October 07, 2006
Following the U.S. elections
If you're looking for an alternative to the talking heads and endless spinning, here are some quantitative tools for following the U.S. elections that I've found useful.
Top of my list would be the betting exchange Tradesports, according to which the Republicans took a real beating this last week. Retaining control of the House is now less than a 50% probability, and Democrats gaining control of the Senate has also become a significantly more plausible event than it was judged to be a week ago:.
The same story is told by the Iowa Electronic Markets, which also has a contract for the now seemingly irrelevant possibility of the Republicans actually gaining seats in either house:
These graphs are live links, so you can check back to this page from time to time as the election draws nearer for updates.
On specific elections, the Senate races in Missouri, New Jersey and Tennessee are very close to 50-50 propositions at the moment, with the odds currently slightly favoring the Republicans retaining their seat in Missouri, gaining a seat in New Jersey, and losing a seat in Tennessee.
In all the other senate races, Tradesports has the favorite with 70% odds or better. These include Democratic pickups in Montana, Ohio, Pennsylvania, and Rhode Island, while Lieberman will likely retain his Connecticut seat.
For following the polls themselves, a nice site is Pollster.com. Also noteworthy is Election Projection, which follows good statistical procedure in giving some small weight not just to the average polls, but also to other useful information such as the recent tendency of the state to vote Democrat or Republican. On that basis, EP does not yet call Tennessee as favoring the Democrat.
And if you still want a talking head, it's hard to beat Larry Sabato.
UPDATE: Election Projection is now joining Tradesports in betting the Democrats will win Tennessee.
Posted by James Hamilton at October 7, 2006 06:39 PMdigg this | reddit
there is also intrade, another electronic market.
Posted by: supersaurus at October 7, 2006 07:04 PM
I always like indices based on people putting their money on the line. And for once in my blog career, I'm going to give Lawrence Kudlow a little credit - he's been following this market information as well. And yes, the GOP's favorite financial talking head is worried that Nancy Pelosi will become Speaker of the House. Unabashed partisan (D) comment - I'm equally worried that Bill Frist will remain Majority Leader (albeit I have as much as respect for Chuck Hagel as I do Harry Reid).
Posted by: pgl at October 7, 2006 07:16 PM
toss up any graphs one wants to
the bottom line is that the people want a change
Posted by: kuros at October 8, 2006 07:35 AM
pgl: since frist is retiring from the senate (and his seat may be lost to a democrat) you can probably relax, although lame ducks can still do some damage.
Posted by: supersaurus at October 8, 2006 07:40 AM
Despite all their flaws, the Republicans are still best for the country. Nancy Pelosi trotted out the usual list of Dem nostrums to remind us of how much damage they can do. With the surge in tax revenues, Bush's goal of halving the deficit has been met three years early. The tax rate cuts need to be made permanent, which will occur only if the Republicans hold on.
I am not throwing in the towel yet, Dem dirty tricks notwithstanding.
Posted by: Rich Berger at October 8, 2006 11:14 AM
Rich: Wow. My old man is a retired marine, member of the NRA, always carried a loaded gun by his side when we vacationed, a life long republican, Goldwater fan: even he voted for John Kerry.
My point: he's about as conservative as you can get, but has always allowed intellect to guide his conservatism. And his intellect tells him that the Republicans are headed in the wrong direction.
I'm an independent. I don't see Dem dirty tricks. I see a corrupt ineffective ideologically blinded party imploding. Time to clean house. Let's hope Democrats take both houses.
Posted by: T.R. Elliott at October 8, 2006 11:40 AM
Further probably worthless comment/thought from me: while jogging, I got to wondering about the accuracy of a betting exchange. Think of it this way. Isn't Las Vegas a sort of betting exchange. People bring their money to Las Vegas and bet that they can beat the house. They can't. And odds are they won't. Yet they bring their money anyway. Same holds for the lottery.
So to the comment above, one I would often make so I'm not deriding, that it's useful to look where people are putting their money, I'm just wondering how much can we really ascertain? How much are people simply betting their particular ideologies, their particular irrationalities, etc.
Or look at it another way. What the value of a price of oil? For a while, I've been hearing that money has pouring into commodities, driving up the price. This has been coming from those who argue that the real price of oil should be lower. So in that case, they argue the opposite: don't look at where people are putting their money.
So now I'm wondering. Should we look at where people are putting their money? Or where they aren't putting their money? Both? Neither?
Why are complex systems so--complex?
Posted by: T.R. Elliott at October 8, 2006 01:00 PM
Some might be interested to hear anecdotes like, e.g., I have always been a conservative, always voted for Republicans, etc., but I'm not one of them. These types of comments are just a form of the argument from authority. I am not a conservative, but have never voted Democrat and certainly will not this time. There's too much at stake.
All indications are that this election will be a close one. I hope the Republicans hold on, but come out chastened.
Posted by: Rich Berger at October 8, 2006 03:09 PM
Allen in Virgina has fallen below 70%. The Washington post has a story today that he has not been reporting infromation about stock options as required by senate rules. Some of them were with companies that depend on government contracts.
Posted by: joan at October 8, 2006 06:31 PM
T.R., I think most people go to Las Vegas understanding they're likely to come away with less money than they brought. But, if you do wish to think of Las Vegas as akin to Tradesports, let me ask you, which side of the Vegas bets would you want, the house side or the suckers' side? In particular, if you had an opportunity to make a large number of uncorrelated 50 cent bets, each of which had a 60% probability of paying off $1, would you take it? If not, let me tell you about the Law of Large Numbers, and the principle that keeps all that neon glowing in Las Vegas.
The relevant question for these prediction markets is, do you think the probability of, say, the Republicans holding the House is 60% rather than below 50%? If so, then why not place the bet yourself? Wouldn't you like to have more money? Even if you don't want more money, there are plenty of other people who would, and will happily step in to profit from anyone who participates in Tradesports for ideological reasons.
And, if these graphs do nothing more than summarize ideology, why do the prices move up and down with news, such as last week's developments about Foley, or Joan's latest point about George Allen? Does ideology change week to week like this? I think not, but the honestly evaluated probabilities do.
It helps me a lot to know what other people think, and I believe this is exactly the information conveyed in the graphs above.
Posted by: JDH at October 8, 2006 07:51 PM
Just to be clear, I'm not maliging the idea of looking at what people think. I'm also a big fan of knowing what people think. I also strongly believe that what person X thinks other people think affects not insignificantly what person X thinks. In an iterative sense. Hence the herding instinct. Hence advertising.
So I'm interested in what Tradesports tells me about what people think. And I'm not at all opposed to the possibility that there might be a correlation between what the selection of people using Tradesports think and what the voting populace at large thinks. I just don't know how the two correlate. I'm not sure about its predictive power.
I don't think your comparison with betting for or against the house is relevant. You are asking me to bet against an average that I suppose I could call deterministic. I know that the house is going to win on average. Just as I know the sun will rise tomorrow. The rules are determined.
The people betting on tradesports are guessing about a complex process, not a well known imbalance between the odds house and the gambler.
Also, I really don't know what, down deep, people think about the state of their bank account and wallet when they return from Las Vegas. Do they know they are going to lose? I'd love to know. I don't and have never gambled in that sense. I do watch people buying lottery tickets. I sense almost desperation, and an irrational belief that they will beat the odds.
I suppose a certain class of economist would argue that the money lost in lottery tickets and other gambling is rational because the losers are purchasing entertainment, thrills, or something like that. What are the users of Tradesports purchasing with their expenditures?
So basically, my observation comes down to the following: when looking at these graphs (predictions), we're essentiall asked to "follow the smart money," to use the phrase often thrown about. It's a great phrase. But I also often hear the smart money referred to, instead, as the greedy money, chasing a bubble, in other words what might be called the stupid money.
When looking at Tradesports, how do I decide which is the smart money, which is the stupid money.
In a poll, the goal is to collect together the predictions of individuals. "Person X, how will you vote in four weeks." So we've got a collection of personal predictions, and with enough, we acquire confidence, if performed correctly, that we've sampled the population. So we now have an estimate of how the population will act in four weeks.
In tradesports, we have a collection of opinions on how others might act in four weeks. A very different question. "What do you think?" versus "What do you think the others think?" So tradesports is a poll, of sorts, of people's opinions on how other people will act in four weeks.
So in summary, I simply don't know how the type of opinion gathering performed by Tradesports compares with polling, and I don't know how adding $$$ to the opinions, in a gambling sense, helps make the opinions more or less predictive.
I just don't know. That's all. But I'm not going to immediately decide that, because money is involved, smart or stupid, the results are more or less accurate than other forms of opinion gathering.
Are there are any good studies on the predictive power of Tradesports for elections?
Posted by: T.R. Elliott at October 9, 2006 07:59 AM
T.R., the studies I've seen suggest they do a pretty good job, certainly a better job than I can. See for example this paper by Wolfers and Zitzewitz.
Posted by: JDH at October 9, 2006 08:39 AM
TR in Las Vegas people are playing games where the odds favor the house but in these type of information market people are not betting against the odds. The individuals on both sides of the bet have an equal chance of wining.
This makes a significant difference and implies that your Las Vegas analogy is not accurate.
Posted by: spencer at October 9, 2006 09:54 AM
Spencer: I agree that the Las Vegas analogy of mine is poor when considering odds of Las Vegas in contrast to Tradesport. I was using Las Vegas more as an example of what I consider to be an irrational component to human behavior. I just picked up "Rationality Gone Awry" by Hugh Schwartz that I stumbled across in a used bookstore. Not sure how good it is. Maybe I'll learn something from it.
JDH: I'll look at the references. I'm not making a statement pro or con. Just raising questions. I'm not sure whether voting produces truth. In a lot of cases, e.g. democracy, it just picks one path out of many. Whether the different paths are more or less true is hard to say. Certainly scientific consensus is a form of voting, and a source of our truths. With Tradesport, we are looking at what might be called voting about the outcome of voting, with a financial incentive for getting the right answer. I understand what it means in a practical sense. But I will retain a certain degree of skepticism. That's all.
Posted by: T.R. Elliott at October 9, 2006 10:40 AM
Interesting that there's no visible bounce off the North Korea bomb. OToneH it reflects a failure of the Bush administration, but OTotherH we usually see new security threats as short-term favorable for the party in power.
Posted by: Hal at October 10, 2006 10:20 AM
You are exactly on point here. Those trying to gainsay the predictive value of open markets are engaging in exercises of futility, hewn from ignorance.
Also, this example makes me recall Poindexter's genius idea of allowing a "market" in terrorist event prediction. Genius, because it too would work, and work well.
Markets, those as unfettered as possible, tell the most Truth, they divine, succinctly, the "wisdom of crowds" and parrot not the prejudice of any.
Posted by: Mark E Hoffer at October 12, 2006 12:35 AM