How Poor are Kenya’s Major Ethnic Groups?
The Kenya National Bureau of Statistics recently released the Kenya Poverty Report based on the 2021 Kenya Continuous Household Survey. The report contains data on Kenya’s poverty rate and a further breakdown of poverty rates per county. Like most government reports, the data presented doesn’t highlight associations between variables, which makes it difficult for readers to establish causation or crucial links between variables. For example, the report suggests that Kirinyaga has a headcount poverty rate of 19.3%, while Turkana’s rate is 77.7%. Without further analysis, it’s difficult for a reader to understand the magnitude of the problem and the reason for their disparate outcomes. Similarly, it is unclear why relatively similar counties like Kirinyaga and Nyeri (26.4%) might have different poverty rates. Therefore, while the data might help with policy-making and implementation, problem-solving is impeded when no further analysis of causation and association between variables is done.
One crucial factor that gets overlooked by these reports is ethnicity. It’s common for different ethnicities to have disparate outcomes in their economic development and rates of poverty. In the United States, where ethnic and racial data is robust, evidence of disparate outcomes in education, health, and socioeconomic status has been observed. In Kenya, it’s also safe to assume that different ethnicities will have different outcomes. The cause of these differences will often go unexamined since the state rarely focuses on ethnic differences. County data is probably the state’s way of knowing how different ethnic groups are doing. However, since it’s possible for ethnic groups to span more than one county, determining commonalities requires further analysis. In this post, therefore, I perform the simple task of highlighting ethnic poverty rates. Afterward, I highlight correlations between poverty rates, stunted growth, and wasting among children under five in each ethnic region.
Overall Poverty Rates
So how poor are the different ethnicities in Kenya? The chart below highlights the differences in mean overall poverty rates by ethnic region in the country. Ethnic areas are determined by grouping counties inhabited by people majorly of the same tribe. For instance, due to the large numbers of Kikuyus in Nyeri and Nyandarua and the large number of Kambas in Kitui, Makueni, and Machakos, the two ethnicities are grouped as Mt. Kenya and Ukambani, respectively. Furthermore, this chart focuses on the country’s major tribes and leaves out the others.
The Kenya National Bureau of Statistics defines overall or absolute poverty as living on less than Ksh 7,193 per month in urban areas or Ksh. 3,947 in rural areas. Nationally, 38.6% or 19.1 million people live in absolute poverty. Nairobi has the lowest headcount rate of 16.5%, followed by the Mt. Kenya region, with a mean poverty rate of 28.2%. Mt. Kenya region consists of 10 counties, including the former central province. For the clearance of doubt, Kikuyus predominantly inhabit central Kenya counties of Nyeri, Kiambu, Nyandarua, Muran’ga, and Kirinyaga. Other Mt. Kenya counties like Meru, Embu, and Tharakanithi are closely related ethnically and politically to the Kikuyu but are a different demographic altogether. Nakuru and Laikipia are politically part of Mt. Kenya and are largely but not entirely settled by Kikuyus. Therefore, subtracting Central Kenya from the larger Mt. Kenya region, we find that the Kikuyu have a lower mean poverty rate of 24.98% compared to the other Mt. Kenya counties, with a mean poverty rate of 31.46%. This mean difference was, however, not statistically significant p=0.090, and we cannot tell for sure whether there’s any difference in poverty rates between the Kikuyu and the larger Mt. Kenya region.
Other ethnic groups of interest in the bar graph above are the Kalenjin, Luo, Kisii, and Luhya. It has been argued that proximity to power by certain ethnicities helps build wealth and reduce poverty in their regions. The Kikuyu have been in power for 35 years, while the Kalenjin have been in power for 24 years. However, poverty rates in the areas settled by the Kalenjin are higher than in regions such as Luo land, Kisii land, Ukambani, and Maasai land that have not been in power. It’s also unclear what explains the poverty levels among the Luhya.
Food Poverty Rates
The Kenya Poverty Report also presents food poverty rates essential in determining how food-secure various ethnicities are. The food poverty rate is determined through a daily adult calorie requirement of 2,250 Kcal. Different food items are brought together to form a food basket priced and updated for inflation. The food poverty lines for urban and rural areas were calculated to be Ksh 2,905 and 2,331, respectively. The bar graph below shows the food poverty rates for various ethnicities and regions.
Kikuyus are from the least food-poor region, followed by the larger Mt. Kenya region, which they also inhabit. Interestingly, areas with very high agricultural productivity, like Kisii land, Luhya land, and Kalenjin land, have relatively high levels of food poverty. What explains this phenomenon? I assume food poverty rates correlate with overall poverty rates. Testing this hypothesis shows a near-perfect and significant correlation of 0.9. Therefore, we can expect that poor regions will also have higher levels of food poverty. The reason for this strong association is partly because KNBS uses the average expenditure on a basket of food commodities in Ksh rather than a direct measure of the number of calories consumed from foods in each region.
A question that comes up, therefore, is whether different ethnic areas have disparate access to nutritious food, despite the food poverty rate. For instance, a typical Luhya meal will consist of certain foods and not others, and so will that of the Kikuyu, Maasai, and Kalenjin. Word on the street is that Kikuyu food is bland, lacks variety, and is less likely to be nutritious. In contrast, Luhya food is considered highly nutritious and contains variety. If the type of food and variety matters, we expect agriculturally productive areas with food cultures that emphasize nutrition and variety, like Western and Nyanza, to have low levels of stunted growth and wasting among children under five. I tested this hypothesis by looking at stunting and wasting in children provided by the Kenya Demographic and Health Survey 2022.
Apart from the Kikuyu, who still maintain low levels of stunted growth, there has been a lot of reshuffling in the chart. Regions such as Mt. Kenya have higher levels of stunted growth despite lower rates of absolute and food poverty. The Kalenjin still maintain high levels of stunted growth, reflecting their high food poverty rates and high overall rates of poverty. Somalis, Kisiis, and Luos have lower levels of stunted growth. The correlation between food poverty rates and stunted growth is moderate (r=0.4) but significant (p=0.005). To wit, it’s still unclear whether regional food culture and culinary behavior influence stunting in children. Nevertheless, food poverty remains an important predictor of stunting. To this point, policymakers should realize there are positive and significant correlations between absolute poverty, food poverty, and stunted growth and that some ethnic regions are more impacted than others. Furthermore, it appears that raising incomes matters if ethnic groups are to escape stunted growth and food poverty.
Unlike the moderate correlation between stunted growth and food poverty, the correlation with wasting is relatively high. Regions with high rates of food poverty also have many children under five who are wasting. The relationship is strong and significant (r=0.7, p=0.00). Wasting also correlates significantly with hardcore (extreme) poverty (r= 0.77, p= 0.00). In contrast, food poverty negatively correlates with overweight rates in children under five. Regions like Nairobi and Central with low levels of food poverty tend to have a larger number of children who are overweight (r=-0.56, p=0.00).
Wealth between different regions also varies. For central Kenya, it is understandable why they would have positive indicators in almost all aspects. The region is not only less poor but also wealthier than other regions. The chart below shows GDP per capita figures for the different ethnic regions.
To sum up, different ethnic communities in Kenya contribute differently to the country’s economic trajectory. As highlighted in this discussion, ethnic differences exist in the rates of income and poverty. I intend to continue updating my dataset, ensuring I have all relevant variables that will help determine causation in the future. It would be interesting to know why some ethnic regions develop faster than others and how to help those doing so slowly. These findings have so far helped to put various data by the KNBS into context, providing more nuance that policymakers can rely on to improve the livelihoods of their constituents. It doesn’t just stop at knowing some regions are poor; of equal importance is knowing how other variables like literacy rates, fertility rates, and various levels of individual and government investment affect a region.
