Sectoral Productivity Estimates

Understanding the productivity puzzle is an important objective of government and policymakers. There are many explanations for the slowdown in productivity growth in advanced economies in recent decades. One hypothesis is that measurement issues have exaggerated the extent of the slowdown. This is not a criticism of standard practice but reflects the idea that economic activity has increased in relative importance in areas that are poorly measured. We revisited the sector-level productivity estimates in this context and in the light of advances in measurement.

The policy of ONS is to introduce double deflation into the National Accounts following the recommendations of the Bean Review. There are a number of concerns about the effects of double deflation. First, the industry-level estimates of real value added may be very different from the current official estimates using single deflation. Second, the double deflation estimates may be more volatile, perhaps implausibly so. Third, double deflation has the potential to change the past. Even using exactly the same data, the growth rate of real GDP may be significantly different from the previously published official figures. We addressed these concerns using a simple approach relying only on publicly available data.

We also analysed new detailed industry-level data developed by ONS to re-examine the UK productivity puzzle. We carried out an accounting exercise that allowed us to distinguish general macroeconomic patterns from sector trends and idiosyncrasies, providing a roadmap for anyone interested in explaining the puzzle. We focused on the UK market sector and found that UK industries that saw the biggest reductions in productivity growth tended to be internationally competitive and more dependent on global demand than other industries. They were also industries where productivity is difficult to measure.

We analysed the effect of double deflation on output and productivity at the industry and whole economy levels. We used publicly available price indices (deflators) at the level roughly of the Supply and Use Tables to estimate double-deflated real value added for seventy-nine industries over the period 1997-2015. Our price indices were the ones employed by ONS to produce the official estimate of GDP from the output side using single deflation. We aggregated our double-deflated industry-level real value-added estimates using the same methodology and data as the ONS to give an alternative estimate of real GDP from the output side.

In analysing the slowdown in productivity growth, we focused on developments in the UK market sector and in particular on the post-recession period of productivity stagnation from 2011 to 2015, compared to the decade before the Great Financial Crisis that began in 2007. We considered where in the economy this stagnation was located, examining productivity patterns across fifty-nine market-sector industries and fifteen broader industry sections. Using internationally comparable data, we considered the UK productivity growth slowdown compared to those in the US and EU. We noted some additional patterns that emerged from the sectoral composition of the puzzle when contrasted with international trade metrics and in light of economic measurement issues.

We found that on average the growth rate of industry-level real value added was substantially lower under double deflation than under single deflation, and the year-to-year volatility of growth rates was much higher under double deflation. We also found that with these estimates, GDP would have grown substantially more slowly than the official estimate for 1997- 2015. Our theoretical work on double deflation suggests that it is possible to implement double deflation in such a way that there is no effect on the headline growth rate of GDP. This requires using a comprehensive set of deflators which are currently not publicly available. It is possible, however, that even on this approach the industry-level estimates of real value added may be quite different from the current official ones.

Our analysis of the productivity slowdown using detailed industry data pointed to the importance of economy-wide and indeed global explanations for the slowdown. Also, the UK productivity growth puzzle is concentrated in sectors where productivity is difficult to measure, such as information and communication and finance, suggesting that measurement issues should not be dismissed as part of an explanation of the puzzle.

This project challenged the methods and delivery of double-deflated estimates of National Accounts, contributing to the ONS decision to delay full implementation of double deflation in Blue Book 2019. The project was influential in helping ONS to understand the variety of options available and methodological challenges as they arose. We also provided input and challenge to ONS work to estimate the intermediate expenditure proportions that feed into the Supply and Use Tables and GDP estimates, following the reintroduction of the Annual Purchases Survey in 2015. Our analysis of the UK productivity puzzle influenced policy debate and the way that ONS presents the real estate sector in industry analyses of productivity.