Sorry, you need to enable JavaScript to visit this website.



Spatial Autocorrelation Panel Regression: Agricultural Production and Transport Connectivity

Note: By clicking the DOWNLOAD button, you will leave the GICA website and be redirected to the source site of the selected document. The source site’s terms of use will govern your use of the selected document.
183 bytes •

Spatial analysis in economics is becoming increasingly important as more spatial data and innovative data mining technologies are developed. Even in Africa, where data often crucially lack quality analysis, a variety of spatial data have recently been developed, such as highly disaggregated crop production maps. Taking advantage of the historical event that rail operations were ceased in Ethiopia, this paper examines the relationship between agricultural production and transport connectivity, especially port accessibility, which is mainly characterized by rail transport. To deal with endogeneity of infrastructure placement and autocorrelation in spatial data, the spatial autocorrelation panel regression model is applied. It is found that agricultural production decreases with transport costs to the port: the elasticity is estimated at -0.094 to -0.143, depending on model specification. The estimated autocorrelation parameters also support the finding that although farmers in close locations share a certain common production pattern, external shocks, such as drought and flood, have spillover effects over neighboring areas.


Iimi, Atsushi; You, Liangzhi; Wood-Sichra, Ulrike. 2017. Spatial Autocorrelation Panel Regression : Agricultural Production and Transport Connectivity. Policy Research Working Paper;No. 8089. World Bank, Washington, DC. © World Bank. License: CC BY 3.0 IGO.


  • connect