Tax policy colloquium week 7: Zach Liscow paper, part 1

Yesterday at the colloquium, Zach Liscow presented his paper “Democratic Law and Economics.” Before getting directly to the paper, I think it’s useful to put it in a broader context that will be familiar to some readers but perhaps not others.

The law and economics movement in law schools, which got going in earnest about 40 years ago, in its early stages made a lot of strides using an approach that some call “Econ 101ism.”  As per Noah Smith:  

“We all know basically what 101ism says. Markets are efficient. Firms are competitive. Partial-equilibrium supply and demand describes most things. Demand curves slope down and supply curves slope up. Only one curve shifts at a time. No curve is particularly inelastic or elastic; all are somewhere in the middle (straight lines with slopes of 1 and -1 on a blackboard). Etc.”

Econ 101ism was a step forward for legal scholarship, but not a step far enough. One had to understand what the logic of a standard neoclassical analysis implied, but the next step required recognizing that its assumptions might not always hold, and then asking what that would imply. It’s a great orienting tool, but its seductive power to purport to answer so many questions, from so parsimonious a starting point, can over-excite the incautious if they forget that they need to take that next step of testing the accuracy and sufficiency of its assumptions.

A more recent trend in what one might call tax or public finance law and economics involves using a very simple model that purports to answer lots of questions. The Atkinson-Stiglitz theorem holds (within the specified terms of its model): “Where the utility function is separable between labor and all commodities, no indirect taxes need be employed.”

Take the Mirrlees model, in which we want to base the tax on ability but can only measure earnings, so we impose a labor income tax that is equivalent to a uniform commodity tax. (Wages are used to buy commodities, and in a one-period model one spends it all currently.) Atkinson-Stiglitz asks whether one might get a better outcome by having non-uniform commodity taxes, e.g., taxing luxuries at a higher rate than the rest. It shows that, within its terms, the answer to this question is NO unless some commodities are leisure substitutes or complements. E.g., if I work more, I may substitute restaurant meals for buying groceries (leisure substitute). Or if I work less I might want to buy a telescope so I can spend hours stargazing (leisure complement). But absent any of that, the uniform commodity tax is best.

Suppose one thought that taxing yachts at a higher rate would have an admitted efficiency cost (discouraging yacht choice vs. other consumption), but also an efficiency gain (raising revenue that permits one to lower the labor income tax rate). So isn’t there a tradeoff, with even an implication that there might be a net efficiency gain, since multiple smaller distortions are often better than just one large one?

Answer: No. You still have to work to get the $$ to buy a yacht, and working is just a way of getting to buy things. So one is still, roughly speaking, discouraging labor supply as much as before, plus one is now also distorting commodity choice. Hence there is now “double distortion” without mitigation of the prior distortion.

Leading figures in tax or public finance law and economics (e.g., Kaplow and Shavell) have shown how broad Atkinson-Stiglitz’s implications might be. For example, one can think of an income tax as imposing higher commodity taxes on future consumption than current consumption, creating double distortion (discouragement of saving) without any mitigation of a consumption tax’s admitted discouragement of work and market consumption. Hence, case closed for the income tax? Yes unless there is more to the analysis, but the point (as with Econ 101ism) is that there might be. (And for what it’s worth, neither Atkinson nor Stiglitz agrees that based on their theorem one should prefer consumption taxation to income taxation.)

Likewise, case closed (within the analysis) with regard to using “legal rules” instead of the “tax system” to redistribute. Here the “tax system” means, not Title 26 of the Internal Revenue Code, but a labor income tax plus demogrant. “Legal rules” means pretty much anything else. E.g., income-conditioned speeding tickets, like they have in Finland. Product liability or tort rules that favor consumers or accident victims on the ground that they’re generally poorer. Mandatory contract clauses to protect tenants. Inducing corporate managers to optimize for “stakeholders,” not just shareholders, under a progressive redistributive rationale.

Under the view, using any of those types of vehicles to address distributional concerns about rich vs. poor merely yields double distortion. You’re still discouraging labor supply, insofar as becoming rich subjects you to expected unfavorable treatment under those rules. Plus, you’re departing from the efficient choice in those areas, which makes things worse overall than if one had just used the labor income tax to address distributional concerns.

The takeaway from this analysis was and is: All analysis of “legal rules” and government policies outside the labor income tax should be driven PURELY by considerations of efficiency. Leave redistribution purely to the “tax system.”

This view has been highly influential in legal scholarship, and also perhaps in regulatory policy as actually done. But it has inspired pushback, both on theoretical grounds (based on whether the underlying assumptions are sufficiently accurate and complete) and in light of concerns that, in practice, it leads to too little progressivity if policymakers adopt it in the realm of “legal rules,” but not when designing the tax system.

This, anyway, is vital background for discussing Liscow’s “Democratic Law and Economics,” to which I will turn directly in my next blogpost. 

-->