Bastiat’s Bastions

I should have purchased flood insurance in 2005. It turns out I made a bit of a mistake when purchasing my house.

Apparently I haven’t learned from my mistake. Maybe I should have purchased flood insurance this year. But I did not. Apparently, I like to live dangerously.

So can I run out and buy flood insurance tomorrow? I could, but there is a 30 day waiting period for the flood insurance policy to become active. If there were not such a provision, one would expect there to be many policies sold this week.

I am stuck bearing the risk of flooding myself. As such, I will be nervous for a couple of days. But enough about my worries.

The other day, Dr. Coats wrote on gambling markets predicting election outcomes. Between the two of us, we have written many times about prediction / gambling markets and the efficiency thereof. If you look over the left, you’ll see in the archives you can read all the post about prediction markets here. Nonetheless, one more post couldn’t hurt.

Guess what else is predicted by gambling markets. Anyone? Weather outcomes. Yes, that is right, you can bet on where Gustav will land.

Some people are going to get all upset – why would people want to bet on where a hurricane will land. It is immoral, rude, whatever. But there is at least one very good reason why these markets are very nice to have. And there others are I won’t even discuss.

But first, some of the background (again). Just as with the political markets, each contract will ultimately expire at a value of 0 or 100. If you purchase the Louisiana contract, and Gustav makes landfall first in Louisiana, and is at least a category 2 hurricane, your contract will be worth 100. If Gustav first makes landfall in some other state, or is not a category 2 hurricane, your contract will be worth 0. You can buy or sell these contracts at any time. You can also buy contracts for the other states that border the gulf, and even some states that don’t. (My Clemson pals are resting assured that the gambling market thinks it is very very unlikely that Gustav will first make landfall in South Carolina.)

Before I get to the market’s current predictions, I should note that currently, this market is what finance scholars like to call a “thin market”. There are not many buyers and sellers. The spread between the bid (what people are willing to pay to buy a contract) and the ask (the price at which people are willing to sell the contract) is quite large. Thus, it is difficult to give you a precise probability – trades are infrequent and the prices noisy. However, I’ll guess it will pick up in the next couple of days.

Nonetheless, at the time of my originally writing this post (3:30 AM on Thursday – which explains any remaining typos), the prices suggest the chances of landfall is:

Louisiana: 40%
Texas: 20%
Florida: 20%
Other: 10%
Less than Category 2: 10%

Click here for updated prices.

A couple of things of note:

If you are an all-star meteorologist, and have a better idea where the hurricane is going than the gambling market, you have a money-making opportunity. But just as Dr. Coats explained with political markets, my bet, pardon the expression, is that the gambling market will get it about right. Again, the caveat is that this is a thinly traded market.

As I mentioned above, some people won’t like this idea. In fact, some pretty smart people though it would be a good idea to set up prediction markets for terroristic events – Dr. Coats has written about this previously.

But one hugely desirable feature of these markets is they allow people to hedge their risk. If I bought a contract today, I can basically buy hurricane insurance, and do so with no waiting period. How?

I could, say, buy contracts that are worth, say $4000 based on the price of 40 today. If the Hurricane lands in Louisiana, while my house might be gone, each contract will have a price of 100, and thus my contracts are now worth $10,009. My gain on the contracts ($6000) would at least partially offset the losses on my house.

If, however, the hurricane peters out or ends up in Florida, my contracts are worth 0. I’ve paid $4000 for contracts that are not worth anything. But I would have gotten to sleep better.

So one last thing – Is the price of 40 (a 40% chance) appropriate?

Just for fun, I went to www.nola.com, and checked out the latest computer models of where the hurricane will go. There are 12 different computer models, and they all disagree to some extent. I have no clue which ones are a good model and which ones are a bad model. However, of the 11 that project far enough out see what state they’d land in, 6 of the 11 predict a Louisiana landfall. If I assume each model is equally good in predicting where the hurricane will go, this, it seems, predicts a probability of 0.545 that the hurricane lands in Louisiana. So you’d think the price of 40 is a good deal, right?

Not so fast. Keep in mind the contract requires landfall in Louisiana and the storm to be a category 2 or higher. If we think the probability that storm will be category is 0.9 (and the intensity of the storm is independent of where it makes landfall), then the probability the storm lands in Louisiana at category 2 or above is 0.545 * 0.9 = 0.50. We are getting pretty close to the neighborhood of what the gambling market says.

What will be interesting to see is how the price evolves as the storm gets into the gulf. If it weren’t for the thin market, we could say more.

My prediction? Bid ask spreads will narrow, and one or two states will see big increases, while the rest will see decreases. Stay tuned.

–CT

In the approaching U.S. Presidential election, there will be discussions about who is better equipped to lead the country for the next four years. Discussions will get heated and opponents demonized. Other discussions will turn to the horse race. Who is ahead? What is the extent of the lead? How can winners be predicted beyond the daily fluctuating and often conflicting polls that are constantly being reported?

Certainly, polls are used to forecast election outcomes, and are crucial in the process. But to see where the smart money is in an election, just look where the smart money is.

In commodities markets, the futures markets provide a best guess as to where the price will be in the future. Speculators abound. And, as I recently discussed (but also originally here at Bastiat’s Bastions), foolish market participants lose money and exit the market while those who are better at predicting are able to increase their market presence.

The market price in speculative markets is also the one that is at the middle of the pack, the price with as many dollars bet above the going price as below, making it a strong predictor.

Speculative markets can be best thought of as betting markets, like sports betting markets. Here is where we see the smart money in the market, where the center of the money being wagered by those who are confident enough with their predictions that they each “put his money where his mouth is.”

This argument is nicely made in a recent Slate.com article, where the markets are explained. Take a look at Intrade.com and at the University of Iowa’s “Iowa Electronic Market, the first election market, (a research and educational tool set up by faculty members at the University of Iowa), but also look at each prospectus in the Iowa markets. Two types of markets to look at are the vote share markets, where the betting is on who will get what share of the vote, kind of like betting in sports markets with point spreads, and the “winner-take-all” markets which simply bet on the winner. The market prices for the vote share markets, then, are predictors at how the shares of the popular vote will turn out, while the price in the “winner-take-all” market is the market prediction of the probability of Obama or McCain winning the election.

At the close of the Iowa Electronic Market on August 20^th, the prices in the vote share market stood at $0.511 for Obama and $0.499 for McCain. The redemption prices at the end of the election will be for Obama will $1.00 times the two-party vote share that Obama receives, and similarly for McCain shares. So, if a speculator thinks that Obama will garner more than 51.1% of the popular vote (among the two top parties), that speculator will be driven to buy shares of Obama, while those who think Obama will get less of the vote will be driven to sell shares.

In the winner-take-all market, the close-of-midnight prices on August 20th were $0.596 for Obama shares and $0.396 for McCain. If a speculator thinks that Obama has better than a 59.6% chance of receiving the most popular votes, that speculator has an incentive to buy shares of Obama in the winner-take-all market. If a speculator thinks McCain has better than a 39.6% chance of getting the most votes, that speculator has an incentive to buy McCain shares, but will have an incentive to sell if the speculator thinks McCain’s chances are lower.

There is a concept in economics called the “efficient markets hypothesis,” which only means that market prices embody all available information. One type of information that is crucial in these markets is the information provided by polls. Speculators in these election betting markets also use every bit of information available not only about which candidates are favored by voters but also about how driven potential voters are to actually make it to the polls to cast their ballots.

So, if you want to find out how your candidate is doing, look beyond the polls and the pundits to the betting markets, particularly those at intrade.com and at biz.uiowa.edu/iem. While no forecast is perfect, these markets have already proven to be better than the polls and television’s talking heads.

(Note: This article, written by Dr. Coats, was first published at BasilAndSpice.com on 8/23/08.)

-MC