IAEE Members and subscribers to The Energy Journal: Please log in to access the full text article or receive discounted pricing for this article.

Noisy Data and Uncertain Coefficients: A Comment

Noisy data are frequently used to estimate regression coefficients for energy demand equations. Although analysts new to the area might expect reported coefficients and forecasts to reflect this source of uncertainty, reported numbers, in general, do not reflect it. Recently, Kher et al. (1987, hereinafter KSS) confronted the problem and proposed a new technique for reporting forecasts based on the estimation of bounds for regression coefficients. Regression coefficient bounds reflect the uncertainty in an estimated model stemming from the use of noisy data. In this comment I identify additional studies that have addressed the estimation of such bounds and also present an example of a more common use of them.

Purchase ( $25 )

Energy Specializations: Energy Modeling – Energy Data, Modeling, and Policy Analysis

JEL Codes: C52: Model Evaluation, Validation, and Selection, C51: Model Construction and Estimation, Q38: Nonrenewable Resources and Conservation: Government Policy, L71: Mining, Extraction, and Refining: Hydrocarbon Fuels, Q35: Hydrocarbon Resources, Q41: Energy: Demand and Supply; Prices, Q43: Energy and the Macroeconomy

Keywords: Noisy data, Uncertain coefficients

DOI: 10.5547/ISSN0195-6574-EJ-Vol10-No1-15

Published in Volume 10, Number 1 of the bi-monthly journal of the IAEE's Energy Economics Education Foundation.


© 2023 International Association for Energy Economics | Privacy Policy | Return Policy