Dairy production is an important economic sector in Austria. In 2019, 527,000 dairy cows produced 3,781,000 tonnes of raw milk (Statistik Austria, 2020). This represents 1,356 Mio € or 18% of the Austrian agricultural production (BMLRT, 2020). Self-sufficiency in the Austrian dairy sector is given for drinking milk, cream, and cheese. For butter, the self-sufficiency rate is 69% (Agrarmarkt Austria, 2020a). Around 5,000 workers were employed in 84 Austrian milk processing companies in 2019 (BMLRT, 2020). As for all sectors of the economy and society, the COVID-19 pandemic had great impacts on the dairy industry. Rapid changes in demand required rapid adaptation of dairy production. Additional hygiene measures and concepts had to be taken. The long-lasting securing of supply chains, including those for food, has once again moved into the spotlight (BMLRT, 2020). In order to understand specific impacts, it is important to gain a general overview of the processes of the drinking milk supply chain and the underlying systemic structure. Therefore, the research question was defined as follows: What impacts does a pandemic that is transmitted from person to person have on the drinking milk supply chain in Austria? To answer this question, the following methodological approaches were chosen: Business Process Model and Notation (BPMN) and a qualitative System Dynamics analysis based on a Causal Loop Diagram (CLD). The case of the COVID-19 pandemic in the year 2020 in Austria was applied to evaluate the findings on the basis of a real-world event.
The BPMN method is used to break down long processes into their individual steps, and thus generate an understanding of the individual procedures and workflows. CLDs represent a systemic approach, whose aim is illustration of complex systems and their interactions in order to develop a better understanding of the system and its interrelationships and dependencies. Due to their holistic view, the combination of both approaches is qualified to generate a comprehensive overview of the Austrian dairy industry when influenced by a pandemic. The data basis for these methods was obtained in the course of literature research and expert interviews. The topic of this paper is the dairy supply chain, with special attention to drinking milk, as it makes an important contribution to national food security. The limitation of food security to the national view leads to a focus on the national dairy industry, dependencies and connections with foreign countries are not considered, except for the export of dairy products in the CLD.
The remainder of the paper is structured as follows: In the next section, materials and methods are presented. After that, the results are described and divided into three parts: the process description with BPMN, the systemic description with a CLD, and description of the use case. In the following discussion, key findings of all three approaches are contrasted and, finally, combined in the conclusion.
This paper is based on the methods of BPMN and CLD. To carry out these methods, the computer programs Bizagi (version 3.4.1.068) and Vensim (version Professional 7.2 [Single Precision] ×32) were used.
BPMN consists of several symbols to describe processes. Developed at the beginning of the 21st century, it is the leading standard in the field of business process languages (Chinosi and Trombetta, 2012). Due to its expressive and simple symbols, the method is broadly applied. Every process begins with a start event and finishes with an end event. In the middle, one activity is followed by another, connected with sequence flows, depicted as arrows. Gateways, squares marked with “x”, stand for decisions, where one of the different options must be taken (Göpfert and Lindenbach, 2013).
CLD is a method of System Dynamics, described by Forrester (1997). A CLD consists of variables connected with arrows. The variables are quantities or states that can increase or decrease. The arrows express the action that happens between the variables, on the one hand by the direction of the arrows indicating a cause–effect relationship and on the other hand by labelling the arrows as either positive or negative. A positive arrow means that if the cause variable
Within this study, a literature review and expert interviews build the data basis for the BPMN and CLD analysis and the use case. On the one hand, the literature research consists of scientific literature, textbooks, and legal regulations on the subject of drinking milk production. This was done using a snowball search in the scientific database BOKU:LITsearch with different combinations of the key phrases
Expert interviews were conducted with two management members of a leading dairy in Austria (Expert interviews 1 and 2) and three representatives of the Chamber of Agriculture responsible for dairy farming in different Austrian federal states (Expert interviews 3, 4, and 5). They were selected based on the research needs identified in the literature review: representatives of dairies as important players in the system of drinking milk production and representatives of the Chamber of Agriculture as those who have a good overview of current developments in Austrian agriculture. The experts were interviewed using an interview guideline based on the most important topics and unanswered questions identified in literature research.
Figure 1 shows how the different methods are related. Literature research and expert interviews form the data foundation for all three applied methods, BPMN, CLD, and use case. First, the BPMN was created to understand the processes of the drinking milk production. Then, the CLD was modelled to understand feedback effects and look at the drinking milk production system under the influence of the COVID-19 pandemic. Both modelling processes were an iterative process in which the models were repeatedly validated against the literature and the expert interviews. Finally, the case study was created to cross-check the CLD with the actual developments in Austria.
The material flow, depicted as a business process model in Figure 2, takes place at the following locations: dairy cattle farm, collection point, transport, dairy, and sales point. After milking, the raw milk is cooled and stored in the milking tank. There are two different channels of milk sales: disposal to a dairy and (legally limited) direct marketing. In the latter case, further processing and sale takes place at the dairy cattle farm. In 2019, 3.2% of the Austrian raw milk delivery was used for private consumption or sold by direct marketing (Statistik Austria, 2020). Due to its minor importance for the national milk supply, direct marketing of raw milk is not further considered. In case of disposal to a dairy, the milk tanker collects the milk either at the dairy cattle farm or at a collection point and transports it to a dairy. After reception, the milk is processed and filled into different packages. The final product is transported to the different sales points: food retailing, central market, export, etc.
The processes of milking and processing at the dairy are complex and consist of many subprocesses. These are not part of Figure 1, but described in brief in the following. To maintain udder health and a high milk yield at the same time, it is important to observe the exact sequence of the individual work steps during milking (Tröger, 2003). There are different milking systems, which can be generally categorized into automatic and non- or semi-automatic milking systems. The non-automatic milking systems are: bucket or pipe milking system, and milking machines of various designs. The individual process steps during non- or semi-automatic milking are carried out by an agricultural worker. The first step is the verification of a cow's eligibility for milking. If an animal is sick, treated with antibiotics, or its milk has a high cell count, it must be milked at the end and the milk must not be sold (Regulation [EC] No. 853, 2004). The next steps are pre-milking and visual quality control of pre-milking. If necessary, a mastitis test is carried out and samples for laboratory examinations are taken. Afterward, udder and teat cleaning, application of milking equipment, and monitoring of the milking process take place. If automation is available, one, several, or all of the following steps can be automated: removal of the milking equipment, disinfection of the teats, and cleaning as well as disinfection of the milking equipment (Savary et al., 2010). If the last cow elected for milking has been milked, the milking equipment is cleaned and disinfected.
If milking is done by an automatic milking system, all process steps are realised automatically by a milking robot. The process steps basically remain the same. During milking, various parameters are measured by the milking robot and used to assess the quality of the milk (Häußermann, 2017). The subprocess of milk processing varies depending on the type of drinking milk produced. To produce pasteurised drinking milk, the process steps are the following: cleaning, preheating, adjustment of fat content, homogenization, pasteurisation and cooling (Becker and Märtlbauer, 2016). Extended shelf life milk (ESL-milk) can be produced either by direct superheating, indirect superheating, or bactofugation. Bactofugation is a process in which spores and vegetative bacteria are largely removed by centrifugal force (Strahm and Eberhard, 2010). The most frequently used method, however, is microfiltration. For filtration, the milk is first cleaned and separated into cream and skimmed milk. The microorganisms are then separated from the skimmed milk by micro-filtration. The cream is heated and homogenised. Finally, the filtered skimmed milk and cream are mixed and pasteurised. UHT-milk (long-life milk) is produced by indirect or direct heating, but heat treatments are performed at higher temperatures than for ESL-milk (Strahm and Eberhard, 2010).
Figure 3 shows the CLD of the impacts of an unspecified pandemic that is transmitted from person to person. The moment of observation is shortly after the outbreak of the pandemic; long-time consequences are not considered. For a better understanding, the most important variables of the CLD are explained in Table 1.
Explanation of the most important variables of the CLD
Tabelle 1. Erklärung der wichtigsten Variablen des CLD
Pandemic intensity | The intensity of a pandemic affecting a society can be measured with different indicators, for example, the number of newly infected people per day |
Human health | The health status of a population, which can be measured with different indicators, for example, the number of patients in hospitals or the number of people who are on sick leave |
Extent of shutdown | Government measures to contain the spread of a virus. The extent varies, according to which and how areas of work and public life are restricted or closed |
Packaging gap | Lack or surplus of packaging material for dairy products. The greater the gap, the greater the lack of packaging material |
Dairy staff gap | Lack or surplus of dairy staff. The greater the gap, the greater the lack of dairy staff |
Raw milk gap | Lack or surplus of raw milk. The greater the gap, the greater the lack of raw milk |
Available staff for milk collection | Available staff for the transportation of raw milk from dairy cattle farms to dairies (milk tanker drivers) |
Dairy production quantity | The quantity of production of dairy products in dairies |
Available dairy products | The number of dairy products available for sale in supermarkets and at wholesalers |
The starting point of this CLD is the
The
The
On the left side of the CLD, the topic of raw milk demand and supply is addressed. The
The
Deciding which dairy products to produce can help to manage a surplus of raw milk supply. As throughout the year, the peak of milk supply does not co-occur with the peak of milk demand, this is a common procedure. As shown in Figure 4, in the decision-making process, the amplified production of milk powder, butter, and cheese products for storage is used to handle a raw milk surplus.
As many parts of the economy and the social life, the Austrian dairy industry was also hit severely by the impacts and consequences of the COVID-19 pandemic in the year 2020. In the following, the use case is described using the key points market changes, supply chain echelons, counteractions, and a special focus on the situation in Western Austria.
The milk price, which Austrian farmers received in the first half of 2020, was slightly below the previous year's level and then rose from July onward, finally reaching 40.33 Cent/kg. In terms of milk deliveries, the total volume in 2020 was 2,608 tonnes, or 0.08%, below the previous year's volume. Organic milk, however, achieved an increase of 2.7% (Agrarmarkt Austria, 2021a). As a first reaction to the pandemic, there were occasional calls from dairies to dairy cattle farms to reduce milk deliveries (Agrarmarkt Austria, 2020b).
The COVID-19 pandemic triggered considerable market changes because there were strong segment shifts due to the closure of restaurants and hotels. The largest production increases in 2020 were in long-life milk (+11.6%), soft cheese (+10.1%), sour cream (+6.2%), and butter (+5.2%). There was a reduction in the production volume for pasteurised milk (−8.3%). Here, the trend has generally been toward milk with a longer best before date for some time, and this was reinforced by the pandemic. There were also production losses for sweet cream (−7.8%) and hard cheese (−4.5%) (Agrarmarkt Austria, 2021b). In particular, there was a slump in demand for cheese products for high-end gastronomy and for products in bucket packs.
The trend in 2020 was increasingly toward regionality and organic products, which is why direct marketing was also able to record large increases (Agrarmarkt Austria, 2021b). However, it is still questionable whether this change will last. In general, the dairy industry in Austria was able to manage the production adjustments well, as capacities for overproduction are available in the dairies. Due to the seasonal nature of the milk supply, the dairies are prepared for fluctuations (Expert interview 1, 2019).
At the beginning of the pandemic, there was a staff problem on farms; but by amending the quarantine order for farmers, this was solved, and farmers were able to take care of maintaining the farm even in spite of quarantine (Expert interview 3, 2020). Milk transport in Austria is handled by forwarding agents. The market of providers is relatively small, but stable. A loss of drivers is difficult to replace, as they need special training to operate the equipment. In addition, a milk tanker driver drives to several farms on a collection tour, so the loss of one driver means the loss of an entire tour. By driving to several farms, milk tanker drivers also pose a great risk in terms of virus spread. However, strict hygiene strategies and measures prevented severe absences and milk collection was maintained (Expert interview 2, 2020).
In the dairies, the already existing hygiene measures were tightened to avoid spread of the virus because spread of infection would mean the loss of an entire shift. Moreover, the possibility of a diversion to other filling lines exists, so that filling is guaranteed. There were COVID-19 infection cases in Austrian dairies, but they did not lead to any production problems. There are hardly any weekly or monthly commuters in the Austrian dairy industry, so there was no impact of entry regulations. However, there are daily commuters who are employed at locations near the border. For them, nevertheless, it was possible to cross the border at any time (Expert interviews 3, 4, and 5, 2020).
In the case of packaging material, there was the risk of a supply chain disruption. However, the supply chains worked well because the borders were also always open for goods transport. As production increases within the usual growth rates, there was always enough packaging material available. In general, however, there are only a few suppliers of packaging materials, and therefore the supply chain is vulnerable even if one supplier quits (Expert interview 2, 2020).
Private storage was a measure taken by the EU to stabilise the market and counter the crash in milk prices (European Commission, 2020). It involved financial subsidies for storage costs in order to be able to hold back goods for a few months. In Austria, it was used from the beginning of May 2020 and skimmed milk powder, butter, and cheese were stored. According to interview partners, this measure worked well and most of the stocks were reduced again by the end of the year (Expert interview 2, 2020).
The sale of cattle cows was confronted with some difficulties, as demand had collapsed, which led to a backlog on some farms, as stated by the interviewed experts 3 and 4 (2020). In some federal provinces, the slaughter of dairy cattle was financially supported through a “culling premium”, which had a relieving effect on the market situation (Expert interview 5, 2020).
The situation in Western Austria (especially Tyrol and Vorarlberg) must be viewed in a somewhat different manner, as there exists a greater dependence on tourism than in the rest of the country. After losses in spring 2020, the summer was very successful for the dairy industry and the losses could be partially compensated. However, as there was hardly any tourism in the winter, large losses were recorded again. Western Austrian dairies suffered sales losses of approximately 25% in 2020; small cheese dairies with their own sales, which are very focused on tourists, lost up to 75% of their sales. Private storage was hardly used, as the warehouses are mostly filled with long-maturing cheeses (Expert interview 5, 2020).
Overall, the Austrian dairy industry was able to cope well with the upheavals caused by the COVID-19 pandemic, the system proved to be stable, and the supply chains worked. Broadly positioned dairies that offer different products and supply different segments managed the crisis more easily than dairies with greater specialisation.
The BPMN approach showed that the dairy supply chain is very complex, and especially, the processes of milking and milk processing consist of many individual process steps. However, the low number of required processing locations shows that milk is a locally produced product, which is also related to its high perishability and the need for quick processing.
As it can be seen in the CLD, self-influencing effects can emerge with regard to packaging material and dairy staff. These two points are, therefore, important determining factors. In expert interviews, the supply of packaging material was also mentioned as a critical point, as the market on the supplier side is very small. The assessment of dairy personnel as a neuralgic point could also be confirmed; dairy specialists are in demand and apprentice positions are available.
In order to mitigate the effects of a pandemic, various considerations can be made on the basis of the CLD. For example, if there is not enough processed milk, policy makers could consider increasing the extent of the shutdown; this means drawing a negative link from
Another link that could be drawn is a positive link between
As shown previously, in 2020, the milk volume in Austria remained relatively stable. The sales problems with the slaughter of dairy cattle were also manageable. However, the dependencies of the underlying system, as presented in the CLD, were present and were also shown to a moderate extent in the use case.
This paper provides valuable insights into the organisation of the Austrian dairy industry. The BPMN showed the complexity of the dairy supply chain and illustrated subprocesses in more detail. The CLD showed the structure of the system when hit by a pandemic like COVID-19. Besides raw milk supply, the availability of packaging material and dairy staff affects the possible quantity of dairy production. Thereby, the availability of packaging material is doubly influenced by the extent of the shutdown, as both personnel and transport can be affected. The availability of staff and packaging material could be identified as particularly critical points by applying System Dynamics. The use case of the impacts of the COVID-19 pandemic on the dairy industry in Austria in 2020 pointed out the actual impacts and developments. It turned out that the dairy industry in Austria was able to react effectively to the changed conditions in 2020. But it is not certain that this is the same in other crises scenarios. For example, there might not be enough storage capacity to store long-life dairy products, a shortage of labour staff could occur, or packaging materials could be scarce. Particularly critical in the case of drinking milk is that there are hardly any packaging alternatives, but raw milk is lost as food if it is not processed quickly enough. The identification of feedback loops helps to focus on the critical points in further observation of the system. Subsequently, ongoing trends, developments, and their effects can be analysed based on the presented system. As some dependencies of a supply chain are not so obvious, these can be filtered out by the applied mix of methods. Especially feedback effects, which can be identified using CLDs, are important to consider in risk analyses, as supply chains do not only consist of cascading effects and feedback loops can develop a strong dynamic of their own. In addition, the empirical knowledge presented in the use case of this paper can be used to deal with other crises.
Further research could use the same mix of methods to analyse other important food supply chains in Austria, such as the supply chain for pork. Another research extension could include a quantitative System Dynamics analysis with Stock and Flow models or computer-based simulation-optimisation. This may be a vital contribution to national food security in the long term.