In a discussion of trade imbalances the Washington Post told readers that: "it was that risk -- of a collapse in the value of the dollar and of U.S. government securities -- that kept many economists up at night."Actually, competent economists were not terribly worried about this nearly impossible scenario.
China and other countries were deliberately propping up the value of the dollar in order to sustain their exports to the United States. While these countries may at one point back away from this policy because they decide it is no longer in their interest, it is almost inconceivable that they would flip overnight to the opposite policy of allowing their currencies to soar against the dollar. The idea that China would allow an exchange rate of say 4 yuan to the dollar or that Europe would tolerate an exchange of 2 dollars to the euro is almost absurd on its face. The market for these countries' exports in the United States would collapse at these exchange rates, while U.S. exports (we still export more $1.7 trillion annually) would become hypercompetitive in other countries, wiping out domestic competition.
Since the story of a dollar collapse was so far-fetched, competent economists did not lose sleep over it. They did lose sleep over the housing bubble, the collapse of which produced the economic disaster the country is now witnessing. (Unfortunately, the folks running economic policy were not among the group of economists paying attention to the housing bubble.)
Strangely, currency prices receive only passing mention in this piece on trade imbalances. This is the mechanism for adjustment. In order to move the U.S. trade deficit closer to balance, the dollar will have to fall against other currencies. There is no other plausible mechanism. It is difficult to understand why this point was not mentioned.
It is worth noting that the savings rate has increased by about 2.0 percentage points more than is implied in this article. This is the result of the statistical discrepancy in GDP accounting. At the peak of the bubble, capital gains income was showing up on the income side as ordinary income. This overstated true income and therefore overstated the savings rate. With the collapse of the housing bubble and plunge in stock prices capital gains income is no longer showing up as income in GDP accounts to the same extent. Therefore the savings rate is no longer overstated. This adjustment means that the savings rate has risen by about 2.0 percentage points more than the official data show.
[Addendum: In response to comments about the capital gains issue -- I am referring to the NIPA measure of income and savings, which is not supposed to count capital gains income. However, capital gains income did show up in this measure during the years near the peak of the stock and housing bubbles. The statistical discrepancy turned strongly negative during these years, which means that measured income side GDP was larger than measured output side GDP. (By definition, they should be equal, although measured output side is usually larger.)
We regressed the statistical discrepancy on lagged increases in stock and housing prices. The fit was extremely strong, with a very simple regression explaining almost 60 percent of the variation in the statistical discrepancy. Based on this analysis, I think it's pretty clear that the official data substantially overstate income and therefore the saving rate both at the end of the 90s and the years 2004-2007.]