Richard Iley of BNP Paribas thinks so. In a comment to a post earlier this week, he noted that BNP Paribas was as bearish as Merrill – and that falling US rates and falling returns on foreign direct investment should help the income balance.
Given Roubini’s extremely well-known bearishness on the US (He now puts the odds of recession at 90% -- above Merrill and BNP Paribas) and my almost as well known bearishness on the path of the US income balance, Richard Iley’s comment prompted me to stress test my assumptions. Can Roubini be right about the US economy and I be right about the US income deficit?
The bottom line – We can both be right. A fall in the average interest rate on US external debt (actually -- a fall in the rate on new borrowing that keeps the average rate from rising) from a US recession won’t be enough to keep the income balance from deteriorating. To keep the US income balance from deteriorating, returns on foreign direct investment in the US have to fall off a cliff.
That is possible: returns on FDI certainly fell to very low levels in the last US recession. But I also suspect that expectations that returns on foreign FDI will fall back to the 1-2% face a headwind of their own: the BEA’s efforts to improve the quality of the US data. Details – with charts – follow.
Let’s start by reviewing the state of play.
By my calculations, US paid roughly 4.2% on its external debt in the first half of 2006, and received 5.05% on its external lending in the first half of 2006. The fact that the US got a higher rate on its lending than it paid on its debt is a big reason why the income balance didn’t deteriorate more in the first half of 2006. I don’t think this reflects the United States’ skill as a financial intermediary so much as the shorter-maturity of US lending. As I discussed previously, when US rates fell in the first part of this decade, the interest rate on US lending fell faster than the interest rate on US borrowing.
US FDI abroad returned a bit more than 6% -- 6.3% to be exact -- in the first half of 2006. The implied return fell a bit from the 7% average between 2002 and 2004 largely because the value of US investment abroad soared in 2005. Actual cash flows didn’t change much. The return on foreign direct investment in the US rose to around 4.7% -- largely because reinvested earnings soared. The implied return on foreign direct investment in the US is still not all that high, but it has trended up recently. My guess is that this reflects the BEA’s efforts to improve the quality of its data as much as anything else. The extremely low returns and absence of any reinvested earnings in 2001-2003 didn’t make much sense.
My baseline forecast assumes that the interest rate on US external borrowing will, over time, rise to around 5%. I am not sure how quickly this will happen, though my guess is that the process will be complete by say 2008. In 2008 all the five years issued in 2003 will need to be refinanced. In the following graph, I assumed it will happen faster – by 2007. The rise in 2008 reflects the ongoing rise from $1 trillion in assumed additional borrowing. In some sense though it doesn’t matter how the rise is distributed between 2007 and 2008 – you end up in the same place. I also assumed that the interest rate on US lending rises marginally, to 5.25%. As the chart shows, the result is a substantial headwind – net interest payments rise by $125b between 2006 and 2008. Actually they rise by $125b from their levels in the first half of 2006 – the 2006 entry comes from annualizing the first half data, the deficit for the full year should be larger.
I assumed that the rate of return on foreign FDI in the US and US FDI abroad won’t change in 2007 or 2008, and that the stock of both US and foreign FDI will rise – with the stock of US FDI abroad rising by slightly more than the stock of foreign FDI in the US.
The result is that that net income from FDI in the income balance rises in line with the recent trend -- but the overall income balance still deteriorates dramatically.
One note – the 2006 income balance is based on data from the first half of the year annualized.
I could easily come up with a scenario showing even more dramatic deterioration in the income balance. Back in 2000, the average interest rate on US external debt topped 6%. In 2008, US gross debt will likely top $12 trillion – and there is a big difference between paying an average interest rate of 4.25% of $12 trillion and paying an average interest rate of 6.25% of $12 trillion. Real big.
What, though, would happen in the scenario Richard Iley outlined in the comments. The US enters into a growth recession. The fed starts to cut rates. And the return on foreign direct investment tumbles.
The fall in rates doesn’t actually help much. I modeled a fall in rates by assuming the interest rate on US lending abroad (mostly short-term) falls back to around 4.25%, and the average interest rate on US external borrowing stays constant at around 4.25%. That is about as good as it gets – in practice, I suspect that the average interest rate on US lending would fall below the average rate on US borrowing. But it still doesn’t keep the income balance from deteriorating – as the income from the United States roughly $5 trillion in lending falls and payments on US external debt rise because, well, the US has to finance ongoing deficits.
But if the return on US FDI abroad stays at around 6.3% and the return on FDI in the US falls to 1.7% (its average level between 2001 and 2003), well, that really does help. The US has received – if one believes the data, and counting reinvested earnings – about $100b-$150 more on its FDI than it has paid for the last five years. Under this scenario, the US gets about $250b-300b more.
It basically keeps the income balance constant.
I don’t think this is a likely outcome – largely because I think the very low historical rates of return on FDI (and extremely low returns during the last recession) reflect bad data, not just a recession – and the BEA seems to be making an effort to improve the data. The average return on FDI has steadily trended up: going from 1.1% in 2002 to 2.35% in 2003, 3.3% in 2004, 3.7% in 2005 and 4.7% in the first half of 2006. Moreover, the return on US FDI abroad didn't fall noticeably when Europe had its recession -- and most US FDI is in Europe.
But a huge fall isn’t totally implausible -- if the US really slows and the slowdown produces a big fall in the profitability of FDI in the US.
So there you have it. If the US proves to be a truly terrible place to invest the income balance won’t deteriorate over the next few years.
OK, in this scenario, investments in long-term bonds do ok – falling interest rates and slowing US economy are good for most bonds. At least if you have a good currency hedge. It is just every other US investment though doesn’t do well.
That highlights the problem I have with most projections showing a relatively benign evolution of the income balance and the net international investment position. If you assume that equity investments outside the US will continue to dramatically outperform US debt (and US equity) in dollar terms, the US NIIP won’t deteriorate that much …
But if investment outside the US does so much better than investment in the US, why would the rest of the world rationally continue to lend almost all their spare savings to the US?
I kind of suspect some folks in China are asking the exact same question. Fan Gang for one:
``The U.S. dollar is no longer, in our opinion is no longer, (seen) as a stable currency and is devaluating all the time, and that's putting troubles all the time,'' Fan said, speaking in English, at the World Economic Forum in Davos, Switzerland. ``So the real issue is how to change the regime from a U.S. dollar pegging to a more manageable reference, say euros, yen, dollars -- those kind of more diversified systems.''
That at least was what he said before he joined the PBoC’s monetary policy committee ..
A bunch of methodological notes – this estimate of the income balance leaves out dividend payments on portfolio equity. They aren’t big, but it is a bit off as a result. And I estimated the implied rates of return by comparing the observed flows in the BEA data (adjusted for dividend payments) with the stocks implied by taking adding 2006 capital flows to end 2005 stocks. I assumed gross debt rises by $1 trillion a year, and gross lending stays constant. That is a strong assumption, but since in both scenarios the gap between lending and borrowing rates effectively disappears, it doesn’t matter much. And I made a set of (reasonable) assumptions to forecast FDI stocks in 2007 and 2008. I treated foreign FDI and US FDI very similarly – net flows of both were $200b, and the market value of existing stocks rose by 5% a year. I used very simple assumptions to try to pinpoint the impact of shifts in the rates of return. You could certainly argue that a US recession would lower the market value of foreign equity in the US – though so far a slowdown in the rate of growth can coincided with an equity rally …