3.1. Evolution of the Research on Behavioural Economics in Agriculture (BEA)
Table 1 shows, in figures, the evolution of the most relevant dimensions in the research on behavioural economics in agriculture (BEA) between 1991 and 2020. We can distinguish two clearly differentiated stages in the evolution of the research on behavioural economics in agriculture during the period analysed. In the first half of the period, the publication of studies on this subject matter was highly irregular. In fact, the publication of studies was intermittent. The maximum number of studies published during this period was three, in the year 1997. Similarly, the variables of the number of authors, journals and countries shows the same trend. Given the nature of the variable of the number of citations, it began to be counted in the year 1993, when the first citation of a study in the sample was made. From then, this variable has also exhibited irregular behaviour although with a clear upward trend, reaching a maximum of eight in the year 2005. The average number of citations per article also grew during the period, reaching a total of 4.7, also in 2005. Before 2005, only 5.7% of the documents making up the sample had been published.
The year 2006 constituted a turning point, whereby a growth trend began which stabilised towards the end of the period studied. In this way, the number of documents that had remained stable at one between 1991 and 2005, rose to 26 in 2020. We can observe that the greatest boost in this subject area occurred in the final years of the period studied, given that more than 55% of the articles in the sample were published after 2016. In order to determine whether this trend is due to an overall inertia in research on agriculture, we have conducted a comparison of the percentage of annual variation in the number of publications in both lines of study. The result is shown graphically in
Figure 2. In this figure we can observe, on the one hand, how research in agriculture experienced a stable growing trend throughout the whole period, with an average annual variation in the number of publications of 8.7%. With respect to research on BEA, we can observe the aforementioned difference between the two sub-periods. From 2008, the variable experienced an almost exponential growth, with an average annual increase during the whole period of 11.9%. These data allow us to conclude that research on BEA is still in the early stages of its evolution. Furthermore, if the trend remains stable, we would expect that over the next few years, the use of behavioural economics applied to research in agricultural aspects will become a relevant line of study.
With respect to the number of authors participating in the research on BEA, we can observe that this variable experienced a more intense variation than the number of documents published. In total, 478 researchers participated in 176 publications. This figure grew from two in 1991 to 81 in 2020. The average number of authors per document has increased slightly over the whole period, increasing from one during the first half of the period of study to three from 2015. However, only 5.8% of the total authors involved in the research on BEA have participated in more than one study. This supports the claim that this line of research is in its infancy. The number of journals in which the studies on BEA have been published experienced an almost identical variation as the number of documents. In this way, the average number of documents per journal remained almost unchanged at one, with a slight increase at the end of the period. This data, again, corroborates the early stage of this line of research, given that the concentration of a high number of articles in a series of journals is more in keeping with a consolidated topic [
2]. The total number of journals in which studies on BEA have been published is 116 and the variety has also grown to 24 in 2020.
The variable that has changed the least is that of the number of countries involved in the research on BEA. In 1991, a joint study by researchers affiliated to institutions in Israel and the USA constituted the first contribution of this topic within the sample analysed. Since then, a total of 40 countries have been involved in this field of study. This figure is low if we compare it with other more consolidated topics, which usually exceed one hundred [
24]. The number of citations that the scientific studies accumulate in considered to be an indicator of the impact that the research has within the field of study. In this case, although the annual number of citations is not very high, this variable exhibits normal behaviour. The number of citations obtained increases in line with the increase in the number of journals publishing on the topic under study. In this way, the maximum number of citations was reached in the year 2020 at 705. The average number of citations per article can be a clearer indicator. This variable is found to have an increasing trend until 16.1 citations per article in 2020. This figure is very similar to that of other more consolidated topics [
27]. This data enables us to determine that this line of research is set to gain relevance within the research on agriculture in the coming years.
3.5. Institutions
The first study to apply behavioural economics to agriculture was developed through collaboration between the Hebrew University of Jerusalem and the School of Environmental and Biological Sciences de la Rutgers University–New Brunswick USA, in 1991. In 2020, more than 50 institutions were involved in developing studies along these lines.
Table 5 shows the 20 institutions with the highest number of publications on BEA. These institutions belong to nine different countries, all included in
Table 3, except for Ireland. There are no institutions representing Canada and China which are among the most prolific countries. This result is due to the fact that some countries participate in this line of research with a higher number of institutions, but with less studies conducted by each of them. Another noteworthy aspect, similar to the case of the countries, is the small number of contributions per institution, with none of them exceeding six studies. Once again, we cannot refer to a group of leading institutions in this subject area, due to its incipient nature.
The institutions with the highest number of publications on BEA are the Wageningen University & Research from the Netherlands and the University of Goettingen in Germany with six studies each. These are followed by the Centre d’Economie de l’Environnement in France with five; and Harvard University (USA), Martin Luther University (Germany), the University of Nebraska–Lincoln (USA) and the Montpellier SupAgro (France) with four. The rest of the institutions only have three publications. The institutions with the most relevant publications, measured by the number of accumulated citations are Harvard University with a total of 450, the National Bureau of Economic Research with 378, the University of Reading with 192, the Wageningen University & Research with 120 and Scotland’s Rural College with 106. However, the institution with the highest average number of citations per article is the National Bureau of Economic Research with 126.1. This is followed by Harvard University with 112.5 and the University of Reading with 64.2.
With regard to the international collaboration of the institutions, the average percentage of studies carried out in collaboration with other institutions is 34.8%, slightly lower than the average per country. However, we cannot observe any regularity in the behaviour of the different institutions. Nevertheless, 100% of the studies conducted by the Australian National University were conducted through collaboration; while others, such as the Martin Luther University or the University of Reading, have not conducted any studies in this way. On average, the studies carried out in collaboration obtained 12.8 citations. Meanwhile, those conducted autonomously obtained an average of 30.4 citations.
3.6. Authors
Table 6 shows the authors who have participated in at least two publications on BEA. This group is formed by a total of 28 researchers who, as already mentioned, account for 5.8% of the total authors included in the sample analysed. These authors are affiliated to a total of twenty different institutions, belonging to eight different countries. In the table, the authors have been grouped based on their co-authorship relationships. The different groups have been differentiated using colours. Within the different colours, we can distinguish different shades based on the amount of studies conducted jointly. The pale shade indicates one document shared, the medium tone indicates that two studies are shared and the dark shade indicates three shared documents. In this way, and by way of example, the first group is represented in violet. Four authors are included in this group. However, two of them appear with a medium shade of violet because they have co-written two studies with other authors, while the other two appear in a dark shade of violet because they share three articles. These co-authorship relationships are relevant given that they help to explain the inclusion of some of the authors in the table.
The author with the most studies published on BEA is Oliver Musshoff, of the University of Goettingen, with a total of five. He is followed by Norbert Hirschauer with four and Denise Peth, Simanti Banerjee, Philippe Le Coent, and Jayson L. Lusk with three. The rest of the authors have only published three studies. The most citations, both in absolute and average terms are obtained by Lucia A. Reisch from the Copenhagen Business School, with 240 and 120.1 citations respectively. This author has participated in just two studies which were published in 2012 and 2013. She is followed by Luiza Toma, affiliated to Scotland’s Rural College, with a total of 101 and an average number of citations per article of 50.5. In third position are Nick Hanley, Laure Kuhfuss, Raphaële Préget and Sophie Thoyer, who accumulate 77 citations in the two articles in which the four share authorship, with an average of 38.5 citations per document. Ada Wossink, from the University of Manchester, is the researcher in the table who has been studying BEA for the longest, with her first article published in 1997. At the other extreme, Peter D. Lunn (Trinity College Dublin) and Sean Lyons (Economic and Social Research Institute) are the most recent newcomers to this line of research with their first article on BEA published in 2020.
With respect to the co-authorship relationships, the group incorporating Musshoff, Hirschauer, Peth and Funke is noteworthy. These German authors are affiliated to the University of Goettingen and the Martin Luther University. The different researchers share the authorship of two or three studies, depending on the case. There is a group of American authors appearing in blue shades who are affiliated to the universities of Nebraska–Lincoln, Michigan and Delaware. Within this group, Burbach, Czap H.J. and Czap N.V share the authorship of the two studies that place them in the table; while only one publication is shared with Banerjee and Kecinski. Le Coent is a unique case. This author has the most extensive collaboration network within the table. He shares a study with Napoleone (represented in yellow), another with Subrevie (orange) and another with the group of authors represented in red (Hanley, Kuhfuss, Préget and Thoyer). Finally, the authors who do not share authorship with any other researchers included in
Table 6 are represented in white.
3.7. Main Topics
An analysis of the keywords allows us to identify the most relevant topics within the research on BEA. We will refer to the most salient aspects of each of them.
Behavioural Economics has contributed to the agricultural field in one way by focusing on animal species. In this sense, Gunnarsson et al. [
33] study the behaviour of laying hens and the elasticity of demand with respect to certain types of litter (straw and feathers). Huijps et al. [
34] explore the sub-optimal economic behaviour in the decision-making of Dutch dairy farmers in order to adopt measures to improve udder health, concluding that the low rate of adoption and of following the advice of the dairy industry is due to a certain level of inertia in the behaviour of the farmers. In order to obtain the desired behaviour, penalties are more effective than bonuses. Kristensen and Jakobsen [
35] identify the role of veterinarians as dairy herd health advisers to farmers, as they have the capacity to translate their knowledge within the farming system. The “irrational” actions of the farmers can be explained as their values, behaviours and risk perception are incorporated in the decision-making process. Instead of assuming that farmers seek to maximise profits, other explanatory factors are included in their utility function, such as animal health, animal welfare or other farmer’s recognition. Toma et al. [
36] focus on the determining factors explaining the biosecurity behaviour of farmers in Great Britain. The results show that the farmer’s perceived importance of specific biosecurity strategies is explained through 13 variables including the attitude towards animal well-being and the organic certification of the farm. Brugere et al. [
37] study aquaculture. The objective of this article is to argue for farmer-based, syndromic surveillance as a way of overcoming the current limitations of the conventional surveillance systems and demonstrate its usefulness in aquaculture. The authors highlight the complex interconnection of behavioural factors (economic and social) underlying farmer’s reporting of disease.
Another important area of study is the field of management. In this respect, Schmid and Robison [
38] conduct a series of experiments in order to verify the existence of social capital and explore its impact on the productivity of firms and individuals. The findings show that the identity of the commercial actors is important as it affects the purchase and sales prices, the choice of share or cash leases in agriculture, the acceptance of catastrophic risk or how the banks invest in social capital to retain customers. Barnes et al. [
16] compare the voluntary adoption of water quality management techniques within a Nitrate Vulnerable Zone (NVZ) in Scotland. They find that different behavioural groups can be observed depending on the acceptance of the regulation, the responsibility towards issues related to water pollution and the degree of compliance with the established regulation. The authors indicate that these behaviours recorded by the farmers are explained by a range of attitudinal alignments and should include across designations in order to change social norms.
Within the field of management, we can identify a sub-group of literature focusing on the study of perceptions. In this line, in order to promote environmentally friendlier techniques that are innovative, Wossink et al. [
39] analyse the degree to which the perceptions of risk and costs of farmers differ from the normative costs and risks and the characteristics that farmers perceive as important for adopting IAFS techniques (Integrated Arable Farming Systems). The results show that farmers consider a minimum level of knowledge, labour requirements and associated risks to be of importance. De Koeijer et al. [
40] conduct a review of concepts in agronomy and in farm and behavioural economics in order to determine which input-output combinations are possible while being optimum in practice. Among other factors, the findings show that as well as behaviouristic aspects, the preferences and perceptions of the farmer should also be considered. Duflo et al. [
41] assume that farmers have low fixed costs when purchasing fertilizer, introducing a stochastic bias. Therefore there are farmers who postpone the purchase of fertilizers until later while there are others who are more impatient. The authors find that, in accordance with the model, many farmers in Western Kenya do not make efficient use of the investments in fertilizers, but they do respond slightly after the harvest to certain small discounts that are limited in time (such as free delivery). They conclude that these types of policies work better than laissez-faire techniques or strong subsidies. Mills et al. [
17] identify the main drivers of farmer’s decision-making in relation to environmental management practices that are sustainable over time. They conclude that there is enormous heterogeneity in farmer’s beliefs and values in relation to custodianship and productivity.
Another area in which behavioural economics has been significant in agriculture is in the field of development. Banerjee [
42] reflects on development economics and examines the conditioning factors involved in people developing their natural talent. He indicates five important issues: contracts theory, coordination failures, political economy, learning and behavioural economics. Drawing from human behavioural ecology and behavioural economics, Tucker [
43], studies how people evaluate activities in their portfolios and the possible alternatives considered. The plans to create the Mikea Forest National Park (Madagascar) considered the elimination of slash-and-burn maize agriculture and the promotion of manioc crops (labour intensive). The analysis revealed that manioc is not a suitable replacement for maize as the two crops are cultivated differently (use of labour, delay-to-reward and rainfall); and the planners should offer alternative sources of proteins and cash to conserve small game. Furthermore, few resources should be dedicated to protecting lemurs, as they are rarely eaten and are never sold. Taking some East Asian countries as examples, Wade [
44] considers that low-income countries and their aid donors should focus more on industrial policy, as this is does not only mean “picking winners”. Industrial policy can be implemented by either leading or following the market. In addition, industrial policy can be adjusted to the available resources and state capacity. Datta and Mullainathan [
45] conduct a review of human behaviour and its application to development policy. Through behavioural economics, they indicate the principal pitfalls faced by policymakers in developing countries when seeking to design effective policies for these problems. Specifically, they use as a case study of agricultural policy the intervention made to promote the use of fertilizers among farmers in Sub-Saharan Africa. Nally and Taylor [
46] consider that the modernisation projects led by the Rockefeller Foundation were based on principles of behavioural economics, imposing a human capital model on the agricultural transformation that it proposed. In this way, they highlight the fundamental role played by philanthropy in the shaping of a new world order. Brune et al. [
47] study the use of policy intervention through savings accounts for the case of developing countries in order to increase the use of agricultural inputs by households. An experiment was carried out in Malawi among cash crop farm households. The results show that offering savings accounts increases the number of bank transactions, but also has a significant and positive effect on measures of household well-being.
Another framework that has received particular interest is economic policy. Bishop et al. [
48] study the attitudes towards adopting new technology in dairy farms. They examine the behaviour, motivations and intentions of the potential adopter. An important implication is that it could be beneficial for decision-makers to guide the policies previously targeting different types of agents. In a study conducted in the Democratic Republic of Timor-Leste, Lover et al. [
49] explore the perceived malaria risk, causes of malaria, net usage patterns, barriers to protection and consistent use within families. The results indicate that there is an overall perception that mosquito nets should only be used by pregnant women and young children and there is a need for sufficient sleeping space under a limited number of nets within households. In conclusion, they emphasise that net usage is important for all members of the family, irrespective of age or gender, which highlights the complex behavioural economics. Pedersen et al. [
50] question whether the search for profit is the only goal, as assumed by traditional economic theory, or whether there are other factors that explain decision-making as suggested by behavioural economics. They research the effectiveness of incentive-based environmental policies. The results reveal two groups that are differentiated in terms of the use of pesticides: there are farmers who are more interested in maximising profits and others who focus more on land yields and who are less sensitive to the implementation of exclusively economic policy instruments. Clarke and Grenham [
51] study micro-insurance markets and their protection against catastrophes. Taking into consideration issues of supply and demand, aspects such as climate change and the associated risks are contemplated, which are typically covered by this type of insurance. In order to increase the demand for acquiring disaster insurance, governments should promote it, maybe using subsidies, with a commitment to limit the subsequent post-disaster financial assistance given to the uninsured. Lusk [
52] highlights the importance of the findings generated by BEA, revealing that the behaviour of the subjects is not consistent with the results of classic economic theory. However, he also indicates that Behavioural Economics cannot be used to justify all market failures. This is because consumers suffer from cognitive biases and, therefore, governments must act in a paternalistic way to conduct policy interventions. Miller et al. [
53] study agri-environmental schemes (AESs) and the factors that determine farmers’ decisions related to maintaining pro-environmental practices that go beyond what is established in contracts. They find that both pecuniary and non-pecuniary factors affect their decisions and that the influence of information regarding social norms is highly significant. Kuhfuss et al. [
54], continue studying the AESs. In order to improve the participation of the farmers and increase land enrolment for lower overall budgetary cost, they contemplate the implementation of a conditional collective bonus. This bonus would be paid in addition to the usual AES payment if a given threshold is reached in terms of farmers’ participation. The authors show that these bonuses increase expectations of farmers on others’ participation, therefore favouring a change towards a pro-environmental social norm and the adoption of less pesticide-intensive farming practices. Bouma [
55] study how, despite a commitment made by 195 countries of the United Nations when they signed the Sustainable Development Goals (SDGs) in 2015 and the research existing in this respect, the soil–water–plant–climate system still poses basic problems regarding soil behaviour that have yet to be resolved. It is necessary to share more information in order to be able to link the existing research with stakeholders and policy-makers. This is even more the case with the information revolution which affects the attitudes of increasingly critical stakeholders, making it difficult to discern between irrelevant and relevant information on the internet and social media.
Another field of study is related to policies referring to food production, food consumption and food security. In this respect, Just [
56] seeks to determine whether the food assistance programmes to combat obesity are more effective through traditional economic policies (such as manipulating information or prices) or policies related to behavioural economics and psychology. They find that the behavioural models are effective, but little is known about how eating behaviours interact with prices and other traditional mechanisms. Roosen and Marette [
57] analyse how the experiments contribute to the regulatory debate existing about the information referring to food quality and safety. They conduct a brief review of how laboratory and field experiments on food are complemented with theoretical analysis, discussing strengths and weaknesses. Goto et al. [
58] study whether environmental interventions can affect the decision-making of elementary school students. They determine whether students who are motivated to choose white milk due to environmental changes alter their total milk consumption. Their findings demonstrate that school-based practices guided by the theory of behavioural economics can offer useful insights and strategies for improving policies related to food selections. Chandon [
59] observes how packaging is important for food manufacturers and retailers as a marketing tool. This author studies how information relating to marketing, health and nutrition creates certain “health halos” leading to the belief that products are healthier than they actually are, with a positive effect in terms of increased consumption and the perception of a lower calorie intake. On the other hand, Reisch et al. [
60] examine the challenges faced by humankind in the near future related to the consumption and production system of current foods. In this sense, agricultural production must withstand the impacts generated by climate change, the growing conflicts related to land use and the social and health costs on an individual and societal level. With regard to nutritional aspects, Réquillart and Soler [
61] study how government policies related to nutrition have focused on informing consumers about the benefits of balanced diets, which has had a positive but modest effect. Recently, the attention of these policies has been directed towards market environments, with an emphasis on the characteristics of the food supply. Aiking [
62] addresses the issue of food security and food sustainability, given that in the next four decades the current food production will have to be doubled. They emphasise that for every kilo of animal protein, six kilos of plant protein are required, leading to concerns related to climate change and sustainability or the loss of biodiversity. Furthermore, intense livestock production is associated with antibiotics resistance and freshwater depletion. Liu et al. [
63] explain how the policy carried out in the USA to reduce obesity levels consisting in providing consumers with nutritional information about the products that they consume has had a very modest effect. They indicate that, among other factors, this is because it is necessary to have a certain level of understanding of the nutritional information and due to a conflict between motivation and lack of control. Lagi et al. [
15] explain the recent increase in food prices, which is affecting the most vulnerable populations around the world. The authors conclude that the principal causes reside in the investor speculations on ethanol conversion, driven by recent changes in the regulation of the commodity markets and in the policies implemented for ethanol conversion. Richards and Hamilton [
64] examine wasting food. They analyse the commercial peer-to-peer mutualisation systems (CPMSs), or sharing-economy firms, as platforms for exchanging food surpluses so that prices generate incentives for all of the actors to manage food surpluses more efficiently. The results show that the secondary markets are an effective way to reduce food waste. Lusk and McCluskey [
65] emphasise how public policies affect the decision of food consumers as their choices shape the food and farming system, with the known impacts on health, environment and food security. They discuss the future challenges such as diet-related illnesses and the efficiency of the policies aimed at improving dietary choices, confidence in the food system, farm technologies, environmental impacts of food consumption and food safety.
Since results on the traditional neoclassic economics have highlighted that these could not be supported in the praxis, the behavioural economics offer some explanations to the observed deviations from the theoretical expectations. In this sense, a great part of the scientific production on Behavioural Economics shows influences from Thaler et al. [
66], who contribute to the debate on libertarianism and paternalism regarding the State role. So can be the relevance of the term “nudge” be understood as a “gentle push to urge into action”, or more specifically “any aspect of the choice architecture that alters people’s behaviour in a predictable way without forbidding any option significantly changing their economic incentives”. This work establishes strong affirmations by which the responsibility over economic measures is shifted to the policy-maker. According to the measure design, nudges can be articulated so that individuals guide their decisions in a specific planned direction.
As the research corpus has increased, works summarizing the most relevant contributions have been identified through literature review. Particularly, Li and Just [
67] research about the factors explaining why behavioural economics has a great influence on the design of agricultural policies. This study focuses its analysis on two questions. On the one side, consumers’ behaviour regarding food and, on the other side, farmers’ decision-making in the food production and distribution phases. This work points out how the field of consumer behaviour is still limited, although many advances have been made in other disciplines, as well as in the development of experimental techniques. Moreover, references of previous works show the relevance of the preference inconsistency for sustaining individual habits. In the same way, some governmental practices based on a greater nutritional information of food have had a modest impact on the population and its caloric consumption. The authors also include previous works that highlight that individuals make a moderate cognitive effort when taking a decision regarding the consumption of food; they rather take advantage of the use of heuristic techniques and rules of thumb. Regarding this question, they offer references applied to school lunches. Related to Agricultural Production, the authors emphasize the challenges to apply behavioural economics on this field. Since each farm faces unique production possibilities and constraints, studies related to decisions under risk become special significance. It is observed how farmers wish to maximize benefits and, at the time, minimize risks through the behavioural model from the expected utility theory. In this sense, in order to manage risk and uncertainties, farmers’ studies offer examples of production diversification, as well some kinds of insurance coverage and governmental production subsidies. The studies cover the different attitudes toward risks that can be adopted (risk-averse vs. risk-lover), as well as entering into contracts. Further specific examples show how individuals tend to group all decisions without regarding if they provide benefits or losses, whereas in other samples they make a clearer difference. Finally, authors compiled research related to how farmers tend to value more changing in risk rather than changing in outcomes.
A further work that should be taken into account is the one by Streletskaya et al. [
68] who completed a literature review on agricultural technology adoption and behavioural economics, in order to better guide economic politics. For this reason, they established definitions, similarities and differences among both of them. Behavioural economics focuses on the study of deviations of decisions from the predictions within traditional economic models. The deviated decisions focus on intrinsic factors like preferences and cognition, among others. They collected these data through controlled economic experiments based on repeated games. This allows a causal analysis of the observed behaviours. The literature on agricultural practice adoption analyses those factors related to technology adoption by farmers and the evolution of uptake within populations. It focuses on extrinsic factors like physical, economic and demographic questions. It studies how populations adopt agricultural technologies through descriptive analyses and regressions. Furthermore, three areas for future research are identified: models of behaviour under risk and deviations from expected utility, behavioural time discounting models, and behavioural models of learning and social preferences.