Back to results list
Show full item record
Please use this identifier to cite or link to this item:
|Title:||Construction manpower demand forecasting||Authors:||Wong, JMW
Multiple regression analysis
|Issue Date:||2011||Publisher:||Emerald Group Publishing Limited||Source:||Engineering, construction and architectural management, 2011, v. 18, no. 1, p. 7-29 How to cite?||Journal:||Engineering, construction and architectural management||Abstract:||Purpose– The purpose of this paper is to examine the performance of the vector error‐correction (VEC) econometric modelling technique in predicting short‐ to medium‐term construction manpower demand.
Design/methodology/approach– The VEC modelling technique is evaluated with two conventional forecasting methods: the Box‐Jenkins approach and the multiple regression analysis, based on the forecasting accuracy on construction manpower demand.
Findings– While the forecasting reliability of the VEC modelling technique is slightly inferior to the multiple log‐linear regression analysis in terms of forecasting accuracy, the error correction econometric modelling technique outperformed the Box‐Jenkins approach. The VEC and the multiple linear regression analysis in forecasting can better capture the causal relationship between the construction manpower demand and the associated factors.
Practical implications– Accurate predictions of the level of manpower demand are important for the formulation of successful policy to minimise possible future skill mismatch.
Originality/value– The accuracy of econometric modelling technique has not been evaluated empirically in construction manpower forecasting. This paper unveils the predictability of the prevailing manpower demand forecasting modelling techniques. Additionally, economic indicators that are significantly related to construction manpower demand are identified to facilitate human resource planning, and policy simulation and formulation in construction.
|Appears in Collections:||Journal/Magazine Article|
Show full item record
Citations as of Nov 29, 2018
Citations as of Dec 10, 2018
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.