 I want to begin this series of recordings by giving a short introduction to microsimulation in general and the Tanzania microsimulation model TASMOD in particular. During the course of this presentation we will briefly explore what is microsimulation, then consider what is TASMOD. Rwyf ni'n gynnwys y tasmod yn ddweud, a'r gwerth ddweud o'r tasmod yn y gweithio, ac mae'r ddweud bryd o'r ddweud o'r cyflwyno yn y training. Rwyf ni'n ddweud o'r micro-simulatio. Efallai ychydig yn wirtwnedol, micro-simulatio'r modelau, yw mycro data ar siqaf o'r gyllid o'r mycro unydd o'r holl. A'r amser yn fawr anhel ar gyfer ond finnwys iawn mae'r rôl yn cael pom ar y mycro. Oherwydd, mae'n mycro data sefydliad dda wedi bod yn Y Cwysig am Tasmod, of the case of the case of TASMOD, this is the Tanzania household budget survey and that micro data set contains information on individuals and households has information on family structure, income and expenditure and that forms the base data set on which the micro simulation model works. Now the micro simulation model itself ond, mae'n ddrygu'r gweithio tansynion y cyfnodau, mae'n ffordd gweithio gwybod a'r ddysgu'r cyffredinol a'r ddysgu'r ddysgu'r cyffredinol, yn ei bod yn ymgyrchu'r ddrygu'r ddysgu'r cyffredinol, ac mae'n gofynu'r cyffredinol o'r gweithio ymgyrsiau sy'n gweithio ddrygu'r cyffredinol. Yn y cwestiynau microbynfadau gwahanol bod yn cael ei MCF yma yn ystod i'r pawb, wrth y tynnu cyfnodwyr cyfrannu sylwadau maen nhw, byddai cyfrannu a bwysig iawn i'r byw, ac mae'n ddiddordeb i ni'n gynnwys ffordd i'r parbyn fel ymddangos a'i cyfrannu pa Allun i'n cyhoeddwyr Nadol Ynnyfodd ac Iechyd Gwynt Bwysig. This will become clearer as I think when we go through the way the model works in detail. But basically this power to be able to simulate policy systems is a very good one, because it is not only simulating the current policy situations, ac mae'n gael'r hyffordd yma i gael y cyfle o'ch gweithu o'r sysgol sydd a'i gael ar gŵr curves. Yn hyfin i'n gweithio i'r gweithio, i'w cyffredigau mylion oedd yn ei gweld ar gweithio hyffordd. Mae angen yw'r cyffredigau mylion i gweithio'n gwahanol i'r gyfan hwodraeth, sy'n gynnwys cwestiynau pwysig hynny cael cerddwyd ar y system ond y cyfan y cyfan ymgyrch yn y cerddwyd. Felly, yn y cyfan y cyfan, mae'n ddweud yn cyffredinol. Gwydwyd yn cyfan y cyfan y cyfan y cyfan y cyfan y cyfan y cyfan sy'n cyfan y cyfan y cyfan y cyfan y cyfan y cyfan y cyfan y cyfan can actually look at the difference between switching on the simulation of the various social benefits such as the PSSN benefits I talked about. And then also look at the distribution with the benefits switched off and then therefore be able to look at the extent to which the existing PSSN benefits actually do reduce poverty and inequality. It also allows us to simulate the cost of implementing a particular social security policy reform. So for example we could modify the rules for the basic cash transfer and work out how much that would cost. Or we could introduce a completely new cash transfer such as some kind of transfer for unemployed youth. And we could actually cost out that. And importantly we could work out what would be the impact on poverty and inequality of introducing a new policy or modifying an existing policy. OK. For we actually get to the nitty gritty it's important to actually say something about where TASMOD fits with a broader family of micro simulation models in Europe and elsewhere in Africa. So TASMOD is a static tax benefit micro simulation model which is based on existing EuroMOD model architecture. And EuroMOD has been developed over a 20 year period by Professor Holly Sutherland and colleagues and is currently using over 25 countries in Europe. However it's moved beyond Europe and there is a current collaboration between a number of parties and the United Nations University to bring the EuroMOD platform to Africa and elsewhere in the developing world. So who are the partners here? Well the partners are the University of Essex, UNU Wider and the Southern African Social Policy Research Insights SASPRI. It's a major research programme in which tax benefit micro simulation models for selected developing countries in Africa and elsewhere are in the process of being built and finalised. And the outcome of the work will be a set of models for individual countries and research papers that contain simulation analysis of tax and benefit reforms and we can talk about some of that in relation to Tanzania later. Now SASPRI is working with local partners to develop models for Mozambique, Tanzania and Zambia and SASPRI is also updating SAMOD and NAMOD as part of the southMOD activity and SAMOD and NAMOD are in a sense African prototypes of using the EuroMOD framework. SAMOD was developed by SASPRI ten years ago as part of a programme of work for the South African government to produce a transparent user friendly micro simulation model for South Africa. And NAMOD is a sister micro simulation model that was developed for Namibia around four years ago. OK, TASMOD itself, the Tanzanian model has a standalone user friendly interface which we will look at in a little while. The important thing is that the model workings are transparent and the user has full control over the simulations carried out. So this is really important and it underpins the EuroMOD philosophy and indeed the SouthMOD philosophy that users of this model can actually make the changes themselves and analyse the output themselves. TASMOD does not rely on continual reference back to consultants or others when a government for example wants to test out some particular tax or benefit. The current version of TASMOD version 1.0 is underpinned by the House House Budget Survey 2011-2012 as I think I mentioned earlier and obviously it can be updated with new data and policy reforms when required. So when the 2017-2018 HBS is released then the model can be adjusted so that it becomes updated with the new data set. It also as taxes and benefits change it can be updated with new policy rules. So the first step then in micro simulation is to collect data of the incomes and expenditures of individuals in a representative survey of households. And as I've indicated TASMOD is underpinned by the HBS 2011-2012 and this represents the characteristics of the mainland Tanzania population in 2011-2012. And it is just the mainland population as people here will know. We do want though in due course to be able to produce a system model for Zanzibar and we're beginning negotiations in this regard. Now the HBS 2011-2012 is a representative sample of 46,593 people living in 10,186 households in the Tanzania mainland. And the weights from the HBS 2011-2012 as adjusted are used to calculate national figures from the simulations when we need to calculate impacts on poverty and inequality and cost. The next step is to have a set of policy rules that can be applied to individuals in the data to determine what social transfers they are entitled to and what taxes they should pay. In order to simulate a policy we have to be able to translate the rules into a format that can be understood by the model. So that's one constraint. So the rules need to be in such a form that they can be translated into rules within the model. The second is that the data set needs to have information that will enable us to apply the rules. Now these two things aren't always straightforward. There's often constraints within the data. The data didn't ask particular questions that are necessary to model a policy. And often here we have to make assumptions but we need to just make explicit assumptions that we have applied. In Tasmod we've produced policy rules for the years 2012 and 2015 and that's, we describe as Tasmod version 1.0. We will shortly be preparing the data rules for 2016 and then 2017 all within the course of this year. What are the actual policies that we have implemented? Well in terms of social assistance and social insurance we've implemented the productive social safety net policies, a fixed basic cash transfer and the variable conditional cash transfer. We've also simulated the public works but in this case we've only produced an eligibility flag. And even with those three types of social assistance we've had to make assumptions which we can talk about in questions. We've also simulated the National Health Insurance Fund, both the employer and employee contributions and will in due course be simulating pension contributions. In terms of direct and indirect taxes we've simulated PAYE income tax for those in receipt of income from salaries and wages. Also presumptive income tax which is payable to small traders if their turnover from self employment is less than 20 million tansony and shillings per annum. We've also simulated personal income tax for account cases, these are the self employed whose income exceeds the presumptive tax threshold. We've simulated a limited number of excise duties, in fact those relating to alcoholic drinks, tobacco products and vehicle fuel. In the case of vehicle fuel it also includes the fuel levy. We have also implemented value added tax across the board. But it's important to say that part of this training will be to review any other policies that need to be implemented for 2016 or 2017 and indeed to get your views on the way that we've implemented the policies that we have implemented. And if there are things that don't appear to you to be correct or that could be improved it would be really important to get that feedback because you are the key stakeholders in this. And then we can amend the model to build a better model. And it's important I think at this point to say that model building is an iterative process. I mentioned that SA mod has been going for 10 years and I think progressively over that 10 year period the rules have been refined within the model and the model has got better over time. It's an organic process. We need to get the model to be as good as it possibly can be and it will inevitably be improved over time. OK, we've talked a lot about what TASMOD can do. What can't it do? Well first of all some existing policies aren't simulated and those are not the ones that are not on the policies list that I showed you in the previous slide. Such as pharma subsidies, pensions and pension contributions, though I did mention we will be simulating pension contributions later in the year. It also doesn't simulate changes in behaviour. It's a static micro simulation model which simulates what would happen if the policy rules were as they are specified in the model, the immediate impact of that. And if you change the policy rules it's the immediate impact of the change. It doesn't take into account for example that if you increase cash transfers that might make people more mobile and more able to get jobs and improve their job search and improve their income. That would be a behavioural change which isn't covered by this model. It also doesn't take into account any macroeconomic effects. The focus is on the direct impact on households. However, the output from the model can be used as a starting point for more complex analyses involving behavioural change and indeed even micro-macro linkages. So, for example, in South Africa we used SA-MOD as a tool for forecasting tax and benefit expenditure into the future. The tax benefit expenditure and tax recovery into the future and fed these into a social accounting model. OK, let's now look at TAS-MOD and give an overview of what the interface looks like. For this I'll use these screenshots but I will also show you the actual live model itself. First of all, the input data set it's worth saving is in text format. It's already supplied with the model but new data sets can be added and existing ones are added and that will happen from time to time. This screenshot is of the model program itself which stores all the model parameters and allows the user to make changes and run simulations. Then it generates output in text format that can be analysed using a statistical package or one of the built-in tools so that the impact on policy reform or poverty and inequality can be examined. Now I'm just going to now move back and look at the model itself in real time. So here is the actual model. I close up all of the expanded parts but this is the model interface and all of the policies are listed that are simulated. As I'll explain in subsequent sessions the policies are of various kinds. There are what's called definitional policies which for the time being I won't mention but there are also the insurance contribution policies. I watch my cursor and then the income tax policies, the different policies for the PSSN benefits and finally excise due to it and value added tax. But going back to the presentation I'll show key parts of the user interface. So first of all there are seven tabs, countries, display, country tools, administrative tools, add-ons, applications, help and info which are along the top and as these sessions unfold you'll get to know how to use those, which ones are useful, which ones aren't useful or as useful. And then there's the representation of Tanzania's tax benefits system for different policy years which are in that column called policy. They're in their collapse state at the moment so you can't see all the workings but you can see the headline taxes and benefits that are simulated. Then there's the run Tasmod, in fact it's now still called run Euromod button and that generates another menu once you press it with additional functionalities. So how might Tasmod be used? Well Tasmod so new it hasn't yet had any uses that we can talk about but we can talk about South Africa so I mentioned SA Mod is now really getting quite mature and it's been used in many different ways. So for example it's been used to simulate a youth benefit for the National Department of Social Development and there have been various different versions of youth benefit that have been simulated and there's actually a journal article which details how it's been used. The National Department of Social Development I should explain is the department in South Africa responsible for social grants, that is social transfers. We've also simulated a carers benefit again for another project for the Department of Social Development. A non means tested universal pension, this was done for the National Department of Social Development and the National Treasury and then most recently a non means tested universal child benefit for the Department of Social Development. So those give you some examples of how the model might be used, obviously there will be different situations in Tanzania but how it might be used in practice. Ok just to conclude the rest of the training course, the next session will be called getting started with Tasmod and we'll go into that interface that I showed you in more detail. We'll then have a special session just looking at tax and benefit policies, then a very important session on introducing new policies and amending existing ones and finally a session on analysing the output data. After this what I will do is take people through the model just in more detail than hitherto, so if I can go back I will just take us back to the model itself and conclude by just walking you through the model in a little bit more detail. I mentioned the tabs, seven tabs, countries simply there is only Tanzania at the moment though as I also mentioned hopefully there will be in the not too distant future Zamsibar to join it. The display tab has all sorts of options that enable you to for example implement conditional formatting which means when there is a change you can see it because it will be highlighted in a different colour etc etc. It's not one that we use very often but as you get more experience at using Euromod you tend to use it more. The country tools allow you to add a system and we'll talk about systems in the next session but a system is something like this column. There will be a new system for 2016 and 2017 in due course, this allows us to add a system and delete a system and those are probably the two most important there. The other important one is operating indices because one of the things that's clear when you've got a data set which is 2011-2012 and we're now sitting in 2017 is that we need to up-rate all of the monetary data in that survey up to 2017 prices and we do that using the up-rating indices. So that's that and the add system and delete system are the most important buttons on that particular toolbar. The administration tools that by far the most important one here is the variables tool, again that will become clear in later sessions. The add-ons will become important when they've been completely and finalised for the South Mod programme. Applications is useful or at least the open output file application is because this enables us to check the data very quickly by using the Excel button because it enables us to examine the output data in Excel. And then the others are yet to be properly implemented for South Mod when they are, summary statistics will allow us to look at the summary statistics for Tanzania. And then finally the help and information but the help particularly will be useful for helping debug a function etc. Now we're not going to go into these things in detail but each of these different policies does unpack into a series of steps and I'll do the operate one just out of interest. There's something called a function and this is the operate function and then the function itself has a series of parameters. And some are more complex than others. Income tax for example has a total of five functions and each one of those functions has a number of parameters attached to it. All of this don't worry about at the moment. It's just to illustrate the interface and the fact that it can be seen in the collapse form but then you can un-collapse it in order to work on the different policies. And that's all for this session. Thank you very much.