Assessing the impact of government policies on food consumption behavior requires knowledge of demand system parameters. In Pakistan where regional consumption habits differ considerably, estimating such a system requires careful modeling of habitual behavior. Previous demand studies have not handled the regional heterogeneity of consumption patterns in a satisfactory manner. Also, some of these studies relied on restrictive assumptions regarding consumer preferences. As a result, there is little consensus on values of demand elasticities for even the key food commodities;This study proposes the hypothesis that regional consumption patterns differ because of habits induced by long-term relative prices. Differences in regional relative prices stem from inadequate and unevenly developed infrastructure. Stability of price differentials between regions causes persistent biases in household consumption towards cheaper local foods and, in the long run, leads to formation of strong habit for consuming those foods;These ideas were operationalized using the theory of price dependent preferences. According to this theory, long-term prices condition household preferences while market prices and income define their budget sets. A Price Conditional Demand System (PCDS) was derived where the expenditure shares were functions of long-term prices and, in addition, depended on current market prices and total food expenditures. This was done without imposing arbitrary restrictions on consumer preferences. The PCDS was estimated using micro-level data from the Pakistan Rural Household Survey conducted by the International Food Policy Research Institute, during 1986-88;The results in this study support the hypothesis that households develop habits for consuming foods that are cheap in the region, on a long term basis. The model also enabled estimation of conventional price and expenditure elasticities for eight food commodities;The estimates of model parameters were used to simulate effects of trade liberalization and infrastructure improvements. The result suggested that in the long run, investments in infrastructure that reduce regional price differentials have a larger impact on food consumption than trade liberalization.