HoC 85mm(Green).tif


Women and Equalities Committee

Oral evidence: Unequal Impact: Coronavirus (Covid-19) and the Impact on People with Protected Characteristics, HC 276

Wednesday 6 May 2020

Ordered by the House of Commons to be published on 6 May 2020.

Watch the meeting

Members present: Caroline Nokes (Chair); Nickie Aiken; Sara Britcliffe; Angela Crawley; Alex Davies-Jones; Peter Gibson; Kim Johnson; Kate Osborne.

Questions 81117


I: Professor Sir Michael Marmot, Director, UCL Institute of Health Equity; Liz McKeown, Director of Public Policy Analysis, Office for National Statistics; Professor Imran Rasul, Professor of Economics, University College London, and Director, Institute of Fiscal Studies.


Examination of witnesses

Witnesses: Professor Sir Michael Marmot, Liz McKeown and Professor Imran Rasul.


Q81            Chair: Thank you to the witnesses for being here today. I am going to give you all the opportunity to give a brief introduction as to who you are and how you can help us with this inquiry. We know that, on Monday, the Health Secretary indicated to us the review that is being carried out by Public Health England into the data and what we know of how Covid-19 is impacting different communities. I would be very interested to know whether you will be feeding into that review. If not, how would you feed into it if you were invited to?

Professor Sir Michael Marmot: Thank you for inviting me to give evidence to this Committee. As I hope you know, I was responsible for the publication—"Health Equity in England: The Marmot Review 10 Years Onit feels like a different geological era, but it was 25 February this year. We were concerned with inequalities in health. Most of our work was inequalities in terms of thinking of socioeconomic inequalities. We referred to ethnic differences where we had data. We pointed out that the data are not routinely collected that way, so there is a problem. We also were concerned, obviously, with gender, but there is much less data on sexual orientation and people with disabilities. These are all important questions.

It is highly likely that much of what we covered applies to groups with special characteristics of concern to this Committee, but, because of limitations of the evidence, we did not say so very much about them. Where we did refer to ethnic differences in life expectancy, we pointed to the lack of uniformity by ethnic group. It depends which ethnic group the people belong to. In general, the social determinants of health and health inequities apply to these groups of special interest.

My general view as to what has happened since is that the pandemic has exposed and amplified the underlying inequalities in society. The response to the pandemicthe lockdownlikewise will have the effect of increasing inequalities. We can talk about that more.

Professor Rasul: Let me start by thanking you for inviting me here today to give evidence. I am a professor of economics at University College London. I am a research co-director of the research centre within the Institute for Fiscal Studies. Most recently, I have been involved in work looking at the short-run impacts of the current crisis on a range of different dimensions, looking at how key workers are being impacted and at differences across different minority groups, as well as some more recent work thinking about differences in the geographical impacts of the crisis in the UK along different dimensions.

My expertise comes from the use of administrative data to document what was happening in the run-up to the crisis. As Michael mentioned, there has been a lot of work done at IFS looking at the drivers of inequality in the UK, how some dimensions of inequality have increased in the run-up to the crisis and how the current situation has amplified some of those differences but also made us more aware of new dimensions of inequality that may well be relevant for policy and the dynamic impacts of the crisis going forward. Those would be the types of things that I would very much like to feed in, both today and if I were invited to give evidence to the wider review.

As a bit of background, I have done some previous work looking at other epidemics and pandemics. In particular, I have work studying the impacts of the Zika crisis in Brazil, as well as the impacts of Ebola, and trying to understand how those shocks impacted those societies in terms of health responses, how people responded to information about those outbreaks, and what the short-term and long-term economic impacts of those crises were. This is following a broader agenda that economists have been interested in, trying to understand how robust our economies are and what the impacts are on different groups from these very aggregate, sudden and uncertain types of shock.

Q82            Chair: When looking at the economic impact on different groups after previous crises, did you specifically break that down into gender and people with protected characteristics, such as the disabled?

Professor Rasul: In the context of the Zika crisis, the most at-risk group were pregnant women. We were able to use administrative records from Brazil to try to understand, if there is a particular group at risk, how they change their behaviour with the supply of new informationhow they change their likelihood to become pregnant and their behaviour during pregnancy. That was very much focused on a particular group that was at risk. We documented how the responses to information that was provided by the Brazilian Government at the time varied across the socioeconomic status, marital status, age of women, educational background and so on and so forth. That gives us one context: how do households respond differently to the same types of information? How long does it take them to get back to the types of behaviour that we observed pre-crisis?

In the context of the Ebola crisis, I was working in Sierra Leone. We were understanding how adolescent girls and young women, aged between 12 and 24, were impacted by the crisis, and particularly how school closures during the crisis may have led to persistent impacts on those girls’ ability to acquire schooling, even after schools had reopened. There are many elements of these crises where short-term policy responses can feed into more persistent impacts. That might be one of the issues we will touch on a little bit later. That is what we studied in that context.

Liz McKeown: I am the director of public policy analysis at the Office for National Statistics. At ONS, our role is very much about providing the data, statistics and analysis needed to serve the public good. In the current Covid-19 pandemic, that means making sure that both the public and policy makers are informed about what is happening across the UK economy and society and giving the information and evidence needed so that we can manage the pandemic in the best possible way. A key part of that is understanding how different groups’ experiences are varying over this period.

We have done a lot of work on the current data and statistics that we have, to understand how that might happen, and then moving on to think about how we can fill evidence gaps that are likely to happen. We have done a lot in a short period of time to stand up some new surveys to give new evidence and to bring in different data sources that can help shed light on some of these issues. Hopefully, there will be a chance to talk about some of those developments over the next couple of hours.

Q83            Nicola Richards: Professor Marmot, your 10-year review was published earlier this year, as you have already mentioned. In it, you said the Government had not prioritised health inequalities, and noted there had been no national health inequality strategy since 2010. What impact, if any, has this had on the Government response to Covid-19?

Professor Sir Michael Marmot: I would imagine it has had very little impact so far, because of the urgency of responding to the pandemic. I was encouraged that after 25 February, the very next day, there was a question to the Prime Minister at Prime Minister's questions, and the Prime Minister responded very positively. My key recommendation was that the Prime Minister should lead a cross-Government health inequality strategy, to fulfil his stated ambition of levelling-up. We had given evidence on the areas where investment was needed.

There was a debate the following week in Parliament—I think it was a three-and-a-half-hour debateabout my report. As I read Hansard, all sides seemed to accept the importance of the findings and that action needed to happen, so I was pretty encouraged that we might get somewhere. Then, of course, the pandemic swept over us, so I do not think too much has happened along those lines directly.

That said, what the Chancellor did, in trying to support the salaries of employed people who were furloughed, is a form of universal basic income. I was far too shy to recommend universal basic income, but the Chancellor just did it. Trying to support the social and economic conditions of the worst off seems to have been part of Government strategy. I do not think it was an explicit response to my report. I would hope, as we start to emerge from lockdown and from the pandemic, that my report will indeed become part of the discussion about what kind of society we want to create.

Q84            Nicola Richards: Are you concerned overall that Covid-19 will entrench existing health inequalities even further?

Professor Sir Michael Marmot: I am terribly concerned. We can see it already. The Food Foundation reported on Monday that, of families with children, 5 million adults and 2 million children had experienced food insecurity since the start of the lockdown. Wow! People do not have enough to eat. Most of us agree that a fairly basic measure of a civilised society is that everybody can eat. In public health, we are concerned about the quality of food, of course, but if people cannot eat at all, are actually going without food or are having smaller meals because they cannot afford it or lack access, this will entrench and make worse existing inequalities. Once the furlough scheme starts to tail offtoday’s news was that the Chancellor is considering how to tail it off—that will again be likely to entrench inequalities.

Imran does not need me to quote the IFS, but I will anyway. Actually, I will quote ONS; Imran can quote IFS. ONS reported that before lockdown, if you look at workers in the hospitality industry, 10% said they had the ability to work from home. Some 53% of people in communication and the like said they could work from home. We know that, the lower your income, the less likelihood there is of working from home. Those people will have to go out to work, which puts them at risk, or become unemployed. If their salaries are supported to 80% but they were at the poverty line, they are now below the poverty line. There are going to be severe difficulties of exaggerating social and economic inequalities. That will not be solved overnight.

Q85            Alex Davies-Jones: My initial questions are for both Professor Marmot and Professor Rasul. We are getting evidence of the disproportionate number of black, Asian and minority ethnic people, as well as men, dying from this virus. It is still emerging. At this point, is this because of the nature of the virus itself or wider social factors? What is your interpretation?

Professor Sir Michael Marmot: There was a paper published yesterday from Oxford University, which I feel I could have written. It said exactly what I thought about the issue, which was two things. When you take the black, Asian and minority ethnic groups overall, the high risk could largely be explained by social disadvantage. On average, these groups tend to be poorer, have more crowded households and multi-generational households with greater likelihood of transmission within that household, lower-status jobs and the things I have just been talking about, such as having to go out to work.

Happily, in the UK we do not think it is a health service issue. If we are talking about the United States, we would be saying that black residents are less likely to get adequate healthcare, but we are blessed with a national health service with equity of access, so we do not think it is lack of access to healthcare. We do not think that is the issue. In fact, this paper published yesterday shows that the proportion of BAME patients in intensive care is higher than the average in the population. In general, we do not have to think about the nature of the virus; we can probably explain most of the excess on social determinants of health.

But what about the doctors? I cannot get my head around that. I do not know. Something like 40% of the doctors in the NHS are from BAME backgrounds, as are 95% of the deaths. I would not imagine doctors as suffering from social disadvantage, poverty or overcrowding; they may be multi-generational households. I do not know why the doctors seem to be at high risk. It is terribly important, and I do not think it is readily explicable on the basis of the kind of things that I was talking about. Maybe we will understand more about this terrible disease if we can explain that.

Professor Rasul: Let me summarise some recent data work that IFS has done particularly on this issue and highlight some of the wider implications of that, thinking about the extent to which we can attribute excess deaths to demographic factors and pre-existing health differences vis-à-vis those that might come from exposure, say through the workforce or labour market.

The first key point is that there are large degrees of heterogeneity across different minority groups. It is not useful to think of all of them as one homogeneous category. The evidence that we have put together suggests that there are some groups that have excess deaths and others that do not. The underlying reasons for this may be very different. In raw terms, the black Caribbean population has a death rate that is about double that of the white British population, whereas Bangladeshis are slightly below that of the white British population. There is a tremendous amount of heterogeneity, looking at the raw numbers.

The second thing I wanted to emphasise was that there are many different outcomes we could consider here. One is death rates as caused by Covid. Another potential outcome that I think has been less studied is how excess deaths—namely the total number of deaths irrespective of cause, relative to some long-run average at this point of the yearbreak down across different minority groups. A fraction of them will be directly due to Covid, but an additional amountperhaps even the majoritywill be caused by other factors related to the re-allocation of resources around the heath service. That may also disproportionately affect minority groups. A lot of the discussion so far has been in terms of differences in Covid-related deaths, but there may be other elements of deaths that become increasingly important, in both the short term and the longer term.

The analysis shows that, when you try to condition and take account of basic demographic factors and the fact that minority populations are typically concentrated in urban areas, where there have been higher caseloads and death rates, and at the same time take account of differences in age structure, you find the raw differences in death rates actually get extenuated even further. You might think the difference was down to geography, and that narrows the gap a little bit, but the fact is that most minority populations have demographic age structures such that it becomes even harder to explain some of these differences across groups. Basic demographics do not seem to do it. They seem to extenuate some of these differences.

There are a whole range of other risk factors that Michael has already touched upon. Families and household structures differ tremendously across different groups. For example, south Asian groups are much more likely to live in larger households. To give you an indication of the extent of those differences, if we just focus on London, around a third of households are single-person households. Among households where the household head is Bangladeshi, Indian or Pakistani, those figures are 11%, 17% and 13%. There is a wide difference in those household structures.

In addition, compared to white British households, minority ethnic groups also tend to live in more overcrowded accommodation, even once you control for a particular region of residence. For example, fewer than 2% of white British households in London have more residents than rooms. In contrast, this figure is around about 30% for Bangladeshi households, 18% for Pakistani households and 16% for black African households. Those types of household structures and density are likely to make self-isolation harder and increase the opportunities for within-household transmission. That probably plays an important role. However, such overcrowding is not so prevalent for black Caribbeans, which is the group that has the most excess deaths relative to the white population.

The obvious remaining factor is what Michael was suggesting, which is pre-crisis differences in health conditions among these groups. You see a very striking pattern there. For many minority groups, their precondition health status tends to be worse, on a whole range of different measures that we might think make people particularly susceptible to Covid, but there is a very strong age gradient there. When we talk about younger people in minority groups, they actually have lower rates of pre-existing conditions than the white British population. Where minorities have worse pre-existing health conditions is among the older members in those populations. For Indians, Pakistanis, Bangladeshis and black Caribbeans, pre-existing health conditions among older members of those communities are much more prevalent than among the white British population. You can see how those two things might interact to be driving some of the differences that we see.

Coming back to my earlier point about distinguishing direct deaths caused by Covid from indirect deaths that might be caused from reallocations in the health service or people delaying going to hospitals, we have seen a huge drop in A&E admissions, for example. Those might be directly related to some of these preconditions that older minority groups suffer from to a greater extent. It is not a matter of demographics. It may be a matter of some of these household conditions or these pre-existing health conditions. More work needs to be done to try to tease apart those elements.

In terms of exposure at work, we can also look at the extent to which minority groups are involved as key workers. There we again see a lot of heterogeneity across different groups. If we look at the share of all workers in a particular group, black Caribbeans and black Africans have a greater share of people who are working within those groupswho are working in key worker sectors. For example, the share of white British individuals working as a key worker is about 20% or 21%. That corresponds to the share of black Africans who are working just solely in health and social care, even ignoring all the other sectors in key work. There are large disparities there. Some other groups, such as Bangladeshis and Pakistanis, are less exposed through key worker sectors and occupations.

There are also differences across these groups in terms of employment and income risk, in terms of sectors they have been exposed to that have been shut down. We know that shut down sectors are more likely to be lower paid sectors. Even among sectors that remain open, lower paid jobs within those sectors are less likely to be able to be done from home. In both dimensions, across sectors and within sector, you are seeing impacts at the low end of the income distribution, where some minority groups tend to be concentrated.

I can talk further about impacts in terms of the ability to save and resilience across these different groups. One final point that I would like to emphasise is that many minority groups tend to be over-exposed to self-employment as well. It has been somewhat harder for the Government to implement policies to provide financial support to that group. For example, Bangladeshis, especially Bangladeshi men, are far more exposed to self-employment than white British men or any other male group among any of the other minority groups.

Maybe I should stop, then we can come back to some of the economic impacts across groups a little bit later. It is essentially a combination of all those factors, but demographics and age would make it harder to explain these differences across minority groups.

Liz McKeown: Could I come in briefly just to add to a couple points we have heard there? I wanted to flag to the Committee that we have some forthcoming publications that I think will shed further light on some of the issues that have just been raised. Tomorrow, we will be publishing an analysis of ethnicity and mortality. In that analysis, we have linked death registrations to data from the 2011 census, where we can get ethnicity and other characteristics. Those other characteristics will allow us to explore some of those questions that were just being discussed. It is a rich dataset. We will obviously share the findings of that with the Committee after publication. That will be very useful to address some of those questions.

The other area I wanted to pick up on was the talk of how different groups are exposed to Covid-19 at work. Next week, we will be launching a publication that looks specifically at that issue. It uses a technique to see how much different occupations are exposed to infection and also how proximate they are to other individuals at work. It allows us to think about the risk of Covid-19 at work for different occupations. We can then look at how different protected characteristics work in those occupations. Some occupations are over-represented by women, ethnic minorities or what have you. That will add to our evidence base in this area and begin to fill in some of the gaps. Taken together, the IFS work last week and these forthcoming publications will give us quite a rich picture of what is happening with ethnic minorities during the pandemic.

Q86            Alex Davies-Jones: Could I come back to the UK Government’s response in identifying these specific at-risk groups? Have your organisations been able to do any analysis and comparisons with other countries in the world who have dealt with the coronavirus and have seen similar results? If you have, where does the UK’s response rank in terms of this?

Professor Sir Michael Marmot: I have just, a little earlier today, come off the telephone with people at WHO headquarters in Geneva. We are trying to look at exactly this question of inequalities in Covid-19. The first thing that is very clear is how blessed we are in the UK with good data. Liz, I am not just saying this just to make you blush, but we cherish ONS. The series of reports and the data that come out of ONS are extremely valuable. Most countries just do not have that, certainly in low and middle-income countries. It is extremely difficult. We want Dr Tedros to be able to see. We want to provide him with the evidence so he can talk about the inequality impact of Covid-19 in his briefings, but getting the evidence for him is extremely difficult. That is the general issue.

It is highly likely that, particularly in low and middle-income countries—we have a little bit of data coming from India—the impact is absolutely catastrophic. If you are a migrant worker, the majority of employment in India is in the informal sector, so there is no such thing as furloughing with 80% of pay. You are an informal worker, a migrant worker. You are out of work, homeless and hundreds of kilometres from your home, with no access to food, medicines or healthcare. It is catastrophic and we do not have very good data. That is the problem.

As I mentioned a bit earlier, when we look at the United States, where we do have better data, the problems that Imran and I have been referring to are amplified by the inequities in access to healthcare. The United States does not have a healthcare system; it has a healthcare sector. It is an economic activity for the country, and it is exposed at times like this, so it is made very much worse. We see the excess among African-Americans. Interestingly, there is the so-called Hispanic paradox in the US: overall life expectancy seems to be better for Hispanic Americans, but they certainly seem to have an excess in mortality associated with Covid-19. That again may be partly a healthcare issue, as well as all the other kinds of issues Imran was talking about.

Professor Rasul: To follow up on your point about international comparisons, there has been lots of discussion about league tables and how the UK’s response is comparing. There are differences in methodology about how data is being collected and what it represents. I wanted to echo what Michael said to Liz. We are in a very fortunate state in the UK with the data we have through ONS and other sources. We are in a tremendously strong position to base our decisions on real-time evidence.

When we are making those comparisons across countries, we have different rates of testing and different types of testing. That is going to lead to differences in case counts. How we count deaths, whether it is in hospitals or in social care, varies across countries, as does whether we use deaths from Covid or excess deaths. Those are all going to be measurement and methodological issues that will make cross-country comparisons difficult.

Imagine we put all of those to one side and we said that everybody supplies the same data. Could we then still make useful cross-country comparisons? The thing to start with is to say that the only thing that is common across all countries is the basic epidemiological model of transmission. Let us assume that rates of transmission are the same and the biology of the virus is the same. How that then mediates or translates into deaths, cases or anything else you are interested in is always going to be mediated by a variety of different factors.

The first is what we have been discussing so far, the underlying social structure in that society. What are the initial levels of inequality? What are the particular dimensions of inequality that differ across societies? Different minority groups might be in very different circumstances in different countries. That is going to lead to some differences naturally across countries.

The second element is going to be the policy response that countries have initiated. The devil is in the detail there, in terms of exactly what is implemented, the timing of that implementation and the extent to which that policy is enforced. As Michael was suggesting, when we look across the global income distribution, the key factor is the ability and capacity of the state to provide resources quickly and in a targeted way to groups that are most at need. We have had very good examples in the UK of how our underlying infrastructure has allowed us to target resources to those most in need in a very quick way, given all the other circumstances.

There are differences in the social structure and policy response, and then differences in how households behave. Households will respond differently to the same types of information across countries, but the information that is being provided is coming from a variety of different sources there. In a sense, to try to disentangle those elements is a very difficult, complicated task. To then say, “This gives a clear ranking of the effectiveness of the policy” is difficult when all of these other factors are also different across countries, in terms of the underlying social structures, which policies were implemented and household behaviour. It may be that, in the longer term, we can make sensible comparisons across countries. That will be rather painstaking work and subject to lots of caveats if we go down that route.

I might also highlight to the Committee that, in the shorter term, it might be as important to put emphasis on trying to understand what is happening within a country, across different parts of the country. We might say, “The policy environment is somewhat similar. What is different in different parts of the country is these social structures, how households might be responding and some of the group differences. That might give us a better handle on the extent to which the same policy is effective in some areas and has not been effective in other areas. In the shorter term, that might shed much more light on how well the Government are doing than trying to make comparisons across countries, which are fraught with methodological as well as the compounding factor differences.

Some of the most recent work we have been doing at IFS has been trying to understand that geographic variation in the crisis, which parts of the country will be hit on different dimensions of crisis and whether one policy can deal with all of those or whether we are going to need a whole range of different policies to try to deal with that geographic dispersion.

Liz McKeown: I utterly agree with what has been said. There are significant methodological challenges when trying to make international comparisons. I agree with Imran that, over time, they will be possible, and they will be possible when we have the sort of depth and breadth of data on deaths registered across different countries that we are striving hard to produce in the UK now. We know that sort of data is not currently being produced in all countries, so those comparisons are not currently possible in that way.

In ONS, we will continue to look at the disaggregations we can produce to help inform policy in this area. Last Friday, we produced a publication looking at geographical variation at different levels, including allowing people to look at how deaths varied at relatively local levels, which has been of significant interest to both the public and policy makers. Our work tomorrow on ethnicity will be another chance to look at how mortality is varying across different groups. Next week, we will be looking at occupation in a similar way and carrying on building up the evidence base in the way that Imran suggested within the UK, while also looking ahead to a period where those international comparisons will be more possible. That is the way we are going at the moment.

Q87            Alex Davies-Jones: Are there any immediate actions that you would like to see the Government implement in order to protect these at-risk groups more? Are there any recommendations that you have for themanything that can be done immediately?

Professor Sir Michael Marmot: The answer is yes. My 10 Years On report gave a framework of understanding of the causes of health inequalities and overall health of the population. Through the life course, we look at early childhood, including poverty and services for early childhood. We know, when we look at education, that what we are indulging in, in a way, is an extended summer-fade experiment. We know that, when you look at inequalities in educational outcome, the disadvantage in learning of children from more disadvantaged backgrounds increases over the summer. They start the new term, when they come back in the autumn, at a disadvantage. The disadvantage tends to lessen over the school year and then increases again next summer. A likely explanation is kids from more privileged backgrounds—I am shy here—have books in the house and parents who can foster learning and the like. The lockdown of schools is like an extended summer-fade effect. It is highly likely that the inequalities in learning will increase.

I could go through the rest of it. The third one was about employment and working conditions, which we have already been talking about. Then there is money, having enough money to live on. The fifth one relates to communities, housing and the like. It is absolutely vitally important and urgent to look at each of those and ask what we need to do with urgency to make sure we are not increasing inequalities, conditions for early childhood and for education. There are things we can do on all of them.

Q88            Alex Davies-Jones: Are there any specifics you would like to mention? Are there any recommendations you would like to see implemented now?

Professor Sir Michael Marmot: I would not like to trivialise it by just pulling things off the top of my head. I am particularly concerned about the likely increase of educational inequalities with the closing of schools. For me, it is a matter of urgency to pay attention to how we can reopen in a phased, graded way with all attention to safety. It means looking at the transport system, having the proper public health system in place in terms of testing, tracking and isolation, which has to be done with the reopening anyway, but giving priority to that, and paying attention to whether there is more we can do to redress the likely amplification of educational inequalities. As I say, I could go through the rest of it in similar fashion, but that is the kind of thing I would like to see us paying attention to.

Professor Rasul: I would emphasise that there is a very important interaction here with age and minority groups. It comes out quite strikingly when we look at the data. We already mentioned that household structures and household overcrowding is different across different groups, to the extent that older people in those minority groups are more likely to find it harder to self-isolate. As we think about policies to move us out of the lockdown, we recognise perhaps the need for older people to remain self-isolated. That is going to be much harder to do for some groups relative to others. That is an important factor to bear in mind.

The second one is the one I already mentioned, so the strong age gradient in health preconditions that is very striking for minority groups as well. Older members of those communities were in worse health pre-crisis than people of a similar age in the white British population. Older people in some of those groups are being hit by a triple-whammy in terms of their exposure to shut down sectors. When we look at whether you work in a shut down sector and how that varies by age, we find starkly different patterns between some minority groups and the majority population. Among the majority population, it is very much younger people who are employed in many of the sectors that have been shut down, whether it is retail, tourism or other sectors in which face-to-face contact is important. That position is completely reversed among Pakistanis and Bangladeshis, where it is older individuals who are more likely to work in some of the shut down sectors, presumably because of channels of self-employment or working in retail. In other groups, such as Indians or black African populations, there is essentially no difference across age groups. All ages are similarly exposed through these economic channels. There are some important interactions between age structures and minority groups that make older individuals, especially among Pakistani and Bangladeshi groups, particularly susceptible on this whole range of dimensions.

We might say, “None of this matters as long as there is somebody else in the household who is able to work during the crisis. When you look at household compositions and the extent to which you have a working partner who is not affected by a shutdown sector, that again varies tremendously across different groups. For example, among white men who are in a shut down sector, about half of them have a partner who is still in paid work at the current time.

That is quite different among minority groups. Among Bangladeshi men, about 47% of them are in a shut down sector and two-thirds of them have a partner in no paid work. For those households, the consequences of being in a shut down sector are even greater than for some other groups. For black Caribbean men in shut down sectors, the majority of them are either single or lone parents and, again, absent another individual in the household who could potentially be earning and helping to smooth some of the shock.

There seem to be important markers that suggest that some households will find it difficult in the current crisis. Also, as we try to come out of the lockdown, different households might be differentially able to self-isolate or to recover back in terms of some of the economic losses as well.

Q89            Alex Davies-Jones: I have one more question, which is for Liz. You mentioned your new report, which is due to come out, where you have had to use census data in order to gather some of this intel. The Home Office has said that it has no plans to include ethnicity on death certificates and death registrations. Do you agree with that? Would it be a lot easier for you if this was included? What is your opinion on this?

Liz McKeown: It is not for the ONS to determine what is on death certificates, but we are obviously really interested in making sure we can produce the information that is addressing some of the big questions at the moment. Of course, one of the big questions at the moment is how mortality is varying by ethnicity. We had to be creative at that point in our analytical approach to think about how we could address that. Our ability to address that was by matching across to the 2011 census information. That matching has been effective and, as I said, the findings from that will be through tomorrow.

There are a lot of different considerations that have to be weighed up around what is on a death certificate. There are costs and benefits of including more data on there, in terms of overall quality of information that you are then able to obtain. Our focus has been on how we can produce the richest possible evidence base with what is available to us.

Q90            Nickie Aiken: Professor Marmot, you were talking about the problems of particular ethnic minorities in educational attainment. I have been very interested to see the Education Select Committee is now looking at the fact that it is white working-class boys who tend to be more affected by poor educational attainment, particularly in south Wales and other parts of the United Kingdom. From my own constituency in the Cities of London and Westminster, my experience has always been that the Bangladeshi, Pakistani and Indian families are very much about educational attainment, and I would say more of their children go to university than white working-class boys. I would be interested in your views.

Professor Sir Michael Marmot: I was referring to socioeconomic differences, not ethnic differences. When it comes to looking at ethnic differences, Imran’s point applies. There are real differences in different ethnic groups. It is not uniform. If you talk to people in different parts of the country, you come upand I think the data confirms itwith a picture that people of Gujarati background do better educationally than the English average. People of Pakistani background do less well than the English average. It is not uniform.

I was referring to socioeconomic differences. As you say, the boys of a white working-class background do poorly. There is a very clear social gradient. It is not just that there is one sub-group that does poorly. It is a very clear social gradient: the more deprived, the worse they do. The ethnic difference is not quite orthogonal to that but it does not map completely on the social gradient because, as I said, there are these differences among different ethnic groups.

Q91            Peter Gibson: This question is directed to Professor Marmot. You talked initially about the fact that Covid-19 has exposed and amplified inequalities in health. Looking at the fact that it seems to have had a pronounced impact on those with certain protected characteristics, do you believe that this was foreseeable? Could you comment on the impact specifically on men and why the impact seems to be falling greater on the male population? You interestingly commented on the differences between our healthcare system and those in America and highlighted the inequalities there. To give the Committee some balance in terms of other good practices, other healthcare systems without the same systems as the United States, could you highlight some other countries and their healthcare systems without those inequalities?

Professor Sir Michael Marmot: As I said at the beginningand I apologise for this because I have already misled two Committee members and probably moremost of my responses were around, in general, socioeconomic differences in health. Those were the inequalities we analysed. As I said in my introductory remarks, it is largely because we did not have the evidence on the other protected characteristics that are the interest of this Committee.

To me, it was predictable. On Friday, after the ONS published its Covid-19 mortality by levels of deprivation, I was asked, “Was this a surprise? My answer was, “No, not at all. I had been imagining that that is what we would see. I just had not seen the data until then, and then ONS published it on Friday and it was very clear. What was clear was the social gradient that I have been talking about. If anything, the gradient for Covid-19 mortality was slightly steeper than the gradient from all causes.

Yes, it could have been foreseen from what we understood about transmission and the importance of social distancing; I certainly predicted it. For all the reasons we have rehearsed already this afternoon, social distancing is far more possible for people at greater social advantage. To the extent that people from different ethnic groups and those who are disabled are less likely to find themselves in a position that makes social distancing possible, they will be at higher risk. That part was all predictable.

What was not predictable to me, but this may just represent the depths of my ignorance, was the excess in men. Please do not misunderstand me if I say I am confused about gender. I mean intellectually; it is not a personal comment. In general, when you look at inequalities in health, men have bigger inequalities in health than women, and we documented that in my 10 Years On review. The social gradient was steeper for men than for women. That much was clear.

Years ago, when we looked at the east-west differences in life expectancy across Europe, the disadvantage of being in the former communist countries of Europe was bigger for men than for women. It seemed to be part of a general phenomenon that, when you have conditions of social disadvantage, as characterised the former communist countries of eastern Europe, men seem to be more susceptible than women. That is what the evidence seems to show.

In my 10 Years On report, we pointed to the fact that women in the most deprived 10% of deciles on the index of multiple deprivation, outside London, actually had a decline in life expectancy. Although the inequalities were bigger in men, it was women in the bottom 10% who actually had a decline. For men in the bottom 10% it did not improve very much, but women in the bottom 10% actually got worse. Now you can understand what I mean when I say that I am slightly confused. I had this nice, clear picture that men are more susceptible to social disadvantage than women.

When we turn to Covid-19, is the excess in men because of social disadvantage or is it, as I have heard biologists speculate, that women are more biologically robust and have better immune systems? We know women live longer than men, which I think is biological, not social. I think everything is social, but I do not think that is. I think women live longer than men for biological reasons. I do not know whether that relates to women’s relative protection from mortality from Covid-19, but it is a very important question.

About the UK versus the US in health systems, the US is unique among high-income countries in not having universal health coverage. The Commonwealth Fund does a regular survey of 11 countries’ health systems, and the US ranks bottom most of the time. The NHS ranks top in equity of access. Health systems across Europe do it differently in different countries, but they all, more or less, guarantee uniform universal health coverage. The US is unique. It is a distinct outlier. The Swiss are very proud of the fact that they have an insurance-based system with private practitioners and so on. They seem to achieve universal health coverage their way. We all do it differently in different European countries, but we all get there in our different ways.

Q92            Peter Gibson: My point was not to ask you to compare with the United States further but to ask you for your views in respect of which healthcare systems have better outcomes and less health inequalities than our systems, rather than lumping all the other European countries together. I am driving at who is better than us. From what you are saying, I accept that the United States is worse than us, but who is better than us?

Professor Sir Michael Marmot: There was a recent report from the European office of WHO. They claim—and I will allow them this claim—that only about 10% of the health inequalities of a country can be attributed to differences in their healthcare systems. Most of the health inequalities we see in different European countries are due to social determinants of health, not to differences in functions of the healthcare system. When we compare inequalities across countries, we are not actually learning very much about inequities in the healthcare system. We are learning much more about inequities in society more generally.

When we look at that, it is a very vexed question. It relates to some of the kinds of difficulties that were discussed earlier. Interestingly, the Mediterranean countries seem to have narrower health inequalities than we do, not the Nordic countries. My prejudice would be that the Nordic countries would have narrower health inequalities. They do, but it is actually Italy and Spain that seem to have narrower health inequalities than we do, which is really interesting. Goodness knows what that is aboutsunshine, olive oil, the Mediterranean dietbut that is what we seem to be observing. I do not think it is because of outcomes of the healthcare system. I think it is much more general than that.

Q93            Peter Gibson: Could I direct my next question specifically to you, Liz, in respect of data? As we see the daily press briefings, we have seen the statistics we are looking at and comparing with other countries. The differences in data-gathering and the quality of data have been highlighted today already. That clearly poses some difficulties for us as a country in comparing the impact of the virus here with other countries that are facing the same challenges. Can you tell me what the ONS is doing to improve its own data-gathering in respect of people with protected characteristics here?

Can you also highlight what work is being done to establish and clarify to a better degree the level of excess deaths that we have had here in this country, in comparison to other countries that are recording their deaths slightly differently? One of the key issues that has come to my mind is that we appear to be recording all deaths with Covid. I believe that some other countries may only be recording deaths because of Covid. What work, if any, is being done by the ONS to extrapolate that data?

Liz McKeown: Thank you very much for that question. There are a number of parts there. What we are doing, as ONS, to improve coverage for people with protected characteristics was the first part. As I have already mentioned, each week we publish a large range of death statistics now, both overall deaths and deaths involving Covid-19. We have those available with age and sex breakdowns, so that is already available. I have previously mentioned that we have done some analysis to link our death registration data to the census 2011 data, which is allowing us to look at more protected characteristics. There is the ethnicity release tomorrow. We will go on and be able to look at, for example, religion in due course as well. It will build up our evidence base on deaths by different protected characteristics. That is some of the work that is going on internally at ONS.

It is more challenging on the international picture and making those comparisons, as we have discussed before. We are in contact with our colleagues in other national statistical institutes, and that is a way of us sharing how we are approaching these issues and seeing what they are doing. Over time, I think we will see more countries producing the sort of disaggregated data that we are producing in the UK. As we were discussing previously, those international comparisons will become more possible over time.

One of the things we have been really keen to do in ONS is try to help explain to people how these different data sources interrelate. Very early on, people were very interested to know the daily figures on deaths produced by DHSC, compared to our weekly death registration figures. It was really important for us to be clear about the differences there. Each week, along with our publication that shows our own weekly death figures broken down by those splits that I have mentioned, we also put out a comparisons piece that shows how the data compares across different data sources and what inferences you can and cannot draw as a result. Being very clear about what is behind each data source and what comparisons you can make is important.

In fact, it is that that is challenging in the international comparisons space, because we do not always have that information from other countries about the exact methodologies that have been used, which means that, when you are trying to compare across countries, you cannot always see straight away what some of those methodological issues will be. Over time, we will resolve a lot of those issues and the comparisons will become easier, but that will take some time.

Q94            Peter Gibson: In terms of doing the disaggregation of the data in order to establish those people with protected characteristics and the impact on them, how much further work is there to do on that? Are you getting the full support of Government in terms of everything you need to do that?

Liz McKeown: The additional work links in to census 2011. We have what we need to do that in the office. It is quite complex. Linking in different datasets is always analytically quite tricky, and we want to make sure that we get that right. Obviously, the first publication from that work is coming tomorrow, so that is really good progress. We will carry on for age and sex, continuing to produce that, disaggregated, every week, so that will continue to be available.

There are some protected characteristics that were not in the 2011 census where it is much harder to think about how one might approach that, and that includes, for example, sexual orientation. That is more of a challenge.

Professor Rasul: Peter, just very specifically on your question about whether this was predictable or not, we know from other times that inequality in some important dimensions in the UK has been increasing since the 90s, say in terms of household earnings. We know that, when we are hit by shocks such as the financial crisis or just economic downturns, that does have impacts differentially across the income distribution. For a health-type shock, it is not that surprising that it is amplifying health inequalities that were there before. It is very difficult, in a short space of time, to think about policies that are going to really mitigate that in the short term.

The crisis has amplified some inequalities but not all inequalities. For individuals who have a private pension, and where a lot of their wealth is tied up to the valuation of the stock market, those households tend to be relatively well off to begin with. They have seen a large decrease in their wealth, and that is reducing inequality in that sense.

We have other dimensions of inequality that we never considered to be that important, which are really being highlighted by the crisis, and going forward, whether it is in terms of the ability to work from home, having access to broadband or having access to green spaces, they might, in the longer term, have very important implications for how we think about the impact on mental health and other types of outcomes to the population.

The real billion-dollar question is about whether we think of this as a once-in-a-century event or whether we think of this as predictable and that something like this could happen again. We have known from other literatures that the frequency and complexity of viral outbreaks has been increasing over the last 30 years. The forces driving that are rises in global temperatures, increases in urbanisation, and increasing contact between human and animal populations. Those trends are going to be very hard to reverse, so we are seeing these outbreaks occur more and more frequently. The question is about the extent to which we want to almost pathogen-proof the economy. If we want to go down that route, how do we think about future crises?

For example, imagine we had an outbreak that affected international travel. You would imagine that that is going to impact different sectors of the economy than if we had a crisis that targeted children or pregnant women or different parts of the population. This is the real challenge going forward. How do we think about all of these dimensions, and what might be a robust set of policies to have in place? It is not just about the particular characteristics of Covid. Obviously, those will dominate in the short and medium term, but in the longer term, it was predictable in the sense that something like this was going to happen eventually. The question is about what we think of as being the important ways that we can respond to these types of crises when the epidemiology of those crises may well vary at the time.

Q95            Kate Osborne: My question is to Professor Marmot. We have heard that some groups, such as LGBT+ people, may be at greater risk from Covid19 due to existing issues and/or co-morbidities being higher than in the general population. Does this follow with any of the research that you have done? Do you think there needs to be more work on this area?

Professor Sir Michael Marmot: I have no data on LGBTQ+ sub-groups of the population in relation to the kinds of analyses that I have done, so I cannot really comment. We kind of understand a lot of the issues, and where they fall on the socioeconomic spectrum will affect what happens, but there are then other issues, such as discrimination and social isolation, that are likely to affect people’s access to the things that are important to health. This is coming from speculation, and I am more comfortable talking about the evidence.

Q96            Kate Osborne: The Committee has heard that charities such as Shelter have reported a high volume of LGBT young people who are in need of help since lockdown, particularly around housing and the risk of them being homeless. I know you say that you do not really have figures around this, but I just wondered what your opinion is as to why this is the case and what could be changed to protect this vulnerable group.

Professor Sir Michael Marmot: First, in my 10 Years On report, we quoted figures from Shelter that suggested that one-third of employed people are one pay cheque away from being homeless. To the extent that belonging to one of these protected groups, particularly LGBTQ, puts you in that vulnerable situation because of discrimination or for whatever other reason, you will have a condition of being in danger of being made homeless. I am only speculating about how belonging to a particular sub-group might put you in a condition of vulnerability.

Again, the responses in relation to the pandemic have made a huge difference—people actually being forgiven their rent or their mortgage repayment. If a third of the population were one pay cheque away from being homeless, without taking steps to forgive mortgage repayments or rent we would have an epidemic of homelessness on top of everything else, which is catastrophic for people’s health—talk about amplification of health inequalities. As I say, I cannot comment on the specifics of your question, but the general issue of homelessness is one of immediate and urgent concern.

Q97            Kate Osborne: I would like to move on to regional inequalities, if I can, and again direct this to you, Professor Marmot. Dr George Rae, the British Medical Association’s regional chairman for the north-east has written an open letter to say that the area is suffering disproportionately, with figures showing that more than 7,000 people in the north-east have tested positive for Covid-19. The figures seem to show a disproportionate amount of serious cases and deaths, and it seems to have shone a light on the health inequalities in the north-east. Can you tell me whether the economic impact of Covid-19 is likely to be felt equally across the UK regions and nations? Do we know which areas have been, or will be, hardest hit by Covid-19 in terms of income, employment and low health outcomes?

Professor Sir Michael Marmot: I have two responses. First, we have known for a long time—and I referred to this in my 10 Years On report—that if you look at the social gradient in life expectancy, it is true everywhere in the country that the more deprived you are and your area is, the shorter the life expectancy, but the slope of that gradient varies by region. The gradient is steeper in the north-east and the north-west than it is in London and the south-east.

Another way of saying that is, if you are in the least deprived decile on the index of multiple deprivation, in terms of life expectancy it does not matter where in the country you live; you have relatively long life expectancy. The more deprived you are, the more it matters that you live in the north compared with London and the south-east, and the greater the disadvantage of being in the north-east and the north-west. The gradient is steeper. As I mentioned when I was talking about gender, in those northern areas life expectancy for women in the most deprived 10% actually went down.

The second part is that, when the Prince of Wales, the Prime Minister, the chief medical officer, the Secretary of State for Health, Professor Neil Ferguson and high-status people such as Tom Hanks were reporting Covid-19 infections, you would say, “Yes, that is what you would expect. There is a lot of mixing, travelling around and shaking of hands being up there. You would expect London to get it more than elsewhere for all those reasons. We are all catching the tube, and it is crowded on buses. It is very crowded, with mixing and people coming from other countries. Just as you would expect high-status people to get it initially, you would expect London to get it initially, and that is what seemed to happen.

Then, as we have been seeing, as the pandemic really bites and social distancing measures are implemented, for all the reasons we have been saying, the lower down you are, the less able you are to exercise social distancing and the more we see this familiar pattern—the greater the deprivation, the higher the mortality from Covid-19 and the higher the admission to critical care units.

Although the figures are very suspect because of the lack of widespread testing, it does look like the infection rate that began in London has now been overtaken by the north-west and the north-east, utterly predictably for the reasons I have just been saying, initially, about the social gradient in mortality and life expectancy being steeper in the north-east and the north-west. There is something about being towards the bottom of the deprivation spectrum and being in the north-east or the north-west that puts you at particular disadvantage in terms of health, so it is predictable that it would put you at particular disadvantage in respect of Covid-19.

Q98            Kate Osborne: Following on from that, if regional inequalities are likely to be exacerbated by Covid-19, what response do you think is needed by the Government to tackle these inequalities?

Professor Sir Michael Marmot: Forgive me that I keep referring to my report of February, but that is what I thought was needed. As I said earlier, when we launched it, I said, “We will give the Prime Minister what he needs in order to implement his levelling-up agenda.We went through what was needed: reducing child poverty; having Sure Start children’s centres or the like for improving the quality of early child development; investments in education; and employment. We went through each of those domains. That is what is needed to change the fortunes of people in the more deprived regions of the country. It is not a mystery as to what we need. We know what we need. The evidence that we compiled says what we need. We have, in a sense, produced it and said, “Right, do it.

I can see why there are other priorities right as we speak but, that said, we should not neglect what is needed both to deal with the longer term underlying issues of health inequalities, which show these big regional differences, and to deal with how we prevent more urgent problems as we emerge from the pandemic.

Professor Rasul: Just to pick up on the geographic patterns of the crisis, this is something that we have started to look at in a bit more detail at the IFS. What makes the crisis unique and different from, say, the financial crisis is that it is affecting families and businesses through so many different channels. We can think about three different dimensions to the crisis: one is the health dimension, one is the economic dimension and the other one could be the impact on vulnerable families and vulnerable children.

We have been trying to document and understand the geographic variations across all three dimensions and how they relate to each other. What we find is that, in terms of vulnerable children or vulnerable families, that is very highly correlated to standard measures of deprivation across local authority areas. However, when we look at the economic impacts of the crisis, say as driven by the share of workers in an area who are employed in a shut down sector, that actually turns out to be negatively related to how we normally measure deprivation. It is actually the least deprived areas that have had the biggest economic shock. The health shock is somewhat in between. A good example of that is London, where a large number of workers are in shut down sectors, but workers in London are more able to work from home, for example.

The normal patterns that we see in terms of economic deprivation do not play out on all margins of the crisis in the same way. What we are finding is that these different elements of the ways in which local areas and local economies are being hit are not very correlated to each other. Those parts of the country that are being potentially impacted the most on the health dimension are being less impacted on the economic dimension. Similarly, those parts of the country that are being most hit in terms of vulnerable families, social care and child referrals are being less impacted, say, by the economic dimension. It actually becomes quite a complicated picture.

There are some local authorities that are being hit on all three dimensionsthose that have old populations in poor health, that are reliant on tourism and that have pockets of deprivation in them. There are a number of local authorities that are really going to be hit on all three dimensions, and we are trying to classify different areas to the extent they are being hit on multiple dimensions.

This then raises the question of what the optimal policy response should be when we have such geographic variation. There has been a lot of support that has been provided to local authorities, and local authorities are being hit in two ways. One is that their revenues, say through business rates, have fallen, and that is related to some aspects of the economic shock, and yet their demands for expenditure related to social care and vulnerable families have increased. It may not be in the remit of local authorities to deal with all three dimensions of the crisis, so there may be new policies and policy instruments that we need to think of.

Certainly, in our preliminary work, the geographic variation of the crisis is very complicated. It depends on which dimension of the crisis you are looking at. They do not always overlap, with some authorities being hit along all three dimensions, and others really being hit along one.

Q99            Angela Crawley: Can I direct my first question to Professor Marmot? You touched briefly earlier on the issue of socioeconomic disparities. The Government have introduced a number of measures to protect public health. Are there particular groups that you think will be most affected by these measures? Could you outline specifically which groups and why?

Professor Sir Michael Marmot: Let me start by picking up on what Imran said. We documented that the reduction in spend by local authorities over the last decade was intimately related to degree of deprivation, which is regional. The poorest 20% of local authorities had a 32% reduction in spend per person, and the least deprived 20% had a 16% reduction. Liverpool had the steepest reduction of any city over a certain size; it had something like a 48% reduction in spend. There are huge regional differences related to deprivation on spending by local government.

We are concerned about social care. If you look at the adult social care element of those figures—Imran will be pleased that I am quoting IFS figures that we produced in our report—the reduction in social care spend was something like 16% in the most deprived 20% of local authorities and 3% in the least deprived, and it was a gradient. Public health budgets have been slashed. The public health spend went down dramatically. Public health was moved into local government and, given that local government was being cut in the way I have just described, in a sharply regressive fashion, that limits the ability of public health to respond. Local government has been sorely restricted, and the public health dedicated budget was cut quite dramatically.

I am not sure quite what specific public health arrangements you are speaking of, because we are in a fairly difficult position nationally on social care, public health and on general ability of local government to respond, because of the parlous financial position that they have found themselves in.

Q100       Angela Crawley: I appreciate that, and I am conscious as well that you are correct that it is the activists and the local volunteers who are, in many cases, filling the gaps, as many healthy local volunteers are providing food and medicine delivery for people who are unable to leave their homes. As many of these volunteers will potentially return to work, what measures would you suggest to Government in order to ensure that those who are the most vulnerable and who will need to continue to shield get that support if there is a reduction in local authority funding available?

Professor Sir Michael Marmot: The Government obviously have huge decisions to make. My own view, for what it is worth, is that I am encouraged by the Prime Minister saying, “No, we do not want to talk about austerity.I was asked that question on “Newsnight” the other night. That is good, because it would be a mistake, for the reasons I have just been pointing out. We need to be looking at how we fund local services properly because, as Imran said, the need is increasing for local government.

I talked to people in local government before the pandemic was upon us. They were saying, “We can provide very little other than for those for whom we have a statutory duty, and even some of those we are having difficulty with. It is vital to look at what local government needs to be able to do and whether it is funded appropriately. Again, as Imran said, its revenue locally will be hard hit by the economic downturn, so it is going to be highly dependent on Government funding. It will have to look at that.

Volunteers are very welcome and heart-warming. The third sector is vitally important, but that has also been cut. Funding to third-sector organisations has been drastically hit. Volunteers need organisation, and the voluntary sector has a very important role to play, but that needs to be funded properly too.

Q101       Angela Crawley: Turning to Professor Rasul, you touched briefly earlier on the impact on black, Asian and minority ethnic groups and the impact on intergenerational households. Those who are designated as key workers during lockdown will experience particular impacts, but what research have you done on the characteristics of key workers and how they might be specifically affected?

Professor Rasul: We have looked at a variety of different dimensions across key workers. Again, the underlying repetitive message is that not all key workers are alike. When we compare key workers as a whole to non-key workers, the differences are not very stark to begin with. When we look at key workers sector by sector, that is when we really start to see that not all key workers are the same. There are very different compositions within them.

One big difference to begin with is just in terms of gender. Twenty-five per cent. of women who are in the labour force are in one of the key worker sectors. Sixty per cent of all key workers are women. Again, there are very big differences across key worker sectors. Women represent the majority of key workers in health and education, corresponding to about 70% to 80% of all key workers in those sectors, whereas, in other key worker sectors, such as public water and transport, they constitute about 20% of workers. The gender dimension across these key worker sectors is very important.

The age structures across key worker sectors are also very different. The food sector really stands out as having quite a different age structure to other key worker sectors. For example, just over 16% of key workers in the food sector and about 40% of farmers are 65 or older. That is one of the key worker sectors that has the oldest population amongst it, and that is almost three times as high as for the workforce as a whole.

In terms of country of birth, again, there are quite large differences across different key worker sectors. Most key worker sectors have a large share of workers who are foreign bornbetween 15% and 30%. Again, the food sector is the sector that has the largest share of foreign-born workers, close to 30%. Within each key worker sector, amongst the foreign born, the majority tend to come from outside the EU. There are clearly going to be implications when thinking about those issues going forward.

Key workers also differ tremendously in terms of their underlying pay and the extent to which key workers are low paid. As a whole, key workers are more likely to be low paid. About a third of them earn £10 or less per hour, and that compares to about 28% of workers in non-key worker sectors being at that part of the wage distribution. In some sectors, the pay of key workers is even lower: 71% of food sector workers and 58% of employees in social care earn below £10 per hour. That is below the long-term target for the national living wage. Conversely, at the other end of the distribution, non-key workers are more than twice as likely to earn £30 an hour or more. There are big differences in terms of the pay for these workers, and that partly drives who is able and willing to enter these sectors to begin with.

There are big differences by gender, age, wages and country of birth across all of these key worker sectors.

Liz McKeown: I just want to flag that we have an additional statistical publication coming out on key workers on 15 May. As well as looking at some issues around protected characteristics, it will also explore some new issues, like how much public transport they use, which is another important element here.

We have also used our Opinions and Lifestyle Survey, which is our new survey looking at the social impacts of Covid-19, to ask about the concerns that key workers have. Unsurprisingly, they are more likely than other workers to be concerned about their work and to raise specific concerns around health and safety and increasing their hours.

Q102       Peter Gibson: Professor Rasul, as you were talking about regional inequalities, you alluded to the need for specific policies to deal with those. Putting you on the spot, if I may, do you have a top five key policies to tackle those regional issues?

Professor Rasul: The point that I wanted to emphasise was that different parts of the country are being impacted in different ways by the crisis. The health, the economic and the vulnerable families dimensions might require different policies, not all of which will be in the remit of local authorities, which are also observing lower revenues, say through business rates, and higher demands for expenditure, say through social care. This might require some policies that we have not had the ability or have not had the need to try to implement before. It will be a complicated picture, given these differences across different areas. I do not have any one policy that I think will deal with all of these elements.

The additional complicating factor is that, presumably, the speed of coming out of lockdown might be different across different regions. Some will gain more economically in the short term and others will have older populations that presumably would rather have a slower exit from the lockdown, everything else equal. That will raise some issues about inequality across areas, how they are being currently impacted and how they might be impacted in the future. That might require a whole new raft of policies that we have not currently really seen, because a lot of the policies have been thinking about one dimension in isolation, but the same parts of the country are being impacted in very different ways.

Q103       Peter Gibson: Do you have any policy ideas for us?

Professor Rasul: The key role of Government here is to make sure that, when policies are communicated, in a sense that they are both communicated credibly and with certainty. It is a bit like how we require the Bank of England to give us forward guidance. Minimising the uncertainty that households have of how the lockdown will be exited will be a key element of how Government respond. If households are uncertain about what is going to happen in the future, that can lead to people taking precautionary behaviours that might undo some of the policy impacts that we are trying to have. We saw that coming into the crisis when some households started to stockpile. You do not want to start those kinds of precautionary behaviours as we start to try to predict how we are going to move out of the lockdown.

It is important to see and obtain data on what a household’s expectations are about what is going to happen in the future. There have been some real-time surveys that have been conducted in the UK asking households for how long they expect to be furloughed and how long they expect their income to be reduced. There is a team working out of Oxford, for example, that have surveyed UK households; they find that, in the UK, workers expect to earn about a third less for the next four months compared to normal times. The average probability that they perceive of losing their job within the next month is about 30% in the UK. That is fairly similar to the US numbers. In Germany, about a quarter of people expect to lose their job in the next few months. There are differences across countries in the expectation of the economic severity, but those kinds of expectations that households have very much need to be factored into the way that information is presented to them. We know that households might respond differentially to the same information, but it is key to make sure that you minimise that uncertainty.

Peter Gibson: Your advice would be clarity and certainty.

Professor Rasul: Yes, with forward guidance coming out of the crisis.

Q104       Sara Britcliffe: Going back to the economic impact, Professor Rasul, how might the overall economic impact of the virus affect certain groups? Do you think there will be some groups that are particularly vulnerable to this?

Professor Rasul: We have mentioned before how groups differ in terms of their exposure to shut down sectors, for example. There is an interaction between minority groups and gender, for example. To pick a few examples, Bangladeshi and Pakistani men are much more likely to be in the shut down sector than white British men. Black African and black Caribbean men are also slightly more likely to be in a shut down sector. For women across different groups, the patterns are less stark.

There are different groups where men are going to be especially hit by being employed in a shut down sector. Some of that is driven by differences in self-employment, as I mentioned. That then feeds through into something that Michael brought up earlier, which is the extent to which households report being able to survive without any earnings or with reduced earnings coming in. We see that about a third of white British households report being able to draw on savings at this time of crisis for up to six months, and that number falls to below 10% amongst Bangladeshi households, black Caribbean households and black African households. Again, the resilience to the crisis is partly a function of the savings that they have been able to acquire beforehand, and that varies across different groups as well.

I would imagine that there will be implications both in the short term and in how quickly households are able to recover from the shock.

Q105       Sara Britcliffe: Just on that, is the data currently available to you helpful in analysing this? If not, what needs to improve with that data?

Professor Rasul: We are in a very fortunate position in terms of the quality of the data and the speed with which we have it. Just to reiterate what we said before, ONS has been an absolute gold standard for providing that, and it has allowed a lot of research to take place.

However, there are a couple of points that I also wanted to stress. The remit of the Committee is to discuss many different protected groups. There is only a subset of those for which we have sufficiently large samples to be able to discuss with any kind of scientific confidence, and those are really gender and different minority groups. For other types of protected groups, it is much harder to say anything, because we simply lack the sample sizes to say anythingbeyond anecdotally or in very select samples to begin with. In order to get those large sample sizes, it is critical that we have a good infrastructure of administrative data in the UK. At the Institute for Fiscal Studies, we work with a number of different Government Departments and their administrative datasets. Those datasets exist. They are available for research. They allow these types of quantifications to take place.

Where I would push would be for us to think about long-term investment and also linking across different datasets to allow us to do cross-tabulation, say, of people’s labour market status with their children’s schooling outcomes. It is somewhat hard to always get enough samples in some of those cross-tabulations, and that might be critical at a time of crisis when, as I mentioned before, households are being hit on so many different dimensions. Any given administrative dataset will only give you a very detailed picture on one dimension. That is what they are designed to be collected for, whether it is from the Department for Education, the DWP, the Ministry of Justice or whatever. When we can actually link those, then we get a fuller picture of all the different ways that households are actually being exposed and all of the cross-variation by age, group, gender and other dimensions, which might be incredibly important for how the crisis actually plays out in the longer term. It is about continually investing in those data linkages.

I am very encouraged at some of the large household studies that we have in the UK, which have been funded by Research Councils UK to have Covid-specific modules now going out to field. They will give us a very rich picture of 40,000 or 50,000 households for which we have information from other time periods, to really get a grasp in real time of how households are responding to the crisis. We have had a lot of investment in good data in the UK. Long may that continue. That really allows us to understand and be a world leader in those aspects.

Q106       Sara Britcliffe: What do you think is working well in relation to the economic response to Covid-19?

Professor Rasul: The ability to target resources to furloughed workers and to target resources to self-employed workers has had a tremendous impact on those families. When you look across countries at the extent to which they have been able to use their pre-existing infrastructures to get resources to those households that need it, it is vital.

The real challenge now is how we reverse those policies to an extent. When and how do we do that? Coming back to what I was mentioning to Peter, what kind of guidance and what kind of certainty is provided to households about how those policies will be rolled back? They will need to be rolled back to some extent, but this might lead us to think about more permanent changes and how we think about redesigning the benefits system, which has really changed a lot over the last few decades to target in-work benefits rather than out-of-work benefits. Potentially, that might be an important issue going forward in terms of the most vulnerable households and communities.

Q107       Sara Britcliffe: Are there any that you believe need to improve right now?

Professor Rasul: You could always argue that things could be done better in hindsight, but it is very hard. We are facing incredibly difficult times and incredibly difficult trade-offs. Anything that we try to do is only based on the quality of the data that we have. That is really the bedrock and foundation of being able to come up with good ideas and understand where problems are emerging. We need to make sure we always have that data available to researchers to provide this as one of many inputs into the political process.

Q108       Sara Britcliffe: My next question is to both yourself and Professor Marmot. Do you think that the safety net of benefits and assistance available at the moment will help to stop existing economic inequalities from worsening? If so, why?

Professor Sir Michael Marmot: The benefits and safety net are of absolutely vital importance. One of the issues has been that they were not generous enough before the crisis, before the pandemic. They were not generous enough and it was, to put it neutrally, a bit clunky in the way it was actually implemented. The well-documented problems with people accessing universal credit are well known, causing great difficultyloans, food shortages and the like. They were not very generous. If you look at the minimum income necessary for healthy living, the benefits available to families in need did not bring their incomes up to the minimum necessary for a healthy life. This was not a way to live.

If the purpose of doing that was to encourage people to get into work, fine, I can see the argument there. If there is no work, having a benefit level set at a level that makes it incompatible with the minimum necessary for a healthy life means that the answer is no, that is not sufficient.

Q109       Sara Britcliffe: What do you think could be changed in order to protect the most vulnerable groups?

Professor Sir Michael Marmot: In a way, Peter Gibson’s earlier question about policies, which is a very important question, made me nervous just thinking about it. Part of how I would respond to him is that at the heart of all policies should be concern for the likely impact on health inequalities. We know what things are likely to have an adverse impact on health inequalities, because we have been studying it. That is what the Marmot review in 2010 summarised and that is what my 10 Years On review summarised. We know what the key determinants are of health inequalities, and we should have regard in all of our policies to the likely impact on health inequality. If people do not have enough money to enable them to live a healthy life, that should have a big effect on the way we design policies in these very strange circumstances in which we find ourselves.

Q110       Kim Johnson: My first question is directed to Professor Rasul. We have heard this afternoon about the disproportionate impact on some of the protected characteristics. I would like to ask you what contribution that 10 years of austerity has had on these groups and the increasing levels of poverty and social deprivation. Were these foreseeable?

Professor Rasul: As we discussed before, there has been a long-running rise in inequality pre-crisis. It is not surprising that some of those dimensions then get amplified during the crisis. As Michael was saying, perhaps it is welcome to think that, in the future, we are not going to go back to austerity, but we are going to have to find some other way to fund the short-term policies going forward.

Linking to the earlier question, the policies I mentioned go well beyond the benefits system and the re-design of the benefits system, much as that may be needed. We also need to think about how we are going to encourage firms to actually re-employ people at the end of the crisis and to maintain current matches. We might also think of this as a potential opportunity to help firms to invest in technologies that allow more people to work from home or to invest in other types of technology that allow the impacts of social distancing to be smoothed out across more workers and more sectors.

Other elements that can be thought about are in terms of making sure that we provide assistance to workers who might need to work in a different sector to where they have previously been and allow those types of re-matches to take place. We have a large set of people who would like to work and are currently unable to work. Is there a way to usefully employ them and to help those kinds of transition?

The other group that we have not really touched upon very much are young people who are currently in the crisis of making a key transition either between school and university or between university and the workplace. We know from a battery of evidence that graduating in a recession tends to be very bad for people’s careers. It leads to a lower likelihood that you are going to find a job upon graduation and to lower earnings once you do find a job. Those impacts, known as scarring or hysteresis effects, tend to last for up to a decade. Numerous studies have shown how that that group might be particularly vulnerable relative to people who graduated two years ago and people who will graduate in two or three years’ time.

Again, we need to think about policies that might be specifically targeted to those individuals who are just transitioning into the labour force, by finding ways that they can switch jobs more quickly—that has been a key mechanism by which people can offset the effects of recessions—or encouraging voluntary work or encouraging them to perhaps acquire more human capital and stay on in education. Those are things that we might need to think about to smooth the shock for those particular groups as well.

The range of policies is beyond benefits. It is also thinking about how we can work with firms as well as other groups that might be at risk. Those groups will be at risk potentially for the longer term. Some of these effects will not have a chance to be offset for nearly a decade.

Q111       Kim Johnson: My other question links back to young people and other groups. Would you say that working in precarious and zero-hour contracts also has an impact on some of these groups? Does the fact that young people get paid less money to do the same work than older peers also contribute to some of these disparities?

Professor Rasul: We have seen the stark pattern that younger people are more likely to be employed in shut down sectors. Shut down sectors are more likely to be low-wage sectors. Even amongst sectors that remain open, it is low-wage workers in those open sectors that have the least ability to work from home. On a whole range of dimensions, young people could potentially be more impacted economically.

That translates into the precarious situation that many households find themselves in at the bottom of the income distribution, where they have very limited resources with which to deal with such shocks. Yes, there are important age gradients here that will affect young people on the economic dimension over and above other dimensions.

Q112       Kim Johnson: Could I also ask about older women in the workforce? Would they be impacted because of the changes in terms of state pension and having to work longer in these precarious and frontline positions?

Professor Rasul: That is really not my expertise, so I would prefer not to comment on that.

Q113       Kim Johnson: My next question is to Liz and is regarding the collection of data. You mentioned a report being issued tomorrow. Public sector organisations have a duty in terms of collecting information from their workforce. I just wanted to know how this information is collected from the private sector and the manufacturing sector in terms of these particular groups that we are talking about today.

Liz McKeown: The report that we are publishing tomorrow does not take data from the private sector or other groups in that way. That is information provided by individuals in the census and then linked to death registration information. In that way, we get that information.

In terms of a lot of what we have discussed today about labour market impacts, we would use our Labour Force Survey as our main source of information to get labour market information by different protected characteristics. One of the things that we have done there, so that we can better understand some of the impacts that we have talked about today, is that we have added questions to that survey on things like whether people are part of the furlough scheme, so that, when that information comes out later this month, we will be able to see in more detail than on the evidence we have so far how different groups have been affected in the labour market across different industries.

One thing that we have already been able to do, because we wanted to immediately see the impact that different businesses were experiencing because of Covid-19, was to stand up a new survey, which we have called BICS—the Business Impact of Covid-19 Survey—which is a fortnightly survey that we send out to businesses. We can find out information from that, such as what proportion of staff have been furloughed. What we have already been able to do is look at how that compares to the make-up of individuals in those areas. There are a number of different ways that we can get answers to the sorts of questions you are posing there.

Q114       Kim Johnson: What gaps are there in the data? What data do you think the Government should be collecting?

Liz McKeown: We already discussed earlier that it is easier to be able to produce an analysis where the size of the protected characteristic group is larger, especially when we are using survey evidence. When you have a survey, if it is a relatively small size, you are constrained by how far you can look at different groups as a result. In essence, the larger the sample you draw, the more likely you are to be able to look at the different groups.

As we look at the different protected characteristics that we have looked at today, it is nearly always relatively easy to look at age and sex differences and also to look at ethnicity and disability differences. It then becomes harder to look at some of the other differences.

There are different things that need to be done there. One is that we have to make sure we are asking for the information, and ONS plays an important role in that respect by making sure that we are asking for the information in the same way. We have harmonisation teams in the office who craft a way of asking the question so that, when you are asking a question about disability or ethnicity in one survey, you are asking it in the same way in different surveys, and the same in administrative data systems. That then allows you to make much more robust comparisons between different sources of information.

Q115       Kim Johnson: My final question is to both professors. If you could ask three things of the Government, for the short term and further into the future, what would they be?

Professor Sir Michael Marmot: If I may, I just wanted to respond to your pension question. If you look at the social gradient, not just in life expectancy but in disability-free life expectancy—healthy life—then, if the pension age became 68 rather than 65, which is Government policy, twothirds of the population do not have disability-free life expectancy as long as 68, looking at the social gradient by deprivation. I said earlier, as you know well, that women have longer life expectancy, but they spend more of their years with ill health than men. The healthy life expectancy, or the disability-free life expectancy, is more of an issue for women. Making the pension age later will have a big impact on women because, as I say, given the social gradient, the more deprived two-thirds do not have disability-free life expectancy as long as 68.

In relation to the ask of Government, the No. 1, which I have already said, is that I would like the Prime Minister to chair a cross-Government strategy, involving the whole of Government, to advance a health inequality strategy. That was the No. 1 recommendation of my report. As you know, the Government have a stated policy of advancing healthy life expectancy by five years by 2035. They cannot possibly achieve that without addressing health inequalities. To meet that stated Government policy, which was already there before Covid-19, the No. 1 should be that the Prime Minister should chair a cross-Government activity to elaborate a strategy for dealing with health inequalities, which will relate to the aspiration to advance healthy life expectancy by five years by 2035.

The second, as I said—and it relates to the first—is that all Government policy should in a sense go through a health equity screen. All Government policy should be looked at by its likely impact on health inequalities. That is environment, economic, employment, education—every sector.

The third is to please read my report.

Professor Rasul: I just have one ask from a research point of view. All evidence-based policy is only as good as the data that it is based on. To reiterate what I said before, we need to continue to invest in both administrative datasets and some of the great survey evidence that we have in the UK, and also to think about how we can best link those different datasets together and give researchers access to that. It is only by looking at the same individual over time or by linking parents and their children that we actually get the whole picture of how the crisis is affecting households. That is what has made this crisis so different from anything that we have seen before. It is affecting households in so many different ways.

It would be wonderful if we are ready for this for the next timesomething like this probably will happen again—so that we can link across these different health, economic and family dimensions. That type of investment, in terms of the ability for us to then think about what the issues are and to have the discussions that we are having today, will generate huge rates of return in just the possibility to design better policies, because we just have a more holistic picture of all the different dimensions through which households are being impacted. That would be my main research-focused ask of the Government.

Q116       Chair: Can I conclude with a final question? I am going to direct it at Imran, but I am very happy for either Michael or Liz to chip in if they would like to.

Right at the beginning you spoke about the Zika outbreak in Brazil and said that the response of the Brazilian Government did not include a gendered response to that crisis, notwithstanding the fact that we knew that it was a condition that affected pregnant women specifically. That led to many women still, traditionally, having to go out and get food and cooking equipment. Knowing what we do about Covidthat it affects men more than womenshould we be thinking about having a gendered response to Covid?

Professor Rasul: One of the biggest puzzles in health economics is how people respond to information about their health and the health of individuals around them. In many other dimensions of life, when we are given information, it is fairly predictable how people are going to use that information. When it comes to information about your own health, all kinds of behavioural biases tend to enter.

What we documented in the Zika context is that not everybody responded to the health information that was provided. That may be for a variety of reasons. For some women, it is incredibly costly to delay pregnancy, especially those at the lower SES end of the spectrum. They are as likely to become pregnant after the Zika outbreak as before. It is higher SES women that choose to then delay pregnancy, perhaps because it is less costly for them to change the timing of births.

Conditional on becoming pregnant, we see similar changes in behaviour during pregnancy of high SES and low SES women. It is not that low SES women do not have information or do not respond to it, but how they can respond to it might vary.

We then also find that there is a group of individuals who became pregnant just before the health alert was made public, and they do not change any of their behaviour during pregnancy at all. They are not more likely to ask for an ultrasound or an abortion or to go for more health checks. It is almost as if some people tend to ignore the health risks that they are subject to when, in a sense, they are already subject to them.

That just fits in with a larger body of evidence that, when it comes to providing information to people that may need to be targeted or may need to be tailored, you might need to utilise other evidence in terms of how exactly you try to persuade people and convince them of certain messages. There will not be a uniform way to do that across different types of households, and that is something that comes up repeatedly when we try to study how people respond to information about health, whether it is that you should not smoke or that you should wash your hands. These are things that everybody knows, but whether they act on that information actually turns out to be quite puzzling in terms of the extent to which people do not.

Those are common findings that we have. I have no doubt that they are also going to apply to the extent that people are adhering to social distancing information that is provided to them or other practices that have been suggested.

Professor Sir Michael Marmot: It is an excellent question. I would respond in two kinds of ways. One is that there are some issues that are clearly gender-specific, such as the apparent increase in domestic violence consequent upon lockdown if people are being cooped up in confined spaces and more women are calling for help. That is a very gender-specific type of issue.

There are then general issues that might have gender-specific effects. I pointed to the decline in life expectancy in women in the most deprived 10% outside London. In my report, I did not offer an explanation for it. I was just troubled, but I did not offer an explanation. It was put to me after we published it, and I looked at the House of Commons Library report from 2017 that claimed that something like 80% of the brunt of the policies of austerity had fallen on women. I did point in my report to single-parent households, because that is usually women, and cuts in child benefit and the like. There are policies that will have bigger effects on one gender than another, not because they are designed that way, but that is the way it turns out.

In other words, there are two types of issues: some that are gender-specific and some that might have differential effects on gender. I would say with gender that we should do the same as I was calling for when I said that all Government policies should be looked at for their likely effect on health equity. I would include gender in that.

Q117       Chair: Thank you very much. Liz, did you wish to conclude with any further comment, or are you happy?

Liz McKeown: I am happy, thank you. The only thing I would add to that is it just shows the importance of having your data disaggregated by those different characteristics, which is what ONS is committed to doing.

Chair: Thank you to everybody who has taken part, particularly our three witnesses. I know these are weird circumstances in which to have to give evidence to a Select Committee. That concludes our meeting.