Keywords

6.1 Introduction

Work schedules are the subject of growing scholarly and public scrutiny in the USA. The reasons for this scrutiny are both structural and idiosyncratic. As in many service-based economies, workers increasingly find themselves straining to keep up with 24/7 operations at work and without a dedicated caretaker at home. Tension between contemporary work and family norms is exacerbated in the USA by an exceptionally low level of welfare state spending and labour protections (Berg et al. 2014; Kalleberg 2018). Employers enjoy tremendous discretion over the working time of their employees and, especially in large service industries such as retail and food service, often employ ‘just-in-time scheduling’ practices that result in instability and unpredictability for hourly workers (Henly and Lambert 2005, 2014). Responding to concerns about work–life conflict in general and just-in-time scheduling in particular, new research and reporting are bringing work schedules into sharper focus as state and civil society actors seek to limit problematic employer scheduling practices.

In this chapter, we discuss the functions, prevalence, and governance of several types of work schedules—mainly on-call and on-demand work—that in recent years have attracted significant research and regulatory attention in the USA. We begin by identifying features of the institutional context that allow employers unilateral control over working time, emphasising how the adversarial and fragmented structure of labour relations contributes to weak and uneven labour standards in the contemporary period. In this context, employers do not need to designate jobs as ‘on-call’ or ‘zero hours’ to schedule workers in unpredictable or erratic ways. While workers with non-standard arrangements are at greater risk of on-call and on-demand work, we show that workers in standard employment also experience significant unpredictability and instability. These features of the US labour market help explain why relatively few workers identify as having on-call or irregular schedules despite widespread use of ‘lean’ and ‘flexible’ staffing strategies by employers (Houseman 2001; Cappelli and Keller 2013).

Our account of on-call and on-demand work in the USA is premised on a critical interrogation of the categories used to measure work schedules. We critique conventional measures of on-call work and other irregular work schedules for confounding multiple dimensions of schedule variation and for failing to specify clear criteria for classification. We present an alternative approach that analyses schedules as constellations of control, advance notice, and consistency with distinct functions for employers and characteristic effects on employees. In particular, we identify a broad constellation of unstable schedules defined by variation in the timing or number of hours which the worker does not control. Within this constellation, we define on-call work as the subset of unstable schedules with short notice and on-demand work as the subset with considerable volatility in the number of hours. This multidimensional approach emphasises functional differences between employer- and employee-driven variation while revealing that many employees effectively work on-demand or on-call, although they may not classify their job as ‘zero hours’ or ‘on-call’.

This analysis of the functions and prevalence of on-call and on-demand work informs our discussion of recent voluntary, contentious, and legislative responses to scheduling issues. In the US context of unilateral employer control, an attractive strategy for would-be reformers is to incentivise employers to voluntarily reduce practices that result in on-call and on-demand work. But many employers resist doing away with familiar scheduling practices, even when presented with a compelling business case for doing so (Lambert 2014; Ton 2014). This resistance elicits more contentious responses from social movement actors and their allies who seek to push individual firms to change certain practices through public pressure campaigns. Yet the ultimate goal of many advocates is encompassing legislation that sets standards for ‘predictive scheduling’ backed up by state enforcement. Legislative responses are currently stymied at the federal level, but gaining traction in states and cities where progressive coalitions and entrepreneurial politicians have taken up scheduling as part of a series of pro-labour reforms. Given the adversarial and decentralised system of labour regulation in the USA, scheduling legislation is resulting in complicated administrative rules that may be challenging for employers to implement and for officials to enforce.

6.2 Institutional Context

The USA is widely regarded as the quintessential liberal market economy, exhibiting in stark relief features common to wealthy countries of the former British Empire (Esping-Anderson 1990; Hall and Soskice 2001). Populated by diverse groups of natives, settlers, slaves, and immigrants; laying claim to a vast territory abounding in resources; and governed from an early date by a majoritarian, federal constitution laden with veto points, the USA did not develop a labour party or welfare state on the model of other industrialising democracies. Instead, it developed a more antagonistic and decentralised political economy in which employers exert tremendous control over their workforce and yet rely on competitive mechanisms to coordinate activities beyond the boundaries of the firm. Even in comparison with other liberal market economies, the USA is distinguished by fragmented labour market institutions, minimal employment protections, and private provision of care, training, and insurance (Huber and Stephens 2001; Kalleberg 2018).

With respect to working time, the institutional configuration of the USA can be characterised as a regime of unilateral employer control (Berg et al. 2014). Employers are generally free to offer or withhold work in the pursuit of business objectives. But the power of employers is not absolute. Individual employers face competitive pressures to satisfy employees’ schedule preferences, particularly in markets where qualified labour is scarce or costly to replace. Employers also operate in an adversarial legal and regulatory system where countervailing forces, however episodic and uneven, can impose punitive and uncompromising terms on employers (Prasad 2012). Although federal labour standards and enforcement have generally weakened since the heyday of the labour-liberal Democratic coalition in the mid-twentieth century, labour groups and their political allies continue to shape working time through protective legislation, lawsuits, and public pressure campaigns—particularly at the state and local levels—in which employers figure more often as opponents than as partners. In this fragmented and adversarial context, we contend that scheduling practices are best understood in terms of their functions and effects rather than their form.

6.2.1 From Adversarial Regulation to New Federalism

Given the substantial discretion that employers enjoy over many aspects of working time, it is tempting to view US regulation as inherently market-oriented or laissez-faire. However, recent scholarship in comparative law and political economy challenges this view, arguing that US regulation is defined more by its adversarial character than its limited scope (Kagan 2001; Prasad 2012). This adversarial character stems from the common law tradition transplanted from England, but is also shaped by the belated development of an administrative state tasked with addressing problems of domestic overproduction specific to the US political economy. The most striking examples of adversarialism in labour regulation are federal agencies such as the National Labor Relations Board and the Equal Employment Opportunity Commission. These agencies were created by reform-oriented governments, mostly Democrats, as part of legislation to establish labour, consumer, or civil rights demanded by social movements. Although subject to legislative, executive, and judicial constraints, these agencies are empowered to set, enforce, and interpret their own rules governing various aspects of employment relations.

The bureaucratic and relatively autonomous power of federal agencies has ironically made them an arena for partisan political battles. In an era of heightened party polarisation and Congressional deadlock, battles over regulation are increasingly fought between the executive and judicial branches, resulting in alternating expansion and retrenchment of labour standards. For example, the Department of Labor under the Obama administration issued a change to the administrative rules of the Fair Labor Standards Act that would have expanded eligibility for overtime premiums to over 4 million employees with annual salaries below $47,476 (McCrate 2018, p. 20). This rule change was opposed by employer associations and Republican State Attorneys General who obtained a federal injunction halting its implementation in December 2016. Now under the Trump administration, the Labor Department is moving to revise the salary threshold for overtime exemption, possibly restoring it to its previous lower level. Federal labour enforcement exhibits similar partisan dynamics—expanding mostly under Democrats and shrinking mostly under Republicans (Weil 2010, 6–7).

The adversarial character of US regulation makes labour standards highly dependent not only on partisan control of government but also on the power of labour unions, further amplifying cyclical dynamics. In the postwar period of tight labour markets, powerful unions bolstered wage and hour standards not only for their members and but also for non-members in their industry or region, as non-unionised employers sought to avoid unionisation and attract qualified employees. But this virtuous cycle turned viciously against labour with the economic crisis and employer mobilisation of the 1970s, leaving weakened unions struggling to maintain even lowered standards (Kalleberg 2018). While some unions maintain contractual schedule protections such as a minimum hour guarantee (Crocker and Clawson 2012), many now include ‘two-tier’ arrangements that provide inferior terms for contingent, part-time, or less senior workers (Weil 2014). The contemporary regime of unilateral employer control over working time reflects this dual movement of deregulation in the political sphere and de-unionisation or fissuring of the workplace.

As unions and their Democratic allies have lost power nationally, they have reoriented regulatory efforts towards more local levels, contributing to what some commentators call a ‘new federalism’ (Nathan 2006; Takahashi 2003). In cities and states with large Democratic majorities, measures to improve ‘bad jobs’ have emerged as winning issues for entrepreneurial politicians and a broader labour movement comprising not just unions but also worker centres, advocacy organisations, and community groups. Among the major developments related to working time are laws requiring employers to provide paid leave, to which tens of millions of mostly non-union workers are now entitled. Regulation of the scheduling process itself is more limited, but has grown significantly in recent years. We discuss the content and prospects of new scheduling legislation later in this chapter. For now, we merely underscore that the decentralised and adversarial character of US labour regulation shapes working time in ways that undermine formal classification of employment arrangements and potentially resist the prevailing regime of unilateral employer control.

6.2.2 Functions and Consequences of Unilateral Control of Working Time

The dominance of employers and fragmentation of labour standards have important implications for the study of on-call and on-demand work. The more discretion that employers have over employee schedules, the less plausible it is to assume a strict correspondence between the form and function of scheduling arrangements. Employers in search of labour flexibility need not rely on formal on-call or zero hours contracts when they can vary the hours of workers even in standard employment. In retail, for example, it is common for employers to maintain a large pool of workers with part-time or reduced full-time hours (less than 40 per week) whose hours can be increased on a weekly or daily basis without incurring overtime pay (Carré and Tilly 2017). To be sure, many retailers also hire workers on seasonal contracts or schedule on-call shifts with the expectation that these arrangements entail more volatile and unpredictable hours. But even when employment contracts stipulate regular work hours, these provisions seldom function as guarantees.

A review of relevant case law and union contracts reveals that scheduling provisions in the USA typically concern how employers allocate hours, not the number of hours offered (Alexander et al. 2015; Crocker and Clawson 2012). Employers are prohibited by law from discriminating on the basis of race, sex, or another protected class in scheduling employees. Conversely, many union contracts specify ‘fair’ criteria (e.g. seniority) for allocating work hours. But so long as employers respect these procedural limitations, they may offer as many or as few hours as they wish. Even in rare cases where employment contracts explicitly prohibit unilateral furloughs, employees have at best mixed success mounting legal challenges to reductions in work hours (Merola 2010). In a context of unilateral employer control, there are no guaranteed hours; every employee has a zero hours contract.

If employers are not bound by formal scheduling arrangements, then researchers may benefit from a more flexible analytic approach. Rather than simply categorising contracts and shifts, we propose to define types of schedules in the light of their functions for employers and effects on employees. Both on-call work and on-demand work allow employers to vary the timing or number of hours in order to meet business needs.Footnote 1 This employer-driven variation results in unstable schedules for employees. On-call work is characterised not only by instability but also short notice of employer scheduling decisions. Short notice allows employers to incorporate more recent information into scheduling decisions, yet in combination with instability it results in unpredictability for employees. On-demand work is characterised by considerable volatility in work hours, which facilitates short-term adjustments in staffing levels but is likely to exacerbate instability for employees. Zero hours contracts represent an extreme form of on-demand work in which employer-driven volatility is not constrained by a minimum number of guaranteed hours. By reconceptualising zero hours contracts in terms of extreme volatility, however, we can analyse functionally similar types of on-demand work without relying on the notion of ‘guaranteed hours’, which is dubious in the absence of enforceable contractual or statutory minimum hours. We offer more precise definitions of ‘short notice’ and ‘considerable volatility’ below in our discussion of recent evidence from national surveys.

6.3 Evidence from National Surveys

National surveys of the US labour force have traditionally captured little detail about work schedules. The Current Population Survey (CPS), conducted by the US Census Bureau, has long been the primary source of official statistics on work hours and occasionally includes supplementary questions on scheduling arrangements such as shift work. However, the conventional approach to working time has been to treat it as a form of economic activity organised according to discrete and generally recognised categories. We argue that this categorical approach neglects the multidimensional nature of working time and requires workers to classify themselves without clear criteria. Fortunately, recently available data from alternative national surveys make it possible to compare estimates of on-demand and on-call work using traditional measures as well as the multidimensional measures we propose.

6.3.1 Categorical Approaches to Estimating On-call and On-demand Work

Our analysis begins with the premise that working time has multiple features, including advance notice, time of day, number of hours, and control over the scheduling process itself. Any analysis of work schedules must conceptualise and measure at least some of these features. Conventional measures, however, tend to collapse or neglect key dimensions of schedule variation, yielding results that are either difficult to interpret or too restricted to capture functionally similar scheduling practices across different forms of employment.

The CPS exemplifies the conventional approach to work schedules. Its primary measure of working time is the number of weekly work hours. This measure is used to categorise workers’ schedules as full-time or part-time and also serves as an indicator of aggregate economic activity, analogous to the number of people employed. The CPS asks workers about their usual hours per week as well as the number of hours worked in the past week. But it is difficult to interpret the difference between these numbers as a measure of hour variation, since questions about the reason for this difference are only asked of those who report fewer than 35 hours per week, the conventional threshold for defining full-time employment. It is also difficult to interpret the responses of workers who, in lieu of a number, volunteer that their hours vary. Because the main CPS questionnaire does not include a measure of schedule control, researchers cannot distinguish between workers offered unstable hours by their employer and workers who set variable hours for themselves.

Supplements to the CPS provide additional information on the timing and regularity of work schedules, but still suffer from the limitations of a categorical approach. The Work Schedules Supplement (WSS), last fielded in 2004, includes a question on ‘flexible work hours’, which can be interpreted as a measure of employee control over start and end times (McCrate 2012). The other questions about timing are designed to capture regular arrangements rather than schedule variation. Workers are prompted to specify the time they begin and end work ‘most days’, although they can volunteer that the timing varies. Workers are then asked whether or not they work a daytime schedule or some other schedule. If not, they are asked which of the following ‘best describes the hours [they] usually work’: an evening shift, a night shift, a rotating shift, a split shift, or an irregular schedule. Because workers are only given the option to report an irregular schedule if they first report they do not usually work a daytime schedule, this measure is likely to underestimate the prevalence of irregular timing (McCrate 2012, 2018).

The Contingent Worker Supplement (CWS) is designed to measure alternative work arrangements such as temporary and on-call work that deviate from the standard of ongoing, dependent employment (Polivka 1996). To measure on-call work, the CWS relies on a distinction between ‘regular hours’, assumed to be typical of standard employment, and on-call hours offered only as needed. The full text of the question reads:

Some people are in a pool of workers who are ONLY called to work as needed, although they can be scheduled to work for several days or weeks in a row, for example, substitute teachers and construction workers supplied by a union hiring hall. These people are sometimes referred to as ON-CALL workers. Were you an ON-CALL worker last week?

According to our correspondence with the Bureau of Labor Statistics (BLS), the emphasis on ‘only’ in this question is meant to exclude arrangements where at least some hours are regularly scheduled. Beginning with the 1997 CWS, this restriction was made explicit by the addition of a follow-up question that asks whether any hours are ‘regularly scheduled’. Nevertheless, the follow-up question is unclear as to whether regularity refers to the timing or number of work hours, or perhaps both. Moreover, the CWS neglects the length of advance notice, which we suggest is a salient and consequential feature of on-call work.

The top rows of Table 6.1 present estimates of the prevalence of on-call work among the US population of current civilian employees aged eighteen and older,Footnote 2 using the categorical measures of on-call work from the 2017 CWS. Unlike previously published analyses of these data, we distinguish between the official and conventional measures of on-call work. Officially, the BLS defines on-call as an alternative work arrangement in which work is offered only as needed and there are no regularly scheduled hours. This official definition of on-call work is akin to what is called a ‘zero hours contract’ in the UK, although that term is seldom used in the USA. Only 0.8% of employees meet this stringent definition of on-call work. However, conventionally the BLS and most US scholars measure on-call work using the ‘only work as needed’ criterion, disregarding the follow-up question about regularly scheduled hours, which was not included in the initial (1995) round of the CWS. This less restrictive measure puts the prevalence of on-call work at 1.9% of employees. While on-call work seems rare by either measure, it is striking that more than half of the workers conventionally counted as on-call also report regular hours, contrary to the official definition. The discrepancy between these measures of on-call work further illustrates the limitations of an analytic approach to working time that relies on underspecified categories such as ‘regular hours’. It also suggests that workers may experience functionally similar types of work schedules in a variety of employment arrangements.

Table 6.1 Schedule prevalence by type and data source
Full size table

6.3.2 A Multidimensional Approach to Schedule Constellations

Our approach to on-call and on-demand work addresses the limitations of conventional accounts through a multidimensional analysis of more detailed data on work schedules. Rather than equating on-call and on-demand work with discrete forms of employment, we analyse unstable, on-call, and on-demand work as constellations of three dimensions of working time: control, advance notice, and consistency. We follow McCrate (2012) in defining unstable schedules as a combination of inconsistency with a lack of employee control over paid working time. However, we broaden our analysis of inconsistency to include workers who classify their schedule as ‘irregular’ or who report variation in the number of hours, as well as those who report varying start or end times. Within this broad constellation of unstable schedules, we distinguish on-demand and on-call work by the type of unpredictability that results from employer-driven instability. In the case of on-demand work, unpredictable schedules result from considerable volatility in weekly work hours, whereas in the case of on-call work, unpredictability is a function of short notice. New items in national surveys allow us to compare estimates of the prevalence of on-call and on-demand work using this multidimensional approach with various thresholds for what constitutes ‘considerable volatility’ and ‘short notice’.

In recent years, several national surveys have introduced new schedule questions that attempt to distinguish multiple dimensions of scheduling and to measure variation in a continuous or at least ordinal manner. Advance notice is a dimension of scheduling that is absent from the CPS but has been measured since 2016 by the Survey of Household Economics and Decisionmaking (SHED) and since 2011 by the National Longitudinal Survey of Youth, 1997 CohortFootnote 3 (NLSY97). The NLSY97 also asks about the most and fewest hours worked per week in the past month at the main job, allowing us to measure the extent of volatility in relative terms as the ratio of the range to the usual hours per week. Since 2014, the General Social Survey (GSS) has asked a subset of respondents detailed questions about work schedules, including advance notice, control, and instability in the number and timing of hours for all jobs. The available national data differ in their relative strengths and weaknesses, with more detailed measures of work schedules available in surveys with relatively small or targeted samples and larger, more representative data available from surveys containing less nuanced measures. In this section, we take advantage of the relative strengths of recent national surveys to estimate the prevalence and distribution of unstable, on-demand, and on-call work schedules.

Table 6.1 presents estimates of the overall prevalence of unstable, on-demand, and on-call work schedules among the US labour force based on the best available data. The population of interest is current civilian employees aged eighteen and older residing in the USA, except for the NLSY97 which represents a cohort born in the USA between 1980 and 1984 (30–36 years old at the time of the interview). The year(s) listed correspond(s) to the field period for the survey. The effective N is the unweighted number of respondents with non-missing data on at least one of the measures included in this table. The other rows of the table are organised around distinct schedule types or constellations, from broadest (schedule instability) to narrowest (on-call work).

6.3.3 Schedule Instability

As discussed above, we define schedule instability broadly as variable hours or timing without employee control. Where it is possible to distinguish between instability in number of hours and instability in timing, unstable hours appear more common. According to the GSS, one in three employees (33.5%) have unstable hours, as compared with one in ten (10.7%) who have unstable timing. Nearly two in five (38.7%) experience one or the other types of schedule instability in their current job(s). The NLSY97 does not ask directly about control over hours and only asks about timing using the conventional shift-type approach, which is likely to underestimate variation. However, if we use control over start and end times as a proxy for control over hours, the data suggest that schedule instability is widespread among this cohort of early-career employees, affecting nearly one in three (31.6%) in their main job. The WSS lacks a comparable measure of variable hours, but we can combine the items on usual shift type, start and end times, and employee-driven flexibility to estimate unstable timing. Using the most recent WSS data from 2004, we estimate that 13.6% of employees experience unstable timing. The NLSY97, GSS, and SHED all include measures of control over timing, although with more or less nuanced questions and response options. The GSS asks separately about variation and control over timing, yielding an estimate of 10.7% of employees with unstable timing. The SHED uses a simplified measure of variation in timing ‘primarily based on my employer’s needs’, which results in a higher estimate of unstable timing (14.9%). The NLSY97 uses a categorical measure of shift timing, yet also draws fine-grained distinctions between levels of employee control, yielding the lowest estimate (5.1%) of unstable timing.

6.3.4 On-demand Work

We define on-demand work as a subset of unstable schedules with considerable volatility in the number of weekly hours. This schedule constellation is comparable to the ‘if and when’ contracts described elsewhere in this volume. The NLSY97 and GSS include continuous measures of greatest, fewest, and usual work hours which allow us to estimate the prevalence of on-demand work using different thresholds of volatility.Footnote 4 At the 25% volatility threshold, the prevalence of on-demand work is 22.7% of employees using the GSS data and 19.7% of early-career employees using the NLSY97. Above the 50% threshold, the estimated prevalence of on-demand work is 10.1% in the GSS and 8% in the NLSY97. In other words, about one in ten employees do not control the number of hours they work and in the past month experienced major volatility equivalent in magnitude to most of their usual weekly hours.

6.3.5 On-call Work

In contrast to the official categorical measure, we define on-call work as a constellation of short advance notice, unstable work hours, and lack of employee control. Both the SHED and GSS measure advance notice with an ordered series of response options, allowing us to compare the prevalence of on-call work using different thresholds.Footnote 5 In the SHED questionnaire, respondents are asked how far in advance their employer usually tells them the hours they need to work, for which the minimum option is ‘one day in advance or less (including on call)’. Using the SHED data, we estimate that 5.1% of employees work on-call with very short notice and unstable timing. For the GSS, our estimate is somewhat higher (6.4%), which may reflect the inclusion of employees with very short notice and unstable hours, a constellation not captured by the SHED. When the threshold for short notice is three days or less advance notice, we estimate that more than one in ten employees (11% in the GSS, 13.5% in the SHED) work on-call. If we extend the threshold to a week or less advance notice, the prevalence increases to 14.9% of employees. As with on-demand work, the NLSY97 yields a somewhat lower estimate of on-call work (10.4%), since the underlying measure of schedule instability is more conservative.

These estimates all suggest a much higher prevalence of on-call work than estimates based on commonly used categorical measures. Recall that the BLS conventionally counts 1.9% of employees eighteen and older as on-call workers, although only 0.8% satisfy the official criterion of ‘no regular hours’ in the CWS (see top two rows of Table 6.1). By contrast, we estimate that 6.4% of employees experience the combination of schedule instability with a day or less notice—our preferred measure of ‘on-call work’ from the GSS. Put differently, less than 1 in 50 employees work only when needed, but more than 1 in 16 work on-call.

6.3.6 Variation Across Categories of Standard and Non-Standard Employment

Table 6.2 provides further evidence that conventional categories of employment arrangements underestimate the prevalence of work schedules that function to create unstable, on-demand, or on-call work. We use the NLSY97 for these analyses because it includes both the new items that capture the magnitude of work hour fluctuations and length of advance notice as well as conventional questions about the type of schedule and employment arrangement. For ease of presentation, we have dichotomised the measure of control, combined several of the advance notice options, and categorised the continuous measure of volatility using three convenient cut points. Columns A through C group respondents according to their usual shift type (regular day, regular evening or night, and irregular or rotating), whereas columns D and E contrast respondents with a standard (ongoing, dependent) employment contract to those with non-standard (temporary, on-call, and third-party) employment contracts in their main job. For each schedule dimension, the percentages represent the estimated share of the cohort population (aged 30–36) conditional on the shift or contract type of the column. Within each column and schedule dimension (i.e. control, volatility, notice), the percentages sum to 100. The last two rows, however, represent the estimated prevalence of particular schedule constellations: on-demand work (considerable volatility without worker control) and on-call work (instability with a week or less advance notice).

Table 6.2 Schedule distribution by dimension and employment category, 2015–2016 National Longitudinal Survey of Youth 1997 Cohort (NLSY97)
Full size table

Given the relatively small sample sizes of non-standard jobs in Table 6.2, we are less concerned with describing the distribution of each schedule dimension (across rows) than we are with identifying patterns of similarity and difference between categories of workers (across columns). Two basic patterns emerge from this comparison: first, unstable and unpredictable schedules are more prevalent among workers with non-standard forms of employment; second, standard employment does not guarantee stability or predictability.

On the dimension of schedule control, we find that workers without regular day shifts or standard contracts are significantlyFootnote 6 less likely to control their starting and ending times (e.g. 45% of non-standard workers vs. 57% of standard workers). The exception is workers with irregular or rotating shifts, who are about as likely as regular day shift workers to report having schedule input or control. On the volatility dimension, the most striking contrasts are in the prevalence of major volatility (more than 50% of usual weekly hours). Such volatility is more prevalent among workers with a rotating or irregular shift than among those with a regular evening or night shift (38 vs. 28%), and least prevalent among workers with a regular day shift (14%). Workers with a temporary, third-party, or on-call contract are nearly twice as likely as those with standard contracts to report major volatility (35 vs. 18%). However, a similar proportion of workers with standard and non-standard contracts report minimal volatility in the past month (26 vs. 20%, difference not significant). On the advance notice dimension, schedule notice of a week or less is reported by two in five workers with rotating or irregular shifts and nearly as large a proportion (38%) of workers in temporary, on-call, or third-party employment. Such short notice is less prevalent among regular evening or night shift workers (31%) and still less among workers with regular day shifts (19%). Nevertheless, some 23% of early-career workers with standard employment arrangements receive their schedules with a week or less advance notice.

The final rows of Table 6.2 reveal a similar pattern in constellations of on-demand and on-call work. On-demand work is widespread among workers with non-standard contracts (37%) and those with regular evening or night shifts (33%). Workers with irregular or rotating shifts have a somewhat lower rate of on-demand work (23%). Yet workers with a regular day shift are not immune to on-demand work (15%), nor are workers with a standard employment contract (18%). On-call work, defined here as schedule instability with a week or less advance notice, is less prevalent than on-demand work, but also shows a disparity between standard and non-standard employment. Workers with standard contracts are much less likely to work on-call than workers with temporary, third-party, or formal on-call contracts (9 vs. 21%). Workers with regular day shifts are half as likely to work on-call as workers with other types of shifts (7 vs. 15%), though workers with regular evening or night shifts are no more likely to work on-call than workers with irregular or rotating shifts.

It is important to underscore that disparities in prevalence are only part of the picture. Although unstable and unpredictable schedules are associated with non-standard forms of employment, a majority of early-career workers in non-standard jobs do not report on-call or on-demand work. Yet a significant minority of workers with standard jobs do. Since the vast majority of US workers classify themselves as having a standard employment arrangement, this group accounts for most of those with on-demand or on-call work. The same holds true for workers with a regular day shift. Even if we restrict our focus to workers with a standard employment contract and a regular day shift (not shown in Table 6.2), this group with traditional daytime jobs still account for a larger share of on-call and on-demand work in the NLSY97 cohort than workers in all other forms of employment. Thus, in order to understand the extent of on-call and on-demand work, it is critical to recognise the considerable instability and unpredictability that occurs in standard forms of employment, even if such employment is relatively more stable than night shift, temporary, or otherwise non-standard jobs (Carré and Heintz 2009).

6.3.7 Variations by Worker and Job Characteristics

Besides workers in non-standard jobs, which groups are most likely to experience constellations of unstable or unpredictable schedules? Table 6.3 draws on data from the SHED to address this question. We use the SHED for these analyses because it represents the entire population of current civilian employees eighteen and older (unlike the NLSY97) and because its large sample (relative to the GSS) allows for more precise estimates of schedule constellations by demographic groups as well as occupation and industry. However, it should be noted the SHED lacks a measure of instability in hours, which other data (see Table 6.1) show to be more common than unstable timing. We are thus unable to analyse the prevalence of unstable hours or on-demand work using these data. Instead, we focus on variation in the prevalence of unstable timing—specifically employer-driven variation in start and end times—and on-call work, defined here as the combination of employer-driven variation with a day of less of advance notice.

Table 6.3 Prevalence of unstable timing and on-call work by demographic and job characteristics, 2017 Survey of Household Economics and Decisionmaking (SHED)
Full size table

We observe a U-shaped relationship between schedule instability and age. Unstable timing is most prevalent (25%) among 18- to 24-year olds, comparably high (18%) among those 55 and older, and least prevalent (10%) among 35- to 44-year olds. On-call work is most prevalent among employees 55 and older (7%), but not significantly different between those middle-aged and younger (4–6%). On-call work is more common among male than female employees (6 vs. 4%), but the prevalence of unstable timing does not differ significantly by gender. We find no significant differences between broad ethno-racial groups on these schedule measures. Other analyses of GSS and NLSY97 data not reported here suggest that there may be race and gender differences in hour instability which the SHED does not capture (see also McCrate 2018).

We find more marked differences in the prevalence of unstable timing and on-call work by industry, occupation, and full-time versus part-time status. Part-time employees are two and a half times as likely as full-time employees to say their schedule varies primarily based on their employer’s needs (30 vs. 12%). They are also more likely to report on-call work with a day or less advance notice (7 vs. 5%, p < 0.05). Unstable timing is most common in sectors that include retail trade (33%), accommodation and food service (31%), and transportation (26%), where the prevalence is three to four times higher than in sectors that include finance, insurance, and real estate (9%) or administrative and information services (8%). On-call work is especially prevalent (17%) in wholesale trade, transportation, and warehousing, but exceedingly rare (2%) in education, health care, and social assistance.

We find similar patterns in our comparison of major occupational groups, although here the SHED data are of lower quality.Footnote 7 Unstable timing is most prevalent (30%) among sales and related occupations and least prevalent (4%) among computer, engineering, and science occupations. We estimate that on-call work is most prevalent (11%) among the broad group that includes production, construction, transportation, and maintenance occupations. On-call work appears especially rare (2%) among computer, engineering, science, and health technician occupations.

These results are consistent with the view that, in a context of broad employer control over work scheduling, the instability and unpredictability characteristic of on-call work is shaped by complex processes of segmentation and decentralised bargaining (Carré and Heintz 2009). There is some evidence of bifurcation between highly skilled jobs, where on-call work is rare, and less skilled jobs, where on-call work is common. This segmentation also seems related to the use of part-time jobs, which are considerably more likely than full-time jobs to involve unstable timing and on-call work. However, employees in full-time jobs and capital-intensive sectors such as transportation, extraction, and construction also experience significant instability and unpredictability. Here again, we observe disparities between more and less advantaged groups of workers, yet significant unpredictability even among relatively privileged groups.

6.4 Regulatory Responses

A large body of research documents how on-demand and on-call scheduling practices make it difficult to fulfil responsibilities for caregiving, school, and additional jobs, fuelling work-to-family conflict, stress, and financial insecurity (Clawson and Gerstel 2014; Gassman-Pines 2011; Henly and Lambert 2014; Schneider and Harknett 2019). Recognition of these deleterious consequences has sparked initiatives by labour groups, reporters, and policymakers to improve work schedules of hourly jobs in which on-demand and on-call scheduling is widespread, particularly retail and food service jobs. While the effects of these regulatory responses remain to be seen, it is clear that these initiatives are being shaped by the institutional context of unilateral employer control and adversarial regulation in the USA.

6.4.1 The Business Case for Voluntary Change

Calls to ‘make the business case’ for improving scheduling practices can be found in the press, policy briefs, and the scholarly literature, attesting to the primacy of employer control in the USA. The goal of these efforts is to convince corporate managers that their firms will materially benefit if they voluntarily reduce schedule unpredictability and instability (Ton 2014). For example, we recently conducted a randomised experiment at the retailer Gap, Inc. to assess the business effects of an intervention designed to improve the predictability and stability of sales associates’ work schedules. Initial results from this research show the intervention increased median store sales by 7% and labour productivity by 5% (Williams et al. 2018). The business-case argument centres on generating enlightened self-interest on the part of corporate officials, not on ensuring the basic human rights of employees or fairness in the workplace. But as Lambert (2014) explains, even a compelling business case for improving work schedules is likely to fall flat when firm profitability is determined more by short-term returns to shareholders than the sustained quality of products and services, job tasks are fragmented so workers are interchangeable, and the true cost of labour is externalised to the public through safety net programmes that supplement low pay with additional income, food, and housing. Not surprisingly, then, voluntary efforts by US employers to improve work schedules in hourly jobs have been modest at best.

6.4.2 Agitation for Adversarial Regulation

Labour groups and advocacy organisations have sought to regulate employer scheduling practices through more contentious tactics, including public shaming, increased enforcement of existing employment laws, and lobbying for new legislation. Journalists have contributed significantly to agitation for scheduling regulation by exposing problematic scheduling practices, particularly erratic hours and short notice at large retail and fast food chains (e.g. Greenhouse 2012; Kantor 2014). These accounts seem to resonate with a broader public, including many who have either worked in retail or food service or know someone who has.

In response to media coverage and public outcry, some government officials have sought to curtail on-call work through more active enforcement of existing employment laws. Eight states and the District of Columbia have ‘show up’ or ‘reporting pay’ laws that require employers to provide some compensation to employees who show up to work a scheduled shift but are sent home immediately or before the scheduled end time (Alexander and Haley-Lock 2015). As originally written, the laws do not cover on-call shifts for which employees wait to be told by their employer whether or not to come to work. In 2015, the New York State Attorney General (AG) announced that the state was taking steps to treat the contact between employer and employee concerning the decision about an on-call shift as actually reporting to work. This would have required New York employers to provide reporting pay (three to four hours of pay, depending on the industry) for cancelled on-call shifts. Following this announcement and related investigations by the AG, six major retailers announced they would no longer use on-call shifts.

6.4.3 Legislative Efforts

Largely propelled by the actions of policy organisations and labour groups, policymakers in a growing number of jurisdictions are moving to introduce and enact legislation that regulates employer scheduling practices. Although scheduling legislation has been introduced at the federal level, as with paid leave, there has been more policy movement at the municipal and state levels. As of October 2018, four municipalities (San Francisco, Seattle, Emeryville, and New York City) and one state (Oregon) have passed comprehensive laws that regulate multiple aspects of work schedules.Footnote 8 These laws regulate scheduling practices of large, service-sector employers, primarily covering customer-facing, hourly jobs in retail stores and fast food restaurants. However, some proposals (e.g. in Chicago) envision broader regulation that would cover most hourly workers as well as lower salaried employees across the private sector.

The current and proposed laws commonly target multiple aspects of employer scheduling practices, providing employees with: (1) a good faith estimate of the number and timing of hours the employee will usually work; (2) a minimum of fourteen days advance notice of scheduled days and times; (3) compensation for employer-driven changes to the original schedule; (4) the right to refuse hours added to the original schedule; (5) the right to rest between shifts (i.e. at least 10 hours off between shifts on two consecutive days) and extra compensation for working more closely spaced shifts; (6) access to additional hours for existing employees before new or temporary workers are hired; and (7) the right to request a schedule adjustment without employer retaliation (such as reductions in future work hours). These provisions, and the administrative rules that define their implementation, represent a novel response to unstable and unpredictable scheduling practices that may have significant consequences for on-call and on-demand work. While more experience and research is needed to evaluate its consequences, the content and implementation of this legislation already illustrate the role of employer discretion and adversarialism in shaping the regulatory process.

The wide discretion of US employers to determine conditions of work is evident in most provisions of existing scheduling legislation. First and foremost, the laws do not directly regulate variation in employees’ hours and thus do not directly address on-demand work. Industry associations have argued vehemently that employers need ample labour flexibility to respond to unforeseen business needs if they are to remain profitable (Committee on Civil Service and Labor 2017). The provisions in current scheduling laws largely concede this battle to employer interest groups. Although the good faith estimate provision may result in more consistent hours for employees, there is little in the laws to prevent employers from modifying or deviating from this estimate. The law does not require employers to guarantee minimum weekly or annual hours. Employees covered under the new scheduling laws are still, in effect, on zero hours contracts and at risk of on-demand work.

The laws do regulate advance notice, a defining feature of on-call scheduling. Indeed, the provision of premium compensation sometimes called ‘predictability pay’ holds the potential to reduce employers’ use of formal on-call shifts. Similar to existing reporting pay laws, employers are required to provide a minimum amount of pay when they cancel a scheduled shift. But the laws stipulate greater disincentives to employers cancelling scheduled hours than adding previously unscheduled hours. Thus, the legislation may limit certain scheduling practices but not necessarily reduce the resulting unpredictability for employees. Moreover, provisions related to premium pay are likely to have different consequences for workers depending on the compliance strategy of their employer: avoid practices that cost a premium or simply pay to continue these practices.

Although the administrative rules guiding predictability or premium pay may reduce the use of formal on-call shifts, they may do less to reduce other changes to the posted schedule that create unpredictability characteristic of on-call work. In addition to some compensation for cancelled on-call shifts, the laws require additional compensation when workers agree to a manager’s request to work additional hours (e.g. one extra hour of pay) or when workers are sent home early from a shift (e.g. half of the remaining hours). However, there are notable exceptions to these provisions. Administrative rules commonly require employers to provide predictability pay only for employer-driven changes to work schedules. In San Francisco, this means that if a manager asks an employee to work additional hours because another employee has called off that day, then no premium pay is due to the employee who agrees to work the extra hours. In Seattle, if a manager offers additional hours to multiple employees via ‘mass communication’, then the employee who volunteers to pick up the hours does not receive premium pay. And in Oregon, no premium is required for additional hours worked by employees who have volunteered to be on a ‘standby list’ of workers who would like to be offered more hours. Given these exceptions, some workers covered by the laws will experience unpredictable schedule changes without being scheduled for formal on-call shifts or receiving additional compensation.

Notwithstanding the wide scope for employer control maintained under existing scheduling legislation, employers’ response to scheduling legislation has been overwhelmingly negative, attesting to the adversarial character of the regulatory process. In legislative meetings and public hearings, employers and industry groups frequently characterise these laws as an onerous intrusion into core business operations (Tu 2016; Committee on Civil Service and Labor 2017). This adversarial approach results in a focus on the letter of the law—especially specific administrative rules that govern implementation—rather than the spirit of reducing unstable and unpredictable work schedules. Employer resistance pushes lawmakers and regulators to grant complex carve outs and exceptions. Moreover, lack of faith in employers to conform to the spirit of the law leads to requirements for extensive documentation, and in some municipalities, a private right of action for workers to sue employers for violations. Most municipalities require covered employers to document the exact date each weekly schedule is released to employees, the reason for each schedule change, whether or not the worker received premium compensation for such changes, and how the availability of additional hours was communicated to employees. In this adversarial context, even employers whose existing scheduling practices come close to the spirit of new legislation align themselves with competitors who oppose statutory requirements, as Costco did in its opposition to Seattle’s Secure Scheduling Ordinance (Tu 2016).

6.5 Conclusion

Our analysis of the functions and consequences of scheduling practices reveals that on-demand and on-call work occurs across the US labour market, including in standard employment arrangements. Employment standards are weakly institutionalised in the USA in comparison with countries where labour parties or civil law established statutory protections tied to formal employment contracts. Because employers have unilateral control over many aspects of working time, they can pass instability and unpredictability onto workers without needing to formally designate workers as ‘on-call’ or in ‘zero hours contracts’.

By adopting a multidimensional approach that analyses constellations of schedule control, volatility, and advance notice, we obtain estimates of the prevalence of on-call and on-demand work that are much larger than official estimates based on a categorical approach. We find that 6% of employees have variable work hours or timing that they do not control and only learn about with a day or less advance notice. By contrast, official government statistics put the number of on-call workers at about 2% of employees. We find that on-demand work—defined as considerable volatility in work hours outside of the worker’s control—affects at least one in five employees, whereas less than 1% of employees report an arrangement with no regular hours (equivalent to a zero hours contract). We show that workers with non-standard arrangements are at greater risk of on-call and on-demand work, though workers in standard employment account for most of those with on-call and on-demand work because of their greater numbers.

Concern with the negative effects of unstable and unpredictable schedules has spurred efforts by labour, advocacy organisations, and entrepreneurial politicians to regulate scheduling in hourly jobs most at risk of on-call and on-demand work, particularly in retail, food service, and hospitality. The most novel and potentially consequential of these efforts are ‘predictive scheduling’ laws which have been enacted or proposed in a growing number of US cities and states. These laws target multiple features of problematic work schedules while preserving substantial labour flexibility for employers. Nonetheless, they reflect the challenges of regulating employer scheduling practices in an adversarial and employer-dominated context. Recent scheduling laws focus primarily on reducing employer-driven unpredictability, but not instability, which is viewed as a crucial managerial prerogative. As a result, these laws have greater potential to limit on-call than on-demand work, although the latter is more prevalent in the USA. Moreover, the adversarial nature of the regulatory process surrounding these laws has resulted in administrative rules that tend either to impose strict requirements and substantial liabilities on employers or issue weak prescriptions with uncertain remedies for non-compliance.

We argue that a multidimensional, functional approach to working time is particularly useful in the US context of labour market fragmentation and unilateral employer control. Yet we believe this approach is also useful for comparative research, given that countries vary not only in the way they officially classify employment but also in the norms and practices surrounding standard employment (Vosko 2010). We hope that a clearer understanding of the functions and forms of unstable, on-call, and on-demand work will advance research on working time and strengthen the empirical basis for ongoing regulatory efforts.