How long will the world keep on financing the US: my new paper with Nouriel Roubini

Nouriel and I went out on a bit of a limb in this paper.

Since the beginning of the year, the dollar has rallied and yields on the ten-year and thirty year treasury bond have fallen: both market moves suggest that the US is not currently having any difficulty financing the US government’s $425-30 billion estimated FY 05 budget deficit (including the costs of fighting wars in Iraq/ Afghanistan), and the US is not currency having difficulty financing its roughly $60 billion monthly trade deficit/ $700 billion plus annual trade deficit. So long as the dollar is strengthening and US interest rates are low, the market is hardly sending a signal that the US needs to cut back. Nouriel and I argue this really is too good to last, and lay out the reasons why the US will be forced to cut back before the end of 2006. Put differently, we argue that while US might be able to finance external and budget deficits of the current scale at current interest rates and exchange rates for a few more months, it won’t be able to do so for two more years.

The value of this paper is that it forced us to lay out our assumptions. Hopefully, it will force those who think deficits of the current magnitude can financed and therefore both policy makers and the markets can continue on their current trajectory to lay our their assumptions as well.

Realistically though, not too many people will want to read the entire paper. Nouriel provided a brief summary of the paper earlier, but I wanted to chime in as well and provide my own guide to the paper’s highlights.Over the past two years, the US has financed a cumulative current account deficit of about $1200 billion. ($530b in 2003, $650 b in 2004). It has done so, in broad terms, by placing about $950 billion in debt with foreign central banks, and by getting about $250 billion in financing from the private markets. Our estimate for the amount of financing that foreign central banks provided the US in 2004 by adding to their dollar reserves is now $475 billion, a bit higher than the $400 billion that I have used previously.

There is no doubt that foreign central banks have provided the majority of the (net) financing the US has needed over the past two years, and have added to their portfolio of US assets at a far faster rate than foreign private investors. But the 950b/ 250b split between central bank financing and private financing also oversimplifies: foreign private investors hold more gross claims on the US can foreign central banks, so even if foreign central banks are adding to their claims faster than private investors, equilibrium still requires that foreign private investors be willing to hold on to their current claims (at current prices), rather than dump their existing claims in the market.

The actions of US investors also matter: if they want to increase their holdings of foreign assets, that would add to the United States aggregate need to attract foreign financing. For example, in 2003 and 2004, US investors bought more foreign stocks than foreigners bought US stocks, and US firms invested more abroad than foreign firms invested in the US. This resulted in a net equity outflow of $200 b in 2003, and an estimated $150 b in 2004.

Add this $350b net (equity) outflow to the $1200 b cumulative current account deficit of the past two years to get the total amount of debt the US needed to place abroad in 2003 and 2004: $1550 billion. Even if $950 billion of that was placed with foreign central banks, foreign private investors still had to buy around $600 billion of US debt for everything to work (foreign private investors have been buying significant quanitities of US corporate debt).

The $100 million dollar question is will this last.

Some numbers. The 2005 current account deficit is likely to be around $800 billion, maybe a bit more. If nothing changes, the 2006 deficit, is likely to be around $900 billion: Remember, just paying interest on the $800 billion borrowed in 2005 might add $40 billion to the overall 2006 deficit. That implies a total two year financing need of $1.7 trillion dollars IF net equity outflows stop, and a need to place around $2 trillion in debt abroad over the next two years if net equity outflows continue at around $150 b a year.

These are large sums by any measure. Betting that the current interest rate/ exchange rate combination can last essentially means betting that the rest of the world will make that much financing available to the United States.

Nouriel and I laid out three reasons why we doubt that much financing will be available over the next two years:

1) The needed inflows from foreign private investors may not materialize. If the Japanese Ministry of Finance/ Bank of Japan stay out of the market the US could only get $400 billion a year in reserve financing from the the world’s central banks. Afterall, Japan added $200 billion to its reserves in 03, and $175 billion plus in 04 -- and almost all of that went into dollar assets/ US treasuries. $800 billion is a lot of money -- it implies that the world’s central banks, and in particular the central banks of emerging economies, would be continuing to accumulate dollars at unprecedented rates -- but it still covers less than 50% of the United States’ estimated external financing need (a lot less than in 2003/2004).

Suppose the Japanese do jump back in, and the US gets $500 billion a year in reserve financing in 05/06, or $1 trillion total. The US still needs to raise at least $700 billion by selling debt to foreign private investors, and perhaps more if net equity outflows continue. That amount of private financing may not be available from private investors around the world at current dollar interest rates.

2) Smaller central banks may reduce the pace of their dollar reserve accumulation. Smaller in this case means anyone other than the Japanese and Chinese. In aggregate, the "smaller" central banks -- think the central banks of Korea, Taiwan, Singapore, India, Russia, Malaysia, Thailand among others -- are not so small. Leave out the People’s Bank of China and the Bank of Japan and the world’s central banks still added over $300 billion to their reserves in 2004, and at least $150 billion to their dollar reserves. Non-China emerging Asia and Russia combined to add $200 b to their reserves, and at least $100 b to their dollar reserves and probably much more. If these central banks started to defect from the dollar financing cartel, and stopped adding to their reserves at their current rates, the Bretton Woods 2 system of financing the US would break down, or, at a minimum, the burden the system places on China would go up substantially.

3. China. We all know, by now, the reasons why China does not want to change its exchange rate regime. But it worth remembering that the current system also imposes substantial costs on China -- costs that China’s central bank, at least, cannot ignore, even if China’s political brass is no more willing to consider changing their peg than the Bush Administration is willing to consider a serious effort to reduce the US fiscal deficit.

Still, the costs to China of keeping the current system are rising. If nothing changes, China probably needs to add roughly $250 billion annually to its reserves in 2005 and 2005, so its overall reserves would rise to about $1.1 trillion by the end of 2006 (53% of China’s GDP). The People’s Bank of China would presumably need to add about $200 billion a year to its dollar reserves (financing about 1/4 of the US current account deficit), so its estimated dollar reserves would need to rise from say $450 billion to $850 billion by the end of 2006. With dollar reserves equal to 44% of its estimated 2006 GDP, China would stand to lose a lot of money in the event the renminbi is allowed to appreciate against the dollar. Sterilizing (selling bonds to keep the money supply from rising in line with rising reserves) that kind of inflow is not easy either -- China issued something like $70 b in sterilization bonds in 2004, but that did not even come close to mopping up the 200 b plus 2004 reserve inflow. And the People’s Bank of China will have to do more in 2005 and 2006 ... keeping this kind of inflows from producing domestic financial trouble inside China is not easy.

In broad terms, the world was only able to provide the US with $450 billion plus in reserve financing in 2003 and 2004 because the People’s Bank of China was willing to overfinance the US. China’s current account surplus in 2004 is probably about $50 billion. China’s reserve accumulation was around $200 billion. If $150 billion of that was invested in dollar assets (that number could be a bit high, but it makes the math easier), China provided $100 billion in financing to the US not from its own trade surplus, but from foreign capital flowing into China. Put differently, in 2004, private investors around the world wanted to finance a large Chinese current account deficit, and provided China with $150 billion (almost 10% of China’s GDP) in net capital inflows. Private investors were willing to finance a small US current account deficit, and provided the US with say $200 billion in net financing, enough for a roughly 2% of GDP US current account deficit. Sustaining a global equilibrium with a Chinese current account surplus and an enormous US current account deficit, given those kinds of private financial flows, required that the People’s Bank of China intermediate between global demand for Chinese assets and the United States need for financing.

It did so in 2004, and it obviously did so for a reason (see Steve Roach, among others). The current system is delivering rapid growth in China. But that does not mean that the current system does not also impose substantial costs on China. Over the next two years, Nouriel and I suspect those costs will become increasingly apparent, and China’s willingness to continue to "overfinance" the US will fall -- forcing the US to start to adjust ...

Obviously, that is a debatable proposition, but given its importance to the global economy, it also something worth debating!