Date on Master's Thesis/Doctoral Dissertation

8-2013

Document Type

Doctoral Dissertation

Degree Name

Ph. D.

Department

Industrial Engineering

Committee Chair

Depuy, Gail W.

Committee Co-Chair (if applicable)

Heragu, Sunderesh S.

Author's Keywords

Regression; Economic indicator; LODI; Index conversion

Subject

Economic indicators--Kentucky--Louisville Metropolitan Area; Business logistics--Kentucky--Louisville Metropolitan Area

Abstract

Indices are popular in many sectors of the US economy and are commonly used by businesses when making important decisions. In this research, two new indices are developed; a regional index for the Greater Louisville area and a national index for the United States. Both indices predict changes in the level of the logistics and distribution activity (as measured by employment values) and can be used by various organizations to plan expenditures that effect their logistics and distribution operations. This analysis utilizes two types of data; raw tonnage and economic factors. Local railway and local barge data are reported by the Ports of Indiana. Airway data for the Greater Louisville area is collected from the Regional Airport Authority. Both local and national roadway data come from the American Trucking Association. National data for air transit is collected from the Bureau of Transportation Statistics while barge data is provided by the Army Corps of Engineers. National railway data is provided by the Association of American Railroads. Both local and national employment data is collected from the Bureau of Labor Statistics. Additional organizations provided secondary data including: the Purchasing Manager's Index (PMI) [y-charts.com], Gross Domestic Product [y-charts.com], crude oil prices [the Energy Information Administration], and the exchange rate of the US Dollar [International Monetary Fund]. Linear regression models are utilized to predict a response variable that is then converted into an index value for both the regional and the national indices. The regression models are tested against historical data to ensure their predictions are valid. Resulting index values are also tested to confirm the changes they indicate coincide with actual changes in employment data. A comparative analysis is completed which verifies that the national index is as useful as (if not more useful than) existing indicators. The regional regression was found to make predictions within 3% of the regional employment values 81% of the time. The national model falls within 3% of the national employment values 67% of the time. These indices are found to be valid, leading indicators of future activity in the logistics and distribution industry, for their specified regions.

Download

Share

COinS