 I think we best start. So in this session, we have two papers. The first is from Rob Calvert-Jump, who is a researcher at the University of Greenwich. And what he's gonna be talking about is neighborhood renewal funds and whether they improve local labor market outcomes. The abstract is there if you want more detail, but I think what we can do is move on to Rob. So I've given the speakers a bit more time if they want it for their talks, but hopefully they won't use lots so we've got a bit of time for discussion as we only have two papers in this session. So over to you, Rob. Yeah, thanks Nigel. Let me mute my video and... Okay, there we go. Okay, perfect. Thanks Nigel, thanks for the introduction and thanks for inviting me. So this is a paper co-authored with a colleague from the US on new labor's flagship attempt to improve outcomes in deprived local authorities in the early noughties. Oh, there we go. A bit of boilerplate, so the views that I'm gonna express are... Right, so what I'm gonna do is I'm gonna give you a bit of background on the National Strategy for Neighborhood Renewal and the Neighborhood Renewal Fund, what they were, how they related to one another, talk about the questions that we are looking to answer and why, and then we'll look at some results. I will try and keep everything right. So what was the National Strategy for Neighborhood Renewal? Basically it was the last serious attempt by any government, at least prior to levelling up. To level up. It was launched by New Labor in 2001. It was a relatively broad-based sub-regional development which targeted 88 of the local authority districts in England. There were similar programs in Scotland and Wales which we're not looking at. So we're focusing on the NSNR as it applied to England. The 88 local authority districts were determined by the Index of Multiplication, which was briefly mentioned in the last session. The Index of Multiplication was actually developed at the same time, pretty much by John Prescott's office. The goal of the strategy was to improve relative outcomes with respect to health, education, crime and employment, worklessness at least, in these local authorities. Although in principle the National Strategy for Neighborhood Renewal applied to the whole of the country, part of its rationale was that mainstream services weren't particularly joined up and they were very badly joined up in deprived parts of the country. If you could somehow join them up and make mainstream services, nationwide services weren't better together than you could probably improve outcomes in health, education, crime and employment in deprived parts of the country. But certainly they focused on 88 very deprived local authority districts and the Neighborhood Renewal Fund was the funding arm of the National Strategy for Neighborhood Renewal and dispersed about 3 billion pounds to those 88 local authorities. Between 2001 and 2008, 2008 was when the programme was wound up and replaced by a similar but slightly smaller programme under Gordon Brown. So we're talking about something that lasted a long time. It targeted quite a number of deprived areas and the amounts of money that were involved are similar to the kinds of sums that the current government is talking about when it's in terms of the town's fund or the leveling up and things like that. I think the town's fund is about 3.9 billion. The local authorities districts that received NRA funding in black and there were similar programmes in Wales and Scotland looking at, what you can see here is that most of the local authority districts by New Labour, by other social districts are either in metropolitan areas or close to them. Exceptional is industrial town districts but even then these were usually quite close to cities which is important when we come to think about policy implications. The really big exception to that was Cornwall which is obviously historically quite into private. So motivation, why are we... Sorry, I'm having trouble with Zoom. Motivation, so why are we interested in this? Well, from the perspective of this conference I guess that most people would be interested in this because we use ONS statistics that aren't used that often in academia. So in particular we use the annual business inquiry data set and we use the old local labour force survey data set as well. My colleagues are obviously Americans so place based policies are a massive thing in America in the US and of course now the Conservative government has turned towards levelling up as a vote winner. That's the narrative. In principle we should be looking back at the things that New Labour did and learning from them, did they work, how well did they work, why did they work so on and so forth. So there's a number of reasons that we might be interested in. This is, I'm almost certain, the first evaluation of the labour market effects of the labour renewal funds since the official Department for Communities and Local Government study in 2010. And there was an interim report in 2008 as well. So it hasn't been looked at for a while at least in the effects of labour market outcomes. So the basic question, the specific question that we're asking is how do improvements in local labour supply through all the various investments that the National Strategy for Labour Renewal and the Labour Renewal Fund promoted, how do they affect labour market outcomes in terms of worklessness and in terms of employment. We largely follow the approach of a paper by Alonso and authors who, to the best of my knowledge, are the only existing academic evaluation of the NRF since the official government evaluations. But they focus on crime. Use a simple difference and differences approach and we basically follow them but we extend their method using a spatial diff in diff to look at spillover effects because we think spillover effects are likely to have existed and a set of identified diff that's robust to various biases that those models can suffer from. But I'll talk about all of that as we go along. So just to recap, before we look at results, we are looking at the labour market impact of the Labour Renewal Fund, which took place on 2008 in England. Sorry, I'm still having trouble with these things. There we go. We're looking at the labour market impact. We're looking at the effect on jobs. We're looking at the effect on employment. We're looking at the effect on self-employment. We're looking at the effect on worklessness, measured by claimants. And we are using a difference in differences strategy. So what you're looking at here, if you are not familiar with them, is basically the graphical output of an event study difference in differences model. What we have on the vertical axis is the percentage effect of the NRF on the dependent variable, which in this case is job counts in local authorities. So 0.02 would be 2%, minus 0.4 would be 4%, minus 4%, so on and so forth. And on the horizontal axis, you've got years in relation to NRF implementation. Zero, year zero, and you can see my mouse. That's here. Year zero is fiscal year 2002. So that was fiscal year 2002 was when NRF started being dispersed. Fiscal year 2001 is normalized to zero. And each of these point estimates, which are the filled circles with their confidence intervals, they are basically point estimates of effect sizes of the NRF relative to 2001. So we take this point estimate at period six at face value. What that's saying is that job counts in neighborhood renewal fund treated local authorities were about 2% lower six years after the implementation of the policy relative to all the other local authority districts in England that weren't treated by the NRF. They didn't receive NRF funds. Of course, that's not statistically significant result. So what this is telling us is that over the lifetime of the policy, the neighborhood renewal fund doesn't seem to have had any effect on local job creation. On the other hand, if we look at employees, the employee counts is from the local area labor force survey and then the elevation survey after that. There does seem to have been a significant treatment effect on employees, right? The end of it, the number of employees moving people employed by firms in NRF treatment areas was about 5% higher than it was in the non-NRF parts of the country. On average, the effect was about 2.5%. If we look at self-employment, it's even higher. It peaks at about 20% four years after implementation and the average affects about 98%. Now putting that into actual numbers, in 2001 in the NRF treatment areas, there were about 78,000 employees. On average, there was about 10,000 self-employed workers. So you're looking at somewhere around 90,000 people employed in these areas with a median population of about 140,000. So the results on employment and self-employment suggest that the neighborhood renewal fund was responsible for the creation of around 3,000 jobs for its resident households in the median treatment district. And 3,000 on top of a pre-treatment total of about 90,000 isn't too bad. Now how does that square with the current null effect on jobs? Jobs are measured on a workplace basis. Employment is measured on a residence basis. But we don't think that the NRF was causing any jobs to be created inside NRF treatment areas, so inside these areas higher back. We do think that it was helping people get employed that lived in these areas. What's the implication? Well, the implication is that they're probably finding jobs outside of the treatment area. That makes a lot of sense if we remember that the NRF treatment areas were in cities or close to cities or in these towns, which were presumably a lot of them, which is in the Adinottes. Moving on to claimants, the impact on claimants looked relatively large, but of course there are much fewer claimants in these areas than there are total employment. Now this result with an average effect of about minus 5.6% on job seekers allowance claimants implies that less than 400 claimants were being taken off the books. It doesn't square massively well with the estimates for employment. Now what's the reason for that? The reason for that is here, basically. Now the identification assumption in difference and differences model is that basically there should be no effect. There should be no difference in the evolution of the treatment variable prior to the treatment date, which is basically year 2002, and then you should get a change afterwards. Now what we see here is that the number of job seekers allowance claimants in treatment areas was increasing relative to non-treatment areas prior to the policy and then started decreasing quite rapidly. Now in principle, we should take account of this trend because if the number of out-of-work benefits claimants was increasing in deprived parts of the country relative to non-deprived parts of the country prior to 2002, in the absence of the NRF, one might assume that it would keep increasing. So these estimates here might downplay the treatment effect of the NRF on worklessness. Indeed, if we implement a technique that controls for this, then we get a much larger effect on claimants and it would appear if we control for differential pre-trains in the run-up to the policy that the impact of the neighborhood renewal fund on job seekers allowance claimants was actually much larger. These numbers imply the neighborhood renewal fund caused at least 1,000 claimants to come off the books for each of the NRF treatment areas somewhere between 1,000 and 2,000 and that's much more consistent with the evidence about 3,000 employment opportunities being created. Okay, so that's our results. That's nearly done really. I'll just summarize what we've got and then talk a bit about policy conclusions. So basically what we're saying is that the neighborhood renewal fund appears to have had no obvious effect on job creation within treatment areas. The total employment in the median NRF district increased by about 3,000 people as a result of the policy, as a result of being found outside of treatment areas. If we control for differential pre-trends we think that out-of-work benefit claimants decreased by at least 1,000 persons as a result of the policy. Now, one of the major contributions that we're making is that that number is substantially higher than the official evaluation concluded. They thought about 750 claimants had been taken off the books in the median NRF district. Presumably the lower number is because they didn't control for differential trends although the model they use is somewhat different. It's a relatively kind of cross section model. It doesn't control for much. Certainly it seems like there was also a bias downwards. We also find that self-employment was the increase in self-employment caused by the NRF was quite high and that is something that was observed to be an unexpected benefit of the Neighborhood Renewal Fund at the time. The people that ran the policies in the local partnerships and the local strategic partnerships were pleasantly surprised by the enthusiasm for self-employment in those areas by all accounts. And that's the outcome of our results. Some actual results that I didn't put in because there's only 20-minute presentations is that we split the effect of the NRF on job seekers claimants age group and the effect seems to be particularly pronounced for younger claimants in education longer but we're not entirely sure. Now at this point in time we use a spatial diff in diff model to tease apart indirect effects and there seems to be indirect effects on claimants i.e. claimant numbers seem to be decreasing in areas that neighbor NRF treatment areas but are not themselves treated. We think that's probably because in the larger cities in which the NRF was implemented and particularly in London people don't necessarily need to go to the job centres that are in their district of residence. It's perfectly possible to live in Hackney and to the closest job centre to be in Tower Hamlets for example or vice versa. So we think it's probably just a simple data definition but we're not entirely sure. There is minor evidence of a potential positive spillover for jobs which would be consistent with people finding employment outside of their area. And finally our results are robust to the inclusion of dummies for different policies implemented by the new labour government which affected similar areas particularly the new deal for communities which is thought to have had some effect on employment and there does seem to be some evidence of a compound effect of the NDC and the NRF on employment in our results but largely our results are pretty robust to that. So final thoughts and I'll wrap up the goal of the NRF and the National Strategy for Neighbourhood renewal broadly was to improve relative outcomes in a variety measures in deprived parts of outcomes education and employment. We already know from Alonso that there was a significant the education, the official education evaluation in 2010 was pretty thorough and from our results we think that the neighbourhood renewal fund also had a pretty significant impact on local labour supply in the sense that local claimants count decreased which is consistent with people getting jobs in those areas but we think that those jobs were being found in outside of the NRF treatment areas. Basically what we're talking about here is something that Helen Ladd used to call a place-based people strategy. So they're place-based in the sense that they target specific parts of the country but they're not place-based in the sense that they mainly focused on say infrastructure which is what the town's fund is focusing on what they did was they focused more on programmes and policies that helped people into education that helped people into employment that asked people what do we need in these areas for young people to stay in school rather than committing petty crime and so on and so forth. That's what the NRF did and it seems to have done it quite successfully. The other characteristic which has been noted to have been important at the time was the fact that there was no tendering, there was no applications there was no lottery system you didn't have to apply for funds to the NRF. The NRF ax anti-identified H8-deprived areas and those areas were given money and people in those areas were included in the conversation about how to spend it rather than the single regeneration budget which preceded it in which local authorities had to apply to central funds and the town's fund and the levelling up funds they were seeing now which has returned to that kind of model. There are clearly lessons here there are clearly lessons here if the current government is serious about levelling up whether or not they want to learn anything from new labour's experiences is not clear but certainly I think that the policy makers would do well to go back and look at how the neighbourhood renewal fund worked and the achievements that it had and that's me done one last point a proposal for absolutely nothing if anyone knows anything of the whereabouts of the old employment record 2 data then possibly know about anyone that works for the Department of Employment in the 80s or anything it's a first and that is me 100% Thanks very much Rob so now I'd like to introduce Tanya from the University of Glasgow who is going to talk about the an analysis of people holding multiple jobs and whether this is a route out of poverty So thank you very much for the invitation to talk today I'm going to be talking about a new piece of work that I'm doing looking at individuals who are in multiple employment focusing specifically on individuals who are in multiple low paid employment and this first piece of work is going to be very descriptive where I'm going to be looking at the trends in this type of employment over time as well as the characteristics of work and the reason that I'm presenting it here is because this work grew from almost three decades of data availability in the labour force survey So the broad research question that has initiated this piece of work is the idea of whether taking on more employment provides a route out of poverty and there was a piece of evidence that one of my project co-members saw from the FRS that suggested that there was a gender imbalance in holding multiple jobs So we went to look in the data we couldn't really find too much in the evidence so in the literature we couldn't find very much in the evidence about multiple job holding but what little we found as I said from the FRS was that it appeared to be for women the poverty rate did not fall when they were holding multiple employment whereas there was some evidence that for men the converse held that either it was able to pull them out or above the poverty threshold and multiple low-paid employment is something that we think may be of particular interest because this type of employment so juggling two jobs as well as other responsibilities may be very difficult for individuals so as I said we started out with basically some scoping work to understand like how large of a phenomenon is this has this changed over time and also trying to go and look at the characteristics of the workers themselves as well as the characteristics of the type of jobs that people work when they're working at multiple employment and as I said before especially in this talk we'll be focusing on multiple low-paid employment so what I'm going to be looking at is the extent of multiple low-paid employment and how it has changed over time going to be looking at the characteristics of workers as well as the features of the work that is done the first conceptual issue that we came across is how would we define multiple low-paid employment defining multiple employment you have more than one employment is very straightforward but how do we define multiple low-paid employment so a standard definition is the OECD definition that it's employment paid below two thirds of median wage one thing that we were thinking about is whether this should be defined defined through hourly wages of each employment or as weekly wages overall and within the LFS and I think this is quite well known that there is because LFS is self-reported data there is a potential for measurement error in that individual's responses may may go and give estimates of their hours of work so we do think that individuals who are paid an hourly rate may have a good estimate of the number of hours that they have worked in the previous week but if individuals who are paid a fixed wage may not actually keep track of their hours as accurately as those who are working on an hourly contract there's also an issue in the LFS that there is a quite a higher degree of individuals who report weekly wages or monthly wages but don't actually report hours so there's more missing observations for hours reported and then we were also thinking one of the issues that whether you fall within a definition of being multiple low paid is also to some degree a choice variable in terms of an individual may be working in a job where their hourly wage rate would not put them into multiple low paid employment but they may only be working a smaller number of hours so they would fall into the definition of low paid employment according to weekly wages and of course it may happen the other way that an individual may be working a job that does on an hourly rate that does fall below but actually works more hours in order to go and bring them above so we have some comparisons that we're doing using both measures at an hourly of hourly wages and weekly wages there are some differences in terms of when we're looking at the prevalence of multiple low paid employment but when we look at the characteristics of workers there isn't as much of a difference when we're looking at the distribution of characteristics of workers and also jobs we also came across and I think again this is well known for people who work with the various survey that capture wages is that there is a discrepancy between wage calculations that use the sources so specifically I looked at the calculation according to the labour force survey the family resources survey and the ASH data the annual survey of hours in earnings the latter is the one that is generally relied on for those headline figures of median and mean wages because it comes from administrative records and so it's not self reported data and it seems to be more objective and also I think the larger size of coverage is larger however when I looked and when I looked into it did see that indeed if I look at the labour force survey especially in comparison to those other sources there is quite a substantial difference so this is close to £1.50 per hour which is quite large and also about £50 per week so we decided to take we're using actually the ASH derived level of median wages in order to do our calculation here so moving on to first of all looking at the trends in multiple low paid employment so as I was saying before because we're using the labour force survey and we're using the quarterly labour force survey as introduced in 1992 we can look at almost three decades of data and this is something that being able to look as far back really goes and gives us a picture of the preference you can see that overall preference so this is the proportion of workers who are working in multiple low paid employment has been coming down steadily over time so it was around 2% like 30 years ago down to about 1% now but this looking at this overall picture really does mask the differences by gender so as we can see that about half a percent of male workers are working within this category whereas for female workers and that level has been fairly constant over time whereas for female workers this is what is driving that decline in multiple low paid employment oh it was something that I was thinking about literally just for the last a few days I do want to go and examine this a little bit further because if I contrast this trend with the overall employment rate which has been upward retrended over time over the past 30 years right and we've seen that this upward trend is slightly higher for women as compared to men so in men over this period this is a difference from of about 5 percentage points between the early 90s and the last data point is 2019 here where an increase of around 5 percentage points so for 30 years for women this is around a 10 percent increase in employment and of course what that is is the denominator of the multiple low paid employment trends that I just showed you so there may well be a compositional effect here such that because employment is rising it may be the case that the number of individuals who are falling within multiple low paid employment is increasing but it's increasing at a lower rate than the number of individuals who are in employment so that is something I'm going to caveat that I haven't checked yet but I'm going to go back into the data to look at that as whether there's this compositional effect whether there's this compositional effect and we can also contrast this with the trend in low paid employment so this has been decreasing over the period we can see a large decrease happen towards the end of the 1990s and this has often been attributed to the introduction of the national minimum wage so and then has been steadily declining over this entire period and again this is something that is really driven by or driven more by that decline in the proportion of women who are in low paid employment but something here to note is that this decline is faster than in the previous graph for multiple low paid employment that we do have a decline that appears to have leveled from the late 2000s for multiple low paid employment although it's carried on decreasing for low paid employment overall so just to summarise those findings there is strong evidence that the proportion of workers in multiple low paid employment has been declining over the past three decades driven predominantly by the decrease in the female rate but I do want to do some more work to look at whether there's some sort of compositional effect here because of that strong rise in the female employment rate we see that the rate of decrease for multiple low paid employment as compared to low paid employment from the mid 2000s which indicates that that was coincidental just tiny wise with the session and the introduction of austerity afterwards so now moving on to the characteristics first of all we're going to look at the characteristics of workers and there's a lot of data points here so I'll just then highlight the ones that I think are quite interesting on this first panel here is all workers so workers in all types of employment and in the second panel on the right hand side I'm focusing on individuals within multiple low paid employment and one of the things that can go and pick that is striking for me is the first indication that male workers in multiple low paid employment are somewhat different to female workers in multiple low paid employment so if we look at cohabitation status or partnership status the indication here is that there is a lower rate of cohabitation and marriage amongst male individuals in LPE compared to all workers whereas the proportions for female are more or less the same when we look at formally married we see a higher proportion of females who were previously married and this could be separated divorced or widowed in multiple low paid employment looking at the impact of children we see that women in multiple low paid employment are more likely to have dependent children but older dependent children and overall more dependent children than their counterparts when we look at all female workers and something that does pop out that for both male and female they are more likely to be in receipt of benefits and this is benefits excluding child benefit payments following on if we look at female workers in multiple low paid employment are different for male workers it really is the younger age groups and you can see this declining proportion of individuals in multiple low paid employment for male workers declining with age well declining with age up to the period just before the age group and the age group and the age group up to the period just before the one up to retirement however for women this is an increase in patterns for women it is older women who are in their 30s and their 40s who are more likely to be working in multiple employment and multiple low paid employment we see little ethnicity differences especially for men there are slight ethnicity differences for women but these are very small numbers and this may also be related to different rates of labour force participation amongst ethnic groups for women and very little difference in terms of UK nationality something that I found a little unexpected was that the relatively high education rates so I will admit that I had a prior that individuals in low paid employment especially multiple low paid employment would have a far lower rate of holding a degree and actually it is fairly sizable of proportions of individuals who work as in who hold a degree and a couple of large differences that we see is individuals in multiple low paid employment are more likely to report having a long term illness and also reporting some type of disability so just to summarise the characteristics that I have been looking at for multiple low paid workers is that male and female MLP workers in the data appear to be younger more likely to not be partnered or never be married they have fewer children and there are no real ethnicity differences whereas female MLP workers are more likely to be older they are more likely to be formally married so separated, widowed or divorced and have more older dependent children and children and as I said there is an indication that they are less likely to be from ethnic minorities but key similarities between the two groups is as I was saying before a higher rate of benefit receipt and more likely to report having some sort of long term illness or disability now moving on to the employment characteristics first of all I looked at work patterns and contract types and what popped out from the data was that the preference of working here in the day was very similar across the groups the one that really stood out is that individuals in multiple low paid employment are far more likely to be working during the evening and this is related to the type of employment that they will be taking and then I looked at the industry sectors and multiple low paid employment is something that is specifically prevalent within arts and hospitality as well as in the service and the caring sectors but consistent with arts and hospitality there is a higher preference of this type of working during the evening compared to the other type of shift patterns we do see a larger proportion of individuals in this type of employment who are working on zero hours compared to all employment but even though it is far higher it doesn't explain this is not the work in multiple low paid employment is within zero hours contract it is still relative to the permanent rates and the majority of individuals report that they have a permanent contract at least in their first job so it is not the case that all multiple low paid employment is on zero hours contract employment. What I did go and find interesting is that when individuals are asked whether they work full time a fairly low proportion in multiple in LLP report working full hours however when I look at the number of hours that they work whether they work the equivalent of full time hours and here the definition of working 35 hours or more is the equivalent of working full time hours we see that although they report less that they work full time when you look at the number of hours that they actually do they are more likely to be working full time hours of course this is more than one job so it may be that they are reporting full time in their first job only they do work on our future then overall for workers in all employment and they report that they are more likely to report under employment so that they would prefer to be working more hours and they are less likely to be reporting over employment also looking at further characteristics of the jobs we see that the union coverage for individuals who are working in multiple low paid employment is lower than compared to overall the rate of working in the public sector is actually fairly similar especially for women slightly lower for men for the first job we see that there is a lower rate of individuals who report working from home and again as I remind you the last date point I have is 2019 this is not related to the changes that happened during the COVID pandemic and they are far more likely to be working in occupations that are at the lower tier of the SOC classification so more likely to be in manual and or elementary occupations looking at the second job in comparison to the first job where there is a higher prevalence of working from home and a similar prevalence of being in the manual and elementary occupations so this really sums up what I was just saying there so as I was saying before this is the first work stream in a larger project where we are looking at women in multiple low paid employment the project contains a number of different work streams so I am working on the quantitative analysis and the first step on that was to be looking at the nature and the extent of multiple low paid employment in the UK and how and whether it has changed over time but the features that have come out of that have led to subsequent questions that I will be looking at in the future I apologise my dog is just coming from his home I am just going to put him outside over here again and one of the things that I want to go into in future work is whether labour markets labour market conditions influence the prevalence of multiple low paid employment specifically what we noticed there was with the levelling out of low paid employment after the great recession is that somehow related to aggregate labour market effects or indeed local labour market effects also over this period of course have been changes to social security measures for instance the introduction of university credit and austerity measures whether that induced any differences to the prevalence of working in multiple jobs also looked at the dynamic nature of multiple low paid employment so is this a short term phenomena for an individual or do individuals move in and out of multiple low paid employment and also relationships with health as I said the indication there was from the labour force survey data is that individuals do report a higher rate of long term illness and disability but is this related in somehow to multiple low paid employment and if so in what direction does this go and also looking at whether caring responsibilities do individuals choose this type of working pattern because they have other responsibilities such as child care or elder care that they have to balance along with their employment but another another role of the quantitative analysis is to inform another work stream which is going to undertake qualitative analysis so this is the first time that I've worked in a multi disability project where we have both quantitative and qualitative analysis and what I have learnt is that analysis is really informed by what we have found so far in the quantitative analysis is good for sharing our prevalence and the extent and associations but it doesn't help us really to understand why so the role of the qualitative analysis is to undertake engagement with individuals who are in multiple low paid employment especially women who are in multiple low paid employment to investigate their lived experiences of this type of employment and the hypothesis so far is to really look at wanting to understand those mechanisms between multiple low paid working and health and also how and whether care responsibilities are at play so thank you very much for your attention there and I'm happy to go and take any questions or discussion now or feel free to email at any time always happy to go and talk about this type of work