Widely, 12-month or 4-quarter average indicators, such as monetary poverty rates, are computed from repeated cross sections of household surveys and interpreted as annual. This is a valid interpretation only when individuals do not change their status within the year; for instance, those observed as poor in the month they are interviewed stay poor the other 11 months. First, we demonstrate that such misinterpretation affects the calculation of several labor market, inequality, and monetary poverty measures. Then, we propose several methods to accurately annualize sub-annual data. Some methods rely on ancillary questions often included in household surveys while others require econometric techniques such as predictive mean matching. Using data for Colombia, we apply the methods to compute annual measures of labor participation, occupation, per capita labor income, average per capita household income, the Gini coefficients of labor income and per-capita household income, and moderate and extreme monetary poverty indices (headcount, gap, and severity). We show that differences with respect to the usual calculations based on monthly averages can be substantial.
