In late 2004, Nouriel Roubini and I wrote that “the tensions created [by the Bretton Woods 2 system] are large, large enough to crack the system in the next three to four years.” In a 2005 Wall Street Journal online debate with Michael Dooley I tried to hedge a bit, and gave the system six years.
“we [Roubini and Setser] never said the system would collapse at the end of 2005. A collapse in 2006, maybe. A collapse before 2008, likely; before 2010, almost certainly.
I used to worry that these dire warnings would prove to be wrong – and that Michael Dooley and Peter Garber would remind me of them at a time and place of their choosing. Now I am starting to worry that Nouriel and I may end up being partially right.
The available trade data for q3 does not show a fall in either the US deficit or China’s surplus. But I increasingly suspect that the combination of falling oil prices and falling demand for imported goods will produce significant fall in the US trade and current account deficit in the fourth quarter, with a corresponding fall in the emerging world’s combined surplus. The Bretton Woods 2 system – where China and then the oil-exporters provided (subsidized) financing to the US to sustain their exports – will come close to ending, at least temporarily. If the US and Europe are not importing much, the rest of the world won’t be exporting much.
And rather than ending with a whimper, Bretton Woods 2 may end with a bang.
In some sense Bretton Woods 2 has been on life support for a while now. China’s recent export growth has depended far more on Europe than on the US. US demand for non-oil imports peaked in 2006. One irony of the past year is that the US was borrowing far more from China that it was buying from China. Campaign rhetoric that the US was paying for Saudi oil with funds borrowed from China isn’t far off – though it leaves out the fact that the US also borrows from Saudi Arabia to pay for Venezuelan, Mexican and Nigerian oil.
If Bretton Woods 2 ends in 2009 – if US demand for imports falls sharply in the last part of 2008 and early 2009, bringing the US trade deficit down – it won’t have ended in the way Nouriel and I outlined back in late 2004 and early 2005. We postulated that foreign demand for US debt would dry up – pushing up US Treasury rates and delivering a nasty shock to a housing-centric economy. As Brad DeLong notes, it didn’t quite play out that way. The US and European banking system collapsed before the balance of financial terror collapsed. Dr. DeLong writes:
All of us from Lawrence Summers to John Taylor were expecting a very different financial crisis. We were expecting the ‘Balance of Financial Terror’ between Asia and America to collapse and produce chaos. We are not having that financial crisis. Instead we are having a very different financial crisis. Catastrophic failures of risk management throughout the entire banking sector caused a relatively minor collapse in housing prices to freeze up global finance to a degree that has not been seen since the Great Depression.
The end result of this crisis though could be rather similar: a sharp contraction in credit, a fall in US economic activity, a fall in US imports and a fall in the amount of foreign financing the US needs.* The US government is (possibly) trying to offset the fall in private demand by borrowing more and spending more — but as of now there is realistic risk that the fall in private activity will trump the fiscal stimulus.
Consequently, this still strikes me a crisis of the Bretton Woods 2 system. In retrospect, Bretton Woods 2 depended on two things: ongoing flows from the emerging world’s governments to the US Treasury and Agency market, and the ongoing ability of the US financial system (broadly defined to include the dollar-based “shadow” financial system operating in London and other offshore centers) to transform these flows into loans to ever-more indebted US households. US investors** effectively sold their holdings of Treasuries and Agencies to the world’s central banks, and then redeployed their funds into private-label mortgage-backed securities. Between the end of 2003 and q2 2007 (three and a half years), the stock of mortgages held by private issuers of asset-backed securities rose from about $1 trillion to around $3 trillion. That demand meant that credit was available to any household that wanted it – even those without much ability to pay if the housing market ever turned.
Or, to put it more succinctly, Bretton Woods 2, as it evolved, hinged both on the willingness of foreign central banks to take the currency risk associated with lending to the US at low rates in dollars despite the United States large current account deficit AND the willingness of private financial intermediaries to take the credit risk associated with lending at low rates to highly-indebted US households.
The second leg of the chain collapsed before the first. And it collapse looks set to deliver a nasty shock to everyone – including the countries that supply the US with vendor financing.
In some sense, the vendor finance analogy never really worked. The “vendor” financers didn’t actually lend directly to the US households that were buying their goods. The big emerging market central banks were willing to take on currency risk associated with lending to the US but not the credit risk associated with lending to US households.
That didn’t matter so long as US financial institutions were willing to take the credit risk.
But now US financial institutions are neither willing nor able to take on the risk of lending even more to US households. For a while the US government was able to ramp up its lending to households (notably through the Agencies) and in the process effectively take over the function previously performed by the private financial system (over the last four quarters, the flow of funds data indicates that the Agencies provided around $800 billion of net credit to US households). But now the US government is struggling to keep the financial system from collapsing. It doesn’t seem like it will able to avoid a sharp fall in the overall availability of credit.
In retrospect, the fact that (reported) bank profits didn’t fall as the Fed raised rates should have been a clue that risks were building. An inverted yield curve isn’t good for institutions that borrow short and lend long – but it initially didn’t seem to have an impact on financial sector profitability. It is now clear how the financial sector kept profits up: it took on more risk, as it shifted from borrowing short to buy safe long-term assets (Treasuries and Agencies) to borrowing short to buy risky long-term assets. Leverage in the system also increased (and for some broker dealers that seems to be an understatement), as more and more financial institutions believed that the US had entered into an era of little macroeconomic or financial volatility. The net result seems to have been a truly explosive concentration of risk in the hands of a core set of financial intermediaries in the US and Europe. Securitization – it seems – actually didn’t disperse risk into the hands of institutions able to handle it.
Many have highlighted the role that loose US monetary policy played in supporting the housing boom. And there is no doubt much truth in this story: pushing rates down to help “clean” up the bursting of the .com bubble and holding them down for several years certainly helped induce the rise in home prices and the housing boom. At the same time, this story leaves out what to me is a crucial part of the story: the housing boom didn’t end when the Fed reversed course and raised long-term rates. The really risky loans were made in late 2005, 2006 and early 2007 – after policy rates had increased. Private institutions kept on lending – in part because they decided that it was safe, in a world of low assumed macroeconomic volatility – to take on more leverage and more credit risk to keep profits up.
Foreign central banks that kept on lending to the US despite large ongoing deficits – and poor returns, after taking into account the currency risk – on their dollars contributed too. If they had scaled back their financing more rapidly, the US would have been forced to adjust sooner – reducing the risk of the kind of severe crisis we are now in.
I hope that the process of adjustment now underway isn’t as sharp as I fear. The US economy gradually can shift from producing MBS for sale to US investors flush with cash from the sale of safe securities to China and Saudi Arabia to producing goods and services for export – but it cannot shift from churning out complex debt securities to producing goods and services overnight. Indeed, in a slowing US and global economy, improvements in the US deficit will likely come from faster falls in US imports than in US exports – not from ongoing growth in US exports.
But right now it looks like there is a real risk that the adjustment won’t be gradual. And it certainly looks like the flow of Chinese (and Gulf) savings to US households over the past few years has produced one of the largest misallocations of global capital in recent history.
US taxpayers are going to be hit with a large tab for the credit risk taken on by undercapitalized financial intermediaries. Chinese taxpayers may get hit with a similar tab for the losses their central bank incurred by overpaying for US and European assets as part of its policy of holding its exchange rate down. The TARP is around 5% of US GDP. There are plausible estimates that China’s currency losses will prove to be of comparable magnitude. Charles Dumas puts the cost at above 5% of GDP:
“Charles Dumas of Lombard Street Research estimates that China makes 1-2 per cent on its (largely) dollar reserves. It then loses up to 10 per cent on the exchange rate and suffers a Chinese inflation rate of 6 per cent for a total real return in renminbi of about minus 15 per cent. That is a loss of $270bn a year, or a stunning 7-8 per cent of gross domestic product.”
I have estimated that the annual cost of adding $600b (15% of China’s GDP) of unneeded reserves to China’s stockpile is roughly 5% of China’s GDP — though the exact loss depends on the size of the RMB’s eventual appreciation. Others have calculated large losses to Chinese households on the basis of the very low rates China has maintained on domestic deposits to support the RMB.
The US taxpayer are currently getting hit with the tab for much of the credit risk that supported the big increase in US household consumption – as Martin Wolf quipped, what looked to be private lending turned out to be public spending. And China’s taxpayers will get eventually have to pick up the bill for the currency risk associated with lending to the US in dollars at low rates …
* In one key respect the scenarios do differ: the collapse of the balance of financial terror would push Treasury yields up. The credit crisis has pushed Treasury yields down.
** If a US money market fund shifted from investing in Treasuries to investing in the dollar denominated paper of a European that was “funding” its holdings of mortgage-backed securities in the money market, the money market fund effectively ended up taking on credit risk – albeit quite indirectly. A money market fund that shifted from holding Treasuries to holding Lehman paper certainly ended up taking on credit risk.