Premium
An Application of spectral analysis in determining crop Rotation Frequencies
Author(s) -
Just Richard E.
Publication year - 1980
Publication title -
canadian journal of agricultural economics/revue canadienne d'agroeconomie
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.505
H-Index - 37
eISSN - 1744-7976
pISSN - 0008-3976
DOI - 10.1111/j.1744-7976.1980.tb01934.x
Subject(s) - frequency domain , econometrics , series (stratigraphy) , regression analysis , regression , mathematics , lag , time domain , time series , ordinary least squares , statistics , linear regression , nonlinear system , harmonic , computer science , physics , mathematical analysis , paleontology , computer network , quantum mechanics , computer vision , biology
Spectral analysis has been applied many times in agricultural economics in determining harmonic characteristics of observed time series. However, spectral techniques have rarely been applied in conjunction with models which stress economic relationships. But two such empirical approaches are possible. One can use either frequency‐domain regression of times series, i.e., general distributed lag estimation, in which time series of all important economic variables are considered [4, 5, 11], or one can use ordinary econometric estimation methods and then study frequency content of estimated disturbances with spectral techniques [1, Chapter 10]. The application of spectral analysis in this paper corresponds to the latter case. Ordinary time‐domain (rather than frequency‐domain) regression methods are used because the general lag relationship estimated here is not a case included in the general frequency‐domain regression model. First, nonlinear regression techniques are used to explain the usual economic content of acreage supply response. Then estimated disturbances are investigated using spectral analysis to determine the importance of cyclic behavior in aggregate supply response. Specifically, this paper demonstrates how a cross section of time series may be pooled to obtain statisticaly significant findings when time series are short. These methods are demonstrated in an investigation of harmonic content in regression disturbances apparently due to crop rotation.