New book argues official statistics need to live up to their promise

Imperial academics lay out how national statistics offices can better engage with the public, businesses, and civil society in a new book.

Official statistics gathered in the UK cover everything from health and education to crime and commerce. But are these statistics living up to the promise that they should be indispensable in democratic society, helping us live better lives?

Professor Paul Allin and Professor David Hand, both from the Department of Mathematics at Imperial, have authored a new book to tackle this question, titled: From GDP to Sustainable Wellbeing: Changing Statistics or Changing Lives?

We spoke to the authors about the current state of official statistics, how they might be improved to better take into account principles like sustainable wellbeing, and how the COVID-19 pandemic has helped shift views on the topic.

What do we currently collect as ‘official statistics’?

David Hand (DH): Official statistics are statistics collected and published by the Office for National Statistics (ONS), government departments, and the devolved administrations. The statistics tell us about all aspects of life in Britain, from health, through education, commerce, crime, transport, and beyond. Many of the statistics appear regularly in the media: you may have recently read about GDP (gross domestic product), CPI (consumer price index), or excess deaths from COVID-19.

All these statistics are collected so that we can understand society, how it is functioning, where it is going, and how it can be best served. Not having the insight into society provided by those statistics would be like driving a car without looking from its windows.

Paul Allin (PA): In total, we found that last year there were 3,570 separate publications releasing new or updated UK official statistics. Since then, the ONS has added more than 150 publications specifically to help society understand the impact of COVID-19.

What function do they serve in society?

PA: They are to help underpin the design, implementation, and monitoring of policy. This could be concerning the labour market, immigration, or building homes and roads, for example.

DH: Official statistics are in fact aimed at “serving the public good”. That means that as well as informing government, they will also inform business and planners in all areas, as well as individual citizens. Official statistics help us make effective decisions and hold government to account. The Statistics Code of Practice says they “are part of the lifeblood of democratic debate”.

GDP is one of the statistics we regularly hear about. It is often compared between countries as a measure of success, but is this the best measure to use?

DH: It is important to recognise that GDP was designed as measure of economic activity. As such a measure, GDP has a clear definition as part of the UN system of national accounts. It means that economic progress can be studied and comparisons between countries can be made. However, GDP necessarily pulls together a vast amount of data from different sources, and there are different ways in which that might be done.

PA: Despite its precise definition, GDP is often taken as a measure of success and progress more generally. The criticism is that this paints too narrow a picture. It ignores, for example, adverse impacts of economic growth on environment, not only within the UK but elsewhere of course. It also hides inequalities in the distribution of income and wealth across the population. There are alternative measures, which we categorize as measures of sustainable wellbeing, that take these dimensions into account.

Has the COVID-19 pandemic changed what statistics we should focus on?

PA: We were completing the book earlier this year, as the pandemic was getting hold. There were already calls that governments, businesses, and us all as citizens, should widen our focus beyond just maintaining a viable economy. If we are to “build back better”, to reset and to adopt a new growth narrative we will clearly be in the territory of looking at wider measures of sustainable wellbeing.

It is still early days to say what this will entail and what will be the things that matter to people from now on. However, our analysis of the role of statistics is that there is much that official statisticians can do to reach out and engage with society, to help in plotting a course beyond the pandemic.

DH: The pandemic has provided a perfect illustration of the power of statistics. It has enabled us to understand how the disease spreads and impacts society, and also how effective different kinds of interventions were. The statistics have also brought into sharp focus the tension between health impact and economic impact.

How else might official statistics be improved to better serve society?

DH: Increasingly in the modern world, large and highly informative data sets are collected by non-official bodies – by corporations, for example. Combining the information from those data sets with official data collected by more conventional means will lead to greatly improved understanding of society, enabling better decisions to be made and benefitting all of us.

PA: We found this idea is already being explored in producing measures of progress towards the UN’s sustainable development goals for 2030, but we believe it points the way for official statistics as a whole. And there is an immediate need for such an approach, to come up with a more coherent presentation of COVID-19 statistics.

At present, for example, official figures (which are not official statistics) on cases found in testing are presented separately from official statistics on the incidence of COVID-19 detected in representative, random samples of the population. A more integrated presentation would explain how the number of cases depends in part on the number of tests carried out.

Can universities play a part in official statistics?

PA: Definitely! There is much that universities can do and are already doing. Here at Imperial, for example, future generations of statisticians are being trained, not only in advanced methodology but also covering the role of official statistics.

It is important that statistical skills are encouraged among all students. And our Imperial colleagues working on statistical modelling of COVID-19 and on the REACT-1 programme, tracking infection rates in the community, provide great examples of the kinds of non-official data that we propose should contribute to the expansion of official statistics into public statistics.

DH: I entirely agree with Paul about training. In addition, university research makes extensive use of official statistics, while also developing new methodological tools.