This optimistic analysis of the capital inflow is,wrong. It results from a raisin-. terpretation of the data provided by the US Treasury in the press release for its monthly Treasury International Capital report. It is easy to see why analysts reach this wrong conclusion. Recent TIC press releases stated that the . capital inflow was $278bn in the third quarter of last year, or $82bn more than the current account deficit for that quarter.. The Treasury also reported that $257bn of this capital inflow came from private buyers.

In reality, there is no excess capital inflow and private investors are almost certainly not the primary source of the funds coming to the. US. The figures in the TIC press release, while technically' correct, are misleading for two reasons. First, the TIC release refers only to transactions in long-term securities, that is to equities and long-term bonds. It excludes- bank deposits and bank lending, and flows of foreign direct investment into the US and by American investors to the rest of the world.

A comprehensive measure of the capital inflow and outflow would show that. the total net inflow is almost exactly equal to the amount needed to finance the current account deficit. Whenever the net long-term capital inflows exceed that deficit, the difference is offset by a net outflow of short term funds and direct investment.

A second source of confusion in the TIC report is an. easily misunderstood classification of whether the funds coming to the US are from governments or private sources. The TIC measure of inflows from "private" sources overstates the actual private investment because it does not distinguish between a purchase by 'a private buyer for its own account and a purchase executed by a private institution on behalf of a foreign government. For example, if the. Chinese government purchases US bonds through JPMorgan or another private bank, these funds will be recorded in the TIC data as a private purchase. Similarly, purchases of dollar assets by governments of the Organisation of the Petroleum Exporting Countries or their investment authorities that are done through British banks would look like private purchases with a British origin.

My own belief, based on widespread conversations with officials and with private bankers, is that the inflow of capital that now finances the US cur-rent account deficit is coming primarily, perhaps overwhelmingly, from governments and from institutions acting on behalf of thosu governments.

The nature of the current capital inflow is very different from the experience as recently as five years ago. In 2000 and before, the current account deficit was financed by a net inflow of equity funds. Some of this took the form of foreign direct investment while the rest was portfolio investment. The. fact that the net inflow was equity implies that it was private foreign investors who were shifting funds . to the US. Now there is very little equity flow to the US (only $37bn of equity securities in the third quarter of 2005) and the corresponding equity outflow ($32bn in the third quarter) has often been as large or larger than the inflow. The very large current account deficits are now being financed by bonds and shorter term fixed-income funds.

Some of this has recently come from Opec governments and other oil producers that are temporarily placing revenue in dollar bonds and bank deposits until they can spend those funds on investment or consumption. Much of the inflow in recent years has come from Asian governments that wanted to accumulate foreign ex-change to eliminate1the risk of speculative attacks of the sort that hurt those countries in the late 1990s. A large amount is coming from China and other Asian governments to stop a falling dollar reducing their net exports. If they decide to buy fewer dollar bonds, the US current account deficit could not continue to be financed at current exchange rates and interest rates.

The US current account deficit increased from $668bn in 2004 to an annual rate of $790bn in the first three quarters of last year and is widely predicted to move much higher in 2006. This unprecedented level is equal to 6.4 per cent of US gross domestic product.

The current small interest rate differences in favour of US bonds are not nearly enough to compensate investors for the fall in the dollar that is likely over the next few years. Investments in 10-year government bonds receive only about 1 percentage point more on dollar bonds than euro bonds and about 3 percentage points more on dollar bonds than on yen bonds. The dollar must fall faster than these small interest differentials in order to prevent the current account deficit from increasing more rapidly than GDP. This means that investors in dollar bonds will eventually have lower cumulative returns, potentially very much lower returns, than investors in the bonds of other currencies.

At some point, that will trigger a shift away from the dollar. Private investors and the governments that are concerned about the total return on their portfolios will inevitably shift at some time from dollars to euros or yen to take advantage of the predicted rise in the value of those currencies and to avoid the loss of value of their dollar bonds. That that has not happened already reflects investors' belief that it is still possible to benefit from the interest differentials before the dollar depreciates. That sanguine belief may, however, reflect a serious misunderstanding of the magnitude and nature of the capital flow to the US.

The writer is professor of economics at Harvard University and president of the US National Bureau of Economic Research.