Matt Nolan scrutinises taxes targeted at social behaviours, separates 'excessive' and 'patently ridiculous' from where efficient correction can occur

By Matthew Nolan*

As a note, so far in this series on tax the following issues have been touched on:  Why do we tax?  What are the distortions of tax?  What are poll taxes, and how does fairness matter?  What are factor/income taxes?  What are consumption taxes?  Where does inequality fit in? Progressivity

---------------------------------------------------------------

After our recent series of articles on tax we are ready to chat about some other forms of taxation that crop up.

Today we’ll look at externality taxes.

These are also termed corrective taxes (as they are seen to be correcting something) and Pigovian taxes (after Arthur Pigou).

When discussing tax previously we talked a lot about the tax creating a “wedge” between the private and social value of an activity – and that being what we term inefficient.

However, what happens when this wedge exists in the first place?  In this case, we can use a targeted tax to correct this wedge – economists call this corrective tax an externality tax.

A common example of an externality that is often used is a factory that pollutes.

The factory will make a product, and so their decision to make the product will depend on the price they can get and the cost of making the product – as a result, the consumers’ willingness to pay is helping to drive this decision.

However, if the factory creates soot, and this soot lands on top of my house and messes up the place and my lungs, this does not influence the factory’s decision to produce – even though it has a negative impact upon me.  This is a negative externality.

We could improve outcomes by making the factory take into account the cost it is imposing on this third party (me with my house and lungs) – and we do this with an externality tax.

Let me give a random example.  I hate it when I’m on the bus and the person next to me starts eating. Buses are small, and I’m being forced into the other person’s meal without wanting to be there. Yes I’m a bit oversensitive, but I can’t help how I am.

Now, there is an externality from the other person’s decision to eat next to me here – it is having a negative impact on someone not involved in the market transaction to buy and consume the food. 

This will show up in the relative price of other market transactions, eg it will change my willingness to pay to go on the bus instead of driving (I’m less willing to pay to go on the bus, knowing there is a chance someone will chow down next to me). As a result, this is an example where the burden of where the “externality” falls is unclear – just as we described last time with taxes.

The appropriate externality tax in this case will capture the cost of the externality on the effected party – in this case me.  How does this work?

The Coase Theorem

Pointing out these ways that the price “can be wrong” is pretty popular with government – it gives them a way to jump in and make things better by taxing them.  However, we have to be careful before running around trying to fix everything.

The problem that exists here is the lack of property rights and a market – the person sitting next to me eating is able to enter my space and impose a cost on to me they don’t take into account.

However, in of itself this suggests a solution, why don’t I just turn around and offer the culprit money to hold off eating until I’m off the bus?  If I would be willing to pay $5 for him not to eat, and he is willing to accept it, then we can trade – and we are both better off!  Furthermore, if I would be willing to at most pay $5 but the person eating was not willing to trade (they valued it at more than $5) then having the eating occur is the efficient outcome – as the other person values eating more than it upsets me.

Remember, we are thinking about the allocation of goods and services when we discuss this – if there is no change in the allocation of goods and services there is much less of an efficiency argument for the tax in the first place.

Now this helps to get the efficient solution (as we had described in the tax article) but there is of course a distributional one – by changing who has ultimate control over the eating (the property right) we are shifting the endowment of wealth.  But this brings us back to standard tax and transfer ideas.

As a result, the key point is that if we could costlessly establish property rights and markets the efficiency issue is able to solve itself.

So what prevents this from occurring?

The answer is a potential inability to enforce property rights, and the transaction cost of establishing the market.

If a market could be easily set up, a tax is likely to be unnecessary – if the market would be difficult and/or costly to set up, there is more scope for government involvement.

What to count

We also need to be a bit careful when we decide something is an “externality” that we, as a society, should “internalise” by taxing. Some people don’t like the look of other people who are a different race, or who do different things to them. In the way we’ve discussed an externality, it sounds as if we should tax people’s existence and choices just because other people inherently dislike them or their habits.

This is not ok.

And it begs the question, who chooses whose preferences are relevant for policy and which are not relevant. This involves making government (and thereby large social groupings) a moral arbiter over the preferences of individuals, which is an area we need to be honest and transparent about rather than hiding the trade-off in the terms “externality” and “efficiency”.

Double counting and caution with social cost studies

While externality taxes are, in principle, an area many economists agree policy action is good – see self-avowed right leaning economist Greg Mankiw and his Pigou Club – it is possible to take these things too far.

In the last decade there has been a proliferation of social cost studies, to try and justify the use of corrective taxation. This is a good thing, as it involves making sure we measure and are aware of social costs involved when setting up policy.

However, these studies can often be misleading in the way they conceptualise the issue of externalities – leading to demands for excessive taxation, and suggestions of patently ridiculous “losses” if something isn’t done.  See this post by University of Canterbury lecturer Eric Crampton on some of these excessive claims.

There are three areas where we need to be especially careful with these ideas.

First we need to be careful not to confuse the right “externality tax” with the “gross social cost” associated with the externality. Remember when we discussed me sitting grumpy on the bus earlier, and I noted that it would change my willingness to pay to use the bus, or it may change my effort or remuneration at work.  The corrective tax merely gets the bus eater to take into account the cost they impose on me.  A study looking into the gross social costs associated with what is going on is an accounting measure articulating the total set of costs associated with the transaction.

If we were to count gross social costs, we would need to include the cost of the eater’s breakfast – but this is irrelevant to our externality, as this person has already taken into account this cost and was willing to buy and eat his breakfast because of a countervailing benefit.

The externality tax shouldn’t be set in relation to these other effects on priced markets – it should be set to close the implicit “price wedge” that exists where the externality occurs. If people eating on the bus is an externality that we decide requires a corrective tax, we want to set the tax so that it makes people behave as if there is a market in place between the people involved.

Setting an externality tax to capture the costs that are accrued when an individual makes a choice to do something (costs that fall into priced markets) involves the worst kind of double counting – as the associated costs and benefits are already priced in, so we are pretending things have no value when they do.

A second important point is that an externality tax pulls up the price faced by everyone by a certain amount – but the actual externality we care about may differ strongly depending on how much people consume or produce. For example, any externality associated with excessive drinking occurs when drinking becomes excessive – but when we introduce a tax we are also hitting people who have a quiet drink with dinner. In this way, the externality tax cannot be a perfect solution.

And finally, we need to be careful as many externalities may be “priced in” without us realising it over a broad economy.  In the case of jumping on the bus, the bus company realises that some people value eating and some people dislike it.

As a result, they can choose to ban (and enforce) or allow eating on the bus based on the inherent willingness to pay of the individuals that ride the bus.

In the end, the bus company is internalising the relative costs and benefits and allocating bus seats in an efficient manner – and my perceived externality is just me whining about something I don’t like.

We may argue it is unfair that I have to deal with that, or state that we have some social preference for no eating and drinking on buses. But here we are arguing about a direct issue of fairness we should be honest about, not an issue of efficiency. Pretending an “externality” exists in the strict economic sense would be misleading regarding the trade-offs involved.

Careful with correction

There is a time and a place for corrective taxation, and I have written favourably on them in the past.

They fit neatly into the framework for allocative efficiency that we have described, when talking about the way tax influences the allocation of goods and services.

However, they are not a panacea for perceived social ills, and they need to be looked at carefully before such a tax is imposed.

------------------------------------------------------------------------------------------------

Note:  Eric Crampton from the University of Canterbury has a neat post running through these issues in more detail, with a bit more technical description, over on this site.

------------------------------------------------------------------------------------------------

Matt Nolan is a senior economist at Infometrics. You can contact him here »