Many good comments; I have time to respond only to a few of them.
One comment quite rightly points out that the estimate of 50 million dead in the Spanish flu pandemic cannot be right if the mortality rate was 1 percent--50 million is 1 percent of 5 billion, which was more than twice the world's population in 1918 and of course not everyone in the world was infected. The studies I've seen give a lower bound estimate of the number who died as 20 million, which if correct surely implies a mortality rate greater than 1 percent. It is estimated that about a quarter of the U.S. population, or 25 million, were infected; and 500,000 died--that is 2 percent.
Several comments emphasize that the problem of vaccine production from the manufacturers' perspective is that property rights are likely to be threatened in a pandemic--hence the threat to Roche of compulsory licensing. That risk cannot and probably should not be eliminated, and so the answer may be to compensate manufacturers for the risk by subsidizing vaccine R&D. Beyond that, as another comment remarks, when a vaccine has to be created in a hurry because a new virus has appeared, the risk of side effects cannot be calculated and incorporated into the sale price; this is an argument for the government's sharing the risk through some scheme for indemnifying some of the liability costs of the manufacturers of the new vaccine.
A number of comments discuss the distinction I made in the Orphan Drugs posting of the previous week between economics and utilitarianism. (In fact the Orphan Drug and avian flu topics are closely related; both are about potential problems in the production of drugs, and they are quite similar problems--in the Orphan Drug context stemming from the small size of the market relative to the fixed costs of a new drug, and in the avian flu context to the small size of the market as a result of the fact that a flu vaccine may be good for only a single flu season, or indeed may have no market at all if the flu that it is designed to prevent doesn't materialize to the point where people will bother to be vaccinated.) Economics is a social science and utilitarianism is a philosophy, but they do share the word "utility," used however in different senses. Economists usually distinguish between "value" and "utility" merely to bring attitude toward risk into the analysis: the value of a 10 percent chance of $100 is $10, but the utility may be less for a risk averse person (hence the market for insurance) and more for a risk preferrer (hence the appeal of gambling). Economic value with or without attitude toward risk is different from utility in the utilitarian's sense, as the former is strongly influenced by wherewithal. If I have only $1 and a life-saving drug would cost $2, then the value of the drug to me is only $1, though the utility in the sense of the happiness I would derive from the drug would be vastly greater. So there would be a utilitarian argument for the government's subsidizing my purchase of the drug, but not necessarily an economic argument. I emphasize "not necessarily," however, because there may be an economic argument for the subsidy, based on insurance or altruism notions, but it would be more complicated than the utilitarian argument.