Portal de Periódicos da CAPES

Measuring productivity in the construction industry

2006; Taylor & Francis; Volume: 34; Issue: 3 Linguagem: Inglês

10.1080/09613210600590041

1466-4321

Paul L. Crawford, Bernard Vogl,

Construction Project Management and Performance

Abstract The paper provides an overview of methods used to measure productivity in the construction industry. The advantages and disadvantages of average labour productivity and total factor productivity measures are discussed in detail, and the relationship between these two measures is established both theoretically and in an application at the industry level. The usefulness of any productivity measurement framework for policy-makers and industry practitioners alike depends crucially on the extent to which it enables the identification of the underlying drivers of productivity. This requirement necessitates an approach that involves formally describing the production process and explaining as much as possible of construction output in terms of the quantity and quality of inputs that are used to generate it. Whilst it is accepted that data requirements are a major constraint to such an approach, it is suggested that by establishing a robust measurement framework, data deficiencies can be defined more easily. Guidance on areas where improvements are needed is provided and it is considered that the focus of future research should be in creating new and improving existing datasets. Cet article donne une vue d'ensemble des méthodes utilisées pour mesurer la productivité dans le secteur de la construction. L'auteur examine en détail les avantages et les inconvénients de la mesure de la productivité moyenne du travail et du facteur total de la productivité; la relation entre ces deux mesures est établie tant sur le plan théorique que dans le cadre d'une application au niveau industriel. L'utilité de tout cadre de mesure de la productivité pour les décideurs politiques et les professionnels de l'industrie dépend, pour l'essentiel, de la mesure dans laquelle elle permet d'identifier les facteurs sous-jacents de la productivité. Cette exigence nécessite une approche qui implique une description formelle du processus de production et qui explique autant que possible les résultats de la construction en termes de quantité et de qualité de facteurs contributeurs qui sont utilisés pour la générer. Bien que l'on admette que les besoins en données soient une contrainte majeure pour une telle approche, on pense qu'en établissant un cadre de mesure solide, on pourrait définir plus facilement les insuffisances en matière de données. L'auteur donne des conseils sur les domaines qui doivent être améliorés et pense que les futures recherches devraient se concentrer sur la création d'ensembles de données et sur l'amélioration de ceux qui existent. Mots clés: secteur de la construction, productivité du travail, fonction de production, mesure des ressources entrant dans la production, facteur de productivité total Keywords: construction industrylabour productivityproduction functionresource input measurestotal factor productivity Acknowledgements The authors thank their colleagues at the Department of Trade and Industry and four anonymous referees for helpful comments, but the responsibility for the views expressed and any errors lie solely with the authors. Notes 1. These measures will be defined and explained in the second section. 2. Products must be similar to control for differences in demand. Capital intensities must be similar to ensure that apparent productivity differences are not due to factor substitution. Other differences are less important since they often do affect the potential of a firm or economy to translate inputs into outputs. 3. If measures can be developed for these 'intangible' inputs, they would join the set of measurable inputs. 4. The production function can be respecified in per-capita terms as: Y/L = Af(K/L). This helps illustrate that an increase in capital intensity is a movement along, not a shift in the production function. 5. As a simple example, if Y = AK α, then log(Y)=y = log(AK α)=log(A) + αlog(K)=a + αk, and we obtain the additive log form: y = a + αk. 6. PPP are the rates of currency conversion that equalize the purchasing power of different currencies by eliminating the differences in price levels between countries. Strictly speaking, they are not exchange rates. In their simplest form, they are the ratio of the prices in national currencies of the same good or service in different countries. They are mainly used in the comparison of national GDP and industrial gross value added (GVA), for welfare purposes. Where performance in the market rather than welfare is the focus, it can be argued that annual market exchange rates might be more appropriate. For a construction-specific discussion of PPPs, see Vermande and Van Mulligen Citation(1999). 7. The direct comparison, unit price comparison, overlap, production cost, and imputation methods have also been developed to control for output quality, but hedonic prices are currently preferred. The US is the leader in developing these, and the availability of comparable international series is still limited. 8. For an exposition of the PIM, see Martin Citation(2003). 9. The NIESR are producing capital stock estimates for the Department of Trade and Industry (DTI) at the three-digit SIC level. 10. Renting and operating/leasing of construction and civil engineering machinery and equipment without operator and renting of scaffolds and work platforms without erection and dismantling is recorded under SIC 71.32 as a service activity, whereas renting of this machinery and equipment with operators is classified as construction in SIC 45.50. 11. For a discussion of these methodologies, see Van Biesebroek Citation(2004). 12. This flexibility may be at the cost of restrictions across time and space, which, however, is not a serious restriction if the focus is on long-run relationships between output and inputs. 13. Best-practice econometrics develops this basic specification to allow for (1) dynamic adjustment, via a lag structure, (2) the non-separability of inputs, via cross-product terms, and (3) the simultaneous determination of inputs and outputs, via the generalized method of moments (GMM) estimator, but this does not need to be explored here. 14. The constant in Equationequation (3) has now evolved into a trend variable. If the data have no time dimension, e.g. if the data are project-based, the time subscripts and the time trend will drop out of Equationequation (5) and TFP will be measured as the sum of ϵ i and α i. 15. For an overview of the earlier work, see Chau and Walker Citation(1988), and more recent reviews in Edkins and Winch Citation(1999a) and Royal Commission into the Building and Construction Industry Citation(2002). 16. The plants produce an identical product, polyethylene terephthalate pellets, which are used in the manufacture of a range of plastic products. 17. For example, Carr and Winch Citation(1999) analyse data that were generated by a computerized activity sampling tool called CALIBRE. 18. The methodology used for this exposition is based on a Cobb–Douglas production function, and relies on the assumptions of perfect competition and constant returns to scale. If perfect competition does not hold, TFP will be underestimated; and if the assumption of constant returns to scale does not hold, TFP will be overestimated (Jorgenson et al., Citation1987; Pentrin and Levinsohn, Citation2004). One solution to this problem would be to estimate a more flexible production function such as a translog specification using a parametric method (Harris, Citation2003). 19. Accidents might be a proxy for management quality