Corrigendum to "How information about foreign aid affects public spending decisions: evidence from a field experiment in Malawi" J. Dev. Econ. 146 (2020) 102522

The authors regret an error in the coding of our treatment variable. Some observations that should have been in the aid information treatment condition were incorrectly coded as control. This issue affects 151 of the 1252 maps in the experiment. We also included in the analysis some schools which were outside respondents’ constituency boundaries.1 This issue affects 83 out of the 1252 maps in the experiment. Below we show the revised tables and figures after correcting the treatment variable and removing affected maps (Table 1, Table 2, Table 3, Table 4, Table 5, Fig. 4, Fig. 5, Fig. 6). Our treatment effect estimates are consistent and generally larger in magnitude after correcting these errors. We stand by the conclusions in the original article. We apologize for any inconvenience caused.