Over the last few days, a storm has arisen over Steve Levitt and Stephen Dubner’s Superfreakonomics. The book hasn’t been released but the chapter on climate change has generated a ton of discussion. That chapter was posted (I suspect not with publisher’s approval) on the net and I have read it. It is entitled “What do Al Gore and Mount Pinatubo have in Common?” It is clearly part of the by-line for the whole book, “Global Cooling, Patriotic Prostitutes and Why Suicide Bombers Should Buy Life Insurance.” So it is clear the authors think this is a key chapter.

What is the controversy? Levitt and Dubner are accused of bias (to the point of denialism) and also stating incorrect scientific facts. Here is an identification of the latter. And then trying to work out what is going on is a mess. But, in reality, that is a couple of pages of the whole chapter. What of the rest of it?

This is one where each will have to read it themselves to decide. But on its face, in my opinion, there is nothing too wrong with it. Not enough for name calling. What the authors are doing is identifying the ‘economist’s angst’ in this whole discussion. They start off with the 1970s discussion of global cooling although it isn’t clear that is a strong prediction either. We are all frustrated that climate models aren’t perfect even if sufficient scientists think there is enough and so I read that discussion as providing an example of that. But then there is a pretty clear discussion of the intertemporal problem we face in dealing with this. Again, all fairly non-controversial if leaning to a sceptical side. Nothing to get upset about. Indeed, Levitt has a point that the entire chapter is mostly about what to do about global warming and so it is hardly a denialist exercise. They aren’t in the Michael Crichton camp. (Paul Krugman’s critique just doesn’t square up on reading the chapter).

What I think the chapter does well is express some hard-headed uncertainties about what we are trying to do on climate change management. I have expressed similar concerns on this blog and noted the attacks that come. Let’s face it, many environmentalists argue with a religious fervour. That might be fine for them but it isn’t for me. And moreover, just because climate scientists have demonstrated human generated warming does not mean that they are the final work on economic and political policy. The chapter really identifies such issues well. Australians should take note that when the book uses us as an example as to why unilateral greenhouse gas reduction would be a personally costly and fruitless idea.

The chapter then moves on to consider what to do about climate change. It too quickly moves through current plans for carbon prices saying they are uncertain and costly. Both true but there is surely more to it than that. After all, there is a ton of pollution going on and surely economists should be nervous about that being unchecked. Levitt and Dubner recognise externalities but then propose to mitigate externalities by eliminating the harm rather than changing behaviour. This is to be done through geoengineering which is discussed at length.

For starters, their claim that geoengineering is work exploring and even experimenting with is pretty persuasive. Why not try it? They argue that a form of repugnance is preventing that being taken seriously and note that environmental leaders will have to be convinced to turn that ship around. They write:

The real heavyweights in this fight are people like Al Gore.

And what does he think of geoengineering?

“In a word,” Gore says, “I think it’s nuts.”

And Levitt and Dubner ‘leave it there.’ You don’t leave it there. Please explain what Gore’s issue is. I still don’t know. They don’t explain and that exposes some underlying bias in the chapter.

But my issue is that they abandon economics at this point as they are mesmorised by the idea we can solve climate change cost effectively. The underlying argument as to why we should pursue this rather than behavioural change is not environmental religion but whether it is possible to put prices on this and get behavioural change. Levitt and Dubner don’t think it is possible. They argue that if you can’t get doctors to wash their hands to stop the spread of infection, how can you expect people to clean up the planet?

But come on. Isn’t the whole point of the Freakonomics project that prices work and behaviour changes in response to incentives? Everywhere else, a few pennies will cause massive consumption changes while when it comes to a carbon price, it is all too hard. My own view is that a carbon price plus some information might have some drastic effects on the behavioural side. This chapter should have been about what it takes to change behaviour and what we do and do not know about that. Instead, it passes through that entire issue in a completely unsatisfactory way. Where is the call for field experiments and randomised trials? Instead, we should stick hoses into the atmosphere and seemingly give up on trying to get people to turn off lights when they exit a room.

And there is another missing bit of economics. It is stated but again ‘left there’ but one concern about geoengineering is that it might be a license to pollute. No guilt or concern about future carbon prices holding anything back. Again, that is not explored but it is an issue capable of economic analysis and it is a gap here that they do not consider it. That is disappointing.

This is all a shame because the point of having a popular face for someone like Steve Levitt is to be able to push agendas they are convinced about. The geoengineering one is a candidate for that but the incompleteness of the economic job has undermined the exercise and may well damage the platform he has built. It would have only taken a few pages. Yes, it would be a little less punchy and counterintuitivey (is that a word?) but the influence factor would surely have been stronger.

When it comes down to it, pollution is occurring. As Ken Arrow pointed out long ago, if there are irreversibilities then there is an option value in curtailing the polluting activity so that we can work out what the right future path is. It is that risk that means we should not let externalities go unpriced. And even if geoengineering works out, there will still be externalities. The safe course is to deal with them now.

And as to whether you should buy Superfreakonomics. I trust Tim Harford’s review on this one and will likely buy the entire book.

21 Responses to The Climate of Superfreakonomics

  1. andrewt says:

    Worse, you don’t know that Gore actually said that geoengineering is nuts or, that he believes that  (but he may well) .  As they’ve inverted Ken Caldeira’s views – this books is clearly a source where you should check every claim carefully.
    And the “solar cells are black”  stuff suggests zero review/proof reading/fact checking.   Blue PV panels are so ubiquitous that any vaguely observant person should know this is incorrect and the quantitative error looks worse. The effect of the solar cell on albedo needs to be compared  to the radiative effects of CO2 emission avoided.   I haven’t seen anything definitive but people are suggesting the second is over 4 orders of magnitude larger
    They are also very misleading about what scientists were talking about in the 70s, which you can see if you read this ref: http://ams.allenpress.com/archive/1520-0477/89/9/pdf/i1520-0477-89-9-1325.pdf

  2. Brad Johnson says:

    My reading of the chapter — isn’t so sanguine — Levitt and Dubner strongly overstate the uncertainty about human involvement, imply climate scientists don’t understand the forcings of various greenhouse gases and the sun, and also strongly overstate the uncertainty of projected damages from global warming. Among a raft of errors, minor and major.
    Putting that aside — you’re absolutely right that their arguments about the political impossibility of carbon pricing are bizarre. The US already has some de facto carbon taxes (gasoline taxes, etc.) and is implementing carbon-limit standards for automobiles. The EU has a cap-and-trade system and many member nations also have carbon taxes on top of their strong gas taxes. It’s certainly fair to say that carbon emissions reduction is a difficult political problem, but it’s certainly not an impossibility.

  3. Richard Green says:

    Whenever Freakonomics I strayed from Levitt’s work it became very weak. Unsourced anecdotes and cool stories that weren’t satisfactorily explanative or simply poorly researched (like the Ku Klux Klan segment, which was largely retracted). This was is very stark contrast to the segments based on Levitt’s work which was sourced in its empirics and already subjected to rigourous scrutiny.
     
    Now Superfreakonomics seems to be all of the weak points of the first book, with no Levitt work. In short, it seems to be Dubner with  Levitt’s name. This does not bode well.

  4. [...] Joshua Gans: There is nothing too wrong with [the geoengineering chapter]. Not enough for name calling. What the authors are doing is identifying the ‘economist’s angst’ in this whole discussion. [...]

  5. Kevin Cox says:

    My main objection to the article is the assumption that models of economic systems are something like models of natural systems and economic thinkers have something they can contribute to the debate. The writers have obviously never looked at or if they have looked at them have misunderstood natural systems modelling.
    If economics used the same sort of modelling as is used by the weather predictors then the GFC would not have been a surprise.
    The reason is that economic modelling models emergent properties not the underlying interactions that give rise to emergent properties. Weather forecasting – as an example – models the dynamic underlying mechanisms and uses “statistics” on the emergent properties to see if the underlying mechanisms description matches reality.
    Economic modelling uses “statistics” on the emergent properties of the actual system and models those.  It is little wonder that economic modelling does such a bad job at explaining what has happened let alone what will happen.
    The climate change sceptics tend to use economic type modelling – correlations and inferred cause and effects – to “model” climate systems. Climate change models – or at least the people who do it well – observe the underlying mechanisms and then make predictions on the emergent properties. The weather guys do a fantastic job but because of the “chaotic” behaviour in most natural systems caused by positive feedback they can NEVER predict the weather or a hurricane in detail. However, they can predict  “emergent properties” of the total system.
    The modellers I know are much much more pessimistic than the public statements they give. They have modelled the underlying physical interactions and they can see a multitude of ways that the systems may go – all for the worst.
    The human race best chance of survival is to try to stop some of these runaway possibilities and that is best done using sound models based on the underlying basic interactions.
    My plea is for economists to take a good look at any of the natural sciences most successful modellers and to emulate them and please please do not use historical data as the basis for your models. Use the historic data to TEST your models. Finally do not critique natural systems modellers on the basis of your techniques because their models are way way better than yours.

  6. [...] Joshua Gans identifies in Dubner and Levitt an odd inconsistency that I’ve identified more broadly: those who go on and on about how people respond to incentives when they’re making a pro-free-market argument suddenly seem to lose all faith in the power of incentives when the goal is to induce more environmentally friendly behavior: But come on. Isn’t the whole point of the Freakonomics project that prices work and behaviour changes in response to incentives? Everywhere else, a few pennies will cause massive consumption changes while when it comes to a carbon price, it is all too hard. [...]

  7. M says:

    Kevin some good points on modelling.  Last I checked economists didn’t require super-computing time to model anything.
    For climate models, we understand a lot of the underlying physics (like radiation and insulation, not so much clouds though), its just a case of adding all the details correctly then throwing massive amounts of CPU power at the problem to get good enough resolution.
    Having seen the level of detail and rigour involved in Astrophysics models I don’t really rate most economic “modelling”.  Its pretty standard to spend 3+ years coding and testing one or two new elements to a simulation (be it cosmological, blackhole, SPH+gas+magnetic fields, etc…).

  8. DP says:

    Kevin,

    Hopefully nobody acccuses me of climate change denial, however my understanding of the current climate models is that they cannot actually predict/model what happened in the past. I believe that the climate models are indeed more complex than economics models, but this lack of matching historical data with predictative capability is a weakness not a strength.
    Finally, the arguments about relative warming versus cooling depends upon which decade the writer chooses to use as the baseline. I think the decades of the 50s, 60s and 70s were actually cooler than long-run average, possibly due to sulfur dioxide and nitrogen oxide pollution prior to the introduction of a (successful) cap and trade system  in 1990. Now there is an irony. Anyone for geoengineering?

  9. Paul Klemencic says:

    Have you guys read page 187 of SuperF ? The number of lies and distortions and outright scientific errors is almost unbelievable.
    Quoting Myhrvold as an expert in thermodynamics, electric power generation, and heat transfer is laughable… an undergrad ChemE or MechE could easily see that the statement about the thermal energy re-radiated by solar panels as a significant global warming issue is clearly wrong. First, the solar panels are covering a surface that likely had a similar albdedo.
    Next, even the best power plants reject half or more of the thermal energy to the environment, but this thermal energy is insignificant compared in the global energy balance. Finally, CO2 ends trapping over 100,000 times the thermal energy released by burning carbon to CO2. Myhrvold’s energy analysis would get a failing grade in any undergrad science course on global warming. Anyone trained in doing energy balances, and understanding the science behind the energy industry would have spotted this howler a mile off. Myhrvold seems to suffer from PowerPoint superficial analysis syndrome. Worse, Myhrvold, then keeps on with the same misunderstanding and explains that solar power plants would actually make global warming worse for years to come, because of this thermal energy release. This whole section is total nonsense.
    Here is a critique of that page in SuperF:
    http://delong.typepad.com/sdj/2009/10/six-questions-for-levitt-and-dubner-more-superfreakonomics-blogging.html
    They also seem to have bought into denier propaganda about global warming forecasts depending on complex computer models, and how these models could be wrong and thus global warming may not be as bad as forecast.  Wrong!!  And I don’t have enough space here to even begin to get into all the other problems with this geoengineering scheme, like ignoring the ocean acidification problem, etc.
    Levitt and Dubner need to save what is left of their reputations and get the publisher to pull the book. Next time, they need to have more creditable and knowledgeable people in the fields they write about, review their material.

  10. Paul Klemencic says:

    It is also clearly apparent that Levitt and Dubner walked into several traps laid by skeptics. This is surprising because these gentlemen appear to be reasonably intelligent and diligent and should have been able to see the traps, particularly since the traps are well-known skeptical ruses.
    I have found the videos by Peter Sinclair effective in exposing several of the traps that caught Levitt and Dubner. The videos are short, entertaining, and to the point in refuting the skeptical lies ( eventually published in Superfreakonomics).
    Here is the link to the video debunking the statements about a 1970’s scientific consensus on global cooling prominently printed on the SuperF cover:
    http://www.youtube.com/watch?v=4nTw0KneNLg
    Here is a link to the video debunking the so-called recent cooling since 1998:
    http://www.youtube.com/watch?v=y15UGhhRd6M
    The fact that the SuperF authors fell for these long debunked lies, even to the point to exclaiming that the scientific consensus on AGW is akin to a religion, shows how little research they did before writing this chapter.
    The idea that by simply adopting a contrarian stance on a subject as well studied and analyzed as global warming, this will suddenly reveal an earth shattering new solution, is laughable.
    There is no substitute for knowledge, and espousing a solution without at least a modicum of knowledge about the factual information in a specific field is a mistake.  Levitt and Dubner need to act fast, and fess up to the mistakes, or else they run the risk of being remembered throughout history as the chumps who fell for the most rudimentary lies about the science of global warming.

  11. Anna says:

    The climate scientist bloggers over at Real Climate have good overview of the science and particularly why geo-engineering is a bad idea:
    http://www.realclimate.org/index.php/archives/2009/10/why-levitt-and-dubner-like-geo-engineering-and-why-they-are-wrong/

  12. Kevin Cox says:

    M DP
    Complex systems with positive feedback mechanisms are indeterminant. The so called butterfly effect comes into play where a small change in the initial conditions means a different outcome.
    Economic modelling is deterministic in the sense that if you start with a small variation in the initial conditions you will end up with the same result.
    Dynamic systems modelling of the economy is not all that hard if only the modellers would look beyond their differential equations and move towards agent based models. Essentially you build a replica of the existing economy which is pretty easy to do as for Australia there are say 20 million economic agents and the way they interact is relatively simple.
    What you do is you model agent interactions. You give me some goods and I give you some money etc.
    This is pretty easy to do as you can watch people do these things. You simplify by giving the economic agents rules of behaviour depending on their characteristics and you start the agents interacting.
    You now observe the behaviour of your model and see if it matches what happens in the real world.
    You can at the same time observe the real system and look for chaotic (unpredictable) patterns.  Unpredictability arises from the construction of the systems so that it has positive feedback mechanisms. The one that causes the business cycle and the GFC is the ability of some agents to create money backed by other money.  Make the system so that this is not as attrractive as creating money (loans given) when there is a real asset backing it now or in the future and we will make our economic system more predictable.
    This is only one of many ways to change the interactions of agents to dampen this positive feedback attempt but it will easier than trying to control the creation of money through the price of money.
    Come to my workshop on Sat 24th Oct http://greeninstitute.org.au/gnd/program/financing-green-initiatives/ to see one way of doing this – which has the added benefit of reducing green house gas emissions with no increase in energy costs.

  13. [...] of the defenders…we have Tim Harford, Bryan Caplan, Robert Waldmann and Joshua Gans.  Some neutral players include Tim Haab and Tyler Cowen.  Brad Johnson on Caldiera.  Even the [...]

  14. DP says:

    Anna,

    Has the relationship between SO2 pollution and the cooling observed mid-century been investigated in any detail?

    I agree that geoengineering on a dynamic system possesses the real possibilities of unintended consequences, however it is hard to discern today’s impacts of civilization going about its business as anything but. (For example, there are various reports on the measured effect of aircraft contrails during the Sept. 11 ban on flights over USA.) Maybe we’ll need to buy some time until mechanisms to reduce (as opposed to halting the increase) CO2 kick in?

  15. [...] it is true that the bulk of the chapter is about how to respond to climate change, and Levitt and others have more or less argued that ’so what does it matter if there are some minor errors in the [...]

  16. [...] the links. The uberbloggers you’ll find for yourselves, but also check out Yoram Bauman and Joshua Gans. Stephen Dubner defends the chapter [...]

  17. Barry says:

    DP:  “Hopefully nobody acccuses me of climate change denial, however my understanding of the current climate models is that they cannot actually predict/model what happened in the past. I believe that the climate models are indeed more complex than economics models, but this lack of matching historical data with predictative capability is a weakness not a strength.”

    Your understanding is incorrect; calibration to historical data has long been something that climate models were judged by.  Read the ‘Discovery of Global Warming’; among many other things it covers calibration to past climactic events.

  18. DP says:

    So it is… However the IPCC site itself is best. Very open discussion on the accuracy of their models as compared to past periods.

    http://ipcc-wg1.ucar.edu/wg1/wg1-report.html

  19. [...] Josh Gans: The Climate of Superfreakonomics [...]

  20. [...] had intended to review Superfreakonomics chapter by chapter but only managed Five, One and Two. Chapter 3 was perhaps the most satisfying of the book dealing with John List’s [...]

Leave a Reply

%d bloggers like this: