Asymmetric Cost Behavior Across Life Cycle Stages

The traditional model of cost behavior has been criticized for its symmetric cost behavior assumption. A new model has been proposed assuming that costs respond differently to upward and downward activity changes. The main objectives of this paper are to investigate the existence, degree, and nature of asymmetric cost behavior (ACB) phenomenon and examine how the organization life cycle (OLC) affects this phenomenon in the context of Egypt. The current study achieves these objectives by employing multiple regression to explore the behavior of cost of goods sold (COGS), selling, general and administrative cost (SGA), and total cost (TC) for 1577 firm-year observations (99 manufacturing firms) during the period from 2000 to 2019. The results demonstrate that all three cost proxies (COGS, SGA, and TC) are sticky with the highest degree of stickiness to TC. In addition, OLC is a conditional factor that affects how costs behave in response to change in activity level. Consistent with theoretical propositions, both COGS and TC exhibit anti-stickiness behavior for firms in the introduction stage and stickiness behavior for firms in the growth, mature, and shakeout/decline stages. However, SGA is only sticky for firms in the mature stage. However, the hypotheses related to asymmetric behavior of SGA were rejected for firms in the introduction, growth, and shakeout/decline stages.


Introduction
Cost behavior is considered one of the most significant analyses of the decision-making process. The traditional model of cost behavior assumes that costs change symmetrically to changes in activity level. However, numerous cost behavior studies provide robust empirical evidence of asymmetric cost behavior (ACB); specifically, costs behave differently to upward and downward changes in cost driver level. Anderson, Banker, & Janakiraman (2003;ABJ hereafter) provide empirical evidence that selling, general and administrative cost (SGA) decreases less when revenues decrease than they increase when revenues increase by an equivalent percentage. They labeled this new phenomenon as "cost stickiness". Costs are sticky if they decrease less as output level falls than they increase as output level rises by an equivalent percentage (Balakrishnan, Labro, & Soderstrom, 2014;Yao, 2018). Other studies prove that costs are anti-sticky in that they decrease more as output level falls than they increase as output level rises by an equivalent percentage (Kama & Weiss, 2013;Weiss, 2010). Both scenarios represent the forms of ACB.
Given that costs behave asymmetrically relying on the traditional cost behavior model causes information distortion even when employing more advanced accounting practices. For example, Noreen (1991) shows that activity-based costing is relevant to allocate costs only when costs change in direct proportion to activity level. ACB model represents a strategic behavior model (Balakrishnan & Gruca, 2008). It reflects both the effect of change in activity level during the current period and the managerial assessment of the past and expected changes in demand.
Literature provides several factors affecting the nature and degree of ACB including, but not limited to, the existence of adjustments costs (Cannon, 2014;Yasukata, 2011), economic growth (Ibrahim, 2015), optimism of managers about expected demand (Yao, 2018), Empire-building incentive (Chen, Lu, & Sougiannis, 2012), the incentive to meet earning targets (Kama & Weiss, 2013), and corporate governance (Ibrahim & Ezat, 2017). On this ground, ACB occurs due to several internal and external determining factors that could be expressed by the organization life cycle (OLC) stages. OLC reflects an organization's development resulting from changes in both internal and external environments (Vorst & Yohn, 2018).
According to the life cycle theory, organizations are just like living organisms in that they go through several anticipated configuration phases of development (Kiani, Aghaee, & Etemadi, 2018;Miller & Friesen, 1984). In this way, the life cycle framework provides managers with guidelines and directions, helping them in decision-making.
One of the most popular OLC models is to group firms, depending on their environmental context, strategy, structure, and decision-making methods, into five primary stages: introduction, growth, maturity, shakeout, and decline (Miller & Friesen, 1984). While firms at the same stage have common characteristics, each stage has considerably different contexts that discriminate it from other stages (Su, Baird, & Schoch, 2015;Vorst & Yohn, 2018). These persistent differences between organizations in different life cycle stages propose that the nature and degree of asymmetric cost behavior can be modeled as a function of an OLC. The current study extends cost behavior literature by investigating the undiscovered relationship between OLC and ACB.
The contribution of this study is threefold: First, prior literature demonstrates that asymmetric cost behavior is affected by several factors such as firm size (Dalla Via & Perego, 2014), ability to access the capital market (Cheng, Jiang, & Zeng, 2018), managerial optimism (ABJ, 2003), and corporate governance (Ibrahim & Ezat, 2017), among others. Varying in these factors across life cycle stages indicates that understanding the effect of OLC on ACB (still undiscovered relationship) offers additional insight on the determinants of such phenomenon. Second, most cost stickiness studies focused mainly on the asymmetric behavior of SGA (e.g., Alavinasab, Mehrabanpour, & Ahmadi, 2017;ABJ, 2003;He, Teruya, & Shimizu, 2010). The current study extends those studies by investigating the stickiness behavior of other costs such as COGS, which represents a large percentage of the cost structure in manufacturing firms, and TC. Third, most cost stickiness studies are conducted in developed countries (e.g., Chen et al., 2012;Weidenmier & Subramaniam, 2003), which leaves a gap to find out how costs behave and what the determinants of such behavior are in less developed countries such as Egypt, which has a different context, especially to generalize the initial results of the ABJ (2003) study and its subsequent studies.
The rest of this study is arranged as follows: Section 2 analyzes the literature review to develop the study hypotheses. Section 3 explains the research methodology by showing the study samples, the empirical model specification, and the statistical techniques employed. Section 4 reports the study's statistical results. Section 5 presents the combined discussion. Section 6 demonstrates conclusions, implications, and limitations.

Literature Review & Hypotheses Development
The current study has two main objectives: The first is to explore whether COGS, SGA, or TC behaves asymmetrically in response to change in activity level. The second is to investigate how the nature and degree of ACB vary across the OLC stages.

The Existence of ACB
ABJ claim that the main reason for ACB is -the deliberate managerial decision‖ that refers to managers' interventions in a way that affects the cost responsiveness patterns to change in output level. Managers respond differently to demand decreases and demand increases. When demand increases beyond the current capacity level, managers usually increase the level of resources and subsequently the cost, but when demand decreases, managers usually hesitate to cut slack resources (ABJ, 2003). Two main groups of arguments are introduced as the primary interpretation for this asymmetry in managers' response and, therefore, the existence of ACB.
Grounded on the previous presentation, it is suggested that the degree of decrease in cost is less than the corresponding degree of increase in costs when sales changes by an equivalent percentage. Therefore, the following hypotheses are considered: H1a. COGS demonstrates sticky behavior.

The Relationship Between OLC and ACB
According to the life cycle theory, organizations evolve in predictable developmental stages, International Journal of Accounting and Financial Reporting ISSN 2162-3082 2021 each reflecting a different context concerning resources, capabilities, competencies, strategic orientations, organizational structures, and operating environments (Dickinson, 2011;Kiani et al., 2018;Miller & Friesen, 1984). The core of OLC theory suggests that managerial decisions and organizational performance are considerably affected by the contexts change across life cycle stages. Numerous studies recognize that OLC explains significantly variation in several accounting variables, such as cost of equity (e.g., Hasan, Hossain, & Habib, 2015), capital structure decisions (e.g., La Rocca, La Rocca, & Cariola, 2011), and profitability (e.g., Dickinson, 2011). This suggests that firms that belong to different stages have different characteristics that may influence the nature and degree of ACB.
Introduction-stage firms are described as relatively small and young firms, with simple structure and systems (Miller & Friesen, 1984), face a highly uncertain ambiguous environment (Jirá sek & Bí lek, 2018), and have no control over its external environment (Hasan et al., 2015). In addition, young and small firms suffer from their limited ability to reach the public markets (Berger & Udell, 1998). Regarding ACB, literature provides empirical evidence that small firms show anti-stickiness cost behavior (Dalla Via & Perego, 2014). Besides, the costs of firms with limited access to capital are more likely to be anti-sticky due to decreasing the downward adjustment costs (Cheng et al., 2018). Following the previous arguments, the following hypotheses are considered:

H2a: COGS of introduction firms shows anti-stickiness cost behavior.
H2b: SGA of introduction firms shows anti-stickiness cost behavior.

H2c: TC of introduction firms shows anti-stickiness cost behavior.
Growth-stage firms are characterized by several characteristics, including increased size, where structure becomes less centralized, departmental, and more complex (Miller & Friesen, 1984). This suggests that growth firms utilize more assets and hire more employees than those firms in the introduction stage, which in returns increases both assets and employee intensity and consequently have higher adjustment costs. In addition, firms in the growth stage have a higher competitive advantage (Kazanjian, 1988) as they have already built their unique capabilities and competencies (Hatane, Gabrielle & Angelina, 2019), where demand is getting increased in a way that exceeds supply (Jawahar & McLaughlin, 2001), these features increase the managerial optimism about future demand. Following the previous arguments, the following hypotheses are considered: H3a: TC of growth firms shows the highest degree of cost stickiness across life cycle stages.
H3b: COGS of growth firms shows the highest degree of cost stickiness across life cycle stages.
H3c: SGA of growth firms shows the highest degree of cost stickiness across life cycle stages.
Mature-stage firms show several features such as stable demand levels (Adizes, 1979), where efficiency substitutes innovation. Consequently, they have narrower product scope compared to the growth stage (Su, Baird & Schoch, 2015). Firms in the mature stage are less proactive (Koberg, Uhlenbruck & Sarason, 1996), focusing on exploiting the existing opportunities rather than exploring new ones (Dufour, Steane & Corriveau, 2018), and allocating more of their resources to corporate social responsibility activities (Hsu, 2018).
These characteristics suggest that the costs of mature firms are stickier than those of introduction firms but less sticky than those of growth firms as mature firms have relatively high adjustment costs and have higher incentives to beat earnings targets and hence reduce costs stickiness. Following the previous arguments, the following hypotheses are considered: H3a: TC of mature firms shows a high degree of stickiness across OLC stages.
H3b: COGS of mature firms shows a high degree of stickiness across OLC stages.
H3c: SGA of mature firms shows a high degree of stickiness across OLC stages.
Organizations in the shakeout/decline stage demonstrate inconsistent characteristics depending on which stage they were before moving to the current stage. If they moved from the growth or mature stage, then they are the largest, facing the highest level of competition (Hatane et al., 2019) and giving more attention to innovation and diversification in both products and markets (Su et al., 2015), but focusing only on significant products and markets (Jirá sek & Bí lek, 2018). However, if they moved from the introduction stage, they may suffer from poor performance and the least innovative activities (Miller & Friesen, 1984), indicating negative earnings per share, return on net operating assets, and profit margin (Dickinson, 2011). These characteristics imply a contradictory effect on cost stickiness as increasing size and innovation orientation positively affect the cost stickiness due to an increase in adjustment cost; however, the aggressive competition and poor performance may provoke the need to cut cost rapidly, showing a negative effect on cost stickiness. Following the previous arguments, the following hypotheses are considered: H4a: COGS of shakeout/decline firms shows the lowest degree of stickiness across OLC stages.
H4b: SGA of shakeout/decline firms shows the lowest degree of stickiness across OLC stages.
H4c: TC of shakeout/decline firms shows the lowest degree of stickiness across OLC stages.

Sample and Data Collection
The main purpose of this study is to investigate the asymmetric cost behavior phenomena in the context of Egypt. Since cost behavior is more pronounced in manufacturing firms (Dierynck et al., 2012), and due to the homogenous structure of the income statement among these firms, we follow Weiss (2010) in restricting the sample of the current study to only manufacturing firms through the period of 2000-2019. The primary financial data used in our ACB estimation include sales revenues (REV), cost of goods sold (COGS), selling, general & administrative cost (SGA), and total cost (TC). All data are extracted from annual reports published on Thomson Reuter DataStream, Egypt. TC is calculated as sales revenues minus income before tax. To mitigate the negative effect of outliers, we winsorized variables at 95%. ISSN 2162-3082 2021 To provide some level of homogenous and minimize the negative effect of outliers, we exclude from the sample firm-year observations with (1) sales revenue less than EGP 10 million and (2) negative total costs, i.e., when income before tax is higher than sales revenues. These procedures of sample selection result in a sample of 1,577firm-year observations from 99 firms. Table 1 shows the sample distribution according to years, listing state, life cycle stages, and sectors.

Life Cycle Stage Classification
Several proxies have been employed to capture the stage in which a firm is such as size, retained earnings, age, assets growth, and sales growth (e.g., DeAngelo, DeAngelo, & Stulz, 2010;Faff, Kwok, Podolski, & Wong, 2016;Owen & Yawson, 2010). Among the limitations addressed to these proxies is that they don't evolve monotonically across life cycle stages. This means that the same level of the proxy variable may give two different classifications (Faff et al., 2016). Dickinson (2011) introduces a new proxy that captures the cyclical nature of OLC by employing signs of the three cash flows components (operating cash flow =OCF, investing cash flow =ICF, financing cash flow = FCF). Table 2 shows how firms are classified into one of five life cycle stages.

Asymmetric Cost Behavior and Life Cycle Effect
The most common model for estimating ACB that is employed by the majority of cost stickiness literature is the model of ABJ (2003), which depends on an interaction dummy variable that distinguishes between activity-increasing periods and activity-decreasing periods to capture cost stickiness as follows: Where; i represents the company i; t represents the year t; Cost represents SGA; Sales represent sales revenues; and DEC is a dummy variable that equales1 if Sales i t < Sales i t and 0 otherwise. Log specification has been used to enhance the comparability and also accommodates economic interpretation of the estimated coefficients. ISSN 2162-3082 2021 Since a DEC variable takes the value of 0 when sales increase, the coefficient β 1 estimates the increasing percentage in costs resulting from a 1% increase in sales revenue, while the sum of coefficients (β 1 + β 2 ) estimates the decrease percentage in costs responding to a 1% decrease in sales revenue. This means that the coefficient β 2 illustrates the average degree of cost stickiness by capturing the degree of cost response relating to sales decreases versus increases. Statistically, cost stickiness (anti-stickiness) is proved when there is a significant negative (positive) coefficient β 2 conditional on a positive coefficient β 1 .

International Journal of Accounting and Financial Reporting
To fulfill the first objective of this study, related to investigate the nature and degree of ACB, we replicate the pioneer model of ABJ (2003) for three proxies of costs, including COGS, SGA, and TC, as follows:

Model (2):
(2) Model (3): Where COGS i,t , SGA i,t and TC i,t refer respectively to cost of goods sold, the selling, general and administrative cost, and the total cost for the firm i at year t, while REVi,t stands for the sales revenue for the firm i at the time t. DEC is a dummy variable that equals 1 if the sale revenue of the current period is lower than the previous period value and 0 otherwise.
This study reruns the pre-mentioned three models separately for firm-year observations in each life cycle stage to explore the effect of the life cycle stage on the nature and the degree of ACB.  Panel C of Table 2 shows the percentage of firm-year observations (frequency of firm-years) when revenues and costs variables decrease in the current period relative to the prior period. The frequency percentage of the firm-year observations when costs fell (from 23.34% to 25.17%) is relatively less than when revenues fell (26.51%). Also, except for total cost, which has a mean value of decrease of about 16.7%, the mean value of reductions in revenue (16.59%) is relatively higher than that of reductions in costs 15.64% and 15.76% for COGS and SGA, respectively, which may provide a sing for the existence of cost stickiness.

Hypotheses Testing Results
Following prior studies in asymmetric cost behavior, we employ Ordinary Least Squares (OLS) to estimate the cost stickiness model for the entire sample first, then we test the effect of the life cycle on the nature and magnitude of asymmetric cost behavior. The following section shows the regression results regarding these two groups of tests. Table 4. Results of regressing annual changes in costs on annual changes in sales revenue

Coef. t-statistic Coef. t-statistic Coef. t-statistic
Using Panel Least Squares for the pooled sample (Note 1). Table 4 demonstrates the results of regressing the annual changes in each of COGS, SGA, and TC on the annual change of sales revenues.
Concerning the COGS model, the estimated value of β 1 is 0.566, implying that COGS increases by approximately 0.57% for each 1% increase in sales revenues. The estimated value of β 2 is negative at -0.142, and the sum of estimated coefficients β 1 + β 2 is 0.424, implying that the COGS decreases only by approximately 0.42% for each 1% decrease in sales revenues which reveals that COGS exhibits stickiness behavior.
Regarding the SGA model, the estimated value of β 1 is 0.246, implying that SGA increases by approximately 0.25% for each 1% increase in sales revenues. The estimated value of β 2 is negative at -0.098, and the sum of estimated coefficients β 1 + β 2 is 0.148, implying that the SGA decreases by approximately 0.15% for each 1% decrease in sales revenues. This means that SGA exhibits stickiness behavior.
Regarding the TC model, the estimated value of β 1 is 0.537 (t-statistic = 34.351), implying that TC increases by approximately 0.54% for each 1% increase in sales revenues. The estimated value of β 2 is negative and equals -0.145 (t-statistic = -5.237), and the sum of estimated coefficients β 1 + β 2 is 0.392, implying that the TC decreases by approximately 0.39% for each 1% decrease in sales revenues. This means that TC exhibits stickiness behavior.
We compute the relative percent decrease to increase RPD in cost to compare the degree of stickiness among the three costs proxies. A lower value of RPD reflects a higher degree of stickiness. SGA has the highest degree of costs stickiness with RPD at 0.60 (0.15/0.25), followed by TC with RPD at 0.72(0.39/.54), and finally, COGS has the lowest degree of costs stickiness with RPD at 0.74 (0.42/0.57). ISSN 2162-3082 2021 Table 5 shows the results of running the three regression models separately for each group of firm-year observations belonging to a specific life cycle stage.

The Effect of OLC on ACB
For introduction firms, the coefficient β 1 in the COGS (TC) model is 0.39 (0.28) indicating that COGS (TC) increase by approximately 0.39% (0.28%) for each 1% increase in sales revenues. The coefficient β 2 is positive at 0.58 (0.70), and the sum of coefficients β 1 + β 2 is 0.97 (0.98), implying that the COGS (TC) of introduction firms decreases by approximately 0.97% (0.98%) for each 1% decrease in sales revenues. This means both COGS and TC demonstrate anti-stickiness cost behavior for introduction-stage firms. Both β 1 and β 2 are significant at a 1% level.
For growth firms, the coefficient β 1 in the COGS (TC) model is 0.38 (0.37) indicating that COGS (TC) increase by approximately 0.38% (0.39%) for each 1% increase in sales revenues. The coefficient β 2 is negative at -0.18 (-0.17), and the sum of coefficients β 1 + β 2 is 0.20 (0.20), implying that the COGS (TC) of growth firms decrease by approximately 0.20% (0.20%) for each 1% decrease in sales revenues. This means both COGS and TC demonstrate stickiness cost behavior in mature-stage firms. Both β 1 and β 2 are significant at a 1% level.
For mature firms, the coefficient β 1 in the COGS, SGA, and TC models indicate that these cost proxies increase by approximately 0.73%, 0.33%, and 0.71 respectively for each 1% increase in sales revenues. The coefficient β 2 is negative for all costs. Both β 1 and β 2 are significant for the three models at a 1% level. Given the sum of coefficients, β 1 + β 2 of three models indicate that the COGS, SGA, and TC of mature firms decrease only by approximately 0.51%, 0.06%, and 0.46 respectively, for each 1% decrease in sales revenues. This means all proxies of costs demonstrate stickiness cost behavior in mature-stage firms. The insignificance of β 2 conditional on the significance of β 1 in the SGA model indicates no significant difference between the increasing percentage and decrease percentage of SGA (i.e., SGA behaves symmetrically) for firms in the introduction, growth, and shakeout/decline stages.
To compare the degree of COGS and TC across life cycle stages, RPD is computed. The RPD column in Table 5 indicates that both COGS and TC have their highest (lowest) degree of cost stickiness for firms in the growth (shakeout/decline) stage at a value of RPD at 0.53 (0.77) for COGS and 0.54(0.72) for TC. Mature firms show a moderate degree of costs stickiness with a value of RPD at 0.70 for COGS and 0.69 for TC.

Robustness Test
To confirm the previous results related to the effect of OLC on ACB, instead of running separate regression, we incorporate dummy variables into the ABJ model to capture. Three dummy variables are included, namely GRTH, MATUR, and SHKDEC taking the introduction stage as a benchmark. Also, to estimate both the increase and decrease percentage relative to one percent change in sales revenue, each stage dummy variable is introduced twice, with and without the decrease indicator. The general model after incorporating these dummy variables of is presented as follows: International Journal of Accounting and Financial Reporting ISSN 2162-3082 2021

OLC Model
(4) Where, COST i,t reflects each of COGS i,t (for model 1), SGA i,t (for model 2), and TC i,t (for model 3). GRTH is a dummy variable that takes the value of 1 if the firm (i) at year (t) is in the growth stage and 0 otherwise. MATURE is a dummy variable that takes the value of 1 if the firm (i) at year (t) is in the mature stage and 0 otherwise. SHKDEC is a dummy variable that takes the value of 1 if the firm (i) at year (t) is in the shakeout/decline stage and 0 otherwise. Table 6 shows the results of the regression of the previous model.
Since the introduction stage is the benchmark for the regression model, the coefficient β 1 captures the estimated increase percentage in the investigated cost for a 1% increase in sales level. The sum of coefficients β 1 + β 2 captures the estimated decrease percentage in the investigated cost for a 1% decrease in sales level for firms in the introduction stage.
The previous interpretation is valid only when the coefficients are significant; however, they should be interpreted differently when the coefficients are insignificant. For example, the insignificant coefficients β 2 and β 8 in the SGA model may be interpreted as there is no significant difference between the increase and decrease percentage of SGA (i.e., SGA behaves symmetrically) for firms in the introduction and shakeout/decline stages. The insignificant coefficient β 3 in the SGA and TC models may be interpreted as there is no significant difference between the growth firms and introduction firms regarding the increasing percentage of SGA and TC. Similarly, the insignificant coefficient β 5 in the SGA model may be interpreted as there is no significant difference between mature firms and introduction firms concerning the increasing percentage of SGA. ISSN 2162-3082 2021  Prob(F-statistic) 0.000 0.000 0.000 ***, **, *significant at 1%, 5%, and 10% levels, respectively

International Journal of Accounting and Financial Reporting
Based on the previous presentation, Table 7 shows the increase (INC) and decrease (DEC) percentages of change in three cost categories for a 1% increase and decrease in sales level, and also the relative percent of reduction to increase (RPD) as follow. The RPD column of Table 7 indicates that COGS and TC are anti-sticky in the introduction stage but sticky in all other stages (growth, mature, and shakeout/decline stages), with the highest (lowest) degree of stickiness in growth (shakeout/decline) stage. In contrast, SGA costs are sticky only in the growth and mature stages but behave symmetrically in the introduction and shakeout/decline stages. The results are so close to those of separate regressions run before, which confirm the previous results.

Discussion
A considerable body of literature reports the existence of ACB. The current study attempts to extend the literature by investigating the ACB in the Egyptian context. The study examines the ACB at two levels of analysis: First, the study explores the nature and degree of cost stickiness using several proxies of costs. The results demonstrate that all the investigated costs (COGS, SGA, and TC) exhibit stickiness behavior; specifically, they decrease less than they increase when the activity level changes by an equivalent percentage. However, the degree of cost stickiness is different across these different cost accounts, which confirms the expectations of . Comparing the degree of stickiness among the three costs examined shows that TC exhibits the highest stickiness degree, while COGS shows the lowest degree. This result may be attributed to several reasons. One reason is that the cost structure of COGS comprises more variable costs compared to SGA. In addition, direct material, which constitutes a high percentage of COGS, is proved to be anti-sticky (Ghaemi, & Nematollahi, 2011).
Second, the study examines the effect of OLC on ACB. Despite the stickiness behavior of COCS, SGA, and TC, ACB differs across the different stages of OLC. Consistent with our conjectures, costs tend to be anti-sticky for firms in the introduction stage but sticky for firms in the other stages. Two main justifications may be discussed here; the difference in both adjustment costs and managerial optimism. As introduction firms are relatively small, young, and have simple structure and systems (Miller & Friesen, 1984), adjustment costs are more likely to be lower than those of other stages firms.
Moreover, the high ambiguity about the environment negatively affects managerial optimism, which correlates directly with cost stickiness (ABJ, 2003). Subsequently, it encourages managers to cut costs rapidly when the activity level falls. In addition, firms in the introduction stage may have a higher incentive to avoid losses, and costs tend to be ant-sticky when managers have higher incentives to avoid losses (Kama & Weiss, 2013).

Conclusion
In this study, we have investigated the nature and degree of ACB and how life cycle stages may affect it. We argue that costs exhibit stickiness behavior in general; however, this behavior is affected by the firm's life cycle stage. With regard to the first group of hypotheses, the results show that all three costs proxies are sticky and therefore accepting each of H1a, H1b, and H1c. With regard to the second group of hypotheses, only H2a and H2c are accepted, while H2b is rejected. Regarding the third group of hypotheses, all hypotheses are accepted. Finally, both H4a and H4c are accepted, but H4b is rejected. Table 8 summarizes the results of hypotheses testing. ISSN 2162-3082 2021