Die Time: In comments yesterday, Ed Foster mentioned "die time". This is exactly why you can't get, for example, .380 right now. At most manufacturers, the machinery they use to load .380 (which uses different "dies" to load different calibers) is only used for that purpose for a small portion of the year; the rest of the time it's used to load more high volume stuff, like 9mm. They churn out .380 for a couple of months, say, at the end of the year, and it's enough to hold the market over 'til the next winter.
Rumor has it that this past year's demand for 9mm FMJ was so great that Winchester didn't bother tooling up for .380 and kept the presses pumping out 9x19 ball. Even if the other two companies didn't do likewise, what percentage of the .380 market do you think that Winchester represents? Federal and Remington certainly don't load enough to take up the slack, and that causes the supply to crash to nothing.
Case Lots: Believe it or not, the kind of people that read gun blogs, post on internet gun forums, go to the range every weekend, and name their gun "Vera", are a small minority of gun owners. For fifteen years I tried to convince Joe and Jane Public to buy ammo by the case. I failed miserably. No matter how much you explain the price savings when buying a thousand rounds at a lick, or the fact that ammo doesn't go bad, most people would look at you and say "I don't know, $100 seems like a lot of money, and what am I going to do with a thousand rounds of 9mm?"
The ammunition manufacturing and supply pipeline is simply not set up for the average consumer to walk into Wal-Mart and buy two cases of ammunition. If your average shooter normally bought 100rds/month to take to the range and decided to buy two cases instead, "just in case", he has just bought more ammunition at one lick than he would normally buy in two years. Think about that for a second, and then multiply it out over several million shooters suddenly buying way outside their normal pattern.
Production Capacity: The manufacturers are running full tilt. The only way they could make more ammunition is to build more plant, and they are not going to do that for several reasons. The first is that this bubble will contract sooner or later. Joe and Jane Sofaspud are going to realize that they really don't need 10,000 rounds of Winchester .45 in the basement, and that minivan payment isn't getting any smaller. They'll sell it to Annie Appleseed and Ivan Ipsc and demand will cool down.
The second reason they won't build more plant is financial. Remember that economy thing? Yeah, well it's still bad. Business loans aren't really easy to get right now, especially for businesses that are square in the middle of the Media-Congressional Complex's crosshairs. When the stroke of a pen could cut your sales by 50%, you are not what lenders call a "good risk".
Supply & Demand: Right now prices are high not because of transport costs or raw materials costs, which drove the price spike of '05(really an honest adjustment, as ammo prices had stayed almost artificially flat for the better part of a decade,) but because of simple supply and demand. If I put my widgets out for $1, and the first guy that walks in the store buys them all, I'm obviously not charging enough for my widgets.
As demand stays high and supply stays small, prices keep going up. They will go up until they become high enough to cool demand. As demand cools, supply will build back up. In order to move the new supply, prices will come back down somewhat until they reignite demand. This is Econ 101, folks, and it's as predictable as 32 ft/sec² or π*r².
UPDATE: Grant Cunningham points out that when the supply chain gets sucked dry back to the manufacturer level, it can have ripples even farther up the line. Ammunition is not made of wishes and unicorn droppings, after all.