In textbook treatments of price theory and of public economics, one invariably sees national defence cited as an example of a public good. But how useful national defence is surely depends upon the expected change in policy if a foreign power took over. If this change is for the better, then clearly national defence is only an impediment to amelioration; if the change is for the worse, the utility of national defence would depend upon how much worse and compare that magnitude to the costs. If the change is neutral, national defence is again a public bad.
One could argue that citizens want a national defence and so their preference for it makes it a public good. But why should citizens qua citizens know what is good for them? Facing virtually no chance of their opinion being pivotal to any national decision, they can indulge in whatever idiocy they want. I do not normally like the distinction behavioural economists make between decision utility and experienced utility, but it seems serviceable in this case. If citizens want national defence, that is just their decision utility talking, but in actual fact it could be that they are throwing money away. Maybe it is a purely normative issue, but it seems to me to be just as normative to refer to something with such ambiguous effects as a public good.
If one goes along and takes the right name to depend upon whether national defence safeguards against changes for the worse or just keep things bad, it would seem to me that national defence is to be referred to as a public bad in the countries which pursue the most horrendous policies (certainly it would seem a normative point to give a name like "public good" to racist polcies such as the nuking of areas containing people of a certain skin colour, wanted by racists). I have made the case before that policy may not actually improve from outsiders invading such countries, but of course if policy is about the same the war was wasteful and so national defence was a bad idea.
With no spending on the military in the countries pursuing the worst policies, the countries pursuing the second-worst policies face no threat of invasion from countries with worse policies anymore. Again, maybe invasions by countries with better policies would not actually result in improvements, but if things do not change appreciably it would seem that these countries also waste resources now by spending them on national defence.
In this way, we regress to a situation in which no country spends anything on national defence. This is an equilibrium because no country now has any incentive to spend money on protection, though the people in the countries pursuing the most horrendous policy arguably still have an incentive to change things somehow - just not by getting invaded. The operative assumption here is that countries which are invaded suffer only if policy is changed for the worse (otherwise why resist?) and that the risk that this happens goes away if the invading country is one pursuing better policy.
In equilibrium, national defence is a public bad. Outwith equilibrium there may be a point to it in countries with good policies at risk of invasion by countries with bad policies (though there are problems with this view, too), but still many countries today do not face any serious risk of that happening (e.g. Switzerland, New Zealand, Canada). If the language of economics textbooks took steps to approach value neutrality by calling national defence, say, a "public-bad-or-good(-but-probably-bad)", my guess is there would be no less national defence anywhere. It would be better language, though, and reflect better thinking. That's something.