Has the broad dollar fallen enough

A standard rule of thumb, mentioned by Paul Blustein in his excellent look at the impact of a falling dollar on US trade last week, is that a 10% fall in the broad dollar leads to a $100 billion fall in the trade deficit. Before last week’s rally, the dollar had fallen by about 15% from its peaks. Will this -- after all the lags work themselves through the system -- lead the trade deficit to fall by $150 billion.

The post on US travel to Europe -- still rising despite a strong dollar -- hints at my answer: No.

Starting points matter. The dollar (using the Fed’s broad dollar index) rose 10% between 2000 and 2002. It then fell by 15%. I don’t think the full impact of the dollar’s rise ever registered in the trade deficit -- lots of other things happened in 2001 and 2002 that got in the way. If the deficit never reached "its equilibrium" level with normal US and world growth at its peak, then the fall back to its 99 level is unlikely to produce major improvements in the trade deficit. Remember that the 99 level of the dollar was still well above the dollars’ level in the first half of the 90s, and the work of people like Cathermine Mann suggested the deficit would tend to expand over time even at this level.

When the Nasdaq fell from 5000 to 4000, it was just less overvalued than before, not fairly valued! The same could be said for the dollar’s rally (v. the Euro) from 88 cents/ euro to $1 or even $1.10 or so per euro ...

Back in 1997, I was in charge of a professional quality trade model. If you plugged the impact of the dollar’s rise from 90 to 95 into the model, the model said the deficit would rise signficantly, particularly if world growth slowed (as one expected at the time). That is what happened: the trade deficit doubled, going from 1.5% of GDP to 3%. A dollar at 95 back then was consistent with a steadily widening trade deficit. Does that mean that a fall in the dollar from 110 to 95 won’t have an impact? Of course not. It will help US exports, and once J-curve effects work themselves through, it should slow the growth in import volumes. But I also suspect that at historic growth rates, a broad dollar at 95 is also consistent with a trade deficit that widens over time.

So there is reason to doubt whether current fall in the dollar in the course of 2004 will lead the deficit to start to fall in 2005. Not unless some other things happen too. As Steve Roach has recently reiterated, higher real interest rates in the US -- which will slow asset appreciation and thus reduce consumers ability to finance consumption by borrowing against their growing assets -- are also essential. Unless import demand growth in the US slows, it is impossible for the US to export its way out of its trade deficit. In October, non oil imports were up [15%] y/y. Given the gap between the US export and US import bases, if imports grow at 16%, exports have to grow at around 24% to prevent the trade deficit from widening.

To use a bit of economic jargon -- there has been an adjustment in the dollar, but no adjustment in the US trade or current account deficit. No fall in imports. This discrepency leads to tortured analysis from many currency strategists. Consider this graft from Andy Xie of Morgan Stanley:

The dollar value against major currencies is 20% below its 30-year average in nominal terms and 10% below in real terms. As I search around the world, the dollar is the only thing left that is still cheap.

Currency strategists often note that "the dollar is undervalued" -- while also mentioning that the dollar likely has further to fall to correct the US trade and current account deficits. That is a bit confusing: how can the dollar both be undervalued and at levels that will leave the US with an unsustainable trade deficit?

If the dollar has to fall further to correct an unsustainable trade and current account deficit, in simple terms, the dollar is still overvalued.

It is pretty clear that the Chinese renminbi, which still follows the dollar on its journey through global currency markets, needs to appreciate v. the euro. But does the dollar? Put diferently, if Asia continues to peg to the dollar (China, Malaysia, Hong Kong), intervene to keep the dollar stronger than it otherwise would be (Korea, Taiwan, India) or scare the bejesus out of speculators with the prospect that it might buy another $300 billion to support the dollar (Japan), is the euro fairly valued at 1.30? Or, put differently, has the broad dollar fallen enough to begin the process of global rebalancing.

Over the past twenty-five years, the US trade balance has trended toward a bigger deficit. Looking at 20 or even 30 year averages therefore may produce a misleading picture of the dollar’s fair value in a world where the US trade deficit HAS to start to trend down over time.

Of course, the dollar has moved far more against some currencies (the euro) than others (the reminbi). The dollar could well fall further on a trade weighted basis and still rise a bit against the euro.

Finally, there is China. it is too big and important not to ignore. Two key points:

1) the current measures of the broad dollar underweight China. China was about 6% of US trade in 2000, when the indexes were last reweighted. It is now 10%. And its share of US trade is rising fast. As we all know, China’s exchange rate has not appreciated v. the dollar.

2) China is investing like mad -- investment is now almost 50% of GDP. Lots of that investment is going into the manufacturing sector: China’s capacity to produce goods for sale abroad is growing extremely rapidly. A lower pace of investment in 2002 and 2003 produced the capacity that let China increase its exports to the US (using US import data) by 25% in 2004. If chines exports to the US grow at this pace in 2005, even if US exports to China continue to grow strongly, the bilateral deficit with China will increase by about $40 billion. Keeping the overall deficit constant therefore requires that the US bilateral deficit with the rest of the world fall by $40 billion.

By the way, a change in the renminbi dollar would produce the mother of all J curve effects: in the short-run, it will cause the trade deficit to rise. Chinese exporters would raise their dollar prices to cover their renminbi costs. Import prices would go up. Import volumes would take some time to adjust -- our import bill would go way up in the short-run. It is still important that this change happen: higher export earnings will lead China over time to import more, and thus increase US exports; and the change in the renminbi dollar will reduce incentives to locate new investment in china, and increase incentives to invest in new US capacity.

If my forecast for a widening bilateral deficit with China is right, getting the overall deficit to fall requires an improvement of more than $40 billion in our trade balance with the non-Chinese world.

US imports from China have doubled since 2000, when the "trade wieghts" used in the dollar index were last rejiggered. Since the dollar has not moved against the renminbi, and since China is now a more important US trading partner than it was in 2000, the standard dollar indexes tend to overstate dollar weakness.

Global capital flows matter. The US trade deficit was relatively small in the mid 1990s both because the dollar was relatively weak against both the world’s major currencues -- the Euro, then the D-Mark, and the yen -- and because most emerging economies were attracting significant capital inflows and were running significant trade and current account deficits. Remember that most emerging Asian economies had significant current account deficits before the Asian financial crisis, and the major Latin economies also had large current account deficits. A boom in capital flows to emerging economies contributed to strong US export growth from 93-97.

The US context matters as well: as long as the budget deficit is staying constant or is shrinking modestly at best, higher consumption and growing investment in the US necessarily lead to a wider current account deficit. Strong consumption -- specifically consumption growth that exceeds income growth -- implies less savings, and thus more need to borrow savings from abroad to finance any given level of investment. That is the danger created when a country with a low savings rate runs a structural budget deficit -- good times tend to lead to a risky widening of the current account deficit. Look at what happened in Latin America in the 1990s.