A Comparative Analysis of the Public and Private Water Supply and Sanitation Service Providers ’ Work in Brazil

This paper aims to evaluate the performance of the private management of water supply and sanitation services in Brazil. We can verify therefore, whether this management model has presented superior performance compared to the model of public management. The analysis was based on various economic-financial and operational indices related to water supply and sewage collection and treatment between 2000 and 2010. The results indicate that private companies outperform public entities in several aspects, like productivity, investment, return and access to these services, pointing out the concessions and public-private partnerships are feasible ways to reach political targets of universalization of Water Supply and basic Sanitation.


INTRODUCTION
he extension and the quality of the infrastructure are determining factors for the development of a country.Ascher and Krupp (2006) state that infrastructure is the backbone of a developed economy and a pillar of quality of life.
The lack of infrastructure investments has been identified in Brazil.According to Pinheiro (in OLIVEIRA et al, 2013, p. 103), it has been 3 decades since the country invests a little over 2% of its GDP, whilst its direct competitors in the international market and Latin American countries invest from 4% to 8% of their GDP.Deficiencies in infrastructure greatly affect the population, representing an obstacle to social welfare.
To provide water supply and sanitation infrastructure is a public health issue.The PNUD (2006) (United Nations Development Program) in their Human Development Report, it is observable that worldwide, one million eight hundred thousand children die each year from diarrhea and other diseases, as a result of unsuitable water for human consumption and the and the absence of sewer installations.This is the second leading cause of infant mortality in the early 21st century.Furthermore, investment in water supply and sanitation provides productivity gains and savings of public resources in health, as for for every US$1 invested it is possible to generate a potential benefit estimated between US$5 and US$28, depending on the country (HUTTON; HALLER, 2004).
In order to meet the investment needs, governments have adopted models that lead to a greater private sector participation in the provision of public services, through privatizations, concessions and public-private partnerships.These models have been disseminated by most countries, becoming a widespread and relevant practice within the public sector reform process, being intensified from the eighties.
In this paper four complementary analysis are developed with the objective to verify whether private management of the water supply and sanitation services have presented a higher performance compared to public management, in operational, administrative and financial terms.In addition, we evaluated the willingness to invest and to achieve the goal of services universalization.For this end, we used a database of 1.203 providers of these services in Brazil between 2000 and 2010, this data coming from the SNIS (National Sanitation Information System).
A Comparative Analysis of the Public and Private Water Supply and Sanitation Service Providers' Work in Brazil 117 BBR,Braz. Bus. Rev. (Engl. ed.,Online), Vitória, v. 13, n. 1, Art. 6, p. 115 -136, jan.-feb. 2016 www.bbronline.com.brThe main contribution of the paper is to verify whether the decisions to delegate management of water supply and sewage services to the private sector provided the expected results, with the objective of universalizing services.Another contribution is to trigger discussion on performance metrics which can be used in periodic reviews of tariffs and contract renegotiations.
The paper is organized over five sections, including this introduction.In the second section we go through the literature review on the theme.The third section presents the methodology and in the fourth section we present the results of the research.In the fifth section are the conclusions of this study.

CONCESSIONS AND PUBLIC-PRIVATE PARTNERSHIPS
According to Evans et al (2004), the introduction of the private sector in providing urban water services provides a unique opportunity to address a service economic policy for the poor and informal communities.For Guasch et al (2007), from the eighties on, the trend in Latin America and the Caribbean to transfer the provision of public services to the private sector began, due to the need to invest in infrastructure, the scarcity of public resources and from the assumption of greater efficiency of the private sector.These transfers took several forms, such as management contract, concession or privatization.
According to Silvestre (2012) the reasoning for the change in the provision of public sector services to the private sector has its genesis in the belief that there would be more efficiency and productivity and focus on users' needs, when that latter were responsible for the services.According to Bennett (20120, since private companies work in pursuit of profits they would always tend to increase their productivity, achieving higher performance. It should be admitted that part of the lower efficiency of the public sector stems from the restrictions issued by regulatory standards imposed on the public administrator, or due to the absence of an incentive structure that induces the civil service to achieve higher goals consistently.In addition, Mello (apud SHINOHARA;SAVOIA, 2008, p. 4) also cites the agency cost of the public sector, whose governance becomes more complex due to the existence of a third agent, politicians who prioritize maximizing their budgets instead of efficiency, because of the political prestige that it entails.
However, as pointed by Ruester and Zschille (2010), there is no consensus on the best way to achieve this goal among the various modes: public-private partnerships, concessions and privatizations.This is due to differences among countries in their legislation, practices and institutions that lead to different results among countries.
In sum, private service providers present as advantages: (i) their greater access to the necessary capital in order to invest, given the fiscal situation of states and municipalities making them unable to obtain credit; (ii) insertion into a more flexible legal and regulatory environment, without the binding nature of the Bidding Law, conferring the private sector greater flexibility in the business management; (iii) labor relations without the inefficiencies of the public sector, allowing greater ability to achieve efficiency; (iv) the non-existence of agency cost of the public sector; (v) access to human and material resources in order to seek new technologies that will result in the improvement of the services provided; (vi) their motivation for efficiency gains; and (vii) the competitiveness of the bidding process, which contributes to the practice of reasonable tariffs.
It must be considered that the provision of public services represents the concession to operate a monopoly, and its surrender to the private sector must ensure that the agent keeps the quality of services and the practice of reasonable tariffs.Da Motta and Moreira (2006), who researched the monopoly power of service providers between 1998 and 2002, obtained evidence that, in the absence of incentives for improved efficiency, service providers relieve productivity and apply higher tariffs.
The need for regulation in the sector and proper accountability is emphasized by several authors, such as Locussol (2006), Kessides (2004) and Brocklehurst and Janssens (2004).
Seen in these terms, it is appropriate to quote the World Bank's recommendation, in which the pertinent regulation regarding water supply must achieve three goals, of economic and social well-being character: (i) efficiency in production and in supplying water, at the lowest possible cost; (ii) equal access to the entire population at affordable prices and with a quality service delivery; and (iii) sustainability from an environmental viewpoint, to minimize damage to natural resources (KESSIDES, 2004).
It should be noted that the delegation of basic sanitation services to private agents is not risk-free.For example, Uruguay, who had started the sector's privatization in 1993, went back on its decision a decade later due to perceived deterioration of water quality and the high tariffs imposed on the population.These facts led to the enactment in 2004 of a law establishing for such services to be an exclusive assignment of public administration (BORRAZ et al, 2013).It is necessary, therefore to check whether the concessions and public-private partnerships (PPPs) have brought tangible benefits to society, and whether this activity has been conducted by private companies more efficiently.Thus, the justification of this work resides in the investigation of the existence of benefits when conceding public water supply and sanitation services to the private sector, considering the possibility of their better performance, including, the achievement of public policies objectives and of universalization.

THE EFFICIENCY OF PUBLIC AND PRIVATE WATER SUPPLY AND SANITATION SERVICE PROVIDERS
In international literature, the results of comparisons between the performance of public and private providers are not conclusive.Walter et al (2009) produced an extensive literature review to investigate the influence of organizational structure on business performance.Bhattacharyya et al (1995apud WALTER et al, 2009) studied 221 water supply companies in the US in 1992 and concluded that the public companies have lower costs, and tend to be more efficient than private providers.Etache andKouassi (2002 apud WALTER et al, 2009) and Kirkpatrick et al (2006WALTER et al, 2009) studied the performance of African water supply companies by means of the Cobb-Douglas production function, using similar data for both studies: as input variables the cost of labor, materials, energy, among others, and as an output variable the volume of water produced; the first study pointed to the greater efficiency of private companies, while the second showed that eventual differences are not significant.García-Sánchez (2006WALTER et al, 2009) conducted a data envelopment analysis of the performance of 24 service providers in Spain considering as predictors variables the staff, the treatment plants, the cost per kilometer; and as dependent variables the amount of water produced and the number of connections; as a result, we found no significant differences between the performance of private and public providers.Renzetti and Dupont (2003), in their study on the relationship between the origin of the capital and the performance of water supply and sanitation service providers in the US, United Kingdom and France, did not obtain strong evidence that private providers outperform the public providers.
Saal et al (2007), studied 10 providers of water supply and sanitation service in the UK between 1985 and 2000, and concluded that efficiency does not stem only from the origin of capital, but also from the regulatory regime.After the sector's privatization, efficiency decreased, due to the changes arising from the privatization itself combined with the less restrictive price limits policy (price cap).With the changes in this policy from 1994 onwards, these limits altered and most efficiency losses were recovered.Brocklehurst and Janssens (2004), when studying the contracts with the private sector in countries from Central and Western Africa, related the period length from the beginning of the PPP contracts to the amount of dedicated staff, per 1000 water connections, and found significant reduction in the workforce, in spite of the distortions caused by illegal connections and the difficulty of measuring the service.
In Brazil, Tupper and Resende (2004) conducted a study on efficiency and regulation, analyzing 20 state companies between 1996 and 2000, by means of data envelopment analysis (DEA).It was found that performance is a function of the system water losses and the density of water supplying services, given by the relation of the population served and the extension of the supply network.The conclusion is that parameters of this nature should be considered in the implementation of mechanisms to encourage competition.Sabbioni (2008), by examining the efficiency of water supply and sanitation service providers in Brazil between 2000 and 2004, concluded that regional state companies, have the lowest specific costs, which stems from the economy of scale.In addition, local private operators have specific costs similar to those of local public enterprises, while other local public providers have the highest specific costs.Ferro et al (2004) studied the cost structure of the water supply and sanitation service providers between 2003 and 2010 and comment that the administrative independence is not relevant to explain the cost structure but the origin of capital does, since private companies have a lower cost structure.Also according to the authors, independent operators have better cost management and professional training; by contrast, the dependents have subsidies, such as office space, information technology, administration, and others.

METHODOLOGY
The research objectives and the methodology used are summarized in Table 1.On topics 1, 2 and 3, the verification of the significance of the differences between the means of the variables grouped by type of service provider, we used the ANOVA test, whose use is subject to the normality of the variables (HAIR JUNIOR et al, 2009;p. 322); otherwise we shall use the Kruskal Wallis' test.The normality test we used the Kolmogorov-Smirnov.
The regression on topic 4 aims to verify the correlation between tariff and productivity indexes, according to the legal nature of the service provider.The dependent variable is the tariff, and the explanatory variables are the selected productivity indexes, individually.We did not consider more than one index per regression given the existing colinearity between the same.Dummy variables were created to identify public sector and mixed capital companies.
The model used is the static panel, indexed by year and by municipality.We used Hausman's test to choose between panels with fixed effects or random effects.Several municipalities do not have information for all years studied, setting an unbalanced panel (STOCK; WATSON, 2004, p.186).The major regression coefficients indicate that the tariff practiced by that provider finds a better explanation in its cost structure, which opens the possibility to share productivity gains with the population served.The existence of significance for the coefficients will be checked by Wald's test.In which: Tariffs refers to the tariff of water and sewage practiced in the city.
Vartest is the tested productivity index. (1) D2 is a dummy for the public sector, including the direct, local authorities and public companies.D3 is the dummy for mixed capital companies with public and private administration.
is the intercept 1 is the slope (declivity) coefficient of Vartest.
2 and 3 are the coefficients for D2 e D3, i.e., the additional intercept for these dummies.
2 and 3 are the coefficients of the interactions between Vartest and dummies D2 and D3 and represent the additional declivity on private providers.Along with 1, they will provide answer to the research question.
The interaction between a dummy variable and a continuous independent variable is used in order to check the declivity associated with the dummy (WOOLDRIDGE, 2012, p. 226; STOCK; WATSON, 2004, p.150).Thus, as we intend to measure the correlation between a given index of productivity and the practiced tariff for each sector (public, private, mixed capital), the coefficients of the interaction variables shall provide the desired result.
The analysis conducted only includes data on water supply and sanitary sewage systems.What does not fall under the scope of this study are the treatment of solid waste, cleaning and systems and rainwater harvesting, although they are included in the definition of sanitation in the Law 11.445/2007, which is the regulatory framework for the sector.
All information has been extracted from SNIS (National Sanitation Information System), relating the entrusted supplier of the service (Table 2).In some situations, the SNIS omits the identification of the private concessionaire and relates only the public entity conceding the services.

RESULTS ANALYSIS
The information from SNIS dates back to 1995 and depends on the effort and willingness of municipalities and service providers to feed its database.As shown in Table 2, the number of municipalities records in 2000 corresponds to a fraction of 1/6 of the 2010 quantitative, which shows the revolutionary construction of the database process.
Table 3 shows the income of service providers.Providers may have local, micro-regional and regional coverage.

ANALYSIS OF THE PERFORMANCE OF WATER SUPPLY AND SANITATION SERVICE PROVIDERS
Initially we will address the number one research question, consisting of a direct comparison between the contents of the SNIS, which reflect the performance of service providers from optics an economic, financial, administrative and operational viewpoint.For comparative purposes between the average rates, we shall only use information from 2010.
Table 5 shows which indexes were used, and for what purpose.By using as a reference, the classification where the least significant legal natures were excluded, we verified that private providers are located in the first or second position in 9 out of 14 possible situations, which demonstrates its differentiated performance in the evaluated aspects.Private providers have better financial performance, are more efficient in charging, have greater operational margin; at the same time, the concessions which they manage provide more comprehensive services in terms of sewage collection and treatment and are the BBR, Braz.Bus.Rev. (Engl.ed., Online), Vitória, v. 13, n. 1, Art. 6, p. 115 -136, jan.-feb. 2016 www.bbronline.com.brsecond regarding the supply and treatment of water.On the other hand they assume the third place in terms of tariffs practiced.Curiously, they assume only the third place in personnel productivity indexes, whether we are referring to productivity based on the number of active savings (IN019) based on the number of water and sewage connections (IN102).

WILLINGNESS TO INVEST
One of the reasons to concede services to the private sector resides in the willingness to make new investments, in view of its ability to raise funds.Thus, one aspect to be investigated is: "Does the private partner, in fact, invest more?".In terms of the amount invested, mixed capital companies are the ones who invest the most.These companies have large dimensions, with average operating revenues of R$ 700 million (Table 3), and manage basic sanitation in the big cities.This comparison should be put into perspective on the municipality's population.We consider only those situations where the investment made in the year differed from zero b) Kolmogorov-Smirnov's test presented Z = 22.67, or Sig = 0.00; therefore, it must reject normality.c) Kruskal Wallis test the variable average investment per capita, grouped by the variable legal nature of the provider rejects the hypothesis that the populations are equal, with significance at 0.0001.Source: SNIS.Developed by the authors.
Of the above, we can claim that private companies invest more, per inhabitant, than all other legal natures, except for mixed capital companies with private management.

SERVICE UNIVERSALIZATION
In the spirit of Law 11.445 / 2007 to ensure the population access to the services of sanitation, it is justified to assess whether private service providers have shown higher performance in relation to public providers in achieving this goal.
The analysis of the access to services of water supply and sewage collection considered  10 presents the means of the presented indexes.The water coverage index stood at around 83% in 2005, which means there is little room for growth in access to these services.The index for sewage collection indicates values close to 60% in 2005, which is a long way to the achievement of universal service.With regard to the sewage treatment deficiency is even greater, where the corresponding index reaches 23% on average of the municipalities.
BBR, Braz.Bus.Rev. (Engl.ed., Online), Vitória, v. 13, n. 1, Art. 6, p. 115 -136, jan.-feb. 2016 www.bbronline.com.brResults are presented in Table 11.and IN055, respectively.Therefore, in all cases we must reject normality.c) Kruskal Wallis' test for the variables IN015 and IN046, grouped by the legal nature of the provider, we reject the hypothesis that the means are equal, with significances at 0.013 and 0.000, respectively.The same test supports the hypothesis that equality of the means for the variable IN055 with significance of 0.634.Source: SNIS.Developed by the authors.
With regards to water supply, there is no significance for the difference between the growth of the indexes of the different types of providers, which is justified by the occurring high level of service in 2005.However, private service providers have provided the largest  2013), have adopted in the contracts a productivity factor, also called Factor X, whose objective is to share these gains with the population served; thus, this factor is a reducer of the tariff practiced.
If the relationship between tariffs in a given index of productivity is inelastic, the possibility of transferring the benefits to the users of the services is limited and there would be other explanations for the level of tariffs practiced.Take for example, the situation in which the water supply and sanitation services are fulfilled by the direct public administration and that part of the collection with the services is used to fund other public expenditures; thus, other factors that explain the practiced tariff level other than the costs incurred by the water supply and sanitation service.
It should be noted that some rates are directly proportional to the tariff, as those indexes Static panels were processed with fixed and random effect.Table 13 presents the coefficients of the static panel with fixed effects among the productivity or expenses indexes (explanatory variables), by the practiced water and sewage tariffs (IN004), the dependent variable.The dummy D2 was introduced, representative of the direct public administration, regional authorities and public companies; and dummy D3, relative to the companies with mixed capital with public and private management.As the table's objective is to verify the correlation between a given index of productivity and the value of the tariff, the information of interest in these regressions are the coefficients of the interaction variables 2 and 3, and not the constants 2 and 3.Prob> 2 0,3333 -0,1551 -0,0629 1,1619*** -0,1030 -0,1236 (0,585) (0,802) (0,942) (0,000) (0,749) (0,800) 2=3 3 0,493 0,000 0,004 0,219 0,000 0,166 0,474 0,000 0,355 0,300 0,000 0,669 0,044 0,060 0,486 0,012 0,545 0,883 Based on the results from Table 13 we can establish a classification whose rationale is presented in the following examples: Regression 4: IN0026 refers to operational expenditure and it should be directly proportional to the tariff charged (the greater the expenditure, the greater the tariff to be practiced).The coefficients 1, 2 and 3 are significant, and their values are equal to 1.1457, -0.2046 and -0.2613, respectively.We conclude that 1 is greater than 1+2, and that 1 is greater than 1+3.Statistically, 2 and 3 are equal, for Wald's test result was 0.3552 for equality between them.We conclude that the private sector has better conditions to reduce tariffs when operational costs are reduced; the public sector though is tied with companies of mixed capital.
Regression 9: IN004, tariff is inversely proportional to IN017, one of the indexes related to productivity or scale (the greater productivity enables the reduction of tariff).The lower the declivity is the greater the possibility to transfer part of this productivity to the user of the service.By construction, 1 reflects the declivity of the private sector and the declivity of the other sectors should be added to this.In Table 13 we read that 1 = 0.1002and 2 = 0.0757, both with significance, and 3 is not significant.Thus, the declivity of the public sector amounts to 1 + 2 = 0.0245, a figure higher than the declivity 0.1002 of the private sector.
As 3 is statistically equal to zero, 1 + 3 is statistically equal to 1 and it is assumed that the declivity of companies is equal to the private sector.Needless to say, that the Wald's test rejects the equality of 2 and 3.Thus, we can conclude that, by classifying the sectors according to their ability to transfer greater reduction to the tariff in the hypothesis of economy of scales that index IN017 represents, so we have: first, in a tie comes the private sector and companies of mixed capital; and last comes the public sector.
3 indexes, namely:  IN015 -Sewage collection index  IN046 -Treated sewage index referring to the water consumed  IN055 -Total water service index Table Growth in access to services will be measured by comparing the 2005 figures with those measured for 2010.Thus, three variables were created:  IN015: difference of the index of sewage collection between 2005 and 2010;  IN046: difference of the index of sewage treatment referring to the water consumed between 2005 and 2010;  IN055, difference of the index of the total water service between 2005 and 2010.
of expenses and losses.In this category are the indexes IN003, IN008, IN026, IN027, IN029 and IN060.Other indexes are inversely proportional to the tariff though, as the productivity and scale indexes.The indexes IN017, IN019, IN048 and IN102 fall into this category.The following tables bring these indexes.

Table 4 -Coverage of the Service Providers -2010
Table 4 provides the profile of the providers on coverage and the percentage of direct operational revenues.In general, mixed capital companies have regional coverage.Other legal natures have eminently local character.

Table 6
presents the descriptive statistics of the listed indexes, as wel as Kolmogorov-Smirnov's test, which rejected normality for all variables.Thus, the significance of differences between means may not be based on ANOVA but in the Kruskal Wallis test.

Table 7
presents the average scores obtained in accordance with the legal nature of the service providers, Kruskal Wallis test and the classification of private companies in relation to the analyzed content.

Table 8
presents information available on investments made in the last five years of the past decade (including from 2006 to 2010) distinguishing on the source of funds (service provider, municipality or state government).Note that when the service provider is a private company, the invested public resources are virtually non-existent.

Table 9 -Average Investment per Capita Made by the Service Provider
Table 9 makes this adjustment, with the per capita investment made by the service provider.By this criterion private companies rank the second place among those who invest the most, second only to mixed capital companies with private management.A Comparative Analysis of the Public and Private Water Supply and Sanitation Service Providers' Work in Brazil 127 BBR, Braz.Bus.Rev.(Engl.ed., Online),Vitória, v. 13, n. 1, Art. 6, p. 115 -136, jan.-feb.2016  www.bbronline.com.br

Table 11 -Average evolution in access to services, by legal nature of the provider
A Comparative Analysis of thePublic and Private Water Supply  and Sanitation Service Providers' Work in Brazil  129growth for the indexes related to the collection and treatment of sewage.It is evident, therefore, that public officials should consider the possibility of delegating to the private partner the provision of collection and treatment of sewage services, in order to achieve the goal of universalization of services that the Law 11.445/2007 deals with.4.4 THE ABILITY TO COMMUNICATE PRODUCTIVITY GAINSThe bidding for hiring a private agent aimed at the provision of public services is a competitive process, causing this agent to reduce their prices until the desired returns limit and pushing them to the constant pursuit of productivity and cost reduction.Public administration has no such motivation, for it prioritizes the bureaucracy, hierarchy and the administrative processes.It is expected that the tariffs practiced by private providers are more correlated to the cost structure than those practiced by public providers.Productivity gains from concessions open the possibility of reducing public tariffs.Based on this, some regulatory agencies, among which ARSESP (Regulatory Agency for Sanitation and Energy of the State of São Paulo State) ( BBR, Braz.Bus.Rev.(Engl.ed., Online),Vitória, v. 13, n. 1, Art. 6, p. 115 -136, jan.-feb.2016  www.bbronline.com.br

Table 13 -Regression Coefficients Between the Productivity Indexes and the Practiced Tariff
A Comparative Analysis of the Public and Private Water Supply and Sanitation Service Providers' Work in Brazil 131