 Thank you, and now we're coming back to the corporate part of tax havens I will talk about the effects of country-by-country reporting and whether it might even increase profit shifting So there's a joint project with my PhD student Ribi Dullerman and Dirk Schindler from Rotterdam So I don't have to say much about the introduction because Casper nicely did most of that for us and So as I said we return to the use of multinational corporations using tax havens We know that a relatively large share of multinational corporations profits are held in tax havens And there has been some policy changes in the last couple years, the global minimum taxation And efforts to increase transparency The main effort here is the implementation of country-by-country reporting Which has been in place now for a couple years since 2016 Quick reminder, what is the country-by-country reporting? What does it do? It provides tax authorities in one country with information on a firm's activities in other countries So let me give you an example Let's think of a US multinational which has three foreign affiliates One here in Norway, one in Nigeria and one in Ireland The Norwegian tax authorities pre-country-by-country reporting were only able to see the profits and the employees, the tax payment in Norway They had no idea what was going on in the other countries With country-by-country reporting, the Norwegian tax authorities can now see that this firm has a ton of profits in Ireland And the profits in Ireland might be an indication that this firm is very actively shifting profits to Ireland And that it might be worthwhile for the tax authorities to target this firm and audit it more closely to figure out what's going on How is it shifting the profits to Ireland This is in place, it's private at the moment, so only the tax authorities get the information And it applies in not all but many countries worldwide to all multinational firms with a global turnover of above 750 million euros What do we want to show in this paper? We want to figure out a bit more about the mechanism how country-by-country reporting might work against profit shifting As I told you in the example, country-by-country reporting helps the tax authorities to identify these firms that are shifting profits very actively So they can target them in audits increasing the cost of profit shifting That's the idea behind it, how it should work However, it doesn't have a second effect When the tax authority sees the Norwegian tax authority sees the high profits in Ireland, it does not see whether these profits originated in Norway Let's see, it's a medium profitability firm in Norway, it has a ton of profits in Ireland The profits in Ireland might also come from Nigeria, right? The tax authority doesn't know that So what country-by-country reporting now changes is that how much profit is shifted from Nigeria to Ireland changes the likelihood of an audit for the Norwegian firm So the profit shifting decisions of the individual affiliates now get interlinked That's the mechanism that I will explain to you in a bit more detail in a moment What we do in the paper, long paper, it has a theory model and presents some empirical evidence I now only want to give you a rough overview of what we do I'll be very brief about the previous literature There's some empirical literature, I think the takeaway message from this slide is that the findings are maybe disappointing to those who have hoped that CBC reporting will be a game changer. There's mixed evidence, nobody has found strong effects There's little real evidence of country-by-country reporting working against profit shifting And we think we have an answer to why that's the case, why the empirical literature has not really found a clear picture Let me explain the mechanism we show in the model in a bit more detail So what we model is a multinational firm with three affiliates A high-tax affiliate, in the example I gave you think of the Nigerian affiliate tax rate, let's say 30%, maybe 35% High-tax country, an affiliate that's really producing something There's a second affiliate that's also really producing something, we call it the low-tax affiliate Maybe we should call it the middle-tax affiliate, this is again in a country where the firm is really active, it's really producing something But does country with lower tax rates think of like 15% to 20% the Norwegian affiliate, in my example And then there's a third affiliate in a tax haven, might be a zero tax haven like the Cayman Islands, might be something like Ireland And the assumption here is that no production is taking place in the tax haven, it's just there to save taxes What this multinational firm can do is that it can shift the profits that it produces by the real production in the two blue affiliates, light blue affiliates It can shift these profits to the tax haven affiliate If it shifts profit it incurs two types of costs The first type of cost for those of you who are somewhat familiar with this literature, that's the standard profit shifting cost We call it tax planning cost, that's the ex-ante cost of shifting profits to the tax haven You need to set up some structures, we heard about the fixed cost of profit shifting You need to transfer patents, you need to hire lawyers, consultants to set up the possibility to shift profits from any of these affiliates to the tax haven And these costs increase somewhat in the amount of profit shifted But they are affiliate specific, so shifting profit regarding the tax planning cost it doesn't When you decide whether to shift more profits from the high tax affiliate to the tax haven, you don't care what the low tax affiliate is doing What we add in the model is that we add a second type of cost that we call audit costs We think of these as the ex-post cost of profit shifting So what could this be? Well sometimes you do things that are in a, let's say, legally grey area, illicit area You may be audited, the audit is costly because tax authorities come to interrupt your business operations You have to provide them with information It might go to court, there might be a settlement, there might be fines to pay The probability of all this happening goes up the more profits you shift to the tax haven, the more you're targeted by the tax authorities These are the audit costs The audit costs is what changes when country by country reporting is introduced With country by country reporting you're more likely to be targeted if you're part of an affiliate that has lots of money in a tax haven So now suddenly for the low tax affiliate, let's say for the Norwegian affiliate, it matters what the other affiliate is doing If this affiliate is shifting a lot of profit to the tax haven, the audit probability for the low tax affiliate goes up too And vice versa, right? The audit probability for the high tax affiliate also goes up if the low tax affiliate is shifting a lot of profit to the tax haven So what it means is this country by country reporting, the transparency over the firm means that the profit shifting decisions from the individual affiliates get interlinked Now that changes the profit shifting decision of the multinational as a whole, so in the model we assume the multinational firm maximizes total after tax profits And what do we find when we introduce country by country reporting? Cost of profit shifting go up because the audit costs go up So costs go up, so there's less profit shifting, the total amount of profit shift that decreases, the tax base and the tax haven decreases Like we would ideally want to see it In this low middle tax country, profits increase, there's less shifting out of this country That is a clear fact in the model, however the effect for the high tax country is ambiguous Why is that happening? There are two effects that work, I think one is the straightforward direct effect Profit shifting decreases because country by country reporting implies increased overall costs But now there's also a substitution effect And that substitution effect arises because of this interlinking of profit shifting costs The audit probability of the Norwegian affiliate goes up because the Nigerian affiliate is shifting a lot of profits out of Nigeria to tax havens And with the substitution, we call it a substitution effect because from the perspective of the multinational as a whole It now becomes more attractive to shift, if you do any profit shifting you should do it from the country with the higher tax rate So the substitution effect means my overall profit shifting goes down, but the remaining profit shifting I do from the high tax affiliate Why do I do that? It gives me the bigger bang for the buck, the tax rate differential is higher Shifting any profit to the tax haven increases my audit costs, so if I shift something I shift it from the high tax country That means the result of country by country reporting, it helps the middle tax countries, it does not help so much for the high tax countries And that resolves some of the mixed empirical findings that previous literature has shown us Let me just go through some of that For example the previous literature finds very small effects on increasing total tax payments or consolidated effective tax rates As this literature usually uses always which over represents relatively high tax countries where we expect this ambiguous effect We also see in the literature that country by country reporting seems to increase investments in European countries that have preferential tax regimes So which seems to be pretty close to the low tax country in the model because they do offer, they want real production but they do offer low tax rates And the literature shows that investments there tend to go up However none of these previous papers have actually looked at high and low middle tax countries separately So in the second part of the paper we do exactly that So we also use the Orbis data set that has been mentioned by Jakob It's balance sheet data at the affiliate level and we do a difference in difference analysis that means we compare the profitability of firms before and after the introduction of country by country reporting And as a control group we use affiliates of those multinational firms that are not subject to country by country reporting Either because they fall under the size threshold or because they are part of a multinational group that's not in one of the countries that introduce country by country reporting The model gives us three hypothesis The first one is that profits and tax havens should decrease which we can't really test because Orbis has notoriously bad coverage in tax havens So we should not focus on that So the main test is what happens with the profits in high tax countries and what happens with the profits in the low middle tax countries For the low and middle tax countries the effect should be clear Profit shifting goes down so the profits in these countries should go up for the affiliates affected by country by country reporting For the high tax country the model does not make such a clear prediction It depends on the size of the direct effect and the substitution effect So the profits could go up indicating less profit shifting or they could go down indicating more profit shifting as this profit shifting substitutes the profit shifted from the other country So that's the setup, here are the results So it's the sample split So columns one, three and five are the low tax countries and we find exactly what we expected Profits in these countries go up after the introduction of country by country reporting less profit is shifted to tax havens Great news for the country by country reporting Less great news when you look at the high tax countries So we see consistently negative coefficients here not when we add controls no longer that significant But clearly there's country by country reporting is not a success story in the high tax countries It even seems to indicate that less profit is in the affiliates in the high tax countries are less profitable after the introduction of country by country reporting In line with the idea from the model that more profits are shifted out of these affiliates instead of also shifting out of the low tax affiliates I was asked to think a bit about what the implication of this paper is for low and middle income countries And that's actually a very interesting question Because I think this paper actually does have some interesting things to say given the relatively high corporate tax rates that we have in many low middle income countries So I think the straight message from the model is that we should not expect country by country reporting to be a major positive development for developing countries It's given the relatively high tax rates they're likely to be in the high tax group and unlikely to profit from the country by country reporting That's the prediction We do have a data set so we thought well maybe we should just have a look It's the Orbis data set that as Casper has already said has not very good coverage of these countries Actually there is no low income country in our data at all so this is actually a result for middle income countries And almost all of our observations here are in our definition high tax affiliates in high tax countries around 95% of the observations And then we just rerun the main empirical analysis in the sample and we find a substantial negative effect Meaning after the introduction of country by country reporting if it works profit should go up as less profit is shifted to tax havens But what happens is profits actually go down relative to the control group So the country by country reporting does not seem to help at all It's a tiny sample, it's 4,000 observations, it was a quick and dirty analysis for the conference so I would not put too much faith into it But I think it does underline the point that we should not put too much hope into country by country reporting Solving the problems arising from profit shifting from multinational corporations in any countries is relatively high tax rates And it seems to hold in particular for low and middle income countries Look at the size of the coefficient, it's a lot higher than the negative coefficients we got for the high tax countries in the main analysis So if we want to put faith in this very small sample, it's bad news on the effectiveness of country by country reporting Great, so I think I'll use my last 30 seconds to wrap up, thanks for listening I hope to have planted an idea in your mind that is country by country reporting might not be as positive as it first seems That it might actually have negative effects that arise because the fact that the tax authority now sees what is happening in other affiliates Changes the profit shifting incentives and might imply that more profit is shifted out from the affiliates that give you the biggest bang for the buck Those affiliates that face the highest tax rates And that's what the model predicts and which seems to hold pretty well in the empirical analysis Thank you