Facing South

Investigation: Business bankrolls study claiming job losses from Employee Free Choice Act

Anne Layne-Farrar.jpgThe battle over the Employee Free Choice Act kicked off this week with versions of the bill being introduced in the House and Senate -- and with it a flurry of claims about what the landmark labor legislation would mean for the country.

One report in particular has been seized on by business groups and the media as proof that the bill would be an economic disaster -- a study released last week which claims that the EFCA would ultimately destroy 600,000 jobs.

CBS, MSNBC and The Wall Street Journal have all covered the study, as have a slew of conservative websites like Pajama's Media. But few have asked who's behind the report -- and even less have looked under the hood at the research itself, which is based on a surprisingly tiny sliver of data from just three Canadian provinces dating back 33 years.

The report is titled "An Empirical Assessment of the Employee Free Choice Act: The Economic Implications." It was published on March 5 by Anne Layne-Farrar, an economist at corporate consulting firm LECG. Boasting offices in 31 countries, LECG posted $17 million in revenues in 2008, largely in fees they receive for delivering research and testimony on behalf of industry clients.

In fact, Layne-Farrar testified against the EFCA [pdf] yesterday before the Senate, drawing on her study to make alarming predictions about the potential raise in unemployment if the bill passes. Fox News' Special Report highlighted Layne-Farrar's Senate appearance, reporting that "One economist warned of what would happen if the bill passed and met its predicted goal of growing unions by 5 to 10 percent." Fox then cut to Layne-Farrar's testimony in which she claimed, "This would result in an increase in the unemployment rate of around 1 and a half to 3 percentage points."

On screen, Fox only identified her as an "economist." Other news reports have noted her connection to LECG, but only to say the firm is "non-partisan." But Layne-Farrar and LECG are far from disinterested parties in the battle over the Employee Free Choice Act.

As the press release that came out March 5 announcing Layne-Farrar's report admitted, "Funding for the Study was provided by the Alliance to Save Main Street Jobs." The release goes on to say the alliance

[I]s chaired by HR Policy Association and includes the American Hotel and Lodging Association, the Associated Builders and Contractors, The International Council of Shopping Centers, the Real Estate Roundtable, the Retail Industry Leaders Association and the U.S. Chamber of

All are groups that have invested large amounts of money and effort towards defeating the Employee Free Choice Act; the U.S. Chamber of Commerce alone plans to spend $20-$30 million in opposition.


Not only have media sources failed to identify who's backing the study -- they've also been slow to look into Layne-Farrar's claims and methodology, which are based entirely on decades-old data from a handful of Canadian provinces.

Layne-Farrar's thesis follows a classic business mantra: The Employee Free Choice Act will increase the number of workers in unions, which will in turn increase costs to business, lower investment and cause jobs to be lost.

First, note that the report takes as a given that more workers will join unions if given the "card check" option under the Employee Free Choice Act -- a very different line of argument than the business claim that EFCA takes away the right to a secret ballot (which it doesn't).

But back to the report's argument. Layne-Farrar builds her entire claim that the EFCA and resulting rise in union membership will lead to mass unemployment around one set of data -- namely, "a panel dataset of Canadian provinces over the twenty-two year period 1976-1997" (page 20).

Why Canada? Because, she says, their economy is roughly similar to the U.S. Canada is also unique in that labor laws differ between the country's 10 provinces -- some provinces use EFCA-like "card check" and others that don't, and one can compare the results.

But buried in page 20, Layne-Farrar herself admits that the data from which her entire argument is constructed isn't so great after all:

While the Canadian dataset is quite rich, it does have its limitations. For example, out of ten provinces that experienced changes in labor institutions (i.e., card check vs. mandatory voting) between 1976 and 1997, only three had enough variation in the card check rules themselves over time to allow for the reasonable estimation of any direct effects.

So instead of comparing 10 provinces, the study is really based on the experience of just three: Alberta, British Columbia and Newfoundland, from which Layne-Farrar proceeds to extrapolate how many jobs will supposedly be lost in the U.S. if EFCA is to pass.

But does even this small sampling of three Canadian provinces tell us much of anything about card check, unions and unemployment?

Not really, given all the other factors that contribute to losing jobs. For example, in Newfoundland, one of Canada's poorest provinces, unemployment has always been high -- but that has more to do with the boom and bust of the oil and fishing industries in the coastal province than anything to do with card check.

Blaming union laws for unemployment in Canada is especially futile given that they're constantly changing within the provinces themselves. Layne-Farrar (page 16) acknowledges that British Columbia changed its card check law "three times" in the period her research covers (1976-1997) -- a fact that would make it nearly impossible to find any long-term proof that card check alone caused jobs to be lost.

But Layne-Farrar doesn't take any of these realities into account. And she ignores the most obvious evidence of all: If unions really were the cause of unemployment, why has Canadian unemployment risen in recent years -- including 241,000 jobs lost in manufacturing along between 2001 and 2007 -- even as union membership has declined?

Even as a piece of business research-for-hire, Layne-Farrar's study is shockingly weak -- based on a thin set of old and irrelevant data that doesn't even bear out her own conclusions.

If businesses bankrolled Layne-Farrar in hopes that her research and testimony would be a silver bullet that could help stop the Employee Free Choice Act, they didn't get their money's worth.

Photo: Anne Layne-Farrar testifies to Senate committee on Employee Free Choice Act on March 10, 2009. Photo by Change to Win. For more information, see Media Matters for America.



People Referenced:


re: Investigation: Business bankrolls study claiming job losses

I'd say this is "junk" social science.

re: Investigation: Business bankrolls study claiming job losses

The "bidness community" and their wholly-owned subsidiaries in the SCUM (SoCalledUnbiasedMedia) will operate in this campaign in ways analogous to how the Bushevik ShiteHouse either stipulated the intelligence outcomes they needed to have to justify their policies, or cherrrypicked items that aided their cause.

re: Investigation: Business bankrolls study claiming job losses

Well, time to take a nap on this article. Seems that this Free Trade Agreement stuff hasn't done so well this far. Take a look around, the country is in shambles. Whether Unions or non-Unions, American workers are being scammed by our business owners so they can make more profits for themselves. Wages are compressing but not for the "rich and famous". Wall Street, AIG, Lehman Bros., CitiGroup and on and on is a good example while the automobile sector had to jump through hoops to get their loans.

re: Investigation: Business bankrolls study claiming job losses

I took a look at the paper. While there is much to criticize in it, the blog posting is incorrect to assert that only data from 3 provinces is used. She notes that only three provinces have enough variation to look at the effect of switching the rules on card check. So, her solution is to look at the effect of union density across all 10 provinces on unemployment rates.

In other words, her study does not examine the effect of the rule change directly, it can't the data isn't sufficient. Then, we can quibble about a lot of the methods and models, but we can set that aside and ask whether her models are fundamentally sound. Basically, when she asks how union density drives unemployment rates, her models necessarily make the assumption that unemployment rates don't drive union density (that's the way the math works). But, that's clearly a fallacy. She tries some adjustments to get around this, but none are convincing. For example, she lags the union density so that she is asking whether union density from a year ago is related to unemployment today. I see what she's trying to do, but it's just not good enough--especially for a paper that draws such sweeping conclusions.

Now, for the labor supporters in the audience, you might consider whether there is any value in conceding for the sake of argument that a marginal increase in unemployment may be a necessary trade off for increased job stability, security, and improved wages and working conditions. Heck, if you told me we could have all these things and bump up unemployment by 1 percentage point, I would argue that's fine let's improve the social safety net for that 1% and make work life a whole lot safer and more rewarding for the 10% plus that will benefit from the policy change. Politically, that might be tough, but it's worth thinking through.