Underwhelming New Ads from Obama, McCain

TPM has three new TV ads out today, two from the Obama campaign and one from McCain. Honestly, I think they're all pretty bad.

The first Obama ad, called "Still", criticizes John McCain as, basically, an old fart who can't use the internet or email and favors tax cuts for corporations. The ad's graphics are kind of crappy, so I assume it's web only. At least I hope for the Obama campaign's sake that it's web only, because I think it would be unwise to put this ad on the air.

I think the "can't send an email" thing mostly bothers young upscale liberals, not swing voters. Dissing McCain for not using email might actually backfire with seniors. And very unwisely, the ad pronounces that "things have changed in the last 26 years, but McCain hasn't." Certainly that undercuts another Obama message, which is that McCain sold out his "Maverick" principles to get the Republican nomination. In fact, "things have changed in the last 26 years, but McCain hasn't" sounds like something the McCain campaign would say to frame McCain as a Reagan Republican, not a Bush Republican.

The second ad ("Real Change") features Barack Obama talking to the camera, saying he's the candidate who will really bring change. This messaging is OK, but I'm not sure it's likely to have much impact; after all, McCain is out there saying essentially the same thing. Since the candidates have similar favorable ratings and are both associated with "change" or "reform" in the voters' minds, I doubt that either one saying "change" over and over is likely to matter.

The last ad ("Disrespectful") criticizes the Obama campaign for its own criticisms of Sarah Palin. Generally, I think whining about attacks against Sarah Palin is a bad idea, because it undercuts her image as a very strong woman who is unfazed by criticism. Mercifully, the ad does not claim that Obama's "lipstick on a pig" remarks were an attack on Palin.

Read More......

Sarah Palin, White Women Voters, and Abortion

Noam Scheiber over at the New Republic notes that, according to the most recent ABC News poll, John McCain has improved his standing among white women voters by 11 points since putting Sarah Palin on the ticket. Scheiber, however, caveats:

Once Palin's extreme anti-abortion positions become known (no exceptions for rape or incest!) and her halo wears off, I fully expect Obama to at least pull even among white women, which would give him a couple-point cushion in this poll.

Implicit here is an assumption that Palin's position on abortion particularly hurts her with women, or at least white women. Why would this be? One reason would be if women were more likely than men to support abortion rights; however, that's not true. Another possiblity is that pro-choice women are less willing to support pro-life (or strongly pro-life) candidates than pro-choice men are; i.e., that there's no aggregate difference in policy preference between men and women, but there is an aggregate difference in intensity of preference. That's plausible, but I wouldn't assume it to be so without evidence. Can anyone point me to polling data on that?

Read More......

What's the deal with Hillary on Intrade?

Warning: wonky post ahead...

I'm a fan of markets generally, and so I'm inclined to like the Intrade prediction market as a way to forecast the election. As a futures market, it asks the right question ("Who will win the election?") while polls ask the wrong one ("Who would you vote for, if the election were held today?"). Polls tend to overstate the importance of the most recent news cycle and are subject to systematic biases, most notably the "convention bounces" that follow the DNC and the RNC. There is debate about whether prediction markets are more accurate than polls; it's clear that they are better than unadjusted polls, but it may be equally good (or better) to adjust polls for known biases, which is the approach taken by the folks at FiveThirtyEight.

Here's a little background about Intrade: contracts on Intrade trade on a scale from 0.0 to 100.0. Each point represents ten cents. If you buy an Obama contract at 40.0, that means you pay $4 today to win $10 if he wins the election. Conversely, if you sell at 40.0, you receive $4 today and agree to pay $10 if Obama wins. The higher a number a candidate trades at, the more likely he is deemed to win the election. For note, the market prices are supposed to track percentage likelihood of winning, not the likely percentage of the vote the candidate will garner.

I see two key peculiarities in the recent activity on the Intrade market. Peculiarity one: the candidates have often been trading at a sum significantly over 100.0. Since the prices represent a candidate's odds of winning, the current trading prices for all candidates should add to 100.0, or a figure very close to it. Howeer, as of 1:37 PM today, here are the most recent prices for the three candidates that had contracts traded within the last day:

52.9 B. Obama
46.0 J. McCain
03.4 H. Clinton

That's a total of 102.3. For note, the bid prices also sum to more than 100 (currently 101.6). That means an arbitrageur could come in, receive $10.16 for selling contracts on all three candidates at the current bidding prices, and know he'd only have to pay out $10.00 on election day (because only one candidate can win), for a sure profit of $0.16 per contract.

So, why isn't arbitrage pushing the markets to an equilibrium where the candidate prices equal 100.0 in total (or, more specifically, one where the sum of the bid prices is always less than 100.0, and the sum of the asking prices is always more than 100.0)? One possible answer is that markets are pretty thinly traded, and there wouldn't be a lot of money to be made with an arbitrage strategy. Unfortunately, that points toward overall inefficiency in the market; because the amounts of money involved are not very high, time itself may impose significant transaction costs that interfere with accurate price movements.

Astute readers may have already spotted peculiarity two: that is, Hillary Clinton continues to trade at a price representing a non-trivial likelihood of winning the presidency. For note, there are a lot of candidates with a last trading price greater than 0.1: Mike Huckabee, for example, last traded at 0.2. However, Huckabee contracts are no longer actively traded, and that most likely reflects a final trade well before John McCain accepted the nomination at the RNC. Hillary contracts, however, retain moderate trading activity; indeed, just today, she has traded as low as 3.1 and as high as 4.1.

Why is her contract trading so high? The contract could be acting, morbidly, as a future on the likelihood of Barack Obama's ceasing to be the Democratic Party nominee for President, either through death or removal. However, there are several key problems with that interpretation:

1. Under this theory, the contract should fall in price with each day that Obama remains the Democratic nominee and the election draws nearer; however, the price had been relatively flat since Hillary suspended her campaign, and actually spiked in the last week. At the beginning of the month, she was trading below 1.5.
2. Such trading assumes that Hillary would be the replacement nominee for Obama. Now that Joe Biden is the VP candidate, he seems like a much more likely choice. However, Biden's contract is traded much more thinly than Hillary's (with no trades at all today) and last traded below 1.0.
3. 3.4 seems like an implausibly high price under this kind of thinking. That's especially true because the contract represents a three-part, composite probability: (1) that Obama ceases to be the nominee, (2) that Hillary is his replacement, and (3) that Hillary defeats McCain in the election. For the conjoint probability of all three events to be 3.4%, the probability of part (1) must exceed 3.4%, assuming that the conditional probabilities (2|1) and (3|1,2) do not equal 100%.
4. There is no similar actively-traded contract (say, on Mitt Romney) to reflect the possibility that John McCain ceases to be the Republican nominee. However, point (4) is weaker evidence, as it is plausible for the market to hold that the probability of a Republican win after McCain's removal as nominee is close to zero; as such, the possibility of McCain's death or removal would be included in the Obama contract price, not some other Republican's price.

So, why else could it be that this contract is still trading, and at such a high price? Are active Hillary supporters artificially driving the price up, either out of a desire to make their candidate look good or a mistaken belief that she might somehow still win? Because the Hillary trading is active but not heavy (much lighter than trading in Obama and McCain) some dedicated supporters could do this pretty cheaply. And, given the efficiency issues discussed above, they could probably do it unmolested by arbitrageurs who would otherwise bring the price down to reality.

One bonus for Intrade is that this explanation of peculiarity two might actually reduce the importance of peculiarity one. The high Hillary price is non-market, but it's not worth the time involved for arbitrageurs to exploit it. However, traders of McCain and Obama contracts are correctly disregarding the recent Hillary trades and treating her price as properly zero. Once you exclude prices on Hillary contracts, the McCain and Obama contracts are behaving as expected: that is, their total bid prices are less than 100.0, and their total ask prices are more than 100.0.

Read More......