Where is the externality here?

A question rightly asked by Matt Nolan over at TVHE. Matt writes

There have been wild debates surrounding the BERL report into the social costs of alcohol. I haven’t read the report, I haven’t read the replies, I have to admit I have been busy.

However, in one of Eric Crampton’s many posts on the issue I see that generally an externality from lost output, excessive unemployment, and forgone wages has been assumed in the discussions. I’m sorry but what?

And he argues

The labour market is a market, how can we have an externality when there is a market with a market price (wages). Yes, alcoholics produce less, less of them are employed, and they tend to have lower wages – but this isn’t an externality it is part of the market process. They are paid less because their marginal product is lower, and they are willing to be paid less because the benefit they receive from consuming alcohol is sufficient compensation – this is a completely internalised decision for the drinker isn’t it, so where is the social cost.

And don’t say it is too the firm – the firm can set a lower wage because of the fact that the marginal product of this worker type is lower.

All of which makes good sense.

Matt goes on to say that market failure could occur due to adverse selection. A firm hires someone without knowing their drinking habits and they could therefore end up with a low productivity/high drinking worker. But the firm can fire the worker, in extreme cases, or adjust future wages to take into account the lower productivity. In either case the bad effects of the drinking appear to have been internalised.

And thus no externality.