Obamacare’s defenders have worked themselves into a tizzy, attacking the recent study published by McKinsey & Co., the world’s leading management consulting firm. The study indicated that 30 percent of surveyed employers were “definitely or probably” planning on discontinuing employer-sponsored health insurance after 2014.
Because McKinsey had refused to release details of the methodology used in their work, Democrats and left-of-center writers accused the company of having something to hide. A “keyed-in source says McKinsey is unlikely to release the survey materials because ‘it would be damaging to them,’” asserted Brian Beutler in Talking Points Memo. Senator Max Baucus (D., Mont.) wrote a letter to McKinsey demanding they release the survey’s methodology, with three House committees intending to follow suit.
Well, lo and behold, McKinsey decided to release the details: the full questionnaire used in their survey, along with a 206-page report detailing the survey’s complete results. Accompanying these details was a thoughtful discussion of the survey’s methodology, one that pops the balloon of those who tried to tar McKinsey as some sort of careless, partisan outfit. Despite reporting which implied that McKinsey wanted to distance itself from its own work, the company declared, “We stand by the integrity and methodology of the survey.”
The survey was funded by McKinsey
One of the silliest criticisms of the McKinsey study was that the company didn’t declare its funding source. It’s a silly criticism because the funding source doesn’t tell you anything about the truth value of the study. If ExxonMobil declares that 2 plus 2 equals 4, is arithmetic suddenly corrupt? At any rate, the company put the speculation to rest by making the entirely unsurprising disclosure that “the opinion survey was paid for entirely by McKinsey as part of its routine, proprietary research.”
The survey was conducted by Ipsos, a well-established opinion research firm
A more plausible criticism of the study was that McKinsey didn’t release the actual questions they used in their survey. In response, as I noted above, the company released the entire questionnaire, which was formulated by Ipsos, one of the largest market and opinion research firms in the world. Ipsos happens to be headquartered in that noted right-wing capital, Paris.
Contrary to those who suggested that the McKinsey survey did not adhere to well-accepted standards for public opinion research, McKinsey points out that “Ipsos fully adheres to the CASRO Code of Standards and Ethics for Survey Research, the ESOMAR International Code of Marketing and Social Research and the American Association of Public Opinion Research (AAPOR).”
The companies surveyed were representative of the broader economy
Another concern: was the survey of 1,329 employers representative of the broader economy, in terms of the size of the companies surveyed, their locations within the U.S., and the industries they represented? I had speculated that the survey may have been biased towards McKinsey clients, which are typically larger companies, implying that the survey actually underestimated the problem of employer dumping. However, McKinsey says, “Respondents were drawn from a panel of nearly 600,000 people maintained by Ipsos, not from McKinsey clients.” The detailed survey results make clear that respondents came from nearly every state, every industry, and every size category. Indeed, the survey contained a large number of smaller companies; 20.5% had less than 20 employees, and 51.6% had fewer than 100 employees.
Respondents were required to be either the “primary decision maker” or “have some influence in the decision-making process” for employee health benefits
Another criticism was: are the people participating in the survey the ones who actually make decisions about employee health benefits? Well, it turns out that the McKinsey survey required participants to have just such a role. Those participating in the survey were asked, in the very first question: “In your job, do you play a role in choosing which benefits your company provides to employees?” If respondents replied “do not play a role in this decision” or “not sure,” they were not allowed to continue answering the survey’s questions.
Of the 1,329 respondents who were allowed to continue on with the survey, 683 (51.4%) reported that they were the “primary decision maker,” while the remaining 646 (48.6%) said they “have some influence in the decision-making process.” As McKinsey writes in their discussion, “The top five reported respondent occupations were: owner, head of human resources, head of procurement, CEO/president, vice president of compensation or benefits director/manager.”
Most tellingly, primary decision makers were significantly more likely to drop employee health benefits: 36.5% of primary decision makers said they “definitely or probably” would drop benefits, compared to 22.4% of those who simply had some influence over the decision.
Respondents were informed of Obamacare’s exchange subsidies in a neutral, factual manner
Yet another criticism of the McKinsey work is that the study informed respondents of Obamacare’s exchange subsidies, which the law’s defenders asserted would bias the results. As I wrote previously, this criticism doesn’t make sense: “It would be one thing if McKinsey were, say, a Republican push-polling firm trying to scare people into voting against Obama. But McKinsey’s only agenda is to give its clients sound strategic and business advice.”
It turns out that McKinsey engaged in the entirely reasonable practice of providing neutral, factual information about the exchanges, before asking respondents what they were likely to do. “No information was offered on whether a particular provision or particular action would benefit or have an adverse impact on the employer or the employee. Nor was information provided on existing benefit rules, regulation, and taxes.”
The relevant survey questions, #41a and #41b, live up to that claim, describing Obamacare’s exchange subsidies in a straightforward manner. Contrary to premature allegations that the survey biased respondents in favor of employer dumping, question 41b directly states, “You would also pay a penalty of $2,000 per employee after the first 30 employees” for discontinuing employer-sponsored health insurance.
Employer surveys and economic simulations are different
One of the White House talking points against the McKinsey study was that the McKinsey study was an “outlier,” because its results are discordant with a handful of cherry-picked economic analyses which suggest that Obamacare will not result in widespread employer dumping.
McKinsey, in an effort to duck from Democratic projectiles, writes in its discussion that their survey “was not intended as a predictive economic analysis of the impact of the Affordable Care Act [and is therefore] not comparable to the healthcare research and analysis conducted by others such as the Congressional Budget Office, RAND and the Urban Institute.” However, as Paul Winfree has noted, the simulation models used in the studies more favorable to Obamacare are significantly flawed. And, as I have pointed out several times, there are a number of other studies that echo the McKinsey survey’s conclusions.
Only 29% of respondents “definitely or probably would not” drop employer-sponsored insurance
The details of the survey are even less flattering to Obamacare than had been previously reported. We knew that 30 percent (29.7%, to be exact) of respondents would “definitely or probably” drop employer-sponsored insurance. But the raw data shows that another 41.6% of respondents were undecided on the subject. Only 28.7% of respondents “definitely or probably would not” drop worker health coverage.
Among companies with under 50 employees, 36.6% would “definitely or probably” drop coverage, compared to 21.5% for companies with 500 or more workers. Among those who had low awareness of Obamacare’s provisions, 24.0% would definitely or probably drop coverage, compared to 54.0% of those with high awareness.
What’s next for McKinsey critics?
As far as I can tell, there’s little to nothing wrong with the way McKinsey conducted this survey. I’ll be interested to hear what Obamacare’s defenders say now. By falsely portraying McKinsey as some sort of right-wing conspirator, they set themselves up for disappointment.
UPDATE 1: Aaron Carroll put up a post alleging that the McKinsey survey was irredeemably biased, because it made no mention of the $2,000-per-employee fine for those who stop providing employer-sponsored coverage. Aaron has corrected his post, but says that his opinion of the survey is "unchanged."
UPDATE 2: Time's Kate Pickert, an early McKinsey critic, agrees that "McKinsey's methodology was sound," whereas Max Baucus is still breathing fire and brimstone, describing the study as "unjustifiable." The White House picks nits.
UPDATE 3: Joe Rago of the Wall Street Journal articulates it well: "The White House method is nonetheless to assail even disinterested analysts as dishonest or motivated by bad faith."
Elsewhere, Grace-Marie Turner points out some other data points on employer dumping:
A PriceWaterhouse Coopers survey of employers found that nearly half of all employers “indicated they were likely to change subsidies for employee medical coverage” thanks to the law.
An Associated Press story from last fall included quotes from a Deloitte consultant saying that “I don’t know if the intent was to find an exit strategy for providing benefits, but the bill as written provides the mechanism” and from the head of the American Benefits Council claiming that the law “could begin to dismantle the employer-based system.”
Former Tennessee Governor Phil Bredesen — a Democrat — in an op-ed said that Tennessee could drop coverage for its state employees, pay the $2,000 per employee penalty to the federal government, give their workers cash raises to compensate for the loss in health benefits, and still come out at least $146 million per year ahead.
- health insurance