PurposeThe present study evaluates the impact of special economic zones (SEZs) on poverty, both rural and urban with special reference to Andhra Pradesh in India, using household consumption expenditure data. In addition to estimating the effects of the SEZs on poverty, the authors explore some of the possible mechanisms generating these effects.Design/methodology/approachThe authors apply a difference-in-differences (DID) technique on a pooled, cross-sectional, district-level dataset based on official annual household surveys for the period from 2001 to 2012 to estimate the average effects of SEZs on household expenditure per capita, a commonly used measure of household poverty.FindingsThe establishment of the SEZs constituted a major exogenous shock to rural economies by creating demand for large chunks of land, which had an immediate impact on the economic and social settings of these economies and aggravated rural poverty. However, over time the poverty aggravating effects of SEZs in rural areas dampened. The effects of SEZs on urban poverty are found to be different from those on rural poverty. It is also revealed that the districts with multiple SEZs experienced larger effects than those with only one or two SEZs. Overall, the SEZs did have positive expenditure effects, but this transition might have been accompanied by heightened inequality between the rural and urban areas.Research limitations/implicationsFirst, the authors did not have access to village or municipal-level consumption data. It is therefore assumed here that district level performance is a reliable proxy for the relevant impacts of SEZ operations. Second, panel data, which would allow more precise measurement of effects than the pooled cross section data used in this study, are not available. Third, the authors’ econometric analysis is essentially comparative statics in nature and does not capture possible spillover dynamics, issues of relocation of economic activity, or migration.Practical implicationsFirst, land acquisition is likely to emerge as a major political and social challenge for the localities that host SEZs. For effective policy implementation, it is necessary to establish legal institutions to address this challenge. Second, governments in developing countries often announce new SEZ programmes on a very large scale and insist that they be implemented over short periods of time. The authors recommend that the government should adopt an experimental approach in implementing the policy. Third, the authors provide evidence that in the long run, effects of SEZs hinge on the success of SEZs in attracting investment and generating additional employment. The policy must therefore be informed by rigorous analysis of the potential of SEZs in the country, as well as alternative policy options.Social implicationsThe authors’ results show that large-scale land acquisitions to implement large industrial projects are likely to result in shocks to the rural economy exacerbating rural-urban inequalities: village communities lose their resource base, are marginalised in the process, and, as a result, face economic deprivations. It may lead to severe economic, social and political consequences. The authors’ study implies that any strategy for large-scale industrialisation should take cognisance of its effects on the affected communities and should be designed to include strategies to improve their economic opportunities and to ensure social inclusion.Originality/valueSEZs are one of the most controversial topics within development policy discourse. Their regional development effects are subject to intense debate. Yet, there is surprisingly little systematic evaluation to inform the debate and to guide policymakers. This is one of the earliest studies to assess the poverty effects of SEZs and is the first for India, using household consumption data.