Friday, October 8, 2010

Well-meaning policymakers and unintended consequences

I'm giving a course on economic inequality these days and as part of it I listened recently to an hour-long lecture on YouTube by Richard Wilkinson and Kate Pickett, the authors of The Spirit Level: Why Equality is Better for Everyone ("Inequality: The Enemy Between Us?" in 8 parts, highly recommended). At one point, Richard Wilkinson mentions the growth of inequality in the UK during the Thatcher years that resulted, in part, from changes in taxes and transfers and a reduction in trade union power. He notes that, according to the evidence put forward in The Spirit Level, this had long-term consequences that those responsible for those policies did not intend or foresee: a growth in social dysfunction, as reflected in growing problems with such things as obesity, violence, and teenage pregnancies.

   The example reminded me of how the textbooks use examples of such unintended consequences in a different way -- to make a rhetorical point of their own in favour of 'free market' policies. Is something like the following familiar? The apparently short-sighted or economically ignorant, policymaker who thought that rent controls were a good idea, but did not realize that it would make it harder to find an apartment as excess demand developed. Or the well-intentioned raising of minimum wages by politicians, not realizing that an excess supply of labour would develop -- at least according to the perfectly competitive model. Or the claim that requiring people to buckle up and use their seatbelts will have a more-than offsetting effect of inducing them to drive more carelessly, leading to an increase in accidents and injury and death. (OK, no text I know of is crazy enough to try that one out, but I could not make it up; it's in the professional literature. See Sam Peltzman, “The Effects of Automobile Safety Regulation,” Journal of Political Economy 83 (1975), 667–725. For evidence that it's not true, see studies such as Anindya Sen, “An Empirical Test of the Offset Hypothesis,” Journal of Law and Economics 44:2 (2001), 481–510, or Alma Cohen and Liran Einav, "The effects of mandatory seatbelt laws on driving behavior and traffic fatalities", The Review of Economics and Statistics, November 2003, 85(4): 828–843.)
   These are cases of what the great economist Albert Hirschman, in his 1991 book The Rhetoric of Reaction: Perversity, Futility, Jeopardy, called 'The Perversity Thesis'. One form that conservative or reactionary arguments take is to say, in effect, 'Yes, policy X is undoubtedly well-intentioned, but don't you realize that it will have perverse effects leading, in fact, to undesirable outcomes?'
   Hirschman comments (pp.39-40):
For example, those who emphasize the perverse incentives  contained in unemployment benefits or welfare payments never mention that large areas of social assistance are fairly impervious to the 'supply response' that is at the bottom of whatever perverse effect may be at work: people are unlikely to gouge out their eyes in order to qualify for the corresponding social security or tax benefits.When industrial accident insurance was first introduced into the major industrial countries of Europe toward the end of the nineteenth century, there were many claims on the part of employers and various 'experts', that workers were mutilating themselves on purpose, but in due course these reports were found to be highly exaggerated. 
    Do some textbooks really offer examples of a 'perversity thesis'? You can see if you can find some examples yourself, but here's a quote from Baumol and Blinder's Microeconomics: Principles and Policy, 10th Ed. (2006) which I found with no effort:
Any government that sets out to repair what it sees as a defect in the market mechanism runs the risk of causing even more serious damage elsewhere. As a prominent economist once quipped, societies that are too willing to interfere with the operation of free markets soon find that the invisible hand is nowhere to be seen.


  1. Interesting. One place I have come across similar examples recently is in the push for greater energy efficiency in homes and in businesses in preparation for a looming energy shortage in the medium to long term, and as a quick way for helping the economy grow in the short term.

    However, many studies have shown that efficiency gains with regards to energy use are prone to the 'rebound effect' which means the lower energy useage is only a small term win. Instead, people invest their savings in another tv, another trip abroad, another car or another energy guzzling piece of equipment for their business. Over the long term, demand continues to rise and the energy problem remains unsolved.

    I think the problem of the rebound effect stands to reason and presents huge problems for policymakers. However, they seem, at present, to be happy with the short term wins, and leaving the longer term issue of increased demand to next round of politicians to deal with. This kind of short termism is a key reason why policies aiming at some social problem fail to grasb and deal with underlying fundamentals which provide huge problems when left unquestioned.

    Wilkinson's work on social inequality as a key underlying diver of social dysfunction highlights this point.

  2. Ronan's comment is a good one. I'd first read a discussion of the 'rebound effect' regarding energy use in George Monbiot's book "Heat" (from p.61). (He refers to a paper by J. Daniel Khazzoom, "Economic implications of mandated efficiency standards for household appliances", Energy Journal, 1980, Vol. 1, 21-39.)

    As energy efficiency improves, there's an income effect, so more real income is available to spend on all things. There is also a substitution effect as the relative price of using energy-using things falls, so more is demanded. The net effect could be an increase in energy use. Hence, I guess the need to set overall caps on things like greenhouse gas emissions, rather than just focusing attention on increased energy efficiency.

    Ultimately, it's an empirical question. Hirschman's comments about 'the perversity thesis' that I quoted in the original post also suggest the same thing. So it's true, for example, that unemployment insurance does increase time spent unemployed, a kind of perverse effect that labour economists have measured. But one then has to consider how big that is and whether there are other offsetting beneficial effects of UI: eg. a better match with the next job due to the lessened need to accept the next job that comes along (an efficiency issue).