The value of primary versus secondary data in interindustry analysis : an Arizona case study emphasizing water resources.
AuthorBoster, Ronald Stephen,1944-
Water resources development -- Arizona.
Water-supply -- Arizona.
Committee ChairMartin, William E.
MetadataShow full item record
PublisherThe University of Arizona.
RightsCopyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction or presentation (such as public display or performance) of protected items is prohibited except with permission of the author.
AbstractInterindustry, or input-output analysis is a widely used economic tool in regional analysis. This study investigates the relative worth of alternative types of input data for such models. Primary data are defined as first-hand, or survey data; secondary data are defined as second-hand, or published data. Procurement of primary data is generally much more expensive than for secondary data. Most regional economists have long held that, in general, primary data are superior to secondary data for regional investigations. This study attempts to assess the value of primary versus secondary data in view of the wide variation in collection costs. Two recently published input-output studies for the 1958-60 Arizona economy were utilized to accomplish this goal. One model used mostly primary data; the second model was compiled entirely from secondary data sources. Following careful considerations of base-period differences, geographical overlap, and sector definitions, two models-- the ARZ and the CRB--were constructed from the two source studies so as to be commensurable for statistical comparisons. The ARZ model was derived from the source study that was compiled entirely from secondary data sources; the CRB model was compiled from the source study that used mostly primary input data. Several nonparametric statistical tests were utilized on several components of the ARZ and CRB models to test for statistical similarities and dissimilarities. In general, analyses indicate strong statistical similarity between the two derived models for aggregative I-0 characteristics such as entire matrices and output multipliers, but less similarity for less aggregative components such as sub-matrices and columns of matrices. For example, comparisons between entire technical and between interdependency matrices fail to reject the null hypothesis of no median difference at the alpha level of 5 per cent. At still more unaggregated component analysis, such as for columns within technical and interdependency columns a reasonable degree of statistical similarity persisted. One chapter (V) focuses on the two derived models as they relate to water resources planning implications for Arizona. Results are similar to those for the non-water analyses; however, in the case of weighted water multipliers the similarities between the ARZ and CRB models are remarkably close. Based on an assumed short-run change of 10 per cent in deliveries to final demand for all sectors in the economy, the models differed in predictions of induced aggregate water requirements by only 9000 acre-feet of water, or less than 2 per cent. Results from the study cast doubt on the commonly held assumption of primary-data supremacy in regional interindustry studies. Results also indicate that the most important component for regional economic analysis is the final demand for each sector rather than the interindustry flows. This follows from the narrow clustering of values for output multipliers consistently observed in regional I-0 studies, and reaffirmed in this study. Therefore, students of regional economies are advised to spend marginal resources (money, time, energy) compiling more accurate final demand vectors rather than developing more accurate endogenous interindustry flows.
Degree NamePh. D.
Degree ProgramHydrology and Water Resources