Evidence for Ratchet Effect among German Electricity Distribution System Operators (DSOs) Michael Hellwig and Dominik Schober ZEW Centre for European Economic Research MaCCI Mannheim Centre for Competition and Innovation
Executive Summary: Motivation Introduction of incentive regulation in Germany in 2009 RRRRRRRRRRRRRR CCCCCC = CCCCCCCCCC nnnnnn cccccccccccccccccccccccc + CCCCCCCCCC cccccccccccccccccccccccc XX iiiiiiiiiiiiiiiiiiii Two regulatory regimes Hybrid, dynamic yardstick (> 30.000 connected consumers) Revenue cap (< 30.000 connected consumers) Same dynamic incentives: reduction of inefficient costs over 5-year period But difference in base year Inflate costs to increase future revenues in revenue cap regime Ratchet effect Especially for rather efficient DSOs (too tough targets resulting from cost simplification) 2
Executive Summary: Results DSOs in the revenue cap regime pile up more expenditures than their counterparts in the yardstick regime in the base year Empirical evidence for ratchet effect Efficient DSOs in revenue cap regime invested most compared to others and lost their initially high level of cost efficiency Ignoring heterogeneity in firm efficiency by setting uniform efficiency targets creates the disincentive to invest inefficiently Strong point for yardstick regulation We add to both the literature on the ratchet effect and on the comparison of incentive regulation regimes 3
Identification DSOs in revenue cap and yardstick have same dynamic incentives over the regulatory period apart from the base year Treatment and control group in diff-in-diff framework However, small DSOs can opt for revenue cap regulation Account for possible selection bias by excluding strategic switchers Revenue Cap 2 nd Period Yardstick 1 st Period Revenue Cap Yardstick < 30.000 consumers < 30.000 consumers > 30.000 consumers < 30.000 consumers > 30.000 consumers < 30.000 consumers 4
Methods Difference-in-differences: fixed-effects OLS regression Identification of disadvantaged DSOs in revenue cap via efficiency analyses Stochastic Frontier Analysis & Data Envelopment Analysis TOTEX = f(grid size, exit points, energy delivered, installed capacity for renewable electricity) Decomposition of total factor productivity into catch-up and technological change 5
Data 105 DSOs with less than 100.000 connected consumers 22 in yardstick, 83 in revenue cap regime Data derived from annual statements and other legal reporting requirements 2010-2013 Dependent variable: net investment in fixed assets Cost approximation follows guidelines of Incentive Regulation Ordinance TOTEX = CAPEX (= imputed equity yield rate + imputed depreciation) + OPEX (= material + personnel + sundry costs + interest on borrowed capital) - non-controllable costs (= concession fees + charges for the use of upstream network levels + feed-in remuneration for renewables) 6
Diff-in-diff results (investment) Dependent variable: Investment ratio Revenue cap x base year Disadvantaged x revenue cap x base year Non-disadvantaged x revenue cap x base year Revenue cap Revenue cap vs. yardstick vs. yardstick (1) (2) 0.0080 (0.0061) 0.0244** (0.0107) 0.0032 (0.0063) Control variables: Exit points, Energy delivered, Network length, Cap. renewable, Growth solar cap., Lag. growth solar cap., Grid acquisition DSOs 105 105 R² within 0.22 0.26 F 2.44 2.33 Notes: OLS estimation with DSO-fixed effects and time-fixed effects. Clusterrobust standard errors in parentheses. Distinction between non- and disadvantaged DSOs using SFA efficiency scores. Years 2011 and 2013. *,**,***: significant at 10%, 5% and 1% respectively. 7
Diff-in-diff results (expenditure measures) Dependent variable: TOTEX (in m euro) OPEX (in m euro) CAPEX (in m euro) Revenue cap vs. yardstick Revenue cap vs. yardstick Revenue cap vs. yardstick Revenue cap vs. yardstick Revenue cap vs. yardstick Revenue cap vs. yardstick (1) (2) (3) (4) (5) (6) Revenue cap x base 1.020* 0.938* 0.082 year (0.536) (0.528) (0.061) Disadvantaged x revenue cap x base 1.374** (0.566) 1.236** (0.551) 0.138* (0.074) year Non-disadvantaged x revenue cap x base year 0.916* 0.850 0.066 (0.540) (0.532) (0.061) Control variables: Exit points, Energy delivered, Network length, Cap. renewable, Growth solar cap., Lag. growth solar cap., Grid acquisition DSOs 105 105 105 105 105 105 R² within 0.41 0.42 0.39 0.40 0.29 0.31 F 11.87 11.62 11.67 11.29 4.50 4.06 Notes: OLS estimation with DSO-fixed effects and time-fixed effects. Cluster-robust standard errors in parentheses. Distinction between non- and disadvantaged DSOs using SFA efficiency scores. Years 2011 and 2013. *,**,***: significant at 10%, 5% and 1% respectively. 8
Efficiency change results TFP decomposition of SFA scores (2010-2013) 9
Conclusion Empirical evidence for ratchet effect: cost inflation in base year Rather efficient DSOs in revenue cap regime piled up most costs and lost their initially high level of cost efficiency Ignoring heterogeneity in firm efficiency by setting uniform efficiency targets creates the disincentive to invest inefficiently Too much simplification comes at a price Strong point for yardstick regulation 10
Thank you for your attention! hellwig@zew.de 11