Covid Lockdown and Employment in the Philippines

Abstract

The COVID-19 pandemic has severely impacted employment worldwide, mainly due to social distancing measures – hard lockdowns especially. Using difference-in-differences (DiD) analysis on Philippine Labor Force Survey (LFS) data, which exploits (1) the preemptive and selective application of a hard lockdown within the country; and (2) the conduct of the LFS coinciding with the imposition of the hard lockdown in April 2020, this study finds that the hard lockdown has a significant impact on employment apart from the general impact of the pandemic. The hard lockdown’s effect falls mainly on the intensive margin (weekly hours worked) rather than on the extensive margin (number of employed) per se. While employment and hours worked were generally down during the pandemic, the hard lockdown reduced weekly hours worked by an additional 18 hours. The most heavily affected workers are young male workers with low to medium educational attainment levels; and workers in sectors with low telework potential: Manufacturing, Construction, and Transportation. These results may inform the scope and form of government assistance given the limited fiscal space; and highlight the importance of developing digital skills and technologies to minimize the adverse employment impacts of hard lockdowns.

Keywords:

• extensive margin
• intensive margin
• hard lockdown
• employment
• COVID-19
• difference-in-difference analysis

Disclosure statement

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

Notes

1 See International Labor Organization (ILO, 2021).

2 For an example of a study that attempts to tease out the impact of the hard lockdown on employment, see Juranek, Paetzold, Winner, and Zoutman (2020), which is discussed in the next section.

3 Government of the Philippines. Executive Order No. 112-Imposing an Enhanced Community Quarantine in High-risk Geographic Areas of the Philippines and a General Community Quarantine in the Rest of the Country from 01 to 15 May 2020. https://www.officialgazette.gov.ph/downloads/2020/04apr/2020030-EO-112-RRD.pdf

4 https://www.rappler.com/newsbreak/in-depth/duterte-shock-and-awe-coronavirus-pandemic-response-generals/.

5 The stringency index is based on nine response indicators on containment and closures, and travel and mobility restrictions (Hale et al., 2021).

6 Unemployment figure for India is from the Center for Monitoring Indian Economy (2021).

7 Department of Labor and Employment. 2019. Guidelines on the Adjustment Measures Program for Affected Workers Due to the Coronavirus Disease 2019 (Department Order No. 209). Accessible at https://www.dole.gov.ph/news/department-order-no-209-guidelines-on-the-adjustment-measures-program-for-affected-workers-due-to-the-coronavirus-disease-2019/.

8 Using Generalao’s (2021) teleworkability of occupations index, Asian Development Bank (2021a) reports a similar finding for the Philippines: the hardest-hit sectors due to the COVID-19 crisis – construction, whole and retail trade, transportation and storage, and accommodation and food services – are also those with occupations with the least telework potential.

9 https://psa.gov.ph/statistics/survey/labor-and-employment/labor-force-survey/title/Employment%20Situation%20in%20April%202020

10 The data (LFS, 2019–2020) will be made available upon request.

11 The approach is similar to Ducanes and Ramos (2022), although in that paper the analysis focused on paid employment outcomes for women with and without children.

12 This and succeeding arguments are also made in Ducanes and Ramos (2022).

13 In this case, it can be said that the Stable Unit Treatment Value Assumption (SUTVA) is not met.

14 This is based on the Gross Regional Domestic Product in 2019, as reported by the PSA.

15 The full tables will be made available upon request.

16 In this sense the DiD coefficient, in the case of paid employment, measures the combined impact of the hard lockdown and the associated government measures to combat the hard lockdown.

17 We do this instead of directly using interaction variables for two reasons: (1) to be clear that DiD analysis by subgroup is valid only if the parallel trends assumption is met for each subgroup and to test for this assumption; and (2) to avoid the lengthy table resulting from using interaction variables. The advantage of using interaction variables is that it allows for formal testing of the difference in the DiD estimates by subgroup. We thus perform the estimation using interaction variables but just cite the results in this paper. The complete regression outputs will be made available upon request.

18 The parallel assumption is however unmet for this sector.

19 This is tested via a DiD regression with a fully interacted male dummy variable. The coefficient of the Lockdown period variable with the ECQ areas variable and the male dummy variable tests for the difference in the DiD impact between males and females.

20 See Box Figure 2.1b in Asian Development Bank (2021a).