Author:
Dimitropoulos Dimitri,Yatchew Adonis
Abstract
This paper emphasizes the use of data-driven techniques to discern trends in commodity prices. Whereas many previous papers rely on parametric assumptions, we take as our departure point the view that such trends areinherentlynonparametric, driven by complex forces of innovation and obsolescence, competition and strategic behavior, resource depletion and discovery, and supply and demand. Our reference specification is the partial linear modelyt=f(t) +ztβ + ϵtwhere macroeconomic variableszenter parametrically, andfis nonparametric, to be discovered using cross-validation.We analyze data on 11 commodities—3 hydrocarbons and 8 metals. For the majority of these commodities, our data-driven estimates offbear close similarity to band pass estimates which include long term trends and super cycles. The OPEC effect is estimated to have increased oil prices by over 50% on average and coal prices by about 25%. U.S. coal mining legislation is estimated to have increased coal prices by 9% to 15%.
Publisher
Cambridge University Press (CUP)
Subject
Economics and Econometrics
Reference64 articles.
1. Semiparametric Regression for the Applied Econometrician
2. World Bank (2014b) World Development Indicators Databank. GDP Growth [Data File]. Retrieved from http://data.worldbank.org/indicator/NY.GDP.MKTP.KD.ZG.
3. U.S. Geological Survey (2013a) Minerals Yearbook [Database]. Retrieved from http://minerals.usgs.gov/minerals/pubs/commodity/myb/.
4. U.S. Energy Information Administration (2014a) Annual Energy Outlook 2014. Coal Supply, Disposition, and Prices. [Data File]. Retrieved from http://www.eia.gov/oiafaeo/tablebrowser/.
Cited by
5 articles.
订阅此论文施引文献
订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献