Estimating industry-level productivity trends in the building industry from building cost and price data

A method of measuring total-factor productivity (TFP) trends in the building industry is described in this paper. This method is an improved version of the approach described in a paper by Chau and Walker in that it requires less restrictive assumptions and is theoretically less biased while requiring only slightly more data. With small modification, the same method can be used to measure other productivity trends corresponding to other productivity concepts. The data used in measuring TFP of Hong Kong’s building industry are also different from those proposed in Chau and Walker. A number of modifications have been made. These modifications have been possible both as a result of work by Chau and the increased availability of statistics in recent years. One of the major difficulties in measuring TFP trends in the building or construction industry has been the lack of data. This is also one of the major reasons for the dearth of empirical studies in this area. Very few attempts have been made to solve the problem of data availability. Lowe recognizes the difficulties in obtaining suitable data for measuring TFP of the British construction industry. His suggestion, however, is to use capital productivity as the second-best alternative to TFP rather than solving the problem. Chau and Walker’s solution is to develop a method for estimating TFP from construction cost and price data. The method is then used to measure the TFP of Hong Kong’s building industry. Since construction cost and price data are usually more readily available in most countries, Chau and Walker’s approach is potentially useful in these countries. There are, however, certain deficiencies in Chau and Walker 's approach in that some of the underlying assumptions are relatively restrictive. As more statistics are now available in Hong Kong, this has rendered Chau and Walker’s approach a crude tool for making full use of the data. This paper presents a modified approach which requires less restrictive assumptions and can make better use of the newly available data.