Modeling of residential property rents in Somalia using two-stage modeling: Hedonic regression and artificial neural network

Purpose: Real estate is one of the fundamental growth engines for developing economies as it contributes urbanization and infrastructure development. In recent years, Somalia has witnessed massive real estate development in both housing and commercial buildings. The purpose of this study is twofold. First, the study examines the determinants of residential property rents. Second, it investigates whether residential property rents are fairly valued. Design/methodology/approach: This study uses two-stage modeling. A hedonic regression model is used in the first stage, while an artificial neural network is applied in the second stage. Findings: After analysis, this study established that size, location and security of a residential property have a significant influence on its monthly rents. Alternatively, the study identified that residential property rents are not fairly valued in Mogadishu and overvaluation is more frequent than undervaluation. Originality/value: This implies that Somalia’s real estate industry is more speculative-driven than real demand-driven. Though Somali real estate is an infant industry with huge potentials in the long run, it may end up disastrously following the well-known bubble-then-burst behavior. To avoid such crisis, this study recommends formulating government policies that regulates, supervises and protects the infant real estate industry without undermining the needs of the poor and low-income citizens.