This paper investigates whether estimating vulnerability to poverty using consumption expenditures and food calorie intake provide consistent policy implications based on the vulnerability as expected poverty framework. Feasible generalized least square (3FGLS) method is applied on large data-set of about 90,000 households in the Punjab province, Pakistan. The results indicate that the analysis based on the two approaches results in considerable geographic differences in the distribution of vulnerability to poverty. The implication of our results is that anti-vulnerability interventions based on monetary indicators of poverty will penalize those regions with low monetary poverty but high nutrition poverty.