Info dearth hampers anti-poverty fight

Summary - In 1994 the ANC committed itself to eliminating poverty and unemployment. Ten years later, how are we doing? Attempts to measure progress are unfortunately hampered by unsatisfactory data. The four main sources of information on poverty - the 1996 and 2001 censuses and the 1995 and 2000 income and expenditure surveys (IES) – are all problematic. The 2001 census showed zero income for over 2,5 million households, which cannot be correct, and the 2000 IES was poorly executed. Analysts are consequently divided over whether inequality has remained about the same or worsened. There are similar difficulties with measuring unemployment. Statistics were compiled by the October Household Survey (OHS) between 1994 and 1999 and by the Labour Force Survey (LFS) from 2000. However, the two surveys are not strictly comparable as the LFS criteria for counting a person as employed are more lenient. Taking everything into account, it seems that approximately 600 000 jobs were created between 1997 and 2002. However, the labour force grew even faster. The picture that emerges is therefore one of a slow loss in the fight against inequality and unemployment. One reason for this is tepid economic growth. Another is population growth. White fertility has been decreasing for decades but black fertility, though dropping rapidly, will not reach comparable levels for 10-20 years. From about 2010, labour force participation rates will start to stabilise. A third reason is that labour supply has increased even faster than the population, with many more women seeking work. Another major cause of poverty and unemployment is inefficient human capital formation. Academic quality in many of our schools remains deplorably low, especially in maths and science. This disproportionately retards African advancement and is also a constraint on economic growth. On the plus side, as enrolment in education stabilises as a result of the declining birth rate, schools will be able to concentrate on improving the quality of education. In the long run, then, there are trends that hold out hope for the battle against poverty and unemployment. In the short run, expanded social grants are an important palliative. Child support grants will soon cover about half the country’s children under the age of 14, and pensions will also rise rapidly as the population ages. By 2012, social grants will approach the education budget in size. Analysts agree that current social grants are well targeted and have considerably reduced poverty. Proponents of the basic income grant (BIG) believe that an even greater fiscal effort would eliminate poverty entirely. However, this is unlikely as poverty is too extensive and taxation is too narrowly based. The government has limited room to add new grants, especially as health expenditure will have to rise as Aids morbidity and mortality increase. Job creation undoubtedly requires more effort. The government is hoping that economic growth will provide the necessary revenue for increased expenditure, but no upswing lasts forever and the economy is likely to experience increased fiscal stress by 2010. We will then have to make important choices between further short-term poverty alleviation and longer-term investment for growth.

Not surprisingly, the African National Congress made the elimination of poverty and unemployment a major part of its policy platform for the 1994 election, and it has remained a priority ever since. So, ten years later, how are we doing?

In the first place, there are substantial uncertainties of measurement. There are four main sources for the measurement of poverty and inequality: the 1996 and 2001 population censuses and the 1995 and 2000 income and expenditure surveys.

The 2001 census income data are more complete than the 1996 data, partly because the 2001 census used imputation techniques to estimate missing information — a defensible and widely used technique — and the 1996 census did not. Even so, the 2001 census assigns zero income to over two and a half million households, which cannot be correct. A lot of small incomes have been missed.

The 1995 income and expenditure survey (IES) is a useful source. But it soon became apparent when the eagerly awaited 2000 income and expenditure survey was released that it had been poorly executed, bearing signs of deficient supervision of fieldwork and inadequate data editing. As a result, the analytical community is divided into those who believe that inequality was about the same in 2000/01 as it had been in 1995/96, and those, like myself, who believe that it has worsened.

There are also difficulties around measuring unemployment. Between 1994 and 1999 the October Household Survey (OHS) measured unemployment. Since 2000, the Labour Force Survey (LFS) has measured it. There is a break between the series, since the LFS asks more searching questions about activities which might count as employment. As a consequence, estimates of employment in the informal sector are much higher in the LFS than in the OHS, reflecting more a change in the instrument of measurement than a change in circumstances. It should be noted that the LFS criterion for counting as employed is very lenient: just one hour of work in the past week, which can be spent on a range of non-market activities as well as work for pay or family businesses.

I have put together all the available information and estimated that employment increased by about 600 000 between 1997 and 2002, 80 per cent of it in the formal sector. This is substantial, but less than labour force growth in the same period. The limitations of the data make the estimate imprecise. Other defensible estimates could range from zero employment growth to double the employment growth of my estimate. There is substantial sampling error in the LFSs, which interview about 30 000 households in each round. They contain much useful information but they cannot detect short-term changes in unemployment as they are meant to do, unless these changes are unusually large.

The picture appears to show a slow loss of the fight against inequality and unemployment in the last ten years. What are the reasons for this? It is conventional to refer to tepid rates of economic growth and to suggest that these need to be doubled. That cannot be faulted as far as it goes, but there are other dimensions to consider.

South Africa’s population increased from five million in 1900 to forty-three million in 2000. The more than eight-fold increase can be ascribed to a stretched-out fertility transition. There is evidence of dropping white fertility in the 1890s, whereas the black African fertility transition — already far advanced — will not be complete for another ten or twenty years. Labour supply has increased even faster. In 1950, less than 25 per cent of women in the 15-64 age range were economically active. Now the rate is more than double that. Among the youngest age group (15-24), labour force participation rates are the same for men and women. Even if we had escaped the economic travails of late apartheid, there would have been a major problem about labour absorption after 1973, when growth internationally slowed.

The twenty-first century will be different. The population growth rate is slowing at present and by the end of the decade will be little above zero. In part this is because of the increased mortality occasioned by Aids, but it would not have happened if fertility had not been dropping rapidly over the past twenty years. Labour force participation rates will also start to stabilise. Labour supply will therefore grow more slowly, though reliance on its slow growth rate alone to mop up unemployment will mean that full employment will still be two or three decades away.

Of course, an undifferentiated approach to labour supply is inadequate. The inefficiency of South African human capital formation is a direct cause of poverty and unemployment. The education system was subject to the demographic pressures of the twentieth century and rapidly rising enrolments that not even Dr Verwoerd was able to arrest. White South Africans born in 1885 received eight years of education and everyone else less than two. Their counterparts born in 1970 received over nine in the case of Africans and coloureds and over eleven among Asians and whites. Again, the twenty-first century will see the stabilisation of enrolments in education generally and, ultimately, in higher education.

No sector has carried out a more thorough Confucian ‘rectification of names’ than education since 1994. Pupils have become learners, teachers have become educators, inspectors have become educator development facilitators, syllabuses have become curriculum statements, wheelchair access facilities and reconciliation programmes have been introduced, and the emblem of colonial and apartheid brutality — corporal punishment — has been abolished. If only progress in imparting the three Rs had been as rapid! Academic quality in a great many South African schools remains deplorably low, especially in arithmetic and mathematics, creating a poor platform for further and higher education. It disproportionately retards African advancement, hence rising inequality. And it holds back growth.

At the demographic level, one further factor needs to be mentioned. High and late apartheid directly moved, or applied pressure to move, on millions of people. As a result, commercial farms, small holdings and recreational areas contain a far smaller proportion of the population than they did fifty years ago. Poverty in these areas is now much less severe than in the heavily populated tribal areas. A disjunction between settlement patterns and employment opportunities remains. The 2001 census shows substantial immigration into Gauteng and the Western Cape during the preceding five years. This is likely to continue for the next couple of decades, with continuing high levels of poverty and unemployment in urban informal areas.

So there are new trends that, intelligently worked with, hold out fresh hope in the fight against poverty and unemployment in the longer run. But what of the short run?

Expanded social grants are an important part of the answer. Since 1994, child support grants have been introduced and will, from 2005/2006, cover about half the country’s children under the age of 14. There is upward pressure on other grants. The population is aging, so that the number of older person’s grants will rise fast. 40 per cent of whites and 90 per cent of everyone else in the relevant age range receive these grants. Aids is putting upward pressure on childcare grants and disability grants.

A study I did eighteen months ago concluded that, while 110 per 1 000 South Africans received social grants in 2002, 250 per 1 000 will receive them in 2012. The real cost of these social grants in 2012 will be 60 per cent higher than in 2002; social grants will approach the education budget in size.

Analysts agree that the current social grants are well targeted and that poverty would be considerably worse without them. The question arises: can a greater fiscal effort eliminate poverty entirely? I do not believe it can; poverty is too extensive and too deeply rooted and taxation too narrowly based for that. Proponents of the basic income grant (BIG) tend to believe otherwise. The Taylor Commission concluded that a BIG of R100 per month, at March 2001 prices, for everyone not receiving a social grant would largely do the job.

Critics of the BIG have pointed to the fact that it would be very expensive and that it would be poorly targeted, in that it would be paid to the richest as well as to the poorest people. BIG proponents meet the latter point by making BIG taxable so that there would be considerable fiscal claw back. But it is in financing that the BIG coalition falls apart. To finance it wholly from personal income tax would require grotesquely high marginal rates of tax from the middle of the range to the top. To finance it largely through indirect taxes, notably Vat, would have the unions up in arms. Their members would mostly be net losers.

Seeing this, the government is opposed to BIG, but leaves the door open for the introduction of other sorts of grants.

The room it has for new grants is limited. The South African Revenue Service has done a remarkable job in improving tax compliance in recent years but still has some way to go. Improvements are likely to come at a slower rate and so increases in collections at a given rate will be more modest in the coming years. As we have seen, it has to finance rapidly increasing social grant expenditure even without any new commitments. Health expenditure too has to rise. The limited response of the government to popular pressure about HIV/Aids is a remarkable aspect of the functioning of our new political system that deserves a full analysis in its own right. But the pressure will rise. Although HIV prevalence rates may be peaking, Aids morbidity and mortality have not.

Employment generation undoubtedly needs more effort. At present, the government is working for more labour-intensive techniques on projects it meant to fund anyway. This needs to be complemented by some new projects. Informal area upgrade in metropolitan areas would be a suitable field for expansion.

All this puts substantial upward pressure on government expenditure. The most recent Medium Term Expenditure Framework reflects this to some extent and relies on optimistic growth forecasts to provide the necessary revenue. In the next couple of years growth may well be sufficient. But no upswing lasts forever and the system is likely to start experiencing increased fiscal stress by the end of the decade. It is then that important new choices will have to be made between further short-run poverty alleviation and longer-term investments underpinning growth.