The overestimation of Pigovian taxes

Over at Anti-dismal Paul Walker points at a paper on Pigovian taxes by John VC Nye from George Mason university. In this paper Dr Nye makes the following point:

A claim about the optimal Pigou tax is a joint claim about the size of the externality and about the optimality of observed outcomes, not just the externality. Measuring the size of the observed Pigovian externality – even if done perfectly — is not a reliable guide to the proper level of the Pigovian tax because in a world of efficient transfers we will still observe some externalities

Now this is very true – if we sit in the world of partial equilibrium analysis (which a lot of Pigovian analysis is based on) for too long we can forgot the fact that markets do not act independently, and this is bound to colour our view on the correct level of the Pigovian tax.  At some level this is a critque of Dr Mankiw’s Pigou Club.

The four factors that the author believes cause current Pigovian tax estimates to be too high can really be boiled down to two:

  1. Coase bargaining,
  2. Government regulatory and tax policy outside of the Pigovian tax.

Now I think that he is completely right that Coase bargaining needs to be taken into account, and government solutions that do not take into account the potential for Coase bargaining will be too heavy handed.

Furthermore, I agree that regulatory and tax systems in place will impact on the level of the appropriate Pigovian tax – however I do not agree that in this case it necessarily implies that the Pigovian tax must be lower (I do however agree that it is an issue that the government should take into account – something the NZ government has reacted to by doing GE modeling of the ETS, which is similar to a Pigovian tax).

Ultimately, Dr Nye believes that if there are goods and services and income taxes in the economy, these will shift the outcome in the market with externalities towards the optimal level, and so we do not need to add a Pigovian tax on top.  This works if we set the Pigovian tax at the level that would have been optimal without GST the current tax would be excessive in the given industry!

My issue with this analysis is that he is going back to partial equilibrium analysis to think about the optimal amount of one good – rather than thinking about the more general issue of the allocation of resources.

The reason we have goods and services and income taxes is because we require some amount of revenue to preform government services.  We use a GST or income tax so that the burden is spread as evenly as possible across the economy – limiting the impact on the optimal allocation of resources.  The ultimate purpose of this is to try to minimise dead-weight loss from taxation.

If this is the purpose of a GST or income tax why do we want to effectively remove the impact of these taxes on goods that produce externalities?  If the burden of these taxes is going to be spread across all goods in the economy then we shouldn’t pay them any heed when working out the Pigovian tax as otherwise we will be forcing other goods to cover up the revenue hole created by the externality good – which will be more distortionary.

This may sound weird, given that the purpose of the Pigovian tax is to correct distortions.  However, in a general equilibrium sense the goal of the tax scheme is to ensure that we raise revenue with a minimum distortion to the economy – which implies setting the marginal (per-tax dollar) dead-weight loss equal among all industries.

Now on complication is that, a flat tax on everything doesn’t actually achieve the lowest level of dead-weight loss – as the dead-weight loss from a tax depends on the elasticities of supply and demand (the more inelastic, the smaller the dwl for a given level of revenue).  However, the administration costs associated with making the tax on every good appropriate would be huge, so we satisfy ourselves by spreading the revenue gathering burden evenly.

However, in the petrol industry we do have a case where both supply and demand is relatively inelastic – and as a result the revenue generation in this industry is relatively more “efficient” than from other industries (allowing us to reduce taxes in more “distortionary” areas).  So if we want to make a case for giving the petrol tax special status there may be a case for setting it even higher than a Pigovian tax would suggest, on the basis of revenue generation concerns.