A predictive model for construction labor supply in the Houston Standard Metropolitan Statistical Area

Abstract

The purpose of his study was to develop a construction labor supply model for the Houston SMSA. This task was undertaken because of the need of business and government for an efficient forecasting mechanism that would lead to better allocation of manpower. The study shows that the methods available for use in developing a model fell into four classes. Each of these methods were examined ad found to be lacking in applicability to the specific objective of this study. For this reason, general regression methods were used in developing the model. The research methodology involved correlation analysis, time plot analysis, MAX R² variable selection techniques and multiple regression. Through the correlation analysis several classes of independent variables were identified. The first class included data on population, labor force, and employment. Wage data in the construction industry made up the second class. The third class dealt with data on the firms in the industry and the final class was made up of data on the dollar value of construction. The time plots revealed that all variables exhibited seasonal variation. Therefore, all variables were adjusted for seasonal variation by using a ratio-to-moving-average technique. Once all seasonal adjustments were made the correlation analysis and time plot analysis was repeated. The variable selection was accomplished using the statistical technique of MAX R². Several sets of variables were used in order to gain insight as to which combination of variables would lead to the best regression models..