Three essays in the efficient treatment of randomness

Channing Arndt, Purdue University


This dissertation consists of three essays. They examine diverse economic phenomena, but they are bound by a need to treat randomness effectively and practically. Each essay seeks to demonstrate enhanced capacity to effectively and practically treat randomness in specific problems of data collection, data analysis, and economic simulation. The economic insights gained in each essay are presented below.^ Essay one introduces a new sampling techniques called Gaussian quadrature (GQ) sampling which, under certain circumstances, outperforms random samples of the same size. The essay examines the circumstances under which GQ samples would be expected to outperform random samples of the same size. The process of collection of many types of economic data seems well suited to GQ sampling. For example, rural household revenue and expenditure surveys, particularly in less developed countries (LDCs), might be more efficiently accomplished using GQ sampling.^ Essay two examines demand systems estimation in the presence of binding quantity constraints. These constraints, particularly non-negativity constraints, are a pervasive problem in individual level data sets. The essay shows that the most widely used estimator in these instances generates biased estimates. In addition, the essay presents numerical methods which render consistent estimation practical using an estimator developed by Lee and Pitt.^ Essay three examines general equilibrium projections of production and trade in the year 2005 under conditions of uncertain growth in gross domestic product (GDP) and capital stock for each region in the model. Distributions for GDP and the capital stock are developed using time series data. The analysis generates logical mean projections for the evolution of the global economy over the period 1992-2005. However, standard deviations on model results appear to be small. This confirms earlier results indicating that the model might not be capable of capturing the full dynamism of the world economy. Nevertheless, analysis of variability in output and trade flows yielded two testable hypotheses. Results suggest that output from export oriented sectors tends to be more stable than import competing sectors and that exports, by commodity, tend to be more variable than imports. ^




Major Professor: Wallace E. Tyner, Purdue University.

Subject Area

Economics, Agricultural|Economics, Theory

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server