4.1. The Impact of Organizational Capital on Cost Stickiness
To examine the relationship between OC and cost stickiness, we first calculated OC from SG&A expenses (
Lev et al., 2009). To address potential endogeneity concerns, particularly reverse causality, firms were classified into high- and low-OC groups based on the median value of lagged OC (t − 1). The median-split approach follows
Venieris et al. (
2015) (see
Section 3.2). Specifically, firms with lagged OC values above the sample median were assigned to the high-OC group, while those below the median were assigned to the low-OC group. Cost stickiness for each group was subsequently estimated and compared based on
Anderson et al. (
2003).
Before conducting the estimations, we performed the RESET test to verify the linearity of the pooled model specification (see
Table A3). Whereas the test indicated potential misspecification in the full-sample regressions, the results for subsample analyses did not reject the linear specification. Overall, these findings suggest that the linearity assumption is broadly reasonable for the purposes of our analysis. The estimations employ the basic, ABJ, and extended models (see
Section 3.1). The estimation results are presented in
Table 6,
Table 7,
Table 8 and
Table 9.
In addition, the Studentized Breusch–Pagan tests (see
Table A4) revealed heteroskedasticity in most full-sample regressions. Accordingly, all pooled OLS results are reported with heteroskedasticity-robust standard errors to ensure reliable inference.
4.1.1. Organizational Capital and SG&A Cost Stickiness
Table 6 reports regression results with the logarithmic annual change rate of SG&A expenses as the dependent variable. Firms are divided into high and low OC groups.
For the low capital group, the estimated coefficient β1 is 0.4687 *** (0.4635 ***; 0.4471 ***), indicating that a 1% sales increase corresponds to a 0.4687% (0.4635%; 0.4471%) rise in SG&A expenses. This implies that SG&A costs rise less than proportionally with sales, reflecting managerial efforts to achieve scale efficiency when sales expand. The coefficient β2 is −0.0844 * (−0.1199 **; −0.1271 **), indicating cost stickiness, though significance is weak in the basic model (10% level) and stronger in the ABJ and extended models (5% level). This means that when sales fall, cost reductions are smaller, suggesting firms are reluctant to immediately cut resources such as administrative staff or overhead, which may involve adjustment costs. The combined β1 + β2 = 0.3843 (0.3436; 0.3200), meaning a 1% sales decline results in a smaller reduction in SG&A expenses.
For the high capital group, β1 is 0.4647 *** (0.4627 ***; 0.4405 ***), meaning a 1% sales increase leads to a 0.4647% (0.4627%; 0.4405%) rise in SG&A expenses. The magnitude is similar to the low OC group, suggesting that both groups exhibit proportional SG&A growth during sales expansion. The estimated β2 is −0.0981 ** (−0.0933 **; −0.1603 ***), indicating stickiness, consistently significant across all model specifications. This suggests that firms with stronger organizational capital are more constrained in adjusting SG&A downward, likely because accumulated human capital, internal routines, and organizational processes cannot be easily reduced without harming long-term capability. The combined β1 + β2 = 0.3666 (0.3694; 0.2802) suggests that a 1% sales decline results in a smaller expense reduction. Since β2 is more negative for the high capital group, SG&A costs are stickier in these firms.
The extended models provide additional insights into factors driving stickiness by incorporating firm-level and macroeconomic variables. For low capital firms, the asset-to-sales ratio (β7 = −0.0120 ***) has a significant dampening effect, suggesting firms with larger asset bases relative to sales show smaller SG&A increases. This indicates that asset-intensive firms tend to rely more on fixed investments and less on discretionary SG&A spending, making SG&A less sensitive to sales increases. β3 is insignificant in both the ABJ model (−0.0095) and the extended model (−0.2675), implying asset intensity does not materially alter stickiness when OC is low. In other words, low-OC firms already exhibit relatively flexible cost structures, and asset intensity adds little explanatory power. By contrast, in high-OC firms, β3 (−0.4213 ** in ABJ, −0.4962 ** in extended) and β7 (−0.0073 ***) are significantly negative, indicating asset intensity strengthens cost reductions during sales declines, thereby weakening stickiness. This suggests that firms with high organizational capital and high asset intensity can still reduce costs more aggressively, potentially by leveraging economies of scale or reallocating resources more effectively.
Sales decline persistence also affects stickiness. In both groups, β8 is significantly negative (Low: −0.0046 ***; High: −0.0049 **), suggesting that firms experiencing prolonged sales declines reduce SG&A costs more aggressively. This finding reflects adaptive behavior: persistent downturns push firms to eventually cut SG&A despite initial reluctance. However, β4 shows opposite patterns: in low capital firms, it is significantly positive in both ABJ (0.5569 ***) and extended (0.2880 ***) models, meaning successive declines partially offset expected expense reductions, reinforcing stickiness. This implies that managers in low-OC firms may hold back on cutting costs too quickly, anticipating that declines may reverse. In high capital firms, β4 is significant in the ABJ model (0.2844 ***) but becomes insignificant in the extended model (−0.0392). This suggests that, once other firm-level controls are considered, sales persistence has a limited additional effect on cost behavior in high-OC firms.
Macroeconomic conditions influence behavior similarly across groups. β9 is significantly positive (Low: 0.0012 ***; High: 0.0013 ***), suggesting that stronger macroeconomic conditions are associated with higher SG&A expenditures. This result is consistent with the view that during favorable macroeconomic periods, firms expand discretionary spending to support growth. Meanwhile, β5 is significantly negative (Low: −0.1465 ***; High: −0.1339 ***), suggesting that, during economic expansion, reductions in SG&A costs are mitigated when sales fall, thereby increasing stickiness. This indicates that firms are less willing to cut SG&A during downturns if the broader economy is performing well, possibly reflecting optimism about recovery.
Finally, financial flexibility as measured by free cash flow reveals contrasting effects. For the low OC group, β10 is significantly negative (−3.74 × 10−7 ***), implying that firms with greater free cash flow cut SG&A costs more readily when sales decline, reducing stickiness. This supports the view that financially flexible firms adjust spending more efficiently in response to downturns. By contrast, for the high OC group, β6 is significantly positive (7.47 × 10−6 *) while β10 is also significantly negative (−6.83 × 10−8 ***). This indicates that although abundant cash flow encourages SG&A maintenance in the short term, overall, firms with higher OC still reduce SG&A when cash flow is high, although to a lesser extent than low-OC firms. This result suggests that although abundant cash flow initially encourages managers in high-OC firms to maintain SG&A spending, overall, these firms still reduce costs when financial slack is high, but the adjustment is weaker than in low-OC firms.
In addition, firm characteristics play a consistent role across both groups. Leverage (Low: −0.0010 ***; High: −0.0009 ***) exerts a negative effect, suggesting that more highly leveraged firms cut SG&A costs more aggressively during sales declines, thereby reducing stickiness. This is consistent with the disciplining role of debt, which forces managers to tighten cost control under financial pressure. By contrast, firm size (Low: 0.0052 ***; High: 0.0046 ***) is significantly positive, implying that larger firms maintain relatively higher SG&A expenditures, which contribute to greater cost stickiness. This aligns with the idea that larger organizations have more administrative complexity and overhead, making rapid cost adjustments more difficult.
In summary, both low- and high-OC firms exhibit SG&A cost stickiness, but the effect is stronger for firms with greater organizational capital. These results support Hypothesis H1, confirming that OC level significantly influences the degree of cost stickiness.
4.1.2. Organizational Capital and Adjusted SG&A Cost Stickiness
Table 7 reports the regression results using the logarithmic annual change in adjusted SG&A expenses as the dependent variable, dividing firms into low OC and high OC groups.
For low-OC firms, β1 is 0.5009 *** (0.4965 ***; 0.4843 ***), indicating that a 1% increase in sales is associated with a 0.5009% (0.4965%; 0.4843%) increase in adjusted SG&A expenses. This suggests that adjusted SG&A rises roughly in proportion to sales growth, consistent with flexible cost behavior. The coefficient β2 is 0.0400 (0.0507; 0.0753) and is not statistically significant in any specification, implying no clear stickiness. This indicates that when sales fall, low-OC firms reduce adjusted SG&A almost symmetrically, showing limited organizational rigidity. The sum β1 + β2 = 0.5409 (0.5472; 0.5596) shows that a 1% sales decline yields a near-symmetric reduction.
For high-OC firms, β1 is 0.4858 *** (0.4808 ***; 0.4772 ***), meaning that a 1% sales increase leads to a 0.4858% (0.4808%; 0.4772%) increase in adjusted SG&A expenses. This again reflects proportional cost behavior under expansion. The estimated β2 is −0.0673 (−0.0546; −0.0946), but it is not statistically significant in any specification. Thus, adjusted SG&A expenses in high-OC firms do not show robust stickiness once adjustments are made, suggesting that organizational capital’s influence is less pronounced when non-discretionary items are removed. The sum β1 + β2 = 0.4185 (0.4262; 0.3826) indicates that although reductions are smaller in magnitude, they are not statistically distinguishable from symmetry.
In summary, both low- and high-OC firms show little evidence of stickiness in adjusted SG&A expenses. This result implies that once SG&A is adjusted to exclude non-discretionary components, the role of organizational capital in amplifying stickiness weakens considerably. These results provide a nuanced view of H1, which posits that OC influences the extent of cost stickiness in SG&A expenses.
4.1.3. Adjusted Organizational Capital and SG&A Cost Stickiness
Table 8 reports results using the logarithmic annual change in SG&A expenses as the dependent variable, separating firms into low AOC and high AOC groups.
For the low AOC group, the estimated β1 is 0.4649 *** (0.4610 ***; 0.4432 ***), indicating that a 1% increase in sales is associated with a 0.4649% (0.4610%; 0.4432%) increase in SG&A expenses. This shows that SG&A costs rise less than proportionally with sales, reflecting partial scale efficiency. The coefficient β2 is −0.0964 * (−0.1142 *; −0.1183 *), suggesting the presence of cost stickiness. The effect is weakly significant in the basic model (10% level) but becomes stronger in the ABJ and extended models. This indicates that when sales fall, low AOC firms are reluctant to reduce SG&A proportionally, possibly due to adjustment costs or managerial inertia. The combined coefficient β1 + β2 = 0.3685 (0.3468; 0.3249) indicates that a 1% decrease in sales leads to a 0.3685% (0.3468%; 0.3249%) decline in SG&A expenses.
For the high AOC group, the estimated β1 is 0.4691 *** (0.4660 ***; 0.4395 ***), indicating that a 1% increase in sales results in a 0.4691% (0.4660%; 0.4395%) increase in SG&A expenses. This again reflects the partial proportionality of SG&A to sales increases. The estimated β2 is −0.0797 * (−0.0880 *; −0.1400 ***), consistently negative and statistically significant, confirming stickiness. This suggests that high AOC firms face stronger organizational rigidity, as accumulated routines and capabilities make cost reductions harder when sales fall. The sum β1 + β2 = 0.3894 (0.3780; 0.2995) likewise indicates a muted reduction during sales declines. The more negative β2 values in the high AOC group imply stronger stickiness compared to the low AOC group.
Overall, both groups display SG&A cost stickiness, with the effect more pronounced in high AOC firms. This supports H1, showing that adjusted organizational capital strengthens the asymmetry in SG&A cost behavior, particularly under sales declines.
4.1.4. Adjusted Organizational Capital and Adjusted SG&A Cost Stickiness
Table 9 reports the regression results using the logarithmic annual change in adjusted SG&A expenses as the dependent variable, comparing low AOC and high AOC firms.
For low AOC firms, β1 is 0.4872 *** (0.4869 ***; 0.4710 ***), indicating that a 1% increase in sales is associated with a 0.4872% (0.4869%; 0.4710 ***) increase in adjusted SG&A expenses. This suggests that adjusted SG&A responds almost proportionally to sales growth, reflecting flexible cost behavior. The coefficient β2 is 0.0738 (0.1741; 0.2145) and is not significant, indicating no clear stickiness. This means that when sales fall, low AOC firms reduce adjusted SG&A almost symmetrically, showing little evidence of rigid cost structures. The sum β1 + β2 = 0.5610 (0.6610; 0.6855) suggests a near-symmetric response to sales declines.
For high AOC firms, β1 is 0.5041 *** (0.4976 ***; 0.4921 ***), meaning that a 1% increase in sales leads to a 0.5041% (0.4976%; 0.4921%) increase in adjusted SG&A expenses. This indicates that SG&A scales proportionally with sales expansion. The estimated β2 is −0.1054 (−0.1387; −0.1969 *), negative and significant only in the extended model, suggesting partial evidence of cost stickiness. This implies that high AOC firms are slower in reducing adjusted SG&A during downturns, likely because organizational routines and embedded capabilities create resistance to cost cuts. The sum β1 + β2 = 0.3987 (0.3589; 0.2952) indicates a smaller decline in adjusted SG&A expenses when sales fall, especially when controlling for firm characteristics.
In summary, low AOC firms show no significant stickiness in adjusted SG&A expenses, while high AOC firms exhibit some stickiness, particularly in extended specifications. These results support H1, suggesting that adjusted organizational capital is associated with greater asymmetry in SG&A adjustments, even after excluding non-discretionary components.
Across specifications, high OC and high AOC firms generally exhibit stronger SG&A cost stickiness, reflected in more negative
β2 coefficients. In other words, when sales decline, these firms reduce SG&A expenses less, sustaining higher expenditure levels. By contrast, low OC and low AOC firms show weaker or insignificant stickiness, indicating more symmetric cost adjustments. Overall, these findings robustly support H1. These results align with
Anderson et al. (
2003), who documented asymmetric SG&A adjustments, and extend their evidence by showing that organizational capital amplifies the degree of stickiness. Unlike
Venieris et al. (
2015), however, low-OC firms in Japan also display cost stickiness, reflecting institutional differences.
4.2. The Impact of Organizational Capital on Cost Stickiness Based on the Magnitude of Sales Changes
To test the hypothesis that the effect of organizational capital on cost stickiness depends on the magnitude of sales changes, this study adopts a comparative approach by dividing the sample into two groups: firms experiencing small sales changes and those experiencing large sales changes.
4.2.1. Organizational Capital and SG&A Cost Stickiness Across Sales Change Magnitudes
Table 10 and
Table 11 report the regression results using the logarithmic annual change in SG&A expenses as the dependent variable, classifying firms by OC level and sales-change magnitude (
Venieris et al., 2015).
For small sales changes (see
Table 10), stickiness is clear. For the low OC group,
β1 is 0.5583 *** (0.5503 ***; 0.5265 ***), indicating that a 1% sales increase leads to about a 0.56% rise in SG&A expenses. This reflects partial proportionality of SG&A with sales expansion. The coefficient
β2 is −0.2576 *** (−0.2574 ***; −0.2879 ***), strongly significant and negative, confirming asymmetry. This means that when sales fall, SG&A costs decrease less, consistent with cost stickiness. The combined
β1 +
β2 = 0.3007 (0.2929; 0.2386) confirms that a 1% sales decline results in a much smaller reduction in SG&A. In the high OC group,
β1 is 0.5124 *** (0.5129 ***; 0.4795 ***), indicating SG&A grows by 0.5124% (0.5129%; 0.4795%) when sales rise 1%. This suggests that SG&A growth remains proportional under sales expansion. The estimated
β2 is −0.1263 * (−0.0832; −0.1653 **), confirming asymmetry but with weaker significance than in the low OC group. This implies that high-OC firms still show some stickiness, although the evidence is less consistent across models. The combined
β1 +
β2 = 0.3861 (0.4297; 0.3142) likewise indicates smaller reductions during sales declines.
For large sales changes (see
Table 11), the evidence of cost stickiness weakens considerably. For the low OC group,
β1 is 0.4290 *** (0.4155 ***; 0.4016 ***), showing SG&A increases by 0.4290% (0.4155%; 0.4016%) for a 1% sales rise. This reflects proportional adjustment under growth. The coefficient
β2 is 0.0568 (0.0462; 0.0486), all insignificant, indicating no asymmetry. This suggests that under larger shocks, SG&A cuts are more symmetric. For the high OC group,
β1 is 0.4536 *** (0.4386 ***; 0.4167 ***), likewise confirming proportional SG&A growth. This indicates that cost structures adjust in line with sales expansion. The estimated
β2 is −0.0677 (−0.0969; −0.0917), negative but not significant. Thus, even high-OC firms show little evidence of stickiness under large shocks, suggesting managers adjust SG&A more flexibly when sales swings are large.
Comparing across sales-change magnitudes, cost stickiness is statistically significant only for smaller changes (up to 10%) and largely disappears when changes fall between 10% and 50%. This pattern applies to both low- and high-OC firms, consistent with the idea that stickiness is more relevant under marginal adjustments but not under major shifts.
In summary, SG&A cost stickiness is evident under smaller sales changes but weakens considerably as the magnitude of changes increases. These findings support Hypothesis H2, which posits that the impact of organizational capital on cost stickiness varies with the magnitude of sales changes. When sales changes are modest, firms tend to preserve organizational capacity and resources, leading to asymmetric cost behavior. In contrast, larger sales changes compel substantial cost realignment, reducing stickiness.
4.2.2. Organizational Capital and Adjusted SG&A Cost Stickiness Across Sales Change Magnitudes
Table 12 and
Table 13 report regression results using the logarithmic annual change in adjusted SG&A expenses as the dependent variable, dividing firms by their level of OC and by the magnitude of sales changes.
For small sales changes (see
Table 12), the results show strong evidence of cost stickiness. For low-OC firms,
β1 is 0.6493 *** (0.6343 ***; 0.6159 ***), indicating that a 1% increase in sales is associated with a 0.6493% (0.6343%; 0.6159%) increase in adjusted SG&A expenses. This suggests that SG&A rises slightly more than proportionally during expansions. The coefficient
β2 is −0.2699 (−0.4198 **; −0.4471 *), consistently negative and significant in two specifications. This implies that when sales fall, cost reductions are smaller, reflecting asymmetric adjustment. The combined
β1 +
β2 = 0.3794 (0.2145; 0.1688) further confirms that SG&A declines much less than sales, consistent with stickiness. In high-OC firms,
β1 is 0.6575 *** (0.6475 ***; 0.6437 ***), showing that SG&A grows by about 0.65% when sales increase by 1%. This indicates a strong proportional relationship in expansions. The estimated
β2 is also negative and significant (−0.3772 **; −0.3794 **; −0.4029 **). This confirms stronger stickiness in high-OC firms, consistent with higher organizational rigidity. The combined
β1 + β2 = 0.2803 (0.2681; 0.2408) confirms stronger stickiness in high-OC firms, as evidenced by the more negative
β2.
For large sales changes (see
Table 13), the evidence of cost stickiness weakens considerably. In low-OC firms,
β1 is 0.5251 *** (0.5066 ***; 0.5098 ***), indicating proportional SG&A growth with sales. This suggests flexible scaling during larger expansions. The
β2 values (0.0774; 0.1376; 0.1135) are small and insignificant, indicating no clear cost asymmetry. This implies that large sales declines trigger symmetric cost adjustments. In high-OC firms,
β1 is 0.4167 *** (0.4096 ***; 0.4155 ***), meaning SG&A rises 0.4167% (0.4096%; 0.4155%) when sales increase 1%. This reflects a more moderate proportional response compared to smaller changes. The
β2 values (0.1170; −0.0741; −0.1199) vary in sign but remain insignificant. Thus, under large sales shifts, even high-OC firms adjust SG&A more symmetrically, with little evidence of stickiness.
Comparing across sales-change magnitudes, adjusted SG&A cost stickiness is statistically significant only for smaller changes (up to 10%) and largely disappears when changes range between 10% and 50%. This pattern holds for both low- and high-OC firms, but the evidence is more consistent and pronounced in the high-OC group.
In summary, adjusted SG&A cost stickiness is evident under smaller sales changes but weakens considerably as the magnitude of changes increases. These findings support Hypothesis H2, which posits that the impact of organizational capital on cost stickiness varies with the magnitude of sales changes. Under small fluctuations, both low and high-OC firms maintain SG&A expenditures asymmetrically, with the effect stronger in high-OC firms. In contrast, larger shifts require unavoidable cost adjustments that diminish stickiness.
4.2.3. Adjusted Organizational Capital and SG&A Cost Stickiness Across Sales Change Magnitudes
Table 14 and
Table 15 report regression results using the logarithmic annual change in SG&A expenses as the dependent variable, dividing firms by their level of AOC and the magnitude of sales changes.
For small sales changes (see
Table 14), the results provide strong evidence of cost stickiness. In low AOC firms,
β1 is 0.5490 *** (0.5437 ***; 0.5094 ***), suggesting that a 1% increase in sales is associated with a 0.5490% (0.5437%; 0.5094%) increase in SG&A expenses. This indicates that SG&A grows somewhat less than proportionally with sales, reflecting partial scale efficiency during expansions. The coefficient
β2 is −0.2481 *** (−0.2459 ***; −0.2491 ***), negative and significant across all specifications. This means that when sales fall, cost reductions are smaller than increases, consistent with sticky SG&A. The combined
β1 + β2 = 0.3009 (0.2978; 0.2603) indicates that a 1% decline in sales results in a smaller reduction in SG&A expenses, confirming asymmetric cost behavior. In high AOC firms,
β1 is 0.5214 *** (0.5193 ***; 0.4931 ***), implying that a 1% sales increase leads to a 0.5214% (0.5193%; 0.4931%) rise in SG&A. This reflects a proportional but slightly smaller expansion response than low AOC firms. The coefficient
β2 is −0.1328 ** (−0.1058; −0.1988 ***), negative and significant in two of the three models. This points to some stickiness, though weaker than in the low AOC group. The combined
β1 + β2 = 0.3886 (0.4135; 0.2943) likewise indicates smaller reductions during declines, but less pronounced than in low AOC firms.
For large sales changes (see
Table 15), evidence of cost stickiness weakens considerably. For the low AOC group,
β1 is 0.4349 *** (0.4214 ***; 0.4063 ***), showing that SG&A rises by 0.4349% (0.4214%; 0.4063%) for a 1% sales increase. This suggests broadly proportional scaling in expansions. The
β2 values (0.0305; 0.0193; 0.0350) are small and statistically insignificant, indicating no strong asymmetry in cost adjustment. This means that large shocks appear to prompt more symmetric SG&A adjustments. For the high AOC group,
β1 is 0.4435 *** (0.4289 ***; 0.3951 ***), again indicating proportional SG&A growth with sales. This suggests managers adjust SG&A in line with larger sales movements. The
β2 values (−0.0336; −0.0555; −0.0401) are negative but not significant. Thus, neither group exhibits clear cost stickiness under larger sales changes.
Comparing across sales-change magnitudes, SG&A cost stickiness is statistically significant only for small changes (up to 10%) and largely disappears when changes range between 10% and 50%. This pattern holds for both low and high AOC firms. However, under small changes, the stickiness is more pronounced in the low AOC group, whereas it is weaker in the high AOC group.
In summary, SG&A cost stickiness is evident under smaller sales changes but weakens substantially when the magnitude of changes increases. The effect is stronger in the low AOC group when sales changes are small, while both low and high AOC firms show little evidence of stickiness under larger changes. These findings support Hypothesis H2, which posits that the impact of organizational capital on cost stickiness varies with the magnitude of sales changes. The results imply that cost asymmetry is more relevant when adjustments are marginal, while major shocks leave managers with few options but to adjust costs symmetrically.
4.2.4. Adjusted Organizational Capital and Adjusted SG&A Cost Stickiness Across Sales Change Magnitudes
Table 16 and
Table 17 report regression results using the logarithmic annual change in adjusted SG&A expenses as the dependent variable, dividing firms by their level of AOC and the magnitude of sales changes.
For small sales changes (see
Table 16), the results show strong evidence of cost stickiness. For the low AOC group, the estimated coefficient
β1 is 0.6277 *** (0.6145 ***; 0.5744 ***), suggesting that a 1% increase in sales is associated with a 0.6277% (0.6145%; 0.5744%) increase in adjusted SG&A expenses. This indicates SG&A grows proportionally with revenue in expansions. The coefficient
β2 is −0.2260 (−0.3370; −0.3172), negative but not significant, implying that cost reductions during declines are smaller, though statistical support is weak. This points to possible but not robust stickiness. The combined
β1 + β2 = 0.4017 (0.2775; 0.2572) indicates that a 1% decline in sales results in a much smaller reduction in adjusted SG&A expenses, suggesting asymmetry in magnitude, but this is not statistically robust. For the high AOC group,
β1 is 0.6790 *** (0.6672 ***; 0.6700 ***), showing that a 1% increase in sales leads to 0.6790% (0.6672%; 0.6700%) rise in SG&A. This suggests stronger proportional growth during expansions. The coefficient
β2 is also negative and significant (−0.4224 ***; −0.4702 ***; −0.5041 ***). This confirms pronounced cost stickiness when sales decline. The combined
β1 + β2 = 0.1973 (0.2055; 0.1715) confirms stronger stickiness for high AOC firms.
For large sales changes (see
Table 17), evidence of stickiness weakens considerably. For the low AOC group,
β1 is 0.5000 *** (0.4877 ***; 0.4802 ***), indicating that SG&A rises by 0.5000% (0.4877%; 0.4802%) for a 1% increase in sales. This shows proportional adjustment in expansions. The
β2 values (0.1336; 0.2140; 0.2229) are small and insignificant, indicating no strong asymmetry in cost adjustment. This means that costs adjust symmetrically under large shocks. For the high AOC group,
β1 is 0.4437 *** (0.4383 ***; 0.4450 ***), suggesting SG&A rises 0.4437% (0.4383%; 0.4450%) when sales increase 1%. This reflects proportional scaling. The
β2 values (0.0599; −0.1424; −0.2074) are not significant. Thus, neither group displays meaningful stickiness when sales changes are relatively large.
Comparing across sales-change magnitudes, adjusted SG&A cost stickiness is statistically significant only when sales changes are small (up to 10%) and largely disappears when sales changes fall between 10% and 50%. This result holds for both low and high AOC firms, although stickiness appears stronger in high AOC firms when present.
In summary, adjusted SG&A cost stickiness is evident under smaller sales changes but weakens substantially as the magnitude of sales changes increases. These findings support Hypothesis H2, which posits that the impact of OC on cost stickiness varies with the magnitude of sales changes. Firms with higher AOC appear more inclined to retain resources and sustain spending under small sales shifts, but when sales shocks are large, cost flexibility dominates, and stickiness dissipates.
This section examined whether the impact of OC on cost stickiness varies with the magnitude of sales changes. Across all model specifications, the evidence indicates that cost stickiness in both SG&A and adjusted SG&A expenses is statistically significant when sales changes are relatively small (up to 10%), but it largely disappears when sales changes are more substantial (10–50%).
Specifically, both low- and high-OC firms exhibit asymmetric cost behavior under small sales changes. In most cases, the effect is generally stronger for firms with higher OC or AOC, as reflected by more negative
β2 coefficients, although in some adjusted SG&A specifications (
Table 14) stickiness appears more pronounced in low AOC firms. This suggests that when sales fluctuations are modest, managers prefer to retain resources and maintain organizational capacity, especially in firms with higher OC, thereby sustaining higher SG&A. For larger sales changes,
β2 becomes statistically insignificant, indicating that costs adjust more symmetrically to sales increases and decreases regardless of OC, as firms are compelled to realign their cost structures under major shocks.
In summary, these results provide strong support for Hypothesis H2, which posits that the effect of OC on cost stickiness depends on the magnitude of sales changes. The findings suggest that OC reinforces cost asymmetry primarily during periods of small sales changes, whereas its influence weakens considerably or even diminishes when sales shocks are larger. This pattern supports the argument of
Balakrishnan et al. (
2004) that cost behavior is contingent on the scale of sales changes. Similarly to
Venieris et al. (
2015), we find stronger stickiness in high-OC firms, but unlike the U.S. context, Japanese firms with low OC also exhibit asymmetry, highlighting the role of institutional and cultural settings.