MODELING THE EFFECT OF DEVOLUTION ON YOUTH UNEMPLOYMENT RATES IN KENYA USING AUTOREGRESSIVE INTEGRATED MOVING AVERAGE- INTERVENTION MODEL

Abstract

Youth unemployment remains a major concern, particularly in African coun tries with the youngest population globally. In Kenya, youth unemployent rate has shown fluctuations despite several government efforts such as Kazi kwa Vijana (jobs for youth), Youth Enterprise Development Fund (YEDF), Kenya Youth Empowerment Project (KYEP) and Youth Employment Scheme Abroad (YESA). There is limited empirical research that has quantified the effect of devolution on youth unemployment trends in Kenya, as seen in the literature. To address this gap, the study modeled the effect of devolution on youth unemployment rates using an Autoregressive Integrated Moving Av erage (ARIMA)- Intervention model. ARIMA- Intervention model was se lected due to its applicability in quantifying the effect of policy changes on the time series of interest. To capture the time varying effect of devolution on youth unemployment while controlling the effect of key macroeconomic in dicators such as Gross Domestic Product, inflation rates, interest income on the Kenya’s public debt and Foreign Direct Investment (FDI), the Dynamic Adaptive Intervention Model (DAIM) was used. Secondary data on youth unemployment rates from the World Bank data bank covering the years 1991 to 2022 was used. Computational analysis was done using Python and R programming software, with model selection based on the Akaike Information Criterion (AIC) and Pseudo R- Squared for ARIMA model and DAIM re spectively. ARIMA(0,0,0)(0,0,1)[4] model with non-zero mean incorporating devolution as an intervention revealed that the effect of devolution on youth unemployment rate was statistically significant at 5% significance level. The average unemployment rate rose from 6.67% during the pre-intervention pe riod to 10.19% during the devolution period. The counter factual projection was approximately 8.583%, suggesting that devolution had a measurable and negative effect on youth unemployment rates. The fitted DAIM revealed that the negative effect of devolution on youth unemployment intensified gradually over time, rather than manifesting as an immediate shock. Additionally, inter est income on the Kenya’s public debt consistently worsened the time-varying effect of devolution on youth unemployment rates across all the considered distributional quantiles. Empirically, the study findings provides a quantita tive basis for evaluating the effectiveness of the devolved governance system in addressing youth unemployment in Kenya.