Market Competition, Downward-Sticky Pay, and Stock Returns: Lessons from South Korea

Abstract

This study examines whether market competition reduces managerial slack under downward-sticky CEO pay schemes, thus mitigating the potentially negative link between downward-sticky pay and shareholder’s value. Using data on the Korean product market, which has been dominated by business conglomerates known as ‘chaebols’, we first find that downward-sticky pay is prevalent in underperforming firms and affects shareholder value negatively. Then, we find that a higher level of market competition alleviates the value-deteriorating effect of downward-sticky pay. Overall, the findings from our study imply that market competition as an external mechanism of corporate governance threatens still highly paid CEOs with worsening performance and motivates them implicitly to work harder. Together with a need for shareholders’ influence on downward-sticky pay, this study sheds light on the importance of market competition regimes in developing countries where legal protection for shareholders and internal governance structures are weak.

1. Introduction

A stream of studies describe downward-sticky pay, whereby a CEO’s annual pay increases when the firm’s performance increases but does not decrease to the same extent when firm’s performance reduces, as evidence of managerial slack. According to agency theory, CEO pay should be proportional to a firm’s performance to align interests between management and shareholders (Jensen & Meckling, 1976; Holmstrom, 1979; Grossman & Hart, 1983; Jensen & Murphy, 1990; Murphy, 1998; Core et al., 1999). Contrary to the intent of this performance-based pay, downward-sticky pay seems to be pervasive in the real world (Gaver & Gaver, 1998; Adut et al., 2003; Ahn & Lee, 2003; Garvey & Milbourn, 2006; Jackson et al., 2008; J. Kim et al., 2017). Related literature relates downward-sticky pay to pay without performance (Bebchuk & Fried, 2004; J. Kim et al., 2017). The downward-sticky pay scheme (which seems to be derived from management’s influence on the board) might allow executives to work passively and reflect both managerial rent-seeking and slack (Blanchard et al., 1994; Yermack, 1997; Bertrand & Mullainathan, 2001; Bebchuk et al., 2002; Bebchuk & Fried, 2004).

This paper examines whether downward-sticky pay reduces firm value and whether market competition, as an external governance mechanism, mitigates this negative effect. The literature on the industry characteristics of competition focuses on competition’s effects on top managers’ efforts and the strength of pay-for-performance schemes. The common assumption is that market competition increases the threat of bankruptcy and CEO dismissal because it may render the firm unprofitable, thereby reducing managerial slack and inducing the manager to work harder to avoid liquidation (e.g., Machlup, 1967; Hart, 1983; Schmidt, 1997; Raith, 2003; Beiner et al., 2011; Chang et al., 2015; Noghani & Noghanibehambari, 2019). On the other hand, some studies argue that the relationship between competition and managers’ effort level is negative or ambiguous (e.g., Scharfstein, 1988; Hermalin, 1992; Graziano & Parigi, 1998; Karuna, 2007; Golan et al., 2015). These studies propose the alternative explanation that increasing competition lowers firm profits and makes firms’ cost reduction initiatives more attractive, thereby motivating their managers to work harder through more incentives. The argument between these two conflicting theories centers on the impact of the liquidation threat on managerial incentives.

However, both studies largely ignore not only the level of firm performance but also the design features of executive compensation. The studies’ discussion of the liquidation threat is deficient in two ways. First, the worsening condition of a firm increases the threat of bankruptcy. Second, the implicitly motivating mechanism of the bankruptcy threat is valid only under non-optimal pay arrangements when there is sufficient managerial slack. To consider the conditions that the two conflicting theories neglect, we posit that threat of liquidation derived from increasing market competition can induce managers to make more of an effort to avoid liquidation under downward-sticky pay schemes.

In this context, our study centers on the role of market competition as the external means of corporate governance of preventing value-deteriorating behaviors by a rent-seeking CEO. Specifically, market competition regimes in developing countries where legal protection for shareholders and internal governance structures are weak are expected to play a crucial monitoring role. A strand of research reports the dark side of the reduction in market competition attributed to conglomerates (Carney et al., 2018; Pereira & Sousa, 2023). This aspect seems to be notably prevalent in the Korean product market, which has been dominated by business conglomerates known as ‘chaebols’ (Lee, 2002; Byun et al., 2012, 2018). So far, sales of chaebols accounts for around 87% of Korean GDP (2022yr). The Korean market data facilitates the observation of the monitoring role of market competition over managerial slack that possibly stems from business monopoly.

Using data on the Korean product market, which has been cornered by business conglomerates, we investigate whether market competition acts as the external means of corporate governance and thus deters value-deteriorating behaviors driven by a rent-seeking CEO. Specifically, this study offers an examination of whether market competition reduces managerial slack under downward-sticky pay schemes and attenuates the possibly value-destroying effect of downward-sticky pay. Using a sample of 6524 firm-year observations from 2018 to 2022 for non-financial firms in Korea’s KOPSI and KOSDAQ market, this study provides evidence that downward-sticky pay negatively affects a firm’s stock returns, but market competition alleviates the value-deteriorating effect of this pay scheme. Using a variety of estimation methods, our results were found to be robust.

This study contributes to the literature in several ways. First, although the agency problem, which is rooted in the dominance of managerial power, is becoming a central issue for business and scholars, empirical evidence is lacking with regard to the relationship between non-optimal performance-related pay and stock returns. This study addresses this core issue by empirically investigating the value-deteriorating effect of downward-sticky pay. Second, the conflicting theories on market competition have been having a continuing argument about whether market competition reduces or increases managerial slack.1 Based on the empirical results of this study, we complement the literature on the linkage of market competition and managerial slack by providing empirical evidence that market competition can reduce managerial slack problem “under downward-sticky pay scheme”.2 Moreover, we add to robust evidence for this discussion by employing not only price-based but also non-price-based measures (advertising and R&D-to-sales ratio). Finally, many Korean studies have raised various issues associated with weak internal governance of business conglomerates known as ‘chaebols’ (e.g., Bae et al., 2006; W. C. Kim et al., 2007; Almeida et al., 2011). This study complements the literature on the dark side of the monopolization in the Korean product market derived by chaebols (e.g., Lee, 2002; Byun et al., 2012, 2018) by exploring the external governance mechanisms of market competition in the Korean market. Based on evidence from the Korean market, this study highlights the importance of market competition in developing countries where legal protection for shareholders and internal governance structures are weak, ultimately providing an insight for policy officials regarding the regimes of product market competition.

2. Theory and Hypotheses

Pay without performance (particularly downward-sticky pay in underperforming firms) might be driven by management’s use of power to influence the board of directors inappropriately when its members are fulfilling their role by setting compensation. Many studies have provided an extensive account of how managerial influence shapes the executive compensation landscape, and how its influence on remuneration might impose substantial costs (Bertrand & Mullainathan, 2000a, 2000b; Bebchuk et al., 2002; Bebchuk & Fried, 2004; J. Kim et al., 2017; Kwon et al., 2017; Yang & Mo, 2018; Olaniyi & Olayeni, 2020; Yang et al., 2020; Sun et al., 2024).

As a number of researchers have recognized, pay-without-performance schemes reflect managerial slack and rent-seeking rather than providing incentives for efficiency (e.g., Blanchard et al., 1994; Yermack, 1997; Bertrand & Mullainathan, 2000b). Related studies suggest the pervasiveness of downward-sticky pay by focusing on the asymmetric relationship between operational performance and CEO pay (Gaver & Gaver, 1998; Jackson et al., 2008; J. Kim et al., 2017).

Meanwhile, Dechow (2006) discussed Leone et al. (2006) and found an association between the asymmetric sensitivity of CEO cash compensation and stock returns; this was explained by the managerial power and rent extraction perspective, by which management has the power to influence their own compensation and to use that power to extract rent. Though it is widely assumed that pay-for-performance schemes based on optimal contracts encourage executives’ efforts, we are concerned with whether downward-sticky pay allows them to work passively and thus negatively affects shareholder wealth.

So, we first explore whether downward-sticky pay schemes allow executives to work passively with sufficient slack, thus affecting shareholder wealth negatively. In other words, downward-sticky pay not only increases excessive pay but also fails to motivate executives to work hard, thereby reducing shareholder wealth. We therefore propose the following:

Hypothesis 1:

Downward-sticky CEO pay is negatively associated with firm value.

The motivating role of industry product competition has been debated. A stream of studies argue that competition substitutes for managerial incentives, acting as a disciplinary mechanism, and reduces managerial slack (e.g., Machlup, 1967; Hart, 1983; Schmidt, 1997; Raith, 2003; Beiner et al., 2011; Chang et al., 2015; Noghani & Noghanibehambari, 2019). Relevantly, Hart (1983) showed that higher competition provided stronger implicit managerial incentives, as additional market players made firms more informed, and thus better able to evaluate managers’ actions. Schmidt (1997) showed that market competition increased the likelihood of liquidation, thereby forcing managers to work harder to retain their jobs. On the other hand, other studies argue that the relationship between market competition and managers’ effort level is negative or ambiguous (e.g., Scharfstein, 1988; Hermalin, 1992; Graziano & Parigi, 1998; Karuna, 2007; Golan et al., 2015). For instance, Karuna (2007) uses three determinants of competition to show that firms have provided strong incentives when industry competition has been greater, therefore suggesting a complementary relationship between competition and managerial incentives. Golan et al. (2015) find that market competition makes it more difficult to infer a manager’s action given the realized output, thus increasing managerial slack.

The argument between these two conflicting theories hinges on whether the threat of liquidation and CEO dismissal, caused by increasing competition, affects the level of managerial effort, and thus improves shareholder wealth. Neither theory looks at the risk of bankruptcy, nor the effectiveness (optimality) of the compensation scheme, which has been thus far assumed to be the catalyst for increased managerial effort.

We posit that increased competition when a firm underperforms encourages managers to ramp up their efforts to avoid liquidation and retain their positions. This might be especially true with downward-sticky pay-for-performance schemes, which result in less-than-ideal levels of management effort. Our rationale is that higher competition increases the potential for a CEO’s dismissal, therefore causing executives to concentrate on the interests of shareholders, while also restraining the value-deterioration effect of moral hazard problems that come with sticky CEO compensation, such as “pay without performance”. Therefore, even if downward-sticky pay during poor performance negatively affects stock returns, market competition acts as a substitute for executive incentives and thereby alleviates the potential undesirable effect of sticky compensation on shareholder wealth. Consequently, we suggest the following hypothesis:

Hypothesis 2:

A higher level of market competition mitigates the negative effect of downward-sticky pay on firm value.

3. Variables, Models and Data

3.1. Main Variables

3.1.1. Estimating CEO Pay Stickiness

Studies on cost asymmetry have shown that the degree of cost stickiness can be estimated at the firm level (e.g., Homburg & Nasev, 2008; Weiss, 2010). Homburg and Nasev (2008) partition their sample into sticky and non-sticky firms to measure cost stickiness as the cost ratio conditional on decreasing sales and on costs falling proportionately to sales. Weiss (2010) estimates the difference between the rate of cost decrease for recent quarters with decreasing sales and the corresponding rate of cost increases for recent quarters with increasing sales by utilizing quarterly financial data. However, since CEO compensation information is provided annually, the measure of Weiss (2010) is not suitable for our analysis.

We estimate firm-level compensation stickiness following Homburg and Nasev (2008). As described in Equation (1), we first measure the CEO pay ratio, representing changes in CEO compensation relative to changes in accounting performance. Following Sloan (1993) and Leone et al. (2006), we employ ROA as the accounting performance measure, where ROA is defined as net income scaled by beginning-of-year total assets. If CEO compensation increases (decreases) scaled by ROA, even though ROA decreases (increases), the CEO pay ratio will show a positive (negative) number. Further, we measure the degree of compensation stickiness by capturing both the decrease in ROA and positive Pay_Ratio. Therefore, the degree of compensation stickiness is determined by multiplying the indicator variable of decreasing ROA (Decrease) and the positive CEP pay ratio (Pay_Ratio ∗ I_Pay_Ratio) as shown in Equation (2):

Pay_Ratio_it: (Total_Pay_i,t/Roa_i,t) − (Total_Pay_i,t−1/Roa_i,t−1)

(1)

Stickiness_it = Pay_Ratio_i,t ∗ I_Pay_Ratio_i,t ∗ Decrease_i,t

(2)

where I_Pay_Ratio_it is 1 if Pay_Ratio is positive and 0 otherwise; Decrease_it is 1 if Roa is smaller than that of the previous year and 0 otherwise.

3.1.2. Measuring Market Competition

We first employ the Herfindahl–Hirschman Index (HHI) as a proxy of industry competition following prior literature (e.g., Harris, 1998; DeFond & Park, 1999; Engel et al., 2003). Here, we compute the HHI to represent the sum of the squared market shares (in percentages) of all firms in an industry (using four-digit SIC codes).

Regarding the threshold used to classify high versus low market competition, prior studies have classified the degree of market competition based on the Merger Guidelines of the U.S. Antitrust Division of the Department of Justice (Ho et al., 2012; Geiger & Schiereck, 2014; J. Kim et al., 2017; Ryu et al., 2017, 2020; Amini et al., 2024). Following the recent guidelines (December 2023), we define industries with an HHI greater than 0.18 as less competitive. In other words, the degree of market competition is higher if the value of the HHI is lower.

Next, we follow the economics literature to measure industry-level competition advertising and R&D-to-sales ratio (Stigler, 1968; Schmalensee, 1992; Chen et al., 2015), where higher ratios correspond to a greater extent of non-price competition. The economics literature provides theoretical arguments for the use of industry-level advertising and R&D-to-sales ratios as measures of non-price competition. Advertising expenses can induce initial trials of high-quality products and services by customers, but to generate repeat sales, non-price competition industries must also spend heavily on new product development with associated R&D spending (Schmalensee, 1978).

3.2. Models

To investigate the existence of downward-sticky CEO pay, we constructed an empirical model as presented in Equation (3). This model, following Leone et al. (2006), demonstrates an association between the asymmetric sensitivity of CEO compensation and a firm’s operating performance. When an α₃ coefficient is conditional on a significant and positive α₁ coefficient, then a significantly negative α₃ coefficient suggests that CEO pay is sticky, which means that compensation reacts by a smaller degree when ROA decreases than when ROA increases.

With regard to the main controls, prior research (Rosen, 1982; Smith & Watts, 1992; Carter et al., 2007; J. Kim et al., 2017, among others) shows that larger firms (Asset) and firms with great cash flow (CFO) and profitability (ROA) pay higher compensation. Also, we expected that CEO compensation would be high when the firm’s financial risks (Leverage) are high (Cyert et al., 2002). Asset intensity (INT), capturing the degree to which managers allocate firm resources to specific business areas, is also included as a control variable, given its potential associated with variations in compensation structure (Anderson et al., 2003). Market type (Market) is included to capture structural differences in the market (KOSPI vs. KOSDAQ) environment (Yoo & Chun, 2023). We include year and industry dummies in all regressions in this paper.

Δln(EXECcomp)_it = α₀ + α₁ΔROA_it + α₂ROA_Decrease_ii + α₃ΔROA_it ∗ ROA_Decrease_ii
+ α₄ln(Asset)_it−1 + α₅ CFO_it + α₆ROA_it + α₇Leverage_it + α₈INT_it
+ α₉Market_it + Fixed Effects + ε_it

(3)

Δln(EXECcomp)	Natural logarithm of total executive compensation/lagged total executive compensation.
ΔROA	Changes in return on assets, which is operating income scaled by beginning-of-year total asset.
ROA_Decrease	Indicator variable, 1 if ROA is smaller than that of the previous year.
ln(Asset)	Natural logarithm of total assets/lagged total assets.
CFO	Operating cash flow scaled by beginning-of-year total asset.
ROA	Return on assets, which is operating income scaled by beginning-of-year total asset.
Leverage	Total debt to total assets ratio/lagged total debt to total assets ratio.
INT	Asset concentration, end-of-year total asset divided by sales.
Market	Indicator variable, 1 if observation firm-year is in KOSPI, and 0 otherwise.

Next, we used the empirical model to analyze the effect of downward-sticky pay on stock returns as presented in Equation (4). We modified the Easton and Harris (1991) regression (common in empirical literature on the value relevance of accounting) by adding a compensation stickiness as a dummy variable (stickiness), or as a magnitude of compensation stickiness (stickiness), respectively. This is in accordance with Homburg and Nasev (2008), who suggested using a dummy variable not only for compensation stickiness, but also for the magnitude of compensation stickiness. In Equation (4), the negative γ₁ coefficient reflects the negative effect of downward-sticky pay on stock returns.

Return_it = γ₀ + γ₁Stickiness_it + γ₂Competition_it + γ₃ROA_it + γ₄ΔROA_it + γ₅ln(Asset)_it−1
+ γ₆ Return_Vol_it + γ₇Leverage_it + γ₈Market_it + Fixed Effects + ε_it

(4)

Return	Cumulative abnormal return, total of monthly abnormal return using market-adjusted model.
Stickiness	Indicator variable, 1 if total CEO compensation scaled by ROA (=EXECcomp/ROA) increases while ROA decreases, and 0 otherwise.
ln(StickyLevel)	Natural logarithm of (Stickiness* Change in EXECcomp/ROA), where the result using ln(StickyLevel) is presented in Appendix B.
Competition	Indicator variable, 1 if HHI is less than 0.18, and 0 otherwise.
ROA	Return on assets, which is operating income scaled by beginning-of-year total asset.
ΔROA	Changes in return on assets, which is operating income scaled by beginning-of-year total asset.
ln(Asset)	Natural logarithm of total assets.
Return_Vol	Standard deviation of monthly abnormal return.
Leverage	Total debt to total assets ratio.
Market	Indicator variable, 1 if observation firm-year is in KOSPI, and 0 otherwise.

Furthermore, we developed Equation (5) to test whether higher industry competition reduces the negative effect of downward-sticky pay on firm value.

Return_it = γ₀ + γ₁Stickiness_it + γ₂Competition_it + γ₃Stickiness_it ∗ Competition_it + γ₄ROA_it           
+ γ₅ΔROA_it + γ₆ln(Asset)_it−1 + γ₇Return_Vol_it + γ₈Leverage_it + γ₉Market_it       
+ Fixed Effects + ε_it                              

(5)

Return	Cumulative abnormal return, total of monthly abnormal return using market-adjusted model.
Stickiness	Indicator variable, 1 if total CEO compensation scaled by ROA (=EXECcomp/ROA) increases while ROA decreases, and 0 otherwise.
ln(StickyLevel)	Natural logarithm of (Stickiness* Change in EXECcomp/ROA), where the result using ln(StickyLevel) is presented in Appendix B.
Competition	Indicator variable, 1 if HHI is less than 0.18, and 0 otherwise.
ROA	Return on assets, which is operating income scaled by beginning-of-year total asset.
ΔROA	Changes in return on assets, which is operating income scaled by beginning-of-year total asset.
ln(Asset)	Natural logarithm of total assets.
Return_Vol	Standard deviation of monthly abnormal return.
Leverage	Total debt to total assets ratio.
Market	Indicator variable, 1 if observation firm-year is in KOSPI, and 0 otherwise.

According to Equation (5), if market competition mitigates the negative effect of downward-sticky pay on firm value, then the γ₃ coefficient is expected have a significant and positive value, while the γ₁ coefficient would maintain a negative value representing the deteriorating effect of downward-sticky pay on stock returns.

Regarding the main controls in Equations (4) and (5), the inclusion of the industry indicator variables controls for idiosyncratic returns within each industry group over the relevant period, and the inclusion of the time indicator variables allows us to measure returns relative to the average return in the sample over the same time period. We include the accounting performance measures (ROA, ΔROA), firm size (Asset), standard deviation of the return over the previous twelve months (Return_Vol), and leverage (Leverage) as independent variables because prior studies have indicated that risk and size are potentially important determinants of firm performance (see Core et al., 1999). Also, market type (Market) is included to capture structural differences in the market (KOSPI vs. KOSDAQ) environment (Yoo & Chun, 2023). All variables and their measurement are presented in

Table 1. Variables.

Dependent variables
EXEComp	Total compensation per executive.	Total executives compensation/Number of executives
Δ ln(EXECcomp)	Change in natural logarithm of total executive’s annual total compensation per executive.	Natural logarithm of (EXECcompit/EXECcompit−1)
Monthly Return	Monthly realized return.	(Pit − Pit−1)/Pit−1 where Pi,t is the closing price at the end of the month t
Return	Cumulative abnormal return, total of monthly risk-adjusted abnormal returns using market model.	∑(Monthly Returnitm − Monthly Market Index Returntm).
Independent variables
Stickiness	Dummy variable of compensation stickiness.	1 if EXECcompit/ROAit − EXECcompi,t−1/ROAit−1 > 0 while ROAt decreases, and 0 otherwise,
StickyLevel	Level of stickiness of compensation.	|EXECcompit/ROAit − EXECcompit−1/ROAit−1| when Stickiness value = 1.
Market Competition
HHI	Herfindahl–Hirschman Index (HHI).	The sum of the squared market shares (in percentages) of all firms in an industry (using 4-digit SIC codes).
Competition	Indicator of market competition.	1 if HHI <= 0.18, and 0 otherwise
EXPtoSALES	Advertising and R&D intensity.	Sum of firm-level advertising expense and R&D expense divided by the sum of firm-level Sales in industry (SIC digit-4),
Controls
ROA	Return on assets.	Operating Incomeit/Total Assetit−1
ΔROA	Changes in ROA.	ROAit − ROAit−1
Decrease	Indicator of decrease in ROA.	1 if ROAit < ROAit−1, and 0 otherwise,
ln(Asset)	Natural logarithm of total assets.	ln(Total Assetit)
CFO	Operating cash flow scaled by beginning of total asset.	Operating Cash Flowit/Total Assetit−1
Return_Vol	Standard deviation of monthly abnormal return.	σ(Monthly abnormal returnsitm)
Leverage	Total debt divided by total assets.	(Debti,t/Total Assetit)
INT	Asset intensity.	Total Assetit/Salesit
Market	Market dummy variable.	1 if observation firm-year is in KOSPI, and 0 otherwise,

Note: This table provides definitions of the variables in this study. Firm characteristics are collected from the Data Guide and cover 2018 to 2022. Firm industry is classified based on four-digit industry classification codes.

.

3.3. Data

Our sample consists of Korea’s KOSPI and KOSDAQ firms during the period 2018–2022.3 The financial data, with identical four-digit SIC codes, were obtained from the Data Guide. The final sample consists of 6524 firm-year observations taken from 1472 firms covering 231 industries. We exclude all financial firms (SIC codes 6400–6699) due to their distinct regulatory environment. We also exclude firms whose fiscal years do not end in December in order to maintain sample homogeneity. To reduce the effect of outliers, we eliminated Δln(EXECcomp) and ΔROA variables at the 1st and 99th percentiles. We include in the sample only those firms for which four-digit SIC codes are available, in order to assess product market competition.

Table 2. Sample selection.

Firm-year observations in Data Guide	16,705
Less: Financial firms (SIC two-digit codes 64, 65, 66)	(627)
Less: Observations with fiscal year-ends other than December	(140)
Less: Observations with missing financial data in Data Guide	(6848)
Less: Top and bottom 1% observations per year from Δ ln(EXECcomp) and ΔROA	(307)
Less: Observations without SIC four-digit code	(2259)
Final sample (firm-year observations between 2018 and 2022)	6524

summarizes our sample selection procedures, starting with 16,705 firm-year observations.

4. Empirical Results

4.1. Statistics and Pay Stickiness Test

Table 3. Descriptive statistics.

Variables	N	Mean	Standard Deviation	Min	25%	Median	75%	Max
ln(EXECcomp)	6524	19.37	0.83	15.65	18.87	19.35	19.81	23.35
Return	6524	0.01	0.47	−1.78	−0.27	−0.05	0.20	5.99
ROA_Decrease	6524	0.52	0.50	0.00	0.00	1.00	1.00	1.00
Stickiness	6524	0.38	0.49	0.00	0.00	0.00	1.00	1.00
ln(StickyLevel)	6524	8.43	10.73	0.00	0.00	0.00	21.18	28.42
Competition	6524	0.42	0.49	0.00	0.00	0.00	1.00	1.00
1/HHI	6524	6.77	6.89	1.00	2.22	4.35	7.64	33.02
EXPtoSALES	6524	0.04	0.06	0.00	0.01	0.02	0.04	0.62
ROA	6524	0.03	0.09	−0.64	0.00	0.03	0.07	0.96
ΔROA	6524	0.00	0.06	−0.32	−0.03	0.00	0.02	0.32
ln(Asset)i−1	6524	26.37	1.48	22.58	25.34	26.09	27.14	33.09
Return_Vol	6524	0.12	0.08	0.02	0.07	0.10	0.15	1.77
Leverage	6524	0.42	0.22	0.02	0.25	0.41	0.57	7.37
Market	6524	0.41	0.49	0.00	0.00	0.00	1.00	1.00

Note: This table summarizes the descriptive statistics of the variables investigated for the empirical analysis of this study. The sample consists of 6524 firm-year observations of Korea’s KOSPI and KOSDAQ firms covering fiscal years 2018 to 2022. Firm financial data are taken from Data Guide. Please refer to Table 1 for a detailed explanation of the variables.

provides detailed descriptive statistics for the variables of executive compensation per person and its stickiness, market competition, and firm characteristics for our sample (consisting of 6524 firm-year observations of KOSPI and KOSDAQ firms for 2018–2022). The mean (median) executive total compensation (EXECcomp) is KRW 381,000,000 (KRW 253,000,000), and the mean (median) cumulative abnormal return (Return) is 1% (−5%). The average degree of ln(StickyLevel) was 8.43 and our sample represents 38% of the compensation-sticky firm-years. Among 52% of observations that showed decreasing ROA, approximately 73% of the firm-years have sticky compensation schemes.

Also, we test the asymmetric relationship between executive compensation and corporate performance and confirm the pervasiveness of downward-sticky pay, as shown in

Table 4. Estimate of downward-sticky pay.

Independent Variables	Predicted Sign	Dependent Variable = Δln(EXECcomp)
		Estimated Coefficient	T-Statistics
Constant		−3.0198 ***	(−24.46)
ΔROA	+	0.4895 ***	(3.14)
ROA_Decrease	-	−0.1029 ***	(−5.91)
ΔROA * ROA_ Decrease	-	−0.4717 **	(−2.16)
ln(Asset)		0.0722 ***	(3.03)
CFO		0.0874	(1.38)
ROA		−0.1322	(−1.61)
Leverage		−0.0682 *	(−1.71)
INT		0.0131	(1.62)
Market		−0.0473 ***	(−4.21)
Fixed Effects		Year and Industry (SIC 2)
Observations		6524
Adjusted R2		0.1093

Note: T-statistics are reported in parentheses under each estimated coefficient. To mitigate any influence from outliers, Δln(EXECcomp) and ΔROA are eliminated at the top and bottom first percentile. Dependent variables: Δln(EXECcomp) in Equation (3). The independent variable of interest ROA_Decrease is an indicator variable set at 1 if the firm’s ROA is negative and 0 otherwise. Please refer to Table 1 for a detailed explanation of the variables. The symbols *, **, and *** correspond to 10, 5, and 1% significance levels for p-value.

.

4.2. Effects of Downward-Sticky Pay and Market Competition

In

Table 5. Effects of pay stickiness and market competition.

Independent Variables	Predicted Sign	Dependent Variable = Return
		OLS Model (4)		OLS Model (5)
		Estimated Coefficient	T-Statistics	Estimated Coefficient	T-Statistics
Constant		−0.6893 ***	(−6.61)	−0.6827 ***	(−6.65)
Stickiness	-	−0.0244 **	(−2.56)	−0.0392 ***	(−3.3)
Competition	-	0.0073	(−0.99)	−0.0064	(−0.44)
Stickiness * Competition	+			0.0366 **	(2.11)
ROA	+	0.8763 ***	(15.62)	0.8753 ***	(15.53)
ΔROA		0.6538 ***	(7.88)	0.6632 ***	(8.07)
ln(Asset)		0.0077 **	(1.88)	0.0077 **	(1.98)
Return_Vol		4.0757 ***	(76.03)	4.0758 ***	(76.15)
Leverage		−0.0647 ***	(−3.07)	−0.0639 ***	(−3.06)
Market		0.0446 ***	(4.05)	0.0445 ***	(4.03)
Fixed Effects		Year and Industry (SIC 2)		Year and Industry (SIC 2)
Observations		6524		6524
Adjusted R2		0.5027		0.5030

Note: T-statistics are reported in parentheses under each estimated coefficient. To mitigate any influence from outliers, Δln(EXECcomp) and ΔROA are eliminated at the top and bottom first percentile. Dependent variable in both models is cumulative abnormal return, total of monthly abnormal return using market-adjusted model. The independent variable of interest in Equation (2) is Stickiness, an indicator variable, which takes a value of ‘1’ if EXECcomp_i,_,t/ROA_i,t increases from the previous year while ROA_i,t decreases, and 0 otherwise. In Equation (3), the interaction term Stickiness * Competition captures the incremental effect of market competition on stock return. The proxy for market competition, Competition, is calculated as 1 if the Herfindahl–Hirschman Index (HHI) is lower than 0.18, and 0 otherwise, where HHI is the sum of the squared market shares (in percentages) of all firms in an industry (using 4-digit SIC codes). Please refer to Table 1 for a detailed explanation of the variables. The symbols **, and *** correspond to 5, and 1% significance levels for p-value.

, we report the OLS regression estimates for Equations (4) and (5), which examine the impact of sticky compensation on stock performance and whether a higher level of market competition mitigates the potential negative link between sticky executive pay and firm value. The results demonstrate the negative coefficients for the dummy variable Stickiness and the degree of stickiness ln(StickyLevel).4 Specifically, the coefficient for Stickiness is significant and negative, indicating that downward-sticky pay has a significantly negative association with annualized cumulative abnormal returns.

In the second column of Table 5, the results shown from testing Equation (5) indicate that the coefficient of this interaction term (Stickiness*Competition) has a significantly positive value (coefficient = 0.0366, t-stat. = 2.11) and suggests that the value-deteriorating effect of downward-sticky pay decreases in an environment with higher competition.

4.3. Portfolio Sorting Approach (By Market Competition Level)

To additionally investigate whether market competition curbs the negative effect of downward-sticky pay on stock performance, we sort groups by market competition level and perform the OLS regression estimates for Equation (4) separately. In

Table 6. Portfolio sorting test.

Independent Variables	Predicted Sign	Dependent Variable = Return
		High Competition (HHI <= 0.18)		Low Competition (HHI > 0.18)
		Estimated Coefficient	T-Statistics	Estimated Coefficient	T-Statistics
Constant		−0.3040	(−1.4)	−0.8392 ***	(−7.13)
Stickiness	-	−0.0109	(−0.68)	−0.0323 ***	(−2.75)
Competition	-	−0.0013	(−0.92)	−0.0103 **	(−2.15)
ROA		0.8531 ***	(9.94)	0.9453 ***	(12.29)
ΔROA		0.5003 ***	(4.05)	0.7957 ***	(7.23)
ln(Asset)		−0.0113	(−1.48)	0.0161 ***	(3.55)
Return_Vol		4.1133 ***	(46.55)	4.0504 ***	(61.05)
Leverage		−0.0435	(−1.43)	−0.0924 ***	(−3.14)
Market		0.0486 **	(2.46)	0.0385 ***	(2.90)
Fixed Effects		Year and Industry (SIC 2)		Year and Industry (SIC 2)
Observations		2736		3788
Adjusted R2		0.4692		0.5397

Note: T-statistics are reported in parentheses under each estimated coefficient. To mitigate any influence from outliers, Δln(EXECcomp) and ΔROA are eliminated at the top and bottom first percentile. Dependent variable in this model is cumulative abnormal return, total of monthly abnormal return using market-adjusted model. The independent variable of interest in Equation (4) is Stickiness, an indicator variable, which takes a value of ‘1’ if EXECcomp_i,t/ROA_i,t increases from the previous year while ROA_i,t decreases, and 0 otherwise. In this table, portfolios are constructed based on higher market competition (HHI <= 0.18) and lower market competition (HHI > 0.18) groups. The symbols **, and *** correspond to 5, and 1% significance levels for p-value.

, the first column (higher-competition group) shows that the coefficient for Stickiness is insignificant, thus implying the mitigating effect of market competition. In contrast, the second column (lower-competition group) exhibits that the coefficient for Stickiness is significantly negative (coefficient = −0.0109, t-stat. = −0.68) under lower market competition. This presents that downward-sticky pay negatively affects stock performance where market competition is low, whereas its effect is not significant under high market competition. This result may be in line with our predictions that downward-sticky pay reduces firm value (Hypothesis 1) and market competition alleviates this negative effect (Hypothesis 2).

4.4. Alternative Tests of Market Competition

We undertake the alternative test by using industry-level competition advertising and the R&D-to-sales ratio (Chen et al., 2015). In

Table 7. Alternative proxies of market competition.

Independent Variables	Predicted Sign	Dependent Variable = Return
		OLS Model (4)		OLS Model (5)
		Estimated Coefficient	T-Statistics	Estimated Coefficient	T-Statistics
Constant		−0.6834 ***	(−6.69)	−0.6779 ***	(−6.63)
Stickiness	-	−0.0245 **	(−2.55)	−0.0353 ***	(−3.33)
EXPtoSales	-	0.0912	(0.50)	−0.0483	(−0.25)
Stickiness * EXPtoSales	+			0.3239 **	(2.38)
ROA	+	0.8793 ***	(15.62)	0.8799 ***	(15.65)
ΔROA	+	0.6521 ***	(7.88)	0.6685 ***	(8.13)
ln(Asset)		0.0074 *	(1.88)	0.0074 *	(1.91)
Return_Vol		4.0741 ***	(76.03)	4.0737 ***	(76.08)
Leverage		−0.0642 ***	(−3.07)	−0.0638 ***	(−3.06)
Market		0.0447 ***	(4.05)	0.0447 ***	(4.05)
Fixed Effects		Year and Industry (SIC 2)		Year and Industry (SIC 2)
Observations		6524		6524
Adjusted R2		0.5027		0.5030

Note: These models are modified from both Equation (4) and Equation (5). Instead of Competition, we use the alternative measure of market competition, denoted as EXPtoSales. Column (1) presents the results of the replaced Equation (4), and Column (2) presents the results of the replaced Equation (5). For the details on EXPtoSales variable, see Table 1. The symbols *, **, and *** correspond to 10, 5, and 1% significance levels for p-value.

, we present the results from the alternative competition measurement when we replaced the HHI in Equations (4) and (5) with EXPtoSALES. The coefficients of the interaction term of Stickiness * EXPtoSALES variables are found to be significantly positive (coefficient = 0.3239, t-stat. = 2.38), consistent with the results presented in Table 4.

In

Table 8. Alternative proxies of market competition (portfolio sorting test).

Independent Variables	Predicted Sign	Dependent Variable = Return
		High Competition (EXPtoSALES > Median)		Low Competition (EXPtoSALES <= Median)
		Estimated Coefficient	T-Statistics	Estimated Coefficient	T-Statistics
Constant		−0.3867 **	(−2.11)	−0.7962 ***	(−6.27)
Stickiness	-	−0.0215	(−1.41)	−0.0287 **	(−2.35)
EXPtoSales	-	0.2969	(1.26)	−2.2499	(−1.49)
ROA	+	0.8239 ***	(11.34)	1.1284 ***	(11.53)
ΔROA	+	0.6069 ***	(5.48)	0.6105 ***	(4.77)
ln(Asset)		−0.0066	(−0.94)	0.0146 ***	(3.01)
Return_Vol		3.9804 ***	(47.99)	4.1412 ***	(59.4)
Leverage		−0.0191	(−0.64)	−0.1172 ***	(−3.89)
Market		0.0692 ***	(3.52)	0.0321 **	(2.45)
Fixed Effects		Year and Industry (SIC 2)		Year and Industry (SIC 2)
Observations		2941		3583
Adjusted R2		0.4714		0.5390

Note: This table is an additional analysis of Table 5, where EXPtoSales replaces Competition in Equation (4), and portfolios are constructed using median of EXPtoSales. Here, if EXPtoSales is greater than the annual median of EXPtoSales, the firm-year is classified into the higher-competition group, and if the EXPtoSales is less than the annual median of EXPtoSales, the firm-year is classified into the lower-competition group. The symbols **, and *** correspond to 5, and 1% significance levels for p-value.

, we present the results from the alternative competition measurement when we replace the HHI in Equation (4) with EXPtoSALES. The coefficients of Stickiness in first column (high competition) are insignificantly negative (coefficient = −0.0215, t-stat. = −1.41), while the coefficients of Stickiness in second column (low competition) are significantly negative (coefficient = −0.0287, t-stat. = −2.35). They are consistent with the results presented in Table 5.

4.5. Endogeneity Effects

To mitigate concerns of endogeneity, we use two econometric approaches. First, we estimate firm fixed-effect regressions, which control for time-invariant unobservable firm-specific factors correlated with both firm performance and the compensation structure. Column (2) in

Table 9. Effects of market competition (using firm fixed effect).

Independent Variables	Predicted Sign	Dependent Variable = Return
		OLS Model (4)		OLS Model (5)
		Estimated Coefficient	T-Statistics	Estimated Coefficient	T-Statistics
Constant		4.8373 ***	(7.92)	4.8488 ***	(7.94)
Stickiness	-	−0.0306 ***	(−2.87)	−0.0473 ***	(−3.58)
Competition	-	−0.0290	(−0.80)	−0.0447	(−1.21)
Stickiness * Competition	+			0.0413 **	(2.14)
ROA	+	0.6399 ***	(5.70)	0.6438 ***	(5.73)
ΔROA	+	0.5774 ***	(5.74)	0.5848 ***	(5.81)
ln(Asset)		−0.2023 ***	(−8.70)	−0.2026 ***	(−8.71)
Return_Vol		4.5369 ***	(72.18)	4.5366 ***	(72.20)
Leverage		−0.0322	(−0.57)	−0.0319	(−0.56)
Market		−0.0634	(−0.38)	−0.0581	(−0.35)
Fixed Effects		Firm and Year		Firm and Year
Observations		6524		6524
R2 (within)		0.5423		0.5428

Note: This table provides an additional analysis based on Table 4. To control for endogeneity, we replace the industry fixed effect with the firm fixed effect. The symbols **, and *** correspond to 5, and 1% significance levels for p-value.

presents the estimation results using a firm fixed-effect regression. After including firm fixed effects, we confirm that firms with sticky compensation structures in harshly competitive industries exhibit higher market returns.

Secondly, to mitigate the endogeneity issue, we use the Propensity Score Matching (PSM) method. This approach reduces the differences between the two groups (treatment variables) to address endogeneity. Here, we apply 1:1 matching based on propensity scores to create a new sample by matching firms in the high-competition group with those in the low-competition group.5 Column (2) in

Table 10. Effect of market competition (using Propensity Score Matching (PSM)).

Independent Variables	Predicted Sign	Dependent Variable = Return
		OLS Model (4)		OLS Model (5)
		Estimated Coefficient	T-Statistics	Estimated Coefficient	T-Statistics
Constant		−0.7487 ***	(−4.73)	−0.7400 ***	(−4.68)
Stickiness	-	−0.0338 **	(−2.57)	−0.0608 ***	(−3.07)
Competition	-	0.0047	(0.25)	−0.0153	(−0.74)
Stickiness * Competition	+			0.0542 **	(2.32)
ROA	+	0.8784 ***	(9.75)	0.8778 ***	(9.79)
ΔROA	+	0.6297 ***	(5.15)	0.6365 ***	(5.22)
ln(Asset)		0.0103 *	(1.68)	0.0103 *	(1.69)
Return_Vol		4.0383 ***	(29.85)	4.0416 ***	(30.01)
Leverage		−0.0898 ***	(−2.84)	−0.0886 ***	(−2.80)
Market		0.0381 **	(2.38)	0.0375 **	(2.33)
Fixed Effects		Year and Industry (SIC 2)		Year and Industry (SIC 2)
Observations		4427		4427
R2		0.4799		0.4807

Note: This table provides an additional analysis based on Table 4. To control for endogeneity, we apply the Propensity Score Matching (PSM) method. Here, we apply 1:1 matching between the high-competition and low-competition groups based on propensity scores. Specifically, we use the following model: Competition_it = β₀ + β₁ln(Sales)_it + β₂ln(Asset)_it−1 + β₃ln(Equity)_it + β₄ln(Marke Value)t_it + θ_it. The symbols *, **, and *** correspond to 10, 5, and 1% significance levels for p-value.

presents the regression results using the 1:1 matched sample. After applying the Propensity Score Matching method, we confirm that firms with sticky compensation structures in highly competitive industries exhibit higher market returns.

4.6. A Sensitivity Test Omitting COVID-19 Effects

To exclude the possibility that our results were produced due to influence from COVID-19, we eliminated observations from COVID-19-affected years and retested the hypotheses. As shown in

Table 11. Effect of market competition (omitting COVID-19-affected years).

Independent Variables	Predicted Sign	Dependent Variable = Return
		OLS Model (4)		OLS Model (5)
		Estimated Coefficient	T-Statistics	Estimated Coefficient	T-Statistics
Constant		−0.5045 ***	(−4.50)	−0.4996 ***	(−4.46)
Stickiness	-	−0.0131	(−1.24)	−0.0291 **	(−2.23)
Competition	-	0.0067	(0.48)	−0.0080	(−0.51)
Stickiness * Competition	+			0.0395 **	(2.08)
ROA	+	0.7730 ***	(12.74)	0.7718 ***	(12.72)
ΔROA	+	0.7340 ***	(8.22)	0.7458 ***	(8.33)
ln(Asset)		0.0022	(0.52)	0.0023	(0.54)
Return_Vol		3.8650 ***	(65.68)	3.8654 ***	(65.71)
Leverage		0.0365 ***	(−2.93)	0.0361 ***	(−2.90)
Market		0.0365 ***	(3.03)	0.0361 ***	(2.99)
Fixed Effects		Year and Industry (SIC 2)		Year and Industry (SIC 2)
Observations		5219		5219
Adjusted R2		0.4843		0.4846

Note: This table presents the retest result regarding Table 5. To eliminate the effects of COVID-19, we excluded the year 2020 from our analysis. In 2020, Korea experienced negative economic growth due to the impact of the pandemic. However, the economy fully recovered in 2021, with nominal GDP reaching KRW 2221.9 trillion—surpassing the pre-pandemic level of KRW 2040.6 trillion in 2019. The symbols **, and *** correspond to 5, and 1% significance levels for p-value.

, the results are consistent with our main results. Therefore, our empirical results were not driven by potential problems that stem from the effects of COVID-19. South Korea experienced negative GDP growth in 2020, reflecting the economic contraction induced by the pandemic. However, the economy fully recovered in 2021, with GDP levels exceeding those of the pre-pandemic period. This exclusion ensures that our analysis of the relationship between compensation stickiness and stock returns is not confounded by the effects of COVID-19 market movements. As shown in Column (2) of Table 11, our results remain robust after excluding pandemic-affected periods.

5. Conclusions

Using data on the Korean product market, which has been dominated by numerous conglomerates (chaebols), this paper examines whether downward-sticky pay negatively affects firms’ stock returns and, if so, whether market competition mitigates the possibly value-destroying effect of downward-sticky pay. We confirmed that downward-sticky pay is significantly negatively associated with shareholder value. Furthermore, we demonstrated that market competition mitigates the aggravating effect of sticky CEO pay on firm value. These results have all been found to be robust in tests using alternative proxies for market competition. The results imply that market competition alleviates this negative effect of downward-sticky pay on firm value and thus protects shareholder’s interests against managerial rent-extracting. This study highlights the importance of market competition in developing countries where legal protection for shareholders and internal governance structures are weak, ultimately providing an insight for policy officials regarding the regimes of product market competition.