Abstract
This paper investigates the relationship between output-based incentives for service quality and the use of capital and non-capital resources to meet regulatory targets in the electricity industry. To conduct the empirical analysis we use a dataset collected with the support of the Italian energy regulatory authority, comprising micro data on monetary incentives and physical assets for the largest electricity distribution operator in Italy (86 % of the market). Our results show that physical assets and operational expenditures do affect service quality. Moreover, when we investigate causality in the relationship between incentives to quality and the use of capital and non-capital resources, we find that incentives Granger-cause capital expenditures (and not vice-versa). Finally, our results reveal an asymmetric effect of rewards and penalties on capital expenditures’ decisions across areas with different quality levels. From these findings, we derive several policy implications.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
As for Italy, see Lo Schiavo et al. (2013).
For an overview of the regulatory regimes in the US electric utility industry on alternative output based goals in the US, see IEE (2013).
Notably, the ownership share held by the state through municipalities in the other Italian distributors is always larger than 50 %, for example, A2A—Milan with 54 %, ACEA—Rome with 51 %, AEM (now IREN)—Genoa and Turin with 54 %.
For the second tariff period (2004–2007) WACC was set 6.8 % and the X factor at 3.5 %. For the third tariff period (2008–2011) the WACC was increased to 7 % and the X factor was decreased to 1.9 %. For the current, fourth tariff period (2012–2015) the WACC is set at 8.6 % for the first 2 years and at 6.4 % for the remaining 2 years; the X factor is set at 2.8 %. Details on the choice of the WACC and X factors in the energy sector in Europe can be found in Cambini and Rondi (2010). Specific investment benefits have been introduced to support, for instance, the deployment of low-loss transformers and to promote automation and control of active grids (Lo Schiavo et al. 2013).
In addition to SAIDI the Italian regulation requires distribution companies to report the average number of long interruptions per customer, known by the acronym SAIFI (System Average Interruption Frequency Index), as well as the average number of short (shorter than 3 min and longer than 1 second) interruptions per customer: this index is called MAIFI (Momentary Average Interruption Frequency Index). The average number of long (short) interruptions per consumer is calculated as:
$$\begin{aligned} \textit{SAIFI}(\textit{MAIFI})=\sum \limits _{k=1}^M {\frac{N_k }{N_{tot} }} \end{aligned}$$where notation is as above. From 2000 to 2007 rewards and penalties were applied to SAIDI only. From 2008 onwards, rewards and penalties apply to SAIDI as well as to another indicator calculated as the sum of SAIFI plus MAIFI (total number of interruptions, long and short ones).
Each of the Enel’s units (Zones) includes two or three districts, typically of different density levels (see also Sect. 4).
In the time span of our analysis, the average household has paid an extra cost due to quality increments of about \(2 \,{\EUR }/\hbox {year}\). The cost of continuity regulation was accounted for in the distribution charges of the electricity bill. For the average household the latter amounts to around \(500\, {\EUR }/\hbox {year}\).
Urban networks present, compared to rural networks, shorter feeders, a higher share of underground cables and a higher level of redundancy. These structural characteristics favor continuity of supply.
In particular, the Italian regulation makes a distinction between interruption events and exceptional interruption events or, better, “exceptional time periods”. Since 2004 these events (time periods) are identified using a statistical methodology which, originally, identified an extreme region in the daily SAIDI (and SAIFI) plane, where such exceptional events (periods) belonged to. The boundaries of this region were originally defined for each district using thresholds of means and standard deviations of daily SAIDIs (for details see Fumagalli et al. 2009). Such exceptional events (thus including extreme weather) are considered as caused by Force Majeure.
See Arellano (2003, Ch. 6) for details regarding the use of Granger causality tests in the context of a panel setting. Granger causality tests were recently used to examine several regulatory issues such as tariff rates, leverage, investment, intensity of regulation, regulatory independence, etc. (Edwards and Waverman 2006; Bortolotti et al. 2011; Cambini and Rondi 2012).
By combining district-wide data into Zone-wide data, the relation between population density and continuity of supply (duration and frequency of interruptions, but also penalties and rewards) becomes considerably less precise.
Even so, in our empirical analysis we do test for differences across the two periods. We thank an anonymous referee for this suggestion.
As already mentioned, the variable SAIDI coincides with the regulated part of total SAIDI (e.g., it does not include notified interruptions, nor events that originated on the transmission network or that were caused by Force Majeure). Also, it was winsorized to exclude an outlier in the first year of observation.
Non-capital resources are represented in Ter-Martirosyan and Kwoka (2010) by two variables, operations expenses (that cover current firm operations) and maintenance expenses (that involve servicing the infrastructure). Our dataset does not allow us to distinguish between the two and we are bound to employ a single variable which, inevitably, aggregates preventive as well as corrective costs.
Ter-Martirosyan and Kwoka (2010) employ average income per capita or per capita consumption with the same purpose. Other potentially interesting control variables, such as zonal density or the average length of feeders per substation report a high correlation with the variable UNDER (respectively, 0.666 and −0.575). Hence, they were not included in the analysis.
The publicly available data from ISTAT are provided per administrative province, which closely matches Enel’s distribution units. Precipitation is defined as rain, snow, sleet or hail that falls on the ground and is measured in mm.
Accounting data typically include only historic cost valuations of fixed assets (capital stock), which usually bear little relation to current replacement cost of long-lived fixed capital assets. Hence, we calculate the replacement cost of the capital stock using the perpetual inventory formula: \(p_{t+1} K_{t+1}=p_{t} K_{t}(1-\delta )(p_{t+1}/p_{t})+p_{t+1} I_{t+1}\), where \(p_{t}\) is the domestic price index of investment goods in period t sourced by the ISTAT (the National Institute of Statistics), \(K_{t}\) is the fixed capital stock in period \(t, \,I_{t}\) is the investment flow in period t, and \(\delta \) is the depreciation rate. The sector specific depreciation rate for the energy sector (4.4 %) is derived from Bureau of Economic Analysis estimates reported in “Rates of Depreciation, Service Lives, Declining Balance Rates, and Hulten-Wykoff Categories”.
When taken as ratios to the beginning of year capital stock, variables INC, REWARD, PENALTY and OPEX are not adjusted for inflation.
Given the presence of a number of potentially relevant time invariant territorial characteristics, we started the empirical analysis by estimating a random effects panel model. However, although the random effects estimates are more efficient than fixed effects estimates, in order to be valid, one must ensure that the individual invariant component in the error term is not correlated with regressors. To test for the consistency of the random effects coefficients we thus employed the Hausman (1978) specification test, but in all specifications the results pointed us to use fixed effects estimation.
As a robustness check, we also estimated the above models normalizing the independent variables with respect to the power sold (MWh). Our results remain consistent with those in Table 4. For this reason, and to save space, we do not report them in the paper, but make them available upon request.
We use the dynamic System-GMM model developed by Arellano and Bond (1991) and Blundell and Bond (1998). This model estimates a system of level and first-differenced equations and uses lags of first-differenced variables as instruments for equations in levels and lags of variables in levels as instruments for equations in first-differences. For the estimation, we used the xtabond2 Stata module created by Roodman (2006).
The set of instruments includes lags of all the variables in the regression as well as a number of external variables that account for the unit-specific environment: the size of the service area in \(\hbox {km}^{2}\) (AREA); the area covered by forests, in ha (FOREST), and a dummy variable denoting proximity to the sea (COAST).
The procedure adapts the two-stage procedure to the GMM-System estimation method and relies on the critical values developed by Stock and Yogo (2005) for testing weak identification as used in the 2SLS framework. Following Wintoki et al. (2012), and adapting to the system structure of GMM-System, we perform the test on the levels of endogenous variables regressed on the instruments in first-differences and obtain the first Cragg-Donald (CD) Statistic. Then, we regress the first-differences of the endogenous variables on the instruments in levels and obtain the second CD statistic. We finally compare the CD statistics with critical values by Stock and Yogo (2005). In Table 7, the values of the CD test are well above the critical value of 10, which is the “rule of thumb” critical value suggested for assessing strength of the instruments. See also Fremeth and Shaver (2014) for a similar implementation of the test.
Differences in quality performance across Italian distribution units are associated with network structure and type of consumers served (Cambini et al. 2014). Since the latter can hardly be modified, it is not surprising that quality improvements in Italy are mainly driven by structural interventions (e.g., grounding of feeders and network automation).
The Sargan-Hansen test is robust, but may be weakened if there are too many instruments with respect to the number of observations (Roodman 2006). Therefore we follow a conservative strategy using no more than one/two lags of the instrumenting variables (i.e. the third or fourth lags in our case), to assure that the number of instruments is not greater than the number of firms. The rather demanding time structure of the Granger test and of the GMM-System estimator is also the reason why the number of observations drops from 690 in Table 4 to 345 in Table 5 (although of course all observations are used in the estimation).
The AR(2) tests of the second-order correlation in the first-differenced residuals could not be calculated by STATA due to length of our time series. However, the purpose of the AR(2) test is to assess the validity of instruments lagged 2 years and in case of invalidity of the instrument, the third lag has to be used. As explained in the footnote above, the third lag of the variable is indeed the earliest instrument that we use, hence the AR(2) test is not relevant to us.
We recall that previous literature has found (weak) evidence that quality standards (not necessarily output-based incentives) result in greater maintenance expenditures (Ter-Martirosyan and Kwoka 2010). We must postpone further investigation until this distinction in the data becomes available.
Recall that revenues from tariffs (i.e. sales) also cover quality-related costs for the provision of target SAIDI.
See, for example, Hubbard (1998) for a comprehensive survey of company investment models estimated with panel data, Fazzari et al. (1988) for a seminal contribution, Lyon and Mayo (2005) for an application to the US electric utility industry and Cambini and Rondi (2010) for an application to EU energy companies.
The set of instruments includes lags of investment, sales, cash flow and incentives (or rewards or penalties) to capital stock ratios as well as a number of external variables that account for the unit-specific environment: the percentage of the non residential energy consumption (PERC_NR); zonal density, measured by the number of LV consumers over network length (DENSITY); the area covered by forests, in ha (FOREST), and a dummy variable indicating Zones in the North of the country (NORTH).
We also checked the fixed effect results from the static version of the investment model and we found that they hold. Results are available on requests.
We thank an anonymous referee for suggesting us this further analysis.
Using a different empirical approach, Poudineh and Jamasb (2015) find that the cost of energy-not-supplied seems relevant in explaining the investment behavior of Norwegian electricity distribution companies. However, their result suggests that investments have mainly been of a corrective nature rather than a preventive one (they are a response to outages in the same time period).
References
AEEGSI (2007). Testo integrato della regolazione della qualità dei servizi di distribuzione, misura e vendita dell’energia elettrica. Regulatory Order no. 333/07. Available from: www.autorita.energia.it.
Arellano, M. (2003). Panel data econometrics. Oxford: Oxford University Press.
Arellano, M., & Bond, S. (1991). Some tests of specification for panel data: Monte Carlo evidence and an application to employment equations. Review of Economic Studies, 58(2), 277–297.
Arellano, M., & Bover, O. (1995). Another look at the instrumental-variable estimation of error-components models. Journal of Econometrics, 68, 29–51.
Blundell, R., & Bond, S. (1998). Initial conditions and moment restrictions in dynamic panel data models. Journal of Econometrics, 87, 115–143.
Bond, S. R., Elston, J. A., Mairesse, J., & Mulkay, B. (2003). Financial factors and investment in Belgium, France, Germany and the UK: A comparison using company panel data. Review of Economics and Statistics, 85, 153–165.
Bortolotti, B., Cambini, C., Rondi, L., & Spiegel, Y. (2011). Capital structure and regulation: Do ownership and regulatory independence matter? Journal of Economics & Management Strategy, 20(2), 517–564.
Cambini, C., & Rondi, L. (2010). Incentive regulation and investment: Evidence from European energy utilities. Journal of Regulatory Economics, 38(1), 1–26.
Cambini, C., & Rondi, L. (2012). Capital structure and investment in regulated network utilities: Evidence from EU telecoms. Industrial and Corporate Change, 21(1), 73–94.
Cambini, C., Croce, A., & Fumagalli, E. (2014). Output-based incentive regulation in electricity distribution: Evidence from Italy. Energy Economics, 45, 205–216.
Cerretti, A., Di Lembo, G., & Valtorta, G. (2005). Improvement in the continuity of supply due to a large introduction of Petersen coils in HV/MV substations. 18th International Conference on Electricity Distribution, CIRED, 6–9 June, Turin, Italy.
Coelli, T., Gautier, A., Perelman, S., & Saplacan-Pop, R. (2013). Estimating the cost of improving quality in electricity distribution: A parametric distance function approach. Energy Policy, 53, 287–297.
De Fraja, G., & Iozzi, A. (2008). The quest for quality: A quality adjusted dynamic regulatory mechanism. Journal of Economics and Management Strategy, 17(4), 1011–1040.
Edwards, G., & Waverman, L. (2006). The effects of public ownership and regulatory independence on regulatory outcomes. Journal of Regulatory Economics, 29(1), 23–67.
Fazzari, S., Hubbard, G., & Petersen, B. (1988). Financing constraints and corporate investment. Brooking Papers on Economic Activity, 1, 141–196.
Fremeth, A. R., & Shaver, J. M. (2014). Strategic rationale for responding to extra-jurisdictional regulation: Evidence from firm adoption of renewable power in the US. Strategic Management Journal, 35, 629–651.
Fumagalli, E., Lo Schiavo, L., & Delestre, F. (2007). Service quality regulation in electricity distribution and retail. Berlin: Springer.
Glachant, J. M., Khalfallah, H., Perez, Y., Rious, V., & Saguan, M. (2013). Implementing incentive regulation through an alignment with resource bounded regulators. Competition and Regulation in Network Industries, 14(3), 265–290.
Granger, C. W. (1969). Investigating causal relations by econometric models and cross-spectral methods. Econometrica, 37, 24–36.
Hausman, J. A. (1978). Specification tests in econometrics. Econometrica, 46(6), 1251–1271.
Innovation Electricity Efficiency (IEE) (2013). State electric efficiency regulatory frameworks. The Edison Foundation Report, July 2013. Available from: www.edisonfoundation.net/IEE.
Jamasb, T., & Pollitt, M. (2007). Incentive regulation of electricity distribution networks: Lessons of experience from Britain. Energy Policy, 35(12), 6163–6187.
Jamasb, T., Orea, L., & Pollitt, M. (2012). Estimating the marginal cost of quality improvements: The case of the UK electricity distribution companies. Energy Economics, 34(5), 1498–1506.
Joskow, P. L. (2008). Incentive regulation and its application to electricity networks. Review of Network Economics, 7(4), 547–560.
Kuosmanen, T. (2012). Stochastic semi-nonparametric frontier estimation of electricity distribution networks: Application of the StoNED method in the Finnish regulatory model. Energy Economics, 34(6), 2189–2199.
Lo Schiavo, L., Delfanti, M., Fumagalli, E., & Olivieri, V. (2013). Changing the regulation for regulating the change. Innovation-driven regulatory developments in Italy: smart grids, smart metering and e-mobility. Energy Policy, 57, 506–517.
Lyon, T., & Mayo, J. (2005). Regulatory opportunism and investment behavior: Evidence from the U.S. Electric Utility Industry. Rand Journal of Economics, 36(3), 628–644.
Ofgem (2010). Handbook for implementing the RIIO model. Available from: www.ofgem.gov.uk.
Poudineh, R., & Jamasb, T. (2015). Determinants of investments under incentive regulation: The case of the Norwegian electricity distribution networks, Energy Economics n4 (201409).
Reichl, J., Kollmann, A., Tichler, R., & Schneider, F. (2008). The importance of incorporating reliability of supply criteria in a regulatory system of electricity distribution: An empirical analysis for Austria. Energy Policy, 36(10), 3862–3871.
Roodman, D. (2006). How to Do xtabond2: an introduction to “Difference” and “System” GMM in Stata. The Center for Global Development, WP n. 103.
Sappington, D. E. M. (2005). Regulating service quality: A survey. Journal of Regulatory Economics, 27(2), 123–154.
Sims, C. A. (1972). Money, Income, and Causality. American Economic Review, 62, 540–552.
Stock, J. H., & Yogo, M. (2005). Testing for weak instruments in IV regression. In D. W. K. Andrews & J. H. Stock (Eds.), Identification and Inference for econometric models: A Festschrift in Honor of Thomas J. Rothenberg. Cambridge, UK: Cambridge University Press.
Ter-Martirosyan, A., & Kwoka, J. (2010). Incentive regulation, service quality, and standards in U.S. electricity distribution. Journal of Regulatory Economics, 38(3), 258–273.
Valtorta, G., Calone, R., D’Orazio, L., & Salusest, V. (2009). Continuity of supply improvements by means of circuit breakers along MV lines. 20th International Conference on Electricity Distribution, CIRED, 8–11 June, Prague, Czech Republic.
Wintoki, M. B., Linck, J. S., & Netter, J. M. (2012). Endogeneity and the dynamics of internal corporate governance. Journal of Financial Economics, 105, 581–606.
Author information
Authors and Affiliations
Corresponding author
Additional information
For comments on earlier versions of this paper we thank the Editor, Michael Crew, two anonymous referees, Alberto Iozzi, Tooraj Jamasb and David E.M. Sappington and participants at the 2nd Conference on the Regulation of Infrastructure Industries, EUI—Florence School of Regulation (Florence 2013), the Mannheim Energy Conference, ZEW, University of Mannheim (Mannheim 2013) and the 7th Meeting of the Network on Economics of Regulation and Institutions (Turin 2014). Technical and financial support from the Italian Regulatory Authority for Electricity, Gas and Water is kindly acknowledged. The opinions expressed in this paper do not represent the official position of the Italian Regulatory Authority and do not commit the Authority to any course of action in the future. Similarly, the opinions expressed in this paper do not represent the official position of Enel Distribuzione and do not commit Enel Distribuzione to any course of action in the future.
Rights and permissions
About this article
Cite this article
Cambini, C., Fumagalli, E. & Rondi, L. Incentives to quality and investment: evidence from electricity distribution in Italy. J Regul Econ 49, 1–32 (2016). https://doi.org/10.1007/s11149-015-9287-x
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11149-015-9287-x