Krugman’s lazy apologia

Paul Krugman can’t stop attacking the McKinsey survey. His filed his latest apologia this morning (“McKinsey Pulls Back the Curtain”). It’s not his finest moment.

He dismisses the report as a mere “poll,” which is presumably a less reliable thing than the economic models that everyone else has been using. But why should a detailed survey—over 50 questions asked of over 1,300 respondents, mostly decision-makers—be less reliable than statistical extrapolations from not very comparable historical data?

Krugman quotes a stat from a Time reporter, Kate Pickert—not from the original document, curiously—with what he thinks is a clincher. The respondents didn’t know what they were talking about!

When asked how much their companies spend on medical and prescription drug benefits per full-time employee – something you might expect a health benefit pro to be intimately familiar with – 58.3% said they didn’t know.

Yes, that’s in the survey (question 15, for those scoring at home). The full question is actually more complicated, and might require a little spreadsheet work to figure out precisely:

Approximately what did your company spend on medical and prescription drug benefits last year per each active, full-time employee (averaged across both single employees and employees with spouses/families)?

I think it’s certain that the “don’t knows” know that they’re spending more than $2,000.

And the respondents weren’t just “health benefit pros.” Only about 10% were human resources execs; far more were owners and CEOs (question 2).

That aside, Krugman forgot to quote this observation from Pickert, from the same article:

Its poll of employers was not a GOP-funded shoddy survey meant to gin up criticism of Obamacare. Rather, the poll was long, complicated, conducted by a well-established polling firm and weighted to reflect the American business community as a whole. (This weighting helps compensate for the fact that the survey was conducted online, which can lead to problematic self-selection.) There are more than 50 questions in the survey, many of them multi-part questions, and the data collected is organized into easy to read cross tabs and breakdowns….That McKinsey initially seemed to release only the most headline-grabbing data subset is disappointing – the full results are full of all sorts of interesting nuggets like that, compared to very small businesses, about half as many large businesses would probably or definitely drop coverage post-reform. But its initial decision to keep its full data trove secret doesn’t mean the company’s motives were evil or partisan.

By contrast, Krugman’s motives, though not evil, sure look partisan.

10 Comments on “Krugman’s lazy apologia

  1. Kate Pickert, who Krugman quotes notes that the poll results are ambiguous:

    How have you heard your competitors will respond/How do you expect your competitors to respond to reform?

    Wait and see what other companies do: 27.7%
    Continue as-is or with minor changes: 27%
    Continue offering medical coverage but likely with significant changes: 24.3%
    Drop coverage altogether: 1.6%
    Drop coverage for some groups: 1.3%
    I don’t know: 31.4%


    How does that square with the other results? Polls are always very sensitive to the way questions are asked. Since there is a major caveat about the poll not being “predictive”, we should interpret the results with caution.

    Before looking at speculative data, it might be worth looking at the effect of the healthcare legislation on Massachusetts healthcare plan (which is somewhat similar to Obamacare) on the job market in that state.

  2. Dr. Rachel Nardin, over at the PNHP blog, on MassCare:

    The reform has done much more to increase access to insurance than to increase access to care. This may explain the Harvard School of Public Health poll finding that among those actually affected by the reform, more believed the reform had hurt than helped the uninsured.

    […] the reform has done nothing to address health care costs, which continue to escalate at alarming rates. Although the governor’s hope that accountable care organizations and other payment reforms will solve this problem, this is based more on faith than solid policy evidence.

  3. A while back, Doug, you predicted – or perhaps hoped with some reason – that the liberal-left would break away from Obama and the Dems and help to open a space for more radical politics. You cited widespread lefty disillusionment with JFK/LBJ as the ideological moment before the “Sixties” and “Seventies” commenced. I hope you are right, but somewhat doubtful since our intellectual/media climate is very different, not to mention the vast improvements in the owner-operating class’s PR and co-optation capabilities. But there are examples of cracks – like Glenn Greenwald. Maybe Cornell West (maybe, maybe).

    No doubt, Professor Krugman is trying to shore up a firewall (aka, the acceptable leftward limit of dissent, particularly against the Democratic Party and Obama) at The Almighty Times. You can tell his game: GOP is crazy and destructive, whilst Obama and the Dems are merely foolhardy and gullable. As if there were some yawning gulf between the “two” parties and their approach-“es” to economics. Uh huh.

  4. I spent some time playing with the subsidy calculator for Obamacare. The incentive for businesses to drop coverage, pay the fine, and perhaps even make up (part of) the shortfall through higher pay is enormous. Of course, the main atraction of being an employee at all is the health care coverage. Without that, you may as well turn your payroll into independent contractors, which allows the employee to deduct all sorts of things and lower its tax bill. There could well be a double whammy of huge subsidies costs and lowered tax income when Obamacare starts. And forget about tinkering with the program with the Reps controlling the House!

  5. The answer is: we’ll find out in 2014.

    McKinsey did initially say that their study predicted one third of employers would drop coverage. No other study had come up with that.

    After being pressed, they admitted their “study” was actually a survey with the results saying:

    “Given this information, how likely do you think your company would be to discontinue employee health coverage?

    Definitely would: 9.2%
    Probably would: 20.5%
    May or may not: 41.6%
    Probably would not: 18.2%
    Definitely would not: 10.5%

    Only 9.2 percent “definitely would” not 33 percent, so McKinsey was in fact being misleading.

  6. Rube Goldbergs only function when the interlocking pieces work as planned. Otherwise, most people won’t be able to see what the engineers intended. But the entire justification for such baroque contraptions is that they perform. So when the engineers, and their apologists, plea incessantly for patience, or argue that their critics are either too ignorant or too malignant to appreciate how the apparatus works, I raise the odds of failure.

  7. I’m assuming you’re also the Brad on Facebook I responded to. Here’s what I said there:

    That’s not a push poll at all. They dispassionately informed respondents of their options, since many of the respondents didn’t know. They’re trying to get some sense of how business will respond as 2014 approaches. Contrast that with the Wikipedia def of a push poll:

    A push poll is a political campaign technique in which an individual or organization attempts to influence or alter the view of respondents under the guise of conducting a poll. In a push poll, large numbers of respondents are contacted, and little or no effort is made to collect and analyze response data. Instead, the push poll is a form of telemarketing-based propaganda and rumor mongering, masquerading as a poll. Push polls may rely on innuendo or knowledge gleaned from opposition research on an opponent. They are generally viewed as a form ofnegative campaigning.

    The point of a push poll is to lie to people under the guise of research – research whose results aren’t reported. McKinsey lied to no one and made every effort to disseminate the response.

    Also, this from Stuart Rothenberg (

    I probably sound like a broken record, but I’ve had it with people who still don’t know the difference between political polls and advocacy telephone calls.

    Polls are methodologically rigorous public opinion surveys of generally 500 to 1,000 people intended to learn about and measure voters’ opinions and test possible campaign messages. Advocacy telephone calls, on the other hand, are made to tens of thousands of people and are intended to create or change opinion.

    It really isn’t all that difficult to understand the difference. Any high school graduate ought to be able to figure it out. That’s why I find it so frustrating that journalists and politicians don’t make distinctions. This year again, a bipartisan group of Members of Congress has introduced legislation to “prevent push polls.”

    The term “push poll” never should have entered our lexicon, since it does nothing but confuse two very different and totally unrelated uses of the telephone.

    As I have argued every year for the past five and apparently will have to continue doing until I have taken my last breath, push polls are really advocacy calls aimed at thousands of recipients. They are like television or radio ads, except they are delivered over the telephone. They seek to convey positive or negative information to influence a voter’s final vote decision.

    Advocacy calls are not, in any shape or form, public opinion surveys.

    Phone banks established to deliver advocacy messages are designed to make large numbers of calls and have no interest in getting responses from those called, except possibly for identifying supporters, opponents or the undecided for the purpose of deciding whether those people should receive still more advocacy or get-out-the-vote calls.
    Hiring a firm to do 500 or 800 advocacy calls in a Congressional district or state — where thousands or tens of thousands or even millions of voters are going to go to the polls — would be idiotic, since it wouldn’t accomplish what advocacy calls are trying to do, which is change opinion.

    Why do I get all hot and bothered about this? Because referring to advocacy calls as push polls adds to public cynicism and, more importantly, discredits a legitimate survey research approach.

    The new legislative push really is no more than an updated effort. Two years ago, Wisconsin Rep. Tom Petri (R) introduced the ill-named Push Poll Disclosure Act of 2005, and he has done so again in the 110th Congress.
    Petri may not be intentionally adding to the confusion between advocacy calls and public opinion surveys, but he certainly is doing so by constantly referring to push polls and even putting those words in the title of the bill.
    Petri seems concerned about telephone calls that include very negative information about a candidate for office. This kind of information can be part of an advocacy telephone call or part of a legitimate poll. When they are in a real survey, they are known as “push questions,” because they seek to measure which questions actually push voter sentiment and which issues can be used by a candidate to win a race.

    Push questions are not the same thing as push polls. Push questions, which are included in a survey of only 500 to 1,000 respondents, are a legitimate part of a public opinion poll that seeks to test effective messages.
    My guess is that someone who didn’t know very much about survey research — I’m putting my money on some twentysomething who worked in the media — heard about push questions and incorrectly confused them with advocacy calls, creating the illogical term “push polls.”

    Serious polls can include push questions that contain some explosive or even incorrect information, but that doesn’t make them advocacy calls. Testing possible messages is a legitimate survey research function, and as long as the question is asked of a small sample and seeks to get a response to know whether the issue is useful in an election, it really doesn’t matter how negative the message is.

    Congress certainly ought not to interfere with legitimate pollsters’ (and I’m including incompetent pollsters here, too) efforts to accumulate data or test messages, no matter how negative they might be.

    If Congress wants to require a disclaimer on advocacy calls, the way it has done on TV spots, that’s certainly reasonable. But messing with polls would be a terrible idea.

    Wisely, in defining the problem, the Petri proposal seems to focus on the number of calls made, which is an important distinction between polls and advocacy calls. But throwing around phrases like push polls doesn’t make me very confident that Members of Congress know what they are talking about.

    A first step toward sounding smart might be to change the language of any bill that seeks to regulate advocacy telephone calls. It also might be nice if members of the media banned “push polls” from their vocabulary.

  8. Oh good grief. Rothenberg is full of shit. And you’re dabbling it that too. You damn well know that question wording can make all the difference in responses. Good survey research will seek to avoid leading or pushing responses. Partisan survey research won’t. Advocacy can be putting TV/radio spots on my telephone. Advocacy can also stick a few zinger questions in an otherwise methodologically sound poll so as to use the poll results to influence opinion or media spin. I don’t know if McKinsey intentionally was trying to enter the fray but as survey research its all ice skates on an pig. You can put ’em on…but why?

  9. Pingback: In the Belly of the Beast: 6/23/11: Feeding Resistance: Food Not Bombs Members Arrested in Orlando for Serving Meals Without a Permit | A New Worlds in Birth

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: