L.A. Times Poll Scandal?
Plus: Recall Wreck Roundup
Up is bad. Down is bad. What's productivity to do? I'm quite willing to believe that "most economists predict that unemployment will remain almost unchanged at nearly 6 percent through the elections in November 2004." But shouldn't the NYT, after making this surprising assertion at the top of its lede business story, produce some actual evidence that most economists think this? Evidence like a survey of economists, or a blue-chip weighting--or at least quotes or views from a plural number of economists who make such a prediction? All the Times's Edmund Andrews gives us is a) his say-so and b) economist Brad DeLong. DeLong's quote:
I don't see where the demand is going to come from to produce a falling unemployment rate. ... Very few people are predicting real G.D.P. growth of more than 4 percent."
But even DeLong's argument is immediately undercut by Andrews himself, who writes, "More than a few economists say that the economy might expand at an annual rate of more than 5 percent in the final half of this year." [Emph. added] ... In case DeLong misinterprets the above: I'm not saying unemployment won't stay up near 6 per cent. (I've been wrong about unemployment before.) I'm not saying most economists don't predict unemployment will stay up near 6 percent. I'm saying the Times doesn't offer much, beyond blind faith in Edmund Andrews, to convince non-economists like me of either of these things. ... P.S.: I'd even argue the Times has a special burden in this regard because it has such an obvious institutional commitment to continued economic malaise. But I doubt Andrews' story meets even normal standards of evidence. ... Update: Gloouis Uchitelle states the full underlying NYT breakdown theory, in which rapidly rising productivity is bad because it means fewer workers are needed to meet a given level of demand. I can see that. But Uchitelle loses me when he switches to arguing that it's also bad if productivity fails to rise rapidly. Up-bad. Down-bad. Capitalism can't win with this guy. Do "most economists" agree? ... More: Even DeLong is unhappy with the NYT's reportorial laziness in a companion story to Andrews'. ... Second Day Update: Alert reader S.M. sends A hot-off-the-press AP story with some of the evidence Andrews could have used--"a panel of 35 professional forecasters" from the National Association of Business Economics" who predict robustt 4 to 4.5 percent growth but think
the unemployment rate would still be around 5.8 percent at the end of 2004.
The panel apparently considers this an upbeat end to the "jobless" aspect of the so-called "jobless recovery." It's just a very slow end. ... 1:47 A.M.
Kf traveled to the LAXMarriott Saturday to survey the distinguished hacks covering the GOP state convention. Hack consensus: Nobody believes the latest LAT poll, at least on the "recall Davis" question, where the Times shows a virtual tie. Other polls, apparently including private ones taken by Democratic interest groups trying to replicate the Times poll, show Davis in significantly worse shape. ... But there's disagreement on why the Times is full of it. 1) Some say the LAT sample included too many liberals. 2) Others say it included too many conservatives, citing McClintock's relatively strong showing. (Why would these conservatives vote against the recall? Because they were spooked by the previous widely-disbelieved Times poll showing Democrat Cruz Bustamante well ahead of Arnold Schwarzenegger! They'd rather keep Davis than get Cruz. That's the theory, anyway.) 3) Others (including kf mystery pollster "M") say the difference is probably the filter for "likely" voters and that the Times results will converge with other polls as Election Day nears (when likely voterhood becomes easier to measure). ... The Times' same-day defense of its own poll somewhat bizarrely adopts both explanations (1)and (2)--the Times had lots of liberals and lots of conservatives but fewer moderates! The paper rejects explanation (3), however. ... P.S.: Some sympathy was expressed for the poor LAT reporters who for reasons of institutional pride have to write their stories pretending that their paper's poll reflects actual reality. ... 10:13 P.M.
Weintraub picks up a contradiction in the van Lohuizen memo (see preceding item) that I missed: Schwarzengger's pollster says the "days when we could count on 20% or more of registered Democrats crossing over are long gone"-- and then he says Arnold will get "up to 20%" of the Democratic vote. "So which is it?" Weintraub asks. ... Were the Schwarzeneggerians lowballing their candidate's crossover appeal in the attempt to panic McClintock out of the race? Not necessarily. Both the Field and LAT polls show Schwarzenegger getting only 10 percent of Democratic votes. ... 7:18 P.M. Friday, September 12, 2003 B reaking Recall News: I had thought the "pressure-on-conservative-Tom-McClintock-to-get-out" story line was overdone and premature, with the California recall election 25 long days away. But tonight, Friday, the night before the state Republican convention in Los Angeles, Team Arnold has sent out what would seem to be a slightly panicky pollster's memo to "interested reporters," saying essentially (if I read it right) that Schwarzenegger will probably lose unless McClintock gets out.
Friday, September 12, 2003
B reaking Recall News: I had thought the "pressure-on-conservative-Tom-McClintock-to-get-out" story line was overdone and premature, with the California recall election 25 long days away. But tonight, Friday, the night before the state Republican convention in Los Angeles, Team Arnold has sent out what would seem to be a slightly panicky pollster's memo to "interested reporters," saying essentially (if I read it right) that Schwarzenegger will probably lose unless McClintock gets out.
The analysis, by Jan van Lohuizen of Voter/Consumer Research, is contained in an email from the Schwarzenegger campaign's Darrel Ng. It notes (and I'm quoting mainly the slightly panicky, negative bits):
... we continue to find that Democratic crossover voting is minimal. To the extent that Democrats do cross over, it is to Arnold; but there are only so many crossover votes to be had. Keep in mind that both in the Lungren and in the Simon elections more than 85% of registered Democrats voted for Davis. The days that we could count on 20% or more of registered Democrats crossing over are long gone. ... [snip]
5. A thorough analysis of the data so far show that Bustamante's share of the votes cast by registered Democrats is extremely unlikely to go below 60% and could go as high as 70%. The remaining 30% to 40% will vote for Arnold (up to 20%), vote for one of the candidates to the left of Bustamante (10% to 15%) or not vote at all in the election to replace (up to 5%).
Simple back of the envelope math shows that under these circumstances it is vital to maximize the share of Republicans voting for Arnold. Even under the most optimistic turnout scenarios Democrats will cast more ballots than Republicans. With more Democrats showing up at the polls and Bustamante getting a greater share of Democratic votes cast than Arnold getting of Republican votes cast, is a clear prescription for losing our best shot at winning the Governorship. ... [Emph. added.]
On the other hand, van Lohuizen says: