The politics of statistics can be made entertaining
JOHANNESBURG – In an article titled “Data fails to capture complexity of SA’s unemployment crisis”, the University of Amsterdam’s Daniel Mugge, a professor of political arithmetic, and Juliette Alenda-Demoutiez, a post-doctoral researcher in political science, make the politics of statistics not only serious, but also entertaining.
Mugge and Demoutiez recently argued that Statistics SA was innovative and experimented with the inclusion of discouraged work-seekers in the compilation of unemployment numbers.
They go further to say their “recent research found the country’s appetite for statistical creativity has slowly waned, thus undermining the gallant effort StatsSA put into the numbers”.
They further argue that measures relating to the labour market that StatsSA uses might not be appropriate, as they might fail to direct where the root cause lies.
The narrative in their political arithmetic makes severe statistical and political errors, and in so doing Mugge and Demoutiez assist to mislead or even apportion suspicion to the usual suspects – politicians – who are innocent this time around.
In South Africa quite the contrary is true. The appetite for numbers has sharpened, including at the political level.
StatsSA, in dealing with the raft of numbers, has also pointed to where the root cause is, thus ensuring that through data the issues are captured comprehensively.
For instance, in the specific case of labour markets, StatsSA using both the official and expanded definition as well as multidimensional poverty measures pointed out that the malaise lies in education.
The agency further pointed to the broad societal issues, which undermine education.
The problem neither lies with politicians nor statisticians.
The argument provokes me to excavate a seminal presentation that Helmut Spinner of Karlsruhe University made in 1998 to the European National Statisticians.
The gap is in what Spinner refers to as the risk profile of those who play in the space of statistics, which in this regard includes statisticians.
Spinner identifies six categories of players and their attendant risks. For purposes of addressing the critique I mention politicians, scientists and statisticians.
He classifies politicians as the mechanics, who have no understanding of the numbers, but they possess two critical keys, which are insight and intervention.
The second category is the scientists who have insight and understanding, but no intervention.
Finally, the statisticians, whom he classifies as the anonymous group: they have an understanding of the numbers, but have neither insight nor intervention.
The challenge I have come to realise is that there is not a planning system in South Africa that provides the right space in which the defects of each of the players are exposed and remedied and subsequently, their strengths displayed.
In his analysis Spinner considers the sixth player as the interventionist, and this is the theatre of knowledge as power.
These players require a convening system, where each one’s shortcomings are fully compensated for by the other’s dialogued strength to achieve a scientifically known and understood outcome.
StatsSA, for instance, continues to publish both measures of unemployment. and explains them to the extent that even the blind of Jericho can see and understand.
However, the theatre where knowledge is power does not exist. The problem is the dismal absence of a theatre for intervention. That theatre is where the evidence is turned into a practical long-term plan that can be tested for robustness and fidelity.
That the measures are Eurocentric certainly does not measure up. There is nothing Eurocentric in the questions and the line of questioning. The design of these methods are participatory and South Africa, especially Statistics South Africa, leads in matters of measurement and it is not a recipient of intellectual largesse from Europe or anywhere.
Mugge and Demoutiez argue that “simply put, international unemployment standards would ignore a big part of South Africa’s labour market problems. This would do an injustice to women and the black population in particular.”
Nothing is further from the truth than what the duo argue.
StatsSA through a raft of data has ably captured the numbers and explained them. The vacuum is the architecture of the policy space to respond to this challenge.
I would submit that an architecture, which has no professional planning capability and modelling competence, is incapable of digesting and responding appropriately to the challenge facing the country.
In such an environment I would agree with Mugge and Demoutiez that the tendency to ignore soft, hard-to-measure issues is amplified by politicians who happily pounce on statistics that displease them.
But I would argue that it is because they do not have the capacity to respond to this challenge and obviously to expect them to do so is a tall order. They have bureaucrats who should, but they don’t. This is the malaise.
In my paper delivered at the ISI in Marrakesh in 2015, I argued that the 21st Century Statistician is one who dances on the serrated side of the saw and not on the plain side. In this argument I agree with the duo that the country’s statisticians should continue to be bold and paint a picture that does justice to the colourful quilt that is the “rainbow nation”. And politicians have a duty to respect these efforts if they are genuinely committed to effective policies.
Dr Pali Lehohla is the former Statistician-General of South Africa and former head of Statistics South Africa. Meet him on www.pie.org.za and @PaliLehohla