Heterogeneity in financial performance caused by the COVID-19 pandemic: Food producers in Czechia

© 2023 Published by VŠB-TU Ostrava. All rights reserved. ER-CEREI, Volume 26: 63–74 (2023). ISSN 1212-3951 (Print), 1805-9481 (Online) Heterogeneity in financial performance caused by the COVID-19 pandemic: Food producers in Czechia Tomáš HERYÁN1 * , Martina NOVOTNÁ2, Petr GURNÝ2, Aleš KRESTA2 1 Department of Finance and Accounting, School of Business Administration, Silesian University in Opava 2 Department of Finance, Faculty of Economics, VŠB-TU Ostrava Abstract This paper focuses on the heterogeneous impacts of COVID-19 on financial performance among middle-sized food product manufacturers in Czechia in dependence on ownership concentration. In particular, this work aims to estimate differences in these pandemic impacts among Czech companies with a major owner. The DuPont framework has partially revealed significant relations of return on assets depending on the return on sales, both reflecting earnings before interest and taxes. Neither is the only novelty in the use of heterogeneous difference-in-differences with cohorts of companies developed by Wooldridge (2021); however, 2018-2022 annual financial data from the Bureau van Dijk Orbis database for 480 companies have also been merged with epidemiological data, particularly the average morbidity per 10,000 inhabitants within NUTS-3 regions based on the open source database from the Nature journal (Naqvi, 2021). The tests revealed a significant difference between middle-sized companies with high ownership concentration and the rest of the market. Keywords COVID-19 economic impacts; Financial performance; DuPont analysis; Heterogeneous difference-in-differences with cohorts; JEL Classification: G32, M21 * [email protected] This paper was supported by the Ministry of Education, Youth, and Sports Czech Republic within the Institutional Support for Long-term Development of a Research Organization in 2023. Ekonomická revue – Central European Review of Economic Issues 26, 2023 64 Heterogeneity in financial performance caused by the COVID-19 pandemic: Food producers in Czechia Tomáš HERYÁN, Martina NOVOTNÁ, Petr GURNÝ, Aleš KRESTA 1. Introduction Manufacturing can be considered a key industry, as it contributes substantially to the gross domestic product and can also be considered a major employer in the economy as a whole (Haraguchi, 2016; Szirmai and Verspagen, 2015). Companies in this sector have long spent large amounts of money on research and development, positively impacting the overall economy (Fernández and Palazuelos, 2018). Manufacturing also plays an important role in international trade, as its products are usually among the main export items (Rinaldi and Bottani, 2023). The general importance of manufacturing in the EU countries is evident from the economic indicators in each country's national accounts, especially the share of the activity in gross value added. In 2022, the manufacturing industry in the EU-27 contributed 17% to the total value added of all sectors, including services and public administration (Eurostat, 2023a). After a recent period of decline, manufacturing production in EU countries is showing an increasing trend. Although there was still a 7% year-over-year decline in 2020 due to the COVID-19 pandemic, production increased by 8% in 2021 and 5% in the following year in 2022. Furthermore, among the most important economic activities within the EU manufacturing industry by the value of products sold are basic metals and fabricated metal products, the manufacture of food, beverages, or tobacco products, and the manufacture of motor vehicles and other transport equipment, which represent 47% of the total value of output sold in the sector in 2022 (Eurostat, 2023b). The COVID-19 pandemic affected several industries and the impact on manufacturing was substantial given its share of total economic output. These impacts have been studied in related literature and can be broadly categorized into three main areas (Ardolino et al., 2022): (i) lockdowns and compulsory closures; (ii) social distance and teleworking; and (iii) changes in consumer behavior. Lockdowns and compulsory closures have had major impacts on production, marketing, closures, material shortages, and restrictions on the distribution of goods (Tang et al., 2021) and have led to the division of operations into surplus and essential (Carletti et al., 2020). This was accompanied by limited availability of human resources, raw materials, and consumables to shut down or suspend capacity in almost all industries (Paul and Chowdhury 2020; Singh et al. 2020; Xu et al. 2020). However, SMEs in particular experienced immediate adverse effects due to logistical challenges, reduced capacity utilization, and demand spillovers (Juergensen et al. 2020). Manufacturing companies were indirectly affected by restrictions imposed on restaurants, cafés, shopping centers, and general recreational and sporting activities (Juergensen et al., 2020; Seetharaman 2020). The COVID-19 pandemic has also reinforced structural changes, especially towards online services (Haapala et al. 2020). This forced shift to online sales has required a revision of inventory plans by a large number of companies and a consequent change in financial performance. Ardolino et al. (2022) further state that one of the key questions that need to be answered in follow-up research is the impact of the application of COVID-19 anti-pandemic measures on the economic performance of firms. In doing so, it is necessary to analyze not only the overall impacts of the pandemic on individual industries but also, in particular, the causes and varying degrees of intensity of these impacts in a heterogeneous setting. Answering these questions can play a key role in the behavior of firms and authorities in the event of similar events with global impact. Given the characteristics of the above pandemic interventions by each authority, it is necessary to rely on quasi-experimental strategies to identify causal effects. The difference-in-differences (DiD) approach appears to be appropriate in this case, as it offers considerable advantages over classical time series methods (Tobías, 2020). The DiD technique has so far been predominantly used in analyses of the impact of non-medical antiviral measures on pandemic indicators in different countries (e.g., Cho 2020; Dave et al., 2021; Flaxman et al. 2020; Hsiang et al., 2020). Although research has deployed just a classical DiD approach so far, the use T. Heryán et al. – Heterogeneity in financial performance caused by the COVID-19 pandemic: Food producers in Czechia 65 of the forthcoming years has to deal with a certain level of heterogeneity. Hence, the heterogeneous DiD technique is crucial in terms of employing different cohorts of companies. This paper aims to estimate differences in financial performance among Czech middle-sized food product manufacturers that have a major owner, particularly caused by the impacts of the COVID-19 pandemic. This study contributes to understanding the specific impacts of the pandemic on the financial performance of middle-sized food producers in Czechia, with a focus on ownership concentration. The paper utilizes the DuPont framework and a heterogeneous difference-indifferences approach to analyze the financial data of manufacturing companies from 2018-2022, merged with epidemiological data. The study further adds to the existing literature on the impact of COVID-19 on the economic performance of firms, particularly in the manufacturing sector. It highlights the need for further research to explore sub-indicators and draw more detailed conclusions. This paper is structured as follows, Section 2 briefly describes the role of the manufacturing industry and how its companies suffered by the pandemic. Section 3 is then divided into three sub-sections, first, including a few examples describing the heterogeneous differencein-differences approach, second, data are descriptively introduced, and third, the estimation output is further discussed. Finally, Section 4 concludes. 2. Manufacturers in the COVID-19 pandemic Szirmai and Verspagen (2015) focused on assessing the impact of manufacturing on economic growth and development by econometrically analyzing data from 88 countries, 21 developed economies, and 67 developing economies, over the period 1950-2005. They not only looked at whether manufacturing positively impacted economic growth in developing countries during the period but also compared this impact with other activities, particularly the service sector. Using a longer period also allowed them to assess whether the importance of manufacturing or services in economic growth changed over the period. The results of their study confirm a moderate positive impact of manufacturing on economic growth, assuming a sufficient level of human capital, which has not been shown for the services sector. However, this direct effect was only found for the subperiod 1970-1990, and since 1990, manufacturing has not contributed to growth as much as in previous years. Haraguchi (2016) examined the market value added (MVA) relative to GDP in selected countries, which are classified as more and less developed. His results show that while the MVA to GDP ratio in developed countries has been declining since 1970, for developing countries, some decline has occurred after 1990. However, this is not a long-term trend, so it cannot be concluded that manufacturing is declining in importance in developing countries. The decreasing share has not been due to changes in the quality or quantity of manufacturing activities but to the failure of manufacturing development in many developing countries compared to the significant development in a small number of countries, which has led to their concentration. Although previous authors have focused on the role of manufacturing in developing countries in particular, Fernández and Palazuelos (2018) examine the contribution of manufacturing to overall productivity growth in European economies and the contribution of individual industries to labor productivity growth. Based on an empirical analysis of 14 selected EU countries, their results do not confirm a lower share of manufacturing in aggregate productivity compared to the service sector. As they point out, aggregate productivity growth in more developed countries still depends on the size and dynamism of the manufacturing industry, and thus any policy to increase productivity in European countries should include a policy aimed at strengthening manufacturing. Clampit et al. (2021) address the impact of the COVID-19 crisis on the performance of firms in manufacturing, and they focus on the impact of selected financial ratios and R&D intensity on sales growth. Their study draws on data from 1,298 US manufacturing firms, comparing their performance over the preCOVID and COVID periods. Their findings show that firms' more risk-taking R&D strategies positively affected their performance during the crisis, pointing, in particular, to the positive impact of higher inventory levels compared to cash holdings. Similarly, Rinaldi and Bottani (2023) provide empirical results on the impact of the COVID-19 crisis on the logistics and supply chain of Italy's five industries: food and beverages, engineering, metallurgy, logistics and transport, and textiles and fashion. The authors synthesized clear evidence that the COVID-19 pandemic had different impacts on the selected sectors, these impacts occurred primarily at the beginning of the pandemic and were mostly of short duration, for example, the food industry and the logistics and transport sector were the least affected, with an overall increase in demand during the period under study. In addition, social distance has led governments and businesses to introduce home offices (Kanda and Kivimaa, 2020; Omary et al., 2020), which has further greatly influenced consumer behavior (Diaz-Elsayed et al., 2020). On the one hand, the spread of the COVID19 pandemic has enabled the reorganization of work, Ekonomická revue – Central European Review of Economic Issues 26, 2023 66 which exposed challenges that needed to be faced in the manufacturing industry (Bolisani et al., 2020). Restrictions on social contact and face masks have affected efficiency in some workplaces, leading to operational problems and adjustments in working hours (Garlick et al., 2020; Kurita et al., 2020; Telukdarie et al., 2020; Weersink et al., 2020). In some cases, teleworking reduces the efficiency of interaction and coordination (Ali Abdallah, 2021). Despite these difficulties, most manufacturers agree that work processes will be redesigned in the future, with social distance and remote work continuing (Moutray, 2020). With the spread of COVID-19, socioeconomic activities have gradually stopped, schools were closed, many manufacturing activities were curtailed, sports and entertainment events were canceled, tourism collapsed, and unemployment increased (Baker et al., 2020; Basilaia and Kvavadze, 2020; Devakumar et al., 2020). The sharp decline in tourism has strongly affected demand in the automotive and aviation sectors (Rahman et al., 2020; Sjoberg, 2020). Lockdown policies and the introduction of home offices have reduced sales in the apparel and fashion industries (Hilmola et al., 2020). On the contrary, most technology companies have seen an increase in demand due to home-office and distance learning (Sharma et al., 2020). Another industry whose consumption was significantly affected by the spread of COVID-19 was the food and beverage industry. For example, beef, which is widely used in restaurants and fast-food services, experienced a significant decline in demand (Telukdarie et al., 2020). The closure of restaurants and cafes led to a reduction in milk consumption, while globally increased demand for snacks and baked goods was observed as people ate more during the day, especially during the lockdown (Fao, 2020). For the same reason, because restaurants and bars can only serve takeaway food, the demand for packaging materials has increased significantly (Liu et al., 2020). According to the above, manufacturing plays a crucial role in an economy. Apparently, among other industries, food producers play an important role in different European countries, including Czechia. However, if the functioning of medium-sized manufacturers is disrupted, that is, due to the COVID-19 pandemic, this would have a negative impact not only within the sector but also in the context of the entire economy. Albeit, what motivates this research first is that a difference in financial performance affects all these macroeconomic issues. For this particular reason, it is desirable to find out how financial performance has changed heterogeneously due to the pandemic. 3. Heterogeneity estimation The empirical part of this paper has been divided into three parts. First, the methodology of the heterogeneous difference-in-differences with cohorts is described, further including a few examples for a better understanding of this innovative technique. Second, data obtained within the estimations are examined descriptively. Third, finally, the results of the heterogeneous impacts of the COVID-19 pandemic on the financial performance of Czech food producers are discussed. 3.1 Difference-in-Differences The theoretical grasp of the hDiD methodology is not difficult in the context of the pandemic. Let us imagine that there are manufacturers whose operating profit has fallen the least compared to the market (i.e. the upper quartile of profitability), or, vice versa, firms whose operating profit has fallen the most compared to the market (i.e. the first quartile of profitability). Wooldridge (2020) describes another great example of this approach. First, imagine there is always a difference between a property situated in the capital and a similar property situated in another town. Then, try to imagine the change of such a difference in price in the post-pandemic period affected by the increasing interest rates on loans or increasing unemployment among those particular regions. Neither just the differences-in-differences technique, but even the heterogeneous DiD with covariates has just been described. In general, the average effect of treatment while investigating cohorts is primarily focused on differences between treated and never-treated (infinite) groups according to eq. (1): 𝑌 𝑖𝑔𝑡 = 𝛽0+ 𝛽1𝑇𝑖𝑔𝑡 + 𝛽2𝑋𝑖𝑔𝑡 +∑𝜏𝑞 𝑄 𝑞=1 𝐷𝑞𝑇𝑖𝑔𝑡 + 𝛼𝑖+ +𝜋𝑡+ 𝜀𝑖𝑔𝑡, (1) where 𝑌 𝑖𝑔𝑡 is the outcome variable, the return on assets (ROA) for individual company i in group g at time t. 𝑇𝑖𝑔𝑡 is a binary treatment indicator, equal to 1 if a company is in group g and treated at time t, and 0 otherwise. 𝑋𝑖𝑔𝑡 is a vector of control variables, the return on sales (ROS) that might affect the outcome. 𝐷𝑞 is a binary indicator for the epidemic cohort q, with q=1,2 in this case. These variables represent the different cohorts described in the next paragraph. 𝜏𝑞 is the treatment effect specific to each cohort. 𝛽0 is the intercept, representing the baseline outcome for the never-treated group within pre-treatment periods. 𝛽1 represents the overall treatment effect for the entire sample. Two-way fixed effects (TWFE) are 𝛼𝑖 for individual fixed effects across the cross-section of the panels, capturing time-invariant individual characteristics that might affect the outcome, and 𝜋𝑡 for time-fixed effects, capturing time-specific factors that might affect the outcome. 𝜀𝑖𝑔𝑡 is the error term. T=5 in this particular case (2018-2022) having two pre-treatment periods 2018 and 2019, two treatment periods, 2020 and 2021, and one post-treatment T. Heryán et al. – Heterogeneity in financial performance caused by the COVID-19 pandemic: Food producers in Czechia 67 2022. Hence, following tests of two assumptions, parallel trends (PT) within the pre-treatment period and no effect in anticipation of the treatment (NA), the twoway fixed effects (TWFE) heterogeneous differencein-differences (hDiD) model with cohort exits and covariates has been deployed according to Wooldridge (2021). Furthermore, the impact of the pandemic has been investigated through the moderating effect of binary variables q, including epidemiological data in the form of cumulative COVID-19 cases per 10 thousand inhabitants within NUTS-3 regions in Czechia. In simple terms, this works either with the most profitable manufacturers from the least affected regions (the first quartile of pandemic impact) or, vice versa, with the least profitable ones from the most affected regions (upper quartile of impact). Either one or the other group, the remaining approximately 75% of the market can thus be described as an infinite group control sample, where two types of differences have been observed (difference-in-differences), not only over time compared to the pre-pandemic period but also differences between two sets of medium-sized companies in Czechia. Heterogeneity arises due to the estimation of differences in 2020 and 2021, for each manufacturer group (cohort) separately, which is crucial. Indeed, between these two years, including the never-treated group 𝑌 𝑡(∞) there may be a total of three different groups g, i.e., the existence of three types of cohorts of affected manufacturers. Either (i) the manufacturers were affected in the same way in both pandemic years or, in terms of the region affected, (ii) only in 2021, when there is only a different start of the observation (staggered case). This tends to be common in medical statistics, which is where the method comes from (we always classify post-operative patients as operated patients). However, in economics, the so-called (iii) exiting cohorts are quite new, in which case it is obvious to detect manufacturers affected by the pandemic in 2020, but whose status changed in the following period and hence have been excluded from the cohort. The current study includes substantial, statistically significant differences against the staggered case implemented in the latest STATA 18, where the estimation of cohort exits has not been implemented. Therefore, such findings are not only significant in the field of corporate finance but also in terms of the development of the method itself. 3.2 Financial and epidemiological data Epidemiological data has been obtained for morbidity, particularly amounts of cumulative COVID-19 cases in 2020 and 2021 per 10,000 inhabitants among NUTS-3 regions in Czechia from Naqvi (2021), the open data source of the Nature Journal. Annual financial data of medium-sized manufacturers from 2018 to 2022 have been obtained from the Orbis financial database of Bureau van Dijk (BvD), a Moody’s analytics company. In particular, total assets, sales, and earnings before interest and taxes (EBIT) to explore both, return on assets (ROA) and return on sales (ROS). According to the BvD definition, a medium-sized manufacturer is a company not exceeding at least two of these three conditions: (i) the company has between 50 and 250 employees; (ii) total turnover is between 8,030 EUR and 32,120 EUR, and (iii) total assets between 4,015 EUR and 16,060 EUR (all in thousands of EUR). The descriptive statistics in the Appendix illustrate interesting findings related to differences among changes of variables described above, exactly between those manufacturers of different NUTS-3 regions with a high concentration of ownership structure and a major owner (67% of companies) against those with dispersed ownership structure (33%), based on the independence indicator reported by BvD. Food producers have been divided into two groups, specifically, one group having a major owner with more than 50% shareholder rights, and the other with a dispersed ownership structure, even from the controlling rights of third parties. Of course, in 2018 and 2019 there were no treatment variables according to the equation above. According to that, we have just one dataset differentiated exactly between these two categories of middle-sized manufacturers. Fig. 3–1 Data distribution of ROA and ROS across NUTS-3 regions, including comparison against a group median (black vertical lines) Nevertheless, foremost, the crucial issue that has to be covered is the exploration of binary variables using ROS as well as the COVID-19 average morbidity. First, the bottom quartile of ROS in Figure 3-1 is created by those companies either having a low level of sales or high costs in 2018-2022. Second, the upper quartile, on the contrary, is created by companies either having a high level of sales or low costs within the estimated period. Then, of course, the binary variables using epidemiological data have affected both groups due to the Ekonomická revue – Central European Review of Economic Issues 26, 2023 68 upper or bottom quartile of the COVID-19 morbidity (highlighted in red or green) within our modeling. That is particularly the way the treatment variables have been constructed only for the pandemic period 2020 and 2021. The exact names of NUTS-3 regions highlighted in red or green are covered by the Appendix. Medians of profitability in both groups of box diagrams have varied around the entire sample medians (black lines). In particular, the ROA median (blue boxes) is 4.78% in the group having the major owner and 4.27% in the group with dispersed ownership. The ROS median (red boxes) is 3.31% and 3.23% in the second group of companies. Surely it does not mean that these particular medians cause between each other anyhow. However, further information is important in Figure 3-1, particularly, which NUTS-3 regions in Czechia belong to those most or least affected by the COVID-19 pandemic measured as the average morbidity per 10,000 inhabitants. All regions have been affected, of course. However, the pandemic situation was different across the country. Hence, precise differences in relations between the profitability ratios of companies from these regions have been estimated through hDiD modeling. Surely, there is no argument supporting the idea related to the differences between ROS and ROA in Figure 3-1. Albeit, the differences in the relation between both profitability ratios concerning the DuPont analysis, caused by the pandemic have been estimated in the next section. 3.3 Discussion on empirical results On one hand side, it is possible to evaluate how the pandemic situation affected the financial performance of those companies with high ownership concentration and a major owner while summarizing new heterogeneity among middle-sized food product manufacturers in Czechia in Figure 3–2. On the other hand, however, the average estimated changes in the financial performance for those manufacturers that have dispersed ownership structures have not been significant at all. That is the main reason, why the comparison is made against the sample with ALL companies. Figure 3–2 illustrates both types of heterogeneous changes in financial performance – negative average treatment effects on those treated companies from the most affected regions by COVID-19 and simultaneously having the lowest profitability, or, vice versa, positive ATET (Average Treatment Effect on Treated) of those manufacturers from the least affected regions in Czechia with the highest profitability – estimated through heterogeneous difference-in-differences (hDiD) models with two-way fixed effects according to Wooldridge (2021). We can see that food producers having a major owner performed better based on positive ATET compared to the whole Czech market. On the other hand, negative ATET suggests deeper negative changes in performance, even though linear trends are violated. Hence, these highly concentrated units with a major owner have been able to perform better even before the pandemic, compared to the rest approximately 75% of food producers in Czechia. Fig. 3–2 Heterogeneous % changes in financial performance among Czech food producers according to their ownership concentration Table 3–1 Heterogeneous DiD with cohorts to investigate changes in financial performance among Czech food producers according to their ownership concentration ALL companies Major owner R O A (DuPont) High R O S vs. Q1 Low R O S vs. Q4 High R O S vs. Q1 Low R O S vs. Q4 A T E T 7.2016 a -4.9356 a 7.7844 a -6.4439 a Linear trend 0.0734 0.0060 0.0888 0.0079 No Anticipation 0.9440 0.6990 0.5260 0.9700 ∑ firms 504 504 309 309 R2 Within 0.3397 0.3316 0.3254 0.3524 R2 Between 0.3909 0.4104 0.4013 0.4315 R2 Overall 0.3712 0.3786 0.3622 0.3869 Note: Symbol a means statistical significance at 0.01 level. Technically, to estimate average treatment on the treated (ATET) in Table 3–1 both assumptions, parallel trend within the pre-treatment period as well as no effect of anticipation have been evaluated. Even though these no anticipation test results are far beyond the rejection threshold of 0.05 in all cases, the other tests incline much closer to rejection of linear trends in both cases with positive ATET, and even rejection of linear trends in both cases with negative ATET. Yet it is not possible to estimate the exact negative changes caused by the pandemic due to this violation of linear trends, those positive changes in financial performance are precisely estimated. T. Heryán et al. – Heterogeneity in financial performance caused by the COVID-19 pandemic: Food producers in Czechia 69 Our findings are crucial. Companies that achieved high operating margins (ROS) in the pre-pandemic period and operated in relatively less affected regions were able to lift their overall performance during the pandemic by more than 7% compared to the rest of the market. Conversely, companies with relatively low prepandemic operating margins relative to the rest of the firms, but which also operated in the most affected regions, reduced their overall performance (ROA) by almost 5% (relative to all firms) and by more than 6% relative to companies with concentrated ownership structures. Thus, the overall dispersion in ROA for the firms analyzed increased significantly driven by the COVID-19 pandemic and, according to our findings, was strongly influenced by the initial level of operating margin. It should be highlighted that in this paper, the effect of the initial level of ROS on the change in ROA induced by the pandemic was analyzed, which turned out to be statistically significant, as well as the effect of a high concentration of ownership structure (majority owner) compared to the rest of the market. However, it is important to note that the effect of ROS on ROA is, according to the DuPont decomposition, also complemented by the level of activity of the firm in question, as measured by asset turnover (the ratio of sales to total assets of the firm). Thus, if there was an increase in ROA for firms with high initial ROS in low-affected regions, this could technically be due to either a further increase in ROS or a higher level of asset turnover. According to Table A-1 in the appendix, it is clear (regions in green) that both of these effects played a role. In several cases there has been detected an increase in EBIT despite a decrease in sales, is caused by a decrease in costs (the state subsidy policy probably played a role in covering personal costs, which affected other revenues, not the costs, nevertheless), surprisingly, the increase in sales was not an exception (which could be justified given the nature of the analyzed firms and will be the subject of further research). In some cases, an increase in asset turnover also seemed to play a role, which, in addition to the increase in sales, could have been caused by a decrease in the balance sheet total, or a combination of both factors. In particular, the decline in assets could have been caused by a reduction in (replacement) investment, a reduction in inventory levels, or the removal of retained earnings. Similarly, for the firms analyzed in regions significantly affected by the COVID-19 pandemic, where the initial low operating margin was shown to have a statistically significant effect on the ROA decline, this decline could have been caused either by a further decline in operating margin or a decline in asset turnover. Again, according to Table A-1, a combination of all these factors can be assumed. Here, however, we need to distinguish between two types of firms with low operating margins: (i) firms that have a business built on low margins, i.e. low prices concerning costs, and (ii) firms where low operating margin indicates poor corporate policy due to a higher level of fixed costs, which leads to a lower efficiency of corporate policy in terms of sustaining price competitiveness. While the first group could theoretically have seen an increase in activity and possibly a positive overall effect on ROA in the pandemic situation, for the second group the pandemic reduced both the return on sales and lifted the turnover of assets. It is clear from the results obtained that the second type of firm is predominant in the food manufacturing sector analyzed. This is driven by the fall in ROA for firms with an initial low ROS. If those firms had low ROS on intention (low margins + high asset turnover), the ROA drop would not necessarily occur (the firm itself is strong, it just focuses on a different type of customer). If the initial ROS was a sign of a poor corporate policy there is a decline in ROA (the second group). The results of this study show that the COVID-19 pandemic had an impact on the financial performance of Czech firms in the food industry, and we find a significant increase in the financial performance heterogeneity in the dataset used over the period. While a statistically significant effect of operating margin on the change in ROA in differently affected regions has been demonstrated, other effects need to be further investigated. However, the significant impact of pandemics on food producers is also evident from the literature review (e.g. Juergensen et al, 2020; Seetharaman, 2020; Telukdarie et al. al, 2020; FAO, 2020; Liu et. al, 2021), however, the intensity of the impact could vary according to the type of production and its ability to respond to sudden changes. For example, food firms using outsourcing services may have a competitive advantage and potential cost reductions in business processes (Hsiao, 2009). This is also consistent with our findings. If the main objective was to assess the impact of COVID-19 on the financial performance of firms in each NUTS-3 region, a sub-objective of equal importance was to examine the role of ownership concentration in the impact of the pandemic crisis on firm performance. This hypothesis is supported by the work of other authors who have confirmed the positive impact of ownership concentration on financial performance or firm value (Alimehmeti and Paletta, 2012; Nguyen et al., 2015; Machek and Kubicek, 2018; Alkurdi, 2021), where higher concentration allows for a better alignment of owners' and shareholders' interests, which promotes firm value growth. Similarly, Song et. al. al. (2021) report that unpredictable managerial situations during the COVID-19 crisis, combined with high levels of managerial and institutional ownership, may have acted as credible indicators for owners (shareholders). However, their study did not confirm the relationship between ownership and stock returns. A similar conclusion was also reached by Laporšek et al. (2020), whose Ekonomická revue – Central European Review of Economic Issues 26, 2023 70 study also failed to confirm a statistically significant relationship between ownership concentration and financial performance. In our study, the statistical significance of the effect of a highly concentrated ownership structure on the change in performance in the different affected regions (with different effects in different affected regions) was confirmed based on the hDiD method, but the effect of a fragmented ownership structure was not found to be statistically significant. One of the main limitations of this paper is the specific focus on two selected ratios, ROS and ROA, and their interconnectedness in the aftermath of the pandemic crisis. We realize that the impact of COVID-19 on financial performance is analyzed in this paper only based on these two ratios, without a deeper analysis of the possible related effects and impacts. In our study, using the TWFE hDiD technique, we have demonstrated the impact of the crisis on the selected financial indicators concerning the extent to which the NUTS-3 region was affected by the COVID-19 pandemic, however, we see great potential in the future by focusing on sub-indicators through which we can draw more detailed results and conclusions. 4. Concluding remarks This paper aimed to estimate differences in financial performance between Czech middle-sized food product manufacturers that have a major owner, particularly caused by the impacts of the COVID-19 pandemic. Thus, the findings in this paper revealed the heterogeneous effects of the pandemic on financial performance among middle-sized food producers in Czechia, especially when it comes to ownership concentration. In particular, the notable differences that appeared between highly concentrated ownership versus the whole market indicate that the ownership structure influences the ability of these businesses to withstand financial shocks during the pandemic. Significant relationships between the return on assets and the return on sales, including earnings before interest and taxes, were found using the DuPont framework. Furthermore, the association between the increase in heterogeneity in financial performance and inflation in the following years suggests that the pandemic would have caused long-term changes to the financial landscape of the Czech food manufacturing industry. These findings have implications for corporate finance policies since they make it clear how important it is for companies to take ownership concentration into account and how it could affect their ability to perform financially in difficult times. Research methodologies are innovative according to the combination of financial and epidemiological data and the application of the heterogeneous difference-in-differences methodology with cohorts of companies. The results suggest that the pandemic has caused long-term changes in the financial system, potentially exacerbating inflationary pressures in Czechia. However, the results cannot be generalized, as only Czech food producers have been investigated. This paper also has several limitations. Without taking into account other possible factors that could affect financial performance, the examination of the impact of COVID19 on financial performance is based partially on the DuPont framework, particularly the association between return on assets (ROA) and return on sales (ROS). Although the results suggested that the epidemic has caused long-term changes in the financial system, potentially exacerbating inflationary pressures, the entire manufacturing industry must be investigated to estimate where the changes were at the highest level. Hence, a few other sub-industries among other European countries, as well as the role of liquidity and costs of debt financing, should be investigated in future research. References ABDALLAH, A. A. (2021). How Can Lean Manufacturing Lead the Manufacturing Sector during Health Pandemics Such as COVID-19: A Multi Response Optimization Framework. Computers, Materials & Continua 66(2): 1397–1410. https://doi.org/10.32604/cmc.2020.013733. ALIMEHMETI, G.; PALETTA, A. (2012). Ownership concentration and effects over firm performance: Evidences from Italy. European Scientific Journal, 22(8), 39 – 49. ALKURDI, A., HAMAD, A., THNEIBAT, H., ELMARZOUKY, M. (2021). Ownership structure’s effect on financial performance: An empirical analysis of Jordanian listed firms. Cogent Business & Management 8(1), https://doi.org/10.1080/23311975.2021.1939930. BAKER, S., BLOOM, N., DAVIS, S., TERRY, S. (2020). COVID-Induced Economic Uncertainty. NBER Working Paper Series, w26983. Cambridge, MA: National Bureau of Economic Research. https://doi.org/10.3386/w26983. BASILAIA, G., KVAVADZE, D. (2020). Transition to Online Education in Schools during a SARS-CoV-2 Coronavirus (COVID-19) Pandemic in Georgia. Pedagogical Research 5(4). https://doi.org/10.29333/pr/7937. BOLISANI, E., SCARSO, E., IPSEN, C., KIRCHNER, K., HANSEN, J. P. (2020). Working from Home during COVID-19 Pandemic: Lessons Learned and Issues. Management & Marketing: Chall Knowl Soc 15: 458– 476. https://doi.org/10.2478/mmcks-2020-0027. T. Heryán et al. – Heterogeneity in financial performance caused by the COVID-19 pandemic: Food producers in Czechia 71 CARLETTI, E., OLIVIERO, T., PAGANO, M., PELIZZON, L., SUBRAHMANYAM, M. G. (2020). The COVID-19 Shock and Equity Shortfall: FirmLevel Evidence from Italy. The Review of Corporate Finance Studies 9(3): 534–568. https://doi.org/10.1093/rcfs/cfaa014. CHO, S.-W. (2020). Quantifying the impact of nonpharmaceutical interventions during the COVID-19 outbreak: The case of Sweden. The Econometrics Journal 23(3): 323–344, https://doi.org/10.1093/ectj/utaa025. CLAMPIT, J., HASIJA, D., DUGAN, M., GAMBLE, J. (2021). The Effect of Risk, R&D Intensity, Liquidity, and Inventory on Firm Performance during COVID-19: Evidence from US Manufacturing Industry. Journal of Risk and Financial Management 14: 499. https://doi.org/10.3390/jrfm14100499. DAVE, D., FRIEDSON, A. I., MATSUZAWA, K., SABIA, J. J. (2021). When do shelter-in-place orders fight COVID-19 best? Policy heterogeneity across states and adoption time. Economic Inquiry 59(1): 29– 52. https://doi.org/10.1111/ecin.12944. DEVAKUMAR, D., SHANNON, G., BHOPAL, S. S., ABUBAKAR, I. (2020). Racism and Discrimination in COVID-19 Responses. The Lancet 395(10231): 1194. https://doi.org/10.1016/S0140-6736(20)30792-3. DIAZ-ELSAYED, N., MORRIS, K. C., SCHOOP, J. (2020). Realizing Environmentally Conscious Manufacturing in the Post-COVID-19 Era. Smart and Sustainable Manufacturing Systems 4(3): 20200052. https://doi.org/10.1520/SSMS20200052. EUROSTAT (2023a). Gross value added and income by A*10 industry breakdowns. [Online], accessed from: <https://ec.europa.eu/eurostat/databrowser/view/nama _10_a10__custom_8713003/default/table?lang=en>. EUROSTAT (2023b). Industrial production statistics. [Online], accessed from: <https://ec.europa.eu/eurostat/statisticsexplained/index.php?title=Industrial_production_statistics#Industrial_production_by_sector>. FAO. (2020). COVID-19 and Smallholder Producers Access to Markets. Rome. http://www.fao.org/familyfarming/detail/en/c/1272443/. FERNÁNDEZ, R., PALAZUELOS, E. (2018). Measuring the role of manufacturing in the productivity growth of the European economies (1993-2007). Structural Change and Economic Dynamics 46: 1–12. https://doi.org/10.1016/j.strueco.2018.03.003. FLAXMAN, S., MISHRA, S., GANDY, A., et al. (2020). Estimating the number of infections and the impact of non-pharmaceutical interventions on COVID19 in 11 European countries. Imperial College London (March), 1–35. https://doi.org/10.48550/arXiv.2004.11342. GARLICK, C., MCMILLAN, M., PETERSON, R., SCHEUERMANN, S., AWWAD, M. (2020). Case Study Review of the Effects of COVID-19 on the Supply Chain of Manufacturing Companies in California. In: Proceedings of the 5th NA International Conference on Industrial Engineering and Operations Management. Detroit: IEOM Society International, 423 – 432. HAAPALA, K. R., KIM, K.-Y., OKUDAN, G. E., KREMER, R. K., SHILKROT, R., SCIAMMARELLA, F. M. (2020). An Open Online Product Marketplace to Overcome Supply and Demand Chain Inefficiencies in Times of Crisis. Smart and Sustainable Manufacturing Systems 4(3): 20200055. https://doi.org/10.1520/SSMS20200055. HARAGUCHI, N. (2016). The importance of manufacturing in economic development: Has this changed? Inclusive and Sustainable Industrial Development Working Paper Series, No. 1/2016. Vienna: UNIDO. https://www.unido.org/sites/default/files/201702/the_importance_of_manufacturing_in_economic_development_0.pdf. HILMOLA, O.-P., LÄHDEAHO, O., HENTTU, V., HILLETOFTH, P. (2020). COVID-19 Pandemic: Early Implications for North European Manufacturing and Logistics. Sustainability 12(20): 8315. https://doi.org/10.3390/su12208315. HSIANG, S., ALLEN, D., ANNAN-PHAN, S., et al. (2020). The effect of large-scale anti-contagion policies on the COVID-19 pandemic. Nature 584, 262–267 (2020). https://doi.org/10.1038/s41586-020-2404-8. HSIAO, H.-I. (2009). Logistics outsourcing in the food processing industry. A study in the Netherlands and Taiwan. https://edepot.wur.nl/11438. JUERGENSEN, J., GUIMÓN, J., NARULA, R. (2020). European SMEs amidst the COVID-19 Crisis: Assessing Impact and Policy Responses. Journal of Industrial and Business Economics 47(3): 499–510. https://doi.org/10.1007/s40812-020-00169-4. KANDA, W., KIVIMAA, P. (2020). What Opportunities Could the COVID-19 Outbreak Offer for Sustainability Transitions Research on Electricity and Mobility? Energy Research & Social Science 68: 101666. https://doi.org/10.1016/j.erss.2020.101666. KURITA, J., LIMOUSIN, M., FERREIRA, N., OZUNA, J. (2020). CFD Analysis on Air Ventilation at a Manufacturing Plant as a Tool for Designing Machine Layout, a Case Study. In: Proceedings of the 5th NA International Conference on Industrial Engineering and Operations Management. Detroit: IEOM Society International, 2412 – 2419. LAPORŠEK, S., DOLENC, P., GRUM, A., STUBELJ, I. (2021). Ownership structure and firm performance –