Policy and heterogeneity: A point

By Matt Nolan 12/02/2014


Via Geoff Simmons came this interesting post about new health policy in New Zealand.  This isn’t my area of expertise, but I found the post really insightful – I’d definitely recommend it as a read.

However, a small part of the post did spark my imagination, and will lead me to write on a loosely related, but important issue.

In the post, when talking about potential issues with the new scheme, the author says:

We hear echoes of the Bolger-led Government adoption of social capital in the late 1990s. Remember that? It placed social problems and their solution-generation with ‘communities’. There is a worthy role for this type of policy in a wider package, but it can also be used as a distancing policy to shield a government and the state from its responsibilities (e.g. on welfare benefits), deflecting the blame and responsibility for solutions to the level of communities.

This is a fair point, words like “community” and “opportunity” are often used by politicians on the right to avoid action.  However, politicians on the left are just as eager to push inappropriate policy at a national level by dismissing these claims.  In truth, the relative importance of “community”/”opportunity” as opposed to nationwide determination of policy depends on our assumption about how different people are – to whip out some jargon, we need an idea about how heterogeneous individuals are with regards to the issue we are looking at.

Take the posts discussion of a 20% sugary drink tax as an example.  Well, we can only really defend that policy is we have perceived a specific need for that policy – why are we taxing sugary drinks?  We could try out the following explanations:

  1. Sugary drinks hurt health, which hurts the person – and they don’t include that as they are time inconsistent.
  2. Sugary drinks hurt, we have public health funding, this is an indirect form of “user pays”.
  3. Sugary drinks are seen as a status good, and are overconsumed by people to achieve “relative status”.  This was sort of implied in the Spirit Level book.

Note:  The “fake food” argument doesn’t mean much unless we have a behavioural mechanism such as the ones above.

Cool, ok.  With time inconsistency we are asking why people don’t precommit (examples of commitment here and here).  With the second one we may view as a fat tax, this involves thinking about the relationship between consumption and health outcomes – a kicker is the tax is linear, while the “costs” are non-linear.  The final one involves stating that these drinks are seen as a type of status good, and we may have a view that increasing the price would reduce this specific coordination failure.

Even at an aggregate level these things are hard.  But food is such a good example as individuals ARE massively different with regards each of these three explanations.  Different groups are time inconsistent in different ways, different people experience negative health effects in different ways, and different communities respond to having something as a status good very differently.  Therefore the appropriate tax/policy on these different groups will be very different!

The status good argument is one that is especially easy to pick on – if you increase the price, this may in fact increase the use of sugary drinks as a status item.  If it is mainly poor households that use this type of food as a “status good”, then it is really a punitive regressive tax!  And even if we were to deliver an income transfer to do this, the fact that the higher price may have made it more of a status good simply implies that we have exacerbated the problem!!!

When discussing “community” the key idea is that we have specific failures that policy can address, and those failures are not equally distributed across society or across class.  National policy that focuses on helping communities and targeting need can in this way make more sense than national policy that is meant to address the problem itself with a single instrument – something like a “fat tax” could well be too blunt a tool to address the specific failures we are concerned about.

Sidenote:  I would more generally point out that, if we find sensible targeted community programs do not change behaviour, this doesn’t so much imply we need to go in harder – it more implies that the actual “market failure” involved may be smaller (or different) than we had previously believed.  We can’t observed the heterogeneity of individuals in terms of their wants and desires, we can only get an inkling through their choices – we must be careful to not, in turn, impose our own preferences on others :)


Site Meter