Eric wonders if there is a serious treatment of intrafamily effects and whether these should be accounted for as externalities. I don’t know of any literature on this (not that I have looked), but a couple of points do occur to me.
First, rather than get caught up in a semantic debate about what is the exact definition of an externality, we need to focus on what is the implication of declaring an intrafamily effect to be an externality when conducting cost-benefit analysis. Typically, we are looking to set the optimal level of a Pigouvian tax. Eric suggests that a husband and a wife are in a contractual relationship, and so, by our normal definitions of an externality, all external effects are contractually internalised. To a non-economist that sounds like the sort of callous, evil thinking we expect of economists, who think that the complexity of human relationships can be reduced down to maximising behaviour and market relationships.
But now consider the alternative view that the costs imposed on someone by his or her alcohol-affected spouse is an externality. By this view, it would be appropriate, if possible, to levy a Pigouvian tax on the drinker equal at the margin to the cost imposed on the spouse. It wouldn’t matter that the revenue goes to the government rather than the spouse, as long as the revenue, in principle, would be sufficient to compensate the spouse for the harm caused. Furthermore, it wouldn’t even matter if the tax had no effect on the behaviour of the drinker. By the logic of externalities and Pigouvian taxes, inelastic demand would simply reveal that the benefits to the drinker were greater than the costs to the spouse, and so it would economically efficient to not abate the behaviour. To me that sounds like the sort of callous, evil thinking we expect of economists, who think that the complexity of human relationships can be reduced down to maximising behaviour and market relationships.
My point here is that we shouldn’t even go down the route of using the language of externalities and cost-benefit analysis to think about intra-family issues. If we want to slot them into the welfare economist’s toolkit, it is much more in the domain of equity issues. This applies equally to costs imposed on children.
This relates to my post last year on accounting for irrationality in cost-benefit analysis. The point there was a world-view that says people are irrational does not automatically lead to a conclusion that self-imposed costs should be accounted for within the cost-benefit framework. Rather, we should estimate appropriate Pigouvian taxes based on narrowly defined external costs, and then ask, separate from the cost-benefit analysis, whether, for paternalistic reasons or for concern for family members, it would be appropriate to make further interventions at the margin, and whether taxes would be the appropriate instrument. Unlike the case of intervening against genuine externalities for efficiency reasons, the elasticity of demand is crucial in determining the appropriateness of tax interventions to deal with these problems.
This relates to my second point. I have enough Irish ancestry to have historical family anecdotes of alcohol consumption by men imposing costs on their wives and children. But these are not anecdotes of Jake the Muss types, but simply of non-violent men who drank away too much of the family income. These costs were real (and were the reason, I believe, that the pre-1991 family benefit was paid directly to mothers). That doesn’t mean, however, that it would make sense to roll them into a measure of social costs and use that to justify higher taxes that would, if demand is inelastic, result in even more of the family income being drunk away.