 So this session, we're looking at from tax data to tax policy steps to take. It should be very exciting. In the past few years, decades, there has been about four major developments in tax systems. I think the first has been the introduction of VAT. That was the major change. And then there has been generally a gradual reduction in tax rates. I remember at one time in my country Malawi, sorry, my name is Ozon Ali Gomega. I'm from Malawi, I think I should have mentioned that at first. At one time, the copper tax was 45% but generally in most countries, these rates have been going down. That has been the second major change in tax systems. The third one has been the introduction of semi-autonomous revenue authorities. And the idea was that once these revenue authority functions are separated from the Ministers of Finance, they should be more effective in raising revenues. Whether that is being achieved, it's another discussion that we can have. But they are more interested in the fourth development that has happened in the tax system. And this is what we will be discussing here in the session. And this is the increase in automation of tax administration functions. And this relates to tax registration, tax filing, tax payments and the whole automation systems. And once the tax administration functions have started to be automated, we have generated a lot of data. Data that before maybe would wait for a survey or some experiments. Now we have generated a lot of tax administrative data. But with this lots of data that we are generating as tax administration, the question comes, how can we level it on this opportunity that we have of having lately available data, tax data, to make impactful tax policy to see how effective a policy is. So I have a panel of four individuals who have experienced in using tax data for policy purposes from different countries. Who discuss this issue and share with us their journey of transforming this administrative data into policies, how they are using it. So allow me now to introduce the panel that will be discussing this issue. First from the far end is Justin Makubu. Justin is chief revenue officer in South Africa. January he is responsible for ensuring tax compliance. And I'm sure to know who is compliant, who is not compliant, has to use data. So we should hear more from that South African perspective in terms of that. Then next to him is Tina, Tina Kaidu. Tina is a manager of research and revenue modeling in Uganda revenue authority. She is responsible for overall coordination and supervision of the research lab activities. So you can't have research lab without using data. We should hear more what Uganda is doing, transforming this data into policies. And then next to her is Josepha Damame. Josepha is chief economist and she is with Tanzania revenue authority in the planning section. Last but not least is Aayanda. Now I have to pronounce this carefully. I hope I have tried right. Aayanda is from South Africa. She is director of secondary sectors at the National Treasury of South Africa. And she is mainly focusing on microeconomic policy unit in the economic policy division. Her focus is mainly on industrial policy, manufacturing and trade. So thank you colleagues for being here. And I think without much talk from me, we should hear from this panel. And I have a few questions that I'll be asking the panel. So the first question I want to ask is what is the importance of using data for tax policy? What has been your experience and what is the importance of using tax data to do policy? So I think the perspective that I bring is through the national treasury and the economic policy division. And so we work with a lot of economic policy, policy making, evaluating policies, assisting whether certain programs are working or not. And so essentially what we've tried to do in the national treasury partnering with SARS and UNUWIDERS have a program that's called the SATIED program, which essentially advocates for more use of tax administrative data for evidence-based policy making and public policy evaluation. So in essence, this was put together to try and unlock the potential for tax policy data beyond just assessing tax policy because we've had studies that look at the employment tax incentive and other tax incentives, but also beyond that and pushing the envelope a bit further. In terms of research, we've also looked at various themes including firm behavior, productivity, innovation and concentration. In terms of firm productivity, for example, if I can just give you a sense of some of the papers that have come out of those work streams. We've looked at how misallocation of labor and capital has affected total factor productivity, keeping it below its optimal levels. We've also had papers that have explored whether their productivity spills over when workers move between firms and this is enabled by the tax data because you have a data set where you can match firms and their workers. I think that's one of the unique aspects that the tax data allows for us and as well as innovation. In terms of innovation, we found that even though South Africa is not necessarily as intensive in terms of innovation investments, we still see quite a jump in terms of the returns to innovation through the data and that's through the R&D tax incentive. We can also analyze because the tax data is structured with the personal income tax data, you have the corporate income tax data, you have the value added tax data, but you also have the customs data. So when we are able to merge all of these components together, we're then able to look at trading outcomes for firms. So we've been able to look at international trade, multinational corporations, as well as export performance of firms in South Africa, which I think is the most granular we've had for some time. So we've learned a lot about the types of firms that operate in the South African market, what they look like. We've learned things like most of the firms that export in South Africa are simultaneously importers and that's data that we didn't have before this. We've also been able to look at whether firms in South Africa actually respond to exchange rate fluctuations in the way that we'd expect them to and we found that actually they don't. It's different, it differs by destination, it differs by product, it differs by firm size. And we've also been able to look at temporary employment services, finding that the impact, the penalty from wages through temporary employment is actually a lot higher in South Africa compared to what the literature would say. I think I'd be wrong to not mention the wage and income dynamics literature that we've also produced because this is one of the most comprehensive sources of income or wage data that we have in South Africa currently, which covers a broad scope of the wage distribution, including especially the high end, which is sometimes not being captured in surveys comprehensively. So this has basically allowed us to rigorously and at a granular level look at worker and firm outcomes in South Africa as well as being able to credibly analyze wage and income dynamics. I think I'll stop there. Thank you very much for that. Just if we are already hearing how tax data is being used to understand various factors in the economy, but I want from your perspective to know why is it important to make tax data available. Why do we have noted for most countries tax data is the sensitive data is data that maybe is not being shared with the researchers and other institutions. It is only available to the tax administration. Why do you think it's important to start sharing this data and how best can we do that from your experience? Thank you so much. First of all, why data is not being shared to other institution or to other parts. It's because of the sensitivity of the data, of the taxpayers data, the information which in taxpayers data are so sensitive. So to share the data you may lose or you may expose the data to other people which is not necessary. But the tax data are very, very important to inform the tax makers. That's when you can share through the Minister of Finance because they need to make some policy changes when they need a rise. That's when you can share. But to other institution is so restrictive because we just want to maintain the tax payers information. I think this is cut across all the revenue institutions that we have a law which restrict to share the information unless it's shared through the revenue authorities. I want to add that tax data is very important especially to tax police people because if I take my example of my country and I work with the Minister of Finance, you cannot change any police without any data. So when you want to change any police in your country, let's say we want to increase some revenue rate, you should ask data from the tax administration people so that the government is informed that I am moving from this rate to this rate. What will be the impact of increasing this rate? Let's say I want to increase the rate of VAT rate from 20 to 20 or from 16 to 18. So that one you have to get it from tax administration so that you make a suitable policy change. Otherwise also if we want the government want to increase the salary to the people, this number has to come from the tax administration people in order to do how much or what rate should the government want to increase and what is the impact will be going to happen. Secondly, if also the government want to reduce sometimes you can attract some people production to agriculture so we want to reduce some taxes maybe import to attract people to import some raw materials. You should have a data to know how much the government is going to give or is going to lose in terms of tax because you are collecting this amount and I want to reduce some amount so that I attract production or I attract production from agriculture sector. You have to know how you have to have the data so that you do your policy change. So it's very important the data is very crucial in our development, in any activity or any policy we want to take in the country. Thank you so much. Thank you very much. I'm still sensing that yes this data tax data is still confidential so it's either you get it through the revenue authority or the ministry of finance. This is where we are sharing it but I want to hear from Uganda Tina. I know that you've partnered with some of the research institutions the ICTD to do a lot of research in Uganda. How are you able to share this tax data with these institutions that are even not Uganda institutions and why did you see that this is important. Maybe you could share the Uganda experience and how you're doing it. Thank you. From Uganda's perspective it's been quite a long journey for us, a journey of collaboration with different research agencies and over time we realized that we needed to structure this process of availing data for research purposes. As you rightly put in your introduction that we've automated revenue administrations have automated and Uganda inclusive and we've generated a lot of data. However this data has to be put in a structured format for it to inform policy. Revenue authorities primarily collect data for government to meet, to exercise our mandate. However this data can also be used to inform research, to inform policy and administrative reforms. So when we realized that we had this resource and we didn't have the capacity we had to partner with other agencies, ICTD, UNUIDA but again in order to manage this in a structured way we had to work through memorandums of understandings, nandisclosure agreements and this took quite a lengthy process for us to actually formalize the arrangement and all that. So over time through sharing with our partners we were able to come up with a research data lab that is stationed at the URA. So in this lab we collect the data, we anonymize it, we document it and make it available for researchers. We work on mutually agreed research areas or sometimes we actually are informed of some areas that are critical for domestic resource mobilization. So we work hand in hand and make this data available. Through the MOUs we also are able to be part and co-author papers with international researchers which has been a plus for us as well. So when research is conducted and insights are drawn out of this we've been sharing across the board with policy makers, with tax administration and this has helped us improve our policy making process. So to us this has been a win on our side because we generate detailed data from the formal sector that is paid through our systems and the frequency of this data is usually on a regular basis as opposed to surveys which are done periodically. So we believe a lot of insights to improve our policy process are being drawn as a result. Thank you. No thank you very much for that. But yes so how are you overcoming the issue of the confidentiality of data and as we've heard from Uganda you know you can have an MOU and maybe the data can be shared anonymized and that as you're saying it's proved to be quite useful to have robust results because you're collaborating with institutions from other countries and even research institutions that can share knowledge and skills so very exciting. How about South Africa? Justin maybe we can hear from you. Why do you think it's important to be sharing the tax data and how would you go about it? Are you collaborating with other institutions and how would you share this data and what is important? Sure I think without repeating what the colleagues have said looking in what the tax administration becomes very important to have a data governance framework because I think without a data governance framework that governs who has got what authority over what data it becomes quite blurry in terms of what governs the data within the tax administration. So as a South African Revenue Service we found it quite fitting that we define the data governance framework within the organization which defines who is the data governor for what part of the data within the tax administration and therefore that person owns the data end-to-end including pronouncing what happens in as far as sharing of the data is concerned. We quite some time back almost a decade ago had collaborated with Univida International Treasurer and as said to create a data lab because we felt that whilst there is many problems that need to be defined that lead to policy formulation those problems are crystallized better in the data. So we pushed through almost monthly to the data lab at National Treasury almost real time taxpayer data and I think without a database or a data lab it would become very difficult to be able to share data of course things such as confidentiality is quite sacrosan to us and that's why we anonymize mask data we are very careful that we don't inadvertently expose taxpayer information especially where you've got very limited number of taxpayers in a particular geography and you are doing geographical analysis analysis as a researcher we want to make sure that we don't inadvertently expose confidential taxpayer information. So we do mask we set thresholds in terms of what's the minimum we should be able to expose in the data so that we avoid inadvertently putting out data in the environment. But I think it's actually a virtuous cycle the more we share information with researchers the more we are able to amongst other things have policies policy decisions that can be conceptualized that help amongst other things with domestic resource mobilization so that we are able to continually feed into the revenue cycle in terms of collection of revenues. So I think there's a symbiotic relationship that exists between us responsibly sharing the data through the data lab as interfacing with the research environment in the academic space specially through Unwider but also other academic institutions that are out there and maybe lastly what made it comfortable for us to want to approach the issue of the data lab is that almost 14 years ago we started publishing tech statistics for the country and that information in the public domain is very rich for researchers but even more better if that information can be puzzled at a very granular level because again if you do publish tech statistics sometimes they are at a very aggregated level and you need it at a disaggregated level. I think I'll stop there for now. Thank you very much. I think what is coming out clear from the panel is that it's important to be sharing this data and what you can do with this data and what you were talking about. But I think two important things that have come out one is have to have you need to have a proper framework of how this data is going to be shared because at the end of the day still taxpayer data it's confidential data how best do you share that data. And second it's this aspect of having a dedicated unit. I'm sure it's doing whole aspects of data governance which I think is coming out quite clear and it seems very important. But now coming back I mean I want to ask you what has been some of the challenges of using this tax administration data that you've experienced. I want to come back to you Tina. Well one of the major challenges we've experienced is the data quality. Like I mentioned this data was originally collected for tax mobilization but however we are now using it to inform policy. So having it out there we have been obtaining a lot of feedback from researchers and in a way this has helped us improve address some of the gaps that are being seen within the data. So for us what would seem like a challenge has been an opportunity to improve the quality of data but also tax administration data is prone to misreporting because of the self assessment regime that we have taxpayers are declaring and telling us what they would want us to see. So when we are using this data to conduct research we need to be cognizant of that so we try as much as possible to go through a rigorous data cleaning process to structure the data in a format that is usable for researchers. Thank you. Aayanda how about you you're using micro you know this data to understand micro aspects of the taxpayers. What has been some of the challenges of using the tax administration data. Okay maybe I'll answer your question in two folds so maybe I'll talk to some of the some of the challenges that we also faced in South Africa in setting up the data lab for example just to take a step back. You know I think data sharing in any case as everyone has said you know you start by finding the you know you have to make a legal case for why you should have the data in the first place. So I think maybe it's important to give that context that you know this is a long journey if you date back to where we started I think Johnston talked about it's over it's been over a decade I think since we started developing this. And so it's a work in process you know in every year I think we learn new challenges we need to learn to solve new challenges. Capacity has been one of the greatest challenges I think we've also faced just in terms of getting the right kind of skills and people that you can work with who have the vision that you want to lay out for the lab. From recruiting the right kind of skills for people who are data scientists who can clean the data who can actually run quality checks on the data. Who can transfer the data in a in a manner that is secure that doesn't compromise the data because it's important for us to ensure the integrity of the data in the data lab. We had to put in place a lot of systems to ensure that the data that's extracted from the lab is of good quality and research quality. And that you know we we try to limit any exposure to taxpayers as Tina had already indicated this data wasn't intended for research. And so you know I think even I as a taxpayer if my information ended up in the public domain I'd be very frustrated. So there are a lot of processes and protocols that we've had to put in place. And I think even with our lab being quite developed at this stage we still find little glitches every now and then where we see that actually we can't make certain calls on a case by case basis. It's important to have frameworks in place policies in place that eliminate human era and so that everyone understands what procedures and processes need to be there. So to the best way possible I think you know those are some of the other challenges that you face in trying to allow data accessibility to to to users and researchers broadly. But also I think some you know some of some of the other challenges that you end up facing and I guess I'm pre empting sort of where we want to get to in the future starts allowing. Is you know with with taxpayer data or tax administration data not having been positioned for research. You'll find that the elements that you'd like to pull into the tax data so that you can better understand the individuals that you're capturing in the firms that you're capturing and other characteristics. That give you a fuller picture of the economy and who you're actually looking at and what else could be driving the results that you're seeing. So I think that you know you then move into a phase as well where when you're thinking of expanding and how you you grow from where we are now incorporating new data sets further risks de identifying data. You know the things that we have to put in place ensuring that we're still compliant with the puppy act which is the protection of personal information act in South Africa. So we've had to learn that their processes on getting exemptions from the information regulator which we're currently undertaking to try and see what's possible. What else you can add unemployment insurance information some procurement data as well. Yeah I think admin data. I guess what I want to highlight from this lengthy speech is that you know I think there's so many challenges it's such a long process. And sometimes when we talk about all of the results that we've seen and that we're able to get from the data it doesn't really show just how much work and effort has gone into it by everyone who contributes to the project. Yeah. Very interesting to look at all these challenges. I'm hearing a lot of data quality issues that we needed to think about even about how do we get the right skills to analyze this data. Do you want to add anything on the challenges that you may have first while using this tax administration data. Yeah of course the issue of data quality is very crucial because we are lacking some training on how to manage the data which we are collecting for taxes. The tax data which has been collected there we are collecting data for taxes which are not meant for research purposes but also we need to have some capacity building in collecting that and keeping this data which we are collecting. So we need a clean data and so we need a capacity building so as to know how are we going to collect that and how are we going to manage the data which we are collecting. That is a challenge to us also because as I see back home that we have a lab and we have some people who are working with the data tax data so those people need to be incapacitated so as to produce a clean data when it at all is needed. Produce a data which can be used all over and the accuracy of the data which is being kept by the tax administration. Thank you very much for that. Now I want to look at another question and this is the linkage from data to policy. How do we make sure now that we are using this data for tax policies that are impactful and that will be taken by the policy makers. You take this information you've analyzed the data and the minister would say yes I think this is a policy that I should take to parliament. Take me through the process how best should we be using this data to make sure that we are having policies that are being taken on board. I want to start with you Janssen. Look I think firstly it starts with an understanding that it's not one paper that is written that influences policy. We often get asked questions to say about the paper as tax administration but we give us information and we publish the paper and when is it becoming policy. And I found that the appreciation and understanding that it's a multifaceted approach to take a number of papers to bring a particular tax policy to pass. I think it's important to make sure that as we do research the likes of national treasury, tax administration and academic institutions are always in the loop because we don't want to wait for the process. We want to make sure that the research that is being done from when it is identified and sanctioned it is meant to address policy challenges that are pressing in the first fiscal environment. So we don't necessarily do research for the sake of doing research. We are doing research in response to some of the pressing policy challenges that we have. I think for me it starts by making sure that the genesis of the research that we do is informed by the fiscal problems and challenges that we are trying to address as sovereign states. I think one that is achieved it makes it quite in fact it leaves in anticipation for that output from a research perspective to come through. We've been pondering in South Africa around the issue of wealth tax and I'm sure if you speak to political authority they want to know what researchers are saying about that particular subject in as far as the nuances of the South African economy are concerned. And therefore research that is tilted towards a burning platform is research that becomes relevant in terms of addressing some of the structural issues that we are faced with at a political level and at a tax administration level if I may add. So I think for me what is important is to make sure that there is relevance in the research at its onset so that there is a clear link to what policy problem you're trying to address through that type of research that then comes out. I think I'll stop there. Thank you very much. I think I'm catching the emphasis on robustness of these findings. It's not just one paper. We've done a research on VAT and it shows that if you increase VAT from 10% to 15% you collect more and you run with it without looking at other aspects of that policy. Thank you for that. You want to jump in? What are some of the steps that we should be looking at when we are transforming from this data with media available and how do we transform it into policy? So I think to add to what Johnston has put forward I'd say I think there are many ways to try and get the research to be top of mind for policymakers. Part of that for us we've seen has been collaboration between policymakers and researchers have been one of the best ways I think to equip to essentially change an institution to be a lot more research driven and evidence based. When the policy makers can almost own the research results and what has been put forward they are a lot more comfortable to be able to infuse that into policy position papers at Treasury. I think a lot of the research that we do falls into the policy position papers that we put forward that help us develop a Treasury view on certain things. Yes everyone does the research in their personal capacity but when enough evidence is there the department is able to say this is our view on the employment tax incentive and whether it's been impactful or not. This is our view on the R&D incentive, this is our view on localization policies or industrial policies. So I think that collaboration can't be emphasized enough and I think another aspect of things is for even when I write in my academic capacity I sometimes have to take a step back and reflect on the policy decisions or policy recommendations that I want to make in a paper. I think marrying these two aspects of things more intentionally when we write academic papers is important. Even taking those papers whether you're collaborating with policy makers or not and presenting them to government departments. It's a very different kind of platform. It forces you to package your research results in a way that almost forces you to crystallize exactly what the policy intentions are and what you expect the people that you've put in the room to take from yours. So I think that the way I've seen the most traction from research and people engaging in research is when you actually have academics coming to present in policy seminars. I wouldn't underlook that. So I think that all of those things show you that incrementally we really do try to infuse a lot of the work specifically with the tax related work because you know Treasury has also invested quite a bit in it. We try to also infuse some of the research into our budget reviews and documents when we have staff members that have produced papers for example and we're able to showcase that in our annual publications as well as Treasury. And I think that that goes a long way into infusing certain ideas that allow policy makers you might not directly see the connection but you're slowly influencing the thinking of the people who work in those institutions and they're able to then make better decisions because you're informing them and giving them more evidence to be able to make the best decisions that they can when they're trying to put together a proposal for a minister or for parliament or a cabinet briefing. Thank you very much. Justin says robustness of our results because it's just not one paper that we should rely on. We should be robust. We should be in collaboration. Let's understand the policy makers. What are they looking for and how do we present it to them in a way that they can consume. Again the same question. We have made this data available. How do we make sure that we're using it well after when we're using it for policy that is impactful. Take us through the process. What do you think are some of the important aspects when we have made the data available so that it's used well for policy. I think the main issue is to include some institution in data management so that we don't depend on one institution for the purpose of making policy. We have a different research institution. We can also collaborate with them for data management so that when we take data from the tax administration we can compare the data. Maybe the minister of finance also can be included in data collection so that when you receive data from tax administration you also compare with the data from the minister. And other research institution which are in the country so that the data which are collected from different parties can also be used and make a comparison so that we don't depend on one area. That's what I can say through that steps to take. Thank you very much Tina. What do you think? Creating awareness to the very relevant stakeholders using available channels. And this could be conducting dissemination workshops where you engage policy makers, decision makers, the civil society and share some of these results that are coming through the papers that have been developed. We've also done prepared policy briefs that I also shared with those stakeholders and this in a way is creating awareness of the data that has been used and the insights that have been drawn from the different studies that have been undertaken. Thank you. Maybe I can add that it's perhaps also important for policy review because I think it's not only when we are formulating policy but this data becomes very important for the purposes of decisions we've made in policy. We've made some decisions in South Africa around some retirement reforms and through our micro simulation at PIT level through the PID mod. We started to see the implications or the impact of the policy decision and I think another close of feedback loop that is important is the policy review feedback loop to say what do we then do based on the evidence that we are seeing post implementation of certain policy decisions. And I think it's important that that loop also be leveraged fully through research because once these policy decisions are made I've found that whether again it's retirement reforms or it's sugary beverage levies or health promotion levies these decisions need to be checked for efficacy. And research comes in quite importantly especially at the back of reliable data that then feeds back. I just want to raise the issue of policy decision reviews utilizing data so that we can be in a continuous improvement mode in terms of our policy environment. Thank you and I think that's important indeed that once we've made the policy we need to go back to see how that policy is operating. And I think awareness is also important so we've conducted this research we've used this data we have these results but so what. Sometimes it's not just for policy or even before we make policy we need to make people aware of what the results are saying so that they can even support the policy. No thank you very much for that. One of the limitations of the tax administration data I think some of them you've already started you've already raised them but also I've noted that for most of our countries developing countries we're learning on this tax administration data but this data is limited because most of the businesses are informal in which case they are not registered with the revenue authority that we do not have data or information about these tax payers that are not registered. How should researchers now start thinking about the informal sector and how we can start extracting some information that we can use for policy purposes because we have always had taxing informal sector how best to do it. So I was thinking this issue of data we are focusing we are taking data from the tax administration but this tax administration doesn't have data from the informal sector. What should researchers or tax administrations be doing in order to have information from this other aspect that the tax administration might not have. Ayanda you want to start? You know I guess we can't expect the tax data to answer all questions so having the tax data doesn't necessarily remove the need for survey lead research for example. So I think that they're still ongoing especially by the statistics departments as well who try to survey the informal economy and you still have that data as well. I think you can just use it in conjunction you know with the tax data it won't have as much coverage because you're obviously then you know you're limited to the areas that you're able to sample to capture informal activity but I think that's not to negate the importance. I think you raise a really good point. It's important to understand that we're only looking at formally registered businesses and formally employed people when we're using the tax administration data which I think has its benefits. I think I'd say that I'd say revenue institutions especially I guess SARS and Johnston probably can come in on this will still leverage the research that comes out of which parts of the informal economy can be better formalized and brought into the system as well. I think that's always an ongoing effort but I think that you know we should always leave room to separate the two where we have formalization because the burden as well of forcing the informal economy to formalize and have to report you know it becomes very challenging for them to survive in that way which is why I think we're always wary of increasing rate for even small firms as well to some extent that are informal. So I would say you continue running survey led research to get insights and understand the size of the formal economy and what can be done to help them and assist them so that one day they can transition and become formal so that you can start receiving revenue from them. But yeah I'd separate the two. Maybe to just add I think but by and large there is an appreciation that you need to shift from structured data return data in your methodology from a tax administration point of view to also incorporate unstructured data. We've for a long time only looked at return data as structured form of data and I think in the days of big data and machine learning we are learning in artificial intelligence we are learning that we also need to leverage big data rather unstructured data for the purposes of creating visibility of the tax ecosystem including the informal sector by the way. So I think for me it becomes important that we don't only take a structured approach to data but also start to incorporate unstructured data. I mean we in the South African Revenue Service we started to ask for answers what machine learning should we introduce that slips the social media to be able to pick up because that's not retained data. It's unstructured data that gives you insights around what's happening if you want to argue what people are posting and what are they saying in their social conversations. This is where you have it unstructured data and we start to say what does this data tells you tell you about the tax ecosystem and the fullness thereof. So I think I'm arguing for yes the structured data element but there's also an unstructured data that you can be able to create machine learning and unsupervised machine learning can assist in terms of making sure that we enrich the data that we have. Of course deal with the issues of reliability of that data and how you can augment it. We are of the view that's the combination of both structured data and unstructured data that will take us to where we need to get to in future. Thank you. Thank you very much. Interagency cooperation and data sharing can also go a long way in enriching the tax data. We've seen this when we on a limited scale obtained data from the local authorities because they have players in the informal sector that are not into the tax net. But of course the limitation of probably merging that data or making that data speak to each other is the lack of a unique identifier. But I believe if data is shared across and probably these gaps are addressed it could go a long way into making enriching the tax data. Thank you. To add in that we can also standardize the data in which we are collaborating with the local government to see what type of data do we need so that they can arrange according to your requirements. For the informal sector. So we can tell them that we need data in this format so the local government can create data in the format which is needed. So the standardization could help us to get data which is not a tax data also. Thank you. Very interesting indeed a lot of collaboration and this the use of unique identifiers so that we can easily do data matching. It's very important and I see that other countries I think Tanzania you do use one number to register bank to do and that helps in data matching. Thank you very much for all that a quick recap of what we've said so far. So first on the issue of why it is important to share tax data. And I think the sentiments are that once we've shared and we collaborate with others outside the skills sharing and you know improvement of the quality of the results that we get. But to share the tax data still remain sensitive so we should have a proper framework and a proper data governance. A proper data governance framework to help us to enable us to actually do this in a lawful manner. I think that's what we've said and on the issue of what has been some of the challenges in using the tax data I think what has been highlighted is one is the quality. We still need to keep working on how to improve the quality of the data especially because it's safely ported so we have to be mindful of that when we are using it. And the issue of confidentiality we should not forget whenever we're using this data and then capacity of the people using the data if it's academia maybe there's capacity there. But when it's the revenue authority or the minister of finance maybe there's lack of capacity and this is where maybe the collaboration comes in and becomes helpful. And then we have talked about the linkages with other institutions where we can obtain data and even data from the informal sector. And lastly how do we make our research using the data now into policy and making it impactful. I think the issues that have been discussed here is first we make sure that the results that we are producing are robust. We produce areas that everyone is like no that doesn't work like that. Next time you come do not listen to you say I know we can't use this research paper so as much as possible we make sure that you know the research we are doing is robust. And second I think you're saying collaborating and I want to agree with that. Sometimes you have some tax incentives and some of these incentives maybe are exemptions to the minister, the president, the members of parliament. And you think it's a bad policy but for you from the minister of finance to say let's remove this incentive it becomes difficult. But when you collaborate with ICTD and ICTD people say these are some of the incentives that are removing money. Maybe it becomes easier to communicate that to say it's not us but this research institution is saying this and maybe this is the direction we should be taking. I think that's one of the importance of collaboration. So thank you for this. Now I would like to open it up for questions from whatever we've discussed here or any other questions related to data and how we move from data to policy. So now it's open online and my colleagues here. Wonderful presentations. I'm Ano Chinfembe from the Zambia Revenue Authority. My question is probably something that Wasiwana covered I think in his closing remarks and Johnston covered it also I think on a policy review. It's more or less in terms of two-way traffic in terms of the way the question will come. So this has to do with tax expenditure. So how do revenue authorities conduct this in order to advise tax policy? And certain policies that come from probably the policy makers to give incentives. How do revenue administrations track the impact of those incentives that are given over a period? So I would also give another example of certain, for example let's say VAT. You have certain products that have been zero rated for quite some time. How do revenue authorities track these tax expenditures over time? And given that you have all the data that you have is sitting in one place, kindly maybe give a comment on that. Thank you. Thank you very much. My name is Esra Mazwanika. I am the research manager at African Tax Administration Forum. I am impressed by the discussion regarding data. Because as ATAF we also handle data for 37 tax administrations in Africa. And basically the data that we have been collecting through our member tax administrations is macro level data. But here we are talking of micro level data through the data labs that I hear the panelists talking about. And so my interest in my question to the panelists is we have been collecting on behalf of the members macro level data. And we also have a vision of actually collecting micro level data. One of the global, for example the global discussions on international taxation issues like the ones happening at the OECD level require us to have micro level data for us to voice our concern as Africa. We need to have tangible evidence to say this is what the data is telling us. This is the data that we have regarding the multinational companies in African countries. So my question to the panelists is how can we collaborate so that we can, because for example Tina said we can collect the data. We can go to Uganda and use the data lab. We cannot transfer that data because it's sensitive and we cannot actually have it but we can go and use it there. But I'm trying to see whether we can not have also a situation whereby we can have this consolidated data like we are doing for the macro level data. Because sometimes also working with panel level data it will be actually be more robust than working with one country level. So that is my first question to the panelists how we can get that data. Second day I wanted to say regarding what South Africa is talking about wealth taxation. I think it's etaf we have already circulated a call for expression of interest for our tax administrations who want to collaborate with us on taxation of INET with individuals. And we have already received like eight expressions of interest from member tax administrations. What I just wanted to say is like what the chair said it's actually better for such sensitive topics where you think maybe politicians are involved. If you work with collaborating partners you can be able to get actually the information that you want without having many challenges. So this is just a comment and encouragement to say let's do it collaborations are necessary. And etaf is an MOU with UNWIDER and so we work together. Last week we actually held the ATRN which is the African Tax Research Network. And I'm happy because Amina was representing the UNWIDER at the ATRN Congress. So collaborations are important and then Karis thank you. Thank you very much for your for the discussion I found it really interesting. I would first of all like to acknowledge the incredible amount of work that's gone into making this data available. All of you there on stage your colleagues in country colleagues in other countries. It's a lot of work and it's been incredibly beneficial and it's not obvious at all. Just 10 years ago this data wasn't available. Now there's so many studies that inform policies and it is to be celebrated. But my question is very similar to what was asked before. What do tax administration need to do to have information on the informal sector. Except I would ask a slightly different question which is what do tax administration need to do to get revenue and information on high net worth individuals. And I would like to particularly ask your views on based on the data that you help put together that you manage. What is the potential in that data to start taxing high net worth individuals more effectively. What are the limitations and maybe following up on what Zera just said. I understand this is a politically sensitive area and we as independent researchers can sometimes do things that are harder to do from within government. And within government institutions. So what do you think is the scope there for people like me as an independent researcher. But there are so many others. So this is not really about me. What can we do as independent researchers to support progressive reform. And especially taxing high net worth individuals more effectively. Thank you. Thank you very much for those questions. We can come for another round and we will check I think soon after this. If we have any questions online. So who wants to take any of the questions. Who wants to start. Maybe let me maybe let me start. I think I think we've been reporting on tax expenditures in South Africa for quite some time now in our tax statistics. As far as quantifying the impact of most of the exemptions whether it's in the automotive sector. Whether it's in the research and development. I think we're quite comfortable that we are able through the collaboration with different institutions to capture the data and report it. Is it complete. I'm not yet convinced that it is as complete as it should. I think if you look at our 14th edition that we published in December last year you'll be able to see that we go to an extent in terms of quantifying this. I think again there is work that has been done out of either the Treasury or Department of Trade and Industry to start to quantify even better the tax incentives or any incentive scheme that is in place. So I think more and more the quantification of these expenditures and therefore making decisions on which ones must be retired and which ones must be sustained is work that is ongoing. But I'm encouraged that the visibility is there. It's available at least in the South African context. And I think it's work that must be done because we spoke earlier on around the bigger implications, macro implications of some of these incentive schemes that are out there. In South Africa as far as taxation of high wealth individuals we decided to set up because I think we believe in the segmentate approach into the tax base. We think that once you segment your tax base you're able to have focus. We segmented that, created a unit that is responsible for the taxation of high wealth individuals and following the global practices that we have seen elsewhere. But I think the leveraging of data such as the CRS. There's a paper earlier on that was presented where we saw that there is around in our case using 2018 data, 420 billion rents of offshore holdings by South Africans. And we are starting to unpack that. It's a lengthy process but we are confident that we are going the right direction in terms of that work. Also leveraging the number of leaks that have come through in terms of the original one, Panama Papers. So we think the focus in segmentation as well as exchange of information in particular CRS to be able to see what offshore holdings our South African citizens have is starting to give us sharper focus in terms of the high wealth individuals that are in our jurisdictions. And we're starting to see some gains in terms of revenues. We found about 6.6 billion rents worth of undisclosed assets in a sample that we've looked at which has generated around 207 million in Texas. So we continue to find momentum as we hit these sweet spots in terms of revenue gains. And I think we continue to want to make sure that the construction of the creation of the high wealth individuals unit does get justified by the return to the fiskus in terms of revenues and tax-based broadening which is the pursuit that we'd like to have. Thank you. Thank you. I want to add on that the tax incentive group provided or granted to some business of institutions or other external institutions as an exemption. So that one is being taken care. For example, in TRA we have a section which is dealing with tax exemption. Why do we have that? Because we have to inform the policy makers that look here you are overdoing on exemption which you are granting to some people. So in East Africa we have agreed that the tax exemption should not exceed one percent of GDP. So we have to make sure that we inform the government that where are we in such exemption which we are granting to some business people. So we do take on an account of that. Joseph, sorry just to ask, do you make this information public? Is it like the minister announced it in parliament or it's something that we just make available to the policy makers? Is it public information, the tax expenditure? It's not the case of the tax administration to announce to the public. If you send to the minister for finance he is the one to announce that what exemption the country has given to issued. But some parliaments they do ask about that. So if he's asked then it is if he's requested it's given but it's not a public issue. Let's say the tax administration send the information to the minister of finance that the exemptions which has been granted has been to this magnitude. Yes. Similar to SARS and TRA, tax expenditures and incentives are tracked and monitored in URA by a dedicated team. We also have quite a number of studies that are ongoing in this area because it's an area that we are trying to rationalize as a country. So we believe it's an area for continuous improvement. I just want to confirm South Africa, are you making this a tax expenditure information public available? At aggregate level we do make it available. Maybe if there was a question about how do we collaboratively make information available. I think it's going to be a challenge in the short term but I think it's something that we should continue to work on in terms of how do we make sure that this micro data as a tax administration does get available. I mean we are working in the era of blockchain and cloud. I think it should be important at the heads of states AU level to decide what is the African cloud strategy for instance so that we can be able to start sharing information. I mean if we are going continental free trade arrangements it becomes very important that we start to also share information. I think in the medium to long term it's something that will be necessary for us as a collective to work on. But I don't think in the short term it's something that I foresee happening but certainly medium to long term it's something that's very crucial for us in the continent. And maybe even other developing countries to start to look at from a block point of view. Thank you. Thank you. I just wanted to add that I think Johnston is right in terms of a long term view it is a more long term process to make data available. But I think one of the things that we can start exploring in the short term while we don't have those structures set up is better enabling virtual access to some of the data labs. I think one of the biggest constraints currently even if the data sets are not merged if you're able to have access to the various databases from where you are we should try and leverage that a bit more in the short term. Just so that you eliminate the costs of travelling between things there's a lot of new technology developments now where you can use facial recognition to ensure that the people logging on to systems are the correct people who are supposed to have access to the data. So I think starting to formalize those processes could go a long way in being able to do comparative studies. The question that was asked sorry you go ahead Tina. I think I'm taking it on from Ayanda on trying to make the data available currently you are and you and your IDA are conducting a feasibility on the same on how we can make the data lab accessible virtually. Of course we have to look at what would be the legal framework what what kind of infrastructure would be would enable us to do this that data security concerns and all that. So yes it's an area that we are exploring so that we can make it available for researchers. Thank you. How about the question on the high net with individuals what is the scope of you know revenue authorities even the ministers of finance open to this idea to say we can collaborate on how based can we tax high net with individuals with independent researchers or research institutions. What are your views. There is a specific area that you mentioned that high net with individuals. It's an area that Africa is also looking at and it's an area with potential revenues. What are the thoughts of your thoughts on collaborating with other institutions to see how best to tax this area. Okay so I think I guess Johnston sort of coveted for South Africa but I'll add even though it's not my area. I think extensive research is always something that we're looking at maybe not even just in terms of how you tax them but also I think there's been a lot of interest in South Africa in conjunction. You know I think a lot of people have raised okay but you're looking at a basic income grant. Let's say what about a wealth tax if you're going to focus on this why not the other. And so I think it's ongoing discussion and we're continuously looking into research into those areas trying to think what would it actually look like in South Africa. Is it something that makes sense for us for something that is that much of a policy shift. If you're going to make any any proclamation or decision you need to be you need it to be evidence based. You need to think through what the potential implications are exactly how you would capture what systems you need to have in place to be able to see. Okay so some assets people will declare but some assets you'll have to find ways in which to make people declare those. So I think it's a it's a very long process where we're open to collaboration. The tax policy team within National Treasury is continuously collaborating with various research institutions together with the revenue receiver in this area. Thank you very much. Joseph I want to hear from Tanzania. I'm not too sure I haven't much collaboration with the outside world with other institutions on research. What are your views is the Tanzania open to have other institutions work with you do on some taxing areas? Yeah in Tanzania so we work with Unwider on the issue of data. Though I haven't mentioned but we have a collaboration with Unwider. Also in our university institution there back home we work together in terms of data and the other institution is national statistic. So we collaborate together also in making that. So you know it is there everywhere. I think we have time for one or two more questions. Do you have any? Yes thank you to all panelists. This has been a very insightful. I'm Yuka Birtele University of Helsinki and Union wider. I just wanted to continue along the lines of institutional collaborations within the countries. And how do you see in particular the role of the statistical institutes or the statistical services in making data available for researchers going forward. Because I understand that the tax revenue authorities are of course the custodians of the tax administrative data. But in a sense I mean it would be logical going forward that there would be a role for the statistical offices as well. So maybe if you have comments on this whether your countries rules and regulations allow you to share data with the statistical offices. Thank you. Thank you very much. Did you also want to ask a question? Thanks. I mean I'm from Union wider. So you can still my question but I tried to change it a little bit. We've spoken a little bit about international collaborations. What is the scope for collaborations with universities and academics and research institutions within your country? I mean it's really nice that Union wider can come and work with the institutions but I guess as a South African. I mean I'm sure I've heard of others researchers or universities complain and say actually we never get access. Is there room and is there space given or priority given for local academics? Thank you. I think behind this you also want to ask a question. I was almost letting go. Mine is not a question, it's a comment on the high net worth individual. Sometimes we talk about high net worth individual but who are these persons? In Uganda what we've done is first develop a criteria to suit what you call a high net worth individual. Then following that criteria run through using data analytics and identify who these persons are. Then create a specialized office to handle this, to focus on them. We're also in plans of exchanging the automatic exchange of information. So we'll add in the exchanged information for the beefing up criteria and then having within our large taxpayer a section that really focuses on this. That is our approach on high net worth individual. Thank you. Thank you very much for that insight. There are two questions really. A scope of collaboration, there are of the national statistics office. Do you see an error in this? Should we start pushing the tax and decision data to the national statistics office? What are your views to making this tax data more available to people? And the scope of collaboration with local universities? Yes, we are collaborating with other institutions maybe outside our country. But locally are we collaborating? So maybe I'll jump in first. In South Africa, I guess we've started a new project which is done together with the statistical department. With SARS as well and national treasury and another program that we have that's called the city support program. Through spatial data developments and trying to geocode our tax data. That's an ongoing project as well. We've reached out to stats SA and they're developing what we're calling an integrated business register. Which will put together the different surveys and censuses that the stats department runs together with the tax data. And a lot of the other data sets that I was alluding to saying, we'd like to incorporate more data from the unemployment insurance fund, credit registry data. Just so that you have a comprehensive integrated register. Which probably would be housed at stats SA. But because we already have data lab facilities, we would like to still maintain having a version of the data that we can also retain. For researchers to use. So we definitely do have collaborations and discussions that are ongoing in that space. In terms of, for example, collaboration with academic researchers, I think the collaborations that I was mentioning included at academic researchers as well. So for example, a lot of us within the national treasury, if I'd say, would usually partner with an academic, a local academic or an international academic. But I think for some of the other research papers, we noticed that there's a gap. Because when you collaborate in that manner, you tend to choose from the same source of people that are in your network. So to try and broaden the scope and to also include previously disadvantaged universities that are not like your top four universities, let's say, in South Africa, we're trying to experiment with having calls that just look for co-authors so that prospective PhD students can collaborate with some of the international researchers that we have through the SA type program. And that way we also have collaboration with different types of researchers and academics. We also have other projects that we collaborate on sometimes with research institutions. For example, some universities we'll have a contract with them in exploring developing the procurement data, for example, which is not necessarily a partnership that's done purely for research, but we do have partner with them to help us set up other databases as well. Yeah, thanks. Thank you. One minute. Yeah. So I think what's important to note is that in our context, various pieces of legislation give certain responsibility to different ministers in terms of certain data sets. So trade data, for instance, trade statistics is in the purview of the Revenue Administration Commissioner by law. And other data would belong somewhere else, but there is a provision in the National Statistics Office, as you call it, to call for any data for statistical purposes. So that then allows the Statistics Office to come in and make certain requests to be provided either of the Central Bank of the Revenue Authority, but again, the volumes of data would just be too huge to be housed there. But I think the portal that has just been launched on special data is an attempt to come in, to get into this integrated register that pulls everything together. Thank you. Thank you very much. For the case of Uganda, we have an MOU with the Statistics Office and we routinely share data with them at aggregate level, at micro level. But the micro level data is for more of creating or developing the macro indicator. So we haven't got to this stage of making the data available for them so that they can share it with other parties. But it would be interesting for us to explore this option. Then regarding collaboration with the universities and the research institutions, we have to a limited extent, but it's something that we're also working on Initially, we've been having the CITPANO data, but we're building on more data sets. So with this, we plan to create more awareness and collaborations with them. Thank you. Thank you very much for all that. And I think the short answer is the role of the National Statistics Office. They still have error, but a little bit. Not much. I think the data that you find with them for the time being, it would be at the aggregate. But if we still want to use this micro data, taxpayer and video taxpayer data that maybe we can track how they've been progressing, things like that, we still have to go through their MOU route and have that. I think that's quite the short answer that I'm getting. That, yes, at the aggregate level, the National Statistics Office will have some data, but at the micro level data, we still have to go through the revenue authority. Collaboration with the local authorities. South Africa is usually the odd one out in Africa, because usually they would say, yes, we have that. But for the other countries, we're quite limited on how we are collaborating with the government. We're quite limited on how we are collaborating with the local universities and things like that. It's quite limited. And I'm speaking about my own experience, about Malawi. So you might be doing it in Tanzania. But in Malawi, we haven't collaborated to do research with the university or things like that. That is not really happening. And I think for most countries, maybe it's speaking now. We've collaborated with outside institutions before they would come say, let's work together on this one. So I think it's an important issue to raise, because sometimes we have the capacity right in the country that we're not utilizing. So thank you very much for raising that. And I think at that point is where we ending the session. Allow me to thank sincerely Johnson, Tina, Josepha, and Dayanda for sharing your experiences in using tax data to make impactful tax policies. So thank you very much.