This paper considers the case that electricity is a weak link in development, meaning that small productivity improvements in the electricity sector lead to large long-run productivity gains in the aggregate. To do so we build a new global database of electricity inputs and outputs covering countries of all income levels. We document that poor and rich countries actually have similar levels of TFP in the electricity sector. We then use our data to estimate a multi-sector model in which electricity is a strong complement to other inputs and subject to distortions that raise electricity costs in poor countries. The model predicts that TFP improvements in the electricity sector have smaller aggregate impacts than predicted by their Domar weights, which is inconsistent with the view that electricity is a weak link in development. Counterfactual scenarios when electricity TFP is low in poor countries are more consistent with a weak link view, but make counterfactual predictions on various other dimensions of the data. In extensions of the model, we explore some channels through which electricity might be considered a weak link, finding some promise in a version of the model when electricity losses lead other inputs to be used less efficiently.
