Brad Setser on Similarities Between Amaranth and the People's Bank of China...

He writes:

RGE - Unintentional irony watch (hedge fund edition): Risk management standards do seem to be slipping. Davis, Sender and Zuckerman in the Wall Street Journal:

The risk models employed by hedge funds use historic data, but the natural gas markets have been more volatile this year than any year since 2001, making the models less useful. They also might not predict how much selling of one’s stakes to get out of a position can cause prices to fall.

The models broke down because this year has been more volatile than any year since 2001? That isn’t so long ago, except in hedge fund time... I suspect Mr. Geithner won’t be happy if stress tests -- and risk management models -- don’t look a bit further back. 1998 might be a relevant data point. Certainly for anyone betting on emerging markets, the yen/dollar (or the yen/ euro, since the dyanmics of an unwinding carry trade are similar) or the Treasury market. 1998 also might be a relevant data point for a few commodity markets. Mr. Geithner also wants – I suspect – those lending to hedge funds to assume that liquidity will dry up when their clients need it the most. Getting out when you don’t need to is easy. Getting out when you absolutely have to is hard. Especially if you have a large concentrated position...

That lesson may be relevant for another set of institutions with large positions. A few central banks might also want to ponder the impact large, sustained purchases can have on market dynamics. Davis, Sender and Zuckerman, again.

According to natural gas investors who traded alongside Amaranth, Mr. Hunter repeatedly used borrowed money to double down on his bets. Buying more futures contracts of the kind his fund already owned supported their price by increasing demand, propping up paper gains, these traders say. But the support only lasted so long as Amaranth and its lenders were willing to spend cash to buy more contracts. Such trades may also have masked growing weakness in market fundamentals.

Substitute dollars, Treasury bonds and Agency bonds for “futures” and “contracts,” and substitute the PBoC (and a few others) for Amaranth...

China is making a huge bet that Stephen Jen and Friedrich Wu are right, and the RMB isn’t really undervalued (and the dollar and euro are not overvalued against the RMB). It is betting that the RMB wouldn’t really rise if it stopped buying. That is a rather large gamble though.

Especially since Chinese productivity is growing faster than US productivity (see the graphs on p. 13 of Feng Lu’s presentation), so a key “fundamental” is moving against China’s financial position...

I think Fan Gang at least understands as much. Yu Yongding as well. And Zhou Xiaochuan.

The Politburo and the State Council? I am not so sure.

The Politburo and State Council may understand it. They may be thinking as follows: "We grow at 8% per year as long as we can keep export-led industriallization going. When export-led industrialization stops and we have to substitute domestic-demand-led industrialization, our growth rate is likely to fall to 5%. Thus each year we keep this juggling act going raises China's GDP--permanently--by about RMB 500 billion a year, an increment to the present value of China's total national wealth of RMB 10 trillion. To keep the juggling act going requires that we spend RMB 3 trillion a year buying dollar-denominated securities that will be worth only RMB 2 trillion when we sell them. That looks like a benefit-cost ratio of 10:1. So let's keep juggling as long as we can.

That maybe what they are thinking in the Politburo and the State Council.

Comments

[wkwillis]

Chinese people don't have medicare and social security, and they weren't allowed to have several children, so the bank deposits are all they have in cash and their apartments in the coastal export zones are their main source of wealth.
If exports break down, the apartment values crash. If the banks close, and they have no cash or it is inflated away...

[anne]

Brad DeLong:

"We grow at 8% per year as long as we can keep export-led industriallization going. When export-led industrialization stops and we have to substitute domestic-demand-led industrialization, our growth rate is likely to fall to 5%. Thus each year we keep this juggling act going raises China's GDP--permanently--by about RMB 500 billion a year, an increment to the present value of China's total national wealth of RMB 10 trillion. To keep the juggling act going requires that we spend RMB 3 trillion a year buying dollar-denominated securities that will be worth only RMB 2 trillion when we sell them. That looks like a benefit-cost ratio of 10:1. So let's keep juggling as long as we can."

Brilliant.

[kharris]

I wonder if the WSJ has identified the right problem in citing volatility. Amaranth blew up over a calendar trade, I believe. While volatility was an issue in a mechanical sort of way, the real break down was because spreads went the wrong way. The assumption was that spreads between two constracts would follow the same pattern this year as in prior years as utilities and other holders of inventories made seasonal adjustments (the seasonal assumption is what led to a calendar trade). Spreads went the other way. Leverage and inadequate secrecy did the rest. If that is, in fact, what happened, it is failure to read up on LTCMs spread trades in off-the-run Treasuries that got 'em. Spreads that reliably widen one day just narrow instead, or vice versa.

As to the similarity with the PBOC,...um...aren't we stretching to make a point? The point that you can be very wrong about the value of assets doesn't need Amaranth as an object lesson. PBOC isn't leveraged 5-to-1, and isn't involved in an exotic trade - it owns dollar-denominated assets outright. The PBOC is risking other peoples' money, that's about it.

[Paul J. Reber]

I'm not quite following the math totally (the numbers listed are clearly being plugged into some models I'm unfamiliar with since I am not an economist), but the concept is clear and sensible enough.

But does this analysis then give some insight into the conditions under which China is likely to need to "unwind" their position? Suppose the US housing market really is faltering and we see slower growth (or even recession) here that leads to an overall reduction in demand for China's exports. Then China can't maintain that 8% growth rate (?), although in parallel they ought to not need to keep taking long positions in overvalued US securities to compensate for the general account deficit.

They'll still be holding a lot of securities and although I mentioned I'm not an economist, it looks like a tough position to get out of. My (uninformed) intuition is that the PBoC is likely to make the meager 5 billion dropped by Amaranth look like chump change (by 2 or 3 orders of magnitude?). Although, I don't think for a second they're being irrational -- I bet the long-term benefits still add up positive along the lines of what Brad suggests. But talk about systemic risk. It sure looks like some mighty big-time gambling going on.

[howard]

prof, i want to second anne: your analysis of the strategic play the politburo is endorsing makes total sense and answers a long-held question of mine.

[anne]

What is important is to watch China use dollar assets for increasingly extensive international aid, beginning near in Asia but extending away to Africa and Latin America. Dollar assets can easily support the Chinese banking system as well.

We can choose numbers to show a relative loss of value of the dollar, simply select a helpful year to show a significant swing. But, I am impressed that on rolling 10 year basis the dollar has been remarkably stable against developed market currencies. The dollar had lost 3% in value against developed market currencies as of the 10 years from September 15. The stability is remarkable to me.

[ed_finnerty]

they are probably also thinking

"If we stop creating jobs for all of these newly urbanised males who can't get dates because of the one child policy we are going to have Tianamen on a gargatuan scale"

[brad setser]

anne -- "dollar assets can help the banking system." Uh, not really.

The Chinese banks have RMB deposits and need RMB assets. Otherwise the PBoC just shifts its currency risk to the banks -- creating a currency mismatch on the banks balance sheet.

China has shifted some dollar reserves to the banks in an accounting move, but that really is just accounting -- the state owned banks now have more regulatory capital -- but they already had a call on the state for capital). The banks also reportedly got a promise from the PBoC to make up any currency losses. The vast majority of Chinese bank recapitalization has been done in RMB, by either shifting bad loans to the AMCs and giving the banks an AMC bond or having the PBoC buy bad loans at par and then shift them to the AMCs ... Guonan Ma of the BIS has the best data on this.

[linebeingwalked]

"prof, i want to second anne: your analysis of the strategic play the politburo is endorsing makes total sense and answers a long-held question of mine."

Completely disagree. China is more concerned with the potential of unleashing domestic inflation than the threat of losing billions in the forex market. Why unwind? As long as the economy does not blow up it doesn't really matter if the state loses paper assets. Chinese policy is not really conceived of a tax on Chinese citizens anyway.

There won't be any political price to be paid by anyone in China for doing nothing, while rocking the boat is dangerous. And topping it all off there is a powerful tendency towards inertia in this sort of policy because forex policy cannot be changed without approval from the State Council.

[anne]

Brad Setser:

"The Chinese banks have RMB deposits and need RMB assets. Otherwise the PBoC just shifts its currency risk to the banks -- creating a currency mismatch on the banks balance sheet."

Yes; I understand, but dollar assets to the extent of the Chinese holdings insulate the financial system beyond domestic currency. Dollars, after all, still do buy things, and curiously buy about as much internationally as they bought in 1996.

[johnchx]

"We grow at 8% per year as long as we can keep export-led industriallization going. When export-led industrialization stops and we have to substitute domestic-demand-led industrialization, our growth rate is likely to fall to 5%."

This is the key assumption. My question: why would this be true? What is the basis for the assumption that export-led industrialization is automatically faster than industrialization focused on production for domestic consumption?

[shah8]

adding to johnchx...

The thing the popped out at me, as critical, and critically wrong, was:

Thus each year we keep this juggling act going raises China's GDP--permanently--by about RMB 500 billion a year, an increment to the present value of China's total national wealth of RMB 10 trillion.

When it unwinds, most of that is going to be a capital overhang, that China will have to *spend* money on to erase. Things that will stay, like roads and communications and lots of educated chinese, all could have been purchased cheaper than having a trillion bucks in reserves go down to 200 billion, or even 700 billion. The opportunity costs are staggering!

Nah, the politburo is thinking in very *political* terms. They may be thinking of economics, but only as that feeds into immediate power relations. Along with linebeingwalked's sentiment, I simply add that forex playing allows the central party leverage over the rest of the chinese polity, and that's why they pursue this policy.

[Jonathan King]

Paul J. Reber writes:
> My (uninformed) intuition is that the PBoC is likely to
> make the meager 5 billion dropped by Amaranth
> look like chump change (by 2 or 3 orders of
> magnitude?). Although, I don't think for a second
> they're being irrational -- I bet the long-term
> benefits still add up positive along the lines of what
> Brad suggests. But talk about systemic risk. It sure
> looks like some mighty big-time gambling going on.

Hi Paul; funny seeing you around here. I think the way to look at this is that the PBoC might be correctly calculating the expected value of what they're doing, but the variability in the outcome is potentially huge, and might not even be knowable. Clearly a major drop in US imports caused by a post-housing-bubble crash could throw the whole Chinese plan into the trash. Similarly, domestic unrest in China caused by any number of things could cause huge problems, and I do think there is a limit to how hard they can bear down to prevent political dissent even if the economy is running great.

But there is no particular reason they couldn't try to keep it going for years, and what Amaranth teaches us is that there are some money managers out there who are willing to take really big risks based on incredibly tiny data sets (seasonal swings in gas futures contracts since 2001??). I don't know if anybody really knows how much leverage there is behind some hedge fund positions that boil down to bets that the Chinese will be able to do *exactly* what they are doing for a long time. I really therefore do not know how much deviation from the status uo the system can really take.

[kco]

"To keep the juggling act going requires that we spend RMB 3 trillion a year buying dollar-denominated securities that will be worth only RMB 2 trillion when we sell them."

If this happens every year, this implies that china will lose 1 trillion every year eventually in her dollar denominated securities. The present value of the total lost is 20 trillion. the benefit:cost ratio is instead 1:2.

pls correct me if i'm wrong

[Maynard Handley]

Let's follow the thinking here for someone simple like myself, and filling in what I consider to be the unspoken blanks

China right now
- sells lots of stuff to the US
- collects lots of dollars
- buys some capital goods, oil etc with those dollars
- mostly buys treasuries with the dollars

Setser thinks that
- buying treasuries is a waste of cash, that China should use its dollars, while they are still worth something, to buy whatever useful it can get for dollars (presumably microchips, airplanes and weapons since it's not clear what else of value the US sells)
- Setser doesn't mention as I saw, but I will, that there would appear to be limits on what China can buy, that (US politically) they cannot get away with spending their dollars on real estate or companies

DeLong thinks that
- the Chinese have to prop up the US economy because
-- it's basically a basket case in the absence of their help
-- there's something of a paper chase going on, world wide and internal to China; that infrastructure, once built, is real, but to get it built requires nationwide confidence and faith in the concept of money, and that the Chinese population believe enough in dollars that a monetary system anchored to US exports keeps them encouraged and going to work in the morning

Setser thinks that
- China could spend more money on US goodies and be holding fewer potentially worthless Treasuries. He does not, however, indicate what those goodies might actually be.
Assuming that such goodies would be various forms of capital equipment (since consumption is presumably not on the table), is it the case that more capital equipment would be to China's advantage right now? Is there enough skill and enough knowledge across the country to buy capital equipment usefully? Spending 100 billion on machine tools that are actually only worth 50 billion in two years when the dust clears and the various mistakes made and overcapacity in certain areas are apparent is not obviously superior to buying 100 billion of treasuries that are worth 50 billion in two years.

I think that the issues Setser leaves unmentioned, to summarize, are
- obviously there is Brad's point about export led growth
- along with that, I'd add that there is probably a psychological issue that gives the individual Chinese some confidence in the value of their money because there are all these dollars in the background. I suspect also that this goes unsaid for the obvious reason that you're going to have to ditch that pyschological crutch, and fast, when the dollar falls. The way you do this, of course, is to make a big song and dance, when necessary (but not earlier), of revaluing the RMB to more dollars (because RMB, *China's* currency, is more valued by the world than $)
- it's one thing to say China should buy more from the US. It's another to point out what should actually be bought that's of value
- even if it's equivalent in terms of keeping the US economy afloat and the Chinese economy afloat to spend the money wastefully on capital goods rather than on Treasuries, Treasuries have the great advantage of always being their in the background, presumably concentrating wonderfully the mind of US politicians.

[piotr]

Why export-driven growth is faster than domestic-demand driven growth, my speculation:

1. Super-fast growth is driven by high rates of return on investment, which drive the said investments, and they, in turn, are driven by wages lagging waaay behind the productivity.

2. What does it mean "domestic demand"? Someone has to be paid more to create the demand, employees, the taxman, or both. Thus the returns on the investments go down, so investments go down, and so does the growth.

3. It is quite possible that the very fast growth is not all that advantageous to the Chinese workers. For starters, the pie is larger, but their share is smaller. Secondly, the pattern of consumption could be more in tune with their needs, say, more healthcare, more education, more infrastructure away from the growth centers, more employement that does not force workers to live away from their families.

4. In an ordinary democracy, say, India, there is some political pressure to increase the domestic consumption, even if it decreases the pace of economic growth, there seems to be a correlation between export-driven growth and authoritarian government. Living under an authoritarian government has many unpleasant aspects that impact the quality of life (say, higher chances of being arbitrarily imprisoned and shot to death, Chinese government delights in big anti-crime sweeps
and the justice system is crude at best in finding real culprits).

5. Is there some zero-sum game here, to wit, are we loosing some opportunities when we sell securities to Chinese to finance our import?