Marcelo Medeiros and Joana Costa from the International Poverty Center (IPC) define feminization of poverty as:
[...] We propose a definition that is in line with many recent studies in the field: the feminization of poverty is a change in poverty levels that is biased against women or female-headed households. More specifically, it is an increase in the difference in poverty levels between women and men, or between households headed by females on the one hand, and those headed by males or couples on the other. The term can also be used to mean an increase in poverty due to gender inequalities, though we prefer to call this the feminization of the causes of poverty.
[...] Feminization is a process, whereas a “higher level of poverty” is a state. Feminization is also a relative concept based on a comparison of women and men, including households headed by them. What is important here is the difference between women and men at each moment. Since the concept is relative, feminization does not necessarily imply an absolute worsening in poverty among women or female headed-households. If poverty is reduced sharply among men and only slightly among women, there would still be a feminization of poverty.
[...]The feminization of poverty can also be defined as “an increase in the share of women or female-headed households among the poor”. In contrast to our proposal, this definition focuses on changes in the profile of the poor and not on poverty levels within gender groups. Thus it has a potential disadvantage. It is difficult to interpret the results from this approach because measures of the feminization of poverty can be affected by changes in the demographic composition of the population. For instance, the impoverishment of female-headed households can be offset by a decline in the total number of such households, and thus the result in terms of feminization can be zero. The definition we propose gives rise to indicators that are not affected by these composition effects, which can be analyzed separately.