 Xasblywch mae'n irweithio gweithio ar gyfan gweithio busnes. Fel fgodig iawn o'r meir oho, rwy'n Fnesa Higgins ac rwy'n ei falch cyfrif yn arferiau i Gwneudu Dath Oedon. Mae'r cysyllt o ddiddordeb, yn gweithi Llyfr Ysbryd yw'r ddysgu ddysgu i gydag ymddars o'r ddweud o gweithio busnes fel ddechrau nhw, oherwydd yr adeb ddysgu gol o gyfifith datigol oherwydd mae'n hollol o ddysgu bofnig o'r ddysgu. Felly, mae'r angen i gyfosiau i gyngorwiynau ychydig, ac mae'n ystod y pethau gwahanolio, ycysylltu cyfosiaeth cwfur sydd yn gweithio pan gennym L'UK Data Service Security Lab, ein Llyfi Gwyrd Gwyrd Gwyff角nod, dweud am y symlchedd dros ei swyddi, gyd-dawr fod yn oed yn yr unrhyw gwybod nhw, o ddiwrs o Gwyrd Gwyrd Gwyrd Gwyrd Gwyrdd. Fe starfodd, mi'n rhaid i atbwynt y Llyfr South o'r Gweithio i Gweithio lŷ. Rydyn ni'n gwneud mewn gangannau Siamgellau Oun S a'r Gweithio'r Lave, bwrs unrhyw gwrs i'r glo yn amlwg. Leiw Llyfr South o'r Lave o'n wneud y gwirlo cyd-ded我要thedol o'r cerdydl yn ddigonol, gweld rwyf yn mwyaf, a'r оna gweithio i'r samell. Y Rhyw Llyfr South o'r Gweithio, pwysig o'r gweithio'r gweithio ar gén y dddefnydd. Rwy'n nhw'n dod yn rhan o gweithio rhannu yn cyfrannu cynhyrch yn gondwy, felly cwrsau y cwrsau'r rhannu yn gilydd. Rwy'n nhw'n nhw'n nhw'n nhw'n nhw'n gilydd. Rhywbeth ar gyfer de timestan incerthau o feithio'r rhan o'n dweud i newid y dŵr i ddaint i gyd-rhywun ymhy leagueol. Bydd y ddweud yn y wneud mewn secure lab... ..y meshir cwrsau, cyfrannu y buser Englandry... ..y cael ei ddwy flyn... ..y y cwrsau yn gwybod yn ddeud... am eu gwirionedd, mae'n gynnwch gemiŽ i gwaith i ni'r gwn iechyd na i'r ddweud a'r cwpowenol yn gwybr adultsen yn y Lundi Gwyllgor Rhaglen. Felly byddwch gweinio gwybod nesaf nawr i gweithio'r gyrraedd hwnnw. Dwi'n fawr. Mae'r Ymddol Llywodraeth wedi'u gweld yma, ac oeddwn ni gofod receive o Lluwyr Côty o Ffoghfaeig o'i bwysig o ddweud y mhwy a oedd hyn – dwi'n gwybod, i ddwy'n gwybod i'w ffoghaffydd. I'm Louise Caughty. I'm Head of Insights Development and Impact at the Secure Research Service and I joined in January, actually from the UK Data Service. I worked alongside Vanessa as one of the directors for 20 years, so really happy to be building on today's fruitful collaboration going forward. It's very exciting and we're absolutely delighted with the huge interest in our data. We have lots of users, we have lots of projects going on the whole time, but having 500 people signed up means we really do have a great interest. The event today provides an opportunity to showcase what unpublished business and economic data we have available and also the opportunities for their use. That's under the legal gateways that we have in place to enable that. As Vanessa said, a lunchtime session, if you want to know a bit more about Digital Economy Act and getting your projects accredited so you can use the data, we have our lunchtime slot where we'll have a short presentation from the Stats Authority and from people from our own research environments where you can hear a little bit and have a chance to ask questions yourself. As Vanessa said, both O&S and UK Data Service do host trusted research environments that are accredited under the research code of practice and accreditation under the Digital Economy Act. That's together with Wales, Scotland and Northern Ireland. We can provide access to hundreds of survey and administrative data from central government. We have around 3,000 to 4,000 accredited researchers using data, which is amazing. In the morning, you're going to hear from O&S colleagues about the transformation programme for economic and business statistics and also some updates from some of our most popular surveys. Following that, please do stay for the day because we have an optional lunchtime slot where you can hear more about getting projects accredited. Straight after lunch, we have a panel where we're interviewing people from BASE and O&S about relationships with the academia, how to facilitate research and facilitate the dissemination of research findings as well. Following that, we've got three parallel sessions where we have nine speakers who will talk to us around how they've used data, which is wonderful. I think, just to say, I hope you have a really good day. It's a really exciting agenda and thank you very much for joining us. Now, I'm going to hand over to Chris Woods, who's going to be sharing this morning's session. Thanks very much, Louise and Vanessa. Good morning, everyone. Welcome to our morning session. I'm Chris Woods and I work at the UK Data Service, where I provide support and training to researchers accessing both business and social survey data sets via our secure lab. I've worked at the UK Data Service for just over seven years now, so I've gained a range of expertise in the business data sets and also in supporting researchers in linking external business data sets, such as FAME, to O&S business microdata in our secure lab facility. I also have a range of experience in secure data access and chair of the Safe Data Access Professionals Group, which is a UK-wide network of professionals working in organisations that provide secure access to confidential sources of data, such as the business data sets. I'm really pleased to be chairing this morning's session where we're focusing on data collection and primary analysis and where we have our keynote presentation. So in that context, I'm delighted to introduce Donna Leung, who's given the keynote today on the Office for National Statistics, Economic and Business Statistics Transformation Program. Donna is the Director of Economic Statistics Change at O&S, with responsibility for overseeing the Economic Statistics Transformation Program. Donna also heads up O&S's work on subnational statistics and analysis, levelling up, and experienced senior civil service leader and economist. Donna has previously worked at the Department for Business, Energy and Industrial Strategy, Her Majesty's Treasury and the New Zealand Treasury. She has an MSc in economics from the London School of Economics. So welcome, Donna, and very much looking forward to your presentation. Thanks very much, Chris, and thank you for the introduction. I think Heather is going to be running the slides, so we'll just kick off. I think hopefully now you all know who I am. I think we're just going to start off now with a few slides to show you how O&S is organised and how Economic and Business Statistics Transformation into O&S's wider programme. So next slide, please. So the slogan for O&S, the statistics for the public good, we have four themes in our work. We wish to be radical in our decisions, ambitious in what we do. Importantly, we want to be inclusive, recognising the importance of counting everyone so that everyone counts and sustainable. And this is particularly, I think, important given where we are in the middle of a spending review through what is obviously will be quite a tough time for the public finances. O&S itself is divided into these three different groups, including data capability, which is, I like to think of as the engine room of how O&S works. And importantly is the incubator for the integrated data service, which I'm sure lots of you will hopefully be using in the future. Health population and methods where many of our work during the pandemic has taken place, as well as the very successful census 2021, which hopefully you will hear more about the outcomes of the census as we start to work through what to actually tell us. And then finally the group that I work in, Economic, Social and Environmental Statistics, we've just recently brought together work on macroeconomic statistics and public policy analysis. And my economic statistics change is a new directorate within O&S that is focused on really driving forward the transformation process of economic statistics. Next slide please. Right, so I think that's just to repeat actually. So going on to the Ares programme, now why is it called Ares? It stands for Ambition Stratical Inclusive Economic Statistics. And as it says, that's the programme for development for economic statistics from 2021-22. The previous economic statistics transformation programme which had a much less catchy acronym came to conclusion in 2021. So Ares is set to oversee the completion of the economic statistics change journey which began, I think, was the Bean review and Johnson review, among others, and hopefully will continue into the future. Okay, next slide please. And here's our vision. That as experts in our field, we will provide clear and insightful statistics and analysis to important decision making across the UK in a dynamic, inclusive and sustainable way. Next slide please. So what specifically are we trying to do in terms of economic statistics transformation? I think the first is obviously to make sure that our core economic statistics offer the insights that you need in order to make decisions, including better use of data, use of new innovative sources of data like administrative data and improved survey data, but also making available the vast suite of different indicators and data to the public through, for example, the vast indicators work. We want to be flexible to be able to respond to the needs of the day, including the COVID-19 pandemic, but also a big part of my own day job is thinking about the levelling up agenda. And what can we do in order to help decision makers decide what is the best thing that they can do to reduce the level of disparities within the UK? And so there's lots of discussion, for example, over there, there's something like 100 different indicators that you can look at the subnational level, which ones of those are the ones that really matter in terms of levelling up. And then the last is just to note that, obviously, because we are in the middle of a spending review, we are a bit dependent on user users to tell us what you would like to see next, as well as for the Treasury to agree that that is indeed what is most value for money. So a lot of what we're doing at the moment is thinking about in the context of what can be quite a long-term process in terms of improving statistics, what is the general direction of travel, and what can we do when there's obviously a bit of uncertainty over the next three years. So next slide, please. This is just a snapshot of some of the things, projects that fall within what we might call the Ares Programme Board. So within the Programme Board, it's chaired by our second permanent secretary and the director general for data capability, but I am the senior reporting officer for the programme as a whole. But you can see that there's a very wide range of things that are covered by Ares, including perhaps a new labour market survey that hopefully will address some of the issues that arose for the labour force survey during the pandemic. Bix, in fact, in case I'm sure you're familiar with the sort of, you know, I think what someone called the blue ribbon approach to double deflation. We're now, I think, world leaders in that area and the incorporation of innovative data sources in terms of measurement of prices. Next slide, please. So we're now going to do a bit of a deep dive into what is happening in terms of business surveys and statistics. And I think while this is focused on business surveys and statistics, there's obviously lots of parallels across statistics transformation generally. So hopefully this will give you a good understanding of what's involved in making sure that our core statistics, you know, genuinely reflect what's happening in the economy. So next slide, please. So there's five different key strands I'm going to take you through, including new production systems, because statistics is ultimately a way of taking collecting data and turning it into something that is much more user-friendly. New survey collection methods, extraordinary use of administrative data. As I mentioned, it's thwarting the leveling up agenda because every statistic has a subnational component. And then finally, what are we doing to make sure that we are responding to your needs? So what I'll try and do is I go through each one of these to give you an idea of what we're already doing, what we're planning to do, and what do those changes mean for the data, for respondents and for users. Next slide, please. So new production systems. So this is the first strand. So one thing to note is that O&S has a lot of what we call legacy systems. It's important for efficiency reasons that we move those on to much more modern approaches, and in particular developing a new cloud-based platform to enhance our process for producing statistics. And the very first of the Short Invested Survey Service Retail Sales is moving on to that platform later this year. Some of the more established surveys are a much bigger technological challenge, and we aim to develop new systems for those over the next few years. This is something which is incredibly complicated. So next slide, please. What should that mean? It should mean that we reduce the risk of error. We have more flexibility, the content of our surveys. And I think, for example, the next survey is a good example of that, because it was built on the new platform, and that gives us a lot of flexibility to add things into that survey and evolve and adapt that survey over time. And also more streamlined production delivery statistics. So I will admit that I started my career quite some time ago when Lotus Notes and Lotus 23 was the state-of-the-art database. There are still some bits of organisation that are dependent, for example, on Lotus Notes. So we definitely want to make sure that we're moving off those legacy systems and on to much more modernised systems. Next slide, please. We're also undertaking new survey collection methods, so electronic questionnaires, thinking about, in particular, the challenges for respondents in terms of completing those questionnaires, and creating different ways in which respondents can put their information in and then collate that information, in particular, I think, use secure electronic file transfers, editable PDFs. So we know that often it's not a simple task to provide information in a way that reflects both what we think of as a sixth concept as well as what might be for the business an accounting concept. Next slide, please. So what does it mean if we're using these new digital collection methods? What that means is it's easier for businesses to complete those surveys. I think that they're more likely to respond. That gives us more time to look at and improve the quality of the data and ultimately better quality statistics. But also, as I said before, more flexibility in terms of the questions that we ask. And we can look at, because we are lowering respondent burden, hopefully that means increased helpful sizes as well. Next slide, please. I think one of the buzzwords I think is administrative data and even in the time that I've been at ONS, which has only been this year, I think there's been lots of questions around how do we best use EpiData to supplement what we're doing alongside surveys, sources. So EpiData can play quite an important role in terms of validating survey data. For example, in the case of the labour force survey, we'll have a new online survey, the labour market survey, but we've also got admin data through the PAYE data. So that gives us three different sources of information about what's happening in the labour market. Also, it can help us think about what our methods are in terms of something apportionment. And it can also provide us with new interesting statistics. For example, the use of card data or data from HMRC. And in some cases it can replace survey data, allowing us to replace existing questions, reducing our size and allowing new questions to be asked that are more suitable to a survey format. Next slide, please. So EpiData does give us a different view. It's obviously far greater coverage of businesses than surveys. And as I said before, it can allow much more granular breakdown in the extreme to the person level. Some sources of admin data are produced more rapidly, but actually that's not always true. If we think about, for example, tax data, that's often provided with a lag, whereas survey data can capture what's happening at the time. And it isn't always possible to find appropriate admin data for some statistics. As I said before, there are differences, for example, between accounting concepts and statistical concepts. But admin data can also provide data that couldn't easily be collected via surveys. Next slide, please. So I think this is very much a hate-to-mouth journey to decide what is the right balance between using admin data and survey data. And I think the best way I think is to see them is very much complementary. There's also important security and disclosure issues, which means that we need to be careful around data that can identify individuals and their restrictions on how we use and produce statistics on that basis. That's obviously got issues for users of microdata service. And then, as I said before, the integrated data service may allow looking at a variety of admin sources and survey sources and bring them together. But that does raise the issue around security and disclosure and how we can best build trust in that service. Next slide, please. So I'm just going to take you through a few examples of how we've used, in particular, corporation tax data to support statistics. So first of all, the new statistical business register, understanding of versions between different measures of GDP and the annual business survey. So next slide, please. The other thing I mentioned was faster indicators. So there is a variety of different data which we call real-time indicators. What can we best use that data for? I think there's a huge, potentially huge amount of information whether that's mobility, membership visits, traffic camera activity. In my previous role, my boss used to watch the traffic camera on the M20 to see what was happening at Dover, all companies' house, the solution activity. What can we best use that data for? Is it how closely aligned are these statistics to what's actually happening in the economy? And I'm pleased to say that the Economic Statistics Centre of Excellence is undertaking research to help identify what is the predictive power of these real-time indicators and see if it can prove their now casting capabilities. For example, that might help us, for example, to produce more time in the regional GDP estimates. Next slide, please. So the lining of other data sources. So I talked before about using HMRC data. So we already have VAT data which underpins our regional and quarterly GDP ends. Estimates, how can we use that data to supplement faster estimates and can it help us think about new statistics for industries that aren't covered by surveys? Tax credits is another source of information, in particular R&D tax credits. I will say, though, that it's obviously really dependent on understanding how the tax system works and how people are using the tax system. So once again, I think that's the question mark with our app and data. How is app and data related to economic activity which is what we're really trying to get at? Next slide, please. And then I talked about living up, you know, we've touched upon this. So what are we trying to do? We're trying, ultimately, I think what we would like to do is offer a service to people so that they could go into a region, think about what are they interested in, whether that's a particular local authority, level of granularity, or even, you know, we've had questions over if we could look at what was happening on a particular high street. How can we bring together data on a consistent basis to allow people to effectively build their own geography and pull out statistics for the area that they are interested in? That obviously requires us both to collect data on a basis that allows that to happen to organise that data in a way that can be flexibly pulled off the system as it were and then made accessible in a way. So some of you may have seen the work that OSP doing on towns and high streets and Scotland telling or even, you know, NOMIS is an example of how we can make, you know, the vast array of statistics more accessible on a geographical basis. Right, so we know the next slide was going to talk about the impact on respondents. I mean, I've talked before about using air questionnaires and additional PDFs and so forth to help reduce the burden on respondents. We also want to make sure that surveys themselves are simple and easy to complete. We need to assure respondents that their data will be held securely so that, ultimately, we can, you know, increase the respondent rate and make sure that, you know, our surveys remain representative of the general population. And then the next slide I would have talked about is user needs, which, you know, the audience here today is very much a slice of that. We want to make sure that our statistics are evolving to meet user needs. For example, I've mentioned the BIC survey which delivers now really detailed statistics faster than any other source. And we are also thinking about the Digital Economy survey, which is a replacement to the Ecommerce survey and hopefully we will launch that in 2022. So, you know, we've already seen our statistics, you know, respond to, for example, looking at the impacts of EU exit on trade, recovery from the pandemic, all the introduction of new technologies to the UK economy. The government has set out, for example, its priorities as well as the lifting up and net zero space. Really interested, I think, you know, to take the opportunity to hear from you, our users, as to what else we might be thinking about in terms of both, you know, how our core statistics are evolving, you know, and is there additional information or statistics that you would find useful as users. So, I think what we're going to do now, and Heather, I think perhaps you can stop sharing your screen, is we're going to enter the Q&A section. I think, Chris, are you going to look at the questions in the chat? I'm going to read out the questions from the audience now and give the panel an opportunity to respond to them. Thanks very much, Donna, for a really interesting presentation. It's really great to hear about the range of work that ONS is doing around this. As we've mentioned, Donna's now joined by four of her colleagues from the Office for National Statistics who will be part of the panel. So, I'll just let each of them introduce themselves now if we start with John Goff, please. Hi, good morning everyone. My name is John Goff, as mentioned. So, I'm currently working in Business Statistics Transformation division within ONS. So, working very closely with Donna and many of her colleagues in terms of helping to produce more sustainable infrastructure, as you talked about, ingress platforms and other technologies that are quite old. So, working with ESG to provide new technologies going forward in that area. Prior to that role, I've worked for many years within ESG as well, so I have some knowledge of the annuals and the big series as well. Thank you. That's great. Thanks, John. And Richard Hayes. Hello, I'm Richard Hayes. I'm responsible for productivity statistics, our research agenda which covers the ESCO and our relationship with the Turing Institute. And I lead work looking at the future of the system of national accounts with UN, OECD, IMF and other national bodies for the ONS. Great. Thank you, Richard. And Ed Parmer. Hi, good morning. My name is Ed Parmer, Deputy Chief Economist at the ONS. I also lead the analysis microdata and engagement division. Thank you. Thanks, Ed. And Heather Bowville. Hi, thanks. Yes, I'm Heather Bowville. I'm the head of Service and Economic Indicators within ONS which is largely the analysis and production of the majority of the business surveys that a lot of people on the call will be users of such as the annual business survey and purchases and prodcom as well as the short term survey such as retail sales and BICS which we'll hear more about later. Thanks, Heather. Thank you all. So I will move over now to the questions and just read these out for the panel to answer. So starting with what outcome variables are you using to measure productivity? Which ones are the posted influences on productivity and what is at the moment the relationship between the two? Okay. That sounds like one for me. So in our productivity suite we're currently measuring labour productivity. That's our headline measure and that is GVA divided by hours worked or we also have a jobs measure there. So we look at GVA output divided by a number of jobs. Alongside that we measure what we call multi factor productivity which is where we try and take account of capital inputs as well as labour inputs in the production process to generate again GVA. It's worth saying our flash estimates which are the fast indicators that we produce do use GDP as a proxy of GVA in the first instance but in the main we use GVA. And then finally we also publish public service productivity numbers which look exclusively at the public services which are delivered by the public sector and those which are purchased by the public sector and again we look at that in the same type of inputs and outputs relationship way but because it is very difficult to get estimates of public sector output in market prices because many of them are delivered without a price health and education are delivered for free. We look at that in a straight output picture so we also have intermediate consumption in as an input. I hope that answers the question. Thank you. That's great. Thank you Richard. And our second question is how are you measuring the introduction of artificial intelligence into UK industry and do you have any reports or data sources on that issue? I'll take that one. So it's quite good question actually in links to what we're going to have more presentation on later. Obviously artificial intelligence is of growing interest and something that we understand users want to know about. We're in the process of redeveloping our commerce survey at the moment which would be the survey where we could collect that kind of information. So I will rather than going into the detail of that now we've got a presentation coming up shortly after this with Dan Robinson who will talk about the commerce redevelopment and what that's looking like at the moment and hopefully that will cover the answer to the question but if not we can pick that up afterwards. Just to add to that on our analytical side we're working with you to develop a set of digital estimates in line with the OECD's new digital supply use framework that should be coming out later this year and will show what we can manage at the moment but the development of that new survey that Heather's describing the digital economy survey really is absolutely essential for us to be able to fully populate those outputs so hopefully you'll see some progress on that quite soon. That's great thanks and we look forward to hearing from Dan then later on on the e-commerce survey. So our next question is around the leveling up agenda so supporting the leveling up agenda sounds like your focus on government policy. How do you ensure the fullest possible range of subnational estimates are available including for those who want to question the performance of government? So I think that one's me. So I think one of the nice things about indicators and so forth is that they are what they are. Government may well posit a particular target. The truth is that the metrics that are important for leveling up are probably will move probably in years rather than in months. One of our aims therefore is to make a range of data available so depending on your view about what is the most important thing for leveling up hopefully what we are doing in ONS is making that data available to everyone so they can see and judge for themselves whether what the impact of those policies are having but my personal belief is that a lot of the trends that we see in terms of regional disparities have emerged over decades and it will probably take a similar amount of time perhaps to see whether some of the issues that the government is thinking about leveling up are really having that impact particularly for example if we think about the importance of early education or health or any of those things the impact will take time but that is precisely one of the things that we are thinking about as part of the sub national data strategy is how do we make sure that we are data available on a consistent basis so that people can if they wish to judge the problems of the government for themselves. On that probably one of the big questions for us that we are working through at the moment is where we can redesign some of our data collections to directly collect data at more disaggregated levels and where we are going to need to use methods to apportion data that we collect at slightly higher aggregations to get that picture and it is a question of there is a question of efficiency there and there is a question of optimising the pictures to try and get as much data at a disaggregated level as we can manage whilst retaining relationships and making sure we can understand what that data is saying for us. But there is a lot of methods work to do there. Don has mentioned ESCO and other and we are doing more work on that in-house but if anyone wants to talk to me about apportionment methods and these sorts of things we are always very willing to talk. Thank you Richard. A follow up question around levelling up as well. When talking about the business insights and conditions survey in the context of levelling up analysis is there an alternative to using reporting offices for larger businesses to collect data without this subnational analysis on BICS can be in exact. Thank you Artit, that one. BICS is one of the services in my area and also we've got a presentation coming upon on BICS specifically as well which might go into more detail but the questions are really valid. It's one of the challenges we have with BICS because it's collected at a reporting office level. We've done a few things already to try and produce subnational analysis from that and there's some further things planned to try and improve the subnational analysis. So we've produced a few articles where we've just looked at single site businesses to see how they move and how that compares with the old business movement and similarly we've taken businesses that have multiple sites and assumed their response applies across all sites. So we've tried to look at the difference between those measures but as the questions sort of indicates that's not an exact approach. Going forward Richard mentioned what we're doing to look at whether there's an improved method of apportionment that we could use and we can look at that in relation to BICS as well but we're also considering if there are any sort of specific questions that we could ask of multi site businesses to try and help understand different performance across their sites. I think with the nature of BICS because of its timely nature and pace we wouldn't be looking to get quite the robust and quality information that we get from the annual business survey at the regional level. But I'm hopeful there is something that we can do to try and improve the subnational data that comes from BICS and so they're the things that we're thinking of at the moment but as we welcome the ideas as well from people on the call if they've got ideas to put forward so we can take that into account when we're considering our options there. I'll just add to that, looking beyond BICS we have had for a number of years now within ONS a team that really specialises in data collection from the very largest businesses which could have very complicated sort of corporate structure and so we do engage with those on a very sort of regular and detailed level just to understand precisely what's going on and make sure that we are apportioning and understanding their economic activity in the correct way. Thank you. That's great, thanks Ed. And now a question around the census release. So will you be consulting on redefining towns based on built up areas following the census release? I'm not sure we've got only one on the call who can answer that one specifically but we can certainly take this away and come back to Richard. Yes, okay that's great. Okay we'll take that away then Richard and come back to you following the conference. Okay so the next question is what different sub-national maps of the UK are used for business data and how do we ensure that these are consistent and ensure that they correspond to the aims of the levelling up agenda? I'll take that one. I think here this is one of the areas where we are trying to engage with stakeholders to identify which levels are most useful for them. We've obviously transitioned from the pre-existing nuts classifications to the new UTL classifications which basically map on to each other so people will see consistency there but I think we clearly are looking to listen to stakeholders to try to identify which of those are most useful and in some areas and for some questions it may not always be a universal picture level one, a UTL nuts one aggregation might be quite useful in one area might be quite less useful in others and it may be that some data sets we end up presenting in a more mix and match approach to try and meet user need whilst obviously always being able to aggregate back up to the higher levels. I think also in here as Donna's suggested this is going to be a long term piece of work where the art of the possible today we hope to very much improve as we move forward. We have an ambitious spending review bid that we've submitted to the Treasury in which we're hoping over the next few years to make some significant progress in this area so hopefully there'll be more to see soon. Great, thanks Richard. Now a question around broadband connectivity levels so which granular data are collected on broadband connectivity levels and with what frequency? Okay I'll take that one so until the pre-pandemic the data were collected annually and it was about whether households had an internet connection and the data in terms of granularity sub-nationally went down to regional level. The data came from the opinion survey during the pandemic has naturally moved to focus on the pandemic impacts and also what's happened is the mode has changed for that in terms of being collected online or digitally so what we're doing as part of the review of digital statistics is trying to identify what the user needs are and where they remain and the connectivity point will be is one part of that. Looking at alternative sources if it pre-exists anywhere else and we know off-com have some data and then where there are gaps trying to consider where they'll be filled going forward given the change in mode and content of OPN. So there is data as I said pre-pandemic but there isn't as a gap because of the change of focus of OPN but we are looking into filling that gap and the needs going forward but we don't have a definite plan for that because it's a little bit lag to the commerce development that I talked about earlier. Just to note there's lots of thanks from the audience for the responses to the questions and the question now, will you be working with private sector or will you be working with the private sector like Data City using company, web page scraping and machine learning to define emerging sectors? Can I take that one? I mean we are already working with a range of private sector data providers at the moment. Donna mentioned some of those under her presentation to carry out our faster indicators suite but we are working with a number of private sector data providers also in the financial transactions data area to understand really a faster or more granular level. What's happening in the economy has been really useful in order to respond rapidly to the central government to help them through the last 18 months but we are also thinking now about as we emerge from the pandemic how we can carry on working with those data providers and use the data that they provide us as a compliment to all the other things we have been talking about today to help build that better picture of the economy so certainly any ideas that people have got and I will take that one away will be very welcome please. Lots of folk we are working with already and we would really like to carry on doing that. The green agenda is very much a priority, green jobs, green businesses etc but it is difficult to know what that might include and how to measure. How is that being approached? I will pick this one up. We recently established a new environment division within the new economic statistics environment group that is led by Ian Townsend and they are working through the set of statistics that they will be generating and they will be generating. I think this is an area that we are going to be collaborating across our economic and environmental sides quite intensively on both to support COP26 but in the follow up and afterwards to try and get into some of these measures and some of these concepts a bit more. Net Zero and the recent Das Gupta review both give us ways of thinking about this which are a little bit different to some of our traditional frameworks so there will be a lot of activity in this area over the next couple of years. Great, thank you Richard. Yes, of interest Heather in the paper as you mentioned on the BICS analysis is so just to let the audience know that we will be sharing both the slides from Donald's presentation and links to those papers following the conference so you will be able to have a look at those if there are any more questions. I wondered if I could ask a bit about the ONS data science campus and its role in measuring economic activity. I'm happy to start off but I think others might want to contribute as well. I think the data science campus is organised and probably a unique way for the ONS in the sense that they tend to first think about what are issues and where they can apply their expertise obviously a collection of people with bigger brains than mine anyway to think about the application of data science so they will look for different issues whether that's for example on the leveling up agenda, one of the things that they're looking at is looking at local news sites and whether they can use that to measure sentiment or thinking about satellite imagery and what does that tell us about economic activity so they're very much almost a kind of experimental sandbox as it were for pushing out boundaries of what you can do with data science and how does that apply to economic activity but they also have a sort of important role in terms of supporting work streams like leveling up for example and they are I think they tend to form themselves into what I call squads where they apply themselves to particular issues and then work with teams with ONS and with our academic colleagues at Escrow and Turing Institute so very much open I think to ideas about what could be done in order to help and push the boundaries of what we might consider statistics You go ahead Richard I was just going to say that if one looks at for example the faster indicators they're a really nice example of where the data science campus has been able to do some experimental work which has put us in the position that we've then been able to take that and develop that quickly and deliver what was a key part of our response to Covid in effectively double quick time because a lot of that methodological groundwork had been done in a quite experimental space but it also meant it made a quite easy transition to some of those new platforms that Donna was describing earlier that it had been done in the right platforms in the right coding languages to allow it to be quickly picked up and used and I think that really has been a key example of where we are now seeing real benefits from the innovation data science campus into core statistics Sorry Ed. I was going to say precisely the same thing on the faster indicators which the campus pioneered a couple of years ago and actually I was looking back over a blog that Data Science Campus produced in 2019 which sort of talked about the value of faster indicators as potentially being able to spot big changes in the economy and we've certainly had some big changes in the economy so very present. The other point I wanted to make is that one of the key roles of the campus not only the sort of things that we see published and the work that we've been talking about just now but also in terms of building capability across government in terms of increasing everyone's data science capabilities and so anywhere that you're seeing data science sort of growing across all parts of government not just for measuring the economy quite often there's a key role there for the data science campus within the ONS in driving that capability and they've got a really strong role in doing that and that's one of the reasons it was set up. Thank you. Great. Thank you. Sounds very exciting and innovative and great development. Okay so I think if there aren't any final questions I'd just like to say a big thank you to Donna for her keynote presentation and to all members of the panel for your contributions. Also thanks to the audience for some great questions. I think it's been a really nice question and answer session so thank you all. Just to note Chris in the slides that we'll give you they also include links to all the different contexts for ONS business statistics so if people want to follow up on any particular aspect the slides should give you the email details. Great that's really helpful. Okay so thank you all and we'll now move on to our economic and business data updates. So we now have economic and business data updates from a range of ONS survey experts and I'd like to start by introducing Emily Hobson who will be speaking to us about the business insights and condition survey and should be interested as I know we have had some questions about this already. Emily Hobson is head of the business insights and condition survey at the office for national statistics where she's responsible for fostering its development from a few questions to one of ONS's highest profile economic surveys. Previously Emily worked as an operational research analyst in the business prices team in the ONS and she graduated from the University of Exeter with a Bachelor of Science in Mathematics. So I'll hand over to you now Emily. Morning everyone just checking everyone can see my screen. So as mentioned my name is Emily Hobson and I head up the business insights and condition survey. So just a bit of background on BICS the ONS implemented the business impact of COVID-19 survey otherwise known as BICS. In a matter of weeks where traditionally the creation of new surveys can take 6 to 12 months. During a great uncertainty it was essential to provide quick and coherent data on the impact of the pandemic with data being used in real time by decision makers to identify the strategies needed to respond to the impact of the coronavirus pandemic on the UK's economy and society. The voluntary fortnightly business survey has provided quality of information on turnover, workforce, prices and trade which has offered further insights around these business dynamics. Final results of each wave of BICS are published in the fortnightly bulletin, business insights and the impact on the UK economy and flash headline figures are presented in the weekly faster indicators bulletin. The name business impact of COVID-19 survey was chosen in March 2020 with no understanding of how long the pandemic would last or the impact the data collected from it would have on policy makers government departments and users. After seeing the success and timeliness of BICS there are requests from a range of stakeholders to add additional questions surrounding a wide range of topics including the EU transition period and net zero. This meant that the name of the survey did not reflect all aspects of the questionnaire so from wave 24 the survey name changed to business insights and condition survey. As you can see the acronym of BICS has stayed but the survey was known for this acronym and also there are so many acronyms out there already so why make a new one. As I mentioned just now BICS is a fortnightly survey and questions are reviewed and updated as part of each fortnightly wave. We now have final results of 38 waves which collect responses from businesses during a certain period. The sampling frame used in BICS was designed to achieve adequate coverage of the listed industries on the monthly business survey for the first six waves. The monthly business survey collects data on businesses turnover to use in the monthly gross domestic product estimates. However agriculture, public administration defence, public provision of education and health and finance insurance are excluded from the survey. The sample size has increased over the waves to improve the coverage of regions and different business sizes to ensure the sample was more representative of all UK businesses. The sample design for wave 7 of BICS was reviewed and refreshed and went live for mid June. From wave 7 the questionnaire went to approximately 24,000 businesses. This sample redesign improved our coverage of smaller size businesses. In wave 17 we increased our sample size again to approximately 39,000 businesses. This increase will allow us to further break down the industry data and produce more granular regional data. This is to meet the user's needs to better understand the impact of local lockdowns. The BICS results and outputs have received wide coverage from organisations such as the Bank of England, Office for Budget Responsibility and National Institute of Economic and Social Research. Coverage also extends to the news and media where BICS is quoted on a regular basis. While an extensive user consultation about the questionnaire topics was not initially possible due to early time constraints, over the course of different waves of BICS, ONS works closely with policy and analytical leads in government departments to help identify priority topics for new waves of BICS. These have included conversations and user requirements from BASE, Department of International Trade, HMRC and the devolved administrations. ONS utilises a microdage approach through the secure research service where academics and accredited institutions can work on the confidentialised BICS data from all the waves. This has allowed detailed policy questions specific to user needs to be addressed in a secure research environment. The earliest iterations of BICS focused on financial performance during the pandemic such as the impact on turnover prices and trade compared to normal circumstances and operational performance relating to the workforce. For example, where the workforce size is increased or decreased and what percentage of staff are working remotely or off sick with coronavirus. Questions covered actual performance in the last two weeks and also expected performance in the next two weeks. New questions are added and refined over time. When the new UK government announced the introduction of various schemes initiatives designed to support businesses and the workforce, questions were developed to capture uptake and issues at the UK level as well as uptake for each country within the UK. The question bank of all the BICS questions for each wave is publicly available and with only a two week window to agree content, design questions and build and test the survey before it goes live, we've had to adapt our standard questionnaire design process. As the pandemic increased, the need for weighted results has been more critical to provide an estimate of how all businesses have been impacted from the start of the pandemic to date. With weighted results comparison across waves is feasible, showing how different times during the last 18 months have affected different industries and different size businesses. Results have three possible waiting methods, weighted by counts, turnover or employment. From wave 13 onwards, publications now include weighted results with invitation and a back series from wave 7 to date. In wave 37 as part of a regular and ongoing quality assessment of our methods, we introduced improved methodology for the waiting smaller businesses within the BICS sample when we use employment and turnover to do so. This has in the main not affected our previous findings, we have though revised some estimates back to June 2020. BICS has been used by a variety of monthly and quarterly surveys produced by ONS, particularly in the quality assurance stage and assessing impacts between the fortnightly data sets and monthly outputs. The published fortnightly results of BICS are broken down by industry, workforce size and region to incorporate all users' needs. For example, other monthly publications that quote BICS results have included monthly retail sales and construction output. An example of where BICS data matches closely with other sources is the BICS proportion for furlough estimates. The figure on the right shows the close match between the BICS proportion furlough estimates against the HMRC official furlough team statistics. This is despite the fact BICS has published four days just after the survey closes, while the CGRS figures results are provisional up to four weeks after the time period, which is a feature of the claims deadline, but the HMRC results do offer a wide range of breakdowns, for example local authority, gender and age. BICS were designed to give an indication of the impact of challenges businesses have faced over the last 18 months and a time year estimate than other ONS monthly or quantity surveys. It's allowed for many different streams of analysis, including longitudinal data analysis following businesses responses over the weeks, linking to other sources to provide additional insights into businesses such as the labour market and to help validate other ONS business surveys and provide supplementary hesitance. The BICS microdata from Waves 1 to 37 can now be accessed through the secure research service. The BICS microdata from each wave are released on a rolling basis in the week following the publication of each wave. Only researchers are credited under the Digital Economy Act for able to access data from the SRS. You can apply for accreditation through the research accreditation service, but you will need to have relevant academic or work experience and most successfully attend and complete the safe research training. To conduct analysis with the microdata from the SRS, a project application must be submitted to the research accreditation panel. To access the SRS you must also work with an organisation connectivity agreement in place. The SRS data is forwarded onto the BICS, forwarded onto the UK data service, and all this information on how to apply is available in our fortnightly BICS bulletin. A few notes and caveats of the BICS microdata available on the SRS. The microdata is made confidential and does not disclose information on any specific business. This means no free text responses are included in the data. From Waves 17 onwards, the weights are included in the microdata. However, we are in the process of updating and backdating the employment and turnover weights back to Waves 7. The main difference between the microdata and the final published results is it does not include the imputes for responding businesses with 230 plus employees. This means there might be slight differences between results created through the microdata on the SRS and the final published BICS results. So just to summarise, BICS has provided, BICS has provided evidence to support both the regular on-garrant additional monthly and quarterly surveys in quality assurance capacity, but has also enabled more timely estimates of economic activity to form part of the evidence base for high policy pressure decisions. The BICS questionnaire has shown itself to be sustainable over time and this survey will therefore continue, at least for the short to medium term, therefore continue adapting to the ever-changing economic picture. We continue to explore the potential for further development of BICS and consider what future requirements may be needed by stakeholders. Now open the floor to questions. Thanks very much, Emily. That was really interesting and exciting to see how quickly you were able to respond and develop the survey. It's really, really great. So we have some time for questions now for Emily. So if you do have any questions then please type them in the chat. I just wanted to start, Emily, if I could ask you a bit about whether there's any information in the BICS data set that enables researchers to link the data to other sources of data. So on the SRS and in the BICS microdata there is a variety of just basic information around the businesses that respond, whether it's industry, size band and I think it goes down to four digit standard industrial classification but it also includes the reporting unit reference which means it will be able to link to other data sources such as the IDPR. That's really useful for researchers isn't it, being able to bring in a whole range of business data and link them. That's great. We have a question. How might non-government researchers suggest new questions to be included? What's the process for that? To request questions within the BICS survey, the process is just to drop an email to the BICS inbox which is bics.ons.gov.uk All question requests are considered and we can get in touch with it if it's something suitable to include in the survey or not. Can I ask a little bit about the response rate as well and the impact that might have on the estimate so I know it has a relatively low response rate I just wondered if you've done any analysis looking at the sort of responding and non-responding businesses and whether there is a non-response bias and if the weights are used to correct for that. Useful just a little bit more about that please. We might seem to have a fairly low response rate to some of the surveys published by the ONS but considering that BICS is voluntary it's pretty good I think, roughly averaging between 23 and 25% response rate. You're correct in saying the results are weighted up to be reflective of all UK businesses and this does take into account the businesses that do not respond to the survey so the results are weighted by two digit SIC and size band also to elaborate on that we are in the process of developing standard areas to hopefully measure any kind of sampling so hopefully towards the beginning of 2022 we'll be able to publish standard areas for variables to measure the impact of maybe low response rate in the sample. Great, thanks Emily. Some of the BICS questions refer to expectations compared to normal times of year. Do you have any thoughts about how or if these questions will evolve when for example post pandemic becomes the new normal? That's a very good question and it's something that we've actually been discussing internally for the last couple of months. It is a hard one to judge because it is also dependent on what businesses think is normal, is now the new normal hot topic at the moment so I think we'll continue with using normal expectations for the moment. It's still relevant even in possibly like a couple of years. It is just slowly dependent on businesses interpretation of normal but BICS is trying to measure the impact of challenges so when there is a challenge it's important to measure against what is considered normal whatever time of year that is. Great, thanks Emily. And then a question, can BICS survey information with other data by ONS including these secured and controlled ones? I guess similar to the question at the beginning around that. Yes, so as I mentioned before, the microdata on the SRS can be linked to other surveys with survey data again on the SRS just through linking possibly of reporting unit RU reference number. And yeah, this is a really interesting question actually. Will BICS continue post pandemic? Will it now feature as a permanent ONS data source? So mentioned on my last slide, BICS will continue in the short to medium term. It has shown itself to be very successful and adaptable not only collecting data on the pandemic but on other kind of challenges that have been faced in the last year so definitely in the short to medium term. Great, thank you. And a nice comment here, congratulations on BICS and have you created time series from any of the questions, for example? Question reduced between waves. Yep, no 100%. So in our published data set that also goes out with our fortnightly bulletin all the data available in there is in a time series format. Great, thanks Emily. And then just one final question. Are you able to speak to any of the analysis that has been completed using the microdata? So any sort of summaries of findings or any of the analysis that's been undertaken to date? I feel like that's come up. Nothing all the top of my head but more than happy if someone would like to know a bit more detail to get in touch via the BICS email address that I mentioned earlier and is on the front of the slides and I can get back to them. Great. Okay, so just say a big thank you to Emily. It's really useful and interesting presentation. So thanks very much for that. And yeah, we'll provide the details as well if people want to have a look at some of the analysis that's been undertaken. So thanks again Emily. And for our second update we now have Daniel Robinson from the Office for National Statistics who's talking to us about the e-commerce survey. Daniel is head of the All Common R&D branch within the Office for National Statistics. He has worked in the O&S for over 20 years and has developed experience in a number of roles covering both business and social surveys spanning both data collection and results production. Welcome Dan. Good morning. Thank you for the introduction. So yeah, as mentioned my name is Dan Robinson and I have the All Common R&D branch. So I'm going to talk to you this morning very briefly about our e-commerce survey. It's currently being paused and I'll kind of talk about kind of the development work we're doing on e-commerce. So I guess to start a bit with a backstory of e-commerce. So the survey was launched back in the year 2000 so it's been a relatively long standing survey with the O&S. It was launched in sort of conjunction with Eurostat. It was a Eurostat led initiative and it was kind of launched kind of capturing the e-commerce and kind of the adoption of ICT technology across the EU. So as I mentioned the purpose was to measure the adoption and use of information and communication technologies within the UK business community. So the definition of e-commerce is and the OEC definition is what we follow and it's any goods or services that were ordered over the internet or other electronic networks regardless of payment and delivery methods. And I guess to expand on that a little bit we kind of historically we've captured information on kind of web sales in terms of e-commerce but also sort of through other orders that have been placed through other electronic networks and I guess more recent technologies maybe apps and perhaps E-POS technology for electronic points of sale. So that's kind of a definition as to what e-commerce or the e-commerce definition we've aligned to. So a survey design we've got an annual sample or we had an annual sample of around 11,000 UK businesses and like the majority of all other business surveys in the ONS we drew a sample from the inter department or business register. So this is a business or it's a sample and framework that we use of all UK businesses that are either VAT or PEY you registered and there's over 2 million UK businesses on there that we drew a sample from. A more recent development was that from 2018 the e-commerce questionnaires started to be captured electronically whereas previously we were using paper questionnaires. And the notable industry exclusions for e-commerce are the financial agriculture, education, health and social work, public admin and defence and some arts and entertainment and recreation industries. So we don't cover the whole of the UK economy. Standards sort of techniques are applied to the data once it's collected. So data will undergo validation to improve accuracy and then in order to elevate our results to the population we will apply standard implication, outlier and anestimation techniques to actually develop some meaningful UK representative estimates. Topics covered. So the since the inception of e-commerce the topics that we've covered have varied year on year. It's been a very fluid questionnaire and given that this is an ever changing space with new technologies being developed the modules that we've included have changed year on year. So as of the last survey that was actually running in the UK these are the topics that we actually kind of covered being use of computers to find out whether business has actually used computers to conduct their business. E-commerce in terms of how much trade was generated or sales generated online. Invoicing use of cloud computing services obviously the technology that we use to actually store information is changing year on year and we're now moving towards more cloud based services. Big data analysis we're trying to find out whether the business is approaching or selling big data in order to develop or improve the services they're delivering. ICT specialists and skills have they been purchasing or supplying ICT specialist knowledge as part of their business activity and some of the topics were around the use of 3D printing technologies and the use of robotics in their production processes and with all of those topics we're essentially trying to find out whether or not businesses are using these types of technologies and knowledge to actually enhance or develop the products they're delivering as a business. Just to give you some sort of recent insights so I think since 2015 there's been a general upward trend in the value of e-commerce sales that we've been capturing. I think in 2019 we reached a peak of around about 693 billion in terms of e-commerce sales which was a 15.2 billion increase on the previous year. If we were to slice that taking a slightly different way looking at industries that have reported e-commerce activity to us it seems to be dominated by two industries primarily being the wholesale and manufacturing industries which in 2019 the wholesale industry had an estimated value of around about 214.7 billion of sales recorded that were generated electronically and the manufacturing industries of 188 billion respectively so two dominant industries that report e-commerce sales to us so some things which users of the microdata I guess should be aware obviously like I've mentioned we had generally a long standing survey it's been running close to 20 years and I think the sample coverage has expanded over that time. In 2014 we actually started the sample small nought to nine size businesses whereas previously they'd been excluded so anybody looking at the data from that period on time that period of time forward is something to bear in mind. Also when we talk about kind of our total e-commerce sales I think it's important to mention that historically we've only asked the business to record to us what proportion of their sales are generated via online sales and what we've actually done is then constrain that proportion to the industry totals from the ABS so historically the e-commerce survey has not been collecting a £1 million value from a business in terms of what the e-commerce activity is we've constrained to previously published or already collected information from the annual business survey also noting the mode of collection changed in 2008 by where we started to adopt electronic methods moving away from the paper question is. As I mentioned the topics that we've had on the e-commerce survey have changed in line with Eurostats requirements so this may again something to bear in mind if using the microdata it may impact on the time series in that particular questions or variables you're interested in may not be there for either long periods of time or they may be breaks and cease but it was just driven by what the Eurostat requirements were at the time that we were capturing the data like Emily mentioned on the bits the microdata for e-commerce can be accessed by the secure research service and the UK data service and that the periods are available from 2001 to 2019 so in terms of our development work so the survey paused so the current survey paused in December 2020 so 2019 was the last reference period in the current time series myself and my team then entered into a discovery phase between January and April by where we met external stakeholders internal stakeholders to re-establish user requirements and policy needs I think one of the important things is that rather than this being a legislative led survey in terms of impact in the questions we've actually gone back to the stakeholders in the user community to find out what information they want us to capture and we're trying to deliver a product that kind of specifically meets their needs so slight deviation from the way we've done things historically in around May time we launched a consultation on a blog to try and direct we were conscious that we met quite extensively with government users, we wanted to widen that reach that went out, there was a mail shot out to 40,000 people and there was a blog to direct traffic to that consultation just to try and capture people's views and wider needs that would not necessarily have been covered by government users so over the summer of this year my team have been working on turning those requirements into questions and at the beginning of September we launched into a cognitive testing stage by where the questions are being asked by the users, I've now been actually tested by the business community to make sure that we can actually take them forward and roll them out on a questionnaire early next year as we enter the end of the autumn into the winter we're going to be looking at sort of methods development because one of the key things that we're actually doing now with the survey going forward is collecting a £1 million value in terms of the business's e-commerce sales moving away from collecting those proportions so we need to kind of review and implement what we have and implement new ones to actually kind of elevate those £1 million results to the population. At the moment we're currently working on the basis of launching or relaunching the survey mid-February and then the timeline in terms of how things will pan out from there would be we would look to close for a results production round towards the end of September with provisional publication date of maybe the end of December this year lodging any sort of micro data to the SRS for the 2021 reference period in January 2023. So I guess key differences between the new and the old surveys so I think the what we've done in terms of added is we've expanded the survey to collect more data on e-commerce imports in that we've had basic information in terms of e-commerce exports out of the UK and I think some of the requirements that were coming in was that we need to actually be capturing information on e-commerce activity coming into the UK so we're going to capture UK imports of e-commerce activity more international geographic coverage again this is still relatively high level but I think it's at continents whereas previously I believe the breakdowns that we'd only captured in terms of geography were UK, EU and rest of the world so we're now actually going to be trying to identify the continents that we're actually trading with in terms of e-commerce activity I mentioned about the value of e-commerce sales now being recorded by respondents whereas previously we recorded proportions only and two sort of new topics have actually kind of come in I say two new ICT security is being reintroduced which I think is essentially the same questions that we had with the Eurostat model but there's also a growing interest in digital intermediary platforms so we have some questions to try and understand if the businesses that are taking part would classify themselves as a digital intermediary platform things that we've had to kind of remove or there wasn't an identified stakeholder need was that with all of the conversations that we've had with internal stakeholders external and kind of wider nobody seemed to actually want the breakdown between EDI and web sales so going forward we'll just be looking to capture the total only and some of the topics that we've had to remove because of lack of interest were things around kind of invoes in use of robotics and 3D printing so next steps that's okay I think I'm on track so I think next steps for us are digital trade by nature changes rapidly the survey will continue to evolve and adapt to user needs so one of the things which we've actually had the benefit of being able to do from the start is to change the modules on e-commerce in line with kind of as things emerge in this space and I think the that is a requirement and something we're going to be able to meet going forward is that we can kind of engage with the stakeholder community to capture any sort of new needs they have and look to introduce questions on to e-commerce questioning look to continue to identify needs and explore other potential sources that impact on wider digital measures I think something that Heather mentioned earlier on was kind of around about the you know e-commerce is just one measure of digital activity and we have some that kind of have actually had to cease because of changes to samples and the way data has been collected so I think we need to kind of continue to explore other sources that may be able to kind of fill the gap from those digital measures that we've lost and I think one of the more recent developments is that we're looking at the use of credit card data to help build a picture of the digital economy and then we're going to look to maybe see how perhaps we could dovetail this with e-commerce and potentially maybe helping with the quality assurance work on e-commerce primarily. I think that does bring me to the end of my presentation so happy to take questions and I think there was one actually kind of there initially was kind of linked with the artificial intelligence I think just to pick up on that so that's probably there's an opportunity within the questionnaire or there's a question on technology and we're actually looking to ask sort of questions to the business community about what types of technology have helped them to develop their business and may help them develop it going forward so there would be an opportunity for them to record artificial intelligence as a software or a method that could help them do that but I think interestingly some of the key things that were coming from the stakeholder community were more around customer relationship management software so sort of CRM software and resource planning software and also online accountancy software so it's not that there wasn't a need for AI I think but there just seem to be more types of software or technologies that were kind of coming through in terms of needing to understand more about than perhaps artificial intelligence initially so I think hopefully that answers the artificial intelligence question we had earlier. Thanks for that Dan and it's really interesting presentation and really nice to hear actually how you consulted with users and making changes to the survey off the back of that so I think that's a really nice feature so we have a question about EDI sales so could you explain what's EDI sales cover? So I think EDI sales would be anything that's kind of non-website based so if you should make purchases via an app I mean we all use apps these days and lots of kind of websites have their own apps so you can make actual purchases through which it's not deemed through their website it's through a secondary app and I think also thinking about how some businesses work they may order stock automatically there may be systems in place where they can monitor stock levels and generate an order automatically via an electronic means that is not physically through a website so it's kind of systems around the periphery such as apps and perhaps more automated electronic forms of ordering. Great, thanks Dan. Can I ask a question about the sort of transfer from the paper questionnaire to electronic forms which you mentioned and I know Donna spoke about a sort of overall move to electronic surveys going forward. Have you seen an improvement in the data quality off the back of that? I think one of the things that we benefit from electronically is kind of the promptness of the response. So I think with surveys moving to kind of electronic data capture I think it kind of generates a quicker response and I think it does kind of build in allow us to build in kind of more quality assurance and production time. I wasn't around with the transition from paper to questionnaire but there may have been some sort of modal effect in terms of the results but I don't think there is anything significant in terms of kind of impact of quality from moving paper to EQ but obviously I think that does give us the benefit of actually being able to kind of capture data more quickly and kind of perhaps process it and respond in a more prompt manner. More timely. The automated ordering is quite common in manufacturing supply chains and I wondered as well that with the sample I know some sectors were excluded. What's the reason for that? I think we have a number of O&S surveys don't include all industries as commonplace. I think a large one be in the financial industries and I think we actually capture information for those industries elsewhere and I guess with a number of those kind of nuances to those industries there may be nuances that mean we perhaps shouldn't capture or survey them because their data may be covered elsewhere so I guess it's trying to balance kind of compliance and overburden in businesses I guess in certain industries where perhaps we shouldn't be capturing them. Yes, that makes sense. So another question, any data or plan to cover sharing economy or platform economy activity such as Uber and Airbnb? So we've actually got a team separately outside of my team that actually cover, I think we do have a publication on the sharing economy and I think it's perhaps I guess in answer to your question there is and I think there's another team kind of within our area who cover the sharing economy so I think perhaps we could actually perhaps link them up with, I can answer questions specifically outside of the remit of my team but perhaps we can link up those teams if they have any questions on the sharing economy to follow up with. Yeah, that's great, we can do that following the conference. And then a final question, is the survey supposed to be representative at the sub-industry level for the included sectors? So I think it's represented a bit, obviously we have a small sample size and I think this is representative as it can be with the small sample size. I think if we were unhelpfully part of the redevelopment work we're sticking with the same sample size so I guess there's not any sort of scope to actually increase sample sizes to give more robustness around some of those estimates but I guess it's as representative as it can be with the relatively small sample that we have. Great, thanks Dan. And then one final question I think we're just squeezing, so do you make any adjustments to the estimates of e-commerce to adjust or remove intermediate sales compared with final sales? No, so I think the each return will kind of go through a validation process. A lot of those validations will go on clean so they won't actually be looked at by my team but if any of the variables in their fail validation they will be kind of checked at source with the business and so there may be actually some changes to the most there may be changes to the microdata when the data was validated at source but in terms of the estimates that kind of fall out of the results production process then we don't make any adjustments unless we identify perhaps some anomalies that have slipped through in the validation stages and we would go back and again correct the data at source with the respondent and then reproduce a set of results so we wouldn't sort of intervene and make adjustments so I guess the answer to that is no we wouldn't, all changes are done with the respondent source. Great, thanks Dan. So I think that brings us to the end of Dan's section then so thanks again Dan for your presentation and to the audience for the questions as well that was great and we're moving on now to hear from Jakob Snibeca who's talking about the management and expectation survey. Jakob heads the firm level business dynamism and management practices research at the ONS before joining ONS he was a research fellow in economics at the University of Oxford. So welcome Jakob. Thank you Chris. All right I am very happy to be here today to talk about perhaps one of the lesser well known ONS business service so I thought I would structure this presentation a little bit like a user guide. I hope this is going to be most useful to this audience and I've kept the presentation relatively brief to allow for ample time for questions. So just by way of introduction what is the management and expectation survey? The management and expectation survey is an ESRC funded collaboration with academics through the economics statistics centre of excellence including just to name a few Rebecca Riley at KCL, Paul Mason at Nottingham, John Van Wienen at the LSE and it is a voluntary business survey of which currently two waves exist capturing three years of data so that's data covering 2016, 2019 and 2020 as well as a smaller precursor survey, the management practices survey which captured manufacturing firms only. The management and expectation survey aims to capture firm level evidence of the use of structured management practices and various correlates to allow to study the role of management in anything from innovation to productivity. The service collected at the reference unit level and as such is easily linkable to other ONS business service and thus as all the other service I think discussed on this panel of the conferences available through the secure research service for accredited researchers. So a little bit of a roadmap for the rest of this brief talk. I'll give you a bit of an overview of the methodology in particular what we capture when we say management practices as well as a very brief overview of response rates and other features of the data. I'll then highlight some of the features that are new in the latest wave of the management and expectation survey and finally I'll conclude with some example user cases from our own use as ONS to give you an idea of what is possible with this data. Alright, so first of all how are management practices captured in the mess so the management practices score is an aggregate of four subcategories of structured management practices the methodology builds on this earlier wave of ONS work and even earlier work by Nick Blum Raphael Sadan and John Van Wienen and John Van Wienen are both collaborators in this project as well so firms are asked about their management practices in four key areas that's key performance indicators so the extent to which firms capture and monitor information about important performance indicators the use and communication of targets throughout the organization the practices related to hiring firing training and finally the way that firms react to problems and whether they use those problems to continually improve their processes so just to put a little bit of meat on the bone straight from the questionnaire relating to continuous improvement so we're asking for a specific year what was the most common response to problems faced within the business was it to resolve the problem but not take further action to resolve the problem and ensure that it would not happen again to resolve the problem and put in place a process of continuous improvement to head of similar problems or on no action at all and again just for you as micro data users the SRS data set computes scores for each of the categories but it also contains all of the individual answers if you are interested in particular aspects of management practices or want to aggregate them in a different way and there's documentation that explains each of the questions sample size and response rates just like PICS the management and expectation survey is a voluntary business survey so the response rates are actually really similar to what we've seen on PICS so response rate during the pandemic 24% that's quite a large drop from the previous wave for pandemic related reasons the sample size went up because we sampled a larger number of businesses but the response rate is done somewhat in some of predictable ways across regions and industries I think on this slide I should perhaps also say a little bit about the sampling so the mess goes out to firms with 10 or more employees in 11 Nats 1 regions so Northern Ireland is excluded in a custom mix of service and manufacturing industries there's a lot more detail in our publications which I will link in the chat after the talk now to prevent maybe a question that Dan and Emily got we actually did spend quite a lot of time investigating selective response rates and the nice thing is that maybe start with the conclusion we find very little evidence of selective response rates and in fact even for the linked sample where we can investigate response rates by management practices score in the previous wave and again there's a lot more about those results in the publications that I will link to at the end so what is new in the latest wave of the mess I highlight four innovations the first one might be most important to you we have novel data on pandemic related firm adaptations such as the use of online sales, home working and changes in the supply chains and due to the design of the sample we are able in the same wave of the survey to capture both a pre-pandemic baseline namely the 2019 responses and the pandemic response which is of course helpful for forming comparisons we've also designed the sample this time around in a way to maximise the number of firms we can either link to the previous wave for longitudinal study or to the annual business survey for wider productivity research at the firm level and finally just like the e-commerce survey we have moved to an electronic questionnaire and this allowed us not only to improve the response rates relative to what the paper based survey in the pandemic would have been able to achieve but also to check the data as it was coming in and therefore improve data quality okay so before we go okay this is excellent perfect so let me just close with a couple of user cases to showcase the type of work that you can do with the management and expectation survey so this is all from our own work so first of all using just the mes we're able to look at changes in the distribution of management practices across types of firms and overtime both in the aggregate but then also at the linked firm level with the 2020 wave of the data we're able to look at the differential response of better and worse managed firms to the pandemic in terms of home working rates on land sales and so on as well as sort of the effects of that on labour productivity then linking the mesh to other services again this is from an OLS publication we're able to bring firm level productivity to the table and finally and finally this is from our most recent piece of work we linked the mesh to birds so the R&D or an SRD survey and the UK innovation survey to look at the correlates of innovation and R&D activity so I think here I'll want to open this up to the question but just to sum up the mesh I think provides really compelling and interesting firm level evidence on management practices and their determinants including some that aren't captured elsewhere like certain types of ownership structures, firm expectations at the firm and the economy level and organisational choices that we might expect to be correlated with management practices and importantly for you as micro data users it's possible both over time across the waves as well as to other business service by the reporting unit reference number and with that I will stop and open the floor to questions Thanks very much Rebecca it's really interesting and great to hear more about the survey so our first question is can you observe any pattern or bias in the fall in response rates in the 2020 data for example regarding geographical, sectoral or characteristics like size? Yes so that's a really good question and I try to sort of head off that question in my presentation already so we've tried to predict response rates based on industry and region fixed effects size of the firm and turnover and our spirit is really really small for the linked sample so firms that were sampled both in 2017 and in the latest wave we're even able to look at their management practices score in 2017 because we thought perhaps worse managed firms were less likely to respond to a survey in the pandemic and again we're finding no strong evidence of that. Can I ask how well the questions in the survey capture management practices did you test the questions in advance of rolling them out? Yes so of course there was cognitive testing but more importantly the questions in the mess have been used in a very similar format by Blum and Van Rinen and other settings including the US and we know both for the UK and for other countries that they correlate well with other measures of firm performance so I think they are a good match although of course it's always important to take a critical look at your questions and see whether they need updating. What do you mean by organisational choice here? To give you a couple of examples particularly with the pandemic we were interested in how organisations adapt and so included new questions that weren't included in the last wave around the adaptation of home working, the switch to online sales supply chain disruptions which as far as I know are not captured elsewhere in O&S business service. Except for the big say I guess. In this survey is there anything collected to know about business owners and managers in terms of their personal characteristics such as ethnicity, education, experience of running businesses et cetera? No so I think that's a potential bit of a blind spot. We do collect some information on the managers. In particular we ask firms about the share of managers that have a degree and potentially this service is linkable to other data sources such as companies house data where if desired somebody could bring in ownership characteristics but they're not in the survey. That's a really good suggestion in terms of being able to bring in the external data. I think that brings us to the end of that presentation. Thanks to the audience for your questions too. Then we have our final update on economic microdata research I'm very pleased to introduce Tom Wickersham Tom leads the economic microdata research team at the O&S. The team specialises in developing linked firm level data sets to improve national accounts estimates and for research into business economic performance such as productivity, business dynamism and innovation. Before joining the O&S Tom works in various various economist roles in the National Infrastructure Commission and Her Majesty's Treasury. So I hand over to you now Tom. Thank you very much. As introduced I'm one of the leads of the economic microdata research team here at the office of national statistics. So today I'm going to speak about the work that my team have been doing and work that we have in development. This is a bit broader than some of the previous items on specific surveys so do forgive me if some of it's a bit high level but I'm happy to discuss more detail in the Q&A afterwards. So just to tell you a little bit about my team, our research is quite varied but a common theme is that we specialise in developing linked data sets and typically at the firm level. And then we use these data sets for analysis. So we publish original research ourselves through the O&S websites and through the economic statistics centre of excellence. And then we also often deposit those linked data sets on the secure research service for external researchers to use. So just to give a few examples of the kind of research that we've produced over the past couple of years. The chart on the left, this is from a piece we had of analysis showing how business dynamism has declined since the turn of the millennium with lower job disruption rates helping to explain employment growth over that period. On the right we've got a chart from a piece we did on firm level labour productivity and that's using the annual business survey microdata and it shows widening dispersion of firm level productivity over the past couple of decades. Preparing this presentation maybe realise how much of my team's research is summarised in density charts. The chart on the left was something we published via the economic stats centre of excellence and this one used expenditure data from value added tax administrative data and how the ratio between turnover and expenditure for firms changed during the pandemic which we concluded like we can the assumption that is often used that turnover from the VAT data can be used, the leading indicator of gross value added in national accounts estimation. Is that something where we're doing a bit further work now to look into that more? And then finally on the right we published a piece showing linked administrative trading goods data to show how different types of trading firms were more productive on average with firms that didn't trade being less productive. So a bit of a, just to give a taste of the variety of work there, I mean the other major piece of work which sits in our team is the management and expectation survey which Jacop has just presented on and I explained that some of our data sets we deposit on the secure research service. A couple of examples of those are the management and expectation survey data, linked trading goods with interdepartmental business register and foreign direct investment which is basically just a marker for business units of FDI activity that can be linked to the IDBR. I thought it would be worth highlighting a couple of challenges that we tend to face with all this work. The first is where we're using administrative data so any of you that use administrative data will be well aware that it's not designed for statistical use and that leads to plenty of challenges so that includes getting access to it in the first place, cleaning it, developing the methodology for linking that administrative data to the units that we would typically use for business microdata for analysis such as reporting units and local units. And then finally once we've done all that cleaning and linking understanding how and why what we see in the administrative data might differ to other sources that we use. So that's one sort of major area of challenge. I mean another one is common really is that we're quite a small team and we need to have this balance between updating some of our past data sets with more recent data but also we want to produce new research and data and methodology improvements so we have to sort of trade those things off a bit. So just talk a bit about our ongoing and future work and go through a few projects that we've got in the pipeline. The first I wanted to mention is the annual respondents database or ARDX which is a data set of longitudinal business microdata so this unites the annual business survey data which I think is from 2008 to the present day with its predecessor survey the annual business inquiry from 1994 and an old version of this data set I think already exists in the research service but we want to do various updates to improve it including bug fixes and improve conversion between the different generations of industry classifications, better documentation for users and an updated code for generating firm level capital stocks from this data and we're aiming to get the updated version of this onto the secure research service by January 2022 so over the next few months. We also have some plans for the productivity microdata research so the first thing we'd like to do is to update that firm level productivity publication using the latest annual business survey from 2019 and alongside it to update the how productive is your business interactive tool in which businesses can plug in their turnover, intermediate consumption and employment to see where they sit in the productivity distribution for the industry and that's something we've had some requests to update. We'd also like to do further research into firm level capital stocks estimates and firm level multi-factor productivity. This will be a bit of longer term work as it's sort of treading new ground I think using the kind of data that we have available. We also will continue to contribute to OECD business microdata projects so for example multi-prod is one of these projects which produces internationally comparable firm level productivity measures so we'll just use that to publish new research. We're also going to co-author a user guide for productivity microdata with the O&S productivity team and the aim of that will be to make it easier for analysts who aren't familiar with business microdata to get started in accessing it and using it for productivity analysis. The next project I wanted to mention was the longitudinal business database so the research that I mentioned earlier on business dynamism was based on an early version data set. The database takes quarterly snapshots of the inter-departmental business register and uses them to construct a linked longitudinal spine if you like with a near population coverage of UK businesses. What this spine does is allows it's a spine of linked business references which will then allow other longitudinal business data to be linked onto it. The expectation is that this database will make it much easier to track active businesses over time in the business data to study things like business dynamism and restructuring so things like mergers and acquisitions and to study how firms productivity varies over its life cycle if we are able to link on the productivity microdata to this longitudinal business database spine. We'll use the spine within the O&S to create these data sets which can then be deposited onto the secure research service and we also want to update then our previous research on business dynamism. Just finally then, I wanted to offer a few thoughts on microdata congruence and how it affects the future of firm level data. A big area that our team is working on is linking administrative data sources, particularly tax so things like corporation tax, VAT data and company balance sheet data from sources such as companies house to the more traditional business surveys such as the annual business survey. The initial aim of this work is to actually improve the national accounts by comparing the information from the administrative data about things like turnover, expenditure and profits to our standard survey inputs for the national accounts and identifying areas where the data sources agree or are congruent or where they might disagree. For example, the administrative data might suggest that a survey measure used to estimate GDP or production approach is overestimating the rate of growth for a particular industry and then we could recommend an adjustment to how we balance the different measures of GDP for a more accurate estimate. A couple of minutes. Thank you. The national accounts are the primary customer for this work but longer term this could transform our firm level data too. The administrative data rather than business survey data it could result in much more comprehensive longitudinal business data so particularly for the majority of small and medium sized firms that wouldn't be routinely captured every year in the sample business surveys. The big caveat here is there's lots of further work to do on accessing the data, developing it, all the things I mentioned earlier around the challenges of admin data work. So this is likely to remain within the ONS for the people future but I think there's really big longer term potential here. That was it. Thank you very much for listening. I'm happy to take questions or my email address is there if anyone wants to contact me afterwards about our microdata work. Thank you. Thank you very much Tom. It's really great to hear. Particularly the RDEX update sounds very exciting. I know business data users will particularly welcome that. There are some questions so I'll just go through them then. The first one is actually around the RDEX. What period is the updated RDEX coming out in January 22 which period will it be covering? I think if I remember correctly it will cover 1994 which I think is what the annual business inquiry data that we have goes up to it will either be 2018 or 2019. We haven't had a chance yet to properly look into the 2019 annual business survey microdata. I know there were some issues around response rates and things so we'll need to quality check that before we put out the microdata but hopefully that's what we'll update it to. Great. Thanks Tom. When do you expect to publish a productivity microdata user guide? Is this something in the pipeline for this year? Again, I think that's something it's probably over the next few months. I imagine we'll do an initial version and then add to it based on feedback from users. I perhaps wouldn't hold us to the end of this year because we probably need to plan some of the detail a bit but it's certainly over the next few months again. Great. Thanks Tom. A question around whether the longitudinal database is available on the SRS. Yes, so as I was trying to explain the spine if you like that has all the kind of connected references will probably at least for the time being have to stay within the ONS because it will contain it'll probably be difficult to publish with something like disclosive information but what we'll aim to do then is use that to construct various linked data sets that would be of interest to microdata users that we could then deposit onto this Cure research service and I think we'll aim to do that probably in the first half of next year to start having some data from that available to users we're still working on the engineering if you like of the various linking that needs to happen for that to be produced. Great. Thank you Tom. A question around what is the scope or limitation to linking productivity microdata with R&D data in bird? Difficult to say without understanding a little bit more of the context of what the research question is. I mean I think just in terms of the practicalities so far as I know it's possible to do using the kind of standard business reporting units and things and I'm pretty sure because I mean as Jakob mentioned we published in research recently that links management practices to innovation practices and that involved linking the management expectation survey to data from bird, the research and development survey so it's definitely possible and we've produced research using the results and I'm sure if they wanted to get in touch we could provide a bit more information on how to do that. That's great. Thank you Tom. So to the moment if you wanted to then talk a little bit more about what we're up with with Tom afterwards and then a question to what extent is sector-based productivity analysis viable in the productivity microdata? Again it's not quite clear what they mean by viable but the productivity microdata has all the different reporting units classified by their SIC code for industry classifications and as I've explained one of the big challenges and putting together that longitudinal business microdata is to do the conversions between the various generations of industrial classification so it's definitely possible and in fact it's a big area a major area that we would usually publish our research a long load of that split of different industries so it should definitely be feasible. I guess that's the great thing with all the business data having the standard industrial classification codes in such detailed form it does enable the analysis at those quite detailed levels. So I think that brings us to the end of our questions then Tom so thank you again it's a really interesting presentation and thanks again to the audience for your questions. Okay it's one o'clock so welcome back everyone I think there's people still trickling but we'll start the panel because we only have half an hour so I want to welcome Nick O'Donnell, Francis Pottier and Grant Bittner who are going to do a Q&A session from research ideas to evidence based on the use of economic and business data I'm going to let them introduce themselves and Nick is going to be asking them some questions and we'll have a short amount of time probably a bit less than 10 minutes for questions at the end. Again if you have any additional questions that you want to answer later just pop them in the chat and we will try and formulate a response to that so I'm going to hand over to Nick and he will get the panel started. Thank you very much Louisa and thank you to everybody for attending this session I do appreciate we're competing with both lunch and a nice sunny day and perhaps some people are doing both but so thank you very much and that really interesting morning so hopefully we'll have the same this afternoon so my name is Nick O'Donnell I'm Head of Engagement Impact and Insight for the O&S Secure Research Service and very recently the Integrated Data Service I'm very pleased to have two fantastic panel members today both of whom have got longer job titles than me and so we've got Francis Potier from the Department for Business, Energy and Industry strategy and Francis is Head of Business Statistics Data and in the data analytics and business stats division and then we also have Grant Bitzner who is the chief economist and director of macroeconomic statistics and analysis in the economic, social and environment group in O&S and I'll hand over to both Pest First Francis and then Grant to briefly introduce themselves. Hi everybody so I'm Francis Potier and as Nick says I'm Head of Business Statistics in Bays my team produces four monthly annual and one bi-annual official and national statistics releases related to business and we also provide some data into the Secure Research Service and the UK Data Service from that we also support the use of business data across the department most noticeably the Interdepartmental Business Register but also just general research, business data use, anything that Bays needs Hi and I'm Grant Bitzner. Thank you Francis and Grant please Yeah sorry Grant Bitzner, chief economist and director of macroeconomic statistics recently announced economic, social and environmental group in O&S which accounts for pretty much everything except census populations statistics and migration at the O&S so about 80-90% of what you see on our website Thank you very much Grant and so I'm going to start with questions, I've got three questions for Italy, Francis and then I'll move on to Grant and then there's time as Louise said for questions from the audience to ask so please do record those in the chat and then we'll work through those following the initial round the questions but also just before I ask Francis also wanted to just say the theme of this session is really one that's very close to my heart, we do an awful lot of work with researchers and policy makers data owners in terms of really trying to showcase the value of the research that takes place and how it informs and provides the evidence for policy making service delivery improve understanding of the economy and society so something that's I think really important to make sure that we really realise the benefits of the fantastic research that takes place so with that engagement perhaps in mind for the research community my first question for Francis is how does Bayes currently engage with the research community please Thanks Nick, so we have a range of ways in which we engage with the research community the first and probably the most substantial is that we directly procure a lot of research ourselves so the latest figure I could find was over £15 million a year on external research now that's not just business that also is the energy and climate change and labour market elements of the department but it's a lot of different projects and that covers just a wide range of things large scale regular surveys so my team for example, it's a lot of business survey every year but also small sale qualitative work one off surveys literature reviews just a whole range of stuff that's a big part of it and that's through a lot of different elements researchers so the academic researchers it might be some of the big research organisations or it might be companies that provide data collection services we're also often approached by researchers directly to support their research work sometimes that's because it's using data that we're providing again like the surveys that my team runs it might be because they're looking for some financial report support or it might be that they are looking for support accessing some hard to find data or just help to get visibility for the research that they're doing help it get some traction actually through events like this although obviously that was a lot easier when we were all in person somewhere but it's quite common for base analysts in particular or policy people to go to a conference or to go to a seminar and get talking to the researchers if it looks like it's useful work for base policy development specific to one of my surveys the long-tune business survey we put aside some of the annual budget to fund a handful of small projects every year that's usually four to five projects using the data set and then working with the Enterprise Research Centre we hold an event to publicise the research done and that's part of getting the data out there and making sure that people are using it and the great advantage of the Enterprise Research Centre apart from the brilliant work they do is that they can host us in the shard in London which is always a very nice venue with lovely views and then finally we are also data producers ourselves so obviously we talk to people who are using our data either the publications that we produce or using the micro data that we provide to do research themselves so I will talk to a lot of people particularly as a long-tune business survey or UK innovation survey which we manage and that might be anything from asking questions about the data pitching ideas to us and seeing if they can get our support for that Thank you Francis you mentioned a couple of data sets there UKIS and the long-tune or small business service I mean maybe you can expand on those and other types of secure data that base you or find particularly I think next frozen I think what you were saying is could you tell us a little bit more about the innovation survey Yes of course thank you Kim or me which was the problem My team research questions Thank you Now just following on in terms of Nick would you be able to mute yourself so Francis can speak is that okay because we are having problems with your connectivity until the next question is that okay So directly the team manages the long-tune or small business survey and the UK innovation survey the long-tune or small business survey is actually a survey of SMEs most small and medium enterprises don't blame me for the name it was named before I was working on it and that is an annual survey which is contracted out it's a telephone survey started off as 15,000 businesses there's been some nutrition over the years we've also had some boosts back up we try and survey the same businesses every year as far as we can with a little bit of top up for survey nutrition but also to make sure that we're still taking on boards of new businesses and it's still representative and that covers just a range of issues that SMEs might need to deal with business finance, energy, innovation, exporting it was mentioned this morning that we have some data on the ethnicity and sex of business owners in there we serve the sort of uses across government the devolved administrations we put in questions for them and then we make the data available through the secure research service and through the UK data service for researchers to use. We also publish results ourselves so last month we published two reports on employing and non-employing businesses off the back of that survey this September we'll be publishing the panel report which is a specific look at those businesses that responded in each of the last three years so that's attempting to make some use of the longitude and element. The other one is the UK innovation survey which is carried out every two years and the data collection is done on basis of the half by the ONS and that follows the OECDs I always get this wrong whichever manual it is that does innovation so it's internationally comparable and that's collecting information about innovation activity by British business our sample is businesses 10 employees and over. What activity they're doing what are barriers to innovation to drive this to innovation the type of people that they employ to do that kind of innovative work The latest survey is in the field at the moment and we'll go on for another month or so. We expect to publish the first reports of that one in sometime probably spring next year detail still to be decided because up until now our timetable has been decided by Eurostat on our behalf and obviously we're not quite so we're holding to that anymore so those are the two data sets that we manage otherwise we make a lot of use of the entity departmental business register which is also available as the business structure database through the data services we hold extracts within Bayes we use it for a survey work so both surveys I've previously mentioned are sampled through it because it's useful for us to be able to use the same survey register as ONS using for their surveys we are able to share it with ONS permission with contractors for carrying out surveys but we also use it for analysis for things like identifying control groups if you're trying to evaluate the impact of intervention it's useful to be able to identify a control group for that sort of thing and also cluster analysis so when the industrial strategy was published in 2017 we've published some cluster analysis based on the geographic information in the IDBR finally we make use also of all the other ONS business microdata annual business survey, annual purchase survey business enterprise, R&D, all of them we're very into obviously an R&D innovation as part of the department's policy objectives but also looking at supply chains we're looking at energy consumption, we're looking at domestic production comparing it to trade data and we're very interested in productivity as well can you hear me okay? Thank you very much, that's very comprehensive and really interesting and the wide use and diversity there from the data so final question really for me, are there any particular aims or areas of research interest that you would like research to answer through the use of business data please? Yes so I had to look at our most recently published areas of research interest document to make sure I have all the coverage for this 2020 which is obviously excellent timing, everything that's happened since then but it is interesting, it's still very relevant so we're very interested in, it's been mentioned before this morning that the levelling up agenda, we've always been interested in elements of sort of place based analysis relating to business, clusters, agglomeration so anything relating to those areas we're interested in supply chain particularly bottlenecks and vulnerabilities which has obviously become an even greater thing over the last year, 18 months or so we're interested in R&D, how you turn innovation R&D into business growth and business success and we're also interested in things like business perception of regulation what businesses think of the policies that we're putting out and how successful they are. Thank you Francis, that was really informative and again I've been scribbling everything down there because colleagues on the call who manage our impact insight work is so valuable for us to understand how the data, how that research subsequently translates into policy or service delivery. Perhaps if I can move on to Grant and thank you Francis I can ask the same question of Grant, are there any particular aims that you would like research to help answer through the use of business data? Well many how long have you got Nick? Let me run through a list of things working on what we're currently doing or we plan to work on or we'd love to engage with colleagues on the call on. I think the starting point though is as with the Department for Business and other parts of Government obviously we're looking to be aligned with Government's top priorities and in case you're wondering what those are at the moment you may have heard that net zero and levelling up the top two priorities and certainly we're doing work on both of those areas in terms of subnational data statistics analysis and net zero and broader environmental issues. On so-called global Britain we're certainly doing some work on trade and we will be looking to do more and then there's a whole issue around recovery from the current pandemic and recession and the impact that that has had or will have as well and I'd say there's a particular focus at the moment on subnational research and analysis there'll be a white paper published by the Cabinet Office later this year around the time of the budget. Shortly after that we're expecting to publish a subnational data and statistics strategy which was kind of set out both what we're doing at the O&S and across Government in this space but also what our ambition is to do in the future and that will I'm sure include a lot more granular local and regional data hopefully available in flexible geographies and a lot more research and analysis in this field because we certainly we want to better understand what the drivers of local and regional job skills and prosperity are. In addition to that some of you particularly those who are familiar with their work with ESCO will be aware that the O&S has been one of the leading national statistical institutes doing so-called beyond GDP work moving beyond the current systems and national accounts so that includes quite a lot of work around natural capital, human capital, intangible investments, environmental assets, digitalisation etc I will put a link in the comments to one or two discussion papers that we've published which give you a broader sense of that but that's quite a broad agenda, a lot of that is new work, we welcome engagement with academic researchers on those. Now in addition to those I wouldn't be a director of macroeconomic analysis without talking about some of the more traditional macro themes so we're really interested and will be doing more work on economic scarring and the extent to which for example Government support through the pandemic may have mitigated the extent of corporate failure we're also relating to that extent of structural transformation in the economy that we are likely to see as a result of the pandemic and the recession now of course some of those questions we would need to wait a year or two until we have more up to date data before we can really start to analyse those but nonetheless those are kind of medium term research questions where I think there's a lot of opportunity to do some really interesting work. Relating to that understanding the long term impact of EU exit we're seeing major changes in global supply chains and supply chain constraints there's lots of talk that people will be diversifying their suppliers there'll be real localisation less use of just in time etc of course it's hard to unpack the effects of leaving the EU from the effects of the pandemic and the recession because that will happen broadly concurrently but certainly that's an area we'll be doing more research on over coming years and just to pick one or two other things up I think there's always room to do more on industry and business dynamics and for those of you who saw Thomas Wickinsons run through things like the longitudinal business data set etc and the ARDX I think will give us really good tools to really be able to answer some of those questions in more ways. Also I think a bit more work around industry concentration an area we haven't done as much work on ASA in the US but I think there's real potential there and then finally because I know this has been quite a long answer we're increasingly publishing real-time indicators a lot of those are not currently available in the SRS but we will be looking where we can to include them and to link them to other data sources and we're really keen at some point to be able to share things like the retail scanner data that we're now receiving from a number of major retailers obviously in an optimized form and potentially even other data sources such as the HMRC pay as you own data that may be at a relatively aggregated level initially but certainly just opening up all those high frequency and real-time data sources where we can and it also pointed out there is some interesting data already that we publish every Thursday morning including things like traffic camera and flight information as well so there is a lot of it is not necessarily ideally suited to academic research unless you want to use things like now casting models but nonetheless it's an area that we plan to grow as well and to make available where we can for research. Thank you Chris and that was incredibly comprehensive really interesting. One of my own observations from the research in the secure research service is how few climate sustainability projects have taken place so it's really exciting hopefully to see the drive towards that and see if we start to see a lot more research based on the data being made available so my second question is by averting a little bit away from the mainstream so O&S has just gained research organization status with the UK research and innovation so could you say a little bit more about the opportunities for collaboration you think this might bring? I don't think we've really thought through what the implications are but let me just give you a practical example. Recently the ESRC awarded a reasonable chunk of money to work on productivity and it saw the establishment of the productivity institute at Manchester University. We had extensive discussions with the consortium ahead of that bid going in but we weren't able to join as an equal partner in that consortium whereas kind of the poor cousin in the sense that we would just blow a lot of the data but we wouldn't need to be a research partner. I think this puts us on a more equal footing. It means that we could not only join consortiums but potentially lead bids for research funding and I think that just gives us a lot more opportunity to partake of the wide range of researchers out there but I think it's going to be a bit of a slow burn. I think first we need to identify what areas we think there are gaps or opportunities and really just start to engage with people and talk through what's possible. We won't be able to do everything but I do think that that would see us involved more actively in a wider range of research than we currently are. Thank you. Staying with the theme of engagement, what avenues do O&S offer for engagement with economists and researchers using O&S business data? Well, I saw that Francis talked about the areas of research interest unusually O&S haven't published anything of that nature for a while now which doesn't help. Although you only have to look at our website to kind of see what sort of areas of research we are interested in. We, following the public sector research establishment status, we are working on a fairly comprehensive research plan and I'm hoping we'll also publish in areas of research interest on gov.uk in the next three to six months. My starting point if I was a researcher would be to have a look at that when it comes out and then have a conversation with us if there's something in there that you would like to work with us on. In the meantime, we are working closely with a number of research partners. The Economic Statistics Centre of Excellence ESCO has a broad range. I think about 20 projects at the moment that we're working with academics across the UK and internationally on if you have a look at their website that will give you a good sense of what we're doing there. We are doing three projects in partnership with the Turing Institute. We're also working with the Carter Research Partnership and of course we engage with researchers through the SRS and other forums. My keepers of advice there is if you have an interesting research idea or you see something on our website that touches on an area you're working on. I think my colleagues have run through quite a number of areas this morning that I think people may find of interest. Drops an email, come and talk to us. Tell us what your research idea is. I mean one of the things that's kind of surprised me in the three years I've been at the ONS is I would get maybe one or two academics a year coming to me by email or in person or over the phone and saying I've got this really good research idea but I can't get the data or I'd like to work with your next. It happens a lot less than I would have expected given the huge amount of data that we're sitting on and which is only going to grow in the future. So I think there are some missed opportunities. Let's have a conversation. Tell us what you're planning to do and if we can make it work and we see value in it then let's get on with it. Thank you, Grant. I think I totally agree with the point there. Each of the events that we run and the research capability event that we're running in a few weeks time, the very much focus is being on that engagement right across the research community and the collaboration and by joining up together and sharing insight, sharing ideas, innovative messages, a method sorry, that's very much where we can collectively be more effective and sharing our understanding and our knowledge so I very much agree with that. Thank you both for answering all my questions so I'm going to open it up to the floor if anybody else would like to ask a question from either or both of our panellists and please do write something in the chat and I will follow up. I can see a question from Brian so does the I think this perhaps is asked for both. Does O&S have a directory that academics can access? As an academic I don't know who to approach. I'm not quite sure what that directory might be but if that would chime in with you Grant. Well we publish quite well with the directory. We publish thousands of bulletins and research and articles every year and as I said if you look at what we publish on business dynamism is on trade, on productivity etc everything we publish has contact details on it. If you want to talk about productivity talk to our head of productivity if you want to talk about some of the work we've done in the industry business dynamism look at the latest article we published on that and get in touch with the authors. We're open, we're accessible we could publish a list of who the contacts are for each particular area of what we do but A that would be a very long list and B it would very quickly get out of date I suspect. Grant I just wanted to add in there that I know that there's a guide coming for working with academics as part of the research that's being written now I'm not sure when it's coming out but I think that will be a very good port of call for academics to go to first. Yeah and I think for one of clarity people should assume that almost everything relating to business and economics statistics and research were involved in one way or another there's almost invariably somebody in the organisation who is the authority on a particular area if you have trouble reaching them find out who that is just ping me an email forwarded on to the relevant colleague. So now I can ask for if Francis has a view so the question that's been asked by Scareis is it would be great if government departments published an end of year review of those areas of research interests with links to relevant published work or a note on which projects are in progress. I mean I think that's probably something where we and ask could work together with the base and other organisations because obviously we a lot of that research takes place in the SRS so if there's anything else you want to add Francis. I said the area of research interest document does have some links to research that was I think in most cases complete and have been published in areas relating to other elements in that document and you know we also publish all the statistics and most of the research that we've carried out as well. I'm not really in quite so organised a passion as I think Karris is suggesting but certainly the areas of research interest is a good starting point and also the research and transparency areas of our website. Francis. Just a small question and then we need to wrap up for the breakouts but one more question is brilliant. The question is from I think Tien Chu if I pronounce that correctly. The question is there are quite a number of data and some of them are overlapping with each other in addition there are so many documents for each. For a new user these information sometimes appears to be confusing for example the business expenditure on research and development stroke innovation survey so I don't know if there's anything either of you can add on those but I think part of the answer there I'm sure is the data catalogue developing for instance for the secure research service which will provide a lot more comprehensive metadata and helpfully guide researchers to what the how the data can be used with some examples of where those data sets are used that we can provide from existing projects but perhaps if there's anything else briefly that grant or Francis might want to add on that. So I have one thing which is my team publishes business population estimates which overlaps quite a lot with a couple of O&S releases the business democracy in the UK business activity size and location. So one thing we've done days working with O&S is to publish a document that says which one you would want to use for which purpose. So I think that kind of thing is quite useful. We also try in the innovation survey and I think the third team do as well to refer to each other's releases in our release and why you might want to when you might want to use our release and when you might want to use the other one. So I think that's quite a useful way of going about it and most stats releases these days O&S and other departments will have some part section of it that is sort of associated information or other information on this topic. Yeah and just a tip from me in terms of finding the right person or the right piece of research or article on a topic the search engine on the O&S website is not great. So just go to google key O&S and whatever topic it is and you'll find probably the most recent article on that and from that you'll be able to work out who you need to speak to. I agree with that the gov.uk website as well. Googlers have often a more successful search engine to use. Thank you very much Francis and Graham it's been a fascinating session. I've learnt an awful lot and I've been there in this a few years but it's been really informative so thank you both and thanks for your time and contribution. Brilliant thank you Nick and thank you Francis and Graham it's really interesting. I think for some of the questions that haven't been answered we'll provide responses afterwards but thank you so much. So hi everyone welcome back so Louise and I just wanted to say a few final words to bring the conference to a close and I just wanted to make three short points because I'm sure you've all got other things to be doing and so my first point is to say that we are so lucky in the UK to have access to these incredible data sets that the O&S presentations have highlighted and without access to these data via the secure research service and secure lab then it would be much more difficult for research is so huge and thank you to people who work within those two services to provide these data. Second thing I wanted to say is that the aim of these user conferences is to provide a forum for that vital link between researchers and data producers such as O&S and that has definitely been achieved today. It's been evidenced by the really lively Q&A sessions so it's been I think it's been a brilliant day and then my third point is to say some big thank yous to everyone so thank you very much to all the presenters who've given their time and all the chairs as well it's been a lot of background preparation for the presenters and shares and to everybody here as attendees for creating such interesting and lively discussions around the presentation topics you've all been really good so thank you again to our colleagues who work in secure lab and secure research service for hard work throughout the year and then very finally an enormous thank you to my colleagues Jill Meadows and Saoirse O'Kellaham from the UK Data Service who work tirelessly behind the scenes to ensure the admin and the logistics of these conferences run so smoothly and they organise so many of these important events not just this one for the research community so a huge thank you to both of them for the work that they do and I'm going to hand over to Louise just to say a few words too. Thank you Vanessa just want to echo it's been a really brilliant day fascinating I think I've learnt something we've all learnt something and thank you all for sticking to time and following the rules and adding the word question when you're asked to do questions so I think it's been really nice and it just shows how these virtual conferences can bring so many really skilled people together just very quickly if you have any additional questions for the ONS teams please root them to my email address on there that's the development impact and we'll root them through to the people and we want to try and answer some of the questions probably in a follow up blog coming out and we'll be posting that on the UK data service and we may have an FAQ on there Vanessa as well together with the presentations. The other thing is both of our teams Vanessa's UK data service and the SRS we both have impact teams that do work and to try and work with researchers to accelerate their impact through case studies through knowledge exchange events like this please let us know if you'd like your impact accelerated or feel that you have produced something where you'd like us to help with just dissemination so that's just an open invitation to get in touch with us but otherwise thank you so much for attending and thank you Vanessa and team for doing such brilliant organisation as well.