Understanding the confidence of agricultural producers in an emerging market

Abstract

In this paper, we address the confidence of the agricultural producers in their prosperity as a driver of development. The agricultural sector development is highly dependent on short term investment decisions for sowing area, breeding, among others, which in turn is mediated by confidence. Agricultural producer confidence is continually measured in Argentina by CREA, a non-profit organisation composed of groups of producers spread across the country striving for innovation and best practices diffusion. We leverage this measurement methodology to propose a structural equation model to explain confidence by an exhaustive collection of factors affecting the business at company, sector or economy level, providing critical insight about the range of policies that would enable agriculture to grow. High regulation and tax burdens prevailing in the market during the past decades have been identified as relevant hindrance factors to confidence. On the enablers side are the respect for property rights and the organisation effectiveness actively fostered by CREA.

Keywords

Agricultural sector producer confidence structural equation model developing market Argentina

Introduction

The agricultural market presents unique characteristics. Thousands of producers of different crops and livestock form a scattered and heterogeneous supply side. The natural cycles set the pace for the industry, generating a dependence on short term investment. Sowing area decisions made every year, and livestock breeding decisions, also short term, favourable climate given, determine the production level. As a consequence of decentralised and frequent decisions, agricultural supply is highly dependent on the confidence of the producers on their business prosperity.

Confidence is a subtle variable belonging to the realm of belief and expectations. A protocol for continually measuring consumer confidence was designed and implemented by the University of Michigan. The methodology was adapted to the case of agricultural producers by the Asociación Argentina de Consorcios de Experimentación Agrícola (Argentine Association of Regional Agricultural Experimentation Consortiums, CREA).

Much has been speculated, but little is known about the true causes of the agricultural producer confidence variation. The purpose of this article is to investigate the factors affecting this subtle variable, to inform policy makers about the drivers of agricultural investment and production.

For that purpose, we have followed a three-phase research process. In the first phase, we explore the agricultural producer's mind by means of qualitative techniques, eliciting an exhaustive collection of potential causes of the confidence. In the second one, we address a representative sample of agricultural producers with a quantitative instrument, measuring confidence and previously identified potential causes. Finally, in the last phase, we create a structural equation model (SEM) to corroborate the causal structure of confidence.

Understanding the direct and indirect effects of different agricultural policies is very important in Argentina, as in most developing countries where the industry holds a prominent role. To the best of our understanding, this knowledge does not exist in the scientific literature yet.

In the next section, we present the theoretical foundations of the confidence index, we focus on the Argentine agricultural producer confidence index measured by CREA, we follow with an examination of agricultural producer confidence index in other countries, and we close with the knowledge gaps and objectives of this paper. In the third section, we describe our empirical strategy for building a causal model for the agricultural producer confidence. In the fourth section, we present the results of our empirical implementation. The discussion follows in the fifth section. Then, we conclude by summarising the results of the present research, their policy implications, and further research opportunities identified.

Literature review

Economic agents not only make decisions on account of objective and observable variables, but also on the basis of cognitive and emotional variables (Becker, 1978; Hosseini, 2011; Katona, 1953, 1974). Understanding agents’ perceptions provides valuable insight to anticipate their decision making (Kinsey and Collins, 1994).

The concept of economic agent confidence is more developed in the literature about the consumer, so we take the applicable concepts from there to adapt them to the agricultural producer case. Consumer confidence measures a highly subjective aspect of individuals’ psychology (Katona, 1979), reflecting a variety of psychological influences, such as financial insecurity, moods, emotions, and social norms associated with spending (Roos, 2008). Consumer confidence is sensitive to changes in economic indicators such as unemployment rate, GDP variation, inflation, among others (Soroka et al., 2015). Indeed, Ou et al. (2014) define consumer confidence as “a psychological construct that measures customers’ perceptions about their recent and future financial situation and economic climate”. Therefore, recognizing the relevance of consumer confidence to output growth, it clearly becomes important to analyse the factors that explain it.

Deniz and Çelik (2009) studied the link between economic growth and consumer confidence in six emerging economies in the European Union, being able to establish a long-term relationship between confidence and consumer spending. Jansen and Nahuis (2003) argued that the relationship between the stock market and consumer sentiment depends on expectations about general economic conditions.

Specifically in Argentina, the Finance Research Center of the Torcuato Di Tella University has produced a monthly consumer confidence index¹ since 1998. The index is calculated out of six questions about the personal economic situation and the economy in general, answered by the general population, following the methodology elaborated by the University of Michigan and by official agencies of the European Community.

Agricultural and livestock industries are very important economic sectors in Argentina, making it a worldwide leading food producer (World Bank, 2024). Agro-industrial activity is quite relevant in the country's economy, as in many other emerging and developing economies, significantly contributing to foreign trade, employment, and the development of communities throughout the country. Therefore, extending the knowledge about consumer confidence to the agricultural producer confidence becomes apparent to understand their behaviour.

Since the agricultural sector is eminently private, its development depends on the coordinated actions of a large number of producers. Therefore, it is necessary that these actors get the right signals so that they concentrate their investment and production efforts in the right direction. In order to anticipate the willingness of this sector to invest and grow, the Agricultural Producer Confidence Index has been developed by the Argentine Association of Regional Agricultural Experimentation Consortiums (CREA). CREA associates, representative of mid-to-large sized producers in Argentina, answer six questions about the national and sectoral economy and the financial and economic context of business that feed the confidence index. CREA has issued the index every four months since November 2012. The evolution of the index quantifies the perception and expectations of agricultural producers.

In this study, we intend to understand the factors influencing the confidence of the agricultural producer. This knowledge will allow the authorities to design more effective public policies for the development of the agricultural sector and to neutralise potential adverse conditions.

Consumer confidence as a precedent

Global economic theory has identified that expecting rationality from economic agents can be a restrictive assumption (Katona, 1953). For this reason, more recently, the Behavioural Economics Theory has been developed to better explain the behaviour of economic agents and to understand decision-making (Hosseini, 2011; Katona, 1951, 1953, 1974). In this sense, Kahneman and Tversky (1979) and Tversky and Kahneman (1986) identify a series of typical human cognitive defects that violate the rationality assumption of the economic theory, particularly the principle of expected utility. Thus, the Theory of Perspectives emerged as a descriptive theory of human behaviour in relation to the economy. The entry of psychology into the field of economics has been a gradual process, still ongoing. This theory focuses on the behaviour of economic agents considering psychological, social and emotional aspects at the agent level (Becker, 1978; Michael and Becker, 1973). This implies the assumption that economic agents make decisions based on objective and observable variables and, also, on perceptual and emotional variables (Becker, 1978; Hosseini, 2011; Katona, 1953, 1974). For this reason, understanding agents’ perceptions of economic variables would make it possible to better anticipate their behaviour (Kinsey and Collins, 1994).

The first consumer confidence index (CCI) was developed at the University of Michigan, for the USA in 1946 (Katona, 1946, 1974). The survey was conducted by telephone and was financed by the USA Federal Reserve with the aim of analysing consumption and people's perception of the future of the economy (Katona, 1974). The theoretical model on which this index was developed was proposed by Katona who, through the analysis of economic agents in the early 1940s, argues that consumption is influenced by two forces: the ability and the willingness to buy (Curtin, 2007; Katona, 1946, 1974). The first of these forces is defined as the real purchasing capacity, explained by the current or past income and accumulated assets of economic agents. The other force is influenced by the consumer's perception of the future of the economy. Then, the willingness to buy could explain that when the perception about the economy is positive, the agent increases consumption; whereas when the perception is negative, savings increase at the expense of consumption as a safeguard for future times (Curtin, 2002, 2007; Katona, 1946, 1953). The University of Michigan started an ongoing study based on a four-monthly telephone survey of 500 USA consumers in 1958. The survey has been conducted monthly since 1978 by the Survey Research Center of the University of Michigan. This methodology has spread throughout the world. Over 45 countries were already surveying consumer confidence in 2017, including most of the developed countries, and some emerging economies, such as Argentina and Brazil (Skikiewicz and Błoński, 2017; Yao et al., 2013).

Although the impact of consumer confidence on economic variables is not fully understood, several studies estimate magnitudes and impacts on macroeconomic aggregates, e.g., consumption, savings, and investment by means of econometric methods (Caballero and Krishnamurthy, 1998; Carroll et al., 1994; Chopin and Darrat, 2000; Cummins et al., 2006; Shackle, 1952). More specifically, there are studies that correlate the CCI with observable variables. For example, Garner (1981) uses the Granger causality test² to relate the CCI with top five stock prices, real disposable income and inflation rate. With a similar methodology, Huth et al. (1994) find a relationship between consumption and investment with the CCI. Furthermore, there are documents that analyse the predictive capacity of the CCI in macroeconomics, studying the relationship of the index with the interest rate, the monetary base, the inflation, the Dow Jones index, the individual income and wholesale sales (Chopin and Darrat, 2000). Also, other researchers relate the consistency of the CCI to household expenditures (Carroll et al., 1994), durable goods and food prices (Kinsey and Collins, 1994), and to consumer expenditures (Carroll et al., 1994; Cotsomitis and Kwan, 2006; Gelper et al., 2007).

Curtin (2007), using the Granger causality test for 37 countries, explains the causal relationships between the CCI and the most important variables of the economy (GDP, unemployment rate, inflation rate, personal income, personal consumption, medium-term interest rate, wholesale sales, consumption of durable goods and vehicle patents). The most significant causal relationships found were the following (Curtin, 2007):

In 59% of the countries, it was found that a GDP drop has a negative impact on the CCI, and vice versa, in 56% of the countries a decrease in the CCI has a negative effect on the GDP.

In the case of the individual's personal income, variations in this variable have an impact on the CCI in 55% of the cases; similarly, changes in the CCI influence personal income in 48% of the countries.

Another relevant variable is the unemployment rate, which has an impact on the confidence in 50% of the countries, and changes in the CCI influence the unemployment rate in 62% of the countries.

With respect to the consumption of durable goods, changes in this variable impact the CCI in 54% of the countries and variations in the CCI affect consumption in 46% of the cases.

Other researchers find correlation between the consumer confidence index and economic variables (Afshar Jahanshahi et al., 2011; Blanchard, 1993; Haugh, 2005; Howrey, 2001; Matsusaka and Sbordone, 1995; Starr, 2012), like growth rate, unemployment, interest rates and exchange rates (Barsky and Sims, 2012; Vuchelen, 2004). As a result, the confidence index has attracted the attention of market analysts, policy makers and macroeconomists (Zafeiriou et al., 2019).

Considering these findings, it is quite clear that consumer confidence holds valuable information to anticipate or to understand their decisions and actions.

Agricultural producer confidence

Consumer confidence has been broadly studied in the literature, however, research on producer confidence, and particularly in the agricultural sector, is less developed. Among the scarce references about this subject is the agricultural economy barometer of Purdue University in the USA (“Survey Methodology Ag Economy Barometer”), which provides a national measure to capture attitudes and sentiment regarding the state of the USA agricultural economy. It is based on a monthly five-question survey of 400 agricultural producers, and it is accompanied by a quarterly in-depth survey of 100 agricultural and agribusiness opinion leaders. The agricultural confidence index in Ecuador (Vera-Gilces et al., 2020), not implemented yet, is designed to gather the expectations of exporters and producers, identifying breakpoints, and anticipating changes in the short term both in the performance of the sector as well as in the pillars and variables of the index.

CREA has adapted the CCI to the agricultural producer case in Argentina³, and implemented it in its regular survey⁴ since November 2012. The methodology consists of six questions for the calculation of three sub-indexes: general economic situation, agricultural sector context, and company situation, which are averaged to obtain the confidence index (Fusco et al., 2017).

Understanding the causes of confidence is also important for the agricultural producer, as demonstrated by Ward et al. (2007) who describe the difficulties in the implementation of an agricultural programme entangled in the subtleties of confidence and cultural differences.

Modelling confidence

Causal analysis is highly developed in the public services domain. The quality of service as measured by user satisfaction and loyalty is investigated by means of statistical models that can confirm or reject proposed causal structures. In developed countries, there are measurement systems known as satisfaction barometers, such as the ECSI (European Customer Satisfaction Index) (Johnson et al., 2001), or the ACSI (American Customer Satisfaction Index) (Fornell et al., 1996), which not only monitor the level of satisfaction longitudinally, but also record a series of variables through a cross-sectional sample of users of the service under study. These variables measure the potential causes of satisfaction: service provider image, service quality, expectations, value perception, user loyalty, among others. Then, structural equation models (SEM) are proposed to link these observed attributes to satisfaction through latent variables, representing psychological states of the user. The structural model is calibrated with the aforementioned empirical basis by means of the maximum likelihood method (Bollen, 1989) or more robust methods, such as Partial Least Squares (Esposito Vinzi et al., 2010) when required by the nature of the variables. The data is not collected directly, but through questions that are easy for users to answer. These manifest variables make it possible to indirectly measure the latent variables, or psychological states, through the measurement model which is part of the SEM. In this way, the intensity of the causal relationships between the variables is determined, highlighting the actions required by the policy makers (government or private providers) to improve user satisfaction.

SEM have been created to understand human attitudes and behaviour in many domains, and they can certainly be used to investigate confidence. There are several studies analysing causality in consumer confidence based on cross-sectional ad-hoc surveys (Lassoued and Hobbs, 2015; Stiawan and Arisandy, 2016). However, we have not found SEM in agricultural producer confidence. Indexes serve only as monitoring instruments, but they are not equipped with the causal structure required for SEM.

Gaps and contributions

According to the literature review, no causal analysis has been performed scientifically on the agricultural sector confidence, in particular on producers. Actually, causal analysis on consumer confidence has been conducted, but not yet in an emerging economy like Argentina.

In this article, we aim at filling this gap with a SEM of the agricultural producer confidence in Argentina. This constitutes a starting point for improving the decision-making process within CREA and government policy design.

Empirical strategy

The measurement instrument

The design of the instrument is based on qualitative research. In-depth interviews were conducted with diverse CREA members in terms of activity (agriculture, livestock, and combined), size and region. The interviews were recorded and analysed by the research team by means of the Brainswarming technique, which proposes silent generation of ideas in a group setting (McCaffrey, 2018). As a result, cognitive maps (Eden, 2004) were elaborated revealing a wide set of areas of interest of the producers potentially impacting their confidence and their relationships, enabling the identification of clusters of the most decisive areas of interest and the elements within them. Questions were designed to measure the degree of satisfaction of the need or desire underlying each element identified in the qualitative research, by means of 0 to 10 agreement scales. The areas of interest identified are economic environment, regulatory environment, supplies and logistics, company management capability, and climate expectations. Specific items in each area of interest are detailed in the Appendix. The questionnaire was supplemented with questions about general business expectation, intention to recommend, and the six CREA confidence index questions. The questionnaire was distributed among all CREA members as a module in the regular survey, obtaining 586 complete responses.

The model

A structural equation model (SEM) is proposed to evaluate the relationship between observable and non-observable or latent variables. This approach is formally defined by two sets of linear equations: the structural model and the measurement model. The structural model postulates a causal relationship among the relevant variables following the qualitative research findings, as in a linear model setting. However, these variables are not assumed to be directly measurable. They are latent variables in the sense that they represent actual concepts or constructs in the agricultural producer mind that are not observed. Instead, we employ the observed variables from the questionnaire to measure the latent variables. This constitutes the measurement model. Each latent variable is measured by a set of coherent observed variables, in the sense that they represent a unidimensional set. The latent variable is the common aspect within its set of observed variables or indicators, as in a factor analysis.

An exploratory factor analysis was performed to the variables identified in the qualitative research and measured by means of the survey, to determine different dimensions occurring within each area of interest. Four dimensions were determined in the economic environment area of interest: the consumer purchasing power, strengthening the demand of livestock products; the international market crop prices; the export duties cutting the latter; and, the overall tax burden, putting pressure on the agricultural business. Two dimensions were determined in the regulatory environment area: the property rights, mainly related to the land; and, the regulation burden derived from different state agencies hindering the business and raising administrative costs. Two dimensions were recognized in the supplies and logistics area: the availability of supplies and services at reasonable costs, like seeds, fertilisers, sowing, harvesting, fumigation, transportation, and others; and, the availability of infrastructure, namely roads, railways, storage, et cetera. Two dimensions were distinguished in the company management capability area: organisation effectiveness, and customer orientation. The climate area of interest was unidimensional. In this way, eleven explanatory variables were identified.

The proposed structural model is designed in three layers of latent variables, as shown in Figure 1. The middle layer has the three dimensions of the CREA confidence index: confidence in the company, the agricultural sector and the general economy. These three confidence dimensions are proposed to explain the general confidence on the right. The left layer holds the exogenous variables presumably driving the three dimensions of the confidence, and eventually general agricultural producer confidence. Confidence in the economy is explained by the exogenous variables related to the macroeconomic environment: regulation burden, tax burden, and consumer purchasing power. Confidence in the agricultural sector is illustrated by the exogenous variables impacting the whole agricultural sector: international crop prices, export duties (specific for the sector), supplies and services, infrastructure, and climate expectations. Confidence in the company is explained by respect for property rights, customer orientation, organisation effectiveness, regulatory burden, supplies and services, infrastructure, climate expectations, being the three latter also explanatory of the sector confidence.

Figure 1.

Path diagram of the structural model.

Figure 2.

Structural model coefficients.

The specification of the measurement model is shown in Table 1, where the indicators reflected by or forming each latent variable are listed (please see the Appendix for description of indicators).

Table 1.

Measurement model.

Latent Variable	Indicators	Model Type
Tax Burden	ECON 8, ECON 9, ECON 10	reflective
Export Duty	REGU 6	reflective
Purchasing Power	ECON 1, ECON 2, ECON 3	reflective
Product Prices	ECON 4, ECON 5, ECON 7	reflective
Property Rights	REGU 1, REGU 3	reflective
Regulation Burden	REGU 4, REGU 5	reflective
Supplies and Services	S&L 1, S&L 2, S&L 3, S&L 4, S&L 6	reflective
Infrastructure	S&L 7, S&L 8	reflective
Organisation Effectiveness	STRA 8, STRA 9, STRA 10, STRA 11	reflective
Customer Orientation	STRA 6, STRA 7	reflective
Climate	CLI 1	reflective
Confidence in the Economy	Q1, Q2	formative
Confidence in the Sector	Q5, Q6	formative
Confidence in the Company	Q3, Q4	formative
General Confidence	C1, C2	reflective

The latent variables inherit the positive formulation of their indicators. For instance, a high level of property rights means they are respected, whereas a high level of regulation burden means regulations are light.

The three confidence dimensions (confidence in the economy, in the agricultural sector, and in the company) were measured by following the definition in the CREA confidence index (Fusco et al., 2017). The general confidence was measured independently of the elements just mentioned, by two indicators: the expectation of a successful business in the near future (Ou et al., 2014), and the willingness to recommend the business to a friend. The former is basically the statement of confidence, whereas the latter is a leading indicator of business growth (Reichheld, 2003), which follows from confidence.

Results

The analysis plan first considers the measurement model and then the structural model where the interplay among the latent variables is discussed, by following the practices recommended by Hair et al. (2019). The analysis of the measurement model takes into account internal consistency reliability, convergent validity, and discriminant validity. The analysis of the structural model takes into consideration collinearity, path coefficients and predictive capability.

Internal consistency reliability occurs when the indicators measuring each latent variable are associated with each other. To examine the reliability of internal consistency, we use the rho-c composite reliability measure proposed by Jöreskog (1971), Cronbach's alpha, and Dijkstra's rho-a (Dijkstra, 2010), which relaxes the restriction in Cronbach's alpha, assuming that all indicator loadings are equal at population level. As shown in Table 2, all measures lie in the [0.60, 0.95] interval as expected for a reliable model. Two exceptions are the latent variables export duties and climate that are measured by a single indicator, hence presenting perfect consistency.

Table 2.

Measurement model - internal consistency.

Latent Variable	N indicators	alpha	rhoA	rhoC	AVE
Purchasing Power	3	0.70	0.71	0.83	0.62
Tax Burden	3	0.83	0.84	0.90	0.75
Regulation Burden	2	0.84	0.85	0.93	0.86
Property Rights	3	0.77	0.78	0.87	0.69
Customer Orientation	2	0.75	0.75	0.89	0.80
Organisation Effectiveness	4	0.76	0.77	0.85	0.58
Infrastructure	2	0.72	0.78	0.87	0.77
Prices	3	0.68	0.71	0.82	0.61
Export Duty	1	1	1	1	1
Supplies and Services	5	0.84	0.84	0.89	0.61
Climate	1	1	1	1	1
General Confidence	2	0.73	0.73	0.88	0.79

Convergent validity occurs when the latent variable explains most of the variance of its indicators. The corresponding metric is the mean variance extracted (AVE) for all indicators of each construct, expected to be above 0.50 (Hair et al., 2021), which is the case for our model as shown in Table 2.

Discriminant validity refers to the empirical difference among exogenous latent variables. The corresponding metric is the heterotrait-monotrait correlation ratio (Henseler et al., 2015), revealing if the indicators of each latent variable are closer among each other than to indicators measuring other latent variables. The HTMT ratio, exhibited in Table 3, lies below 0.90 in all cases supporting discriminant validity. Its upper bootstrap interval extreme is shown in the Appendix, underlying this conclusion.

Table 3.

Measurement model - discriminant validity.

	Purch. Power	Tax Burden	Regul. Burden	Prop. Rights	Cust. Orient.	Org. Effectiv.	Infrastructure	Prices	Export Duty	Sup. and Serv.
Tax Burden	0.77
Regulation Burden	0.53	0.66
Property Rights	0.62	0.61	0.84
Customer Orientation	0.20	0.11	0.20	0.27
Organiz. Effectiveness	0.13	0.08	0.12	0.18	0.57
Infrastructure	0.41	0.35	0.33	0.36	0.14	0.17
Prices	0.40	0.38	0.19	0.33	0.11	0.38	0.25
Export Duty	0.52	0.63	0.79	0.76	0.14	0.09	0.29	0.19
Supplies and Services	0.25	0.14	0.12	0.23	0.09	0.28	0.50	0.66	0.08
Climate	0.12	0.12	0.09	0.21	0.06	0.12	0.15	0.13	0.10	0.17

The first level of endogenous latent variables is specified as formative. The confidence in the economy is indicated by the assessment of the current and the expected situations. These indicators, which may not be consistent, reveal the two aspects of the latent variable, then it is specified as formative. A similar case occurs for the confidence in the company. The confidence in the sector is indicated by the perception of the opportunity to invest and the expectation about product prices. These indicators, which may also not be consistent, form the aforementioned latent variable. The validation process is different for formative variables, covering coefficient significance, low collinearity, and convergent validity. The formative coefficients are all significant as shown in Table 4, and the collinearity level inside each formative model is negligible. Convergent validity is founded on Fusco et al. (2017).

Table 4.

Measurement model - formative variables validation.

		Coef.	VIF
Confidence Economy	Current Situation	0.77	1.07
Confidence Economy	Expected Situation	0.83	1.07
Confidence Sector	Time to invest	0.79	1.02
Confidence Sector	Expected prices	0.71	1.02
Confidence Company	Current Situation	0.75	1.02
Confidence Company	Expected Situation	0.77	1.02

As a consequence of the precedent validations, we conclude that all latent variables are properly measured. This allows us to proceed to the analysis of the structural model, where the relationship among them is assessed.

The structural model is composed of four predictive models, one for each endogenous latent variable. The results are presented in Table 5, including the path coefficients, their bootstrap confidence intervals, and the R² fit assessment indicator. The confidence intervals indicate that all path coefficients are significantly different from zero, confirming the influence of the explanatory variables on the dependent ones. The magnitude of the coefficients reflects the strength of this influence.

Table 5.

Structural model.

	R²	Explanatory Variable	Coeff.	Bootstrap Interval
	R²	Explanatory Variable	Coeff.	Q 5%	Q 95%
Confidence Economy	0.32	Purchasing Power	0.22	0.15	0.31
		Tax Burden	0.22	0.11	0.32
		Regulation Burden	0.25	0.17	0.33
Confidence Sector	0.31	Prices	0.21	0.15	0.28
		Export Duty	0.18	0.12	0.24
		Supplies & Services	0.22	0.14	0.29
		Infrastructure	0.19	0.13	0.26
		Climate	0.06	0.00	0.11
Confidence Company	0.40	Regulation Burden	0.12	0.05	0.19
		Supplies & Services	0.27	0.21	0.33
		Infrastructure	0.11	0.05	0.17
		Climate	0.13	0.07	0.18
		Property Rights	0.26	0.20	0.33
		Organization Effectiveness	0.10	0.04	0.17
		Customer Orientation	0.08	0.02	0.13
General Confidence	0.59	Confidence Economy	0.28	0.23	0.32
		Confidence Sector	0.38	0.33	0.43
		Confidence Company	0.37	0.32	0.42

The structural model and results are depicted in Figure 2, where the coefficients linking the latent variables are shown. Exogenous variables are represented by white ovals, while blue ovals indicate endogenous latent variables. The measurement model is omitted in the interest of clarity.

Robustness of the model

Several model alternatives were evaluated to arrive at the final one described above. The first proposed model linked all exogenous variables directly to general confidence. Although this model showed a promising chi-square of 861 with 370 degrees of freedom and a RMSE of 0.046, only 5 out of 11 path coefficients were significant, with one even contradicting the theoretical expectations. Subsequently, the partial confidence variables (confidence in the economy, sector, and company) were introduced between the exogenous variables and general confidence, similar to the arrangement in Figure 2 but with additional connections. This adjustment increased the chi-square to 1170 with a RMSE of 0.050. However, issues were identified in the measurement model. The indicators for the partial confidence variables were not entirely consistent with each other, instead capturing different aspects of the corresponding constructs, suggesting that they should be modelled formatively. This issue led us to shift from Maximum Likelihood to PLS estimation, which is more robust in handling formative variables (Rigdon et al., 2014) and less reliant on distributional assumptions (Hair et al., 2021). The revised model performed fairly well, with R² values of 0.59 for general confidence, and 0.32, 0.31, and 0.44 for the partial confidence constructs. Unfortunately, PLS estimation does not provide reliable overall goodness-of-fit indicators (Hair et al., 2021). However, it was evident that there were opportunities for improvement, as some structural coefficients were not statistically significant. The final model was refined through several iterations.

Discussion

The structural model, as depicted in Figure 2, contains four predictive models to explain the endogenous latent variables. The confidence in the economy is explained by the regulation burden, the tax burden, and the purchasing power. The three path coefficients are statistically significant with positive sign and similar magnitudes (see Table 5). The tax pressure significantly increased in Argentina since 2002, soaring from 20% in 2000 to 33% in 2023 (Organization for Economic Cooperation and Development (OECD), 2024). On top of that are hidden taxes due to differential exchange rate for exporters⁵ and inflation⁶, which would take it around 40%. The highly efficient agricultural sector bore a disproportionate share of this pressure⁷, facing a plethora of different taxes among which the most significant are⁸ eight taxes at national level, four at provincial level and even some more at municipal level depending on the location. In fact, the agricultural sector in Argentina has suffered negative state support rather than promotion as in most countries (Organization for Economic Cooperation and Development (OECD), 2023). The positive coefficient shows that relaxing the tax pressure on the agricultural sector would increase the confidence of the agricultural producers. The regulation burden has also been increased along with taxes in the past two decades. The regulations agencies at the national level are the tax agency (AFIP), the seeds agency (INASE), and the health and food quality agency (SENASA). AFIP requires periodic filing for each tax, and INASE and SENASA, not without imprudence, announce a “simplified” system requiring respectively 38 and 123 different formalities to be fulfilled by producers, including registrations, trip specific permits for transportation of crops and livestock, among others⁹. Additional regulations exist at the provincial and municipal level. The positive coefficient shows that relaxing the regulation burden would bring back a freer business environment and increase the confidence of the agricultural producers. We also find that confidence is sensitive to the purchasing power of consumers. This variable increases the internal demand for livestock products, whereas not that much for crops which are mainly exported. However, the overall impact of purchasing power on confidence is positive and statistically significant. A public policy increasing the purchasing power of consumers would contribute to the development of the agricultural sector.

Narrowing down to the agricultural sector, more specific variables enter the model: export duties, international crop prices including credit availability and infrastructure, supplies and services, and climate expectations. All path coefficients are statistically significant with positive signs (see Table 5). The magnitudes are similar with the exception of the coefficient of climate that is weaker. International prices are obviously beyond control of the government, but export duties, heavily cutting crop prices, are within control and represent a strong driver for agricultural producer confidence and sector development. Supplies and services operate in a free market; however, the state has indirect influence on this market through import restrictions and duties on agrochemicals, as well as fumigation regulations. They should benefit from a better macroeconomic environment, lower inflation, higher economic growth, credit availability, less capricious regulations, et cetera. On the other hand, the state is highly involved in infrastructure development, both at provincial and municipal levels where the road improvement decisions are made, and at the national level where the decisions about railways take place. A consistent policy at different state levels should be undertaken to facilitate agriculture sector growth. Climate expectation has a weaker influence on sector confidence, but it is more relevant at company level as discussed below.

The third predictive model focuses on confidence in the company. Explanatory variables are infrastructure, supplies and services, climate expectations, property rights, customer orientation, organisation effectiveness, and regulation burden. The path coefficients are all positive and statistically significant (as shown in Table 5). The highest influence on confidence is produced by the availability and affordability of supplies and services, and by the respect for property rights. The former was also important at sector level. The latter refers mainly to real estate rights, which have eroded in the past two decades. The governments of the last two decades led to excessively lenient treatment of land squatters, who made progress in the Buenos Aires province, very rich in agriculture production, as well as in other agriculture involved provinces. This violent activity frightened landowners, bubbling up to a social concern by 2020 as illustrated by Smink (2020). This is clearly hindering agricultural producers’ confidence, as demonstrated by the statistically significant coefficient. Climate expectation comes next, showing a stronger effect on confidence at company level than at sector level, reflecting geographic heterogeneity. Agricultural producers are heavily impacted by climate events like drought, floods, and hail. Even though these are low probability events, and they are beyond government control, a more developed insurance market would help agricultural producers neutralise these risks, and the state is involved as an enabler of this highly regulated activity. The regulation burden loading on agricultural activity also shows an impact at the company level. Then come infrastructure and organisation effectiveness, and finally customer orientation, which tends to be of lesser importance, probably due to the fairly standardised products. We have discussed infrastructure development policies above. With regard to organisation effectiveness, it is worth mentioning that it is not a trivial capability to develop, as pointed out by Shankar and Maraty (2009), who propose a government policy to create farmer groups. In Argentina, this is taken care of by CREA, a non-profit private organisation that formed bottom-up with the mission to share experience, generate knowledge and promote ideas among group members for the sustainable development of agricultural companies and the country in general. CREA is a strong network of agricultural producers present across the country that became references of innovation and sustainability. This has certainly been a robustness factor that helped to keep agriculture up and running during decades of adverse conditions; and should be considered for replication in other developing nations.

The final model relates general confidence to the three partial dimensions of confidence. We find a strong contribution of each dimension to the independently measured general confidence, demonstrated by positive and statistically significant coefficients (see Table 5). Confidence in the sector and in the company have the greatest influence on general confidence, whereas confidence in the economy seems more detached from the agricultural producers’ day to day. This enhances the importance of the exogenous variables influencing confidence in the sector and in the company discussed above. In terms of public policy, the respect for property rights or legal security, the relief in the regulatory and tax burdens and export duties, and infrastructure development are the most promising areas of focus for the development of the agricultural sector.

Conclusions and further research

In this article, we analyse the confidence of the agricultural producers in Argentina in their business prosperity. The confidence has been linked to investments and future production level, however its causes were yet unexplored at scientific level.

The confidence is analysed into three dimensions, as defined in the literature: confidence in the economy, in the agricultural sector, and in the company. We find that the three dimensions contribute to general confidence. Confidence in the company and the sector have a more concrete and intense influence, compared to confidence in the economy, which is a more distant concept.

In order to increase the agricultural producer's confidence in the company and the sector, there are several public policies that show promising potential. The regulatory and tax burdens on the sector have reached a very high level during the past two decades. Relaxing them would open positive perspectives for the agricultural industry. The legal security protecting the property rights on the land is another of the main factors for confidence. Land squatting, even if scattered, has produced a high impact on landowners confidence. Better legal handling of these events would help to strengthen the confidence. Infrastructure development is another factor that would enhance sector development and should be actively pursued by the government, either directly or by means of private investors.

We have proposed the most natural causal model following qualitative research, where the exogenous variables influence the different dimensions of confidence consistently with economic theory and qualitative research, and the endogenous variables interact among themselves in a straightforward manner. However, other structural models could be envisioned, with latent variables interacting in different ways. Further research should be pursued to better understand the causal relationships of the confidence of the agricultural producer.

Another avenue for understanding the confidence of agricultural producers is by delving into the psychological dimension. Confidence is fundamentally a belief in one's own abilities and capacity to accomplish tasks. By identifying and measuring these beliefs, deeper insights could be obtained. The methods developed in this article primarily target the population at an individual level, which provides an opportunity to incorporate psychological factors into the analysis. It could be argued that individual characteristics, such as personality traits, are less actionable and might appear irrelevant in the context of policy-making. However, this perspective overlooks the potential for policies to be designed in ways that positively influence the psychological state of producers. For example, initiatives could be developed to foster a growth mindset, resilience, and self-efficacy among agricultural producers, which in turn could lead to an indirect yet significant increase in their confidence levels. These psychological interventions could complement existing agricultural policies by addressing the mental and emotional factors that contribute to a producer's overall confidence, thereby creating a more holistic approach to supporting their success.

Finally, another line to be pursued in further research is the geographical extension. It would be interesting to investigate the applicability of the confidence model in more developed markets, like Europe and the USA. We expect them to be different between each other, given the higher land ownership fragmentation in the former. It would also be interesting to know the results for African markets, where agriculture is a larger part of the economy. The methods developed for this article can be applied to any market with minor adaptation of the indicators.

Footnotes

Acknowledgements:

We appreciate the collaboration from CREA (Asociación Argentina de Consorcios Regionales de Experimentación Agrícola) for the empirical implementation.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Ethics approval and informed consent

This article does not contain any studies with human or animal participants. There are no human participants in this article and informed consent is not required.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

David Montemurri

Valeria A Gogni

Notes

Appendix

Table A1. Indicators in the Instrument.

References

Afshar Jahanshahi

Hajizadeh Gashti

Mirdamadi

, et al. (2011) Study of the effects of customer service and product quality on customer satisfaction and loyalty. International Journal of Humanities and Social Science 1: 253–260.

Barsky

Sims

(2012) Information, animal spirits, and the meaning of innovations in consumer confidence. American Economic Review 102(4): 1343–1377.

Becker

(1978) The Economic Approach to Human Behavior. Chicago, IL: University of Chicago Press.

Blanchard

(1993) Consumption and the recession of 1990-1991. The American Economic Review 83(2): 270–274.

Bollen

(1989) Structural Equations with Latent Variables. Structural equations with latent variables. Oxford, England: John Wiley & Sons.

Caballero

Krishnamurthy

(1998) Emerging Market Crises: An Asset Markets Perspective. NBER Working Papers 6843: 1–45.

Carroll

Fuhrer

Wilcox

(1994) Does consumer sentiment forecast household spending? If so, why? The American Economic Review 84(5): 1397–1408.

Chopin

Darrat

(2000) Can consumer attitudes forecast the macroeconomy? The American Economist 44(1): 34–42.

Cotsomitis

Kwan

ACC

(2006) Can consumer confidence forecast household spending? Evidence from the European commission business and consumer surveys. Southern Economic Journal 72(3): 597–610.

10.

Cummins

Hassett

Oliner

(2006) Investment behavior, observable expectations, and internal funds. The American Economic Review 96(3): 796–810.

11.

Curtin

(2002) Surveys of Consumers: Theory, Methods, and Interpretation. In: NABE 44^th Annual Meeting, Washington, DC, USA, 30 September 2002. Available at: https://data.sca.isr.umich.edu/fetchdoc.php?docid=26396.

12.

Curtin

(2007) Consumer sentiment surveys: Worldwide review and assessment. Journal of Business Cycle Measurement and Analysis 2007(1): 7–42.

13.

Deniz

Çelik

(2009) An empirical investigation of twin deficits hypothesis for six emerging countries. SSRN Electronic Journal: 1–13. DOI: https://doi.org/10.2139/ssrn.1389242.

14.

Dijkstra

(2010) Latent variables and indices: Herman wold’s basic design and partial least squares. In: Esposito Vinzi

Chin

Henseler

, et al. (eds) Handbook of Partial Least Squares: Concepts, Methods and Applications. Springer Handbooks of Computational Statistics. Berlin, Heidelberg: Springer, 23–46. Available at: https://doi.org/10.1007/978-3-540-32827-8_2 (accessed 3 March 2024).

15.

Eden

(2004) Analyzing cognitive maps to help structure issues or problems. European Journal of Operational Research 159: 673–686.

16.

Esposito Vinzi

Chin

Henseler

, et al. (2010) Handbook of Partial Least Squares: Concepts, Methods and Applications. Berlin, Heidelberg: Springer.

17.

Fornell

Johnson

Anderson

, et al. (1996) The American customer satisfaction Index: Nature, purpose, and findings. Journal of Marketing 60(4): 7–18.

18.

Fusco

Pederiva

Barelli

(2017) Índice de confianza de los empresarios agropecuarios en Argentina. Revista de Investigación en Modelos Financieros I: 16.

19.

Garner

(1981) Economic determinants of consumer sentiment. Journal of Business Research 9(2): 205–220.

20.

Gelper

Lemmens

Croux

(2007) Consumer sentiment and consumer spending: Decomposing the granger causal relationship in the time domain. Applied Economics 39(1): 1–11.

21.

Hair

Hult

GTM

Ringle

, et al. (2021) Partial Least Squares Structural Equation Modeling (PLS-SEM) Using R. Cham, Switzerland: Springer.

22.

Hair

Risher

Sarstedt

, et al. (2019) When to use and how to report the results of PLS-SEM. European Business Review 31(1): 2–24.

23.

Haugh

(2005) The Influence Of Consumer Confidence And Stock Prices On The United States Business Cycle. CAMA Working Papers 2005–03. Available at: https://cama.crawford.anu.edu.au/publication/2363/influence-consumer-confidence-and-stock-prices-united-states-business-cycle.

24.

Henseler

Ringle

Sarstedt

(2015) A new criterion for assessing discriminant validity in variance-based structural equation modeling. Journal of the Academy of Marketing Science 43: 115–135.

25.

Hosseini

(2011) George katona: A founding father of old behavioral economics. The Journal of Socio-Economics 40(6): 977–984.

26.

Howrey

(2001) The predictive power of the Index of consumer sentiment. Brookings Papers on Economic Activity 32(1): 175–216.

27.

Huth

Eppright

Taube

(1994) The indexes of consumer sentiment and confidence: Leading or misleading guides to future buyer behavior. Journal of Business Research 29(3): 199–206.

28.

Jansen

Nahuis

(2003) The stock market and consumer confidence: European evidence. Economics Letters 79(1): 89–98.

29.

Johnson

Gustafsson

Andreassen

, et al. (2001) The evolution of future of national consumer satisfaction Index models. Journal of Economic Psychology 22: 217–245.

30.

Jöreskog

(1971) Simultaneous factor analysis in several populations. Psychometrika 36(4): 409–426.

31.

Kahneman

Tversky

(1979) Prospect theory: An analysis of decision under risk. Econometrica 47(2): 263–291.

32.

Katona

(1946) Psychological analysis of business decisions and expectations. The American Economic Review 36(1): 44–62.

33.

Katona

(1951) Psychological Analysis of Economic Behavior. New York, NY, US: McGraw-Hill.

34.

Katona

(1953) Rational behavior and economic behavior. Psychological Review 60(5): 307–318.

35.

Katona

(1974) Psychology and consumer economics. Journal of Consumer Research 1(1): 1–8.

36.

Katona

(1979) Toward a macropsychology. American Psychologist 34(2): 118–126.

37.

Kinsey

Collins

(1994) Index of consumer expectations: Food price effects and durable goods expenditures. The Journal of Consumer Affairs 28(2): 255–277.

38.

Lassoued

Hobbs

(2015) Consumer confidence in credence attributes: The role of brand trust. Food Policy 52: 99–107.

39.

Matsusaka

Sbordone

(1995) Consumer confidence and economic fluctuations. Economic Inquiry 33(2): 296–318.

40.

McCaffrey

(2018) Overcome Any Obstacle to Creativity. Lanham: Rowman & Littlefield Publishers.

41.

Michael

Becker

(1973) On the new theory of consumer behavior. The Swedish Journal of Economics 75(4): 378–396.

42.

Organization for Economic Cooperation and Development (OECD) (2023) Agricultural Policy Monitoring and Evaluation 2023: Adapting Agriculture to Climate Change. Report, Agricultural Policy Monitoring and Evaluation (OECD). Available at: https://www.oecd-ilibrary.org/agriculture-and-food/agricultural-policy-monitoring-and-evaluation-2023_b14de474-en (accessed 3 May 2024).

43.

Organization for Economic Cooperation and Development (OECD) (2024) The design of presumptive tax regimes in selected countries. Report, OECD. Available at: https://www.oecd.org/countries/argentina/the-design-of-presumptive-tax-regimes-in-selected-countries-58b6103c-en.htm (accessed 3 May 2024).

44.

Y-C

De Vries

Wiesel

, et al. (2014) The role of consumer confidence in creating customer loyalty. Journal of Service Research 17(3): 339–354.

45.

Reichheld

(2003) The one number you need to grow. Harvard Business Review 81(12): 46–54.

46.

Rigdon

Becker

J-M

Rai

, et al. (2014) Conflating antecedents and formative indicators: A comment on aguirre-urreta and marakas. Information Systems Research 25(4): 780–784.

47.

Roos

MWM

(2008) Willingness to consume and ability to consume. Journal of Economic Behavior & Organization 66(2): 387–402.

48.

Shackle GLS (1952) Expectation in Economics. Cambridge: Cambridge University Press.

49.

Shankar

Maraty

(2009) Concerns of India’s farmers. Outlook on Agriculture 38(1): 96–100.

50.

Skikiewicz

Błoński

(2017) Economic sentiment level versus the quality of life in European union member states. Prague Economic Papers 27: 1–18.

51.

Smink

(2020) Por qué hay multinacionales que abandonan Argentina y se habla de una ‘masacre de empresas’ peor que en la crisis de 2001. BBC News Mundo, 3 August. Available at: https://www.bbc.com/mundo/noticias-america-latina-53618912 (accessed 2 May 2024).

52.

Soroka

Stecula

Wlezien

(2015) It’s (change in) the (future) economy, stupid: Economic indicators, the Media, and public opinion. American Journal of Political Science 59(2): 457–474.

53.

Starr

(2012) Consumption, sentiment, and economic news. Economic Inquiry 50(4): 1097–1111.

54.

Stiawan

Arisandy

(2016) The effect level of regional security toward level of local investment by mediation consumer confidence Index. In: The First International Conference on Law, Economics and Education Muhammadiyah. Indonesia: University of Metro, 17–22. Available at: https://repository.ummetro.ac.id/files/artikel/Evan%20Stiawan%20,%20Yosy%20Arisandy.pdf.

55.

Tversky

Kahneman

(1986) Rational choice and the framing of decisions. The Journal of Business 59(4): S251–S278.

56.

Vera-Gilces

Silva

, et al. (2020) Índice de confianza agrícola: Un enfoque al sector agroexportador en el Ecuador. X-pedientes económicos 4(10): 57–62.

57.

Vuchelen

(2004) Consumer sentiment and macroeconomic forecasts. Journal of Economic Psychology 25: 493–506.

58.

Ward

Minja

Blackie

, et al. (2007) Beyond participation – building farmer confidence: Experience from Sub-Saharan Africa. Outlook on Agriculture 36(4): 259–266.

59.

Wooldridge

(2010) Econometric Analysis of Cross Section and Panel Data. Cambridge: The MIT Press.

60.

World Bank (2024) Argentina Overview: Development news, research, data. Report, World Bank. Available at: https://www.worldbank.org/en/country/argentina/overview (accessed 2 March 2024).

61.

Yao

Partington

Stevenson

(2013) Predicting the directional change in consumer sentiment. Australian Journal of Management 38(1): 67–80.

62.

Zafeiriou

Katrakilidis

Pegiou

(2019) Consumer confidence on heating oil prices: An empirical study of their relationship for European union in a nonlinear framework. European Research Studies Journal XXII(1): 63–90.