Research Article: 2021 Vol: 25 Issue: 1S
JeongHo Koo, Kumoh National Institute of Technology (KIT)
Daecheon Yang, ChungAng University
Initial Public Offering (IPO), Cost Stickiness, Empirebuilding Behaviour
The literature on managerial accounting documents that cost stickiness is asymmetric cost reduction not responding to demand declines, which is driven by empirebuilding behaviour undesirable for shareholders. We posit that cost stickiness is likely to be rampant around Initial Public Offering (IPO) firms, particularly in an emerging IPO market that the corporate governance structure weakens. This paper focuses on Korea's KOSDAQ market where IPOs have been frequent during recent decades because of the rapid growth of information technology and bio industries. By using a sample of 14,424 firmyear observations from 2002 to 2016 in Korea's KOSDAQ market, we explore whether IPO issuers aggravate the degree of cost stickiness after the offering in comparison with their peer groups. Our results suggest that the cost stickiness of postIPO firms is greater than that of nonIPO firms.
Anderson, et al., (2003) empirically suggest the coststickiness phenomenon, in which costs act asymmetrically, falling less when demand decreases than rising when equivalent demand increases. As an agency explanation, it has been well documented that cost stickiness could result from empirebuilding behaviours for increasing managerial personal utility from status, power, and resources (Chen et al., 2012). Managers' empirebuilding tendencies could cause their companies to grow up beyond the optimal level, retain overslack, and thus lead to excessive costs that should otherwise be cut.
Specifically, this cost stickiness phenomenon might be pervasive around initial public offering (IPO) firms. During the IPO offering, high information asymmetry occurs between issuers and investors (Teoh et al., 1998a, b) and issuers may have great opportunities to pursue empirebuilding under the pretext of the expansion of firms. PostIPO firms might be highly susceptible to cost stickiness. Moreover, IPO environments might provide IPO issuers with a possible source of unwarranted optimism in the growth of companies (Teoh et al., 1998a; Shu et al., 2012). Empirebuilding issuers are expected to overestimate the growth of their young companies and overinvest beyond the optimal level with strong reluctance to downsize, thus inducing greater cost stickiness.
This paper centres on Korea's KOSDAQ market, where IPOs have been frequent during recent decades because of the rapid growth of information technology and bio industries. The data obtained from the KOSDAQ’s IPO firms facilitates the observation of issuers' empirebuilding. In this study, we explore whether IPO issuers trigger the cost stickiness problem arising from empirebuilding opportunities after the offering. By using a sample of 14,424 firmyear observations covering postIPO’s 1,308 firmyear observations during the period 2002–2016, we find that the cost stickiness of postIPO firms is greater than that of nonIPO firms. This paper contributes to the literature by suggesting new evidence of IPO issuers’ reinforcing cost stickiness, which might result in excessive costs associated with overslack so as not to be suboptimal to shareholders. We believe that the high asymmetry during the IPO offering raises the necessity of an active governance structure that restrains the wasteful managerial spending of valuable economic resources.
By Anderson et al. (2003), coststickiness is defined as the phenomenon that operational costs increase when sales increase but does not decrease to the same extent when sales decline. Subsequent empirical studies demonstrated that cost stickiness is pervasive under various settings (e.g., Balakrishnan et al., 2004; Banker & Chen, 2006; Anderson et al., 2007; Weiss, 2010; Chen et al., 2012; Chen et al., 2013; Banker et al., 2014; Via & Perego, 2014; Yang, 2015). Relevant studies describe cost stickiness as asymmetric cost reduction that is derived from managers’ decisionmaking that pursue private utilities (e.g., Balakrishnan & Gruca, 2008; Chen et al., 2012; Dierynck et al., 2012; Kama & Weiss, 2012). In particular, Chen et al., (2012) specify that managers' empirebuilding, which reflects an unwarranted taste for larger firms, induces them to retain unutilized slack.
These empirebuilding preferences, addressing the agency problem, are expected to be greater under the context of high information asymmetry. It is well known that startup and growth companies are the most prone to asymmetric information (Fazzari et al., 1988; Hoshi et al., 1991). In particular, high asymmetry of information exists between IPO issuers and investors (Teoh et al., 1998a, b). This high asymmetry causes the issuer's empirebuilding behaviour to be undesirable to investors, particularly with the pretense of the expansion and growth of firms going public.
During and after the offering, all the participants tend to be overwhelmed by the atmosphere of rapid growth and the IPO issuers be intoxicated by postIPO' prosperity. Our proposition concerns the possibility of the stickier costs of postIPO firms compared with those of nonIPO firms. IPO issuers may pursue an excessive tendency to run large firms, using considerable information asymmetry. We expect that IPO issuers are more likely to run an oversized firm by retaining overcapacities and induce a greater degree of cost stickiness. Thus, we verify whether the cost stickiness of postIPO firms is greater than that of nonIPO firms.1
Based on Anderson et al. (2003)’s model, we estimate a timeseries regression at the firmyear level as the following:
Where, ΔCosti,t is the natural logarithm of change (Costi,t/Costi,t–1) in the cost elements of firm i in year t; ΔSalei,t is the natural logarithm of change (Sale,t/Salei,t–1) in sales in year t relative to year t–1; and DDi,t is the dummy variable that takes 1 if sales decrease (Salei,t<Salei,t–1), and zero otherwise.
According to this model, if β1 coefficient has a significantly positive value, a significantly negative β2 coefficient captures downwardly sticky costs where sales decrease.
Then, we separately include an IPO variable in Eq. (1) and interact it with other independent variables (see Eq. 2). In our test, the timeline of the post IPO periods is the fiscal year including the IPO date and the subsequent year. In Eq. (2), the negative ��4 coefficient on Δ��������i�� * ����i�� * �������� − ������i�� implies that a firm's IPO intensifies the cost stickiness (if ��2<0) or a firm's IPO pushes SG&A costs in a downwardly sticky direction (if ��2>0).
Where, postIPO is a dummy variable that takes the value of 1 if the period t is the fiscal year including the IPO date or the subsequent year in postIPO firm i and zero otherwise (i.e., industry control firm) for the equivalent periods. We control for the oneyear change in asset intensity (ASTit), book value of assets (SIZEit), and book value of debts divided by the book value of assets (LEVit) (see Chen et al., 2012; Yang, 2015). ASTit is calculated as the ratio of total assets divided by sales revenue; SIZEit is calculated as the book value of assets; and LEVit is calculated as the book value of debts divided by the book value of assets. All the control variables are converted to the natural logarithm value to be normally standardized.
Next, we alternatively employ a coststickiness model developed by Homeburg and Nasev (H&N) (2008). Homburg and Nasev (2008) measure cost stickiness (Stickiness) as the positive costtosale ratio (SG&A ratio) conditional on decreasing sales. In Eq. (3), the positive value of Stickiness means the cost stickiness and the degree of cost stickiness increase as its value increases.
Then, we adopt the following fixedeffects H&N regression by using Stickiness as the dependent variable. In Eq. (4), the value of the β_1 coefficient on PostIPO is expected to be positive, thus suggesting the greater cost stickiness in PostIPO firms.
Here, postIPO is a dummy variable that takes the value of 1 if the period t is the fiscal year including the IPO date or the subsequent year in postIPO firm i, and zero otherwise (i.e., industry control firm) for the equivalent periods. Refer to Eq. (2) for the definitions of other variables.
Our data strategies are as follows. We obtain IPO data together from the KIND system of the Korea Exchange and DART system provided by the Financial Supervisory Service of Korea during the period 2002–2016. Our cost and financial data are drawn from the intersection of KisValue II, TS 2000, and Dataguide Pro. (ver. 3) of FnGuide for companies listed in the KOSDAQ market. Our final sample consists of 14,424 firmyear observations covering the postIPO sample of 1,308 firmyear observations.
In relation to the postIPO firm sample, Panel A of Table 1 provides descriptive statistics for all the variables used in the regressions. The average of ΔSale is 0.1137, representing an 11.37% sale increase from the previous period. Also, the average of ΔCOST is 0.2068, indicating that the SG&A cost increased by 20.68% from the previous year. Meanwhile, the natural log of assets/sales, which is one of the main controls, shows an average value of 0.2285.
Table 1 DESCRIPTIVE STATISTICS  

<Panel A> Descriptive statistics of the postIPO sample  
Mean  Std. Dev.  1st Quartile  Median  3rd Quartile  
ΔSale  0.1137  0.4034  0.0435  0.1118  0.2537 
ΔCost  0.2068  0.3038  0.0577  0.1835  0.3224 
DD  0.2973  0.4572  0  0  1 
HN Stickiness  0.0196  0.0545  0  0  0.005 
AST  0.2285  0.5431  0.1546  0.2106  0.5599 
SIZE  24.563  0.7285  24.0497  4.5323  25.0239 
LEV  0.3055  0.178  0.1627  0.2809  0.5494 
<Panel B> Descriptive statistics of the full sample  
Mean  Std. Dev.  1st Quartile  Median  3rd Quartile  
ΔSale  0.128  0.44409  0.0531  0.0924  0.2677 
ΔCost  0.1218  0.37234  0.0264  0.0986  0.2495 
DD  0.3222  0.46734  0  0  1 
HN Stickiness  0.0174  0.0602  0  0  0.0011 
AST  0.1928  0.65807  0.2202  0.1379  0.5352 
SIZE  24.5868  1.15058  23.9408  24.6949  25.3017 
LEV  0.4173  0.21931  0.2461  0.41111  0.5678 
This table summarizes the descriptive statistics of the dependent and independent variables for the full sample of 14,424 firmyears and the postIPO sample of 1,308 firmyear observations, sourced from the intersection of KisValue II, TS 2000, and Dataguide Pro. (ver. 3) of FnGuide over the period 2002–2016. 
If IPO issuers pursue their empirebuilding preferences after their firms go public, we can expect a greater level of cost stickiness in postIPO firms. To examine this conjecture, we run fixedeffects panel estimations based on Anderson et al. (2003) and H&N (2008), respectively.
Table 2 presents the estimates for panel regressions based on Eq. (2). In Col. (1), the results of the baseline model show the significantly positive (��1) coefficient estimate of ΔSale and positive (��2) coefficient estimate of the interaction of the change in costs and the sale decrease dummy (ΔSale × DD). The results suggest that SG&A costs are downwardly elastic in response to sale change in firms listed in the KOSDAQ market.
Table 2  

EFFECT OF IPO ON COST STICKINESS  
Baseline model Col. (1)  PostIPO Col. (2)  
Predicted signals  Estimate (tstat)  Estimate (tstat)  
Intercept  0.4925 (8.46)***  0.4919 (8.47)***  
∆Sale  0.4237 (54.00)***  0.4207 (53.48)***  
∆Sale × DD  (+/)  0.0283 (1.74)*  0.0444 (2.71)*** 
∆Sale × IPO  0.2300 (5.23)***  
∆Sale × DD ×postIPO  ()  0.6084 (8.48)***  
AST  0.0282 (6.6)***  0.0276 (6.48)***  
SIZE  0.0168 (7.19)***  0.0170 (7.27)***  
LEV  0.0115 (0.95)  0.0049 (0.40)  
All the interaction terms  Yes  
Yearfixed effects  Yes  Yes  
Industryfixed effects  Yes  Yes  
N  14,424  14,424  
Adjusted R2  0.2946  0.2979  
Note: This table presents the estimation results for the panel regression for the effects of IPO on the degree of cost stickiness. Cost stickiness is measured based on the method of Anderson et al. (2003). postIPO is a dummy variable that takes the value of 1 if the period t is the fiscal year including the IPO date or the Subsequent year in postIPO firm i, and zero otherwise (i.e., industry control firm) for the equivalent periods. ***, **, and * indicate statistical significance at the 1%, 5%, and 10% levels, respectively. 
Next, in postIPO periods, the results of Col. (2) show that β4 of ΔSale × DD × postIPO is significantly negative (coefficient=−0.6084, t=−8.48) given a significantly positive β1 of ΔSale and positive β2 of ΔSale × DD. Thus, the results reveal that the SG&A costs in post IPO firms are downwardly sticky, unlike the case of industrycontrolled firms.
We run the alternative coststickiness model developed by H&N (2008) using Eq. (4). Our fixedeffects H&N specification could directly capture the impact of IPO on the change in the magnitude of cost stickiness without using many interactions. Table 3 reports the estimates for fixedeffects H&N regressions. The result shows that the β1 coefficient of postIPO has a significantly positive value (coefficient=0.0065, t=2.96), thus revealing greater cost stickiness in postIPO firms. The result is qualitatively are same compared with that of prior test in Table 2.
Table 3  

IPO AND H&N'S STICKINESS  
Predicted signals  H&N's Stickiness 

Estimate (tstat)  
Intercept  0.1746 (16.82)*** 

PostIPO  (+)  0.0065 (2.96)*** 
AST  0.0289 (38.71)*** 

SIZE  0.0069 (16.57)*** 

LEV  0.0180 (8.1)***  
Industryfixed assets  Yes  
Yearfixed effects  Yes  
N  14,424  
AdjustedR2  0.097 
The literature on managerial accounting addresses that cost stickiness is a asymmetric cost reduction which might stem from by empirebuilding behaviour. We extend the literature by investigating the existence of the stickier costs in postIPO firms. Hence, we postulate that IPO issuers are likely to pursue their empirebuilding behaviours by utilizing considerable information asymmetry during and after the offering and thus induce a greater degree of cost stickiness. As expected, the results suggest that cost stickiness is greater for postIPO firms than for non IPO firms. This important conclusion is robust to alternative methodologies for measuring the degree of cost stickiness.
• This work was supported by the Ministry of Education of the Republic of Korea and the National Research Foundation of Korea (NRF2020S1A5A2A01044196)
• Address correspondence to Daecheon Yang, School of Business Administration, ChungAng University, Seoul Korea. Email: [email protected]