![Sample our Global Development journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5929/TOC-Subject-Sample-2021-Global-Development.jpg"})
Abstract
Kalb A. What determines local governments' cost-efficiency? The case of road maintenance, Regional Studies. This paper investigates the determinants of local governments' cost-efficiency in road maintenance for a panel of German counties using a broad variety of estimation approaches. More specifically, efficiency indices were calculated using non-parametric and parametric reference technologies, and were examined by estimating and comparing four different regression models. The results show that the disposable income of the counties' citizens, intergovernmental grants and the municipalities' payments to the counties influence efficiency negatively. Concerning political variables the results show weak evidence that efficiency decreases with an increasing share of seats of left-wing parties in the county council.
Kalb A. 什么因素决定地方政府的成本效益？道路维护之案例，区域研究。本文透过各种广泛的评估方法，探讨德国郡县地方政府专门小组的道路维修成本效益决定性因素。更明确而言，本研究运用非参数与参数指数技术计算效益指标，并透过评估、比较四个不同的回归模型加以检验。研究结果显示，该郡县市民的可支配收入、跨层级政府补助款以及市政府缴纳至郡县的款项，对效益有着负面的影响。有关政治变因方面，研究结果证实，没有显著的证据支持效益将会随着郡县议会中左翼政党席次的增加而递减。
Kalb A. Qu'est-ce qui détermine le rapport coût-efficacité des collectivités locales? Étude de cas de la voirie, Regional Studies. Employant une grande variété de méthodes d'estimation, cet article examine sur le plan des collectivités locales les déterminants du rapport coût-efficacité de la voirie pour un panel de Länders allemands. Plus précisément, on a calculé des indices d'efficience à partir des technologies de référence paramétriques et non-paramétriques, et on les examinent en estimant et en comparant quatre modèles de régression différents. Les résultats laissent voir que le revenu disponible des citoyens des Länders, les aides intergouvernementales et les paiements des municipalités aux Länders influent de façon négative sur l'efficience. Pour ce qui est des variables politiques, les résultats laissent voir de faibles preuves que l'efficience diminue au fur et à mesure que la part des sièges des partis de gauche augmente dans les Länders.
Kalb A. Wovon hängt die Kosteneffizienz der Kreisverwaltung ab? Der Fall der Straßensanierung, Regional Studies. In diesem Beitrag werden die Determinanten für die Kosteneffizienz der Kreisverwaltung bei der Straßensanierung anhand eines Panels von deutschen Landkreisen mit Hilfe eines breiten Spektrums von Schätzungsansätzen untersucht. Insbesondere werden die Effizienzindizes mit Hilfe von nicht-parametrischen und parametrischen Referenztechniken berechnet und durch Schätzen und Vergleichen von vier verschiedenen Regressionsmodellen untersucht. Aus den Ergebnissen geht hervor, dass sich das verfügbare Einkommen der Bürger des Landkreises, der Finanzausgleich und die Zahlungen der Gemeinden an die Landkreise negativ auf die Effizienz auswirken. Hinsichtlich der politischen Variablen weisen die Ergebnisse wenige Anzeichen dafür auf, dass die Effizienz bei einem höheren Anteil von linksgerichteten Parteien im Kreistag abnimmt.
Kalb A. ¿Qué determina la rentabilidad de los gobiernos locales? El caso del mantenimiento de las carreteras, Regional Studies. Con ayuda de una amplia variedad de métodos de estimación, en este artículo se investigan los determinantes de la rentabilidad de los gobiernos locales en el mantenimiento de las carreteras para una muestra de distritos alemanes. En concreto, se calcularon los índices de eficiencia mediante tecnologías de referencia paramétricas y no paramétricas, y se examinaron calculando y comparando cuatro modelos diferentes de regresión. Los resultados indican que los ingresos disponibles de los ciudadanos de los distritos, las subvenciones intergubernamentales y los pagos de los municipios a los distritos influyen negativamente en la eficiencia. En lo que respecta a las variables políticas, los resultados muestran pocos indicios de que la eficiencia disminuya al aumentar la proporción de escaños de los partidos de izquierdas en los ayuntamientos de los distritos.
Acknowledgements
The author would like to thank Uwe Brixy, Lars P. Feld, Paul Frijters, Friedrich Heinemann, Benno Torgler, the participants at the International Summer Conference in Regional Science 2009 in Lübeck, Germany, and two anonymous referees for helpful comments and suggestions. Support by the Deutsche Forschungsgemeinschaft (DFG – German Research Foundation) within its Priority Programme 1142 ‘Institutionelle Gestaltung föderaler Systeme’ is gratefully acknowledged.
Notes
The results of such a specific approach need not necessarily be the same as the results of a global approach since in a global approach there can be ‘offsetting effects’ between the different services provided by a local government. To see this, suppose, for example, that, for the sake of simplicity, a local government only delivers road maintenance and education. Moreover, assume that the output indicators of the services are constant and that the overall budget is fixed. As a result, less spending on road maintenance and more spending on education would imply an increase in road maintenance efficiency, a decrease in education efficiency but no change in global efficiency.
Strictly speaking, in the case of DEA, technical efficiency scores are estimated in the context of a production function framework, whereas the stochastic frontier analysis is based on a cost-function framework yielding cost-efficiency indices. The difference between the two concepts is that the cost-efficiency contains – apart from the ‘technical’ component – additionally an ‘allocative’ component.
In the present context, the decision-making units are the counties of Baden-Württemberg, or, more specifically, the civil servants of the counties authorized to make decisions.
For an introduction, see, for example, Fried et al. Citation(2008), pp. 3–92.
For the descriptive statistics of the data, see the third section.
For a review of two-stage approaches, see Simar and Wilson Citation(2007).
Simar and Wilson Citation(2007) also proposed a single bootstrap procedure where they corrected only for the biased standard errors of the estimated coefficients in the second-stage regression. The double bootstrap procedure, however, outperforms the single bootstrap procedure (Simar and Wilson, Citation2007, p. 55). Therefore, only the results of the double bootstrap procedure are presented in the third section. The results of the single bootstrap procedure are, however, very similar and are available from the author upon request.
To be in line with the authors, the number of bootstrap replications in the first (second) procedure L1 (L2) is set equal to 200 (2000) (also Simar and Wilson, Citation2007, pp. 42–43).
For more details, see Kumbhakar and Lovell Citation(2000).
Parts of this section draw on Kalb Citation(2010b).
According to the Nomenclature des Unités Territoriales Statistiques (NUTS) nomenclature established by Eurostat, the counties correspond to NUTS-3 regions.
For more details, see Neser Citation(2004).
In Germany there are basically four different types of roads (with regard to the financing of the roads): roads that have to be financed (1) by the federal state (Bundesfernstraßen),
Equation(2) by the states (Landstraßen),
Equation(3)
by the counties (Kreisstraßen), and
Equation(4)
by the municipalities (Gemeindestraßen).
considers only rural counties since there are differences between rural and urban counties concerning the revenue and expenditure structure. This is due to the fact that urban counties are county and municipality at the same time and therefore also have to represent municipal interests.
The collection rates can differ from county to county. In 2004, the average collection rate was 33.62%.
The difference between the municipalities' payments to the counties and ‘usual’ taxes is that the payments to the counties are not paid by the citizens (of the jurisdictions) themselves but by the municipal governments belonging to the counties.
For more details on the sources of the data as well as the calculation of some of the variables, see Appendix A.
This assumption implies that all roads are treated as homogenous products. The surface of roads – and also their quality – is, however, heterogeneous: there are bituminous, paved/concrete asphalt or gravel roads. This, in turn, influences the maintenance costs; engineering studies show that high-quality roads are far less expensive to maintain than low-quality roads. Chicoine et al. Citation(1989), for instance, solved this problem by creating an index which scales the roads into a homogeneous equivalent. Unfortunately, such data are not available for the county roads in Germany. However, to take account of this problem, the number of accidents due to bad road conditions was included as a proxy for the quality of the roads to the estimation specification (see below).
This variable also contains accidents caused by other bad road conditions (for example, by slippery roads). Unfortunately, a more detailed indicator was not available. In addition, the number of accidents caused by bad road conditions was available only for the total road network. But given the fact that county roads constitute 43% of the total road network (outside of built-up areas) in Baden-Württemberg (see the third section this is a minor problem.
This interpretation implicitly assumes that the driving behaviour of the (car) drivers is independent of road quality. If instead it is assumed that better road conditions induce (car) drivers to drive faster, and if higher speed results in a higher number of accidents, an increase in the number of accidents could also be caused by (generally) good road conditions.
For more details on the calculation of this indicator, see Appendix A.
For more details on the classification of the counties, see Appendix A.
Baden-Württemberg, the state considered here, consists only of agglomeration as well as urbanized areas (). This is due to the fact that Baden-Württemberg is one of the most industrialized states in Germany.
In Germany, people obtaining long-term unemployment benefits (Arbeitslosengeld II) are usually allowed to own cars up to a maximum value of €7500.
Gross domestic product (GDP) (per capita) would be a better approximation of business fluctuations. However, since information about GDP at the county level is not very reliable (the counties are very small units), the unemployment rate was used as a proxy for business fluctuations.
Since the urban counties do not receive payments from municipalities (because they are county and municipality at the same time), the revenue was taken from the trade tax to proxy the (main) income source of the urban counties.
Free voters' unions are loose federations of persons that do not belong to specific political parties and exist only at the local level.
The descriptive statistics of the efficiency indices along with their correlation matrix are provided in and in Appendix B, respectively.
For the results of the first-stage stochastic frontier regression, see in Appendix B.
The DEA-based efficiency estimates were estimated separately for every year; in other words, the DEA frontier is allowed to shift over time in order to control for technical change (as in the case of the parametric frontier).
The differences in the interpretation of the signs in and are due to the fact that either the efficiency scores (lying between 0 and 1) or their reciprocals (lying between 1 and ∞) can be used as independent variables in the (second-stage) regression. Both values were used as independent variables (for the non-parametric as well as the parametric efficiency estimates) and those models with the highest adjusted pseudo-R2 were chosen. The results of the (two) remaining models, however, are very similar and are available from the author upon request.
In the truncated regression of , total population (along with its squared term) was removed to obtain a negative coefficient (which, however, is insignificant) for the kilometres travelled on county roads.
Since the width of county roads changed only marginally in recent years, this is a minor problem.
Doll et al. Citation(2002) calculated costs of €0.02/km for motorcycles, of €0.04/km for cars, and of €0.05/km (€0.30/km) for trucks with a maximum total weight of below (above) 12 tons for federal highways in Germany (in 2003).