Paul Krugman may have misled readers of his blog yesterday with the comment:
"the trade balance is a macroeconomic phenomenon, determined by the excess of savings over investment."
As an accounting identity the trade deficit is equal to the excess of national investment over national savings. However it would be wrong to conclude that the U.S. trade deficit is caused by our failure to save enough, especially in the current context where the economy is well below its potential level of output.
To take a simple example, suppose that we all become virtuous savers and reduce our consumption by an amount equal to 1 percent of GDP (@ $170 billion annually). This would reduce demand in the economy by $170 billion. In more normal times we might tell a story where this fall in demand would lead to a drop in interest rates, which would in turn spur additional investment. Lower interest rates should also lead to a lower valued dollar (fewer people want to hold dollar denominated assets at a lower interest rate). The lower valued dollar would lead to more exports (our goods are now cheaper to foreigners) and fewer imports (foreign goods are now relatively more expensive than domestically produced goods).
In this story, the end result is that we have the same level of output with higher levels of investment and net exports replacing the lost consumption. We have a somewhat higher level of national savings (the increased investment partially offset the rise in savings) and a lower trade deficit.
That would be the standard story of how a savings-investment balance determines the size of the trade deficit. However, no one can tell this story in today's economy. If everyone started saving more as described above, it would mostly just lead to a fall in output and employment.
The reason is that the adjustment process would not come close to offsetting the loss in demand. With the short-term interest rate already at zero we would see no help there. Long-term rates could fall some, but the reduction in longer term rates would at best have a trivial effect on investment. The dollar may not move at all, both because interest rates will have changed little and also because many countries (yes, China is the biggest) have a policy of targeting the price of their currencies against the dollar. If market forces started to push the value of the dollar down against their currencies they would respond by buying more dollars to keep up the value of the dollar.
In this story, savings will actually rise by considerably less than the initial $170 billion increase in savings because GDP will have fallen. This means that people who had been saving instead find themselves unemployed and spending from past savings (dissaving). The government will also be saving less (running larger deficits), since it is collected less in taxes and paying out more in transfers like unemployment benefits. There would be some reduction in the trade deficit since at lower levels of GDP we buy less of everything, including imports, but for the most part the trade deficit and national savings balance is maintained by lower savings from the reduction in GDP offsetting most of the increased in intended savings.
By contrast, if foreign countries suddenly started buying more of our stuff (say the dollar fell by 20 percent) then we would see an increase in employment and output. This would lead to more savings as formerly unemployed workers get jobs and can now start putting money into the bank. Also government savings increases as increased employment means more taxes and less money paid out in transfers. The net effect is that a lower trade deficit leads to more net national savings.
When considering these accounting identities it is important to keep the stories on causation straight, otherwise you get some really bad policies. Paul Krugman of course knows this and has made the same point many times (here for example), but we must work hard to prevent confusion on the topic.
Note: link fixed, thanks Squeezed Turnip. Also, typo corrected.