Over the past few years, US imports have grown faster than US exports. The result — to no one surprise: an expanding trade deficit. Because of the gap between imports and exports, US exports need to grow around 60% faster than US imports to keep the deficit from growing. Even if you exclude oil, that hasn’t happened on a sustained basis.
The trade deficit grew even as the dollar slumped v. the euro. That doesn’t mean that changes in exchange rates don’t matter, as some argue. Recent US export growth has been over 10% — well above its long-term average. The dollar’s slide (and Boeing’s recovery) is central to that. But import growth was still stronger than export growth. US demand growth has propelled global demand growth, and foreign producers have captured a large share of the increase in US demand.
What hasn't happened? At least not yet? Interest payments on the United States growing stock of external debt haven't contributed to the growth in the current account deficit. Yhe US income balance probably is mismeasured – something Daniel Gros has emphasized. The reported return on foreign direct investment in the US is a bit to low to be believable. But even with adjustment to correct for the differences in how reinvested earnings on FDI appear in the data, the US income deficit hasn’t grown even as US external debt has increased.
Why? Largely because the average interest rate on US external debt has fallen. The interest rate on interest bearing debts fell above 6% in 2000 to around 3% in 2003 and 04. That drove the average rate the US pays on all its liabilities – including foreign direct investment – down as well. That started to change in 2005 – and, in my view, a growing deficit on income payments (think interest on the US debt) will start to drive the expansion of the US current account deficit.
That is one place where I disagree with Richard Berner. I don't think the improvement in the income balance in the first quarter of 2006 is sustainable. And I looked at it in some detail. The increase in China's May trade surplus and Japan's May current account surplus may also tell us something about the course of the US trade deficit. Finally, I worry that the fuel that the dollar's past decline has provided for export growth is beginning to run out — and that without further declines in the dollar, current, strong US export growth won't be sustained. Should US demand growth slow, the world may slow too … slowing US export growth absent further dollar declines.
More generally, the dynamics of adjustment are likely to be daunting — something I explore in some length below (with a couple of graphs)
The following graph shows what happens if the pace of import and export growth moderate a bit – and both retreat to their long-term averages. As a result, the expansion of the US trade deficit slows. But the expansion of the US current account deficit doesn’t slow. Existing debts get repriced at higher interest rates as they come due. And borrowing $1 trillion a year at 5% plus starts to add up.
It is often noted that the US is in a better position than most big debtors because it borrows in its own currency. That’s true. I would be really freaking if the US was borrowing in RMB. It also means that the US gets a boost from dollar depreciation, which increases the dollar value of (some) US external assets. That boost is particularly strong if the dollar depreciates v. European currencies – since most US investment is in Europe.
But the US has a different external vulnerability. With gross debts of nearly $8.6 trillion at the end of 2005 – nearly 70% of US GDP (the end 2005 estimate is mine; the formal data will be out soon) and an estimated 2005 net international investment position of around $3.2 trillion (25% of GDP), the US is increasingly vulnerable to an interest rate shock. Most US lending to the world is in dollars and is tied to US rates – generating offsetting income should US rates rise. But the US would still be hurt on its net debt, and borrowing abroad to invest in equities and the like would be less profitable. With a net (interest bearing) debt position of $4 trillion (gross debts of $8.6 trillion v $4.6 trillion in external lending), a big rise in rates would really hurt. And even a little rise can have an impact. The 2005 interest rate was around 3.6%. It will go to 5% or more. That would push interest payments on existing US debts up by $65b – and the US is adding about a trillion in new debt a year.
The sharp rise in net interest payments from 2005 to 2009 consequently is payback in some sense for the absence of an income deficit from 2000 to 2004. The average interest rate on all US external liabilities — counting the low returns on portfolio investment and foreign FDI in the US — fell from 3.7% in 2000 to 2.5% in 2003. In this projection, it rises from an estimated 3.4% in 2005 to 4.8% in 2006, 5.25% in 2007 and then stays at 5.5% after 2008.
Why will the rate rise to 5.5% — well above the 3.4% rate in 2000? Two reasons, other than 5.5% is nice round number. First, the implied return on US FDI in 2000 was very low – only 2%. It will be above 4% in 2005. Despite all my complains about discrepancies in the way reinvested earnings are counted in the US data, the BEA has gotten a bit better at collecting this data. Without a recession, the implied return on foreign investment in the US is going to presumably remain at 4% or more. That pushes the rate up. The dividend yield on US portfolio equities held abroad has also increased – though it remains pretty low and pulls the average down. Second, the composition of US debt held abroad is changing. In 2000, it was about ½ short-term bank lending and ½ long-term US debt securities. Now longer-term US debt securities are about twice as big as the short-term stuff. Foreign holdings of corporate bonds – to the extent that they don’t disappear from the survey data – will tend to push up the average rate. John Kitchen –whose forecasts have a lower average interest rate going forward – doesn’t account for the changing composition of US external liabilities. I don’t think 5.5% in 2008 is unreasonable …
Footnote: I also assumed a slightly higher rate of return, around 6.5% once US rates normalize, on US assets than on US liabilities in these forecasts – enough to generate a $100b bonus for the US. Call it $2 trillion in dark matter. Personally, I think that is too much, since it basically counts the difference in reinvested earnings as dark matter. But rather than fight the dark matter fight, I just made a simple, standard assumption. The US has gotten a roughly $100b bonus from FDI for the past few years. I project it will continue to get that bonus – but don’t assume that the US will get any more.
Remember, this expansion of the US current account deficit is what happens if the pace of US import growth slows. This projection is consistent with a weaker dollar. Or with slightly slower US and global growth. It most certainly doesn’t represent a continuation of the current import boom. And even in that case, the US current account deficit is headed toward 9% US GDP in 2008.
Compare that graph with the following graph, which shows what happens if US exports start to grow significantly faster than US imports. Roughly twice as fast. I assumed that change starts in 2007. Given the momentum in year over year data (growth in the tail end of 2005 is currently pushing up 2006 imports, for example) it is hard to see how the basic pattern could shift much faster.
Such a change would represent a significant shift in the world economy. Rather than growing, the US trade deficit would start to fall. US import growth is around 5-6% in the projection – basically in line with US GDP growth. It is an environment where those countries now relying on exports to propel their growth have a hard time. That’s you, China. And a few others.
But the fall in the US trade deficit back to its 2000 levels doesn’t bring about any fall in the US current account deficit. At least not if the average interest rate on US external liabilities rises toward the shockingly high (note the irony) nominal rate of 5.5%. Rather, the interest payments that come from the repricing of the United States $8 trillion plus on its existing debts and paying a bit over 5% on roughly a $1 trillion in annual borrowing the US needs keep the US current account deficit above 7% of GDP through 2010.
That, folks, is my definition of soft landing. US growth presumably slows. Something has to change to bring US import growth down. The dollar weakens. Remember, the projection assumes US export growth is nearly twice as high as it was from say 1995 to 2005. Rather than growing in line with US GDP, exports rise as a share of GDP – while imports stay constant as a share of GDP. Global growth presumably is a bit weaker without as much impulse from the US. And so on.
To sustain that soft landing equilibrium, a few things have to happen.
- First, those foreign investors who now hold the $13.9 trillion in claims on the US have to be willing to hold onto those claims. That includes central banks and oil investment funds that now presumably hold well over $3 trillion in claims on the US – personally, I think the actual number might be closer to $3.5 trillion, counting all the Gulf investment funds.
- Second, American investors have to be willing to continue to keep the vast majority of their wealth in the US, and in dollars. That cannot increase their $10.7 trillion or so of (estimated) external assets. If Americans want to add to its investments abroad, the financing the US needs from the rest of the world goes up.
- Third, a combination of foreign central banks and private investors abroad has to be willing to add $1 trillion a year, give or take, to their holdings of US assets. And they need to be willing to do so for nominal rates of around 5%, no matter what happens to the dollar.
This is all just basic math. If you are running deficits, you need to borrow – so someone has to lend to you. And the US will be running big deficits for a long time even if it starts to adjust.
The US isn’t the banker to the world, despite what some still say. It is a borrower from the world.
Getting rid of these borrowing requirements faster implies a hard landing.
This analysis has one interesting implications. At some point in the future, the world’s central banks will no longer be financing export growth when they add to their reserves. Remember, in an adjustment scenario, US import growth slows – and US export growth rises. US imports rise in line with US GDP. Nothing exciting.
Rather, central banks will increasingly be financing interest payments back to themselves.
In some sense, central banks may have no choice but to do this – in the same way that Citibank (and others) found itself forced to lend new money in the 1980s to help big Latin debtors pay interest on the principal they already owed Citi. But it does imply a bit of change. The countries governments – and their electorates/ populations/ businessmen – may not enjoy financing (minimal) payments on their existing debts as much as they enjoyed financing a surge in US imports. Far more constituencies benefit from a surge in their exports to the US than from a surge in interest payments to the central banks of the emerging world.
I’ll have more to say on how this forthcoming change will change the political economy of the Bretton Woods 2 system of central bank financing of the US on another day – I already have gone on for too long.