Food loss and waste in community-supported agriculture in the region of Leipzig, Germany

ABSTRACT

Community-supported agriculture (CSA) has the potential to reduce food loss and waste (FLW) due to a direct connection between consumers and producers, lower standards regarding product aesthetics and consumer-related values. Accordingly, CSA could compensate for possible yield gaps compared to conventional agriculture. Here, FLW and yields were assessed for one year in four CSA initiatives in the area of Leipzig, Germany, and compared to reference data. The results show that FLW during production, distribution and consumption were on average 22–70% lower compared to reference data. On average, due to lower FLW, the CSA initiatives could compensate for yield gaps of 32%, which reflects yield gaps typically observed between organic and conventional agriculture. Actual yield gaps in CSA might be lower, as agroecological approaches to increase productivity are widely adopted. The co-evaluation of the results with CSA gardeners showed that FLW could be further reduced in all stages. Reducing FLW was mentioned as an important sustainability goal of the investigated CSA initiatives, yet they also indicated a focus on biodiversity, education, food quality or the reduction of fossil fuels. This pilot study indicates that CSA can contribute to the transition towards sustainable and resilient food systems.

KEYWORDS:

• Alternative food systems
• community-supported agriculture
• direct marketing
• food loss and waste
• food supply chain
• food system transformation
• resource use efficiency

1. Introduction

Given that the current industrial food system is a major driver of climate change and social inequalities, a transformation towards sustainable, resilient and productive food systems is urgently needed (Haack et al., 2020; Pigford et al., 2018). In the last decades, food production and productivity have been largely increased at the expense of the environment due to the increase of agricultural inputs such as herbicides, fertilizers (including the unsustainable extraction of phosphate) and irrigation, as well as the homogenization of landscapes (fewer, larger fields) (G. A. Baker, Gray, et al., 2019; Liu et al., 2020; Ray et al., 2013; Sánchez-Teba et al., 2021; Seppelt et al., 2020; Tilman et al., 2011; Yuan et al., 2018). Around 26% of global greenhouse gas (GHG) emissions can be attributed to food production (Poore & Nemecek, 2018). Alternative approaches, e.g. organic farming or conservation tillage, which also aim at preserving or improving environmental aspects (e.g. biodiversity or water quality), have recently gained importance (Eriksson et al., 2020; Lipinski et al., 2013; Lülfs-Baden et al., 2020). However, it has been criticized that organic farming might not be able to feed the world due to lower yields (Connor, 2008, 2013; Ponisio et al., 2015; Ponti et al., 2012; Schrama et al., 2018; Seufert et al., 2012; Seufert & Ramankutty, 2017).

Besides agroecological approaches, which have the potential to reduce the yield gap between conventional and organic farming directly (Kremen & Miles, 2012; Ponisio et al., 2015), the reduction of food loss and waste (FLW) provides opportunities to reduce demand for agricultural products and thus to minimise additional area demand due to lower yields, ultimately increasing the sustainability of food systems (Muller et al., 2017; Ogunmoroti et al., 2022; United Nations Environment Programme, 2021). FLW at the production stage mainly occur due to quality sorting to comply with size and aesthetic standards, gaps in infrastructure, overproduction to ensure that supplier demands can be met, and pricing by supermarkets (Göbel et al., 2015; Gustavsson et al., 2011; Porter et al., 2016; Soma et al., 2021). During distribution (i.e. at wholesale or retail markets) and consumption (i.e. at the household level), FLW is related to excessively high and inflexible standards (e.g. regarding expiration dates), overstocking, the removal of ‘imperfect’-looking foods, insufficient planning and communication, oversized portions, and poor storage management at home (FAO, 2019; Lemaire & Limbourg, 2019; Lipinski et al., 2013; Themen, 2014; Yu & Jaenicke, 2021). Across the entire food supply chain, long travel distances are another important source of FLW (Chauhan et al., 2021; Mena et al., 2011).

Community-supported agriculture (CSA) is increasingly recognized for addressing some of the challenges described above. CSA originated in Japan in 1965–1971 (Henderson & van En, 2007; Tang et al., 2019) and was established to reduce pesticide use and strengthen local agriculture practices (Kondoh, 2015). The solidary and participatory structure of CSA initiatives is intended to create greater proximity between food and consumers (Kreutzberger, 2017; Kropp & Stinner, 2018), where costs and risks, as well as the harvest are shared (Falk & Madsen, 2015; Henderson & van En, 2007). CSA initiatives mainly focus on promoting local economies and reducing negative environmental impacts instead of generating profits (Bloemmen et al., 2015; Wellner, 2017). Food is not distributed via the market, but directly to consumers. The farms are thus more independent from market prices and subsidies, which also enables more need-based and flexible working conditions (Netzwerk Solidarische Landwirtschaft e.V., 2021). Due to the short supply chains, lower standards regarding product aesthetics and consumer-related values, CSA initiatives have the potential to reduce FLW (Göbel et al., 2012; Haack et al., 2020; Priefer et al., 2016). While scholarly literature on social aspects, including heterogeneity, behaviour and health of CSA members is growing, quantitative studies on ecological aspects including productivity and resource efficiency in comparison to other farming systems remain scarce (Egli et al., 2023).

To our knowledge, the potential of CSA to reduce FLW and thereby compensate for possible yield gaps has only been estimated in one case study (N. Baker, Popay, et al., 2019). In the present study, this gap is addressed by (i) investigating FLW during production, distribution and consumption in CSA initiatives in comparison to reference systems, (ii) quantifying the potential to compensate for potentially lower yields due to lower FLW and (iii) analyzing strategies to deal with FLW and other sustainability aspects. Therefore, a mixed-methods approach was used, including intensive field work in four CSA initiatives, data synthesis and quantification of FLW and yields, as well as a focus group discussion with CSA gardeners to put the quantitative results in a broader perspective. The findings will foster a more comprehensive understanding of the sustainability potential of CSA in the context of FLW, agricultural yields and resource use efficiency.

2. Materials and methods

2.1. Methodological summary

In four CSA initiatives in the region of Leipzig, Germany, intensive field work following a standardized protocol was conducted to collect data on food loss and waste (FLW) during production, distribution and consumption, as well as on yields for major crops for one year. This data was used to calculate the average FLW in relation to supply and then compared to reference data to evaluate, whether FLW are comparatively lower in CSA. Furthermore, it was analysed to which extent lower FLW can compensate for possible yield gaps typically observed in organic agriculture. Intermediate results were then presented in a group discussion with CSA gardeners for a broader evaluation of the findings, practical implications, as well as the identification of other sustainability aspects.

2.2. Definitions

2.2.1. Food loss and waste

Despite large national and international efforts, no standardized definition of food loss and waste exists (Spang et al., 2019). According to one of the most commonly used definitions, food losses occur from harvest up to, but not including, the retail level, while waste occurs at the retail and consumption levels (FAO, 2019). In the context of this study, the term food loss and waste (FLW) was used to describe all different types of food loss and waste, regardless of their origin or their further usage. Furthermore, (vegetable) FLW was defined as discarded or unused plants or parts of plants that were originally intended for human consumption. This included both their edible and inedible parts (e.g. core, stalk), because they could not be distinguished at the farm level and during distribution, and unharvested plants and losses that occurred during harvesting. This also included all different sizes and shapes of vegetables that were left on the field or at the storage after harvesting or delivering, respectively, at the distribution point after the collection period was over and that were wasted at the household stage.

2.2.2. Food supply chain

In this study, the food supply chain (FSC) included all processes between production and consumption. To enable comparability between the CSA FSC and reference data, this study included the three stages: production, distribution and consumption. The production stage in both cases included on-farm processes until the food was harvested. In the case of the CSA initiatives and in some reference studies this included on-farm storage processes. The distribution stage in the CSA FSC only consisted of the distribution points where food is delivered and picked up directly by the consumers. The reference distribution stage instead included different substages between production and consumption, including grading, storage, packing, distribution and retail. The consumption stage included the handling of food products at the household level, such as (failed) conservation, preparation, cooking and eating.

2.3. Site description

This study was conducted in four CSA initiatives located in the region of Leipzig, Germany (Table 1). The main characteristics of the investigated CSA initiatives were derived based on a standardized survey (Table S1). The CSA initiatives were managed according to self-defined ecological principles but without organic certification. Despite different sizes, all farms cultivated at least 40 crops, showing the high diversity and small-scale cultivation approach particularly of the two smaller CSA initiatives. Average price per share ranged between 80 and 95.5 €. However, all four initiatives conducted a yearly bidding round, where members could indicate how much they are willing to pay. Therefore, actual prices per share ranged between 40 and 150 €. While the two smaller initiatives almost exclusively relied on handwork, the bigger ones also used machinery.

Table 1. Characteristics of the investigated CSA initiatives.

Characteristics	CSA 1	CSA 2	CSA 3	CSA 4
Location	Taucha, bordering district of Leipzig	Leipzig	Taucha, bordering district of Leipzig	Taucha, bordering district of Leipzig
Size (ha)	4.5	0.35	0.8	9
Employees/farmers (no.)	6	2	2	6
Tunnel cultivation (m^2)	280	37	0	500
Products	Vegetables, herbs	Vegetables, herbs	Vegetables, herbs	Vegetables, herbs, fruits
Cultivated crops (no.)	50	40	40	50
Vegetable shares (no.)^a	105	34	70	237
Average price per share	86 €	80 €	95,5 €	93 €
Management practices	Mainly Handwork, partially machinery use	Handwork	Handwork	Handwork, machinery use
Type of distribution	6 distribution points in the Leipzig region.	1 distribution point at the farm gate.	2 distribution points in the Leipzig region.	13 distribution points in the Leipzig region.

^a Typically, the amount of one share is shared between two to three people.

2.4. Crop selection

Due to the large number of crops (40–50), some were grouped into main categories (Tables S2 and S3). Only crops that were cultivated on a minimum of three CSA initiatives were included. To reduce complexity, only the most important crops regarding the harvested area were considered. Therefore, the specific crop land areas across all CSA initiatives were calculated and the crops which on average made up two-thirds of the total harvested area were selected. Open-field cultivation and cultivation in tunnels were considered (Table S2). This resulted in 16 crops (beetroots, broccoli, bush beans, carrots, courgettes, fennel, leeks, lettuces, onions, potatoes, pumpkins, savoy cabbages, spinach beets, tomatoes, turnips, white cabbages).

2.5. Data collection

2.5.1. Harvest data

Harvest data for this study was collected from 8 April 2021 until 31 March 2022. The harvested amount (kg) per crop was recorded once a week per CSA on the weekly harvest day. The type and amount of these crops were recorded by the CSA initiatives themselves per crop in kg or pieces. For all crops that were reported in pieces, conversion factors were collected by determining the average weight per piece by weighing 3–10 randomly picked pieces per crop after these were harvested. The weight was measured by using digital scales (±1 g accuracy). The weighed pieces complied with the type (=product definition) of vegetable that was delivered. For example, if a carrot was delivered with its leaves, these were also weighed. In some cases, when quantities could not be recorded due to unavailability of human resources, the FLW was determined by extrapolating the data from the previous and subsequent week for the same crop. The total harvested areas (m²) of the investigated crops were provided by the CSA initiatives.

2.5.2. Production FLW data

Data on production FLW was collected directly on the field from 8 April 2021 until 31 March 2022. On the harvest day of the corresponding CSA, losses per crop were measured in the respective harvest areas of the day. For this purpose, the vegetables left on the field were collected and weighed using digital scales after harvesting was finished. The collected and weighed FLW corresponded to the delivered products on that specific day.

If possible, FLW was measured on the entire harvested area of the corresponding day. However, if this was not possible due to time constraints or large vegetable fields, the survey was carried out on randomly selected plots of 1 m length, encompassing the entire width of a vegetable plot. Due to the small-scale cultivation of the investigated farms with typical field widths of 0.8–1.5 m, it was not possible to select plots without potential edge effects. Depending on the size of the field and the workload of the specific day, the subplots amounted for 10–25% of the total harvested area of the specific crop and day. The quantities collected were then extrapolated to the total harvested area of the day.

If a crop was harvested constantly over a longer period (several weeks or months) on the same field, only products that were initially picked by the gardeners and then left on the field was collected during the ongoing harvest period. After the last harvest, the biomass that has not been picked by the gardeners and therefore remained on the plant (e.g. beans) or in the soil (e.g. carrots) was additionally recorded. If crops were harvested and stored afterwards, FLW which occurred during the storing process was put aside by the gardeners and measured at the following harvest day.

The definition of FLW in this study integrated vegetables that were grown and intended for human consumption but turned inedible or were not delivered because of flowering. Because flowering changes the type and weight of the produce, the FLW weight for flowering crops was measured as follows. Whenever crops that were delivered as pieces (such as fennel) flowered, the number of flowering plants was recorded and converted into kg with the measured conversion factors (see section 2.5.1). For all other crops where flowering occurred, FLW in nonflowering areas was extrapolated to the entire crop-specific harvest area based on the recorded FLW after the last harvest in nonflowering areas.

In some cases, when quantities could not be recorded due to unavailability of human resources, the FLW was determined by extrapolating the data from the previous and subsequent week for the same crop.

2.5.3. Distribution FLW data

FLW data at the distribution points of the CSA initiatives was collected from 8 April 2021 until 31 March 2022. To investigate the FLW in the distribution points, the amount of vegetables, which was left after the collection period was over (either on the same or subsequent day), was weighed weekly or every second week. In total, data from eight distribution points in the four investigated CSA initiatives was collected. Depending on the CSA, 23% (CSA 4), 50% (CSA 1 and 3) and 100% (CSA 2) of the distribution points were investigated. The surveys were conducted either by members of the research group or by trained CSA members. For weighing, digital scales were used. The delivered quantities of the respective distribution points were provided by the CSA initiatives. If the total loss quantity on a specific day for a specific crop was greater than the recorded delivery quantity on that day, these data points were removed due to errors in the data documentation.

2.5.4. Household FLW data

FLW data on the consumption side was collected in the period between mid-May and mid-October 2021 by conducting 7- or 14-day diary studies with CSA members. Both CSA and non-CSA products were documented to increase the sample size and comparability. In total, primary use and FLW data from 47 households were included in the study, also including members from seven other CSA initiatives to increase sample size. To calculate FLW ratios, the weight of utilized products was recorded within the diary additionally to the amount of vegetables that were wasted. By requesting the members to use a kitchen scale, precise data (in g) was obtained. For this study, all fresh vegetables (or parts thereof) that were disposed were considered FLW, while utilization of products was defined as vegetables that were consumed directly, processed at home and eaten immediately or stored for later consumption.

2.5.5. Reference data

FLW data of the reference FSC was provided through literature research in peer-reviewed literature following a snowballing approach. Studies were included if they (i) contained detailed FLW data for one or more of the three FSC stages (production, distribution, household) or for their substages, (ii) investigated single food categories or crops (a minimum of three) that were in line with the investigated crops of this study, (iii) investigated the amount of FLW as percentage of the volume entering each stage and (iv) investigated high-income countries (preferring Europe). The goal of this approach was to provide reference data with a minimum of four studies per FSC stage. This yielded nine studies in total (Table 2).

Table 2. Overview of the studies providing FLW data for the different stages of the reference FSC. Only crops that were included in this study are listed.

Reference	Stage(s)	Region	Investigated crops or food categories
Beausang et al., 2017	Production	Scotland	Broccoli; carrots; lettuces
G. A. Baker, Gray, et al., 2019	Production	California, USA	Broccoli; bush beans; lettuces; tomatoes; white cabbage
Parfitt et al., 2021	Production	High- /middle-income countries	Fruits and vegetables; roots, tubers and oil crops
Terry et al., 2011	Production, distribution	UK	Broccoli; lettuces; onion; potato; tomato
Gustavsson et al., 2011	Production, distribution, household	Europe	Fruits and vegetables; roots and tubers
Beretta et al., 2013	Production, distribution, household	Switzerland	Potatoes; fresh vegetables, storable vegetables
Themen, 2014	Production, distribution, household	High-income countries	Fruits and vegetables; roots and tubers
WRAP, 2009	Household	UK	Pumpkins; lettuces; potatoes
van Dooren et al., 2019	Household	Netherlands	Vegetables; potatoes

When only food categories were reported, category-specific values were used for all crops belonging to this category, following the categorization of the underlying study. If single (but not all) crops were reported, average values of these were used for the missing crops.

2.6. Data analysis

2.6.1. Loss ratios

Regarding CSA data, loss ratio for each crop c, in stage s and CSA f was calculated as the total losses over the entire study period relative to the supply (equation 1). (1) $\mathrm{L}\mathrm{o}\mathrm{s}\mathrm{s}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{i}{\mathrm{o}}_{f,s,c}={\displaystyle \frac{\mathrm{L}\mathrm{o}\mathrm{s}{\mathrm{s}}_{f,s,c}}{\mathrm{S}\mathrm{u}\mathrm{p}\mathrm{p}\mathrm{l}{\mathrm{y}}_{f,s,c}}}$(1) In the production stage, the amount of supply included harvest and loss, i.e. the entire biomass of the target product. During distribution, the amount of supply corresponded to the delivery. In the household stage, the amount of supply included the loss and the biomass consumed. Due to the low sample size, household data was analysed for all households and not for the specific CSA initiatives.

In the next step, mean ratios across CSA initiatives and crops were calculated. Regarding reference data the mean ratios over all reference studies and crops were calculated.

2.6.2. Yields

Yields for CSA f and crop c were calculated based on the crop-specific deliveries divided by the respective harvested area (equation 2). (2) $\mathrm{Y}\mathrm{i}\mathrm{e}\mathrm{l}{\mathrm{d}}_{f,c}={\displaystyle \frac{\mathrm{D}\mathrm{e}\mathrm{l}\mathrm{i}\mathrm{v}\mathrm{e}\mathrm{r}{\mathrm{y}}_{f,c}}{\mathrm{H}\mathrm{a}\mathrm{r}\mathrm{v}\mathrm{e}\mathrm{s}\mathrm{t}\mathrm{e}\mathrm{d}\mathrm{a}\mathrm{r}\mathrm{e}{\mathrm{a}}_{f,c}}}$(2) In a next step, yield was combined with the loss ratios during distribution stage d and household stage h to calculate net yield for CSA f and crop c (equation 3). (3) $\mathrm{N}\mathrm{e}\mathrm{t}\mathrm{y}\mathrm{i}\mathrm{e}\mathrm{l}{\mathrm{d}}_{f,c}=\mathrm{Y}\mathrm{i}\mathrm{e}\mathrm{l}{\mathrm{d}}_{f,c}\times \left(1-\mathrm{L}\mathrm{o}\mathrm{s}\mathrm{s}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{i}{\mathrm{o}}_{f,d,c}\right)\times \left(1-\mathrm{L}\mathrm{o}\mathrm{s}\mathrm{s}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{i}{\mathrm{o}}_{f,h,c}\right)$(3) Net yields were additionally calculated using the average crop-specific loss ratios of the reference data. Then, the ratio of original net yields to reference yields was calculated to evaluate the potential of CSA to reduce yield gaps due to lower FLW.

2.7. Qualitative data

2.7.1. Group discussion

To get qualitative insights into the practices, knowledge and thoughts related to the quantitative findings, a group discussion was conducted in a workshop with CSA gardeners (one from each of the four sampled CSA initiatives; all between 30 and 40 years; two males, two females) in May 2022. First, interim results were presented to provide an overview of the research outcomes. These results also included comparisons with average vegetable yields in Saxony (Statistisches Landesamt des Freistaates Sachsen, 2022). Second, the interim results were discussed to allow for a qualitative evaluation in the context of the research questions.

The group discussion was facilitated by a moderator and a co-moderator, who followed an interview guideline. The guideline was based on the structure suggested by Krueger and Casey (2015) and was open-ended allowing adjustments during the discussion process. During the group discussion, the depth of the questions increased. Opening and introductory questions were asked first, followed by transition, key and closing questions (Krueger & Casey, 2015). When formulating the questions, it was ensured that the respondents were given the greatest possible degree of self-determination and that the respondents were able to react to contributions from others.

The group discussion was recorded on video and audio. Key points were noted during the workshop. Based on the recordings, the workshop was transcribed based on the literal transcription method according to Dresing and Pehl (2015) (Table S4). The focus of the transcription was on good readability, easy learnability and a limited implementation time (Dresing & Pehl, 2015), representing a verbatim transcription (as opposed to summarizing or phonetic). A translation of the transcript was made from German into English, using ATLAS.ti 22 Qualitative Data Analysis (QDA) software for the structuring and coding process.

2.7.2. Documentary method

For analyzing the results of the group discussion, the documentary method according to Bohnsack (2014, 2021) was applied. The actors interpreted the quantitative research results themselves, which was needed to distinguish subjective assessments from objective facts. The strength of the documentary method lies in the contextualization by the participants themselves: It allows to identify typical sets of implicit knowledge and pre-assumptions of the participants (Bohnsack, 2021). In contrast to hypothesis-testing procedures, this reconstructive interpretation considers the social interactions within a group (as well as the positionality of the researchers and the moderators) and considers the utterances of the participants in the discussion in a larger context. This context includes not only the situation of the group discussion, but also the roles and self-reflection within the practices and discourses of gardening, food provisioning and food loss prevention. In particular, the participants’ ideas about economics and the environment, as well as social and political constraints and opportunities for change, were important. Thus, the qualitative evaluation enables both a connection of the quantitative research results with specific practices of production and distribution on the part of the gardeners and an embedding of their expressions in the social context of their self-reflection and their anticipation of the motives and desires of the consumers.

After transcription, the discussed topic was divided into main and sub-categories. This was already a first step in interpretation, which was made accessible by the researcher through the ‘translation’ of a mostly milieu-bound language without explicit knowledge of the publication (formulating interpretations) (Bohnsack, 2021). In the next step of the analysis, the group's collective practice and orientations that occurred within the group discussion were identified. Therefore, selected passages were first translated according to their wording and then interpreted. The passages selected for reflective interpretation were either thematically relevant to the entire research question or particularly striking due to their interactive or metaphorical density (focusing metaphor) (Bohnsack, 2021). The obtained results were summarized and condensed, considering collective orientations as well as interactions and modes of communication (Bohnsack, 2021).

3. Results

3.1. FLW

The calculated FLW ratios along the FSC ranged from 7.3% (CSA, distribution stage) to 27.1% (reference, consumption stage) (Figure 1). Within the three stages, the CSA FSC showed constantly lower FLW ratios compared to the reference data, although variation between crops was high at the production and consumption level. Regarding single crops, the CSA data showed constantly lower FLW ratios in the production and distribution stage except for savoy cabbages and tomatoes in the production stage (Table S5). In seven out of 16 crops, the reference data showed lower FLW ratios at the household level.

Figure 1. Mean food loss and waste (FLW) at each stage of the food supply chain (FSC) for community-supported agriculture (CSA) and reference data as the ratio of the amount of the FLW to the supply in the corresponding stage. Error bars indicate the standard deviation (SD) of different crops.

3.2. Yield gap compensation potential

On average, net yields were 26.9% lower than actual yields (Table S6). If average FLW ratios of reference data instead of CSA-specific ratios were used to calculate net yields, they were 31.5% lower. Accordingly, the CSA initiatives had the potential to compensate almost on third of possible yield gaps by reducing FLW. For potatoes, courgettes, white cabbages and tomatoes the potential was higher than 50%, while for broccoli, fennel, savoy cabbages and turnips it was below 10% (Figure 2; Table S6).

Figure 2. Mean potential of community-supported agriculture (CSA) to compensate possible yield gaps due to lower food loss and waste for different crops. Error bars indicate the standard deviation (SD) of the investigated CSA initiatives.

3.3. Group discussion

Severeal topics and subtopics were derived based on the statements of the CSA gardeners (Table 3). Reasons for FLW were identified at all stages of the CSA FSC by the gardeners. On the production side, FLW occurred due to the cleaning of the vegetables (especially savoy cabbage and white cabbage). Regarding distribution, the logistics could be improved to reduce FLW, therefore the communication between gardeners and consumers could be essential to plan quantities and types of food products delivered. On the consumption side, knowledge of the correct storage and preparation of food could reduce FLW. For example, one participant of the group discussion commented as follows:

‘There have been surprises with the onions. Onion waste is relatively common, possibly due to a certain ignorance of storage capacity, some of which were stored in the refrigerator by members, for example’. (line 205).

In addition, a gardener stated that lifestyle changes (e.g. dietary changes) of some users are needed to avoid FLW. Standards were mentioned as the main cause of FLW in the conventional FSC, e.g. in relation to the shape or colour of the food. One participant noted:

I would say that is a well-known topic. So, the figure that a third of the world's food is lost is kind of common knowledge. And it is also known that standardization counts as a major cause of losses, but I have never dealt intensively with the topic myself. I tend to have a stronger connection through working on the market and there are mostly really pointless reasons for loss, such as a crooked carrot. (l. 147).

However, the gardeners sometimes noticed differences in terms of the quality requirements of food among CSA members, so that certain standards were also prevalent there. In the CSA initiatives, for example, ‘crooked carrots’ were sometimes left at the distribution points, with some members preferring either particularly small or particularly large carrots. One gardener was convinced that CSA members are more aware of nutrition and ecology, but they still make similar mistakes as consumers fully relying on supermarkets.

Table 3. Topics and subtopics based on the statements of the CSA gardeners.

Topic	Subtopics
FLW	Reasons for FLW Personal assessment vs. reality
Responsibility	Gardeners Consumers Communication/ exchange Consumer vs. gardener responsibility
Research method	Assessment of the project Definitions Weaknesses Derived strategies and integration into future work Suggestions for improvement
Sustainability	Priorities in sustainability aspects Comparison of CSA and conventional agriculture

One gardener was positively surprised that CSA initiatives could achieve similar yields as conventional agriculture, particularly if FLW are considered (net yields), despite the manual labour. Where CSA reached lower yields, this was related to the choice of seeds and varieties. One participant noted:

Well, I think there is also a big difference here, which varieties, which seeds are used. That is how it is for us, we often use Bingenheimer, for example, it is an old variety, sometimes it is just about the taste and not the yield. Sometimes you cannot keep up with that. (l. 233).

While the CSA initiatives primarily focused on taste, conventional agriculture mainly focused on yield potentially at the expense of quality.

Responsibilities for decisions that ultimately affect FLW amounts and other sustainability aspects were sometimes unclear. Based on several statements, it became clear that the gardeners often see themselves as responsible for making decisions, especially regarding food quality and cultivation planning. Some of the gardeners stated that they may have to grow fewer vegetables to minimize the development of FLW at the distribution points. In addition, some growers were wondering whether more sorting-out should be done to reduce FLW among consumers:

But it is also interesting that there are still further differences. So, between the CSA initiatives and with the individual gardeners and project members. So, it is crazy that we then have to decide, can you still eat it or not? How much do the vegetables need to be cleaned? (l. 127).

However, one grower argued that consumers have the responsibility to reduce FLW, e.g. by improving food storage conditions:

Okay, you could somehow optimize that now, grow less white cabbage and stuff like that. But on the other hand, I also think people should get used to the fact that this is ‘Krautland’. If you want to survive here in these latitudes, you can simply rearrange your diet a bit (…) That is why the question here is always, do we adapt to society or do we change society? (l. 254).

The gardener also stated that consumers need to reconsider their quality standards. To reduce FLW, an exchange between gardeners and consumers was seen as helpful, for example, to adjust the cultivation plan, yet opinions among members might diverge. Moreover, the question arose, to what extent gardeners are responsible for meeting the wishes and requirements of members. Nevertheless, the gardeners saw a great advantage in the exchange with CSA members, e.g. to recognize their satisfaction due to high-quality seeds and varieties:

We also had these picture book tomatoes that looked great and did not taste like anything. And you have a lot. But that is not what we want to offer our members either, because then they can run to the nearest Rewe store, like what our members see, where the sparks are in their eyes, where they say oh that tastes good. (l. 237).

At several points in the group discussion, the gardeners draw comparisons between conventional agriculture and CSA. While in conventional agriculture a direct exchange between producers and consumers is not possible through the long FSC, a direct contact and exchange is essential in CSA. Further, conventional agriculture typically focuses on yields, while CSA initiatives prioritize various sustainability aspects. The ecological sustainability aspects identified include resource efficiency, e.g. the low occurrence of FLW due to the reduction of aesthetic standards, preserving biodiversity, e.g. by creating habitats in the field, and reducing emissions, e.g. by reducing transport routes and working without machines or fossil fuels:

‘It is kind of nice to notice that it is not just about getting out the largest possible amount of vegetables. It is about somehow supplying people with vegetables and somehow doing it in harmony with the place’ (l. 281).

Social sustainability aspects mentioned include consumer satisfaction, e.g. through direct exchange (either face-to face or via online-platforms), and educational aspects, e.g. through acquiring knowledge about the correct storage of food or getting to know different types of vegetables.

During the reflective interpretation, two key aspects emerged, namely the responsibility of CSA gardeners vs. consumers and sustainability strategies regarding losses. The gardeners sometimes differed in their opinion of how much responsibility they have compared to those of CSA members (see above). In all four CSA initiatives, minimizing the amount of FLW and foster a circular economy were very relevant aspects. Possible solutions to further reduce FLW included leftover boxes (where products such as ‘too small’ or ‘crooked’ carrots will be made accessible to members separately) or suggestions for the preparation of ‘unfamiliar’ food products. Another strategy for dealing with FLW from traditional agriculture mentioned by a gardener was animal husbandry, e.g. by feeding FLW to chickens, which was done by one CSA initiative. However, in almost all CSA initiatives, compost appears to be the best alternative for using FLW while closing the nutrient cycle. For a gardener, composting food seemed problematic, as it does not justify the time and effort for growing and harvesting. Another concept of a CSA to reduce FLW is the ‘AG Residual Utilization.’ Within this working group, the same three to four members of the CSA try to reduce FLW, e.g. by preserving or processing food and giving it to the consumer or by teaching the consumer to do the same. One gardener mentioned the benefit of a ‘buffer system’ given the proximity of the Leipzig CSA initiatives. This can reduce losses by exchanging surplus vegetables between CSA initiatives.

4. Discussion

The results of this study show that FLW in the investigated CSA initiatives were generally low. In particular, during production and distribution, FLW was around 50% and 70% lower, respectively, compared to the reference studies. According to the results of the focus group discussion, reasons for FLW on the production side included weather-related crop failures, inadequate adaptation measures, e.g. tomatoes were completely lost at one studied CSA, and overproduction. Differences in management between the CSA initiatives and different responses of crops, e.g. to humidity, are likely to explain the large variation on the production side, while variation in the distribution points might be related to members’ preferences (e.g. lower preference for fennel). Most of the field losses on the production side were composted or tilled under, which is a common strategy to deal with food losses at the farm-level (Soma et al., 2021). On the consumption side, waste was still around 20% lower compared to reference studies, yet with large variation across crops and individuals. Possible reasons for FLW at the consumption level included insufficient knowledge about the correct storage (e.g. in the dark and at low temperature) or preparation (e.g. some edible parts of a vegetable are not used), as well as partly high or divergent quality standards among CSA members. Similar reasons partially apply for FLW at distribution points. This indicates that despite a higher awareness among CSA members (Chen, 2013; Pole & Kumar, 2015), similar behavioural patterns still exist for some CSA members and further reductions might be possible.

Different approaches may help to reduce FLW in CSA. On the production side, losses could be minimized by adapted management (e.g. earlier harvest to prevent flowering, protected cultivation). During the group discussion, CSA gardeners proposed a stronger personal contact between gardeners and CSA members to adjust the cultivation plan, if necessary, both in terms of quantity and crop selection. Moreover, more educational work and knowledge exchange were suggested as a way to reduce FLW on the consumption side, e.g. to increase knowledge about the correct storage or preparation of food. CSA memberships have been shown to hold the potential to promote behavioural change. For example, a CSA membership can increase the preference for seasonal products or change eating and cooking habits, as well as increasing the appreciation for farming (Russell & Zepeda, 2008). Whether these changes in behaviour can promote a reduction in FLW among CSA members would have to be examined more closely in follow-up studies. Our findings are in line with N. Baker, Popay, et al. (2019) and support the assumption that direct marketing initiatives reduce FLW (BMEL, 2019; Cattaneo et al., 2021). In contrast, reducing FLW in conventional systems might be more difficult, as long as questionable standards regarding size and shape persist (Göbel et al., 2015; Gustavsson et al., 2011). However, circular economy is an approach to generally reduce and recycle FLW (Liu et al., 2023; Wang et al., 2021).

Our findings indicate that due to lower FLW, CSA initiatives have a large potential to compensate for yield gaps between organic and conventional agriculture (Ponisio et al., 2015; Seufert et al., 2012). Over all crops considered, a yield gap of around 30% could be compensated, which is higher than average yield gaps of c. 20% (Ponti et al., 2012). However, yield gaps are higher if the larger area demand due to a higher proportion of legumes and green manure for nitrogen supply in organic agriculture is considered (Kirchmann, 2019), and largely depend on the local context and the crops considered (Mayer & Mäder, 2016; Ponisio et al., 2015; Seufert, 2019). Regarding vegetables in Germany, yield gaps between conventional and organic agriculture were between 8.9 and 67.2% (BMEL, 2018). According to the findings of this study, even the relatively high yield gap of tomatoes (c. 50%) could be compensated. Only for pumpkin, the observed yield gap could just be partially compensated. The compensatory potential could be further increased by dietary changes towards a more plant-based diet to reduce food demand (Muller et al., 2017), as the production of animal products uses around 83% of the world's farmland and contribute to more than 50% of food-related emissions (Poore & Nemecek, 2018; Willett et al., 2019). In this context, CSA can promote behavioural changes towards more sustainable lifestyles, including a more plant-based diet (Allen et al., 2017; MacMillan Uribe et al., 2012; Rossi et al., 2017).

Actual yield gaps in CSA might be lower compared to organic agriculture in general, as agroecological approaches are widespread in CSA. These approaches have been found to substantially reduce the yield gap between conventional and organic farming while fostering agrobiodiversity and other ecosystem services such as soil quality, pest control, pollination, nutrient management, water retention and carbon sequestration (Garbach et al., 2017; Kremen & Miles, 2012; Ponisio et al., 2015). Such approaches include intercropping (Li et al., 2007; Snapp et al., 1998; Vandermeer, 2012), organic management (Lotter et al., 2003; Pimentel et al., 2005) and a high variety of crops and crop rotations including a high share of legumes (Lammerts van Bueren et al., 2011; Murphy et al., 2007).

In addition to the reduction of FLW, other strategies and sustainability aspects that are established in the CSA initiatives were mentioned in the focus group discussion. The gardeners stated that ecological, social and economic aspects need to be combined to farm sustainably. Ecological measures include the promotion of biodiversity or species diversity (e.g. by creating wet areas or deadwood hedges on the field), as well as the greatest possible renunciation of machines and fossil fuels and the prioritization of short transport routes. In other CSA studies, it was found that especially environmental sustainability plays an overarching role for CSA members (Bobulescu et al., 2018; Farr-Wharton et al., 2012; Medici et al., 2021; Schnell, 2013). Social sustainability can be achieved by fostering education about seasonal and regional foods as well as correct storage and preparation. CSA memberships could further have impacts on the attitude and behaviour of the CSA members due to farm exposure, interaction with the gardeners and by introducing them to seasonal and local food products (Russell & Zepeda, 2008). Another aspect is the de-standardization as well as the de-commodification of food products, including pre-financing of the products that the CSA initiatives provide (Blättel-Mink et al., 2017). While some members are satisfied with understanding that the food products might look different and do not need a ‘perfect’ shape, other members prefer clean and standardized food products (Goland, 2002). Furthermore, economic sustainability aspects mentioned in the group discussion include the creation of regional networks of gardeners, members and other actors. These findings are in line with other CSA studies showing that CSA initiatives do not primarily focus on economic profitability but sustainability and community building (Bobulescu et al., 2018; Galt, 2013). However, it was shown that self-exploitation among CSA gardeners and insufficient income is a challenge in CSA (Galt, 2013; Lass et al., 2003).

To our knowledge, this is the most comprehensive research on CSA yields and FLW based on actual measurements. Nevertheless, this pilot study faces three major limitations. First, research was restricted to four CSA initiatives and one single growing season. Given the high importance of climate (variability), multiple seasons should be considered, especially in the light of climate change (Challinor et al., 2014). Moreover, no comparisons with individual and comparable reference farms were possible. Second, multiple uncertainties arise from the fieldwork. Although a standardized sampling protocol was developed, potential subjectivity of the five researchers and several CSA members involved in fieldwork particularly affected product definitions, which in turn affected the amount of recorded FLW. Further, certain FLW may have remained undetected (e.g. regarding potatoes, only unharvested biomass at the surface and in the upper soil layer was detected) or were biased (e.g. different weather conditions and their influence on the weight of the product were not considered). Third, inconsistencies regarding the definition of products and FLW between and within the studied CSA initiatives and the reference data existed.

In terms of possible policy implications and recommendations, integrated strategic development concepts of urban and peri-urban regions have to explicitly include agriculture to help approaches such as CSA or other community-oriented food production systems out of the niche in which they still remain (Egli et al., 2023). For example, many cities already have elaborated green development and management plans for local recreation and health benefit of their urban dwellers, yet the links to food production in terms of both food supply and as a social activity for urbanites is mostly missing (Thornton, 2018). Including and supporting CSA and other approaches in this context could be a promising governance solution to improve health (Allen et al., 2017; Rossi et al., 2017), civic engagement (Sumner et al., 2010), as well as other ecological, social and economic sustainability aspects (Egli et al., 2023), and to reduce food loss and waste, which is in line with the Sustainable Development Goal 12.3 (halving food loss and waste by 2030), and which in turn contributes to climate change mitigation and other environmental benefits as less food needs to be produced (Read et al., 2020; Shafiee-Jood & Cai, 2016).

This research opens new avenues for future research. Future studies could include more farms, more crops, different regions, a longer temporal extent and in-field comparisons of CSA and other systems, which would be helpful to better understand, whether the patterns observed here are generalizable. Further studies should also integrate other sustainability aspects, for example, related to non-harvested areas (e.g. green manure), agricultural inputs (e.g. organic vs. synthetic fertilizers), biodiversity measures, emissions, as well as social and economic benefits and trade-offs, for example, related to higher labour input due to manual and organizational work. Future research could also address the governance of agro-food systems, as well as ethical approaches and questions (power relations and attribution/distribution of responsibilities and moral ethics between growers, consumers and authorities). This information will be essential to comprehensively understand the contribution of CSA to the transition towards resilient and sustainable food systems.

5. Conclusion

This study indicates that community-supported agriculture (CSA) is a promising approach to reduce food loss and waste (FLW). In the four investigated CSA initiatives, average FLW during production, distribution and consumption were lower compared to reference data. Therefore, CSA could compensate possible yield gaps compared to conventional agriculture. Beside FLW, the examined CSA initiatives focus on biodiversity, education of CSA members, food quality and the reduction of fossil fuels and machines. Including and supporting CSA in the context of integrated development plans could be a promising governance solution to improve health, as well as other ecological, social and economic sustainability aspects, and to reduce FLW. Further research should integrate more farms, a longer time period and other sustainability aspects as well as questions related to governance and ethics to better understand the transformative potential of CSA.

Acknowledgements

We thank the four involved CSA initiatives (Ackerilla eG, Allerlei e.V., Gemüsekooperative Rote Beete eG, Kleine Beete e.V.) for their continuous support and collaboration during study design, field work and analysis. We also thank Anna Spiller and Judith Rüschhoff for supporting field work, and various CSA members for supporting field work in the distribution points, as well as the supervisors of the related master theses.

Disclosure statement

No potential conflict of interest was reported by the authors.

Data availability statement

Major datasets derived during the studies and relevant codes are openly available here: https://doi.org/10.5281/zenodo.8211114.

Additional information

Funding

This research is part of the project InnoLand-Saxony – Innovative models for sustainable and regional value chains: synergies and potentials of community-supported agriculture in Saxony. This measure is co-financed with tax funds on the basis of the budget passed by the Saxon state parliament (FKZ:100595134).