Air transport demand and economic development in sub-Saharan Africa: Direction of causality: Fid-Bau Portal

Air transport demand and economic development in sub-Saharan Africa: Direction of causality

Tolcha, Tassew Dufera / Bråthen, Svein / Holmgren, Johan

Abstract Air transport may be a key tool to advance economic development. However, it is uncertain whether air transport boosts economic development, or vice versa. Both views have theoretical and empirical support. In some countries and regions, air transport is important for initiating development, for example by attracting foreign direct investment or granting access to lifelines. Elsewhere, economic development drives air transport demand. Establishing the direction of causality for regions/countries segmented by income level may inform pragmatic policy. This study analyzes the causal relationship between air transport demand and economic development for six sub-Saharan African countries for the period 1981–2018. Vector error correction and vector autoregression models are employed to identify long- and short-run causalities. The results reveal heterogeneous, context-specific causal relationships. In the long-run, for South Africa, Nigeria and Kenya, the direction of causality runs from economic development to air transport demand; for Ethiopia, causality runs in the opposite direction, with increased demand for air transport promoting economic development; and for Senegal and Angola, the relationship is too weak to infer causal directions. Possible explanations for this heterogeneity include differences in per capita income, low-cost carriers' share of national aviation markets, the presence of large home-based airlines, and comparative geographical advantage as a natural hub.

Highlights • It is uncertain whether air transport boosts economic development, or vice versa. • This study analyzes the causal relationship between air transport demand and economic development for six sub-Saharan African countries for the period 1981–2018. • Vector error correction and vector auto-regressive models were employed. • The results reveal heterogeneous, context-specific causal relationships. • Possible explanations are provided for the heterogeneity.