New Developments in Productivity Analysis

Citation

Vecchi, M. (2002), "New Developments in Productivity Analysis", Journal of Economic Studies, Vol. 29 No. 6, pp. 446-449. https://doi.org/10.1108/jes.2002.29.6.446.1

This book is a collection of papers presented at the 1998 Conference on Research in Income and Wealth, held in Silver Spring, Maryland. It provides a rich forum of discussion of the main issues related to the evaluation of productivity, and it contains both new and old ideas on productivity analysis. The introductory paper by the editors discusses the main implications of the 15 papers that appear in the volume, as well as presenting the key differences between this conference and a former session held in 1975.

The first impression, looking at the book, is that there have not been very many new development sin productivity analysis. Total factor productivity (TFP) is still the “star of the show”, 40 years after Solow’s seminal contribution, and it is still widely used in productivity studies. However, a closer look at the several contributions included in the volume, shows that, like most stars, TFP has been subject to “cosmetic surgery” that has enabled the application of the index number approach to the analysis of new issues. The methodological and measurement problems related to these new issues is the focus of most of the chapters. The end product is a rich overview of what productivity analysis has achieved since 1975 and what still needs to be done.

The structure of the volume makes a very useful and accessible tool, even for a researcher new to productivity analysis. In fact, the first three chapters provide a review of the main theoretical and empirical issues behind TFP calculations. Hulten (Chapter 1) gives a clear exposition of the index number approach, as well as attempting to provide an answer to two key issues, the 1970s productivity slowdown and the recent productivity paradox. Chapter 2, by Dean and Harper, is an historical account of the work carried on by the Bureau of Labour Statistics (BLS) in analysing productivity, while Diewert, focuses on measurement issues in the context of the theory of production (Chapter 3).

Staring from the BLS experience, most of the contributions to the volume extend the analysis in different directions, by modifying the theoretical framework or by presenting an alternative methodology to TFP, such as the econometric approach. The latter finds a marginal space in the volume. This is mainly discussed in Chapters 4 and 7. Chapter 4, by Nadiri and Prucha, analyses the contribution of dynamic factor demand models. These models can provide a rich representation of the factors affecting productivity, such as adjustment costs, expectations and technical change. Unlike the index number approach, this methodology has successfully integrated R&D capital and R&D spillovers in the analysis and hence provides an interesting analytical tool. Its complexity however has played against a wide application in empirical studies. A more straighforward econometric approach is the one presented by Basu and Fernald in Chapter 7. They attempt to distinguish between competing theories of the business cycle by using an extended version of the standard Solow’s model that allows for imperfect competition, increasing returns to scale and variable factor utilisation. Unlike the standard residual (see Chapter 1), this modified version can provide a measure of technical change.

Solow’s contribution to the productivity debate provides an evaluation of the work that has been done following his 1956‐1957 model (Chapter 5). He also introduces the embodiment problem, which is further investigated by Greenwood and Jovanovic, with reference to the endogenous growth literature (Chapter 6). The models they use are based on the endogenous growth theory. The embodiment problem implies that technology does not fall on the firm like manna from heaven but it is embodied in the new capital. Hence new technology needs new investments, in both tangible and intangible assets, in order to influence productivity. The authors provide an interesting review of the extensions of Solow’s original model that accounts for the embodiment problem. They also use this new analytical framework to explain some of the features of the post‐war experience of the USA, such as the 1970’s productivity slowdown and the rise in wage inequality.

Chapters 8‐10 focus more on the use of micro‐economic data from productivity analysis. As emphasised several times throughout the volume, an important development since 1975 has been the increasing availability of firm/plant level data, both within manufacturing and non‐manufacturing. This allows for a wider heterogeneity in the analysis and for a better understanding of what causes productivity fluctuations. Moreover, the availability of micro data allows the investigation of further issues that cannot be analysed when dealing with aggregate variables. For example, we can account for phenomena such as the reallocation of outputs and inputs across individual firms which, according to Foster, Haltiwanger and Krizan (Chapter 8) has an important effect on aggregate productivity growth. By reviewing and existing empirical literature and providing some new evidence for the US manufacturing and service establishments (the automobile repair shop sector), the authors are able to evaluate the different contribution of reallocation over time. Using micro data also allows one to evaluate where technological advances are more likely to affect productivity and how a particular industry has coped with the increasingly high technology environment. Finally, the authors suggest a change in the way the collection and processing of data is carried on, shifting the focus on the micro‐level data and allowing the analyst to connect the macro aggregates to the microstatistics. This suggestion is probably welcomed by economic researchers, if not by the statistical offices given the amount of resource that its implementation entails.

Other examples of productivity analysis in non‐manufacturing can be found in Chapter 9, by Ellerman and Stoker, that looks at the sources of productivity growth in the US coal industry, and Chapter 10, that analyses the performance of five service sectors. In the latter, Baily and Zitzewitz summarise the main findings of the analysis of productivity performance of retail banking, telecom, retailing, public transportation and airlines, carried on by the McKinsey Global Institute. The study, that looks at various countries with very different institutional structures and economic development (the USA and Brazil are in the sample), addresses several issues such as the complexity of output and productivity measurement in the service sector. The authors also note that the development of these industries has been possible thanks to the adoption of ICT capital. However, a methodology that correctly accounts for the impact of this new capital is still missing and, consequently, the improvements in the service sector are not fully captured by the productivity statistics.

As mentioned above, an important development in productivity analysis has been the availability of micro‐level data. However, the range of countries for which such data is available is quite limited. At the same time, some variables, such as human capital, are only available at the aggregate level. Hence, international productivity comparisons have to rely on macro data. In Chapter 11 by Islam, different approaches to TFP measurement are applied to developed and emerging economies, in order to focus on issues such as convergence. Chapter 12 by Jorgenson and Yip, show how the inclusion of endogenous investments in education greatly increases the explanatory power of the economic growth theory. As well as presenting the major results from the application of this approach to OECD countries, the authors provide a useful review and a set of references for anyone wishing to study the issues of capital and labour heterogeneity in more depth.

In Chapter 13, Ball, Fare, Grosskopf and Nehring pursue two objectives: first, they present an alternative methodological tool to TFP calculations, i.e. they base their analysis on the use of Malmquist index of productivity; second, they account for the effect of undesirable outputs, such as the effects of the agricultural use of chemical pesticides and fertilizers on groundwater and surface water quality. In doing so, the authors introduce another new issue in the productivity debate, the importance of the environment. This theme is further developed in the following chapter by Gallop and Swinand, where environmental factors are included in a modified TFP measure, defined as total resource productivity (TRP). The authors show anothers new application of the standard growth accounting framework, that is its extension to the analysis of environmental issues.

The paper by Griliches concludes the volume. In one of the author’s last contributions to economics he reminds the reader of the work that still needs to be done in order to refine the analysis. The main direction of the improvement is towards a better treatments of intangible assets, such as R&D and human capital. Despite the fact that the importance of R&D capital on productivity is emphasised in several papers, a specific contribution on this topic does not appear in the volume. Given the extensive literature on the topic, a summary of the main results would have added to the completeness of the book. Nevertheless, the volume provides a rich contribution to the understanding of productivity analysis and provides several routes for future research.