Principal investigator: Eduarda Miller de Figueiredo
Article title: Does Privatization Benefit the Poor? The Effects of Privatizing Basic Sanitation on Infant Mortality
Author of the article: Thomas Fujiwara
Location of the intervention: São Paulo and Rio de Janeiro, Brazil
Sample size: 436 Municipalities
Sector: Health Insurance
Type of intervention: The Effect of Privatizing Basic Sanitation
Primary variable of interest: Municipal Mortality Rate
Evaluation method: Differences-in-Differences
Policy Problem
Basic sanitation is a service linked to the fundamental hygiene needs of every human being, with infant mortality being one of the areas it impacts. Furthermore, studying infant mortality is important in three ways:
i) Infant mortality is seen as one of the eight "Millennium Development Goals" to be pursued by all UN countries;
ii) Access to basic sanitation is linked to infant mortality, since waterborne diseases are among the leading causes of infant mortality;
iii) the existence of econometric evidence of significant impacts of sanitation privatization on reducing infant mortality (Galiani, Gali and Schargrodsky, 2005).
However, it is difficult to establish what the effects of sanitation privatization will be. Given that greater efficiency in the private sector will result in greater access to treated water and sewage services and better service quality, the author argues that the negative impact on access and quality cannot be disregarded, as private companies may not internalize the positive externalities of water on health and may focus their operations on wealthier and more profitable areas.
Implementation and Evaluation Context
In Brazil, virtually all basic sanitation has been the responsibility of the public sector since the National Sanitation Plan (Planasa) in 1971. In 2005, the year of the study, 90% of Brazilian urban households had access to water; however, less than half of the urban population had access to the general sewage network, which treated only 27% of its waste (Motta and Moreira, 2004).
Planasa enabled a significant increase in access to treated water, but by the 90s evidence was emerging that state budget constraints would not allow basic sanitation to continue developing without private sector intervention. In 1998, only 2% of the Brazilian population had privatized basic sanitation services, and in 2004, only 63 of the more than five thousand municipalities had concessions or permits for the private sector to operate (Parlatore, 2000; Turolla, 2004).
Furthermore, children have weaker immune systems compared to adults and lack the knowledge to avoid the risk of contamination by diseases transmitted through contaminated water, thus suffering more from poor sanitation. In this context, it is known that diarrhea is responsible for 15% of infant deaths worldwide, and the average infant mortality rate in 1991 and 2000 was 6,2 and 4,2 deaths per thousand inhabitants of children aged 0 to 4 years, respectively.
Policy Details
Privatization still lacks studies that assess its impact on society, especially on the less privileged segment of the population. Basic sanitation allows us to test the impacts of privatization on the well-being of these individuals, since its services are directly linked to basic hygiene and health needs, and its structure varies over time, from public to private, allowing for the creation of treatment and control groups.
Rio de Janeiro and São Paulo were the states that had more than one municipality with privatized basic sanitation services in the year 2000. By using data from these same municipalities in 1991, it was possible to obtain data for the pre- and post-intervention periods, enabling the application of the econometric model that allows for the estimation proposed by the study. Thus, the treatment group consists of those municipalities that had their basic sanitation service subsidized by the private sector. While the control group had the same service provided by the municipal administration, public companies, or mixed-economy companies.
Methodology Details
The objective of this article was to measure the average effect of the privatization of water and/or sewage services in a municipality on its under-five mortality rate. Given the existence of a similar study conducted in Argentina by Galiani, Gali, and Schargrodsky (2005), this article uses the same estimation technique, the difference-in-differences model with fixed effects, allowing for a comparison between the results. Therefore, following the model of the Argentinian study, data from 1991 and 2000 were used, the pre- and post-intervention periods, respectively, allowing for a comparison of the results with the neighboring country.
Data on location and type (water or sewage) were obtained from the Brazilian Association of Private Concessionaires of Basic Sanitation Public Services – ABCON. For the year 2000, there were 28 municipalities with some type of private concession in this service; however, three of these municipalities were only created after 1991, and three did not have observations for some covariates. Therefore, 22 municipalities with privatization in basic sanitation were used. The author also used mortality data from the Mortality Information System (SIM) to collect the infant mortality rate, since this database allows disaggregation by cause of death.
To estimate the results, panel data with fixed effects were used, where the variable of interest is the municipality's mortality rate. Since there is self-selection for the intervention, meaning the municipal government chooses to privatize, a series of covariates related to the municipality's economic and educational situation were added to the model. Municipal spending on health, per capita sanitation, and the percentage of the population with access to garbage collection were also used as indicators of the importance the local government places on public health.
Main results
Estimates have shown that the privatization of sanitation services has an average effect of reducing infant mortality by 10,3% and 12,4% in 1991. Comparing this with a study conducted in a neighboring country, the effect is greater, as Argentina shows a reduction of 4,5% and 9,7% in infant mortality.
However, the results obtained could be subject to some type of bias given the lack of municipalities comparable to those that privatized sanitation services in the control group and the different distributions of the covariate vector. To eliminate possible bias, the author used the Propensity Score The model was used to estimate the probability of a municipality privatizing basic sanitation in the next decade. This new sample, called the common support, contains a panel of 369 municipalities. The results of this model also indicate an average effect of a 12% reduction in infant mortality.
Furthermore, the author proposed testing the impacts of privatization on different causes of infant mortality. Estimates indicated that privatization had a 22% reduction in infant mortality from infectious and parasitic diseases, but no impact on infant mortality from other causes. Therefore, access to treated water reduced deaths from diseases transmitted through contaminated water.
However, the results also indicate that privatization did not cause an increase or decrease in the proportion of households with access to the general sewage system, but it did have a significant negative effect on the proportion of households with access to the general piped water system.
Lessons in Public Policy
The study demonstrates that the privatization of basic sanitation affects infant mortality among children under five years old, since increased access to treated water and sewage reduces contact with waterborne diseases. Furthermore, when compared to results found in Argentina, it serves as a warning to Brazil that good regulation is necessary for privatization to produce benefits for society.
Reference
FUJIWARA, Thomas et al. Does privatization benefit the poor? The effects of the privatization of basic sanitation on infant mortality. Proceedings of the XXXIII National Meeting of Economics, 2005.
