The politicization of a newsvendor problem

So last year everyone was nervous about a flu pandemic — and with good cause. The H1N1 virus was an unknown protagonist and it was not out the realm of possibility that it could lead to a devastating public health crisis. We now know it didn’t. Yes, some people got sick but the forecasts of overwhelmed hospitals and massive numbers of deaths never really materialized.  Hey, we even ended up with some excess vaccines.

The last point has now led to some finger pointing as people go back and examine how various governments handled the crisis. An independent commission in Britain has just released a report evaluating the UK’s response and generally has good things to say except for all those left over vaccines (Swine flu vaccine contracts ‘lacked get out clauses’, Jul 1, BBC). Dame Deifre Hine, the commissions chair, discusses the findings:

Here are the details on the excess vaccines:

More than 30m doses are thought to be left over after one of the manufacturers, GlaxoSmithKline (GSK), refused requests for the contract to be torn up.

The other manufacturer, Baxter, agreed to a “break clause” allowing the government to cancel its order. …

A spokeswoman for GSK said because of the demand at the time for its vaccine from governments across Europe it would not have been “ethical” to offer one country a break clause when others were not able to have all the doses they would have wanted.

So what would a “good” supply chain contract look like in this setting? There are really (at least) three sources of uncertainty in this problem.  One is the severity of the virus. Another is how much effort the health service (national or independent) will have to put in to get a significant number of people vaccinated. A third factor is how able the manufacturer is in getting the vaccine out. The first two clearly interact. If everyone is worried about a horrible flu, the health service can get people lined up out the door just by announcing a flu clinic. The third depends on a number of factors since there is a lot of uncertainty in vaccine production. One also needs to consider how full the vaccine maker’s order book is.

A good contract has to balance these various risks without creating bad incentives. I can’t imagine that a vaccine makers would agree to a blanket break clause without charging an exorbitant price. Such a clause undermines the health service providers incentive to vaccinate lots of people — especially if it becomes clear that the flu is not too severe. This is particularly true if the vaccine is sourced from multiple firms. If the health service is holding to piles of vaccines that have differing break clauses, then it should first use up the doses that have less generous terms. That means any manufacturer naive enough to provide generous terms is going to bear an undue amount of risk.

At the same time, the lack of a break or return clause shifts risk off Glaxo when they are late delivering the goods. The Brits would have far fewer doses if the vaccine had been widely available early on when the virus’ severity was uncertain.

That seems to suggest a sensible contract. Why can’t a contract include target delivery dates with shifting terms depending on whether those dates are met or not? If the manufacturer gets the vaccine to the health service in a timely fashion (i.e., when the flu’s severity is still uncertain), it is the health service problem to make sure that the population gets vaccinated.  If the vaccine comes late, it becomes more and more the manufacturer’s problem to deal with unused supplies.  That seems like a fairly sensible way to balance the needs of all the parties.