by Michael Pettis
In the last issue of my newsletter much of the first half was dedicated to a discussion of recent events in Spain and Italy and why they reinforce the argument that several countries will be forced to leave the euro and restructure their debt. The most worrying, but expected, fact was the amount of capital fleeing the afflicted countries. I cited an article in Spiegel that claims that in the past year an amount equal to nearly 30% of Spain’s GDP had left the country. Flight capital is both a major result of declining credibility and a major cause of further declining credibility, and because it is so intensively reinforcing it is a major warning signal.
This matters for China for at least two reasons. First, a worsening Europe will make it harder than ever for China to rebalance growth away from investment, and second, China itself is experiencing capital flight.
To address the first, any sustained increase in the growth rate of Chinese consumption – if indeed this occurs, which in my opinion is very doubtful – will not only have to compensate for a reduction in the growth rate of Chinese investment, but might also have to compensate for a reduction in China’s current account surplus. What is more, the crisis in Europe will only make the global trade environment tenser and nastier.
Notice already what is happening in commodity-exporting countries like Indonesia. According to an article Last month in the Wall Street Journal:
Indonesia’s trade deficit hit an all-time high in June as exports from Southeast Asia’s largest economy fell sharply, a sign that weaker demand from China and the West is affecting some of the few countries still growing at a considerable clip.
A third straight month of trade deficits in one of the world’s biggest commodity producers bodes ill for Indonesia, which had become a darling of foreign investors looking for fresh opportunities, but has struggled to contain the damage from a sharp fall in its currency in recent months that has rattled investors.
Countries whose growth depends either on growth in Chinese investment or growth in European demand are going to see significant deterioration in their trade accounts. This will almost certainly lead to even more trade intervention, currency wars, and all the other beggar-thy-neighbor polices typical of a global demand contraction. I think we should expect to see a lot more articles like this.
In addition China itself is seeing noticeable capital outflows as business owners and other wealthy people begin disinvesting and withdrawing deposits. Capital flight from China began surging in early 2010, and it seems to be getting worse, with some monthly withdrawal estimates as high as $40-50 billion, and this can’t help but put increasing pressure on China’s ability to finance the infrastructure, manufacturing and real estate bubbles that have driven the economy.
Over the long term, and in the name of rebalancing, this is probably a good thing for China. The sooner liquidity-driven over-investment stops, the less debt will pile up and the less painful the deleveraging process will be. But in the short term it will aggravate the slowing down of the economy.
Again not everyone agrees. I have found it interesting that two of my favorite financial periodicals, both British, have such opposing views on China. While the Financial Times has been mostly in the skeptic’s camp for many years, the Economist has been noticeably bullish – although generally the most reasonable and intelligent of the bulls. For example the current issue of the Economist sees the recent retreat from the difficult rebalancing process as a good thing:
China’s unruly property market was once dubbed, with excusable hyperbole, the “most important sector in the entire global economy” by Jonathan Anderson, then at UBS, another Swiss bank. It remains the biggest fear hanging over the world’s second-biggest economy. Home prices began falling about a year ago. The declines have depressed investment and curtailed economic growth, which slowed to 7.6% in the second quarter, its slowest rate since 2009.
The only economic anxiety to rival property is local-government debt. Estimated at 10.7 trillion yuan at the end of 2010 by official auditors (and a lot higher by unofficial ones), much of it was held at one remove by so-called local-government financing vehicles. When one such, Yunnan Highway Development and Investment, told creditors in 2011 that it would not repay the principal on their loans, it was described (with equal hyperbole) as the “default heard around the world” by Business Insider, a news website.
Both worries have roots in the stimulus spree on which China embarked in November 2008. State-owned enterprises began bidding enthusiastically in land auctions, and local governments let their pet projects run wild. Ever since, the two problems have preoccupied China’s central government. In April 2010 it put curbs on speculative homebuying and spent much of last year tidying up local finances. This spring, Wen Jiabao, the prime minister, boasted that local-government debt had grown by a mere 300m yuan in 2011.
Although at the time the Economist was less worried than many of us about the 2008-9 “stimulus spree”, and still usually downplays concerns about China’s debt, they go on to say that things have gotten better:
Signs, however, are growing that both property prices and local-government borrowing are rising again. China’s National Bureau of Statistics (NBS) reports that new home prices rose in June in 25 of 70 cities it tracks. They fell in only 21. Sales volumes also strengthened. Prices are still lower, on average, than a year ago. But according to our weighted average of the (new and existing) home prices reported by the NBS, the pace of decline appears to be bottoming out (see chart). By next month (thanks to a lower base), prices may be rising again, year-on-year.
Local governments meanwhile have been given leave from their debtors’ prison. Reports suggest that China’s banking regulator has told banks to increase lending to “better qualified” financing vehicles. These vehicles have also increased their bond sales, issuing over 420 billion yuan-worth of paper already this year.
Is debt a problem?
The rest of the article argues in part that China can easily manage its debt problems because debt levels are actually relatively low and China has room to increase its net indebtedness:
China’s economy does need help, and its government has ample scope to provide it. Some local governments took on more debt than they could handle. But their liabilities never endangered the fiscal position of the country as a whole. The combined debts of China’s central and local governments add up to about 50% of the country’s GDP (including bonds issued by the Ministry of Railways and China’s policy banks, intended for state-directed lending). Even if local debts are understated, China has fiscal room for error.
I am not sure I agree. First, to make a minor point, I don’t think real estate has necessarily been the “biggest fear hanging over” China. I have always argued that the biggest worry is the unsustainable increase in debt, which historical precedents suggest is an almost automatic consequence of an aging investment-driven growth miracle. While the real estate bubble gets most of press, I would argue that several of the analysts who have been in the skeptic’s camp for many years, like Logan Wright of Medley Advisors or Victor Shih, now with Carlyle, usually agree that debt is the most worrying problem.
Of course borrowing money to fund a real estate bubble is an important source of bad debt, but I have argued for many years, and continue to believe, that economically non-viable infrastructure investment has been a much greater source of bad debt, by which I mean debt whose servicing cost (excluding of course interest rate repression) exceeds the debt servicing capacity created by the investment (excluding subsidies and including externalities). Empty buildings may be much easier to visualize, and much more photogenic, and many people still have an impossibly tough time understanding why it is possible to over-invest infrastructure (isn’t all infrastructure spending good?), but I would argue that sharply reducing infrastructure investment, or at least diverting it into more useful – if less glamorous – projects, is more important than reducing excess real estate development, although this too is clearly a problem.
But that aside, my disagreement with the article is really about whether or not China has a low enough debt level that we can relax about the “fiscal room for error.” Is the relevant debt really just 50% of GDP?
I doubt it. There are three important corrections that need to be made. In the first place total direct obligations of the government must include any net indebtedness within other sovereign or quasi-sovereign institutions, most notably the PBoC (People’s Bank of China). We don’t really know what the PBoC’s balance sheet looks like, but remember that they run – by definition – a hugely mismatched balance sheet in which, since July 2005, the value of liabilities has risen relative to the value of assets by roughly 30% (the amount of nominal appreciation of the RMB against the dollar).
With total reserves equal to roughly half of GDP, it isn’t outlandish to suggest that the PBoC’s net indebtedness might add at least several percentage points of GDP’s worth of debt to the direct liabilities of the government, and perhaps a lot more. A thorough analysis of other relevant entities, such as the national development banks, is likely to add more debt.
Some analysts argue, very bizarrely, that since entities like the PBoC are guaranteed by the central government, and since no one knows the extent of the net indebtedness, it doesn’t really matter at all if they are or are not insolvent. This, however, is an idiotic argument. There are only three things you can do with your debt. If you are solvent you can service it out of operating income. If you are not, you can either default, or you can transfer income from some other source to cover the difference. The cost of the debt, in other words, one way or another must be borne by someone. Invisible debt is still part of the debt burden.
The second important correction has to do with contingent liabilities. It is not just the direct obligations of sovereign and sovereign-guaranteed entities that matter. More importantly, and as we saw very clearly in the last two years in Europe (and in dozens of other cases in recent history), it doesn’t make sense to talk about government debt levels without including contingent liabilities that emerge through the banking system. Here the numbers are potentially significant.
Total loans in the banking system, including the various off-balance sheet transactions engineered in large part to get around lending restrictions, are at least 180% of GDP, and we have no idea of how much potential bad debt there is on the balance sheet of the SOEs (state owned enterprises) or in the informal banking system (estimates are that total loans in the informal banking system are equal to 15-25% of total loans in the formal banking system). Whatever the number, we don’t need to see anywhere near the 40% of total loans that were estimated to be non-performing in the last banking crisis a decade ago for total government debt, including contingent liabilities, to be very high.
Inverted balance sheets
The third important correction is usually the hardest for people to grasp. This has to do with “inverted” balance sheet structures, which I discuss extensively in my 2002 book, The Volatility Machine (Oxford University Press). When you are concerned about a borrower’s credibility, you should not just look at outstanding obligations under current conditions. You should also worry about outstanding obligations in case of a likely adverse shock.
When we think of Spanish government debt, for example, we don’t just think of the current debt load of 69% of GDP (or wherever it has climbed to since the end of last year), or even of the contingent liabilities from provincial debt and non-performing loans. We must also be concerned about the self-reinforcing relationship between the debt and the currency.
If Spain were to leave the euro, in other words, because much of its debt is external debt denominated in euros, any devaluation of the new currency (let’s call it the peseta) would cause a corresponding increase in the debt burden. The peseta could easily lose 50% of its value, for example, in which case the external debt burden would double its share of GDP.
There is a self-reinforcing relationship between the two. Since investors are aware of the risk, the worse the debt-enhancing impact of a devaluation, the more likely the devaluation is to occur, and the higher the external debt, the greater the actual devaluation will turn out to be. We saw this most spectacularly in Argentina, whose debt-to-GDP ratio, if I remember correctly, was “only” around 53% in the period just before the December 2001 corralito and debt default, which was itself followed immediately by the January 2002 devaluation. An earlier, equally spectacular case was that of South Korea in 1997. Its relatively small external debt burden before November 1997 became unbearable just one month later after the forced devaluation of the Korean won.
What does this have to do with China? Perhaps quite a lot, given the many pro-cyclical structures embedded in the country’s economy and balance sheet. If we assume that China will have no problem sailing through its economic rebalancing, the European crisis, and everything else, then clearly we don’t need to worry about anything. But if China’s rebalancing is accompanied by a sharp slowdown in economic growth, or if it occurs during a worsening of the European crisis – both very likely scenarios – then we need to think about what the debt burden will be under those conditions.
So, for example, will commodity prices drop? I think they will, perhaps by as much as 50% over the next three years, and to the extent that there is still a lot of outstanding debt in China collateralized by copper and other metals (and there is), our debt count should include estimates for uncollateralized debt in the event of a sharp fall in metal prices . Will slower growth increase bankruptcies, or put further pressure on the loan guarantee companies? They almost certainly will, so we will need again to increase our estimates for non-performing loans.
Will capital outflows increase if growth slows sharply? Probably, and of course this puts additional pressure on liquidity and the banking system, and with refinancing becoming harder, otherwise-solvent borrowers will become insolvent. Will rebalancing require higher real interest rates, a currency revaluation, or higher wages? Since rebalancing cannot occur without an increase in the household income share of GDP, and since these are the biggest implicit “taxes” on household income, there must be a net increase in the combination of these three variables, in which case the impact on net indebtedness can be quite significant depending on which of these variables move most. Since I think rising real interest rates are a key component of rebalancing, clearly I would want to estimate the debt impact of a rise in real rates.
In a later article I will try to list more systematically areas where I think we should be concerned about inverted balance sheet structures in China, but my main point here is that very often – in fact in the majority of cases – debt crises catch us by surprise because there is a sudden and unexpected surge in debt caused by factors we hadn’t thought about. It is the sudden and unexpected explosion of contingent liabilities that generally precedes debt crises, and not the actual debt burden a year or two before the crisis, that ends up triggering the crisis.
Just remember the finger wagging and the self-satisfied lectures on banking and debt given by senior Spanish government officials and bankers to US and European bankers as late as 2009. No one thought Spain had a problem until debt suddenly emerged from every nook and cranny as a response to the adverse shock Spain was undergoing. Some analysts will complain that it is very difficult to figure out all the contingencies in any country, so acknowledging the possibility adds nothing to the quality of our analysis. But they are dead wrong. An experienced balance sheet analysts can easily tell when overall a country’s balance sheet is more inverted or less inverted, and in the former case he must always assume that the potential for a debt crisis is much greater than the raw debt numbers suggest.
By the way I am not suggesting by any means that Beijing’s current debt level is unsustainable, but I do think it is hard to argue that it has not been rising at an unsustainable pace in recent years. What is more, in every single case in history that I can remember, during a great liquidity-driven bubble, debt structures became increasingly inverted and risky, especially in poor, developing countries, and more especially in countries whose rapid growth is driven by rapid investment growth funded by a financially repressed banking system. The reason should be obvious – when the cost of capital is artificially repressed, economic entities tend to overuse capital as an input. Perhaps that has not happened in China in the past decade, but if it hasn’t, China would represent a truly unique case in history.
We need to be worried about debt, in Europe and the US of course, but we need also to be worried about debt in China. The deleveraging process in any country always results in much slower growth than during the period in which debt was rising quickly, and what matters is overall deleveraging, not just government debt. At some point we will see deleveraging in China, and this must affect growth. Misallocated investment funded by debt means that losses have occurred and one way or another they will eventually be recognized. The recognition of the losses can be postponed for a time, by the simple expedient of not recognizing non-performing loans, but at some point, and usually at the worst possible time, they will be recognized.