 Right, we should probably start. So, good morning everyone, and a very warm welcome to this 2023 edition of the labor force and annual population surveys, user conference. Now, I am aware that most of you know already about the UK data service, but we also think that perhaps some of you don't. So I am going to just very quickly summarize who we are and what we do. So basically the UK data service is a comprehensive online data resource funded by the SRC Now UKRI. It is meant to be, and we can say it is a single point of access to a wide range of secondary social data science. It is probably the main one for the UK. And in addition to being a provider creator of data, it is also a provider of support training and guidance for users. Some of you may have attended some of our trainings. What kind of data do we hold? UK surveys, large scale government surveys such as indeed DLFS, cross sectional or longitudinal surveys. So either major population snapshots, again, as some of the DLFS is, or survey that follow individual people over time. We also have international data with a multinational survey or aggregate database. We have a treasure trove for researcher of census data with a modern, the recent ones and historical, business and administrative micro data. And what some people know probably a little bit less, some also qualitative data. Our users used to be mostly academic researchers and students, but increasingly have also become government analysts, charities, research and charities and the voluntary sector, business consultant, independent research center and think tank. Now about the labor force and annual population surveys, again, a very quick word for those of you who are entirely new to them. The labor force and annual population surveys or rather the labor force survey is actually celebrating its 50th birthday or anniversary this year. So it was first data collection was carried out in 1973 under regulation derived from the Treaty of Rome following the UK accession to what was then the common market. It is used, it's one of the most widely used surveys by researcher, government department and government agencies. And it is regularly mentioned in the media. So it is used for a wide variety of purposes and not just following the trends in the labor market strictly. And I think the presentations we are going to see today will demonstrate that it comes in several flavors. And I will just mention the most common ones here. So the cross sectional quarterly labor force survey whether as an individual level data set or household data set is the best known one. We also have the longitudinal version of the LFS, so two and five quarter longitudinal data set. And finally the annual population survey which is made of component from labor force survey data and booster samples. You may want to know that the quarterly labor force survey series is actually the most downloaded data series that we hold at UKDS. For example, there were more than 12,000 downloads for half of the year of 2022. And it's not for load. So the five quarter LFS is not very far from that number. Another reason why this year is a special year for LFS user is that the LFS is changing and is going to become in summary, if I'm correct, the transform LFS. So this is something that will be presented in detail by our guest and co-organizer the Office for National Statistics in a little while. About the program of the conference. So in a minute, I will introduce our keynote speaker and after the keynote presentation and the discussion at 11 o'clock, we will start the presentation by the ONS, the data producers who will give an update on current still LFS issues and most importantly, the transformed labor force survey. We will then break up for lunch and in the afternoon, we'll have the breakout sessions. So I'm not obviously going to go through all the papers here but just allow me to mention that there will be four sessions on mobility and transitions, long-term trends and structural dynamics in the labor market, methodological innovations and subjective wellbeing. So the time has now come for me to introduce our keynote speaker. I am delighted to introduce you to Richard Blunder who's Ricardo Professor of Political Economy at University College London and also researcher at the Institute for Fiscal Studies. It would be impossible for me to summarize Richard's career. He's one of the big names in labor economics in the UK and he has worked on many topics, too many to summarize here but the presentation he's kindly accepted to do for us today is focusing on inequality, registration and wage progression, especially from point of view of training for low qualified workers and a piece of research that was done among other with labor force survey data. So Richard, I will give you the floor now. I will stop sharing my screen and we are listening to you. So I will open the floor for discussion when you finish. Thank you very much, Pierre. Thank you, everybody. It's fantastic to be here. Well, Manchester is one of my favorite cities so I'd vote for in-person but I can see the advantage of the hybrid and online too. It's great to be here, particularly thinking through the value of ONS and UK Data Archive and the data you produce, which is incredible and happy 50th LFS, I hadn't realized it and I hadn't realized about the transformed LFS so I'm looking forward to the rest of the day as well. A very exciting program. My talk is fairly broad. I hope that's okay. Anyway, I've sent the slides. They're clicked on the program if you want to look at them and there's a whole set of references at the end if you would like to follow up. So what I'm going to do is share my slides and see if it's all working and how is that, Pierre? It's all working so everything is fine. Wonderful. Thank you very much. I just turned my video off, especially as my line might not be perfect, although I hope it is okay at the moment. You can see here my beautiful UCL background here. The first thing to note here is that I've put reflections from the Deaton Review because this work is happening really as part of the ongoing Deaton Review which has been going since 2019, just before the pandemic but we've kept going through it with all the excitement at least from the research point of view of the events that have then happened. So there's a click to the website there if there are things you want to follow up on there. I'm going to give a little bit of background to the review to kind of set the scene and then dig into the detail of what I want to talk about here. It's a five-year study as I said, started in 2019 and we're coming up, I'll show you in a minute to the second phase where the panel are writing a review. So it's an absolutely ideal time to be talking to the data and research community because one, we've used the archive and particularly the LFS and APS data throughout and it's been remarkably important of course. It's also the case that we're looking for input now and so it's particularly valuable if you think there are things that you could contribute as I go through some of the key ideas. Do get in touch with us or particularly with me. The review itself is very broad and that would take hours and hours to go through. It's set the scene about what inequality is matter most. My remit is really around the labor market and redistribution given my interest generally and the overall aim is to come up with what's the right mix of policies perhaps not the particular tax rates and minimum wage levels and so on but to get a broad idea of the balance in a post COVID world. Although most of what I'll talk about is 25 years before COVID and we're using that to set the scene in terms of labor market and what have you as we run through a post COVID world. It's not just for the UK by the way although the UK is the focus it's a running example. It's an interdisciplinary project and so that's particularly useful to get your input because I know this group particularly is an interdisciplinary empirical group and that's very important to us. I won't go through everybody but just to show that we're serious about this or at least as a majority of economists but it is true that we've apart from Angus in the chair I'll point to a few people like Lisa Berkman who's an epidemiologist, Kath Kien and many of you know from York who studies families. Lucinda Platt is sociologist and Deborah Satt who's a brilliant philosopher of inequality in Stanford and the group of economists is not just those that you might automatically think of looking at inequality. We've got people who study trade and study markets as well but my main emphasis here will be on the labor market and redistribution. There's a whole set of studies which are all now published. The first round was to commission studies that are all now available on the website which is what makes it such a useful time to be discussing this and the panel are now writing the starting to write their overview. They have one year to do it or six months really and my interest here is gonna be on inequality, redistribution and progression, particularly progression in the labor market and particularly progression of those that come from less educated, formal education backgrounds and that's important progression is a key area to think of in terms of gender, in terms of geography and place, in terms of labor markets and in terms of redistribution and in all those areas of course we've extensively used the data that we're discussing here today. There's two volumes, they're all online of course and these first studies are all there for you to look at. We're writing a volume with a bit of help from 17 other studies across countries that we also set up with the hope of UKRI and the Nuffield Foundation. As I said, the LFS APS has been obviously a central plank so has many other data sets and I'll be combining in fact in a sense the linkages and combining are I guess to all of us really empirical users perhaps the most exciting thing that's happening in the world of empirical analysis and so the big collaborations that we've done through UNS and the data service have been essential and I'll point to some of those. There's a whole set of studies, I'm not gonna go through them. I've focused a few that I'm gonna be pointing to both studies that I've put together myself including a recent paper but there's a set of ongoing studies on these issues of progression. They're all fairly econometric studies but I'm not going to go into the detail of that here but I'd be delighted to pick up on any points. Through the review itself, the commissioned studies that might be particularly relevant here that you may have seen published over the last six months have been the study on labor market inequality which I'll draw on extensively by Steve Machen and Julia Gioccapponi and the geographic inequalities that Henry and Xiaowei and IFS have produced very much using LFS and APS data. There are many other here including the gender study and firms and what have you all available online. The long-term challenges that we think of here but if everybody can see it, I've got the thing right at the top of the screen but it doesn't block off much. Anyway, of course the kind of key underlying issue in a way is earnings inequality which has gone in waves but still we're facing over the last 25 years we've seen increasing earnings inequality and poor late wage progression for the lower educated workers in particular those in part-time work and we have some fantastic studies today I notice already that are looking at part-time work because it is key. It's always been a key issue for women and it's become more of a key issue than it used to be for men especially those towards the bottom of the hourly wage distribution and it's become a component of the growing solar self-employment which I'll come back to. This means that we've had diverging life cycle wage profiles by education and compounding that is the low levels of on-the-job training for those less educated workers and both progression and training go together and they're very complementary with your initial levels of education and that's gonna be a key point here and I'm going to look at the end of geographical variation and the importance of having a mixed labor market really around you when you're in your initial stages of the labor market. The growing solar self-employment platform work and outsourcing is a key feature across many economies but we know that alongside those goes very low rates of on-the-job training and what turning out to be fewer parts to what you could call good jobs I'll be more specific about what we're thinking of as good jobs but it's basically jobs that produce some reasonable career progression for non-educated and non-university educated workers so either those that don't leave school effectively before completing A levels are equivalent or don't go to college. I'll come back to definitions in a minute and what we've seen over the 25 year period is increasing in work poverty. If you're as old as me, you may remember the days when we thought employment was gonna be the thing that got everybody out of poverty and into self-sufficiency and welfare to work that kind of thing. It was not very successful. We've certainly got reasonable levels of employment but we have rather low levels at the bottom of hourly wages and job quality and so employment alone has increasingly been not enough to escape poverty and low earnings which was a surprise I think to many who designed welfare to work policy in the beginning of this 25 year period. The increasing family earnings inequality even though we've seen that female labor supply increasing I look at that briefly using LFS and FRS actually. The part time work and gender gap, child penalty and the assortative miscoupling of course it's very assortative on wages and earnings. It does little in fact to reduce earnings inequality and of course not surprising almost the main topic of policy today is the large differences in prosperity across regions and that's where the detail the geo detail of LFS and APS has become so key in trying to understand what's going on. And in a sense the way we're thinking about this is that we can't address all these concerns through the tax and welfare system alone. So we have to think of a balance of human capital, minimum wages, regulation, place space those kind of things. So that's the kind of remit. I hope everyone can see Pierre if there's any problem to let me know and I'll try and solve it. But I'm going ahead as if everything is hunky-dory as they say. So, perfect, great. Yes, so if we go just look over the last 25 years this is study this basic use in FRS I've got the same figures with LFS. But it's just to look at the growth in now weekly earnings it's the kind of annualized real percentage growth from five to 25 percentile. And you can see overall there's increasing earnings inequality certainly at the bottom there's been relatively little increase at the top there's been quite reasonable increase. Of course this varies over decades and more recently the minimum wage particularly has seen a boost at the bottom. In fact, if I put on hourly wages here this is the first remarkable picture. Hourly wages at the bottom have seen reasonable growth especially recently. And you can see that it looks rather different between hourly wages and male wages. And that's the kind of first surprise in a way although nobody here will find it surprising. As I said, there's a great paper later on looking at part time because what's happened is at least until around 2011-12 there was a growth in part time work among the bottom quintile of the distribution of male workers. And that was a surprise. It's still there at a relatively high level. It hasn't increased particularly much recently but it's there especially among younger people overall and low wage quintiles over this whole distribution which tend to be less educated. In addition, if you focus in on self-employed especially solo self-employed you'd find quite a big increase there as well. The other lines here for the other quintiles where you see very little growth in this increase in relatively low hours among men and that's obviously a key issue because if one policy is the minimum wage the minimum wage doesn't do much good if you'll work in relatively few hours at least to household overall earnings and I'll come back to that in a minute. But let me just dig dive into what's happened to employment composition which is very key for us and it's a common feature across many economies. This group, the solo self-employed which is self-employed but without any employees so it's kind of new modern solo self-employed and not the old type of self-employed who would have their own employees. Of course, this is bimodal. Hedge fund managers are typically solo self-employed but there aren't that many of them. They're right at the top. The big growth is at the bottom of solo self-employed which often involve contracted out work, some forms of gig economy and many others and I'll come back to that. But you see the growth there is really key in the composition. And if we look at the well-known increase in self-employment in the UK over a long period it's completely driven by the increase in solo self-employed, quite a different group. And as we know, these individual workers are not covered by the minimum wage, they're not covered by sickness benefit and they don't pay NI. And so there's a whole set of issues around that and they get very low, almost no rates of training. If I look across the world, you can see that we're not alone in this. We're quite a high number. If you look right on the right-hand side, GBR is the name that Julia and Steve gave but you can see this is a fairly common phenomena among both the high levels of self-employed but solo self-employed. If you then bring women into the picture it's quite different. It's a kind of interesting symmetry if you want in that female weekly earnings over this period have grown quite rapidly at the bottom from a very low base. And their wage rates, that's the kind of stronger red line hourly wages have also grown, partly because they're, of course, are much more influenced by the growing rate of the minimum wage and together with increasing hours of work at the bottom for women, you can see that that's offsetting and you might think, oh, that solved the problem. At least we look at couples, perhaps inequality has fallen in earnings, not after tax and transfers, but earnings. But because of surditiveness and the low, still the low share of female earnings in household earnings, when you put the picture together and the green is the annualized growth over this period of household earnings before tax and benefits, the red line is what we think of as disposable income if you want over this period, you can see there's a big drop at the bottom in real terms and it's completely fixed at least in a rough sense through transfers, in particular in work transfers because most of employment has been pretty robust but wages are relatively low, earnings as we saw relatively low and this is propped up through WFDC, the tax credits are now universal credit and if I look at the 9010 here, just read off the 9010, you can see why generally it looks like the 9010 and even the genie actually haven't moved hugely over this period but if you look at the 9010 for household earnings, at least for this set of with a working person in the family, you can see that that will have grown much more substantially. So what you look at naturally is important here and so that's a kind of setting there and what it means is that as we look at least pre-pandemic we're in a position where there were many, many more higher number of those working in work poverty and that's been of course the key, one of the key issues that we think of in dealing with inequality and poverty. What I want to do in bulk of what I've got left which isn't very long so I'm gonna be rather speedy but give you a flavor for this is look a bit more deeper into wage progression because it's the key story about labor market inequality or one of the key stories and concerns about it and what to do about it and what we've been doing and it's been a kind of major plank of my work over this period and I can see it in the papers today is the role of education, labor market attachment, part-time work is key and that's been very important. I'll come back to this but when we look at women of course a huge increase in education and a relatively faster increase in education for women than men of course but still relatively low levels of full-time work by still quite high levels of part-time work and still some occupational segregation which have left the gender gap once you control for education still quite large. In fact, we've seen no change in the gender gap in the UK once you control for education and that's been one of the most worrying studies, worrying things about the UK and so one thing we've been looking at is the role of human capital investments during working life. I'll show you the low progression and part of that low progression is partly occupational and geographic segregation but also the opportunities for training and the type of firms that low less educated workers are in and so that takes us to the final issue and that is the role of skills and firms. Once upon a time we used to think all these policies were very individual based but increasingly we see a role for firms and of course this takes an extremely geographic look because the distribution of what we're going to call good firms and bad firms is very geographically concentrated as is the geographic decomposition of educated workers. So what we've done is update our work exploiting household data, LFS particularly but also employee match data, the ASHI data but what's been key is matching through four digit occupations and geography data sets like LFS with that data set and that's where I said the power of linking data has become the driver of much of our work enabling much of our work and we want to look at policies here. If I take the LFS and just average over this pre-pandemic period there's no econometrics here, nothing fancy but this just gives you for hourly wages like look at earnings it's not that different it's more accentuated actually. Across three groups those that leave school with GCSEs or less but no not staying for A-levels kind of A-level group and then degree group. We use these three groups because internationally it's a comparison people like to make but obviously we can dig in and we do dig in much more detail at qualifications but at the bottom you can see two features one is that there's a small amount of if you could call progression at early ages but for women that stops very flattens off very early and for men it's not great it doesn't flatten off so much and it wouldn't surprise you to learn that where it starts flattening off for women is exactly when children come along and so it's lower level and it kicks down the distinction the gender gap really opens up so gender gaps and progression are very important as well as you can see education differences but it's not that the gender gap is less for university educated in fact it's very important there the progression gender gap is really key at the top as well because progression is much more accentuated high progression for university educated workers and women do particularly poorly there and again it's typically the type of occupation and part-time work that they end up in and so this has been this very simple point it's been very important in trying to understand what's going on and design policies to do something about it a little bit on the gender gap in fact the blue line is the gender gap over this period it just falls uniformly so we've had success in the gender gap but actually most of it's coming through education so if I control for education here this is 20, 55 year group that's the red line it's pretty flat and so what it really says is that education's been the great reducer of inequality for women but within education there's been very little that's happened and that's a very important point it's not been that case for other economies by the way although there's something of that here we can see using an affestator again that the gender gap in education levels has really changed across all groups so we know that women have done better in education uniformly over this period and in all groups now have either equaled or overtaken their male counterparts but when we look at the kind of work they're doing you find that employment rates are much much lower of course among those with I'm gonna call it less than high school think of that as the GCSE group and i.e. they don't have the kind of A-level qualifications so we got this three group split you can do it in many ways you're gonna see this happening all along is that the reduction in employment is very related to education and related to children and the lower educated groups have their children on average earlier and the degree groups have them later and you can see the dip in employment much less severe of course for educated but it's still the case it's quite shocking in a way that the part-time employment is very prevalent across all these groups particularly for the less educated but even so for the degree group and that's gonna be important part of that is just the way our system has worked and it's partly the way our labor market is warranted and the way our benefit system works for those of you who remember the great expansion of WFTC to get people into work and to support their incomes this was a typical budget constraint as it came in in 2000s very nice for some econometric analysis because you can track what happens over time but there's a big supplement at part-time work 16 hours actually and a little bit at 30 in the WFTC reform which kept going right the way through until recent Universal Credit reforms Universal Credit looks potentially good for encouraging hours but if you look at the rates in fact it's rather better at encourage low hour jobs and we haven't seen any big shift over the recent period with Universal Credit to more hours of course different folks in the data this is why you need to use some data that can incorporate taxes and welfare some folks are going to be facing depending on their housing costs very different constraints and so that gives you some variation which allows us to look at very incentives that drive part-time work for different households in different points of the income distribution and across different geographies actually because of housing costs and we use that kind of variation in the way it works out over time to try and get some causal analysis of part-time work and experience and all those things on wages and wage progression but if I look at the data across this period you can see it really works I mean whatever data set I look at LFS including if I look at best educated women with children versus those single women who weren't eligible for that incentive you can see the pile up at 16 becoming very clear as that reform came in place and there's not some work on that of course what we find is that in fact that does a lot of explaining and not everything but part-time work and occupational sorting you could say does quite a bit in explaining the raw gender gap but not everything but that's an important component so if I was to summarize a little bit of what the wage progression has suggested if I can have five women that would be good Yes, of course what we show is returns to experience a strong and complementary with education and there's implications for welfare reform because just getting people into work isn't enough anymore what we need to get them into work that has good progression at least that's one important thing and what we do in the two bits of work that the next bit of focus which I'll go over a little bit is first look at the role of training and here again this is using UK HLS another data archive sort of data source to look at incidents of training and I've got the exact same thing from the LFS as well which is particularly useful and it just shows that low educated or those that leave secondary is the GCSE group I've drawn this from a general labor economics paper that we wrote last year and what it shows is a very low rates of training this is at 50 hours or more of training typically work related training we're looking at here which is quite well recorded in that data you can see women have a slightly different profile there's a little bit more when they come in back to work or increase their work after children and we use these variation match with the variation in training intensity across industry and geography over this period from the LFS and to get some if you want to call it shift share variation to look at impacts on wage progression and what we find is that perhaps you'll think that should be the case but there was certainly some view that training wasn't very effective we find that workplace training and we think qualification is another key here has a really important impact on wages and wage progression and given the low levels of training and obviously the market failure that happens for individual training by firms there's a strong role there and we know that and it's particularly worrying after Covid because we've seen huge drops in vocational training apprenticeships starts over that period and then finally the role of firms I'm not going to detail here because this is a big matching of lots of data sets here both using the European Work Condition Survey the labour force survey matched by at occupation level to ONET and the ASHE data because you have to match all these things to really get a progression when you want to look with it inside firms but what we find is the kind of new result perhaps is looking at less educated workers working in occupations where soft skills are particularly useful seems to give both higher wage progression and more likely to receive training and of course soft skills are becoming increasingly important what we define as soft skills and again I won't go into detail but we match this across LFS definitions of occupation with ONET and ASHE it's about working with the group working with teams things that are fairly hard to train people and fairly hard to identify but they seem to really pay off and in fact they're the ones that have been increasing this is just looking at LFS in fact three at the top that we think of as social skills it's the employment in those particular those particular things where they're high that's high levels of social skills teamwork and problem sensitivity they're the areas where we've seen a lot of growth for less educated workers and that's a very important part of what we think of as important here and indeed it's very very related to what people think of as good jobs progression and soft skills seem to go together here and more training as well so something that and here's the results where if we look at those in soft skill if we look at the wage progression that I put up before this is for men and look at less educated men if we look at soft skill occupations they're way getting way more progression same for women by the way so that's the way it is and what we've done there is kind of think about why that's the case and what we find is the type of firm matters and this is where the matching really matters because it turns out that if you're working in firms where there's a good mix of educated groups that's where soft skills really matter so it's a kind of complimentary here and this ties in with the last point which is about geographic geography because we know from LFS that over this period even though the share of graduates has increased regional disparities in education levels are huge really quite extraordinary large and they've been maintained and perhaps the biggest thing we know is that there's both lower levels of education and higher education qualifications and also education flight so the three places I like to pick on just partly for fun is Grimsby, Skagnes and Butte we know them politically have been very important think of Brexit, think of the Red Bull all those things and if we look at this combines Leo data with LFS if we look at not only the low levels of education qualification but the leaving of educated groups to the thriving cities like London and Bristol, Brighton those kind of things it's really key and so this has been very important because in those areas it's going to be particularly hard to get wage progression to work because there's not the type of skills locally that can do this so we know that's about agglomeration I'll give you some examples but if I go to somewhere like Skagnes you can see the share of people with degrees is low but it's the share of adults in the same cohorts just using LFS and APS to follow the code is much less so there's a huge net loss to other areas once people are educated so that's important so that's where I'm going to end sorry Pierre for going over but I wanted to give you a feel for the way we're thinking about this because we want to move away from just relying on in-work benefits and tax transfers and minimum wages they're all key they've all worked of course but they haven't worked well for progression and what we've got to think about is putting a policy agenda that can flesh out this good jobs idea around three things one is about the kind of skills and training the other is about the form of work and solar self-employment and new world of work and the other is about productivity and place-based policies that can build the environment both to reduce education flight but also to enhance wage progression for those who are left in those areas so I'm going to finish here in the slides there's a whole set of what I call post-COVID policies because obviously this work was done prior COVID but it sets the scene and almost everything I've said here gets accentuated during COVID and I can go into that but I better shut up Pierre because I've gone on too much so thank you very much Richard