• Review of VAT zero rate of food items
Articles:

Review of VAT zero rate of food items

18 June 2018

In the February 2018 National Budget Speech the Minister of Finance announced a VAT rate increase from 14 to 15 per cent that took effect on 01 April 2018. The increase is expected to raise an additional R22.9 billion in 2018/19. The increase’s potential impact on poor and low-income households, lead to concerns. The Minister of Finance, through the Davis Tax Committee (DTC), appointed a panel of experts to consider and review the list of zero rated food items. The terms of reference include three specific areas to consider, namely evaluating the current zero rated food items; considering the inclusion of additional food items, and considering mitigating measures. The panel will evaluate the current 19 zero rated food items and assess whether they provide efficient and sufficient relief to the poor. The panel will also consider whether the policy objective underlying zero rating may be better achieved through disaggregation of those items to more specific targeting of products. The panel will consider food items that are not currently zero rated that may be considered for zero rating that will provide relief to poor and low-income households taking into account absolute and proportional benefit to low-income households; market structure and likelihood of producers passing benefit to customers; ease of administration; potential for abuse, and estimated revenue loss. The panel will also explore whether the effect of zero rating can better be achieved by a government expenditure programme; whether a government expenditure programme is more efficient in targeting poor and lower income households than the zero rating of food items; and whether specific current government programmes as determined by or agreed with National Treasury, can be better tailored to achieve the same or a better outcome than the zero rating of food items.

The Margo Commission (1986) recommended replacing the General Sales Tax (GST) with a Comprehensive Business Tax (CBT) or, as a secondary alternative, an invoice-credit VAT system. Following Margo, government released a White Paper (1988) indicating its intention to implement a VAT system. VAT was introduced on 30 September 1991 at 10%. The VAT Act, 1991, was promulgated on 12 June 1991 with no provisions to zero rate merit goods or services (but for exports). On 17 July 1991, it was decided to zero rate brown bread and maize meal. On 29 September 1991, eight additional basic food items were added to the zero rated list. Public protests resulted in additional nine basic food items becoming zero rated on 7 April 1993. The VAT rate was increased to 14% on the same day. Zero rating of food items was introduced to provide relief to low-income households. Pressures to subject luxury goods to a higher VAT rate resulted in the extension of ad valorem excise duties on luxury goods such as air conditioners, cell phones, video cameras, jet skis, and domestic dishwashers. Motor vehicles were later also subjected to a progressive ad valorem formula, which excluded lower end vehicles.

The VAT system uses two mechanisms to “free” the supply side from VAT, namely exemption, and zero rating. Exemptions are generally applied to “difficult-to-tax” supplies, and do not allow input tax deductions to the supplier. The VAT system exempts non-fee related financial services; educational services supplied by approved educational institutions; residential rental accommodation, and public road and rail transport. The VAT system generally applies the zero rate principle to “remove” VAT from supplies, by “freeing” the supply from VAT and allowing input tax deductions to the supplier. The zero rate is generally applied to give effect to the destination principle of taxation (taxing domestic consumption and freeing foreign consumption), and to remove VAT from meritorious supplies (such as supplies consumed by low-income consumers, as is the case in South Africa). The latter is not generally not regarded as a design-purpose of VAT, and is not supported theoretically. The VAT system zero rates exports and international transport services; certain government grants; farming inputs; Illuminating paraffin (since 2001); goods subject to the fuel levy (petrol and diesel); sales of going concerns; and 19 basic food items. The VAT Act zero rates 19 basic food items, namely brown bread; maize meal; samp; mealie rice; dried mealies; dried beans; lentils; pilchards / sardinella in tins; eggs; rice; vegetables; fruit; vegetable oil; milk; cultured milk; milk powder; dairy powder blend; edible legumes and pulses of leguminous plants; and brown wheaten meal.

The DTC (rightfully so) argues that zero-rating is a blunt instrument to assist the poor but also believes that in practice it would be very difficult to do away with all zero-ratings. The DTC recommended that only those zero ratings that clearly benefit the poor should be retained, and the others be done away with. VATCOM (February 1991) recommended that exemptions, zero-ratings and exceptions should be kept to a minimum to ensure simple and low administration for vendors and SARS. Assistance to low income earners should preferably be done outside of the tax system. Treasury published a VAT paper (2007) arguing strongly that an economic case exists to remove all zero-ratings (other than on exports) with the exception of maize meal and mealie rice. Treasury cautioned that such a step should be directly linked to the implementation of direct fiscal measures to offset the impact on the poor.

VAT expenditures (VAT revenue foregone due to zero rates or exemptions) are estimated at R56 billion (2015/16). VAT on food items (zero rated) makes up approximately R23 billion thereof, or roughly 41 per cent (equating to approximately a 1% increase in the VAT rate). The spending patterns of the very poor and poor show that they spend approximately 40% of their income on basic foods (add paraffin) that appear on the zero rated list, compared to the 11% of high and very high income earners. Of the 40% spent, approximately half is spent on only two products, namely mealie meal and brown bread, affording them almost half of the food zero rate relief. Also. approximately 80% of the VAT “saving” received by the poor and very poor through zero rating, is afforded them through their spent on only seven products of a list of 19. It seems wasteful that government would afford the zero rate to all 19 (and more if all fruit are included) under these circumstances. It would arguably be more sound to apply an 80:20 principle, and only zero rate the top seven products which provide 80% of the benefit. This argument could actually also favour only zero rating the top two products, as this still results in about half of the benefit. See Table 1 below.

Table 1: Zero rated spending weights as % of total household expenditure by all groups


Notes:

  1. Only “brown wheaten meal” actually zero rated. Figures did not allow such analysis, figure for bread flour used. Estimated that only 5% - 10% of all retail flour sales for brown wheaten meal.
  2. Assumed relative expenditure patterns remained constant since National Treasury study.
  3. Adjusted data for CPI per annum since National Treasury study.
  4. Further adjustment to data to adjust to VAT zero rated expenditure on food items and paraffin (R23.329 billion) as presented in Budget 2018 Summary.

Source: Own data manipulation, basis used: National Treasury, The VAT Treatment of Merit Goods and Services, 15 October 2007.

Taking the analysis further, and considering the actual Rand expenditure highlight an even more interesting result. The two top relief-bringing-products bring relief to the very poor and poor of R3.3 billion, compared to total relief of R7 billion by all 19 products. This equates to almost half of the relief to the poor. The total “relief” (or “saving”) that the high and very high income earners receive through the zero rating is approximately R13 billion.

In other words, it costs government R23.3 billion (including the middle income earners as well) to bring relief of R7 billion to the poor and very poor (spending R3 to bring relief of R1). This seems non-sensible. Even worse, the top seven relief-bringing-products afford relief of about R5.7 billion (of the R7 billion) to the poor, or some 80% of the total relief afforded them through the list of 19 items. The high and very high income earners “saves” almost R9.3 billion through the zero rating of the top seven products. Add to that, the R2.4 billion “saving” to middle income earners, this amount increases to R11.7 billion. This illustrates that government brings relief of R5.7 billion to the poor and very poor through the zero rating of the top seven products, but at a cost of R11.7 billion, again seemingly not a very economical decision.

South African studies, on the other hand, show that direct poverty relief is less costly and more effective. A study by Jansen and Calitz (Considering the effectiveness of VAT zero-rating as a pro-poor policy – a case study of South Africa, May 2016:20) infers that R1 allocated to direct targeted social investment can result in R0.55 benefit. This far outweighs the R1 spent through zero rating that potentially only result in R0.3 benefit.

Table 2: VAT zero rating “savings” accruing to different expenditure groups (in Rand Million)


Notes:

  1. Only “brown wheaten meal” actually zero rated. Figures did not allow such analysis, figure for bread flour used. Estimated that only 5% - 10% of all retail flour sales for brown wheaten meal.
  2. Assumed relative expenditure patterns remained constant since National Treasury study.
  3. Adjusted data for CPI per annum since National Treasury study.
  4. Further adjustment to data (from R20.515 billion) to adjust to VAT zero rated expenditure on food items and paraffin (R23.329 billion) as presented in Budget 2018 Summary.

Source: Own data manipulation, basis used: National Treasury, The VAT Treatment of Merit Goods and Services, 15 October 2007.

The analysis demonstrates in simple terms the following. Assuming the revenue foregone (by government) as a result of zero rating (food and paraffin) is R21 billion (close enough and easier to demonstrate). Thus, R21 billion brings relief of R7 billion to the poor, and R14 billion to the middle income earners and the rich. It therefore costs government R21 billion to bring relief of R7 billion if one excludes the cost of administration by SARS and compliance costs of taxpayers (to administer their zero rated supplies) (see Table 3, left-hand side). This means that R14 billion effectively represents the cost of providing the relief, over and above direct cost and cost of compliance. Standard rating food and paraffin would result in relief to the poor of R14 billion (after taking into account that the poor will now pay VAT of R7 billion), before taking into account cost of administering a direct relief programme. This leaves R7 billion to administer a direct relief programme, or one third of the money made available by removing the zero rate. The current zero rating system costs R3 to deliver R1 relief. If government could deliver the direct relief programme for less, i.e. not spend the full R7 billion to bring relief, it would make financially sense to abolish these zero rates. That, of course, does not take into account policy consideration such as those discussed above, and the infrastructure available to government to provide direct relief.

Table 3: Comparing relief through VAT zero ratings and Direct Relief

Hopefully the VAT panel and the Minister will consider the economic merits of using the VAT system to alleviate the circumstances of the poor, consider the cost of using the VAT system, and consider the alternatives available. It is, however, concerning that the VAT panel’s report is expected to be done and dusted in a month or so. This will not do this very important topic justice and may have many undesired consequences.

Read more BDO Insights