 What is corporate profit shifting? In its general sense, the broadest sense is the illegal movement of money from a high-tech jurisdiction into a low-tech jurisdiction. So the aim of all this is of course to lower the total tax that is due of the multinational group of companies or multinational corporation. Now there are lots of ways to do so. The major ones are listed here, but the companies are getting innovative. So every time there is a policy that combats one of these ways, the firm finds another. All right, so on the simplest example of how it works, imagine a firm A that resides in the Czech Republic, my home country, and imagine a firm B that resides in Cyprus, which is a very popular tax haven for Czech firms. Now imagine that the firm B owns firm A, right? So there is some kind of an ownership link between the two companies that is translated into data as a foreign direct investment of firm B into firm A. And what firm A does is that it shifts the profit from the Czech Republic. So the profit that is made in the Czech Republic shifts it illegally into Cyprus, into firm B. Now why would a firm do so? Of course, one of the major reasons is that in Cyprus the corporate income tax rate is lower, but there are also other reasons like increased financial secrecy in Cyprus and with respect to the Czech Republic and other favorable conditions and minimum holdings and so on. I'm gonna sum up the motivation of my paper or our paper in three main points, three main areas. So the first one is gonna be, is there even actually corporate profit shifting? Now the cases in the media and some evidence of individual cases suggest that there is. So the answer to this question is obvious maybe, but it serves as kind of a robustness check of our methodology if we get the same answer as we would anticipate, right? So the second question is much more difficult to answer. It asks how much corporate profit is actually shifted and from which countries? And the third question is how much do these countries lose on tax revenues as a result of this profit shifting? The methodology that we employ in this paper is called the FDI approach. It has been pioneered by YoungTab in his 2015 World Investment Report. Now the basic idea of the approach is that if there is profit shifting, so we assume that there is, then it's gonna show somehow in the data, right? Which is what we want. We want to observe the practices in the data. So the way it shows in the data is that as deflated reported profits in higher texture restrictions. So basically what we could say is that if there is more FDI from tax havens in one country, the reported profits in that country are gonna be deflated somehow because the profits were shifted before they were taxed in that high texture restriction. So coming back to my example with the Czech Republic and Cyprus, we're focusing on this channel. Is it gonna show out in the reported profits of Czech firms as decreased, right? And the report profit in Cyprus has increased. All right, so what kind of data is available on this to evaluate this problem, to analyze this problem, is the first one, the first category is the FDI stock. So the amount of money that is invested from tax havens or from these bad countries into the good countries, right? The well-behaved countries. So on FDI stock we use the IMF Coordinated Direct Investment Survey that contains some kind of a matrix of firms that have the investing countries as rows and the invested in countries as columns, right? So there's kind of a bilateral matrix. We also complement this by Ongtec's own FDI statistics, which is a unilateral database. So there is only the amount of inward FDI and we don't know from where that FDI comes. Then the second part of the relationship that we want to estimate is the rate of return on this investment. Now to do so we use the IMF balance of payments data which contain the rate of return on FDI that is invested in a country, okay? We also use a component of this as a general process, the quasi-component which is a subset of the rate of return, all right? What we need is some kind of a measure of how much foreign direct investment in a country comes from these bad countries, right? The non-compliant countries, you can say. We define as, we call this the offshore indicator, which is the share of the more of the FDI from these risky countries. We classify those in two categories, the tax havens, so you're placing the China Islands, China Sea, Cyprus, countries like this. We say that 100% of the investment from these countries is risky because the electronic activity in these countries is very low compared to the level of outward FDI from these countries. The same group of countries is those that report data of special purpose entity's activities. What are, for example, networks of the submerged areas, yeah? Some countries, central banks, report how much profit or how much money of the FDI figure that is reported is actually just money flowing through that country and doesn't stay actually in that country, okay? So for this purpose are exactly special purpose entities created. All right, so imagine the offshore indicator as, this is FDI in country X, this is FDI from tax havens, and this is FDI from these SP enabling countries. The other FDI is what we call non-risky, okay? So the offshore indicator is the first two parts. All right, so the basic idea that I talked about at the beginning is that if the offshore indicator is high in one country, then the reported profits in that country on this FDI are gonna be deflated so lower in some way. So let's take a first look at the data, the offshore indicator on the horizontal axis. We see that the relationship on the global scale is negative, so that is kind of a confirmation of the anticipated answer to the first question. But there are lots of differences between countries, right? To put them all in one graph is easy and nice to see, but it doesn't really tell you some real numbers, okay? So that's why we include in the estimation of this relationship, we have to include some country fixed effects and year fixed effects and also income group fixed effects. I'm gonna talk about that more in the later part. All right, so let's move to the second question. So how much corporate profit is actually shifted and from which countries? Now, to estimate that, we estimate what we call the profitability gap, right? So it's the difference in the rate of return on FDI that we would expect and the actual reported rate of return on that FDI, okay? So that's the profitability gap. We assume that it is due to profit shifting because we assume that investment from Cyprus should be no more profitable than investment from Spain. There is no reason for an investment that is vested in one country to be more profitable based on which country it comes from, okay? So we say that the only reason that the profits could be lower is that the profits were shifted out of that country. All right, so let's look at the results of the estimation of this relationship. So the offshore indicator itself, as I said, is a negative significant predictor of the rate of return. Now, imagine that you're looking at the Czech Republic, for example. So it's gonna be the sum of these three numbers that the offshore indicator itself and then the category, it's a high-income country in the OECD and it's in Europe, okay? So the sum of these three is gonna make up the final estimate of the profitability gap for the Czech Republic. For Mozambique, for example, which is a low-income country in the Sub-Saharan African region, it's gonna be the sum of these three. So in this case, coincidentally, we omitted the base as we started from low-income countries from Sub-Saharan Africa as the base for this estimation. But it's an equivalent. From the profitability gap, it's just one easy step to get to the missing profits, right? So the missing profit due to this profit shifting can be easily derived as the product of the profitability gap and the amount of risky FDI in dollars. So if you're not lost yet, the amount of risky FDI is the product of the offshore indicator times the total FDI in a country. The last step is how much do these countries lose on tax revenue as a result? Now again, the step from the missing profits to the tax revenue results is kind of easy. So we just multiply the missing profit by the effective corporate tax rate that is applied in that country. And what we get are estimates of tax revenue effect of profit shifting based on this method. All right, so the first thing we remark is that there is a lot of missing data, okay? So that's one big issue that we face and that we aim to overcome in the coming years. We hope that the method will be improved by more and more data being available. Now you see that in USD billions, the biggest losers are some developed countries in Western Europe and some less developed countries in Latin America and Mexico. Looking at the numbers as shares of GDP, we see two columns. The first one is for the rate of return. The second one is for the equity component, which we like to view as kind of a lower bound and upper bound of our estimates, okay? So there are two estimates of the final effect and we say that it's somehow in that interval, okay? So you see the biggest losers on this table will have to look in the future in specificity on countries that lose the most like Saint Martin. It's supposed to lose 6% of their GDP, which seems quite high and we're gonna be looking into reasons why it is so. Coincidentally Mozambique is second. Falls well into this conference. All right, so let me zoom in to Europe for a second, which has much better data coverage and if we look at numbers in USD and US dollars, you see that the biggest losers are the developed countries of Europe, right? But if I invert this as a share of GDP, we're looking at a whole different map with lower income countries in Europe losing the most, okay? So that's another observation from the results that we made. What we achieved is additional support for previous relatively high numbers of corporate profit shifting. The tax revenue losses in total in the whole world resulting from this method are in the order of $200 billion, so it's a lot of money and as I said, the lower income countries in the relative terms, so as a share of GDP or their tax revenues lose much more, okay, than the developed countries. Now at the same time for these lower income countries, it's much harder to fight back because there's this big multinational and lower income countries are not in the position to fight back on these practices. All right, so my final slide is talking about the drawbacks. As you see, there is a lot of them and I'm gonna divide those into two categories, the ones that we think that might be solvable and those that are not solvable because of the method, because of the nature of the method. All right, so the first one that we will do is to derive shares of corporate taxes because now with the new GRD, the new addition, our estimates can be linked because our estimates are for 2015, so we could not link them to tax revenues because data was not available up until a couple of weeks ago. So that's the first step that we're gonna do to improve the results. Now the limited coverage, as I said, maybe it will improve in the future. We might look into some additional data sources that would improve the coverage. Now another problem is that the methodology still partly relies on a 01 definition of tax havens. So if we say that there are 40 tax havens that we classify as tax havens, then this is an arbitrary decision at all times. All right, so maybe we'll look into use of some spectral measures of being a tax haven like the financial secrecy index is gonna be probably the best choice in this case. Also we're gonna look at some unique features of some countries as you saw Belgium was for some reason read. And we might be looking also into including tax treaties data as an explanatory variable. Now what is not solvable is that these estimates only include those profit shifting practices that require a foreign direct investment link. Okay, so if there are two entities that are not linked in the Dayton foreign direct investment, our results do not cover that, okay? So what this means is that our estimates can be taken as a lower bound, okay? Some kind of a subset of all the profit shifting that is taking place. The second one is that it's an inherently imperfect method. We talked about this before the presentations that the only thing we know about our estimates is that they are wrong, okay? But it's just to present some kind of an order of magnitude of how much of this is actually happening. And another one is that we're not sure whether to use in the last step of the methodology whether to use effective or nominal corporate tax rates because we can never be sure that if there was more foreign direct investment in the country what the effective tax rate would be because this is an indigenous decision of the state which can decrease taxes for some investment opportunities and so on. And so on. And so on. And so on. And so on.