The Covid pandemic has greatly exacerbated existing poverty and unemployment in South Africa. What are the best policies to alleviate it? Targeted social grants may aim to reach the poorest of the poor but there are many inefficiencies in the targeting method. A Universal Basic Income Guarantee is more costly, yet simpler to implement, will reach those in need more quickly, and is financially feasible.

The Covid Social Relief of Distress (SRD) grant is set to end in March 2022, and with it, the only source of income support for more than 9 million people. The burning question is what, if anything, will replace it? The recent Medium Term Budget Policy Speech (MTBPS) has given us no answer. Instead, Finance Minister Enoch Godonwanga has stated that, while this may be reconsidered by Cabinet, no decision has been taken regarding potential replacements, nor any budget allocation made for its extension currently.

It is untenable to leave this question unanswered. That the Covid SRD grant of a meagre R350 has provided a lifeline for millions is a stark testament to how dire the conditions are that many live in. The Covid pandemic has exacerbated already dangerous levels of food insecurity, poverty, and unemployment in South Africa. To believe these conditions will no longer exist in 2022 is to deny reality. Such denialism led to the grant’s initial termination in April 2021, a move that cannot be delinked from the desperation underlying the subsequent unrest in July, regardless of how its political factors are understood. While job creation and economic growth are key goals, to pretend these kinds of medium- to long-term solutions answer the question of immediate crises, is likewise untenable.

Here, direct cash transfer programmes have been proven to be effective and fast-response measures for poverty alleviation. Yet who should be the beneficiaries? If deciding such a policy, would you choose to assist everyone or only those considered ‘poor’ or ‘extreme poor’? In theory, the obvious answer may seem to be the latter. Yet as the adage goes: in theory, there is no difference between theory and practice, in practice, there is.

This choice speaks directly to the ongoing debate. On one hand there are calls for a Universal Basic Income Guarantee (UBIG) as one possible answer; an unconditional monthly cash payment to each individual aged 18 to 59, with the benefits being clawed back from the wealthy. On the other, there are arguments that a UBIG is unaffordable, inefficient, and has undesirable effects. Instead, many propose the most effective solution would be to target social assistance specifically to those in need, identified by their level of income, unemployment status, or other indicators of poverty through some form of means-testing. Examples of this can be found in the current Covid SRD grant, the World Bank’s recently suggested ‘job-seekers’ grant, or National Treasury’s idea of a ‘family grant’.

On the face of it, the logic of poverty-targeting seems straightforward:

• By restricting eligibility criteria to limit the number of beneficiaries, the policy becomes far more cost-effective.
• Since each Rand spent in transfers goes directly to the poor, its allocative efficiency of expenditure on poverty-reduction is greater.
• Instead of spreading a limited budget too thinly, larger transfer values can be given to a smaller pool of beneficiaries, hence increasing its poverty reduction impact.
• Furthermore, it is argued that linking eligibility to economic status ensures the support is only temporary and falls away when beneficiaries ‘graduate’ out of poverty, which avoids entrenching a culture of dependency.

In comparison, the proposal of universality is often treated as naïve idealism.

Yet here is the hook. We do not argue for a UBIG simply because it is the morally just choice, but because it is the pragmatic and beneficial design choice. This question is explored in greater detail in our recent policy brief on Designing a Basic Income Guarantee (BIG). Here we briefly consider some of the issues.

The common assertion is that universality is simply unaffordable, rendering cheaper targeted options the only feasible alternatives by default. Yet this knee-jerk reaction errs in several respects. Firstly, while constrained, the current fiscal space is not so crudely fixed, with a myriad possible financing measures shown to be able to fund a UBIG.

Secondly, the real cost of these options is not equivalent to the total annual value of their transfers, but to the end cost of their implementation. For example, a proposed UBIG taxation claw-back mechanism (a social security tax) would account for roughly a third of its cost. On the other hand, a family grant would likely carry the highest implementation costs because it requires an entire system South Africa does not possess, namely a national household social registry and a civil service capacitated to identify and monitor up to five million beneficiary households. This is in addition to the multiple reasons civil society organisations have identified that render the family grant inappropriate and unimplementable.

Thirdly, social spending is not wasteful expense. Research suggests a multiplier effect of between 0.7 and 1.9 on GDP for investment in social policy, with positive returns on total tax revenue and employment, and reductions in inequalities. When taking into account the positive knock-on effects, the economic benefits of investment into social protection tends to offset its costs, at least partially. When it comes to policy choices, it is always cheaper to do less. However, the question is whether that is desirable.

But what if the desired purpose can be achieved while spending less? Here, comparative modelling by SALDRU (Southern Africa Labour and Development Research Unit) and ADRS (Applied Development Research Solutions) both confirm that a UBIG will have the largest impact on poverty reduction. However, the SALDRU research shows that the family grant performs nearly as well in closing the extreme poverty gap (a 73% vs 86% reduction), while costing far less. This also speaks to its higher efficiency (percentage of expenditure going to close the absolute poverty gap) compared with the UBIG and other measures. In terms of value-for-money, the SALDRU research finds that a family grant is theoretically the best option to address absolute poverty, particularly if the target is only those living in extreme poverty- below the food poverty line- and not those living in poverty, broadly defined.

So why consider a UBIG? Firstly, because isolating our response to addressing extreme poverty (those who live under the Food Poverty Line) is short-sighted. Moreover, modelling policy in the abstract is simply worlds apart from implementing it in practice.

All the logic of poverty-targeting rests upon one assumption: that poverty can be easily targeted. Yet this is not the case. In researching the efficacy of poverty-targeting in programmes across 23 low- to middle-income countries, Development Pathways proves high rates of exclusion of those supposedly eligible, inclusion of those ineligible, and flaws in the registries supposed to differentiate between them. Brazil’s Bolsa Familiascheme itself, the inspiration for the proposed family grant, carries an exclusion rate of 44% of its intended beneficiaries. Out of the 25 poverty-targeting programmes studied, 12 reflected exclusion errors over 70%, while five reflected above 90%.

The relationship here is clear: the narrower the group the social policy targets, the higher the rate of exclusion of precisely those it targets. This is because it requires an increasingly complex system of identification and means-testing to isolate smaller and more specific groups. This creates a multitude of barriers to access, as seen in the current Covid SRD grant with outdated databases, lack of administrative capacity, technological barriers, and more.

Notably, for policy-makers less disturbed by potentially high exclusion errors (because this means less cost), the issue of inclusion errors that inflate the budget are unavoidable. As pointed out by some of the SALDRU researchers, if we assume a likely 40% error, the cost of a family grant balloons from R59bn to at least R120bn (or R270 bn with a 70% inclusion error).

The related key assumption is that poverty can be easily defined. But poverty is not static. This is true especially in the context of high levels of poverty and highly dynamic incomes found in low- to middle-income countries. Instead, to try to target the poorest or the unemployed is to target a group in constant flux. This issue is exacerbated when targeting a ‘family’ or ‘household’, considering there is no easily useable definition for these in South Africa[xvii]. Instead, research has shown high levels of porosity and fluidity in household formation. In other words, it is a moving target, economically, socially, and geographically.

Furthermore, means-testing takes time, and the family grant requires this in excess. To announce one will reduce extreme poverty by 73% for R58.7bn is to be silent on the years this will take to roll out, and the fact that the likely cost of implementation is estimated to be much higher (up to double and possibly more). Indeed, the proposal of National Treasury (according to a government document leaked to Ground Up is to replace the Covid Grant which currently has 9.5 million beneficiaries, with a family grant restricted to only 1 million families, which it aims to gradually expand. What must those who cannot eat today do until then?

Herein lies the beauty of universality. It does not require a complex system to regulate access. It is much simpler to implement and possible to finance. As currently proposed, it is actually ‘targeted’ towards those in need, and should be seen as an effective delivery mechanism to achieve this end; this is achieved through the allocative use of taxation claw-back mechanisms rather than means-testing. Not only can it significantly address poverty and inequality if progressively financed, but create people-centred development and provide a new right to citizenship premised on socio-economic inclusion. As for fears of a UBIG making people dependent on it, and refusing to work, evidence repeatedly suggests that ‘dependency’ resulting from social grants is little more than a myth.

That the family grant was not announced at the MTBPS is a testament to the last-minute pressure from civil society for common sense to prevail. Yet it remains deeply concerning that National Treasury would even consider pushing this proposal despite the warnings of non-feasibility from the very researchers who have investigated it.

Ultimately, while the 2021 MTBPS did not answer the question of the Covid SRD grant’s possible termination and what replaces it, this is a challenge that government cannot avoid. The danger is for government, in responding to this challenge, to present a less costly solution that works only on paper, aimed purely at restricting expenditure rather than resolving the crisis of poverty. It is naïve to think that society as a whole will not bear the cost of such myopic short-termism. While the family grant proposal may have fallen by the wayside, we can expect other proposals aimed at cost-cutting to be put on the table. But the nettle of a Universal Basic Income should be grasped if a sustainable solution to poverty is to be found.

Written by Nathan Taylor, Econ3x3