Should I buy domestic food? What influences Hungarian customers in this matter? Open Access

The concept of “ethical consumption”, which used to mean environmentally friendly behaviour, is now being used in more and more areas of life and in more and more different ways (e.g. human rights, animal welfare, country of origin, fair trade, anti-globalisation, less meat, zero waste, etc.). Ethical consumers make informed choices when shopping, based on their personal and moral values, and seek to influence corporate practices by buying or not buying a vendor’s products (Auger and Devinney, 2007). Food consumption is a frequently studied area in the literature on ethical consumption (Dowd and Burke, 2013) driven by several reasons. In addition to the fact that food systems must face more and more challenges, as they must satisfy the needs of the growing population with increasingly limited resources (e.g. land, water, transport), an important reason is that global food systems are significant sources of greenhouse gases (Striebig et al., 2019; Tandon, 2022). The GHG emissions of food-related global freight transport are almost twice the amount of greenhouse gases released during production (Li et al., 2022).

The supply of food via shorter transport distances is therefore an important endeavour, which drives the attention towards locally grown and processed foods. Focusing on local food consumption, in addition to its environmental effects, it is also worth mentioning its food safety and health aspects. Consumption of domestic food is encouraged by the freshness, seasonality, traceability and controllability of food produced close to the consumer compared to products that pass through other countries. In addition to its effects on the environment, food safety and healthy diet, it should not be forgotten that increasing the popularity of domestic food is also an economic interest. The conscious purchase of domestic products strengthens the competitiveness of domestic producers and processors.

Considering these arguments, a broadened understanding is needed of factors that encourage buyers to choose food from domestic sources. From this perspective, this paper seeks to deepen the knowledge of the behaviour of the “local food purchase” by exploring the factors that affect this purchasing intention and practice regarding Hungarian consumers.

In addition to arguments justifying a deeper understanding of consumers’ motivations for local food consumption, there are several motivations – derived from the characteristics of the country – for investigating the topic in the Hungarian food market. One of these is that Hungary’s agricultural characteristics are above average, both in terms of topography and climate (Fehér, 2019). The gross added value of Hungarian agriculture is significant compared to EU member states (Figure 1; KSH, 2020).

In addition to the country’s characteristics, it is also an important aspect that, according to Hungarian Central Statistical Office data, a quarter (24%–26%) of our annual expenditure per capita is spent on food and non-alcoholic beverages (KSH Stadat, 2022), which means that where and in what proportion this money is spent is an important question. Buying domestic products makes domestic producers and processors more competitive.

Based on these, this study addresses the following research queries:

To answer the queries, a quantitative online survey was conducted with 700 respondents representing Hungarian customers over 18 years old to investigate local food purchasing intention and behaviour of Hungarian buyers. Partial least square-based structural equation modelling (PLS-SEM) was used to assess the effects of different factors.

Identifying factors affecting consumers’ local food purchasing intention is essential for marketers to develop appropriate marketing strategies, positioning and promotion, and for policymakers to support domestic businesses, environmentally conscious consumption and food safety by encouraging local food consumption.

Section 2 provides a brief literature review of the predictor variables that can influence consumers’ intention and behaviour to buy domestic food; in Section 3, the research methodology is described; results are presented in Section 4; and implications are discussed in Section 5.

The focus of the research is the identification of the factors that can affect the investigated behavioural intention (BI) and the behaviour itself. One of the most popular and widely used social psychological models for predicting human behaviour, which seeks to predict and explain the occurrence of a given behaviour in a specific context, is the theory of planned behaviour (TPB) by Ajzen (1991). The TPB posits that behaviour is best predicted by behavioural intention, i.e. someone is more likely to perform a certain behaviour if they have already formulated their intention in advance. According to the TPB, BI is driven by three factors: favourable or unfavourable attitude towards the behaviour; subjective norms arising from the perceived social pressure of significant others; and perceived behavioural control (PBC; self-efficacy), which refers to the perceived control over the implementation of the behaviour (Ajzen, 1991). In the model, attitudes and subjective norms influence behaviour indirectly through BI, while PBC can have an indirect effect through intention and a direct effect on behaviour.

Studies using the TPB model – from different research areas (health, communication, tourism, etc.) – consistently support the predictive power of the model (Berkes, 2015; Dowd and Burke, 2013; Kirmani et al., 2023; Kumar, 2021; Ogiemwonyi and Jan, 2023). However, these studies also explore the potential for extending the original TPB model and examine the effect of additive factors specific to the area under study on the predictability of behaviour. It means that, in addition to the simplicity and efficacy of the TPB model – proven in many studies and in a wide context – the extension of the TPB framework provides an opportunity to customise the model to the specific research context.

In the context of this study, which focuses on predicting local food purchase intention and behaviour, three additional variables were found to be relevant, which were thought worth customising in the original model. These are the ethical or moral obligations (MOs), environmental concerns and collectivism, assuming that these variables may also play an important role in the development of purchasing preferences for domestic products.

“Moral reasoning is defined as an intra- and interpersonal psychological phenomenon that is important in individual and collective moral judgement and behaviours” (Jensen, 2020, p. 222). Based on Auger and Devinney (2007), the ethical consumers are driven by their personal and moral values. The term “ethical consumption” itself suggests that moral considerations play a decisive role in sustainable consumption behaviour. In the case of domestic food consumption, moral considerations cover several areas because this behaviour affects the local economy in addition to the natural environment.

Despite not being included in the original TPB model, the influence of moral or ethical considerations on sustainable behaviour has been confirmed by several studies (Dowd and Burke, 2013; Manstead, 2000), and later, Ajzen himself also found that the consideration of moral aspects in certain contexts significantly increases the predictive ability of the model (Beck and Ajzen, 1991).

The factor of moral or ethical considerations is represented in the models in two ways. First, as a positive sense of self-reward, such as Arvola et al. (2008) or Kirmani et al. (2022), as a positive feeling from following moral principles in the context of organic food choice. On the other hand, negative feelings of guilt or moral anxiety when moral values are violated, such as McEachern et al.’s (2007) research on the impact of animal welfare concerns on meat consumption, or Sharma et al. (2021), which investigated the association between the use of takeaway food delivery apps and guilt over food wastage. In this study, in accordance with the study of Kirmani et al. (2022), positive sense was used by measuring the level of agreement with statements that formulated morally correct behaviour.

Another potential predictor in the context of local food consumption as a manifestation of sustainable behaviour is the concern about the natural environment. In recent decades, there have been increased social and political concerns about environmental problems because of the alarming signs of climate change, such as extreme temperatures, and more frequent extreme weather events, such as heat stress, drought stress and floods. The visible and obvious consequences of growing environmental pressure and climate change call for the immediate and widespread implementation of sustainable behaviour. Such behaviour is future-oriented and not directly benefiting the person performing the behaviour; unlikely to deliver instant personal gain or gratification, it is likely that underlying concepts related to people’s beliefs about their ability to influence future outcomes and their desire to provide benefits to others may influence environmental protection (Kim and Choi, 2005; McCarty and Shrum, 2001). In addition to meeting current needs, pro-environmental consumer behaviour seeks to minimise environmental, social and economic consequences to achieve justice within and between generations (Hosta and Zabkar, 2021).

Concerns about the natural environment also appear in many studies examining the topic of sustainable consumption as an extension of the TPB model (Abdelradi, 2018; Konuk, 2019; Liu et al., 2022; Sadiq et al., 2021; Weber et al., 2020). The results of these studies confirm the hypothesis that our concerns for the environment have a significant influence on our sustainable consumption decisions. In other words, individuals who are highly concerned about the environment are more likely to engage in environmentally friendly activities and make sustainable consumption choices (Nagarajan et al., 2022).

Collectivism was the third predictor factor that the research investigated in addition to the original TPB factors. Collectivism is a belief that focuses on the well-being of society, thus attracting the attention of researchers studying this sustainable individual practice. Researchers have hypothesised – and their findings support this – that collectivist beliefs encourage people to engage in sustainable consumer behaviour (de Morais et al., 2021; Khan and Kirmani, 2018; Sreen et al., 2018).

The significant relationship between collectivism and environmental concern is supported by several research studies. The study of Arısal and Atalar (2016) concluded significant relations among collectivism, environmental concern and ecological purchase intention. Collectivists were more concerned with environmental issues, and environmental concern influenced individuals’ environmental purchase intentions. Naiman et al. (2023) found that collectivism was a significant positive predictor of environmental attitudes and intention to behave environmentally friendly. Based on this, H7c was formulated.

On the other hand, behavioural research proves that there is a connection between collectivism and the subjective norm. Park (2000) and similarly Trongmateerut and Sweeney (2013) found that the subjective norm scores of members of collectivist cultures are usually higher. The study of Al Zubaidi (2020) found that collectivism increases green product purchase intention, and subjective norms play a mediating role in this relationship. H7b was formulated based on these results.

MOs, environmental concerns and collectivism are individual conditions that can be strengthened from early childhood through education, values and norms; thus, the results of the study cannot only help to influence current customers in the direction of sustainable behaviour but can also provide useful information from public education institutions to higher education to increase the commitment of the younger generations to a sustainable environment.

The aim of the study was to reveal Hungarians’ preference for domestic products when purchasing food according to the planned behaviour theory model by Ajzen (1991) extended with collectivism, environmental concern and MO attitudes. The proposed model of the study is shown in Figure 2.

Based on the theoretical background, the first 4 hypotheses of this study were formulated according to the original TPB model in relation to Hungarians’ domestic food purchasing behaviour:

The study sought to outline and test a framework for understanding the role of psychological factors such as collectivism, environmental concerns and MOs on the intention of the consumers to buy domestic foods. H5–H7 were formulated based on the TPB model extension:

In addition to direct effects (for H1–H7), this study also examined mediating effects (for H4a and H7c).

An online survey was conducted with non-probability quota sampling (Table 1). The target group of this study was the customers over the age of 18 who do their own grocery shopping. In the database of the Central Statistical Office, the age group breakdown included the population size of the age groups under 14 years, 15–64 years and over 65 years. Based on these age group categories, the Hungarian population over 15 was considered as the target population size, with 8,206,689 people according to the data in 2023. Besides 99% confidence level and 5% margin of error, the needed sample size is 666 respondents. For simplicity, the targeted sample size was 700 respondents. Quotas were determined by adjusting the ratio of gender, age groups, education, regions and settlement sizes within the sample to the ratio of the entire Hungarian population.

The questionnaire was available between 2023 November 13 and 20 to the panel members of a Hungarian market research company. During this week, the data-cleaned responses with a composition and number corresponding to the quotas were collected.

The research instrument started with demographic questions. The screening in the first step helped to compile the sample corresponding to the quota composition. Because the target group of the research includes people who purchase food products themselves, the next block of the questionnaire started with a filter question to exclude those who rarely or almost never purchase food themselves (someone else – e.g. a family member – does the grocery shopping).

After the screening and introductory questions, the items related to the TPB model followed. According to Figure 2, eight research constructs were intended to be measured, for which 28 items were compiled (Table 2). Because of the heterogeneous target population and relatively big sample size, measurements on a seven-point Likert scale were considered more appropriate than a five-point scale because a wider scale allows for a greater spread of answers.

After evaluating the items of each construct, the respondent could no longer return to the previous items to prevent the possible distortion effect arising from the later items, e.g. the actual behaviour – measured at the beginning of the model examination – before we called the respondents’ attention to the moral, environmental and collectivist aspects.

The analysis was conducted in two stages, consisting of the evaluation of the measurement model and the evaluation of the structural model according to Ringle et al. (2017).

The measurement model was evaluated by the structural equation modeling (SEM) of the SMART PLS 4 program package. PLS-SEM (partial least square-based SEM) estimates the parameters of a set of equations by combining principal components analysis and regression-based path analysis to test or confirm the theory.

Because PLS-SEM is non-parametric in nature, it does not require the data to be normally distributed. Highly robust as long as the missing values level is less than 5%. Metric and quasi-metric (ordinal) scaled variables are suitable as inputs, and statistical power is high even for small sample sizes, and larger sample sizes further increase the precision (i.e. consistency) of the PLS-SEM estimations (Hair et al., 2021).

The bootstrapping procedure of SMATR PLS 4 was used to evaluate the structural model, which tests and estimates the hypothesised relationships in the model.

The original TBP model is a reflective (or effect) measurement model, where causality flows from the latent construct to the indicator items. The three additional latent variables (moral or ethical obligations, collectivism, environmental concerns) also were assessed as reflective models. The reflective model evaluation process covers indicator reliability, internal consistency reliability, convergent validity and discriminant validity (Hair et al., 2021).

Based on reliability and validity analysis, two modifications were necessary:

1. One item from the three was excluded from the PBC construct because of the small (0.376) value of factor loading.

2. Discriminant validity tests showed that the MO and BI constructs are not empirically different. Discriminant validity measures the extent to which a construct is empirically distinct from other constructs in the structural model. In other words, discriminant validity coefficients (correlation between constructs) should be noticeably smaller in magnitude than convergent validity coefficients (correlation within constructs). Therefore, removal of the MO construct from the model was decided, and thus H5 was not investigated further, but it is a thought-provoking result that such a strong positive correlation was shown among the items of MOs and items of BI. If we assume that the items formulated for each construct were appropriate (which is supported by the fact that the items come from studies tested by scholars), then it is suggested that positive ethical statements about the environment and local businesses are very closely related to the intention to buy local food.

The extended TPB model constructed from the retained items is shown in Figure 3. The PLS-SEM calculation model shows the path coefficients, the AVE values of constructs and the factor loadings of the items that build up the constructs.

Table 3 summarises the items of the proposed model and results of reliability and validity analysis. Indicator reliability – examining to what extent the variance of each indicator is explained by their construct – measured by factor loadings (FL) for all items, was higher than 0.708, which is the cut-off value of FL. Internal consistency reliability – shows the extent of correlation between multiple items measure the same construct – measured by Cronbach’s alpha (CA) ranged between 0.787 and 0.962, as well as composite reliability (CR), which had the lowest value of 0.903. These values meet the reliability criteria: the recommended value of both indicators is greater than 0.7. Convergent validity – the extent to which the construct explains the variance of its indicators – measured by average variance extracted (AVE) – shows 0.778 as the lowest value; the recommended value is higher than 0.5, which establishes the convergent validity of the model. The threshold values of indicators shown here are based on Hair et al. (2021).

The extent to which a construct empirically differs from other constructs in the structural model is measured by discriminant validity. Table 4 shows the discriminant validity measures of the structural model with the square root of the AVE on the diagonal – must be greater than the inter-construct correlation (Fornell–Larcker criterion) – below the diagonal, and above the diagonal the heterotrait–monotrait (HTMT) values are shown. A HTMT value smaller than 0.90 indicates the presence of established discriminant validity. The HTMT values in Table 4 (the highest is 0.873) confirm the validity of the model (Henseler et al., 2015; Hair et al., 2021).

The stage following the evaluation of the measurement model was the evaluation of the structural model, which tests and estimates the hypothesised relationships in the model. For this, the bootstrapping procedure of SMART PLS 4 was used, with settings of 5,000 subsamples, bias-correlated and accelerated (BCa) bootstrap as the confidence interval method and complete (slower) set of results.

SRMR is an approximate measure of model fit. The SRMR value of the structural model was 0.060, which is less than the cut-off value of 0.08, as proposed by Hu and Bentler (1999).

Table 5 includes the adjusted R-square values – explanatory power of the variables included in the conceptual model – and the conclusions based on these values.

Bootstrapping standard errors as a basis for calculating t-values of path coefficients or alternatively confidence intervals (a path coefficient is significant at the 5% level if the value zero does not fall into the 95% confidence interval) provides assessment for the significance and relevance of path coefficients (Hair et al., 2021). Table 6 shows direct positive and significant relationships are existing (based on both P values and confidence intervals) between each pair of the latent variables of the structural model.

Figure 4 shows the result of the bootstrap calculation. The inner model values are path coefficients, with t-values in brackets; within constructs, the adjusted R-squares are visible, and the outer models show t-statistics.

The bootstrapping procedure of SMART-PLS 4 enables the evaluation of the mediation effect in the structural model. The structural model in this study includes two mediation relations. H4a: BI mediates between PBC and B, and H7c: EC mediates between C and Att.

Table 7 shows that there is only a weak (10.49% and 14.49%) significant partial mediation effect in the case of both relations.

The study intended to test empirically what factors have significant effects on the Hungarian buyers’ domestic food purchasing intention and behaviour by use of the TPB framework extended by three individual conditions (MOs, environmental concerns and collectivism). Based on discriminant validity tests, the MO and BI constructs were not empirically different; therefore, the removal of the MO construct from the model was decided. Although our model was thus not suitable for testing H5, this result also provides information according to which the degree of respondents’ agreement with the items of the MO construct, which formulate positive ethical statements regarding pro-environmental behaviour and local businesses, is closely related to the items of the construct of intention to buy local food.

The factors affecting the intention and behaviour of Hungarian customers to buy local food were examined by extending the original TPB model with two individual constructs: environmental concern and collectivism. The results supported that all predictors remaining in the model after the reliability and validity check had significant positive effects on the local food purchasing intention. Examining the predictors of the original TPB model (attitude, subjective norms and PBC), an important finding is that the direct positive effect of PBC on behaviour itself is stronger than its direct effect on intention, which proves that in our model, PBC – which reflects the perceived ease or difficulty of local food purchasing – mainly directly influences the shopping behaviour of Hungarian customers.

Among the extended factors, collectivism had an outstanding predictive role, which exerted a strong, significant positive direct effect on environmental concerns (pc = 0.661** and t = 2.530) and a moderate significant positive direct effect on attitude (pc = 0.430** and t = 9.077) and subjective norms (pc = 0.497** and t = 15.135).

In addition to the fact that the purchase of domestic products strengthens the competitiveness of domestic producers and reduces the environmental impact associated with long-distance transport, in Hungary, encouraging the purchase of domestic food has a special role because Hungary’s characteristics are favourable from the aspects of agriculture. The research revealed how and to what extent the predictors included in the model encourage the Hungarian customers to buy domestic products and provides new insights into the understanding of Hungarians domestic food purchasing preferences. The elements of the proposed TPB model can help marketers and policymakers to find the effective way for the promotion of domestic food shopping, and thereby increase the Hungarian population’s commitment to domestic food consumption.

Because of the strong direct impact of PBC on the purchase itself, it is important to consider the possibilities that result in the perceived ease of purchasing domestic food. Perceived self-efficacy can be improved, e.g. with easy identification of domestic food products, competitive pricing and dissemination of the environmental and economic impact of their choices.

In addition to the constructions of the original TPB model, the study included the examination of the effect of factors that are individual conditions. These factors can be developed through formal and informal social influences, norms and expectations (family, school education). In addition to actions that can be implemented as immediate support for domestic food consumption, long-term desirable social guidelines have also become visible.

The outstanding predictive role of collectivism in the model of domestic food purchase intentions indicates that it is worth developing the responsibility felt for each other and the feeling of belonging to the community, e.g. with local, regional and national programmes and events, as well as to strengthen this feeling already in school education to achieve deeper commitment.

The strength of the study is that the sample was representative of the Hungarian food purchasing population over the age of 18. Hungary is considered as an individualist society with a score of 71 based on (Hofstede Insights, 2023). Compared to the EU member states, with this value, Hungary is in the middle (the highest score is 100 of the Netherlands, and the lowest score is 46 of Romania). The results of the study proved that collectivism plays an important predictive role in domestic food purchase intentions; future direction of research could be the comparison of the role of collectivism in the extended TPB model in countries with higher and lower individualism scores.

In this study the range of psychological factors included in the model was limited, which was further narrowed by the fact that the MO construct was not empirically different from the BI construct, so it was excluded from the model. In the future, on the one hand, it would be worthwhile to examine the effect of additional psychological predictors. On the other hand, it would be worthwhile to reformulate the items of the MOs construct to get an empirically different construct from the others in the model to make it possible to examine the role of MOs in the domestic food preference of Hungarian buyers.

In the present study, the demographic characteristics of the respondents were used to ensure the representativeness of the sample, and we focused exclusively on the effect of psychological predictors. In the future, in addition to psychological predictors, it would be worthwhile to examine the effects of some demographic parameters.

The research from which the study was carried out was supported by the Centres of Excellence of the Budapest Business University, connected to the BBU’s Centre of Excellence for Sustainability Impacts in Business and Society.

Statements and declarations: The authors wish to confirm that there are no competing interests, there have been no financial or non-financial interests that are directly or indirectly related to the work submitted for publication.