Your post is very interesting, especially as I have heard alot about the financial incentives driving U.S. forign policy. Currently the U.S. has a huge trade deficit and a massive budget deficit (and consequently national debt).
Since the U.S. enjoys the favor of being the World Reserve Currency, such deficits have been allowed to exist and grow. However, current U.S. spending is so large that the dollar is begining to become seriously devalued. Eventually countries will see that it is not in their best interest to use the dollar (Iraq, soon Iran). So far the Bush administration's answer has been to invade such a country. We shall see what happens as more countries decide to make the switch.

I want to comment about the inflation calculation issue. I don't know of any problem with food calculations in the CPI; energy and metals feed into the CPI through their impact on finished goods prices. Thus I would not agree that the price of oil and commodities goes unnoticed in the CPI. Yet the CPI calculation is important because the government pegs many transfer payments to the CPI, and price index calculations figure into interest paid on inflation-adjusted government bonds. That being said, the CPI calculations are flawed in a number of ways, and each of them would seem to lead to underestimation. It is a while since I worked seriously with the CPI (I helped forecast it for awhile when I worked for the consulting firm Data Resources, years ago) but in keeping up with the news coverage of the issue, here are some points that get my attention (if all you care about is my direct answer to the poster's question about oil and dollar devaluation, skip my numbered points and read the final paragraphs of this post):

(1) The Bureau of Labor Statistics (BLS) which calculates the CPI, uses "hedonics." The purpose of hedonics is to to measure price changes of a CPI component at constant utility.

According to this methodology, for example, IT products are getting cheaper because the cost per floating point operation (FLOP) is decreasing. I would take issue with that approach though. For the longest time a solid desktop machine cost $2,000 or so, no matter what upgrades it included. Despite the increase in product quality, if you wanted a decent desktop, that is what you paid. OK, that is finally changing, but you get the point. Consumers (and the calculation of consumer welfare is what counts in a CONSUMER price index) had to pay the same amount for "a desktop computer that did solid basic calculations, word processing, and communications," but at that time, BLS used increasing product quality per dollar, to factor in price "deflation" in the computing sector. Price per flop is not the component that drives consumer expenditure; the price a consumer needs to pay for consumer goods is what we are interested in, or should be, when we calculate the CPI.

(2) The BLS calculations for owner-occupied housing prices are notoriously flawed. One would expect BLS to measure housing prices by monitoring home price transactions and deriving from them changes in the average price per square foot of residential space. (There is a separate component of the CPI for residential rentals; that is a different calculation.)

(3) According to a recent article in, I believe, Barron's Financial Weekly, the medical expense portion of the CPI does not count the employer-paid portion of health insurance. I would argue with this for two reasons. First, it would seem absurd to argue that health insurance in effect costs less because someone else pays for a portion of it. Second, we all pay for health insurance, either through our own outlays or from pass-through of employer-paid costs into consumer prices. We cannot say that because we are not laying out the funds directly, the cost of healthcare is less. Is the cost really going up as people lose insurance, or have to make higher co-payments? No, just more people are paying for it out-of-pocket.

I apologize for the length of this post but if you want to understand the CPI you must know that it is an index composed of subindices, each of which comprises a piece of the theoretical market basket of a representative consumer.

Now we can address the poster's main point. The US current account deficit does require that other countries finance it by purchasing US treasury securities (or buying US companies or hard assets, or something else priced in dollars).  I'm not sure inflation is the big factor here. What matters to me is that we are exporting dollars in exchange for goods, and to the extent the quantity of US dollars exceeds the demand for things that are priced in dollars (at constant exchange rates) one would expect the dollar to be devalued (price of dollars in exchange of other currencies falls).

Now, if we denominate international trade of oil in dollars, oil producers will be loth to accept dollars (or at least hold them for any length of time) without some kind of compensation if they expect dollars to be devalued in the future. To the extent that the dollar is expected to fall, oil producers might try to raise prices, or get oil trade denominated in another currency, or at least a currency basket that is less heavily dollar weighted.

An increase in the price of gold shows that investors' preference for holding gold, which has a cost to hold in the vault, is increasing relative to their wishes to hold dollars, that pay interest. In my view that is a vote against the dollar. I would remind you, however, that the dollar is at the moment strengthening against other currencies. So maybe what investors are saying is that for the moment they would rather hold increasing amounts of gold, rather than the paper currencies of most countries.