Behavioral finance: a four-decade bibliometric exploration

Over the past four decades, behavioral finance has evolved from a fringe idea to a mainstream field of study, reshaping the understanding of financial markets and investment behavior (Prorokowski and Roszkowska, 2014). This evolution has been stimulated by a growing body of research spanning numerous academic disciplines, including psychology, economics and finance (Moosa, 2011; Wang, 2020). As the field has matured, scholars have produced a wealth of literature investigating different aspects of investor behavior, from the impact of emotions on investment decisions to the role of cognitive biases in market anomalies (StatMan, 2018). Behavioral finance attempts to explain how financial choices are made in real-world contexts and why their choices might not appear rational every time and therefore have unpredictable consequences (Loang and Ahmad, 2021). This contrasts with many traditional theories that assume investors make rational decisions (Agarwal et al., 2025; Albastaki et al., 2020; Barber and Odean, 2000).

This bibliometric paper aims to provide a comprehensive overview of the trends, patterns and themes that have characterized behavioral finance research over the past forty years. By systematically analyzing the literature, the study seeks to identify key milestones, influential authors, seminal works and emerging areas of interest within the field (Alam et al., 2025; Do Nascimento Sancheta et al., 2021; Calma, 2019; Kumar and Kumar, 2023). Through quantitative techniques such as citation analysis, co-authorship networks and keyword clustering, this research aims to uncover the underlying structure of behavioral finance research and shed light on its trajectory over time (Mongeon and Paul-Hus, 2016; Makarenko et al., 2022; Barber and Odean, 2001; Bollen et al., 2011; Nguyen et al., 2024).

While prior literature reviews (Chaudhary and Sura, 2025) have offered valuable insights, a comprehensive scientometric review spanning the entire four-decade evolution of the field (1984–2024) is currently lacking in the literature. The extended horizon proposed in this study provides unique insights by tracing the path from the field's genesis (prospect theory) to its contemporary convergence with FinTech. This paper builds upon and benchmarks against earlier bibliometric contributions to behavioral finance (Ohlan and Ohlan, 2022; Hsu and Marques, 2022; Shaikh and Khan, 2025) by offering increased methodological depth and temporal scope.

Despite the proliferation of behavioral finance research, two significant gaps remain. Theoretically, while existing bibliometric studies have focused on specific sub-periods or narrow themes, there is a lack of a longitudinal synthesis tracing the intellectual evolution of the field across a continuous four-decade horizon (1984–2024). This study addresses this gap by mapping the transition from foundational heuristics to modern digital paradigms (Barbić et al., 2019; Beketnova, 2020; Bialkowski and Starks, 2016; Bikhchandani et al., 1992). Practically, existing reviews often fail to translate complex science-mapping results into actionable insights for the digital finance era. This research bridges that gap by providing a framework that practitioners can use to understand the behavioral underpinnings of FinTech, robo-advising and algorithmic trading (Singh and Saini, 2025b; Thaler, 2016; Thaler and Sunstein, 2008; Venkatesh et al., 2003).

The significance of such an analysis lies in its potential to offer valuable insights to researchers, practitioners and policymakers alike. By mapping out the intellectual landscape of behavioral finance, it is possible to gain a deeper understanding of the foundational concepts that have shaped the field and the directions in which it is heading (Disatnik and Steinhart, 2015; Do Nascimento Sancheta et al., 2021; Donthu et al., 2021; Ohlan and Ohlan, 2022; Beketnova, 2020). This, in turn, can inform future research agendas, guide investment strategies and facilitate the development of policies aimed at fostering more informed and rational decision-making in financial markets.

The contributions of this study are categorized into three distinct pillars to address the multifaceted needs of the field:

1. Empirical contributions: This study provides a large-scale data synthesis of over 2,000 documents, utilizing co-citation and co-word analyses to objectively quantify growth trends and identify burst topics. By employing quantitative metrics such as the compound annual growth rate for specific clusters, it offers a rigorous, evidence-based map of the literature's expansion.

2. Theoretical contributions: The conceptual understanding of the field is advanced by documenting the evolutionary trajectory from classic heuristics and prospect theory to the current convergence with neurofinance and behavioral artificial intelligence (AI). This provides a unified theoretical framework that connects foundational psychological tenets with modern digital financial behavior.

3. Policy-related contributions: The findings indicate a strategic roadmap for financial regulators and policymakers to implement soft nudges and behavioral guardrails. By identifying the persistent nature of specific biases in digital environments, this study facilitates the design of policies aimed at enhancing investor protection and promoting consumer financial well-being.

For decades, the financial world operated under the assumption of the rational investor: a mythical being that makes flawless decisions based solely on logic (Barbić et al., 2019; Leoni, 2009; Kahneman and Tversky, 1979). However, persistent anomalies in financial markets have challenged this notion. These anomalies hinted at a reality where markets were not perfectly efficient but susceptible to biases that sway investor behavior. This fertile ground gave rise to behavioral finance, a revolutionary field that delves into the intriguing world of investor psychology.

Prior to the emergence of behavioral finance, modern finance reigned supreme for most of the 20th-century. This dominant theory, championed by authors such as Sharma (2023), Shiller (2003), Paule-Vianez et al. (2020), Hussein (2021) and Falkowski (2011), assumed markets to be efficient, with prices reflecting all available information. Nevertheless, the cornerstone of modern finance–investor rationality–came under fire. Psychologists (Flin, 1997) argued for the limitations of rational choice models, highlighting the influence of human heuristics and biases. Building on this foundation, Olsen (1998) delivered a powerful critique of expected utility theory, drawing on prospect theory to account for irrational behavior under uncertainty.

Nonetheless, how has the understanding of investor psychology evolved? While previous literature reviews have offered valuable insights, the shear volume of behavioral finance research can make it difficult to grasp the broader picture (Campbell and Ramadorai, 2024; Çera et al., 2021; Farooq, 2022; Feldman and Liu, 2023; Hendershott et al., 2011; Hirshleifer, 2008). This study aims to bridge this gap by employing a powerful tool: bibliometrics.

Over the next few sections, the study traces a 40-year trajectory of behavioral finance research using bibliometric analysis. By examining large-scale research data, identifying key themes and revealing the intellectual currents shaping the understanding of investor behavior, the study explores how the field has evolved over the past four decades. This approach goes beyond traditional literature reviews by employing specialized software tools to conduct a quantitative assessment of research data (Gu et al., 2020; Gulati et al., 2025).

One pivotal discovery lies in the prevalence of cognitive biases among investors, frequently leading to suboptimal choices (Seamster, 2018; Compen et al., 2022). Extensive research has elucidated the impact of biases such as overconfidence, loss aversion and anchoring on investment behavior (Li et al., 2021; He and Liu, 2018). These biases often result in investors engaging in excessive trading, holding onto losing positions for too long, or making decisions based on irrelevant information, ultimately undermining their financial objectives (Blackburn and Ukhov, 2013; Goudarzi et al., 2015; Barber and Odean, 2000).

Besides, behavioral finance scholarship has a significant role of emotions in driving financial decisions (Hitt and Tambe, 2016). Studies have revealed that emotions such as fear, greed and regret influence investor behavior and market dynamics (Franco and Mahadevan, 2021). During market downturns, fear of losses frequently triggers panic selling among investors, exacerbating downward pressure on prices (Shiller, 2003). Conversely, periods of exuberance fueled by greed can precipitate speculative bubbles and subsequent market crashes (He et al., 2019; Agarwal et al., 2025). Recognizing the influence of emotions on investment decisions is vital for investors and financial professionals alike, as it enables them to develop strategies to manage emotions effectively, maintain discipline during turbulent market conditions and avoid succumbing to irrational exuberance or unwarranted pessimism (Yasmin and Ferdaous, 2023).

A prior bibliometric analysis suggested several avenues for further exploration within the field of behavioral finance. One area of interest highlighted is the application of advanced quantitative methods, such as machine learning and network analysis, to enhance the current understanding of investor behavior and market dynamics (Hirshleifer et al., 2006; Hsu and Marques, 2022; John et al., 2019). These techniques offer the potential to uncover intricate patterns and relationships within financial data that traditional statistical methods may overlook, thereby providing new insights into the underlying mechanisms driving market phenomena.

Figure 1 shows connecting lines between nodes that indicate a relationship between the papers. Thicker lines signify a stronger connection, suggesting that the papers are highly relevant to each other. These connections help visualize the intellectual conversation between different research efforts. Each node represents an academic paper related to the original paper that the study is exploring. These papers are not arranged chronologically (by publication date) but rather by similarity. Papers addressing related topics and potentially employing similar methodologies are clustered in the graph.

Table 1 summarizes the most influential prior studies in behavioral finance, highlighting seminal contributions, publication outlets, citation impact and primary research focus. The studies listed illustrate the foundational themes of market inefficiencies, investor biases and limits to arbitrage that have shaped behavioral finance research over the past four decades. Moreover, the analysis proposed examining the implications of technological advancements, such as algorithmic trading and robo-advisors, on investor behavior and market efficiency (D’Acunto et al., 2019; Hendershott et al., 2011). These innovations alter the financial landscape, potentially affecting the prevalence of behavioral biases. The present study advances the field by employing a bibliometric analysis of more than 7,053 peer-reviewed articles indexed in Scopus over a 40-year period. This comprehensive dataset enables the examination of underlying patterns and trends to address the following key research questions (RQs):

Substantively, this paper provides a critical synthesis of conflicting findings within the behavioral finance literature, analyzes regional heterogeneity in collaboration and defines a structured research roadmap based on the rise of algorithmic behavioral finance and financial literacy.

1. Methodological contributions

   • Enhanced data rigor and scope: A large, filtered Scopus dataset of 7,053 peer-reviewed articles spanning four decades (1984–2024) is leveraged, thereby providing an unprecedented temporal scope for quantitative analysis.

   • Integrated science-mapping approach: A novel combination of co-authorship networks, thematic burst detection and refined VOSviewer thematic mapping is employed, justified by explicit parameters, to visualize the intellectual structure with greater precision.

   • Transparency and reproducibility: The exact, filtered search query is provided, along with a detailed description of the criteria used for database selection (Scopus vs Web of Science (WoS)) and data filtering, enhancing the study's methodological transparency.

2. Substantive contributions

   • Critical synthesis of conflicting findings: A critical synthesis of conflicting findings within the behavioral finance literature is provided, leading to a detailed perspective on core behavioral anomalies across diverse contexts.

   • Mapping regional heterogeneity: Regional and institutional heterogeneity in collaboration networks is examined, highlighting differing thematic focuses (e.g., a US focus on micro-level biases versus an Asian focus on FinTech applications).

   • Identifying future avenues: A structured and empirically-driven roadmap for future research is defined based on the identified emerging themes, particularly the integration of algorithmic behavioral finance and neurofinance.

This research employs bibliometric analysis techniques to study the evolution and current landscape of behavioral finance research.

1. To evaluate publication trends over a 40-year period to understand the evolution of key themes, scientific production, emerging trends and shifts in focus within behavioral finance research.

2. To identify key influences and networks of the most impactful publications, influential authors and co-citation networks that have shaped the field's development.

Behavioral finance represents a paradigm shift from traditional finance by integrating insights from psychology, economics and sociology to explain the often irrational behaviors of market participants (Mushinada and Veluri, 2019). Unlike traditional finance, which assumes rationality, behavioral finance acknowledges that decisions are influenced by cognitive biases, emotions and social factors (Dugdale, 1996; Muttar et al., 2021; Disatnik and Steinhart, 2015). The field is defined as the study of how psychological factors and biases affect financial decision-making, asset prices and market outcomes. This approach is crucial as it provides insights into market anomalies (Nwogugu, 2019) and serves as a bridge between disciplines (Singh and Saini, 2025a, b), fostering collaboration and innovation by integrating bibliometric insights into traditional financial frameworks (Clemente-Almendros et al., 2023; Hsu and Marques, 2022).

The research methodology follows a stringent, multi-stage process, as visually represented in Figure 2, methodological framework. This framework ensures that the study is systematic, transparent and reproducible, which are critical requirements for a high-quality scientometric review. This approach is grounded in established best practices for bibliometric analysis, ensuring the rigor required for mapping an interdisciplinary field over an extended period.

The design is based on a funnel approach, where the process begins with a broad search and progressively filters the data to arrive at a highly relevant and refined corpus suitable for advanced analysis.

This study employs a scientometric review based on a bibliometric analysis design. The quantitative approach is well-suited to mapping the intellectual landscape of a mature and expansive field such as behavioral finance over a four-decade period (1984–2024). The method is adopted to provide a rigorous, objective and reproducible analysis of the structure and evolution of the literature. The design aligns with the latest standards for methodological rigor (Clemente-Almendros et al., 2023; Ribeiro-Navarrete et al., 2022) and utilizes the robust analytical framework outlined by leading contemporary studies (Hsu and Marques, 2022; Singh and Saini, 2025a, b). This enables the identification of influential authors, journals and thematic shifts through statistical analysis of citation patterns and keyword co-occurrence.

The research was performed following the systematic framework presented in Figure 2. Methodological framework. Figure 2 illustrates the multi-stage methodological framework adopted in this study, outlining the systematic process of data identification, screening, refinement and bibliometric analysis used to examine behavioral finance research.

• Stage 1: Search with multiple criteria. In this stage, a comprehensive search is conducted using Boolean operators to combine search terms related to the research topic. Keywords such as “behavioral finance”, “investor psychology” and “financial decision-making” are used for retrieving relevant documents. The search query results in a total of 18,729 documents matching the specified criteria.

• Stage 2: Search with the Scopus database. The search is refined by using the Scopus database, which offers a vast collection of scholarly literature across various disciplines. Scopus allows for more precise filtering and access to high-quality research publications.

• Stage 3: Refinement criteria. Further refinement of the search is performed based on additional criteria to ensure the selection of relevant and high-quality research articles. These refinement criteria include limiting the publication date range (1984–2024), restricting the search to specific subject areas (economics, business, management and accounting), specifying the document types (articles, book chapters) and limiting the search to a particular language (English).

• Stage 4: Co-citation, network, co-word and co-authorship analyses. Many bibliometric analyses are conducted to explore relationships between documents, with the aim of identifying connections among different research areas within behavioral finance, investor psychology and financial decision-making. This approach provides insights into the interconnectedness and evolution of research themes and concepts within the field (Jianakoplos and Bernasek, 1998; Chauhan and Yadav, 2024).

• Stage 5: Research articles selected for the study. After applying the various filters and refinement criteria, a final set of research articles is selected for further analysis. From the initial pool of 18,729 documents, a subset of 8,729 documents, 7,053 peer-reviewed articles may be identified as meeting the specified criteria and deemed suitable for inclusion in the study.

• Stage 6: Exporting the data. The final selected data is exported in a suitable format, such as Comma-Separated Values (CSV), for further analysis using appropriate tools, including R and VOSviewer for bibliometric analysis. This step ensures that the data are organized and ready for in-depth examination and exploration of research trends, patterns and relationships within the chosen research domain.

To ensure the robustness and comparability of this analysis, the methodology is specifically benchmarked against leading literature in scientometrics and business research:

• Data sourcing and filtering: The systematic approach to data collection was informed by precision-focused protocols demonstrated in recent large-scale mappings (Clemente-Almendros et al., 2023; Ribeiro-Navarrete et al., 2022).

• Analytical procedures: The technical execution, including the selection of co-word analysis and network thresholds, adheres to the authoritative guidance of Singh and Saini (2025a, b) and .

• Mapping and interpretation: The approach to constructing and interpreting thematic maps and collaboration networks aligns with rigorous models recently applied in the behavioral domain (Hsu and Marques, 2022; Singh and Saini, 2025a, b).

This methodical, multi-stage design ensures that the data collection is systematic, the scope is precisely defined and the resulting analysis is robust and directly addresses the established RQs.

Table 2 outlines the criteria for inclusion and exclusion in a literature search related to behavioral finance, investor psychology and financial decision-making within the fields of business, management and accounting.

Search query: ((TITLE-ABS-KEY(behavioral AND finance) OR TITLE-ABS-KEY(investor psychology) OR TITLE-ABS-KEY(financial decision-making)) AND PUBYEAR >1983 AND PUBYEAR <2025 AND (LIMIT-TO (SUBAREA,“BUSI”) AND (LIMIT-TO(DOCTYPE,“ar”) AND (LIMIT-TO(LANGUAGE,“English”) AND (LIMIT-TO(SRCTYPE,“j”) OR LIMIT-TO(SRCTYPE,“b”) OR LIMIT-TO(SRCTYPE,“p”)) AND (LIMIT-TO(PUBSTAGE,“final”)).

The search query was specifically designed to target only peer-reviewed journal articles to ensure the highest quality and comparability of the analyzed corpus, thus excluding conference papers, book chapters and books. The search query provided ensures the inclusion of documents containing specific keywords related to behavioral finance, investor psychology and financial decision-making, while adhering to the specified criteria for inclusion and exclusion across various aspects such as publication date, document type, source type, language, subject area and publication stage.

The Scopus database was selected as the sole source for data extraction. This choice is grounded in its superior handling of metadata necessary for comprehensive scientometric mapping, a preference supported by recent methodological benchmarks in the field (Singh and Saini, 2025a, b; Ribeiro-Navarrete et al., 2022). The search covered a four-decade span from 1984 to 2024 to trace the field's evolution comprehensively.

Limitation and sensitivity check: This paper acknowledges the methodological limitation of using a single data source, which may omit influential works exclusively indexed elsewhere. However, this focused approach maintains consistency in citation indexing and data quality standards (Ribeiro-Navarrete et al., 2022). To validate the dataset's representativeness, a parallel search of the field's core keywords was conducted in the WoS core collection. This sensitivity check confirmed no major shift in the top-cited foundational documents, suggesting that the Scopus dataset is a reliable representation of the field's core intellectual structure.

This initial search, using a broad document-type query, yielded 8,729 documents (articles, books, chapters and conference papers) published in English within the business, management and accounting subject areas.

For rigorous quantitative analysis, the dataset was refined by applying strict inclusion criteria regarding document type, a necessary step to focus the analysis on peer-reviewed, completed research articles.

Table 3 presents the inclusion and exclusion criteria applied to different document types in the Scopus database, outlining the methodological rationale for retaining only peer-reviewed journal articles in the bibliometric analysis.

Table 4 illustrates that the database covers a 40-year period, ranging from 1984 to 2024. The information is drawn from 2,187 different sources, including journals, books and conference proceedings. A total of 8,729 documents were selected for analysis after applying the filters. The number of documents showed an average annual growth rate of 8.95% within the selected timeframe. On average, the documents are 8.88 years old. Each document was cited 22.58 times, with a total of 393,672 references across all documents.

An examination of the author details shows that 18,394 unique authors contributed to the documents. Notably, 1,724 are single-authored. Collaboration seems to be fairly common, with an average of 2.56 authors per document. Additionally, 20.17% of the documents involve international co-authorship, indicating collaboration between researchers from different countries. Finally, the database includes a variety of document types. Articles constitute the majority (7,053), followed by books (327), book chapters (624) and conference papers (725).

The collected dataset of 7,053 documents was processed and analyzed using a combination of statistical and visualization software to quantify performance metrics and map the intellectual structure of behavioral finance research.

The primary software tools utilized were RStudio–specifically the bibliometrix package–for comprehensive statistical processing and calculating performance indicators and VOSviewer for constructing and visualizing co-citation, co-authorship and co-word networks.

The analysis was structured around three key methodological approaches:

1. Performance analysis

   • This involved calculating quantitative metrics to assess the output and impact of the research over the 40-year period. Key metrics included:

   • Scientific productivity: Measurement of annual publication volume (Figure 3) and the output of the most prolific authors and sources (Table 6).

   • Impact metrics: Calculation of total citations (TC), mean total citations per article and mean total citations per year since publication for the entire dataset (Table 4) and for individual highly cited articles (Table 5).

2. Network analysis and visualization

   • VOSviewer was used to construct maps of the social and intellectual structures. To ensure analytical robustness and methodological rigor, the following specific parameters and counting methods were applied for network creation. These parameters, particularly the use of modularity-based clustering and fractional counting, are consistent with the high-precision workflows recently validated by and Singh and Saini (2025a).

   • Co-authorship analysis: Calculated using full counting to assign equal weight to each document, with a minimum threshold of three documents per author. The network was normalized using the association strength method.

   • Co-citation analysis (documents/sources): Calculated using fractional counting to accurately reflect interdisciplinary connections, with a minimum threshold of three citations per document or source.

   • Co-word/keyword analysis: Calculated using full counting with a minimum threshold of five keyword occurrences. The resulting thematic map was clustered using the modularity-based clustering algorithm for optimal thematic grouping.

3. Thematic mapping and trend analysis

The bibliometrix package in R was used to construct a thematic map (Figure 19). This mapping visualizes the conceptual landscape based on two key dimensions: centrality (measuring the degree of external connections and the theme's developmental importance) and density (measuring the strength of internal connections and a theme's maturity). Additionally, a trending topic analysis (Figure 18) was performed on Keywords Plus and author keywords to identify shifts in research focus over time, including the emergence of big data and FinTech.

To ensure a logical progression from the methodological framework, this section presents the empirical findings derived from the scientometric analysis of 7,053 documents. The results are structured to address the study's research objectives, specifically evaluating longitudinal trends, impact metrics and the shifting thematic landscape of behavioral finance.

The field's growth is first evaluated to address Research objective 1, which focuses on understanding the evolution of scientific production and emerging trends over four decades.

Research objective 1: To evaluate publication trends over 40 years to understand the evolution of key themes, scientific production, emerging trends and shifts in focus within behavioral finance research.

The analysis then shifts to publication venues to identify where the intellectual discourse is most concentrated. Figure 4 examines the average citations per year, which shows a generally increasing trend from 1994 to 2014, despite minor fluctuations. This indicates a growing, sustained impact of research in the field over time.

Table 5 details the evolution of citation metrics (1984–2024), providing quantitative evidence of the field's longitudinal influence. While total citations (TC per article) show volatility in the early years–peaking in 1986 (88.76) and 1998 (146.97)—this pattern reflects the enduring influence of a small number of seminal foundational papers.

To provide a more rigorous assessment of contemporary impact, the mean total citations per year since publication (ACPY) is included. This metric normalizes citation counts by article age. Two critical phases are identified based on this measure:

1. Peak foundational impact: The highest period of sustained influence occurred between 1998 (5.44 ACPY) and 2001 (4.61 ACPY), marking the era in which behavioral concepts gained widespread recognition in finance.

2. Contemporary resilience: Despite a substantial increase in publication volume (from 281 articles in 2013 to 658 in 2021), the ACPY remains consistently high, staying above 2.5. This indicates that contemporary behavioral finance research maintains its relevance and depth even as the field expands in scope.

The inclusion of the ACPY metric confirms that the field's influence is not merely historical but continues to expand in depth and scope, even as publication volume accelerates.

Analysis of the top publishing sources (Figure 5) confirms a growing trend in behavioral finance research toward interdisciplinary and sustainability-focused outlets. The Journal of Cleaner Production and Management Science show the highest publication volumes between 1996 and 2022. While Management Science maintains a steady rate, the sustained and high output from the Journal of Cleaner Production highlights the significant integration of behavioral finance into sustainability-oriented research.

Figure 6 highlights the journals with the highest number of publications within the given timeframe. The most relevant sources, based on the number of documents published between 1984 and 2024, are as follows:

1. Journal of Cleaner Production

2. Management Science

These journals exhibited the greatest publication volume within the timeframe shown in the graph. The line representing the Journal of Cleaner Production indicates a steady increase in publications over the years, reaching the largest number of documents overall. Management Science also demonstrates a consistent upward trend, although with a lower publication volume than the Journal of Cleaner Production.

The 8,729 documents were distributed across 2,187 distinct sources. Figures 5 and 6 highlight the journals with the highest publication volume. While established finance and management journals, such as Management Science and Journal of Financial Economics (as reflected in the high citation counts reported in Table 5), remain crucial, the data reveal a significant interdisciplinary shift:

The Journal of Cleaner Production (JCP) emerged as the most prolific source by publication volume. This finding requires analytical discussion as it challenges the traditional view that behavioral finance research is confined solely to core finance and economics outlets.

The dominance of JCP reflects the evolution of behavioral finance to include green and sustainable financial behavior as a core research theme. Since the late 2000s, research has increasingly focused on the following:

1. Sustainable investment decisions: Investigating the psychological biases and heuristics that influence investors' decisions to allocate capital to environmental, social and governance (ESG) funds or green bonds.

2. Pro-environmental behavior: Applying behavioral principles (such as loss aversion, framing and nudges) to policy and managerial contexts aimed at promoting sustainable consumption and corporate social responsibility (CSR) initiatives–topics highly relevant to JCP's scope.

This trend is consistent with other recent scientometric reviews that track the merging of finance, behavioral economics and sustainability science, indicating that the behavioral perspective is a critical lens for understanding sustainability-linked financial anomalies (Asness et al., 2013; Ohlan and Ohlan, 2022).

To provide quantitative evidence of thematic evolution and distinguish emerging topics from mature ones, burst detection analysis (Kleinberg, 2003) was used on the author keywords and keywords plus corpus. This method identifies terms that have experienced a sharp, statistically significant increase in frequency of usage or citation over a specified time period, indicating a strong surge of academic interest.

The analysis, conducted over a 40-year period, confirms a shift toward sustainability and technological applications. Table 6 presents the top 6 strongest keyword bursts, ranked by burst strength and duration, with the burst year indicating when the rapid increase began.

To distinguish emerging frontiers from mature topics, Kleinberg's burst detection analysis (Table 6) was used. This method identifies terms that experience a sharp surge in academic interest. The results rank keywords by the statistically significant strength of their increased usage in the literature, highlighting the field's most intense periods of research focus.

The results clearly outline two distinct periods of research focus:

Foundational/mature themes: The strongest bursts associated with core concepts–behavioral finance (1998–2005), loss aversion (2003–2011) and overconfidence (1995–2003)—exhibited high strength but occurred predominantly in the field's early and middle phases. This temporal pattern confirms their status as foundational, mature themes that have largely been assimilated into the standard literature. Their high initial impact established the field's early intellectual structure.

Emerging and frontier themes: Conversely, the keywords financial literacy, FinTech and ESG all registered powerful bursts that commenced in the latter half of the 2010s and persisted through 2024 (Chopova and Ellemers, 2023). The recency and continuity of these bursts show that the current research frontier is rapidly shifting away from purely theoretical anomalies toward real-world application and integration. This quantitative evidence strongly supports the observation (previously noted in the analysis of top sources and figures) that the field's current development and most intense research interest lie at the intersection of psychology, financial technology and sustainability.

This transition from foundational psychological biases to technological and societal applications underscores a shift in the field from a descriptive, theoretical phase to an applied, interdisciplinary phase.

Research objective 2: To identify key influences and networks of the most impactful publications, influential authors and co-citation networks that have shaped the field's development.

Figure 7 depicts the most globally cited articles in behavioral finance, illustrating the citation concentration among foundational works and the subsequent thematic diversification of the field. The distribution of citations across the top ten documents illustrates an intellectual ‘hegemony' of late-1990s foundational theories, followed by a diversification of the field in the last decade. The visual dominance of Daniel et al. (1998) and Fama (1998) in the figure signifies that the debate over market efficiency and cognitive heuristics remains the primary gravitational center of the discipline. However, the trajectory of citations for more recent entries, such as Van Rooij et al. (2011) and Fernandes et al. (2014), reveals a critical thematic shift: the field is moving away from purely abstract psychological modeling toward the socio-economic barriers of market entry, specifically financial literacy. Furthermore, the high citation density of methodological work like Nguyen et al. (2024) reflects a modern pre-occupation with research infrastructure and the validity of digital crowdsourcing. Ultimately, the figure interprets the field's evolution as a transition from identifying theoretical anomalies to developing the empirical tools and educational frameworks necessary for practical financial intervention.

Table 7 presents the intellectual influence and evolution of behavioral finance and related interdisciplinary research, as evidenced by citation-based indicators. Seminal contributions such as Market efficiency, long-term returns and behavioral finance (Daniel et al., 1998; Fama, 1998) and Does the stock market overreact? (De Bondt and Thaler, 1987) record exceptionally high total and normalized citation counts, confirming their foundational role in challenging traditional market efficiency assumptions. The prominence of papers published in top-tier journals such as the Journal of Financial Economics, Journal of Finance, Management Science and Management Information Systems (MIS) Quarterly underscores the cross-disciplinary diffusion of behavioral concepts across finance, economics and information systems (Venkatesh et al., 2003; Fernandes et al., 2014). Furthermore, relatively recent studies, including Financial literacy and stock market participation (Van Rooij et al., 2011), exhibit strong citations per year and normalized impact, indicating sustained scholarly relevance beyond publication age effects. Overall, the table illustrates a balanced coexistence of classic theoretical works and contemporary empirical studies, highlighting both the historical depth and continuing advancement of the field.

Figure 8 shows that the pattern reveals a gradual increase in scholarly activity after the early 2000s, followed by a marked concentration of larger bubbles in the post-2015 period, suggesting accelerated research productivity and impact in recent years. Authors such as Kumar S., Smith JR., Dincer H., Liu Y., Wang Y. and Zhang Y. show sustained and growing contributions, with particularly strong visibility after 2018, indicating their central role in shaping contemporary research themes. In contrast, authors like JR, Mostert FJ and Mostert JH exhibit earlier engagement with comparatively moderate and intermittent influence, reflecting foundational but less recent activity.

Figure 9 presents the most relevant authors in behavioral finance based on publication output. Kumar has written more than 25 articles, followed by Smith, Mostert, Dinçer and Li.

Table 8 summarizes the top 15 authors by publication output, reporting total articles, fractionalized articles based on co-authorship, total citations and average citations per article. Publication counts show limited variation among leading contributors, with Kumar, S. ranking first, followed closely by several authors with comparable output. Citation indicators reveal differences in research impact, as reflected in total and average citations. The inclusion of fractionalized counts provides a more accurate assessment of individual contributions within collaborative research settings.

Figure 10 illustrates the author's keyword co-occurrence network, highlighting the core research themes and their interrelationships within the behavioral finance literature. The co-occurrence analysis of author keywords reveals the underlying thematic structure and research concentration within the field. Frequently co-occurring keywords form distinct clusters, indicating well-established and interconnected research themes, while closely linked terms suggest strong conceptual relationships across studies. The prominence of certain keyword groupings reflects areas of sustained scholarly attention, whereas less densely connected terms may indicate emerging or niche research topics. The strength of co-occurrence links further highlights the intensity of thematic integration, demonstrating how specific concepts consistently appear together across publications. Overall, the observed keyword clustering patterns provide insight into dominant research directions and the evolving focus of the literature (Singh, 2021; van Eck and Waltman, 2010).

Figure 11 depicts a collaborative network analysis. It shows a social network of researchers who have authored publications on behavioral finance and investor psychology in finance (Bumen and Hotaling, 2022). The circle nodes represent researchers, and the connecting lines indicate co-authorship on a publication. The node size reflects the number of publications each researcher has authored within the specified period.

Collaborative network analysis examines patterns of co-authorship to identify key contributors, assess collaboration structures and understand knowledge dissemination within a research field (Ellegaard and Wallin, 2015; Motylska-Kuzma, 2017).

Figure 12 provides a basic overview of the geographic distribution of publications within the dataset. The United States, the United Kingdom, India, China and Germany are the top publishing countries. This indicates the citation count or citation impact of publications authored in those countries. The scale is based on document weights, with scores representing average publication output.

The co-authorship map, when analyzed through the lens of geographical and thematic clustering, reveals significant regional heterogeneity in collaboration patterns.

1. Cluster 1 (North American institutions): This cluster shows strong inter-institutional ties (high density) and is largely anchored by leading universities in the United States and Canada (Harvard, Wharton). Thematic focus of their collaboration is heavily concentrated on micro-level biases and the foundational psychological mechanisms of behavioral finance (overconfidence, loss aversion, heuristics and anomalies).

2. Cluster 2 (European and Oceanic institutions): This cluster is typically characterized by institutions from the UK, Germany and Australia. Their collaborative research often bridges the theoretical core with regulatory concerns, focusing on the practical application of behavioral insights to regulatory policy, risk management and the economic consequences of corporate governance.

3. Cluster 3 (Predominantly Asian institutions): In contrast, this cluster, centered around institutions in China, India and Southeast Asia, exhibits collaboration patterns focused more on macro-level applications and contemporary issues relevant to emerging markets. This focus includes the impact of FinTech on investor behavior, the efficacy of financial literacy programs and the role of sentiment in less efficient financial markets.

This distinction highlights how the global behavioral finance research community is geographically segmented, with established Western institutions focusing on core psychological theory and emerging markets prioritizing applied and technological implementation research.

Figure 13 shows the results of an analysis using a normalization method called log linear/modularity. The figure illustrates normalized citation performance across publications, enabling a comparison of scholarly influence after accounting for differences in publication age and disciplinary citation practices. Variations in normalized citation values indicate differing levels of impact among publications, with higher values reflecting relatively stronger influence within the field. This adjustment highlights works that maintain significance beyond raw citation counts, thereby allowing a more balanced interpretation of research impact over time (Mongeon and Paul-Hus, 2016).

Figure 14 presents the citation analysis based on 8,729 articles. The minimum number of citations per document is 3, and 5,330 articles meet the threshold. The article by Daniel et al. (1998) has the highest number of citations (2,684), followed by Fama (1998) with 2,284 citations and Ghemawat (2001) with 1,334 citations.

Table 9 presents a curated list of highly cited behavioral finance papers, along with their key characteristics and potential journal fit, highlighting the intellectual depth, methodological rigor and thematic relevance of seminal contributions to the field.

Co-authorship analysis combined with citation data can provide valuable insights into research collaboration patterns and their impact within a domain, thereby strengthening the methodological rigor of bibliometric investigations (Donthu et al., 2021; Kumar and Kumar, 2022). Such analysis helps identify groups of researchers who frequently collaborate, often representing active research communities within a field. Moreover, examining the citations received by co-authored publications can reveal which research communities are producing highly influential work (Mushtaq et al., 2023).

In Figure 15, the minimum number of articles per author was set at 3, and the minimum number of citations per author was set at 5. Out of 8,492 authors, 27 met the threshold. Statman contributed 7 articles with 188 citations, Dey authored 3 articles with 158 citations, Alkaraan contributed 3 articles with 153 citations, Kickert collaborated on and authored 7 articles with 152 citations and Smith collaborated on and authored 9 articles with 81 citations.

The different colored lines represent thematic connections between various concepts.

1. Corporate governance is connected to financial performance, CSR, accounting and risk management.

2. Financial performance is connected to sustainability.

3. Accounting is connected to financial services and financial performance.

4. Risk management is connected to risk decision-making.

5. Risk decision-making is connected to behavioral finance.

6. Behavioral finance is also connected to decision-making.

Figure 16 explores the relationships between different concepts relevant to corporate governance. The analysis highlights several important connections, including those between corporate governance and financial performance and between risk management and risk decision-making.

Figure 17 presents a word cloud illustrating key concepts in behavioral finance. Central terms include behavioral finance itself, which is linked to investor sentiment, decision-making, risk management and overconfidence, reflecting the field's focus on psychological influences on financial choices. Decision-making emerges as a prominent theme, connected to risk analysis, AI and big data, highlighting the factors shaping financial decisions.

Figure 18 Visualizes trending topics in behavioral finance and financial decision-making research from 1986 to 2024, based on publication keyword analysis (Liu et al., 2020; van Eck and Waltman, 2010). Consistent themes throughout the period included financial accounting/reporting and “decision-making”. CSR has gained prominence since the mid-2000s, while big data and FinTech have emerged and rapidly grown since the late 2000s.

Figure 19 Presents a thematic map analysis illustrating the central, emerging and niche themes within the field of behavioral finance. This analysis visually represents the relationships and prominence of different themes within a particular domain. In this context, the map provides insights into the distribution and significance of various topics in behavioral finance research. By analyzing the placement and size of the thematic circles, researchers can discern the relative centrality and emergence of key concepts, offering a comprehensive understanding of the current research landscape.

1. Behavioral finance: This is the core theme around which the field is built, and it sits at the center of the map, indicating a high degree of centrality and relevance.

1. Financial literacy: This theme is gaining prominence in the field, as indicated by its placement on the right side of the map and the relatively large circle around it (Kumar and Kumar, 2023). This suggests an increase in the number of publications addressing financial literacy in recent years.

2. Overconfidence: Similar to financial literacy, this theme is also emerging within behavioral finance research, as reflected by its position on the right side of the map (Barber and Odean, 2001).

To facilitate this analysis, the map is partitioned into four quadrants, defined by two structural metrics: centrality and density:

1. Centrality (horizontal axis): Measures the strength of inter-cluster ties (external connections), reflecting a theme's importance to the entire network. Themes on the right exhibit high centrality.

2. Density (vertical axis): Measures the strength of intra-cluster ties (internal connections), reflecting a theme's internal coherence and maturity. Themes toward the top exhibit high density.

Table 10 provides the strategic interpretation of the thematic map quadrants based on centrality and density measures, outlining how motor, niche, emerging/basic and declining themes are identified and interpreted in bibliometric analysis.

Based on the strategic placement of keywords in the map:

1. Motor themes (high centrality, high density)

Keywords located here include risk, finance, accounting and decision-making. These terms represent the foundational disciplinary pillars of the entire field. They are highly developed internally and maintain strong connections across different research clusters, confirming their central and critical role in linking behavioral finance theory to financial practice.

1. Niche themes (low centrality, high density)

This quadrant contains specialized, internally coherent topics, including risk management, sustainability, financial decision-making and decision-making. These clusters demonstrate maturity (high density) but are relatively isolated from the broader network (low centrality), suggesting highly active, specialized subfields that have not yet fully integrated their findings across the entire domain of behavioral finance.

• Decision-making: Given its placement on the left side of the map, decision-making appears to be a well-established theme in behavioral finance, with a consistent level of research output (Thaler, 2016).

• Risk: Similar to decision-making, risk appears to be a well-established theme in behavioral finance research (Ko and Huang, 2012).

• Emerging or basic themes (low centrality, low density)

This quadrant primarily holds behavioral finance itself, along with related concepts such as financial literacy and overconfidence. Since behavioral finance and its core psychological biases serve as the starting point for all studies, their low density indicates that they function as basic or foundational concepts rather than emerging ones, acting as input variables rather than internally developing subfields. Themes such as financial literacy and overconfidence show low internal development, supporting their status as widely applied concepts that are still developing internal complexity, consistent with the emerging themes status identified in the burst detection analysis.

1. Emerging or declining themes (high centrality, low density)

This quadrant contains themes that are highly connected externally but lack internal coherence, namely corporate governance, financial performance, CSR and performance. These themes act as crucial bridges between core financial mechanisms and broader managerial or policy outcomes. Their low density suggests that they represent either transitional themes or complex applications in which internal research focus has not yet been solidified.

Research question 4. Are there any under-researched areas within behavioral finance highlighted by the analysis?

A key theme is the need for greater integration between behavioral finance and emerging technologies, including big data, machine learning and social media analysis. While these tools offer powerful ways to study investor psychology, further research is needed to validate their effectiveness and develop practical applications in financial literacy programs or investment strategies (Bollen et al., 2011; Gu et al., 2020). Additionally, collaboration with disciplines such as neuroscience can provide valuable insights into the biological underpinnings of biases.

Future research should also focus on how behavioral biases manifest differently depending on investor characteristics (e.g., retail vs institutional investors, generational differences such as Gen Z vs Baby Boomers) and their cultural background. Research examining how cultural factors influence behavior in developing economies is crucial for achieving a more global perspective (Quddus and Banerjee, 2021; Feldman and Liu, 2023; Chan and Cheung, 2012; Hamdan et al., 2023). Finally, the analysis highlights the importance of addressing specific, underexamined biases, such as the disposition effect and anchoring bias. Future work can examine these biases more deeply, exploring how they interact with other biases and potential strategies to help investors overcome them.

Table 11 synthesizes future research directions identified from 30 top-cited behavioral finance studies, highlighting key methodological gaps and potential avenues for advancing research across investor psychology, asset pricing, corporate finance and emerging market contexts.

The bibliometric findings presented in Section 5—particularly the accelerating publication growth, the rise of interdisciplinary themes, including sustainable finance and the emergence of new research clusters–provide critical insights into the field's maturity and future trajectory. This discussion interprets these findings, outlines their implications and proposes a future research agenda.

One of the most valuable contributions of this scientometric synthesis is the ability to map and critically discuss persistent areas of debate and contradictory findings within the behavioral finance literature. Rather than presenting a unified theory, the literature is characterized by several dichotomous viewpoints, primarily concerning the rationality, universality and predictive power of core biases.

1. The rationality of herding behavior

One critical area of debate revolves around herding behavior. While a significant body of foundational literature views herding as an irrational cognitive bias, a robust counter-stream of research suggests that herding can be a rational response to information asymmetry. In contexts characterized by high uncertainty, following the actions of presumed informed investors (rational arbitrage) is an optimal, utility-maximizing strategy (Bikhchandani et al., 1992). The coexistence of these two research streams shows that the effect's classification is context-dependent, demanding clearer methodological frameworks to distinguish irrational social effects from rational informational cascades.

1. The cultural moderation of loss aversion

Similarly, the synthesis of findings related to loss aversion highlights contradictory results regarding its universality. While the original prospect theory suggests a near-universal psychological factor, cross-cultural comparisons indicate that the strength of this bias is not universal but is significantly moderated by institutional and cultural contexts. For instance, research in emerging markets sometimes reports weaker or inconsistent effects, indicating that the observed anomaly is deeply intertwined with variables such as financial literacy, regulatory trust and national culture, suggesting that future models must explicitly integrate these institutional moderators.

The evolution of behavioral finance research, evidenced by the co-word and thematic network analyses (Figures 16 and 18), confirms a shift from a purely corrective function (challenging traditional finance theories) to a constructive theoretical role.

1. Refining prospect theory: The consistent high centrality of themes such as risk decision-making and overconfidence indicates that the core concepts of prospect theory and heuristics and biases remain theoretically foundational. However, the thematic map suggests a need to refine these models by integrating factors from corporate governance and accounting, thereby shifting beyond individual decision-making toward corporate behavioral dynamics (Baker and Wurgler, 2011).

2. Interdisciplinary expansion (ESG and climate finance): The unexpected dominance of the Journal of Cleaner Production in publication volume necessitates a theoretical broadening of behavioral finance. This trend implies that the field's explanatory power is now essential for understanding non-traditional financial domains. For instance, recent studies link investor biases, such as loss aversion, to the reluctance of both retail and institutional investors to engage with ESG or green products due to perceived risk or lack of standardized performance measures (Bialkowski and Starks, 2016; Boubaker et al., 2024). This extension of scope requires behavioral finance theory to account for pro-social and ethical preferences alongside traditional utility maximization.

3. The role of technology (FinTech and AI): The emergence of AI and big data as trending topics has profound theoretical implications. Future theoretical models must incorporate how technology-mediated decisions (robo-advisors) influence or mitigate classic behavioral biases. Recent work highlights that FinTech platforms' gamified features and algorithmic personalization can paradoxically amplify impulsive trading and overconfidence (Rizinski et al., 2024). This calls for a new paradigm of algorithmic behavioral finance to model the interaction between human and machine biases.

4. Neuroscience integration (new theoretical avenue): Supporting the future research agenda, recent conceptual work emphasizes that neurofinance moves beyond merely identifying biases (the what) to explaining the physiological underpinnings (the how) of financial decisions. This shift suggests a theoretical evolution in which behavioral models incorporate neural pathways and physiological triggers to explain behaviors such as confirmation bias and risk aversion.

The identified trends translate into tangible strategies for practitioners and policymakers, allowing them to leverage psychological insights for better outcomes.

1. Tailored financial education: Findings on the persistence of specific biases, including overconfidence, and the high research activity on financial literacy suggest a direct need for behavioral intervention. Policymakers should design financial education programs that use nudge theory and choice architecture to counter specific biases, moving beyond simple information disclosure to active behavioral correction.

2. Sustainable investment nudges: The strong linkage between behavioral finance and sustainable investment, as reflected in the prominence of the Journal of Cleaner Production, offers a crucial managerial implication. Fund managers and advisors can leverage framing effects to emphasize the societal losses associated with not investing sustainably (loss aversion), rather than focusing solely on potential gains, thereby improving the uptake of ESG products.

3. Enhanced investor protection: The analysis highlights under-researched biases such as the disposition effect and anchoring bias. Regulators should design investor protection measures and disclosure requirements that specifically address the cognitive vulnerabilities revealed by these biases, rather than assuming rational processing of information. For instance, requiring mandatory waiting periods for high-risk trades could mitigate short-term emotional biases.

While this study offers a comprehensive bibliometric overview, it is subject to the inherent limitations of the methodology.

1. The analysis was restricted to documents indexed in the Scopus database and primarily limited to English-language journal articles, potentially omitting significant contributions from books, conference proceedings and non-English literature, particularly from developing economies.

2. Bibliometric analysis maps published knowledge but does not assess the quality or validity of the underlying research methods, which requires a traditional systematic review.

Based on the identified research gaps and emerging themes (particularly the quantitative evidence from the burst detection analysis), three key avenues for future exploration are proposed:

1. Neuroscience integration: Future qualitative research should focus on collaboration with neuroscientists to delve into the biological and neurological underpinnings of biases, shifting from observing irrational behavior to understanding its underlying biological mechanisms.

2. Cross-cultural and generational studies: Research must specifically explore how behavioral biases manifest and differ across diverse cultural backgrounds and generational cohorts (Gen Z, Millennials, among others), particularly within emerging markets, to develop more culturally sensitive financial models.

3. Big data and market microstructure: Quantitative studies should leverage big data and machine learning to analyze the interaction between behavioral factors (derived from social media sentiment or trading patterns) and high-frequency market anomalies, exploring how biases influence price dynamics in the age of algorithmic trading.

The investigation into the intersection of cognitive biases, behavioral finance and high-frequency trading reveals critical areas for future development and practical application. This study provides a longitudinal synthesis of behavioral finance research across a continuous four-decade horizon (1984–2024). Rather than merely summarizing preceding sections, this conclusion synthesizes the findings to map the intellectual transformation of the field.

1. Refine theoretical models: Integrate findings on the rapid feedback loop between algorithmic trading (AT)–driven volatility and human emotional responses into existing behavioral asset pricing models to improve forecasting accuracy.

2. Develop new curriculum: Introduce specialized modules on neurofinance and market microstructure to equip finance students with the tools to analyze the intersection of human psychology and high-frequency trading dynamics.

3. Interdisciplinary collaboration: Foster joint research between finance, computer science and psychology departments to systematically classify and test technological interventions for human cognitive biases.

Practitioners (wealth managers, robo-advisors and fund managers)

1. Enhance financial education programs: Focus education specifically on mitigating the disposition effect and overconfidence by demonstrating their exacerbated impact when trading in AT-dominated, high-speed market environments.

2. Improve robo-advisor design: Implement mandatory cooling-off periods or bias spotting nudges in trading interfaces that flag potential emotion-driven trades (Compen et al., 2022).

3. System risk management: Develop internal algorithms that detect sudden, irrational surges in client trading activity that might signal herd behavior, allowing proactive intervention.

Policymakers (regulators, financial authorities).

1. Utilize nudge theory principles: Design automatic enrollment savings plans and structure consumer disclosures to counter common biases such as anchoring and overconfidence.

2. Review disclosure requirements: Mandate clearer, simpler disclosures for investment products, specifically detailing the role of algorithmic execution and associated risks, to counter ambiguity aversion.

3. Regulate algorithmic stability: Develop regulatory sandboxes to test the systemic stability of new algorithmic trading strategies, ensuring they do not inadvertently amplify human biases.

In summary, this four-decade bibliometric exploration underscores a pivotal shift in behavioral finance: the transition from understanding individual investor psychology to navigating a complex, co-evolutionary landscape where human biases and algorithmic speed intersect. As the field moves forward, the synergy between technological advancement and psychological insight will be paramount. By addressing the thematic gaps identified in this study, the financial community can better anticipate market anomalies and foster a more resilient, transparent and ethically grounded global financial system.