American Society of Civil Engineers

Effects of Omitted Variable Bias on Construction Real Output and Its Implications on Productivity Trends in the United States

by Bryan Dyer, (corresponding author), M.ASCE, (Assistant Professor, Dept. of Technology, Eastern Kentucky Univ., 307 Whalin Technology Complex, 521 Lancaster Ave., Richmond, KY 40475-3102. E-mail:, Paul M. Goodrum, M.ASCE, (Professor, Civil Engineering Dept., Univ. of Kentucky, C151C Raymond Building, Lexington, KY 40506. E-mail:, and Kert Viele, (Associate Professor, (Formerly), Statistics Dept., Univ. of Kentucky, Lexington, KY 40506.)

Journal of Construction Engineering and Management, Vol. 138, No. 4, April 2012, pp. 558-566, (doi:

     Access full text
     Purchase Subscription
     Permissions for Reuse  

Document type: Case Studies
Abstract: An accurate measure of the U.S. construction industry’s productivity remains a consistently perplexing challenge. The thread of previous research has produced contradictory results, which contributes to the differences in opinion regarding whether construction productivity has been improving or declining over the past several decades. Typically, macroeconomic studies involving industry aggregated output measures have shown a decline in construction productivity, while micro studies involving activity based data have shown an increase. A dominant theory to explain these opposing results is that current measurements of construction inflation do not adequately consider the change in quality and amenities provided in today’s structures as compared to when the measurements were first established. The U.S. Census Bureau (Census) New Single-Family Houses under Construction Price Index is one of the industry’s major measurements of construction inflation. Previous research suggests that omitted quality variables in the Census price index leads to an omitted variable bias which overestimates construction inflation, leading to both an underestimate of construction industry output and productivity. However, previous research has not formalized necessary changes to the price index to avoid this bias, nor has the bias been actually measured to quantify its direction or magnitude. Utilizing sales data of new home construction in Bowling Green, Kentucky, this study examines the magnitude of the bias for one geographic location, and provides a direction for potential changes to improve the accuracy of the price index as a whole. The writers compare the current hedonic model with a proposed model on the basis of the Bowling Green data, which includes quality variables that are not part of the current hedonic model. While this study is obviously limited to the differences measured on this small data sample, this paper’s primary contribution to the overall body of knowledge provides the first known documented evidence of omitted variable bias in the current Construction Price Index, albeit limited to a very small geographic region. Regardless, these findings have important ramifications on the accuracy of existing industry output measures and related efforts to use those output measures for measuring the productivity of the overall U.S. construction industry.

ASCE Subject Headings:
Construction industry
Quality control
Case studies

Author Keywords:
Construction inflation indexes
Construction quality measurement
Construction productivity