Poverty: definitions and direction of change

The debate on whether poverty in South Africa has increased or decreased in roughly the last decade is continuing. The South African Institute of Race Relations claimed that peopleliving in absolute poverty (defined as less then $1 per day) increased from 1.9 million in 1996 to 4.2 million in 2006. Pres Thabo Mbeki refuted this claim as discussed in an earlier contribution to this blog.



Today the director of the SAIRR, John Kane-Berman, defended their publication. A few points were made:


1. Poverty is not a question of service delivery. Service delivery success is well documented leading to increased standards of living for millions.


2. Absolute poverty is rising, but it seems to be beyond its peak:
Some of the data we have published are based on a definition of absolute poverty, living on the equivalent of less than a dollar a day. On this criterion, the number of poverty-stricken South Africans rose from 1,9-million in 1996 to 4,2-million in 2005. Poverty peaked at nearly 4,5-million in 2002, but it has still doubled since 1996 (at constant prices). Mbeki finds this assertion “absurd”. But it is a consequence of the fact that unemployment has doubled. On the strict definition, it has risen from 2-million in 1996 to 4,3-million last year. On the expanded definition, it has risen from 4,2-million to 8-million, after peaking at 8,4-million in 2003.


3. Definitions of poverty as well as a focus on the direction of change seems to drive perceptions on the problem:
There are various definitions of poverty. The president refers to one published by his own office in June in which the poverty line is defined as R3000 per head per year. He says that on this measure, the proportion of people living in poverty has dropped from 53,1% in 1996 to 43,2% last year (in constant rands). Translating the latter percentage into actual numbers means there were nearly 21-million people living in poverty last year — a far higher figure than that of 4,2-million cited by our institute (for 2005).

Listening to all of this, were are we? My interpretation:

First, the difference between $1 per day or R3000 per annum seems very small to me (give and take a bit on the exchange rates). So depending how one defines it we have between 4 and 21 million people absolutely or just above absolutely poor in the country. That's a huge difference for a small change in definition. Definitions thus drive magnitudes and ultimately policy decisions on how to deal with the problem.

Second, there may be some evidence that social grants start to reach some of the poorest of the poor. Or is this decline artificially driven by an appreciating Rand? (have to check on that one). Social grants reach 25% of the population already and may have reached its plateau. It does look (and this needs to be verified) as if social grants are largely responsible for 'filling the gap' between the two definitions of poverty.

Third, it also seems that it is not due to sustained economic growth and job creation that poverty is reduced, but to some extent through the creation of a social safety net which may have reached its limits.

Verified information on this is vital for understanding the dynamics of the South African socio-economic system and vital for using the correct policy levers to influence better development outcomes.