Who can work from home in MENA?

ABSTRACT

The COVID-19 pandemic has had a profound impact on the world economy. The need for social distancing, lockdowns, or complete curfews has meant that this impact varied significantly across segments of society. Those unable to work remotely, or who work in settings necessitating close contact with others faced a trade-off between their lives and livelihoods. This trade-off was especially pronounced early on during the pandemic when vaccines had not yet been developed, hospitals were overwhelmed and governments were resorting to strict social distancing measures to mitigate the impact on their already strained healthcare systems. In this study, I examine the extent to which jobs can be successfully performed remotely in five MENA countries: Algeria, Egypt, Jordan, Palestine and Tunisia. I develop a teleworkability index using micro data on occupational characteristics. I find that relatively few jobs in MENA countries are compatible with teleworking and this share varies considerably by industry, gender, age and the formality of employment. I further investigate the ability to work from home in practice by considering the digital divide (a lack of reliable access to vital tools for teleworking, such as a personal computer and reliable internet access) as well as actual work from home behavior during the pandemic using real time surveys. I find that even for those who have high telework potential only few have access to computer and internet. Surveys conducted during the pandemic suggest that our measure of teleworkability was quite close to actual work from home behavior in each country.

KEYWORDS:

• Covid-19
• occupations
• work from home
• remote work
• MENA

JEL classification:

• J21
• J22
• O1

Acknowledgement

I wish to thank two anonymous referees, and the Editor, Prof. Leila Baghdadi, Meltem Dayioglu Tayfur, Vito Intini, Walid Merouani, and participants at the ERF Annual Conference in 2021 for valuable comments and suggestions on earlier versions of this study. The views expressed in this publication are those of the author and do not necessarily represent those of the United Nations, including UNDP, or the UN Member States. An earlier version of some of the analysis presented here appeared in a UNDP Regional Bureau for Arab States working paper on the impact of Covid-19 on Arab Labor Markets. The present paper is a substantial revision, update and extension of the analysis in several directions, using additional data that became available since the first study was produced and expanding substantially on the methodology. This work was sponsored in part by the Economic Research Forum (ERF) and has benefited from both financial and intellectual support. The views expressed in this work are entirely those of the author and should not be attributed to ERF, its Board of Trustees or donors.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 The views expressed in this publication are those of the author and do not necessarily represent those of the United Nations, including UNDP, or the UN Member States. An earlier version of some of the analysis presented here appeared in a UNDP Regional Bureau for Arab States working paper on the impact of Covid-19 on Arab Labor Markets. The present paper is a substantial revision, update and extension of the analysis in several directions, using additional data that became available since the first study was produced and expanding substantially on the methodology.

2 The Occupational Information Network (O*NET) is developed under the sponsorship of the U.S. Department of Labor, and provides information on daily activities and tasks for over 1000 occupations, https://www.onetcenter.org/.

3 For example, the monthly jobs report from the US Bureau of Labor Statistics in September 2020 reported that women had been leaving the labor force at four times the rate of men, primarily due to the increased burdens of parenting with children home from school, the lack of child care services for very young children and increased household chores of cooking and cleaning due to families being home all day (Hsu, 2020). Studies for the region using the ERF COVID Monitors, have also documented larger rates of transition to being out of the labor force for women than men during the course of the pandemic (Hlasny and AlAzzawi, 2022).

4 These are a series of short panel phone surveys that were conducted at several time periods during the span of the pandemic, between 2020 and 2021, in 3 of the countries under study: Egypt, Jordan and Tunisia.

5 These estimates are based on the ILO’s ‘nowcasting’ model which is a ‘is a data-driven statistical prediction model that provides a real-time measure of the state of the labour market, drawing on real-time economic and labour market data’. For more details see https://ilostat.ilo.org/resources/concepts-and-definitions/description-ilo-monitor/.

6 Estimates of working hour losses due to the Covid-19 pandemic by quarter available from ILOSTAT (as of November 2022) are only provided for country groups, not individual countries. Annual estimates are available by country, but do not show as much detail in terms of the changing impact of the pandemic and accompanying workplace closures, on employment, over the year. The ILO classifies ‘Arab States’ as comprising Bahrain, Iraq, Jordan, Kuwait, Lebanon, the Occupied Palestinian Territories, Oman, Qatar, Saudi Arabia, the Syrian Arab Republic, the United Arab Emirates and Yemen. It defines Northern Africa as comprising Algeria, Egypt, Libya, Morocco, Sudan and Tunisia.

7 See for example the World Bank’s COVID-19 High-Frequency Monitoring Dashboard, containing preliminary results of rapid phone surveys to assess the impact of Covid-19 on households. At the time of this writing only results for Djibouti, Iraq and Yemen were available.

8 UNDP report ‘Compounding Crises: Will COVID-19 and Lower Oil Prices Lead to a New Development Paradigm in the Arab Region?’ (UNDP, 2020) provides detailed summaries and references to several rapid assessments conducted to examine the impact of Covid-19 on key sectors and groups including health, poverty, labor markets, migrant workers, small and medium enterprises, women, among others.

9 See, for example, the O*NET database Content Model (https://www.dol.gov/agencies/eta/onet) or World Bank Group (2018).

10 The complete set of statements from the ‘Work Context’ questionnaire are (1) the average respondent states they use email less than once per month; (2) the majority of respondents say they work outdoors every day; (3) the average respondent says they deal with violent people at least once a week; (4) the average respondent says they spend the majority of time wearing common or specialized protective or safety equipment; (5) the average respondent says they spent the majority of time walking or running; (6) the average respondent says they are exposed to minor burns, cuts, bites, or stings at least once a week; and (7) the average respondent says they are exposed to diseases or infection at least once a week. The complete set of statements from the ‘Generalized Work Activities’ are (1) performing general physical activities is very important; (2) handling and moving objects is very important; (3) controlling machines and processes [not computers or vehicles] is very important; (4) operating vehicles, mechanized devices, or equipment is very important; (5) performing for or working directly with the public is very important; (6) repairing and maintaining mechanical equipment is very important; (7) repairing and maintaining electronic equipment is very important; (8) inspecting equipment structures or materials is very important. If any of these statements are true the occupation is coded as one that cannot be performed from home.

11 Calculations from the Algerian Labor Force Survey 2014 were performed on my behalf by Walid Merouani, who had access to the micro-data files for the survey, as it is not publicly available.

12 Dingel and Neiman (2020) provide the following example to explain this process: ‘if a particular SOC has 100 US employees and is associated with two ISCOs that have respective totals of 3000 and 1000 employees in a country, we allocate 75 of the SOC's US employees to the larger ISCO and 25 to the smaller one. Those values of 75 and 25 are then used as that SOC's weight when calculating the average across all SOCs within each ISCO for that country’ (Dingel & Neiman, 2020, p. 6).

13 The sample includes all those who were employed in each survey, including subsistence workers, as long as they had an ISCO occupational classification.

14 I also estimated a simple OLS regression using TWI as a continuous variable and the results were qualitatively almost identical. I chose to report the results based on the binary low teleworkability variable instead to avoid misinterpretation since the measure cannot necessarily be interpreted linearly. Using the binary variable provides a benchmark and allows us to answer the question: ‘what are the characteristics of workers least likely to have a teleworkable job?’. The results for the continuous TWI OLS regressions are in the appendix, Table A2.

15 A similar analysis was performed, defining $y_{ij}$ relative to the median instead of the mean. The results were similar in terms of signs but the results based on the mean had a higher significance. In the interests of brevity, only the results based on the mean are reported. The other results are available upon request.

16 These are a series of short panel phone surveys that were conducted at several time periods during the span of the pandemic, between 2020 and 2021, in 3 of the countries under study: Egypt, Jordan and Tunisia.

17 The Labor Market Panel Surveys for Egypt, Jordan and Tunisia include a question on current place of work. This can give insight into jobs that are already performed at home, even though they might not be ‘teleworkable’ per se, such as a home-based catering business or craft business. The response to this question includes many categories such as ‘own home’, ‘shop/restaurant’, ‘workshop/factory’, etc. I calculated the share of workers in each survey who responded that their current place of work was ‘own home’. That share was 2.9 percent for Egypt, 1.6 percent for Jordan and 1.5 percent for Tunisia. For those whose place of work is home, the share of teleworkable jobs according to the TWI is 15 percent in Egypt, 12 percent in Jordan and 10 percent in Tunisia. Women are more likely to have responded that their place of work is home in Egypt and Tunisia (in Egypt 5.8 percent for females versus 1.4 percent for males, in Jordan 1.3 percent for females versus 1.7 percent for males, in Tunisia 3.9 percent for females versus 0.6 percent for males). While this distinction between work from home and telework is important to consider, those who were actually working from home in those surveys represent a very small share of workers in all three countries (under 3percent) and therefore do not substantially alter our results. Thanks to an anonymous referee for raising the issue of current work place location and how it relates to the TWI.

18 Employment shares are reported in all tables and figures in conjunction with the TWI to give perspective on how large a particular group of workers is compared to overall employment, and hence the overall weight that group’s TWI would have on the countrywide TWI. For example, while jobs held by women in Algeria had a high TWI of 0.62 according to the index, women only made up 14 percent of those who worked at the time of the survey and hence this still translated to a relatively low TWI for Algeria of 0.23.

19 It is worth noting while education jobs may be teleworkable in principle, their teleworkability depends critically on access to appropriate devices and a reliable internet connection for both students and teachers, which varies drastically by country and socioeconomic group within each country.

20 In practice, social security was the more important factor in defining formality, since very few workers had social security but no contract, while many had a contract but no social security benefits. This makes the results comparable to those for Algeria.

21 Irregular wage workers, who make up about 30 percent of all informal workers in Egypt, Jordan and Tunisia, are likely to be particularly vulnerable given the nature of their employment. I found that the share of jobs that would be teleworkable for irregular wage workers is even lower at just 5 percent in Egypt, 7 percent in Jordan and 2.5 percent in Tunisia; compared to 23, 33 and 19 percent for regular wage workers, respectively. Thus, the share of teleworkable jobs for irregular wage workers are just one third to one half the shares of teleworkable jobs for all informal workers in these three countries as presented in Figure 4.3. For perspective, irregular wage workers’ share of total employment is 15 percent in Egypt and Jordan and 12.5 percent in Tunisia. Data on irregularity of employment was not available for Algeria and Palestine but I expect their teleworkability levels to be similarly low.

22 The surveys for Jordan and Tunisia only included a subset of the questions that could be used to construct the index and even where the questions were included, only very few observations had responses to these questions, with the majority missing. I therefore had to rely exclusively on the labor market panel survey for Egypt for the construction of the index.

23 Dey et al. (2020) examined the relationship between teleworkability as measured by Dingel and Nieman’s O*NET based classification, and the actual incidence of working from home as measured by two US Bureau of Labor statistics surveys and found that the ‘margin of error’ (percent of jobs classified by Dingel and Nieman’s measure as not teleworkable that actually were performed at home) was quite low, ranging from 4 percent to 6 percent depending on the survey used. Furthermore, in May 2020, at the height of workplace closures and lockdowns in the United States, Bloom, Barrero and Davis conducted surveys of 20–64 year old workers who had worked full time in 2019 and found that 42 percent of the U.S. labor force were working from home full time (quite close to Dingel and Nieman’s estimate of 37 percent), while another 33 percent were not working at all due to the Covid induced recession, and the remaining 26 percent were working on their business’s premises, most of whom were essential workers (Bloom, 2020). These two sets of results confirm the usefulness and applicability of the Dingel and Nieman classification used in the current study since the margin of error for jobs that cannot be performed remotely was very low, while the actual share of jobs that were performed remotely during the strictest lockdowns (42 percent) was very similar to their estimate of 37 percent.

24 See for example Diwan and Haidar (2017); Assaad et al. (2019a,2019b) ; Assaad et al. (2020), among others).

25 Several recent studies have shown that politically connected firms in the region that tend to capture the lion’s share of the profits as well as preferential access to capital, resources and regulatory privileges are less productive and create few jobs, while crowding out smaller more efficient firms (El-Haddad (2020), Diwan et al. (2015), Diwan and Haidar (2017)).

26 See for example Morsy et al. (2015).