Are natural gas spot and futures prices predictable?

Research output: Contribution to journalArticleResearchpeer-review

12 Citations (Scopus)

Abstract

We answer two questions concerning natural gas spot and futures prices. The first is: Can natural gas futures prices predict natural gas spot prices? The second is: Are natural gas spot and futures prices weak form efficient or can they be predicted based on examination of historical data? To answer these questions, we use daily data for Henry Hub natural gas spot and futures prices. Our answer to the first question is that natural gas futures prices do not predict the magnitude of future natural gas spot prices any better than what would be predicted by a random walk model. This result has important implications for many financial analysts and policy institutions that have used commodity futures prices to predict movements in spot prices. The answer to the second question is that when we apply a unit root test that allows for heteroskedasticity and two structural breaks, natural gas spot and futures prices are predictable. We then simulate a contrarian trading strategy for spot and futures prices to show under what circumstances trading in spot and futures prices are also profitable. The results point to the need to accommodate heteroskedasticity when applying unit root tests to model energy spot and futures prices with high-frequency data, such as daily data.
Original languageEnglish
Pages (from-to)178 - 186
Number of pages9
JournalEconomic Modelling
Volume54
DOIs
Publication statusPublished - 2016

Cite this

@article{67d693e10c534ae097a31bcd87560fa0,
title = "Are natural gas spot and futures prices predictable?",
abstract = "We answer two questions concerning natural gas spot and futures prices. The first is: Can natural gas futures prices predict natural gas spot prices? The second is: Are natural gas spot and futures prices weak form efficient or can they be predicted based on examination of historical data? To answer these questions, we use daily data for Henry Hub natural gas spot and futures prices. Our answer to the first question is that natural gas futures prices do not predict the magnitude of future natural gas spot prices any better than what would be predicted by a random walk model. This result has important implications for many financial analysts and policy institutions that have used commodity futures prices to predict movements in spot prices. The answer to the second question is that when we apply a unit root test that allows for heteroskedasticity and two structural breaks, natural gas spot and futures prices are predictable. We then simulate a contrarian trading strategy for spot and futures prices to show under what circumstances trading in spot and futures prices are also profitable. The results point to the need to accommodate heteroskedasticity when applying unit root tests to model energy spot and futures prices with high-frequency data, such as daily data.",
author = "Vinod Mishra and Smyth, {Russell Leigh}",
year = "2016",
doi = "10.1016/j.econmod.2015.12.034",
language = "English",
volume = "54",
pages = "178 -- 186",
journal = "Economic Modelling",
issn = "0264-9993",
publisher = "Elsevier",

}

Are natural gas spot and futures prices predictable? / Mishra, Vinod; Smyth, Russell Leigh.

In: Economic Modelling, Vol. 54, 2016, p. 178 - 186.

Research output: Contribution to journalArticleResearchpeer-review

TY - JOUR

T1 - Are natural gas spot and futures prices predictable?

AU - Mishra, Vinod

AU - Smyth, Russell Leigh

PY - 2016

Y1 - 2016

N2 - We answer two questions concerning natural gas spot and futures prices. The first is: Can natural gas futures prices predict natural gas spot prices? The second is: Are natural gas spot and futures prices weak form efficient or can they be predicted based on examination of historical data? To answer these questions, we use daily data for Henry Hub natural gas spot and futures prices. Our answer to the first question is that natural gas futures prices do not predict the magnitude of future natural gas spot prices any better than what would be predicted by a random walk model. This result has important implications for many financial analysts and policy institutions that have used commodity futures prices to predict movements in spot prices. The answer to the second question is that when we apply a unit root test that allows for heteroskedasticity and two structural breaks, natural gas spot and futures prices are predictable. We then simulate a contrarian trading strategy for spot and futures prices to show under what circumstances trading in spot and futures prices are also profitable. The results point to the need to accommodate heteroskedasticity when applying unit root tests to model energy spot and futures prices with high-frequency data, such as daily data.

AB - We answer two questions concerning natural gas spot and futures prices. The first is: Can natural gas futures prices predict natural gas spot prices? The second is: Are natural gas spot and futures prices weak form efficient or can they be predicted based on examination of historical data? To answer these questions, we use daily data for Henry Hub natural gas spot and futures prices. Our answer to the first question is that natural gas futures prices do not predict the magnitude of future natural gas spot prices any better than what would be predicted by a random walk model. This result has important implications for many financial analysts and policy institutions that have used commodity futures prices to predict movements in spot prices. The answer to the second question is that when we apply a unit root test that allows for heteroskedasticity and two structural breaks, natural gas spot and futures prices are predictable. We then simulate a contrarian trading strategy for spot and futures prices to show under what circumstances trading in spot and futures prices are also profitable. The results point to the need to accommodate heteroskedasticity when applying unit root tests to model energy spot and futures prices with high-frequency data, such as daily data.

U2 - 10.1016/j.econmod.2015.12.034

DO - 10.1016/j.econmod.2015.12.034

M3 - Article

VL - 54

SP - 178

EP - 186

JO - Economic Modelling

JF - Economic Modelling

SN - 0264-9993

ER -