This paper explores the role of constraints in agriculture in preventing farmers from entering export commodity markets and thus from escaping poverty. Major export crops, like tobacco and cotton in Malawi, are usually considered high-yield crops because they are more profitable than subsistence activities (like production for home consumption). Hence, we expect households engaged in export markets to be less likely to be poor than households restrained in subsistence agriculture. If it is true, however, that export commodities are more profitable than subsistence activities, then promoting entry into export market sounds like a plausible policy recommendation, especially in terms of poverty reduction. In addition, observing higher returns in export agriculture raises the question of why farmers are not switching production activities into the more profitable ones. Although part of the differential return surely compensates for special attributes required to produce for exports, such as skills, another part may be due to large distortions in the economy. Indeed, entry into commodity export markets usually requires special knowledge, credit, start-up capital, infrastructure, and labor supply. With imperfect markets, wide margins in the returns to export agriculture and to subsistence agriculture may arise in equilibrium between otherwise similar households. Faced with higher tobacco prices, for instance, some farmers may be unable to enter markets if they do not have sufficient credit or wealth to cover any start-up costs (in seeds, fertilizers). Similarly, faced with higher cotton prices, for instance, farmers may be unable to expand current cotton production if family labor supply is limited or if rural labor markets are thin. In this paper, we propose to study the role of household constraints in agriculture on entry into commodity export markets. We begin by quantifying income gains in export agriculture. Then, we link these income gains to specific constraints in agriculture, mainly infrastructure variables and household determinants. There are two sets of determinants of commercialization that we explore. There are micro-determinants at the household level and more aggregate determinants at the village level. Once these links are identified, we will be able to dictate broad guidelines to assess policies that are conducive to poverty reduction.