Friday 24 July 2009

Voter preferences and electricity generation

Seamus posted yesterday a political second-best argument in favour of maintaining public ownership of Meridian: while a first best would be private ownership, that's not a politically stable equilibrium because the median voter would demand regulations on a private owner even more inefficient than the initial public ownership.

I wondered in a comment there, and in a comment over at Anti-Dismal, whether the political second-best argument might well not justify a rather wide range of inefficient status quo positions. Paul takes me to task, but I think he's conflating my extension of Seamus's argument with my actual worry about potential use of the political second best argument against change.

I'm certainly not against political second best arguments in particular application. I tend to think, for example, that the New Zealand government has sufficiently proven its utter inability to avoid expropriating Telecom's investments in broadband infrastructure, through mandated sharing with competitors, that there's little chance of substantial private investment in fiber absent government making that investment directly.

Does the political second best argument hold in the power example? First, what do voters think about government ownership or regulation of electricity generation? Well, we have some data on this from the NZES. 40% of respondents want to have full government ownership of generation and distribution; 18% want "mixed" ownership (the current situation); 27% want a private but regulated industry, and 4.5% want a private industry (rest didn't know). So, it looks like voters are strongly against having a private, unregulated industry. That suggests that a private, unregulated industry is not a political equilibrium. Even the current "mixed" ownership situation isn't particularly stable.

Now, some interesting covariates. Probability of supporting full government ownership is significantly affected by:
  • Economic thinking (-): a standard deviation increase in economic thinking reduces support for full government ownership by 8%
  • Political ignorance (-): a standard deviation increase in political ignorance reduces support by 4%
  • Listening to National Radio (+): 3.3%
  • Being male (+): increases support 12%
Other covariates didn't matter much. Probability of supporting complete privatization is increasing in economic thinking, political ignorance, being female, not being a superannuitant, and being of asian ethnicity. By and large, the magnitudes of the coefficients is tiny (less than two percent increased probability) except in the case of asian ethnicity which increases the predicted probability by 6%. Economic thinking most strongly predicts support for a private but regulated system.

I'd expect a private but strongly regulated system to be an equilibrium; I can't see a private, unregulated system being a political equilibrium. If I could push a button that would privatize the electricity system while insulating it against stupid regulations, I'd push that button. Unfortunately, such buttons don't exist.

Now, is this a generalized argument in favour of any political status quo? I don't think so. I think there's reasonable evidence consistent with that privatization would result in a highly regulated system, and I can buy that such a system could be worse than what we now have. Paul says this may make me a Marxist, but I'd say rather Coasean.

I do worry that in some cases, building popular support for economically efficient policies requires government to be the first-mover in adopting currently unpopular moves and then building support for the shift by highlighting the successes of the reforms. Economists haven't a theory of the path to equilibrium, and much less a theory of the path from inefficient but politically stable equilibria to efficient stable equilibria that avoid worse-case outcomes. Hard problems.

In any case, if it takes substantial political capital to make any of these kinds of moves, why not start with ones that are more likely to lead to stable and desirable equilibria?


  1. I find the 'political ignorance' and 'sex' covariates counter-intuitive...

  2. I completely agree on the latter. But note that I'm also correcting for economic thinking, which is strongly predicted by gender. So what it's saying then is the portion of preference not already covered in economic thinking relates to a reduced preference for privatization.

    On the former, a lot of my findings on political ignorance are consistent with a generalized redneckism that is more skeptical of government ownership than of government spending.

  3. The first point makes sense.

    The second point still intrigues me...people that 'know more' about politics have relatively more faith in the government?! I don't think I'll ever understand that one! (although I guess public servants would feature strongly here).

  4. EC. I guess that if you took a survey of people in 1983 and ask them if they supported the types of reforms we got post-84 most would have said no. Most economists at the time didn't support the reforms. But those reforms have survived, by and large, since then. So what may have been a politically unstable equilibrium then is now stable. Let us see the government move and see if we can build support ex post.

  5. I'd say rather that folks with less political knowledge may say that some risks are worse than they really are.

    Caplan finds, for example, that the general public is a lot more worried about the risks to the US economy of too much spending on welfare than are economists, on average.

    And again, I'm correcting for a whole ton of covariates that themselves predict ignorance. The ignorance measure is then telling us what's explained by ignorance beyond that which would be predicted from the overall set of covariates.

  6. Paul: that's exactly the kind of case I was thinking of. However, in that case, it's very difficult to imagine how "crazy voter" results could have delivered things much worse than the ex ante situation.