Brad Setser in his latest posting (http://www.rgemonitor.com/blog/setser) wonders about the real exposure Chinese institutions might have to subprime mortgages. Chinese institutions -- largely the PBoC and the largest banks -- have had to invest over $800 billion abroad just in the past two years and it would be surprising if very little of this was exposed to assets adversely affected by the crisis. He is a little skeptical that we have heard the end of the story, and I can't help but agree.
The official numbers suggest, as best I can remember, that total subprime exposure is under $15 billion. That is not a lot, but I want to suggest two reasons for caution.
The first and most obvious is that in a very complex financial market it is not always clear where the exposure lies. I remember in 1995 that one of the US investment banks most severely hit by the Mexican crisis was caught completely by surprise over the extent of its exposure. It had assumed that the bulk of its exposure to Mexico would be on the Latin American trading desk, and an examination of the books indicated that its losses there were manageable.
It turned out however that the repo desk, an area of the firm that had almost no expertise in or knowledge of the risks of trading Latin America, had been doing a roaring business in Latin debt and had been earning very high spreads largely, I suspect, because they had unwittingly underpriced risk. They had no idea that their "good" collateral could turn bad so quickly. When the bank's managers finally realized the extent of the losses on the repo desk, the bank came close to failure.
This story is all based on market gossip and rumors because very little of this was made public, but any experienced trader recognizes the story. While managers were diligently monitoring risk in the area where it most obviously belonged, they hadn't noticed that a completely separate business had taken on the same risk, and that this businessdid not fully realize its extent. I don't know for a fact that this has occurred in the case of China, but I assume and hope that financial authorites and bank risk managers are looking not just at the desks that are formally permitted to take subprime exposure, but also any desk that may have indirectly assumed such exposure.
The second reason for caution is that I suspect there have been a lot of subprime-related losses in other non subprime areas that have not been recorded or explained. For example based on anecdotal evidence I believe Chinese institutions have been big buyers of structured notes. Many of my former students from Peking University and Tsinghua University have joined trading or capital markets desks at major foreign and Chinese investment banks, and from what they tell me the structured products groups have been among the most profitable and active groups in each of their banks. In fact there is one French bank for which, I understand, most of their revenues comes from the sale of derivatives and structured products.
Without knowing exactly what these products are, I have no way of knowing whether or not there have been big losses, but I do know that at least one European bank (I spoke to one of my students there) has booked huge profits in the last few weeks from being on the sell side of the product, and unless the buyers were hedged (almost inconceivable, for many reasons) they must have equivalent or greater losses.
What would explain this? Most investors buy structured notes for their high coupons, especially in a market in which interest rates are low and investors eager for a spread. It is very easy to structure a note with a high coupon. All you have to do is imbed options into the structure such that the buyer of the note is effectively selling the note-seller one or more options. The higher coupon is, in effect, the price of the option amortized in the coupon.
That means buyers of most structured notes are effectively short options -- or short volatility, in the lingo. If there is little volatility in the market, the buyer keeps his higher coupon and gives up nothing in return. Therein lies happiness. If volatility spikes upward, however, or the underlying asset moves in the wrong direction, the value of the option immediately rises and, since the buyer of the structured note is "short" the option, the value of his note drops.
Without knowing exactly what Chinese institutions have purchased, it is tough to say whether or not the fair market values of their structured notes have dropped, but I would be very, very surprised if there wouldn't be some pretty substantial losses on a mark-to-market basis. I don't think most of the buyers mark to market and in fact there really is no market for most of these notes. I suspect that the buyers also don't have the pricing formulae needed for figuring fair market value (no selling bank would give away the pricing formulae because these are usually proprietary and very valuable), but I would bet heavily that if they asked the selling banks to buy them out of their positions, they would book very large losses.
It is the impact of the subprime crisis on underlying volatility that probably will account for the biggest losses Chinese institutions (and non-Chinese institutions too, for that matter) will take from the events of this summer. Unfortunately we don't really know how much that is.
In the case of China, there is some central risk management in the form of foreign exchange controls. Because SAFE has to approve capital transfers, they are in a position to national risk management. Yes it is possible to circumvent those controls, but it is hard (not impossible) to do it in a way that is macroeconomically significant.
As far as structured notes, it really all depends on the type of structured notes. Like all derivatives, you can use structured notes to assume risk or reduce risk. If structured notes were used to get higher rates of overseas return while cushioning against currency fluctuations and interest rate risk, this is good. If they were used to get around risk management procedures and regulations, this is bad. Again, I don't think this is that much of a black box, one can quietly ask people about the types of structured notes that were bought and sold, and one can also do studies of the motives that cause people to behave in the way that they do.
Finally, "we don't know" is not the same thing as "it's bad news." There is a tendency in dealing with China to assume "we don't know and since the government is obviously incompetent and corrupt, it has to be bad news." The trouble with this model of events (which Gerald Segal explicitly stated) is that over the last twenty years, it has led people to consistently underestimate the Chinese government's ability to deal with problems. The people I've seen that are involved with economic public policy in China have all been sharp thoughtful people, and so where there is an "unknown" I tend to be more optimistic than a lot of people that the real situation isn't as dire as people assume.
Most of the notes, as far as I understand it, were for yield enhancement. When you get a high coupon you have to give up something else in return. Usually you implicitly sell options -- and unfortunately I don't think we can assume that banks have always done a great job of explaining exactly how risky that can be. The exotic derivatives markets thrives on lack of understanding.
One thing that did concern me is that apparently many of the positions held by the selling banks have been very profitable recently. Usually derivatives desks hedge the delta and keep the volatility risks, and with volatility having increased, if they are profitable I guess it is because they are long. Since derivatives are zero-sum, that suggests that their clients are short volatility and have equivalent losses (or more likely greater, since the selling banks are likely to be at least partially hedged), unless they themselves hedged, which I very much doubt.
Although I agree that many of the people involved in economic policy are indeed sharp and thoughtful, their track record in risk management has been quite poor, which is not surprising given that the only way to get good at this is through bad experiences. The copper trading scandal last year, the soybean oil bankruptcies in 2003-4, the jet fuel losses in Singapore, all show how weak the understanding of market risks has been.
I agree that no news doesn't necessarily mean bad news, but in my experience lack of news tends to be sort of pro-cyclical. When times are good we assume the best and when times are bad we assume the worst. That itself becomes part of the problem during a difficult period.
By Michael Pettis - 9/6/2007 1:22 PM
It's not just structured synthetics that are potentially underreported - among Wall Street syndicate desks it's no secret that spreads on Alt-A RMBS and super-senior tranches from ABS CDOs were pushed in by Chinese capital. Subprime is only part of the problem.
By Sami - 9/7/2007 12:20 AM
But the type of yield enhancement that would be attractive to Chinese companies wouldn't necessarily put them at risk. Chinese interest rates are extremely low, so one could get higher returns with a structured note funded by overseas bonds. Since everyone is counting on a revaluation upward, it would make sense if the note had an embedded currency option, and perhaps a floor should overseas interest rates fall below local ones. In order to hedge the risk of this note, the bank would probably sell a complex derivative with the opposite position to someone else, and then make money charging a transaction fee to both sides.
Banks can make money from complex derivatives in two ways. The can serve as a market maker finding multiple parties on opposite sides of trades and the charge a transaction fee, or they do proprietary trading. If a bank is operating as market maker, then it shouldn't be making any bet on the direction of the market. As far as proprietary trading, one would be an idiot to buy a derivative direct from a bank's prop trading desk. These two functions are kept very separate from each other. Also derivative desks when the are doing market-making *don't* keep volatility risk. The attempt to either set things up so that the error from delta hedging is concave or they engage in volatility hedging.
I seriously doubt that the people who Chinese corporations are talking to are the prop traders. I can't imagine the Chinese banking regulators allowing that to happen. The reason market-making derivative desks need to be market neutral is that otherwise there is a huge conflict of interest with the client. If the derivative desks are selling the type of product that I mentioned earlier, then they may be racking up huge transaction fees, and that would be a good thing.
As far as derivative disasters go, bear in mind that it is the plane crashes that make the news, and not the flights that arrive safely. (And disasters are hardly unique to China.) Investment banks have lots of internal processes to keep disasters from happening, and a lot of this involves functional separation of prop trading and market making and also separation of front/mid/back office roles. Chinese banks might be a bit new to this, but I would be shocked and alarmed if foreign investment banks weren't doing the kindergarten stuff needed to prevent disasters from happening. One should *always* be alarmed if a bank makes too much money because that suggests that they are making bets on the direction of the market, and eventually the market will turn. The question of *how* banks are making money and what types of structured notes they are selling aren't black boxes.
Also it's not clear that there is underreporting. Right now we have about $15 billion in reported exposures to mortgage CDO's from two of the big banks. That's consistent with reports of "lots of Chinese money coming in."
Sami, what do you hear the Chinese are buying, and what kinds of institutions?
By Michael pettis - 9/7/2007 8:30 PM
Twofish, I am not sure what you mean by prop traders. Typically the clients (corporations or insitutional investors) speak to salesmen on the structured product desks. In some cases the desk purchases the underlying instruments from the prop desks, who are then left with the responsibility of hedging, and in other cases (more and more) it is the structured product desks that buy and hedge the underlying positions. These desks often explicitly take volatility risk. I ran market making desks at several Wall Street institutions and quite frankly I don't think there are any major market-making desks that don't take fairly significant risk positions. The information you get as a market maker is too valuable not to use for speculative purposes.
At any rate, the point is that if the recent volatility has resulted in big profits among the structured notes groups, they MUST be taking positions, and there MUST be losses on the other side. I don't know how those losses are distruibuted, but someone in China is taking them.
By Michael Pettis - 9/7/2007 8:37 PM
What I've seen is that the salesman at the structured product desks then talk to the traders and the quants who then hedge the underlying positions. You then have risk managers look over the shoulders of everyone to limit risk exposure, What I've seen is that the prop traders are kept separate from the market making desks with the latter given comparatively small "risk budgets" to limit the amount of speculation that the latter can take. The market makers are supposed to make their money for the firm by transaction premiums and not by guessing the direction of the markets.
I should note that it is quite possible that I've misinterpreted what I'm seeing. It's also possible that the places I've seen are not representative of how the industry works in general. But it does seem a massve conflict of interest to be selling a client a product that will make you money, only if the person you are selling it to losses money (and vice versa). I suppose that the different things we've seen first hand, might explain why I think derivatives are a grand thing, and you don't.
I disagree with the syllogism. Derivatives are often much to complex to attempt to make money by taking market positions, and the money made off of them are often transaction fees for creating the product. The money that the structured note groups are making might be the result of a premium that the customer is willing to pay over what the costs are to the bank to create and hedge the product. I'd imagine that this premium could be quite high in China.
One other question, is it absolutely essential that the people taking losses be in China? If we do have a zero-sum situation with some of the parties outside of China, there are plenty of people who could be taking losses.
A massive conflict of interest? Pardon my cynicism but yes, of course. In my previous Wall Street job, where I headed Latin American Capital Markets and Liability Management, because I had done a lot of work in derivatives one of my unofficial jobs for salesmen friends on the floor turned out to be reverse engineering and pricing some of the derivative products before they were shown to clients, to make sure the clients weren't being reamed too badly.
The salesmen took it as a matter of course that the derivatives guys had very different incentive structures from themselves, to put it politely. I am not opposed to derivatives. In fact I designed and introduced many that became standard in Latin Amerian markets in the 1990s. I just think that they need to be handled very carefully in unsophisticated and rigid markets such as China's, and I am pretty certain that a very high level of risk management is required to account properly for them, and I am skeptical that we have much of this in China.
I am pretty sure that most of the stuff sold by the Beijing and Shanghai offices are only to Chinese clients. Regulations make exceptions difficult and anyway other Asians would be handled out of HK or Singapore. Also what my student have been telling me recently is not that the desks are profitable -- they always have been -- but rather that they have taken large mark-to-market profits in this summer's turmoil. That can only come from positioning.
By Michael Pettis - 9/8/2007 2:15 PM
Yikes.
The problem with a lot of the derivative instruments is that they act reasonably well in normal times, but in a crisis, they turn deadly in ways that even the people selling them can't forsee. The basic problem is that capitalism works well when you have snake oil salesmen selling snake oil to other snake oil salesmen, but really bad things happen when you try to sell the stuff to widows, orphans, and third world nations.
One big difference between Latin America and China is that Latin America was a net borrower of capital whereas China is a net exporter of capital. This means that Chinese banks are not so much trying to lend to the corporate sector but rather borrow from them, and this means that a lot of the signs are the opposite of what one would have in Latin America.
Off the top of my head I can think of a lot of things other than volatility that could cause a big run up...
1) The RMB has stopped appreciating and has started depreciating in mid-July has the PBC moved the exchange rate to stop hot money coming in. Things could have been priced expecting that the RMB would appreciate at the rate it's been, and a depreciation could cause the banks to make money depending on how the hedges were set up.
2) RMB interest rates have gone up, USD interest rates have gone down. If someone were long USD bonds and short RMB ones, that would cause position profits.
3) Finally, there might be some panic selling. I can imagine some Chinese corporations suddenly being wary of structured notes, and wanting to sell them at a deep discount to their net present value. Once a bank buys the note and then resells them at NPV, it gets a nice profit.
Personally, I'm not that concerned about direct losses done to the summer's events. What worries me are indirect effects. The summer's events are going to change the economic realm in ways that aren't clear now. One obvious thing is that if the US cuts interest rates to prevent recession and the PRC has to increase interest rates to prevent inflation, then this is going to put much more stress on the RMB exchange rate.
Michael Pettis is a professor at Peking University's Guanghua School of Management, where he specializes in Chinese financial markets. He has also taught, from 2002 to 2004, at Tsinghua University’s School of Economics and Management and, from 1992 to 2001, at Columbia University’s Graduate School of Business. He is a member of the board of directors of ABC-CA Fund Management Co., a Sino-French joint venture based in Shanghai.
Pettis has worked on Wall Street in trading, capital markets, and corporate finance since 1987, when he joined the Sovereign Debt trading team at Manufacturers Hanover (now JP Morgan). Most recently, from 1996 to 2001, Pettis worked at Bear Stearns, where he was Managing Director-Principal heading the Latin American Capital Markets and the Liability Management groups. He has also worked as a partner in a merchant banking boutique that specialized in securitizing Latin American assets and at Credit Suisse First Boston, where he headed the emerging markets trading team. Besides trading and capital markets, Pettis has been involved in sovereign advisory work, including for the Mexican government on the privatization of its banking system, the Republic of Macedonia on the restructuring of its international bank debt, and the South Korean Ministry of Finance on the restructuring of the country’s commercial bank debt.
Pettis is a member of the Institute of Latin American Studies Advisory Board at Columbia University as well as the Dean’s Advisory Board at the School of Public and International Affairs. He is the author of several books, including The Volatility Machine: Emerging Economies and the Threat of Financial Collapse (Oxford University Press, 2001). He received an MBA in Finance in 1984 and an MIA in Development Economics in 1981, both from Columbia University.
Subscribe to this Site
) wonders about the real exposure Chinese institutions might have to subprime mortgages. Chinese institutions -- largely the PBoC and the largest banks -- have had to invest over $800 billion abroad just in the past two years and it would be surprising if very little of this was exposed to assets adversely affected by the crisis. He is a little skeptical that we have heard the end of the story, and I can't help but agree. 
Comments (9) for "Are there more losses from t...
January, 2008