The effect of government expenditure on income inequality in Sri Lanka

ABSTRACT

The government of Sri Lanka has constantly allocated a substantial budget to provide free health care, education, and other community services, intending to reduce income inequality. To see whether government spending has reduced inequality, we analyse data from Sri Lanka. We find that public expenditure is ineffective for the policy objective; instead, it increases inequality.

KEYWORDS:

• Government expenditure
• income distribution
• inequality

JEL CLASSIFICATION:

• D30
• D63
• H50
• O15

Disclosure statement

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

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/13504851.2022.2073328.

Notes

¹ Doerrenberg and Peichl (2014), Anderson et al. (2017), Hailemariam, Sakutukwa, and Dzhumashev (2020), and Sidek (2021) obtain similar results analysing different groups of countries.

² We obtain the Gini coefficient from the standardized world income inequality database; https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/LM4OWF. We collect the government expenditure from the Central Bank of Sri Lanka; https://www.cbsl.gov.lk/sites/default/files/cbslweb_documents/publications/annual_report/2020/en/16_S_Appendix.pdf.

³ We collect the government revenue, budget deficit, and debt from the Central Bank of Sri Lanka; see https://www.cbsl.gov.lk/sites/default/files/cbslweb_documents/publications/annual_report/2020/en/16_S_Appendix.pdf. We obtain the GDP per capita from the World Bank; see https://databank.worldbank.org/source/world-development-indicators. Finally, we use the globalization index of KOF Swiss Economic Institute; see https://kof.ethz.ch/en/forecasts-and-indicators/indicators/kof-globalization-index.html. This index represents the social, economic, and political aspects of globalization. The websites are last accessed on 19 November 2021.

⁴ We also consider a dummy for the United National Party (UNP) to be governing (1977–1993, 2001–2003, and 2015–2016). The UNP is a dominant party along with the Sri Lanka Freedom Party. The estimated coefficients on that dummy are near zero and statistically insignificant (10%) for all specifications, but all giving similar predictions about the effect on inequality of government expenditure.

⁵ The tsunami has wiped out a significant portion of the land, demolishing hundreds of thousands of properties and a substantial portion of socio-economic infrastructure. It caused a huge economic loss; see https://www.worldbank.org/en/news/feature/2014/12/23/lessons-learned-sri-lanka-tsunami-reconstruction.

⁶ For similar topics, Padhan et al. (2020) and Cevik and Correa-Caro (2020) study Indian and Chinese, respectively, yearly time series data from 1980 to 2013, and Raza, Shahbaz, and Reddy Paramati (2016) analyse Pakistan yearly data from 1972 to 2012.

⁷ The first-stage $F$ statistics are above ten for all regression specifications where we use the instrumental variables, and we do not reject the hypothesis that the instruments are exogenous at the five percent significance level for all regression specifications where we use these instruments; see Tables S2 and S3.