 maennawna gwiriaia, kautآ teitya kama tūpa kulaidinaid? kızım pabrij pallibai Darvet Pehysek sorry melaidande Niko Bene is also working at Sars and we were looking at tax expenditures. Like the previous speaker, Nadina's mentioned that we've actually got a very broad system with limited deductions, but one of our large deductions that we still have is the incentive to actually provide for your retirement when you retire one day. So what we know is that tax systems are actually there to raise revenue efficiently. But there's also socioeconomic objectives that are driven by objectives in the tax system. And tax expenditures are usually used to achieve such goals. And so it is just like an incentive that is actually coming through the tax system and not directly on the budget expenditure. And then in South Africa, we only publish a tax expenditure statement, not a complete tax expenditure budget. And if you calculate the cost of that tax expenditures that we do cost, it's about 4.5% of our GDP. And the personal income tax system is actually the largest amount and constitute about 57% of this total tax expenditures. And if you look at the two main tax expenditures that we have, our tax exemptions are very little and it's mainly on interest income and it's also CAD. So it's Retirement Contribution Deductions and Medical Tax Credits. And the Retirement Contributions constitute about 69% of the value of that tax expenditure in 2021. So the focus of this research paper is actually then to look at the personal income tax expenditure and do we see if it's that we know that in the initial design of the personal income tax system, we always had a provision for retirement contributions but it was like a split system that only allowed deductions like for formal registered pension funds and not provident funds. And so there was a lot of, I think, anomalies within the system and this was then changed so that we actually have a much more broader system where if you provide for your retirement that it will be deductible against your income. And then we also know that we've had actually quite a lot of distributional gains when we were converting our medical expenses to a tax credit in 2012 and we want to know if we do a similar adjustment to our contribution to retirement funds, whether we can actually also then have any possible gains if we are going to convert our deduction in the personal income system. Then so we analysed the personal income tax system with the focus on the retirement contribution deductions and we converted in the retirement contributions into a tax credit and for two reasons we also looked at what will be the revenue gains and what will be the distributional equity implications. This is just what the current deductions are allowed in the system and you can see is it's the lesser of 350,000 rand per annum or 27.5% of your gross income. So it's a really very generous incentive to actually encourage formal workers who belongs to retirement funds to actually provide for their old age. We've used the PITMOD micro simulation model that has been also funded by UNWIDE to do this simulation that we did and the latest date that we had administrative data available was in 2019-20 and we also then looked at taxpayers below the minimum threshold and above the maximum threshold because you will see if you do anything on tax deductions or taxable income changes so we could pick up that dynamic impact that's happened but it's a static model so there are no behavioural changes. On the development of the model is really a very big data set that we're working on. We've got about 15 million records and with about 2,000 variables in this database and so we were really fortunate to be able to actually do the simulations and then to see what is the feedback in the data. I think there was also very good quality control that was done in the data. We used three data sets to actually compile this data and I think that is the difference between the data that Nadine used and what we used so we used the IRP-5 data that's coming from employers mainly on employee income and then we used the ITR-12 that is the declarations that were done by taxpayers and we then also use the assessed data to actually then to have one complete database as accurate as we can. I think once you work with actual data you do realise that they are like a lot of anomalies in the data that you actually have to work with to make sure that your data is as accurate as you could get it. So when we just calculate this is what is the cost of the tax expenditure. It's about 275 billion that is deducted by about 7 million taxpayers and that is about 10.1% of the total taxable income declared by South African taxpayers and it's close to 48% of taxpayers with taxable income that contribute towards retirement funds. If we calculated what does the tax expenditure cost it's about 92 billion or 16.9% of total final tax liability. Because of the concentration of those that are actually claiming pension contributions the values are actually the average values are really low. Remember I told you that you can deduct up to 350,000. What we've seen is that the average value contributions is about 39,000 per annum and if we look at the median it's actually only about 21,000 and that is just because of this heavy concentration that we see of workers in the low-medal and middle income groups contributing to pension contributions. Then you can as we expected we can see as though that earn above one and a half million per annum the average contribution was about 175,000 per annum and we know that if you're in a system where you can deduct the marginal tax rates to matter because then you are taxed at the 45% rate so the value of that deduction is much more valuable than say for instance if you deduct if you deduct at the 26 or 31% or 36% rate in the simulations that we that we have done for this exercise. So on the reform scenarios that we did is we firstly abolished the retirement contribution deductions so that we could see actually what happens to the taxable income and the tax liability and so you actually then determine where are the concentration and how is the distribution of the retirement contributions in your tax system and then for the second scenario we actually did three conversions so the conversion rate is just saying is that we're using the marginal tax rate that you are actually going to benefit from the rate the credit rate that you're going to benefit if we actually now going to do a retirement contribution credit system instead of a deduction system so the first one that we did is that if we want no extra revenue that we don't want to mobilise more revenue the revenue neutral rate will be then 35% at the 35% conversion rate we will we will not have additional revenue and then we took the 26% as a lower marginal tax rate because that is actually where the average deductions are per annum and we did a second one at the 31% so that will be the third income bracket just to see is what happens to the guinea coefficient in the system and to the distribution of taxable income and tax liability. So if we if we do the scenario one we we actually then eliminate the tax expenditure like i've said it's about 275 billion that we that the taxable income actually increases from from 2.7 trillion to about 3 trillion and then you will also see that the taxable income of taxpayers in the 500 000 to 750 000 income groups are simulated to increase the most namely by 102 103 billion or 2.1 percentage points and the increase in taxable income is equal to 37% of the total income in in taxable income so you can see how concentrated these deductions are and and where the impact will be most if we are actually going to eliminate the tax incentive you know for retirement contribution the deduction for retired contributions so the the graph on the left is just showing you the distribution of taxpayers and then the change in that distribution to the base case and then it on the right hand side you'll just see the distribution of the taxable income the changes to taxable income and then in the first scenario where we do the revenue neutral scenario where we do the conversion rate at at 35 oh this is the tax liability sorry so it's still the baseline case with the reform scenario one so what happens with tax liability and once again you can see the 500 to 750 000 the most revenue will actually come about 24 billion from from that income group and the second highest will for those will be above 1.5 million at taxable income rendering almost the same with 23 billion and then there's like you know minor tax relief that are given you know to to lower lower income groups below 150 000 so it's it's very minor there but you can really see where because this is actually just showing you what will the impact on tax liability if we don't have deductions if we don't have the tax incentive then the simulation if we now do the revenue neutral situation where we actually got the conversion rate at 35 percent then we see this is that most increase in average taxable income is simulated to be for taxpayers earning more than 1 million taxable income per annum and that's also like high income earners if you look at the distribution of income in South Africa so there are tax at the marginal rate of 41 percent then it's increasing by 2.3 percentage points and then for those above 1.5 million per annum it's taxed at 45 percent marginal rate that increases 2.7 percentage points and then you will see is this once again where we actually see if we do the 35 percent conversion rate and those that will benefit with a lower tax liability not just because currently their marginal rate are less than less than 35 percent so they will benefit with a higher credit deduction relative to what they had when they were doing a deduction at the marginal rate then similar if we do a 26 percent now taxpayers less taxpayers will actually benefit because the conversion rate is smaller at 26 percent so now we also our increase in our revenue mobilisation is that we will actually almost 23 billion or increase of 4.2 percent in our tax taxable income our tax liability and you will also now see is this is that the number of taxpayers that will benefit will be less because it will only be those taxpayers that have got a marginal rate of less than 20 percent to 26 percent and on average we see that there's a the tax liability reduces for taxpayers with taxable income less than 500 000 per annum so what we're just saying is is that once again you know we simulate that the tax liability increases for those in the higher higher income groups and what we see with the guineco efficient for final tax liability increasing by 0.3 percentage points so the system do become more progressive so similarly if we do a 31 percent marginal tax rate so that will be a little bit more taxpayers in the lower income groups will be benefiting from this because there's more taxpayers with marginal tax rates less less than 31 percent so now we will actually our yield and income more tax income will only be 9.5 billion and because more taxpayers are now paying less there's a less advantage for lower income taxpayers and we also say is that the total number of individuals with a final tax liability reduces by 180 300 or 2.5 percent and in the average tax liability once again for taxpayers less than 500 000 per annum is reduced and the system is also slightly slightly more more progressive um so what is the main findings um that we have if we um so the graph on the right hand side is just showing the effective tax rate and and what happens at the different income groups for the for the for the scenario so this first scenario um just having um when we we don't have the that's the bar graph that just shows us is what happens um with the effective tax rate and the percentage point difference uh if we don't have a deduction anymore and then we've got the three scenarios at at the different conversion rates um so i think this is where our main findings is is that to incentivise the provision for old age uh by allowing by allowing a tax deduction is costly i mean 200 and 70 92 billion in tax liability 17 percent of our total tax liability is quite high and it's like the our chairperson mentioning our high unemployment rate um our low growth uh the problem structural problems that we have in the economy so we actually have a very generous tax incentive for old age provision of old age and that's also quite important aim for government to have because we don't want people when they retire to be dependent on on government for for social grants but we also have to take into account how how we actually distribute this incentive and if we if we know what is the pattern of are we your most taxpayers are actually providing for for old age and at what rates and and where they are concentrated it seems that the 350 000 or the 27 and a half percent might be too generous and in the context of of of the data that we're looking at and then um we know that because of a tax deduction system is that high income and is with higher marginal tax rate actually benefits more and then um i think is is that it was quite surprising to actually calculate the the revenue neutral rate uh 35 percent that is actually quite also a high high conversion rate to use and then just switching the to the 26 percent will yield the the most effective revenue if we if we're looking at uh you know raising revenue out of the system uh with about 23 billion uh it's quite a sizable amount revenue um that we could actually raise uh and to actually distribute uh directly through other means or finance government or reduce the debt um and then yes like i saying is if we do a change in the conversion rate at only 31 percent and then we will yield much more revenue but i think the impact will will then also be less um then what we will have at the at the 26 percent you know for middle and middle high and higher income earners um so in our main findings we say if we want to look at this policy reform option and to mobilise tax revenue then consider 26 percent conversion rate and that will protect the the low income earners from an increase in tax liability and reduce the tax liability of low income earners contributing uh to pension funds so it shouldn't alter uh any behavioural response from those that that are actually currently providing and and and you know and that's low and low middle income uh taxpayers um so the way forward is this we thought this is a little bit more analysis to refine the impact analysis on low income earners contributing to pension funds because we do see the movements of taxpayers at the different rate conversion rates across the minimum tax threshold and then um we also want to combine it with the SA mod that will simulate what we can do in the tax benefit system uh with with uh if we've got actually increased the social assistance grants um by this increase in tax revenue and then I think lastly I think we need to consider any changes to what we currently provide uh the in in the deduction system um you know the cap at 350 000 is relatively high and also the percentage deduction of 27.5 percent and that's all from my side thank you you