Mike Konczal has a new white paper out theorizing on when public provision of some good is superior to handing out coupons that people can use to buy the good privately. What follows are two fairly unstructured thoughts about it.
First, technically it is the case that coupons and public provision are very different from cash transfers. The former two involve government-directed consumption while the latter simply modifies the overall distribution without directing consumption. So Konczal makes a wise choice in not including cash transfers into the mix: they are seriously just different kinds of things.
In practice though, many government coupon and public provision programs exist (usually those focused on the poor) because people do not want poor people to have the cash. The programs are intended to modify the overall distribution, but with a paternalist bent (usually based on bankrupt theories about how “it’s the taxpayer’s money“). For instance, why do we have means-tested food stamps instead of cash transfers? Why do we have means-tested public housing projects and section 8 vouchers instead of cash transfers?
My point here being that much of what passes as government-directed consumption is really just paternalistic redistribution, and that class of programs is very different from the others. I think the way you analyze programs in each group should be different. In almost all cases if what you are really after is redistribution, then you should just do it with cash (with perhaps exceptions for those with grossly diminished capacity).
Second, on the public provision side of things, there are at least two different kinds of public provision, and they deserve being distinguished. There is free public provision (or public provision with heavy subsidies) like libraries. Then there is public provision that relies entirely upon user fees (aka paying customers), and generally competes against other market players (what I would call a public option, although Konczal seems to use the term to refer to both types).
These two types of public provision require completely different sets of arguments to support. In fact, the user fee public option is arguably not even public provision at all: sure the government owns the business instead of another actor, but if they aren’t subsidizing it, there is no directed consumption. Generally speaking, it is going to be hard to justify the first type of public provision unless you intend the service or good being provided to be compulsory and universal (primary education and healthcare are good examples). For my money, the user-fee public option is basically always a good idea (Konczal gives a whole list of why) so long as it actually does sustain itself with its own collected fees. Anyhow, my point is that these two different types of public provision are so different from one another that separating them out may be more appropriate.