Search






Jeff's Amazon.com Wish List

Archive Calendar

November 2024
M T W T F S S
 123
45678910
11121314151617
18192021222324
252627282930  

Archives

Polls and barometers and markets, oh my! [Karl]

Yesterday, I noted that the current polling trends were not showing any sign yet of a significant “bounce’” from Barack Obama’s world tour.  At almost the same time (accounting for time zones) TNR’s Jonathan Chait was noticing the same tightening of the race, under the title “Maybe Pundits Need to Spend More, Not Less, Time Following the Polls,“ juxtaposing the current trend with this lede:

I see that John has already beaten me to this, but I continue to be amazed at how many people seem to believe that Barack Obama is crushing John McCain. Kevin Drum writes, “McCain is pretty obviously doomed this year.” Howard Fineman writes, “You can’t make up how bad things are going for McCain.”

I would add Time magazine’s Michael Grunwald to the list of doomsayers.  The establishment media is likely fully aware of what the horserace numbers are and probably remains too focused on them.  So why does there seem to be an assumption that John McCain is failing badly?  One obvious explanation would be the obvious pro-Obama tilt of most of the establishment media.  However, there is likely more to it than that, which is worth exploring.

For all of the media’s focus on polling data, the punditocracy is all too aware that these polls are merely snapshots and that November is far, far away in political time.  Considering that a root cause of poll mania is a desire on the part of the press to prognosticate, as opposed to simply analyzing poll results, pundits tend to bring additional information into the mix.

A common source — and the one to which I most often fall prey when I slip out of analytical mode — is history.  John Fund is probably not the only person in the press seriously considering that this cycle may resemble 1980, with Obama playing the role of Reagan.  In this mindset, the race may remain close for some time, but that ultimately late deciders will all break to Obama once they see him as a credible alternative — as happened with Reagan after debating Carter (this is why I am interested in whether Obama’s trip could reverse the current narrowing trend of the race).  Camp McCain has tried to paint Obama as Carter II — there are substantive reasons to support this, but Obama may be too charismatic (and McCain so uncharismatic) for this tack to be successful.

Some may move beyond mere historical analogy to creating a “barometer” or model for predicitng the outcome of elections with historical data.  Tom Maguire recently discussed the two best-known of these, developed by Alan Abramowitz (who predicts Obama wins in a blowout) and Ray Fair.  Maguire notes that Fair’s model overestimated Bush’s 2004 victory; I would add that 1992 was an even worse year.

Abramowitz’s “electoral barometer” weighs: the approval rating of the incumbent president; the economy’s economic growth rate; and whether the president’s party has controlled the White House for two terms (the “time for a change” factor).  It has correctly forecast the winner of the popular vote in 14 out of 15 postwar presidential elections.  However, as Kevin Drum has noted, the “time for a change” factor alone correctly predicts 12 out of 15 postwar presidential elections. (This is, incidentally, one reason I like the 16-year cycle of “change” elections as a guideline, though I would never state it as a formula.)  Furthermore, the “barometer” predicts a winner, but a given reading does not necessarily correlate to a particular popular vote margin.

More broadly, these types of models suffer from the same sorts of flaws one sees with simple programs used to predict the direction of stock prices.  The variables may backtest well, but there are not enough real world results in the sample set to validate the predictive ability of the formula.

Speaking of markets, the increasing visibility of political prediction markets may also influence how pundits evaluate the state of the campaign.  After all, we tend to believe in the efficiency of markets to pool available information.  In addition, there is research suggesting that these markets in fact predict election outcomes better than polls this far away from Election Day.

There is also recent research showing that polls may be better predictors than the election prediction markets, if one accounts for voters tending to be more strongly anti-incumbent 3 1/2 years into an administration than they are on Election Day.  However, there is again the problem of having too few elections in the sample set to confidently conclude this research is correct.  Moreover, Robert Erickson — one of the researchers on this point — is conducting new research suggesting that the polls are slow to reflect economic information, which he believes would give Obama an edge in this cycle.

At least, that was Erickson’s belief at the beginning of this month.  It could be that the US economy will look better that thought when Q2 GDP numbers are announced.  Indeed, to compound matters, the odds of a 2008 US recession have fallen dramatically on the prediction market at InTrade in recent months and weeks.  If voters end up seeing the economy more favorably by Election Day, it could redound to McCain’s benefit. 

On the other hand, the economic optimists could be wrong.

All of which is to say that when pundits evaluate the current general election campaign, their views are not merely a product of bias or the latest poll, but generalizations from — or formulizations of — history, as well evaluations of the candidates’ respective strengths and weaknesses.  To the degree that even this sort of common sense forecasting itself depends on predictions about factors as influential as the economy, a shaker of salt and more than a modicum of modesty would be in order.

17 Replies to “Polls and barometers and markets, oh my! [Karl]”

  1. happyfeet says:

    I think the coverage being out of synch with the polling is just to hamper McCain fundraising. The media will have to tighten up the race come September or risk suppressing Baracky’s turnout. Obama Girl and all that.

  2. happyfeet says:

    The challenge that presents for Sen. Obama is illustrated by a second question. When voters were asked whether they could identify with the background and values of the two candidates, 58% said they could identify with Sen. McCain on that account, while 47% said the same of Sen. Obama. More than four in 10 said the Democratic contender doesn’t have values and a background they can identify with.*

  3. D. Travers says:

    Here’s a radical idea… Why don’t we just stop polling altogether? What purpose does it serve, other than for the campaigns themselves?

    I fear that in some ways polls can distort the process, and become a leading, not lagging, indicator. If everyone is inundated with poll data showing Candidate X is ahead, his supporters are more likely to stay home since he’s “safe” and Candidate Y’s supporters are more likely to go out to the polls.

    But forget all that. What purpose does it serve to report on -daily- polling results, ad nauseum? It’s just more noise drowning out what should be substantive conversations about the candidates positions.

  4. Obama’s World Trip A Flop…

    Obama was supposed to get a bit of a bounce from the fawning media coverage surrounding his whirlwind gaffe tour (is everybody’s legs tingling?). But he is not getting anything at all (just check out Gallup and Rasmussen). The fact is the man is …

  5. The Lost Dog says:

    Great post as usual.

    But there are two things we should keep in mind.

    First,I have to agree with D. Travers about polls becoming a distortion factor. I can’t stand polls, because the drive bys now use them as a driver instead of an indicater. If you have ever been polled, you know that there is absolutely no room for “shading” when you answer a question. Many of the questions are loaded, and far too many polls are completely out of balance between Dems and Repubs. And guess which way they are slanted?

    And second, it doesn’t matter about the economy. The MSM has decreed that a better economy than Clinton ever had was the worst economy since the great depression, and no matter what, they will continue to do so. I haven’t seen one article saying that we would be “Smokin'” if not for the price of oil (thank you, Democrats).

    I have nothing but contempt for the press and the Democrats who, in 2000 were having a hissy fit at Bush for “talking down the economy”, even though it was obvious to a blind person that the tech bubble had burst, and the whole ball of string was unwinding. And BTW, no one will ever convince me that Clinton didn’t know exactly what would happen when he went after MicroSoft.

    “But don’t you Repubs dare to state the obvious!” was the cry of the day, as I recall. Today it is a constant (and false) drumbeat from the drive bys and the Democrats that the economy stinks.

    And now? A slow (but still growing) economy is the worst since blah blah blah.

    We’re screwed. The truth has very little outlet now.

    The left owns the conversation, and all I can do is sit here and fume, and tell the sheep that some facts might do wonders for their intellect.

    As if they were listening…

  6. ajacksonian says:

    There is one thing that can be confidently given over the last 40 years: America prefers harshly divided government. This is an effect of voters not turning out to vote, and that set of long term relationships will drive this year, as well. Some of this is America returning to its pre-WWI trend of divided government, but the drivers are different today than then. One of the harder ones to examine is that when the top 1% of earners (the super-rich) have less than 9% of the wealth overall, there is a steep polarization of US politics. When the rich have trouble getting richer, which regressive taxation causes, the rest of America sees their opportunity as capped. Instead of being ‘fairer’ such taxes may change voting patterns to limit the amount of regressive taxation that can take place through gridlock.

    Indeed, gridlock is the aim of not voting: causing those left to be ever more partisan and diametrically opposed to each other. Instead of a worrying point, this is a net positive for not voting as it marginalizes how much government can add to itself over time. That is because government does not generate wealth, and limiting its size while no longer supported by any political party, is now something done via the franchise right and not utilizing it. If that regression towards the mean analysis is correct, then whoever wins will have very little time before Congress turns against them in at least one house. An Obama Presidency would quickly effect the bottom line and cause a huge increase in regressive taxation which would polarize the remaining electorate. Similarly a McCain Presidency can expect that the Democratic majority will remain and take hard positions against him beyond anything ‘bipartisanship’ can paper over.

    I’m betting on gridlock: it has been the source of what has moved politics into shrill irrelevency. If either party could pick up just 10% of the disaffected, they would have a landslide majority. They can’t and neither party can reach to that plurality today… and my bet is that it becomes a majority this year. And that is not a good place for democracy to be in.

  7. mojo says:

    Killed and gutted a white goat today.

    The auspices are good.

  8. B Moe says:

    Why don’t we just stop polling altogether? What purpose does it serve, other than for the campaigns themselves?

    What would the MSM talk about then?

  9. Karl says:

    Polling is not the problem per se, if you’re jsut looking at it as you should — as a snapshot. The problem, as alluded to above, is the tendency to predict, and the potential of a “Heisenerg principle” of polling — the study itself affects the results.

    I don’t think this happens re the economy, ultimately. Yes, the MSM slants the economic news a particular way. But people can read a gas pump, their bills, their real estate estimate, etc. They know their circumstances, which drives things more than a general perception of others (not that people necessarily are honest about this with pollsters). The effect is more insidious with subjects on which people do not have their own direct experience. But there is no magic wand that will stop the MSM from doing what it does (just the market), so it is something all campaigns have to live with, just like the weather.

  10. The Lost Dog says:

    As far as markets go, I was delighted yesterday to see that the NYT is getting pantsed(?) by their own arrogance.

    I always thought that “Pinch” was a monkey playing football, and the NYT’s bottom line is finally proving me right.

    (sigh) It’s about time. But I still like buying my New York Post every day, and glancing at the NYT right next to it, and thinking: “Eat me, Pinch!”

    That litltle ritual always brightens my day.

  11. TmjUtah says:

    Data point for you –

    This past week, local Utah media reported that state highway contracts now under construction may face unplanned interruptions due to costs associated with petroleum products. Diesel hurts a lot, but the killer turns out to be humble asphalt. Private jobs… well, I can’t comment on that.

    Local and state entities across the state are reporting budget gymnastics to cover fuel charges; some school districts have doubled the distance a child must live from their school to qualify for busing.

    Utah has enjoyed a sweet decade and change riding the economy. Our government spending has increased at a rate somewhat south of the ongoing obscenity that is California’s public expenditures… but The Teachers’ Best Buddy Huntsman got into the game late. There is certainly no hint of the usual surplus the hacks in Salt Lake have grown used to divvying up for their benefit.

    Mark me down on the hard bottom side for an outlook. Hard Bottom, with debris.

    I wonder what the cloak room gossip is like up on Capitol Hill? I see, at the very least, a redux of the 1992 “Throw The Bum’s OUT!” on the very near horizon. The fiscal conduct of our government is criminal, and has been for decades, across the majorities of both parties. When the house is cold, the larder empty, and the kids are crying, somebody pays.

    I don’t know if the people will wait for elections…

  12. Fresh Air says:

    Karl–

    Surely you remember when the media tried (successfully, in part) to paint George H.W. Bush as an evil rich white guy–so out of touch he didn’t know what a supermarket scanner was. They definitely try to make people feel bad whenever Republicans are in the White House. It’s become standard M.O. ever since.

    But I have a question: What percentage of the people polled realize the Democrats are running Congress? This is deadly serious. How can a Congress with a 9 percent approval rating be favored in generic match-ups? Yes, I’m afraid the people of this country are really that stupid.

    I blame their parents.

  13. Fresh Air says:

    BECAUSE OF THE PREGNANCY!

  14. Salt Lick says:

    Obama is not going to win this election. In the last years, I’ve listened to these MSM assholes tell me:

    1) The world would end at Y2K.
    2) Bird virus would kill billions.
    3) SAR virus would kill billions.
    4) If we invaded Iraq, a billion Muslims would rise in the Arab Street and we’d have WWIII.
    5) We’d lose thousands of soldiers before we ever got to Baghdad, and once we got there the urban fighting would be like Mogadishu squared.
    6) The year after Katrina would see 9 category 5 hurricanes.
    7) 2007 — Iraq was going down in civil war.
    6) Etc, etc, etc.

    We still have 3 months before Election Day. Obama and McCain are presently dead even. The more the American people see the inexperienced, untested little Affirmative Action mascot named Barry Obama, the more they are going to doubt they can hand their future over to him.

    Fuck the MSM, who’s job it is to create anxiety and tension so they can charge greater advertising rates. I will not let them make me doubt my countrymen.

  15. MarkD says:

    The Dems won’t even allow a vote on drilling in ANWR because they know it will pass. Any GOP candidate who runs on this will win.

  16. Karl says:

    Fresh Air —

    No argument on the general point about the media bias. But the polls tend to show that increasing numbers of people are aware of the bias.

    But I have a question: What percentage of the people polled realize the Democrats are running Congress? This is deadly serious. How can a Congress with a 9 percent approval rating be favored in generic match-ups? Yes, I’m afraid the people of this country are really that stupid.

    Regardless of which party controls Congress, people tend to like their Rep. and Senators more than they like Congress in general. Right now there are more Dems in Congress, so Dems do better generically. Indeed, if you look iat the internals on the generic questions you find more people approve of Dems in Congress than GOPers in Congress (which is not surprising given that a lot of conservatives are also disenchanted with the GOP). Moreover, the generic Congress number is similar to the Party ID number — a general inclination untethered to a concrete choice. Similar factors explain why the McCain-Obama numbers are where they are viz-a-viz the generic numbers.

    That’s what makes the generic numbers useful — they are a baseline from which one can gauge what happens when actual names get attached. From another angle, the generic Congress numbers reflect how badly Pres. Bush & the GOP members have managed to damage the GOP brand, in part with things I support, in part with things I oppose.

  17. households now owe 100% GDP – as much as the entire US economy can produce in a year

Comments are closed.