Have you ever wondered why food is so cheap? Food - including restaurant meals - is only about 15% of the average American's budget, according to the
Bureau of Labor Statistics. Contrast that with housing, at 42%, or transportation, at 18%.
It didn't used to be that way. It wasn't that long ago when food was the average person's biggest expense. What changed?
What changed was the industrialization of agriculture. The labor input for farming was reduced by the use of machinery, replacing the work of men and animals with fossil fuels. More importantly, "green revolution" technology in the middle of the last century increased yields by using large amounts of fossil fuel based fertilizers - replacing the work of the sun with fossil fuels. With labor, draft animals, and the sun no longer limiting factors, we could produce food as fast as we could pump oil out of the ground. We can produce it much faster than we can eat it.
As a result, the price of food plummeted. It plummeted to the point where the government started paying farmers not to grow anything. It plummeted to the point where we give food away free to starving third world nations to help exacerbate their population problems. We even started turning food - in the form of edible corn kernels - into ethanol, and mixed it with gasoline to substitute for some of that fossil fuel.
Those familiar with thermodynamics will notice a problem here. Fossil fuels can help us grow more food, but it won't grow enough food to provide more fuel energy than the original fossil fuel used. The sun does still make a contribution, but not much. Fossil fuels provide about three quarters of the energy content of corn based ethanol; the sun only adds the last quarter. Ethanol is much friendlier to oil companies than to the environment.
The good news is that some senators seem to be realizing that the numbers here don't really support ethanol production, and are trying to get the ethanol subsidy suspended. Maybe we can go back to eating the food - and coming up with better fuel alternatives. We could start with using the nonfood parts of plants - stalks, leaves, and other parts that currently go to waste - instead of the edible parts.
Sugar-based ethanol is silly, but it's not quite as bad as all that. Fossil fuels aren't entirely interchangeable - we don't use gasoline in cars because it's efficient per se, but because it's very convenient engineering-wise. (We could go back to coal-fired steam cars, but I don't see that catching on.) It's doubtful you could run a whole civilization on corn ethanol, but you don't have to - it's a special-purpose fuel.
So from that perspective, trading natural gas or coal or hydroelectric power for automotive fuel might be fine, even at a bad exchange rate. On the other hand, once you admit that's the plan, you may want to just bite the bullet and do it directly via gasification (as the Germans did in WWII when oil was short) rather than fiddling about with the corn.
And as you say, in the long run cellulosic ethanol, or better yet something other than ethanol (octane? biodiesel?) could be entirely workable.
I agree that the overall issue of biofuels is more complex. In practice, though, I think the inputs to corn based ethanol in the U.S. are largely crude based; the farming areas aren't generally particularly close to the fossil fuel extraction locations, and oil is the most convenient to transport. I'm also pretty sure the subsidies are for ethanol as a general purpose motor fuel, and cause overproduction for the applications for which ethanol is well suited from an engineering standpoint.
Converting hydroelectric power would not make sense energetically. The efficiency of converting fuels - including ethanol and fossil fuels - to work is only 20-40%; it's over 90% for electricity. Once you've got electricity, you want to keep it as electricity.
Now, sugar cane based ethanol from Brazil may make more sense than corn based ethanol from the U.S. Brazil gets a lot more sun, and perhaps they use less fertilizer as well, so the energy balance may be much better. Ethanol from Brazil doesn't get the subsidy, though - in fact, it's subject to heavy import duties. That again illustrates that the ethanol subsidy is really just an agricultural subsidy; if it were an environmental issue, we wouldn't be discouraging importation.
As you hint at, ethanol isn't as good as oil as a motor fuel, anyway. Its specific energy is much lower - about 7 kcal/g versus about 10 for oil - so you have to cart around more of it. That makes the transportation overhead higher for ethanol, which may be why one can only easily find 80% ethanol pumps in the same farm states that produce it.
Biodiesel, or better yet, simply refining vegetable oil into gasoline, would be a better solution. However, with present methods, it would still be subject to the energetic inefficiencies I described for ethanol; that's why vegetable oil at the grocery store still costs more than gasoline. It does make sense as a way of recycling used cooking oil, but for it to make a big impact on our fuel use, we'd need to switch to a different feedstock - algae has been suggested - or a different conversion method. Ultimately, I think such a switch will be the way to go, but the fact is, the ethanol subsidies actually retard progress in that direction.