Corporate Governance and Employee Productivity: Evidence from Jordan

Abstract

This research paper aims to investigate the impact of ownership concentration, insider ownership, and board size on employee productivity for 136 Jordanian public shareholding firms listed on the Amman Stock Exchange (ASE) from 2012 to 2021. Ownership concentration has been measured by Herfindahl–Hirschman Index (HHI), whereas insider ownership and board size have been represented as the proportion of shares held by insiders and by the number of board members, respectively. Lastly, employee productivity has been measured using a data envelopment analysis (DEA) tool. We employed ordinary least squares regression (OLS) including firm-year-fixed effects. Our empirical results indicate a non-linear relation between ownership concentration and employee productivity, whereby the productivity of employees increases in firms with a proportion of ownership concentration less than 60%. In addition, we found a non-linear relation between insider ownership and employee productivity, whereby the productivity of employees increases in firms with proportion of insider ownership less than 50%. Moreover, we found a non-linear relation between board size and employee productivity, whereby the productivity of employees increases in firms that have less than 11 board members. Our outcome contributed to the knowledge found in the previous literature, as it is the first to highlight the productivity of employees in emerging economies, such as the economy in Jordan. Furthermore, our findings could be useful for the Jordan Securities Commission (JSC) and the ASE on their continuous process to improve and develop corporate governance instructions.

1. Introduction

Corporate governance (CG) has become a crucial issue (Parveen 2021). Therefore, as per the Organization for Economic Cooperation and Development (OECD), many countries and organizations have become interested in implementing governance, as it helps to control corruption by creating an environment of transparency, trust, rule of law, and accountability, thus leading to the protection of financial stability, investments, and an increase in the growth rate (OECD 2015). Akomea-Frimpong et al. (2022) argued that the objective of CG is to improve the monitoring and accountability of managers in order to decrease corruption and increase transparency, which would lead to greater benefits for shareholders. Accordingly, better CG is associated with a firm’s better financial and non-financial performance and value (Aljifri and Moustafa 2007; Tulcanaza-Prieto and Lee 2022; Tulcanaza-Prieto et al. 2020).

This research paper contributes to the previous literature in several ways. To the best of our knowledge, this is the first article to consider employee productivity as a proxy of measuring a firm’s performance in emerging economies and developing countries. Unlike the previous literature (Jaafar and El-Shawa 2009; Qadorah and Fadzil 2018; Sundarasen et al. 2024), which considered both a firm’s profitability and market performance as proxies for measuring its overall performance. In addition, we utilized a data envelopment analysis (DEA) tool to measure employee productivity. DEA is a non-parametric method to measure productivity within a group of homogeneous decision-making units, and to make observations by considering multiple inputs and multiple outputs to evaluate its efficiency (Zhang and Li 2020; Song et al. 2020). This is unlike the previous literature (Ngo and Le 2019; Mia et al. 2023), which employed DEA to measure the financial efficiency of banks. Therefore, this paper aims to bridge the research gap by highlighting the productivity of employees measured by DEA as an important proxy of a firm’s performance in developing countries, with a focus on the ownership structure and board size as significant CG mechanisms. Furthermore, this paper contributes to extend the theoretical perspective of CG and considers financial firms along with non-financial firms, providing that the former also play a considerable role in developing countries such as Jordan.

In this research paper, the reason for selecting and examining the corporate governance in emerging economies, such as the economy of Jordan, is that, unlike developed countries, developing countries such as Jordan are under novel research on corporate governance. Additional reasons include the availability of data, as well as the high ownership concentration characterized by the Jordanian economy, considering that Jordan depends on significant foreign investment (Alzoubi 2016). A supplementary reason is to examine the performance in the post-privatization period, bearing in mind that during the 1990s, the Jordanian government started a privatization process, which was an unprecedented reform to move towards increasing the weight of the private sector in the Jordanian gross domestic product (GDP). As a result, a decrease was observed in the Jordanian government’s shareholding in listed companies from 15% in the 1990s to less than 6% in 2012 (Amman Stock Exchange 2012). Moreover, Jordan is seeking to grow and develop its economy and capital market; hence, the JSC and ASE have improved the transparency and the disclosure of annual reports for the listed firms for the last two decades, starting with the roles of CG implemented in 2002 (Berg and Nenova 2004). New governance instructions were issued in 2017, in which every public firm listed in the ASE became obligated to generate a separate, detailed CG report in their annual disclosures, as published in the ASE by the end of the financial year of each firm (Securities Depository Center 2017). Thus, the findings of this paper would be useful to add value to the aforementioned process.

Accordingly, this research paper aims to measure the impact of two important parts of ownership structure: ownership concentration and insider ownership, along with board size on employee productivity for 136 Jordanian public shareholding firms listed on the ASE, from 2012 to 2021.

The basis of selection for the aforementioned mechanisms in relevance to the CG is due to the significance of large block holders, managerial shareholding power, and board size on the decision-making process and monitoring, which would naturally affect employee productivity and, in turn, reflect on the firm’s overall performance.

The remainder of this paper is structured as follows. Section 2 shows the theoretical framework. Section 3 shows the literature review and hypotheses. Section 4 shows the methodology and sample explanation. Section 5 shows the discussions and results. Finally, Section 6 shows the conclusions and limitations.

2. Theoretical Framework

In this research paper, we consider agency theory and stewardship theory, which are theories of CG. We aim to demonstrate the relationship between the parties involved in the functioning of a firm, which should lead to maximizing the wealth of the shareholders and stakeholders (Squires and Elnahla 2020). Agency theory is the description of protecting the interests of shareholders from the separation between ownership and management functions (Jensen and Meckling 1976). On the other hand, stewardship theory is the description of motivating managers to be executives rather than non-executives by providing them with more power, freedom, and responsibility to achieve the shareholders’ interests (Turnbull 2000).

Agency theory aims to limit any agency problems via CG mechanisms (Panda and Leepsa 2017). Ownership concentration is considered one of the important mechanisms of CG to reduce agency problems by changing the paradigm of the investors; from only a capital supplier to agents with more managerial power and influence in the decision-making process (Reis and Pinto 2021). Likewise, Burkart et al. (2014) argued that ownership structure and large block holders provide powerful and effective monitoring, as well as reduce the possibilities of takeovers. In addition, agency theory refers to the emergent conflict caused by managers who aim to increase their power in the firm. Therefore, the increase in holdings by insiders (board members and executive managers) means an increase in management power, which would lead to more agency problems (Fama and Jensen 1983; Demsetz 1983). Furthermore, agency theory emphasizes the monitoring role of the board of directors on behalf of the owners for the performance of managers to meet the interests of shareholders. Therefore, a larger board of directors is better for monitoring performance (Vitolla et al. 2019). On the other hand, stewardship theory focuses on increasing managerial power and authorities as a motivation to combine their interests with shareholder interest to maximize the wealth of the firm (Ho 2005), in addition to minimizing the agency problems (Jensen and Meckling 1976). Therefore, it facilitates an increase to insider holdings. Demsetz and Villalonga (2001) argued that the increase in the holdings of insiders would not lead to agency problems, because insiders are already representing their institutions as well as outsider shareholders who possess a large block of shares in the firm. Therefore, the insider’s interests are identical to those of the outsider shareholders. Moreover, stewardship theory focuses on effective management. Therefore, there is no need for a large board size as this will decrease the efficiency of the decision-making process, while a smaller board size will enhance performance (Kalsie and Shrivastav 2016). Table 1 summarizes the chosen CG mechanisms in our research paper and their impact on employee productivity based on the above-discussed theories.

Table 1. Corporate governance (CG) mechanisms and theories.

CG Mechanism	Agency Theory	Stewardship Theory
Ownership Concentration	Solution to reduce the agency problems	Focus more in improving the insider ownership
Insider Shareholding	Obstruct the monitoring role of board of directors	Increase managerial power is better for monitoring firm’s performance
Board Size	Bigger size is better for monitoring firm’s performance	Smaller size is better for making efficient decisions

3. Literature Review and Hypotheses

3.1. Employee Productivity

Employee productivity is an important indication of a firm’s performance, as it reflects the work value contributed by each employee (Cortés et al. 2017; Claessens and Djankov 1999). Productivity is defined as the relation between the quality and the size of job tasks performed by employees to achieve the firm’s goals (Rusdiyanto 2021). In addition, it is the process of evaluating the performance of the employees in a similar job by comparing the output between them or the number of units or products that employees could accomplish within a preset schedule (Dorothy et al. 2020). Likewise, it is the level of employee performance based on having accomplished their duties and responsibilities, and it is calculated by dividing the number of products by the company’s input (Saluy et al. 2021). It is considered to be the output divided by the input during a specific period of time (Obeng and Boachie 2018). Edeh and Acedo (2021) identified productivity as the sales per worker, which is the definition that we considered in this research paper. Furthermore, Chiang and Lin (2007) measured productivity by considering labor and capital invested as inputs and sales as an output.

3.2. Ownership Concentration

Ownership concentration is an important mechanism of CG, in which large block holders reduce agency problems. Table 2 summarizes the previous literature of the impact of ownership concentration on a firm’s performance.

Table 2. Literature review of the impact of ownership concentration on firm’s performance.

Author	Sample	Country	Impact
(Horobet et al. 2019)	3506 European firms from 2008 to 2016	Europe	Positive
(Duong et al. 2022)	103 energy firms from 2007 to 2020	Vietnam	Positive
(Omran et al. 2008)	304 firms from 2000 to 2002	Arab countries	Positive
(San Martin-Reyna and Duran-Encalada 2015)	75 listed firms from 2005 to 2011	Mexico	Positive
(Jaafar and El-Shawa 2009)	103 listed firms over from to 2005	Jordan	Positive
(Zeitun and Tian 2007)	59 listed firms from 1989 to 2002	Jordan	Positive
(Drobetz et al. 2019)	126 listed shipping firms from 1997 to 2016	Worldwide	Positive
(Huang 2020)	All listed banks over from to 2018	China	Positive
(Iwasaki and Mizobata 2019)	Meta-synthesis of 1517 estimates 69 studies	Worldwide	Positive
(Claessens and Djankov 1999)	706 firms from 1992 to 1997	Czech	Positive
(Lai et al. 2022)	1658 entrepreneurial firms from 2004 to 2011	Worldwide	No impact
(Demsetz and Lehn 1985)	511 firms from 1976 to 1980	USA	No impact
(Janang et al. 2015)	31 listed firms from 2001 to 2012	Malaysia	No impact

Based on this previous literature, we propose the following hypothesis:

Hypothesis 1.

Ownership concentration is positively related to employee productivity.

3.3. Insider Ownership

Several studies linked insider ownership with a firm’s performance. Table 3 summarizes the previous literature of the impact of insider ownership on a firm’s performance.

Table 3. Literature review of the impact of insider ownership on firm’s performance.

Author	Sample	Country	Impact
(Sheu and Yang 2005)	416 listed electronics firms from 1996 to 2001	Taiwan	Negative
(Han and Suk 1998)	5500 firms from 1988 to 1992	Worldwide	Negative
(Jones and Klinedinst 2012)	490 manufacturing firms from 1997 to 2001	Bulgaria	No impact
(Pant and Pattanayak 2007)	1833 firms for the years 2000, 2001, 2003 and 2004	India	Non-linear *
(Bhabra 2007)	54 firms from 1994 to 1998	New Zealand	Non-linear **
(Jain et al. 2020)	199 firms from 2007 to 2018	India	Non-linear *
(Park and Jang 2010)	251 restaurant firms from 2001 to 2006	Worldwide	Positive
(Rose 2005)	All listed firms from 1998 to 2001	Denmark	Positive
(Hrovatin and Uršič 2002)	488 firms in 1998	Slovenia	Positive

*: Low proportion of insider ownership positively impact firm’s performance; **: The impact is positive when the proportion of insider ownership is below 14% or over 40%.

Based on the previous literature, we propose the following hypothesis:

Hypothesis 2.

Low insider ownership is positively related to employee productivity.

3.4. Board Size

Several studies linked board size with a firm’s performance. Table 4 summarizes the previous literature of the impact of board size on a firm’s performance.

Table 4. Literature review of the impact of board size on firm’s performance.

Author	Sample	Country	Impact
(Bermig and Frick 2010)	294 firms from 1998 to 2007	Germany	No impact
(Guest 2009)	746 listed firms from 1981 to 2002	United Kingdom	Negative
(Cheng et al. 2008)	500 firms from 1984 to 1991	USA	Negative
(Nguyen et al. 2015)	1141 firms from 2001 to 2011	Australia	Negative
(O’Connell and Cramer 2010)	77 listed firms in 2001	Ireland	Negative
(Shahrier et al. 2018)	200 Shariah-compliant listed firms from 2014 to 2017	Malaysia	Negative
(Jaafar and El-Shawa 2009)	103 listed firms from 2002 to 2005	Jordan	Positive
(Kalsie and Shrivastav 2016)	145 non-financial firms from 2008 to 2012	India	Positive

Based on the previous literature, we propose the following hypothesis:

Hypothesis 3.

Large board size is negatively related to employee productivity.

4. Methodology

4.1. Data Collection

Our sample contains firms listed on the ASE in the post-privatization period from 2012 to 2021. Firms for which annual reports are not available during the complete aforementioned period are excluded. Accordingly, the final sample consists of 136 financial and non-financial firms listed on the ASE, out of a total population of 172 listed firms as of 2021. Therefore, the sample represents 79% of the population.

4.2. Definitions and Measures of the Variables

Regarding the dependent variable, employee productivity (EP), we employed a data envelopment analysis (DEA) tool. A DEA is a non-parametric tool to measure the productivity within a group of homogeneous decision-making units by considering multiple inputs and multiple outputs to evaluate efficient productivity (Zhang and Li 2020; Song et al. 2020). Moreover, it is a tool to measure the efficient productivity of the decision-making units (DMUs) based on the proportional change in inputs and outputs (Yong-bae and Choonjoo 2009). Thus, a DEA is one of the most efficient and accurate tools to measure the efficient productivity scale of employees by considering the constant returns to scale (CRS) and the variable returns to scale (VRS), based on the input-oriented DEA model or the output-oriented DEA model. In this paper, we employ the input-oriented DEA model because the goal of this model is to minimize the input to obtain an efficient score of productivity. As shown in Figure 1, the number of employees (EMPNO) is considered as the input, and total sales (SALES) is considered as the output (Chiang and Lin 2007; McConaughy et al. 1998; Tian and Twite 2011).

Figure 1. Number of employees is the input, and total sales is the output of the Data Envelopment Analysis (DEA) model.

Equation (1) shows the input-oriented CRS DEA model, whereby the change in the input results in a constant change in the output:

$$\max \sum _{\mathrm{r}=1}^{\mathrm{m}}{\displaystyle \frac{y_r \mathit{SALES}}{v_r \mathit{EMPNO}}}$$

(1)

Subject to

$$\begin{array}{c}{\displaystyle \sum _{\mathrm{r}=1}^{\mathrm{m}}}{\displaystyle \frac{y_r \mathit{SALES}}{v_r \mathit{EMPNO}}} \le 1 (\mathrm{r}=1,\dots \dots ,\mathrm{m})\\ y_r\ge 0 (\mathrm{r}=1,\dots \dots ,\mathrm{m}) \\ v_r\ge 0 (\mathrm{r}=1,\dots \dots ,\mathrm{m})\end{array}$$

Equation (2) shows the input-oriented VRS DEA model, whereby the change in input may result in either synergistic, constant or antagonistic change in the output:

$$\max {\displaystyle \sum _{\mathrm{r}=1}^{\mathrm{m}}}{y}_r \mathit{SALES}+y0$$

(2)

Subject to

$$\begin{array}{c}{\displaystyle \sum _{\mathrm{r}=1}^{\mathrm{m}}}{y}_r \mathit{SALES} - v_r \mathit{EMPNO}+y0 \le 0 (\mathrm{r}=1,\dots \dots ,\mathrm{m})\\ {\displaystyle \sum _{\mathrm{r}=1}^{\mathrm{m}}}{v}_r \mathit{EMPNO}=1 \\ y_r\ge 0 (\mathrm{r}=1,\dots \dots ,\mathrm{m}) \\ v_r\ge 0 (\mathrm{r}=1,\dots \dots ,\mathrm{m})\end{array}$$

In order to measure (EP) of each DMU, the following Equation (3) is considered (Ngo and Le 2019; Mia et al. 2023; Yong-bae and Choonjoo 2009):

$$EP={\displaystyle \frac{\mathrm{VRS} \mathrm{model}}{\mathrm{CRS} \mathrm{model}}}$$

(3)

where y_r and v_r are the weights of output and input, respectively; EP = Employee productivity; CRS = Constant returns to scale; VRS = Variable returns to scale; m = Number of DMU; SALES = Total sales (output); EMPNO = Number of employees (Input).

In our sample, we have a total of (1360) DMUs, as every DMU represents one firm-year observation. The results for (EP) are between 0 and 1, where 1 is the efficient and most productive DMU, i.e., achieving the highest increment of the proportional change of sales from the lowest increment of the proportional change of employees number. The data has been obtained from the firm’s annual reports and disclosures from the ASE. Figure 2 shows the average employee productivity by year.

Figure 2. Average employee productivity by year.

It is notable that the decrease in employee productivity in 2020 was due to the closing and restrictions that the government implemented to lessen and control the spread of the COVID-19 pandemic.

Regarding the independent variables, the first independent variable is ownership concentration (OC), for which we implemented the Herfindahl–Hirschman Index (HHI) (a method considered one of the best to measure the concentration) by summing the square values of the top 5 major shareholdings who held 5% or more of shares, then we divided the results by 100 to interpret them as a percentage between 0 and 100%, as per the following Equation (4) (Claessens and Djankov 1999; Demsetz and Lehn 1985; Waheed and Malik 2019):

$$\mathit{OC}=\left(T{1}^2+T{2}^2+T{3}^2+T{4}^2+T{5}^2\right)/100$$

(4)

where OC = Ownership concentration; T = The proportion of shares held by the shareholder.

The data have been obtained from the firm’s annual reports and disclosures from the ASE. The second independent variable is insider shareholding (INSH), which is the percentage of insider shareholding (board members, executive management, or their relatives) who hold 5% or more of shares (Welch 2003; Sheu and Yang 2005). The data have been obtained from the Refinitiv Eikon Database. The third independent variable is board size (BMS), which is the number of members on a board of directors (Jaafar and El-Shawa 2009; Kalsie and Shrivastav 2016). The data has been obtained from the firm’s annual reports and disclosures from the ASE.

Regarding the control variables, we employed several; firm size (SIZE), measured by the natural logarithm of total assets (Adams and Hardwick 1998; Hoi et al. 2020); firm leverage (LVRG), measured by total debt divided by total assets (Hoi et al. 2020); firm age (AGE), measured by subtracting the year of incorporation from the current financial year to capture a firm’s age and reputation (Choi and Hong 2022); capital expenditure (CAPEX), measured by dividing the net investment on property, plant, and equipment on total assets, as capital expenditure positively impacts a firm’s transparency as payoffs from tangible assets are clearer to shareholders (Roy et al. 2022); market value to book value ratio (MVBV), measured by dividing the market capitalization on the total equity to capture a firm’s ability to grow (Alzoubi 2016); gross domestic product per labor force growth rate (GDPL), which is the growth rate of the GDP divided by the total labor force to capture the impact of any change of macroeconomics and employment rates on a firm’s performance and labor force (Fidanoski et al. 2018; Bolt et al. 2012). The data have been obtained from the World Bank national accounts data (World Bank 2024); financial firm (FIN), which is a dummy variable, takes the value of 1 if the firm is related to financial sector, 0 otherwise, to avoid any bias between financial and non-financial firms, knowing that we considered the ASE classification for the sectors (banks, insurance, diversified financial service and real estate) as financial. In this research paper, we have resorted to the ASE to obtain the data from the firm’s annual reports and disclosures published by the end of each firm’s financial year, from 2012 to 2021. Table 5 summarizes the variables’ description, descriptive statistics and data source.

Table 5. Variables’ description and descriptive statistics.

Variable Name	Description	Obs e	Mean	Std. Dev.	Min	Max
Variable Label
Dependent Variable:
Employee Number	Number of employees	1360	462	836	2	7191
EMPNO b
Total Sales	Non-financial firms: total revenues; Financial firms: total Interest income; Insurance firms: total premiums income	1360	109 × 106	413 × 106	29,930	6510 × 106
SALES ($) b
Employee Productivity	Measured by DEA tool, EMPNO is input and SALES is output	1360	0.73	0.285	0.008	1
EP a
Independent Variables:
Ownership Concentration	HHI of Top 5 Major Shareholders who held 5% or more	1360	20.6	22.33	0.25	99.99
OC (%) a b
Insider Shareholding	Proportion of shares held insiders who held 5% or more of shares	1360	24.3	25.00	0	99
INSH (%) c
Board Size	Number of board of directors members	1360	8	2	3	13
BMS b
Control Variables:
Firm Size	Natural Logarithm of Total Assets	1360	18.1	1.774	14.79	24.88
SIZE a b
Leverage Ratio	Total Debt/Total Assets	1360	41.7	28.09	0	100
LVRG (%) a b
Firm Age	Number of Years: Financial Year—Year of Incorporation	1360	29	17.52	3	92
AGE a b
Capital Expenditure	Net property, plant and equipment/Total Assets	1360	25.7	27.05	0	98.9
CAPEX (%) a b
Market to Book Value	Market Value of Equity/Book Value of Equity	1360	1.13	2.969	0.13	104.8
MVBV a b
GDP Growth per Labor GDPL (%) a d	Annual Growth rate of (Jordanian GDP/Total Labor force)	1360	−0.32	5.290	−8.05	7.69
Financial Firm	Dummy: 1 if the firm is related to financial sector, 0 otherwise	1360	0.49	0.50	0	1
FIN b

Data Source: ^a: Own calculation from ASE firm’s annual reports; ^b: ASE firm’s annual reports; ^c: Eikon Database; ^d: World Bank national accounts data; ^e: 136 firms × 10 years = 1360.

4.3. Research Model

We developed an ordinary least squares (OLS) regression model. In addition, firm-fixed effects were employed to capture the variation within firms, and year-fixed effects were employed to capture the variation over time; for instance, COVID-19 pandemic crisis in 2020. This is represented in Equation (5):

EP = β0 + β1 OC + β2 INSH + β3 BMS + β4 SIZE + β5 LVRG + β6 AGE + β7 CAPEX + β8 MVBV + β9 GDPL + β10 FIN + YEAR FE + FIRM FE + Ɛ

(5)

where EP = Employee productivity; β0 = the intercept; βn = the coefficients; OC = Ownership concentration; INSH = Insider ownership; BMS = Board size; SIZE = Firm size; LVRG = Firm leverage ratio; AGE = Firm age; CAPEX= Capital expenditure; MVBV = Market value to book value ratio; GDPL = Growth rate of GDP per labor force; FIN = Financial firm; YEAR FE = Year fixed effects; FIRM FE = Firm fixed effects; Ɛ = error term to capture the uncertainty and chaos of financial markets (Omane-Adjepong et al. 2024).

5. Results and Discussion

5.1. Correlations

We employed the variance inflation factor test (VIF), and the results show that there is no multicollinearity problem. Table 6 presents the results of pairwise correlations and VIF test results.

Table 6. Pairwise correlations and multicollinearity VIF test.

Variable	VIF	EP	OC	INSH	BMS	SIZE	LVRG	AGE	CAPEX	MVBV	GDPL
OC	1.16	0.12	1
INSH	1.08	−0.04	−0.06	1
BMS	1.74	0.43	−0.24	−0.22	1
SIZE	2.20	0.58	0.07	−0.23	0.56	1
LVRG	1.73	0.51	0.10	−0.16	0.35	0.61	1
AGE	1.38	0.35	−0.10	−0.11	0.41	0.45	0.38	1
CAPEX	1.08	0.13	0.005	0.04	−0.14	−0.20	−0.25	−0.09	1
MVBV	1.02	0.07	0.08	−0.004	−0.03	0.05	0.10	0.01	0.04	1
GDPL	1.00	0.004	0.003	0.002	0.02	−0.01	0.01	−0.02	0.01	0.03	1
FIN	1.84	0.27	0.02	−0.01	0.08	0.16	0.24	0.01	−0.66	−0.09	−0.001
Mean VIF	1.50

Bold number indicate <5% significance.

5.2. Regression Results

To examine the associations between ownership concentration, insider ownership, board size, and employee productivity, we employ a robust standard errors ordinary least square (OLS) regression, as the P value of Breusch–Pagan/Cook–Weisberg test for heteroscedasticity (Het. test) is (0.00) for all regression models to avoid the heteroscedasticity problem and obtain more accurate results for our analysis, of which the results are presented in Table 7. Furthermore, all the regression models include firm-year-fixed effects to capture the variation over time within firms. In the first model, we examined the impact of ownership concentration, control variables and firm-year-fixed effects; the results indicate a positive non-significant impact of ownership concentration on employee productivity. In the second model, we performed the likelihood ratio test for linearity, the P value result is (0.00), which means that the relation between ownership concentration and employee productivity is a non-linear relation.

Table 7. Regression analysis.

Model	1	2	3	4	5	6	7
EP	Coefficient	Coefficient	Coefficient	Coefficient	Coefficient	Coefficient	Coefficient
	t-value	t-value	t-value	t-value	t-value	t-value	t-value
OC	0.001	** 0.003					** 0.003
	1.65	2.2					2.41
OCQ		** −0.0001					** −0.0001
		−2					−2.02
INSH			0.0001	*** 0.002			*** 0.002
			0.23	3.1			3.16
INSHQ				*** −0.0001			*** −0.0001
				−3.4			−3.59
BMS					0.003	* 0.032	* 0.032
					0.84	1.92	1.96
BMSQ						* −0.002	* −0.002
						−1.89	−1.91
SIZE	*** 0.179	*** 0.179	*** 0.180	*** 0.181	*** 0.178	*** 0.177	*** 0.177
	11.4	11.52	11.43	11.85	11.18	11.15	11.71
LVRG	*** −0.103	*** −0.098	*** −0.106	*** −0.126	*** −0.105	*** −0.100	*** −0.113
	−2.82	−2.67	−2.88	−3.37	−2.85	−2.7	−2.98
AGE	*** −0.007	*** −0.006	*** −0.007	*** −0.007	*** −0.008	*** −0.007	*** −0.005
	−7.34	−6.02	−8.42	−8.64	−8.03	−7.58	−4.97
CAPEX	−0.026	−0.03	−0.029	−0.029	−0.029	−0.031	−0.031
	−0.33	−0.38	−0.36	−0.36	−0.36	−0.39	−0.39
MVBV	*** −0.002	*** −0.002	*** −0.002	*** −0.002	*** −0.002	*** −0.002	*** −0.001
	−4.1	−4.21	−4.3	−4.33	−4.28	−4.23	−4
GDPL	0.004	0.011	−0.003	−0.004	−0.006	−0.003	0.013
	0.39	0.94	−0.31	−0.36	−0.53	−0.3	1.08
FIN	*** −0.367	*** −0.4	*** −0.366	*** −0.385	*** −0.36	*** −0.371	*** −0.425
	−10.25	−10.01	−8.66	−9.45	−10.32	−10.46	−9.38
Obs	1360	1360	1360	1360	1360	1360	1360
R-squared	0.94	0.94	0.94	0.94	0.94	0.94	0.94
Firm FE	Included	Included	Included	Included	Included	Included	Included
Year FE	Included	Included	Included	Included	Included	Included	Included
Het. test	0.00	0.00	0.00	0.00	0.00	0.00	0.00
Likelihood-ratio test		0.00	0.00		0.04		0.00

*** p < 0.01; ** p < 0.05; * p < 0.10.

Therefore, we added the quadratic ownership concentration (OCQ) along with ownership concentration, control variables and firm-year-fixed effects in order to have an accurate result for the impact of ownership concentration on employee productivity; the results indicate a positive, significant impact of ownership concentration on employee productivity at a 5% level of significance (t-value = 2.2), and a negative significant impact of quadratic ownership concentration on employee productivity at a 5% level of significance (t-value = −2). In the third model, we examined the impact of insider ownership, control variables, and firm-year-fixed effects, whereby the results indicate a positive, non-significant impact of insider ownership on employee productivity. In the fourth model, we performed the likelihood ratio test for linearity, and the P value result is (0.00), which means that the relation between insider ownership and employee productivity is a non-linear relation.

Therefore, we added the quadratic insider ownership (INSHQ) along with insider ownership, control variables and firm-year-fixed effects in order to have an accurate result for the impact of insider ownership on employee productivity, and the results indicate a positive, significant impact of insider ownership on employee productivity at a 1% level of significance (t-value = 3.1), and a negative, significant impact of quadratic insider ownership on employee productivity at a 1% level of significance (t-value = −3.4). In the fifth model, we examined the impact of board size, control variables, and firm-year-fixed effects, whereby the results indicate a positive, non-significant impact of board size on employee productivity. In the sixth model, we performed the likelihood ratio test for linearity, and the P value result is (0.04), which means that the relation between board size and employee productivity is a non-linear relation.

Therefore, we added the quadratic board size (BMSQ) along with board size, control variables, and firm-year-fixed effects in order to have an accurate result for the impact of board size on employee productivity; the results indicate a positive, significant impact of board size on employee productivity at a 10% level of significance (t-value = 1.92), and a negative, significant impact of quadratic board size on employee productivity at a 10% level of significance (t-value = −1.89). In the seventh model, we examined the impact of ownership concentration, quadratic ownership concentration, insider ownership, quadratic insider ownership, board size, quadratic board size, control variables, and firm-year-fixed effects, whereby the likelihood ratio test for linearity, and the P value result is (0.00), which proves the non-linear relation between ownership concentration, insider ownership, board size, and employee productivity. The results indicate a positive, significant impact of ownership concentration on employee productivity at a 5% level of significance (t-value = 2.41), and a negative, significant impact of quadratic ownership concentration on employee productivity at a 5% level of significance (t-value = −2.02); thus, these results are not compatible with our expected hypothesis (1). By contrast, these aforementioned results are in line with the results found in the previous literature (Claessens and Djankov 1999; Drobetz et al. 2019). In addition, the results indicate a positive, significant impact of insider ownership on employee productivity at a 1% level of significance (t-value = 3.16), and a negative, significant impact of quadratic insider ownership on employee productivity at a 1% level of significance (t-value = −3.59); thus, these results are compatible with our expected hypothesis (2). Likewise, these aforementioned results are in line with the results found in the previous literature (Sheu and Yang 2005; Pant and Pattanayak 2007; Bhabra 2007; Jain et al. 2020). Moreover, the results indicate a positive, significant impact of board size on employee productivity at a 10% level of significance (t-value = 1.96), and a negative, significant impact of quadratic board size on employee productivity at a 10% level of significance (t-value = −1.91); thus, these results are compatible with our expected hypothesis (3). Similarly, these aforementioned results are in line with the results found in the previous literature (Cheng et al. 2008; Guest 2009).

In addition, a positive, significant impact of firm size on employee productivity was found at a 1% level of significance (t-value = 11.71), which indicates that employees in large firms are more productive than employees in smaller firms. A negative, significant impact of firm leverage on employee productivity was found at a 1% level of significance (t-value = −2.98), which indicates that employees in firms with high debt ratios are less productive. A negative, significant impact of firm age on employee productivity was found at a 1% level of significance (t-value = −4.97), which indicates that employees in older firms are less productive. No impact was found of capital expenditure on employee productivity. A negative, significant impact of market value to book value ratio on employee productivity was found at a 1% level of significance (t-value = −4). No impact was found of GDP per labor force growth rate on employee productivity. A negative, significant impact of financial firms on employee productivity was found at a 1% level of significance (t-value = −9.38), which indicates that employees in financial firms are less productive.

Furthermore, as presented in Table 8, to check for possible endogeneity between ownership structure and employee productivity, we employed dynamic panel-data estimation, two-step system generalized method of moments regression (GMM), since a GMM model deals with heteroscedasticity, simultaneity, reduces errors over time, and controls endogeneity (by internally transforming the data and by including lagged values of the dependent variable). The P values results of the post estimation of the GMM regression are as follows: Arellano–Bond test for AR (1) is (0.01), Arellano–Bond test for AR (2) is (0.11), Sargan test of overid is (0.12), and Hansen test of overid is (0.21); which means that the model and the instrumental variables are well specified and valid. Moreover, the P value result of GMM test for endogeneity is (0.18), which means that we cannot reject the null hypothesis of the test that the variables are exogenous. Therefore, we consider the results of the seventh regression model presented in Table 7 as the final results of our hypotheses.

Table 8. GMM regression analysis.

EP	Coefficient
	t-value
EP t-1	*** 1.011
	4.94
OC	** 0.004
	2.43
OCQ	** −0.0001
	−2.46
INSH	* 0.001
	1.92
INSHQ	* −0.00002
	−1.89
BMS	0.014
	1.05
BMSQ	−0.001
	−0.87
SIZE	0.001
	0.32
LVRG	0.028
	1.27
AGE	−0.0001
	−0.11
CAPEX	−0.008
	−0.41
MVBV	*** −0.002
	−3.72
GDPL	0.00004
	0.07
FIN	−0.013
	−1.09
Obs	1088
Year FE	Included
Arellano-Bond test for AR(1)	0.01
Arellano-Bond test for AR(2)	0.11
Sargan test of overid	0.12
Hansen test of overid	0.21
GMM Hansen test	0.30
GMM Difference (null H = exogenous)	0.18

*** p < 0.01; ** p < 0.05; * p < 0.10.

Considering the found result of the non-linear relation between ownership concentration, insider ownership, board size, and employee productivity, we employed the margins analysis for further interpretation of this non-linear relation finding, as presented in Table 9. The results of this margins analysis indicate a positive, significant impact of ownership concentration on employee productivity when the proportion of ownership concentration is less than 60%. Furthermore, the results indicate a positive, significant impact of insider ownership on employee productivity when the proportion of insider ownership is less than 50%. Lastly, the results indicate a positive, significant impact of board size on employee productivity when the number of board members is less than 11. Figure 3a–c, generated using the margins analysis, visualizes the U-shape non-linear relation between ownership concentration, insider ownership, board size, and employee productivity, respectively.

Table 9. Margins analysis.

OC		INSH		BMS
EP	Coefficient	EP	Coefficient	EP	Coefficient
	t-value		t-value		t-value
OC = 0%	*** 0.003	INSH = 0%	*** 0.002	BMS = 3	*** 0.042
	3.01		3.57		3.65
OC = 10%	*** 0.003	INSH = 10%	*** 0.002	BMS = 4	*** 0.038
	3.15		3.44		3.9
OC = 20%	*** 0.002	INSH = 20%	*** 0.001	BMS = 5	*** 0.033
	3.34		3.12		4.25
OC = 30%	*** 0.002	INSH = 30%	** 0.001	BMS = 6	*** 0.029
	3.55		2.25		4.71
OC = 40%	*** 0.001	INSH = 40%	0.0001	BMS = 7	*** 0.025
	3.63		0.33		5.27
OC = 50%	*** 0.001	INSH = 50%	−0.0005	BMS = 8	*** 0.020
	2.93		−1.59		5.49
OC = 60%	0.001	INSH = 60%	** −0.001	BMS = 9	*** 0.016
	1.39		−2.55		4.34
OC = 70%	0.0001	INSH = 70%	*** −0.002	BMS = 10	** 0.011
	0.19		−2.98		2.48
OC = 80%	−0.0003	INSH = 80%	*** −0.002	BMS = 11	0.007
	−0.5		−3.19		1.14
OC = 90%	−0.001	INSH = 90%	*** −0.003	BMS = 12	0.002
	−0.9		−3.31		0.31
OC = 100%	−0.001	INSH = 100%	*** −0.003	BMS = 13	−0.002
	−1.16		−3.39		−0.21

*** p < 0.01; ** p < 0.05; * p < 0.10.

Figure 3. (a) The non-linear relation between employee productivity and ownership concentration; (b) The non-linear relation between employee productivity and insider ownership; (c) The non-linear relation between employee productivity and board size.

6. Conclusions and Limitations

This research paper aims to investigate the impact of ownership concentration, insider ownership, and board size on employee productivity for 136 Jordanian public shareholding firms listed in the ASE from 2012 to 2021.

Our empirical results indicate a non-linear relation between ownership concentration and employee productivity, whereby the productivity of employees increases in firms with a proportion of ownership concentration less than 60%. This result is compatible with the stewardship theory, which posits that large block holders negatively affect a firm’s performance. In addition, we found a non-linear relation between insider ownership and employee productivity, whereby the productivity of employees increases in firms with a proportion of insider ownership less than 50%. This result is compatible with agency theory, which posits that higher insider ownership will lead to agency problems and thus, negatively affect a firm’s performance. Furthermore, we found a non-linear relation between board size and employee productivity, whereby the productivity of employees increases in firms that have less than 11 board members. This result is compatible with stewardship theory, which posits that a smaller board size is better and more efficient for monitoring and making decisions.

Herein lies the importance and the value added by this research paper to the knowledge found in the previous literature; our research is the first to highlight the productivity of employees in emerging economies such as the Jordanian economy. Moreover, our empirical findings provided evidence about the impact of CG on employee productivity for a large sample of Jordanian listed firms in the ASE. In addition, our findings could be useful for the JSC and ASE in their continuous process to improve and develop CG instructions.

Finally, this research paper has some limitations, that being the exclusion of some mechanisms of CG, such as the independent, non-executive board members and committees emanating from boards of directors, because the majority of the selected firms in our sample reported information about independent, non-executive board members and committees emanating from the board of directors in their annual reports and disclosures since 2017; that is, after the issuance of governance instructions by the JSC. Another limitation is the geographical boundary of the sample, because taking the sample from Jordan renders it as somewhat difficult to generalize the obtained results to a broader context. Regarding further research, this analysis could be applied to other countries in the Middle East.