Zhenjiang, Trump and China's still-expanding local government debt problem
Zhenjiang（镇江）City in Jiangsu province is in a rough spot. Officially, it has a debt of CNY 70bn (some USD 10bn), but unofficially the number is well over five times that, at least CNY 400bn (USD 60bn). We can see that figure on the balance sheets of the city’s local government financing entities (LGFEs). But on top of that there are other future fees owed on projects and miscellaneous bills owed by government departments. This for a city with general tax revenues of only CNY 30bn (4bn), and land sales of CNY 10-20bn (USD 1.4-2.8bn) a year (of which only some 50% makes it back into government coffers). The LGFEs build all the infrastructure - and might have some toll-road revenues - but generally they rely upon the municipal coffers to repay their debts. In other words, the city is not really a going concern.
After what appears to be a failed attempt to restructure that debt with the help of China Development Bank (CDB), Caixin went digging (here for English and here for Chinese). (There’s a lovely moment when the reporter visits an office skyscraper in the New Area of Zhenjiang, to find only one floor occupied…by a LGFE!) The story of Zhenjiang is depressing. It’s a problem thats repeated around much of Tier 3/4 City China. And its illustrative of a problem that Beijing just cannot fix, however many times it tells itself it has. And, in likelihood, local government debt is going to get worse in 2019 as the infrastructure stimulus ramps up.
You have to feel some sympathy for Zhenjiang’s current hapless mayor. His predecessors loaded up the credit card and then usually, after hitting their kaohe zhibiao (performance points) (and never, ever paying off seniors’ with project skim), got helicoptered into the provincial leadership. He’s left holding the bill. In late 2018, the city cooked up a scheme with local CDB officials to combine the debts of a handful of LGFEs into one entity, and then get CDB to lend CNY 20bn, alongside some other banks’ CNY 20-30bn. Thus, some CNY 40-50bn (USD 6-7bn) of short-term, high-cost shadow debt could be refinanced at lower rates and at longer terms. A classic “buy some time” strategy, but magic nonetheless - postponing explosions is, after all, the job of every local official in China.
News about this “debt swap” leaked before Chinese New Year, and we all got excited (or even more cynical) about how CDB was (apparently) being set up as a new LGFE savior. (Many were excited since while Beijing has ruled out saving local governments again, this looked like just that. Others were cynical since CDB’s balance sheet would only ever be able to absorb CNY 2-3tn a year max, unless it really started tapping its People’s Bank of China facility big time, against estimated local debt of some CNY 50-60tn). But then, those dastardly kids at CDB HQ were not filled with joy by the plan and didn’t sign off. Back to first base for Zhenjiang.
The Wufengshan Yangtze River Bridge, Zhenjiang. Source: Caixin
We should all be familiar now with the basic problem. Off balance-sheet financing of infrastructure by China’s local governments boomed after the 2008 “CNY 4tn” (a.k.a “Spend as much money as you can as fast as you can”) stimulus. (Ex-Premier Wen Jiabao gets a lot of flack for that - but personally I think it was a pretty good response, given the circumstances. The problem I have was that he didn’t restrain borrowing in late 2019, when it was clear the economy was heating up again.)
Beijing’s efforts to make LGFEs commercial since then have fallen very flat. Or rather every time Beijing tries, growth slows and Beijing backs off. Happened in 2012, happened in 2015, and lo-‘n-behold its happening again today. In 2014, Lou Jiwei’s Ministry of Finance (MoF) attempted to come up with a final solution. The central government officially recognized some CNY 18tn of de facto local government debt after a big audit of the provinces, and “swapped” most of it for new MoF-backed provincial bonds. From then on, the official li(n)e was that local government debt did not exist, no more ‘hidden’ local debt would be recognized as Beijing’s liability, and that the borrowing entities, the LGFEs, were liable for all their own interest and principal.
Cue a number of very predictable problems, primus inter pares: moral hazard. At the central level, facing a slowing economy, Premier Li Keqiang started watering down Lou’s 2014 reform in the Spring of 2015, ordering banks to lend to “priority projects”. Curiously, localities had a lot of those ready to go. Banks, finding it harder to lend on-balance sheet to LGFEs, invented innumerable ‘shadow’ channels by which they could lend, either dumping the LGFE loans into Wealth Management Products (WMPs) and selling them onto punters, or booking the loans as “investment receivables” on the balance sheet. Local governments still had growth targets to meet and roads to build, so they continued nod-nod-wink-winking that they’d stand behind their LGFEs. Offshore rating agencies play the same game with US dollar bonds the LGFEs are now issuing hand over fist - they write about ‘government support’, when in legal theory there is no obligation. Just the fear of a huge black-mark on your record.
And lenders have a trick or two up their short-sleeves. A couple of years ago I had beers with a leasing company agent who was totally unafraid of lending to really poor county governments at really-high rates. Aren’t you scared of defaults, I asked. (Sometimes I really don’t know when I’m playing dumb, and when I’m really just being dumb.) “No, if they don’t pay, I’ll just hire some guys to protest outside their government offices - and they’ll be so scared of their leaders in the province, they’ll find the funds.” (We’ll come back to that game again below, as its not quite so ingeniously simple.)
To make matters worse, some bright spark at the MoF heard about Public-Private Projects (PPP) and proposed that as the new debt-free template. Everyone in the provinces and at central SOE construction firms then fell right over themselves to do PPPs during 2016-18. But there was never any ‘private’ (almost only SOEs took part), the ‘equity capital’ (zibenjin) portion of these projects was usually borrowed via shadow bank products (with an ‘in-the-drawer’ agreement to repay the funds with 7-8% interest after a couple of years when no one was watching), and future payments for the projects were still owed by local governments, who usually did zilch medium-term budgeting. None of the fake equity (minggu shizhai) or the future payments counted as debt. Local governments also set up innumerable “industrial funds” (chanye jijin) which (of course) borrowed from banks and then financed LGFEs - and that, of course, was not counted as local debt either.
A vision of Zhenjiang’s New Area. Source: here
So local government debt ballooned again in 2015-17, while Beijing was in denial. Come mid-2017, with the economy recovering, some brave souls in Beijing decided shadow banking had to stop, and PPPs had to be tidied up. The MoF sent yet more teams around China to do audits (and are rumored to have found some CNY 45tn (USD 6tn) in hidden debt - but I think that’s still likely an underestimate, I’d put the number nearer CNY 60bn (USD 8.5tn)). LGFEs and SOEs were told to cut their debt-equity ratios. Local officials were told that this time Beijing really, really meant no more LGFE uncommercial borrowing. Banks were told, again, that LGFEs would have to rely upon themselves for loan repayment. We went back to the delusion that LGFEs should and could be completely commercial. Cue another big squeeze on the LGFEs. And, then of course, an infrastructure investment slowdown. Which might have been fine.
But then Trump happened. In Q2 2018, the US-China trade war suddenly started, with US tariffs being introduced on a small share of imports from China. Despite the small impact, everyone and his grandmother, from Beijing to Zhenjiang, was much less happy about the future. (And at the same time, the Communist Party was letting loose with its anti-capitalist urgings, as I wrote about here).
So Beijing blinked. In July 2018, the State Council told banks to “guarantee LGFE funding needs on a market basis”. Double-speak at its finest. The State Council wanted, again, to have its cake (avoid a wave of LGFE defaults) and eat it (pretending that anyone could lend to these things on a market basis). In October, it issued additional instructions which doubled down, telling banks it was not-OK to pull back loans from insolvent LGFEs and it was very-OK (read: required) to roll-over loans to LGFEs. And then at the end of last year, Caixin reports a ministerial meeting at which a MoF representative argued for banks to extend LGFEs loans and issue new LGFE loans, and for securities brokers to help LGFEs issue bonds to repay maturing shadow loans, and for banks to do (god in heaven) “Debt-Equity Swaps (DES)”. There are reports that the bank regulator was not delighted by the proposal.
MoF, desperate not to oversee a second official debt swap with its own balance sheet, basically pushed the problem deeper into the formal financial sector, thus postponing, not solving, the problem. It also let its official deficit expand a bit for 2019 (by 1-2ppts of GDP). It ramped up issuance of targeted bonds (zhuanxiang zhai), provincial bonds which are used for non-commercial projects (things like poverty alleviation and the environment). These, of course, are mostly being bought by banks, using household deposits. And more recently, we’ve got provincial bonds being issued directly to retail investors. Reuters [here reports they’re flying off the shelves, with a 3.3% interest rate.
(A quick aside is required here. We talk about ‘provincial bonds’ since that’s officially what they’re called - but China is not a federal government, so there’s no way for a local government to file for bankruptcy even in legal theory, let alone in practice. Banks were told to buy them with a 40-50bps spread on MoF bonds. This is basically central government debt, with a notional requirement for the issuing province to pay them back out its future budget. OK, aside over.)
The hope, I guess, was that LGFEs could be re-financed with super-long-term bank loans, bonds and short-term paper, while repaying all their higher-interest shadow debts (which could cost 15%+), all the while not increasing the total amount of debt. At the same time, the official story, that this was not government debt, could be maintained. But, of course, all the lenders and buyers are acting on exactly the opposite assumption.
In defense of this policy, its not all bad. One banker working in a poor southern province I had lunch with recently told me he thought it would be put more pressure on LGFEs to meet their obligations. “I’ve had LGFEs threaten not to pay interest a few times - and what can you do? Nothing! At least with bond defaults they create some market panic and officials have to pay.” And indeed, we have seen a number of “technical” defaults of LGFE bonds in the market, where the furor of a missed payment forces provincial leaders to step in to pay interest for one of their LGFEs a few days late. Remember a few years ago, Liaoning tried to default on a CDB loan - and all hell broke loose as bankers downgraded their assumptions about “provincial” support, making it harder for all LGFEs in that province to raise funds.
Now, we have seen quite a few LGFE defaults on ‘shadow’ products - trust loans, bank loans wrapped up in WMPs, loans from leasing companies, etc. I wonder how my friend is doing with his rent-an-angry-scam. Sometimes these appear to get eventually repaid, either by the province or by the lender (to the ultimate borrower). But you might have to wait a few years. And no doubt, there are lots of defaults we never hear about.
Its really murky down there in local government finances. When you give them an inch and a GDP growth target, they’ll take a few miles. The market is back to assuming that no LGFE will really default on a bond. So credit spreads have collapsed again, and everyone is happy to buy them. So, its hard to believe this policy will not lead to more debt. Especially as Beijing is more concerned about growth again and has told itself that deleveraging has been successfully completed.
A vision of Suning Plaza Tower, a skyscraper in Zhenjiang. Its been built, apparently.
And so we return to Zhenjiang, which interpreted Beijing’s new policy as a license for doing the CDB debt swap. Who can blame them? The problem appears to be that the city was going to inject previously undisclosed debt into the new LGFE vehicle. And putting that on CDB’s balance sheet would have meant de facto official recognition of ‘commercial’ debt. Whether the CDB bosses were more afraid of breaking the formal rules or taking on debt that was never going to repaid is unclear.
At the same time, Caixin reports that MoF was pushing Zhenjiang to cut its normal expenditure and to sell assets to meet some of its obligations. This is eminently sensible, but needs some fine-tuning to work (I’m coming up with a Youshu reform package, stay tuned Zhenjiang!). Last year, when the pressure was really on, I did meet LGFEs who were selling assets. But it does not appear the scale was significant - and I’d imagine there’s less pressure today to do that.
And this is the essential knot of the problem. One needs to hem local governments in tightly, and enforce a hard budget constraint, so that they have to sell assets and repay their debts. But local governments’ first reaction to the pressure is to slow capex, and growth slows. Throw in an external shock (Trump), Beijing blinks, the pressure eases, credit becomes available again, and the budget constraint turns to doufu. Officially, MoF debt is 40% of GDP. Unofficially, its nearer 100% - because we all know that the “LGFEs are commercial entities” is not reality. And this year it will go up again. And then next year…