No, Obamacare does not mean you are headed to this line. Photographer: Tim Boyle/Bloomberg.
No, Obamacare does not mean you are headed to this line. Photographer: Tim Boyle/Bloomberg.

Here’s the next conservative argument against President Barack Obama's health care law: It’s causing employers to shift full-time workers into part-time. Too bad it’s wrong.

The law requires employers to sponsor health insurance for all full-time employees. It defines “full-time” as 30 hours per week or more. To avoid that burden, conservatives are saying, businesses are replacing full-time positions with part-time ones.

Christopher J. Conover, a researcher at the right-leaning American Enterprise Institute, and Jed Graham, an economic-policy reporter, have recently stated the case in Forbes and Investor’s Business Daily. (You can also find it in the seedier parts of the right-wing Internet, such as this post on Zero Hedge.)

In the first sixth months of 2013, Conover finds, 4.3 part-time jobs were created for every one full-time one. That's unusual, he says, as the U.S. has historically created three full-timers for every part-timer. And Graham points to a 2.7-percent increase in the number of people working 25 to 29 hours per week and a 1.4-percent drop in the number working 30 to 34 hours per week. Both conclude Obamacare is to blame.

Both analyses are flawed. There’s no substantial evidence yet that employers are switching from full-time to part-time in response to health-care law. I do think that will happen sometime -- the cutoff is indeed a strong incentive for employers to reduce work hours -- but Conover and Graham are misleading their readers.

Let’s look at Conover’s evidence first. Here’s his key graph:

There are two big problems with it. First, his method of comparison is susceptible to creating extreme values. Second, the data set Conover uses, the Bureau of Labor Statistics' "household survey," is far too volatile to draw reliable conclusions from small samples.

Conover takes the number of new part-time jobs created over a six-month period, divided over that period's new full-time jobs. Here's the problem: If the number of new full-time jobs created is close to zero, it doesn't much matter how many part-time jobs are created -- the ratio will still be enormous.

Sure enough, the household-survey data Conover uses reports unusually slow full-time job growth since January: only 130,000 positions, versus 557,000 part-time ones. But this is just regular volatility in the household-survey data, and it’s not common practice among economists to rely on the household survey for just that reason.

Conover, to his credit, acknowledges the issue -- that's why he uses yearly, or in the case of 2013, six-month samples. But it is nowhere near enough to save his methodology. His data comes from a monthly survey of 60,000 households. That may sound like a lot, but it's small enough that no change in employment less than 436,000 people is statistically significant. Combined with his method, the volatility produces extreme yet insignificant results.

When we zoom out and look at a graph of the number of full-time and part-time jobs, we don't see Conover's claimed surge. It’s supposedly that last upward wiggle in the blue line since 2013:

It’s indistinguishable from noise. Another way to see why his method is misleading: Consider how his ratio has popped and dropped over the last decade. It’s so volatile that there have been 20 six-month periods since 1968 with a similar part-time surge. There’s something pretty clearly wrong with his method.

Graham comes at the issue from a different angle, but makes a similar mistake: Using volatile data, a sensitive method of analysis and not a whiff of context. Here’s his key graph:

To me this seemed convincing -- until I redid his work and found that, no, there hasn’t been any surge in work weeks just under 30 hours, nor any drop in weeks just over 30 hours. Here’s my graph, which indexes the total number of jobs by their weekly hours in January 2011. Again, no switch from full-time to part-time can be seen:

The biggest piece of evidence against Conover comes from the establishment survey's reading of average worker hours. This survey draws from 557,000 businesses counting some 45 million workers. (That’s 750 times bigger than the household survey Conover and Graham used, in case you’re keeping score at home.) And as you can see, there's been no noticeable drop in worker hours, contradicting his argument that more people are being shifted into part-time roles.

I think the 30-hour rule in Obamacare is bad public policy that will eventually hurt full-time work, as do Conover and Graham. In fact, I’d like to get rid of the employer mandate entirely. But I also believe in thorough, objective use of data. And they show no evidence of a part-time surge yet.

(Evan Soltas is a contributor to the Ticker. Follow him on Twitter.)