AER0065

Written evidence submitted by the Royal Statistical Society

 

Executive summary

The economic impact of the Covid-19 pandemic has been felt unevenly: potentially exacerbating existing inequalities and causing new inequalities to emerge. The RSS’s concern – as an organisation whose mission includes campaigning to promote the role of data in decision-making – is that existing data used to inform and assess the economic impact of policy decisions is not sufficient to allow an assessment of the characteristics of post-pandemic inequality.

We see a vital role for data and evidence in delivering a more equal recovery: it can enable a better understanding of the differential economic impact of the pandemic as well as providing a timely assessment of the impact of post-pandemic policy on incomes, education and skills, and wellbeing overall. We do not currently have the data infrastructure to do this post-pandemic inequality cannot be adequately understood solely using pre-pandemic data. We must rethink what data the government is collecting and how it is used.

There are three areas where we see particular challenges:

  1. Data sharing. The pandemic has shown the importance of being able to consider health data alongside other indicators – there are currently artificial barriers that prevent this and these need to be addressed, Central government also needs to do more to share data with regional and local authorities: this will give local and regionally-based policy-makers information that they can act on in a more timely manner.
  2. Collecting the right data. There have been societal changes as a result of the pandemic that we do not currently have data about– this prevents us understanding the extent of issues and being able to assess the impact of policy interventions to address them. The increase in homeworking and the impact of the pandemic on young people, ethnic minority groups and disabled people are all areas where more information is required.
  3. Reassessing economic measures. While reassessing the types of data that are needed to assess the post-pandemic recovery, there is also an opportunity to think more broadly about how we measure the recovery: namely, rather than thinking about the recovery using GDP as the single indicator of growth there is an opportunity to move towards wider measures of sustainable wellbeing using data from a diversity of sources.

There is work underway at the UK Statistics Authority (UKSA) to tackle some of these challenges. Our view is that this work needs to be reviewed in light of the new data that will be required to assess the particular impact of the pandemic and then accelerated. There are also areas where longer-term analysis is required. Our recommendations are:

Recommendation 1:              Introduce legislation to ensure that research access to deidentified unit record, survey and administrative data on health and all other data are governed by a single set of rules and standards.

Recommendation 2:              Encourage further data sharing between central government and regional and local authorities, particularly where controlled access is required, so that problems can be proactively identified and tackled at a local level.

Recommendation 3:               The UKSA – working with the Treasury – should conduct a rapid review of the data that is needed to assess how equal the recovery from the pandemic is and the Treasury should invest in the new data sources required. This should be part of a comprehensive data gaps exercise, which also considers a review of classifications.

Recommendation 4:              Invest in new longitudinal studies of the young people who have been affected most by the pandemic to assess how their education and transition to work has been impacted.

Recommendation 5:              In assessing the equality of the recovery identify measures to consider that go beyond purely economic measures such as GDP and incorporate wider measures of wellbeing.

1. Introduction

The economic impact of the Covid-19 pandemic – as the committee recognises – has been felt unevenly: potentially exacerbating existing inequalities and causing new inequalities to emerge. The RSS’s concern is that the existing data used to inform and assess the economic impact of policy decisions is not sufficient to allow an assessment of the characteristics of post-pandemic inequality.

There is a vital role for data and evidence in delivering a more equal recovery: it can enable a better understanding of the differential economic impact of the pandemic and help to provide a timely assessment of the impact of post-pandemic policy on the recovery of incomes, education and skills, and indeed on wellbeing overall. But for this to be possible, we must rethink what data the government is collecting and how it is used: post-pandemic inequality cannot be adequately understood solely using pre-pandemic data.

To illustrate the scale of this challenge, consider the various impacts of the pandemic on young people. There are various questions that we ought to be able to answer about the impact of the pandemic on young people that we are not currently in a position to: will the trend towards increased remote working provide a more difficult pathway into the labour market for future school and college leavers? Is there a difference in employment and/or educational prospects for those who have received their A-Level and GCSE qualifications in 2020 or 2021? What does home working mean for young people who live in areas with reduced internet connectivity?

We do not currently have sufficient statistical information in the system to answer these types of questions: either because there is a lack of data available on a topic or because data infrastructure is insufficiently joined up to allow us to answer them.

In this submission we identify three main ways in which we think that data infrastructure needs to adapt so that it can be effective in helping to promote an equal recovery:

        Improving data sharing – so that different data sets can be integrated and local decision-makers have access to the data they need.

        Collecting relevant data with the right frequency, including collating data from different sources  – ensuring it is published in a timely manner and with an analytical narrative.

        Reassessing the measures that we emphasise in the context of the recovery – looking at factors beyond GDP.

These changes are needed to ensure that the government has information about the groups that have been most affected by the pandemic. While it is possible to make educated judgements about where support is likely to be most effectively targeted without investing in data infrastructure, doing so would leave government unable to properly assess the impact of any support and to re-target it in a more effective manner. The level of investment in order to address these issues is relatively modest, but it would allow a step-change in how effectively funding could be targeted to promote an equal recovery.

2. Improving data sharing

The RSS’s data manifesto emphasises the importance of data sharing. Greater data sharing between government departments and between central and local government could greatly strengthen public services and drive an equal recovery. The public tends to be supportive of data sharing where it is done for the public good by organisations who are trusted to safeguard privacy and confidentiality. The recent concern about the sharing of GP data[1] indicates that the public is live to these issues, but from our perspective this means that it is all the more important to make the positive case for data sharing and ensure that people know how their data will be used and feel that they have the power to access their data and withdraw their consent to it being used in that way if they wish. It is important that the public are prepared to trust government bodies to use their data and many people are. However, there is a need for better education to explain the benefits of using data and to provide reassurance that the procedures and systems used are secure, so as to improve public trust in government.

There are three ways in which we suggest data sharing can be improved in a way that would help the government to assess how equal the post-pandemic recovery is: improved access to healthcare data, more effective sharing of central data with local government and better sharing between government departments.

The pandemic has shown that there is a need to consider health data alongside economic data – the government’s decisions around lockdowns have had to consider health metrics alongside economic considerations. However, for statisticians and researchers there is currently an artificial barrier between health statistics and other statistics – this is because they are governed by different pieces of legislation. Health and care data falls under the Health and Care Data Act 2008, while other statistics come under the Statistics and Registration Service Act 2007 and the Digital Economy Act 2017.

If it was possible to introduce a commonly used identifier for individuals across multiple data sets (such as a national insurance number) then this would have the potential to unlock a variety of questions for study: what is the impact on incomes of people who are affected by long Covid? How have disabled people’s employment opportunities changed post-pandemic? How long has it taken people to recover who lost their jobs due to the impact of the pandemic on their employers? How do increased waiting lists vary by area and between social groups? This would require legislation to ensure that research access to all data – health or any other data is governed by a single set of standards.

There is also a need for central government to get better at sharing data with local government this is particularly acute when considering income deprivation. This cannot usefully be measured at a regional or local authority level – it is quite common for very deprived areas to be next to very well-off areas within a local authority. The indices of multiple deprivation provide a starting point for understanding these dynamics, but tackling them is going to require a more considerable data sharing partnership between central government and local government. There is clearly a question of how we can ensure that the data is accessed in a way that protects respondents’ privacy, which requires people accessing the data to pass various qualification tests. Supporting the ONS to accredit external users of the data more rapidly and to provide more secure sites for accessing the data would be a helpful step in this regard.

Relatedly, there is a need to make data and information available in a timelier fashion. The Census of Population & Housing has taken over a year in the past to deliver full outputs with a knock-on consequence for statistics that use these data as an input. This includes, for example, the Index of Multiple Deprivation (IMD), which has the capacity to identify pockets of deprivation in otherwise well-off areas making this a useful resource for local decision makers.  However, the IMD can only be as good as the data it is built on and it will only reflect recent trends if those data are timely. We know that the Office for National Statistics (ONS) is keen to enable more rapid access to the Census data from the start of this year, which potentially provides a valuable resource to local authorities in understanding their areas. It is not just about having timely data, but also about using it and being clear which data are being used for policy decisions. For example, there is currently a debate about exactly what “levelling up” means. Data needs to be made available for those metrics selected by the government to design, implement, and monitor policy under this aim, as well as for other metrics, such as the IMD, which would generally be seen as helpful in understanding the selection and the impact of policy choices around geographical inequalities.

Improving data sharing between central government and regional and local authorities in this way, would increase the ability of regional and local bodies to proactively tackle challenges as they emerge.

The ONS is currently developing an integrated data platform. This would enable cross-government teams (and wider communities) to provide complex analyses by bringing together analysts, data, information governance and domain expertise within a safe, secure and trusted infrastructure. It would link data from across government departments and enhance decision making by providing a quality evidence base to promote the public good and drive an equal recovery. It is our understanding that during the pandemic the ONS has not always been able to easily access data from government departments. We would like to see this addressed by departments being incentivised to include their data in the new integrated platform so that data can be shared safely and securely across central and local government. With better knowledge the right policies can be developed and operationally the right actions can be taken at the right time.

Recommendation 1:              Introduce legislation to ensure that research access to deidentified unit record, survey and administrative data on health and all other data are governed by a single set of rules and standards.

Recommendation 2:              Encourage further data sharing between central government and regional and local authorities, particularly where controlled access is required, so that problems can be proactively tackled at a local level.

3. Collecting the right data

The pandemic has dramatically changed the way that people live and work. It is not yet clear how much of this will continue as the UK emerges from the pandemic, but some changes are likely to persist to some extent. The move towards homeworking in particular has wide ranging implications for how equally the recovery from the pandemic is distributed across the population and country. Further, the pandemic has had an especially adverse effect on young people – those in school, transitioning to further education or training and those joining the workforce. In both cases decision-makers need to have timely and detailed information about the impact of post-pandemic policy. We have also seen evidence of health inequalities in how the pandemic has affected different ethnic groups – the amount of time that it took for these inequalities to be clear in the data highlights the need to collect more comprehensive data around ethnicity – and it is important to understand how the recovery is felt by different groups. The sooner that more comprehensive data is available in these areas, the more effectively decisions can be guided by the evidence.

In this section we highlight some of the questions raised by these three aspects of the pandemic and set out the types of information that are needed to answer them. This is not intended as an exhaustive treatment – these are examples intended to demonstrate the potential benefits of investing in a data infrastructure which targets the policy questions emerging from the pandemic.

Homeworking

The increase in homeworking – if sustained even partially – potentially has a wide range of impacts and there are many policy questions that government may want to answer: how does the trend impact on productivity? What is the impact – both in terms of spendable income and wellbeing – on workers in different sectors? How is the change affecting town and city centres? What is the impact on people in areas with poor digital connectivity, such as rural areas? All of these questions bear on how equal the recovery is, but we currently do not have the data to adequately inform policy.

It would be useful – both for government and businesses – to understand how productivity changes in businesses and sectors where there is a sustained increase in home working. The productivity data that is currently available provides an estimate of output per hour across nine English regions, Scotland, Northern Ireland and Wales for broad aggregate industry categories. Due to the limited sample size of the Labour Force Survey, which is used to estimate hours worked in different sectors, it is not possible to develop more granular statistics that would have a detailed breakdown of sectors or assess the levels of homeworking among them.

There are a number of ways in which the ability to work from home impacts upon the cost of living. Compared to pre-pandemic, people who are able to work from home are likely to see: a reduction in transport costs as the frequency of their commute decreases; a reduction in spending on lunches and drinks during the working week; potentially a reduction in the cost of housing as they are able to move further away from their place of work (or an increase in wellbeing due to being able to afford a larger property more distant from their place of work). These changes will be complex, however, as there will also be some increases in costs: understanding the impact of working from home and how it varies will be complicated and vary based on where people live. It is important to be able to quantify this change – it is likely to impact people working in specific sectors in and around cities. It is equally important to understand the impact that this has on prices of transport and other amenities for those working in sectors where homeworking is not possible: there is a risk that people in lower-paid sectors and poorer regions may see an increase in cost of living while others see a decrease. Similarly, and related to the points we make in section 3, there are likely to be differing impacts on wellbeing. We do not currently have good enough data to assess these types of changes and how they affect how equal the recovery from the pandemic is.

Understanding how office space utilisation has changed – and how that impacts urban centres in different parts of the UK – is going to be complex. It may involve, eg, collecting and bringing information about the average number of desks per employee, number of people travelling into urban centres, changes in spending habits in shops, cafes and restaurants. This type of data will be important for understanding how town centres can adapt successfully to the increase in homeworking. Relatedly there should be serious analysis of home working spaces, and the inequalities in access and need across regions, sectors and demographics, which will be affected by location.

A shift towards homeworking also poses questions for those who live in areas with poor connectivity: slow internet connections may start to impact how effectively people can perform their jobs. To allow us to assess this productivity data would need to be more granular so that it could be correlated with information around connectivity. More granular detail would also be needed relating to changes in employment levels and income, so that it would be possible to track any impact of changes in productivity levels. This will be a problem primarily in rural areas, where other challenges are likely. There is already an indication that more people are moving to rural areas and pushing up house prices, potentially causing alterations in measures like the indices of deprivation. This might, if analysed bluntly, be viewed as a reduction in inequality, but it may be artificial as more people moving into rural areas makes it harder for local people to afford housing.

Relatedly, it is also worth considering the breadth of information sources that are available concerning employment generally. In particular there is a lack of information directly from employers and concerning the self-employed. It would be beneficial to establish a survey of employers that could effectively shed light on the challenges that they are facing and how they are intending to respond to them. This should go beyond another round of the Skills and Employment Survey.

Young people

The shift towards homeworking may also bear directly on young people, who – as they enter the workforce – need to develop experience in order to advance their careers. In the past this process has been driven in part by informal interactions with senior colleagues – if these are reduced it may mean that young people are slower to progress through the workforce, especially compared to their colleagues who did gain a grounding in offices prior to the pandemic. There is also a question of the mental health impact as young people are unlikely to be able to afford accommodation with a separate office space – making it hard to separate their work and personal lives.

The government recognises that there are a host of issues affecting people who have attended school during the pandemic – allocating £1.4 billion to help this group to catch-up. The effectiveness of this investment will need to be monitored and the data needed to effectively do so is not collected. Due to the change in how pupils were assessed in 2020 – and again in 2021 – it will be important to have granular information about employment for young people of each age – ie, that is capable of assessing the difference in underemployment levels for a nineteen-year-old compared to an eighteen-year-old, and that is capable of measuring whether any disadvantages persist into future work. This will be important so that the exclusion such as identified for the NEETs can be identified and the measures introduced to address them.

There is not currently sufficient statistical information in the system to address these types of issues. At present we rely on tried and tested survey methods which are not detailed enough to show up these types of trends and are not produced quickly enough to enable policy to be able to respond to changes in a timely manner. It is not just these areas where there are issues -- as set out in the Royal Statistical Society’s response to the UKSA Inclusive Data Taskforce[2], the system importantly lacks relevant data to assess the varying impact of the pandemic on ethnic and faith communities, and those with disabilities. In addition, multiple job holding has now become very commonplace across the population and we are unlikely to be reflecting increased levels of volunteering in our current economic estimates. If we are to be able to gauge how equal the recovery from the pandemic is, we need to urgently review the type of information that is needed and scale up data collection. In addition, the statistical system could benefit by drawing on a wider range of sources, beyond its surveys and censuses and the administrative data drawn from departments. ONS’s Data Science Campus is providing some promising developments, for example, which illustrate that useful data can be mined from new sources.

The challenges faced by young people are especially important to understand and are so wide-ranging that they can only be addressed through longitudinal studies of the relevant cohorts. The sample size for these studies would need to be substantial to provide the granularity of detail that is required to properly assess these questions and ensure that the policy response can be adjusted so that the people who have been at school, university or in their first years of employment during the pandemic are not permanently affected.

Inclusive data

The Committee is interested in how the pandemic has impacted different groups – especially different ethnicities and disabled people. While there is information and evidence available to suggest that some groups will be more affected by the pandemic and its economic effects than others, we do not believe that the data available here is detailed enough to enable a detailed understanding of how different groups have been – and continue to be – impacted.

We believe that a comprehensive "Gaps" exercise is needed, to assess the statistical requirements to support all strands of the government's equalities policies, covering all the protected characteristics, as well as identifying areas where more data is neededeg around the impact of homeworking – to assess the equality of the recovery from the pandemic. There are also other minority groups about which there is a need for official statistics, such as the homeless and those living in deprived areas.

There are a few areas where we think there is a pressing need for change:

  1. A review of ethnicity classifications: The classification of ethnicity in official statistics (as used in the Census) has been largely unchanged since the 1980s and needs to be completely reviewed. It is based partly on colour (Black, White, Mixed) and partly on geography (Indian, Pakistani, Bangladeshi, African, Caribbean). Many significantly different ethnic communities are hidden within these categories and are "invisible", for example Jews and Sikhs. In particular, the Indian and Black African categories are so broad that they include many communities which should be, but are not, counted separately. We propose that a “blank sheet” approach should be taken to look at how ethnicity should be classified now, looking ahead rather than to the past. In order to identify some significant communities, it may be necessary to use a combination of ethnicity and religion in such a classification.
  2. Collecting data by religion: Religion has been a protected characteristic since the Equalities Act 2010, and public sector organisations should be mandated to collect data on religion, in the same way as they do for other characteristics such as age, sex, disability, ethnicity, etc. Many organisations currently do not ask respondents for their religion. In some cases, where religion is collected for operational reasons, it is then not retained in digital systems for further analysis.
  3. Disability: Statistics on disability need to be improved in terms of detail and timeliness. A few changes would be especially helpful: disaggregation by other protected characteristics is important, especially by ethnicity and religion; the Family Resources Survey provides timely and robust data on disability, impairment and receipt of benefits, but does not cover wider aspects of disabled persons’ lives; we would like to see a framework for identifying and reporting on different types of disability needs to be developed, for example breaking down to Physical, Sensory (deaf, blind), Mental (mental and learning impairments) and Emotional (eg autism) disabilities. It would also be beneficial to restart the Life Opportunities Survey, which ended in 2014, and provided comprehensive statistics for disabled people on work, education, social participation, transport, and use of public services.

Recommendation 3:               The UKSA – working with the Treasury and other relevant bodies – should conduct a rapid review of the data that is needed to assess how equal the recovery from the pandemic is for various population groups and the Treasury should invest in the new data sources required. This should be part of a comprehensive data gaps exercise, which also considers a review of classifications.

Recommendation 4:              Invest in new longitudinal studies of the young people who have been affected most by the pandemic to assess how their education and transition to work has been impacted.

4. Reassessing economic measures

While reassessing the types of data that are needed to assess the post-pandemic recovery, there is also an opportunity to think more broadly about how we measure the recovery: namely, rather than thinking about the recovery using GDP as the single indicator of growth there is an opportunity to move towards wider measures of sustainable wellbeing using data from a diversity of sources.

GDP is indispensable to economists. However, recent UK research shows that public understanding about GDP is very limited. GDP may be in the news, but people see it as an example of economic jargon, inaccessible and not important to their everyday lives.[3] A post-pandemic recovery that is judged to be equal, for example, on the basis of regional GDP alone will not necessarily be a recovery that is felt equally within or between regions.

Further, and despite its precise definition, GDP is often taken as a measure of success and progress more generally, which paints too narrow a picture. If we consider GDP alone, we potentially miss adverse impacts of economic growth on the environment and on people’s wellbeing generally. GDP as a measure also hides inequalities in the distribution of income and wealth across the population. While there is clearly a case for using GDP when thinking about the recovery, our view is that it should only be a part of what is taken into account and that wider measures of sustainable wellbeing also have a role to play.

This might mean including measures such as:

        People’s health

        Personal finances and levels of disposable income and of wealth

        Affordability of housing

        Available recreational facilities

        Total volunteer contribution to the economy

        Whether people feel like the place they live is thriving

        Concerns around climate change and the environment, eg air pollution

        Measures of subjective wellbeing

These types of measures are not easy to develop and they are not all currently in a perfect state. However, we believe that public statistics need to feature even more in debate and decision-making, so that we are not constrained to considering progress – or other aspects of our societies – simply in terms of the indicators we have. We need to think about the measures that we want to have. As Gus O'Donnell, former UK cabinet secretary, puts it: “Of course, measurement is hard, but roughly measuring the right concepts is a better way to make policy choices than using more precise measures of the wrong concepts”.[4] Delivering a major change in direction and helping us all go beyond GDP and other established measures will be challenging for official statistics systems – but it is essential.

Recommendation 5:              In assessing the equality of the recovery identify measures to consider those that go beyond purely economic measures such as GDP and incorporate wider measures of wellbeing.

 

July 2021

9

 


[1] https://www.bbc.co.uk/news/technology-57555013

[2] https://rss.org.uk/RSS/media/File-library/Policy/2021/RSS-Response-to-IDTF-Consultation-FINAL.docx?ext=.docx

[3] Economic Statistics Centre of Excellence, Public Understanding of Economics and Economic Statistics, November 2020

[4] https://www.ft.com/content/e3b356b4-dbcc-42ef-811d-74d649139916