to financial difficulties affecting LGFVs and LGs. According to an IMF report, “Even a small LGFV default rate of 5 percent would be equivalent to a roughly 75 percent increase in banking system NPLs. The impact would likely be concentrated on smaller local banks, which have weaker buffers and are believed to be most exposed to LGFVs with limited state support.”. The financial weakness of LGFVs is therefore a potential source of systemic risk.

How did we get there:  a road of thirty year ending in an impasse

Prior to the economic reforms of 78-79, the bulk of government revenues came from the profits of State-Owned Enterprises (SOEs), as well as from various taxes. The CG received a variable part of revenues, according to a formula renegotiated frequently, but which left it with enough resources to cover its own expenses as well as to reallocate resources from rich provinces to poorer ones. This system broke down with the reforms, as the weight of the state sector decreased, and the transformation of the economy resulted in many activities falling outside of the traditional tax net. This caused a decline of the overall government share in the GDP, with the CG being much more affected than LGs. This situation was of course unacceptable for the central authorities, afraid of losing their capacity to manage the economy and their control over the richer provinces.

The overall fiscal system was then thoroughly reformed, with in particular the introduction of a tax on profits for corporates and of a VAT, which soon became the main source of government revenues. New sharing arrangements were put in place in 1994, with fixed allocation keys to govern the collection and sharing of revenues between the various levels of government, and LGs made responsible for a much larger part of overall government expenses than their share of global revenues. LGs’ budgets were therefore systematically unbalanced, with the difference supposed to be made up by transfers from the CG to LGs (this is still the case today: the LGs get around 50% of the revenues, but are responsible for more than 80% of the expenses. See here for more details). LGs were in addition forbidden to raise new types of taxes or to borrow directly by themselves, so as to avoid any risk of fiscal profligacy at local level.

This 1994 system has worked well for a long time, and many of its key structural features are still in place today. It has however progressively become out of whack by various shocks and changes in the economic environment, leading to the current crisis.

Over time the gap between revenues and expenses increased, even taking into account central transfers, as the Provinces were made responsible for the many social programs which were introduced in the 2000s in order to address the issues of rural poverty and growing inequalities. Provinces soon found themselves short of the resources they needed in order to feed their economic development. Local leaders were under tremendous pressure to boost GDP growth, which became the main criteria by which their performance was assessed. The most expedient way for them to achieve this was to promote investment in infrastructure or housing, which have a strong stimulative impact on many sectors upstream and downstream, but they could not go fast enough by relying on transfers from the center, and were limited by their incapacity to borrow.

LGs could capitalize on this for instance by purchasing land from villages in suburban areas at a low price, transforming and requalifying it for other uses, and selling it to property developers or industrial users at a higher price. But they also had to invest in order to transform the land before they could resell it, and the transformation cycle takes long. They needed financing, and this is where LGFVs come into play: the basic mechanism is that LGs set up a LGFV, and transfer land usage rights to it. The LGFV can then use these rights as a collateral in order to raise money from banks (and later other types of borrowers as well), thus circumventing the prohibition on direct borrowings by LGs. LGs can then invest the funds it has raised, and the game is on, with the LGFVs as enablers and accelerators.

It turned out to be a very successful and long-running game: it is estimated there are at least 12,000 active LGFVs today, and as we saw above their total liabilities are estimated at 70 Tr RMB. They provided LGs with the funds they needed to fuel the investment-led economic growth of the last few decades in China, with the same basic mechanism being applied to all kinds of infrastructure projects, such as highways, railways, airports, even hospitals and so on.

LGs could raise funds and keep the liabilities off-budget, but this did not happen under the radar screen of the central authorities, who knew very well what was going on but never tried very hard to stop it. They even at times encouraged the use of LGFVs: the massive 2008-2009 stimulus plan for instance was financed mostly by having the banking system lend to LGFVs. It has even been argued that back in the nineties when the new fiscal deal was negotiated, turning a blind eye to semi-legal borrowing at local level was part of "grand bargain" between Central and Local authorities, to get the latter to consent to sharing of revenues and expenses heavily skewed against them.

The banks were enthusiastic players in the game, as they saw lending to LGFVs as an opportunity to grow their balance-sheet and revenues by taking on what they saw as quasi-sovereign risk, a much safer proposition in their eyes than the risky business of lending for instance to privately-owned businesses which nobody would support in case of trouble. They did not bother analyzing in depth whether the assets of the LGFVs they lent to would generate enough cash flow to repay the loans, and relied primarily on shareholders’ support. The banking regulator did worry at times about the uncontrolled growth of these loans, and introduced various measures to try to curb it, but every new measure was met by a new "financial innovation" by the banks to work around it so that the game could go on (this is one of the drivers of the widespread "shadow-banking" business in China).

A crash in slow motion

Such unbridled enthusiasm was bound to lead to excesses. LGs launched more and more projects and became over-dependent on land-based revenues and taxes. The share of these revenues in their overall revenues (including other taxes, and fiscal transfers from the center) was close at its peak in 2021 to 40%. When the real estate market turned cool, these revenues plummeted (minus 30% in 2022 over 2021, and the slide still continues, albeit at a slower pace). Part of the shortfall for the LGs was compensated by an increase in transfers from the CG, but obviously this is no more than transferring the problem from one pocket to another, and even with increased transfers, some provinces came under tremendous financial pressure.

In March 2024, the Government took the unusual step of publicly instructing several provinces to scrap some high-profile investment projects due to their excessive debt levels. The overall operating cash flow of the LGFVs is negative, which means that as a whole they need to borrow just in order to pay the interest on their existing debt, or paying back the part of their debt that reaches maturity (see IMF chapter IV 2023).

How to defuse a ticking bomb

The Chinese authorities did not wait until today to identify and try to address the dangers of an uncontrolled growth of local liabilities, and there have been several attempts in the last decade to put things back in order. The task has however so far proven almost impossible to achieve.

Part of the problem is that the exact size of these liabilities has proven very difficult to assess, due to a lack of transparency at local level and the sheer number of debt-issuing entities: there are almost 12,000 LGFVs, some of them buried deep down in the lower layers of the local bureaucracy.  The existence of LGFVs was itself not a secret, but it took two national surveys conducted by the National Audit Office in 2011 and 2013 for the extent of the explosion in LG undisclosed liabilities following the bank-financed investment boom of 2008-2009 to become clear. The issue of so-called “hidden debts” emerged and took centerstage during the third plenum of the 19th congress in 2013, which called for a full revamping of the fiscal system. A revised budget law was introduced in 2014 and Provinces were for the first time allowed to borrow directly, by issuing bonds (subject to strict quotas). This was meant as a way to solve the hidden-debt problem, as the new debt issued by the Provinces had to be used in priority to refinance part of their LGFV debt. There have been several successive similar “debt swapping” plans since that time but none of them has been entirely successful. In 2019 for instance, the (then) Finance Minister Liu Kun declared  that “China is very serious about hidden debts. I have to admit that there are still some local governments borrowing through financing platforms, which is illegal as it is beyond their statutory [borrowing] limit …. “We are taking strict measures and won’t allow any new cases to emerge.” Obviously, these objectives were not met: as we saw above, LGFV debt has continued to increase from 13% of GDP in 2014 to 48% in 2023.

No easy fix

In theory, the Government could solve the issue at one stroke, by stating that LG debt is indeed sovereign debt, and that the CG will stand behind the LGs. This would immediately increase the official debt-to-GDP ratio of China, but this would not be a disaster: the revised level is still manageable. The CG however has repeatedly stated it will not do so, and there are strong reasons for this.

There is also a strong desire among Chinese leaders to project internally and externally the image of China as a financially and fiscally responsible country, in contrast with the US in particular which they describe as taking advantage of the dominant role of its currency to borrow and spend unsustainably. They also like to point out the regular progress of the country's sovereign rating (now comparable to Japan’s, and higher than some European countries). It helps for this narrative to remain credible that the Government balance-sheet is kept as clean as possible, even though this has become largely cosmetic as the question of LG debt is well known by professional investors and rating agencies (and already reflected in their rating decisions).

The Government needs to handle this issue very cautiously: if it goes too fast, and lets too many LGFVs default, it takes the risk of triggering a major confidence crisis and of aggravating the current economic difficulties. On the other hand, doing nothing is not an option, as the problems will just become larger and larger, and may become unmanageable. There is not yet on the table a detailed and comprehensive action plan, although the third plenum of the 20th congress of the Party held in July has announced one was coming.

The Government has nevertheless already taken a number of steps in the past few years to mitigate the problems. The most visible one is the refinancing of LGFV liabilities by provincial bonds, as mentioned above. This transforms previously off-balance sheet liabilities into on-balance sheet, which also lowers their cost for the issuer (bonds are considered safer than LGFV loans by investors as there is no doubt about their sovereign quality). It amounts to a partial bailout of the LG debt, and will not be an acceptable solution for the totality of the debt involved, for the reasons pointed out above. LGs have issued more than 40 tr RMB of bonds since they were allowed to do so in 2014. This is undoubtedly a very large amount but it is still less than the roughly remaining 70 tr of LGFV liabilities in the form of loans (and for the ones which will continue to be booked, in spite of government instructions…).

An alternative is of course to restructure the loans taken by the LGFVs. They are today very widespread, with an acceleration since 2022. In a recent and well publicized case, “a troubled local government financing vehicle (LGFV) in Guizhou, one of China’s most indebted provinces, has secured approval from banks to delay repayment on 15.6 billion yuan ($2.3 billion) in loans by uniformly changing their term to 20 years”. Restructurings avoid a formal default, they are the most common tool to help spread the recognition of losses over time, and they keep alive the notion that the loans will be repaid at some point in time. This in turn helps the lending banks to avoid declaring these loans as non-performing, which would require them to pass provisions.

Officials also talk about attracting private investors to work under some kind of Public Private Partnership schemes, but this seems also difficult as many of the assets held by LGFVs are not commercially viable (see here for practical examples)

LGs are also invited to tap some other assets resources available to them, such as the State-Owned Enterprises (SOEs) they have under their jurisdiction. The largest and profitable SOEs tend to be managed at the central level, but there are nevertheless many good and viable ones at local level. Local SOEs can help LGs in at least three different ways. They could be directed to increase their dividend pay-out, but given that the total dividends paid by SOEs amounted to 500 bn RMB in 2022 (including the central ones), and that their dividend payout ratio is already close to 30%, this would be helpful but would not make a big dent in the 70 trillion debt mountain. LGs could also sell down their stakes in some SOEs in order to raise funds and repay their debt. This already happened in the past, with for instance Guizhou selling in 2019 a small part of its very valuable majority stake in the famous Moutai company. It could happen elsewhere, but it would remain to be seen whether local authorities would accept to part with some of their most profitable assets, and whether politically acceptable buyers for these assets could be found in significant numbers. A perhaps more promising avenue (from the standpoint of the LGs) is to have their SOEs participate in the restructuring of the ailing LGFVs. This is similar to what is done already in respect of the property market, where SOEs and other local entities have been called upon to take over projects which private developers could not complete.

The government is also taking some preemptive measures to protect the banking sector. It is well aware of the fact that many small or medium size regional banks may not be strong enough to withstand the impact of the combined real estate and local government crises. Having been surprised by the failure of four banks in 2019 and 2020, the regulators are now better prepared for banking emergencies. They have launched a reform plan for the country’s rural banks, which are the weakest link in the banking system and represent 14% of total banking assets. Already, 36 small banks in Liaoning were ordered to be absorbed by a larger one in June 2024, and two other banks were closed (see here).

None of the above is easy or painless, and the ongoing weakness of the real estate market does not help. But what has been done so far is more about avoiding an explosion and sharing the costs of the mess between various categories of stakeholders than about solving the underlying core issues. There are even more formidable tasks at hand.

An obvious issue is about intergovernmental relations: In a country the size and diversity of China, an efficient set of rules and incentives to govern the way fiscal responsibilities are shared between the various levels of government is an essential tool for economic development. The system put in place in 1994 has been very successful for a long time, but it is now broken, leading to an unsustainable accumulation of debt at local level as LGs scramble to find the resources to continue to finance an unsustainable investment binge. The real estate crisis has laid bare the problems and made them worse, but the situation of the LG had started deteriorating from the time they were assigned the burden of financing the stimulus plan in 2008-2009.

It will however be extremely difficult to reform the current fiscal system. The technical issues are challenging, as the economy is today much more complex and sophisticated than in 1994, and there are risks: fiscal mechanisms can work in very unpredictable ways, with unintended consequences. But the main difficulty is political: the question of how to share revenues and expenses cuts through the heart of the political and financial interests of many different powerful bureaucracies at both national and local levels. It took a “grand bargain” to settle the issue in 1994, and the discussions which started in 2013 have led to nowhere yet. It is unlikely that things will be easier this time round.

With the property market still depressed, and financial distress rampant among LGFVs, China is facing a complicated combination of crises. The two crises are of course interconnected, and the principles used to manage them are similar: avoid a large-scale bailout, encourage local restructurings whenever possible, and let the banks and other financial intermediaries take the share of the costs they can afford without risking a financial systemic crisis. The authorities have enough resources and room to maneuver so that they may succeed in spreading the costs of these crises over time and avoid a cataclysmic breakdown.

It is impossible at this point in time to quantify these costs. We can estimate (with some difficulties!) how much has been invested, but there are assets in front of these liabilities and their value is almost impossible to assess. Many of the assets built during the boom years, such as empty or unfinished apartment blocks nobody will ever move in are clearly destroying values. But what is the value of a highway, underutilized today, but which has the merit of connecting a remote area with the rest of the country?

The risk of a full-blown systemic banking crisis appears remote in the short term: like most debts in China, LG’s debts (loans and bonds) are almost exclusively denominated in local currency, and held by domestic banks, which the government controls. The authorities can encourage the banks and financial institutions to restructure and reschedule their loans to LGs and LGFVs, and at the same time manipulate the rules according to which provisions have to be made. But the costs are there, and they will have to be absorbed. They will be a further drag on economic growth, at a time when it is already slowing down, and will add to the financial constraints that China will increasingly feel.

Copyright image : Dale DE LA REY / AFP