In recent years, a new research area in finance has emerged and it has revolutionized the financial theory. Continuing anomalies which were found in financial markets are leading to determine that markets are not as efficient as had previously been thought over the last century, but are influenced by several biases that demonstrate their imperfection, opening a new theory known as "Behavioural Finance".

We have witnessed from the contribution of the theory of expected utility of Von Neumann and Morgenstern (1944) to the establishment of several models whose aim is to explain the behaviour of financial markets. During most of the twentieth century, Modern Finance has been the prevailing trend, based on the assumption that the conditions of a market in equilibrium can be implicitly or explicitly expressed in terms of expected returns (Fama, 1970). One of the most relevant authors regarding Modern Finance has been Fama, who has most defended the efficiency of markets, defining an efficient market as a market in which "prices always 'fully reflect' available information" (Fama, 1970, p. 383).

Investor rationality has been highly criticized, as well as the efficiency of financial markets when prices are allocated. Simon (1956) is one of the first authors who criticized the economic and statistical theories of rational behaviour, arguing that learning theories explain observed behaviour better. In his study, Simon (1956) confirms that human beings adapt well enough to satisfy their needs at a specific level, but they are not able to find the optimal way to maximize the ‘utility function’. However, Tversky and Kahneman (1974) had a greater impact on defending the irrational behaviour of investors, supporting a better understanding of heuristics and biases in order to improve judgements and decisions in environments of uncertainty. Kahneman and Tvesky (1979) also put the expected utility theory into question. They defend that individual preferences breach the axioms of this theory, which leads to the development of the prospect theory.

Statman and Caldwell (1987) defined Behavioural Finance as a descriptive theory of choice under conditions of uncertainty. This concept encompasses a broader view of social sciences, including both psychology and sociology (Shiller, 2003). In general, Behavioural Finance defends that ‘irrational expectations or non-standard preferences affect asset prices’ (Campbell, 2000, p. 1551) and thus, deviations between prices and asset values depend on investors’ psychological motivations (Ofek, Richardson, & Whitelaw, 2004). Due to a growing interest in this new theory, especially in this century, it is absolutely necessary to analyse the evolution, the issues addressed and the current state of Behavioural Finance.

In order to explain the performance and productivity of Behavioural Finance, as well as to understand the themes that it is comprised of and its state, we have developed a bibliometric analysis. We found other bibliometric analyses of Behavioural Finance, however, none of them address this research area on its own (Fonseca, de Melo, de Melo, & Willer, 2017; Fonseca, de Melo, & de Melo, 2019). In addition, a list of all the relevant themes covered by this research area has not been found in any of the studies. Therefore, to provide a broader vision of Behavioural Finance, we performed a co-word analysis using SciMAT software (Cobo, López Herrera, Herrera Viedma, & Herrera, 2012). This enabled us to find the most important themes of the research area and to classify them according to their development and importance.

This work is structured as follows. In the next section the methodology and the process used in this study will be presented. In the third section, we will explain the data used in the analysis. The fourth section presents the results obtained relating to productivity and performance, as well as the analysis of scientific maps. In the fifth section, the conclusions will be explained. In the last section, we will highlight the limitations, which we had to face in this work.

In this work, a bibliometric analysis will be carried out in order to define the intellectual structure of Behavioural Finance. We used SciMAT software (Cobo, López‐Herrera, Herrera‐Viedma, & Herrera, 2012), which was chosen due to its flexibility when selecting measurements to obtain and visualize bibliometric networks. This tool also provides a wide range of pre-processing techniques that allow us to refine the results obtained from the database, improving their quality.

Bibliometrics analysis is defined as a part of scientometrics, which uses mathematical and statistical methods to analyse scientific activities in a research field (Aparicio, Iturralde, & Maseda, 2019; Callon, Courtial, & Laville, 1991). We conducted a bibliometric study from its two key procedures: evaluation and performance and scientific productivity analysis, and scientific maps (Cobo, 2012).

The analysis of scientific maps is one of the techniques whose aim is to monitor a scientific field in order to understand its structure, development and its main participants (Noyons, Moed, & Luwel, 1999). The different types of information that can be used are known as units of analysis. In this work, the chosen unit of analysis is the keyword.

Among all units of analysis, different types of relationships can be established in order to create different bibliometric networks. In this work, we have used a co-occurrence relationship (Callon, Courtial, Turner, & Bauin, 1983). The co-occurrence relationship occurs when two elements appear together in a document. This analysis allows us to identify the main themes of a scientific field, showing its conceptual and cognitive aspects (Cobo, 2012).

In order to obtain significant information about the area through the analysis, we should normalize the bibliometric network. Normalizing the network allows us to relativize the relationships between two analysis units. The normalization measure used in this work was the equivalence index (Callon et al., 1991).

Once the network has been normalized, we can extract the themes by using clustering techniques with which to divide the set of elements into different subsets, whose nodes are strongly linked together, and scarcely linked with others. In this paper, we use the clustering algorithm based on simple centres, which has the advantage of automatically returning labelled clusters with the most central node in the group (Cobo, 2012).

In order to evaluate the position of each theme in the area, we represent the different clusters on the well-known Strategic Diagram established by Callon et al. (1991). The Strategic Diagram allows us to place the themes according to their centrality (axis x) and density (axis y). Centrality measures the degree of interaction of a theme with the rest; in other words, it measures the strength of the external links of a theme with the rest. Centrality allows us to measure the importance of a theme in the global development of the scientific field (Cobo, 2012). Density measures the degree of internal cohesion of a theme, so it measures the internal strength of the different links of the nodes within a cluster. This value can be understood as the measure of the development of the theme (Cobo, 2012).

In this context, we can classify the themes into four categories in the Strategic Diagram:

• •

  Motor-themes: they are located in the upper-right quadrant and they show strong centrality and density. They are well developed and important in the research field.

• •

  Highly developed and isolated themes: they are found in the upper-left quadrant and show low centrality, but high density. These themes are of marginal importance for the field.

• •

  Basic and transversal themes: they are in the lower-right quadrant and show strong centrality, but low density. They are interconnected with the other themes but are not well developed.

• •

  Emerging or declining themes: they are in the lower-left quadrant and they show low centrality and density. They are both weakly developed and marginal.

The process followed to define the intellectual structure of Behavioural Finance can be observed schematically in Fig. 1.

Fig. 1.

Workflow of science mapping analysis.

(0.05MB).

The data set used for the bibliometric analysis was extracted from the Web of Science (WOS) of Thomson Reuters. WOS was preferred to other platforms like SCOPUS because of a greater number of scientific publications it has provided since 1900 (SCOPUS has only collected scientific publications since 1966).

The search was made on 1 January 2018 by choosing only scientific papers that were published until 2017 (including) in journals indexed in the database of the Social Sciences Citation Index (SSCI).

Many concepts applied in Behavioural Finance have been developed in other areas and disciplines, and therefore, have not been included in the search, such as:

• •

  Prospect Theory: the work developed by Kahneman and Tversky (1979) has been one of the most cited in Behavioural Finance. Nevertheless, this theory has also been used on other areas of the economy and has a special impact on psychology.

• •

  Mental Accounting: this term has had a strong impact on Behavioural Finance; however, it has also been applied in other economic disciplines, especially in the Marketing area, due to its consumer behaviour studies.

• •

  Cognitive Bias: this concept is related to studies on investor expectations, but which stands out mainly in health studies. In the economic field, its study stands out in the area of entrepreneurship.

The main concepts studied in this work are “Behavioural Finance”, “Behavioural Portfolio” and “Investor Sentiment”. The search by topics was carried out, so the selected articles contained the previous concepts on the title, summary or keywords. Therefore, we carried out the following search: TS= ("behavioral financ*" OR "behavioural financ*" OR "behavioral portfolio" OR "behavioural portfolio" OR "investor sentiment*"). Thus, a total of 1.214 published articles from 1987 to 2017 were obtained.

With this study, it has been possible to observe a high growth that has taken place in Behavioural Finance research, especially since 2009. This may be motivated due to an increase in anomalies detected in financial markets since the 2007 crisis, highlighting the inefficiency of these markets and the influence of individual biases on them.

When analysing the most productive and cited authors in this area, it was found that the most influential authors in the area by number of citations are Malcom Baker and Jeffrey Wurgler, who, among their contributions, promoted research on the influence of investor sentiment on stock returns. The article “Investor sentiment and the cross-section of stock returns” of these authors, has been positioned as the third most cited article. However, the most cited article was "Market efficiency, long-term profitability and behavioral finance" by Fama (1998), which is positioned in favour of the efficiency of financial markets and may have been the subject of many discussions of studies published on Behavioural Finance.

As for the journals in which Behavioural Finance research has been published, "Journal of Banking & Finance” stands out as the journal that has published the most articles. This demonstrates how until 2017, this area had been disseminated mostly in journals that publish papers of the financial field with a certain generality. However, it was expected that in the coming years, specialized journals such as the “Journal of Behavioral Finance”, would show a greater number of published articles on Behavioural Finance. It is necessary to take into consideration that this journal is relatively new, it emerged in 2003, and still occupies the second position in number of published articles on Behavioural Finance.

Regarding the productivity of Behavioural Finance by country and organization, it has been observed that research in this area is mainly concentrated in the United States, highlighting the “National Bureau of Economic Research” organization.

It should be noted that a major contribution of the study has been obtaining the main themes in which this area of research can be divided. Although articles such as those by Fonseca et al. (2017) and Fonseca et al. (2019) have performed a bibliometric analysis of Behavioural Finance, they have not performed this analysis in isolation and they have not extracted the main themes that make up this research area. To address this issue, a co-occurrence analysis of keywords has been performed using the SciMAT software (Cobo et al., 2012), in which we obtained 13 outstanding themes in Behavioural Finance, which are Investor Sentiment, Disposition Effect (which can be called attitude towards risk), OPI, Overconfidence, Portfolio Selection, Expected Return Stocks, Arbitrage, Model, Bias, Attention, Weather, Mutual Funds and Strategies. When analysing their level of development and their importance within the area, the themes of Investor Sentiment, Disposition Effect, Overconfidence and Expected Stock Returns are the most prominent themes in Behavioural Finance, positioning themselves as motor themes. As basic or transversal themes of the area, we have found the themes of Arbitrage, Bias and Model, which, although they are not themes that are highly developed internally, they do have a great influence on the rest, being usual for them to be dealt together with other themes. On the contrary, it has been found that the themes Weather, Mutual Funds and IPO are highly developed, but hardly linked to the rest, so they are very specific themes within Behavioural Finance. Finally, it should be noted that the themes Attention, Strategies and Portfolio Selection have been the least developed and related to the rest. Within these, the theme of Attention is an emerging theme in Behavioural Finance.

Finally, it is convenient to mention the productivity and impact of each theme, finding that Investor Sentiment stands out over the rest, both by number of articles published, as by the H Index and the number of total citations received. It is also worth mentioning the Arbitrage theme, which with only 59 articles has a total of 1943 citations, showing to be a theme with a high average impact per publication.

With the purpose of identifying and visualizing the intellectual structure of Behavioural Finance, in this work a bibliometric analysis has been carried out in which the performance of the scientific production of the area has been evaluated, as well as the themes that comprise it through an analysis of co-occurrence of keywords. The results contribute to the future development of Behavioural Finance.

We have observed how Behavioural Finance is a research area that is in full swing, substituting Modern Finance for new research based on the irrationality of investors, and trying to explain the various anomalies detected in financial markets, being of special relevance the bubbles occurred in recent years.

However, the great contribution of this work has been to clarify the different themes that make up this research area, obtaining a total of 13 themes: Investor sentiment, disposition effect, Overconfidence and Expected Return Stocks as motor themes; Arbitrage, Bias and Model as basic and transversal themes; OPI, Weather and Mutual Funds as specialized themes; Strategies, Attention and Portfolio Selection as emerging or declining themes.

Regarding the productivity and the impact of each theme, we have observed that the Investor Sentiment theme stands out clearly both in productivity and impact.

With all this, we have detected a hot research area in the financial field, highly influenced by psychology and sociology. The most important theme in Behavioural finance is the study of investor sentiment on the decision-making process. Future research in this line will allow us to advance in a global theory that unifies studies in Behavioural Finance.

None.

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.