 Hello everyone, welcome back from the break. It's that time of day. Time for the final session in what I'm sure has been a long day with lots of information and ideas being exchanged. And of course this session is titled, The Policy Impact of Revenue Authority Research in Sub-Saharan Africa. I'm Denise Wall. If you were here yesterday, you do know that. I think everyone in this room agrees that promoting fair and equitable tax systems is crucial for stable growth. That's because, as we all know, leakages drain much-needed resources and have the knock-on effect of weakening long-term sustainable development. And that's why improving the capacity for tax and other domestic revenue collection is the key target of SDG 17, which has to do with multi-stakeholder partnerships. In this afternoon's discussion, we want to focus the spotlight on the global south where many countries struggle with efficient tax revenue collection. We have a panel of experts who, I daresay, believe that research collaborations can help revenue authorities build stronger tax systems in Sub-Saharan Africa to help promote more equitable and sustainable development in the region. Before we engage our panelists in the discussion at hand though, I'd like to invite Professor Yukka Pirthila to background the session by providing an insight into the kinds of initiatives taking place at the institution. That's UNU-Wider. And contributing to revenue authority collaboration in Sub-Saharan Africa. Please Yukka, and I will sort of give you a fanfare by talking about you for a bit. Yukka Pirthila is a professor of public economics at the University of Helsinki, and he also holds the position of non-resident senior research fellow at UNU-Wider. He's previously worked for the University of Tampere, the Labour Institute for Economic Research, that's in Helsinki, and the Bank of Finland. He conducts research on topics related to taxation and social protection in developing countries, and his research is widely published in journals such as the Journal of Public Economics, the Journal of Development Economics, the Economic Journal, and the European Economic Review. Yukka, the floor is yours. Thank you, Dennis, for the kind introduction, and welcome to the session also from my behalf. So we thought that I would give some background to the session to sort of frame the discussion. So, yes, I work together with the UNU-Wider colleagues and several colleagues from African Revenue Authorities in utilizing tax data for research. Okay, so let me start off with providing you with some of the goals of our work. So, of course, tax authorities need data, and they need data to actually administer taxation, so levy taxis on the firms and individuals. And this data is by design at the taxpayer level, so that's micro data. The information as a by-product is also useful for research purposes looking at various issues in tax research, and also perhaps more widely as well for understanding the economy. And this research has mainly two objectives. The first one is to understand taxpayer behavior. So taxation, of course, is needed to gather revenues, but it also influences the economic agencies in many ways, so it can lead to employment changes, it can lead to changes in investment levels, and understanding these responses is crucial also from the revenue authority perspective. But of course, and this is of course the main goal for the revenue authorities, is to seek ways to improve tax compliance and reduce tax evasion, and big data and combining various data sources can also be helpful there. So I would in a sense summarize this slide saying that this is a combination of big micro level administrative data and then the appropriate research designs that hopefully then provide credible research findings. So the style of the wider work together with the revenue authorities is really to co-create research. Research topics that we have been working on, the topics have been chosen jointly with the revenue authorities. For example, in the Zambian case, we learned that the Zambia Revenue Authority wanted to study, wanted to have new estimates of the tax cap and then we tried to estimate them together with colleagues from ZRA. In order to do so, we need to have access to the data, and the data that researchers use don't need to have the actual identifiers. So it's anonymized what the researchers get access to, and then it's shared with the researchers in a secure manner. So we feel that there are synergies in this sort of collaboration style, because it could be the case that the international researchers from the wider network can bring in some new fancy methodologies. But without the deep institutional knowledge that the local researchers or policymakers have, the research findings wouldn't have relevance and they could even be perhaps misleading. So we think that working together here is key. We have co-authored papers together with revenue authority and researchers and analysts and local academics as well. On top of that co-authoring, there have also been various sorts of actual capacity building activities. So this is a map of Africa and it highlights the countries with whom you knew wider is conducting research using tax administrative data, and the set of countries includes Rwanda, South Africa, Tanzania, Uganda and Zambia. And then there are of course various other research projects going on with several other African countries, but these are the tax administrative research collaborating partners. Let me talk a little bit more about two highlights of the joint research or capacity development work in the area of tax. The first is our work together with the Uganda Revenue Authority in Building a Research Lab. And the second is a capacity development initiative that you and you have run in 2022 together with Stellenbosch University and the Southern SA Tide Research Program in South Africa. Okay, so the data lab in Uganda. UU Wider has collaborated with Uganda Revenue Authority for more than five years. And during this collaboration it became clear that it would be in the interest of URA to open up data access more broadly. And that was why then the two institutions decided to cooperate and a data set was prepared. It was then made available in a secure facility in Kampala. There was a launch event in 2022 and that was also then supported by a request for research proposals. So UU Wider gave some support for research teams utilizing the data set also because the research team members had to travel to Kampala to this physical lab to work on the data. And in fact some of the papers selected from that call have now been presented in this conference and some are available as wider working papers. And so we are very grateful to the Uganda Revenue Authority for working together with us to make this happen. And we are very keen on continuing to develop further the lab and then do similar sort of activities going forward. Oops, now I need to get back. How do I do that? Thank you. So I just wanted to show this. So UU Wider has started this sort of making data, admin data available for the broader public or researchers in South Africa. So the Uganda one is the second on the continent in our understanding. And there's also a short video explaining what the lab is about. All right, so the Winter School was the second one I wanted to talk about. So this was an initiative that the institution did together with Stellenbosch University and the Southern Africa Towards Inclusive Economic Development Program, as they tied for short. So there was an open call. We had more than 300 applicants. Only 10% or less were selected, so that was very competitive. We had, of course, academic, early career researchers, but also representatives from Revenue Authorities, Finance Ministries and Research Institutes. The course consisted of an intensive data and statistical methods bootcamp that was online and then one week of lectures and practical sessions in South Africa in July 2022. So we were happy to have some of the course participants participating also in this event and some of the people who participated in the course actually presented posters over the posters here in this event as well. When it comes to the type of research topics that we have been doing, we have done work of course on compliance, tax compliance. So at this list here, some of the papers, these are by no means the only papers and now I notice that there should be also South Africa papers here. But these are just examples of what can be done. So there's a paper examining the revenue impacts of tax examinations in Tanzania. Another one on looking at an administrative intervention called the withholding value-added tax in the Zambia case and its impacts on revenues from the VAT. As you know, informality is a key bottleneck in taxation in African countries and the African revenue authorities have had various programs to try to get more firms and individuals to the tax net. Maria Jouste and colleagues have analyzed the success of the Yucandan taxpayer registration campaign in her paper that I listed here. And of course in the beginning I mentioned how tax influences taxpayer behavior is also key. Here's one example of such studies where we together again with Maria we look at the responses to a tax reform where the progressivity in the personal income tax system was increased and how that then influenced revenues and inequality, post tax inequality in the country. So that was my hopefully short enough introduction and I'm now looking forward to hearing from the panelists on their thoughts on these issues. Thank you. Thank you Yuka for those introductory remarks and actually for opening up a bit for those of us who may not have known the details a bit more about how UNU Wider has been working with tax authorities in terms of research in the sub-Saharan Africa region and perhaps on a broader scale as well. It's now my role to introduce you to our panelists and I will introduce you one by one and when I do please make your way up to the stage. But perhaps I should start with our one remote participant and that is Ezekiel Piri. He is the director of research and corporate strategy with the Zambia Revenue Authority. He's a tax administration and policy expert with over 50 years experience. He's a short-term expert for the International Monetary Fund in tax administration and he's previously worked in academia and lectured in undergraduate courses. Welcome Ezekiel. Okay, let us hope that Ezekiel has heard our welcome. Hi Ezekiel, you're there. Thank you very much. Good afternoon. Wonderful. Great to have you. I'm delighted to see that the communication link has been properly established. That's one down. Next I'm going to call on Allen Nassanga who is the assistant commissioner for research and innovations at the Uganda Revenue Authority. She's worked on major ICT projects such as the integrated tax administration system which led to the introduction of e-filing and online payments and the data warehouse business intelligence project that aimed at integrating data and enhancing data analytics and usage in URA. Allen has a strong personal vision of using research as a driver of innovations and transformations in tax administration. Please join us on stage Allen. I guess those of you who were present in the previous parallel session that took place in this room, you're no stranger to Waziona Ligomeka. Wazi as he calls himself. He's a tax economist and a member of the United Nations Committee of Tax Experts in International Cooperation. He's got more than 15 years experience in tax policy, tax administration and macroeconomic analysis. And he's also focused on modernization of tax administration functions in developing countries. Please join us on stage Wazi. Last but certainly not least again, we have Ingrid Woolard. She's a professor of economics at Stellenbosch University and honorary professor of economics at the University of Cape Town in South Africa. She served as a member of the Davis Tax Committee in South Africa which advised consecutive ministers of finance on tax reform for inclusive growth. Ingrid is strongly committed to providing research-led policy advice. It's great to have you all. I'm not sure why we have an extra chair, but that's okay. Let's imagine that Ezekiel is there in spirit. Well, let's just get warmed up with an opening question that I hope each of you can address. And again, let's start with Ezekiel, which would be great to see on one of the screens here if our technical team can figure that out. Ezekiel, can you hear us? Yes, I can hear you. Okay, okay. I can hear you. Great. So here's a question for you. Great, we can see you now as well. I'd like for you to give me your impression on the whole question of tax authority research collaboration as it stands in your region now. Well, it's quite a loaded question. Thank you very much for that question, but quite loaded. And I hope I can do a bit of justice in terms of explaining the perspective that you've just asked me to do. So I think the groundwork has already been laid down by the professor in terms of the collaborative work that UNEWIDER is doing with revenue administrations. Indeed, I can confirm that there is some work that is going on currently between ourselves and UNEWIDER and several other guys for that matter that we are collaborating with in terms of trying to utilize tax administration data for research purposes. There are just a few constraints that we have, which I believe that across many countries where this is happening, this could be a bit common. The first thing that I would like to make reference to is the first of all in our setup, we are privileged to have a dedicated research department with a dedicated unit responsible for statistics, which was set up sometime back in maybe 2018-2019, but the research department has been existing for quite a while. But in terms of the doing the research that should benefit both internal and external users, it's been quite a difficult journey because there are various constraints that we tend to face in this space. The first one that I might mention first of all is the one to do with issues of taxpayer confidentiality. In our law, for example, the law that establishes the revenue authority, there are some very stringent provisions that deter the sharing of data with unauthorized persons. And that is embedded in the law, for example, to call the Zombie Revenue Authority Act. It's very expressed in terms of who can benefit or who can get the data to use either for academic purposes or for other purposes. But what is clear is that the taxpayer information is expected to be kept very confidential. And so that has tended to restrict the sharing of the data across the different uses from academia or the way up to a policy-making level. Recently, there have been other laws that have been introduced. For example, the Data Protection Act, which was enacted way back in 2018 in our case. Again, it has got very stringent requirements and it protects micro data from being shared with unauthorized persons and the penalties are quite stiff. So when it comes to that, you'll notice that over time we've had challenges with the aspect of, first of all, beginning to share the data for wider use. But this has been overcome lately by the signing of the MOUs and also putting into place of instruments such as the non-disclosure agreements. When you do this memorandum of understanding, you tend to put in place the non-disclosure agreements that tend to make sure that you are able to share this data without worrying so much of the breach that you can cause to the data protection laws that are in place. So that's been really the first and foremost the most difficult aspect in terms of sharing. But also the aspect of having very small research departments or research units has also contributed because we've tended to focus mostly on internally driven research that is just maybe meant to benefit management and also maybe policy makers to a limited extent. But we've not been able to really share or utilize the data that we have to a very large extent in terms of now making use of it in the manner that we are beginning to make use of it, especially like in the collaboration that we're having with institutions like UNEWIDE and thanks to the recent collaboration that we've had and we've done like the professor indicated in the introduction, we've done quite a number of studies together and we have a number of them planned for implementation this year and getting into the other year. But what immediate also comes to mind? May I? Yes, please. Yes, thank you so much. I think that you've laid out very clearly what are the primary challenges that you face in terms of the research collaboration that we're talking about here today. For example, access to taxpayer data, which of course is subject to very stringent confidentiality rules and the ability to effectively share data. And Yuka actually did talk about the fact that in the research protocols, you tend to use anonymized data. So I'm quite sure that there are solutions possibly around the difficulties that you face. But now I would like to turn to Ingrid and Wazi and Alan in turn to tell me about some of the challenges that you face when it comes to engaging in the research and the sharing of data, for example, that Ezekiel referred to and effective collaboration. You know, it's interesting to think back to where we started this process in South Africa. So it's been almost 10 years since we started discussions with the South African Revenue Services around why making data available to researchers would be useful. And I happen to be in some of those conversations because we'd set up a secure data center at the University of Cape Town for our household survey that I had been the PI for. And I still remember that very first discussion with SARS where the question was, well, why would one want to use tax data for something that isn't a question just about tax compliance? And I think where we are today, 10 years later, we've worked through all of these issues around how do we anonymize the data properly? How do we ensure that there's proper capacity? We're now in a completely different space. So perhaps to talk a little bit about where we are today. So today there is a secure data center which sits at the National Treasury with collaboration from the Revenue Services, a lot of support from UNU wider. They're data scientists. There's research staff on the ground. There's this amazing facility that researchers can go into and get fully supported. But the point I'm trying to make is that that didn't just happen, right? I mean, it sounds easy. But it was 10 years of gradual work of working up to this point where we are today. But we're now in a position where the facility exists. Everybody can see the value. We're not still having conversations around, you know, why would one want to make this data available to non-revenue authority staff? But it required changes in the legislation. It required an enormous amount of effort to think about ways of anonymizing, for example, top incomes to ensure confidentiality. So there was this huge amount of effort that required political will. And I have to say at some points along the process, we went through, in that 10-year period, we've had eight ministers of finance, I think. Some of them certainly would have supported this program. Not all of them would have supported the program. And so it really required a monumental amount of effort from the research community, from the revenue authorities, and from the Ministry of Finance to actually realize this vision. But today, you know, we have this facility. We have many papers not only being written on taxation, but papers that are of enormous value to policymakers and the research community. And yeah, I think we can come back to some of those issues, I'm sure, Denise. Absolutely, we definitely will. Nancy, what's your view on the ground? Thank you very much. Yeah, I think I share the same sentiments in terms of how do we start sharing this data to collaborate with others outside the tax administration. And I have to say, for Malawi, we haven't quite reached that stage yet. You know, we are still trying to look at our legulations, our laws, to see will we be able to start sharing the micro data at the individual level with others. So that has been a challenge in collaborating with others. But I want to add other aspects that we've made in using the administrative data. And the others have been already highlighted in terms of the capacity of the people sitting in the research department. We are generating all this data, but do we have the capacity to analyze it? What is this data saying? So it becomes very important when you are collaborating with the research institution or the academia in terms of understanding what the data is saying. The other challenge that we are facing is the linkages between the data in the tax administration. So you would have domestic data that is not linked with the customs data. And sometimes you would want to see if, for example, someone has imported a shop, 10 TVs. And then when they are reporting, they have reported domestically, they have sold two TVs, but they don't have any TV remaining in the shop. Where did the ATVs go? But we really don't have these linkages even within the tax administration of the data on what is happening. So these are some of the limitations. These linkages would also go even to other institutions. There is an institution that is registering businesses and you are collecting tax data, but you are not linked. So you don't know who is doing business because some are in the formal sector. Yet they have a business registration. We are not quite linked and we don't know how to find them. Lastly, I think what I want to highlight is the issue of the tax data, the administrative data, is data that we will find because someone is registered with the revenue authority. They are filing a tax retain, they are making tax payments. So we are generating this data. But for developing countries like my country, you have a whole load of individuals who have a lot of income, but maybe they are not registered with the tax administration. So who will be missing out on this aspect? So the data is there, it's very useful, but sometimes you would want to get more, but you don't have it. So it's still limited to some extent because it's just covering those that are formal and registered. I think for this I should stop there. Thank you. Thank you so much, Wazi. We actually were discussing, we had a mini panel before the panel where we discussed the fact that a lot of the issues that we might be getting into during this particular conversation would already have been covered during the previous parallel session, but I think it's bear repeating some of these issues and perhaps opening them up from different perspectives. Alan, tell us about your challenges. Wazi talked about fragmentation of data, also perhaps fragmentation in terms of the way institutions work. Is that something that you're seeing in terms of your work? Thank you so much. Good afternoon, everybody. Yeah, before going into the challenges, talking about how you are, we started on the journey of collaborating. For us it was more of a mutual interest and mutual benefit. As you are, we are always challenged to do more, to collect more, raise the tax to GDP and inform policy. We are always asked questions by our supervisors. And on the other hand, we always had these people, different agencies, different researchers coming to request for our data. So for us it was a natural kind of realization that we needed each other. We needed the collaborations. And the way we collaborate with individuals as well as institutions, and all we do this is we normally work through a memorandum of understanding and we jointly agree on what we want to do. And that's what we're doing. We're doing research. We're doing research agenda. So once we set the research agenda, which is mutually beneficial to the authority as well as in the interest of the party, then we can set it moving. And then with studying, of course, there have been a couple of challenges. We've done many joint researches, but yes, then you realize specifically the issue of data confidentiality and privacy. We've done this without disclosing the individuals. And of course, this, we went through nandisclosure agreements, MOUs, but still at the beginning, most of the data would be sharing, would be at aggregate level, so that the individual is not, you can't easily tell, the individual taxpayer. But then as we kept going forward, but maybe before that, when you think about data, it's an output from a process. Revenue authorities generate this data as they are just managing their normal process. So there are moments when you change the process. They are, for example, you have a new process, a new way for efficiency, and that would impact the data. So there would be challenges of data quality or at least the collaborating agencies would think that now there are these fields missing, but the actual point is that the process has changed. And so to walk around that, you needed to work with a local researcher, a person within the authority who appreciated the process to let you know that now there's a new process and from this new process, this field is no longer a requirement, and that goes on from time to time. Yes, out of that, the need of working together and making sure that we have this resource of data available without disclosing the individual, that's how then the concept of the research lab came to birth. So it was, we both were benefitting because in each collaboration, you first, we have guidelines, and you have to share your mandate to share your findings with us before you put it out. So that way we could be able to undertake the IT research. You would complement our resources and I must say that within our research division section, we have very few people, yet we have a lot of work to do. So it was a good opportunity. Then you work with the different research networks and you just agree on what agendas, what to research about, and you help them make sure that they appreciate the data from the process perspective, the taxpayer details are not revealed. So it's a win-win situation. We quickly get many researches completed and many tight knowledge, some information really coming out very fast with our limited resources. So that has been our journey. That's how we came to realize that the way to go is really collaborating and we collaborate with individuals as well as different agencies. Thank you. Thank you so much, Alan. I think that's got a really positive feel to it that there are these challenges but in the end there can be a win-win or win-win-win depending on the number of partners that you're working with. But one thing that I'm hearing coming out of this conversation is one of the issues, if not challenges facing individual authorities and bodies is one of capacity. So I think it was Wazi who talked about a shortage of analysts people to really interpret data. And Ezekiel, is this something if I turn to you that is affecting you in terms of the work that you're doing in your country? Yes, please, moderator. Thank you very much. This is really one of the issues that is affecting us. First of all, the number of people that need to work on these complex assignments is one of them on its own but also just the experience and the capacity of the officers that we have. So we require constant training of officers but even in the presence of training that the institution offers not much the capacities that have been built by specialised institutions that we've been collaborating with lately. So I think the way we've tried to overcome this challenge both in terms of numbers and in terms of the experience in terms of what you can do with the data we've really tried to maximise working with collaborating partners so that the experienced researchers that come from these institutions that have dedicated resources to deal with this will then help our researchers to begin to learn some of the techniques that they may not exactly be having. So indeed, this has been one of the limitations but we've overcome it lately by collaboration with institutions like UNewider and the others that have come on board to try to get to use some of the very rich data that we have except that some of this data in terms of the periods that we cover it may also be a bit shorter than what you would normally want to have because we've been changing systems from one period to the other and so in terms of consistency of quality we end up having a lot more challenges but I think we've overcome that by the kind of collaboration that we've been having. Wonderful, thanks for that Ezekiel. If I come to you Ingrid, I feel like I'm just going... So please feel free it's the thing on my left to the left, to the left please feel free to interject and add your contributions to your colleagues' comments. One question I do have in mind when we're talking about tax administration or tax authority research collaboration in sub-Saharan Africa is that I would like for you to share with our audience concrete examples of that kind of collaboration that has resulted in tax policy, either new tax policy or changes in tax policy. Is that one for me? Yes, for example. It's a tough one, you know tax policy changes very, very slowly, right? Even if one has a robust finding you know to then and even if there's political will to change something, it takes time. So even after 10 years of this type of work I'm not sure that there are so many examples but let me give a couple of small ones. So one is that in 2014 South Africa introduced an employment tax incentive for young people and that work had been based on a randomized controlled trial which has subsequently been quite contested and I don't want to criticize the author of that paper but it wasn't a real world example of the rollout of the actual program that happened. So this, you know, an employment tax incentive for young people in a country with massive youth unemployment is obviously an attractive thing to do. There have now been quite a number of papers using the tax data sitting in the Secure Data Center which are contesting the original findings and saying well actually it's not obvious that this tax incentive is doing what it's meant to be doing. The econometric difficulties are certainly there and I think what's been exciting is that we now have several papers using quite different methods and quite sophisticated which are all questioning whether there's a real result. Now that hasn't yet resulted in the employment tax incentive being binned. I mean I think that's is an incredibly difficult thing to do but I think it's certainly adding to a stock of evidence that might not have been possible in the absence of the Secure Data Center. So perhaps that's an example of something where I think it's going to eventually lead to some adjustments but it hasn't happened yet. Right. Please go ahead Alan. Thank you, I could come in there. It may not be, to change tax policy it could be more than one paper to do that but there are some benefits from an administrative perspective from our research findings there are those things that as a revenue authority we take on and we implement and they result into benefits really, additional revenue. The most outstanding one is that the issue of the research lab out of realizing that we have many data quality issues and we could each time you carry out a research you can do some data cleaning you can then make it available for the different researchers. There's another research we did on the high net worth individual just to see how well are we taxing them and that the recommendations from that were really implemented so we ended up having a criteria thinking through a criteria for defining who is a high net worth individual and of course reviewing it from time to time and then setting up a high net worth individual section so they may not policy changes sometimes take time they are those obvious low hanging fruits which really come from the research recommendations and findings. Sometimes for policy you need more than one papers and there are moments when we just pilot or run a proof of concepts of some of the recommendations and based on the findings from there then bigger government and policy makers is able to move forward and implement change. So yeah there have been so many benefits but it's difficult to say this paper then this policy change that is known to are very possible. Thank you. Please go ahead Wazi and I do understand that it's difficult to draw a straight line between one bit of research and yes this new law this new policy came into effect please go ahead Wazi. Thank you. As I said the collaboration that we have in Malawi it's not mostly with the you know someone outside but we have a code collaboration between the tax administration and the minister of finance and I find this to be very useful I'll give you an example sometimes the minister would say can I increase the their bracket a for pay as you aim by this amount and maybe the level you lost is say 100 million then the revenue authority would say that's too much will not be able to meet this target you're giving us then together we say okay how can we generate the same revenue from somewhere else then we say well maybe let's introduce VAT on some financial transactions that are IT in nature they are not really financial in nature so we have been working together the research department within the tax administration and the policy division within the minister of finance we do collaborate and work together especially on analyzing policy options and I found this to be very beneficial because we understand where they're coming from and what we are trying to do so it makes it easy to present it to the minister so that's a bit of a collaboration that I find useful but hopefully we'll get where we are collaborating with the outside that's great I love that example because it sounds like you're working backwards Alan yes some of the collaborations are really again across the different revenue authorities we have organizations like the ATAF we always collaborate so what the collaborations have done is avail this wealth of information knowledge which you can always reference because there are occasions when you have a challenge and all you do is to research within the different networks is any research that has been done it could have been done in a different revenue agency or within a different collaborating body and you pick this information and you reuse it here and there and it's very efficient yes of course some collaborations are directly with the minister with intergovernment agencies but as well as the different revenue authorities sorry please go ahead Ingrid I wanted to jump in on this example about the simulation models because I think it's such a great example of where one can use data that comes out of this confidential data that sits in the research lab or at the revenue authority but you can then quite easily translate that into an Excel based model that you can then distribute very broadly so there are these wonderful examples now across sub-Saharan Africa and there was a workshop the day before this conference about how to build those models and that's a beautiful example of where you're not requiring people to come in and use the confidential data but you build something based on that that's quite easy to update and then you disseminate it broadly and so we've certainly seen wonderful dissemination Nuki Stain is here in the audience where we've been able to do training sessions across other government departments such that they begin to understand much more about how the tax base is constructed and how you can simulate these tax changes so I think that these great examples across the region of things that you can do that don't necessarily require this big scary data center necessarily It sounds like there's already a great deal of very positive collaboration among the different kinds of authorities and ministries of finance and so on but do you feel that there might be more room or is there room for improvement when it comes to the uptake of research in terms of policy making the floor is open please go yeah for the others that have already started you know these collaborations I think they are really open but for for countries like mine I think there's a lot of room one is this whole aspect of creating dedicated units that will be conducting this research managing the data, cleaning it making sure it's in a well usable format the data lab in other countries that they are creating I think that would produce data or put the data in a format that even can be used by others and develop the models that can be shared with others so definitely there's that room for increased development of things like that capacity building it's always a good area where we can improve on the use of data for countries and I love this collaboration why is the you in for countries that did not see I was like why when are you coming so that's the kind of collaboration that we need as well and it's really helpful I like that idea working with institutions even on the how do we start using this sensitive data the institutions like this can help us this is how other countries are doing it we've heard it's a long process but maybe we don't know where to start but this is how you can and then we start that discussion so there is room I think for improvement especially in these areas capacity building and collaborating with other institutions for countries that are not yet doing it okay please go ahead in terms of room for improvement yes there's room for improvement one specifically with the research lab why do researchers have to move to Kampala to use the lab why can't we avail it online why isn't it virtualized make it easy for them as well that is a room an area for improvement and we are already working with you and you widen a feasibility of having that but then another one is for example in your area we collaborate with a couple of organizations you and you wider then you have ICTD then you have RSE and sometimes we have researchers from all these different institutions at the same time so another room for improvement and I've always wondered is there a way these other bodies who we are collaborating with can also harmonize kind of like the kind of researches they want to do is it even possible so we all sit on one table with all our four people we are collaborating with and develop one big research agenda but that never happens everyone comes on their own now the reason why this is important is we already said that as a Revenue Authority we have very few we have resource constraints in terms of persons managing the research and for each collaboration of course the internal staff have to be involved for purposes of clarifying issues about the process etc. we have one common table where we can all sit and we are different institutions and we talk about this it will be easier for the Revenue Authorities to manage these multiple collaborations I think that's an area of improvement thank you you don't need a conference for that you need a forum and I think the question that I have in my mind next relates to the issue of that Alan raised about why does someone have to go to Kampala to use the data and Wazi you talked about about data and data analytics and so on and leapfrog but be sure you land that's not mine I heard it yesterday thank you someone said it yesterday and I thought it was an interesting statement so that expression I heard and of course it conveys the idea that it's possible for developing economies to just bypass all of those laborious steps that other more developed countries have taken in terms of technology usage for example and just use the next best thing that's already out there and I think that technology is increasingly moving to the center of the work that you do in terms of research and in terms of using large data sets and so on so I'm coming to the question hold on how would you describe your current capacity for using technology and data and very cutting edge technologies and data in your respective areas Ezekiel saved by when you're in the virtual environment you can easily be forgotten but I still have a hangover in terms of the other question that was being discussed in terms of the areas of improvement before I come to this question if I may be allowed briefly there's this aspect of dissemination effective dissemination I think the last workshop we had with UNEYDA which involved a number of government institutions and also was quite a bit publicized in the media is one area where we can keep improving on to ensure that there's proper buy-in from other stakeholders government and non-governmental stakeholders to ensure that they understand what is going on and what benefits can come out of these collaborative arrangements which are very beneficial the gap studies that we did and where and the outcomes were publicized I think attracted quite a lot of attention from the policy makers and even as we were making some decisions regarding some restructuring within the organization I think the gap studies turned out to be very useful in very useful input in informing the decisions on where to deploy resources better because we already have these studies that have informed us where the gaps are in terms of collecting so I think they've been quite useful on a practical end but coming to the question of use of technology and also the issues that come in there we've noticed that also there's a lot of opportunity especially when you do the you warehouse your data you put it in a warehouse so for example we have a very flaged project that we are calling BAUC Data Intelligence Analytics System which has put together data from our internal revenue tax administration system and also from our customs administration system and putting it together and you know that to analyze this data you cannot analyze it or on simple technology I think you require more advanced technology to begin to make good use of this information so lately we've built this engine and we are having to move to a situation where we now need to utilize more technology to begin to connect the information I think one of my colleagues had mentioned the issue of making sure that the data is speaking to each other so to get to that level where you need to put this information together to begin to make sense across different sectors especially with interfaces that we are doing now with other institutions talking about use of third-party information and how we can make good use of it I think technology emerging technologies like AI I think come in handy to begin to simplify some of the issues that you may be facing in terms of analyzing the information so I think there is wider opportunity in that space but it also if I may be allowed it also created a bit of a challenge in terms of the governance framework around all this so we realized that when we had all this data put together we had no framework for governing the use of the information as the information is getting used who has access to this data and so on so we were working from we were kind of working upside down like you are beginning to now do some kind of firefighting we are putting in a data governance framework where after we've had all these issues already put in place so I think there's a lot of opportunity to sanitize the environment and make sure that we are making good use of it so there's scope for utilizing modern technologies to deal with this Excellent thank you so much so I'm hearing Ezekiel saying that there's lots of opportunities to use data and technology but it's a bit of a double edged sword because you have to have the right frameworks in place to manage it correctly Ingrid would you like to weigh in here? Yes I mean I actually think it was one of the mistakes we made at the outset when we started talking to the revenue authorities about the data I think we were a little bit as researchers we were quite cavalier and we were like you've got this amazing data we can do all these things with it and I think the revenue authorities were rightly very cautious about that and said well this is highly sensitive data we are obliged to keep all this information confidential and we need to think very very carefully about how we manage that and the more data you start to merge in and I think that is the exciting frontier is to say well let's also merge in the education records, let's merge in the unemployment data, let's merge in the death records and the birth records and then you can start to do all these exciting things with the observation with records for people that are not in the tax data but the more you do that the more complicated it becomes the more you know there are capacity constraints here but one needs this is tricky work it's tricky data to work with and one has to be responsible in ensuring that the researchers are using it appropriately and so I don't want to sound as though I don't think one should put data out there but it's not an easy overnight task so I think one has to move slowly I think we'd all agree that it would be wonderful if more researchers could access the data but one needs to think very carefully about how one does that Yes you may Thank you, talking about technology yes at URA we've tried to optimize technology to make it easy for taxpayers to simplify our processes and create efficiencies of course with each process you generate a lot of data and what we've done really is to try and make use of this data internally by drawing through data analytics draw some indicators BI indicators to allow you to help you manage compliance better we realize that without technology it's almost impossible to know who the taxpayer is who are they dealing with to have that visibility of who they are and also to identify who is not compliant because everyone we have a self declaration regime everyone will submit their return and then how do you quickly know that this return is likely to be wrong so Alice technology has helped us do much data here and there and throw flags or indicators of possible non-compliance but as you say technology is always evolving there's always a better way a new tool a new idea for managing the data so it's a continuous improvement if I may say and also then the skill sets you need to continuously grow the skill sets of your team to be very analytical to optimize this data and with time you realize that the data you have internally as your is not sufficient then you need to bring in other data sets from other maybe utility companies maybe other now we're talking about the automatic exchange of information to kind of make the best optimize and identify where the missing revenue could be now technology in terms of researching yes the research has helped us because the different researchers international sometimes have more skills in analyzing data they are more advanced in some packages they are working with so again we take we leverage that we learn from them here and there that's what we've done thank you thank you so much I think not much for me to add really because as I've said we are generating this data we have this wonderful technology we have electronic fiscal devices so we know when people someone buys goods somewhere we have the information there and we have e-filing e-payment but really I think we haven't quite started utilizing it but the information is available it's there so looking forward to the collaboration I'm sorry that's fine it's there now I am mindful of the fact that we've gone over the time in terms of the clock not in terms of the time allocated for the conversation and that's because we did start a little bit late and I took the liberty did not ask for permission asking for forgiveness that I claimed back the 10 minutes so now I'd like to hand over to the floor for anyone who has any questions immediately in rule 4 thank you very much I mean I hope it's okay that there's a slight element of commenting in it my name is Finn Taab I'm professor at the University of Copenhagen I'm reacting to sort of a couple of dimensions of the discussion one is why should one go to Kampala and the other thing is oh can you point to your specific research having changed the policy and I'd like to sort of try to see whether I can say that look at the very core of these types of collaboration there needs to be an element of trust mutual trust and you are not going to get that mutual trust if you don't go to Kampala as a collaborator from the north or whatever you want to formulate it so you need to be there and that cannot be done in virtual conferences virtual conferences can be great tools for communication in many ways but the actual building of that trust will only happen once you are there so that's sort of one reaction to why do you need to go to Kampala it's not just a question that there are confidentiality things and so on where you need to be in the lab and so on but it's that broader how do you actually develop effective collaboration and mutual respect now let me try to move that sort of as aspect up to the relationship between the researcher and the policy maker how can you actually have a trustful relationship if the researcher will constantly run around looking for taking the credit for what is actually the policy maker that should take that credit so I mean however much I do understand from years of experience in an action with donor agencies and so on and so forth and the pressures they are on and so on I really am reluctant here I as a researcher I try to do my research I publish it in as good journals and books and so on as I can but I leave the credit to the policy maker let he or she take that credit also take the blame for when things go wrong but I really need in order to have a trustful relationship with senior policy makers I need to leave them to take the credit it's not for me as a researcher to say ah, I change your mind because that's not going to lead to a trustful relationship and I do believe that at the core of these exercises that are going on also in relation to what Ingrid said about the sensitivity the difficulties, the confidentiality and so on you are only going to get over those bumps or over those challenges if you have that fundamental trust among each other and African countries have a lot of reason to be skeptical also in relation to the international research community because we do see hijackers running around being there just taking things out and then running away and not being actually part of the capacity of this policy making process so I hope it's okay I added this as a comment rather than as a question thank you thank you so much for your contribution before I let the panellists respond to your comments we'll take a question from Rose I want to thank the panellists I think there's a lot to learn I just want to try to see the thinking that would be there in terms of doing or creating a research work having a system that would help the institutions to identify policy equations and work together with the policy makers and I like what Finn has said and I'm saying this because I know we have a project that we are implementing and because we are working together with policy makers of course the data is not as much as you are saying but we have kind of a framework of saying this is how we are going to identify policy questions and this is how we are going to move on in terms of research and I think if you have maybe such kind of a system put in place I think it would help you in terms of ensuring that it is not anybody just bringing in even a research agenda that we ourselves are able to define a research agenda that will be beneficial to the institutions but just wanted to find out whether you have that system of co-creating this research work that could benefit from the data that you have. The second element I wanted to find out is I'm imagining 10-year work of setting up the assets cleaning up and the like what kind of investments do you do in terms of the systems because I'm imagining that you have a system where anonymity is not necessarily done but you also have a kind of data management system where anonymity elements are done I'm just trying to understand how whether there is additional investment that comes in as the two are being done and then the final thing I want to ask is it's very true this aspect of systems changing, processes changing and some of us have actually been victims of that, that you go you're being told about maybe data for 10 years ago cannot be gotten because the system was different as you go about in implementing this kind of a project what have you learned that going forward you'll be able to have continuity in the way data is actually put together in as much as systems are changing is there in year 40 even to even have an in-house built kind of a system that you know will not be changing much but you'll be changing one or two things rather than overhauling and having just something very different which is maybe brought in, I'm just curious that the 10-year work that has been done and a system now comes in would it have to erase all that effort that has been put, thank you Thank you and we'll take one more question Thank you Amine Ibrahim from UNewider so my question to you is about the future what do you want to see in your research collaborations or internally with the data that you have what do you want to see in 5 years time where do you want to be in 10 years time what is the vision that you have I'll leave it there Wow Thank you for that question as well let us rewind and start with the first comment there I think some very interesting observations that he made regarding the issue of trust the question of trust and actually declaring yes, we have to go to Kampala if we want to build trust and if we want effective collaboration and mutual respect I see lots of heads nodding so I think that's a no contest kind of issue but I think the more perhaps more provocative intriguing issue had to do with the question of researchers and I researchers claiming credit for moving the needle as opposed to a policy maker making a decision that I will take this research data on board and I will make a policy that will have X impact on revenue collection and of course on people's lives please do Thank you so much for all the comments the first one about the mutual trust I'll take it we've been in discussions with Finland they have an online lab and in their case study they first have the memorandums of understandings and the discussion at institutional level so you find that it's faster institutions agreeing to open up their data virtually and then within the institutions institutions have persons for example we have an agreement and then it's you and your wider to second their different persons but yes in many relationships a virtual country plays the face to face sometimes when you know who you're talking to it's much better that way but for bridging that gap we've been thinking about really doing it at institutional level then later on within the institutions we have the persons then talking about taking credit for research work and policy we have guidelines and what our guidelines have always been is that for many of these collaborative researches you need the local staff within the authority who understands the processes who understands the data etc and our requirement has always been for any publications but you've got to recognize to add either as a co-author the local person who you worked with because yes in a collaboration still you need the local person who appreciates the processes and the actual data to successfully complete it and then I'll move on to the other question was about what happens if a process changes and you've been working for 10 years and you have all this data yes data is an output of the process and the process really is our intent is to simplify tax administration so as we keep continuously simplifying tax administration we change processes and data sets change what we've done or what we intend to do or maybe we have a data warehouse internally where we again aggregate this data for use of BI but what we do is that you take note of the moment when the process changes happened so for example you say for the last 10 years the structure of the data is like this but from this moment forward because of this change in maybe providing efficiency there's a process change so then the researcher and everyone knows that from this moment on the data sets there's a change in the data set so in a way you are not losing what you did before in terms of aggregating the data and storing it but you're taking note so whoever is using knows that there's been a change in the process so the data is different there are some fields that no longer exist moving forward then I think Rose's concern had to do though with ensuring continuity based on those changes and Ingrid you'd like to chime in here I'll be very quick but it builds on this point what Finn was saying and what Rose was saying I think economists tend to think of data as it's a public good and once you've put together the data set you should just put it out there and everybody can use it and I think the sophistication and the intricacy comes around these issues around the fact that tax systems change the variables change you need metadata to explain to you how data is generated tax credits in a different way and there's an enormous amount of that nuanced local information that you need in order to use the data so I think that goes to this point partly about data harmonization but it also goes to the question about the collaboration and why you can't just put the data out there and then for Amina's question about how does one institutionalize this and make sure it continues I think the amount of investment that's gone into setting up these systems is not insubstantial and yeah unless the revenue authority themselves is able to commit in the long run to doing that I don't really see a future without that kind of commitment there's now been this huge investment from external partners into these centers but now there needs to be local ownership because otherwise all that information about how systems have changed etc will be lost to some extent but it does go to the point of why you have to go to compiler to do it because otherwise you really aren't going to do justice to the exactly to it and one last question about formulating the questions I think it's that's something we've really struggled with in South Africa there's been a strong perception from some groups of researchers that there's been gatekeeping that you can't get access to the data unless you've worked through the system and I think that's again comes to this point of why were these systems set up what is their purpose and ultimately the purpose of the centers is to ensure that there is better policymaking and you can't have better policymaking without the involvement of policy makers all right I would like to bring Ezekiel back into the conversation you've been sitting sort of sitting things out for a while and Ezekiel I'd like you to address the question that was raised by Rose about the idea of co-creating research work and also the kind of investments required into research data systems and finally also what's your future looking like what do things look like 5-10 years down the road okay so I think the best starting point is what the first speaker raised the professor the issue about trust mutual trust, mutual respect I remember it takes me back to the time we were formulating the memorandum of understanding with you anyway we spent a lot of time on the MOU to a point where it nearly collapsed the agreement nearly collapsed because we I must be honest here we were not very sure of the motives behind the collaboration up until there was a give and take some explanations and so on because what we put on the table is what's in there for us so I really like the point the way it's been summarized why do you have to go to Kampala add another town why do you have to go to Lusaka it's for the same reason that you have to gain that trust so up until we've been sitting down with the teams we've discussed the concept we've discussed the ownership of the product then it becomes very clear that if the researcher is going to take credit then it becomes a bit problematic so I like the way it's been summarized leave the credit to the policy maker again you would also leave the credit to the to the DAX administration and that's what's going to create benefit for the administration of course there's a way in which the collaborators are also going to get benefit and that's where co-creation comes in you create something together but then you call on it and so there's that there's that trust that you you need to have initiative about the data and how it's going to be used looking at the investment like you have said we've invested into these systems for quite a long time it's not cheap to invest in the systems but it's not so much about the cost I think it's about what is contained in that information and what interpretations and who is going to take the credit out of it so what we would really like to see is really what the professor summarized it very well so we want co-creation we want to ownership and we also want to get the credit we want the benefit to accrue to our researchers that as they are working with these experienced researchers they're also getting the benefit of the training of the of just making sure that there's understanding of this information that we have at our disposal so I think there's no better way to put it in terms of what we would like to see would like this mutual respect I think it's a very important thing that as we work on this there's this mutual respect and there's the trust we have to see the real persons I think that those have been the issues like you cannot create this issue over a virtual a virtual arrangement there must be people behind you must have clear agreements around these things so that you are able to generate that trust that is required so for going forward I think we've established these relationships but what will be more critical is to ensure that even within the revenue administrations themselves we are looking to making sure that even the subject matter experts are heavily involved in this it's not just the preserve of the research department it's everyone else in the institution to call on whatever comes out so I think there's need for the teams involved in these products to get involved so that there's ownership, there's benefit there's simplification because what comes out of what we do I think we would like to do things much better so if we focus on solutions I think that's what will work best for us so maybe that's how I can put it in a nutshell thank you thank you so much Ezekiel and Wazi your thoughts on a bunch of issues I hope you've been taking notes okay thank you I think my colleagues have talked a bit more on the first questions but I want to start with Amina's question what is the vision for us for me I think the key ones that I see are very problematic that I would really love to see happening especially in Malawi one is if we are able to have these unique identifiers for individuals and taxpayers because we have the TPN for the tax administration and then we have the national ID that I used so to integrate the data it becomes difficult in the next 10 years I think it's happening in Tanzania where they're using the national ID as the tax identifier I would love to see more countries it would make life and you know at this analysis a bit more easier that's one second I would want to see more of this collaborative work there's a lot of potential in raising revenues in most countries when you look at the policy gap and the administration gap in most countries it's quite huge and how do we close those it's more of this research but as you've said maybe the research we don't have the capacities and so this collaboration adds to that capacity adds to the robustness of the results which will help us now have policies that will close these gaps so a more collaboration I think that I would do very much support and lastly I think one of the tax that has a lot of potential for most African countries is the VAT the vaccine efficiency for Malawi is at 14% but I understand for other countries as much as 50% for African countries around 30 something percent there's a lot of potential still to collect more revenues from VAT so I would want more of this research where are we missing it how best do we ensure that when you go to the shop some are not asking do you want a receipt or you don't want a receipt how do we get more from VAT I think we need a bit more maybe reset but I would want to see in the next 10 years that we have improved on how we are collecting the revenues from VAT because I think there's great potential there for most countries quickly now I think one more is the process has changed how would we ensure that we are keeping the data this is a big question the my revenue authority now in the process of introducing the ITAS the integrated tax administration system we are moving from we had different systems we are using about four of them to keep taxpayer data but we realize once we started developing the ITAS we couldn't migrate the information it was very difficult so a decision was made to say we are starting from today using ITAS the historic data we forget it we just move forward here so that from say 2019 that's when we are collecting the data we go forward we have lost all these years the my revenue authority started in 2000 but all this data up to about 20 years of data we have lost it so it's a good question indeed and it's a big consideration I don't have the answer now because we felt how to handle that as well other than someone manually inputting the data but I should take those too thank you thank you so much we've now reclaimed our 10 minutes that we lost during the break so I think it's a good time and sort of a natural a natural opportunity to close this afternoon's conversation which has been outstanding as always thanks to our wonderful panelists thank you so much Ingrid Warlad thank you Alan Nassanga and of course thank you so much Ezekiel Piri we've heard about the challenges that are faced by researchers tax administration actually tax administration and their research collaboration in sub-Saharan Africa many challenges but also many opportunities as well and I feel overall a positive sentiment going forward in terms of finding new ways to close the gaps when it comes to revenue collection so thank you so much all of you I know the conversations and the work will continue and we look forward to hearing more of what will come next thanks everyone have a good evening