Date of Award
12-1980
Document Type
Thesis
Degree Name
Master of Science (MS)
Recommended Citation
Wolters, Theodore P., "Estimation and forecasting gross state product: An application to North Dakota" (1980). Theses and Dissertations. 7595.
https://commons.und.edu/theses/7595
COinS
Comments
This investigation was conducted for two purposes. The first purpose was to develop a method for estimating Gross State Product (GSP) for North Dakota on a quarterly basis. Secondly an extrapolation procedure was to be developed for forecasting GSP.
Quarterly estimates of GSP are desired for several reasons. For example, comparing quarterly estimates of GSP with quarterly GNP would enable planners and decisionmakers to determine how sensitive the state's economy is to the nation's business cycle or to federal policy actions. Also quarterly estimates of GSP would provide more data points for a forecasting model.
Forecasts of North Dakota's economy are necessary because of data lags. To obtain the necessary information to estimate GSP for any given year the researcher has to face a delay of approximately two years. Reliable forecasts would be valuable to both the public planner and decisionmaker as well as to business forecasters. Public planners can use forecasts to estimate the effects of policy decisions on the economy or as a variable to input into. their revenue forecasting models·. Business forecasters can incorporate GSP estimates into their forecasting models as an exogenous variable rather than using their own resources to develop such estimates.
Chapter 2 contains a review of the literature for estimating GSP. The chapter explains the estimation technique as first used by Kendrix and Jaycox and ad hoc modifications to the basic technique. The modifications came from Bryan Adair for farm GSP, Albert Niemi for modifying manufacturing GSP, L'Esperance et al for modifying the procedure for estimating government GSP and Mark Henry for his modifications in the Mining Sector.
Chapter 3 is a survey of selected forecasting techniques. This chapter covers two explanatory forecasting processes and four time series forecasting processes. The explanatory models covered are Input-Output Models and econometric models. The time series techniques covered are Naive forecasting models, Exponential Weighted Moving Average models, and Generalized Adaptive Filtering processes and Box-Jenkins methodology.
In Chapter 4, quarterly estimates of each sector's GSP are given. Also included in Chapter 4 is an explanation of how these estimates were obtained on a quarterly basis for each sector.
The final chapter explains why the Box-Jenkins method was selected as the forecasting technique for North Dakota GSP. Sector forecasts and sector forecasting models are also presented in this chapter. Also included in this chapter is an analysis of the North Dakota GSP forecasts and conclusions concerning future directions for research in forecasting a state's GSP.