Would you use Predictive Markets for Weather?

[JYarsh]

Hello all,

 

I'm posing this question as part of a research endeavor into the possibility of launching a predictive market for weather and climate events.

 

Many on this forum are familiar with the weather derivatives market, in which companies can hedge against the types of weather that cause them financial harm. The market never achieved significant liquidity during the 2000s, and today resembles an over-the-counter, parametric weather insurance industry.

 

This question concerns the rise of predictive betting markets in areas such as politics and daily fantasy football. I am interested to know whether people on this forum, who comprise a chunk of the weather enthusiast community in the US, would use a predictive market that would allow users to bet on different weather and climate events happening. With sufficient activity, such markets could advance the field of weather and provide an interesting discourse on climate and society.

 

Examples of potential sub-markets include:

- Will 2016 be the warmest year on record?

- Will Boston receive more than 35 inches of snow next winter?

- Will there be more than 5 named Atlantic hurricanes in 2016?

 

I would greatly appreciate any and all answers!

[dmillz25]

It doesn't sound like a bad idea but i feel like its easier to bet on things like the examples you posted due to different indices and how the pattern would set up. Btw No, Yes, Yes for your example questions.

[Ellinwood]

I'm gonna draft El Nino. He's having a stellar preseason.

[Huffwx]

I'm gonna draft El Nino. He's having a stellar preseason.

 

SO did the Eagles, just saying

[eurojosh]

No.  Predictive markets work best for things in which human perceptions have a role in shaping.  For example, political elections and the price of oil a year from now are both subject in great part to human perceptions, which is why prediction markets for these sorts of things can be (sometimes uncannily) accurate - they provide an insight into the wisdom of the crowd, which is the same wisdom that plays a role in shaping the outcome.  On the other hand, humans play no role in shaping the weather, so the market's insights are not valuable as they are not relevant to the outcome.  Crowdsourcing a weather forecast from qualified experts might be a useful alternative, however.

 

Source: I managed an annual global politics prediction competition for several years.

[eurojosh]

Also, FWIW, prediction markets work best for questions on which there is an interval response available rather than a nominal scale.  For instance, from a year out, they do better on predicting a stock market number, which can be anywhere between 0-20,000 than on a much smaller number of distinct, separable options, such as who the Republicans will nominate.  So if you're structuring your questions, may be better to use scaled options than binary or multiple choices.

[Derecho!]

I've actually already participated in a predictive market for weather - The University of Miami/University of Iowa Hurricane Futures Market back in the mid 2000s.

 

There was some controversy about it, I dimly recall NHC was upset about it and it was shut down; I believe Max Mayfield simply didn't understand it, similar to the way US government research on predictive markets for things like terrorism and such were misunderstood and demonized.   Here are a few links, you may wish to investigate more fully:

 

http://www.rsmas.miami.edu/pressreleases/20050815-mahem.html

 

 

http://bus.miami.edu/magazine/fall2012/thought-leadership/learning_hurricane.html

 

"To conduct their research, Kelly and colleagues at the University of Miami and the University of Iowa established the Hurricane Futures Market. It allowed participants to trade securities — an instrument representing financial value — based on where a hurricane would make landfall.

The securities each pay $1 if the hurricane lands in a particular region. The exact price was set by the buyers and sellers. If traders believed a hurricane was likely to make landfall in a particular region, buying drove up the price. If they believed the hurricane wasn’t likely to make landfall there, selling forced the price down. Therefore, the price gives an indication of the traders’ beliefs that the hurricane will make landfall in the region.

“In short, participants needed to figure out how likely it was for a hurricane to hit a particular region,” Kelly explains. “The more accurate they were, the more money they earned.” Traders made their decisions by analyzing hurricane-tracking forecasts from the National Hurricane Center, run by the National Oceanic and Atmospheric Administration (NOAA), and from other sources. The forecasts often differ and use different methods.

How did the securities market participants do? Quite well — in fact, better than NOAA. “The prices predicted where a hurricane would make landfall with 90% accuracy,” Kelly reports. “The National Hurricane Center was right only twothirds of the time.” Why did the traders fare so much better? They were less biased. Traders quickly figured out that official sources erred on the side of caution when a hurricane was expected to make landfall in a densely populated area. Traders took that bias into account, and as a result they were more accurate. Kelly notes that firms are beginning to recognize the practical applications for such prediction markets."

 

 

http://journalstar.com/business/new-hurricane-futures-market-gets-a-stormy-reception/article_dc2dab01-8322-5bc6-9a13-5538b5fe1174.html

 

 

"But the hurricane project, expected to begin next week, has drawn criticism from some top meteorologists, who say forecasts should come from a source the public has come to trust, the National Hurricane Center.

"If they think they're going to help us forecast hurricanes, I don't see how," said Max Mayfield, the center's director. He said the center uses at least a dozen sophisticated models.

"You'd think if they're serious, they would have contacted the director of the National Hurricane Center, but nobody's called me," he said. Mayfield also warned: "Forecasters are not allowed to dabble in this sort of thing."

[raindancewx]

Hi all,

 

Long time lurker on these forums. I've been working on developing a crowdsourcing plan to see if there is mass public interest in this very idea. If anyone would like to join forces, by all means. I've worked out a fair number of details already, a name, verification, what to forecast, how to intake the data, and so on, I'm currently working on a logo and trying to find someone with more contacts on social media than me to help with the crowdfunding element.

[JYarsh]

Thank you all for your feedback!

 

I think there are two separate, but perhaps related concepts here: 1) a predictive market for weather and climate that would convey market expectations of different large scale events (droughts, sea level rise, global temperatures, etc.); and 2) a hedging/futures market in which companies could hedge against specific future weather events. 

 

The challenge for both seems to be liquidity, and the question is whether the nature of weather and climate risks (specific in time and geography) lend themselves to these sort of markets. 

[raindancewx]

I have launched my Indiegogo campaign if anyone is interested. The main reason I am trying this is that my co-workers - and they aren't that into weather, and there weren't that may of them - were able to produce a decent snow forecast for Albuquerque five out of six months.

 

https://www.indiegogo.com/projects/raindance-weather/x/12141322#/

 

Check it out and contribute if you are interested. If nothing else I have a cute animated video.