Skip to main content

Advertisement

SpringerLink
Log in

The role of externalities in fiscal efficiency

  • Published:
Empirical Economics Aims and scope Submit manuscript

Abstract

Fiscal efficiency is determined not only by the municipality’s local characteristics but also by externalities from the neighborhood. In this paper, we study municipalities’ fiscal efficiency and assess the importance of such externalities in shaping local fiscal efficiency. The quantification of externalities is challenging because municipalities interact in complex ways. We build a municipality-specific indicator of externalities that considers the effect of the geographic and social neighborhoods. Externalities tend to increase when municipalities are close and have more significant differences in social development. To analyze the role of externalities in public spending efficiency, we propose a novel way to compare efficiency models by ensuring similarity in their structural properties (inputs, outputs, and orientation). In this situation, when one does not consider the network effect that a specific municipality receives, one could be misevaluating the municipality’s actual efficiency level. We apply our model to Brazilian municipalities and find that socio-economic externalities play an essential role in local public spending efficiency. Many municipalities are deemed efficient when they are not if we do not explicitly consider the effect of externalities on the municipality’s local performance. Our findings highlight the importance of socio-economic cooperation and integration among municipalities as a tool to achieve better fiscal performance.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8
Fig. 9

Similar content being viewed by others

Notes

  1. For example, economic dependence can arise when an agricultural firm sells primary goods to an industrial firm or individual located in another municipality. This dependence paves the way for spillover effects. In our example, if a natural disaster strikes the agricultural firm’s location, the adverse effects could propagate to the industrial firm or individual, even if they are very distant from the natural disaster. Usually, the dependence of municipality i on municipality j tends to be proportional to the volume and quantity of economic transactions between their economic agents.

  2. For instance, Alañon-Pardo et al. (2018) find that the features of neighboring municipalities help explain the industrial location decisions in Spain.

  3. Acemoglu et al. (2015) argument that the reasons for the strategic choice include the porosity of municipalities, voters’ demand for public goods existing in the neighboring municipalities, lack of local capacity to increase state’s capacity due to problems such as criminality or diseases contagion, and judicial system inefficiency.

  4. In our methodology, the density of the network of intermunicipality interactions is defined as a function of the distance of each pair of municipalities A and B: \(\exp (-\frac{d_{AB}}{k})\). Higher values of k imply a slower decay, generating more extensive neighborhoods and a denser network. If we choose a small k, we effectively consider small regions only (sparse network). In this case, our work would approach Acemoglu et al. (2015)’s methodology, who considers that externality arises only from (physically) bordering municipalities. Our approach is also flexible in considering other approaches, such as one analysis similar to Santolini (2020), who considers the interaction between municipalities with a fixed distance of up to 20 km, i.e., not necessarily bordering municipalities. However, considering Brazil’s continental size and the municipalities’ sizes heterogeneity, we believe that considering externalities arising only from bordering municipalities is inadequate. For example, in the state of São Paulo (Southeast region), there are medium-sized municipalities with solid interaction with small municipalities around them, even if they are physically bordering each other. On the other hand, in the state of Amazonas (North region), municipalities are large, and transport is often more viable by plane, imposing the interaction with more distant municipalities. Considering the Brazilian reality, we used different values k to analyze the robustness of our results.

  5. To evaluate the proximity of two municipalities in our case study using Brazilian data (details will follow in the paper), we consider the median latitude and longitude of all municipality postal codes. Then, we evaluate the distance between two municipalities as the cosine distance between their respective median latitudes and longitudes. This distance metric is the most widely used in the georeferencing literature, given the globe’s spherical shape.

  6. In our empirical application using Brazilian data, we proxy the social development of each municipality by the FIRJAN Municipal Development Index (IFDM), which measures human development in these three dimensions.

  7. Sousa and Stoši (2005) estimate efficiency for Brazilian municipalities and find that efficiency correlates with the municipality’s size.

  8. The size distribution of municipalities in Brazil follows roughly a power law, with very few large municipalities—mostly capitals—and many small municipalities. To understand how significant the differences are, São Paulo is the largest municipality in Brazil (and one of the world’s ten largest municipalities) with more than 12 million people. Serra da Saudade, in contrast, has a population of fewer than 800 persons. The population of both municipalities is estimated for 2019 and is available in https://cidades.ibge.gov.br/brasil/sp/sao-paulo/panorama and https://cidades.ibge.gov.br/brasil/brasil/mg/serra-da-saudade/panorama, respectively.

  9. We also re-run our experiments to analyze whether our findings are robust to an output-oriented approach. Comparing the input- and output-oriented perspectives, the Kendall pairwise correlation for the efficiency models without externalities is 1.00 and with externalities is 0.66.

  10. The empirical literature shows evidence of the relationship between fiscal policy, efficiency, and economic growth (Arin et al. 2019; Bergman et al. 2016; Muinelo-Gallo and Roca-Sagalés 2013; Quigley 1998; Zhang and Zou 1998).

  11. Smith (1992) argues that, when the outputs employed in efficiency models are well-publicized performance indicators by the government (such as development indices on employment, health, and education), the relationship between performance indicators (outputs) and public expenditures (inputs) is more likely to be surveilled by the society. Therefore, governors may use these performance indicators more systematically.

  12. The authors use a Monte Carlo simulation to show that inefficient units using low levels of the endogenous resource may be set tougher efficiency targets than equally inefficient units using more of the resource. Therefore, DEA estimates would become severely unfair as the sample size decreases.

  13. For instance, Conley and Ligon (2002) discuss the importance of geographical distance in the context of cross-country spillovers, especially for neighboring countries. Also, Orlando (2004) and Adams and Jaffe (1996) study the importance of geographical proximity as an environment that fosters human capital spillovers in the context of R&D research by facilitating learning and physical meetings. Disdier and Head (2008) surveyed the literature on distance effects on bilateral trade. They find that distance negatively affects trade and that this result remained persistently high since the middle of the century. Geographical distance also affects firm-to-firm relationships. Su and Moaniba (2020) find that inventor distance has a positive correlation with R&D intensity at a firm level. Cross-border alliances are more probably to occur between firms from shorter geographical distances (Jha et al. 2019). Opie et al. (2019) find that the efficiency of Chinese state-owned enterprises is negatively related to corporate control’s geographical distance. Even for international finance, geographical distance plays a vital role because of informational asymmetries and cultural differences (Brei and von Peter 2018). In sum, the geographical distance envelops other features that can affect municipal social and economic relationships.

  14. For instance, Newman (2003) shows that the network structure can unveil essential functions of different economic, biological, and physical networked systems. Silva et al. (2021b) use a unique business management dataset that contains firm-to-firm controls to construct a bilateral corporate control network and estimate spillovers in the labor market during the global financial crisis. Silva and Zhao (2016) survey complex network measures and different construction approaches.

  15. In our application using data from Brazilian municipalities, we evaluate the similarities between pairs of municipalities using \( k = 100 \). Robustness tests with other values show that the results do not change qualitatively.

  16. Conurbation is an urban phenomenon that occurs when two bordering municipalities expand to meet each other, forming a single urban core.

  17. Conley and Ligon (2002) use a similar idea by employing transportation costs of human factors (mobility) in the context of cross-country spillovers to account for this social distance. Other types of social spillovers are studied in Dinkelman and Schulhofer-Wohl (2015), Burzynski et al. (2020), and Combes et al. (2020).

  18. To exemplify how the socio-economic externalities enter as inputs, suppose three municipalities with the following positive (first term inside parenthesis) and negative externalities (second): A (1, 0.4), B (0.5, 2), C (0, 0). The average positive and negative externalities in the sample are 0.5 and 0.8. We replicate the average positive and negative externalities across municipalities A, B, and C when evaluating DEA without externalities. In contrast, we use deviations from the sample average to evaluate the positive externalities (\(A = 1 - 0.5 = 0.5\); \(B = 0.5 - 0.5 = 0\); \(C = 0 - 0.5 = -\,0.5\)) and negative externalities (\(A = 0.4 - 0.8 = -\,0.4\), \(B = 2 - 0.8 = 1.2\), \(C = 0 - 0.8 = -\,0.8\)) when evaluating DEA with externalities. Recall that, before evaluating the DEA, we shift every distribution to the right to ensure non-negativity of the inputs and outputs (see Sect. 3.2 for further details).

  19. Section 4.2 provides an example of a municipality with a significant discrepancy in efficiency when we compare the model with and without network effects.

  20. Simar and Wilson (2007) criticize semi-parametric two-stage approaches that combine DEA efficiency estimates followed by a regression analysis that uses DEA estimated efficiency as dependent variables. The second stage is typically a censored (like Tobit) regression to account for the bounded nature of DEA efficiency scores or just a simple OLS estimator. Such naïve two-stage estimators have two main shortcomings. First, there is no clear theory of the underlying data generating process that would justify the two-stage approach. Second, treating the DEA efficiency estimates as independent observations is not appropriate due to serial correlation.

  21. In the same vein, Banker and Natarajan (2008) also provide meaningful reasoning and conditions under which second-stage OLS provides consistent estimates. Simar and Wilson (2011) compares Simar and Wilson (2007) and Banker and Natarajan (2008) methodologies and conclude that bootstrap methods provide the only feasible means for inference in the second stage. In this way, we opt to use Simar and Wilson (2007)’s methodology in this paper.

  22. The use of truncated regression explains that the substantial share of fully efficient DMUs typically found in DEA is an artifact of the finite sample bias inherent in DEA, unrelated to the underlying data generating process.

  23. Another test for efficiency robustness is offered in Destefanis and Sena (2005). In turn, Daraio and Simar (2007) propose a robust alternative version of DEA conditional estimator using external-environmental variables which are neither inputs nor outputs under the control of the producer.

  24. We acknowledge that the socio-economic externality may not be entirely exogenous, as neighboring municipalities’ conditions may affect a municipality. However, compared to the municipality’s local fiscal conditions, the socio-economic externality is a second-order effect. Therefore, endogeneity concerns are somewhat alleviated.

  25. We should not interpret the coefficient estimates in this section as causal. Instead, they are partial correlations that give valuable insights into which municipalities are most affected by socio-economic externalities.

  26. For robustness, we also use a less saturated model with time fixed effects. This fixed effect captures nationwide macroeconomic idiosyncrasies.

  27. IFDM data are available on the FIRJAN System website: https://www.firjan.com.br/ifdm/.

  28. Another way of understanding the role of externalities on public spending efficiency is to interpret two different municipalities delivering the same level of public services with the same amount of public expenses. In this case, the most efficient municipality is the one performing this transformation with the fewer help from socio-economic externalities, i.e., less positive (normal input) or more negative (reverse input) externalities.

  29. This is the case for the deviations from the mean and the sample mean of the negative externality (reverse input) and positive externality (normal input), and the per capita GDP growth (normal output).

  30. The detailed IFDM methodology can be found on the FIRJAN System website https://www.firjan.com.br/ifdm/.

  31. For robustness, we also re-run our public efficiency DEA model without the debt cost as input. The pairwise Kendall correlations between DEA estimates without and with externalities are 0.94 and 1.00, respectively, when comparing such model with our baseline. In addition, the average efficiency in the DEA without and with externalities when we exclude the debt cost is 0.77 (compared to 0.78 with our baseline) and 0.63 (0.63), respectively. Therefore, the results remain very similar.

  32. We also run a further robustness test in which we include all municipal public spending besides personnel expenditures, debt costs, and investment. We call this the residual input, given by the residual part of the total public spending as a share of the total public spending. Then, we re-run our public efficiency DEA model with the residual input. The pairwise Kendall correlations between DEA estimates without and with externalities are 0.88 and 1.00, respectively, when comparing such model with our baseline. Also, the average efficiency in the DEA without and with externalities when we include the residual input is 0.80 (compared to 0.78 with our baseline) and 0.63 (0.63), respectively. Therefore, the results remain very similar. We use the model without the residual input to maintain a parsimonious specification as our baseline.

  33. The Fiscal Responsibility Law came into force in 2000 to prevent high and unsustainable public debt levels faced by local, state, and federal governments. The FRL imposed limits on personnel expenditure and public debt and established rules to enforce responsible fiscal management and transparency to society.

  34. According to Brazilian legislation, net current revenue measures the effective municipal capacity of collecting taxes. They are used in the FRL to define the personnel expenditure and total public debt indexes. Net real revenues are employed to determine the limit of municipal debt service. Based on these rules, we use current net revenue to build the personal expenditure and public investment indexes and the net real revenue to calculate the debt cost index.

  35. For example, let us compare two municipalities in the metropolitan region of São Paulo: Francisco Morato and the city of São Paulo itself. IBGE data available at https://cidades.ibge.gov.br/ show that both cities have similar health and education indicators (approximately 96% of schooling rate from 6 to 14 years of age and 12 deaths per thousand live births). However, they differ significantly in other social indicators: sewage (58% and 93%), employed population (6% and 47%), and income (2.2 and 4.1 minimum wages), respectively for Francisco Morato and São Paulo. In addition, Francisco Morato generates 20% of its own revenues and São Paulo, 70%. The DEA efficiencies without externalities of Francisco Morato and São Paulo are, respectively, 0.83 and 0.98, while the DEA efficiencies with externalities are 0.56 and 0.96. These results show that Francisco Morato and São Paulo receive, respectively, net benefits of the order of 48% (\(0.83/0.56 - 1\)) and 2% (\(0.98/0.96 -1\)) arising from spillovers from neighboring municipalities. The vast majority of municipalities receive benefits through network effects. There are also municipalities with zero network effect, such as São Caetano do Sul-SP (DEA without and with externalities equal to 1) and Barcelos-AM (DEA without and with externalities equal to 0.70). There also municipalities with a small negative network effect: DEA without and with externalities of 0.996 and 1 for Louveira-SP and of 0.65 and 0.75 for Eirunepé-AM.

  36. In numerical terms, the average DEA from 2006 to 2014 (input-oriented) in the five Brazilian regions are: North (DEA without externalities: 0.71, DEA with externalities: 0.56, Differential DEA: 0.15), Northeast (0.71, 0.55, 0.16), Midwest (0.83, 0.68, 0.15), Southeast (0.86, 0.70, 0.16), and South (0.87, 0.70, 0.16).

  37. The coefficients in this section should not be interpreted as causal estimates. Instead, they are partial correlations that give valuable insights into which types of municipalities are most affected by socio-economic externalities and how they affect DEA efficiency levels.

  38. Since more economically developed municipalities have more resources to invest in infrastructure and public services (Bierbrauer and Winkelmann 2020), it is reasonable to assume that economic development can be related to externality and efficiency. For example, a municipality can invest in roadways, and its neighbors can benefit from it. Also, it can attract companies because of better infrastructure and generate jobs not only for its citizens but also for those in other municipalities. From the perspective of negative spillovers, the municipality can invest in more security, causing crime to increase in neighboring municipalities (Dell 2015).

  39. The vast majority of Brazilian municipalities are small and have a tiny share of their revenues originating from local taxation. These municipalities depend on federal and state transfers of resources, such as the Fundo de Participação dos Municípios—FPM quota. In many of these municipalities, the dependence on financial resources (FPM quota/GDP) is high because local managers do not institute local taxes within their constitutional competence to prevent wearing the local government out politically. Therefore, they also have poor fiscal management. Thus, the variables FPM quota/GDP and Taxes/GDP can be correlated with externalities and the differential DEA.

  40. Mesoregions encompass neighboring municipalities whose purpose is to integrate the organization, planning and execution of public functions of common interest. For instance, Brazil has state-level complementary laws defining mesoregions.

  41. In theoretical terms, we could propose a more granular division within each of these three economic sectors. However, in practical terms, the classification of a firm’s activity becomes subject to misclassification as we drill down into the hierarchical tree of economic sectors. Therefore, we opt to use a more aggregate classification that is prone to a lower level of misclassification.

  42. The maximum value of per capita GDP is R$ 815,094, which is equivalent to US$ 311,105, considering the December 2014 USD/BRL exchange rate (2.62).

  43. We identify outliers by inspecting observations that deviate more than 1.5 times the interquartile range (similar to identifying outliers in boxplots).

References

  • Acemoglu D, García-Jimeno C, Robinson JA (2015) State capacity and economic development: a network approach. Am Econ Rev 105:2364–2409

    Article  Google Scholar 

  • Adams J, Jaffe A (1996) Bounding the effects of R&D: an investigation using matched establishment-firm data. RAND J Econ 27(4):700–721

    Article  Google Scholar 

  • Afonso A, Schuknecht L, Tanzi V (2006) Public sector efficiency: evidence for new EU member states and emerging markets. European Central Bank Working Paper Series 581

  • Alañon-Pardo A, Walsh PJ, Myro R (2018) Do neighboring municipalities matter in industrial location decisions? Empirical evidence from Spain. Empir Econ 55:1145–1179

    Article  Google Scholar 

  • Arin KP, Braunfels E, Doppelhofer G (2019) Revisiting the growth effects of fiscal policy: a Bayesian model averaging approach. J Macroecon 62:103158

    Article  Google Scholar 

  • Baicker K (2005) The spillover effects of state spending. J Public Econ 89(2–3):529–544

    Article  Google Scholar 

  • Balaguer-Coll MT, Brun-Martos MI, Márquez-Ramos L, Prior D (2019) Local government efficiency: determinants and spatial interdependence. Appl Econ 51(14):1478–1494

    Article  Google Scholar 

  • Banker RD, Natarajan R (2008) Evaluating contextual variables affecting productivity using data envelopment analysis. Oper Res 56(1):48–58

    Article  Google Scholar 

  • Bergman UM, Hutchison MM, Jensen SEH (2016) Promoting sustainable public finances in the European Union: the role of fiscal rules and government efficiency. Eur J Polit Econ 44:1–19

    Article  Google Scholar 

  • Bertrand M, Mullainathan S (2003) Enjoying the quiet life? Corporate governance and managerial preferences. J Polit Econ 111(5):1043–1075

    Article  Google Scholar 

  • Bierbrauer F, Winkelmann J (2020) All or nothing: state capacity and optimal public goods provision. J Econ Theory 185:104955

    Article  Google Scholar 

  • Brei M, von Peter G (2018) The distance effect in banking and trade. J Int Money Finance 81:116–137

    Article  Google Scholar 

  • Brueckner J (2003) Strategic interaction among governments: an overview of empirical studies. Int Reg Sci Rev 26(2):175–188

    Article  Google Scholar 

  • Brunelle C (2013) The growing economic specialization of cities: disentangling industrial and functional dimensions in the Canadian urban system, 1971–2006. Growth Change 44(3):443–473

    Article  Google Scholar 

  • Burzynski M, Deuster C, Docquier F (2020) Geography of skills and global inequality. J Dev Econ 142:102333

    Article  Google Scholar 

  • Combes PP, Démurger S, Li S, Wang J (2020) Unequal migration and urbanisation gains in china. J Dev Econ 142:102328

    Article  Google Scholar 

  • Conley TG, Ligon E (2002) Economic distance and cross-country spillovers. J Econ Growth 7(2):157–187

    Article  Google Scholar 

  • Costa-Font J, Moscone F (2008) The impact of decentralization and inter-territorial interactions on Spanish health expenditure. Empir Econ 34:167–184

    Article  Google Scholar 

  • Dahlberg M, Edmark K (2008) Is there a “race-to-the-bottom” in the setting of welfare benefit levels? Evidence from a policy intervention. J Public Econ 92(5):1193–1209

  • Daraio C, Simar L (2007) Conditional nonparametric frontier models for convex and nonconvex technologies: a unifying approach. J Prod Anal 28:13–32

    Article  Google Scholar 

  • Dell M (2015) Trafficking networks and the Mexican drug war. Am Econ Rev 105(6):1738–1779

    Article  Google Scholar 

  • Destefanis S, Sena V (2005) Public capital and total factor productivity: new evidence from the Italian regions, 1970–98. Reg Stud 39(5):603–617

    Article  Google Scholar 

  • Destefanis S, Sena V (2009) Public capital, productivity and trade balances: some evidence for the Italian regions. Empir Econ 37:533–554

    Article  Google Scholar 

  • Dinkelman T, Schulhofer-Wohl S (2015) Migration, congestion externalities, and the evaluation of spatial investments. J Dev Econ 114:189–202

    Article  Google Scholar 

  • Disdier AC, Head K (2008) The puzzling persistence of the distance effect on bilateral trade. Rev Econ Stat 90(1):37–48

    Article  Google Scholar 

  • Fisher RC (2015) State and local public finance. Routledge, New York

    Google Scholar 

  • Geys B (2006) Looking across borders: a test of spatial policy interdependence using local government efficiency ratings. J Urban Econ 60(3):443–462

    Article  Google Scholar 

  • Jha A, Kim YJ, Gutierrez-Wirsching S (2019) Formation of cross-border corporate strategic alliances: the roles of trust and cultural, institutional, and geographical distances. J Behav Exp Finance 21:22–38

    Article  Google Scholar 

  • Johnson A, Kuosmanen T (2012) One-stage and two-stage DEA estimation of the effects of contextual variables. Eur J Oper Res 220:559–570

    Article  Google Scholar 

  • Kerstens K, Van de Woestyne I (2011) Negative data in DEA: a simple proportional distance function approach. J Oper Res Soc 62(7):1413–1419

    Article  Google Scholar 

  • Muinelo-Gallo L, Roca-Sagalés O (2013) Joint determinants of fiscal policy, income inequality and economic growth. Econ Model 30:814–824

    Article  Google Scholar 

  • Narbón Perpiñá I, De Witte K (2018) Local governments’ efficiency: a systematic literature review-part I. Int Trans Oper Res 25:431–468

  • Newman MEJ (2003) The structure and function of complex networks. SIAM Rev 45(2):167–256

    Article  Google Scholar 

  • Opie W, Tian GG, Zhang HF (2019) Corporate pyramids, geographical distance, and investment efficiency of Chinese state-owned enterprises. J Bank Finance 99:95–120

    Article  Google Scholar 

  • Orlando MJ (2004) Measuring spillovers from industrial R&D: on the importance of geographic and technological proximity. RAND J Econ 35(4):777–786

    Article  Google Scholar 

  • Orme C, Smith P (1996) The potential for endogeneity bias in data envelopment analysis. J Oper Res Soc 47:73–83

    Article  Google Scholar 

  • Panayotou T (2016) 14. Economic growth and the environment. In: Environment in Anthropology (Second Edit: A Reader in Ecology, Culture, and Sustainable Living, edited by Nora Haenn, Allison Harnish and Richard Wilk, New York University Press, New York, USA. pp 140-148). https://doi.org/10.18574/9781479862689-016

  • Portela MS, Thanassoulis E, Simpson G (2004) Negative data in DEA: a directional distance approach applied to bank branches. J Oper Res Soc 55(10):1111–1121

    Article  Google Scholar 

  • Quigley JM (1998) Urban diversity and economic growth. J Econ Perspect 12(2):127–138

    Article  Google Scholar 

  • Revelli F (2005) On spatial public finance empirics. Int Tax Public Finance 12(4):475–492

    Article  Google Scholar 

  • Revelli F, Tovmo P (2007) Revealed yardstick competition: local government efficiency patterns in Norway. J Urban Econ 62(1):121–134

    Article  Google Scholar 

  • Santolini R (2020) Spatial effects on local government efficiency. Pap Reg Sci 1:183–200

    Article  Google Scholar 

  • Silva TC, Zhao L (2016) Machine learning in complex networks, vol 1. Springer, Berlin

    Book  Google Scholar 

  • Silva TC, Hasan I, Tabak B (2021a) Financing choice and local economic growth: evidence from Brazil. J Econ Growth 26:329–357

  • Silva TC, Muniz FJ, Tabak BM (2021b) Indirect and direct effects of the subprime crisis on the real sector: labor market migration. Empir Econ. https://doi.org/10.1007/s00181-021-02051-1

  • Simar L, Wilson P (2002) Non-parametric tests of returns to scale. Eur J Oper Res 139(1):115–132

    Article  Google Scholar 

  • Simar L, Wilson PW (2007) Estimation and inference in two-stage, semi-parametric models of production processes. J Econom 136(1):31–64

    Article  Google Scholar 

  • Simar L, Wilson PW (2011) Two-stage DEA: caveat emptor. J Prod Anal 36(2):205–218

    Article  Google Scholar 

  • Smith P (1992) Negative political feedback: an examination of the problem of modelling political responses in public sector effectiveness auditing. Account Audi Account J 5(1):5–20

    Google Scholar 

  • Solé-Ollé A (2006) Expenditure spillovers and fiscal interactions: empirical evidence from local governments in Spain. J Urban Econ 59(1):32–53

    Article  Google Scholar 

  • Sousa M, Stošić B (2005) Technical efficiency of the Brazilian municipalities: correcting nonparametric frontier measurements for outliers. J Prod Anal 24(2):157–181

    Article  Google Scholar 

  • Stiglitz JE, Rosengard JK (2015) Economics of the public sector: fourth international student edition. WW Norton & Company, New York

    Google Scholar 

  • Su HN, Moaniba IM (2020) Does geographic distance to partners affect firm R&D spending? The moderating roles of individuals, firms, and countries. J Bus Res 106:12–23

    Article  Google Scholar 

  • Zhang T, Zou HF (1998) Fiscal decentralization, public spending, and economic growth in China. J Public Econ 67(2):221–240

    Article  Google Scholar 

Download references

Author information

Authors and Affiliations

  1. Universidade Católica de Brasília, QS 07 - Lote 01, EPCT, Brasília, DF, CEP 71966-700, Brazil

    Thiago Christiano Silva, Solange Maria Guerra & Marcus Vinicius B. dos Santos

Authors
  1. Thiago Christiano Silva
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Solange Maria Guerra
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Marcus Vinicius B. dos Santos
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Thiago Christiano Silva.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

We thank Editor-in-Chief Subal C. Kumbhakar and the Associate Editor for handling the paper’s revision process and the two anonymous referees for the insightful comments. Thiago C. Silva (Grant Nos. 308171/2019-5, 408546/2018-2) gratefully acknowledges financial support from the CNPq foundation. The authors declare that there are no conflicts of interest.

Appendix A Geopolitical division and particularities of Brazil

Appendix A Geopolitical division and particularities of Brazil

Fig. 10
figure 10

Geopolitical map of Brazil in 2014. There are five regions. a North (in yellow), composed of the states of Acre, Amazonas, Rondônia, Roraima, Pará, Amapá, Tocantins. (b) Northeast (in brown), comprising the states of Maranhão, Ceará, Piauí, Rio Grande do Norte, Paraíba, Pernambuco, Alagoas, Sergipe, Bahia. c Midwest (in red), consisting of the states of Mato Grosso, Mato Grosso do Sul, Goiás, Distrito Federal. d Southeast (in green), composed of the states of São Paulo, Rio de Janeiro, Minas Gerais, Espírito Santo. e South (in blue), comprising the states of Paraná, Santa Catarina, Rio Grande do Sul. (Color figure online)

Full size image

In light of its vast territorial extension, Brazil has 27 federative units (26 states and 1 federal district) grouped into five central regions: North (7 states and 450 municipalities), Northeast (9 states and 1794 municipalities), Midwest (3 states, 1 federal district, and 467 municipalities), Southeast (4 states and 1668 municipalities), and South (3 states and 1191 municipalities). Figure 10 depicts the geopolitical organization of Brazilian regions and associated states as of 2014. Brazil has a large heterogeneity in terms of income and development indices. Figure 11 shows the IFDM for Brazilian (a) municipalities and (b) states. We observe that the North and Northeast regions have the lowest development indices, while the Southeast and South have the highest ones. Within the Midwest, there are substantial discrepancies in terms of development indices: the Distrito Federal (small square-shaped territory in the map) has one of the largest IFDM, while neighboring states have mid-to-low development indices.

Fig. 11
figure 11

Map of development indices—measured with the IFDM—at the a municipality and b state levels. The values correspond to the simple average over the period 2006–2014

Full size image

Concentration on a particular activity reflects the degree of specialization of the economy. While activity concentration may be positive in terms of increasing returns to scale, it increases the external dependence of municipalities. Figure 12 presents the average regional concentration degree of economic activities in Brazilian municipalities from 2006 to 2014. We calculate the regional concentration of municipalities using the Hirschman–Herfindahl Index (HHI), in accordance with (8). We can observe in Fig. 12a that, in general, municipalities in the South and Midwest tend to have less specialized regional economies (lower HHI, higher activity diversification). In contrast, municipalities in the North and Northeast regions have higher specialization of regional economic activities (higher HHI, lower activity diversification). Figure 12b displays state averages of regional activity concentration levels. Rio de Janeiro and Rio Grande do Norte have the highest level of activity specialization, particularly a high concentration of services activities. In contrast, Rio Grande do Sul, Santa Catarina, Goiás, and Tocantins have the most diversified regional economies.

In Fig. 13, we are able to see how regional activity concentration changes over time in the five Brazilian regions. The South region has the lowest variability of regional activity concentration around the mean and over time. However, it has the largest number of outlier municipalities.Footnote 43 We can see in Fig. 12a that these outliers refer to a few municipalities in Rio Grande do Sul and Paraná whose economies are highly specialized. The Midwest region presents similar behavior, with outliers originating from Mato Grosso and Mato Grosso do Sul. The Northeast is the region with the largest variation in the concentration of activities, both in terms of variation around the average and in temporal variation. There is an increase in the concentration of municipalities in this region from 2012 to 2014. The Southeast region presents similar behavior but to a lesser extent. The North region maintains the average concentration over time stable, but with variation around it asymmetrically.

Fig. 12
figure 12

Map of regional concentration of activities by a municipality and b Brazilian states. The regional concentration of activities is constructed from the preponderant activities (agriculture, industry, services) of the municipalities in the same mesoregion as the reference one. We consider as preponderant activity the one that generates the highest value-added to the municipality, i.e., contributes the most to its GDP. The regional concentration by HHI calculated using the distribution of these preponderant activities. The values correspond to the simple average over the period 2006–2014

Full size image
Fig. 13
figure 13

Boxplots of regional concentration of activities by Brazilian region during 2006–2014. The regional concentration of activities is constructed from the preponderant activities (agriculture, industry, services) of the municipalities in the same mesoregion as the reference one. We consider as preponderant activity the one that generates the highest value-added to the municipality, i.e., contributes the most to its GDP. The regional concentration by HHI calculated using the distribution of these preponderant activities

Full size image

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Silva, T.C., Guerra, S.M. & dos Santos, M.V.B. The role of externalities in fiscal efficiency. Empir Econ 62, 2827–2864 (2022). https://doi.org/10.1007/s00181-021-02124-1

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s00181-021-02124-1

Keywords

JEL Classification

Search

Navigation