Thursday, December 20, 2012

Geoengineering, 'gobs', free riders, and free drivers

The Economist has a comment on a recent paper by economist Martin Weitzman who speculates about the prospects for geoengineering (focussing only on aerosol injection). Starting from a common distinction made in economics, between public goods and their externalities, he states that we face two different problems with regard to climate change. The first is the interest of everyone (read: nation states) to avoid dangerous warming but not to bear the costs. This is known as the free rider problem. The second problem is the possibility of single nations tinkering with geoengineering because of the perceived benefit to them, regardless of the harm that might occur to others. This is the free driver problem. The high cost of climate change mitigation makes free riding likely and international agreements unlikely. However, geoengineering (aerosol injection) is cheap and therefore in addition to the free rider problem in mitigation policy we get the free driver problem of geoengineering. Weitzman then poses the question what the likelihood and desriability of international cooperation about geoengineering will be. He develops a model which specifies several assumptions (I won't go into this). A central concept for the model is the gob, a new term coined to define goods that can be both 'goods' and 'bads' at the same time. Here is a lengthy quote:
A pure public good is typically defined as a commodity that is both nonexcludable (no one can be excluded from consuming it) and nondiminishable (one person’s consumption does not alter the amount available to others). A public good is standardly considered to be good, meaning that almost everyone thinks more of it is better, at least throughout the domain having policy relevance. Usual examples are police and fire protection, national defense, weather predictions, and the like. A pure public bad is typically defined as a commodity that is both nonexemptable (no one can be exempted from consuming it) and nondiminishable (one person’s consumption does not diminish the amount that others must consume). A public bad is standardly considered to be bad, meaning that almost everyone thinks more of it is worse, at least throughout the domain having policy relevance. A standard example of a public bad is pollution. Because it is costly to increase the level of a public good or to decrease the level of a public bad, such situations are plagued by the free-rider problem. Instead of paying their fair share, everyone wants to free ride off the payments of everyone else. The problem of a geoengineering externality has a different structure. 
I now want to introduce the idea of a gob. A “gob” is a commodity that may be good or bad depending on who is consuming it and how much they are consuming. A pure public gob is a pure public good (more of it is better) for some people under some circumstances and a pure public bad (more of it is worse) for some other people under some other circumstances. Throughout this paper, the primary example of a pure public gob is geoengineering in a future world sufficiently impaired by climate change that some countries would want to do some of it on their own if allowed to act unilaterally. The key issue is that parties differ in their attitudes toward whether more or less gob is desirable and some mechanism is required to reconcile these differences.
In the technical part of the paper Weitzman develops preference and distribution functions, based on two concerns which are tightly connected, optimized social welfare and a simple voting implementation mechanism on the level of global geoenineering governance.

Another approach to the problem might lead to different results. It does not share the characterization of climate change as being a public good but defines it as a common pool resource. A public good can in principle be provided by one powerful actor without the free riding of others adversely affecting it. In contrast, unilateral action cannot produce or protect a common pool resource, but can harm or destroy it. The substitution of the central terminology has consequences for policy making. GHG emissions of one large emitter could spoil the efforts of many small. And the injection of aerosol particles by one country could do the same with regard to geoengineering. In this framework no gobs are needed.

On the basis of public goods theory Weitzman is looking at an abstractly defined social optimum which somehow needs to be enforced globally. On the basis of common pool resources theory we are able to perceive the problem in a different way. We are in the land of Elinor Ostrom's management of common pool resources through cooperation on different levels (mainly local and regional). 

But how does this approach deal with the free driver problem? There are several possibilities, of which the most likely will be the one which follows in the default logic of the current international system. This means that it is easier to agree at UN level to ban geoengineering projects (because they are not allowed as such). If a country was to prepare for a project of this kind it would run against the opposition of other countries and the matter would probably end up at the UN security council. 

Maybe I am missing something here, but it seems to me that attempts to allow geoengineering projects of this kind face the same problem as attempts to reach global agreement on mitigation targets and timetables. This might change if there is a universal perception that the ‘social welfare function’ of geoengineering was clear cut.


Anonymous said...

Our understanding of the earth's climate system is still poor and far away from being complete. The amount of what we don't understand is much bigger than what we do know about it.

And yet some people are starting to discuss geoengineering solutions to solve what they just
b e l i e v e to be a problem?

Ignoring the lack of a proof in empirical research or at least a convincing theory. All we got are models and they do not fit very well with real world observations.

Following the sensitivity discussion at JC's blog, one might wish to find Pielke's famous BS button on the keyboard when he hears a call for geongineering projects ...

V. Lenzer

ob said...

if we see weather-modification as a form of geoengineering, then geoengineering is older than the whole climate-debate, isn't it?

@ReinerGrundmann said...


Indeed, but weather modification is small scale. And weather is not climate. Interesting that main definitions of geoengineering are about interventions into the climate system.

richardtol said...

The paper is a bit strange. It starts as a paper about public gobs -- a non-cooperative concept -- and then goes on to define a voting rule -- a cooperative concept. That is, halfway through, Wietzman assumes away the problem.