Skip to main content
January 16, 2019

Poor data no excuse for our bad policies

A view of the containerized data center at supreme court. /FILE
A view of the containerized data center at supreme court. /FILE

The ability to make good decisions is contingent on the availability and utilisation of sound evidence. The overarching assumption is that the evidence is derived from reliable data. Moreover, it is expected that the data be collected by rigorously established procedures.

Morten Jerven’s book Poor Number: How we are misled by African development statistics and what to do about it provides an insightful analysis of the production and use of data for development in Africa. The book concludes that the capacities of national statistical agencies have fallen apart.

A World Bank report, Poverty in a Rising Africa, published in 2016, argued that lack of reliable and comparable data masks complex realities and makes it difficult to assess Africa’s progress. Hence, sustained and joined efforts are urgently needed to improve the quality of and timeliness of statistics in the continent.

The concerns Morten Jerven and the World Bank raise are spot on, and are strong indictments of the incapacity of the statistical offices in African countries. But there is some data.

The real concern, in my view, is how little use we make of existing data. Whatever little data we have and however old or outdated it is, provide invaluable insights about the long-term impacts of national policy priorities.

At the East African Institute, we looked at about 48 variables ­– from publicly available data – across all of Kenya’s 47 counties. Using both basic and fairly sophisticated data analysis and modelling techniques we uncovered insightful patterns about the differences and similarities among the 47 counties. What is exciting about the insights is that they are most indelible fingerprints of the policies were have implemented for over half a century.

From our analysis Kenya’s 47 counties divide into four neat groupings. One group comprises Turkana, Marsabit, Samburu, Garissa, Tana River, Wajir, Mandera and West Pokot. The defining characteristics of these counties are: High fertility rates; high maternal mortality rates; low levels of mothers’ education, poor access to health facilities and stunting.

Another group of counties comprises Kiambu, Nyeri, Murang’a, Kirinyaga, Machakos, Uasin Gishu, Meru, Nakuru and Nyandarua. A high density of health facilities, high levels of literacy, and high per capita access to grid power characterise these counties, unlike the first set of counties.

A child born in Meru is three times more likely to celebrate her fifth birthday compared to a child born in Mandera. Moreover, a pregnant woman in Turkana is six times more likely to die of pregnancy related complications than a pregnant woman in Kiambu.

Paucity of data can no longer be used as an excuse for making bad public policy of investment decisions. And yes there is so much we can learn from half a century of policy experiments, which have led to divergent and unequal human well-being outcomes.

While counties such as Meru or Kiambu are not perfect, they have something we can learn and replicate in Mandera or Tana River. Certainly, a child born in Mandera must have the same life chances as one born in Meru.


Alex O. Awiti is the director of the East African Institute at Aga Khan University


Poll of the day