SA-TIED Seminar Rethinking corporate tax and what the global minimum tax means for South Africa

chat and with that over to you Pierre. >> Thank you. Thank you very much for inviting me to the essay seminar series. First let me say a word of how great I think this series and and what has been done at the treasury has been in South Africa. It's really an example that I site often now when I travel to other countries meet with other partners of you know what you can do when data is open and yet very secure and you create a whole a whole program. So we've been lucky enough that you know South Africa is part of this study thanks to thanks to the the Treasury and the data lab. But this is a wider study that had the goal to inform especially in light of the global minimum tax that has that has now been implemented by some countries. What is likely to happen not so much for rich countries, not so much for tax heavens, we often talk about these two two groups, but for the actually the majority of countries that are neither the rich countries with a lot of multinational headquarters nor the tax heavens that have benefited immensely from uh from profit shifting. And so this is joint work with my colleague at the World Bank and Brock Meer, with Royal Dom who's now at Bugal the Think Tank, and with Kam Se who's finishing her PhD in Munich. So as you I'm sure know the corporate income tax remains a key source of tax revenue. About 20% of total revenue comes from corporations. Yet it's increasingly and it's been now for a while let's say the past 20 years under pressure and there's at least the perception but the data is starting to biting that large firms increasingly pay less tax than they used to. What has been discussed a lot is that international tax avoidance and profit shifting has allowed multinationals with operation across many countries to really take advantage of tax differential and in addition there's been competition between governments trying to react to this behavior and to attract firms and so a race to the bottom in the statutory tax rates that used to be in many countries above 30% you know two decades ago and in a majority of countries now around 20 between 20 and 25%. In addition and harder to study has been effects on the base, right? So the rates have been going down but at the same time the tax base has typically been narrowed. Yeah. Many tax incentives which can be tax credits, income exemptions, specific reduced rates, special economic zones and the like. And so we study this in the context of the partial implementation of the 15% global minimum tax since 2024. South Africa is has signed the the full OECD package. Uh and if I understand well, I'd be happy to to to to learn more in the discussion. It's becoming really effective in 2026. We have three research questions in this paper. The first is to what extent do tax incentives actually lower effective tax rates of firms below the statutory tax rate? How do effective tax rates vary with firm size? And is it indeed the case that larger firms end up paying less taxes than medium-sized firms? And then how much revenue might a minimum tax raise? And then we're going to talk about kind of a broad minimum tax and then look at the specific rules uh that apply to the global minimum tax. So this paper really does three things. It's first going to leverage full population corporate income tax data from 16 middle low and middle inome countries. We we span quite a spectrum of economic development. I would say you know from the richest country might be Greece in the data and the poorest uh maybe Uganda or Senica. We're going to measure for each firm the effective tax rates in a way that can be calculated that is relatively consistent and comparable across countries. Right? So we're going to be able also to make some crosscountry comparison. We're then going to look at how these effective tax rate vary as a function of firm size along the distribution. Are there some generalizable patterns and which types of tax incentives might explain the differences in effective tax rate and then we'll estimate the revenue potential of the 15% minimum tax. And let me give you a preview of some of the key results. We'll go in detail. First, there's a very clear effective tax rate firm size relationship which displays an inverted U-shaped pattern. Right? So the smaller firms typically pay less taxes. Tax rates increase for the midsize and upper midsize firms up to about the top dile and then at the top start falling and actually drop quite a bit at the very top of the firm size distribution. That means that the top 1% firm pays typically 2.5% less effective tax rates than another firm within the top 10%. Right? So we're comparing the very large firms to the upper upper middle large firms. We'll see that there's different explanation for why the effective tax rate is lower for small and larger firms. Second, you know, we can quantify what's the fiscal cost of corporate tax incentives. It's large. It's about 1% of GDP on average from the micro data. You know, so in some countries, it matches very well the tax gap analysis that have been done. In others uh we have some differences and you know we could discuss why uh this is really fully micro database and there's also some limitation. I'll show you how we comput. But maybe more importantly for the debate on the global minimum tax we notice quite a consistent feature that in most countries and all these countries have tax rates that are always above 15% statutoily typically around 25%. Yet a quarter of the largest firms are currently paying less than 15%. Okay, so that kind of gives you an idea of what is a potential taxable base of low tax profits uh in in the countries we study. Yet we're going to see that despite this, the expected revenue from a global minimum tax, consistent minimum tax, and I'll bring you the acronym that people use, QDMT, which you might have seen, qualified domestic minimum tax. It's actually a lot smaller than what the numbers above make you think. This is due to a much more limited scope, which firms it will apply to, is due to some deductions called carveouts. It's also due to consolidation at the group, right? And so we'll see and we'll try to decompose uh a little bit some of these effect. So that's it for the for the preview. A quick word on the on the literature. So there's a big literature by now on tax minimization strategies often focused on tax evasion and avoidance. We're going to look let's say at the third uh element which is legal tax incentives actually provided by the government but that's a big reason why in practice effective tax rate can be can be low >> certainly in 2025 between the US >> I'm sorry there's a yeah there's some noise in the background yeah thank you um the question is we have some evidence that tax evation typically decreases with firm size in particular because there's a whole set of third party information that can be collected by larger firms, think banks, think suppliers and clients and so on. So it's much easier for smaller firms to hide their income. However, tax avoidance tend to increases as bigger firms have access to more sophisticated technology might be multinational and able to to shift profits around right the relation between tax incentive and firm size have been relatively unexplored. I think that's one of the big contribution showing that that's another reason why tax rates are really not uniform uh across firms. And finally, what does it mean for corporate tax collection? Well, despite where I said the corporate income tax is becoming seemingly less important because of all these uh feature to the law in many countries, a lot of the revenue is still reliant on a few firms, both the other corporate tax and because these large firms retain taxes on many other agents. And so it's very important to understand what what would happen in the context of a minimum tax. And so we have this revenue estimation of the impact of minimum tax. So let me move now to the to to the core. Uh first a few words on data and methodology. So we have the full corporate tax return for 16 countries. We have actually a good representation in Africa. You can see Ethiopia, Ronda, Sagal, South Africa and Uganda. and then a very good representation of Latin America uh including big countries such as Colombia and Mexico and then two countries in Europe the Albania and Greece. So you might wonder why this set of countries it's not random. It's partly the countries where we were working and then countries either that had data labs like South Africa uh or that were very interested in collaboration because they wanted their own estimation. That was the case of Costa Rica or Jamaica for example. Okay. So you notice there's no country for example in Asia that do have a lot of tax incentives too. that might be a reflection both of the limited data openness in these countries or maybe the less interest to to to consider um the global minimum tax. Now just uh you know to to to mention I mean several countries have since actually reached out and so while we're not going to update the paper the analysis is being done after about 18 or 19. This group of country doesn't have a lot in common except one feature. There's no countries that people would typically consider as tax heavens. What would it what would a corporate tax heaven be? It mean that the statutory tax rate itself would be below the global minimum tax uh of 15%. Right? So the lowest tax rate country in this case is Albania with a 15% tax. But actually, and I'll show you the rates later, almost all countries have rates that are quite above the 15% minimum tax, right? So the the median I think is 25 or 26 in. Okay. So what's an effective tax rate? And here there's not one definition, right? But first we can think of what's the concept we like to approximate. The numerator is relatively easy. We want the CIT liability. That is the taxes that are owed this year, right? And here it's important because in the tax return sometime you don't see exactly liability. You see the taxes that need to be remitted or owed after some withholding or the deductions have been have been already applied or advanced payments. For example, here we really want the total corporate income tax liability corresponding to a given fiscal year. And then we want to divide this by a measure that should approximate economic profit. Ideally, we would have true economic profit. That's not really something that exists. And you know, we have to know that both accounting data or tax data have biases in what's being exactly reported. What we do in practice is we try to approximate it with something that has the flavor of the EBT that is a standard accounting concept to which we remove dividend income. Right? Right. So the way to think of this is we're going to take total sales and remove total cost and the dividend income. So this is a concept of profit that you know applies before some of the key elements for example income exemptions or special allowances and deduction. And so therefore it's going to be able to capture in the effective tax rate. It lowers the role that all these special features of the of the tax system uh imply for for for the tax rate. Right. It's important that basically in the back of our mind when we're doing this, we have the idea that there's kind of a benchmark corporate tax that's the same that people do when they do a tax gap analysis, right? Which would basically be to say there's a standard statutory tax rate that should apply to everyone. There's a standard set of allowed deductions including financial deductions. In this case, for example, you can deduct interest, but anything that's kind of a special allowance or deduction on an exemption on income that should that is not in in a classical tax system. This is what's counted as a tax incentive, right? And this lowers your tax rate. Of course, you know, two reasonable people could could debate of what's the right benchmark, but most countries actually apply relatively similar definition to to to to their standard corporate income tax and the deduct deduction that should be allowed. So here I repeat myself. So net profit is total revenue minus let's call them standard deductible cost which would be material labor operational normal depreciation interest and the likes. Note that it's already somewhat of a conservative measure because someone could say well actually there's some special provision linked to what interest can be deducted and so on. Right? So we think already by going this if anything we may be underestimating somewhat the importance that um tax incentives and tax expenditure could have. This is what it looks like. Now you know this is a little bit boring but so I'll go fast over this. So we have total revenue minus total cost and then that allows us to get this approximation of economic profit or losses for some firm. Now this is different than the taxable profit because the taxable profit is going to remove exempt income. It's going to allow for some tax allowances for example some investment incentives, some accelerated depreciation or other special deductions. There's something important though that I need to mention which is in our calculation it's very hard to remove systematically loss carry force that is losses that have happened in previous periods and that are being brought forward. So this is one important limitation is that what I'm going to call tax incentives typically has built into it the loss carry forwards because we couldn't do it in a systematic way across all countries without and you know if you're interested we can discuss a little bit more in the chat why um but I'll show you that for the main results especially for the patterns this is not key right so it matters a little bit for the levels of total tax expenditure but not for the patterns actually then what's the next type of tax incentive a firm can get well it's a reduced rate right most firm will have this type or normal rate. But then some firms are going to be applied a reduced rate. We'll count that as another type of tax incentives. That's going to generate right the taxable profit time the tax rate the gross tax liability of the firm to which now the firm can also apply some special tax credits. So there's some investment credits, some export promotion credits and so on. Now once we've applied those credits, we end up with a net tax liability which is a numerator. And so I repeat, the effective tax rate would be this net tax liability divided by profit. Anything that looks more like a double tax payment, for example, foreign tax credits, right? These are taxes supposedly you've already paid in in other countries are not going to be counted in a net tax liability, right? And that's important because depending on the tax form they might appear in different different places, right? So our goal is to again be relatively conservative in in our approach. The second measure we're going to show you systematically is firm size. Now firm size people a bit might have different opinion what's the best way to measure it. The way we could do it in every country because it's always reported and in a standard way is just to use total revenue. In the paper we show alternative for example to total assets that's available in about twothirds of the country of payroll really the results are not don't hinge much on this right anytime I'm going to show you firm size I'm never going to make a comparison of firm size across countries it's always going to be within countries I'm going to look at the largest firms that operate within a country and we're going to rank firms as a function of their size within their country right so when you're going to look at the top you're going to look at the top 1% firms for example in South Africa relative to other South African firms. I'm not going to I can't make a direct comparison of these firms to those say in Uganda. I'm going to go fast on this. It's in the paper. We have validate this or try to validate with the treasury uh that this was the best approximation we could do from the variables in the in the in the data lab um from the treasury. But I would be very happy again to get feedback on if there would be some way to to potentially improve So what does it look? The first thing we can do is just say let's now plot rank within rank within each country the firm by size. The gray area in every case is going to be those top 1% largest firms where now I've zoomed on the top 1%. Right? So the rest are basically percentiles going from 1 to 99. And then I wanted to zoom in on the top 1% firms. Why is that? Because they're so important for the economy and for total tax collection. They often account for about 50% of total tax collection. So what's happening the rate that applies to this firm is reduced. The second thing I want you to notice in this graph I know there's a lot of information is that we've plotted the statutory tax rate that applies in each of these countries right and so you can see what I mentioned the lowest tax country is Albania at 15%. And then you can see that basically all the other countries are fairly high tax rate countries right so starting at you know 25 then Guatemala and 25 applied to a lot of countries and then I've ranked countries by the statuto tax rate all the way to 30 what is that 34 for col now this figure is a little bit specific because I've allowed first just to show you that the firms that have zero or losses remain in this figure so they have not been removed Now, it's a bit awkward because when you have zero profits or losses, you don't really have an effective tax rate. You're paying zero taxes this year. And so, I've assigned I've imputed, if you will, an effective tax rate of zero. Right? So, in here, you have also firms with zero tax rate that are participating in this in this block. We can see though a very quite clear pattern that is actually applying to most countries that is smaller firms pay lower tax rate. You can see that the effective tax rate rises almost everywhere and then typically peaks not everywhere but in many country it peaks somewhere around the 10th or 15th percentile you know so let's look at Costa Rica where it's very evident for example here Costa Rica small firms are paying very low rate it increases peaks at not too far from the statutory tax rate maybe 25% or so and then starts decreasing very clearly for uh for larger we're going to come back to South Africa uh in a minute. Now you may say it's a bit awkward. You've left you've left in there the firms with effective tax rate with a with a losses or of or zero tax liability. What about if you remove those firms? Okay, so now I'm not changing my x-axis. I still have all the firms, but any firm that doesn't pay taxes in a given year, I remove them. And so if you want, there's no more zero. There's only firms actually paying taxes with a well- definfined effective tax rate. So what happens? You can see that in quite a few countries now the relationship at the bottom at the left part of the distribution has flattened maybe not the best example. So let me let me go back and find another one. So let's look at Guatemala here. Guatemala had quite a lot tax rate for the small firms but now when I look only at the profitable firm you can see that this has become basically flat. Okay. So that tells us that basically a lot of smaller firms are lossmaking and so that's why you know when I have the zeros in there a lot of the small firms are not paying taxes because they don't report any profits that can be because of tax evasion or truly because smaller firms you know they're discovering themselves they're making investment and so on they're still in a growing phase. Okay the pattern that stays more consistent though even here is that typically at the top you still have a large drop in the effective tax rate. Right? So we've kind of flattened the curve. Now we can do one more step of controls which is to say many countries including South Africa actually have special regimes for smaller firms. Often the tax rate might be a function of their sales of their profits of a mix of the two of their assets. So what happens if you try to control for statutory tax rate preferential regimes dedicated to small and medium firms and you can see now let's look back at Costa Rica which we've studied once that was what will happen in Costa Rica by low smaller firms were having much lower rates and so once you control for the rates you see that actually small firms now don't have a gap relative to their own statutory tax rate yet you can still see a very large gap at the top relative to the statuto tax rate of the large corporation >> transfer and obligation So what about South Africa? We can do the same. I'm already going to start with directly the profitable firms in South Africa. And I'm going to plot the statutory tax rate. I think the year we use if it's 19. I think the rate was 28. I think it might have come down to 27 now. And so this is the effective tax rate already for profitable firms. But I told you we could control. South Africa has a special regime for small and medium enterprises if we you know which so so it looks like this and so if we control for this now we can do the gap between the effective tax rate and the statuto tax rate we've basically erased it for about 80% of the distribution but then you can see that there's now still a big gap that appears for the top firms right for the top actually about 20% of firms in South Africa and that is pretty large within the top 1% So you might ask you know there's a few things you do how robust it is to controlling for sectors to computing this over multiple years. You've talked about losses. So I want to show you a few things. So because it's very hard to do it to show you graph for every country. I've created here basically a synthetic country. What does it mean? You take I've taken the average at every quantile every percentile of all the countries in the day. Okay. And now you see this pattern quite clearly that when you take this synthetic country this average of all the countries in the sample smaller firms are paying more around 20% in in effective tax rate this rises for this let's call them upper middleiz firms you know around the P80 P90 of the top distribution so those those are firms that are not small but they're not the largest either this rises to about 23% and then you can see it decreases quite clearly to the point that the very largest firm in the sample are paying even a bit less than the small firm. Okay. Now once we've done this it's easy to do some decomposition for example to look at how it varies across sectors I've separated in the four key sectors if you want primary which is agriculture secondary which would have the industry uh retail and service right and you can see I think two two patterns are interesting first the general pattern is quite consistently there but especially the big drop in effective tax rate at the top of the distribution is very pronounced for industry Right? Probably won't surprise you. It's part of the special economic zones and other things to try to attract capital. It's also though quite present in a sector like services, right? So you can see a fair drop in service. What about if we build the effective tax rate a bit differently? Instead of building it on one year, one year of corporate tax return, we actually take the average across something like five years. That means I'm going to take the total tax liability paid over five years over the total profits and losses accumulated over five years. Know that when I once I've done this, I've really unless a firm is constantly making losses, I've really kind of averaged out the role of of losses, right? And of um of loss carry forward and the likes. So what you see when we do this is that basically now that explains a lot of also the reason why smaller firms are paying lower tax rates and um and it's in part because they're also claiming some of these loss carry forward. We flattened the bottom of the distribution but we have not done much at the top right you can see that the top this drop in percentage relative to the the P90 is very small okay so loss car for matter for small firms in explaining why they also facing lower lower tax rate uh but not so much at the top uh what about South Africa I just wanted to show you the same graphs what's happening in South Africa the drop at the top you can see is quite clear in the secondary sector maybe a little bit in services too not so much in retail and and in the agricultural sector similarly in South Africa if you build the effective tax rate over multiple year of data you see that while it raises a bit the levels it really doesn't change anything to the pattern right these two terms now you might wonder you know is some of these due to reporting at least at the top of the distribution so here one way to say that we think this measure is pretty good is that if you look at something like net profit over assets and now we're going to basically make the assumption that assets is a verifiable relatively well-reported variable. You can see that this ratio is very constant across firms, right? So it's not that there's some funny reporting going on when you look at the very big firms relative to smaller firms that maybe are not reporting as well. So let me move now to try to explain a little bit uh what's going on at the very top of the firm distribution. what tax incentives explain this drop especially we're going to focus on a very simple model just say decompose the effective tax rate with a dummy for the top 1% firm right so I've put P99 these are the 1% largest firms in your country and to make this interesting because I really want to focus on the top I'm only going to keep firms that are in the top 10% anyway so really I want to make a comparison within large firms but what explained that drop at the very top in that gray zone when I was showing you the And then what we're going to do is just add controls one by one in this model from different tax return variable that could be explaining or mediating this effective tax rate firm size relationship. Very concretely, I'm going to add a dummy for if there's exempt income, if there's a reduced rate that applies to this firm, if they have applied for tax credits, if they apply for loss carry fors, and then some firm characteristics that we can observe in systematic ways such as sectors and geography. So this is the table. Let me a little bit summarize it for you. So we start with the baseline that dummy of the one top 1% average across countries which is minus 2.2, right? So the top 1% firms relative to the other firms in the top B size, so relative to also large firms are paying 2.2% less effective tax rate. You can see that in 14 of the 16 countries in the data, this is a negative coefficient, right? I've averaged it, but it's negative in 14. And here is the inter intal range across countries, right? So at the P25, it's -4.2. At the P75, it's -0. The first thing I'm going to do is put a set of dummies for firm characteristics. They're quite simple. The geography and sectors I'm going to show you later discuss a bit geography matters because in some tax returns we cannot directly observe special economic zones and we think this is capturing that right and so you can see that once we've put a dummy for geography and sectors there's quite a big drop in this dummy for the top 1% right. So now if you control for this there's only a preferential advantage of top firms of 1.46 percentage points. The next column is going to put a special damage for tax allowances, special deductions. You see that this drops a little bit the coefficient but not by so much. Then we have reduce rate. This almost does very little, right? And that's normal. It's rare that the very largest firms have reduced rate except if they're part of special economic zones. Then we have a dummy for uh oh I think this has been reversed. Sorry that's a mistake because it's the tax credit dummy that is oh no no sorry I'm reading I'm reading my own graph wrong with those little crosses. Sorry. So it's the tax credit dummy that is doing a lot of the lifting. Right. You can see in column six um where now the coefficient has dropped to minus 1.54 and finally and something important I've mentioned you what happens when you put the loss carryover dummy you can see that it basically doesn't change very much the baseline contriution right so loss carryover matter for the total level of tax expenditure you're measuring but they don't mediate the relationship between firm side and tax expenditure in the final model I can throw in all of those dummies together right to see How much we explain? We explain most of this difference not not it fully in particular because you know our model is a bit simple. These are just dummies and not continuous measures and so on. But you can see that once you've accounted for all these type of tax incentives or characteristics you you've mediated most of these gaps. So let me summarize maybe in in in a few words what we seeing here. Top firms in 14 out of 16 country have an rate advantage. I've shown you the gap for the advantage of the top 1% but actually if you focus on the top 0.1% even larger from the the gap becomes even larger and closer to 3 percentage point and then which types of tax expenditure might explain this gap tax credits matter a lot about a third of the advantage of larger firms and then firm characteristics account for another third and in some countries where we can see this it's basically mediating for special economic zones that offer a lot of tax advantages but often are not direct not always directly observable on the T. Now I want to put two caveats which might come back in the discussion. First, there's a limit to observability of tax incentive especially if you want to do this in a totally systematic and comparable way across country from the tax returns. The reason is that many incentives on the tax return >> many incentives on the tax return they come in with maybe in one place of the tax return and they come in all together to actually observe them and observe the rationale and so on. You will need the annexes that we have only in a handful of countries they much you know sometime they don't even digitize or fully digitize this data. You also will need details on the law because often on these annexes what you will have is you know we're claiming low number 1534. Okay. So what does that correspond to what concept and right so it's very hard to do this in a comparable ways across many countries and so that's why we didn't go I've already discussed it's also difficult to remove the impact of past losses. So what about South Africa? South Africa we can do the same type of decomposition. Uh the gap between the top 1% and top the rest of top 10% is minus0.9% in South Africa. So not as large. But there's something I want to show you for South Africa which is this is really a bit of an underestimate because if you go back to South Africa here South Africa compared to other countries the drop in effective tax rate starts much earlier in the distribution. So in other countries we saw that you know this gap was pretty flat up to the P90 and then started to drop in South Africa start dropping already from the P8 right? And so that means that in the comparison I'm doing in this table, I'm really comparing, you know, firms here at the P90 or P95 to firm at the P99.9. And so you see that's why this gap is actually mediated. The true gap and we could we do this exercise for South Africa to be more accurate should be to take firms maybe at the P80 and look at look at what's happening to top firms. Okay. Now we can still ask you know what about the very largest firms what's happening in South Africa. What's explaining the fact that they have an advantage? What's quite significant is special economic zones in South Africa where I know there's a reduced rate I think of 15%. Um we can see that exempt income is also going to explain a little bit but not as much actually of of the variation. But then special deductions, special allowances are really what's explaining most, right? And once we've accounted for them, we can see that this dummy on the top 1% actually if anything is now positive, right? So if you look at how the coefficients changing here, the two things that are really making it drop is special economic zones and special detach. Uh and then you know in the paper we've listed some of the main incentives. I think this is not by size but by the number of firms claiming it. uh and you know you some of you probably know this a lot better than me but these are the key um provisions um that that that explain this gap in the in the effective tax rate. So I have 10 minutes now to try to cover maybe a little bit less uh the global minimum tax. So first we can do a very simple extract now that will compute the effective tax rate for all firms and including for top firms. How many of these firms are paying a low rate? Now what does it mean to pay a low rate? depends on each country but the global minion tax asset low rate at 15%. And so we can see across country what share of the top 1% firms are paying less than 15%. So that's the figure to the left and I think really this is quite a staggering figure because you can see that maybe with expression exception of Albania none of these countries the low tax rate country right all these country have statutory tax rate of 25 27 28 30%. Yet in all countries you have at least 10% and very often 20 or even 30% of the largest firm that currently face an effective tax rate below 15%. So that gives you an idea of a potential pool of under tax profits. Of course these are all firms not just subsidiaries or multinationals as long as they part of the top 1%. And so we can do a very simple exercise which is to say imagine you actually apply a domestic minimum tax that just says all firms all large firms you have a threshold that applies to the top 1%. That are currently paying less than 50% have to pay a minimum of 50%. This is not a totally crazy policy. For example, Colombia has something very similar to this and already implement and implemented it last. How much revenue could you collect? So the figure to the right is showing you the increase in the corporate income tax relative to baseline that this simple domestic minimum tax policy could bring. And you can see right here we have a bit more variance. It depends on you know on how many firms uh and on on the stock of profit. But you can see that in a lot of countries this is not trivial. We're talking about substantial revenue if you were to tax this low tax profit. Where is South Africa? South Africa is about at the median, maybe a little bit below with potentially about 10% higher corporate income tax collection if you were to apply such a broad-based domestic minimum. Now, in practice, the global minimum tax is very different, right? So, it's been called a milestone in the regulation of globalization, a corporate tax revolution by the financial time. It's an agreement of about 140 countries for 15% minimum tax that would applies to large multinationals, right? So there's a threshold of global sales that the multinational has to to reach and then the tax rate would apply to all its subsidiaries. Who has implemented it? most high- income countries with the exception of the United States and a set of low and middle inome countries which includes South Africa which includes Brazil and then other countries considering it has a set of quite smart connection rules that let's you know we can review quite quickly and then I'll come to our exercise right and the idea of this rule is that you have an ultimate parent entity that might control an intermediate parent entity that itself has a set of affiliates that might be in low tax places and other in high tax places. Low tax and high tax in this case is determined by the effective tax rate that the subsidiary is paying in the country. So anytime it's going to pay less than 15% in a country, it's going to be considered a low tax appeal. Now there's three set of rules in the global minimum tax. The first one is this qualified domestic minimum tax. It's saying the country where the low tax affiliate subsidiary is is located is allowed basically a first bite of the apple. Concretely can apply and decide to tax this firm up to 15%. And then this voids any other taxes or claims from other countries. Step one and that's what we're going to model. It's important to know that there's two other set of rules. One is the income inclusion rule where the parent company of the subsidiary is headquartered is located and the parent company if there's no QDMT is allowed to say ah actually I'm going to collect the missing profit from the headquarter country. Now what about if the headquarter country doesn't do it then there's a third rule which is that countries where other affiliates all the subsidiaries are located which are already overt taxed if you want tax more than 15%. they're now allowed to go and claim the under tax profit in the low tax countries to a proportion of how much of the worldwide sales they have of of the okay so this means that really countries if if this whole mechanism is well applied most countries should choose to enforce the QDMT because this way they're the first to tax otherwise their subsidiaries in their countries are going to be taxed anyways but either by the headquarter country or by countries which already taxed uh the same multinational subsidiaries that are higher. Okay. So in terms of data, what are we going to do and I'm going to accelerate here a little bit. We're going to do two things in in a subset of countries where we could do this. Something very important is we need to identify which firms are affiliate of multinationals and of those largest multinationals. We're going to ask a set of partner tax administration to match the data set with Orbeez. Orbeez is this private data set but that is pretty good for having ownership links especially for large multinational headquarter in the United States or in Europe and so we from that we're going to be able to observe in the corporating contact in each country which firms seem to be subsidiaries of large multinationals. The other approach now we're going to get an imperfect merge when we do this and so that means that we're probably getting a lower bound of the number of firms in scope. The second method we can do is actually say we know from the country bycountry report how many firms are supposed to be in scope in each country right so we know that there might be 100 firms in what country so what we're going to do is just use an indicator for if a firm is a multinational in a country when those exist and just say we'll take the 100 top firms top foreignown firms as to match the number of firms that appear in the country by country this is likely to yield an upper bound Now because we've taken the top firms, the ones with the highest top of tax, right? So we're basically getting two methods. One that gives us a lower band and another an upper band. So let me go to calculations. Once we've identified the firms in scope, right? So that's going to limit the pool of firms. We then going to calculate the tax base under what is called globe, the rule of the global minimum tax, which allow for some quite generous deduction. you're allowed to deduct 10% of your payroll and 8% of your tangible assets. Okay? So the tax base is going to shrink quite a bit. So when we say it's a 15% minimum tax, it's not really true. It's a 15% minimum tax if you have no substance, no payroll, no tangible assets. In practice, it's going to shrink the the the top of tax. The other thing that's important is one needs to consolidate firms within a group. Okay? Because this applies at the group level within a country. So we've done this for five countries. So here are the five countries including South Africa. And so first how many firms become liable at the top you have this uh the first method uh the the the upper bound method at the at the bottom for each country you have the lower bound method and you can see that you know it's not a whole lot of firms we're talking about South Africa which is the largest country among these among these you know it's about 130 or 140 firms to which we estimate this would apply let me move on here and show you the composition uh and I'm going move directly to the South African kids. Okay. So, we're going to move scenario by scenario and say sorry I need to remember the the legend. So, okay, I'll do it for Costa Rica and then go to to to South Africa. So, first what happens when we only considered at the group level we've excluded small firms in Costa Rica? We estimate that the corporate income tax could rise by 24%. Now let's apply those deductions I just told you about. Then we lower this to about 21%. In practice, those deductions actually are going to become less generous over time. There's a phasing out of those deductions. 10 years out, we estimate about a 20% increase in corporate income tax. And now at the bottom, we have two counterfactual scenarios that we think that are interesting. The first one is what happens if all your current tax credits become what's called refundable. So there's a provision in the global min that says if if a tax credit is refundable, meaning you're allowed to go and the tax administration owes you money in the end, then actually doesn't count in the calculation. So if Costa Rica was to transform it tax credit to refundable, you see that basically the global min would go to zero. The last counterfactual scenario is one where we look at what was initially proposed as part of the global minimum tax that was a rate of 20% and immediately the year 10 carve outs that are less generous and you see that that would make actually a difference then you would collect around 30% in Costa Rica. So what about the same type of scenarios in South Africa? So we started 1.6%. So it's not huge the corporate income tax top up you could get once we apply the deductions for payroll and assets. You see that we've have this and we have less than a percent. 10 year out as deductions become a bit less generous we would go back to slightly above a percent. Mechanically if tax credits become refundable then there's no there's no gains for South Africa and under a 20% rate this would raise to about 2%. Okay. So relatively modest uh gains in South Africa because actually the number of firms in scope is not so large. So let me let me just open this because I think some of these questions you might have in the discussion. We've documented large economywide gaps between effective and statutory tax rate. Tax incentives allow large firms to lower the effective tax rate quite a bit lower than that of side firms. And the global minimum tax only collects a modest fraction of potential revenue especially when you compare it to a broader minimum. There's still many unknowns in what we've done. The first is the implementation of the global beon tax. Which countries will adopt? What about the US that is negotiating a sidebyside deal that would partly protect US multinationals? If the global million tax works well, it's also going to shape deeply firms reporting production and maybe even location decision especially out of tax 7. So that could be another type of gains to tax revenue. So that's on the global tax. On the tax incentives and efficiency front, you know, there's a question, why are tax incentives highest for the largest firms? Is this partially about take up that small and medium firms are not as aware maybe and finally you know there might be in a world where this is quite justified that the relationship between firm size and tax incentives can be explained by efficiency cost but I think many countries really lack good research on that I think South Africa is actually getting some research on that and that's great and I think you know the the size of the tax advantage really mean that we have some of the most powerful actors paying a lot less taxes than other everyday firms you know and So we need to be sure that this is just I'll stop here and open for questions. Thank you. >> Thank you Pierre. Um we've now moved on to the we're going to move to the Q&A segment of the seminar. I know that we're joined by an who is one of Pier's coers. Um to ask any questions you're welcome to type questions in the chat or you can raise your virtual hand. I see that we already have one question in the chat here and it is from Gossini Lamini. He asks, "Is there a risk that tightening effective tax rates at the top could discourage investment particularly in sectors where incentives are currently concentrated? If there any other questions, please raise your hands. I don't see any hands at the moment. Yeah, I don't know if you'd like to take that or should defer it to an and do you want to answer this or >> happy for you to go ahead. >> Uh so this is a very good question. I think that was our last point and especially in a world that sometime you know industrial policies kind of making a comeback and so on you could say you know it makes sense to protect our largest champions right and so that would be with lower tax rates. I think though you know we've documented that this is really systematic across countries yet very few countries do this type of real estimation of are our tax incentives justified and there's something a bit perverse because a lot of countries are also saying well we want to encourage monomial firms to grow and so on and yet you can see that initially the profile of tax rate is typically increasing over most of the size distribution right and so you know these are complex problem because there's issues of fairness, there's issue of competition between firms and so on. And so there's a world where this can be justified, but I think in most cases we don't have actually enough evidence to to to to say so. And I think that's really an encouragement for more studies to go and say, you know, does it make sense that it is the largest firm benefiting or would you be better off spending this money on something else for startups of of, you know, creating an ecosystem of actually younger firms that eventually will grow. Oh, we have one more. Um, okay. We have one more from. He asked, "Given that the estimated revenue gains from the global minimum tax are relatively modest for South Africa, um, where should policy attention shift if the goal is to strengthen revenue mobilization? So this is a very good question. You know, we were a little bit surprised because a lot of estimations are were larger for the gains. Um so it's important maybe to put some caveats. The gains we measure are the mechanical gains, right? They're literally applying the higher tax rate. Some more positive scenarios and analysis say by the OECD or the international tax observatory. What they have tried to do is say okay but if these rules are really well applied actually a lot of the profit shifting motives putting your profits in low tax countries are going to diminish right and so that could mean that you know a corlary of this is that there's more reported profits in higher tax countries which what we've seen South Africa is mainly a higher tax country even though it has some pockets of low tax firms right so these indirect effects they're very hard as you can imagine to to estimate you But there is a world where they are not trivial and at least you know for South Africa they might double you know to three 4% the corporate income tax revenue right so so that means that applying this rule is not you know it's not just about the mechanical effect it's also about participating in this kind of global public good that countries have actually taken a decade to try to implement to limit this profit shifting and I think in that sense it does make sense for a large country like South Africa to to participate But I think what we're showing is we should not expect you know miracle increase in in in tax collection uh coming from this. And so to your question where does revenue mobilization comes from? Well, it might come from really rethinking some of the tax incentives. If they're overly generous and they're not reaching their purpose and South Africa has many as we've seen, that's maybe where there's an easier an easier way to to to increase revenue mobilization. But as your other question said, you know, these have to be estimated because some might be very efficient and others not as much. >> Thanks, Pier. And before we take another one of the questions, are there any questions from the floor? >> Okay, not um the last one Pier is the analysis uses. >> There's a question by Hey, I think she missed her. Oh, okay. Haley, I didn't see your hand. Sorry. Please go ahead. >> No problem. Um, thanks very much for the presentation, Pier, and for the work. Um, it's certainly been very interesting for us. Maybe just a couple of comments. I have quite a lot of it's a it's a very interesting topic, and I could talk at length, but maybe just a few points. Um, on the incentives question, maybe let me start by saying I'm not a fan of tax incentives. I don't necessarily think that they change firms tax behavior. And from a South African perspective, we are trying to evaluate uh their success etc. and see if they're aligned with government objectives and see whether they whether they should stay in place. I also think it's unfortunate with the recent change in design of the GMT that effectively, as you mentioned, one of the measures that um that reduces revenue potential is refundable tax credits. Um, and it's unfortunate that substance-based tax incentives and the way that it's been watered down in effect almost removes the disincentive for countries to introduce tax incentives. I think that's quite unfortunate. Um, a question and and so one of the things I guess is similar to one of the questions you raised in terms of the take up on incentives. The question for me at least is does that mean that medium firms don't take up incentives and is it a question of tax incentives being designed specifically for large firms? We had an incentive in place like that in South Africa that that was removed because of that reason from a fairness perspective and competitive reasons. um or is it that there should be better information so all firms can take them up and then a a quest it was also interesting to see how the sectors are different so for example the secondary sector seems to have the bigger downturn in ETRs for larger firms and I guess that part of the GMT design where there's a lot of substance then they then they're not liable for as much topup tax um and then a question for interest on the domestic minimum tax and you mentioned that Colombia IA is a country that that has one of these. I'm interested from a policy perspective how likely you think it is that that countries will adopt such a measure given that they're obviously concerned about competitiveness etc. Um and some countries might there might be a view that tax incentives are useful tool for attracting investment. How likely do you think it is that there would be a wide ranging take up of something like a domestic minimum tax? Uh thanks very much. Sorry for the long intervention. >> Thank you Haley. And you know first a big thank for Haley that has helped us has helped us a lot in this paper. You've talked to us a few times and that has been very valuable and something we really appreciate also by the Satite program the fact that we've been able to talk to to to people that came. So uh I mean I think it's a great intervention. You know better than us. I think you know these issues need to take up. We know that from some of the literature in in in high income countries actually take cap is an issue for small and medium firms and if anything some of the papers we have seen show that condition on take up actually small and medium firms often benefit more from incentives and then that's actually where they're quite efficient right so to that point of efficiency if you could target it well to these firms communicate them well maybe that's where you have the bigger bang for the back um I think you your analysis in South Africa is right you know It's the fact that it's going mainly to firms that actually have substance mean that there's not a whole lot of potential for the the global minimum tax. And on the issue of firm size, we looked a bit at the special economic zones and actually there's a fair distribution of sight within the special economic zones. Right? So while this is a big tax incentive, it's not one of the most sight dependent in South Africa. And I was a bit surprised when seeing this number because in a lot of other countries it's almost exclusively very large firms that I know. So that might be you know I mean I don't know enough about that context but I know there's some papers uh from essay tit on this that I should read and there might be more context and now you know to your question what's likely to happen uh with those refundable credits the evolution of the sidebyside deal I think it's less likely than a few years ago that countries would want to do domestic minimum taxes especially the kind of broader ones that Colombia has implemented there might still be a case I think to start doing more tax incent let's put it this way more industrial policy maybe a bit outside the tax system you know and then we might want to think what what that means right but sometimes the tax system is being used because it's an easy way really to connect to firms it's often the the best administration in many countries and so that's the way to reach firms but it's really not clear that lowering the tax liability is the way to achieve goals if they're linked to say employment or capital formation right and so um on this I invite you my colleagues at the World Bank have just published a very big report on industrial policy that you might have seen it's generating some noise because I know the the Financial Times and others have said you know it's really changing 30 years of of the Washington consensus but you might find that interesting we are left with four minutes um I think we've given a bit of air time so I'm going to take a question from Enoch he has two we should have enough time to cover those. Um, the first is which specific tax incentives are the biggest culprits for the U-shaped tax burden and should they be scraped entirely or crap? I guess yeah, scraped. Then the second one pier is if South Africa implements a domestic minimum tax that is more aggressive than the GMT, do they risk losing foreign direct investment to the neighbors who stick strictly to the 15% OECD standard? Then I think we'll cap it at that and >> and do you want to take those? [clears throat] >> Sorry, I was just typing a response to Kosinaki's third question. Um so I'll >> okay so so so I'll try to answer not question so so so thank you for the question that's one you know we've we've received a few times and I think that's the limit of what we've done right we've covered many countries and find a good balance but I think in the end of course when you're trying to reform those you have to go very deep in the legislation of each specific country right we've seen that something that matters are the tax credits in particular so I would say within the cate categorization is often the one that the large firms are most able to to to take use of. Often tax credits are linked to specific investment behavior, right? So you might think this is part of kind of a negotiation for the firm to set up in a country and so on. So I think that ties back very clearly to this efficiency question. Uh and so it's very hard to answer your question. Should they, you know, should they be scrapped or I think I'd prefer would you say it be capped? You know, there should be a limit and there should be a lot of transparency. there should be evaluation, right? I think that's really all the things that matter that currently few countries have, right? So, we've lived in a world of kind of lack of transparency on everything linked to tax incentives. Um, and so I think hopefully that is changing slowly and uh now I think you are right going more aggressive means some risk for for FBI that is that is certain. I think that was the idea of negotiating a global deal that everyone would apply, right? It's really a coordination problem that was partially uh resolved. But as we seen, you know, especially in this a bit more erratic world that we are currently in, it seems that some bits of this agreement are are you know won't be applied. I still think it's a big step forward and I'm really happy that countries you know that South Africa are choosing to implement because we can there's many reason to criticize the OECD global minimum tax too but it I think it's a key step forward not so much in what it's going to do for revenues with but in the logic that you can actually have a coordinated policy uh with some global rules and we can imagine later I don't know in five or 10 years if the kind of global cycle is more prone again to cooperation that this would be strengthened in a more significant