Despite the government's policies and efforts to boost fertility rates in Iran, the total fertility rate has been declining over the past two decades. This research aims to identify the weight of determinants of fertility in Iran's provinces and provide relevant policy recommendations. The study utilizes the mathematical methods of Shannon entropy and K-means clustering. The findings reveal that per capita gross product holds the highest weight coefficient among fertility determinants across the provinces. Furthermore, the influence of economic, social, and demographic indicators on fertility varies among provinces. Provinces with higher coefficients in social and demographic indicators tend to exhibit higher fertility rates. Notably, provinces such as Sistan and Baluchistan, North Khorasan, and South Khorasan, despite low economic indicators, maintain fertility levels above replacement level. In conclusion, this research emphasizes the significance of understanding localized determinants of fertility and advocating for tailored policy interventions at the provincial level. Hence, a nuanced approach to regional politics becomes imperative. Furthermore, provinces like Alborz, Semnan, Gilan, Kermanshah, Ilam, and Isfahan, which score low on the overall index, should be prioritized for targeted population policies.
 
Extended Abstract
Introduction
The population and fertility rates have long been a subject of concern for policymakers and researchers. Total fertility rate (TFR) has significantly declined over the past three decades in Iran. TFR dropped from 7 children per woman in the 1960s to 1.65 in 2023. This decline is attributed to various factors including increased female education, delayed marriages, greater independence for women, and the success of family planning programs. Despite government efforts to reverse this trend through pronatalist policies, such as financial incentives and family support programs, fertility continues to decline. This suggests that a more nuanced, regionally tailored approach is needed, as fertility determinants vary across Iran’s diverse ethnic and cultural regions. The study aims to explore these regional differences and examine whether uniform policies can effectively raise fertility levels across the country.
 
Methods and Data
This research is quantitative and employs the Shannon entropy method. The data was sourced from the Statistical Center of Iran, specifically the 2016 Population and Housing Census. The statistical population includes all provinces of Iran (31 provinces), with each province serving as the unit of analysis. Initially, the key factors influencing low fertility were identified through brainstorming and a review of previous research. Using Shannon’s entropy method, the weight of each factor was calculated and interpreted for all 31 provinces.
The weights assigned to each index were multiplied by the corresponding index values to calculate the score for each province in that specific function. Finally, the provinces were categorized based on their evaluation scores using k-means clustering. The clustering was then performed based on the provinces' scores in economic, social, and demographic functions.
 
Findings
Compared with 2016, the TFR in 2021 is much lower, although the government has taken some serious measures toward increasing fertility over recent decades. TFR decreased from 1.65 in 2016 to 1.2 in 2021, and such a reduction also has taken place for all provinces except Khuzestan. In 2021, Sistan and Baluchistan, South Khorasan, and Khuzestan had fertility rates above the replacement level. The provinces of Golestan, North Khorasan, and Razavi Khorasan had fertility rates near replacement level (2 to 2.1 children), while the remaining 25 provinces had fertility rates below replacement.
Table 1 presents the calculated weighting of fertility determinants using the entropy method. The results indicate that GDP per capita holds the highest rank and carries the most weight. Conversely, the ratio of women to men, with the lowest weight, is ranked last.        

Table 1- Normalized weight for each index using the Shannon entropy method

 
The provinces of Tehran, Khuzestan, and Bushehr have the highest economic ranks, while Chaharmahal and Bakhtiari, Sistan and Baluchistan, and North Khorasan rank the lowest. Of Iran's 31 provinces, only 11 have an economic score above the national average. Regarding economic clustering, 23 provinces fall into the third cluster (indicating poor conditions), while only six are in the first cluster, representing the best conditions. These findings highlight significant economic disparities across the provinces.
Regarding social function, the provinces of Sistan and Baluchistan, North Khorasan, and South Khorasan rank the highest, while Tehran, Semnan, and Alborz rank the lowest. Regarding social clustering, West Azarbaijan, Bushehr, and Qom are in the first cluster (better conditions), whereas Tehran, Semnan, and Alborz fall into the third cluster (weaker conditions). Regarding demographic function, Sistan and Baluchistan, Hormozgan, and North Khorasan hold the highest ranks, while provinces like Tehran and Isfahan rank among the lowest. In total function, Sistan and Baluchistan rank first, while Gilan, Semnan, and Alborz rank the lowest.
Finally, it is important to note that, given the significant ethnic and religious diversity across Iran's provinces, it was expected that ethnic and religious minorities would cluster together based on economic, social, and demographic factors. However, the observed clustering appeared to be more geographically based.
 
Conclusion and Discussion
Policies aimed at increasing fertility should consider both the economic and social diversity among provinces, as uniform policies may not yield the desired results. Given the provincial differences in fertility rates, each province should adopt policies that align with its specific demographic realities. Therefore, the unique conditions of each province or region, based on various indicators, must be considered when formulating policies. The second point is that, based on the overall index's function, we can conclude that in provinces where a balance among the indicators is maintained (such as Khuzestan), there is hope that the fertility rate will remain above replacement level and not decline. Additionally, considering the strong correlation between social indicators and total fertility rates, it can be inferred that if national policy shifts towards a regional focus—emphasizing the most influential indicators in each province—the potential for increasing TFR is within reach.
Based on these findings, it is recommended to implement a population policy targeting the central and northern provinces, with a focus on prioritizing and planning efforts to improve key social indicators. These include increasing the overall marriage rate, reducing divorce rates, lowering the average age of marriage for both men and women and providing support packages for families. These efforts should be coordinated across various departments. Second, the economic and demographic indicators in the eastern and western provinces of the country were less favorable. It is therefore necessary to closely monitor and address these indicators in these regions to stabilize and improve the total fertility rate.
 
Acknowledgments
This article is based on the master's thesis of the third author in the field of demography, completed at the Faculty of Economic and Social Sciences, Bu-Ali Sina University. We would like to express our sincere gratitude to the esteemed referees of both the thesis and the article.