Who Will Pay for China’s Bad Loans?

Since this is a very long post, it may make sense first to provide a quick summary of what I am going to argue.  As I have discussed often in earlier posts, pessimists are starting to worry about excessive debt levels in China, about which they are very right to worry, and many are predicting a banking or financial collapse, which I think is much less likely.  Optimists, on the other hand, are blithely discounting the problem of rising NPLs and insisting that they create little risk to Chinese growth.  Their proof?  A decade ago China had a huge surge in NPLs, the cleaning up of which was to cost China 40% of GDP and a possible banking collapse, and yet, they claim, nothing bad happened.  The doomsayers were wrong, the last banking crisis was easily managed, and Chinese growth surged.

But although I think the pessimists are wrong to expect a banking collapse, the optimists are nonetheless very mistaken, largely because they implicitly assumed away the cost of the bank recapitalization.  In fact China paid a very high price for its banking crisis.  The cost didn’t come in the form of a banking collapse but rather in the form of a collapse in consumption growth as households were forced to pay for the enormous cleanup bill.

When US leverage was rising and the world growing quickly, the cost of that collapse in consumption was easily masked by China’s surging trade surplus, but it was real nonetheless. The bank recapitalization resulted in a brutal exacerbation of China’s already unbalanced growth model, and made it all the more vital for consumption in China to surge, especially as the world’s appetite for Chinese trade surpluses is dwindling rapidly.  As happened in Japan after 1990, when households were forced to clean up their own massively insolvent banks, the consequence could be a slowdown in consumption growth just as the country is being forced to rebalance its economy towards consumption.

If there is another surge in NPLs and government debt, once again the banks will need to be recapitalized, but the cost this time will be much more difficult to manage. If NPLs surge, in other words, don’t expect a banking collapse.  Expect further downward pressure on consumption growth.


Since 2004-5, I have been arguing that the Chinese national balance sheet includes a lot more debt than most analysts realize, and that it is structured in a way that I defined as “inverted” in my book, The Volatility Machine. Among other things, inverted debt structures tend to result in a surge in debt at the worst possible time, when the economy is already struggling, usually through an explosion in contingent liabilities.

This means that even if countries with inverted balance sheets don’t currently have very high debt levels, in many cases they should nonetheless be considered and analyzed as highly leveraged because at exactly the time when leverage becomes a worry, debt levels will automatically rise. This is why I have argued (predicted?) for the past five years that “within a few months” the market was going to become obsessed with China’s debt structure.

Unless you define a “few” months as forty to sixty months, clearly I have been wrong for many years – calling things way too early is perhaps an occupational hazard for those who read too much financial history – but it seems that debt levels are finally becoming an issue. In the past six months the market has become much more passionate about figuring out what China’s debt structure really looks like, and much more worried with what it sees.There is widespread recognition that Beijing’s total debt is not the 20-25% officially recorded, but a lot higher.

In fact going through my calculations I think it is hard to come up with a number less than 60-70% of GDP, perhaps much more, and this is almost certain to rise sharply in the next few years.  And there may be stuff out there that I haven’t even considered: For example just how much bad debt is there in the SOEs? Are all current non-performing loans in the banking system correctly identified? How sensitive are NPLs to rising interest rates, or to a rising RMB? Is the PBoC currently solvent, and what would be the impact on net indebtedness of a currency revaluation? Is there municipal and provincial indebtedness that has not been captured in the visible debt, including the guaranteed funding vehicles that Victor Shih famously identified? How much bank debt is collateralized by potentially overvalued real estate? I could go on.

But although there are definitely things to worry about when we examine China’s balance sheet, I wonder if now the worriers, after ignoring the problem for so long, may not be getting a little overexcited about the consequences, or at least about the wrong consequences. Beijing definitely has a lot of debt, and much of it inverted and so highly pro-cyclical, and normally this is a toxic combination, but there are also some stabilizing factors within the country’s balance sheet that are being ignored. A number of very smart people are now warning that China is on the verge of a banking or financial collapse, but I don’t think this is likely.

Rising NPLs? No problem

Let me quickly insist that I am not in those camps that argue that the problem is much less severe than we think, or that China can costlessly grow its way out of the debt as easily in the future as it has in the past. This last point is one that is made very often, I think, by the more optimistic of China analysts,who have pointed out perhaps too many times that the last surge in non-performing loans a decade ago was also widely cited by doomsters as a sign of impending collapse. And yet, they cheerfully claim, nothing terrible happened – China grew its way out of the loan mess at little apparent cost, and it can do so again.

Even The Economist, a lot more skeptcial about miracle cures when it discusses other countries, takes the view that China’s last banking crisis was relatively painless. They also have been resistant to claims that debt levels are much higher than reported, and recently approvingly quoted one analyst as saying that the very worst-case scenario was debt levels of less than 40-50% of GDP (with which I strongly disagree). In fact I was reading an issue from, I think January, in which, after expressing a great deal of doubt in one article about the higher debt numbers some analysts were proposing for China, just a few pages later, in an article about bad debt in the US, approvingly quoted Carmen Rheinhart (co-author of This Times is Different) as saying that contingent debt levels almost always turn out to be worse than even the pessimists expected. Their skepticism is pretty variable, I guess.

But while there is certainly a legitimate and intelligent debate about how much Chinese government and bank debt there really is, the commonly-repeated argument – that high debt levels don’t matter and the doomsayers are wrong to worry because they were wrong in the past – does not qualify, for me anyway, as a very plausible argument. I think anyone who makes this claim has failed to understand how Beijing paid for its earlier banking crises. In fact the cost of resolving the previous surges in non-performing loans actually exacerbated China’s domestic imbalances and left China in a perilous position, and the current build-up of bad debt may very well do more of the same.

How so? The first and most obvious point to make is that if a highly insolvent banking system is cleaned up, you cannot simply assume away the cost without identifying who actually paid for it. Here is where the confusion resides. The optimists perhaps assume that the only way that a banking crisis gets resolved is through a banking collapse or an explicit bailout. Since there was no banking collapse in China in the past decade, and what looked like a fairly small and manageable bailout, then clearly there was no real banking crisis, right?

Not necessarily. There are many ways to resolve banking crises, some more visibly and some less so – just no way to resolve them costlessly, and the key is to figure out the true cost and how it was paid. As I see it there were mainly three sets of tools Beijing used to manage the sharp increases in bad loans that threatened the banking system a decade ago, and of the three, the two most important were not explicit and so not easily measured or noticed. All of these required forcing down interest rates so as to pass the bulk of the cost onto bank depositors, and so all of these had an adverse impact on the quality of Chinese growth. In other words the previous cost of the banking crisis was not a banking collapse, but that doesn’t mean the cost was easy to absorb.

The role of interest rates

The first of the three tools used to manage the banking crisis involved reducing the accumulation rate of NPLs, basically by keeping borrowing rates low. The PBoC actually has been explicit about this policy. Low borrowing costs make it easier for struggling businesses to roll over the debt, and effectively reduce the real value of debt payments. This slows the growth of NPLs by passing on part of the cost to someone else.

Remember that if you reduce the coupon payment on a loan, it is economically the same thing as forgiving part of the principle amount, but this forgiveness is effectively disguised.  Those who remember the Brady debt restructurings of the 1990s fully understand how this works. In the main Brady restructurings, creditors were offered equivalent exchanges in which either principle was explicitly forgiven (the so-called Discount Bonds) or, alternatively, for those who found it difficult to recognize or acknowledge the principle discount, coupons were set at very low fixed rates (the Par Bonds). Similarly, by repressing interest rates, the PBoC was able to transfer part of the principle cost onto the banks that made the loans and so obtain debt forgiveness for the borrowers.

But while this helped the borrowers, it did not of course help the banks – unless the banks themselves were able to push the cost onto depositors, which of course they did. The PBoC repressed both lending rates and deposit rates to allow struggling borrowers debt forgiveness and some breathing space. Of course households paid for this in the form of very low returns on their savings (and with few alternative investment opportunities, they had no choice but to accept the cost).

The second of the three sets of policy tools, and the only very explicit one, involved infusing the banks with additional equity. Part of this occurred directly with the sale of bank equity to government institutions, and part of this capital infusion occurred indirectly by creating AMCs to purchase bad loans at well above their liquidation value. In both cases the capital infusion was financed by government borrowing, which at artificially low rates, to repeat what I said above, has the effect of passing the repayment burden onto lenders. Since most of these bonds were held by banks, once again the cost of the capital infusion was passed on through the banks to depositors. (As an aside, because equity infusions were so explicit, and because the banks are no longer fully owned by the government and are even partly owned by foreigners, I suspect future recourse to this particular form of recapitalization may be limited.)

Finally and most importantly, the third way of cleaning up the banking crisis involved the central bank mandating a wide spread – probably around 1.5 to 2.5 percentage points more than the normal spread – between the bank lending and the deposit rate, which increased bank profitability substantially and so helped to recapitalize the banks. In other words not only were depositors “taxed” for the clean-up by having to fund the very low lending rates, but they were taxed a second time to guarantee sufficient bank profitability to rebuild capital. With all these transfers from the household sector to the banks, amounting to several percentage points of GDP every year, households were forced to clean up the Chinese banking system.

Beijing’s strategy to clean up the banks was very successful, and certainly prevented the banking crisis that many expected, but there was nonetheless a significant cost to the economy. The bailout implicitly required that bank depositors subsidize the cleaning up of the banking industry. This in effect represented a large transfer of income from the household sector to the banks, to government and to businesses, equal annually to several percentage points .

NPLs and household consumption

How much was the value of the transfer?  It’s hard to say, but we can make some estimates.  Over the past decade nominal lending rates in China have been about 6% while nominal GDP growth rates have been 14%.  Economic theory tells us that nominal interest rates should be equal to nominal GDP growth rates if providers of capital are to earn their fair share of growth, and in fact in developed countries the relationship holds pretty well.  However Jonathan Anderson at UBS put together a very interesting analysis in a November 12, 2009, report that argued that it was wrong to assume Chinese nomnal interest rats should be equal to its nominal growth rate.  He looked at the case of other developing countries and found that there was no obvious relationship between the two.

But I am not convinced.  First off, if nominal interest rates are much lower than nominal growth rates, then almost by definition the providers of capital are getting less than their share of the benefits.  Since the providers in China are mainly households, and the users of capital are businesses, speculators, and the government, this must represent a real transfer of wealth from households – which I think Anderson acknowledges, although he argues that the high savings rate is an independent variable that drives the low interest rate, whereas I think that it is one of the consequences of low interest rates and other policies that force households to subsidize production (and so force up the gap between production and consumption, which is the savings rates).

Secondly, his sample includes a lot of developing countries with closed or sticky capital accounts, and who intervene in their currencies, most especially the countries that followed the so-called Asian development model.  These countries have systematically repressed interest rates – in fact that is for me one of the definitions of the Asian development model.  This makes their inclusion in a statistical sample to determine the “correct” level of interest rates very questionable. He also includes a lot of OPEC countries who for totally different, and explainable, reasons have very low interest rates, and these too create a downward bias in the statistical sample.

Finally from his own numbers, even with the possibility of significant statistical bias, I would say that there does seem to be a reasonable relationship between nominal interest rates and nominal growth rates.  On average nominal interest rates have been roughly two-thirds of nominal growth rates, although there is wide dispersion around the mean, which we would expect if interest rates were repressed.  With nominal growth rates of 14% during the past decade, this implies that nominal interest rats should have been nearly 10% or a little less, versus the actual 6%.

This is not a very scientific way of going about it, but my very back-of-the-envelope estimate suggests that interest rates in China, without financial repression, would have probably been anywhere from 300 to 800 basis points lower than the appropriate equilibrium level during the past decade.  Add this to the excess spread between deposit and lending rates, which is anywhere from 150 to 250 basis points, and we could easily argue that the deposit rate is at least 450 basis point lower than it should be, and perhaps an awful lot more.

How much is that in GDP terms?  A quick call to my friend Logan Wright at Medley Advisors gave me the following data. Total banking deposits in China are around RMB 64 trillion.  Around 60% of the total represent household deposits (an estimate, since there is some ambiguity in the numbers).  Total GDP is nearly RMB 34 trillion.  Inputting all of that into my trusty Excel Spreadsheet suggests that at a minimum, households have “paid” in form of excessively low rates on their deposits a minimum of 5% of GDP every year, and possibly up to two times that amount, during the past decade.

This is, to me, an astonishing number.  Every year households may have transferred at least 5% of GDP to the banks, and possibly a lot more.  Now of course they are paying for a many other things than simply recapitalizing the banks.  They are also paying to keep the cost of capital low so as to make viable a whole series of investments – manufacturing investments, real estate investments, infrastructure investments, PBoC sterilization bills, other government bonds, etc – that might be considered non-economic investments and that would otherwise show negative returns (in fact excessively low interest rates, as the various recent US bubbles clearly indicate, almost always lead to misallocated investment).  But since a lot of this investment occurs through the banking system anyway (for example banks directly or indirectly buy most sterilization bills), much of this ends up as part of the bank clean-up.

By the way forcing unlucky households to clean up the banks is pretty standard in the annals of banking crises, and for example has occurred in the US with the recent bank bailouts (which of course were paid for with taxpayer money), but not only was China’s total bill over many years much higher, because of its domestic distortions the impact in China was worse than it would have been in the US (because forcibly reducing consumption in China is much worse than doing the same might be in the US). Added to the other major transfers from the household sector (the undervalued exchange rate, and slow wage growth relative to productivity growth), and given the sheer size of the clean-up, it is perhaps not surprising that during the period of the bailout, household income, already a relatively low share of GDP, declined to alarming levels. This happened even in spite of explicit and much-publicized attempts by Beijing to raise the household consumption share of GDP.

This, then, is the real risk of another bout of rising non-performing loans in China. It is not that China’s banks are likely to collapse. It is illiquidity that causes bank collapses, and unless capital controls are sharply undermined we are not likely to see this happen in China. Debt levels are certainly high and highly pro-cyclical, but even if the banks are insolvent Beijing largely controls domestic funding and domestic interest rates and can protect itself from the bank runs that plagued US and European banks. We saw the same thing in Japan thirty years ago, when it was able to fund the massive banking bailout and soaring government debt levels, to what would earlier have seemed like unimaginable levels. Like Tokyo in the 1990s, Beijing is in a strong position to continue to fund its rising bank-related liabilities and will not have a debt problem any time soon. Government debt levels are indeed very high, but they can go much higher.

This doesn’t mean however that we don’t need to worry about the debt, and it certainly does not mean that if China runs up more bad loans as a consequence of the recent lending spree it will simply “grow” its way out. In the past China could certainly grow its way out, even with household consumption declining as a share of GDP, because one effect of declining relative consumption – a rising savings rate along with a rising trade surplus – was easily absorbed by a rapidly growing world economy. As long as debt levels in the US and other deficit countries could easily rise to counteract the adverse employment effect, the world, and especially the US, had no trouble with absorbing China’s rising trade surpluses.

Rebalancing household consumption

Things may be very different now. Unemployment is high in trade-deficit countries and debt levels are being forced down. If the world can no longer absorb rising trade deficits, and especially if over the next few years trade tensions increase, China must reduce its excessive reliance on exports and investment to fuel its continued growth. The only healthy way it can do so is if household consumption rises as a share of GDP because of surging consumption.

And household consumption will indeed rise as a share of GDP – with such a low current level of household consumption, and rising global concern over the employment effects of China’s trade surplus, China has no choice.  But since growth in household consumption has always been constrained by the growth in household income, it may be unreasonable to expect a surge in consumption when households are also required to clean up another sharp increase in non-performing loans.

So as a consequence of the global crisis, China’s growth will rely more than ever on the growth of household consumption. The good way this can happen is by a surge in household consumption that will allow economic growth to remain high. The bad way is by lower growth in household consumption matched by a very sharp decline in economic growth. If the worriers are right, and non-performing loans surge, China can nonetheless easily avoid a banking collapse, but that does not mean the cost of cleaning up the banks will be negligible. On the contrary, it will put even more downward pressure on low-consuming Chinese households and will make the inevitable rebalancing of China’s economy much more difficult than many expect.

As I discussed in a posting last month, Japan showed how difficult. In the past two decades Japanese consumption growth has slowed from its headier pace of the 1980s. Consumption growth has limped along at 1-2% annually from 1990 to now as Japanese households were forced indirectly to clean up their own bad loans using almost identical mechanisms – repressed interest rates and an undervalued currency. Whereas in the 1980s, when Japanese economic growth exceeded its consumption growth thanks to its large and rising trade surplus, in the past two decades Japan’s economic growth – less than 0.5% annually – has been less than its consumption growth as Japan slowly and painfully rebalanced its economy towards consumption.

Likewise perhaps with China. Unless the rest of the world is willing to absorb rising trade deficits and supply it with rising trade surpluses, rebalancing for China means that instead of being the lower limit of economic growth, consumption growth will now be the upper limit. If future Chinese consumption growth also slows, as it did in Japan, because households are forced to foot the new bad-debt bill, we may see the real cost of the current explosion in bad loans – several years of sub-par growth.

It turns out that banking crises might not be costless, even if they don’t lead to banking collapses. In the case of China they may instead lead to a collapse in consumption growth.  As part of the trade dispute that China is facing with the rest of the world, this should give some indication of how little room China has for its adjustment.  Anyone who is too impatient with the glacial pace of Chinese adjustment must recognize just how difficult it will be for China quickly to reorient its economy towards household consumption.  The risk is that China, like Japan in the 1990s, will rebalance in the form of a sharp contraction in GDP growth as households struggle to pay for the misallocated lending boom.

Opinions and comments on RGE EconoMonitors do not necessarily reflect the views of Roubini Global Economics, LLC, which encourages a free-ranging debate among its own analysts and our EconoMonitor community. RGE takes no responsibility for verifying the accuracy of any

One Response to "Who Will Pay for China’s Bad Loans?"

  1. Roland   April 8, 2010 at 10:41 pm

    Another limit to “growing out” of a debt problem is demographics. China’s population is aging in manner similar to most of the developed nations.