It is always painful to see an economist do battle with an accounting identity. Martin Feldstein takes on the task in a piece that explains that the dollar could fall sharply in the future, if investors come to expect its decline.

Most of what Feldstein says in this piece is exactly right. Investors do not need to hold dollars as a safe haven and in fact it is not a safe haven if it is falling in value against other currencies. Also, the vast majority of foreign dollar holdings are not necessary as foreign exchange reserves to finance trade. This means that we could in fact see large declines in the value of the dollar in the future.

The conflict with national income accounting comes in Feldstein's takeaway:

"And, despite a more competitive exchange rate, the US continues to run a large current-account deficit. If progress is not made in reducing the projected fiscal imbalances and limiting the growth of bank reserves, reduced demand for dollar assets could cause the dollar to fall more rapidly and the interest rate on dollar securities to rise."

Okay, if we reduce our fiscal imbalances and there is no decline in the dollar (remember, that is the bad news in this story) then what will happen to employment? Feldstein can't seriously believe that investment in equipment and software will increase substantially from what is already a reasonably high level measured as a share of GDP. He also can't expect that consumption will rise substantially given that the savings rate is already at a very low level compared to the post World War II average. And with enormous overbuilding in both the residential and non-residential sector, he can't think that construction will fill the gap in demand that will result from smaller deficits.

This would mean that the policy he is advocating is that we should deliberately slow growth and raise unemployment because if we don't the dollar will fall in the future. This would seem to contradict both common sense and what Feldstein himself has advocated in the past: allowing the dollar to fall to boost net exports.

Fans of accounting identities know that the only way that the United States will be able to have something resembling full employment without large budget deficits is with a sharp decline in the trade deficit. This in turn will only happen with a sharply lower dollar, precisely the event that Feldstein wants us to fear in this piece.

Of course the main obstacle to a sharply lower dollar now is the efforts by China and other countries to prop up the dollar against their own currencies to protect their export market in the United States. Perhaps these countries will all agree to raise the value of their currencies, as we have ostensibly been requesting, but that is not a change that is likely to occur anytime soon. If and when it does it would be good news for the U.S. economy, leading to a surge in manufacturing and employment, not the disaster Feldstein portrays.

It is always sad when economists take issue with accounting identities. It is even sadder that, in U.S. policy debates, they often win.