递增阶梯定价政策评价与优化设计——基于充分统计量方法

Policy Evaluation and the Optimal Design for Increasing Block Pricing:The Sufficient Statistics Approach

本文通过求解阶梯定价下福利损失与再分配效应的充分统计量,构建评估与优化阶梯定价的半结构式充分统计量模型。利用从国家电网随机抽取的2003—2011年杭州市5000户居民的月度电量消费数据,评估样本期内实施与调整阶梯定价的政策效果;在此基础上对最优阶梯定价机制进行优化设计,进一步评价了杭州市2012年最新阶梯定价调整政策。研究发现:(1)2004年与2006年杭州市实施与调整阶梯定价的福利损失分配效应转化率逐步上升,阶梯定价再分配的边际福利损失与初次分配的税收边际福利损失相当;(2)分时阶梯定价分配效应转化率比纯阶梯定价低,嵌入分时定价的阶梯定价不能有效地实现阶梯定价目标;(3)与最优阶梯定价机制相比,2012年的最新阶梯定价调整方向正确,但第一阶梯长度扩展过多、第二阶梯扩展不足,而阶梯加价太小;(4)较之结构与缩减式模型,充分统计量方法极大节约数据信息量并能进行福利分析。本文的研究为评估与优化我国各类政府公用事业改革开拓了一条新思路。

Summary： It is important for both regulatory agencies and researchers to create policy evaluations and optimal designs for increasing block pricing. First, we survey the nonlinear pricing policy evaluation literature. By comparing a structural econometric model with a reduced-form econometric model, we use the sufficient statistics method to evaluate increasing Chinese block pricing policy. In theory, we create a semi-structured model of sufficient statistics that can evaluate and optimize increasing block pricing by solving the sufficient statistics of welfare loss and the redistribution effect of increasing block pricing. Based on 5,000 Hangzhou residents＇ monthly electricity consumption data, randomly selected from the State Grid from 2003 to 2011, we first evaluate the policy effect of implementing and adjusting the increasing block pricing during the sample period. We then create optimal designs for the pricing mechanism to further assess the latest price adjustment policy of Hangzhou in 2012.The theoretical and empirical results show the following. （1） Block pricing brings an efficient reallocation of income. The welfare loss allocation effect rates from the implementation and adjustment of Hangzhou＇s block pricing in 2004 and 2006 rise from 69.52% to 73.30%. The marginal welfare loss of block pricing declines from 0.44 yuan to 0.36 yuan, which almost equals the marginal welfare loss from tax. This shows that the efficiency of the redistribution is improved by introducing the block pricing system. （2） The welfare loss allocation effect of time-of-use block pricing is lower than that of pure block pricing. Time-of-use block pricing cannot effectively achieve the objective of block pricing. The aims of time-of-use and block pricing are contradictory. Furthermore, the welfare loss allocation effect of time-of-use block pricing is lower than that of pure block pricing, regardless of whether it is at a peak or valley period. In particular, the welfare loss allocation effect of block pricing during valley periods is much lower, resulting in larger welfare losses. The government should use time-of-use block pricing very carefully. （ 3 ） The sufficient statistics approach can be used to improve the design of block pricing, and the current block pricing system can be improved in terms of the rates of its welfare loss allocation effect. The latest block pricing adjustment in 2012 is going in the right direction compared with the optimal block pricing mechanism. However, some improvements are still needed, as the 2012 block pricing system has too much length expansion in the first block and too little length expansion and a price increase in the second block. （4） The sufficient statistics method greatly saves the amount of information data and can conduct better welfare analysis than the structure and reduced-form model. It can save the costs relating to policy decision making, and provide a feasible path for the mechanism＇s design. Finally, it can be widely applied to policy evaluations of the government＇s public utilities. In conclusion, the sufficient statistics method has a lower data requirement for policy evaluation. It can also connect the advantages of the structural and reduced-form models. This paper opens up a new way to reliably assess and optimize public utilities.