Denialism may be too strong a term.[1] But there seem to be a lot of people arguing that Obamacare has little or nothing to do with the rise in part-time employment. Some deny the rise is even happening, while others are content to deny that Obamacare is the culprit. Admittedly, it takes a little detective work, but if we systematically review the available empirical evidence in an even-handed fashion, the conclusion seems inescapable: Obamacare is accelerating a disturbing trend towards “a nation of part-timers.” This is not good news for America.
Why There is Doubt About Obamacare’s Impact on Part-Time Employment
Part-Time Employment for Economic Reasons Surged Due to the Recession. Let me start by conceding that Slate’s Matt Yglesias is quite correct about broad trends in those working part-time involuntarily. That measure surged during the recession and subsequently has declined at an admittedly erratic pace that is not uncommon post-recession. Whether the most recent uptick in this indicator is statistical noise or the start of an Obamacare-induced “surge” may not be obvious for another year. It should be clear that barring another recession (which, unfortunately, some are predicting for 2014), seeing this measure return above 9 million within the next year would be historically anamolous. Should that happen, it would be hard to avoid the conclusion that Obamacare’s employer mandate had something to do with this reversal.
Part-Time Employment Has Grown for Decades. Economist Scott Anderson of the Bank of the West also is correct that in June part-time jobs accounted for 19.5 percent of total employment, "exactly the average share ... since January 2009." He’s right that part-time jobs sometimes surge for a few months, but then the rapid gains have been reversed (in my graph, I calculate the ratio of PT to FT workers, which is slightly different than the percent of total employment accounted for by PT workers, but both measures tell the same story). So again, we cannot be certain that the most recently uptick is statistical noise or a signal of Obamacare’s impact.
Obamacare-Induced Part-Time Employment is Too Trivial to Worry About. Researchers at the Center for Economic and Policy Research (CEPR) offer a more pernicious argument. Sure, some workers are losing hours of work as employers seek to evade Obamacare, but this impact is too small to worry about. It’s essentially invisible in a statistical sense. I will credit CEPR researchers with having done a well-conceived study. They used Current Population Survey data to compare the first four months of 2013 with the first four months of 2012, examining the numbers and percent of workers who reported working 26-29 hours a week. Their reasoning was that if Obamacare were pushing more employees into part-time status, it would show up as a surge in the number of workers in this work-hour category. They found that only a small number of workers (0.6 percent of the workforce) report working just below the 30-hour per week threshold used to determine eligibility for employer-provided coverage under the Obamacare mandate. Perhaps more surprisingly they found that the number of workers who fall in this category was actually lower in 2013 (852,296) than in 2012 (853,124). Likewise, as a percentage of the workforce, the 2013 figure (0.5968%) was slightly lower (0.6037%) than in 2012. Ergo, no adverse effect of Obamacare.
Proof Positive Obamacare is Accelerating the Move Towards Part-Time Employment
Anecdotal Evidence. But there’s been a tsunami of stories about real-world employers who in fact have shifted to part-time workers precisely to avoid the Obamacare mandate. These include:
- Restaurants and fast-food establishments: Buffalo Wild Wings, CKE Restaurants Inc., owner of the Carl's Jr. and Hardee's hamburger chains, Del Taco, Subway franchises, Wendy’s franchises, White Castle
- Colleges that rely on part-time instructors: Community College of Allegheny County (PA), Kean University (NJ), Northern Virginia Community College, Palm Beach State College (FL), University of Arizona in Tucson, Youngstown State University, (OH)
- Movie theaters: Regal Entertainment Group, which operates more than 500 theaters in 38 states
- State government employees: Virginia
- Local government employees: Long Beach (CA), Dearborn (MI)
The CEO’s of these organizations are on record as attributing their shift to part-time employees to a desire to avoid the added costs of complying with the Obamacare mandate.[2] Each of these decisions affected real live workers who now will have to manage with working fewer hours for less pay. It’s certainly possible that some are exaggerating the influence of Obamacare on making an unpopular decision, but it seems improbable that all are.
Ratio of New PT Workers to New FT Workers Explodes in 2013. For the most part, an examination of metrics measured in millions (e.g., involuntary PT workers or total PT workers) masks what is really going on. A much better sense is given by comparing the changes in PT employment to the changes in FT employment. Because the monthly Current Population Survey are so volatile, it is easier to see what is going on by calculating an average monthly figure for each calendar year to get a sense of whether the number of PT or FT is rising or falling. We only have six months of data for 2013, but this method allows us to compare the average monthly count for the year to date with the average monthly count from prior years on an apples-to-apples basis. We can then calculate the ratio of new PT workers in an average month to new FT workers in an average month. Obviously this ratio will turn negative in years that either FT or PT workers have declined on average. So over the past decade, there’s only 4 other years with which to compare the 2013 experience.
What should immediately be obvious to even someone without a shred of statistical training is how deviant the 2013 experience is compared to the past. For every new FT job added to the economy, there were 4.3 PT jobs added! In most (non-negative) years, the ratio is the reverse: that is, there are typically 5 FT jobs added for every new PT job. Even in 2004—the year with the second-highest ratio during this time-frame--there were 2 FT jobs for every PT job, yielding a ratio of 0.5. Even if growth in PT vs. FT workers reverted to its historic pattern for the balance of 2013, the year’s average monthly ratio still would be four times as large as the 2nd highest ratio from 2004.
As U.S. News and Report’s editor in chief Mort Zuckerman argues ”At this stage of an expansion you would expect the number of part-time jobs to be declining, as companies would be doing more full-time hiring.” As of July 2013, average gross domestic product growth rate annualized over the past 15 quarters (2%) was the weakest GDP growth since World War II. Similarly, the civilian workforce-participation rate has dropped by 2.2 percentage points since the recession ended. “Such a decline amid a supposedly expanding economy has never happened after previous recessions.” So something unusual is clearly happening, and the timing of the 2013 surge in new PT employment relative to FT employment suggests that Obamacare is a culprit even if not the only one.
How Did CEPR Get it Wrong? The CEPR study I cited earlier concluded: “While there may certainly be instances of individual employers carrying through with threats to reduce their employees’ hours to below 30 to avoid the sanctions in the ACA, the numbers are too small to show up in the data.” Conversely, I’ve just pointed to evidence of a pretty clear effect of Obamacare’s impact on PT employment. I would argue that CEPR’s analysis was well-conceived but done too early (i.e., covering only the first 3 months of 2013). I’m willing to predict that were they to replicate this study to cover the first 6 months of the year (or, ideally, the entire year), they would obtain a very different story.
Under the rules issued for the employer mandate, employers can select a “look-back” period ranging from 3 to 12 months in order to determine which of their employees is working 30 or more hours weekly on average and hence must be offered affordable health benefits to avoid a penalty. Thus, there was no need for any employer to “lock-in” reduced working hours in the first three months of 2013. Which explains why so many of the anecdotal reports of cutbacks in hours occurred after March (Del Taco, Subway, Dearborn, MI). And which further explains why the average monthly number of new PT employees in April, May and June (206,000) was so much large than the average monthly number of new PT employees in January through March of 2013 (-20,000).
Is Obamacare-Induced Part-Time Employment Too Trivial to Worry About?
To be sure, the average monthly gain in PT employment during 2013 to date has been only 93,000 workers. In a workforce of 135.9 million nonfarm workers, even annualizing this figure would produce a number (1.1 million) that might seem like rounding error when expressed as a fraction of that large total (0.8%). Moreover, not all of that 1.1 million growth in PT workers could be attributed to Obamacare since obviously there would have been growth in PT employment even without the health law. Since 2000, PT employment has grown 1.6% annually, meaning that we’d expect to see 450,000 additional PT jobs annually anyway. Even so, that would imply 650,000 workers who involuntarily had their hours reduced due to Obamacare.
And the actual figure might well end up higher than this. Note that the 93,000 figure is an average of 3 months of PT job losses at the beginning of the year followed by 3 months of sizable PT job gains subsequently. If July through December look more like the second quarter than the first, we’d end up with a gain of 1.8 million new PT jobs in 2013 meaning that 1.35 million could be attributed to Obamacare-related hours reductions.
And the figure could be much higher. A study of Hawaii’s employer mandate found a 1.4 percentage point increase in the share of employees working less than 20 hours a week as a result of the law (Hawaii requires firms to provide health insurance to employees working 20 or more hours a week).[3] That would translate into 1.9 million workers nationwide who could lose hours of work due to Obamacare. And a UC-Berkeley study estimated there are an estimated 2.3 million workers nationwide who are at greatest risk for work hour reduction under Obamacare (at-risk workers were defined as those working 30-36 hours, below 400% Federal poverty level who do not currently have insurance through their own employer). This would appear to be the upper boundary of the plausible number of workers adversely affected by reduced hours of work under Obamacare.
Whether the actual number turns out to be 650,000 or 2.3 million, forcing workers to cut back their hours of work and income is not good for them and is not good for the health of the American economy. Sure, one can trivialize these numbers by saying they are a tiny fraction of the entire U.S. workforce. But I encourage anyone tempted to do so to talk to someone who has gotten their hours slashed to see how they feel about their situation. My guess is that very few enduring this pain will judge that the purported benefits of Obamacare made their cutback in hours worthwhile.
And remember that this adverse effect is above and beyond that created by the huge work disincentives created by “ACA subsidy cliff” and Obamacare’s Medicaid expansion (which alone could reduce employment by nearly 1 million). We already have a “labor market stuck in the mud” with an employment-to-population ratio that has declined to levels not seen since the early 1980’s. Why in the world would we want to aggravate such a bleak situation by moving forward with a Rube Goldberg contraption of carrots and sticks that will reduce work effort even further?
The delay of the employer mandate for 1 year was a very good start. Let’s delay the whole law for a year to give lawmakers time to ponder how to fix this terribly misguided law.
UPDATE 1: July 31, 2013
Spencer England at Angry Bear argues that "the average workweek has been amazingly stable over the past year, not at all what you would expect if Obama care was causing employers to shift their employees from full time to part time work." I have 2 observations. First, imagine that in the worst case, 2.3 million workers lost 5 weekly hours of work due to Obamacare. That's 11.5 million hours of lost ouput, but when divided among 135.9 million workers, it would reduce the average workweek by less than 1/10 of an hour. It turns out that the average workweek has declined by this amount since March (from 34.6 to 34.5 hours). Second, the figures for May and June are preliminary estimates that might well get adjusted downward once they are finalized. In short, the available evidence on weekly workhours is a soft reed on which to support a case that nothing is going on. More importantly, it's a crude measure insofar as you could do a lot of damage to literally millions of individual lives without this statistic budging at all.
One other observation I should have made in the post itself is that the BLS defines PT as under 35 weekly hours whereas the ACA defines it as under 30 hours. To the degree employers are cutting back hours of existing PT workers (as BLS defines them), say by changing a 34 hour per week worker to 29 hours, the BLS count of PT workers will remain completely unchanged. This is yet another instance of how our available metrics may well be hiding a phenomenon that is much larger than it might appear.
UPDATE 2: July 31, 2013
This blog post has focused on the pressures facing employers to move towards part-time labor since I assume most can agree that forcing a worker to accept a reduction in hours is not a good thing. There's a companion set of pressures on workers eligible for Exchange-provided coverage. Casey Mulligan has explained how in light of the massive subsidies available for coverage on the exchange some workers will discover that working 30 hours at a given wage will generate as much after-tax, after-health expense income as working 40 hours at that same wage. This implies a net hourly wage (after taxes and health spending) that is one third higher for the part-time worker. This will result in at least some workers voluntarily moving to part-time status knowing that it will then allow them to purchase subsidized Exchange coverage (which can be worth in excess of $10K for families). Implicitly, the Exchange subsidy structure is using taxpayer funds to encourage people not to work. I've been a persistent critic of the employer-based health coverage system, but solving this problem by encouraging people to work less quite clearly to me is the wrong answer.
UPDATE 3: August 7, 2013
Joseph Lawlor's Obamacare implicated in sharp shift from full-time to part-time work has a nice discussion of this issue, including a graphic from Investor’s Business Daily’s Jed Graham showing that in the second quarter of 2013, the number of workers putting in 30 to 34 hours at their primary job fell by a monthly average of 146,500 while the average monthly number working 25-29 hours per week in their primary job rose by 119,000. If the employer mandate is ever fully implemented, it will be interesting to examine year-over-year changes in this metric to ascertain more precisely how many workers were adversely affected by having their hours cut.
UPDATE 4: August 8, 2013
Evan Soltas, at Bloomberg, Is Obamacare Forcing You to Work Part-Time?, has done a nice rejoinder. He makes 2 points. First, that the data I used are very volatile, which they certainly are on a month-to-month basis. But he seems to ignore that when averaged out, even very volatile data can give a pretty good picture of general trends. For example, look at the first chart in this post showing the number working PT for economic reasons. The month-to-month figures jump around a bit, but one can certainly tell that on average, this indicator tends to decline in the months/years following the end of a recession. Thus, while it stretches the data to point to a single month's change (up or down) and draw hard conclusions from it, it is certainly legitimate to compare year-over-year differences in monthly averages since by averaging, most of the volatility in this metric gets smoothed out: sharp swings up or down get averaged out etc. Indeed, it makes no sense to squander tax dollars (not to mention the time of survey respondents) collecting these data if we're going to dismiss away the results as meaningless noise too volatile to be meaningful. Many of the basic metrics of our employment situation come precisely from this very same household survey: the unemployment rate, the number working PT for economic reason, the split between PT and FT work etc.
Second, Mr. Soltas does make a good point by claiming my metric (the ratio of net changes in PT workers to net changes in FT workers) is "so volatile that there have been 20 six-month periods since 1968 with a similar part-time surge." Thus, over 45 years, he's claiming the ratio was at or about the 4.3 figure I show in my chart on 20 different occasions, or roughly once every 2 years. It's regrettable he didn't elect to show us a chart of these results as that might have helped deduce what was going on in the general economy to have created such conditions (I don't doubt that this might have happened with some regularity during the abysmal stagflation years of the 1970's, but that is by no means the typical behavior of the economy in normal times etc.). I concede I only plotted the last 11 years since those are the data I grabbed from BLS. If Mr. Soltas is correct, that means that the unusually high ratio I observed occurred 19 times in just 34 years. So either the years from 2003-2012 were extremely unusual in their absence of such a pronounced spike (which should have occurred 5 times during this period had it followed the historical average) and 2013 merely represents a return to "normalcy" or the incentives created by Obamacare resulted in an out-of-the-ordinary spike in this ratio.
Third, Mr. Soltas tries to drive a nail in the coffin of Obamacare-induced PT employment by pointing to the average workweek: "The biggest piece of evidence against Conover comes from the establishment survey's reading of average worker hours....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." There's only two problems with this. First, his own data belie the claim itself. There was an initial uptick in worker hours at the start of the calendar year, but since March, the average workweek for private workers (the metric Mr. Soltas relies upon) dropped from 34.6 hours to 34.4 hours. Of equal importance, the decline in the workweek over the same period has been larger (0.3 hours) in precisely those industries researchers expect to be most adversely impacted by the employer mandate: leisure and hospitality and retail trade. Second, as I demonstrated clearly in my first update, 2.3 million workers could lose 5 weekly hours of work and it would only reduce this measure by less than 1/10 of an hour. Thus, the fact that such a crude measure dropped by "only" 0.2 hours in the space of 4 months should actually be of great concern and certainly is no proof that Obamacare is not having serious consequences for a sizable number of workers. In the interests of further contributing to this debate, I encourage Mr. Soltas to share a chart of his findings rather than leave us with his unsubstantiated claim that the 4.3 ratio I uncovered has been a rather common occurrence over the last half century.
Update #5: August 16, 2013
Jed Graham at Investors Business Daily has documented further evidence that "ObamaCare Fuels Sharp Workweek Drop In 4 Industries."
FOOTNOTES
[1] The blog Denialism.com asserts: “Denialism is the employment of rhetorical tactics to give the appearance of argument or legitimate debate, when in actuality there is none. These false arguments are used when one has few or no facts to support one’s viewpoint against a scientific consensus or against overwhelming evidence to the contrary. They are effective in distracting from actual useful debate using emotionally appealing, but ultimately empty and illogical assertions.”
[2] I have codified the reported rationales for each of these decisions at this page.
[3] There’s a parallel study of the employer mandate in Massachusetts which found no evidence of a disproportionate shift towards part-time work compared to the rest of the U.S. However, in Massachusetts, employers with more than 10 full-time equivalent employees (FTEs) are required to contribute a “fair and reasonable” amount toward health insurance premiums for employees. Firms that fail to make “fair and reasonable” contributions face financial penalties of $295 per FTE per year, which is much smaller than in the ACA ($2,000 per worker excluding the first 30). Thus, the Massachusetts experience appears much less relevant to predicting the impact of Obamacare than Hawaii’s.
