Everything We Think We Know About Chinese Finances is Wrong

China has long faced doubts about the veracity of its economic data and concerns about its rapidly rising level of indebtedness.  While defaults and individual incidents raised questions about debt discrepancies, there was no systematic evidence that the financial system faced systemic misstatement. The People’s Bank of China changed that with a few sentences.

By some estimate, the widely watched debt to GDP metric in China has already surpassed 300%. While this is level is worrying given financial stress associated with countries that reached similar levels, this is only half the story.  There have long been suspicions that Chinese debt numbers are not entirely accurate but data that would demonstrate a systemic difference from data has never emerged.  However, every time a company collapsed, there would inevitably come out a mountain of undeclared debt. While this raised suspicions, there was never systematic evidence.

The Financial Stability Board (FSB), formed after the 2008 Global Financial Crisis, aggregates data for major countries that includes a broader measure of assets by banks, insurance companies, and other major asset holders.  According to their data, at the end of 2015, China financial system assets had already reached 401% of GDP.

This put them only 11% (5100 basis points) behind Germany and 200-300% ahead of comparable emerging markets like Brazil, Russia, India, and Mexico.  By this measure, at the end of 2015, China was already worrying and a distinct outlier, but not completely absurd.

China itself, gave us evidence that its financial data is wildly off.  The annual PBOC Financial Stability Report with little fanfare more than doubled its estimates of financial system assets.  In a little noticed paragraph the PBOC noted that “the outstanding balance of the off-balance sheet of banking institutions….registered 253.52 trillion yuan.” To provide some perspective, official on balance sheet assets were only 232.25 trillion yuan.

The PBOC report matches extremely closely official data for the on balance sheet portion of bank assets, but matches no known official data for the off balance sheet portion of assets. Nor does the PBOC provide many clues as to what these off balance assets are holding.  They do note that roughly two-thirds of the 253 trillion is held as “financial asset services” which may mean everything from structured products sold to clients who believe the bank will stand behind the product, special purpose vehicles holding non-traditional assets, or certain types of financial flows.

If we revise our earlier estimate of financial system assets to GDP based upon the new PBOC numbers, China’s position changes dramatically.  The FSB estimate of all financial systems published only in May 2017 jumps from 401% of nominal GDP to 653% of GDP at the end of 2016 for just banking system assets.

If we take the FSB data, add in the new PBOC data, and estimate forward to 2016 Chinese financial system assets are equal to 833% of nominal GDP ahead of Japan at 657% and behind only international banking center United Kingdom at 1008%.

This level of asset accumulation imposes real costs. Where as Japan and Europe have close to zero or negative interest rates, China has significantly higher. If we make the simple cheap assumption that these assets earn the short term interbank deposit rate of return of 3.5%, this would imply a financial servicing cost to the economy of 29% of nominal GDP. Conversely, Japan with financial assets of 657% of GDP but using the higher long term loan rates of 1% instead, would need only 6.6% of GDP to service its asset costs.  Prof. Victor Shih at the University of California, San Diego wrote in a recent report that “Total interest payments from June of 2016 to June of 2017 exceeded incremental increase in nominal GDP by roughly 8 trillion RMB.”

What makes this disclosure concerning is how extreme the numbers are. Even the FSB placed China among developed country financialization and well outside the range of other emerging markets. The new numbers place China on the extremity of all major economies behind only a major international banking center even in front of Japan who has run strongly expansionary monetary policy for years to try and push inflation.

Many analysts have raised concerns about asset bubbles and debt growth in China but even the most bearish would have had trouble believing this level of financialization.  Even the risks are more than hypothetical.  In bankruptcies or defaults, it is common to find enormous amounts of undisclosed debts or asset management products sold by banks to clients they are expected to make good even if technically off balance sheet.

There are a handful of key points to remember:

  1. We do not know what these assets hold other than three broad categories comprised of guarantee, commitment operations, and financial asset services which even then only comprise 79% of the total 253 trillion.
  2. These are not simply bank to bank flows. It is likely this number includes some financial to financial flow, but significant amount clearly out in the real economy.  The PBOC includes under these assets entrusted loans as well as guarantee operations both of which indicate real economy activity.
  3. Even if the off balance sheet assets are just bank to bank flows this actually makes the banking system worse. This happens because that means official bank borrowing is much higher than official data indicates lowering already strained capital adequacy rates to very concerning levels. Total on balance sheet bank capital is 15.5 trillion or 6.1% of the 253 trillion in off balance sheet assets.  If any sizeable amount of the 253 trillion in off balance sheet assets is lent to the banks for on balance sheet activities, this destroys the banks capital base.  In fact, depository corporations in China only list 28.6 trillion in liabilities to either depository or financial corporations.  So either the off balance sheet assets are not flowing to banks in large amount or official on balance sheet financial figures for China are wildly wrong with disastrous consequences. I personally lean to the idea that most of these assets are not flowing to banks but do want to emphasize that if you are going to make the counter argument, the implications are probably even larger and worse.
  4. There are two primary ways in China that assets end up off balance sheet. First, the Enron model. In this scenario, accounting sleight of hand is used so that SPVs are used so that an entity does not have to consolidate finances of entities it effectively controls. It should be noted that this does not mean that the bank or other institutions have done anything technically illegal, only that while control may legally lie elsewhere and finances are not consolidated up to a known parent, the financial risk never leaves.  Many bad debt management schemes are where a major bank acts as manager but holds less than the controlling amount so that they can claim the debt is off their balance sheet.  In some instances, they work with other banks who contribute the capital required to ensure the manager is not aggregating financials upwards.  I even know of some instances where the banks are buying debt from other banks where the clients who are the bad debtor are contributing the majority of capital as the bank buys bad debt from other banks as the manager of a fund.  The key point is that Chinese banks are technically meeting accounting requirements to move debt off balance sheet but not transferring the risk.
  5. The second most likely source is banks selling asset management products to other clients. These products are widely spread throughout the economy from corporate China looking to store cash for 30 days, wealth management firms, or individual bank clients.  What is important to note is that in this case, the bank typically does not technically/legally carry the legal risk of the product purchased by clients.  Most of the products are unguaranteed.  However, pragmatically, this simply is not an accurate assessment of the reality.  Take an extreme example.  Assume a significant portion of these off balance sheet assets sold, even say 10%, defaulted and went to zero.  This would cause a major problem.  Where we have seen large losses attempt to be imposed on retail type investors, they have almost always been bailed out.  Beijing and defenders can claim all day long that neither Beijing or the state owned banks guarantee these products but when Beijing starts imposing large losses on investors rather than bailing them out, then I will believe it. To date, that has not happened.
  6. It is important to note that given the size of these off balance sheet assets, this obfuscation of financial data has been occurring for many years. Even China does not go from 0 to 253 trillion RMB in one year. This implies that we need to rethink the entirety of Chinese development and finance since probably about 2000.  One truism has been that when true pictures of financial health are obtained, typically in a default, there is always enormous amount of undeclared liabilities.  We can no longer exclude that these are not isolated cases but as the PBOC has admitted, the norm rather than the exception.
  7. We do have some scant evidence of how rapidly this off balance sheet side of the banking system has growth. In the 2015 FSR, the PBOC listed off balance sheet assets at the end of 2014 as equal to 70.44 trillion RMB or equal to 40.87% of “Chinese banks aggregated balance sheets”. In the 2016 FSR, the PBOC said it was equal to 82.36 trillion RMB and equal to “42.41% of the total on balance sheet assets.”  The reason the 2017 exploded to 253 trillion was because “Starting in the first quarter of 2017, the PBC would count the off-balance-sheet wealth management products in banks’ total credit in the MPA framework, which would urge the banks to strengthen off-balance-sheet risk management, so that the macroprudential framework would be more effective when conducting countercyclical adjustment and guiding the economic restructuring.” Put another way, it knew the risks were there before but it was not reporting them. This means that we can assume the on and off balance sheet assets are two distinct pools of capital/assets and not overlapping as it might be rightfully asked.  This means the on and off balance sheet assets for Chinese banks total 232 trillion plus 253 trillion.
  8. The absolute size and growth of assets imply there will be enormous (as in Biblical) costs to deleverage. Let me give you a simple example. Let’s assume a flat rate of economic financialization by which I mean that nominal GDP and systemic financial asset growth are equal.  For our case here, I’m going to use similar but round stylized numbers.  In our world, financial system assets are equal to eight times nominal GDP.  Now, let’s assume that both financial system assets and nominal GDP grow at 10%.  In this stylized but similar world, financial system assets will have grown by an amount equal to 80% of GDP. If this both nominal GDP and financial system assets grow at 10%, by 2025, China will have financial system assets equal to approximately 1,900% of nominal GDP.  Because total banking system assets are so much larger than nominal GDP, simply growing both at the same pace will continue to lever up the economy.
  9. This might actually explain one unique data point which no one has a good explanation for, including myself. For a number of year, fixed asset investment in China has been above 80% of GDP.  Through the first three quarters of 2017, it is only3%.  It has been puzzling to many how FAI could top 80% of GDP even with the growth in debt that we saw. That was simply an amazing number.  Well if there was unseen asset growth of equal to twice official banking system assets, this would explain how FAI could comprise that amount of GDP.  However, this implies that China has been much much more dependent on credit and money growth to drive GDP than anyone, myself could have believed.
  10. This further implies that much of this economic boom has been driven by a hidden expansion of money and credit. As research has noted, it is much easier to stimulate activity with hidden monetary loosening than with expectations.  If the numbers the PBOC note are real, this would imply many years of hidden loosening.
  11. This further implies there is a large (read Biblical) asset bubble. At first glance this seems to match the data.  If we look at the data on the major asset for households, real estate in tier one cities is the most expensive in the world and even the average tier two and tier three city has higher per square foot price than most of the United States.  The median price in the United States for real estate is $139 per square foot. Tier two cities in China are currently $170 with Tier three cities a more pedestrian $110.  Using conservative extrapolations of national housing prices in China yield a current average price per square foot of $191 per square foot.  To provide some perspective, residential real estate in China is 38% more expensive on a price per square foot basis but nominal per capita GDP in the United States is 608% higher.  We could point to a variety of other assets which appear vastly overvalued but given the increase in financial assets appears prone to a significant asset revaluation.
  12. This also has significant implications for foreign exchange policy. It implies that China will maintain strict capital control measures in place for the quite some time. Let’s take a simple example that we could expand to other sectors of the Chinese economy. Assume that markets have pressure to equalize prices. Chinese citizens and firms have a very real interest in switching into similar foreign assets while foreigners have very little interest in switching into Chinese assets.  I have long noted that there is fundamentally, absent controls, a much larger structural non-cyclical interest in purchasing foreign assets by Chinese than in purchasing Chinese assets by foreigners.  Unless China is will to accept a much lower value for the RMB, they cannot allow change to foreign exchange policy.
  13. Though I am always loathe to bring politics into discussions about Chinese economic and financial policy because politics is too unknowable in China, I think there is a little worth commenting on here though this is mostly speculation. This nugget of information was dropped in the middle of a report in an almost off handed way.  However, the magnitude of the revelation is akin to saying over dinner “I just killed five people before I arrived would you mind passing the salad dressing?” The reason this matters is that PBOC head Zhou has been making the rounds talking about a variety of things like Minsky moments and slowing corporate debt growth. I don’t think it was any coincidence that this nugget of information was dropped into conversation as Zhou appears to be heading out the door and making the rounds using language he knows will raise concern.  While it is fair to question his reformist intent, how long he will stay, and other issues, he clearly knows that discussing these issues in this manner and dropping this piece of information raise concern. If I can speculate, it appears Zhou is trying to raise the pressure to reform, without burning it down.  It does make one think that the information was released to pressure Beijing.

There is way too much we do not know about the details of this revelation. However, it is without a doubt the largest and most altering revelation to come out of the Chinese economy probably this decade. It will require a major rethink to what we think we know about the Chinese economy, how it developed, and what the future holds.

I would like to thank Chris Aston who originally Tweeted about this in July from the Chinabankingnews.com website and the appropriately named Deep Throat blog who wrote about this topic and does great work on  a variety of issues who drove me to revisit this issue.  I originally chose not to write about this topic because the numbers were so outlandish I figured I had to seriously missing something that caused them to be much more normal.

My Last Word on FX Swaps on Chinese Banks Net Foreign Asset Position

Just a few last words on FX swaps and Chinese banks based upon the recent Brad Setser follow up and next time we move on to new material.

First, we cannot say exactly who is taking the other side of this swap position but that in no way negates the point that they are being made and volume according to bank and market financials has risen enormously.  Based upon multiple data points it is clearly happening in large number.

Second, swaps are not “off balance sheet”. They are very much on balance sheet.  However, banks do not record the full cash or “notional” amount of the swap only some type of value at risk or expected gain/lost amount.  For instance, assume I buy a $1 million FX swap that returns the domestic currency at a future exchange rate that imposes a 1% loss plus say 0.5% transaction fee. I would say the “notional” amount is $1 million but I would carry as a liability the 1.5% (1% loss plus 0.5% transaction fee). This distinction will come important later.

Third, there is no explanation to the most fundamental of questions: how are Chinese banks obtaining foreign currency to fund this growth in foreign currency assets?  Again, how does a bank fund asset growth if it cannot access, via deposits or debt, capital to fund asset growth?

Fourth, the preferred explanation of higher foreign currency deposits in Chinese banks does not come remotely close to explaining the growth in foreign assets.  This is explained by noting that foreign currency deposits increased during this time.  However, during a period when net foreign assets increased $300 billion, foreign currency deposits went up by a total of $123 billion. This leaves a significant amount of unexplained foreign currency asset purchases even using those numbers.

What makes this increase in foreign currency deposits even more interesting is that it lines up perfectly with the preferred explanation of a move in to swaps to facilitate outflows.  Let me explain.  Though foreign currency deposits increased by $123b, $54b of this increase comes from overseas foreign currency deposits into Chinese banks. In other words, domestic Chinese foreign currency deposits have only risen by $69 billion since January 2015.  When a Chinese bank sets up an overseas operation, depending on whether they are a legally a branch or a subsidiary office, technically two different entities but typically sharing significant overlap, the financials of a branch are technically credited back to China.  Consequently, China is able to credit foreign currency deposits in overseas offices to its own balance sheet.  It is not uncommon for products to be combined for clients between branch and subsidiary balance sheets.

Here is how this explains the “swap” aspect.  Using the example of a client that has RMB on the mainland but can’t get the money out, when a Chinese bank offers via swaps to move this money, banks virtually always require this money be deposited with them in their overseas entity. For example, let us assume that Company A has a Chinese subsidiary with 700m RMB that they cannot get out of China they want to use for other purposes let us assume in London.  Chinese Bank A will facilitate a swap to “lend” money to Company A in London. However, as part of this transaction, Company A will be required to deposit 700m with Chinese Bank A in China. They will also be required to deposit some portion of the corresponding loan in London with the Chinese offshore entity.  For simplicity sake, assume Company A deposits $25m USD in London and uses the remaining $75m for other purposes, this would show a corresponding rise in the foreign currency overseas deposits of Chinese banks.

Last major point is that the numbers presented match the swaps story very closely.  For instance, if companies with mainland operations want to move capital and the Chinese bank requires them to deposit some percentage of the proceeds in an overseas account, this would match the growth in overseas foreign currency deposits and the growth in assets based upon a “foreign currency” loan being made overseas.

The discrepancies noted between the net asset positions and growth in foreign currency deposits are not just inconveniences or rounding errors but significant problems with the story that this net asset rise.  What is important to note is that everything that has been shown is perfectly consistent with an increase use of swaps to fund the growth of foreign asset purchases.  While it remains perfectly valid to ask who is the counterparty and he right in noting that list is pretty short, that in no way changes any part of the analysis and actually

Are Chinese Bank Recapitalizations Monetary Neutral?

So a couple of people that I know and some that I don’t know zeroed in, in my last post, on a couple of monetary issues.  They raised some important questions and so I think it is important answer them as best I can based upon what I think we observe in China.

The basic idea that is being objected to is that bank recapitalizations can be monetary neutral.  Before we even discuss the mechanics of bank recapitalizations, it is important that everyone knows what we mean by monetary neutral.  Assume country A has a fixed exchange rate and decides to recapitalize their banks. If they increase the base money supply by a non-trivial amount that could cause pressure and ultimately some form of a devaluation/depreciation.

Now it is very important to note that a bank recapitalization can be monetary neutral but can also violate the concept of monetary neutrality.  So in other words, it is entirely possible that they are right that a bank recapitalization could be monetary neutral, but it could also be false.

Let me give you two very simple examples to illustrate the difference.  Assume a bank needs to increase its capital base, for any number of reasons, and does a secondary rights offering selling shares to the market to meet capital adequacy ratios.  If they offer the shares to the market and the market buys the shares, there has been no increase in the money supply. Investors with existing capital chose between   different investment options. This simple example could be expanded to cover a pre-emptive, hypothetical, type of recapitalization where the Chinese Banking Regulatory Commission (CBRC) orders all banks in China to sell shares to the market to ensure high capital levels. In these instances, there has been no increase in the Chinese money supply. We have not violated the principle of monetary neutrality.

However, it is also very easy to violate the principle of monetary neutrality.  Assume now that a bank has made a bad loan but the government wants to ensure continued lending and investment growth.  The government does not want the market to buy the shares because that would divert capital used for other investment purposes and it would dilute the governments shareholding.  To solve this problem the central bank prints money to buy assets of some kind from the bank to give them capital continue lending. This results in a tangible and could be material increase to the money supply.

To make this example tangible, assume the bank has $1,000 in loans, $900 in deposits, and $100 in capital (I am being very very simple here). If the bank has a shock with NPL’s rising to 10%, assuming depositors lose nothing, the banks capital of $100 is wiped out.  However, the central bank prints money and offers to buy the bad loan at face value of $100. The bank gets $100, returns its NPL ratio to 0%, and can resume lending. The money supply has gone up but the objectives of continuing to lend with functioning banks has been achieved.

Let’s briefly consider similar but very importantly not identical situations.  Both the Bank of Japan and the Federal Reserve have engaged in quantitative easing whereby they print money to buy sovereign debt issued by their respective governments.  The European Central Bank has engaged in a similar strategy buying a variety of sovereign and high credit quality corporate debt.  Absolutely no one disputes these actions are not monetary neutral. They are in fact quantitative easing.  If the PBOC is printing money engaging in balance sheet expansion to fund monetary easing, even if it is purchasing assets from banks or engaging in quasi lending to banks, this will count as monetary easing violating monetary neutrality.

Forecasting into the future is always difficult and it is entirely possible that if there were some type of “event” where these mechanisms would be discussed, it is possible that China could choose a mechanism that did not violate monetary neutrality.  However, if we look at recent Chinese behavior, we have a very good example that clearly violates monetary neutrality.

In what I believe is one of the most overlooked events in recent Chinese history and will likely in time occupy a more central focus of analysis, Beijing conducted a full fledged bailout of local governments and the bad debts Chinese banks held.  The banks who held vast sums of debt, with even much of it now unlikely to be repaid, were ordered to convert short term high interest loans into 10 year low interest bonds.  As a simple example, a 1 year 7% loan became a 10 year 3% bond. If these debts blew up, this would have had an enormous negative impact on bank capital levels and restricted their ability to lend but also the bailout plan would have restricted their ability to lend.

Beijing came up with a solution when the bankers resisted. Local government bonds could be sold to the PBOC for money that would then be used to make new loans.  This solution effectively wiped out local government debts and “recapitalized” banks by relieving them of bad debts allowing them to speed up new lending.  It should come as absolutely no surprise that lending in China really surged roughly 6-9 months after this plan was first announced.

However, and very important to return to our earlier discussion, it completely violated the principle of monetary neutrality.  The PBOC was printing money to buy assets from the banks.  How do we know this? Chinese data tells us this is exactly what is happening.

In January 2015, prior to any discussion of a local government bailout, PBOC claims on other depository corporations stood at 2.6 trillion RMB but by April 2017 that stood at 8.45 trillion. That is an increase of 5.9 trillion RMB or $852 billion.  In other words, the PBOC has spent the last two years buying large amounts of assets from Chinese banks and importantly exactly as it said it would. This was announced and agreed to by Chinese banks to sell the PBOC bad debts. This is not a shock.

Let me put this number in a little perspective for you in a variety of ways. This 5.9 trillion RMB is equal to 21% of the growth in total loans during this time frame.  This is equal to 38% of net capital for the entire commercial banking industry in China.  This is equal to 1,098% of the growth in M0 over this time frame.  As a slight tangent here, I use M0 here rather than M2, or other potential measures, as the PBOC controls the printing presses to print RMB but they do not directly control for our purposes here broader money measures such as deposits which are also related to history and asset prices. These broader measures are outside the immediate and direct control of the PBOC.  In short, as we can see the purchases of the PBOC are significant by any related financial measure.

Probably the biggest impact of this shadow “recapitalization” is that the banks did not have to declare bad loans reducing their capital base and lending growth. By selling via some form of a repurchase agreement, the banks were able to maintain that loan on their books as a standard loan.  Just as other forms of asset purchases by central banks keep capital costs low and stimulate investment/public spending, so the PBOC purchases here are designed to do this using the banks as conduits.

Now I can already hear an understandable objection. This is not a recapitalization because the PBOC is just holding assets as a collateral they are not recapitalizing the banks.  Possible (which I will return to in a moment) but in the short term, irrelevant for what we are discussing here.  In the short run, the PBOC is clearly violating the principle of monetary neutrality.  Just think of how big the drop in lending would have been, not even assuming second order/dynamic effects, from just backing out the PBOC purchases.  Assuming a not insignificant numbers of these pledged assets are bad assets, think of what that does to bank capital.  Banks are making loans with money that did not previously exist printed by the PBOC to further stimulate lending. We have violated monetary neutrality.

The question I briefly circle back to is whether the PBOC is actually recapitalizing.  I would humbly submit a couple of points of importance here that violate the presumption of standard central bank lending that lead us to the conclusion this is a type of recapitalization.  For starters, we cannot consider 228% growth in just over two years as standard and normal growth.  This is clearly far outside the bounds of normal financial growth even by Chinese standards.  Then, and though we cannot say for certain, given that the most likely scenario is that the PBOC is buying distressed, bad, or low quality loans, this absolutely has to count as non-normal lending practice.

However, probably the most important question is what is the nature of the capital here? By that I mean, does the PBOC seem likely to pull credit and the what happens when the underlying loan is either repaid or is defaulted on? On the first part, I believe it is extremely unlikely that the PBOC will pull the credit facility because this was the whole point of the local government bailout.  Banks would only go along if they had a place to effectively dump these low yield junk/NPL bonds. More importantly is whether this is a “recapitalization” or just standard asset lending by central banks. Given that the PBOC is accepting, most likely, very low quality debt, this is not standard central bank lending.

The question then focuses on the capital supplied by the PBOC.  If the underlying debt is repaid, then the PBOC is repaid and no “recapitalization” has taken place.  So then what about the scenario if the underlying debtor defaults?  In most every system I am aware of and I would assume the same for China, though I cannot say for sure, during a repo, which is likely the type of transaction taking place or a similar transaction when a lender pledges a fixed income security as collateral to borrow if the debtor of the fixed income security defaults while the security is used as collateral for borrowing, the original lender can be held liable for the bad debt. Put another way, if ICBC holds a bond of province X, ICBC takes that bond to the PBOC and sells that bond agreeing to repurchase it in say 5 years, if the province defaults during those 5 years, the PBOC can pursue ICBC to make good the bad debt portion.  Here is what I think is important: assume province X defaults on the bond ICBC sold to the PBOC, I think the probability PBOC would pursue ICBC for damages to recoup losses as above zero but very very low.  In this scenario, the PBOC has effectively recapitalized a bank absorbing the loss they should have suffered.

Circling back to our original questions, while I think it is possible that recapitalizations can be monetary neutral, in China this is clearly not the historical case and would I believe be unlikely in the future. Furthermore, while not all of the new money supply will be “recapital” into banks as some of the securities held by the PBOC sold to them by banks will be repaid, I would deem it highly unlikely that the PBOC would pursue bad debt claims against Chinese banks in the event of default. Banks would in this case receive a backdoor recapitalization by not suffering losses they should have suffered.  It is quite likely, the PBOC is the new Superbad asset management company for China.

Will China have a Crisis Part I

Probably the most common question about China these days is whether China will undergo a financial crisis? The China bulls argue that China has lots of FX reserves, can print its own money, high savings, and a strong regulator that will ensure China can contain a crisis. The bears point to a factors like the speed in the increase of the credit to GDP and the level of credit to GDP so support their case.

I find points of validity in both cases but neither one ultimately satisfying.  I think the major problem with each is that they find broad headline points of commonality or difference with either 2008 subprime or 1997 East Asia financial crises and claim that China is just like or totally different.  This is part of why I find some aspects that are valid in each, but also fundamental shortcomings.

What I am writing here is an attempt to talk through or think out loud about what will happen to China.  Let me emphasize that these are not predictions but rather trying to work with a combination of economic theory and Chinese empirics what may happen, teasing out more detail from the two major sides of this debate.  Today I will start with the bear case that China will ultimately have a financial crisis or hard landing.

The major reason not to believe the bear case is political: Beijing will not allow a crisis is political due to the potential blowback ramifications.  In 2008, the United States and other countries made clear and conscious decisions to not bailout firms and households.  We can argue over whether they should have, whether the divergent approach to Fannie and Freddie vs. Lehman, or whether it should have targeted asset levels via home prices for consumers, but the take away is simply that the United States made a clear and conscious decision to not broadly pursue such policies.  The United States generally allowed asset prices to fall, firms to fail, and households to be evicted or declare bankruptcy.

I do not believe Beijing is willing to incur the risk of suffering such a financial downturn running the risk of allowing such an event.  Assume for one minute a financial crisis hits China. That is literally a once a century event.  Probably bigger economic and financial event that the fall of USSR with larger international consequences.  Beijing is acutely aware that Moscow made it to the 13th 5 year plan and Beijing is in the middle of its now.  Xi has built his entire administration around preventing a weak China and this type of event.  If China suffered a financial crisis, this would likely end Communist Party rule in China with major consequences for ruling elites.

This does not mean that Beijing will make good policy in the interim to prevent such an event, in fact quite the opposite and we should expect Beijing to take all steps to avoid a crisis without addressing the fundamental problems.  In fact, this matches very closely how we see Beijing behaving.  For all the talk of how they intend to deleverage, Beijing has clearly prioritized growth stability above deleveraging.  For all the talk of improving risk pricing and allowing defaults, has always in practice resulted in government and SOE bank led bailouts of companies in default.  Concerned about how a bankrupt firm with large losses imposed on banks and investors would be perceived in the market place, Beijing acting to avoid a crisis without addressing the fundamental problem.

In fact, this policy path, which I believe broadly fits what we see Beijing doing, delays inevitable adjustments but stores up increasing large amount of risk.  Again, this broadly fits what we see happening.  Capital is being spent to delay ultimately inevitable reforms but in virtually every case, it is merely storing up risks.  This makes the financial position increasingly tenuous and risky as we move forward in time.  Now we have a point in the bears favor, there may come a time at which the risks become simply unmanageable provoking a crisis, but currently it seems unlikely we will have a Chinese crisis in the near future.  There are clear signs of stress across a variety of sectors in the economy, however, I do not believe these signs are so dire that Beijing cannot prevent a crisis for the forseeable future.

There are however, a host of smaller reasons that the bears could be wrong.  In real order, they could be:

  1. Estimates of Chinese non-performing loans are overstated. Even Chinese securities firms have come up with estimates of 10%, which I would personally use to establish the baseline estimate.  In reality,  we simply have opaque ways of estimating what might the true number of NPLs be.  Could they be the higher range estimates of 20%+? Sure but do we really know for sure? No, we don’t and we need to leave open the possibility that many are wrong on this.
  2. The structure of debt within the economy matters and may signal less risk of crisis than is understood. Many analysts focus on the total debt level but omit more commonly that most of this is corporate with relatively small levels of government and household debt.  What if corporate debt as a percentage of GDP stagnated but household and government continued to rise over the next 5-10 years? That would imply that total debt as a percentage of GDP could continue to rise for some time.  This would also allow investment and consumption to rise as a percentage of GDP if the public sector assumes greater responsibility in investment and household consumption increases.  The problem with this story is that it implies enormous debt levels in say 5-10 years with very high levels of financial fragility.
  3. Real estate is less of a financial risk and more of a social risk than is appreciated. While implied marginal leverage rates on new purchases of housing in China is rising rapidly, the overall debt associated with housing in China, especially when placed against the current estimated value, is minimal.  Financially, this would seem to imply that there is little actual risk of a crisis being caused by a downturn in the real estate market.  However, it is a poor analysis to conclude there is no risk from a real estate price decline.  There are two specific factors.  First, real estate prices might be the most concerning trigger for social instability.  If for instance, there was a 30% decline in real estate prices in China, I have little doubt that there would be wide spread social urban instability.  That presents a wide range of risks that the Party is simply not willing to tolerate and consequently will do everything to prevent.  Second, depending on the exact estimate you believe an astounding amount of Chinese economic activity is tied to real estate.  On average over the past few years, probably almost 50% of government revenue, 20-30% of GDP, and tied to a grossly disproportionate share of lending in different ways.  Consequently, while real estate does not represent the first order financial risk that the 2008 subprime crisis did, it absolutely represents second order or indirect impact on potential downturn in real estate development, lending, and potential defaults from colleateralization drops.
  4. The transition to a service and consumer driven economy is better than presumed. I find this argument unsatisfying.  Empirically, there appears strikingly little growth in consumption service focused industries.  For instance, travel and hospitality within China, which represents approximately 98% of Chinese travel, flat to low single digit growth.  Virtually the only service sectors enjoying demonstrable real and nominal growth are financial services (for very concerning reasons) and logistics/supply chain/postal services.  However, while it is wonderful for the Chinese consumer, the growth in logistics/postal services are doing little more than cannibalizing activity from brick and mortar retailers.  The marginal boost to growth, after accounting for the cannibalization, is minimal.  There is no evidence of double digit or near double digit wage growth that would drive the consumption/retail sales growth Beijing touts.

The picture we are left with, seems to be an economy that has a wealth of problems, is driven by credit, but one that is not as of January 2017 on the verge of eminent collapse.  Furthermore, each of the supposed arguments of why China will continue to thrive have major problems.  Additionally, if we carry forward the counter argument of the bulls to counter the bear crisis argument, we left within 1-3 years of astoundingly perverse outcomes.

China may be able to prevent a financial crisis through capital controls but that would require hard draconian capital controls.  China may be able to prevent a financial crisis by having the PBOC intervene but that would require widespread debt monetization which brings a whole host of problems on its own and assumes as “soft” or “semi-controlled” debt crisis.  Absolutely neither of these should be considered positive outcomes.  That would be like saying someone had a quadruple bypass and is bedridden for 6 months but didn’t die.

Next week I’ll consider the argument, this is all overblown and China will continue to grow rapidly for the next 10-25 years.

Some Thoughts on Chinese Financing Growth: Playing Whack a Mole

The focus on the recent strong growth in headline financing growth has raised concerns about underlying demand for credit and continued reliance on investment to drive growth.  However, the headline data fails to capture the important underlying trends that are important grasp the change in Chinese financing.

Beginning with the aggregate YTD changes in financing, in Figure 1, we actually see that most types of financing are up strongly.  Only two categories of financing are down in 2016.  Undiscounted bankers acceptance and foreign currency loans are down YTD in 2016 in absolute terms.

Virtually all of the decline in aggregate financing to the Chinese has come from the decline in bankers acceptance.  All other sources of financing are up robustly to strongly.  What makes this precipitous drop in bankers acceptance notes is the lack of evidence as to where it is going.  Bankers acceptance should be used as a type of receivable’s financing.  Consequently, if the outstanding amount of bankers acceptance is falling so rapidly, we should see a corresponding drop in outstanding receivables, however, there has been no drop in receivables.  In fact, net receivables according to official statistics are up year to date 9.4%.  This makes the supposed drop in bankers acceptance rather puzzling.

If we plot this on to growth in various forms of financing growth, RMB loan figures which grow very closely to the total financing numbers are the smallest number with other forms of financing exploding.

For instance, the combination of trust and entrusted loans have more than doubled through August from the same period in 2015.  Foreign currency loans, as a share of the total new financing in 2016, have dropped by an almost insignificant amount.  While this has a not insignificant impact on FX related flows and pressures on RMB, it is almost irrelevant to the stock of financing in China.  Foreign currency loans dropped by 412 billion RMB against total new financing YTD of 11.75 trillion RMB or only 3.5%.

It may be possible, though we have no hard evidence to support this, that the decline in bankers acceptances are being made up for increases else where in the total pool of finance.  The total absolute increase excluding bankers acceptances and loans comes very close to matching the absolute decline in bankers acceptances.  If we sum trust, entrusted, bond, and stock financing change from August YTD 2015, we have a number of 2 trillion RMB compared to the decline in bankers acceptance of 1.8 trillion RMB.  If this is what is happening, this appears to signal that a lot of bank based capital is being shifted into non-bank financial institution lending.

Given that total lending in 2016 to non-bank financial institutions has totaled 1.8 trillion RMB, nearly matching the decline in bankers acceptance, there is some reason to believe that banks are shifting their lending practices  to meet new regulatory requirements about bankers acceptances.  Again, we cannot say this for certain, but there is some evidence that indicates it may be happening.

Follow Up to BloombergViews: Shadown Banking Risks in China

I wanted to write a follow up to my piece from BloombergViews on shadow banking risks in China with a little more technical focus on a few things.  As usual start there, and every thanks to them  and my wonderful editor, before coming here.

  1. Shadow banks is really a catch all phrase for non-bank financial institutions that really encompasses quite a variety of lending types. Trust companies which actually make a couple different types of trust investments.  Wealth management products that can both be created and sold by mainline banks as well as on behalf of third party firms. P2P firms which hold little or no capital but act merely as a platform facilitating financial flows profiting by taking some type of fixed fee.  There are a myriad of products that are designed in a myriad of ways so talking about this industry as a monolith makes no sense.  There are only a couple generalities that are worth mentioning at this point: the products are short term and can cover almost any underlying asset in almost any form from debt to equity.
  2. Many people think of shadow banks in one area and traditional banks in another but this is absolutely not the case. In many ways, shadow banks and traditional banks are almost indistinguishable from each other.  There are two specific ways this happens.  First, shadow banks get large amounts of funding from traditional lenders.  This happens through a variety of different funding agreements from bank purchases of investment securities from shadow banks to wholesale funding agreements.  Bank holdings of the non-loan investment holdings via a number of specific balance sheet line items have exploded over the past few years.  Many banks even admit to making large strategic shifts away from direct lending and into these new products.  Second, commercial banks even have agreements to act as a distributor or sales staff for shadow banking firms.  Consequently, not only are banks purchasing, funding, in some cases directing the funding/locating borrowers, they are then selling for the shadow banks.  Even if there is not a direct ownership agreement, this creates an enormous conflict of interest whereby the bank and shadow bank are essentially almost indistinguishable entities.  (If you want the best example with details that explain the financing practices I am talking about and the overlap, read this example of China Credit Trust from 2014)
  3. Mainline banks are engaging in this behavior for two very simple reasons: financial and regulatory arbitrage. Financial arbitrage means banks earn a higher rate of return from lending to shadow banking customers or the shadow banks themselves.  Even though lending rates have officially been liberalized, in practice this has not really taken place.  There has been little movement in the bank lending rate and customers are not being judged on the risk they present.  As an obvious example, local government debt which banks are being forced to buy is trading at yields lower than sovereign.  If this debt was not priced at a government mandated price, it would clearly be yielding significantly higher.  Banks have even put into IPO prospectuses that they are moving into holding a higher level of these shadow bank products to earn higher rates of returns.  Banks that cannot earn what they deem to be a reasonable risk adjusted return are moving into these other products as a Chinese version of chasing yield.  Regulatory arbitrage is much simpler. If a bank makes a 100 RMB loan to a coal company, they have to report to the banking regulator that they have 100RMB at risk and set aside the appropriate capital to meet their capital adequacy ratios.  However, if they purchase a 90 day security from a shadow bank for 100 RMB, they do not have to set aside anything because Chinese banking regulators allow loans or investment security purchases from financial institutions to have a 0% risk weighting.  In other words, a bank can make the same 100 RMB loan but if they make it to a coal company they have to set aside capital but to a shadow bank, they do not need to.  This has led to many loans made via trust companies that arrive at a specific company through a trust company that the bank does not weight as a loan because the loan is technically made to a trust company rather than the coal company with the trust company lending the money to the coal company.
  4. All of this detail leads to a handful of risks. First, shadow banks are intricately linked with the entire financial sector.  One way to think of this might be similar to the subprime problem in that there were spillover effects from the non-bank financial institutions to the banks.  It is absolutely incorrect to think of Chinese mainline and shadow banks as having some type of dividing line separating them when they are in reality incredibly linked in a variety of ways.  The spillover risks are enormous.  Second, commercial banks enjoy a variety of privileges their non-bank financial institutions do not such as deposit insurance to state ownership.  This provides an implicit government guarantee which shadow banks do not have.  It is dangerous for the government and the general investment population from institutions to retail to think of such a neat dividing line given the overlap in funding, sales, and clients.  In other words, Beijing will ensure that any type of crisis at a major bank like ICBC never becomes a problem maybe never even heard of.  Shadow banks do not have that luxury and could easily trigger liquidity or spillover risks onto the larger banking sector. Third, shadow banks relying primarily on short term funding face very real liquidity risks.  If they cannot get new funding every 30-90 days they will collapse.  Fourth, the large variety of shadow banks could very easily trigger a bank panic given their widely recognized problems of funding causing liquidity to dry up across the shadow banking sector.  This would force Beijing or the banks to step in and guarantee the shadow banks.  Fifth, there is a reason that banks are moving so many of their risks off their balance sheets and on to the shadow banks.  As the old saying goes: if you are at a poker game and you don’t see a sucker, you’re it. There are amazing cases of banks doing this, some of which they ultimately bailed out but anytime a bank moving risks that fast, pay attention.  Sixth, bank risk management practices are weak at best, so you can imagine what the non-bank financial institution risk management practices are like.  There are stories almost daily of different shadow banks going bust and private financing disputes were up more than 40% in 2015 with 2016 expected to be another bumper crop for Chinese lawyers.  Whether it is straight up fraud or simply non-existent due diligence practices on loans, there is a reason that finance is so much more difficult for smaller players.

What is Really Worrying About the Chinese Credit Bubble

Lots of ink or ones and zeros in today’s world has been spilled about the explosion of credit propping up the Chinese economy.  So much so that I really will not rehash this plot of land.  I think what is interesting, and really most worrying, is how little impact the credit explosion has had on real economic activity.

The current credit explosion we are witnessing in China is bigger in absolute term than some of the post-GFC pops in credit and by some relative measures even bigger in relative terms.  Only the most resolute of perma-panda and the Communist party press office Xinhua editorial board dare argue that China isn’t propping up growth with more credit.  Most everyone understands the credit side of the story.

What is most interesting is how this impacting the real economy. Here is what’s happening with that explosion of credit: nothing. All this money that is being tossed around like an after party  at a rap concert is not generating any notable uptick in activity.  In fact not only is economic growth not accelerating, activity continues to decelerate (grow but at an increasingly slower rate).

Let me rephrase what is happening: China is stepping on the gas pedal, they put jet fuel in the gas tank, and the car does nothing but rev the engine and coast forward slower and slower.  All this credit is accomplishing is a slower rate of decline.

This has a number of important implications.  First, this implies that the Chinese economy is in much worse shape that most outsiders wish to acknowledge.  You don’t give an economy this much stimulus unless you are really worried about the fundamental level of activity.  The PBOC may not give the type of minutes and commentary that the Fed does, so if we look at their monetary actions as a reflection on the confidence in the economy, they are telling us that they believe the Chinese economy is incredibly weak.  If you think this overstates the degree of weakness, imagine that the PBOC and Beijing had not decided to drop cash everywhere.  What would have happened?  My favorite statistic is that total social financing rose almost 16 trillion RMB in 2015 but nominal GDP grew only a little more than 4 trillion RMB.  That is a tiny boost to GDP relative to the amount of money that was poured in.  Imagine what GDP would have been if Beijing had not been dropping money from helicopters.

Second, the near complete lack of real economic response to monetary stimulus is telling us very clearly what is and is not the problem in the Chinese economy.  The problem is most definitely not a lack of access to credit, investment, and financing.  The problem is that money is not being put into tangible projects but rather being used to keep old loans from going bad and speculating in commodity/real estate/stocks/bonds/egg futures/online purse startups/Kanye’s new record/hotel chains/or whatever new investment fad the giant ball of money targets this week.  It is such standard practice in modern monetary economics to when the economy slows just push money but I think there are so many micro-problems in the Chinese economy that are not being addressed.  The problems are fundamentally different and shoveling money and most importantly new debt are not solving anything.

Third, this then tells us about the policy response which implies that more money is not the solution. Part of the transition that China needs to face is that some industries need to seriously reduce their capacity while others need to increase and potentially most difficult, labor needs to transitions to new productive activities.  Arguably, the biggest signal in any transition (i.e. pushing people/capital away from some activities and towards others) is the price mechanism which China is avoiding as much as possible.  The credit boom which drives up commodity prices keeping dying firms in business a little longer and unproductive on the payroll of a dying firm a little longer is a short term solution.  The flowing money is merely suppressing the price signal that tells labor and capital where to go for new opportunities.  Does anyone seriously think that the recent run up in steel prices is anything more than a blip on the screen over the next five years? Of course not, but it gives a misleading signal about the health of the industry, long term prognosis, and labor/capital allocation in a transitioning economy.

Fourth, the lack of real economic activity despite the flood of credit implies that the Chinese economy might be entering an increasingly concerning state.  As an example, Total Social Financing has increased through Q1 2016 37.4% over Q1 2015 and by 10% in 2015 but evidence of this flow of credit into real activity in corresponding amount is difficult to find.  While real estate development has ticked up, it is nowhere near the level similar to this flood of credit.  We do not see a corresponding increase in output of industrial outputs that would accompany such an increase in total social financing or fixed asset investment.  For industries with data, FAI is up 11% Q1 2016 from Q1 2015 but physical output of most industrial output is up minimally if at all.  Crude steel and steel material output is down 3.2% and flat at 0%.  We see similar numbers.  Given the explosion in TSF and significant growth in FAI but utter lack of pick up in real activity, this implies that financing is being used simply to prop up historical investment.  It also implies that there is simply no demand anywhere in the economy.  We know cash flow is slow with rising NPLs and rapidly growing receivables.  That simply isn’t positive.

The rapid growth of credit is starting to worry even the most bullish but what is even more worrying is the near complete lack of responsiveness of the real economy to the monetary stimulus.

Follow Up on Chinese NPLs

I wanted to make some brief follow up points that will be a little more technical than what appears in my BloombergViews piece.  As usual start there and finish up here.

  1. I am decidedly pessimistic about the state of the Chinese economy and finances, but I really do not see a near term risk of what we would think of as a financial crisis. China is saddled with enormous sums of bad debt, people are taking their money out of China, surplus capacity is simply astounding, and cash flow growth through the economy is hovering around zero.  There is very little to be positive about, but I see little risk this year of some type of crisis.
  2. I do not believe we have reached or are even at a near term inflection point, barring some exogenous shock, where financial conditions so escape policy and bankers control that the path is set.
  3. Maybe my biggest concern is the inability of bankers and regulators to face the problem. I actually should just say regulators as they are clearly the driving force at play.  As one simple example, after all the talk about deleveraging in December, debt growth has pretty much exploded.  Not only are they not deleveraging, they are adding fuel to the fire.
  4. The supposed 1 trillion RMB debt for equity swap does not even begin to address the size of the problem. Caixin highlights one steel producer, let me say it again one steel producer, with 192 billion RMB in loans in cannot pay with its largest subsidiary not having paid interest since 2011.  Let’s make a simple assumption that you wanted to keep this steel company in business and complete a 100% debt for equity swap.  You would immediately use up 20% of the 1 trillion in capital on one company.  Even the Xinhua has noted the larger problem of unprofitable steel firms which made only a small profit in 2014 and a large loss in 2015.  Accounting for the debt and overstatement of financial health, 1 trillion in capital will not even begin to address the problems in one industry.
  5. There are so many problems associated with some of proposals I have heard floated but one of them what happens to the firms and industries after completing the debt for equity swap. For instance, if all firms stay in business with lower debt burdens the only thing that will happen is an even more heated round of cost cutting.  That is not a real solution.
  6. Furthermore, many are counting on shutting down a firm and selling the assets at the price they are booked for on the company’s balance sheets. History tells us in these situations corporate assets on sale in a bankruptcy type situation even when bought free and clear simply are not worth anywhere near their booked valued.  Then add in the massive surplus capacity and there would likely be little secondary market for these assets.  Anecdotally, from people I have talked to the going recovery rate seems to be about 20-25 cents on the dollar.  This would imply sales, to use a round number, at 10-15 cents on the dollar to investors.  The steel company in the Caixin article supposedly has 290 billion in assets.  Let’s suppose these are all real assets, which as the story notes does have some very real questions, this would imply a sales rate of 30-40 billion RMB, an enormous haircut.
  7. The basic strategy I think of Chinese policy makers is to try and replicate their 2003 strategy of growing their way out of the problem. As I wrote in a paper, there were banks going public in December 2014 with long term bad debt obligations that they used IPO proceeds to payoff.  Economic growth post 2000 was such an ahistorical event that I think it is a highly risky expectation to grow your way out of a mountain of bad loans two times in a row.  Even just think of the strategy of accomplishing this.  Will these bad loans on steel companies be paid off in 5-10 years just putting them off to the side? Unlikely.
  8. The ones that get mentioned the most are coal, steel, and real estate but this same concept holds for virtually any industry in China. Shopping malls, chemicals, and energy suffer from high debt and over capacity.  Even assuming asset prices underpinning assets remains high, which is highly unlikely if there is a significant push to reduce over capacity or reign in lending, large amounts of the Chinese economy need to address the NPL problem.
  9. One of the biggest issues that no one has addressed is what do Chinese banks do with the equity? What secondary market would they sell the equity into?  If they are swapping debt for equity at par, will they be required to mark to market?  What requirements will there be on lending as the equity owner?  This seems like a bad alternative as much of the equity will be near worthless and require haircuts of 90%.  Why then go through the debt to equity swap process if the likely outcome for either the loans or business is large write downs?  This implies that the intention is to try and salvage the value of these assets which would depend on large amounts of new lending.

Why China Does Not Have a Trade Surplus

Life has few certainties except for death, taxes, and large Chinese trade surpluses.  The expected large Chinese trade surpluses are always referred to as both proof of the strength of the Chinese economy and its financial foundation as money continues to flow in.  In nominal RMB terms, the trade surplus amounted to 5.5% of GDP or 79% of total GDP growth.  In other words, in 2015 China is almost entirely dependent on maintaining a large trade balance to drive GDP growth.

However, what if the assumed trade balance did not actually exist?  In fact, how would it change our understanding of the Chinese economy and financial markets if the assumed trade surplus was actually a trade deficit?  Unfortunately, this is not a counterfactual but the reality.  China is running a small trade deficit.

The widely cited international trade data is provided by Chinese customs records.  The value of goods leaving and entering and China is recorded by the Customs Bureau.  According to Customs data, China imported $1.69 trillion (10.45 trillion RMB) and exported $2.27 trillion (14.14 trillion RMB) for a resulting trade balance of $593 billion (3.7 trillion RMB).  These often repeated numbers form the basis for why China is running a large trade surplus.

Before explaining why China has no trade surplus, it is important lay some related groundwork.  By now China watchers knows about the practice of trade misinvoicing.  This is the practice where, as originally executed, capital was either moved into or out of the country based upon fraudulently invoicing an import or export.  For instance, by over invoicing an export, capital can flow into China as the foreign counter party is over paying for the good and vice versa for imports.

To take one example, of trade between Mainland China and Hong Kong, there are significant discrepancies between the value reported to Chinese customs and Hong Kong customs.  Hong Kong reported imports from China worth $255 million USD but China reported exports to Hong Kong of $335 million USD.  The 31% difference in customs prices, or $79 million, is too large to be unintentional and acts as a capital inflow into China.  Conversely, China reports $12.8 billion USD of imports from Hong Kong but Hong Kong only reports $2.6 billion USD of exports to China.  The 385% difference is far in excess of the low mid to single digit invoicing discrepancies that are standard in global trade.  Consequently, the $10.1 billion USD in over invoiced Chinese “imports” acts as a capital outflow from China.

Misinvoicing contributes a not entirely insignificant share to unrecorded capital inflows and outflows.  However, Chinese authorities have become much more aware and concerned about these issues and  gone through various waves of cracking down over this issue.  Furthermore, the aggregate sums here are not enough to move the RMB and cause the currency pressures we are currently seeing.  In fact, misinvoicing is merely the beginning of the financial flow problems in trade with Chinese innovation taking it a step further.

China, as a country with strict currency controls, maintains records on international financial transactions sorted by a variety of categories.  For instance, there is data on payment or receipt of funds by current or capital account, goods or service trade, and direct or portfolio investment.  For our purposes, this allows us to compare in a relatively straightforward manner, how international payments are flowing compared to the customs reported flow of goods.

The differences in key data surrounding trade data is illustrative.  Chinese Customs data reports goods exports valued at $2.27 trillion, with SAFE reporting goods exports of $2.14 trillion but Chinese banks report receipts of $2.37 trillion.  In other words, funds received for exports of goods and services or about $100 billion higher than reported.  At 4-11% higher than the Customs and SAFE reported values this is slightly elevated, but given expected discrepancies in the mid-single digits, this number is slightly elevated but not extreme.

The differences between import and international payment data, however, is astounding. Whereas Chinese Customs reports $1.68 trillion and SAFE report $1.57 in goods imports into China, banks report paying $2.55 trillion for imports.  In other words, funds paid for imported goods and services was $870-980 billion or 52-62% higher than official Customs and SAFE trade data.  This level of discrepancy is extreme in both absolute and relative terms and cannot simply be called a rounding error but is nothing less than systemic fraud.

If we adjust the official trade in goods and services balance to reflect cash flows rather than official headline trade data as reported by both Customs and SAFE, the differences are even worse.    According to official Customs and SAFE data, China ran a goods trade surplus of $593 or $576 billion but according to bank payment and receipt data, China ran a goods trade surplus of only $128 billion.  If we include service trade, the picture worsens considerably.  China via SAFE trade data reports a $207 billion trade deficit in services trade.  Payment data reported via SAFE actually reports about $42 billion smaller deficit of $165 billion.  In other words, the supposed trade surplus of $600 billion has become a trade in goods and services deficit of $36 billion.  Expand to the current, through a significant primary income deficit, and the total current account deficit is now $124 billion.

There are two very important things to emphasize about these discrepancies.  First, the imports customs and payment discrepancy is responsible for essentially all of the discrepancy between payments and customs.  Neither goods exports or differences between service imports at customs and payments explain the difference.  In fact, service is underpaid according to payment and customs data.  Second, if there was a more benign explanation, we would expect to see symmetry between various categories.  Rather, we see most categories reconciling close enough and one channel, conveniently enough one that funnels capital out of China, enormously mis-stated.

This discrepancy between official reported trade data and bank payments is a relatively new phenomenon but has been growing rapidly and reveals important details about flows into and out of China.  For instance, since 2010 China has an aggregate trade in goods and services surplus based upon payments of 1.9 trillion RMB; however, since 2012 an aggregate deficit of 120 billion RMB. 2010 and 2011 were the only years where China ran a trade in goods and services surplus using payments data rather than customs data.  Expanding to consider the current account significantly worsens the outlook.  From 2010 to 2015, China has run a current account surplus of 462 billion RMB but from 2012 to 2015 ran a deficit of 1.44 trillion RMB.  The reason for the shift is simple.  In 2012, China freed international currency transactions made through the current account creating an enormous asymmetry.

There are a number of important conclusions and implications of the data presented here.  First, if we adjust the Chinese traded good surplus on a cash flow basis and include the trade deficit resulting in a net export deficit, Chinese GDP growth in 2015 grew only 0.3%.  If a positive trade balance in economic accounting directly adds to GDP growth then a deficit directly reduces it.  Consequently, swinging from a goods trade surplus of 5.5% of GDP to a goods and services trade deficit of negative 0.3% of GDP has an enormous impact on GDP growth rates.  There is a key distinction here that is important to note and that is on a cash flow basis.  Economic accounting holds that GDP grows because when running a trade surplus, additional cash flow is received than is expended.  This leads to higher investment through savings. In 2015, financial flows indicate this did not happen and there was not trade surplus on a cash flow basis due to the discrepancy between Customs and SAFE reported trade in goods and services values and what banks paid.

Second, the impact on real GDP and output is currently unknown.  There are numerous reasons to question the veracity of numerous aspects of the data which would change our understanding of the data.  For instance, there are examples of goods round tripping into and out of China designed solely to facilitate implicit capital transactions.  Given the enormity of the discrepancy we see in payments for imports, we cannot rule out that a not insignificant amount of trade was either round tripping or phantom trade.  As physical output of many products from industrial to consumer only increased in the low single digits, this would match closer the implied Chinese growth rate of 0.3%.

Third, this sheds new light on the state of Chinese finances and RMB outflows.  For instance, the differential between Customs and bank data reveals rising outflow discrepancies since 2012.  While many have begun to worry recently about rising pressure on the RMB, it is clear that outflows from China are long lasting, large, and completely domestically driven.  In 2015 the capital account maintained healthy levels with the outward direct investment balance in a small deficit of 28.3 billion RMB while the securities investment balance was in an even tinier deficit of 2.9 billion RMB.  Consequently, calls for “temporary capital controls” or attributing it to a recent increase in outward direct investment reveal a profound misunderstanding of what the problem is. There is nothing temporary, foreign, or speculative about RMB outflows.  In fact, quite the opposite.  It is domestically driven long term capital flight which should change the framework of what solutions are called for in managing RMB policy.

Fourth, the change in the current account deficit is a major driver in changes to PBOC foreign exchange reserves.  While these are disguised capital outflows, for accounting purposes it is showing up in the current account statements.  Consequently, while China shows only small capital account deficit of $75 billion and a cash flow current account deficit of $121 billion, this shift largely explains the currency pressures on the RMB.  If you look simply at the Customs reported trade surplus, it would understandably be puzzling why the RMB is under so much pressure when China continues to run a $593 billion trade surplus.  However, in reality official flows are negative to the tune of about $200 billion in 2015.  Add in official net errors and omissions outflows in 2015 of $132 billion and it becomes quite clear why the Chinese RMB is under pressure.

Fifth, regardless the impact on GDP, it is quite clear that cash flows within the Chinese economy are very tight.  The boost from surplus payments that is typically seen from a trade surplus is not present and firms are struggling to pay bills.  Payables and receivables continue to rise rapidly as liquidity deteriorates.  Again we cannot say for sure whether this is actual production being purchased or simply phantom production, though it is likely some blend of the two. What is important to note is that liquidity is much tighter within the Chinese economy than understood.

Sixth, the nature of capital flight from China cuts directly to the heart of why capital controls would be a poor remedy.  Capital is not leaving through the capital account.  Rather with a restricted capital account and a relatively free international transaction via the current account, enterprising Chinese are moving capital via the current account.  To arrest the flood of capital leaving this way, it would require China to bring goods and services trade in the world’s second largest economy to a complete standstill.  Every transaction would have to be verified for units, market price, agreement between importer and exporter, and accurate payment matching the invoice.  It is simply not feasible to impose currency controls that would arrest disguised capital outflows via international goods and services payment without bring international trade in China to a halt.

It is likely the PBOC is aware of the discrepancy between Customs and SAFE reported trade data and what the banks are paying via the current account.  In his interview with Caixin, PBOC Governor Zhou Xiaochuan was very careful to say that China ran a “surplus in the trade of goods” rather than current account, trade surplus, or payments and receipts for international trade.  Many foreign and Chinese agencies and analysts confuse these multiple categories referring to them as one category but they are not.  His mention indicates he likely understands how capital is leaving the country and why capital controls would be a poor remedy which is also indicated.

It is quite clear that the expected $600 billion trade surplus is not hitting the Chinese economy for reasons and some implications that are still unclear.  What we can say, is that this is negatively impacting GDP growth and liquidity.

Unpacking Financial Services in China

Most debate about the Chinese economy now does not revolve around whether Chinese GDP is accurate but the importance and take up of the rebalancing into services.  In fact, as of today, I know of only one (non-official DC) institution or firm that estimates Chinese GDP remotely close to official Chinese data.  A primary driver of the supposed service sector shift is financial services responsible for roughly 17% of the service sector or about 9% of total nominal GDP.  Consequently, due to its size, especially in the service sector, financial services are key to our understanding of the supposed rebalancing and what is driving this shift.

Let’s try and unpack the various aspects to the financial services industry.  First, despite the attention paid to the investment banks and securities firms, financial services remain dominated by much more pedestrian firms.  Commercial banks and insurance comprise roughly 92.5% of revenue among listed firms with similar patterns including unlisted firms.  This does not mean at all that these other industries are irrelevant at all, but we need the proper perspective when looking at this important industry.  Along with that, employment is even more dependent on the commercial banks and insurance.

Second, there are vast differences between revenue size or growth and profit size or growth between sub-sectors.  By net profit share of the financial sector, the division is even more stark.  At the end of Q3 2015, commercial banks were responsible for 80.4% of all net profit in financial services with the remainder divided between capital market and insurance at 10.4% and 8.9% respectively.  What is important to note is the growth rate differentials, for the moment focusing on share, between these industries.  At the end of Q3 2014, commercial banks were responsible for 89% of all FS net profit.  In that same period capital market and insurance accounted for 3.8% and 7.2% respectively.

Third, over the past year, the growth rate differentials have been quite stark.  For instance, while commercial banking revenue grew at a healthy 10% it was also the slowest of all other sub-sectors by a large margin.  For instance, capital market and insurance revenue grew at 173% and 27% YTD YOY respectively with all financial service growth through Q3 growing at 20%.  If we focus on net profit growth, there is an enormous discrepancy.  Commercial bank net profit is up 2% (more on that later) while capital market and insurance net profit is up 213% and 40% respectively with all financial services up 13%.

On the surface, financial services looks rather healthy but as I always urge, let’s look beneath the surface.  There are in fact numerous issues that disguise the weakness of the financial services industry.  Let’s look into some of these.  First, Chinese banks are in much worse shape than their top line revenue growth indicates.  As I have covered elsewhere, Chinese banks appear to be rolling over loans, capitalizing unpaid interest, and counting the unpaid/capitalized amount as interest revenue.  Let me put the perversity of this behavior in some perspective: by refusing to recognize a bad loan but counting the capitalized  unpaid loan amount as higher revenue, unrecognized NPL growth is contributing positively to revenue and Chinese GDP growth. Let me repeat that: unrecognized NPL growth is positively contributing to the Chinese GDP growth. We see another outgrowth of this the change in profit.  Commercial bank net income from operating activities grew by 0.1% and net income grew by only a slightly better 2.2%.  The difference between revenue and profit growth stems from rapidly rising NPL growth which directly impact profit growth.  In other words, neither the revenue or profit side of commercial banks is encouraging.

Second, insurance growth was initially very puzzling.  There was no obvious shift into insurance products throughout Chinese firms and consumers nor were they allowed to invest in the stock market.  However, a look beneath their financials provides us the key details.  Insurance premium revenue was up 20% and listed insurance revenue was also up 27%.  However, as premium revenue is a relatively small percentage of total revenue and with insurance companies in China declaring operational losses every year, profit is gained from investments.  Insurance investment is typically quite conservative  with strict limits on equity market investments.  Even in China only recently was regulation floated that would allow insurance companies to invest in the stock market.  From April 2013 through August of this year, the last date available for insurance data, bank deposits and bond assets held by insurance companies have grown by 13% total.  Conversely, non-bond  assets have grown by an astounding 159%.  While we do not know for sure what assets these are, we can say that Chinese insurance companies were not allowed to directly purchase stocks.  What is most likely happening, as commercial banks did the same thing, they purchased various trust and wealth management product securities.  Though there is no such thing as a standard wealth management product if we generalize we can say that a not insignificant money indirectly made its way into and was tied to the stock market, significant leverage was added, and in higher risk debt type offerings.  Commercial banks have rapidly expanded their purchases of the same types of assets.  To provide some perspective, as non-bond assets of insurance companies rose 159%, the Shanghai index would have gone up 47% over the same time frame.  Now what is worrying is that insurance holdings of these non-bond assets have not declined with the stock market and from the end of June to end of August have declined only 0.7%.  This would seem to imply that the underlying assets are at high risk of large losses which have not been recognized financially by the accounting firm and if they are purely some type of collateralized debt obligation limiting downside but providing higher upside, significantly higher risk of the underlying borrower defaulting.  In other words, the non-bond investment has not been marked to market because there is no market and they can continue to price it at face value even if the underlying risk is enormous.

Third, capital market revenue essentially tracks stock market turnover and is up enormously.  Given that most additional revenue translates very closely into profit as there is little marginal input cost growth, the pass through impact has been large.  Capital market revenue is up 173% with net profit up 213%.  This tracks very closely with what they do and our understanding of their business.  As they have very little risk exposure to assets but are much more closely tied to turnover, there is little macro-industry risk here.

Fourth, if we remove the specifically accounted for net income from investment, we get a very different view of financial service profitability.  For instance, commercial banks would have realized a net profit of less than 0.6% even though net income from investment rose 71% and all financial services investment income grew 91%.  I should note that just as I emphasized insurance companies saw rapid increases in non-bond investments that was most likely invested in trust and wealth management products.  Banks and insurance companies saw identical growth in investment income at 71%.  While the change in commercial bank investment income does not make a major impact, it has an enormous impact on insurance net profit.  Due to the 71% investment income growth, net profit saw a rise of 40%.  However, let’s make a simple assumption that investment income grows only 10% instead of 71%.  In that case, net income actually falls 97%.  Let me repeat that: if insurance income from trading does not increase 71% but rather 10%, net income would 97% lower.  In fact, if we strip out commercial bank and insurance investment income growth of 71%, their combined net income would fall 7%.

Fifth, as this passes through to GDP, China appears to be using a measure that pretty much just tracks operating revenue.  Financial intermediation was up 23% YTD and listed financial services revenue was up 20%.  Given increases in the unlisted firms, specifically securities and shadow banking activities, the difference between 23% and 20% is either small enough to ignore or disappears completely.  This method of nominal GDP accounting however would appear problematic not as much for financial services but for other areas, where revenue growth is firmly disconnected from the nominal GDP growth.  Also remember that perversely, rolling over unrecognized bad loans is increasing GDP growth

Broadly speaking, it is probably accurate to say that financial services grew at the approximate rate the NBSC claims.  However, there are enough underlying problems in the financial services sector, revenue and profit accounting, and nominal GDP measurement in other sectors using the FS method to give one real concerns.