Does Working from Home Increase the Gender Wage Gap? Insights from an Italian Survey of Occupations

ABSTRACT

This article investigates to what extent the working from home (WFH) feasibility of occupations can influence the gender wag gap (GWG) at the mean and along the wage distribution. Based on Oaxaca–Blinder decompositions and unconditional quantile regressions, results show that the GWG is greater among women working in an occupation with a high level of WFH feasibility. We find evidence of both sticky floor and glass ceiling effects for employees with high WFH feasibility and only a sticky floor effect for the group with low WFH feasibility. The positive association revealed between the level of WFH feasibility and the GWG appears particularly strong among older and married women employees. These results underscore that the WFH feasibility may play an important role in exacerbating future gender gaps in wages, as WFH is expected to remain a normal practice beyond the COVID-19 pandemic.

HIGHLIGHTS

• Working from home (WFH) feasibility of occupations influences the gender wage gap (GWG) in Italy.

• GWG is overall wider among employees with high WFH feasibility than employees with low WFH feasibility.

• The GWG is higher among older and married women employees.

• A possible explanation is Italy’s ungenerous work–family policies and poor childcare availability.

• There is a need for policies aimed at regulating WFH from a gender-equality perspective.

KEYWORDS:

• Working from home
• gender wage gap
• Oaxaca–Blinder decomposition
• RIF
• regressions
• COVID-19

JEL Codes::

• D31
• J31
• I20

ACKNOWLEDGMENTS

We thank participants at the International ASTRIL conference (November 2020) the 2021 ESPE Conference (June 2021), the Social Situation Monitor Research Seminar (September 2021), the ICEA “After the Pandemic” Conferences (November and December 2021), and the LabMinerva Seminar (March 2022) for their valuable comments and suggestions. The opinions expressed in the article reflect only the authors' views and not their institutions'. The INAPP is not responsible for any use that can be made of the present results. The other usual disclaimers apply.

SUPPLEMENTAL DATA

Supplemental data for this article can be accessed online at https://doi.org/10.1080/13545701.2024.2326509.

Notes

1 https://www.upwork.com/resources/how-to-adjust-to-the-new-normal-of-remotework.

2 For the sake of clarity, there is a glass ceiling effect when the GWG at the 90th percentile is significantly greater than that estimated at the 50th percentile (that is, the median). Instead, there is a sticky floor effect when the GWG at the 10th percentile is significantly greater than that estimated at the median.

3 The “spoils system” is a practice in which civil servants take turns as the government changes, otherwise the “merit system” consists in a process of promoting and hiring of employees on the basis of their ability to perform a job.

4 Rather surprisingly, Jiří Balcar and Veronika Hedij (2019) show that a manager’s gender influences the wage level but not the GWG in the Czech Republic.

5 More information on the data is provided in the next section.

6 Bobbi Thomason and Inmaculada Macias-Alonso (2020) argue that caregiving – where women are overrepresented – is relevant and underpaid work.

7 However, the pandemic is expected to impact women more severely than men not only through a work/income point of view. For instance, Graziella Bertocchi and Arcangelo Dimico (2020) find that among African Americans, women face a much higher probability of death from COVID-19, and Kathryn J. Holland et al. (2020) bring attention to the fact that sexual harassment and discrimination can still take place even with remote work. In addition, Colleen Flaherty (2020) and Philippe Vincent-Lamarre, Cassidy R. Sugimoto, and Vincent Larivière (2020) argue that the ability of women to innovate or contribute to research appears has been at risk (or has at least narrowed) during the COVID-19 crisis. Gender differences may also be related to the fact that women tend to take the pandemic more seriously and to be more compliant than men (Galasso et al. 2020). Finally, Sandeep Mohapatra (2020) points out that gender differences in the pandemic’s economic impacts are not solely seen in developed countries but are also observed in developing countries.

8 Claudia Goldin (2014) makes clear that during the last century, the converging roles of men and women are among the greatest novelties in society and the economy, following on from several momentous chapters in a history of gender equality. According to her, the “last” chapter does not (necessarily) involve government intervention and it need not make men more responsible in the home. It must be related to changes in the labor market, and particularly how jobs are structured and paid, with a view to improving temporal flexibility. The GWG would be significantly reduced if firms did not have an incentive to pay individuals who worked long or odd hours significantly more.

9 One of the key elements of this survey, which is performed using a computer-assisted telephone interviewing (CATI) approach, is the absence of proxy interviews. In other words, to reduce measurement errors and partial non-responses, only survey respondents are reported in the dataset. The questionnaire was distributed to a sample of residents according to a stratified random sampling of the Italian population.

10 The ICP survey ensures representativeness with respect to sector, occupation, firm size, and geographical domain (macro-region). On average, twenty workers per each Italian occupation are included, providing representative information at the 5th digit. Similar to the American O*Net, occupation-level variables in the ICP survey are collected relying on both survey-based worker-level information and post-survey validation by expert focus groups.

11 An example of an analysis of the effects of WFH practices on income levels during the pandemic in Italy is Carmen Aina et al. (2021).

12 Additional information on the literature regarding the definition of WFH can be found in the section on the explosion of WFH during COVID-19.

13 A sensitivity analysis on the age restriction adopted was implemented, however, and the results are presented in the section on robustness checks.

14 For the sake of brevity, in Figure 1 we do not provide the same graphs for the other four items of the WFH feasibility index. To be clear, however, the group of occupations where physical proximity or maneuvering vehicles, mechanical vehicles or equipment is not important shows a positive association with gender segregation and the group for which dealing with external customers/public or spending time standing is not important has a negative association with gender segregation.

15 The selection of these covariates is confirmed by a search for relevant confounders of the probability of having an occupation with a high WFH feasibility level. Further estimates on the available sample, based on logit regression models, show that all of these variables except the dummies for marital status and household size significantly increase the probability of performing a highly WFH-feasible occupation. This evidence then strongly indicates for controlling for these covariates in our main econometric analysis to avoid potential biases in the final outcomes. More details are available upon request to the authors.

16 For the sake of clarity, the effects on income distributional statistics related to a marginal (or not full) swap of women to men in the labor market can be obtained by scaling coefficients illustrated in the results section (for example, in Table 2 or Figure 3). For instance, if we hypothesize a 10 percent swap of women to men in the labor market, the estimated coefficients of the variable of interest (that is, the GWG in this case) have to be multiplied by 0.1, while their statistical significance remains the same.

17 Although the number of hours worked in a week is available in the INAPP-PLUS data, the survey questionnaire does not collect any information on the number of weeks (or months) worked. This lack of information prevents us from correctly calculating the hourly wage earned by workers.

18 The test of equality of coefficients (Wald test) confirms that the overall difference and the explained portion are different between “high WFH feasibility” and “low WFH feasibility,” while the unexplained portion is the same in the two estimated decompositions.

19 See https://ec.europa.eu/commission/presscorner/detail/en/qanda_20_935.

Additional information

Notes on contributors

Luca Bonacini

Luca Bonacini is Assistant Professor at the University of Bologna and is a research fellow at the Global Labor Organization (GLO). After a period of internship training at the Bank of Italy, he received his PhD in Labour, Development and Innovation at the University of Modena and Reggio Emilia in 2021. He was visiting scholar at the Luxembourg Institute for Socio-Economic Research (LISER) and the University of Luxembourg. His main research interests concern income distribution and inequality of opportunities, with a focus on education and gender issues.

Giovanni Gallo

Giovanni Gallo is a Senior Assistant Professor at the University of Modena and Reggio Emilia, and previously was a postdoctoral research fellow at the Sapienza University of Rome and the National Institute for Public Policies Analysis (INAPP). He earned his PhD in Labour, Development and Innovation at the University of Modena and Reggio Emilia in 2018. He was a visiting scholar at the Tilburg University and the Luxembourg Institute for Socio-Economic Research (LISER) to work on financed research projects. His research interests include evaluation of social policies, poverty, inequality, redistribution, household finance, and pensions.

Sergio Scicchitano

Sergio Scicchitano is Associate Professor at John Cabot University in Rome and Chair of the Department of Economics and Applied Economist at National Institute for Public Policies Analysis (INAPP), Rome, Italy. He is also Leader of the “Coronavirus” thematic research group and of the “Italy” country research group at the Global Labor Organization (GLO) and Section Editor for “Covid-19” in the 2022 “Handbook of Labor, Human Resources and Population Economics.” He was Visiting Lecturer at the Queen Mary University of London and Adjunct Professor of Economics at the Sapienza University of Rome. He earned his PhD in Economics at the Sapienza University of Rome. His personal website is https://sergioscicchitano.wordpress.com/.