Corruption in public health and its effects on the economic growth of Brazilian municipalities

Abstract

This study’s objective is to estimate the effects of corruption in the public health sector on the economic growth of Brazilian municipalities. To build three corruption measures, data from audits conducted by the office of the comptroller general (Controladoria Geral da Uniao, CGU henceforth) in 2009 and 2010 in the health and sanitation sectors were used. Two analysis steps were performed. The first verified the relationship between the performance of the audit and the economic growth rate of the municipalities, using the Ordinary Least Squares (OLS); the second analyses the effects of corruption on public health on the economic growth of the audited municipalities, using OLS and Quantile Regressions. First, in a sample of 5547 municipalities, the evidence indicates that being audited is related to slower economic growth. From this, when the sample is restricted to the 180 municipalities audited in 2009 and the corruption variables constructed from the audit reports conducted in the year, the results indicate negative effects of corruption on economic growth. The results show that in the larger quantiles of economic growth, the adverse effects of corruption are felt more significantly. Both methods tested with the three corruption variables created provide similar evidence, showing robustness of results. Therefore, the study allowed us to conclude that corruption in the public health sector hampered the economic growth of Brazilian municipalities, which is a delayed effect: Corruption in 2009 had negative effects on growth in 2011.

Appendix

Economic growth and corruption

This section presents a version of the neoclassical economic growth model applied to the public health sector. Building on neoclassical growth models such as those of Barro [4] and Robert [32], Pereira and da Cruz Vieira [29] develop a model that considers the effects of bureaucrat corruption on economic growth.

The economy is made up of agents, firms, and the public sector. In this case, non-bureaucratic agents, m, are the society, and bureaucratic agents, n, are politicians and policy makers. Firms form the private sector of the economy, consisting of engineering offices, laboratories, and other entities. Agents are optimizers, so they seek to maximize their usefulness over time:

Agents are optimizers, so they seek to maximize their usefulness over time:

$${ \hbox{max} }\mathop \smallint \limits_{0}^{\infty } u\left[ {c\left( t \right)} \right]e^{ - \delta t} dt.$$

It is assumed that a fraction of these agents, more specifically bureaucrats, are corrupt; that is, they divert public resources to benefit their private consumption. Each corrupt bureaucrat has access to a ${\raise0.7ex\hbox{${\tau y}$} \!\mathord{\left/ {\vphantom {{\tau y} n}}\right.\kern-0pt} \!\lower0.7ex\hbox{$n$}}$ amount of public resources that can be diverted for his own benefit. He or she will choose to be corrupt if:

$$\varPhi \left( {w - \sigma + \frac{\tau y}{n}} \right) \ge w - \sigma ,$$

where Φ is the probability of the agent not being corrupt. By acting corruptly, the agent longs for greater utility. That is, given that utility is a function of consumption, the bureaucrat chooses to divert public resources and act corruptly knowing that the income from this behavior is greater than the income from acting honestly, since the term ${\raise0.7ex\hbox{${\tau y}$} \!\mathord{\left/ {\vphantom {{\tau y} n}}\right.\kern-0pt} \!\lower0.7ex\hbox{$n$}}$ is predominantly positive.

It is also assumed that such diverted resources were intended exclusively for the provision of public goods and services—in this case, of the public health sector. Therefore, the corruption of this dishonest fraction of bureaucrats diminishes the public provision of such services, harming society (non-bureaucratic agents). In addition, in the presence of corruption, the rate of capital and asset accumulation is lower. Thus, the marginal productivity of capital is reduced, reducing the level of investments as well as the level of consumption of agents. As a result, there is a reduction in consumption growth rates, capital accumulation, and economic output growth. Therefore, in this study’s case, the presence of corruption supposedly causes a reduction in the public supply of goods and services related to the public health sector. Thus, the economy is affected in many ways, with the reduction of the economic growth rate the final effect.

See Figs. 6 and 7.

See Tables 9–15.

Fig. 6

Total municipalities audited by state—2009 and 2010. Source CGU data, worked by the authors

Fig. 7

Effects of corruption occurrences (2009) and irregular fraction of the amount (2009) on the municipal. GDP growth rates quantiles (2011)—models with controls. Source Own elaboration

Table 9 Definition and sources of the variables used

Table 10 Correlation matrices between corruption indices—2009 and 2010

Table 11 Descriptive statistics of municipal GDP growth rates, 2009–2011, Municipalities audited in 2009

Table 12 Descriptive statistics of control variables—Municipalities audited in 2009

Table 13 Relationship between corruption indices (2009) and municipal GDP growth rate (2008)

Table 14 Effects of corruption (2009) on the quantiles of municipal GDP growth rate (2011)

Table 15 Wald test results—Quantile regressions