The effect of public expenditure composition on unemployment dynamics in developing economies: evidence from Kenya

Abstract

This paper investigates the impact of disaggregated government expenditure composition on national unemployment in Kenya from 1993 to 2023. Grounded in Keynesian fiscal multiplier dynamics and endogenous growth theory, the study addresses the empirical puzzle of fiscal policy effectiveness in developing economies. We apply Principal Component Analysis to construct composite expenditure indices and estimate an Autoregressive Distributed Lag bounds testing framework. The empirical findings yield three key insights. First, Kenya's labour market exhibits high structural unemployment persistence, with the error correction coefficient being statistically significant. This indicates that there is a slow adjustment to equilibrium. Second, fiscal expenditure components exert opposing effects: development expenditure is significant and negatively associated with unemployment. Further, recurrent expenditure is significant and positively associated with unemployment, reflecting crowding-out effects. Real output growth confirms Okun's law with a statistically significant negative coefficient. These results challenge the premise that aggregate fiscal expansion automatically drives job creation. Instead, sustainable labour absorption requires strategic budget reallocation from administrative consumption toward productivity-enhancing capital investments.