Job creation or labor absorption? An analysis of private sector job growth in Egypt

ABSTRACT

Creating jobs, especially good jobs, is one of the greatest challenges facing Egypt. This paper investigates the nature of job growth in Egypt, including the firm, industry, and worker characteristics that are related to job growth. Using data from Egypt’s establishment censuses linked to various firm and labor surveys, we examine job growth in private sector establishments over 1996–2017. We find that job growth has primarily followed a labor absorption paradigm, with job growth unrelated to productivity and highest for firms with more informal employment.

KEYWORDS:

• Employment
• job creation
• job growth
• Egypt

JEL Codes:

• J21
• J23
• D22
• N35

Acknowledgments

The authors acknowledge the support of the Economic Research Forum under the ‘Egypt Labor Demand Project.’ The authors appreciate the comments of project workshop participants and Bob Rijkers on earlier paper drafts. The contents and recommendations do not necessarily reflect the views of the Economic Research Forum.

Disclosure statement

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

Notes

1 This figure is based on the Egypt Labor Market Panel Survey (ELMPS) 2018. Figures from the official labor force survey are slightly higher. Krafft and Assaad (2014b) and Krafft, Assaad, and Rahman (2019) discuss the difference in the unemployment estimates.

2 Based on authors’ calculations using data from ELMPS 2006, 2012 and 2018.

3 Based on authors’ calculations using data from ELMPS 2018.

4 By the end of 2018, the credit extended by Egyptian banks to small and medium enterprises at concessionary interest rates had reached EGP 136 billion, benefitting 522 thousand firms (Omran, 2019).

5 An establishment is defined as a fixed location in part of a building, a building, or a group of related buildings, which is dedicated to the execution of an economic activity (CAPMAS, 2017). As censuses, these data capture all fixed establishments where economic activity takes place. If establishments are temporarily closed, they are enumerated as non-operating and fall out of our sample.

6 Other characteristics in the data included whether the establishment was an independent entity or a branch and its legal status, but these were not measured consistently over time and are therefore excluded from our analyses. Other variables, such as district, exist in the original data but were not provided to researchers in the microdata.

7 We received only the 10% sample for Suez in 2006 and therefore give each establishment there a weight of 10.

8 The 10% random sample in 1996 is a simple random sample of establishments, which may create random variation in total employment, particularly in industries where larger firms are more common. To account for receiving a 10% random sample, we used the tabulated results from the 100% data to create worker counts in governorate and 3-digit industry cells, which are used to create expansion factors for our 10% sample.

9 The way in which the linking is done is described in more detail below.

10 See Ismail and Abd El-Kader (2016) for a detailed description of the 2012/2013 EcC sampling methodology.

11 We use the 2012 wave of the ELMPS because it falls in between the 2006 and 2017 rounds of the EsC.

12 While 91.2% of firms retained their original four-digit coding, 0.4% moved to three-digit coding, 0.9% to two-digit coding, 2.0% to one-digit coding, and 5.5% were recoded to the most common four-digit industry within their one-digit industry-region combination.

13 The two-digit industry level was selected as the level at which firms were likely to compete.

14 The translog production function, estimated at the establishment level, has as its dependent variable log value added. The explanatory variables are the log value of capital and log number of workers. These are included as main effects, squares, and interactions between capital and labor. The predicted value of log value added from this model is subtracted from the actual log value added to calculate TFP as a residual.

15 When an industry was missing from the LFS or had fewer than three observations we used the next highest level of economic activity available as a proxy.

16 This sampling strategy is adopted to be compatible with the EsC and the EcC data sets. The establishment and economic censuses cover employment only inside fixed establishments in public enterprises and in the private sector. Our sample excludes the public enterprises leaving us only with workers in fixed establishments in the private sector.

17 Professional/managerial includes: Managers and Professionals. White collar includes: Technicians and associate professionals; clerical support workers; and service and sales workers. Blue collar includes: Skilled agricultural, forestry, and fishery workers; craft and related trades workers; plant and machine operators and assemblers; and elementary occupations.

18 When an industry was missing from the ELMPS 2012 or had fewer than three observations we used the next highest level of economic activity available as a proxy.

19 The standardized variables are standardized across the region-industry cells and pooled over the three years.

20 Results for the average product of labor are similar.

21 We checked for a non-linear relationship, but there appeared to be no relationship of any kind.

Additional information

Funding

This work was supported by the Economic Research Forum (Egypt).