In a widely circulated commentary for Sygnia investment managers reflecting on South Africa’s recent Budget debacle, Professor of Economics from UCT Haroon Bhorat concluded that: “VAT is not regressive and is the least of the three tax hike evils that the Treasury has available to it!” (his exclamation).
From media reports, it appears a similar message is being communicated within the Cabinet, where there is now a push for a smaller VAT increase.
It is worth subjecting Bhorat’s arguments to scrutiny as they echo those of the National Treasury which is ultimately responsible — it appears from the outcome of the recent Cabinet process — for making this decision.
The evidence base for this conclusion is weak and we should not take such claims at face value.
First, Bhorat correctly reports World Bank initiated research that showed VAT to be slightly progressive in South Africa. This means that the poor spend a slightly lower share of their income on VAT than the rich, in large part thanks to the zero-rating of items consumed heavily by the poor.
This is correct, although the difference between the proportion of disposable income spent on VAT by the lowest-earning 50% (just below 10%) and highest-earning 50% (just above 10%) is not large.
Zero-rating limit
Zero-rating as a means of shielding the poor, while accounting for this limited slight progressivity, also appears to have reached a limit, with the National Treasury reporting that additional zero-rating implemented in 2018/19 was inadequately passed on to consumers.
What arguments that VAT is mildly progressive fail to acknowledge is that a VAT increase would make the tax mix in South Africa more regressive (or less progressive). That is, a VAT increase would increase the share of taxes paid by the very poorest in our society, and reduce the proportion paid by the highest earners.
This is self-evidently true, as everyone pays some VAT, whereas only the highest-earning 30% or so pay any personal income tax.
Second, Bhorat relies on the findings of the Davis Tax Commission to argue that household welfare — really household expenditure — will be least negatively affected by a VAT increase, compared with a personal income tax or corporate income tax increase.
Going back to the Davis Tax Commission VAT Report one finds that this pseudo-scientific “finding” comes from the commission asking the National Treasury to model the impact of tax increases across the three main tax types.
In a break from accepted good practice, the Davis Tax Commission Report provides no information whatsoever on the model, assumptions, and methodology used by the National Treasury. After a little digging and triangulation of similar findings from the National Treasury, one can fairly reliably assume that these results were generated using its standard Computable General Equilibrium macroeconomic model.
The rub here is that the reason the National Treasury finds that VAT is the least harmful to household expenditure is because the model assumes that it will be less harmful in the first place. That is, the model is designed to generate this conclusion because of how the equations account for the impact of personal income tax and corporate income tax.
Only possible outcome
You cannot take seriously a finding that “x will be greater than y” when the finding is generated from a model that is built in a manner in which that is the only possible outcome.
Even if the aggregate finding was robust, the fact that the model fails to distinguish between households of different incomes and expenditures means its usefulness to concrete policy decisions is questionable.
It might be that the pain to households from a personal income tax increase is greater than from VAT, but in the case of a personal income tax rate increase, this pain is borne disproportionately by higher-income households. As higher-income households tend to save more and spend less on domestically produced goods than lower-income households, the impact on overall growth may be worse from a VAT increase.
Bhorat also conveniently leaves out that the same research found that a VAT increase was more inflationary than the alternatives, and worsened inequality more than the alternatives. In the current environment of disastrous levels of inequality and global inflationary pressures, this should also weigh against a VAT increase (to the extent to which these findings are credible).
Third, those making the argument for a VAT increase are incorrectly juxtaposing it against a blanket personal income tax and corporate income tax rate increase. While returning the corporate income tax rate to at least its previous level of 28% would be a very sensible option, many of us posing alternatives are not necessarily suggesting an across-the-board personal income tax rate increase.
Tax giveaway
For instance, one good option is the removal of retirement fund tax breaks for the highest earners. Removing this tax giveaway from only those earning above R750,000 annually could raise more money than the proposed 2 percentage point VAT increase.
Other suggestions, including drawing from the Gold and Foreign Exchange Contingency Reserve Account or postponing Government Employees Pension Fund contributions, hold no implications for households. The only way to make VAT a preferred option is to artificially narrow the range of options available to us.
Finally, Bhorat’s article repeats what appears to be becoming accepted wisdom — that the initial VAT increase was needed in order to pay for the “extension of the social relief of distress grant (SRD) and an additional 1 percentage point increase in the public sector wage bill from 4.5% to 5.5%”.
Such framing provides cover to the argument that it may be unfortunate to raise VAT, but if it is used to pay for social grants, then it is overall a progressive fiscal policy (as ex-Budget Office chief Michael Sachs has argued).
However, it is not true that tax increases are “needed” for these purposes. After last year’s Medium-Term Budget Policy Statement, the National Treasury itself said that the SRD grant was already funded under a provisional allocation line item, calling into question whether it is now “additional” expenditure that requires “new” funding.
Least progressive tax instrument
But even if that were true, the fact that the National Treasury failed to include agreed-upon policy commitments in the medium-term budget in the first place because it is ideologically opposed to them (the SRD grant and public sector wage bill increase) doesn’t justify the use of our least progressive tax instrument to raise additional revenue.
What it shows is that we need to raise additional revenue because the National Treasury failed to do its job properly the first time around. Under this framing, trying to make the poorest pay for this failure loses much of its lustre.
If the evidence is slim, then we must conclude that arguments in favour of a VAT increase are politically and ideologically motivated based on an economic playbook that seeks to shift the burden of tax to ordinary people.
This is particularly objectionable in an economy as depressed as ours when this “budget gap” (around R32-billion or 1% of the R2.6-trillion planned budget expenditure) can be met in multiple ways that would have virtually no negative impact on the poor or the economy more broadly.
A VAT increase, by contrast, would not only have immediate detrimental effects, but would also be a permanent drain on the people who can least afford it. DM