The OECD Productivity Manual at 25
This June, statisticians, economists and policymakers gathered in Paris to reflect on the legacy of the OECD Productivity Manual, which marked its 25th birthday in 2026, and discuss the future of productivity measurement.
The event held particular significance for ESCoE. The original manual was written by ESCoE Research Director Paul Schreyer during his time at the OECD, while ESCoE Director Rebecca Riley and ESCoE researcher Josh Martin contributed to discussions on the future direction of productivity statistics.
In this blog, they share their reflections on the past, present and future of productivity measurement.
The OECD Productivity Manual was developed at a time when productivity statistics were gaining prominence as tools for understanding economic growth and informing policy. Yet there was a gap between the rapidly expanding academic literature on productivity and the practical realities of producing official statistics.
One of the ambitions of the manual was to bridge that gap. It sought to provide a framework that was grounded in economic theory while remaining accessible and implementable for national statistical offices. Productivity measurement depends on many core economic statistics, including National Accounts, labour market data and capital measures, but its value extends further in helping policymakers understand economic performance and living standards in a coherent way.
I believe the manual was a success because it brought together different perspectives. It worked for statistical offices across a range of countries by providing a practical, pragmatic and non-prescriptive guide to measurement. It worked for academia by clearly pairing the theory with the practice. And it worked for policymakers by maintaining a clear link to the analytical questions productivity statistics are intended to answer.
At the same time, the manual reflected the concerns of its era. Its primary focus was on aggregate and industry-level productivity measurement. It did not attempt to cover areas that have since become central to productivity research, such as the use of firm-level microdata, international productivity level comparisons, or alternative approaches to analysing efficiency and performance.
While the manual stands the test of time, and the core principles remain relevant today, the discussions in Paris highlighted how much the economic environment has changed over the past quarter century. Digitalisation, AI, intangible assets, globalisation and environmental sustainability have all raised new questions that productivity statistics must increasingly address.
The challenge for any future update is therefore not to replace the foundations of the original manual, but to build upon them. The core framework remains valuable. The task now is to strengthen interpretation and communication, while extending the framework to address new measurement challenges and emerging policy needs.
Productivity measurement has come a long way in 25 years. Therefore, several of the measurement challenges highlighted in the manual are no longer considered relevant. Measuring labour input with hours worked instead of jobs and constructing capital services estimates are now commonplace. More complex measures, such as compositionally-adjusted measures of labour input (which account for the mix of skills in the workforce), are also increasingly produced.
The availability of productivity statistics for analysts has also expanded considerably. Industry-level estimates are now widely available, and there has been substantial progress in developing regional and sub-national productivity measures. Together, these developments have greatly improved our understanding of productivity dynamics across sectors, places and time.
Yet the context in which productivity is discussed today is very different . At the time the original manual was published, productivity growth across most advanced economies was relatively strong and dependable. Since the mid-2000s, however, the now well-documented productivity slowdown has raised questions about the validity of economic measurement, including the measurement of productivity.
Productivity measures are only as good as the underlying measures of input and output. If the economy is changing and measurement is not keeping pace, the resulting disconnect is often reflected in the productivity data. For instance, are we fully capturing the value generated by digital goods and services?
Despite 25 years of progress, productivity measurement can’t stand still; as the economy keeps changing, measurement must keep pace. Discussions at the conference highlighted that the measurement of many fundamental variables remains challenging in practice, reflecting the changing economic reality – labour quality may now be better reflected in skills than education; price measurement and quality adjustment are complicated by digital and intangibles; national estimates and international comparisons are obscured by the activities of multinational enterprises.
At the same time, the demands of users are evolving as the priorities of policymakers and populations change. A changing world order, a climate crisis, renewed use of industrial policy, and the emergence of AI all place renewed interest on aspects of productivity measurement. Decisions will be made with or without reliable productivity statistics, but the availability of reliable productivity data can inform the debate and the direction of economies. This makes the careful work of economic measurement more important than ever.
The discussions over the two days generated a long list of possible priorities for a future update of the OECD Productivity Manual. Many important topics have emerged since the original publication, and there is no shortage of ideas about what should be included. Below, I have outlined two areas I think deserve particular attention.
While there are many new conceptual issues that deserve attention, it is equally important to focus on areas where existing measurement challenges continue to affect the usefulness of productivity statistics.
Price measurement and quality adjustment are particularly important examples. Over the course of the workshop, participants repeatedly highlighted concerns about the measurement of prices and quality change, especially in business services. These issues can have substantial implications for measured productivity growth and, consequently, for the policy decisions that rely upon those statistics. A future update could usefully place greater emphasis on the practical consequences of mis-measurement and the benefits of addressing these challenges.
Over the past 25 years, productivity analysis using business microdata has expanded dramatically. These data have generated important new insights into productivity performance, innovation, competition and business behaviour. They have also helped researchers understand phenomena that are difficult to identify using aggregate statistics alone.
This is particularly evident in areas such as globalisation. Research using firm-level data has improved our understanding of multinational enterprises, profit shifting and international production networks. It has also helped support the development of more detailed measures of trade in value added and global supply chains, which are increasingly important for policymaking.
The growing importance of microdata also reinforces the need for continued investment in data infrastructure. Statistical agencies and government departments play a critical role in providing secure access to high-quality business datasets. Future guidance could usefully highlight both the opportunities these data create and the practical considerations involved in their use.
Looking further ahead, there are also opportunities to make greater use of new sources of information. Time-use data, for example, may become increasingly important in understanding how digitalisation is shifting activity between the market economy and household production. As economic activity evolves, our measurement frameworks will need to evolve alongside it.
The OECD Productivity Manual remains a landmark contribution to economic measurement. 25 years after its publication, many of its core principles continue to underpin productivity statistics around the world.
At the same time, the discussions in Paris highlighted that productivity measurement faces challenges that were only beginning to emerge when the manual was first written.
For ESCoE, the workshop was both an opportunity to celebrate the achievements of the past and to contribute to discussions about the future. The task ahead is not simply to update a manual, but to ensure that productivity statistics continue to provide an accurate, relevant and useful picture of economic performance in a rapidly changing world.
ESCoE blogs are published to further debate. Any views expressed are solely those of the author(s) and so cannot be taken to represent those of ESCoE, its partner institutions or the Office for National Statistics.