Management Discretion and Political Interference in Municipal Enterprises. Evidence from Italian Utilities

Abstract

The attention of scholars and policy-makers is shifting from full privatisation to alternative measures as a means to improve the efficiency of public services. This article focuses on three restructuring measures adopted by local governments: partial privatisation (without transfer of control rights), inter-municipal joint ventures, and the presence of outside directors on the Board. Divergent hypotheses on these measures have emerged from two economic theories, namely public choice and the agency perspective. An empirical analysis investigates the effects of restructuring processes on a sample of Italian municipal utilities. Partial privatisation has been shown to play a significant and positive role, while Board independence from management has a less significant influence. Multi-government utilities have been found to suffer from severe coordination costs, which are not outweighed by scale benefits. Our findings reveal that management discretion is a significant source of inefficiency in municipal enterprises.

Key Words:

• Municipal utilities
• restructuring
• management discretion
• political influence

Acknowledgments

We are indebted to anonymous referees, Massimiliano Piacenza, Davide Vannoni, and participants in the XVIII AiIG Conference in Milan and the DIG Lunch Seminars held at the Politecnico di Milano for their helpful comments on this and other related work. The responsibility for any errors lies solely with the authors. This article is based upon a work that was supported by a PRIN 2007 grant (long-term and short-term models for the liberalised electricity sector; Italian Ministry of University).

Notes

 1. In most countries, public authorities are the reference model, while with very few exceptions Italian municipal utilities have been transformed into private law corporations since the early 1990s (Law 142/1990; see Section 2).

 2. Unioncamere (2008) carried out a census of Italian utilities that are fully or partly owned by local governments. These total 1130, and are generally larger than the mean Italian utility (i.e. on average 94 versus 32 employees per enterprise).

 3. In 1994, an increasing awareness about the critical status of water services induced Parliament to issue a comprehensive sector reform (Carrozza 2010). Aside from management models, major provisions concern horizontal integration of service operations within pre-defined territorial areas (provinces in most cases), vertical integration between distribution and sewerage activities, and a regulatory authority to be established by local governments. In compliance with the 1998 internal gas market EU Directive, the gas market was liberalised in 2000 (Dorigoni and Portatadino 2009). Distribution was legally separated from retailing, and the supply market was progressively opened up to new entrants. At the same time, water and gas reforms delegated price regulation to local and national authorities. Tariffs were required to recover full costs and to favour productivity increases. Although cost pass-through factors are still present, and public subsidies have not yet abandoned the water sector, revenue and price cap schemes are now the dominant rule.

 4. A number of local governments have launched privatisation programmes, but control rights are still, in most cases, in public hands (988 out of 1130 local utilities, Unioncamere 2008).

 5. Our dataset is limited to monitor equity arrangements. In contrast, the dataset does not monitor non-equity agreements and mono-service equity agreements (e.g. consortia targeted at gas procurement or refuse collection).

 6. Small towns generally submit deals that will achieve favourable governance agreements, because the citizens and elected representatives are unlikely to concede their cities’ withdrawal from utility operation.

 7. This is why the principal-agent approach emphasises the noxious side effects of those restructuring policies that are designed without proper care of the multiplicity of tasks and stakeholders (Dixit 2002; see also Fumagalli et al. 2007 on service quality after privatisation).

 8. The agency theory also enriches predictions made by property right theories of privatisation (Schmidt 1996, Hart et al. 1997). The efficiency advantage of private service providers is jeopardised if management is separated from external owners, which is quite common in privatised enterprises. The agency theory also sheds light on the traditional ‘expense-preference’ theory (e.g. Oswald et al. 1994).

 9. Along with legitimate political objectives and difficulties contracting performance requirements (Bel et al. 2007, Warner and Hefetz 2008, Hart et al. 1997, Levin and Tadelis 2010), vested interests are in fact a main explanation of the resistance to privatising public services.

10. The creation of larger and jointly owned enterprises could in addition be described as a pre-emption strategy, which discourages large potential rivals from entering the market.

11. If one considers private firms, predictions by the agency theory are clear. Outside (i.e. non-executive) directors are recognised as having stronger incentives than inside (i.e. executive) directors to control financial performance (e.g. Baysinger and Hoskisson 1990), and to limit inefficiencies within the firm (Shleifer and Vishny 1997).

12. In more general terms, the role of outside directors is not limited to financial controls in firms that have political stakeholders; for instance, non-executive directors with a political background are prevalent even in private firms that interact with government agencies (e.g. utilities) (Agrawal and Knoeber 2001).

13. In principle, a full representation of each output could be required to include both the number of users and physical quantity (i.e. delivered, collected or treated quantities). This specification of output structure is not feasible, because we have a limited number of degrees of freedom. Fortunately the physical quantities are closely correlated to the number of users. The correlation statistics are 0.76, 0.81, 0.97 and 0.94, respectively for gas, water, electricity and waste.

14. The ratios between the number of user connections and the network length for, respectively, gas, electricity and water, have been weighted by the corresponding share of users (i.e. WATSH, ELESH, GASSH). This measure does not include the users of waste services, due to the absence of a physical network.

15. The Appendix offers additional information on the econometric analysis, and reports the complete estimates for baseline, stand-alone and augmented models (Table A1).

16. In this article we refrain from labelling returns to the user number as ‘network effects’, in order to avoid any confusion. Several industrial organisation studies use ‘network effects’ and ‘network externalities’ in an interchangeable manner, but it is well known that water, electricity, gas and waste management do not exhibit direct network externalities.

17. It should be noted that Table 5 reports the coefficient estimates that are obtained after having normalised SIZE and other variables to the sample means.

18. In the same way we simulate the combination of two medium-sized utilities (i.e. SIZE varies from 605,292 to 1,210,584 users; MUNHF varies from 1 to 0.5). Cost variation at the sector level is equal to 4.78% (positive at a 99% confidence level).