Time series analysis applied to construct US natural gas price functions for groups of states

V. V. Kalashnikov, T. I. Matis, G. A. Pérez-Valdés

Research output: Contribution to journalArticle

13 Scopus citations

Abstract

The study of natural gas markets took a considerably new direction after the liberalization of the natural gas markets during the early 1990s. As a result, several problems and research opportunities arose for those studying the natural gas supply chain, particularly the marketing operations. Consequently, various studies have been undertaken about the econometrics of natural gas. Several models have been developed and used for different purposes, from descriptive analysis to practical applications such as price and consumption forecasting. In this work, we address the problem of finding a pooled regression formula relating the monthly figures of price and consumption volumes for each state of the United States during the last twenty years. The model thus obtained is used as the basis for the development of two methods aimed at classifying the states into groups sharing a similar price/consumption relationship: a dendrogram application, and an heuristic algorithm. The details and further applications of these grouping techniques are discussed, along with the ultimate purpose of using this pooled regression model to validate data employed in the stochastic optimization problem studied by the authors.

Original languageEnglish
Pages (from-to)887-900
Number of pages14
JournalEnergy Economics
Volume32
Issue number4
DOIs
StatePublished - Jul 2010

Keywords

  • Natural gas
  • Regression
  • Time series

Fingerprint Dive into the research topics of 'Time series analysis applied to construct US natural gas price functions for groups of states'. Together they form a unique fingerprint.

  • Cite this