A less objectionable greed? Work-life conflict and unjust pay during a pandemic

Abstract

Perceptions of unjust pay represent a central feature in research on distributive justice. Prior studies document that work-life conflict (WLC) is a strong predictor of unjustly low pay. We extend that work by asking: Did the social and economic changes associated with the coronavirus pandemic 2019 (COVID-19) modify the relationship between WLC and perceptions of unjust pay? In September 2019, we collected data from a nationally representative sample of workers to profile the quality of work and economic life. Then, during a critical period of widespread economic and social shockwaves, we re-interviewed these same study participants in May 2020 to evaluate change. We observe that the strong positive association between WLC and unjustly low pay decreased overall in the population—but the strength and direction of that association differed significantly across several dimensions of social stratification. Specifically, we found a weaker relationship among visible minorities, younger workers, and individuals with lower socioeconomic status. We interpret these patterns as suggesting that—at least among more vulnerable groups—the “greed” represented in the process of work interfering with non-work was unevenly experienced during peak period of the COVID-19 pandemic.

1. Introduction

One of the most compelling narratives to emerge from the coronavirus disease 2019 (COVID-19) pandemic is that it has amplified many existing social inequalities—and some of the most glaring examples relate to the restructuring of the work-family interface (Thomason & Williams, 2020). Prior research has established that work-life conflict (WLC) is a pervasive stressor that has deleterious personal, social, and organizational consequences (Bellavia & Frone, 2005). In a study of the link between strains in the work-home interface and distributive justice, Narisada (2020) found that individuals who report greater WLC are more likely to describe their pay as unjustly low. When work spreads beyond the work role, individuals must sacrifice their involvement in other life domains––a sacrifice associated with an expectation for greater monetary rewards. This basic proposition aligns with Coser’s (1974) characterization of “greedy institutions” that “seek exclusive and undivided loyalty and they attempt to reduce the claims of competing roles and status positions on those they wish to encompass within their boundaries. Their demands on the person are omnivorous” (p.4, italics added).

Another concept—the ideal worker norm—emphasizes that workers should prioritize work over family or personal needs in order to display undivided devotion to the work role (Williams, Blair-Loy, & Berdahl, 2013). For the ideal worker, some degree of “greed”—translated here as work interfering with life—should be deemed acceptable. And yet, the discovery that WLC elevates the perception of unjustly low pay suggests that many workers experience this inter-role strain as quite objectionable. High levels of WLC violate the psychological contract in which a reasonable amount of work-related effort should be performed within the temporal, spatial, and psychological parameters of the work role. This psychological contract identifies a moral dimension of the work-life boundary, aligning with normative expectations about role segmentation versus role integration. When work bleeds into other domains and hinders functioning, many find this spillover displeasing. In turn, this violation of the psychological contract likely marks one’s current level of remuneration as deficient.

In the present study, we begin with this basic premise: Individuals who report more WLC will tend to describe their pay as unjustly low. We then elaborate on this foundational association by asking: Did the COVID-19 pandemic modify the relationship between WLC and unjust pay? And, if so, how do those changes reproduce inequalities along key dimensions of stratification—especially gender, minority status, age, and socioeconomic status? We propose two competing hypotheses, both of which emphasize magnified vulnerability during the pandemic—but map opposite conclusions about the effects on the relationship between WLC and unjust pay.

The stress amplification hypothesis predicts that the positive relationship between WLC and unjustly low pay should have increased during the pandemic for those who have traditionally been in more disadvantaged or vulnerable locations. This view aligns with claims that COVID-19 has expanded the scope and perniciousness of existing social inequalities (Kristal & Yaish, 2020; Qian & Fan, 2020). The nature and experience of WLC—a prominent stressor—likely changed during the socially and economically tumultuous early months of another macro-level stressor: the COVID-19 pandemic. An alternative view—the tolerable permeability hypothesis—also draws on the idea of differential vulnerability to multiple or combinations of stress exposures, but it instead posits that more disadvantaged or vulnerable groups might have tempered their expectations regarding the psychological contract in ways that tolerate a more porous boundary between work and nonwork roles. A more permeable boundary implies greater spillover and the likelihood of role interference (Glavin & Schieman, 2012). Greater vulnerability to socioeconomic turbulence and its threats to livelihood might have increased some workers’ efforts to adhere to the ideal worker image in ways that signal the prioritization of work over personal and family needs. The tolerable permeability hypothesis therefore suggests a greater willingness—particularly among more vulnerable workers—to accept the greedy institution’s breach of the psychological contract. This dynamic should empirically manifest as a weaker association between WLC and unjust pay among vulnerable groups during the pandemic.

1.1. Data

To test these ideas, we analyze data from the Canadian Quality of Work and Economic Life Study (C-QWELS). In September 2019, we collected data from a nationally representative sample of workers. The online survey was fielded from September 19th to September 24th (N = 2,524). We then followed up during the period of May 17th to May 24th of 2020—the second full month of the society-wide lockdown. All participants are members of the Angus Reid Forum (ARF; see http://angusreid.org).¹ Sample selection started with creating a balanced sample matrix of the Canadian population. Randomized samples of ARF members were then selected to match this matrix to ensure a representative sample. For the September sample, the response rate was 42 percent. Of these participants, 1807 (72 percent) were retained for the May re-contact. After removing cases that were no longer employed in May and missing on study variables, we have an analytical sample of 1276. The analytic methods described below combine the September baseline sample with working respondents from the May follow-up, while adjusting analyses for repeated observations of the same individuals (total number of groups = 2472; total number of observations = 3748). We weighted results according to the most current gender, age, education, and region Census data to ensure broad representation of working Canadians. The Appendix reports descriptives for focal variables.

1.2. Variables

We measure perceived unjust pay with the following item: “When you think about the pay you get for your job, do you feel your pay is unfairly low, unfairly high, or is it a fair amount?” The measure is coded such that higher values indicate more underreward: −5 (Unfairly high) to +5 (Unfairly low) with 0 indicating “a fair amount.” Work-life conflict asks three items: “How often did your job keep you from concentrating on important things in your family or personal life?” “How often did you not have enough time or energy for the important people in your life because of your job?” “How often did your work keep you from doing as good a job at home as you could?” Response choices are: (1) “never,” (2) “rarely,” (3) “sometimes,” (4) “often,” and (5) “very often.” We averaged responses to create the index (alpha = .90).

We test if six dimensions of stratification modify changes in the relationship between WLC and unjust pay during the pandemic. Sociodemographic variables are gender (men = 0, women = 1), age, and visible minority status (not a visible minority = 0; visible minority = 1).1 Socioeconomic variables are education, income, and financial strain. Education contrasts respondents with less than a college degree (0) to those with a college degree (1) or a graduate degree (2). For personal income, we compare individuals earning $50,000 or less to those earning more than $50,000.2 We measure financial strain with three items. The first two ask: “How often did you have trouble paying the bills” and “How often did you not have enough money to buy food, clothes, or other things your household needed?” Responses are: (1) “never” (2) “rarely,” (3) “sometimes,” (4) “often,” and (5) “very often.” The third asks: “How do your finances usually work out by the end of the month?” Response choices are: (1) “a lot of money left over,” (2) “a little money left over,” (3) “just enough to make ends meet,” (4) “barely enough to get by,” and (5) “not enough to make ends meet.” We averaged the items to create the financial strain index (alpha = .85).

All models include a set of control variables: marital status, children at home, occupation, union membership, salaried, job authority, schedule control, job autonomy, job challenge, work hours, and the frequency of working from home during the pandemic. Full details on all study variables are available upon request.

1.3. Analytic strategies

We employ mixed models that adjust analyses for repeated measures of the same individuals in September 2019 and May 2020. The basic form of the mixed model is as follows:

$$Y_{it}={\gamma }_{00}+{\gamma }_{10}WL{C}_{ij}+{\gamma }_{20}tim{e}_{ij}+\sum _q{\gamma }_{0q}{Z}_{qi}+U_{0j}+R_{ij}$$

Within these equations, Y_it is the unjust pay response variable for study participant i at time t, for i = 1…, n and t = September or May survey wave. These analyses model unjust pay at each wave as a result of WLC for respondent i at the same wave, with this association represented by γ₁₀. The variable time indicates whether the time-varying measures correspond to the September or May survey, with the result that γ₁₀ indicates the average change in unjust pay over time. The random error terms R_ij accounts for unobserved influences on unjust pay due to time-varying factors. The mixed model sets the intercept as random, which allows unjust pay to vary as a result of time-stable characteristics. The coefficient γ_0q represents the associations between unjust pay and time-stable factors (e.g., gender), while U _{0 j} is a random error term that takes residual variation due to unobserved time-stable characteristics into account, and therefore adjusts standard errors for repeated observations of the same individuals over time. An interaction between time and WLC in a subsequent model tests whether the association between WLC and unjust pay differs between September 2019 and May 2020.

2. Results

Model 1 of Table 1 confirms that WLC is positively associated with unjustly low pay. However, the interaction between WLC and survey wave indicator (“May”) in model 2 shows that the positive relationship weakened between September and May. Fig. 1 illustrates that the link between WLC and unjust pay during the pandemic differs from just eight months prior.

Table 1.

Unjust Pay Regressed on Work-Life Conflict and Survey Wave (and Control Variables).

	Model 1	Model 2
Work-Life Conflict	.191***	.233***
May	.009	.338*
Work-Life Conflict × May		−.123*
Women = 1	.130	.132
Visible Minority = 1	−.258*	−.260*
Age	.004	.004
Education (REF = Less than College)
College Degree	.204*	.198*
Graduate Degree	.185	.187
Income (REF = High income)
Low income	.527***	.529***
Missing income	.688***	.690***
Financial strain	.228***	.223***
Marital Status (REF = Married)
Single, Never Married	−.125	−.127
Living with Significant Other	−.134	−.131
Previously Married	−.085	−.088
Children at Home = 1	−.273**	−.271**
Professional/Admin/Technical = 1	−.055	−.057
Union = 1	−.122	−.125
Salaried = 1	.008	.004
Job Authority	−.125**	−.126**
Schedule Control (REF = None/Very Little)
Some	−.184*	−.180*
A Lot/Complete	−.306**	−.305**
Job Autonomy	−.122*	−.123*
Job Challenge	−.060	−.061
Work Hours (REF = 50+ hours)
<30 h	−.205	−.208
30–39 h	−.195	−.192
40–49 h	−.099	−.099
Work from Home (REF = Never)
Little/Some of the time	−.163	−.159
Most/All of the time	−.271**	−.268*
Already working from home	−.134	−.136
Unknown	−.087	−.083
Constant	.348	.250

Note: *p < .05, ** p < .01, *** p < .001. Unstandardized regression coefficients are shown in the table.

Fig. 1.

WLC and Perceived Unjust Pay Before and During the COVID-19 Pandemic.

The next set of analyses evaluated contingencies for each dimension of stratification. The following are statistically significant: visible minority status, age, education, income, and financial strain. We did not find differences by gender, so we exclude those findings from the tables. All of the models include the control variables but they are excluded from the tables for the sake of space.

In Table 2 , the interaction between WLC and May is negative and significant only among non-minorities. Pooling minorities and non-minorities, we find a significant three-way interaction between WLC, May, and minority status (b = −.404, p < .05).

Table 2.

Unjust Pay Regressed on WLC by Survey Wave Across Dimensions of Social Stratification.

Minority Status and Age
	Minority Status	Non-Minority Status	Younger workers	Older workers
Work-Life Conflict	.405**	.211***	.281***	.176***
May	1.140**	.227	.559*	.075
Work-Life Conflict × May	−.522**	−.069	−.240**	.016

Education
	Less than College	College Degree	Graduate Degree
Work-Life Conflict	.332***	.229***	.054
May	.695*	.445*	−.707*
Work-Life Conflict × May	−.282*	−.149*	.218*

Income and Financial Strain
	Low Income	High Income	Low Financial Strain	High Financial Strain
Work-Life Conflict	.356***	.161***	.212***	.304***
May	.781**	.098	.083	.714*
Work-Life Conflict × May	−.348***	−.005	−.021	−.285*

Note: *p < .05, ** p < .01, *** p < .001. Unstandardized regression coefficients are shown in the table. All models include the full set of control variables.

Table 2 displays the results separately by age group, split at the median (age 43). Among younger workers only, the interaction between WLC and May is negative and significant. We find a significant three-way interaction between WLC, May, and age (b = .012, p < .05), indicating that the change in the relationship between WLC and unjust pay differed by age.

Moving down Table 2, among workers with less than college or with a college degree, the interaction between WLC and May is negative and significant; by contrast, this interaction is positive and significant among those with a graduate degree. The three-way term for WLC, May, and graduate degree is significant (b = .500, p < .01).

The bottom rows of Table 2 show results for income and financial strain. Among those with low income only, the interaction between WLC and May is significant. We confirm a significant three-way interaction between WLC, May, and low income (b = −.336, p < .01).3 And, the final models show the results for workers with low versus high financial strain (split at the median). Among those with low financial strain, the interaction between WLC and May is not significant; by contrast, that interaction is significant among workers with high strain. There is a significant three-way interaction between WLC, May, and financial strain (b = −.168, p < .01).

Table 3 summarizes the average marginal effects, showing that the positive association between WLC and unjust pay diminished between September and May for visible minorities, younger workers, individuals with less education and low income, and those with high levels of financial strain. To illustrate these patterns, each of the figures present these contingencies (Fig. 2, Fig. 3, Fig. 4, Fig. 5, Fig. 6 ).

Table 3.

Predicted Marginal Effects for the Relationship Between Work-Life Conflict and Unjust Pay for September 2019 (Pre-Pandemic) and May 2020 (During Pandemic).

	September 2019	May 2020
Minority	.405**	−.117
Non-Minority	.211***	.143**
Younger worker	.281***	.041
Older worker	.176***	.191**
Less than college	.332***	.050
College degree	.229***	.080
Graduate degree	.054	.272*
Low income	.356***	.008
High income	.161***	.155**
Low financial strain	.212***	.191***
High financial strain	.304***	.018

Note: *p < .05, ** p < .01, *** p < .001. All models include the full set of control variables.

Fig. 2.

WLC and Unjust Pay Before and During the COVID-19 Pandemic by Visible Minority Status.

Fig. 3.

WLC and Unjust Pay Before and During the COVID-19 Pandemic by Age.

Fig. 4.

WLC and Unjust Pay Before and During the COVID-19 Pandemic by Education.

Fig. 5.

WLC and Unjust Pay Before and During the COVID-19 Pandemic by Income Level.

Fig. 6.

WLC and Unjust Pay Before and During the COVID-19 Pandemic by Financial Strain.

3. Conclusions and discussions

Workers who experience conflict between work and non-work roles feel unjustly underpaid. This pattern is consistent across different dimensions of stratification before the onset of the COVID-19 pandemic. Moreover, during the pandemic there was an overall weakening of the average association between WLC and unjust pay. And, the diminishing strength in this relationship varied across visible minority status, age, education, income, and financial strain.4

Collectively, these observations align with the predictions of the tolerable permeability hypothesis. We posited that, because of differential vulnerability to the shockwaves of the pandemic, disadvantaged groups might have lessened their expectations regarding the psychological contract in a manner that allows for a more porous boundary between work and nonwork roles. For those with greater vulnerability to socioeconomic turbulence and its threats to livelihood, we also suspect there might have been enhanced efforts conform to ideal worker norms. That is, workers seek to express unambiguously the prioritization of work-related demands over personal and family needs. The tolerable permeability hypothesis suggests an inclination to perceive the greedy institution’s violation of the psychological contract as less objectionable. We interpret the weaker association between WLC and unjust pay in May 2020 among vulnerable groups as an empirical manifestation of this more tolerable permeability.

Decades of research finds that WLC is related to unfavorable outcomes. This underscores that WLC is a stressor that most workers would prefer to avoid. The fact that we document a uniformly positive relationship between WLC and unjust pay in September suggests that workers who experienced WLC felt they should be better compensated. By May of 2020, however, things changed. Many workers—especially more vulnerable ones—might have perceived fewer choices in navigating excessive work demands. When those created strain in the work-nonwork interface during the pandemic, individuals with less agency might have shifted their perceptions of the injustice of WLC. In the frame of the “greedy institution” idea, these workers might have found the greed of WLC less objectionable—at least during the turbulent times of a pandemic.

Funding

*Funding from the University of Toronto COVID-19 Action Initiative 2020 and Tri-Council Bridge Funding supports this research (Scott Schieman, PI). We thank Alex Bierman for his guidance and support with this paper.

Appendix A

See Table A1 .

Table A1.

A Weighted Descriptive Statistics for Focal Variables.

	September 2020 Mean or proportion (SD)	May 2020 Mean or proportion (SD)
Perceived Unjust Pay	.781 (2.063)	.606 (1.927)
Work-Life Conflict	2.866 (1.116)	2.557 (1.093)
Women	.484	.473
Visible Minority	.128	.124
Age	41.937 (13.721)	43.211 (13.355)
Education
Less than college	.303	.267
College degree	.540	.559
Graduate degree	.157	.174
Personal income
Low income	.368	.344
High income	.551	.586
Missing income	.080	.070
Financial strain	2.287 (1.049)	1.875 (.868)