A persistent gender gap in inflation expectations, i.e., women systematically reporting higher expected inflation than men on average, has been documented across countries and over time, yet its underlying causes remain under debate. Using more than half a million responses from the ECB Consumer Expectations Survey and a double machine learning framework that provides valid inference on many dimensions of heterogeneity at once, I show that there is no single gender gap: individualized gaps range from roughly −1 to +5 percentage points. This heterogeneity is shaped primarily by subjective belief-formation variables, such as forecast uncertainty and the rounding of inflation beliefs, and only to a much lesser extent by objective ones, such as financial literacy. When inflation is high enough to attract consumers’ attention, the average gap narrows, but the distribution of individual gaps fans out, widening precisely for economically vulnerable women. Since the women who diverge most from men hold the most imprecise beliefs, communication that reduces ambiguity, rather than solely providing information, is a promising policy lever.
