Reliable Electricity Access, Micro and Small Enterprises, and Poverty Reduction in Indonesia

Abstract

While subsidising electricity tariffs, the Indonesian government has launched a massive electrification campaign to improve the quality of life of Indonesians. However, whether these efforts have increased access to energy for all and whether this access reduces poverty is unclear. We empirically test the causal effects of electricity access on poverty reduction. We also investigate the potential role of micro and small enterprises (MSEs) as the transmission channel for poverty reduction. To isolate any endogeneity, we use the instrumental variable approach. We exploit the village’s proximity to the nearest power plant in 1985 as the exogenous variation of the historical least-cost distance by the state-owned electricity company, PLN, to instrument the endogenous nature of current-time electricity access. Our results show that expanding reliable electricity services contributes significantly to poverty reduction. Nevertheless, we find no evidence that the development of MSEs has an influential mediating role in the poverty reduction effects.

Keywords:

• electricity access
• poverty
• instrumental variable
• Indonesia

JEL classifications:

• H54
• L94
• O15
• O18
• R23

ACKNOWLEDGMENTS

The authors thank PLN, the Ministry of Energy and Mineral Resources and BPS for providing data. We also thank Professor Djoni Hartono and Dr. Qisha Quarina for their valuable inputs to improve the analysis of this study and Marcha Violeta for providing excellent research assistance. We thank the three anonymous referees for their generous and careful inputs and comments on the earlier drafts of this article.

FUNDING DETAILS

Universitas Indonesia funded this study through the Pendamping Publikasi Internasional (PPI), Q1 2021, (research grant number NKB-628).

Notes

1 The unconditional scatter plot shows a weaker correlation than the conditional plot when we control for the unobservable, which indicates the endogeneity problem addressed in the empirical section.

2 The Podes data are available only for 2014 and 2018. However, our data set includes 2014, 2015, 2017 and 2018. Therefore, we merged 2014 data on covariates from Podes with our data from 2014, 2015 and 2017. We then merged the 2018 data on covariates from Podes with data from our 2018 set.

3 Appendices can be accessed online at http://dx.doi.org/10.1080/00074918.2023.2175782

4 Tier 1 is defined as electricity access where the maximum amount of energy required to support different power loads is 1–50 watts, where electricity is available for only 25%–50% of the average required operating hours (1–2 hours of electricity access divided by the 4 required hours), where the number of unscheduled outages per week is more than 4 and the cumulative length of unscheduled outages per week is more than 2 hours.

5 Each wave of Podes data showed that 45% to 46% of the excluded villages had electricity access. This almost even composition suggests that the data are missing at random, limiting the omission’s ability to bias the estimation.

6 Appendices can be accessed online at http://dx.doi.org/10.1080/00074918.2023.2175782

7 The Stock-Yogo critical value was 16.38, calculated for one endogenous regressor, one instrument, a 10% maximum instrumental variable relative bias and independent identically distributed errors.

8 Appendices can be accessed online at http://dx.doi.org/10.1080/00074918.2023.2175782