Research Article
Print
Research Article
Counting feminist economics: Areas, geography, and trends of research
expand article infoIrina E. Kalabikhina‡, Sofia M. Rebrey§
‡ Lomonosov Moscow State University, Moscow, Russia
§ MGIMO University, Moscow, Russia
Open Access

Abstract

This study employs bibliometric analysis to map the intellectual structure and evolution of feminist economics, using articles published in the journal Feminist Economics (1995–2021). Natural language processing and big data analysis — including descriptive statistics, frequency analysis, TF–IDF, LDA, and agglomerative clustering — reveal the patterns of knowledge production in the field. We apply two complementary clustering techniques: LDA, a probabilistic topic model, and agglomerative clustering, a hierarchical method. This cross-validates the research themes identified and mitigates the limitations of each approach. The result is a classification into five key areas: theory and methods, household and care economics, labor economics, open macroeconomics, and development economics. Analysis of key words, abstracts, and author data maps global knowledge production and reveals a highly uneven distribution, concentrated in the U.S. and other advanced English-speaking countries. Context analysis also captures distinct usage patterns for the terms “gender” and “feminist.”

Keywords:

NLP, bibliometrics, gender, context analysis, TF–IDF, LDA, Global South, Global North

JEL classification: B41, B54, J16.

1. Introduction

One of the key emerging heterodox approaches in economics is gender, or feminist, economics. It has been developing since the 1960s–1970s, although inquiries into the role of women in the economy began long before. According to A bibliography of female economic thought to 1940, there are “over 10,000 articles­, books and pamphlets on economic issues, written by more than 1,700 women, published between 1770 and 1940. There are works from academic journals and from popular magazines, treatises on public finance and on the ‘science’ of household management…” (Madden et al., 2004). Yet earning a PhD, writing academic papers, monographs, and articles, and collaborating with the most prominent economists of their time was not enough for women such as E. Kyrk, M. Reid, and many others to be included in mainstream academic discussion. Not until the women’s movement gained enough power in the second half of the twentieth century did research on women’s — and later gender — issues begin to be institutionalized through special journals, conferences, and eventually departments and faculties (Kalabikhina and Rebrey, 2022).

Hence the economics of gender, spurred by and developed within gender studies­ rather than economics, is now recognized as an economic discipline and has been incorporated into mainstream economics — yet much remains unclear, starting with the name of the discipline. The differences between feminist economics, gender economics, and the economics of gender are still a matter of discussion among economists (Jacobsen, 2020; Kalabikhina and Rebrey, 2022; Lundberg, 2022). According to Advanced introduction to feminist economics, “the economics of gender is about how gender differences lead to different economic outcomes… Feminist economics is a much broader intellectual project… that also considers how economics… has been shaped” (Jacobsen, 2020, p. 1).

The JEL classification system places feminist economics in category B5, Current Heterodox Approaches, and the economics of gender in J1, Demographic Economics (specifically J16, Economics of Gender; Non-labor Discrimination). Whether this classification reflects the actual area of inquiry, or instead constrains the field by confining gender research to demography, remains debatable.

This research traces the development of feminist and gender economics through a bibliometric analysis of the field’s key journal. The bibliometric ­approach identifies the key areas of feminist and gender studies in economics­, the key terms of the discipline, and its trends. Feminist Economics is an international peer-reviewed academic journal published since 1995 with a very high impact factor (the first quartile in SCImago). Sen (1998) called it “a leading journal in economics, vastly enriching the understanding of important economic issues” in his Nobel Prize lecture. It is treated here as the prime arena for inter­national academic discussion of gender economic research and perspectives, and therefore as a sufficient basis for analyzing the field. Gender and feminist scholars also suggest that “the publication of its journal Feminist Economics in 1995” symbolizes the “self-identified inception” of feminist economics, together with the formation of the International Association for Feminist Economics (IAFFE) in 1992 (Berik and Kongar, 2021, p. 3). Other researchers likewise find the journal a sufficient source for analyzing feminist economics: Tejani (2019), for example, uses Feminist Economics to assess the methodologies used in the field.

Our main contribution is to map the overall picture of research in the field. The key questions are:

RQ1: What are the key research areas addressed in Feminist Economics over 26 years of publishing activity (1995–2021)?

RQ2: What is the difference between feminist economics and gender economics?

RQ3: How do researchers’ countries of residence relate to their research geographies?

RQ4: What are the key trends in the areas of research in Feminist Economics?

2. Methodology and data

Bibliometric analysis has been developing since the last century; data analysis methods, and natural language processing (NLP) in particular, have broadened the paraphernalia of researchers and sharply increased the number of such studies. In economics and the other social sciences, bibliometric analysis is used “to grasp the global structure and morphology” (Claveau and Gingras, 2016). Our procedure is as follows: (1) data mining; (2) exploratory data analysis; (3) statistical analysis and visualization; (4) NLP; (5) discussion and future directions.

Data mining was performed with the web scraping library Beautiful Soup, parsing the journal archive at https://www.tandfonline.com/journals/rfec20. The dataset includes research articles published from 1995 to 2021 and omits reviews and other publications, as well as items published without abstracts and keywords. It consists of 708 rows (one row per research article) and the following fields: title, abstract, authors, authors’ country of residence, number of views, number of citations, keywords, JEL codes, whether the article is the first in the issue­, date of online publication, issue number, volume number, year, and whether the issue is a special issue (and its theme).

Exploratory data analysis was performed in Jupyter Notebook (Python). It covers the structure and dynamics of keywords and JEL codes, the most viewed and most cited authors and topics, and the topics of first articles (those prioritized by the editors). The classification of research areas by keywords is manual and rests on logic and expert judgment. The unit of observation is the keyword occurrence. Each keyword is attributed to exactly one area. Where a keyword contains markers of several areas, it is assigned to the first area in the ordering household, labor, development, globalization, theory and methods. This ordering is applied for reproducibility, and it affects 18 % of the matched keywords. Shares in Table 1 are computed over 2,455 attributed keyword occurrences and sum to 100% by construction.

Table 1.

Classification of key topics by area.

Theory and methods Household and care economics Labor economics Development Open macroeconomics
gender feminist inequality equity equality economics theory method model analysis household care time bargain family unpaid work-life balance reproductive (incl. fertility and family planning) marriage, couple, partnership violence labor force participation employ wage division of labor segregation sex work home-based workers labor union job firm economic development human development (education, health (incl. mortality), poverty, inequality) rural development and land rights gender budget and finance technology ecology discrimination (gender, age, nationalities, sexual orientation) agency austerity poverty globalization migration (and labor) export-oriented economy international organizations global south global pandemic migration industrialization crisis macro

To establish how researchers’ countries of residence relate to their research geo­graphies, the countries of residence were extracted from the author information, and the countries in the focus of the research were extracted from abstracts and keywords.

Statistical analysis and visualization were performed with the libraries pandas, NumPy, seaborn, Plotly, and spaCy, among others.

Natural language processing is a set of methods at the intersection of linguistics and computer science used to process and analyze large volumes of text (Bates, 1995). The term frequency–inverse document frequency (TF–IDF)1 algorithm reflects the importance of a word to a document within a collection (Rajaraman and Ullman, 2011). It is used as a weighting factor in search engines, text mining, content analysis, and sentiment analysis (Beel et al., 2016; Liang and Niu, 2022; Mee et al., 2021). Here it is used to identify the most important words and n-grams (sequences of two, three, or four words) in each abstract.

Context analysis is performed by counting the words that surround a term (four words before and four words after). Keywords and JEL codes are also used to identify the key research areas and their dynamics.

The development of machine learning, and of neural networks in particular, has sharply increased the capabilities of NLP and the popularity of the method (Goldberg, 2016; Goodfellow et al., 2016). One of the most popular clustering techniques in NLP is latent Dirichlet allocation (LDA). It builds a matrix of latent semantic indexing and probabilistic latent semantic indexing and then reduces the number of dimensions between words and their documents by adding a latent layer (latent clusters). The output is a clustering together with the most frequent and most representative words for each cluster (Duong et al., 2022). However, LDA works better with large datasets, which is why a second clustering technique is also applied. Agglomerative clustering2 recursively merges pairs of clusters of the sample data using a linkage distance (Tokuda et al., 2022).

The key research areas are identified through several techniques and indicators: statistical (the most used keywords and JEL codes, the most frequent words in abstracts), NLP (TF–IDF and context analysis), and machine learning (LDA and agglomerative clustering).

3. Results

3.1. RQ1: What are the key research areas in Feminist Economics?

“Gender” and “Economics of gender” are, respectively, the most frequent keywords and the most frequent JEL codes, since they identify both the area and the method of study. Together with such widely used keywords as “gender inequality” and “discrimination,” they form the basis for both philosophy and activism and are classified here as theory and methods.

Apart from theory and methods, there are four distinct areas of research: household and care economics, labor economics, development, and open macroeconomics3 (the global economy; see Fig. 1).

Fig. 1.

Areas of research by keywords. Source: Authors’ calculations.

The top ten JEL codes (Table 2) mirror a similar structure:

Table 2.

The top ten JEL codes.

JEL code Frequency
J16 Economics of gender • Non-labor discrimination 125
B54 Feminist economics 62
J22 Time allocation and labor supply 31
D13 Household production and intrahousehold allocation 24
J13 Fertility • Family planning • Child care • Children • Youth 22
J31 Wage level and structure • Wage differentials 19
D1 Household behavior and family economics 17
J21 Labor force and employment, size, and structure 16
J71 Discrimination 16
D63 Equity, justice, inequality, and other normative criteria and measurement 15
  1. • theory and methods: the economics of gender as the key method of gender analysis, together with non-labor discrimination, feminist economics, discrimination, and equity, justice, and inequality as the fundamental theoretical discussion;
  2. • microeconomic analysis: the household (time use and bargaining power) and the labor market;
  3. • macroeconomic analysis, represented by demography.

TF–IDF analysis (Table 3) shows the most important sets of words in an abstract and thereby identifies its topic. The labor market is the most popular area of study in the journal, accounting for almost two thirds of the abstracts and leaving far behind all the other sets of words, including such basic collocations as men/women.

Table 3.

TF–IDF n-grams.

2-grams 3-grams 4-grams
((“labor”, “market”), 97), ((“men”, “woman”), 63), ((“labor”, “force”), 63), ((“united”, “state”), 55), ((“woman”, “men”), 52), ((“woman”, “labor”), 47), ((“force”, “participation”), 44), ((“feminist”, “economics”), 43), ((“gender”, “inequality”), 42), ((“gender”, “gap”), 40), ((“paid”, “work”), 36), ((“gender”, “difference”), 34), ((“unpaid”, “work”), 30), ((“feminist”, “economist”), 30), ((“gender”, “equality”), 29), ((“woman”, “employment”), 29), ((“developing”, “country”), 26), ((“woman”, “worker”), 24), ((“bargaining”, “power”), 24), ((“care”, “work”), 23), ((“married”, “woman”), 23), ((“paid”, “unpaid”), 22), ((“time”, “use”), 22), ((“gender”, “wage”), 20), ((“economic”, “development”), 20), ((“economic”, “growth”), 20), ((“labor”, “force”, “participation”), 44), ((“woman”, “labor”, “force”), 25), ((“paid”, “unpaid”, “work”), 17), ((“woman”, “labor”, “market”), 14), ((“gender”, “wage”, “gap”), 13), ((“unpaid”, “care”, “work”), 10), ((“demographic”, “health”, “survey”), 9), ((“labor”, “market”, “participation”), 9), ((“time”, “use”, “survey”), 9), ((“woman”, “labor”, “supply”), 8), ((“gender”, “division”, “labor”), 8), ((“labor”, “market”, “outcome”), 8), ((“woman”, “bargaining”, “power”), 7), ((“current”, “population”, “survey”), 6), ((“intimate”, “partner”, “violence”), 6), ((“difference”, “men”, “woman”), 6), ((“issue”, “feminist”, “economics”), 6), ((“force”, “participation”, “rate”), 5), ((“gap”, “labor”, “force”), 5), ((“negative”, “effect”, “woman”), 5), ((“force”, “participation”, “lfp”), 5), ((“woman”, “labor”, “force”, “participation”), 22), ((“labor”, “force”, “participation”, “rate”), 5), ((“gap”, “labor”, “force”, “participation”), 5), ((“labor”, “force”, “participation”, “lfp”), 5), ((“woman”, “labor”, “market”, “participation”), 5), ((“intimate”, “partner”, “violence”, “ipv”), 4), ((“gender”, “gap”, “labor”, “force”), 4), ((“female”, “labor”, “force”, “participation”), 4), ((“household”, “income”, “labour”, “dynamic”), 4), ((“income”, “labour”, “dynamic”, “australia”), 4), ((“data”, “german”, “socioeconomic”, “panel”), 4), ((“german”, “socio-economic”, “panel”, “soep”), 4), ((“health”, “retirement”, “longitudinal”, “study”), 4), ((“china”, “health”, “retirement”, “longitudinal”), 4), ((“retirement”, “longitudinal”, “study”, “charls”), 4),

This set of words deserves special attention. Feminists usually pay attention to the order of the words “woman” and “man,” since “woman” conventionally comes second, and they make a point of putting it first — and they are close to success: “woman” comes first in 52 cases, against 63 cases in which it follows “man.”

The second area of feminist inquiry is the household level, which covers time use (the imbalance of paid and unpaid work, bargaining power) and intimate partner violence.

The third area is economic development and growth, which covers education, health, and retirement.

NLP and machine learning. LDA (Fig. 2) identified only labor economics as a distinct topic, finding it in 71% of the articles. Care economics, development, and globalization are split into smaller categories because of the specifics of the language used. On the one hand, LDA needs more texts to identify topics, which is why the topics here are not clearly separated. On the other hand, this classification still yields meaningful conclusions. The first and largest topic concerns the labor economy; the second is related to marriage and partnership in the urban environment in the context of women’s participation in the labor market. Trafficking is associated with girls and literacy rates, and immigration with population aging and satisfaction (with the use of migrant services).

Fig. 2.

LDA topics. Source: Authors’ calculations.

Agglomerative clustering identified six clusters (Fig. 3). The clusters contain similar sets of words but differ in the frequency of those words. In two clusters the most frequent word is “employment”; another cluster starts with “care.” The remaining clusters point to the macroeconomic level: “country,” “policy,” “social.” The other important words of the clusters can also be divided into these three groups, and each cluster contains most of them.

Fig. 3.

Agglomerative clustering. Source: Authors’ calculations.

What the editors consider most important is examined through the articles placed first in an issue and through the themes of special issues. The top twenty keywords of first articles (Fig. 4) are very similar to the general structure; the focus, however, shifts from labor economics to care economics, taking in such terms as time use, altruism, economics of the family, intrahousehold bargaining, and bargaining models, whereas labor economics is represented only by employment. Globalization and development, including empowerment, form another group of areas. The focus also shifts geographically: Africa overtakes China, and India drops out.

Fig. 4.

Top twenty keywords of first articles. Source: Authors’ calculations.

The list of special issues (Table 4) confirms the shift of focus to care, develop­ment, and globalization, and to the problems of the Global South, including Muslim communities.

Table 4.

Special issues

Topics of special issues Year
Rethinking Macroeconomic Modeling of Gender and Care 2022
Feminist Economic Perspectives on the COVID-19 Pandemic 2021
Reproductive Health, Care, and Employment 2020
The Care Economy, Gender, and Inclusive Growth in China 2018
Sex Work and Trafficking 2017
Voice and Agency 2016
Engendering Economic Policy in Africa 2015
Gender and Economics in Muslim Communities 2014
Land, Gender, and Food Security 2014
Critical and Feminist Perspectives on Financial and Economic Crises 2013
Gender and International Migration 2012
Unpaid Work, Time Use, Poverty, and Public Policy 2011
Unpaid Work, Time Use, Poverty, and Public Policy 2010
Inequality, Development, and Growth 2009
AIDS, Sexuality, and Economic Development 2008
Gender, China, and the World Trade Organization 2007
Women and Wealth 2006
Gender and Aging 2005

3.2. RQ2: What is the difference between feminist economics and gender economics?

Three terms have important peculiarities: economics of gender, gender economics, and feminist economics. The term “economics of gender” denotes a method of study that includes gender. The term “gender economics” denotes an economic approach that emerged from women’s studies in the late twentieth century by bringing men into the research lens and applying a feminist approach to stereotypically “normal” behavior and relations in society (Benería et al., 2016; Kalabikhina and Rebrey, 2022). Today, however, gender economics has become a mainstream approach that does not necessarily apply a feminist lens, whereas feminist economics is built on feminist philosophy.

M. Power proposes the term “social provisioning” as a basis for a feminist economics rooted in other heterodox traditions such as institutionalism, political economy, and the capability approach (Berik and Kongar, 2021).

According to Power’s approach, its five main components are “incorporation of caring and unpaid labor as fundamental economic activities; use of well-being as a measure of economic success; analysis of economic, political, and social processes and power relations; inclusion of ethical goals and values as an intrinsic part of the analysis; and interrogation of differences by class, race-ethnicity, and other factors” (Power, 2004). Feminist economics has remained a pluralistic project, encompassing mainstream and nonmainstream feminist economists and feminist scholars from diverse disciplines (Berik and Kongar, 2021). As Becchio puts it, “feminist economics, mainly influenced by women’s studies and feminism, rejected neoclassical economics, while gender neoclassical economics, mainly influenced by home economics and the new home economics, adopted the neoclassical economics’ approach to gender issues” (Becchio, 2021).

The journal Feminist Economics chose its name to underline the feminist approach as the central one. Is it possible to trace the difference between gender and feminist economics by bibliometric means? Analysis of the frequency of the terms and of their contexts sheds some light. The term “gender,” however, does not always imply feminist content. Feminist content is found in terms that describe feminist methods or feminist theories, such as “gender stereotypes,” and in terms that also rest on a feminist basis, such as “gender gap” and “gender inequality.” On the other hand, there are cases of neutral usage of the term “gender,” such as “population gender structure” and “gender differences.” Moreover, the word “gender” is used as both a noun and an adjective, whereas the word “feminist” is an adjective only.

Hence the term “gender” is twice as frequent as the term “feminist,” owing to the greater number of meanings and forms of the former and to the popularity and theoretical load of the latter. The frequency of the terms therefore does not really help to gauge the level of feminism in the journal.

The annual dynamics of frequency (Appendix, Figs A1–A6), however, show how the terms are used, and these trends are revealing. From 2006 onward the term “feminist” falls behind the term “gender.” This may signal a rise in empirical research and a decline in theoretical work. The observed divergence between “feminist” (theoretical) and “gender” (empirical and methodological) aligns with the broader shift of the journal toward empirical research identified under RQ4.

The context of the term “gender” (Appendix, Fig. A7) does reflect our classification into feminist and neutral usage, and the journal uses the term in a feminist context more often (gap, inequality, equality, wage, labor, and so on).

Over the entire period of the existence of the journal we observe only varying ­degrees of intensity of one context or another and minor shifts in the priority of the same contextual words — gap, inequality, equality, and so on (Appendix, Fig. A8).

The context of the term “feminist.” The term “feminist” refers more to methodology and theory, hence the proximity of such contexts as economics, economist, economic, research, perspective, and theory (Appendix, Fig. A9). There is no significant change in this context over time (Appendix, Fig. A10).

The context of other important terms. The feminist context is visible not only in the term “gender.” Analyzing the terms important for feminist analysis, we see contexts loaded with feminist meaning in the journal (Appendix, Fig. A10).

The context of the term “fertility” (Appendix, Fig. A11) is closely related to the employment of women, which reflects the underlying conflict between the private and the public spheres of a woman’s life. The annual use of the term “fertility­” is growing, though not steadily, reaching its maximum in 2020 (Appendix, Fig. A12). The last decade is the most focused on fertility issues, as they keep growing in importance on a global scale. Contextual analysis points to the labor market: how to combine employment and fertility, and how to strike a work—life balance. The second substantive point is teenage fertility. This is both a research question and one of the key policy issues.

The context of the term “violence” (Appendix, Fig. A12) is related to the experience of women and to the intimate partner as the source of violence. This discussion is viewed as a feminist achievement. Discussion of violence in the journal has intensified over the last decade (Appendix, Fig. A4). The context analysis supports the view that most episodes of domestic violence involve an intimate partner and predominantly affect women. Employment is not the least of the factors behind violence.

The issue of austerity was discussed most vigorously during the period of reduced economic growth after the 2008 crisis and in the context of the new normality and growing inequality (Appendix, Figs A6, A13).

Comparing the frequency of use of the terms “inequality,” “equality,” and “equity” (Appendix, Figs A1-A3), the use of the term “equality” is growing, which means that the positive connotation is becoming more prevalent. Concepts of fairness are discussed less frequently, as they have yet to be worked through on the rapidly growing body of empirical studies.

As for the use of JEL codes by the authors published in the journal, their frequencies contradict the idea that feminist economics is broader than the economics of gender. Among the 689 articles that carry JEL codes, J16 is considerably more common, appearing in 125 articles, while B54 appears in only 62, 40 of which carry both codes. Conversely, the large majority of the dataset (425 articles, or approximately 71,8%) contains neither B54 nor J16.

3.3. RQ3: How do researchers’ countries of residence relate to their research geographies?

Where are the researchers from? Data on the country of residence of researchers are extracted from the author information and are available for all the articles. Data on the country studied are extracted from keywords and abstracts. Not all papers focus on a single country, and not all mention the country studied in the abstract (the U.S. in particular), which is why these data are available only for a fraction of the articles.

More than a quarter of the articles are written in the United States, and they mostly concern domestic issues. A further 7% of the articles are written by researchers from the U.S. in cooperation with other countries, such as India (9 articles), China (6), the EU, and South Korea. The rest also come predominantly from developed English-speaking countries and European states.

Researchers from the South are a minority, particularly without a co-author from the North. Most of these articles are written by researchers from China and India, mostly in cooperation with Canada and the U.S., from Turkey (solo), and from African and Asian countries. These articles too mostly concern domestic issues. Regional or global analysis and international comparisons, particularly among the countries of the Global South, are, however, conducted by researchers from the South more often than by those from the North.

Which countries are more often represented among the first articles? The largest­ number of first articles are written in the U.S. (22.2%). A few papers are written by single authors from other developed and predominantly English-speaking countries. However, 60% of the articles are written in international collaboration, and most of them include emerging economies. The only emerging, non-English-speaking country with solo-authored first articles is Turkey.

Which countries are studied? The countries studied are extracted from the keywords and the abstracts, and these two methods yield different structures (Fig. 5). Whereas the keywords are dominated by countries of Asia (35% of the countries extracted from keywords, and 11% of the articles) and sub-Saharan Africa (20% and 6%, respectively), the abstracts are dominated by the United States and the other countries of North America and Oceania (27% and 18%) and by Europe (24% and 16%). The difference rests on the fact that most articles written in the U.S. concern domestic issues and do not mention the country in the keywords.

Fig. 5.

The regional structure of the countries studied, according to keywords (left) and abstracts (right). Source: Authors’ calculations.

More than half (51%) of the papers analyze the North and are written by researchers from the North (Fig. 6). The second group (19%) comprises articles in which the North explores the South. The South is also explored by the South (14%) and by the South together with the North (13%). There are also a few cases in which the North and the South together study the North (for example, researchers from Gabon and France study Belgium) and in which the North and the South together conduct a global analysis (1 article, 0.26%).

Fig. 6.

Who researches what? Note: North–North — researchers from the North explore the North; North–South — researchers from the North explore the South; South–South — researchers from the South explore the South; NS–S — researchers from the North together with researchers from the South explore the South; NS–N — researchers from the North together with researchers from the South explore the North; NS–NS — global analysis — the North and the South

Source: Authors’ calculations.

The areas of research in feminist economics change over time, and the research agenda mirrors the economic environment and economic shocks (Fig. 7, Appendix, Fig. A14). In the first year of the journal the focus was on family and childcare, whereas the market ranked only fifth. The late 1990s were focused on labor markets, until quality of life and childcare drew attention in 1999–2000. In 2003 the capability approach of Amartya Sen spurred the discussion of gender equality and development. Since then the labor market has come first only three times: in 2009, as the aftermath of the crisis, and in 2018–2019. Otherwise, time use and care economics have remained the key areas of research of the journal.

Fig. 7.

Key area per year. Source: Authors’ calculations.

The most used keywords per year (three times or more) show that the focus on family and care economics has been present throughout (Fig. 8). Development is represented by agency and empowerment. Labor economics stands out only in 2012–2013 and 2018.

Fig. 8.

The most used keywords per year. Note: The data were collected in 2022; therefore, the list of publications for 2022 is incomplete. Source: Authors’ calculations.

As the titles and keywords of the articles show, the research is becoming more and more empirical. Whereas the 1990s saw the creation of a theoretical foundation and the collection of data, the inquiry is now mostly empirical and increasingly intersectional. The bibliometric analysis conducted by the editors of Feminist Economics reaches the same conclusion in their own assessment and underlines that the trend toward quantitative approaches thrives despite the lower likelihood that such articles will be published: “However, quantitative empirical submissions, relative to other methodological approaches, are more likely to be desk rejected and less likely to be eventually published if they proceed to peer review” (Small and Braunstein, 2024). Tejani (2019) interviewed the editors of Feminist Economics, analyzed the shift of methodologies toward econometric approaches, and concluded that this shift is driven by feminist empiricism, ­institutional barriers to methodological pluralism, the background and training of researchers, and professional pressures.

4. Conclusion and discussion

The bibliometric analysis of 26 years of Feminist Economics (1995–2021) reveals­ a stable yet dynamic intellectual architecture. Contrary to claims that feminist economics is a monolithic or purely critical enterprise, we identify five core, coexisting research areas: theory and methods (27%), development economics (27%), household and care economics (23%), labor economics (18%), and open macroeconomics (5%). This pentagonal structure, validated through both manual classification and complementary machine learning techniques (LDA and agglo­merative clustering), demonstrates that the field has matured into a pluralistic yet coherent heterodox tradition. Notably, the near-parity between theory and methods and development economics suggests that the field does not separate abstract thinking from real-world problems. However, the marginal share (5%) of open macroeconomics — despite the stated priorities of the editors­ — reveals­ a persistent challenge: the field has yet to integrate feminist analysis sys­tematically into the core mechanisms of monetary policy, international finance, and global governance.

Our context and frequency analyses show that the term “feminist” has been in relative decline since 2006, while “gender” has proliferated. We interpret this as evidence of a strategic — yet potentially perilous — division of labor. The term “feminist” remains anchored in theoretical, philosophical, and normative discourse (for example, capabilities, social provisioning, power), whereas the term “gender” indicates a method of research that implies comparative analysis of male and female data. This means that gender analysis has gained legitimacy and technical sophistication. The mainstream, however, has absorbed “gender” as a method while bracketing feminist structural critiques. We therefore argue that the declining frequency of “feminist” within our own pages — correlated with the rise of empirical articles — risks losing the critical edge. The editors of Feminist Economics have noted that quantitative submissions face higher desk-rejection rates (Small and Braunstein, 2024), yet our data show that published articles have become more empirical. This paradox suggests a filtering mechanism that favors methodologically conventional work after it has passed peer review. The challenge is to ensure that methodological pluralism does not become a hierarchy in which the theoretical is implicitly devalued as “less rigorous.”

Our most politically consequential finding concerns the geography of research­. The United States — together with the other advanced English-speaking economies — dominates both the location of researchers (over one-quarter of articles) and the geographical focus of research, particularly when measured through abstracts. While keywords overrepresent the Global South (Asia at 35%, sub-Saharan Africa at 20%), abstracts reveal that the North (North America, Oceania, and Europe) remains the implicit universal subject of analysis. This divergence is itself a finding: the Global South is often tagged (through keywords) but not integrated into the main theoretical narrative (abstracts). Furthermore, we find that 51% of papers involve Northern researchers studying Northern contexts, and only 14% represent Southern scholars studying Southern contexts without Northern co-authors. This is not merely a descriptive inequality; it is a structural pattern of epistemic dependence. When the North studies the South (19% of articles), it often does so through Northern theoretical frameworks and research questions. Our analysis of first articles (editorial priorities), however, offers a counter-narrative: special issues and lead articles disproportionately focus on Africa, Muslim communities, and global crises. This suggests that the editorial vision is postcolonial, even if the submission and publication pipeline remains Northern-dominated. We conclude that feminist economics has not yet decolonized its production structures, even as it has decolonized its thematic agenda.

Moreover, according to research conducted by the editors of the journal and based on the comparison of submitted and published papers from both the Global North and the Global South, authors from the Global South engage more in ­empirical, mixed-methods, and intersectional analysis (Small et al., 2025).

The temporal analysis confirms that the research agenda of feminist economics­ is responsive to real economic shocks and policy regimes. The 2008 global ­financial crisis spurred a surge in work on austerity, which our context analysis links directly to rising inequality. The COVID-19 pandemic (visible in our data for 2020–2021) recentered care economics and the household as the locus of economic vulnerability and resilience. The sustained growth of fertility-related research over the last decade reflects global demographic anxieties and the persistent work–life conflict. And the term “violence” — increasingly discussed since 2010 — has been rightfully framed not as a private crime but as a labor market issue, an economic variable, and a violation of well-being. This responsiveness is a strength: feminist economics is an empirically grounded, politically engaged heterodoxy. Yet it also reveals a reactive posture. The 5% share of open macroeconomics suggests that the field is better at critiquing austerity than at proposing alternative macroeconomic frameworks ex ante. The exception is the emerging work on gender-responsive budgeting and care-inclusive macroeconomic modeling, which our data show to be concentrated in special issues rather than in the regular article flow.

Feminist economics has successfully built a recognized heterodox field with a distinct intellectual structure, rigorous methods, and global relevance. However, our bibliometric autopsy reveals three strategic vulnerabilities that require deliberate intervention: geography, macroeconomics, and methodology.

The dominance of the North in knowledge production is unjust and intellectually limiting. We call for a decolonizing praxis: (a) funding for South–South collaboration; (b) editorial policies that require authors to situate Northern theories as provincial rather than universal; (c) translation and open-access subsidies; and (d) the systematic citation of Southern scholars even when the case study is Northern.

The underrepresentation of open macroeconomics is a critical gap. If feminist economics cannot speak authoritatively to monetary policy, sovereign debt, capital flows, and global financial governance, it will remain a corrective to microeconomics rather than a full alternative.

The rise of quantitative methods is not a problem; the problem is the narrowing of what counts as evidence. Our agglomerative clustering and LDA results show that qualitative, historical, and institutional analyses still produce distinct and valuable clusters of knowledge. The journal should consider explicit methodological quotas or rotating special issues to protect this pluralism.

Finally, we acknowledge a limitation of our study: by focusing exclusively on Feminist Economics, we cannot fully trace how feminist ideas diffuse into mainstream journals. Future bibliometric research should compare our findings with the treatment of “gender” in the American Economic Review or The Economic Journal in order to measure epistemic seepage versus epistemic capture. Additionally, a qualitative, interpretive meta-analysis of the most cited and the most desk-rejected papers would illuminate the tacit criteria that shape the boundaries of the field.

References

  • Becchio, G. (2021). A history of feminist and gender economics. New York: Routledge.
  • Beel, J., Gipp, B., Langer, S., & Breitinger, C. (2016). Research-paper recommender systems: A literature survey. International Journal on Digital Libraries, 17(4), 305–338. https://doi.org/10.1007/s00799-015-0156-0
  • Benería, L., Berik, G., & Floro, M. (2016). Gender, development, and globalization: Economics as if all people mattered. New York: Routledge.
  • Duong, Q. H., Zhou, L., Meng, M., Nguyen, T. V., Ieromonachou, P., & Nguyen, D. T. (2022). Understanding product returns: A systematic literature review using machine learning and bibliometric analysis. International Journal of Production Economics, 243, 108340. https://doi.org/10.1016/j.ijpe.2021.108340
  • Goldberg, Y. (2016). A primer on neural network models for natural language processing. Journal of Artificial Intelligence Research, 57, 345–420. https://doi.org/10.1613/jair.4992
  • Goodfellow, I., Bengio, Y., & Courville, A. (2016). Deep learning. Cambridge, MA: MIT Press.
  • Jacobsen, J. P. (2020). Advanced introduction to feminist economics. Cheltenham; Northampton, MA: Edward Elgar.
  • Small, S. F., Sinha, A., Mackett, O., & Braunstein, E. (2025). Diversity in feminist economics research methods: Trends from the Global South. Journal of Economic Methodology, 1–25. https://doi.org/10.1080/1350178X.2025.2460221

Appendix

Fig. A1.

Frequency of the term “equality.” Source: Authors’ calculations.

Fig. A2.

Frequency of the term “inequality.” Source: Authors’ calculations.

Fig. A3.

Frequency of the term “equity.” Source: Authors’ calculations.

Fig. A4.

Frequency of the term “violence.” Source: Authors’ calculations.

Fig. A5.

Frequency of the term “fertility.” Source: Authors’ calculations.

Fig. A6.

Frequency of the term “austerity.” Source: Authors’ calculations.

Fig. A7.

Context analysis of the term “gender.” Source: Authors’ calculations.

Fig. A8.

The context of the term “gender” per year. Source: Authors’ calculations.

Fig. A9.

Context analysis of the term “feminist.” Source: Authors’ calculations.

Fig. A10.

The context of the term “feminist” per year. Source: Authors’ calculations.

Fig. A11.

The context of the word “fertility.” Source: Authors’ calculations.

Fig. A12.

The context of the word “violence.” Source: Authors’ calculations.

Fig. A13.

The context of the term “austerity.” Source: Authors’ calculations.

Fig. A14.

The most frequently used words in abstracts per year. Source: Authors’ calculations.

1 TF–IDF is a statistical measure used to evaluate how important a word is to a document in a collection, or corpus (a large set of documents). Its importance increases proportionally to the number of times the word appears in the document but is offset by how common the word is across all documents.
2 Agglomerative clustering is a “bottom-up” approach to cluster analysis in which each data point starts in its own cluster, and pairs of clusters are merged as one moves up the hierarchy, until all the points are in a single cluster.
3 Open economy macroeconomics refers to the study of how macroeconomic variables such as incomes, prices, and exchange rates are determined in an economy that interacts with other economies, focusing on the effects of international trade and finance (Melvin and Norrbin, 2023).
login to comment