Did McKinsey Repudiate their Employer Health Care Survey?

As I scrolled through the RSS reader this morning, my attention was grabbed by this guest-post on Jonathan Cohn's blog.  The author is Jonathan Gruber, health care consultant to the wonk-stars, and the subject is defending ObamaCare from a recent report by the House Committee on Oversight and Reform.

I won't get into the substance of the disputes right now, since most of it strikes me as Olympic-class quibbling, on both sides, over whether tax breaks count as spending.  But one claim rather jumped out at me:

Claim #3: The ACA Will Lead to a Large Erosion of Employer-Sponsored Insurance

To make this claim the committee leans on a report from McKinsey so flawed that the firm itself ultimately repudiated it. The committee also presents some calculations of financial incentives that might lead employers to drop health coverage. But, in so doing, they ignore a major incentive for employers to keep (or start) offering coverage: the free rider penalty.

I found this rather startling.  I thought I had followed the debate over the McKinsey survey (which showed large numbers of employers saying they'd drop their coverage after ObamaCare went into effect) rather closely, and I didn't remember them repudiating it.  I asked the Official Blog Spouse, who had also followed the debate rather closely.  He had seen the same article, and been puzzled by the same thing.

"But I didn't click on the links yet," he said.  I had.  They also didn't show McKinsey repudiating the survey.  Rather, they referenced Brian Beutler's story in which several anonymous McKinsey employees of uncertain seniority or involvement in healthcare and/or HR projects, bashing the study.

"This particular survey wasn't designed in away that would allow it to be peer review published or cited academically," said one source familiar with the controversy.

All sources were granted anonymity, in order to be able to speak candidly about the controversy.

Reached for comment today, a McKinsey spokesperson once again declined to release the survey materials, or to comment beyond saying that, for the moment, McKinsey will let the study speak for itself. However, McKinsey notes that the survey is only one indicator of employers' potential future actions -- that the conclusions remain uncertain and employers' future decisions will ultimately depend on numerous variables. The three authors of the report were not immediately available for comment.

Another keyed-in source says McKinsey is unlikely to release the survey materials because "it would be damaging to them."

Both sources disagree with the results of the survey, which was devised by consultants without particular expertise in this area, not by the firm's health experts.

A third source speculates that the firm may have reached its outlying conclusion by basing its questions on the firm's own advice to clients on how best to arbitrage the new reforms. Specifically, under the law, employers could devise their benefits packages in ways that makes them unappealing to lower-income employees, who would then have to enter the exchanges. Though TPM could not confirm this, the conclusion is supported by a disclosure within the McKinsey study itself.

McKinsey has about 9,000 employees around the world.  I am sure more than three of them were unhappy that the company had released a report critical of ObamaCare, since hundreds of them donated a total of $207,296 to Obama in the 2008 campaign cycle.  But this hardly proves anything, especially since at least one of these insiders was clearly not very well plugged in:  ten days later, McKinsey released their survey materials.  And they were hardly "damaging".  Kate Pickert, one of the reporters who first raised questions, wrote:

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.
It's obviously possible to argue about what the results mean--and in fact, Jonathan Cohn did, at length.  But it would be extremely funny to describe three griping employees as "the firm itself ultimately repudiated it".

So where was the repudiation?  Googling "McKinsey repudiates health care survey" turns up only this:

In February 2011, McKinsey & Company commissioned a survey of 1,329 U.S. private sector employers to measure their attitudes about healthcare reform. The opinion survey was paid for entirely by McKinsey as part of its routine, proprietary research.

We stand by the integrity and methodology of the survey.
I've emailed McKinsey for clarification, but haven't heard back yet.

Unless both the Official Blog Spouse and I are misremembering--entirely possible, of course--this seems like a very strange way to characterize the study.  Saying that a firm has repudiated something is a strong factual claim with a fairly narrow meaning, which does not include "overhyped some of the survey's findings in their press release" or "had a few employees who disagreed with how the survey was conducted". (Especially since the complaints seem slightly weird.  As far as I know, surveys are not usually designed to undergo "peer review".  And since this appears to be an HR survey, why would the health care practice have necessarily been involved?)  Jonathan Gruber is a widely respected health care economist; I find it surprising that he would make such a claim without solid support.  But then why can't I find it? And why didn't he link to it?

Update:  McKinsey confirms that their June 20th release was their last word on the subject.  In which case, I think that TNR should probably correct the story.

Update II:  And they have!  Jonathan Cohn emails to say that they have corrected as follows:

To make this claim the committee leans on a flawed report from McKinsey that contradicted the projections of CBO and other independent forecasters. McKinsey insiders criticized that study anonymously and the firm itself later said the report was "not intended as a predictive economic analysis."*