Institute of Transportation, MOTC

Chung-Szu Chiu ，2007.01
Graduate Institute of Traffic and Transportation Engineering and Management
Feng Chia University

　　Along with the promulgation of the government procurement act, the procedures of government procurements are much more rigid and complete. According to the 58nd article of the government procurement act, there are some mechanisms designed in the government procurement act to prevent from the awarding a contract to low price bidding tenderer. However, when there is a remarkable gap between the tender awarding value and the government estimate, it is hard to judge that it is a case of low price bidding or the original setting of the government estimate is too high in practices. Besides, the bidding prices of tenderer might be significantly affected by many case-specific and environmental factors, such as number of competitors, financial status of the tenderer, number of contract already awarded s of the tenderer, economic growth, and unemployment rate, etc. It deserves to closely investigate the actual influences of these factors and to provide a reference for setting the government estimate and judging the reasonability of bidding price.

　　Based on this, this study aims to identify the key factors affecting the gap between the tender awarding value and the government estimate. This study analyzes a total of 2251 construction procurement cases of the Second Maintenance Office of Directorate General of Highways, MOTC, during 2001 to 2006, by the techniques of stepwise regression analysis and cluster analysis, by considering the possible explanatory variables of category, number of bidders, amount of procurement budget, government estimation, tender awarding value, quality of procurement, and some macroeconomic status indexes. A total of 1984 cases during 2001 to 2005 are classified into 12 categories, and then used to respectively calibrate the regression models. The procurement cases during January to November, 2006 are used to validate the model. The results show that except for the category of bridge construction, number of bidders is the most significant variable in all of regression models within 11 categories, indicating that the gap get larger as the number of bidders increases. Construction cost index (CCI) is the second significant variable, which negatively affect the gap. Among these regression models of 12 categories, the category of marking construction works exhibits the highest explanatory capacity with R2=0.75, followed by the categories of decorating and planting construction works and lighting construction works with R2=0.64 and 0.54, respectively. The category of signal construction works shows the lowest explanatory capacity with R2=0.07. The validation results also show the applicability of the models. Apparently, the analytical results of this study can provide an important reference for committee members or procurement personnel to set the government estimate or to analyze the reasonability of tender price