Economic Policy Uncertainty, Managerial Ability, and Cost of Equity Capital: Evidence from a Developing Country

Abstract

This study investigates the relationship between economic policy uncertainty (EPU) and the cost of equity capital (CoEC). It also reveals the moderating role of managerial ability (MA) in the relationship between EPU and CoEC in Saudi Arabia. The study sample consists of listed non-financial firms in Tadawul from 2008 to 2019. We analyzed data using STATA, depending on Pearson correlation analysis, two independent sample t-tests, OLS regression, and OLS with robust standard errors clustered by firm. Our study shows a positive effect of EPU on the CoEC. In addition, the results confirm that MA mitigates the positive effect of EPU on the CoEC. This is the first research to investigate the influence of the relationship between EPU on CoEC in Saudi Arabia, one of the largest emerging economies in the Middle East and Gulf countries. Our findings motivate decision-makers to strengthen their MA and establish a safe and stable investment environment to ensure better financing and investment decisions during uncertain times. Lending agencies, investors, and other stakeholders should consider the MA of corporations when making investment decisions.

1. Introduction

Economic policy uncertainty (EPU) has attracted research attention in recent decades to understand its impact on both the financial markets and the real economy (Xu 2020; Barraza and Civelli 2020; Goodell et al. 2021; Chahine et al. 2021). EPU refers to changes in the government policy environment and other uncontrollable shocks, such as fluctuations in energy prices, in addition to financial and political crises (Sarwar et al. 2020; Sharif et al. 2020). Put differently, EPU indicates the unexpected changes that can affect the economic system and results in changes in official policies; i.e., it can bring economic fluctuations at the national level because of uncertainty concerning fiscal and monetary policies (Al-Thaqeb et al. 2022). It can be distinguished from other uncertainty measures at the firm level. In contrast to uncertainty specific to individual firms, EPU can be considered exogenous, posing a larger hedging difficulty for funds and investors (Attig et al. 2021; El Ghoul et al. 2021).

By adopting the agency theory perspective, previous research has stressed the existence of negative firm-level spillovers due to economic policy environment’s changes (Ghasemi et al. 2022). For example, EPU can raise the level of information asymmetry between corporations and investors (Nagar et al. 2019). According to the increased risk that information asymmetry brings to equity holders and creditors, they may require greater returns or impose tougher restrictions, which would drive up the cost of debt and equity financing for corporations (Waisman et al. 2015; Ashraf and Shen 2019; Kim 2019). Further, corporations are less motivated to invest during high EPU periods (Baker et al. 2016; Gulen and Ion 2016; Gupta et al. 2024). This is because policy-related economic uncertainty has a negative relationship with a firm’s earnings and investment, demonstrating that corporations are less motivated to develop their operations when economic policy is unclear (Feng et al. 2023). Therefore, Zhang (2018) indicated that during higher EPU periods, the investors’ risk factor increases, which ultimately increases the required return rate by investors during these periods (Sarwar et al. 2020). A decline in capital investment could be a sign of a lower anticipated cash flow or a higher cost of capital (Pham 2019). Thus, in the current study, we are motivated to investigate the relationship between EPU and cost of equity capital (CoEC).

Additionally, according to the resource dependency theory, higher-ability managers are one of the most important resources in firms (Holcomb et al. 2009). Higher-ability managers are more likely to mitigate information asymmetry problems, providing higher earnings quality (Petkevich and Prevost 2018) and improving transparency and disclosure levels (Baik et al. 2018). In addition, high levels of managerial ability can support firm performance, investment efficiency, and innovation (Gulen and Ion 2016; Bhattacharya et al. 2017; Nguyen and Phan 2017; Chen et al. 2019; Muriithi et al. 2020), which ultimately decreases CoEC (Ghosh et al. 2020; Arslan 2022; Hmaittane et al. 2022; Jang et al. 2023). Accordingly, we expect high-ability managers to affect the relationship between EPU and CoEC. Thus, the current study also examines the moderating role of managerial ability (MA) in the relationship between EPU and CoEC.

Emerging markets recover more quickly than established markets from exogenous uncertainty shocks (Carrière-Swallow and Céspedes 2013; Khelil et al. 2023). Further, it is recognized that EPU can be roughly one-third greater in emerging nations than in advanced countries (Baker et al. 2016). Hence, it is worthy to examine EPU and CoEC in a distinct emerging market with higher EPU, such as Saudi Arabia. Saudi Arabia was selected as the context for this study because of its unique institutional settings (Nurunnabi et al. 2020, 2022). The Saudi government has recently paid significant attention to developing their economic and business regulations, including the development of governance requirements, as evidenced by their vision 2030. Despite sharing some economic features with the surrounding countries in the Arab region, the Saudi economy has recently witnessed an accelerated growth rate, representing 25% of total Arab countries’ GDP (Diab and Hamdy 2022). Further, although the hydrocarbon sector is considered the largest contributor to GDP, and the main source of income in Saudi Arabia (Guizani et al. 2023), the country is more vulnerable to any change in the price of oil that increases volatility in its economic operations and, as a result, the level of uncertainty over economic growth (Al-Khouri and Dhade 2014). Besides, most Saudi businesses depend on Islamic business concepts that have special financial features. Previous research indicates that adhering to Islamic Shariah places limitations on choices made regarding investments, finance, operations, and risk management, affecting the EPU level (Alnori and Alqahtani 2019; Guizani et al. 2023).

We examined a sample of 1188 firm-year observations and found robust evidence that CoEC is positively correlated with EPU. In addition, the results showed that MA mitigates the positive relationship between EPU and CoEC. The current findings indicate that high-ability managers could add value during high uncertainty periods by reducing the CoEC, which is consistent with the resource-based theory. The findings of this study contribute to the literature in some respects, as follows: Firstly, this study highlights the implications of EPU for corporate financing decisions. Secondly, this study bridges the research gap regarding the relationship between EPU and the CoEC while considering MA as a valuable tool for efficient risk management. In contrast to other research that focuses on the negative business implications of EPU, we confirm the positive role of MA in mitigating such negative implications. Thirdly, we expand the discussion on the implications of EPU on CoEC in Saudi Arabia, as one of the most important emerging markets in the Middle East and Gulf countries. Finally, to our knowledge, none of the previous research has examined the moderating role of MA in the relationship between EPU and CoEC. Consequently, our study enriches the growing body of knowledge on the effect of EPU and MA on the CoEC. This is crucial, because economic uncertainty is an external variable that is difficult to control by management. However, higher-ability managers, with their wise economic decisions, can reduce corporate investment risks and, hence, reduce the potential positive influence of economic uncertainty on the cost of capital. By doing so, the findings of this research encourage corporate decision-makers to strengthen their MA to ensure better financing and investment decisions in the case of EPU. They also motivate financial and economic policymakers in Saudi Arabia to establish a safe and stable investment environment.

The remainder of the paper is arranged as follows: Section 2 introduces the theoretical background and hypotheses development; Section 3 presents research methods; Section 4 displays empirical results; finally, Section 5 concludes the paper.

2. Theoretical Background and Hypotheses Development

Following the agency theory perspective, some studies have indicated that higher EPU is associated with lower information asymmetry (Ashraf and Shen 2019), higher investment risk (Zhang 2018), and lower earnings (Gupta et al. 2024). This context makes investors and creditors require higher returns or impose tougher restrictions, which increases the cost of debt and equity financing for companies (Kim 2019). This unbalanced context, as informed by the resource dependency theory, may require the existence of higher-ability managers to mitigate information asymmetry problems, allow higher earnings quality (Petkevich and Prevost 2018), support investment efficiency (Muriithi et al. 2020) and, hence, reduce the CoEC (Arslan 2022; Hmaittane et al. 2022; Jang et al. 2023), as further explained below.

2.1. Economic Policy Uncertainty and Cost of Equity Capital

Policy uncertainty arises due to potential economic policy changes in the future, which affects the overall economy and corporations (Chan et al. 2021). For instance, EPU might reduce the quality of internal controls and increase information asymmetry (Barth et al. 2013). Thus, an increase in EPU might cause information gaps between investors and corporations, making it harder for investors to obtain information. Consequently, investors would require higher returns, and the cost of capital might increase accordingly (Liu and Wang 2022). Further, recent research has indicated the influence of EPU on corporate investment (Kwabi et al. 2024). For instance, Wang et al. (2023) ensured that during higher EPU periods, investors might modify their portfolios, shifting funds from higher-risk to lower-risk investments.

Thus, the relationship between EPU and cost of capital has been recognized among recent researchers, whether due to firm-level uncertainty, such as uncertainty in cash flow (Chay and Suh 2009), or uncertainty in the external environment, such as financial crises related to uncertainty (Bliss et al. 2015), political uncertainty (Huang et al. 2015), and significant changes in tax laws (Buchanan et al. 2017).

In particular, using data from the US, Pham (2019) noted a positive relationship between risk and CoEC. Xu (2020) revealed that US firms’ cost of capital increased when economic policy was uncertain. Therefore, firms issue capital less frequently and in smaller amounts, and the effect is more pronounced when there are high financial constraints and competition. These results are consistent with Chan et al. (2021), who showed an increase in the CoEC during higher EPU periods. Further, using international evidence, Kwabi et al. (2024) reported a positive relationship between risk and cost of capital. In the same direction, Kwabi et al. (2022) found that the cost of capital is positively related to EPU. Focusing on the Chinese context, Li et al. (2018) showed a positive relationship between uncertainty and the CoEC. Similarly, Xu and Liu (2022) found a positive relationship between EPU and CoEC. In the same direction, Liu and Wang (2022) reported a positive relationship between EPU and CoEC rather than the cost of debt.

As noticed from the above discussion, the literature is focused mainly on established economies such as the US and China. Although examining the relationship between EPU and CoEC has been attractive to recent researchers, there is not much empirical evidence on the influence of EPU on firm-specific CoEC in emerging markets such as Saudi Arabia. We extend the current research, expecting that EPU might significantly affect CoEC. Consequently, our first hypothesis is set as follows:

H1: 

EPU has a significant effect on CoEC in Saudi Arabia.

2.2. Economic Policy Uncertainty, Cost of Equity Capital, and Managerial Ability

Previous studies indicating a positive relationship between EPU and CoEC have investigated some factors that could mitigate this positive relationship, such as foreign equity portfolios (Kwabi et al. 2022), political connections (Pham 2019), dependence level on government spending, informative stock price (Chan et al. 2021), financing constraints (Xu and Liu 2022), and information asymmetry and internal control quality (Liu and Wang 2022). However, none of the previous studies have examined the effect of MA. According to the resource-based theory, higher-ability managers are considered effective resources for their corporations (Simamora 2021). Previous studies indicated that MA can add value to firms through several channels, such as improving firms’ information environment and enhancing disclosure quality (Baik et al. 2018), reducing information asymmetry (Petkevich and Prevost 2018), increasing firms’ credit ratings (Bonsall et al. 2017; Cornaggia et al. 2017), providing financial statements in a timely manner (Abernathy et al. 2018), and enhancing firm value (Atawnah et al. 2024).

Further, higher ability managers have more understanding of the operational environment, market dynamics, and product demand forecasts, which is anticipated to yield more stable earnings, less volatile future stock returns, and more efficient future revenue growth (Bonsall et al. 2017). Moreover, the greater knowledge, expertise, and information provided by higher-ability managers may lead to maximizing profits and minimizing the expenditures associated with taking risks (Simamora 2021). Consequently, firms with higher-ability managers are more likely to have less cost of equity compared to other firms (Arslan 2022; Hmaittane et al. 2022; Jang et al. 2023).

Additionally, firms with higher MA might increase acquisitions and enhance cost efficiency during exposure to higher EPU levels (Jiang et al. 2023). In other words, the negative effects of uncertainty on analyst performance might be lower in firms with higher MA. This is because higher-ability managers might be able to act as moderators to reduce the adverse influences of overall uncertainty on the accuracy of analysts’ earnings forecasts (Chen et al. 2020). In this regard, Yang (2022) found that the increased cost of obtaining information, the resultant information asymmetry due to EPU, and the negative influences of investment are more pronounced and important in corporations with lower managerial ability. Accordingly, we expect that MA might have a significant effect on the relationship between EPU and CoEC. Consequently, our second hypothesis is set as follows:

H2: 

MA has a significant effect on the relationship between EPU and CoEC in Saudi Arabia.

3. Research Methods

3.1. Sample Selection Method and Data Collection

The initial sample consists of all Saudi-listed firms on the Saudi Exchange (Tadawul) during the period 2008 to 2019. The data was hand-collected from the financial statements published by corporations, and stock price data available at https://www.saudiexchange.sa/ (accessed on 10 September 2023), http://argaam.com/ (accessed on 10 September 2023), and http://mubasher.info/markets/TDWL (accessed on 10 September 2023). Banking, financial, and insurance firms were excluded. We also dropped corporations with a fiscal year that does not end on December 31 or with missing data necessary to measure the study variables. The final sample consisted of 99 corporations and yielded 1188 corporate years’ observations, classified according to industry, as shown in

Table 1. The sample of study.

		Observations
Panel A: Sample selection methodology
Corporations listed on Tadawul (2008–2019)		1962
(–) Corporations belonging to Financial, Insurance, and Banking		(348)
(–) Corporations whose fiscal year does not end on 31st of Dec and corporations with missing data		(426)
Final sample		1188
Panel B: Sample distribution by industry	No. of Firms	Observations
Energy	4	48
Basic materials	36	432
Capital goods	9	108
Commercial and professional services	2	24
Travel	4	48
Long-term goods	4	48
Consumer services	5	60
Media and entertainment	2	24
Retail of luxury goods	5	60
Food production	12	144
Food segmentation	2	24
Health care	2	24
Pharmaceutical	1	12
Telecommunications	3	36
Utilities	2	24
Real Estate Mgmt & Dev’t	6	72
Total	99	1188

below.

3.2. Variables Measurement

3.2.1. Economic Uncertainty Policy

We followed Ahir et al. (2022) in measuring EPU according to the World Uncertainty Index (WUI). The WUI index is a metric that uses text mining of the Economist Intelligence Unit’s (EIU) countries’ reports to track uncertainty worldwide. It reflects how frequently the term “uncertainty” and its synonyms appear in the national EIU reports. The authors scale the raw numbers by the total word count in each report to make the WUI comparable across national borders (Guizani et al. 2023). The data about the WUI index are available at http://www.policyuncertainty.com (accessed on 10 September 2023).

3.2.2. Managerial Ability

To measure managerial ability, we used the data envelope method (DEA) presented by Demerjian et al. (2012). DEA is a nonparametric technique that calculates the relative effectiveness of decision-making units (DMUs) to maximize output-to-input ratio. DEA employs linear programming to establish an effective frontier of observed production points. The most efficient DMUs, or those on the frontier, are given a value of one by the DEA, while inefficient DMUs are given a value of less than one. As a result, the distance of the DMU from the frontier determines the efficiency score that the DEA approach assigns to inefficient units (Baik et al. 2020). The DEA score shows how well a firm’s management team uses its corporate resources to optimize outputs (Demerjian et al. 2012). According to this method, MA can be measured by capturing efficiency in the sample using the following method:

$$\begin{array}{l}Max\theta =Sales\\ \times [v_{1}CoGS+v_{2}SG\&A+v_{3}PPE+v_{4}OpsLease+v_{5}R\&D+v_{6}Goodwill\\ +v_{7}OtherIntan{]}^{-1}\end{array}$$

(1)

where Sales is output measured by net sales revenue, and inputs include the following seven variables: the cost of goods sold (CoGS); selling, general and administrative expenses (SG&A); net property, plant, and equipment (PPE); capitalized operating leases (OpsLease); research and development expenses (R&D); purchased goodwill (goodwill); and other intangible assets purchased (OtherIntan). Due to shortcomings in disclosure about some previous inputs in financial statements, such as OpsLease, R&D, and goodwill, we dropped some of these variables and applied the model depending on four inputs as follows:

$$Max\theta =Sales\times {[v_{1}CoGS+v_{2}SG\&A+v_{3}PPE+v_{4}OtherIntan]}^{-1}$$

(2)

The efficiency score captured from the optimization procedure in the previous step reflects efficiency for firm characteristics and managers. To get the MA score, we estimated the following Tobit model by industry to exclude firm-level characteristics:

$$\begin{array}{l}{FirmEfficiency}_i\\ =\alpha +B_{1}FrimSize+B_2{MarketShare}_i+B_3{FCFIndicator}_i+B_{4}Ln({Age)}_i\\ +B_5{B.S.Concentration}_i+B_6{ForeignCurrency}_i+B_{7}Year\\ +{\epsilon }_i\end{array}$$

(3)

where FirmSize is the firm size measured as the natural logarithm of total assets. Market share is the percentage of sales of the corporation to total sales of the industry in year t. FCF indicator is a dummy variable that equals one if the firm has positive free cash flows and zero otherwise. LnAge is the natural logarithm of firm age, and B.S. Concentration is business segment concentration computed as the sum of the squares of each industry’s sales in year t as a percentage of total corporate sales. The Foreign Currency indicator is a dummy variable that equals one if the corporation reports a non-zero value for foreign currency and zero otherwise. The residual from model 3 is our main measure of MA attributed to the management team (Demerjian et al. 2012).

3.2.3. Cost of Equity Capital

We follow Omran and Pointon (2004) and Ezat (2019) in measuring the cost of equity capital as the inverse of the PE ratio as follows:

$$CoEC=1/PERatio-\left[\right(EPS-Div)/EPS]$$

(4)

where CoEC is the cost of equity capital, PE ratio is the current share price scaled by earnings per share (EPS), and Div is the dividends per share.

3.2.4. Control Variables

We included some control variables to ensure that the relationship between EPU, MA, and CoEC (models 5 and 6) is not driven by confounding variables. Following previous research (e.g., Khlif et al. 2015; Ezat 2019; Xu 2020), we controlled for firm size (FS), return on assets (ROA), firm leverage (LEV), firm age (FA), market-to-book ratio (MtB), auditor size (AS), and GDP growth rate (GDPG).

Table 2. Variables measurement.

Abbreviation	Variable	Measurement
CoEC	Cost of equity capital	The inverse of PE ratio as mentioned in model 4.
EPU	EPU	The average of WUI during the year.
MAscore	Managerial ability score	The residual from model 3.
FS	Firm size	The natural logarithm of total assets.
ROA	Return on assets	The net profit scaled by total assets.
LEV	Firm leverage	Total debt scaled by total assets.
FA	Firm age	The natural logarithm of firm age.
MtB	Market-to-book ratio	The market value of equity scaled by book value of equity.
AS	Auditor size	A dummy variable equals one if the firm is audited by one of Big 4 and zero otherwise.
GDPG	Gross domestic product growth	The gross domestic product growth retrieved from worldbank.org (accessed on 16 August 2024).
Market share	Market share	The percentage of sales of the corporation to total sales of the industry in year t.
FCF indicator	Positive free cash flow indicator	A dummy variable that equals one if the firm has positive free cash flows, and zero otherwise.
B.S. Concentration	Business segment concentration	Business segment concentration computed as the sum of the squares of each industry’s sales in year t as a percentage of total corporate sales.
The Foreign currency indicator	The foreign currency indicator	A dummy variable that equals one if the corporation reports a non-zero value for foreign currency, and zero otherwise.
PE ratio	Price-earnings ratio	The current share price scaled by earnings per share (EPS).
Div	Dividends payout	Dividends per share.
Variables used in measuring firm efficiency
Firm efficiency	We depend on Data Envelopment Analysis (DEA) (Model 2) using one output (net sales revenue) and four inputs (the cost of goods sold (COGS), selling, general and administrative expenses (SG&A), net property, plant, and equipment (PPE), and other intangible assets purchased (OtherIntan)).

defines the variables used in models 5 and 6.

3.3. Research Models

To investigate the first hypothesis (H1), which claims “EPU has a significant effect on CoEC” we regress EPU and control variables on CoEC depending on the following ordinary least squares (OLS) regression:

$$\begin{array}{l}{CoEC}_{it}=\alpha +B_1\left({EPU}_t\right)+B_2\left({FS}_{it}\right)+B_3\left({ROA}_{it}\right)+B_4\left({LEV}_{it}\right)+B_5\left({FA}_{it}\right)+B_6\left({MtB}_{it}\right)\\ +B_7\left({AS}_{it}\right)+B_8\left({GDPG}_t\right)+B_9\left(IndustryEffect\right)+B_{10}\left(YearEffect\right)\end{array}$$

(5)

To investigate the second hypothesis (H2), which claims that “MA has a significant effect on the relationship between EPU and CoEC”, we regress EPU, MA, the interaction variable of MA and EPU (MAscore ∗ EPU), and control variables on CoEC depending on the following OLS regression:

$$\begin{array}{l}{CoEC}_{it}=\alpha +B_1\left({EPU}_t\right)+B_2\left({MAScore}_{it}\right)+B_3\left({MAScore}_{it}\ast {EPU}_t\right)+B_4\left({FS}_{it}\right)\\ +B_5\left({ROA}_{it}\right)+B_6\left({LEV}_{it}\right)+B_7\left({FA}_{it}\right)+B_8\left({MtB}_{it}\right)+B_9\left({AS}_{it}\right)\\ +B_{10}\left({GDPG}_t\right)+B_{11}\left(IndustryEffect\right)+B_{12}\left(YearEffect\right)\end{array}$$

(6)

4. Empirical Results

4.1. Descriptive Statistics

Panel A of

Table 3. Descriptive results.

Panel A: Descriptive Results for the Sample (Full Sample = 1188)
Variables		Mean	SD	Median	Minimum	Maximum
CoEC		0.159	0.121	0.102	0.00	1.00
EPU		0.138	0.065	0.115	0.05	0.26
MAscore		0.001	0.021	0.000	−0.54	0.59
FS		14.638	1.522	14.528	11.05	17.83
ROA		0.055	0.079	0.045	−0.12	0.23
LEV		0.385	0.214	0.381	0.01	0.94
FA		3.183	0.572	3.295	1.95	4.17
MtB		1.856	1.1243	1.502	0.00	4.22
AS		0.586	0.492	1.00	0.00	1.00
GDPG		3.592	3.221	3.439	−2.06	10.99
Panel B: Univariate statistics (Full Sample = 1188)
Variables	Higher EPU period (n = 594)		lower EPU period (n = 594)		t-test	Sig.
	Mean	SD	Mean	SD
CoEC	0.2687	0.197	0.049	0.081	24.961 ***	0.000
FS	14.602	1.529	14.674	1.516	−0.817	0.723
ROA	0.046	0.080	0.064	0.078	−3.782	0.799
LEV	0.390	0.220	0.380	0.208	0.770	0.177
FA	3.197	0.571	3.169	0.573	0.869	0.845
MtB	1.701	1.109	2.011	1.118	−4.804	0.383
AS	0.537	0.499	0.636	0.481	−3.491 ***	0.000
GDPG	4.287	3.722	2.897	2.438	7.613 ***	0.000
Panel C: Univariate statistics (Full Sample = 1188)
Variables	Higher-ability managers (n = 594)		Lower-ability managers (n = 594)		t-test	Sig.
	Mean	SD	Mean	SD
CoEC	0.142	0.174	0.176	0.197	−3.180 ***	0.000
FS	14.635	1.504	14.641	1.541	−0.059	0.882
ROA	0.081	0.078	0.029	0.072	11.922 ***	0.002
LEV	0.372	0.210	0.399	0.217	−2.202	0.865
FA	3.169	0.580	3.197	0.564	−0.833	0.313
MtB	1.978	1.155	1.734	1.080	3.759 ***	0.002
AS	0.592	0.491	0.580	0.493	0.412	0.411
GDPG	3.894	3.281	3.290	3.134	3.245	0.587

* The results are significant at a level < 10%. ** The results are significant at a level < 5%, *** The results are significant at a level < 1%. For the variable definitions, see Table 2.

presents descriptive results of our sample. The CoEC varies between 0.00 and 1.00, with a mean value of 0.159, a median of 0.102, and a standard deviation of 0.121, which is comparable to some previous studies (Omran and Pointon 2004; Ezat 2019). EPU ranges between 0.05 and 0.26, with a mean value of 0.138, a median of 0.115, and a standard deviation of 0.065. MAscore ranges between −0.54 and 0.59, with a mean value of 0.001, a median of 0.000, and a standard deviation of 0.021, which is comparable to some previous studies (Abernathy et al. 2018; Cho et al. 2018). The mean values of FS, ROA, LEV, FA, MtB, AS, and GDPG are 14.638, 0.055, 0.385, 3.183, 1.856, 0.586, and 3.592, respectively.

Panel B of Table 3 shows variance statistics between corporations in higher EPU periods and lower EPU periods. The results indicate that corporations have significantly higher CoEC in higher EPU periods (at the 1% significance level) than in lower EPU periods. In addition, GDPG is higher in higher EPU periods (at the 1% level) than in lower EPU periods. Finally, corporations in higher EPU periods are less likely to assign Big 4 auditors (at the 1% level) than firms in lower EPU periods.

Panel C of Table 3 reports variance statistics between corporations with higher-ability and lower-ability managers. The results show that corporations with higher-ability managers have significantly lower CoEC (at the 1% level) than corporations with lower-ability managers. However, corporations with higher-ability managers have significantly higher ROA and MtB (at the 1% level) than corporations with lower-ability managers.

4.2. Correlation Analysis

Table 4. Pearson’s correlation matrix.

Variables	1	2	3	4	5	6	7	8	9	10
CoEC	1
EPU	0.638 ***	1
MAscore	−0.104 ***	−0.099 ***	1
FS	−0.004	−0.019	0.025	1
ROA	−0.152 ***	−0.135 ***	0.355 ***	0.043	1
LEV	0.082 ***	0.034	−0.024	0.441 ***	−0.307 ***	1
FA	0.147 ***	0.041	−0.072 **	−0.246 ***	0.067 **	−0.228 ***	1
MtB	−0.068 **	−0.177 ***	0.129 ***	−0.197 ***	0.157 ***	0.006	0.163 ***	1
AS	−0.140 ***	−0.147 ***	0.058 **	0.427 ***	0.255 ***	0.268 ***	−0.149 ***	0.048	1
GDPG	−0.363 ***	−0.093 ***	0.130 ***	−0.018	0.139 ***	−0.037	−0.095 ***	−0.011	0.098 ***	1

* The results are significant at a level < 10%. ** The results are significant at a level < 5%, *** The results are significant at a level < 1%. For the variable definitions, see Table 2.

presents the Pearson correlations between the main variables. There is a positive relationship between EPU and CoEC at the 1% level. However, there is a negative relationship between MAscore and CoEC at the 1% level. These findings reflect a rise in the CoEC when EPU is higher and a decline in the CoEC when MA is higher. In addition, the findings indicate that ROA, AS, and GDPG are negatively correlated with the CoEC at the 1% levels. MtB is negatively correlated with the CoEC at the 5% levels. However, LEV and FA are positively correlated with the CoEC at the 1% levels.

4.3. Regression Analysis

We employed the fixed-effects models in the regression methodology because the data collected is from nonfinancial industries. These models control for unobserved heterogeneity that could influence both EPU and CoEC (Moody and Marvell 2003). In addition, we examined the appropriateness of the random effect estimation vs. the fixed-effects one, depending on the Hausman test. The results revealed that unobserved firm-specific variables have an insignificant relationship with the other variables. In addition, we depended on robustness checks (Section 4.4) to test whether the main results from the fixed-effects models have been severely influenced by the potential presence of endogeneity problems (Blundell and Bond 1998; Gerged et al. 2020).

Table 5. Regressing EPU and MA on CoEC.

Variables	Model (5)		Model (6)
	Coefficient	t-Value	Coefficient	t-Value
Constant	−0.414 ***	−3.48	−0.406 ***	−3.44
EPU	3.889 ***	28.67	3.887 ***	28.92
MAscore			0.001	0.00
MAscore ∗ EPU			−3.070 ***	−5.09
FS	0.003	0.43	0.001	0.22
ROA	−0.033	−0.27	−0.020	−0.15
LEV	0.152 ***	3.04	0.174 ***	3.47
FA	0.079 ***	5.04	0.079 ***	5.06
MtB	0.007	0.94	0.005	0.62
AS	−0.026	−1.27	−0.023	−1.13
GDPG	−0.038 ***	−13.97	−0.037 ***	−13.37
Industry Fixed Effect	Yes		Yes
Years Fixed Effect	Yes		Yes
R-Squared (%)	51.60		52.64
VIF	˂4		˂4
Observations	1188		1188

* The results are significant at a level < 10%. ** The results are significant at a level < 5%, *** The results are significant at a level < 1%. For the variable definitions, see Table 2.

presents the results of regressing EPU and MAscore on the CoEC. Model 5 reveals a significant positive effect of EPU on the CoEC (β = 3.889; p < 0.01). This result indicates that the CoEC is higher when EPU increases, which is consistent with previous research reporting an increase in risk premium demanded by investors due to the undiversifiable (political) risk during higher EPU periods (Pham 2019; Xu 2020; Liu and Wang 2022). Hence, this result supports the acceptance of H1.

Model 6 conveys a significant positive effect of EPU on CoEC (β = 3.887; p < 0.01) and an insignificant effect of MAscore on the CoEC (β = 0.001; p > 0.1). However, the results show a significant negative effect of MAscore ∗ EPU on the CoEC (β = −3.070; p < 0.01), which reflects that MA can mitigate the positive effect of EPU on the CoEC. This finding supports the view that an increase in the equity risk premium stemming from a rise in policy uncertainty would have different effects on the CoEC at the firm level (Xu 2020). It also confirms the view that MA is one of the most important resources in corporations. Hence, the positive effects of higher managerial ability, such as managing internal resources more effectively, improving the disclosure level and quality, and providing more accurate forecasts (Demerjian et al. 2013; Chronopoulos and Siougle 2017; Baik et al. 2018; Arslan 2022), can lead to a lower CoEC during higher EPU periods. Hence, this finding supports the acceptance of H2.

4.4. Robustness Checks

We performed some additional analyses to verify that the main results are robust, considering potential simultaneity concerns. Firstly, we followed some previous studies in re-estimating models 5 and 6 depending on a one-year lag for EPU and independent variables (e.g., Pham 2019; Lou et al. 2022; Wang et al. 2023). This analysis allows time’s impact on EPU to be discerned in the CoEC. The results are similar to the previous ones reported in Table 4.

Secondly, considering the reverse causality issue, models 5 and 6 are re-estimated after adding lagged CoEC as independent variables (e.g., Harjoto and Laksmana 2018). The results shown in Table 5 and

Table 6. Regressing lagged EPU and MA on the CoEC.

Variables	Model (6)		Model (6)
	Coefficient	t-Value	Coefficient	t-Value
Constant	−0.517 ***	−3.43	−0.503 ***	−3.34
EPUt−1	2.315 ***	14.17	2.268 ***	13.83
MAscore			−0.017	−0.31
MAscore ∗ EPUt−1			−1.456 ***	−2.67
FS	0.016 **	1.84	0.015 *	1.70
ROA	−0.296 *	−1.87	−0.249	−1.44
LEV	0.192 ***	3.07	0.210 ***	3.32
FA	0.133 ***	6.69	0.131 ***	6.62
MtB	0.003	0.32	0.002	0.21
AS	−0.077 ***	−3.00	−0.079 ***	−3.07
GDPG	−0.039 ***	−11.74	−0.039 ***	−11.62
Industry Fixed Effect	Yes		Yes
Years Fixed Effect	Yes		Yes
R−Squared (%)	33.26		33.70
VIF	˂4		˂4
Observations	1089		1089

* The results are significant at a level < 10%. ** The results are significant at a level < 5%, *** The results are significant at a level < 1%. For the variable definitions, see Table 2.

are qualitatively similar to those presented in

Table 7. Re-estimating models 5 and 6 considering the reverse causality issue.

Variables	Model (5)		Model (6)
	Coefficient	t-Value	Coefficient	t-Value
Constant	−0.192 **	−2.20	−0.190 **	−2.20
EPU	3.679 ***	32.72	3.652 ***	32.81
MAscore			0.040	1.24
MAscore ∗ EPU			−2.513 ***	−5.39
FS	−0.013 **	−2.48	−0.013 ***	−2.62
ROA	0.407 ***	4.32	0.339 ***	3.36
LEV	0.154 ***	4.21	0.166 ***	4.51
FA	0.034 ***	2.89	0.036 ***	3.12
MtB	−0.002	−0.47	−0.004	−0.80
AS	0.019	1.30	0.024	1.62
GDPG	−0.018 ***	−9.08	−0.018 ***	−9.09
CoECt−1	0.532 ***	28.50	0.532 ***	28.85
Industry Fixed Effect	Yes		Yes
Years Fixed Effect	Yes		Yes
R-Squared (%)	77.12		77.78
VIF	˂4		˂4
Observations	1089		1089

* The results are significant at a level < 10%. ** The results are significant at a level < 5%, *** The results are significant at a level < 1%. For the variable definitions, see Table 2.

, confirming that EPU has a significant positive relationship with the cost of equity. This positive relationship is mitigated by MAscore.

Thirdly, to control for heteroscedasticity and serial correlation, we re-ran models 5 and 6 following OLS regression, with robust standard errors clustered by the firm (Pham 2019; Xu 2020; Makosa et al. 2021; Feng et al. 2023). The results reported in

Table 8. Re-estimating models 5 and 6 with robust standard errors clustered by firm.

Variables	Model (5)		Model (6)
	Coefficient	t-Value	Coefficient	t-Value
Constant	−0.414	−1.76	−0.406	−1.69
EPU	3.889 ***	4.81	3.887 ***	4.98
MAscore			0.001	0.00
MAscore ∗ EPU			−3.070 ***	−3.16
FS	0.003	0.25	0.001	0.12
ROA	−0.033	−0.11	−0.020	−0.07
LEV	0.152 **	2.39	0.174 ***	2.94
FA	0.079 **	2.87	0.079 ***	2.93
MtB	0.007	0.65	0.005	0.42
AS	−0.026	−0.76	−0.023	−0.65
GDPG	−0.038 **	−2.36	−0.037 ***	−2.37
Industry Fixed Effect	Yes		Yes
Years Fixed Effect	Yes		Yes
Cluster by Firm	Yes		Yes
R-squared	51.60		52.64
VIF	˂4		˂4
Observations	1188		1188

* The results are significant at a level < 10%. ** The results are significant at a level < 5%, *** The results are significant at a level < 1%. For the variable definitions, see Table 2.

are identical to those shown in Table 5, confirming H1 and H2.

Finally, we computed the cost of equity by industry-adjusted CoEC per year from the period of 2008 to 2019 (e.g., Mishra 2014; Eliwa et al. 2016). The CoEC of each firm is calculated as the difference between the CoEC of the firm and the median CoEC of the industry to which the firm belongs in that year. The results in

Table 9. Regressing EPU and MA on the CoEC adjusted by industry.

Variables	Model (5)		Model (6)		Model (5)		Model (6)
	Coefficient	t-Value	Coefficient	t-Value	Coefficient	t-Value	Coefficient	t-Value
Constant	−0.925 ***	−7.93	−0.916 ***	−7.94	−0.925 ***	−4.00	−0.916 ***	−3.89
EPU	3.858 ***	29.00	3.860 ***	29.32	3.858 ***	4.85	3.860 ***	5.03
MAscore			0.026	0.64			0.026	0.64
MAscore ∗ EPU			−3.193 ***	−5.40			−3.193 ***	−3.29
FS	0.015 **	2.24	0.013 **	2.05	0.015	1.35	0.013	1.19
ROA	−0.099	−0.82	−0.116	−0.91	−0.099	−0.33	−0.116	−0.42
LEV	0.096 **	1.96	0.115 **	2.36	0.096	1.45	0.115 *	1.83
FA	0.090 ***	5.83	0.091 ***	5.91	0.090 ***	3.22	0.091 ***	3.36
MtB	0.001	0.14	−0.001	−0.25	0.001	0.11	−0.001	−0.19
AS	−0.036 *	−1.81	−0.032	−1.62	−0.036	−1.10	−0.032	−0.93
GDPG	−0.037 ***	−14.08	−0.037 ***	−13.89	−0.037 **	−2.37	−0.037 **	−2.38
Cluster by Firm	No		No		Yes		Yes
R-squared	52.61		53.76		52.62		53.76
VIF	˂4		˂4		˂4		˂4
Observations	1188		1188		1188		1188

* The results are significant at a level < 10%. ** The results are significant at a level < 5%, *** The results are significant at a level < 1%. For the variable definitions, see Table 2.

are qualitatively similar to those reported in Table 5, Table 6, Table 7 and Table 8, confirming that EPU has a significant positive effect on the CoEC, which can be mitigated through higher managerial ability.

5. Discussion and Conclusions

This study aimed to examine the relationship between EPU and CoEC, in addition to the moderating role of MA in the relationship between EPU and CoEC in Saudi Arabia. A sample of non-financial firms listed during the period 2008 to 2019 was investigated using univariate and multivariate analyses. Our results reveal a positive relationship between EPU and CoEC. In addition, we found robust evidence that MA mitigates the relationship between EPU and CoEC, which is consistent with the resource-based theory view (Holcomb et al. 2009). Our findings highlight the importance of higher-ability managers in responding to economic changes in the external environment and mitigating the negative implications of EPU for firm value. By doing so, this study contributes to the literature on the relationship between EPU and CoEC and the moderating role of MA by focusing on a distinct emerging market vulnerable to increasing EPU, Saudi Arabia (Guizani et al. 2023).

Our findings have some practical implications. Firstly, they provide a better understanding for corporations, academics, and regulators in Saudi Arabia regarding the potential implications of EPU on the CoEC. Secondly, they encourage corporations to hedge for financing their investments through internal rather than external resources, especially during higher EPU periods. Finally, they highlight the positive implications of MA during higher EPU periods. Accordingly, we recommend that the Saudi government should adopt more stable economic policies to reduce EPU. In addition, it should consider appropriate political and economic procedures to provide external financing to small corporations and startups to avoid the negative impact of EPU on the cost of capital. Further, investors, lending agencies, and other stakeholders should consider MA when making investment decisions.

However, our findings are not without limitations. Firstly, our results are not generalizable to banking and financial firms, as these industries were not included in our sample. Secondly, we did not investigate the research hypotheses during the COVID-19 period, and hence, future research can consider this issue. Thirdly, future research can re-examine the influence of moderating variables, other than managerial ability, on the relationship between EPU and CoEC. Finally, we did not examine the mediating effect of MA concerning the association between EPU and CoEC, which can be examined by future research.