Reflections on the forthcoming System of National Accounts revision

Thank you and good morning everyone.

On behalf of the UK Statistics Authority, it is a great pleasure to welcome you all to London for what I am sure are going to be a fascinating few days. As well as listening to – and engaging with – the excellent range of presenters and panellists on the conference agenda, I hope that you will have time to catch up with old friends, make some new ones and see something of the city. Kings is a great base from which to do so.

For those of you who don’t know, the Authority is the umbrella body that oversees the governance of the UK statistical system, spanning the Office for National Statistics (our NSI), the Government Statistical Service and the Office for Statistics Regulation, which is responsible for setting, promoting and policing our code of practice for official statistics.

Before I begin my remarks, let me thank Rebecca Riley and her team at ESCOE, who have organised this event on behalf of the ONS as this year’s NSI sponsor. Rebecca and her colleagues have been working closely with Andrew Sharpe, Catherine Van Rompaey and the IARIW team to ensure that it will be a great success.

As everyone here knows, this is a big year for economic measurement. In June the complete draft of the proposed new revision of the System of National Accounts was released for consultation. Stakeholders have until the 9th of September to make their views known and, given that the draft weighs in at a thousand pages, if you haven’t started yet you probably should do soon – having taken inspiration I hope from the conversations here this week. Once comments have been taken on board, the new SNA should be signed off at the UN Statistical Commission in February.

Now, I am not going to give you a blow-by-blow account of the many changes in the new revision – you don’t have the time and I certainly don’t have the expertise. As Chair of the Authority, I was appointed to bring a user perspective to the governance of the UK statistical system as a former journalist, think tank director, IMF official, economic forecaster and public finance watchdog. With Ian speaking before me and Richard Heys on the panel to follow, I am very much the non-technical filling in the Authority’s rhetorical sandwich this morning and they are the ones to pick up any difficult questions.

That said, I am delighted that my more knowledgeable Authority colleagues have made a significant contribution to the proposed changes. The UK member of the editorial team, Sanjiv Mahajan, is here this morning. He has laboured long and hard with the rest of the team, led by Peters Harper and van de Ven, and is well-placed to help with any questions you may have on the process or outcomes. Do try to grab him during one of the coffee breaks.

The SNA is a common framework and designed to facilitate international comparisons. But NSIs will still have important decisions to take on whether and how its rules should be applied. To help do that in the UK, Ian as National Statistician can now draw on the wisdom of a dedicated Committee for Advice on Standards for Economic Statistics, chaired by Martin Weale. NS-CASE is already playing a valuable role in ensuring that the international standards we apply and the way we do so appropriately reflect the characteristics of the UK economy.

As the SNA revisions near completion, it is easy to think the heavy lifting is over. But challenging and time-consuming as the conceptual and methodological work has been, there is a new phase of hard graft to come. Not just implementation but disseminating and sharing the changes with user communities and helping them understand what the outputs do and don’t mean and where they should or shouldn’t be used in favour of alternative data. This will be important for finance ministries, central banks, academics, the media, civil society and citizens.

Communicating and winning confidence for this iteration of the SNA may be more challenging than for its predecessors – in part because methodologists and the more engaged users of National Accounts had differing views about what it should be aiming for. Some were largely comfortable with the current approach, which they saw as broadly fit for purpose, well understood and by-and-large trusted. Others looked at the modern economy and modern society and saw a range of important issues where there was a need for new concepts, new definitions and new data to help decision-makers tackle new issues and to improve broader public awareness and understanding.

Over the last 10 years, the editorial team and the wider ecosystem of task-teams, experts and consultation respondents have forged a compromise between these views. In the end, most practitioners and users in the UK National Accounts community would probably have favoured a more ambitious set of revisions, but everyone recognises that a balance has had to be struck. In part that balancing act reflects the fact that some conceptually desirable reforms involve wider use of imputation and judgement, creating greater scope for less scrupulous national authorities to massage statistical outputs in convenient directions. And that understandably weighs more heavily as a consideration in some countries and with some institutions than others.

Nonetheless the version before us delivers valuable improvements to the National Accounts to reflect the impact of globalisation and digitalisation, such as accounting for data as an asset (although deciding not to do the same for brands) and giving a clearer representation of that portion of natural capital which is already captured in the National Accounts.

Alongside these improvements, the UK’s move from the European System of Accounts to direct adherence to the SNA gives greater flexibility here to improve measurement of public services by allowing quality adjustment of output metrics. The Public Service Productivity Review, which Ian is leading, is developing potential improvements to the way we measure health, education and other services in ways that policymakers and others could use and understand.

At the most recent meeting of National Statisticians in Geneva, there was broad agreement on the need to think through input and output measurement in public services, and I am proud that Ian and colleagues are blazing a trail. Let me thank Debra Prestwood and Richard Heys and their teams and partners across the Government Statistical Service for their contributions. You will hear more about their work later this week and I would encourage anyone interested in the quality of public services measurement to talk to them and get involved.

From the maximalist pro-reform perspective, the new SNA has shortcomings. It does not make radical changes to the core National Accounts framework, such as integrating wider definitions of natural and human capital onto the balance sheet, tackling atmospheric degradation or proposing formal methods to integrate wellbeing into new aggregates and metrics.

Path dependence probably plays a role here. The earliest work on the National Accounts began by focusing on flows rather than stocks. Which is understandable given the circumstances of the time: many countries in the late 1940s had seen their physical capital stocks devastated by war; investment in human capital was relatively modest, with fewer than 5 per cent attending university; and natural capital was in great abundance so the marginal cost of depleting these assets was thought to be close to zero. But this mindset has held for the past 75 years, even as the economy, society, the environment and the focus of policy has evolved.

We are all aware of Thomas Piketty’s work showcasing the importance and impact of capital in the modern economy. The new SNA makes some important changes here – I mentioned the recognition of data as an asset – but this is an area where there are big questions to answer and the failure to bring human and natural capital more fully into the framework will be seen by some as a lost opportunity.

We can see this from the relative valuation of different forms of capital that emerges from the current version of the SNA – including its satellite accounts and the System of Environmental-Economic Accounting.

Here in the UK, the National Accounts capture £11 trillion of capital used in economic production. Natural capital is increasingly scarce and society places growing value on ecosystems, biodiversity and natural resources. But the economic exchange values, or imputed equivalents, that we use to value the UK’s natural assets give it a value of only £1.7 trillion. Can it really be right that our headline estimate of the value of the environment in this country is only one sixth that of its machines, roads and buildings, when the former would be impossible to replace?

One issue is that the £11 trillion figure for productive capital sweeps up £6 trillion of land and undoubtedly some of the value of ecosystems too. But the methods we use to calculate the value of these ecosystems and to impute market equivalent prices may also contribute to an undervaluation of natural capital. As the new SNA revision attempts to capture environmental depletion within Net Domestic Product, the impact may become quickly apparent. Those arguing for or expecting more decisive changes might have expected estimates of economic growth to be reduced significantly as they are adjusted for the depreciation of natural assets through environmental degradation. But any impact here looks likely to be relatively modest.  For some vocal users, this will be seen not merely as an everyday concern about the accuracy of the figures but a failure to address a fundamental unfitness for purpose.

The treatment of natural capital will inevitably be a much-debated area of revision. For example, should the measured stock of natural assets be capable of being enhanced, for example by turning a brownfield site into a park or discovering some new mineral reserves? And should such enhancements in natural capital count towards GDP growth like other forms of investment? What value should we place on new natural capital, and how will we source data on both depletion and acquisitions?

The ONS has participated in international discussion on this but has yet to take a firm view. Conceptually, allowing natural capital to be enhanced would seem to make good sense. But the challenge is in sourcing data and prices. If the end result is a set of data which tells policymakers the environment is worth very little, policy decisions may reflect this. So bad implementation could be worse than no implementation.

Turning from natural to human capital, the ONS’s headline estimate of the value of the latter is currently £25 trillion. So we value humanity roughly 14 times more than we value the environment. Intuitively you would expect the valuations to be closer. Colleagues at the ONS will present a paper in this area later in the week.

Whatever the precise valuation, human capital should be a major focus when thinking about the modern economy. As one of my colleagues argues:

‘trying to understand the UK economy without taking account of human capital is like trying to understand the Himalayas without taking account of the mountains’.

And yet, under the new revision, the SNA is not going to incorporate human capital into the core sequence of economic accounts. Within the current framework that is entirely logical: who has economic ownership of human capital, for example? But this means the National Accounts cannot help us consider key policy questions in an integrated way: why does £100m of spending on artificial intelligence count as investment when £100m of spending on actual intelligence does not? And when we measure the economic output of the education system as the number of children educated, rather than considering say the rate of return on the qualifications they earn, we are missing something. Which is not to say of course that we should only consider the economic return to education.

If Kuznets, Stone and Meade were in this room today, and had the ability to start work on the SNA afresh, they might well start by producing a coherent conceptual framework capturing all forms of capital first, and then build the rest of the accounts around it. Flows would remain a key part of the story, but we and users would not be focused as much on GDP – a partial flow measure that pays insufficient attention to the depreciation of these assets or to investments that might augment them.

For all we like or dislike about the upcoming revisions, they confront NSIs with two engagement and communication challenges: to explain the substantial changes which have been agreed to our users and how they affect and augment the figures they are familiar with, but also to explain why some of these changes have not gone further.

Notwithstanding its limitations, we will need to introduce users to the concept of Net Domestic Product, which does take account of the depreciation of natural capital, and encourage them to make use of it. But we will still be producing Gross Domestic Product, with some changes but still fundamentally the aggregate that central banks, finance ministries and economic commentators are familiar with. And when their growth rates diverge significantly, that will need to be explained. We are a long way from being able to provide concrete details of the likely impact, but we do need to start to calibrate expectations: the GDP numbers may not move as users might expect or for the reasons they may presume. Changes to public sector output measures are potentially going to have a greater impact on the growth rate of UK GDP than the the proposed changes to the way we treat the environment, for instance.

How we manage this in the UK is at the heart of our thinking at ONS and the UKSA. We have to balance being ‘Fit for Purpose now’ and ‘Fit for the Future’. In today’s constrained financial climate we have to prioritise our core deliverables: GDP, employment, prices, population, migration and health data. But it is increasingly clear that new areas of statistics are going to be increasingly in demand.

Well-being, the environment, crime, sub-national data, household consumption, income and wealth, unpaid household work, labour accounts and human capital are all areas which need new or enhanced data that will lie outside the SNA, or at least the sequence of economic accounts. We need to continue to develop these complements to GDP to keep pace with changing user needs. Being able to place these data alongside those generated under the SNA such that valid and consistent comparisons can be made is clearly going to require careful thought and methodological development, as well as resources.

It is important to place this in a wider context of statistical demand and development. From the very top, the international direction of travel is clear: the data we produce are not enough. The UN Secretary-General’s ‘Our Common Agenda’ document called in 2022 for new ‘complements to GDP’ to inform decision-making. September’s UN Summit for the Future in New York will discuss the need for new measures at the highest political levels. The European Commission is considering sustainability and wellbeing metrics to support a new generation of continental policymaking. The US is investing more in Natural Capital Accounts. And our new Prime Minister has stated that

“with every pound spent on your behalf, we would expect the Treasury to weigh not just its effect on actual income but also its effect on wellbeing”.

The mood-music is clear: key users want change and the changes proposed in the new revision of the SNA will not yet fully meet their needs.

In this context we cannot see the System of National Accounts as the answer to all questions. High quality economic data will always be required, but they are not the only data that are required, and the SNA should not be expected to serve as the single source for well-being, sustainability, environmental or social data, particularly as so much good work has already been undertaken beyond it.

In the 1993 and 2008 versions of the SNA, topics of new interest were increasingly brought into partial scope through satellite accounts, but the 2025 revision and the limitations it has to live with show that we have run out of road in making the SNA carry the whole load. In the environmental space, the SEEA makes a bold start to share the burden, and under the Friends of the Chair of Social and Demographic Statistics, social statisticians are also discussing how to provide the complements the international community and domestic policy-makers are seeking.

Of course, that does not mean starting from scratch. In the UK, there has been great work developing Natural Capital Accounts, we have published a lot of insightful material on wellbeing and we have created new methods for calculating Inclusive Income, aligned with the concept of Inclusive Wealth, as recommended by the Dasgupta review. We need to work with academics and colleagues from other NSIs to develop a ‘coalition of the willing’ to move beyond the existing and forthcoming SNA frameworks. As those new concepts, methods and statistics become more widespread and accepted, we can revisit the SNA framework and look to extend that work to other NSIs.

I hope in your discussions over the coming week you will not only celebrate and debate the 2025 SNA but also ponder what we will need above and beyond it. I hope your sessions this week will bring innovation and bold ambition to these challenges, and I am delighted to commend the conference agenda to you. Thank you.