What this report finds: An ongoing structural shift toward more intensive use of part-time employment by many employers is driving the elevated rate of involuntary part-time work. Over six years into an economic recovery, the share of people working part time because they can only get part-time hours remains at recessionary levels. The number working part time involuntarily remains 44.6 percent higher than it was in 2007. This growth is being driven mainly by a few industries.

Why it matters: 6.4 million workers want full-time jobs but are working only part-time hours. Involuntary part-time workers are not only earning less income than they would prefer, but suffer because part-time jobs offer relatively lower wage rates and benefit coverage, and have more variable and unpredictable work schedules.

How we can fix the problem: In addition to traditional expansionary policies that would heighten demand for more hours of labor, here are seven policies that would help curb the excessive use of part-time employment and address the harmful effects of involuntary part-time working.

Introduction and key findings

While average annual working hours of all workers rose 9 percent from 1979 to 2013 (Mishel 2013), this statistic masks a hardship faced by many workers in the United States. Particularly for middle- and low-wage earners, the key problem is often too few hours, and/or too variable hours. In other words, they would prefer to be working more hours, and to not have to navigate through erratic work hours or schedules.

Not getting enough hours is the “time-related” type of underemployment, a phenomenon where people may be working but not up to their desired amount, and it is a sign of labor underutilization in the economy.1 The monthly rate of workers in the U.S. labor market who are working “part time for economic reasons”—who are considered “involuntary” part-timers because they want to and are available to work full time—is the most consistent indicator of such underemployment. That rate is higher now that it was before the Great Recession and during the depths of the early 2000s recession. That it remains stubbornly high indicates that there is more labor market slack than is captured by the unemployment rate alone.2

Over six years into an economic recovery and economic expansion, as the unemployment rate has fallen, inadequate work hours are still a concern, as noted by many reporters, commentators, economists, and the chair of the Federal Reserve Bank.3 Indeed, the economic weight of involuntary part-time work has been an issue in the presidential campaign.4

There is not only an incomplete recovery in the labor market—which is likely inhibiting the strength of economic expansion—but greater labor market hardship for many workers than is apparent on the surface. Part-time employment generally comes with many disadvantages vis-à-vis full-time jobs, such as lower rates of overall compensation per hour and work schedules that are often less stable or predictable. When working part time is involuntary, the harms are compounded.

This report suggests that, in addition to cyclical forces (in this case, lingering effects of the recession), there is an ongoing structural shift in many businesses toward more intensive use of part-time employment, driving the elevated rate of involuntary part-time employment. Increased employer use of part-time positions is particularly evident in industries in which part-time jobs are already more prevalent, such as retail, and hotels and food service.

The report identifies and explains the monthly and annual trends in involuntary part-time work, the role of key industries driving much of those trends, the kinds of workers and industries most affected by part-time work, and the myriad challenges that workers in part-time jobs face. Following are a summary of the key findings:

Key findings

Trends and causes of involuntary part-time employment

  • The share of people working part-time involuntarily remains at recessionary levels. In 2015, there were 6.4 million workers who wanted to work full time but were working part time, accounting for 4.4 percent of those at work; this is roughly 2.0 million more involuntary part-time workers, or a 1.3 percentage-point increase in the rate of involuntary part-time employment prior to the recession. In fact, data from 2007 to 2015 show that involuntary part-time work is increasing almost five times faster than part-time work and about 18 times faster than all work.
  • It is this rise in involuntary part-time work that is driving an overall increase in part-time employment generally, as the share of the workforce working part time voluntarily has been stable since 2007. Thus, the “new normal” of underutilized labor primarily reflects the increased employer use of part-time employees and not any increased preference among workers for part time employment.
  • The currently elevated level of part-time work—and of involuntary part-time work in particular—is no longer “cyclical,” i.e., it does not reflect a delayed and slow recovery, although reaching full employment could eventually yield a diminution in part-time work as workers are able to secure full-time employment.
  • The structural nature of today’s involuntary part-time employment is evident in the decrease in workers who say they are involuntarily part time due to slack work. Involuntary part-time work has gradually decreased since 2009 but almost entirely because fewer workers are working part-time hours due to “slack work or business conditions,” which had ballooned during the Great Recession. Slack work is an indicator of cyclical business lows. In contrast, the share of those working involuntarily part time because they “could find only part-time” work (i.e., employers were offering only part-time work, indicative of structural factors) is just as high as it was at the end of the recession in 2009.
  • Involuntary part-time work and its growth are concentrated in several industries that more intensively use part-time work, specifically, retail and leisure and hospitality. Retail trade (stores and car dealers, etc.) and leisure and hospitality (hotels, restaurants, and the like) contributed well over half (63.2 percent) of the growth of all part-time employment since 2007, and 54.3 percent of the growth of involuntary part-time employment. These two industries, together with educational and health services and professional and business services, account for the entire growth of part-time employment and 85.0 percent of the growth of involuntary part-time employment from 2007 to 2015.
  • Trends in the reason for part-time employment by industry also suggest structural factors in play. In 2015, involuntary part-time workers made up 7.8 percent of all those at work in the retail sector. That is 3.4 percentage points higher than before the recession started, in 2007. Roughly 60 percent of this growth in involuntary part-time work reflects those who “could find only part-time work.” Involuntary part-time work was an even higher proportion of employment, 10.4 percent, in the leisure and hospitality industry in 2015, up 3.6 percentage points from 2007. Roughly half of this growth in involuntary part-time work reflects those who “could find only part-time work,” indicating structural factors were at least as important as cyclical factors.
  • The suggestion that the “shared responsibility provision” of the Affordable Care Act (ACA) is behind some of the shift toward part-time work is not supported by the data. The provision requires that certain employers pay a fee if they don’t offer a minimum level of health insurance to employees working 30 or more weekly hours. Had these health care–related labor costs prompted employers to reduce more positions to part-time hours, there would be a number of trends in the data that suggest a structural change in involuntary part-time working or hours worked, and these trends do not appear.

Certain groups of Americans are most vulnerable to the burdens of involuntary part-time work

  • Hispanic and black workers have been hardest hit by the structural shift toward involuntary part-time work. Hispanics and blacks are relatively much more likely to be involuntarily part-time (6.8 percent and 6.3 percent respectively) than whites, of whom just 3.7 percent work part time involuntarily. And blacks and Hispanics are disproportionate shares of involuntary part-time workers: together they constitute just 27.9 percent of those “at work,” they represent 41.1 percent of all involuntary part-time workers. The greater amount of involuntary part-time employment among blacks and Hispanics is due to their both having a greater inability to find full-time work and facing more slack work conditions. Black and Hispanic women (and women of “other race/ethnicity”) are the groups most likely to experience involuntary part-time employment and represented 21.1 percent of all involuntary part-time workers in 2015.
  • Prime-age workers are a disproportionate share of involuntary part-time workers. Workers ages 25 to 54 comprised 57.8 percent (3.5 million) of all involuntary part-time workers (6.1 million) despite being only 44.0 percent of all part-time workers.
  • Men and women are similarly afflicted by involuntary part-time working—with an incidence of 5.1 percent among women and 4.0 percent among men.
  • The service occupations (e.g., healthcare support, food preparation, building and grounds maintenance, personal care, etc.) contribute the most to involuntary part-time employment, followed by sales. Service occupations provide 17.2 percent of all persons at work but represent double that share (34.5 percent) of all involuntary part-time workers. Sales and office occupations account for 26.8 percent of all involuntary part-time employment.
  • Part-time workers work about half as many hours per week as full-time workers, with the clear adverse consequence of a corresponding reduction in one’s weekly earnings.
  • The biggest disadvantage that part-time workers face is their relatively lower rates of pay and their benefit coverage. Prior research shows that the part-time wage penalty (the percent less in hourly wages that part-timers make relative to similar full-timers) is 19 percent for men and 9 percent for women. Part-time jobs, particularly those with the fewest weekly hours, also provide relatively less access to benefit coverage. Part-timers have only one-third the access to health insurance coverage as do full-timers—22 percent compared with 73 percent.
  • Part-time workers must navigate varied and unpredictable hours. Part-time workers are much more likely to have work hours that vary from week to week—at a rate 2.5 times higher than among full-time workers. (Involuntary part-time workers face the most variability). Surveys show that part-timers also face greater irregularity in their work shift times and unpredictability in their work schedules.

Clearly part-time employment, especially involuntary part-time employment, has various adverse consequences. In addition to traditional expansionary policies that would heighten demand for more hours of labor, we need policy innovations to help curb the excessive use of part-time employment by many employers and address the harmful effects of involuntary part-time work. Remedies explored at the conclusion of this report include compensation parity for part-time jobs, reforms to unemployment insurance systems, an employee “right to request” changes in hours, and laws giving part-time workers priority access to increased hours of work that become available. Moreover, laws requiring that workers receive minimum pay for shifts that are canceled or schedules that are changed, along with unemployment insurance reforms would both incentivize reductions and mitigate the adverse impacts of involuntary part-time working.

Structural factors from the employer side are driving the rise in involuntary part-time work over the last decade and a half

The number and share of workers who are involuntarily working part time is higher now than it was before the Great Recession and during the depths of the early 2000s recession. The data suggest that the elevated level of involuntary part-time work is due less now to lingering effects of the recession (i.e., to “cyclical” forces) than to a structural development driven by employers that are shifting more jobs and work to part time.

We have an elevated level of part-time work

Table 1 presents the number of persons at work total and at work part time overall and for the two types of part-time employment—voluntary and involuntary. The involuntarily part-time workers are further broken down into the main reasons for working part time, either due to “slack work” or because a worker “could find only part-time work.”5 All these categories are also presented as a share of total persons “at work” during the week of the survey. In other words, the bottom panel shows, of all the persons at work, what share were part time, involuntary part time, etc. The definitions of these categories are explained in the text box, “Defining involuntary, voluntary, and total part-time employment.”

Table 1

The table takes the monthly numbers and averages them for key years including the bottom of the last recession, 2002 and 2003, the end of the last recovery, 2006 and 2007, and the most recent years, 2014 and 2015 plus the last 12 months of available data ending July 2016. The monthly numbers of involuntary part-time workers and the two subcategories are presented in Figure A and the corresponding shares of total at work are presented in Figure B.

Figure A
Figure B

Table 1 highlights two key trends. First, part-time workers are making up an increasing share of all workers, and second, that this growth is being fueled by disproportionate increases in the share of workers who are part time involuntarily.

In 2015 there were 6.4 million involuntary part-time workers, and they made up 4.4 percent of the total workers at work and almost a quarter of the total 26.4 million working part time (see Table 1).6 They are the individuals who indicate wanting full-time work, but working fewer than 35 hours during the survey reference week, for reasons that are considered to be “economic,” in contrast to individuals working shorter workweeks for reasons that are considered to indicate it is by choice (voluntary).

This is an increase of roughly 2.0 million from just before the last recession. In 2007 there were 4.4 million involuntary part-time workers, and they made up just 3.1 percent of the total workers at work, and less than one-fifth of the total 24.2 million working part-time.

It is also an increase from the bottom of the early 2000s recession. In 2002 there were 4.2 million involuntary part-timers, constituting 3.2 percent of total workers and less than one-fifth of the total 23.1 million working part time.

Thus, the rate of involuntary part-time employment in 2015 was 1.3 percentage points above its prerecession level in 2007, and so far in 2016 it is still 1.1 percentage points above its prerecession level (also see Valletta and van der List 2015).

Put another way, since 2007, the number of people working part time involuntarily has increased 44.6 percent, while total part-timers increased 9.1 percent and total at work increased 2.5 percent. Involuntary part-time work is thus increasing almost five times faster than part-time work and about 18 times faster than all work.

Defining involuntary, voluntary, and total part-time employment

Through the Current Population Survey, sponsored jointly by the U.S. Census Bureau and the U.S. Bureau of Labor Statistics (BLS), the federal government collects an array of data on part-time and full-time work. The BLS reports this data monthly and researchers use CPS microdata to investigate employment trends. Some of the data are based on “usual hours worked” and some are based on “actual” hours worked in the last week in the “reference week.” The focus in this study is a subset of those reporting actual hours worked in the reference week, specifically those whose answers classify them as working “part-time for economic reasons.” They are considered to be working part time involuntarily because they indicate that they are willing and able to work full time but have had to settle for a part-time schedule. Part-time workers who are classified as working “part time for noneconomic reasons” are considered voluntary part-time workers.

These voluntary and involuntary part-time workers are part of the universe of individuals who were “at work” in the reference week, and they are part of the overall number of workers employed. Survey respondents who said they were not “at work” or were “absent” but have a job are also counted as employed. (Workers employed but not at work would include those who were on vacation, or sick, or absent for some other reason.) Those who are at work and those who are not at work/absent together comprise everyone “employed.” The BLS applies statistical methods to the monthly survey responses to compute overall numbers of people at work and employed in the U.S. economy. Yearly BLS reports and tables annually average those data. In 2015 there were 148.8 million employed workers, of whom 143.8 million were “at work,” and 5.1 million (3.4 percent of total employment) were absent (Current Population Survey labor force statistics).

BLS identifies those who are working part time voluntarily as those who worked from 1 to 34 hours in the last week and usually work part-time, 20.0 million in 2015 (Current Population Survey labor force statistics). This group of those working part time voluntarily excludes the 9.8 million who usually work full time but worked 1 to 34 hours in the last week primarily because of vacations, holidays, weather, family obligations or other “noneconomic reasons.” On the other hand, those working part time involuntarily includes all workers who worked from 1 to 34 hours in the last week (whether they “usually work full time” or “usually work part time”) because of what BLS terms “economic reasons.” Economic reasons include “slack work or business conditions,” “could find only part-time work,” “seasonal work,” or the “job started or ended during the week.” There were 6.4 million such workers in 2015.

The BLS definition of voluntary part-time work includes many workers, especially women, who seek part-time work “voluntarily” only in the sense that they receive little if any supports for child care, sick leaves, and other social obligations that, if supported, would help enable them to work full-time hours. Thus, if one works part-time hours on a regular basis because of “childcare problems” or “other family/personal obligations” one would be considered doing so voluntarily, for “non-economic” reasons.7

In this study, which uses CPS microdata rather than the published BLS CPS tables, the total number of part-time workers is measured as the sum of those working part time voluntarily and involuntarily, 26.4 million in 2015.

This categorization surely provides an undercount of those working part time involuntarily and voluntarily, because it excludes the 3.4 percent of workers who were absent from work (Current Population Survey labor force statistics).8 How much of an undercount is unknowable. People can be working part time regularly because they have failed to find a full-time job and be absent in the reference week and therefore may be excluded from the total count of those who are part time for economic reasons. In addition, others could be working part time one week due to slack work and be absent (not “at work”) during the reference week. Nevertheless, we can calculate what the total number of part-time involuntary workers would have been in 2015 if those absent had worked (had they been present) part time involuntarily at the same rate as those at work: this “upper bound” on the total number of involuntary part-time workers would be 223,000 higher (4.4 percent of the 5,066,000 absent persons).

This paper examines aggregate patterns and trends using the universe of persons “at work” rather than “total employed” because the at-work workers are the only ones identified as working part time voluntarily or involuntarily. Part-time employment is set as the sum of involuntary and voluntary part-time employment. These aggregate data are obtained directly from published BLS tables or from downloaded BLS data series (identified in footnotes as data are discussed in the text). The paper also examines trends by demographic group, industry, and occupation using the universe of wage and salary workers (i.e., excluding the unincorporated self-employed and unpaid family members) who were “at work,” a group corresponding most closely to “payroll employment” that is the subject of the various legislative policy initiatives addressing part-time underemployment and work scheduling. The data in our demographic, industry, and occupational analyses are primarily examined as shares of total persons at work in various groups (e.g., race, ethnicity, gender, occupation, and industry). These data are drawn from EPI tabulations of Current Population Survey data.

We first examine the recent pattern of involuntary part-time employment to assess how much is cyclical—associated with the general slack in the economy due to the Great Recession and subsequent slow recovery—and how much is a secular or structural change reflecting new ways of organizing work. The proportion of workers working part time involuntarily has declined, gradually, over the recent recovery period. As of July 2016, averaging the data from the 12-month period ending in July, the number stands at 6.1 million employed (Table 1).

The currently elevated level of involuntary part-time working is no longer cyclical

After peaking in 2010, a year after the Great Recession officially ended, and averaging nearly 9 million between 2009 and 2011 (Sum and Khatiwada 2010;Mayer 2014), involuntary part-time employment only recently (in the summer of 2016) fell to just under 6 million. However, it has declined notably more slowly than the unemployment rate. Moreover, this form of underemployment in the U.S. labor market is not only high relative to its prerecession low of 4.4 million workers in 2007, it hit historically higher peaks in this recession and recovery than in previous ones (BLS 2008; Mayer 2014). If the labor market in 2015 were completely back to its 2007 share then the number of involuntary part-time workers would have been 4.5 million rather than the 6.4 million that we had, suggesting a rise of 1.9 million adjusted for growth in overall employment.9 Thus, in addition to those who are not in the labor force but can be expected to return, in 2015 there were about 1.9 million extra people in part-time jobs who would have preferred full-time work.

The share of total persons at work who were working part time involuntarily reached a high of 6.7 percent in November 2009, in contrast to peaking below 4 percent in the previous recession (Figure B). Indeed, the current share of 4.2 percent in the most recent twelve months thus remains at what could be considered comparable to the level in a mild recession–and clearly above the rate of 3 percent observed once the economy completed the recovery from the 2001 recession (see Canon, Kudlyak, and Reed 2014). Over the course of the Great Recession, the share of workers involuntarily employed part time more than doubled to 6.7 percent. We explore the potential sources for the lingering high rate of working part time for economic reasons by contrasting the two main “economic reasons” given by CPS respondents—“slack work or business conditions” and “could find only part-time work.”10 Those working part time involuntarily due to “slack work or business conditions” account for about 60 percent while those who “could find only a part-time job” represent only about 35 percent of all involuntary part-time workers in 2015 (see Table 1).

Compared with 2007, by 2015 the U.S. labor market had about 2.2 million more people working part time and about 1.2 million more people working full time (not shown, but full-time workers are those at work but not part time, Table 1). Indeed, until October 2010, the growth in part-time employment exceeded the growth in full-time employment (Hotchkiss 2015). However, since then, the part-time share of employment has declined, down to 18.3 percent for 2015 (Table 1) from its peak in early 2010 (Hotchkiss 2015). Some have noted that involuntary part-time work has declined along with this overall decline in part-time work (Bernstein 2016).

In summary, the currently elevated level of involuntary part-time working is no longer due largely to cyclical forces, and appears to reflect more than just the delayed and slow economic recovery. It is possible that future economic expansion and further reduction in rates of unemployment will eventually yield a diminution in the number of involuntary part-timers able to secure full-time employment, but this cannot be assumed or certain.

Employer preferences are behind the structural factors driving the increase in involuntary part-time work

A BLS report recently asked, “Is involuntary part-time employment here to stay?” (Karageorge 2015). And the trend in the involuntary part-time component of part-time employment has become the subject of two lively debates among labor market analysts—whether there is a “new normal” of a “structural shift” in the labor market beyond the usual cyclical pattern of economic expansions and contractions, and, if so, what factor or factors are primarily responsible for a structural change toward higher rates of part-time work and involuntary part-time employment (e.g., Cajner, et al. 2014; Valletta and van der List 2015; Even and MacPherson 2015). Some suggest the currently higher rate of involuntary part-time work mainly reflects the greater severity of its cyclical source (e.g., Kroll 2011; Valletta and Bengali 2013; Jolevski and Sherk 2014), while others do see new, structural factors at play that will keep involuntary part-time work above its prerecession level. As the previous section showed, cyclicality does not explain the current elevation of involuntary part-time employment. According to recent literature, structural factors that may explain the persistently high level of involuntary part-time working include changes in industry composition, population demographics, and higher labor costs for hourly workers (Fallick 1999; Karageorge 2015; Valletta and van der List 2015).

Looking first at industry composition, involuntary part-time employment has been observed to be generally higher in the service sectors, in which full-time employment fluctuated less in both the recession and recovery than full-time employment in goods-production, and where the demand for part-time workers is relatively higher, particularly in the retail and wholesale trade and in leisure and hospitality industries (Karageorge 2015; Valletta and van der List 2015).

Demographic factors may have also increased the numbers of involuntary part-timers. Workers younger than 25 are both falling in number and more likely to be voluntarily part time, leading employers to either hire workers for part-time jobs who are more likely to prefer more hours, even full-time hours, and/or hire more relatively older workers for part-time positions (Valletta and van der List 2015).

Regarding labor costs, another potential structural factor is federal tax rules regarding exclusion of part-timers from many types of benefits, which make it attractive to employers to hire more part-time than full-time employees if benefits are increasing in cost more than wage compensation. Moreover, new technologies that reduce the cost of scheduling part-time workers “on demand” and lax regulation of such on-call scheduling facilitate the ability of employers to “call off” employees, right before or even during shifts, and this might also be increasing the proportion of workers employed part time even though they would prefer to work more hours (see Lambert, Fugiel, and Henly 2014; Alexander and Haley-Lock 2015).

Finally, there may be structural changes also on the labor supply side. As much as one-third of the increase in involuntary part-time employment during the recession represented a shift of workers, transitioning from voluntary part-time to involuntary part-time employment (Cajner et al. 2014; Canon et al. 2014). A weakened labor market and stagnant hourly earnings could have spurred an “income effect” leading some part-time workers to seek more weekly work hours, for example, if a household’s “secondary earner,” who had been voluntarily working only part-time hours developed an unfulfilled preference to work longer hours (thus categorized as working part-time for economic reasons).11 On the other hand, the increased flow from voluntary to involuntary part time work during the recession could represent a real behavioral change.12

Two of these forces cited above—the growth in industries already using high part-time to full-time staffing ratios, and higher fixed costs for full-time jobs relative to part-time jobs—refer to the employer creation of part-time jobs that underlies this trend.

Thus, involuntary part-time work has grown not only because of the cycle, and rising unemployment, but also because of a secular increase that appears to reflect in great part an increased employer use of part-time work.

It is instructive that the growth of involuntary part-time employment has been driven more by workers’ inability to find full-time work and not by an escalation of slack-related part-time work: from 2007 to 2015 involuntary part-time workers as a share of all those at work rose by 1.3 percentage points and “could find only part-time employment” contributed more to that growth than increased part-time work due to “slack work.” Although a long and lagging recovery was predicted to occur after any recession triggered by a financial sector collapse, the return of unemployment (currently 4.9 percent) but not involuntary part-time employment, to prerecession levels is both puzzling and worrisome.13 It is instructive that the rate of working part-time voluntarily actually has edged down a bit since 2007, reinforcing that the growth of part-time employment reflects employer-side actions rather than workers increasing their desire for part-time schedules.

From the recent research and behavior observed in the time series data, we consider developments, besides a slack economy, that might help explain the recent trends related to part-time employment. One is whether there are a few sectors driving this increase, a development reflecting new employer strategies that emphasize part-time employment to produce goods or services. Another potential factor inhibiting a return to previous lower levels of involuntarily part-time work could be a structural change due to the ACA “shared responsibility fee” (Even and MacPherson 2015). However, an ACA effect would be evident only if the rate of involuntary part-time work changed more or less equivalently across all industries and both reasons for working part time, and the timing of such a change coincided with the (actual or expected) implementation of the ACA.14 Thus, it is key to observe the trends and patterns within industries—for jobs and hours by reason for working part time—to discern whether there is more than typical cyclical behavior of part-time work recently.

Retail and leisure and hospitality are among the key industries driving the structural shift toward involuntary part-time work

There are at least two ways that industry trends could explain the changes in the behavior and character of involuntary part-time employment. First, the proportion of the workforce that is part time—for voluntary and involuntary reasons combined—may simply differ across industries so that a disproportionate growth in part-time-intensive industries could elevate part-time employment overall and involuntary part time with it, i.e., through a “composition effect.” Second, involuntary part-time work might have increased during the recovery in some industries but not (much) in others because of employer strategies that increase the creation (or holding) of part-time jobs.15

The aggregate increase in involuntary part time evident in the data suggests that at least some if not most industries indeed experienced a greater share of involuntary part-time workers after 2007 (Robertson and Terry 2014). While at the industry level, the “could find only part-time work” component is quite positively correlated with the “slack work” component, the recent, relatively higher involuntary part-time employment may be attributed to changes in industry composition (Valletta and Bengali 2013). Indeed, the growth in (particularly involuntary) part-time employment occurred within just a third of detailed industries—particularly in food service and production and segments of the retail industry. About 20 percent of all workers are employed in those industries that experienced an increase (between 2012 and 2013) in involuntary part-time employment (Valletta and Bengali 2013).

Indeed, the analysis in this report below shows that among major industries, the percentage of involuntary part-time employment is highest in retail trade and in leisure and hospitality.16 However, since 2007, the increase in involuntary part-time employment is explained more by increasing rates within industries—in particular, within two major sectors, leisure and hospitality and retail trade—rather than a change in the composition of employment by industry (i.e., rather than leisure and hospitality and trade generating jobs at a faster clip than other industries).

As this example shows, the percent of the workforce working involuntarily part-time might remain elevated because of either (or both) greater part-time job creation or involuntariness in certain industries. The growth could reflect labor supply reasons: people who normally work full-time jobs may be still grasping those part-time jobs, until something else becomes available. But the growth is more likely to reflect labor demand reasons: greater use of part-time jobs and workers by employers in those industries.

Thus, we next engage in some further investigation at both a major and more disaggregated industry level. First, we identify the sectors most responsible for the recent trends in the growth of both total part-time and involuntary part-time employment. Then, we contrast some selected major and detailed industries intended to be representative industries—with a focus on retail and leisure and hospitality, examining construction as a point of comparison.

Table 2 examines the changes in part-time employment among wage and salary workers at work overall and in four key sectors. The bottom panel shows, of all persons at work part time (and involuntarily part time), what shares are in four key sectors. The analysis focuses on wage and salary workers, thus excluding unincorporated self-employed and unpaid family members, because it is these workers that are the subject of policy discussion regarding scheduling practices and part-time employment. (That is why the totals don’t match the totals in Table 1.)

Table 2

Table 2 shows the concentration of part-time work in certain industries. Roughly three-fourths of all part-time workers (and well over three-fifths of all involuntary part-time workers) are working in four key sectors—and these shares have increased markedly since 2007.

Total part-time employment grew by 2.2 million overall between 2007 and 2015, the most recent full year of data; during the same period, unemployment returned to its earlier rate. The growth of part-time work was concentrated in four sectors: retail trade, professional and business services, education and health services, and leisure and hospitality (such as hotels and food service). In fact, the 2.2 million growth in part-time employment in these four sectors fully accounts for all the growth of part-time jobs in the economy, as outside of these sectors part-time employment actually fell (by 89,000). The contribution of these four sectors is clearly disproportionate since they accounted for all the part-time employment growth from 2007 to 2015 but represented just 51.7 percent of those “at work” in 2007 (not shown in Table 2 but using the same data source). These four key sectors use part-time workers more than other sectors as almost three-fourths (74.6 percent) of all part-time workers worked in these sectors in 2015. Not surprisingly, these key sectors employed nearly 70 percent (69.5 percent) of all involuntary part-time workers in 2015.

That fact that growth of part-time work is concentrated by industry, rather than across the board, lends credence to a strong role played by ongoing structural changes beyond the purely cyclical reason of a painstakingly slow recovery. Table 2 also shows that almost all the growth of part-time work between 2007 and 2015 was among those who work part time involuntarily—2.0 million workers. Moreover, the growth of involuntary part-time work was also concentrated in those same four sectors, which contributed 1.7 million, or 85 percent, of the total growth in involuntary part-time work.

Table 3 presents a different analysis. Instead of showing the shares of all part-time workers who are in key industries, it examines the shares of all workers in key industries who are working part time, and part time involuntarily. The totals in the top data row, which show the shares of workers “at work” who are working part-time and what share are doing so involuntarily, do not exactly match Table 1 because for this industry-specific analysis we are looking at wage and salary workers, not unincorporated self-employed and unpaid family members.

Table 3

The table shows two keys things. First, these four key industries do indeed make heavy use of part-time work, particularly involuntary part-time work when you compare the share of sector workers who are part time with the share of workers overall who are part time and the share of workers in other sectors who are part time. Second, this heavy use of part-time work has increased since the Great Recession, virtually all because of the increasing use of involuntary part timers.

The share of all those at work who work part time rose from 16.9 percent in 2007 to 18.0 percent in 2015, a rise of 1.1 percentage points. The rise of involuntary part-time workers’ share of all at work, from 3.1 percent to 4.5 percent, a rise of 1.4 percentage points, more than explains the growth in part-time employment as a share of those at work.

In 2015, involuntary part-time workers made up 7.8 percent of all those at work in the retail sector, a rise of 3.4 percentage points from 2007 and more than the 2.6 percentage-point increase in the share of retail workers working part-time overall. Involuntary part-time workers constituted an even higher share (10.4 percent) of all leisure and hospitality workers, up 3.6 percentage points from 2007, an increase that exceeds the 2.9 percentage-point increase in the share of leisure and hospitality workers working part time overall. This validates that a large share of the growth of part-time work in these sectors reflects employers systematically creating part-time jobs, with workers filling those jobs despite wanting full-time work.

As Table 2 showed, these two sectors, leisure/hospitality and retail, contributed over half (together, 63.2 percent) of the growth of all part-time employment from 2007 to 2015. Table 3 shows that the rising share of part-time work within these sectors is totally explained by the rising share of involuntary part-time work. These two sectors make intensive use of part-time work, 39.8 and 29.5 percent, respectively, in leisure/hospitality and retail trade in 2015. Thus, it is the part-time intensive sectors, with their increasing use of involuntary part-time work, that are driving the growth of involuntary part-time work. In these four key industries combined, involuntary part-time work grew twice as fast as part-time work generally between 2007 and 2015.

Table 4 shows even more clearly the stark contrast between the two sectors in which involuntary part time rose most—retail and leisure/hospitality—with another sector that behaved differently over the last eight years—construction. While in construction the use of part-time workers rose 0.2 percentage points between 2007 and 2015, it rose far less than in retail (3.4 percentage points) and leisure and hospitality (3.6 percentage points). The entire rise in construction is due to workers working part-time because they could find only part-time work, and not due to any rise in slack work. In leisure/hospitality the growth in involuntary part-time work was equally due to slack work and the inability to find full-time work. In retail the growth of involuntary part-time employment was mostly driven by workers wanting full-time work but finding only part-time employment.

Table 4

Next, we examine actual paths in the monthly time-series data by industry since 2003 in the same sectors discussed in Table 4. Specifically, the trend in involuntary part-time employment as a share of employment (by major reason for working part-time) is shown for construction, retail, and leisure/hospitality in Figures C, D, and E.17 While all industries exhibit commonalities such as a clear cyclical pattern, with involuntary part-time employment increasing in the recession, the recession and recovery had a stronger impact on construction (Figure C). Involuntary part-time employment spiked in 2008–2009 in all three industries and the subsequent decrease, unsurprisingly, is far more prominent in the “slack work or business conditions” reason.18 The most revealing pattern is that the proportion of those at work part time involuntarily because they “could find only a part-time job” has yet to recede much in any of the sectors.19 Although involuntary part-time employment in construction rose to greater heights during the recession, by the end of 2015 involuntary part-time constituted a greater share of workers in retail than in construction (true for 2015 as a whole, as Table 4 shows).

Figure C
Figure D
Figure E

In the retail trade, total involuntary part time has only nudged downward during the recovery period, and with a far longer lag time than in the construction sector. In fact, Figure D shows how involuntary part-time employment is barely relenting in retail trade. The share of retail workers involuntarily working part time more than doubled from its values just prior to the start of the Great Recession to a little over 9 percent at many points between 2009 and 2013. Indeed, the share of retail workers involuntarily working part time because they “could find only part time” was only slightly lower at the end of 2015 than at its peak levels in 2010 and 2011. As Table 4 shows, in 2015, it was up about 2 percentage points from its prerecession share in 2007. Moreover, in retail, “could find only part time” now actually exceeds “slack work” as the main reason for working part time involuntarily. So, whereas involuntary part-time work in construction has largely returned to prerecession levels, albeit with a lag of almost eight years and mainly because of the decline in slack work, the trend in retail represents more than just changes attributable to the cycle.20

Figure E shows employment patterns within the leisure and hospitality industry, a sector in which employment has grown by more than one percentage point since the start of the recession (from 8.7 to 9.8 percent of total employment).21 In any case, the leisure and hospitality industry somewhat mirrors the retail sector. While involuntary part-time work actually escalated to as high as 14.3 percent of employment in January 2011, it has gradually receded to about 8.5 percent. After receding from its recessionary spike, the share of workers who are involuntarily working part time because of slack work has gradually declined, although remains a little elevated above its prerecession level. As in retail, the share of leisure/hospitality workers who are involuntarily working part time because they could find only part-time work escalated substantially during the recession; less like retail trends, this share continued elevated for years, until edging only slightly downward since 2014.

In sum, the data suggest both cyclical adjustment, plus some kind of ongoing “structural” change that has increased the share of workers who are working part time involuntarily since the Great Recession. This appears to be concentrated in at least two key service sectors, retail and leisure and hospitality. Time series estimation techniques, indeed, show that there were at least some unique, albeit temporary, changes in the monthly time series behavior of the part time for economic reason data series.22 The jury is still out on whether there has been a permanent structural change in either or both involuntary and voluntary part-time working patterns, and whether this change is shaped by worker responses in addition to the observed, presumably employer-led adjustments toward greater use of part-time jobs.23

Demographic breakdowns show that race, age, and occupation affect likelihood of working part time involuntarily

This section examines the demographics and occupation and industry of part-time employment (voluntary and involuntary) using the 2015 Current Population Survey data sample that reported being “at work in the last week” and identified their weekly hours.24 This leaves out those who were absent from work all week because of illness, vacations, or other reasons.25 This analysis follows the Bureau of Labor Statistics and focuses on wage and salary employment when providing industry breakdowns, leaving out unincorporated self-employed and unpaid family members. This focus reduces involuntary part-time employment estimates by 10.4 percent, or by 655,000.26

Table 5 presents the number (top panel) and share (bottom panel) of persons in various gender and racial/ethnic subgroups who are at work and working part time, broken down into type of part-time status (voluntary and involuntary, and, for involuntary, for what major reason—either because they “could find only part-time work” or because of “slack work”). The table shows each demographic group in turn. For example, the third data row in the bottom panel shows what share of all at-work women work part time, what share work part time involuntarily, and what share work part time voluntarily.

Table 5

As the table shows, women are roughly twice as likely to work part time as men (24.5 percent do so, compared with 12.4 percent of men), but this is almost entirely driven by the differential rates at which men and women voluntarily work part-time hours: 19.5 percent of women are voluntary part-time workers while only 8.4 percent of men are. On the other hand, both men and women have similar rates of involuntary part-time employment: 4.0 percent of men and 5.1 percent of women work part time involuntarily. Women are more likely than men (2.1 versus 1.4 percent) to be involuntarily working part time because they could find only part-time work.

A similar share of whites, blacks, and Hispanics work part time, around 18 percent, but Hispanics and blacks are relatively much more likely to be involuntarily part time (with involuntary part-time shares of 6.3 percent, 6.8 percent, and 3.7 percent for blacks, Hispanics, and whites respectively). Although blacks and Hispanics comprised just 27.9 percent of those at work, these groups represented 41.1 percent of all involuntary part-time workers (calculations based on employment counts in table but not shown). There is a greater incidence of involuntary part-time employment among blacks and Hispanics because they are more likely to be unable to find full-time work and to face more slack work conditions. Black and Hispanic women (as well as women of “other race/ethnicity”) are most likely to experience involuntary part-time employment, and though not shown in the table, they represented 21.1 percent of all involuntary part-time workers in 2015 (author’s analysis of CPS microdata).

In summary, Table 5 and related data show that the groups most likely to be working part time involuntarily are blacks and Hispanics, particularly black and Hispanic women (who together made up about a fifth of all involuntary part-time workers in 2015).

Table 6 presents similar information to that of Table 5 but the subgroups examined for their likelihood of part-time work are men and women of different ages. Young male and female workers are much more likely to work part time, especially teenagers, and also are most likely to be working part time involuntarily. Older workers are more likely to work part time than prime-age (25 to 54 years old) workers but are less likely to be working part time involuntarily.

Table 6

Nevertheless, author’s analysis of data on the share of part-time and involuntary part-time workers falling into the various age categories show that prime-age workers comprised 57.8 percent (3.5 million) of all involuntary part-time workers (6.1 million) despite being only 44.0 percent of all part-time workers. Women and men comprised, respectively, 31.2 and 26.6 percent of all involuntary workers. Prime-age women comprised a similar share of involuntary part-time workers as they did of all part-time workers, around 31.0 percent.27 Prime-age men, on the other hand, comprised just 13.1 percent of all part-time workers but 26.6 percent of all those working part-time involuntarily with their high share caused by slack work.28

In summary, Table 6 and related data show that while young workers are most likely to be working part time involuntarily among all age groups, prime-age workers make up a substantial share (57.8 percent) share of the involuntary part-time workforce, a share that exceeds their 44.0 percent share of part-time workers overall.

Table 7 presents the likelihood that workers in given industries work part time and part time involuntarily. Part-time work is most intensive, by order of intensity, in leisure and hospitality, retail, other services, and education and health services (although the share is actually greatest in jobs found not in firms but in private household type services, such as cooks, maids and gardeners). It is least intensive in public administration, manufacturing, and mining. Involuntary part-time work is most prevalent in leisure and hospitality, retail, and private household work.

Table 7

As noted earlier in the discussion of the shares of part-time workers falling in certain industries, roughly 75 percent of all part-time workers and 70 percent of all involuntary part-time workers were in the four sectors discussed in the previous section of this report: retail, leisure and hospitality, education and health services, and professional and business services. Although not in these four sectors, construction stands out as an industry with a high share of involuntary part-timers due to slack work. Retail and leisure and hospitality make up a large and disproportionate share of workers who report that they are working part time because that is all they could find: these two sectors account alone for half (51.4 percent) of all of such workers able to find only part-time work.29

Last, Table 8, like Table 2, reviews the share of persons at work part time (all and involuntary only here the goal to understand which occupations contribute most to involuntary) part-time working. Service occupations, which provide 17.2 percent of all persons at work but represent double that (34.5 percent) of all involuntary part-time workers. The largest two are food preparation and serving, and cleaning and maintenance jobs. Sales and office occupations have one of the larger concentrations of involuntary part-time employment (comprising 6.5 percent of total, not shown). Sales and office occupations account for 26.8 percent of all involuntary part-time employment. Together the service and sales/office occupations account for 70.9 percent of all the workers who work part time because they could  find only part-time work. Because these same occupations also provide a substantial amount of the voluntary part-time employment (61.9 percent), that means involuntary part-timers are often concentrated alongside the voluntary part-time workforce.30

Table 8

The Affordable Care Act’s employer mandate did not cause the rise in involuntary part-time work

The ACA defines “full-time” employment as working an average of 30 hours per week. Larger employers not offering health insurance coverage for full-time workers are assessed a fee to help finance federal subsidies that eligible workers could receive when enrolling in individual insurance exchanges (see Kavaliro 2015). This “employer-shared responsibility” provision in the law, known generally as the “employer mandate,” could have increased involuntary part-time employment if employers had responded to the mandate by reducing the hours of existing or new employees to just below 30 per week.31 However, if the ACA had boosted part-time employment, we would see a number of trends in the data that suggest a structural change in involuntary part-time working or hours worked. First, the universe of workers vulnerable to having hours cut back due to the incentives of the mandate would have to be significant, in the aggregate or by industry. Second, there would be a clear decrease in average hours worked per week at the time of implementation (or perhaps its announcement) that is not accounted for by overall economic conditions. Third, there would be a greater cluster of workers appearing at or near 29 hours per week. Fourth, there would be a shift in the composition of net job growth, toward involuntary part-time work (for both reasons) and away from full-time work.

Finally, a separate but related trend could be at work and affect our observations. There is some evidence that the ACA has reduced a form of job-lock; workers who had felt compelled to accept full-time work because it was the only way to obtain affordable health insurance coverage are now free to take the fewer hours of work they desire. Thus, alternatively, we might observe a slightly increased number of voluntary part-time workers, representing a net gain in worker well-being, due to the ACA.

Because the implementation dates of the ACA were announced back in 2010, and the act was implemented in 2015 (after a delay of one year by size of employer), these effects should be observable in time-series data. Generally, data do not support the hypothesis that the ACA has been a primary driver of trends in involuntary part-time work, as explained below.

Studies predicting the effects of the ACA are mixed

One study predicted that the effect of the per full-time employee fee on large employers not offering healthcare would introduce new labor costs that would prompt employers to reduce both employment and hours per person, and thus aggregate working hours, by about 3 percent (Mulligan 2015).32 Another study predicted the ACA would lead covered employers to increase their use of workers averaging less than 30 hours per week, particularly in industries with the largest concentrations of these workers (e.g., retail trade, hotels and other accommodations, and food services), given that health care reform in Massachusetts modestly increased part-time employment, at least among the less-educated segment of the workforce (Dillendera, Heinrich, and Houseman 2016).

But findings from other studies suggested little impact. Under health care reform in Hawaii, employers must offer health insurance to employees who work more than 20 hours per week. Although more inclusive than the ACA, the Hawaii law was found to have led to an increase in the share of workers working fewer than 20 hours by only 1.4 percent. Indeed, the Massachusetts law, with a much smaller per employee fee ($295) than the ACA, had no discernable effects on either average hours or part-time working (Buchmueller, DiNardo, and Valletta 2011).

The universe of potentially affected workers is small

A study by Graham-Squire and Jacobs (2013) quantifies how many workers are exposed to the risk that their employers, looking to evade the ACA employer mandate, would reduce workers’ hours because they could relatively easily reduce these hours. According to the study this universe includes only workers at large firms that do not currently offer health insurance. This subset further narrows to workers who are close to the ACA hours threshold (which the authors define as working between 30 and 36 hours per week.) It turns out that this is a small subset of workers, mostly because a very large share of large employers already offer health insurance to their full-time workers (most uninsured workers tend to work for smaller firms, not affected by the mandate).33 Thus, only about 1.8 percent of the workforce meets these criteria, about 2.3 million workers. Moreover, they are concentrated in a few industries—retail, restaurants, accommodations, and health care. Thus, Graham-Squire and Jacobs (2013) show that the empirical bite of the ACA provisions on involuntary part-time work is likely to be small.

Still we might expect that if there are observable increases in the hiring of part-timers at just below 30 hours, the splitting of existing full-time positions into part-time jobs, or the hiring of more employees who are part-timers rather than full-timers, then the ACA might be responsible for some of the rise in the share of people who are working part time for economic reasons, i.e., involuntarily.

There is no clear decrease in average work hours or part-time job creation coinciding with ACA implementation that is not accounted for by overall economic conditions

There is little evidence that the ACA employer mandate provisions are the driving force behind the trends in involuntary part-time work. As noted above, the case for the ACA being a driver is quite weak, as only a small number of workers are even potentially vulnerable to easy hours adjustment by employers responding to the ACA. Here we show that direct estimates of the ACA’s effects, holding other economic trends constant, generally fail to show consistent, large effects. And as the next section shows, examining the composition of involuntary part-time employment by hours worked reveals that much of the increase in involuntary part-time work is actually accounted for by employees working more than 30 hours per week, and thus their schedules are unlikely to be dictated by employers wishing to avoid the mandate.

Between 2010, when the Affordable Care Act was passed, and 2015, when the employer mandate was implemented, the number of part-time jobs decreased (ADP 2015). And in recent years, there has been a clear, steady increase in full-time jobs and a declining overall share of the workforce employed in part-time jobs (Hotchkiss 2015). Few known studies have found a statistically significant effect of the ACA on involuntary part-time work. Even and MacPherson (2015) find that involuntary part-time employment was significantly higher (by roughly 1 million workers) in 2014 than would have been predicted based on historical relationships (since 1994) between overall economic conditions and the composition of jobs and workers in the labor market. They find a higher probability of being involuntarily part-time employed, since 2010, in those industries and occupations most likely to be affected by the ACA mandate but that the probability was unchanged in industries and occupations with the smallest number of workers likely affected by the mandate.

However, Valletta, Bengali, and van der List (2015) find that cyclical factors can account for all of the rise and fall of involuntary part-time work in the years during and following the Great Recession. They trace any changes in involuntary part-time employment during the period primarily to separations of workers from full-time employment, not any conversion by employers of full-time jobs into part-time jobs for the same workers.

Importantly, if involuntary part-time work was on the rise because employers were adopting new employment and scheduling practices in the wake of the ACA, the behavior of workers cycling into and out of involuntary part-time work would be much different. However, the data show that involuntary part-time employment is a remarkably transitory state, and most workers who move in and out of that state come from or leave to full-time employment, or cycle in and out of part-time work in the same job (Borowczyk-Martins and Lalé 2014; 2016). These are not patterns consistent with a change in employer behavior suggested by those arguing the ACA mandate has increased the incidence of involuntary part time.34

Finally, even though employers may have said they would hire more part-time workers, the numbers don’t suggest they actually did follow through on that. A 2015 survey by the National Small Business Association found 12 percent of small employers said they were hiring more part-time workers due to the ACA, and 7 percent said they were reducing employee hours (NSBA 2015).35 But in 2014, the year before the mandate went into effect these intentions were actually higher—15 percent of employers had stated that they would be hiring more part-time workers, and 10 percent said that they would be reducing employee hours. The most direct evidence of employers’ intentions in response to the ACA is from surveys conducted in late summer 2014 by human resource management services company ADP (ADP 2015). About 38 percent of respondents indicated that they would be adjusting employee hours. Yet, if employers did indeed reduce hours to keep employees below the 30-hour threshold, one would expect to see an increase in part-time jobs and fewer hours worked per job. However, employment data from the ADP shows that the part-time workforce actually grew more slowly (2.7 percent) than the full-time workforce (3.8 percent) between the first quarter of 2014 and the first quarter of 2015. In short, the evidence gleaned from asking employers about their intended behavior and from looking at their actual behavior shows little impact of the ACA.

More workers are not clustering near the 29-hour mark

Because part-time is defined as working less than 35 hours by the BLS but as working less than 30 hours by the ACA employer mandate, studies that focus just on changes in the under 35-hour category may miss additional information gleaned from focusing more specifically on the under 30 category.36 Mathur, Slavov, and Strain (2015) focus on trends not only in all involuntary part-time employment, but employment just above and below the 30-hour threshold below which employers are not subject to the ACA mandate. They tested whether shifts occurred equally or not across two categories of part-time work, as defined by the BLS. They find some evidence of a relative shift from the 31–35 hour category to the 25–29 hour category after the passage of ACA in March 2010. However, unlike Even and MacPherson (2015) found, this shift was no more pronounced among workers in industries and occupations most likely to be affected by the mandate (Mathur, Slavov, and Strain 2015). This sheds doubt on whether the mild increase in clustering they find below 30 hours is driven mainly by the ACA mandate.

Furthermore, the ADP survey of employers finds little to no change in the percentage of workers in any of the three categories of part-time hours worked: less than 30 hours, 30 to 34 hours, and more than 34 hours. This distributions of hours worked were quite similar across all quarters in 2013 and 2014. Indeed, there was no observed increase in part -time employment relative to full-time employment; in fact, the reverse occurred.37 Thus, at best, if the ACA is pushing employers to reduce hours, these effects are being well cloaked by economic improvements driving reductions in part-time work overall (ADP 2015).

Using the same CPS data as Even and MacPherson (2015), for the period 2005–2015, Moriya, Selden, and Simon (2016) find only a very small increase in the adjusted probability of working either 25–29 hours or fewer than 25 hours in 2013, 2014, or 2015. Any upward trend in such probability began before passage of the ACA. In addition, they observed no large reductions in the frequency of working 30–34 hours.

Finally, Garrett and Kaestner (2014; 2015) also find little evidence of increasing involuntary part-time work immediately before the ACA employer mandate was implemented. They do find a significant but small increase in part-time work during 2014 relative to what would be expected at that point in the economic recovery based on prior experience (since 2000). However, while this overall increase in part-time work is fully attributable to an increase in involuntary part-time working, it is not specific to the category of part-time work most likely to be affected by the ACA employer mandate—part-time work of less than 30 hours. The small change identified by Garrett and Kaestner (2014) applied just as much to part-time work between 30 and 34 hours per week as it did to workers clocking in fewer than 30 hours per week. Finally, they find that transition rates of workers moving between full-time and part-time work indicate that any increase in part-time work in 2014 is not ACA-related, but more likely due to a slower than normal recovery of full-time jobs following the unusually deep and long recession. Thus, despite a timing coincidence between higher involuntary part-time employment (working less than 35 hours per week, as defined by the BLS) and the appearance of the ACA, there is no evidence that employers are hiring workers for fewer than 30 hours due to the ACA employer mandate (Mericle and Reichgott 2015).

Response to the ACA suggests it is a worker-friendly policy

One additional potential outcome from the ACA stipulation that 30 hours per week is the threshold above which workers are considered full time and thus eligible for health care is that the hours of work of “full-timers” might voluntarily drop down.38 Some workers who prefer to work closer to 30 hours no longer need to stay in a job with 40 hours simply to qualify for health insurance coverage. In addition, because the ACA mandated that those under the age of 26 be eligible for coverage by their parents’ insurance, this may have reduced “job lock” (not leaving a job due to fear of losing health coverge) for young adults and allowed them to reallocate their time less toward work and more toward other activities (Colman and Dave 2015). Indeed, the biggest increase (of 5 percent) of voluntary part-time work in the first seven months of 2014 was among workers age 16 to 35 (Jorgensen and Baker 2015). Gooptu et al. (2015) try to measure a similar potential labor supply effect of the Medicaid expansion included in the ACA, implemented in 2014. They find that there might have been a slight reduction in the amount of hours worked per week (i.e., labor supply), presumably from those low-income workers who had been working more weekly hours than they actually wanted in order to be eligible for their employer’s health insurance coverage. Thus, there may be a well-being enhancing effect of the ACA, facilitating better matching between workers’ preferences and their weekly hours of work, and consequently, some work hours could be shifting toward those involuntarily working part time who would prefer extra hours and income.39

Analysis of trend data supports research findings on the absence of significant ACA effects

In sum, today’s persistently high level of involuntary part-time employment is attributable less to the ACA and more to the typical cyclical sensitivity of involuntary part-time working (Valletta, Bengali, and van der List 2015; Borowczyk-Martins and Lalé 2014, 2016) and to some other structural forces at work.

The recent time series trend data examined in this report support the conclusions from the research reviewed above. In 2014, when the ACA employer mandate was originally supposed to go into effect, and 2015, when the employer mandate began to go into effect, there were no noticeable breaks with trends in part-time job creation, workweeks, ratio of involuntary to voluntary part-time working, and reason for working part time. Indeed, the share of workers working part time involuntarily due to the “inability to find full time” has dropped a bit in 2014 and 2015, after remaining stubbornly high in the recovery. Moreover, the average workweek of part-time workers increased from the recovery through 2015. Thus, any effects of the ACA have been dwarfed by cyclical forces or other structural changes that appeared prior to 2014. Finally, because about 70 percent of persons working part time because they “could find only part time” are in just a few industries (see also Glosser and Golden 2016), this industry concentration, rather than dispersion, reinforces the conclusion earlier in this report that structural changes outside the ACA are behind the painstakingly slow decline in involuntary part-time work after the Great Recession, and the escalated rate of workers who are part time because they could find only part-time work.

Part-time workers have lower earnings, work less than they’d like, and often work irregular hours

We care about the level of involuntary part-time employment not only because the afflicted people are earning less than they would prefer but because part time jobs offer relatively less pay and weaker benefits, and more variable and unpredictable daily or weekly working hours.

Part-time jobs come with a wage and benefit disadvantage

Part-time jobs come with a wage disadvantage relative to full-time jobs (see Hirsch 2005; Glauber 2013). Indeed, the most salient disadvantages faced by part-time workers are in their relatively lower rates of pay (what they are paid per hour) and benefit coverage. Generally, part-time jobs pay little more than half the hourly rate paid in full-time jobs (CPD 2016). More precisely, the part-time wage penalty (the percent by which hourly part-time wages lag full-time wages) for men is between 46 and 49 percent and for women it is between 22 and 26 percent. When adjusting for differences in personal, educational, locational, industrial, and occupational characteristics of workers, the part-time wage penalty for men is 19 percent and for women, 9 percent (Hirsch 2005). The part-time pay penalty is somewhat larger for those working only 1 to 20 hours per week relative to those working 21 to 34 hours.

The wage penalty appears even steeper for involuntary part-time workers. Among those who are paid by the hour, voluntary part-time workers earn $15.61 per hour on average compared with only $15.11 for those working part-time involuntarily. Among those who could “find only part time work” their hourly earnings were even lower, $14.53.40

Part-time workers have considerably less access to benefit coverage than full-time workers (Shaefer 2009; BLS 2015; Bishow 2015).41 For example, part-time workers have only one-third the access to health insurance coverage as do full-time workers—25 percent compared with 73 percent. And only 14 percent of part-time workers actually participate in an employer plan because the take-up rate is relatively lower among part-time than among full-time employees. And the degree of part-time work matters. Access to health insurance coverage for those working 30–35 hours per week is 20 percentage points higher than for those working in the less than 30 hours category—44 percent compared with 24 percent. Similarly, access to retirement benefits has a noticeable cutoff at 35 hours (Bishow 2015). Moreover, in industries with relatively “lower average hours” per job—such as accommodation and food services; arts, entertainment, and recreation; and retail trade—the differences between the coverage rates of full- and part-time workers are even starker.42

In 2015, Rutgers University published a survey of almost 1,000 part-time workers (Zukin and Van Horn 2015).43 Among other things, it provides useful information on the economic hardships of part-time workers, and involuntary part-time workers in particular. Part-time workers were considered “involuntary” if they answered “yes” to the question, “Do you want a full-time job—35 hours a week at one place?” (By reason for working part-time involuntarily, “cannot find full-time work” was the predominant “main” reason—54 percent compared with 34 percent citing “not enough work or poor business conditions.”44) By income level, the involuntary part time are overrepresented in the $30,000 and below earnings category and underrepresented in the $100,000 plus categories. As many as seven in 10 of involuntary part-time workers say they earn less money than they need to survive and support families—which the study’s authors characterize as experiencing an “economic nightmare.” Moreover, about 44 percent said they make a lower wage than full-time colleagues doing the same work.

Part-time workers work fewer hours than they want

An obvious adverse impact of being an involuntary part-time worker is that you work far fewer hours than you want with a corresponding reduction in your weekly wages. In 2015, involuntary part-time workers had an average workweek of 22.2 hours, barely over half that of the 42.4 hour workweek of full-time workers.45 Those who are involuntary part-time because they could not find part-time work averaged 23.6 weekly hours while those suffering from slack work had just 20.0 weekly hours. Involuntary part-time workers do work more than voluntary part-timers, who averaged 19.0 weekly hours in 2015.

In the Rutgers University survey, median “hours a week [they] normally work, on average” was 25 for the entire sample, 24 for voluntary part time workers, and 29 for involuntary part-time workers. Those with one part-time job report working a median of 20 hours per week voluntarily, but 25 if involuntarily. Reported hours rose to 35 hours if one were involuntarily part-time and had more than one job and to 50 hours per week if one of those other jobs were full time (Zukin and Van Horn 2015). Among those involuntary workers who came into a part-time from a full-time position, 32 percent report their current employer has reduced their hours, almost double the rate of only 17 percent of voluntary part-timers who say this happened to them.

Effectively being “half unemployed” has consequences in the form of reduced income and well-being, relative to that provided by full-time jobs in similar industries or workplaces (Friedland and Price 2003; Otterbach 2010; Anderson and Winefield 2011; Wilkins and Wooden 2011; Bell and Blanchflower 2013; Angrave and Charlwood 2015; Bassanini and Caroli 2015). Underemployed workers also may adversely affect employers, in terms of performance, absenteeism, and turnover rates (McKee-Ryan and Harvey 2011; Wang 2014).

Part-time workers are more likely to work irregular workweeks and shifts

The Current Population Survey asks respondents to identify their “usual weekly hours” of work. While they are prompted to report a specific number, if they do not, their response is entered as “hours vary.”  Table 9 shows that part-timers experience hours too variable to specify at a rate 2.5 times higher than among full-time workers: The rate is 4.3 percent among full-timers but 10.9 percent among part-time workers. Moreover, if the part-time work is involuntary and the reason is slack work, the variable hours rate is 21.5 percent, which is considerably higher than the part-time rate overall. In contrast if the part-time work is involuntary because they could find only part-time work, the rate is a modestly higher 11.3 percent.46

Table 9

A recent poll (EINet 2015), conducted by Public Policy Polling (PPP) includes “hours vary too much to say” as a response to a question about one’s usual hours per workweek, whereas in the CPS, it is a category created after the fact, for respondents who didn’t specify a number of hours. Relative to the CPS’s 7 percent, a much higher percentage—55 percent—of those polled by PPP report that their weekly hours are variable.47

Figure F shows that an astonishingly high 55 percent of those surveyed by PPP said their “hours vary,” i.e., are not the same every week. Although not shown in the table, those paid hourly and or paid some “other way” are more likely to report that their hours vary than are salaried workers.48 Indeed, in the PPP poll, part-timers comprise 28 percent of those whose “number of hours worked vary from week to week” although they make up only 22 percent of the poll’s national sample.49

Figure F

Table 10 also shows that while workweeks are most variable for those working in agriculture, they are also highly variable for those employed in food services and production, retail and wholesale trade, and transportation and utilities. Food services and retail are two sectors where part-time employment is most intensive. Thus, part-time employees are most prone to having their hours vary, in part because such jobs are concentrated in industries where part-timers are disproportionately concentrated.

PPP has conducted further polling in individual states, with differing wording of survey questions. In New Mexico and in Virginia, roughly a third of workers have variable hours. While this state estimate is lower than the 55 percent national finding, it suggests a level of instability in workweeks still far above the 6 percent figure produced by the CPS survey, which, again, does not explicitly ask workers if their hours vary too much to say.

In Minnesota, the pattern for part-timers is characteristically different in another key respect: while almost half of full-time workers’ schedules never change, only one in five part-timers’ schedules never change.50

Part-time workers were also asked what were their recent “fewest hours worked in a week.” Among just regular part-time workers (who constituted 21 percent of the sample), the majority—53 percent—worked less than 20 hours, and 84 percent worked shorter than 30 hours.

As many as 9 percent of “regular part-time” workers report that 35 or more hours were the “fewest hours worked” in a week. Thus, one in eleven workers who are full time according the BLS definitions consider their job “regular part-time.” On the one hand, this full-time level of weekly hours might be welcome news for those who are in part-time jobs but prefer full-time, but it speaks to the high variability of part-timers’ workweeks generally. Most part-timers on average tend to get a week or two advance notice of their upcoming schedule; however, they are less likely to get four weeks or more notice than full-timers.51

And part-time workers experience more frequent schedule changes once schedules have been posted. When asked if schedules change after posting, over 40 percent of part-timers answer “regularly” or “sometimes,” versus only 23 percent of full-timers.52 Such just-in-time scheduling practices are one of the key sources of the variable, unstable hours that part-time workers are more likely to experience (Watson and Swanberg 2013; Luce, Hammad, and Sipe 2014; Ruan and Reichman 2014; Alexander and Haley-Lock 2015).

The National Institute for Occupational Safety and Health’s General Social Survey (GSS) provides data that probes deeper into the irregularity of workers’ schedules. The GSS Quality of Work Life (QWL) Supplement asks respondents how their “usual work schedule” is “best described.” Workers in part-time jobs are more likely to report being on an irregular shift/on-call schedule, rotating shifts, or split shifts than are full-time workers. Sixteen percent of part-time workers said they work irregular shift/on-call, double the rate of 8 percent for full-time workers.53 In a very real way, those reporting irregular shift/on-call, rotating shifts, and split shifts can all be considered workers with irregular schedules. When including all three responses, about one-fourth of the part-time workforce has an irregular schedule, compared with only 15 percent of full-time workers.54 Thus, part-time jobs are much more likely to be characterized by irregular schedules.

Why does this matter? As studies have shown, employees who work irregular shift times, in contrast with those with more standard, regular shift times, experience greater work-family conflict, and sometimes experience greater work stress. Working part time is only associated with a lower level of work-family conflict if it is voluntary (Golden et al. 2011; Golden 2015a).55

Work-family conflict is also suggested by Zukin and Van Horn (2015), who found that as many as one-third of involuntary part-timers cited getting “less favorable work schedules,” such as on holidays and weekends, in contrast to just 10 percent of voluntary part-timers. Respondents were presented with seven possible advantages of part-time work. The advantage of being “able to set” one’s own “schedule” was agreed by most workers to be the greatest potential benefit, which was shared equally among involuntary and voluntary part-timers, above 60 percent.

Nonetheless, as many as 55 percent of all involuntarily part-time workers regard it “difficult to schedule work and family obligations on a daily or weekly basis,” whereas only 23 percent of voluntary part-timers felt this way. Thus, not surprisingly, about three in 10 of those working part time involuntarily are “dissatisfied” with their jobs—this is about three times the proportion of “dissatisfied” in part-time jobs voluntarily—while only about one in 10 involuntary workers are “very satisfied” with their jobs, far lower than the 43 percent of voluntary part-time workers who were very satisfied. Involuntary part-timers were also more likely to experience feelings (during “a lot of the day yesterday”) of anger, worry, stress, sadness, and physical pain, with less enjoyment and happiness. This speaks to the likely lower subjective well-being experienced among persons working part time involuntarily.56

Seven policies to address causes and consequences of involuntary part-time work

Involuntary part-time workers who cannot get the additional work hours they want from their current employer(s) often change jobs or take on multiple jobs. However, from a policy standpoint, helping these workers get the additional hours they prefer at their existing job or employer would be more effective and far less disruptive to them, their employers, and the labor market. Involuntary part-time workers are also more economically vulnerable (Glauber 2013).

Recognizing this, this section describes the landscape of policies that are being innovated to help improve the prospects for part-time workers, including those wishing full time jobs. This landscape is fluid, with ongoing, state, local and national proposals. The first set of policies attempt to reduce the num­ber of Americans in involuntary part-time employment.

A second set of policies would improve the quality of part-time positions. A third set would reduce the economic vulnerability of part-time workers whose hours are shortened or take part-time jobs in lieu of preferred full-time (e.g., Glauber 2013 outlines the ways part-time workers are economically vulnerable).

These three sets combine in a comprehensive, seven-pronged approach: Pay and benefits parity for part-timers; innovative soft-touch regulations providing workers with rights to request minimum and maximum weekly hours; fair scheduling practices and rights to access additional hours; minimum reporting pay and predictability pay; unemployment insurance reform; overtime regulation reform; and creating “steps” rather than a single “cliff” in the ACA Shared Responsibility provisions to cushion any potential unintended consequence of employers cutting work hours, even with the limited evidence of its occurrence.

Guaranteed hours, pay and benefits protections for part-time employees

This report and others (see for example Glauber 2013) have highlighted the economic penalties associated with involuntary part-time employment. Congress should alleviate these penalties with policies that enhance the quality of part-time work. First, it should pass legislation requiring that part-time workers receive a benefit package equivalent to that of full-timers, with all benefits prorated to reflect the differences in hours worked. This would reduce the number of Americans working in involuntary part-time positions, as employers would no longer stand to gain as much from employing part-time workers (Tilly 1990). The International Labor Organization (ILO), the international body that issues and monitors minimum standards for treatment of workers, adopted a Part-Time Work Convention in 1994 (No. 175), which inspired the European Union to pass its own prohibition of discrimination against part-time workers in 1997 (as part of the EU Council Directive 97/81/EC). This in turn inspired other countries (including South Korea, Mozambique, Russia, Bulgaria, Sweden, and Turkey) to improve the quality of part-time work. (Messenger and Wallot 2015). The key is to practice prorating of compensation so that part-time employees receive pay and benefits proportionate to their weekly hours.

Related to assurance of parity in pay benefits are policies that set minimum and maximum work hours that jobs must offer. Some countries require that part-time workers get a minimum of work hours. In Algeria, part-timers must receive not less than half of the statutory working time. Denmark allows collective agreements to prescribe a minimum of 15 hours per week for part-time work. France provides a minimum target of 24 hours per week for part-time workers. Some countries also have maximum thresholds for part-timers—such as not to exceed 30 hours per week or 120 hours per month—which may prevent “involuntary full-time” working, i.e., unwelcome additional hours. Overtime is sometimes outright prohibited for part-time workers (Brazil), or allowed only in emergency cases (Argentina). In South Korea it is allowed but requires the agreement of the part-time worker, cannot exceed 12 extra hours per week, and comes with a pay premium for work beyond the normal work hours of full-time employees. Finally, some laws aim to facilitate the transfer from full-time to part-time work and vice versa. In Romania, employers must, as much as possible, take into account the demands of their employees to be transferred either from full-time to part-time work or vice versa.

In some countries, a part-time employment contract must indicate the number of working hours and their distribution; additionally in Romania that contract must include cases where the work schedule may be amended, while in France and Portugal employment contracts must include a provision that if these requirements are not fulfilled, the contract is deemed to be a full-time contract (Messenger and Wallot 2015).

For the entitlement to annual paid leave, countries have instituted various formulas. Japan for example has a threshold of 30 hours per week entitling the employee to the same amount of annual leave as full-time workers, plus a minimum number of days of leave for those who work less than 30 hours per week. Others, such as Brazil, specify the number of days of leave to which part-time employees are entitled according to the number of hours they work per week. In France, the same rules are used to calculate holiday pay for both part-time and full-time workers. Some countries require a minimum number of work hours (or earnings) to qualify for retirement pension contributions also received by full-timers, as in Ireland. In South Africa, employees who work less than 24 hours per month for an employer are not covered by the majority of their law’s working-time provisions.

The ILO provides three pertinent specific recommendations to improve the treatment of part-time employees while curbing the incidence of involuntary part-time work (Messenger and Wallot 2015):

  • Set some basic minimum work hours standards (similar in principle to current ILO and EU standards regarding maximum hours), stipulating appropriate penalties in the event of noncompliance.
  • Mitigate the vulnerabilities of “marginal” part-time workers (part-time workers who work very short hours, generally less than 15 hours a week) by using either premium hours pay after a standard “part-time workweek,” a fixed minimum compensation rate for “on-call” times not worked, or partial unemployment benefits.
  • Eliminate contractual provisions such as “exclusivity clauses” under which part-time workers might be involuntarily required to be available for work at all times.

In the US, the Washington, DC City Council recently passed the country’s first guaranteed hours law establishing a 30-hour minimum workweek for janitors in large commercial buildings. There are similar proposals in the City Council in Jersey City, NJ for janitors, security guards, maids and the State of Connecticut for its State building-mainentance workers.

Rights to request minimum and maximum work hours

Legislating a “Right to Request” that includes an employee right to request—free from fear of retaliation—a minimum floor of weekly work hours would reduce the incidence of what is effectively “zero hours contracting” (work arrangements such as on-call scheduling that employers use to meet fluctuating market demands). Right to request laws also typically involve a right to request shorter or part-time hours. Thus certain employees would be allowed to request both a floor and a ceiling—effectively a more stable number of weekly hours—and also block out days and times. The right to request modification of one’s schedule could be based on the need to care for a child, an elderly parent, or a family member with a serious health condition and—importantly for a growing share of the population—to accommodate schooling, job training programs, or a second job. Employers are simply required to interact with employees to determine if their requested schedule modifications are feasible, and if not, suggest an alternative.

Granting this request could include a reward for employers who accommodate and implement such a request, in recognition of the potential burden even if it did not to rise to the level of denying the request. Some advocates prefer to establish a minimum hours floor for all employees, such as guaranteeing hourly part-time workers a minimum number of hours when they are hired, whether it is 15, 20, 25, or 30 hours or something else (e.g., Cauthen 2011). However, this number may be better left to the input of the employees, in the event that they are not the typical involuntarily part-time worker but wish merely to remain part time with a few extra work hours or shifts per week. The key would be not so much to specify a set minimum but to provide workers some option to access additional hours of work at their existing job before resorting to multiple job holding or searching for a new job. In addition, granting requests of employees for shorter hours (at proportionally reduced pay) might free up available work hours for employees who are part time involuntarily, whether in the same workplace or in the industry generally (Golden 2015b).

Access to hours legislation

State legislation, and municipal ordinances such as those enacted in San Francisco and proposed in the District of Columbia, Seattle, and San Jose, California, would grant employees in certain industries (such as commercial cleaning crews in the District of Columbia) a right of first refusal if additional hours of work become available at the workplace (CLASP 2014). “Access to hours” would mean additional hours would have to be offered first to existing part-time staff, with some time lapse (e.g., one to three days), before hiring new part-time employees, temporary workers, or contractors. An “Access to Hours” system engenders a more transparent process for distributing additional work hours and enables many underemployed workers to realize their preferences for more work hours at their primary job.

The San Francisco Predictable Scheduling and Fair Treatment ordinance (2015) requires covered employers (large retail chains) to offer extra work hours to current qualified part-time employees in writing before hiring new employees or using contractors or staffing agencies to perform additional work.

Before hiring new employees or using contractors or a temporary services or staffing agency to perform work in a Formula Retail Establishment, an employer shall first offer the additional work to existing part-time employee(s) if: the part-time employee(s) are qualified to do the additional work, as reasonably determined by the employer, and, the additional work is the same or similar to work the employee(s) have performed for the Formula Retail Establishment. The employer need only offer the number required to give the employee 35 hours of work in a week, and maintains the discretion to divide the additional work hours among its part-time employees. The part-time employee may, but is not required to, accept the offer of additional work hours within some time frame (e.g., 72 hours) in writing, after which time the employer may hire, from the time they posted the additional hours in a conspicuous location, which includes electronic notice on their internal web sites.

In San Jose, the “Opportunity to Work Ordinance” was approved in April 2016 to be included by the city clerk as a November 2016 ballot measure. If passed, it would require San Jose employers to offer newly available working hours to qualified part-time employees before hiring more part-time staff or temporary employees. The employer would be free to set as short a time period as it prefers between announcing the availability of hours and making a new hire. Although San Francisco, Seattle, and SeaTac, Washington, have adopted similar ordinances, the San Jose initiative would be the first measure of its kind in the country to apply across all industries. The proposal was partly in response to survey findings that about 162,000 people employed in hourly jobs in San Jose have part-time or variable schedules at their main job, increasing from 26 percent in 2006 to 43 percent today (Center for Popular Democracy 2016).

In the District of Columbia, the “Hours and Scheduling Stability Act of 2015” would require not only compensation parity and advance notice of scheduling, but that employers must first offer additional hours of work to existing employees before hiring additional employees or subcontractors. In addition, the “Building Services Employees Minimum Work Week Act” requires janitorial or building maintenance service workers who work in an office building or office park (with over 350,000 net rentable commercial office square feet) to be provided with a minimum workweek of no fewer than 30 hours of work per week.

Stable scheduling, minimum reporting pay and predictability pay

Following in the footsteps of San Francisco, Seattle and Emeryville CA (with its large presence of retail and food companies) have adopted the most comprehensive work scheduling stability ordinances. Workers in retail and food service chains will now have greater input into their work hours, at least two weeks’ advance notice of shifts, compensation when shifts change unexpectedly, access to more hours when available and a minimum rest time between work shifts. San Francisco’s Retail Workers Bill of Rights, covers only chain retail stores, but it requires employers to provide: (1) a “good faith” written (but nonbinding) estimate of an employee’s “expected minimum number of scheduled shifts per month,” including the expected days and hours of those shifts, prior to the start of employment; (2) two weeks’ advance notice of scheduling changes that were not requested by the employee; (3) “predictability pay” (a premium of one to four hours of pay at the employee’s regular rate) if an employee is given less than seven days’ notice of a scheduling change, except in certain limited circumstances; and (4) written offers of additional work to qualified part-time employees before hiring new employees or using contractors or staffing agencies to perform additional work. Predictability pay would require employers to pay from one to four hours of compensation to an employee who is obliged by the employer to be on-call for a scheduled shift but not given any work.57Seattle and Emeryville adopted even tighter standards in terms of reporting and predictability pay.

To date, eight states have enacted laws requiring that if a shift, or remainder of a shift, is cancelled within 24 hours of its start, the employer must pay the employee for showing up.  In 2015, New York State Attorney General Eric Schneiderman spearheaded an inquiry into the on-call scheduling practices at 12 major retailers. Subsequently, 7 major retail brands announced an end to their use of on-call shifts (affecting 250 thousand employees). In 2016, New York was joined by eight other Attorneys General from California, Connecticut, the District of Columbia, Illinois, Maryland, Massachusetts, Minnesota, and Rhode Island, sending letters to 15 more retail companies inquiring about their use of on-call scheduling.

Unemployment insurance reform to reduce the frequency and incidence of underemployment

Part-time workers are disadvantaged in three ways by current unemployment insurance laws. First, because state laws require minimum earnings to qualify for benefits, many part-time employees do not earn enough to qualify. Part-time work, with its irregular hours, volatile earnings, and/or high turnover makes it less likely that a worker will have the “consistent work history” necessary to qualify for unemployment insurance (UI), if laid off (Cauthen 2011; Cauthen, Case and Wilhelm 2015; White House 2016). Weekly benefit amounts could be raised by states and the eligibility requirements relaxed to enable more part-time workers to qualify for benefits.

Second, employees can be disqualified from benefits in many states if they seek only part-time work, even if their qualifying employment and earnings came exclusively from part-time work. The UI Modernization Act gave grants to states to encourage them to allow otherwise monetarily eligible claimants with a part-time work history to search and be available only for part-time work. Nonetheless, as many as 21 states still require UI claimants to search for a full-time job. Eligibility for UI should be extended to more part-time workers and, indeed, to anyone who wants to reduce his or her work schedule for compelling reasons, including personal health and child care responsibilities. The federal government should enact a minimum standard in which workers qualify for UI benefits as long as the work being sought is for at least 20 hours per week (Glauber 2013).

Third, unemployed workers receiving UI benefits lose too much when they accept part-time jobs. Although all states allow workers to receive partial UI benefits while employed part time (as a way to encourage employment and cut state UI outlays), the rules vary enormously from state to state and in some states there is little to be gained by taking a part-time job that pays barely more than the claimant’s weekly benefit amount. Weekly UI benefits are always reduced because of earnings from part-time employment, but the best-designed policies cap earnings at 50 percent above the claimant’s full weekly benefit amount and disregard earnings up to 50 percent of a claimant’s full weekly benefit amount, or one-third of weekly part-time earnings. (Ben-Ishai, McHugh, and McKenna 2015; Ben-Ishai 2016; NELP 2015).

The Middle Class Tax Relief and Job Creation Act of 2012, which has expanded UI “work-sharing” programs to 29 states, provides another way to increase the income of involuntary part-time employees. Employers may shorten workweeks of employees in lieu of instituting layoffs, allowing employees to qualify for “short-time compensation” (STC) replacement, usually half of lost income (or that state’s income replacement ratio), if the work-sharing plan conforms to federal standards. Thus, work-sharing arrangements could act as partial UI compensation for those experiencing involuntary part-time workers during periods of “slack work.” Without such compensation for affected workers, reduced work hours would tend to slow the progress of an economic recovery (Reynolds and Wenger 2010). Work sharing with STC reinforces the built-in, automatic stabilizer role of UI. Indeed, in other countries where work sharing is more widespread, while work-sharing might induce a higher rate of underemployment during the first throes of a recession, those economies which rely on STC the most wind up with lower involuntary part-time working in the longer run (Messenger and Wallott 2015).

Overtime work regulations reform

Implementing the Department of Labor’s recent increase in the threshold for exemption from overtime pay requirements for salaried executive, administrative, and professional employees could indirectly help relieve the incidence or degree of underemployment.58 Salaried workers who earn below the threshold must be paid “time-and-a-half” for each hour worked beyond 40 hours per week (hourly workers in most service and blue-collar occupations already enjoy these protections), so raising the threshold raises the number of workers who qualify for overtime. To the extent employers start constraining salaried workers’ weekly hours to 40 to avoid paying an overtime premium, it may induce transfers of those workloads and thus work hours over to the firm’s or organization’s part-time workforce (Boushey and Ansel 2016). To the extent some employers do limit salaried employees to 40 hours per week, many of those hours could be channeled to hourly paid, part-timers. Many of these part-timers might prefer the additional hours, thus alleviating some underemployment, especially if there were a “right to request” or “access to hours” to facilitate hitting the target. The potential of the overtime rule to shift hours to part-timers who want them provides another good reason to support facilitating and protecting individually requested adjustments to work shifts, number of hours, and schedules (Golden 2015b; Boushey and Ansel 2016).59

Adjustments to the Affordable Care Act

The Patient Protection and Affordable Care Act delimits full-time employment at 30 hours, with employers required to pay a per-full-time-employee fee if they don’t offer a minimum level of health insurance. The fee could occur at steps rather than a cliff and be prorated, so that the fee gets higher the more hours the employee works. This could discourage even the small effects on the level of involuntary part-time working that some have attributed to the ACA. The state of Hawaii’s coverage requires contributions for all employees above 20 hours—this would avoid the complexity of prorating by “normal” hours and be even more inclusive.60There also could be one level above, at 35 hours, and another level below 30, at 20 hours, etc. For example, not only a threshold at 29, but a step at 19, too. A company would have to pay half a premium for up to 19 hours. Alternatively, the subsidies to workers could be reduced as the hours level rises, to counter the incentives indicated by Mulligan (2014). Alternatively, the pay “penalties” for each could be set at $1,500, $2,000, and $2,500, respectively. Thus, while adding some complexity, gradually stepping the penalty from 20 to 25, to 30, to 35 hours might avert most of any unintended side effects.

About the author

Lonnie Golden is a professor of economics and labor-employment relations at Pennsylvania State University, Abington College. He is also a senior research analyst for the Project for Middle Class Renewal, School of Labor and Employment Relations at University of Illinois as well as an affiliate for the Employment Instability Network at University of Chicago. He holds a Ph.D. in economics from the University of Illinois-Urbana. His research has centered on the labor market and hours of work—specifically the economic and non-economic determinants of hours, including legal, organizational, and individual preferences, and their effects on well-being, including work-life and worker happiness, and on the level of employment. He is co-editor of the books, Working Time: International Trends, Theory and Policy Perspectives (Routledge Press) and Nonstandard Work: The Nature and Challenge of Changing Employment Arrangements (Cornell University Press), and author of research articles that have appeared in journals such as Industrial Relations and Monthly Labor Review.


The author would like to acknowledge Lawrence Mishel, Stuart Glosser, Jin Dai, Susan Lambert and EINet, Carrie Gleason, Maggie Corser, Elianne Farhat, Jaeseung Kim, Ross Eisenbrey, Josh Bivens, Nicholas Blair, Teresa Kroeger, and Lora Engdahl. BLS economists Gaim Medhin and Lisa Williamson provided valuable assistance.


1. Underemployment is a measure of labor underutilization in the economy. The Bureau of Labor Statistics’ most comprehensive measure of underemployment, the U-6 measure, includes unemployed workers, people who would rather and could work full time but are not for economic reasons (“involuntary part-timers”), and people who want and are available to work but have given up (Mishel et al. 2012). But this measure does not capture workers who are highly skilled but working in jobs that do not use their skills. A forthcoming report will focus on alternative, broader measures of “underemployment,” of any workers, part time or full time, who have a stated preference for more hours of work and income as opposed to their same hours and income.

2. For example, Schneider, 2015; Bernstein 2016; Gillespie 2016.

3. See for example Yellen 2016.

4. Inadequate hours of work for many workers even appeared in politicians’ campaign talking points criticizing the state of the economy. Jeb Bush, campaigning in 2015, remarked, “It means that people need to work longer hours and, … gain more income for their families…That’s the only way we’re going to get out of this rut that we’re in.”

5. Data on total persons at work are obtained from CPS Table 19, select years, http://www.bls.gov/cps/tables.htm, and recent monthly data (CPS Table A-24, select months, http://www.bls.gov/opub/ee/2016/cps/monthly.htm and http://www.bls.gov/opub/ee/2015/cps/monthly.htm). Other series are from downloaded BLS data (LNS12032194, LNS12032195, LNS12032196, and LNS12005977).

7. See Sen and Razza (2015:10)

8. Data from unpublished Table 14 show an absence rate of 5.3 and 3.0 percent, respectively, for part-time and full-time employment.

9. This is derived from taking the share of total persons at work who were involuntary part-time workers in 2007 (3.1 percent) and applying it to the total at work in 2015, yielding 4,457,000 and comparing this number with the actual involuntary part-time employment of 6,365,000 in 2015, yielding an increase of 1,908,000.

10. “Slack work” and “business conditions” are reason numbers three and four in the survey. The other reasons for being on part-time hours for economic reasons, which presumably show little if any cyclical pattern, are “job is seasonal” and “job started/ended during week; while the remaining reasons (numbers 5 through 9) are considered “voluntary part time.” The 3 million workers who had time off during the week about which survey respondents answered are not treated as either voluntary or involuntary part time.

11. This effect represents the hours of labor supply component of the “added worker” effects often occurring during recessions (Appelbaum, Boushey, and Schmitt 2014). In contrast, a corresponding substitution effect of falling real hourly earnings leads at least some other part-time workers to disengage from the labor force, disappearing from the pool of employment, including part-time employment. Indeed, not participating in the labor force, particularly among younger workers, is a well-noted, recent trend (Toossi 2013; Blau and Kahn 2013). In addition, CPS respondents might have become more likely to attribute their part-time hours more to economic than to non-economic, preference-related reasons (Cajner et al. 2014).

12. Nevertheless, persons who usually work part time for non-economic reasons are considered “voluntary part-time” workers, since their reasons for working shorter than full-time hours are attributed to childcare problems, family or personal obligations, school or training, retirement or Social Security limits on earnings, and other reasons. This group includes those persons who usually work part time and were at work one to 34 hours during the reference week for a “non-economic reason.” This excludes persons who usually work full time but worked only one to 34 hours during the reference week for reasons such as vacations, holidays, illness, and bad weather. It also excludes those who are absent from their jobs for the entire week. Many of these voluntary part-timers may not be so voluntary in a broader context: if the U.S. had a more comprehensive child care system many who now choose to work part time might well choose full-time employment.

13. A more lagging return may reflect in part the Reinhart and Rogoff (2009) prediction that economic downturns whose source is mainly a major financial crisis will be deeper and more protracted. Thus, the recoveries from such downturns may not only be slower, but creation of full time jobs is especially slowed.

14. Even and MacPherson (2015) estimate that the ACA announcement and implementation has led to as many as 1 million of the increase in workers who are working part time for economic reasons.

15. A third role would be if there were a reduced share of voluntary part-time workers within the industry, so even with an increased share of employment being involuntary part time, the overall proportion of part time may be no higher in that industry.

16. In 2015, retail trade and leisure and hospitality had the largest shares of employment in the United States, by quite a margin, at almost 20 percent in each.

17. These data are seasonally adjusted, to remove usual, typical monthly fluctuations (which are common in these particular industries over the course of a year), to illustrate its actual behavior over time.

18. Similar trends were present in the educational services industry; analysis is available upon request. Indeed, in the educational services industry, there is only a hint of slight decline in the two components of working part time for economic reasons.

19. Indeed, the longitudinal feature of the CPS reveals that there is a characteristic difference between the two reasons for involuntary part-time working: of all those who in June 2010 “could find only part time work,” as many as 24 percent were unemployed in a different industry 15 months earlier. In contrast, of all those working part time due to slack work and business conditions, in June 2010, only 5 percent were unemployed in a different industry 15 months earlier (Glosser and Golden 2016).

20. This despite the share of retail employment actually having dropped from 2004 to 2014 as a share of total employment (after peaking in the late 1980s) from 10.5 to 10.2 percent of total employment (BLS 2015, Table 2.1) and remaining flat since around 2009. For an additional contrast, we explored also the pattern in the durables manufacturing industry—a typically highly cyclically-sensitive industry. There, the number who are involuntary part-time exhibits the same spike during recession and return approaching prior prerecession levels, as observed in construction. More than in construction, however, a trend line suggests at least a slight increase in the inability to find full-time work that started during the recession and perhaps even prior, although subsiding after 2013.

21. See Valletta and Bengali (2013, 23). Since leisure/hospitality employment has been growing since at least the late 1960s, growth in this sector alone is unlikely to be causing a sudden increase in aggregate involuntary part-time employment any higher now than in the past. This would have to explain why it takes until after 2010 for the growth of this sector’s employment share to have a structural effect, but not before. While there may be certain occupations in these leisure or hospitality industries that are growing rapidly with more part-time hours than full-time hours, the industry composition of jobs is not likely to explain the stubbornly high rates of workers who are part time for economic reasons, even the reason of finding only part-time employment.

22. Using Auto-Regressive Integrated Moving Average (ARIMA) time series analyses, to examine its month-to-month dynamic behavior from 2006 to 2014 (see Glosser and Golden 2005 for its application to prior time series). No discernable trend or increased variance appear to be present in the time series for part time or part time for economic reasons. These suggest the series have no permanent structural “break” point after some date. However, the response in the wake of the recession did involve significant lag time, suggesting there was at least a temporary change in employer behavior attributable to the unprecedented severity and a protracted nature of the recession. While the intercept term was positive during both the recession period and early stages of the recovery—suggesting a possible shift in time series behavior—its return to trend before the 2014–2015 date of the introduction of the ACA casts substantial doubt on the ACA as a source of structural change that resulted in permanently higher involuntary part-time working.

23. Allowing workers to take jobs with less than 40 hours but still above 30 hours, while maintaining health insurance coverage, could be behind the recent upward trend of  the rate of voluntary part-time working. For example, there was an unprecedented spike in voluntary part time, a one-year increase of 3.9 percent between the first quarter of 2014 and the first quarter of 2015 (Baker and Jorgensen 2015).

24. The CPS tabulations produce estimates of part-time work and its components in the at-work workforce that are somewhat higher than what BLS publications show, about a half percentage point higher in 2015 as well as in 2007 (0.3 and 0.2 percentage points higher for involuntary and voluntary part-time work, respectively). The fact that this difference is comparable in 2007 and 2015 suggests that the trends in this analysis based on CPS data are consistent with those of BLS.

25. The absence rate in 2015 was 3.0 percent for full-time workers, 5.3 percent for part-time workers, and 3.4 percent overall.

26. According to BLS, unincorporated self-employed and unpaid family members comprised 5.8 percent of nonagricultural workers at work. This group comprised 10.4 percent of involuntary part-time workers so their omission understates the total intensity of part-time work. See: http://www.bls.gov/cps/cpsaat21.pdf

27. In probit estimation models (using the last four months of the Current Population Survey Outgoing Rotation Group, available upon request), which include all demographic characteristics of workers (including education level), results confirm the significance of the observed cross-sectional differences and similarities controlling for the other worker and job factors. Working part time for economic reasons is more likely for blacks and Hispanics. It is also more likely for 25- to 34-year-olds, and 35- to 44-year-olds, though this is attributable to workers in these age groups being more vulnerable to “slack work.” Being involuntarily part time because one “could find only part time” is only slightly more concentrated among those age 25 to 44 and among black and, especially, Hispanic workers. By gender, when controlling for industry and occupation distribution, men are actually slightly more likely to be part time for economic reasons—but not for the reason of finding only part-time work. Thus, men appear to be relatively more likely to be on shorter hours for slack work reasons, strictly because of their occupation and/or industry where they tend to be employed. Finding only part-time work is no different between the genders when controlling for their industry and occupation distribution.

28. A breakdown of part-time employment by marital status and parental status (available upon request) shows that single parents were the group most likely to be working part time involuntarily (7.4 percent). Among single mothers, 8.1 percent were part-time involuntary workers.

29. Probit estimations (controlling for education, race, ethnicity, gender, and age) confirm that the three major industries most likely to use part-time involuntary workers (in order of magnitude), are retail/wholesale trade, construction, and professional services, Involuntary part-time working, including taking a part-time job because that is all that was found, is most associated with being employed by detailed industry in retail, and also somewhat in food services, accommodations, professional services, and educational services. Both types of involuntary part-time working are less prominent in the goods-producing industries.

30. Probit estimation results show that working part time for “voluntary” reasons is somewhat more likely among those under 25 and over 64. By gender, women are far more likely to be voluntary part time. By industry, working voluntarily part time seems to be relatively more likely in the retail and public sectors, but also in educational services and leisure and hospitality and somewhat more in food production and services.

31. Further, the ACA mandate carries some adverse potential distributional incentives as well. Because the employer fee is not based on wage level, it may be larger as a share of total labor costs for low-wage workers not offered coverage (Jolevski and Sherk 2014).

32. One study predicted shorter workweeks mainly for the workers in the bottom quintile (Jolevski and Sherk 2014).

33. Comparisons for quarters in 2013 to 2014 were consistent, so the time frame chosen was not confounding nor consequential (ADP 2015). Employers could conceivably be moving selected part-timers up to full-time hours, if their talent is valued.

34. This in turn might reflect the cessation of unemployment benefits after 26 weeks, or whatever the extended period is (Glosser and Golden 2016).

35. There is also a running tally by Investors’ Business Daily, of businesses that state that they have reduced hours or relied more on part-time jobs as a result of the ACA rule (Graham 2014). Most interestingly, the bulk of employers stating an intention of reducing employees’ weekly hours to the high 20s were public and nonprofit sector institutions, such as municipal agencies and school districts, and for employed students or adjunct teachers—not beleaguered private-sector businesses just above the 50 employee threshold.

36. The BLS defines part-time as less than 35 hours, while the ACA mandate defines it as 30 hours. This difference in definitions means that the link between the ACA’s employer mandate and involuntary part-time work is not clear.

37. Comparisons for quarters 1 or 3, in 2013 to 2014 were consistent, so the time frame chosen was not confounding nor consequential (ADP 2015).

38. Some regard this reduction in labor supply as a downside of the ACA (Mulligan 2014), stemming from the tax incentives that might lead some workers to prefer to work no more than 29 hours per week (Mulligan 2015).

39. Note that this could reduce the non-trivial percentage of those who usually work part-time but actual weekly hours exceeded 35 in the previous week—4 percent of all part-timers (who could be considered “involuntarily full time”), despite their designation as “part time” (Golden and Wiens-Tuers 2002).

40. Tabulations of the 2015 Current Population Survey.

41. Bishow (2015) uses data from the 2013 National Compensation Survey, Bureau of Labor Statistics.

42. See forthcoming EPI report, Part-Time Disadvantage—Minor or Major Penalty? (working title) for what we know about pay and benefits penalties by gender, race, occupation, etc.

43. The survey sample size of 944 contained 504 involuntary part-time workers (thus oversampled), plus another 440 voluntary part-time workers. The survey was fielded between March 25 and April 6, 2015. Of the total sample of 944, 623 were working part time as paid employees, 233 were self-employed plus another 98 were working but had retired. Almost two-thirds, 600, had a single part-time job, another 165 had multiple part-time jobs, and the remaining 179 also had one full-time job.

44. This was roughly the reverse proportions that are found with the CPS approach (the remainder identify seasonal reasons). When allowed to provide multiple reasons for working part time, fully three in four in the “involuntary” category cited that they “could not find full time,” but four in five did mention “poor business conditions.” Another quarter of the involuntary group mentioned “family or personal obligations” (the predominant reason mentioned by those working part time voluntarily). This suggests that many workers regard “obligations” as an imposed constraint, not purely a voluntary “choice.” Only one in 11 gave “prefer to work part time” as one of the many possible reasons. Thus, “part time for non-economic reasons,” in the CPS may not be synonymous with “voluntary part-time” working.

45. Data in this paragraph are based on tabulations of CPS microdata for 2015, specifically the average for all workers at work in the reference week.

46. Those who usually work full time, but are working fewer than 35 hours in the most recent week, experience variable weekly hours more so than those who usually work part-time hours, if the reason for their shorter hours is either “slack work” or “seasonal work” conditions. Earnings instability for households is attributed at least in part to instability of work hours over the week (Gottschalk and Moffitt 2009) or the year (Koo 2016).

47. See Henly and Lambert 2015. In a survey of 293 retail employees, respondents were randomly assigned to receive either the question with or without the option of “or too variable…” as a explicitly listed response option, which allowed respondents to indicate this possibility. See Dickson, Bruno, and Twarog (2015) for a focus on fast food workers.

48. In the sample, 56 percent of the workforce is paid hourly, while 42 percent are paid by salary, and the rest “some other way.”

49. Table 11 of PPP poll.

50. Scheduling instability is more common among nonmanagerial employees in the retail trade industry. A survey with 364 respondents found that as many as 55 percent of these nonmanagerial employees have had their work scheduled with less than one week’s advance notice, and 17 percent reported that their manager reduces their hours without their consent (Alexander and Haley-Lock 2015).

51. Tables available by request–Public Policy Polling, March 27-31, 2015, State of Minnesota, shared by EINet, University of Chicago.

52. Part-timers are far less likely to “never” work on-call shifts, and more likely to on occasion work on-call, relative to full-timers. For example, in the Minnesota survey group, 77 percent of full-timers said they “never” work “on-call shifts,” whereas only 47 percent of part-timers respond “never” to the question of “how often” they work on call. Indeed, 21 percent of part-timers “regularly” or “sometimes” work on-call shifts, over double the proportion of full-time workers.

53. General Social Survey Quality of Worklife Supplement (National Institute of Occupational Safety and Health), pooled years 2002, 2006, and 2010– the sample was 4,518 including 806 part-time.

54. Multivariate model estimations confirm that the likelihood of working part time is enhanced for those who report their usual schedule as irregular/on-call (or afternoon/night shift). Also, part-time workers are significantly more likely to be on irregular/on-call shifts.

55. However, controlling for the actual number of working hours, the effect of part-time work loses its statistical significance. Thus, the associations reflect the work hours commitment more than the part-time status per se if one’s work shift is irregular/on-call; working part time makes the irregularity less harmful for work-family conflict. This suggests that working part-time hours cushions the adverse effects of trying to balance work and family otherwise experienced by those of irregular work schedules

56. Thus, not surprisingly, four in five voluntary workers are at least “somewhat satisfied,” compared with only half of involuntary part-time workers saying they are “somewhat satisfied.” Perhaps this is related also to the finding that involuntary part-time workers are doubly uncertain about the likely duration of their employment: 30 percent indicate they are not sure about whether their job will continue compared with 15 percent of voluntary part-time workers. Finally, fewer than one in eight of those involuntarily part time thought it was “very likely” or “optimistic” that they would get a full-time job with their current employer, or find one in the next 12 months. Respondents were asked whether part-time employees ought to receive overtime premium pay for working hours over 35 in a week: fully two-thirds of involuntary part-time workers were in favor, compared with 56 percent of voluntary part-time workers, while about one in five or six of each group were opposed (See Golden and Okulicz-Kozaryn 2015).

57. Seattle’s Secure Scheduling ordinance addressed a potential shortcoming—for both the employer and for the employee—by providing an exception to predictability pay when an employee’s new shift is the result of voluntary shift-swapping. Future crafting of fair scheduling legislation could accommodate this exception to a penalty, when shifts are added on short notice, when workers clearly welcome the additional work, since this would be a case of helping workers alleviate underemployment (Rachel Deutsch, in Center for Popular Democracy 2016).

In addition to these actions by federal and state legislators, David Weil, the U.S. Department of Labor, Wage and Hour Division administrator, stated in an interview earlier this year that his agency is “looking very actively at” the question of whether workers should be legally entitled to predictable scheduling, as a protection to be covered by the Fair Labor Standards Act (Alexander and Haley-Lock 2015; Weil 2014)

58. The National Retail Federation’s sponsored study (by Oxford Economics), predicts that the new overtime regulations could lead retail employers to hire up to an estimated 117,100 part-time workers to fill their FTE labor needs, mostly going to underemployed part-time workers, as work is shifted away from salaried managers who become entitled to overtime pay. (National Retail Federation n.d.). Zukin and Van Horn (2015) advocate for an overtime pay premium for part-time employees that kicks in at relatively fewer hours per week than for full-timers, e.g., at over 35 hours per week, which would further redistribute hours toward the underemployed.

59. Because underemployment is surprisingly prevalent among those working full time (Li and McCully 2016), some suggest that this new overtime threshold for low salaried workers might have the unintended side effect of exacerbating underemployment, to the extent employers limit such newly covered employees’ weekly hours or adjust downward their base salary—a response that is predicted to occur among 13 to 26 percent of employers (Mathur 2016). Note, however, that both provide workers more time off per week away from work, from which they can choose to pick up extra work hours elsewhere. Note also that in earlier experiences with “work-sharing” (Fuchs and Jacobsen 1991) found that workers who had compulsory reductions experienced little if any reduction in well-being, depending on how they allocated the new time off; 12 percent actually increased their unpaid hours of work in their factory.

60. The full effects of the ACA’s 30 hours would be equal to the employment level net change plus the utility level with both increased security and health, and greater disposable income due to hours stability, and that underemployment is preferable to unemployment, with job lock removed.

News Letter Signup