June 14, 2005
Betting on Climate
Posted to Author: Pielke Jr., R. | Climate Change
Placing bets on the future state of the climate makes sense, but in a research mode, not just in public displays of "calling out" particular opponents. Several commentators here on Prometheus have directed our attention to discussions of "betting on climate," that is, putting one's money where one's mouth is when making claims about the future state of the climate. James Annan, a scientist at the Frontier Research Center for Global Change in Japan, has been trying unsuccessfully to get prominent climate skeptics on the record in the form of a bet. Mark Bahner has responded by trying unsuccessfully to get James Annan on the record with some sort of bet.
This recent flurry of calling people out (reminds me of elementary schoolyard brawls - "I'm faster than you!" "No you're not!" "Prove it!" "Meet you after school on the playground!") no doubt has a high element of drama. Ronal Bailey takes credit for starting the betting flurry when he apparently misquoted MIT's Richard Lindzen discussing what odds he would take on a bet on the future state of the climate. Underlying all this, as Bailey observes, is the now legendary story of the Erlich/Simon wager which in spite of assertions to the contrary is widely viewed as a vindication of Simon's perspectives over Ehrlich's - of cornucopian correctness and catastrophist error. This is where the drama comes from, as bettors who call people out by name probably see themselves immortalized in future history with a publicly-visible betting victory like Julian Simon experienced.
I think that while such chest thumping displays are certainly entertaining, they tell us little about the broader state of uncertainty among experts or the public. But there is something important here that should not be overlooked. Betting, or more accurately a betting market, can tell us a lot about the state of uncertainty across a large body of experts. In 2001 I wrote an essay on "prediction markets," here is an excerpt:
"Nobel Prize winner Nils Bohr once remarked, "Prediction is difficult, especially of the future." Recognizing the difficulties of prediction, what if there was a way to integrate all available information about the future into a single forecast that would instantaneously incorporate and make available new information? Sound too good to be true? Research in economics and recent developments in the private sector point to a novel way to think about forecasting in the earth and atmospheric sciences.
The basic premise of the approach lies in the efficient market theory from the field of economics and most closely associated with the work of Eugene Fama at the University of Chicago in the 1960s. The efficient market theory holds that the current price of a commodity in an exchange market reflects all available information. When you hear the phrase "you can't beat the market" it is referring to the perspective that the "market" (usually the stock market) is efficient; i.e., if information were available that would allow someone to gain a trading advantage, this information would be reflected instantaneously in the price of the commodity through the actions of buyers and sellers in the marketplace. Whatever advantage the trader thought may have existed is absorbed into the market...
What if a "prediction market" were created that would allow trading based on specific predicted outcomes such as the weather? Could such a market be created that operated efficiently and had sufficient participation to integrate all available information about the future? Here is a research hypothesis: An efficient "prediction market" generally will outperform all competing prediction methodologies in the earth and atmospheric sciences...
Several policy issues come to mind. First, there is an implication for the research community. It would appear to make sense to test the hypothesis about whether "prediction markets" can outperform other approaches to forecasting phenomena related to the earth and atmospheric sciences. Perhaps an equivalent to the Iowa Electronic Market could stimulate such research related to weather forecasting, climate forecasting, and, in principle, any area where predictions are made. Second, if the marketplace can provide skillful forecasts, then perhaps a mechanism might be created for this information to be provided systematically to consumers of forecasts through the public or private sectors, or through a partnership."
There is a brilliant, straightforward (and fundable!) research proposal waiting to be written that develops the idea of prediction markets in the context of climate (and something that with a serious PI I'd be interested in collaborating on). The model to emulate is the Iowa Electronic Exchange, which is funded in part by the U.S. NSF as a research project, to explore how well prediction markets perform in the context of elections and certain measures of economic performance.
I can imagine a whole bunch of interesting datasets that might be collected through a climate prediction market. For example, it would be useful to create a market that is only populated by IPCC authors, a market populated by people with the same expertise as the IPCC authors but not part of the IPCC, different markets for different disciplinary areas of expertise, a market open to the public, and markets based on various metrics of interest to decision makers (beyond global average temperature) such as rainfall in particular locations or the number of hurricanes, etc. etc. Such markets would allow for a robust empirical examination of a wide range of obvious and not-so-obvious questions. (James Annan has set up here a market open to the public focused on global average temperature.)
On the climate issue it is often difficult to separate out the political theater from the serious attempts at research. Prediction markets offer a great opportunity to dramatically advance understandings of expert perspectives on the future state of the climate, as well as the diversity of views within the expert community. But to gain such understanding will require going beyond the schoolyard brawl mentality to real, rigorous research. Who thinks this is going to happen? Wanna bet? I'll meet you on the playground at noon.Posted on June 14, 2005 06:32 AM
An excellent suggestion...hopefully it will be developed.
Analyses of the differing groups you propose would also provide data on the feelings of those who actually know someithing about the subject vs. those who know it only peripherally.
However, to get an idea of the uncertainty of the prediction, I think a way would have to be found to relate the size of the wager to the size of the bettor's assets. In your example of the Erlich/Simon wager, the Erlich group (headed by a multi-million selling author) bet a maximum of $1000 (if the price of the goods went to $0) while Simon (an academic - a group not noted for their high incomes) had an unlimited bet (in reality less, but no mathematical limit). Clearly, there was little uncertainty in the direction of prediction in Simon's mind while the Erlich group agreed to a betting structure that limited their losses. This implies one side was betting what he considered a near certainty while the other side had significant doubts about the veracity of their prediction. I'm not sure how to work this into a market structure, but I believe it needs to be there.
Posted by: Jim Kanuth at June 14, 2005 09:49 AM
Roger, you wrote, "James Annan, a scientist at the Frontier Research Center for Global Change in Japan, has been trying unsuccessfully to get prominent climate skeptics on the record in the form of a bet. Mark Bahner has responded by trying unsuccessfully to get James Annan on the record with some sort of bet."
I think it's too early where my discussions with James Annan will lead. But I can tell you right now that I suspect my position is somewhat different from James Annan's, in that I'm not intersted in the slightest in making any money.
That's why I posted these two bet challenges at a charitable site called "Long Bets".
1) Bet challenge #180 is that Michael Chrichton's prediction of 0.81 degrees Celsius warming over the 21st century will be more accurate than the prediction in the IPCC Third Assessment Report, or TAR. That prediction (as interpreted by Wigley and Raper in Science, 2001) was for a 50 percent chance of warming of 3.06 degrees Celsius.
2) Bet challenge #181 is that my predictions of methane atmospheric concentrations, CO2 emissions and atmospheric concentrations, and resulting temperature changes, for 2030, 2070 and 2100, will be more accurate than those in the IPCC TAR.
Both bets are open to any of the primary or secondary authors of the IPCC TAR.
I want the IPCC TAR authors to step up and take responsibility for their (ludicrous) projections. I won't earn a penny from forcing them to do so, since all proceeds from the bet go to charity (and not until the year 2100!).
Posted by: Mark Bahner at June 14, 2005 10:52 AM
By the way, all registered visitors at the Long Bets site can officially "vote" one way or the other on all bet challenges.
Posted by: Mark Bahner at June 14, 2005 10:57 AM
A final word, as lunch is ending, "There is a brilliant, straightforward (and fundable!) research proposal waiting to be written that develops the idea of prediction markets in the context of climate (and something that with a serious PI I'd be interested in collaborating on)."
I'll cc you on an email I plan to send in the next few days.
Posted by: Mark Bahner at June 14, 2005 11:01 AM
Considering that Bahner only made his initial bet offer to Annan two days ago (see http://markbahner.typepad.com/random_thoughts/2005/06/proposed_bets_f.html and http://julesandjames.blogspot.com/2005/06/3-more-non-bets-on-climate-change.html#comments), I think it's inappropriate to say he's been "trying unsuccessfully to get James Annan on the record with some sort of bet." Annan clearly has an ongoing interest in negotiating a bet, unlike the skeptics he contacted in the past.
The discussion has gone here: http://groups-beta.google.com/group/sci.environment/browse_thread/thread/a8e644b272ee890c/fb54fd193a838298?hl=en#fb54fd193a838298
Seems to me Annan has accepted Bahner's offered bet on CO2, additionally offerred a higher-stakes version of the same bet, and is trying to negotiate a fair bet on temperatures. Bahner doesn't appear to have noticed that one of his bets has been accepted.
Posted by: Brian Schmidt at June 14, 2005 11:46 AM
All- Sorry if I mischaracterized the Annan and Bahner exchange. Good luck with the negotiations and feel free to keep us posted.
Posted by: Roger Pielke, Jr. at June 14, 2005 12:21 PM
Simon, like Erhlich, was an academic who had published books, some of which sold well and was associated with the Cato Institute, so I doubt he was destitute. http://en.wikipedia.org/wiki/Julian_Simon
BTW, if you want an example of why you have to carefully define such bets here is one which Simon lost http://www.forestry.auburn.edu/sfnmc/web/bet.html
Posted by: Eli Rabett at June 14, 2005 06:07 PM
Eli- Very cool, thanks for the link to the Simon-South bet, I have never heard of it. South looks to be on a hot streak.
Posted by: Roger Pielke Jr. at June 14, 2005 08:54 PM
I think calling the bet offers a "schoolyard brawl mentality" misses the value they have in public policy. A comment posted on RealClimate nails it:
Posted by: Brian Schmidt at June 14, 2005 09:07 PM
In the end, though, bets on climate change metrics are missing the forest for the trees. Specific temperature values, GHG levels, etc. mean nothing without the larger context of the climate they are forcing or are forced by. These bets would be far more interesting to me if the bettors (not myself, ever, Reno has enough of my money already) would pick some measurable socioeconomic outcomes, such as:
1- number of people/$$ spent on human migration
2- measurable increase in catastrophic storm damage
3- measurable increased stress on water delivery systems
Otherwise, why not just bet black holes and keep it friendly and apolitical? Obviously the reason this is happening is *precisely* because of politics.
Posted by: kevin vranes at June 14, 2005 10:19 PM
Thanks for your comment. I use the phrase "schoolyard brawl" to refer to bets with names attached to them. The value of a prediction market lies in the market. Efficient market theory holds that it is the integration of many perspectives that provides the best available take on uncertain information (hence, efficient). I'll grant you that calling individuals out (Put up or shut up!) is great political theater, and as Simon/Ehrlich tells us can have lasting and perhaps disproportionate impact, but it is something different than a prediction market.
Posted by: Roger Pielke Jr. at June 15, 2005 05:55 AM
I posted this over at Real Climate in their discussion on climate betting:
One interesting point relevant to policy about bets focused on global average temperature (GAT) -- it appears from this discussion that there is unanimous agreement among those who disagree on what future GAT will be that there is no need to discount for policy interventions. That is, the IPCC projections are not conditioned upon assumptions about policies being implemented related to emissions. A functioning market would make this more apparent (e.g., if the market price is equal to the IPCC midpoint scenario), of course, but it appears to me that in the debate motivated by James so far that there is no expectation that emissions reductions policies will have a discernable effect on the climate (as measured by GAT) by 2030 (the terminal point of the bet). From a policy perspective, this suggests a number of possibilities, among them, (a) that the benefits of emissions reduction (as related to GAT) lie beyond 2030, (b) that everyone implicitly agrees that emissions reductions are unlikely (I doubt this is the case), (c) that emissions reductions of any conceivable amount will have no noticable effect on GAT, and so on. If bettors expect that emissions reductions will have a discernable influence on GAT over this period, then it should be apparent in the market value for GAT, which would be less than the IPCC's non-policy mid-point projection.
Posted by: Roger Pielke Jr. at June 15, 2005 07:30 AM
I just posted this comment on "Real Climate." I'm not sure when, if ever, they will publish it. (They don't like words like, "lying"...at least not when applied to the authors of the IPCC TAR.) Anyway, I thought you might be interested.
Roger Pielke writes, "One interesting point relevant to policy about bets focused on global average temperature (GAT) -- it appears from this discussion that there is unanimous agreement among those who disagree on what future GAT will be that there is no need to discount for policy interventions. That is, the IPCC projections are not conditioned upon assumptions about policies being implemented related to emissions. A functioning market would make this more apparent (e.g., if the market price is equal to the IPCC midpoint scenario),..."
I don’t think the market price will equal the IPCC Third Assessment Report (TAR) midpoint scenario, unless the market was seriously ignorant. The IPCC midpoint scenario is completely bogus. And the IPCC TAR midpoint scenario becomes INCREASINGLY bogus the further one goes out in time. The explanation for why the IPCC midpoint scenario is bogus, and becomes INCREASINGLY bogus the further one goes out in time is a simple matter of economics:
1) There were no penalties to the people who made the IPCC TAR projections for lying, and
2) There were no rewards to the people who made the IPCC TAR projections for telling the truth.
No penalties for lying, and no rewards for telling the truth. So the people who made the IPCC TAR projections performed the economically rational act: they lied. (Note: Robert Watson was later dismissed from heading the IPCC, so possibly there was one person who was penalized for the fraudulent projections.)
Some evidence for that fact that the IPCC TAR projections writers lied can be seen simply by looking at the results of the Second IPCC Assessment Report (SAR), versus the TAR. The SAR projected a warming from 1990 to 2100 of 0.9 to 3.5 degrees Celsius. The TAR projected a warming for the same period of 1.4 to 5.8 degrees Celsius.
But stronger evidence that the IPCC TAR projections writers lied can be seen from the fact that some of the highest projections were added AFTER PEER REVIEW (apparently at the urging of Robert Watson):
On the charitable website known as “Long Bets,” I have placed my OWN “midpoint” values for 2030, 2070, and 2100, versus the midpoints in the IPCC TAR. I have challenged any IPCC TAR author to bet that their projections are better than mine:
Further, on my own weblog, I have offered James Annan and William Connolley $4, up to a maximum of $200, for every IPCC TAR author they get to come on the Long Bets website, and “vote” on my projections versus the IPCC’s. (I will pay them $4 for every vote, regardless of whether it’s a vote for my side of the bet or against my side of the bet.) Finally, also on my weblog, I have challenged them to make me the same offer ($4 for every IPCC TAR author vote, up to a maximum of $200, regardless of the side on which the vote is cast.)
Posted by: Mark Bahner at June 15, 2005 02:50 PM
I just realized that I made a serious mistake in that last comment. My comment implies (well, states) that there were no penalties for lying, and no rewards for telling the truth, and the IPCC TAR projections writers lied.
But I did NOT mean that, absent any signals at all, the IPCC TAR projection writers lied. They actually had a very, very large *reward* for lying, in the form of vastly more research money.
That's why they lied. It doesn't excuse what they did, but it explains that they didn't just lie for the fun of it.
Posted by: Mark Bahner at June 15, 2005 03:03 PM
Over the long run (2100) it seems to me that these bets on future climate are really bets on which emission scenerio will be followed. I don't think that anyone rational considers all of the SRS equally probable, but there is probably a lot of disagreement exactly where the future will lie and on a long projection it makes a difference.
On the other hand twenty years is not long enough to significantly separate most of the scenarios including those with various ameloriations and in any case there appears to be a decades long induction period related to the oceans acting as heat sinks. Thus a 20 year bet is more or less a bet on climate sensitivity and Prof. Pielke's point, though well taken on a hundred year long scale, probably makes little difference.
To me the two big uncertainties that could effect a 20 year bet are a large volcanic eruption or a huge methane burp from the clathrates.
Posted by: Eli Rabett at June 15, 2005 09:58 PM