Some breweries (like Rogue) are quite proud of their beer, and regardless of whether or not Dead Guy cost significantly more than Inversion they want to project an image high quality, expensive beer, so they price it higher.
Nobel prize winner Michael Spence won his prize for developing the theory of signalling. The idea is this: if you are a high productivity person (because perhaps you are especially smart), how to you credibly convey this to a prospective employer. You can't just say "I'm very smart" because if this increased the likelihood of you being hired or getting a higher wage, everyone would say it even if it were not true. So you have to find other ways to let them know, or 'signal' your type. One way to do this is to obtain more education. This works because acquiring this education is easier for those that are smarter. Less smart individuals will decide not to invest in education, because it is costly and, after employed, employers learn your true type. So in equilibrium there is a natural sorting, smart people get educated and less smart people don't and, here is the punchline, employers can use education as a real measure of smarts.
[By the way I tested this model (with Mike Conlin now at Michigan State) using new players in the NFL and found that players with some private information about their ability can signal this through holding out and delaying agreement on a contract. The idea is that missing some training camp is more costly for less able players than for more able players.]
This basic model has been translated into advertising. Why do some manufacturers spend a lot of money on ads that tout the superior cleaning ability of their laundry detergent? Any manufacturer can claim this but consumers figure out whether this is true. Since what manufacturers of laundry detergent want is to get customers to become loyal to their brand, if spending a lot on ads does not get consumers to keep coming back, it is pretty wasteful relative to those brands that do get consumers to become repeat customers. Thus, again, there is a natural equilibrium, high quality brands will do a lot of advertising and lower quality bands will not and, again, the punchline is that the ads themselves are credible signals of the true quality.
What does this have to do with beer? Well, if Chris of Belmont Station is right, high price is itself a signal of true quality. Since consumers figure out beer quality by trying it once, price itself may be a reliable signal of the quality of the beer. To put it another way, if consumers didn't think Rogue was exceptional beer, they could not get away with their high prices, so the fact that they can is a signal that they do indeed make exceptional beer.
So it is not necessarily just a matter of projecting high quality, but informing consumers about quality as well...