Using machine learning to select variables in data envelopment analysis: Simulations and application using electricity distribution data

PublikationArtikel (med peer review)
Magnus Söderberg

Sammanfattning

Agencies that regulate electricity providers often apply nonparametric data envelopment analysis (DEA) to assess the relative efficiency of each firm. The reliability and validity of DEA are contingent upon selecting relevant input variables. In the era of big (wide) data, the assumptions of traditional variable selection techniques are often violated due to challenges related to high-dimensional data and their standard empirical properties. Currently, regulators have access to a large number of potential input variables. Therefore, our aim is to introduce new machine learning methods for regulators of the energy market. We also propose a new two-step analytical approach where, in the first step, the machine learning-based adaptive least absolute shrinkage and selection operator (ALASSO) is used to select variables and, in the second step, selected variables are used in a DEA model. In contrast to previous research, we find, by using a more realistic data-generating process common for production functions (i.e., Cobb–Douglas and Translog), that the performance of different machine learning techniques differs substantially in different empirically relevant situations. Simulations also reveal that the ALASSO is superior to other machine learning and regression-based methods when the collinearity is low or moderate. However, in situations of multicollinearity, the LASSO approach exhibits the best performance. We also use real data from the Swedish electricity distribution market to illustrate the empirical relevance of selecting the most appropriate variable selection method.

The article in total can be read here.

Duras, T., Javed, F., Månsson, K., Sjölander, P., & Söderberg, M. (2023). Using machine learning to select variables in data envelopment analysis: Simulations and application using electricity distribution data. Energy Economics, 106621.


Liknande innehåll

Working Paper No. 381: How social norm feedback can reduce unsorted waste and increase recycling in the residential sector
Working paperPublikation
Ek, C., & Söderberg, M.
Publiceringsår

2024

Publicerat i

Ratio Working Paper No Series.

Sammanfattning

The EU waste legislation requires member states to prepare 55% of municipal waste for re-use and recycling, to recycle 65% of all packaging waste by 2025, and to limit landfilling municipal waste to 10% by 2035. A large majority of the member states are at risk of missing one or more of these targets. Thus, there is a need to identify additional policies beyond command-and-control and market-based instruments that can effectively contribute to these targets. This policy brief describes one such policy: social norms feedback. This has been trialled with nearly 20,000 households in Sweden that faced Pay-as-you-throw schemes. In this setting, the unsorted waste fraction was reduced by around 10% and three quarters of that was due to increased recycling. The large trial sample and wide-ranging socio-economic characteristics suggests that 10-20% reduction of unsorted waste can be expected in jurisdictions with flat tariffs. If local governments collaborate and share the cost of waste truck equipment, then the policy is likely to generate a substantial economic surplus.

Norm-based feedback on household waste: Large-scale field experiments in two Swedish municipalities
Artikel (med peer review)Publikation
Ek, C. & Söderberg, M.
Publiceringsår

2024

Publicerat i

Journal of Public Economics

Sammanfattning

We conduct two large-scale randomized controlled trials to produce the first evidence that Home Energy Report-type norm feedback letters can be used to reduce household waste. We explore several feedback variants, including a novel short-run dynamic norm that emphasizes ongoing changes in waste behavior. Waste reductions are on the order of 7%–12% for all treatments, substantially larger than usually found in the energy or water domains. Effects are mostly driven by increased recycling of packaging and remain largely intact a year after the intervention ended. Feedback is highly cost effective compared to alternative non-price waste policies. However, net social benefits depend on household preferences for receiving feedback, which we elicit in a valuation survey, and whether existing waste fees internalize the marginal social cost of waste.

Ratio Working Paper No. 372: Customers’ value-for-money for a regulated service across differen towners
Working paperPublikation
Biggar, D., & Söderberg, M. (2024). Customers’ value-for-money for a regulated service across different owners (Ratio Working Paper No. 372). Ratio.
Publiceringsår

2024

Publicerat i

Ratio Working Paper Series.

Sammanfattning

What are the best ownership and governance arrangements for a natural monopoly facility? There are three broad approaches: (a) private ownership, coupled with arms-length public utility regulation; (b) some form of government (central, state, or local) ownership; and (c) customer or community ownership. While there is a substantial literature comparing outcomes under private and public (i.e., government) ownership, there is relatively little literature comparing private and/or government ownership with customer ownership. One of the obstacles of performance comparison is that different businesses may choose a different price-quality trade-off, making direct comparison impossible. In this study we cut through this problem by comparing customer perceptions of value-for-money. The study is based on interviews of more than 600 randomly selected electricity distribution customers in Sweden, approximately 150 in each ownership category (municipal, customer, private, and state). These distributors are subject to an identical regulatory framework. The results show that those owned directly by customers are perceived to deliver significantly more value for money than those owned by the government or by private investors. These results lend weight to the view that a well-governed customer-owned utility may lead to better outcomes than other owners.

Visa fler