Imputation methods for adjusting SOUTHMOD input data to income losses due to the COVID-19 crisis
This note sets out two different methods on how to adjust incomes in the microdata underlying the standard SOUTHMOD models to reflect a sudden shock, in this case the COVID-19 shock, as done in the accompanying working paper by Lastunen et al. (2021).
The note first describes how industry-specific GDP shocks are calculated. Next, it describes how these shocks are randomly allocated to individuals’ incomes in the microdata. Using the World Bank Phone Survey for Uganda, an alternative, regression-based method is described that models labour market transitions in a more detailed manner as it takes into account individuals’ characteristics.
The method could easily be replicated for other countries provided such alternative microdata eventually become available. We further track and compare variables between the survey data underlying the UGAMOD input data, the 2016/17 Uganda National Household Survey, and the World Bank Phone Survey for Uganda.