Document Type : Research Paper
Authors
1 Associate Professor, Department of Economics, Faculty of Literature and Humanities, Ilam University, Ilam, Iran
2 Assistant Professor, Department of Economics, Faculty of Literature and Humanities, Ilam University, Ilam, Iran
3 Associate Professor, Department of Economics, Faculty of Economics, Allameh Tabataba’i University, Tehran, Iran
Abstract
Universal subsidies have a limited impact on the welfare of vulnerable groups, while substantially increasing government expenditures and disproportionately benefiting high-income households. This highlights the urgent need to reconsider the mechanisms for subsidy payment and the identification of poor households. One proposed approach is the regionalization of the subsidy system. In this respect, the current study compared national and regional subsidy targeting at the urban level. Using the data from the Urban Household Income and Expenditure Survey (2022/2023), the study measures the poverty line and poverty indices. On the basis of these indices, urban areas across provinces were classified into eight regions using a hierarchical clustering method. Then, a numerical optimization method was employed to compare the outcomes of subsidy targeting at the national and regional levels. According to the results, subsidy targeting based on both individual and combined characteristics of the most vulnerable groups leads to significant differences at national and regional levels. These differences are evident in optimal subsidy amounts, poverty reduction outcomes, efficiency of targeting, and associated errors. Therefore, even when policymakers target households with identical characteristics at both national and regional levels, subsidy payment amounts must still vary across regions and population groups to achieve effective targeting.
Introduction
Untargeted subsidy payments in Iran have not only failed to improve the welfare of vulnerable groups but have also disproportionately benefited higher-income households. This outcome prompted reforms to certain commodity subsidies; however, the cash subsidies introduced as replacements were likewise not targeted. Consequently, universal commodity subsidies were effectively replaced by universal cash subsidies, which had little impact on poverty reduction in Iran. In light of these challenges, the need to reconsider subsidy payment mechanisms and improve the identification of poor households has become increasingly evident. Targeting subsidies by distinguishing the poor from the non-poor can reduce resource waste and increase the share of benefits received by the poor within a fixed budget. However, such benefits depend on the government’s ability to accurately identify poor households. In practice, structural weaknesses in the tax system and social provision institutions limit the government’s capacity to do so (Darvishi et al., 2019). To address this limitation, the current research aimed to propose a numerical optimization method for targeting the poor under conditions of a fixed budget and missing information about household welfare. The study also compared subsidy targeting outcomes in Iran at the national and regional levels.
Materials and Methods
The study relied on three main categories of data: (a) data on per capita household expenditure excluding received subsidies, household size, and household population weights; (b) information on the economic and social characteristics of households; and (c) the poverty line. The data for the first and second categories was obtained from the Household Income and Expenditure Survey for urban and rural areas of the country for the year 1401 (2022/2023). The third category concerned the estimation of the poverty line, for which several points merit clarification. First, a daily intake of 2,300 kilocalories was used as the standard calorie requirement. This decision was based on the calculations that account for the calorie needs of different age and gender groups and their respective population shares. Second, the cost-of-basic-needs (food basket) approach was employed to calculate the poverty line. Third, using household equivalence scales was deemed unnecessary since the 2,300-kilocalorie threshold was derived from population-weighted calorie requirements across age and gender groups.
The Foster–Greer–Thorbecke (FGT) class of poverty indices was employed to measure poverty, which is defined as follows:
where z denotes the poverty line, yi represents the income of the i-th individual, and α reflects society’s degree of poverty aversion. For α = 0, α = 1, and α = 2, the index corresponds to the headcount ratio, the poverty gap, and the squared poverty gap (poverty severity), respectively.
In the next stage, the estimated poverty indices were used to classify urban areas across the country’s provinces using hierarchical clustering methods. Finally, a numerical optimization model was applied to compare subsidy targeting at the national and regional levels. Specifically, drawing on the existing literature, the study adopted a numerical optimization method to target poor households under conditions of a fixed budget and missing information about individual welfare. This approach could determine optimal transfer payments that maximize reductions in any additively decomposable poverty index, such as those in the FGT family of poverty measures.
Results and Discussion
According to the results, the characteristics associated with efficient targeting are identical at both the national level and across the eight urban regions—namely household size, the education level of the household head, and the number of children under six years of age. However, substantial differences exist in targeting efficiency as well as in inclusion and exclusion errors between the national and regional approaches. Among the characteristics, household size is the only variable that shows a strong and consistent correlation with poverty across all regions examined. Therefore, it should receive serious consideration in poverty alleviation policymaking, and greater coordination and alignment should be established between population policies, which are aimed at encouraging population growth, and poverty reduction policies.
Moreover, a comparison was made between national and regional targeting outcomes based on individual and combined household characteristics. The former included gender and marital status of the household head, household size, the education level and employment status of the head, housing status, and household demographic composition, such as the number of members aged six years or younger and those aged 65 years or older. Combined household characteristics consisted of the following: 1) divorced women, widows, or married women who are female household heads for any reason; 2) illiterate household heads; 3) households with five or more members; 4) households with at least one member under six years of age; and 5) households with unemployed heads. The comparison revealed substantial differences. These differences are evident in the population shares of target groups, the amount of subsidies paid, targeting efficiency, and the outcomes, particularly poverty reduction. Therefore, even when policymakers seek to target identical households at both the national and regional levels based on individual or combined characteristics, subsidy payment amounts to the same population groups must still vary across regions.
Conclusion
The findings indicated that several necessary conditions exist for replacing national targeting with regional targeting. However, implementing regional targeting requires careful consideration of the political, technical, and migration-related challenges, yet it is also essential to take into account the following key factors. First, a fundamental prerequisite for geographic targeting is the development of poverty maps at the smallest feasible geographic units within the country. Second, because of limitations in the precision of geographic targeting, this approach is rarely used in isolation for transfer payments involving large amounts. Geographic targeting should therefore be combined with other targeting methods to enhance efficiency and reduce errors. Third, it should not be assumed that poverty can be eradicated—or even substantially reduced—solely through subsidy payments. Actually, meaningful poverty reduction requires the implementation of comprehensive social development policies. In this regard, geographic targeting based on poverty mapping is not limited to the allocation of transfer payments; it can also be used to design and implement programs aimed at improving infrastructure and expanding access to social services at the regional level.
Keywords
Main Subjects
)