Gordon Tullock had two standard lines for the (many) anarchists in our grad programme. One was that government was needed to mandate interconnection between private roads. The other was that we need government to enforce quarantines during pandemics.
And so it's great that Otago's Nick Wilson has been doing some work on pandemic preparedness. The Press reports on his work on the topic; here's his blog post. In the 1918 influenza pandemic, some communities remained clear with strict internal entry controls. He also recommends closing the borders and emptying communal living facilities in advance of a pandemic hitting New Zealand, depending on the pandemic's seriousness.
If "closing the borders" does not imply shutting down international trade, it's pretty plausible that this could pass cost-benefit for a sufficiently serious pandemic. If the flu (or other) virus can only survive outside the body for n days, then continue to allow ships to offload shipping containers, put them into holding areas, then release them post-quarantine. It'll be a big costly hassle, but potentially better than losing a reasonable fraction of the population. If it means shutting the border to trade entirely, well, the "how serious a pandemic" hurdle gets much higher. Were some version of ebola to come out with a week-long incubation period, followed by a few days' high infectiousness, followed by death, then temporary autarky wouldn't seem so bad by comparison.
It would be a fun (but likely infeasible) honours project to work out some ballpark numbers on expected mortality rates sufficient to justify different pandemic quarantine levels, ranging from nurses / health checks for incoming flights at the airport, to quarantine facilities for incoming visitors, to the complete sealing of the borders. I expect that working out the likely real costs of some of these measures would be a bit beyond the scope of an honours project. But it also seems the kind of thing where the government should have set plans: if the combination of communicability and morbidity hits different thresholds, then different measures get rolled out.