 Okay, so thanks a lot, it's a great pleasure to be here. Many thanks for attending this session. So I talk about an ongoing project in South Africa where we try to quantify multinational tax avoidance in South Africa and think about the effect of countermeasures against multinational tax avoidance in the South African context. It's very much work in progress as you will see and we just share what we have up to now. So as we heard many times in this conference, multinational tax avoidance is very high on policymakers agendas and we have seen all these reforms unilaterally, multilaterally that aim to constrain profit shifting to low tax countries and developing nations and emerging markets have played a very active role in that process. But looking from an academic perspective, many questions are arguably still unanswered. So in the context of developing countries, how prevalent is multinational profit shifting actually and how effective are anti profit shifting regulations? So we have some evidence and the evidence mainly, especially on the second question is on the developed world and there's reason to believe that these results may not necessarily be externally valid for less developed countries, among others because the tax capacity in these nations is lower. Okay, so what we do here is be a very rich data from South Africa tax administrative data where we can see the universe of corporate tax returns can link that to the universe of firm level trade data in a very fine grained way. So we can see trade in very fine grained six digital product categories between South African firms and partner countries and that allows us to get some notion about profit shifting behavior and we can distinguish between multinational and national firms and to some degree. And the main questions are here and bold on that slide is so we want to quantify profit shifting and mainly we also want to get a notion on how has it changed over time? So we've seen a period of a lot of effort to constrain profit shifting. Did that show any result? Is there any evidence that this problem has gone down over time and can we link that to anti profit shifting regulations that have been implemented? So if you think about how to quantify multinational profit shifting then the whole problem is about low tax reporting of multinationals in a high tax country like South Africa. South Africa is arguably a high tax country levies a tax rate of 28%. So what you basically want to track is do these multinational report enough profit in that country and one approach is just to compare them to national firms that are arguably kind of similar but do not have at their disposal this option to use international tax avoidance. Now using the strategy is very difficult because if you think about multinational and national entities they are arguably very different. Dominika has a nice paper that makes this point on a theoretical basis. So what the literature does is basically to match on size and industry characteristics that might not be enough so you might also match on productivity on market power. And the problem with these measures is that they are affected by profit shifting so we do not have clean productivity measures and one medium run aim of that project is actually to come up with productivity measures that are cleaned by profit shifting. So here in the first step we look at pre-tax profitability of multinational firms relative to some comparison firms that are arguably similar and to look at something that we call zero tax reporting. This is something that Niels has written like a couple of years ago or established which is the observation that some companies actually constantly report very low profits in high-tax countries every year. Again zero or very close to zero. And this is very suggestive that they have profit shifting strategies in place where they have fixed cost that they expend at the beginning to set up a scheme and that allows them then to shift like much of their profit out every year. And the question is, do we see that in the data? And I show you here the distribution of pre-tax profitability of two types of firms in South Africa, so in blue this is multinational firms so pre-tax profitability range between minus one and one and the zero in the middle and in red you see national entities arguably kind of similar. And you see this big spike on the zero so that is arguably something like zero tax reporting and you see that it's much more pronounced for the multinational firms. And this is not any multinational firms but this is multinational firms in the data that have a parent company in a haven country. So arguably firms that by their structure are already kind of suggestive in engaging in profit shifting behavior. What is very interesting is if you do the same graph for multinational firms that do not have a parent in a tax haven country and this is actually the large majority of firms in South Africa. So haven parents, it's something like 700 or less depending on the year you look at firms with a haven parent. So this is like the bulk of multinational firms that do not have a parent in a haven country and here this graph looks arguably quite similar for the national entities and the multinational entities. So I guess what and others have written about that as well is that you can take away from that not every multinational firm engages in profit shifting or at least suggestive from this graph but it's a smaller number of entities that do so but the smaller number of entities tends to be large. So it's the largest firms that do that and so in terms of revenue loss that may still be substantial in your economy. And we kind of see the same pattern if we look at pretext profitability rates. So here in the upper part we compare multinational firms with a parent in a haven relative to control firms and you see they are less profitable in terms of reported pretext profitability by around two percentage points. This is quite a bit. And you do not see the same type of pattern if you compare the pretext profitability of multinational firms without a haven parent and comparable entities. So what we then do in the project is look at specific profit shifting channels and what I show you here is infra firm trade mispricing which is arguably also reflected in the literature and one major channel perceived to be a major channel where firms strategically distort prices for infra firm trade to shift profit from high-techs to low-tech countries. And our data is exceptionally well suited to assess these practices because we see these fine grain product categories and can then basically ask if the profit shifting incentive changes because the partner country, for example, changes the corporate tax rate, do we see that the price for this very fine-gain product category, the price that the firms charge, does it change in a way that is consistent with profit shifting? And the answer is yes, we see that in the data. So this is like a regression output and what you need to look at is this negative coefficient here which just tells you if the corporate tax rate in the partner country increases and this is import rate, then this reduces infra firm import rate prices and this is consistent with profit shifting adjustments that you see in the firm. So if you think about a low-techs country, then import rate would be overpriced for profit shifting purposes. If the corporate tax rate goes up, the price goes down, this is consistent with less profit shifting incentive. So arguably where we actually want to go with that project is to understand a bit better how did these incentives change or have changed over time. Can we say something about the effectiveness of anti-profit shifting regulations? And I think the my general perspective would be that we have seen like a decade at least or even longer of activity against multinational profit shifting or at least the topic has been very high on the public agenda, on the policy makers agenda. There have been many reforms and kind of the landmark was the 2013 kickoff of the BEPS project of the OECD. And so arguably there was some anticipation and actual tightening and profit shifting regulations and there has been a lot of investigative journalism with individual firms being exposed by journalists of profit shifting practices. I think there's kind of an awareness especially in consumer facing multinationalist that this is really a risk being exposed or getting negative reports about multinational profit shifting. So I guess you can think about the whole decade or all these years about something happened in this profit shifting domain and did that show up in general in the behavior of multinational firms? And then of course like more specifically you can ask did any specific measure add to constraining profit shifting and if you think about the South African context this could be measures taken by the South African government. And if I think about my sample frame then there have been two main measures that stand out. This is the introduction of country by country reporting in 2016 and kind of related to that a tightening of transfer price documentation regulations where multinational firms with very high levels of intra-firm trade were actually required to give like a detailed transfer price documentation. And I can ask that that have any effect on the behavior of companies. And then from a South African perspective of course there were all these other countries I'll show you a graph in a minute tightening their rules made this also have had an effect. And here you see a suggestive graph that does not look that as nice as I would like it to look so it starts in 2013 and shows you the development of intra-firm trade prices for goods. So unfortunately I do not see service trade so this is only goods trade between South Africa and between tax havens in South Africa. So this is import trade and the green line is the prices for intra-firm import trade and the blue line is for extra firm import trade from tax havens so both from tax havens. And this is everything is conditional on the firm and conditional on the product category. So this is really, this is not compositional effects but price effects. And what we see is that the intra-firm trade prices starting in 2013 they tend to decline and we do not see the same type of movement. For the extra firm price we of course want to extend like in earlier periods to see when this actually started but we haven't done that yet. But the data actually is available from 2009 actually. So but this is kind of suggestive that there was kind of a adjustment at least if you look at tax haven trade and this is like typical tax haven countries on the rice and Heinz list or Damer-Paller and Heinz list. Okay, so if we think about what happens to trade volumes with tax havens then we see that the propensity to trade with tax havens also came down in the same period. So here you see aggregate trade volumes at the outset of my period and at the end of the period and you see like a mild decline of trade with tax haven countries. And here you see the number of firms and this is here in blue, the number of firms in South Africa that have a parent in a tax haven country. And this also came down here in the middle of the sample period by this is all firms in South Africa. It's scaled differently. So this is very few firms and this is several thousands of firms. But you see that the trend is like different in these lines. Suggesting somehow that like the tax haven presence of these firms at least to some extent declined. And the question is can we somehow link that to policy actions in the country? And as I said, one prominent policy action was in 2016, the introduction of country by country reporting and the tighter transfer pricing provisions for certain firms that have very high levels of intra-firm trade. And so what we did in a second analysis is that we asked the question again, this is here import prices for haven trade so it imports from haven countries. And we see this would be like the general trend that these prices declined within that timeframe. And the question is do we see any additional effect driven by countries that companies that are actually treated by this intervention? And in terms of the point estimate, this looks a bit suggestive but it's not very precisely estimated. So that would not be statistically different from zero in a statistical sense. But here we can go back to that in the discussion if you're interested. We need to kind of simulate and make assumptions which companies are treated by that. Because we see intra-firm trade only on the import side and not on the export side. And so we somehow need to model who's actually treated and perhaps we do not model that in an adequate way right now and this is where the noise comes from. So then the last question or one of the last questions would be did it have any effect that many, many other countries in the same time period also tightened their anti-profit shifting regulations? And what you see here is the number of countries with transfer pricing rules being in place. So green would be general transfer pricing rules and red and blue would be transfer price documentation requirements starting in the early 2000s until about today but there's like a steep increase across the world. Should we expect that this somehow impacts South Africa? So in the first step you may think about transfer pricing regulations. They constrain mispricing behavior at the firm level so it might affect intra-firm trade. But if you now think about these reforms happening in other countries and South Africa being a relatively high tax nation you might actually not expect something really happening at the bilateral level because if it's a higher tax country that tightens its transfer pricing regulation the profit shifting incentive is actually small. So it may nothing happen for that reason. And if it's a low tax country the low tax country does not have any incentive to enforce that regulation. So you might just expect that something indirect and general may happen that of course if other countries tighten their rules that may hamper profit shifting to low tax countries and if there are some kind of complementarities it may also constrain profit shifting from South Africa for consistency reasons for example. So it's probably very difficult for a firm to charge on the same product vastly different prices in tax haven trade from one high tax country and the other. However, one interesting thought is there's this very nice paper and many of you may know that by Toslov et al. who say or argue that the pricing behavior of firms may in part be driven by tax authority incentives and what they document for Denmark is that the tax authority in the transfer pricing domain does not go after the tax haven trade but after the trade between high tax countries because the tax authority always has an incentive to tilt the price in its favor. And that may give funny incentives to the firm if there's one tax authority with very tight regulation or very tight enforcement and another very lax tax authority and the firm does not want to have hassle or there are a hassle cost then it might tilt the price towards the tax authority with the tighter regulation and then if these transfer pricing rules change it may tilt the prices in favor of the tightening jurisdiction. And the question is do we see that in the data and the answer is no. So here we basically look at South African partner countries tightening their transfer pricing regulation either introducing legislation or documentation requirements and the question is so if this is in portrait again if it would tilt the prices in favor of the partner countries we would expect that these prices go up but we do not see that in the data. Okay, so what I showed you so far and as I said this is like ongoing and preliminary to some extent I showed you at least some suggestive evidence that this tax haven presence of multinational firms has to some extent mildly declined over the last decade but what you see here in the graph and this is again the zero tax reporting for at the beginning of the sample frame and at the end of the sample frame and you see that this has not massively changed. So unfortunately the scale is not like right so this spike is actually smaller than this spike so we have remote access to that data and I didn't manage to get this right but you still see it in the data so it's not gone away so the problem has not vanished so our intermediary conclusion as firms seem to adjust to some extent but it didn't go away the problem and we want now to dig deeper into other profit shifting channels and as I said would also like to clean up the productivity and market power measures from profit shifting effects in the medium run. Okay, thanks for listening.