JSDEWES: Using Multivariate Statistical Analysis for Examining the Relationship between Food Waste Generation and Socio-economic Factors

Food waste is a complex issue influenced by various factors, including the perishable nature of food, unpredictable supply and demand dynamics, and limited control over production factors [1]. Additionally, social elements such as household composition, lifestyle choices and eating habits play a significant role. Both private and public stakeholders often prioritise other concerns, such as profit maximisation or regulatory compliance, which further exacerbates the problem of food waste. Previous research has shown that consumers' perceptions, behaviours, attitudes, beliefs, and values regarding food and waste are the primary factors driving food waste and loss [2]. The inefficient use of technology in the supply chain, organisational weaknesses, inadequate legislation and lack of awareness about proper consumption further worsen the problem [3].

In the near future, food production will face numerous challenges as the world's population continues to grow, and per capita intake of calories, protein and cereals is expected to increase further. According to the World Resources Institute, feeding 9–10 billion people by 2050 will require a 70% increase in food calories compared to 2006 levels. Mitigating food waste is a potential approach to achieving a sustainable food supply, but it alone cannot completely solve the problem. According to the United Nations Environment Program (UNEP), in 2016, a total of 931 Mt of food was wasted from households, retail establishments, and the food service industry, which suggests that 17% of total global food production may be wasted [4]. It is reported that 61%, nearly 570 Mt, of this waste comes from households, 26% from food service, and 13% from retail. The average global per capita food wastage stands at 74 kg per year, with minor differences between various income level countries (high-income countries: 79 kg/capita/y, upper-middle-income countries 76 kg/capita/y, lower-middle-income countries: 91 kg/capita/y). This indicates the necessity of improvement in most nations irrespective of their economic status [4].

The EU generates over 58 Mt of food waste annually [5], with a market value estimated at 132 billion EUR [6]. According to Eurostat, approximately 10% of food available to EU consumers may be wasted across retail, food services, and households. At the same time, more than 37 million people cannot afford a quality meal every second day [7]. The total food waste for EU-27 is 130 kg/capita in 2020, including 12.30 kg/capita (9.4%) for primary production, 26.85 kg/capita (20.6%) for manufacture of food products, 8.95 kg/capita (6.9%) for retail and other distribution, 11.86 kg/capita (9.1%) for restaurants and food services and 70.25 kg/capita (54%) for households [7].

Food waste has numerous environmental, economic and social consequences. The unnecessary wastage of resources used in food production, such as water and energy, places a significant burden on the environment. Food waste accounts for approximately 6% of total greenhouse gas emissions [8] . Economically, food waste represents a substantial loss for both producers and consumers. Socially, it deepens food insecurity and poverty in communities where it is a serious problem. Much of this wastage occurs at the point of consumption in the developed world [9]. These factors lead to the discarding of edible food that could have been consumed, resulting in unnecessary waste and its associated negative impacts [10].

The bioactive compounds present in food waste offer promising opportunities to convert this waste stream into value-added products across diverse fields [11], such as nutritional foods, bioplastics, biosurfactants, biofertilisers and single-cell proteins, which have identified food waste as a novel and promising source material [12]. Diverse food waste streams, such as fruit, vegetable and lipid-rich residues, can be utilised for the synthesis of various biopolymers, including polyhydroxyalkanoates, starch, cellulose, collagen and others [13]. Studies conducted in Malaysia have demonstrated the public's favourable reception of rice crackers produced from rice food waste [14]. Furthermore, the people's willingness to pay a specific amount to process their food waste suggests a high potential for the marketability of food waste recycling [14].

Repurposing food waste into energy also offers potential economic and environmental advantages within a circular economic framework. Food waste can be addressed through a range of treatment methods, encompassing thermochemical (incineration, pyrolysis, gasification, torrefaction and hydrothermal carbonisation), biological (anaerobic digestion, composting, aerobic fermentation, dark fermentation and photofermentation) and chemical processes (transesterification) [15]. The environmental impacts of the various methods were thoroughly examined through life cycle assessment studies. Based on this comparative analysis, anaerobic digestion emerges as one of the most effective conventional approaches, while esterification, gasification and hydrothermal carbonisation prove to be the superior thermochemical methods [16].

Reducing food waste is crucial not only for environmental sustainability, such as resource exploitation and greenhouse gas emissions [17], but also for economic and social reasons as well as addressing food insecurity [18]. Various methods can be used at the household level to reduce food waste, including meal planning and mindful shopping [19], proper food storage such as refrigeration or freezing, creative repurposing of leftovers into new meals, composting food scraps instead of discarding them, and donating excess food to local food banks or shelters [20]. Increasing the efficiency of food production and transport is also crucial in prevention efforts [21].

Community dining venues such as school canteens and university cafeterias can employ various strategies to mitigate food waste, including testing spoons, awareness campaigns, plate waste tracking, guest forecasting tools, reducing plate and utensil sites, offering menu options, adapting online pre-booking meal system, etc. [22]. According to research conducted in Swedish school canteens, awareness campaigns proved most effective in reducing plate waste (by 13 g per portion). In contrast, forecasting and plate waste tracking interventions were most successful in reducing serving waste (by 34 and 38 g/portion) [23].

The paper focuses on examining geographical variations in food waste quantities within EU countries and their possible connection with various environmental, social, and economic factors, including selected indicators of Sustainable Development Goals. Our study recognises that while prior research has examined the issue of food waste, the exploration of multivariate relationships and geographical differences across EU countries has been limited. Recent research lacks a comparison of food waste behaviour across countries and regions [24].

In our previous work [25], correlation coefficient calculation was used to find the relationship between food wastage and predefined environmental, economic, and social indicators. However, the results did not reveal any nontrivial connections between food waste and social-economic parameters. It is assumed that there should be some connections, but a more sophisticated method is needed to uncover them because the effect of unknown parameters and measurement error bias them. Two methods were used to find nontrivial multivariate connections between food waste and other parameters: the K-means clustering method and the Principal Component Analysis (PCA) method. In our previous work, the K-means clustering method was applied to find the groups (clusters) of countries with similar food waste characteristics and collect the social parameters for every group [25]. It could be proven by cluster analysis that the historical past has a significant role in food waste parameters, and several important differences between socio-ecological parameters exist. In this paper, the results of the examination based on the PCA method are introduced.

The novelty of this research lies in its introduction of a new methodology for uncovering the relationships between food waste and related parameters, as well as for conducting a more integrated, multidimensional analysis of food waste profiles across EU member states. These findings are not only relevant to the academic community but also hold practical significance for professionals and policymakers in the fields of circular economy and waste management, facilitating the development of more effective strategies.

The left diagram of Figure 1 shows the five food waste parameters of the 27 EU member states.

While there are noticeable variations in the quantity of various components, it is evident that different types of countries can be identified: some exhibit a high value of food waste from the manufacture of food products and beverages (MFP) but relatively low food waste from retail and other distribution of food (RFS). In contrast, others show values above the average for both parameters. Nevertheless, recognising patterns in this diagram proves to be challenging.

Figure 2 shows EU members' data in the P₁-P₂ plane after conducting PCA on the dataset. The origin (0, 0 point) is near Romania because this country's food waste parameters are the closest to the EU average. Three countries (Denmark, Belgium, and Cyprus) have significant positive P₁ values, which means that they are highly above the average in MFP (food waste from manufacturing of food production), FP (food waste from primary production), and RDF (food waste from distribution of food production), according to the s₁ base vector. Meanwhile, they have approximately zero P₂ values, which includes food waste from restaurants and other food services (RFS) as well as from households (HHA). Three more countries (Greece, Ireland, and Netherlands) have positive P₁, and all of them have moderate P₂ values. This reveals a nontrivial connection between food waste components: if a country's first PCA component is over the average, the second component is close to the average. On the other hand, if P₁ is negative, then P₂ can be a significant negative or positive value. This special connection between PCA components results in large empty areas in Figure 2 over and under the green dotted lines. Accordingly, the countries can be divided into three groups:

1. P₁>0: Cyprus, Belgium, Denmark, Netherlands, Ireland, and Greece. Their difference from the average food waste characteristics is mainly in s1 direction, which means high excess in food waste from the manufacture of food products (MFP), and significant excess in food waste from primary production (FP) and food waste from distribution of food production (RDF) components.

2. P₁<0 and P₂>0: Portugal, Italy, Malta, Luxembourg, Lithuania, Latvia, Austria, Germany. They have below-average values in FP, MFP, and RDF, but there is excess food waste from households (HHA) and food waste from restaurants and other food services (RFS) values.

3. P₁<0 and P₂<0: Romania, Czech Republic, Slovakia, Hungary, Sweden, Poland, Estonia, France, Croatia, Bulgaria, Finland, Slovenia, Spain. All of their food waste components are smaller than or equal to the average.

Uncovering the reasons behind these results is challenging, particularly because the correlation coefficients did not reveal a clear connection with the examined geographic, economic and social indicators.

Belgium, Denmark, Netherlands and Ireland are geographically close to each other and have similar geographical features such as seashores and weather conditions. These similarities may have a comparable impact on their agriculture systems, similar to the case of Greece and Cyprus. In the mentioned countries, the primary production, manufacturing and distribution of food generate more food waste than in the others. This raises intriguing but challenging questions that encourage further investigation. Economic factors associated with food waste on farms and throughout the supply chain encompass price fluctuations, workforce insurance costs, supply chain constraints, commercial and client performance, agreement incentives and other existing policies [1]. Avoiding the harvesting of fields due to low market demand or low prices is one of the primary causes of food waste [2].

Geographically and culturally, Portugal, Italy and Malta share similar features. It is reasonable to assume that eating out is more common in these countries. This could be linked to a higher P₂ value associated with a higher amount of food waste from restaurants and other food services (RFS). There are also cultural and/or historical similarities due to the geographical proximity between Austria-Germany-Luxembourg and Lithuania-Latvia. In the mentioned eight countries, the amount of food waste from households (HHA) and food waste from restaurants and other food services (RFS) are higher compared to other countries, indicating more wasteful eating and cooking habits. This is particularly noteworthy in the case of the three Mediterranean countries.

The history of the countries must have a significant effect on food waste parameters because most of the former Eastern Bloc countries are in the third group. This finding is in good accordance with the results of our former cluster analysis results [25].

The connection between the PCA components (P₁, P₂) and the selected indicators was investigated. At first, the linear correlation coefficient was calculated for every possible pair comprising a socio-economical parameter and a PCA component. The result is negative: there was no case with at least 0.7 correlation coefficient. The investigation was then extended to multiple dimensions by considering the hypothesis that a linear combination of a few socio-economic parameters is suitable for reconstructing PCA components. Again, the result is negative: with ten or fewer components, the maximum correlation factor did not reach 0.7.

After the negative results of linear regression, plot pairs for all 33 geographical, social and economic parameters and the two PCA components were generated and visually checked for any pattern. In a few of these plots, the members of the first group introduced above (P₁>0) form a special set. Figure 3 shows the mountain type area percentage (REG_MOU) case as an example. Here, 5 of the six members of the first group have almost zero mountain-percentage value, which means that no region has a dominant mountain-dwelling population.

Similar regularity is present in the IAF (Inability to afford a meal with meat, chicken, fish (or vegetarian equivalent) every second day) plot, Figure 4. The only exception is Greece, which has the smallest P₁ value in the first group. It can be established that the five countries with the largest P₁ value (Ireland, Netherlands, Denmark, Belgium, and Cyprus) have a below-average percentage of mountain areas and a small percentage of persons who cannot afford meat or fish every second day. Moreover, these 5 countries have near to zero P₂ values, which means the average amount of food waste from households (HHA) and the average amount of food waste from restaurants and other food services (RFS).

Despite the many uncertainties, the quantity of food waste, and especially its RFS (food waste from restaurants and other food services) and HHA (food waste from households) components that determine the P₂ component, can be related to cultural and historical events affecting the countries. According to a study on the determinants and country differences of food consumption in the EU during the last decade of the 20th century, food consumption in the European Union countries (countries that joined EU later than 2000 were not part of the study) can be summarised by four major trends [53]: a decrease in the proportion of expenditure allocated to food, already at very low levels; reaching maximum level in total food consumption; a shift in the structure of food consumption; and an increase in the proportion of food consumed outside home. This latter trend is common for all countries but varies widely in intensity based on labour circumstances and social aspects. Single-person households, younger individuals and families with young children may be more inclined to seek out various meal options outside the home [1], which can be related to the increased amount of food waste from restaurants and other services. Palatability, pricing, portion size, temporal limitations, dishware dimensions, suboptimal catering operations, fluctuating diner populations and caloric content can represent the prevailing originating factors of food waste in the case of restaurants and other food services [2].

According to previous studies, factors contributing to individual food waste include personal habits, interests, and perceptions regarding food appearance, quality, and the ability to consume large quantities and a diverse range of foods, regardless of geographic location or season [1] [2] [54]. Consumers' tendency to engage in excessive purchasing can be attributed largely to their impulsive buying behaviour and lack of mindfulness when shopping. Conversely, consumers' heightened disgust sensitivity may lead them to discard products beyond the recommended best-before date without verifying their edibility [55] [56]. The main causes of food waste in private households are food overprovisioning, improper storage,transportation challenges, overcooking, large portions, difficulty managing leftovers, lack of knowledge about assessing food edibility and packaging issues [2].

In a previous study [57], the analysis focused on whether Europeans (including only member states at that time) were homogeneous or heterogeneous in terms of food behaviour and attitudes. The findings indicated that despite the globalisation of food processes, Europe could not be conclusively seen as a uniform bloc with regard to food culture. The research suggests that national boundaries combined with language barriers remain strong indicators and best predictors of variations in food-related behaviours [58].

Research utilsing primary data from the United States, Canada, the United Kingdom and France examined and compared consumers' food waste behaviours. The findings indicate that age, dining outside the home, and using expiration dates and appearance to assess food spoilage are associated with increased frequency of food waste, and these trends are consistent across the studied countries. Furthermore, the researchers observed notable differences in food waste behaviour among the countries, suggesting that cultural factors play a significant role in shaping such practices [24].

Cultural values, acquired preferences, and ways of life all play a substantial role in shaping food consumption. The formation of habits influences attitudes and inclinations towards food products, which persist over time. Variations in dietary patterns across different countries can largely be attributed to historical factors such as local production and physical availability [59].

Governments can play a significant role by incorporating education campaigns to change consumer habits and preferences [60], implementing municipal composting programs, and providing tax credits to farmers who donate excess produce [61]. These actions help individuals and households reduce their food waste while contributing to the national target of reducing food waste by 50% by 2030, aimed at improving overall food security and conserving natural resources. Collective action is required to tackle this pressing global problem [62].

Findings from our present study and previous research [24] also suggest that researchers and policymakers should account for cultural factors when developing strategies and policies to mitigate food waste within their respective countries. It is crucial to recognise that an effective food waste reduction policy requires country-specific tailoring, as consumers exhibit varying food waste behaviours across different nations. The successful approach implemented in one country or region may not necessarily be applicable to others. Consequently, research and policies should focus on addressing the unique circumstances of each individual country or region to ensure contextual relevance and efficacy.

The current research aims to explore the relationship between indicators of food waste and various environmental, geographical, social, and economic factors, using multivariate statistical methods to reveal their interdependencies. Due to the complexity of this field and the limited data availability, previous works could not find any measurable connection.

Our investigation produced both positive and negative results. The analysis presented in this paper indicates that food waste indicators do not exhibit a linear association with socio-economic factors. Univariate linear analysis revealed no significant correlation between any of the examined five food waste parameters and the 33 socio-economic parameters. This finding aligns with preceding studies. The Lasso method with 2 to 10 variables was applied, but the current multivariate linear regression investigations did not find statistically significant correlations.

Principal Component Analysis demonstrated that the five parameters characterising food waste within EU member states could be reduced to two PCA components. These two principal components summarise key food waste characteristics and enable cross-country comparisons. Furthermore, this result may be significant for further research as dimensional reduction can aid general visualisation and understanding. The plane of these two components is not evenly covered by the countries, which indicates a hidden relationship behind these parameters. The recognised pattern was described in the article, but the exploration of the underlying reasons is the subject of further research.

The present study identified three distinct clusters in the PCA-based classification of countries. The first group, characterised by high food waste in the early stages of the supply chain, includes six countries where primary production, manufacturing, and distribution waste is predominant. In contrast, the second group (8 countries) experiences lower waste generation in these domains but more significant waste generation in restaurants and households. The third cluster, comprising 13 countries, can be characterised by reduced food waste across the studied sectors. The analysis determined that the identified patterns are shaped by geographic, cultural, and historical factors rather than by direct socio-economic influences. The multivariate linear correlation between the PCA factors and socio-economic parameters remains weak, but it was possible to identify a few nonlinear regularities. For instance, five of the six countries in the first group above are characterised by the population settled on flatland and an above-average supply of meat or fish. Another pattern observed is that former Eastern Bloc countries belong to the third group. Further research is necessary to investigate the factors driving these trends.

The obtained results provide important information for decision-makers, environmental specialists, and practitioners in the circular economy and waste management fields. Understanding the country-level patterns may contribute to developing more effective food waste mitigation strategies tailored to regional characteristics.