The market reaction to syndicated loan announcements before and during the COVID-19 pandemic and the role of corporate governance.

This study examines the wealth effects of syndicated loan announcements before and after the onset of the COVID-19 outbreak. Using a sample of 637 loan announcements by European borrowers, we find significantly higher wealth gains during the pandemic compared to the pre-pandemic period. The results suggest that the certification of multiple lenders loans conveys a positive signal for the borrowers' creditworthiness during the pandemic-driven economic meltdown. We further show that certain corporate governance mechanisms, such as board size, gender diversity, and CEO duality and compensation, are related differently to borrowers' excess returns before and after the COVID-19 pandemic. The results are robust to alternative model specifications that control for different estimation models and event windows. They also hold after addressing self-selection bias.

Introduction

The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) and the disease it causes (COVID-19) that was first identified in December 2019 in Wuhan (China) has infected over 250 million people and has led to more than 5 million deaths worldwide1 . As a result of the high infection spread, the World Health Organization (WHO) declared COVID-19 a global emergency on 20th February, 2020, and a pandemic on 11th March. Apart from the infections and deaths caused by this disease, COVID-19 has also led to unprecedented repercussions on daily life and the economy (Goodell, 2020).

In most countries around the world, policy makers implemented a kind of lockdown which proved to be, as expected, particularly harmful for the financial markets and the economy overall. During the first quarter of 2020 the S&P 500 lost around 30 % of its peak in February, while the escalated uncertainty caused significant volatility in global stock markets (Baker et al., 2020; Pástor and Vorsatz, 2020; Shehzad et al., 2020). A major consequence of COVID-19 is that global growth is projected at −3.0 % of the GDP in 2020, which is far worse than during the 2008−09 financial crisis (IMF, 2020). The pandemic-driven crisis motivates research to focus on the impact of COVID-19 on stock markets as well as on financing and cost of capital (Goodell, 2020). A number of studies evaluate the impact of the COVID-19 pandemic on stock markets (Al-Awadhi et al., 2020; Ali et al., 2020; Baker et al., 2020; He et al., 2020) and the existence of safe haven investment opportunities (Conlon and McGee, 2020; Conlon et al., 2020; Goodell and Goutte, 2020), while others investigate its effects on credit markets (Halling et al., 2020; Ito, 2020), investor behavior (Harron and Rizvi, 2020; Ortmann et al., 2020) and liquidity of firms (De Vito and Gomez, 2020; Li et al., 2020). However, there is no empirical evidence with regard to the impact of the COVID-19 pandemic on the valuation effects of firms’ financing decisions.

Firms often use syndicated loans as a means of financing in order to satisfy their mid- and long-term financial needs. A syndicated credit facility is a loan provided by one or more arrangers (i.e. large financial institutions and institutional investors) who are responsible for screening and monitoring (Dennis and Mullineaux, 2000; Sufi, 2007; Focarelli et al., 2008). The facility is split into brackets (often of different size) and is offered to the subscribers, who therefore assume the credit risk associated with the borrower. After the loan origination, the lenders cancel or underwrite part, or all of the loan to other investors so they can increase their liquidity and decrease their exposure to an individual borrower. Syndicated loans have significant flexibility as they can be tailored to the borrower’s needs, while they can raise substantial funds since they are placed among a large number of potential borrowers at harmonized conditions for all. In this context, syndicated loan agreements represent a hybrid of private and public debt (Focarelli et al., 2008).

Prior empirical evidence provides inconclusive results with regard to the economic consequences of syndicated loan announcements for the shareholders of borrowing firms. For instance, Fungáčová et al. (2020) suggest that bank loans are associated with positive excess returns for borrowers, while Huang et al. (2012) argue to the contrary. Also, the research findings only partially support the hypothesis of a positive certification effect upon loan announcement during an economic meltdown. The studies of Gasbarro et al. (2017) and Li and Ongena (2015) find that a syndicated loan is considered by the market to be a positive signal during the global financial crisis, while Marshall et al. (2019) and Godlewski (2014) suggest that loan origination became less informative for the creditworthiness of borrowers during the crisis period. The ongoing pandemic disrupts the revenue streams of firms, which have to deal with their fixed costs, debt expenses as well as declining cash balances (Hasan et al., 2020). De Vito and Gomez (2020) suggest that, in the worst case scenario, the average firm would exhaust its cash holdings in about two years and its current liabilities would increase, on average, by eight times. Therefore, firms are expected to resort to the debt market in order to effectively address the financial challenges of the pandemic. Li et al. (2020) show that in the first quarter of 2020, firms faced a liquidity shock and this translated into the largest increase in liquidity demands ever observed in the banking industry. In this context, the valuation effects of bank loans during the COVID-19 pandemic are of primary importance for borrowers, banks and investors, and this is a research gap that needs to be addressed.

Many scholars highlight the role of corporate governance inassisting firms to deal with the economic meltdown brought upon by the COVID-19 pandemic (Ding et al., 2020; Xiong et al., 2020; Mazur et al., 2020; Hu et al., 2020). A particularly important corporate governance mechanism is the board of directors whose primary duty is to protect the interests of shareholders (Fama and Jensen, 1983). The board’s main responsibilities are to provide advice and monitor management (Bonn et al., 2004; De Villiers et al., 2011). The board of directors is a crucial internal corporate governance mechanism in promoting good governance in terms of strategic decision-making (Chen, 2015). Empirical evidence suggests that the structure of the board as well as the executive compensation policies have a significant effect on corporate performance (e.g., Boone et al., 2007; Belkhir, 2009; Aggarwal et al., 2012; Beltratti and Stulz, 2012; Murphy, 2013; Schwartz-Ziv and Weisbach, 2013; Guo and Masulis, 2015; Rau, 2015; Arora and Sharma, 2016; Hermalin and Weisbach, 2017). Using the European context, we first examine the market reaction upon the announcement of syndicated loans and, then, we explore the role of board attributes in explaining the borrowers’ excess returns before and during the pandemic.

Our study intends to shed light on the value implications of syndicated loans during the COVID-19 crisis. For the purposes of our study, we examine the wealth effects of 637 syndicated loan announcements by European firms from January 1, 2018 to July 31, 2020. We find an overall positive market reaction upon the announcement of syndicated loans; however, the examination of the excess returns before and during the pandemic yields interesting findings. More specifically, we provide robust evidence for significant wealth gains for the European borrowers during the COVID-19 pandemic, which are sufficiently higher compared to the excess returns in the pre-pandemic period. The results suggest that the announcement of syndicated loans during the coronavirus-induced financial meltdown has positive implications for the shareholders of borrowing firms, which provides evidence forthe hypothesis that loan origination has a positive certification effect. We further examine the significance of certain corporate governance mechanisms on the wealth effects of borrowing firms before and during the pandemic. The results show that, in both periods, board size has a non-linear relationship with borrowers’ excess returns; however, the relationship is inverse between the two periods. Specifically, increasingly large boards are associated with shareholder value creation before the pandemic, but not during the pandemic. Gender diversity is negatively related to abnormal returns before the pandemic, while there are indications that female directors enhance shareholder wealth during the pandemic. The link of CEO compensation to shareholder return and the combination of the CEO and chairman roles (i.e. CEO duality) both have a positive and significant effect on borrowers’ excess returns before the pandemic only. With regard to the percentage of non-executive board members and the CEO being a board member, the estimated coefficients of both governance mechanisms are insignificant, suggesting no relation with the abnormal returns of borrowers. Methodologically, the results of the study are robust to alternative model specifications with regard to the estimation of abnormal returns. They also hold after addressing sample self-selection bias using propensity score matching.

The findings of our study contribute to the existing literature on several fronts. First, we contribute to the current literature by investigating the market reaction upon syndicated loan announcements before and during the COVID-19 pandemic. In times of high market volatility, this study analyzes whether loan announcements have a favorable impact on borrowers’ stock returns. We do so using different asset pricing models and alternative model specifications in order to fully estimate the information content of the syndicated loan agreements. Second, we advance an emerging field of business research which focuses on the effects of corporate governance mechanisms on firm performance. More specifically, we extend the literature by demonstrating that certain governance practices are significant drivers of value creation for borrowers. In this respect, we provide evidence that the effect of certain governance mechanisms differs prior to and during the pandemic. Therefore, we enrich current literature by illuminating the economic impact of corporate governance practices during periods with different economic conditions. Third, our study contributes to the debate on gender diversity by finding that female directors can have a positive impact on shareholder value during uncertain times, such as the COVID-19 pandemic. The effect of gender diversity is negative before the COVID-19 pandemic, suggesting an ineffective board function when more females are present. However, the negative effect disappears during the pandemic, while there is partial evidence for a significantly positive effect on borrowers’ gains. Women on board can mitigate agency costs, enhance connection and cooperation, and develop positive relationships with other stakeholders, which can be key value attributes during a crisis period. Furthermore, women are more tied to moral, ethical and sustainability issues, and, thus, help businesses to build trust in turbulent periods. Considering the market and regulatory pressures on European firms to improve gender balance on boards (Article 60, Directive, 2013/36/EU; Arnaboldi et al., 2020; Greene et al., 2020), our study adds important knowledge to the literature on the effects of gender diversity on value creation. Finally, the examined sample of European firms can bring useful insights on the nexus between corporate governance and firm performance. The corporate governance landscape in Europe is diverse, considering that the corporate governance framework in each country is a combination of national laws and codes that consist of recommendations for good governance. Since 2003, the European Commission has made considerable efforts through action plans, proposals and directives (i.e. EU Action Plan, 2003 & 2012; Directive, 2006/46/EC; Directive, 2007/36/EC; Directive, 2017/828), towards convergence, aiming to establish a unified corporate governance system in Europe. However, there are still significant differences on corporate governance practices and norms among the European countries. Our study provides robust evidence for the wealth effects of certain corporate governance mechanisms and, thus, can provide useful insights to policy makers, managers and other stakeholders. Furthermore, the different findings concerning the value effects of many governance mechanisms before and during the COVID-19 pandemic, highlight the need for a more flexible and market oriented corporate governance framework that will enable firms to accommodate the dynamics of the external environment. For instance, the results show that during the pandemic, an effective decision-making process, in terms of shareholder value creation, requires small boards, higher participation of female directors and a separation of the CEO and chairman roles. However, large boards, CEO duality and the link between CEO compensation and shareholder return are significant variables of value creation before the pandemic.

The remainder of our paper proceeds as follows: Section 2 presents the relevant literature. Section 3 provides the employed methodology. Section 4 presents the empirical findings while Section 5 runs a variety of robustness tests. Finally, the main findings, limitations and suggestions for further research of this paper can be found in section 6.

Related literature and hypotheses development

Financial consequences of COVID-19

The outbreak of the COVID-19 pandemic caused a new wave of empirical research investigating the consequences of the pandemic on the economy and financial markets. More specifically, Ali et al. (2020) and Shehzad et al. (2020) examined the returns and volatility of global financial markets, finding that all financial securities suffered heavy losses due to COVID-19, while market volatility increased dramatically. The European markets registered the largest negative returns compared to other markets, while financial volatility in the US and Europe increased more than during the global financial crisis. In contrast, the Chinese market showed a lower decline and remained less volatile, providing better opportunities for portfolio optimization. He et al. (2020) suggest that COVID-19 had a severe but short-term effect on stock markets, while the impact of COVID-19 in the Asian stock markets had spill-over effects on European and American markets. Baker et al. (2020) focused on the US stock market, finding that the market plunge caused by the COVID-19 pandemic was more forceful than in previous pandemics.

In this context, Ashraf (2020) evaluated the economic impact of government emergency actions during the COVID-19 pandemic around the globe, such as social distancing measures, public awareness programs, testing and quarantining policies, and income support packages. The results show that social distancing measures have both positive and negative effects on stock markets, while public awareness programs, testing and quarantining policies, and income support packages are related to positive market reactions. Ortmann et al. (2020) find that investors increased their trading activities, reduced the use of leverage and took more short positions during the outbreak of COVID-19. Under the new and unprecedented economic and financial conditions caused by the COVID-19 pandemic, Conlon et al. (2020) and Conlon and McGee (2020) investigate safe haven investment assets, focusing on cryptocurrencies. The results suggest that Bitcoin and Ethereum do not show safe haven properties for equity index investors. In contrast, Tether can be considered a safe haven investment as long as it is pegged to the US dollar.

The wealth effects of syndicated loan announcements

The announcement of syndicated loans entails valuable information for borrowers, causing a subsequent market reaction. Empirical evidence fails to reach a consensus regarding the value implications of loan origination for the shareholders of borrowing firms. The results are also inconclusive as to whether bank loans are indicative of the creditworthiness of borrowers during a crisis period. For instance, Fields et al. (2006), using a sample of 1,111 uncontaminated loan announcements in the US from 1980 to 200, find a positive market reaction in the 1980s; however, the positive returns disappear by the latter part of the sample period. Utilizing the sample of Fields et al. (2006); Byers et al. (2008) examine the relationship between the wealth effects of loan announcements and borrower governance mechanisms. They show higher excess returns for borrowers with weak governance structures, suggesting that banks appear to be substitutes for corporate governance, but only for borrowers whose external governance is weak. Gasbarro et al. (2017) analyze 8,867 syndicated loan announcements by US firms during the period 2004–2012 and find significant positive abnormal returns for borrowers. The authors further show that credit loans generate value in the pre-crisis (01/2004-07/2007) and crisis periods (08/2007-11/2008), while term loans create value in the post-crisis period only (12/2008-12/2012). Li and Ongena (2015) also focus on US corporate borrowers, investigating the value implications of 351 syndicated loan announcements between 2005 and 2009. They find positive and statistically significant abnormal returns during the financial crisis (09/2007-12/2009), but for the period prior to the crisis (01/2005-08/2007) they calculate close to zero excess returns. The authors argue that the certification of corporate borrowers by banks in a financial and economic turmoil plays a significant role for the market, while in an economic boom the banks’ role diminishes.

In contrast, Marshall et al. (2019) and Godlewski (2014) argue that loan origination became less informative regarding the creditworthiness of the borrowers from the global financial crisis onwards. The former examines the market reaction to 1,537 debt announcements (syndicated loans, bilateral loans, public debt, privately placed debt) in the UK from 2001 to 2013, showing significant positive excess returns for borrowers upon the announcement of syndicated loans in the pre-crisis period (2001–2007) and insignificant returns in the post-crisis period (2008–2013). The latter, considering 253 loan announcements by French firms over the 2000–2009 period, finds a negative market reaction upon the announcement of bank loans in the crisis period (the crisis period starts 09/2008), while the market reaction is insignificant before the financial meltdown. Fungáčová et al. (2020) consider 7,136 debt announcements (syndicated loans and bonds) made by borrowers from 17 Western European countries over the period from 1999 to 2012 and find that debt issuance enhances shareholder wealth. The results show that a loan issuance is associated with higher excess returns for borrowers compared to a bond issuance. The authors also argue that the Eurozone crisis has a negative and statistically significant effect on announcement abnormal returns.

Huang et al. (2012) investigate 424 bank loan announcements from Chinese firms over the period from 2001 to 2006 and find significant negative excess returns for borrowers. The results are worse for firms that are perceived to be more vulnerable to expropriation by controlling shareholders. However, loan origination by more efficient banks mitigates the negative relationship between excess returns and vulnerability to expropriation. Gan et al. (2014) also study a sample of bank loans in China, showing a significant negative market reaction to loan announcements for the 1996–2004 period, which disappears in the 2005–2009 period due to a series of reforms in the banking system. Finally, Boscaljon and Ho (2005), examining128 bank loan or borrowing agreements (loan initiations and renewals) in Hong Kong, Korea, Thailand and Taiwan between 1991 and 2002, find positive and significant excess returns for borrowers after the Asian crisis, while they document insignificant returns prior to the crisis. The results support the hypothesis that more information derives from a loan arrangement in times of economic uncertainty.

To formulate our expectations, we rely on prior empirical findings, since there is no extant theory regarding the wealth effects of syndicated loan announcements. Considering that during a crisis period the majority of empirical studies find positive results (i.e. Boscaljon and Ho, 2005; Li and Ongena, 2015; Gasbarro et al., 2017; Fungáčová et al., 2020), while only a few single-country studies present negative findings (i.e. Godlewski, 2014; Marshall et al., 2019), we formulate our first hypothesis as follows:

The announcement of syndicated loans during the COVID-19 pandemic generates positive wealth effects for borrowers.

Corporate governance mechanisms and borrowers’ gains

Corporate governance intends to address the agency problem and protect investors against expropriation by corporate insiders (Shleifer and Vishny, 1997; La Porta et al., 2000). Towards this goal, a number of corporate governance mechanisms have been developed within firms. To examine the effectiveness of these mechanisms, empirical studies investigate their impact on shareholder value around corporate announcements, such as mergers and acquisitions, debt issues, equity offerings, stock splits, as well as earnings or dividend announcements. Both the literature and empirical evidence highlight the board of directors as a crucial internal corporate governance mechanism, as its main responsibilities are to monitor the management and provide counselling and resources (Fama and Jensen, 1983; John and Senbet, 1998; Becht et al., 2003; Tricker and Tricker, 2015). In this paper, we examine whether certain board attributes, such as board size, board gender diversity, CEO duality, CEO board member and the percentage of non-executive board members, have an effect on shareholder wealth upon the announcement of syndicated loans. We detail the theoretical framework and potential value implications for each of the five board attributes below.

Board size

Large boards give rise to agency problems due to coordination issues, internal conflicts among directors and director free-riding (Lipton and Lorsch, 1992; Jensen, 1993; Aggarwal et al., 2012). Therefore, boards become less effective and fail to make value-enhancing strategic decisions. However, drawing on resource-dependent theory, large boards are likely to include more experienced and talented directors and, thus, they can provide better advice and more qualitative counsel to the management (Dalton et al., 1999; Bonn et al., 2004). On the other hand, smaller boards are thought to be more focused, cohesive and efficient, and, therefore, improve the decision-making process of a business organization (Jensen, 1993; Eisenberg et al., 1998; Coles et al., 2008). Considering that the origination of a syndicated loan is a complex process, requiring timely response and flexibility, we conjecture that large boards may be unable to carry out effectively such a financing arrangement, while smaller boards may be able to communicate and coordinate better to arrange a syndicated loan under the conditions of the pandemic-induced financial meltdown. On these grounds, the following hypothesis is suggested.

Borrowers with larger boards are associated with lower abnormal returns.

Board gender diversity

Women and men directors may differ in terms of risk-aversion, trust, ethics and friendliness (Ali et al., 2014; Zalata et al., 2019). Women increase board independence, provide better advice to management and have the ability to better monitor executives (de Cabo et al., 2012; Cumming et al., 2015; Farag and Mallin, 2017). Therefore, the inclusion of more women on board can minimize agency costs (Reguera-Alvarado et al., 2017). According to resource dependence theory, women on board can bring greater connection, cooperation and understanding to build positive relationships with others (Bear et al., 2010; Liao et al., 2019). Compared to male directors, female ones are more likely to improve risk management, contribute to compliance with gender and legal issues, advance sustainability objectives and enhance corporate governance (Kim and Starks, 2016). In addition, stakeholder theory suggests that firms should not focus exclusively on shareholders’ interests, but they should also meet the interests of multiple stakeholders, such as employees and customers, in order to achieve long-term success (Prado-Lorenzo and Garcia-Sanchez, 2010; Chakrabarty and Bass, 2014; Harjoto et al., 2015). Considering that women tend to focus on social and moral values, corporate social responsibility and voluntary disclosure of non-financial information, the inclusion of more females on board is further supported by stakeholder theory (Chakrabarty and Bass, 2014; Hussain et al., 2018; Liao et al., 2019). Hence, borrowers with a larger fraction of female directors can have more effective and better performing boards, while they can develop better relationships and stronger connections with the leading institutions. The discussion leads us to propose the following hypothesis.

Borrowers with a larger fraction of women on board are associated with higher abnormal returns.

CEO duality

The combination of the CEO and chairman roles by the same person may reduce objectivity and accountability in decision-making. As argued by Jensen (1993), CEOs who are also chairmen are more likely to dominate their boards, which become ineffective in monitoring managerial opportunism. Duality promotes CEO entrenchment and that may force CEOs to pursue their self-interests at the expense of the interests of shareholders (Elyasiani and Zhang, 2015; Defrancq et al., 2021). In contrast, stewardship theory argues that CEO duality reduces communication conflicts and information asymmetry; CEOs who are also chairmen are able to respond more effectively to external events, make better decisions and provide a clear strategic direction (Rechner and Dalton, 1991; Krause et al., 2014; Pham et al., 2015). The decision for a syndicated loan requires reliable and timely information, while it is also considered under the overall funding strategy of a firm. In addition, a syndicated loan should be decided once all other funding solutions have been examined thoroughly by the top management and the board as well. The above arguments suggest that the decision of a firm to combine (or not) the roles of the CEO and chairman would have a significant effect on value creation at the announcement of a syndicated loan; however, the direction of that influence is unclear. We, therefore, formulate the following hypothesis.

The combination of the CEO and chairman roles is associated with borrowers’ abnormal returns.

CEO board member

According to agency theory, CEOs who participate in the board are likely to exert influence over board decisions and behave opportunistically due to their information advantage about the operations of the firm (Fama and Jensen, 1983; Jensen, 1993; Hermalin and Weisbach, 2001). As a result, CEO board participation may reduce the board’s independence and hinder the effective monitoring of the CEO by the board. Resource dependence theory asserts that the participation of the CEO on board strengthens the flow and the quality of information to other board members and, thus, it facilitates the effective functioning of the board (Pfeffer, 1973; Hillman and Dalziel, 2003). Stewardship theory (Davis et al., 1997; Tosi et al., 2003) also argues in favour of CEO participation on board, suggesting that it reduces the costs of information transfer and improves board collaboration on various corporate matters. This advances the quality of decision-making and enables effective strategic planning and implementation. Based on the above, the decision of a firm to be financed through a syndicated loan may be due to the effective functioning of its board resulting from the participation of the CEO on it. However, such a decision may be taken due to managerial entrenchment and poor CEO monitoring by the board. Indeed, we expect this corporate governance mechanism to have a significant (positive or negative) effect on shareholder value on the announcement of syndicated loans. Therefore, we make the following conjecture.

The participation of the CEO on board is associated with borrowers’ abnormal returns.

Percentage of non-executive board members

The recruitment of non-executive directors on board aims to improve monitoring of managerial behavior, actions and performance, and to reassure shareholders that the decision-making processes on board are aligned with their interests (Hart, 1995; Mura, 2007). Non-executive directors are effective monitors of top management and tend to control any opportunistic behavior of other directors on board. Thus, they contribute to mitigating agency costs, while they also bring valuable resources, specialist knowledge and experience to the board (Khan et al., 2021). The appointment of non-executive board members is an effective means of balancing interests on board for any firm, irrespective of its formal board structure. A syndicated loan is a major strategic financial decision and, thus, is expected to be decided after an effective and independent supervisory function carried out by the non-executive directors. On that basis, the percentage of non-executive board members is expected to have a positive impact on borrowers’ wealth. Hence, we formulate the following hypothesis.

Borrowers with higher percentage of non-executive directors are associated with higher abnormal returns.

Link of CEO compensation to total shareholder value

In addition to the above analyzed board attributes, we include in our analysis the link of CEO compensation to shareholder value. Firms apply various compensation packages as a means of mitigating the conflicts of interest between management and shareholders (Ozkan, 2011). Equity-based compensation is considered an effective mechanism to align the interests of managers with those of shareholders (Jensen and Meckling, 1976). Empirical evidence suggests that there is an association between executive compensation and a firm's debt financing policy (Almazan and Suarez, 2003; Coles et al., 2006; Brockman et al., 2010; Castro et al., 2020). More specifically, firms choose to get financed with syndicated loans when CEO compensation is tied to stock performance (Albring et al., 2011). Equity incentives encourage CEOs to use bank financing when the profitability of the underlying projects is expected to be high. On the basis of the positive link between CEO equity compensation and the issuance of syndicated loans, the following hypothesis is proposed.

The link of CEO compensation to total shareholder value is associated with higher abnormal returns for borrowers.

Data and methodology

Syndicated loans sample

We retrieve data on syndicated loan announcements using Thomson One DealScan2 . The final sample consists of 637 such announcements from January 1, 2018 to July 31, 2020 that comply with the following criteria:

Borrowers are domiciled in Europe and are listed firms in a European Stock Market.

Borrowers have available stock price data 270 days before and 20 days after the announcement day available from Thomson Reuters Datastream and sufficient financial statement information at the year-end prior to the loan announcement date from Thomson Reuters Worldscope.

Borrowers have available corporate governance information at the year-end prior to the loan announcement date available from Refinitiv.

We exclude from the sample financial firms (SIC 6000-6999) and utilities (SIC 4900-4999) due to the highly regulated nature of these industries (John et al., 2011).

We exclude contaminated dates and compounded events (Fungáčová et al., 2020) to capture the “pure” effect of each loan announcement.

Description of borrowers

Table 1 provides a cross-country description of the financial characteristics of borrowers and the syndicated loan announcements during the examined period. The majority of loans were announced by UK firms (190), while firms from France and Germany follow with 87 and 79 announcements, respectively. On the contrary, firms from Poland, Ukraine and Hungary show the lowest number of loans with 7, 5 and 3 announcements, respectively. UK firms also show the highest number of leveraged/high yield loans (35), followed by French (19) and Spanish firms (18). Interestingly, in Denmark and Hungary no such loans were announced. The highest average loan amount is found in Luxemburg (3,077.3$m.) and the lowest in Ukraine (237 $m.). The average size of borrowers as proxied by their market value is 17,734 $m. Borrowers from Hungary are by far the largest; however, there are only three loan announcements by Hungarian firms. Borrowers from Sweden, Denmark and Switzerland are also large, while borrowers from Austria, Ukraine and Finland are the smallest ones. A similar picture emerges for borrowers when their size is proxied by their total assets. It is notable that Ukrainian firms are ranked at a high level based on their total assets, while they are second to last based on their market value. Glamor borrowers (i.e. firms with high price-to-book value of equity ratio) come from Switzerland, Denmark, UK and the Republic of Ireland, while value acquirers (i.e. firms with low price-to-book value of equity ratio) are domiciled in Luxembourg, Ukraine and Hungary. Considering the total debt to total assets, the total debt to common equity and the long term debt to common equity ratios, the most indebted borrowers come from Norway, Spain and Italy, while borrowers from Luxembourg, Hungary, and Sweden show the lowest debt burden. Examining the level of profitability as proxied by the return on assets ratio, borrowers from Denmark and Sweden are the most profitable, followed by the Ukrainian ones. In contrast, borrowers from Spain, Norway and Austria are ranked at the bottom in terms of profitability. Danish and Swedish borrowers are not only highly profitable but also have high growth opportunities based on Tobin’s Q ratio. The level of price volatility does not fluctuate significantly among borrowers; however, borrowers from Norway show the highest risk, followed by borrowers from Spain and Luxemburg. The less risky borrowers are from Hungary, Switzerland and Ukraine. Finally, borrowers from the UK have the longest history, while borrowers from Austria, Belgium and the Netherlands are ranked at a high level too. In contrast, borrowers from Ukraine, Norway and Poland have the shortest history.

Table 1

Cross-country statistics.

Borrower Nation	N	Highyield	Loanamount	Marketvalue	Totalassets	Price/book value	Total debt/total assets	Fixed assets/common equity	Total debt/common equity	Long term debt/common equity	Return onassets	Tobin's Q	Pricevolatility	Age
Austria	12	4	1,359	3,753	7,634	1.66	30.92	113.55	104.79	84.61	3.83	0.84	0.31	31.75
Belgium	10	3	823	5,436	6,056	2.00	33.31	77.46	87.38	73.35	4.51	1.24	0.27	29.70
Denmark	12	0	1,096	46,729	82,692	3.03	22.33	62.64	66.78	57.67	8.86	1.64	0.33	27.58
Finland	25	4	526	5,319	5,679	2.20	24.64	67.28	68.74	40.47	4.95	1.21	0.33	23.88
France	87	19	2,207	20,951	26,098	2.25	27.11	55.84	76.63	76.19	4.54	1.14	0.27	27.98
Germany	79	17	2,389	20,120	39,189	2.40	22.63	75.43	80.05	64.11	5.66	1.20	0.28	27.03
Hungary	3	0	667	231,337	367,600	0.93	15.81	119.31	36.11	21.27	7.69	0.73	0.24	23.67
Italy	32	12	1,032	5,781	15,127	2.54	32.06	47.32	120.33	93.08	4.02	1.25	0.29	22.03
Luxembourg	11	4	3,077	13,022	62,906	0.58	19.55	93.43	42.57	34.83	5.55	0.53	0.34	22.82
Netherlands	29	7	1,498	13,053	31,034	2.72	27.41	77.27	94.02	77.22	5.13	1.37	0.30	29.07
Norway	25	8	827	35,966	40,074	2.75	34.36	163.36	126.95	99.77	3.13	1.41	0.40	12.88
Poland	7	2	1,248	19,329	33,232	1.13	26.55	68.41	59.61	48.34	6.32	1.00	0.30	13.57
Republic of Ireland	14	4	1,048	8,724	8,876	2.98	21.73	48.28	50.76	45.18	5.08	1.56	0.30	20.00
Spain	33	18	1,246	7,700	10,577	2.56	36.46	66.49	87.58	139.13	1.01	1.17	0.34	13.70
Sweden	40	6	713	52,083	50,458	2.34	26.54	9.19	7.05	51.33	8.91	1.61	0.27	21.03
Switzerland	23	3	2,648	45,292	34,517	3.99	30.95	80.06	103.93	87.64	7.00	1.83	0.24	27.87
Ukraine	5	5	237	4,012	55,437	0.82	32.92	50.56	59.85	43.32	8.63	0.81	0.26	11.60
United Kingdom	190	35	958	5,971	10,047	2.96	24.01	68.91	56.69	68.31	6.40	1.40	0.31	33.26
Total	637	151	–	–	–	–	–	–	–	–	–	–	–	–
Average	–	–	1,416	17,734	25,779	2.57	26.51	68.18	71.89	72.51	5.54	1.31	0.30	26.77

The table presents the cross-country descriptive statistics for the loan-specific and the borrower-specific characteristics. The sample of loans is distributed according to the number (N) of the total loan announcements included in the sample, the number of loans that considered as leveraged/high yield in DealScan database, the average deal loan amount in million $, the average market value in million $, the average total assets in million $, the average ratio of price to book value of equity, the average ratio of total debt to total assets (%), the average ratio of common equity to total assets (%), the average ratio of fixed assets to common equity (%), the average ratio of long-term debt to common equity (%), the average ratio of return on assets (%), the average Tobin’s Q ratio, the average price volatility, and the average borrower’s age in years. All continuous variables are winsorized at the 1% and 99% levels.

Measurement of abnormal returns

The abnormal returns of European borrowers’ loan announcements are estimated using the market model, as shown in Eq. (1).Where is the expected return of borrower i at day t, α and are the model’s intercept and slope respectively, is the return of the market portfolio and is the error term with zero mean and constant variance .We use a specific benchmark index as proxy of the market portfolio for each country3 . To enhance the robustness of our results, we also use two other indices as proxies of the market portfolio, namely the STOXX Europe 600 index and the EURO STOXX 50 index.

The daily abnormal returns are estimated as shown in Eq. (2):Where AR is the abnormal return of borrower i at day t, R is the realized return of borrower i at day t and is the expected return of borrower i at day t calculated from Eq. (1).

The announcement period cumulative abnormal returns (CARs) are the sum of the daily abnormal returns over the length of the applied event window surrounding the loan announcement day (day 0), as shown in Eq. (3).

The statistical significance of the CARs is tested using the parametric Pattell-Z test as well as the nonparametric Corrado's rank test4 .

Cross-sectional analysis

Prior empirical studies suggest that borrower-specific factors as well as loan-related characteristics have an effect on shareholder wealth upon the syndicated loan announcements. To capture the impact of the COVID-19 pandemic on borrowers’ gains after controlling for the effect of other variables, we estimate Eq. (4) as follows.Where the dependent variable, is the cumulative abnormal return of borrower i for the period (t. The intercept β measures the excess returns of borrowers after controlling for the effects of the COVID-19 pandemic and a set of m control variables included in vector X. B is a vector containing the estimated coefficients of all control variables.

To capture the effect of COVID-19 on shareholder wealth we use a dummy variable COVID which takes the value 1 for the events that were announced after the declaration of the disease as a pandemic. We define the pre-pandemic period as the period from 01 January, 2018 to 10 March, 2020 and the pandemic period from 11 March, 2020 to 30 July, 2020 following the declaration of the WHO.

Based on the relevant literature, we control for various borrower-specific characteristics that are thought to affect the market reaction to loan arrangements. More specifically, we use the borrowers’ age as an indicator of information asymmetry (Leary and Roberts, 2010), the natural logarithm of borrowers’ total assets as a proxy of their size (Zhang, 2008), the borrowers’ return on assets as a proxy of their profitability (Huang et al., 2012) and the borrowers’ Tobin’s Q as a proxy of their growth (Fotak and Lee, 2020). The ratio of fixed assets to common equity is used as a proxy of tangibility (Graham et al., 2008) and the ratio of price volatility is used as a proxy of risk (Focarelli et al., 2008).

Furthermore, we control for loan-related determinants and specifically we control for the loan amount (Huang et al., 2012) and for the high yield/leveraged loans (Huang et al., 2012; Kim et al., 2018).

With regard to the borrowers’ corporate governance mechanisms, we use the board size, the percentage of non-executive board members on board, the CEO duality, the CEO board member and the link of CEO compensation to total shareholder value (Liu et al., 2012). Furthermore, given that gender board diversity has a crucial role in organizational outcomes (Sila et al., 2016; Sarhan et al., 2019), we also investigate its impact on value creation upon syndicated loan announcements. As a gender diversity proxy, we calculate the Blau Index as shown in Eq. (5) (Owen and Temesvary, 2018):Where, P denotes the proportion between male and female directors on board and g is the gender variable. Therefore, the maximum value of Blau Index equals 50 % and represents a board with equal representation between male and female directors.

Table 2 presents the definitions and the descriptive statistics of the above variables.

Table 2

Summary statistics.

	Unit	Definition	N	Mean	Q1	Median	Q3	Std. Dev.
Panel A. Borrowers financial-related Variables
AGE	Natural Logarithm	Natural logarithm of the days between the loan announcement date and the first date of the company’s first record in Datastream	637	8.92	8.55	9.07	9.67	0.82
SIZE: Total Assets	Natural Logarithm	Natural logarithm of Borrowers total assets at year-end preceding the loan announcement.	633	15.73	14.51	15.65	16.87	1.70
Total Debt to Common Equity	%	Borrowers Total Debt as a percentage of Common Equity at year-end preceding the loan announcement.	633	71.89	32.62	62.23	111.22	157.02
Return on Assets	%	Borrowers’ return on assets ratio at year-end preceding the loan announcement	628	5.54	2.92	5.02	7.90	5.65
Fixed Asset to Common Equity	%	Borrowers Fixed Assets as a percentage of Common Equity at year-end preceding the loan announcement.	633	68.18	20.14	50.41	100.83	106.66
GROWTH: Tobin’s Q Ratio	%	Borrowers Tobin’s Q ratio at year-end preceding the loan announcement.	632	1.31	0.76	1.03	1.64	0.85
RISK: Price Volatility	%	Stock average annual price movement to a high and low from a mean price at year-end preceding the loan announcement	637	0.30	0.22	0.27	0.34	0.12
Panel B. Loan-related Variables
Loan Amount	Natural Logarithm	Natural logarithm of the loan amount ($)	637	6.44	5.54	6.45	7.32	1.29
High Yield/Leveraged	1/0	Dummy variable that is assigned a value of 1 for a loan considered Leveraged/High Yield in DealScan database and 0 otherwise	637	0.24	0	0	0	0.43
Panel C. Corporate Governance Variables
Board Size	Natural logarithm	Natural logarithm of board size at year-end preceding the loan announcement	637	2.27	2.08	2.30	2.48	0.34
Non-Executive Board Members	%	Percentage of non-executive board members to board members at year-end preceding the loan announcement	637	80.99	71.43	81.82	91.67	13.95
Blau Index for Gender Diversity	Ratio	Index of Gender Diversity calculated as shown in Eq.5 at year-end preceding the loan announcement:	637	0.39	0.34	0.43	0.47	0.12
CEO Board Member	1/0	Dummy variable that is assigned a value of 1 for CEOs who are board members and 0 otherwise at year-end preceding the loan announcement	599	0.65	0	1	1	0.48
CEO Compensation Link to Total Shareholder Return (TSR)	1/0	Dummy variable that is assigned a value of 1 if the CEO's compensation is linked to total shareholder return (TSR) and 0 otherwise at year-end preceding the loan announcement	637	0.55	0	1	1	0.50
CEO duality	1/0	Dummy variable that is assigned a value of 1 if the CEO simultaneously chairs the board and 0 otherwise at year-end preceding the loan announcement	637	0.20	0	0	0	0.40

The table defines the variables used in the empirical analysis. All continuous variables are winsorized at the 1% and 99% levels.

Difference-in-Differences analysis

To further examine whether the abnormal returns of acquirers during the COVID-19 pandemic (treatment group) differ significantly from the abnormal returns of acquirers before the pandemic (control group), we employ a difference-in-differences (DID) approach. This approach is estimated using the following model:

In the above specification, the interaction term between the COVID-19 pandemic and corporate governance mechanisms constitutes the DID estimator. With regard to the continuous corporate governance variables (i.e. board size, percentage of non-executive board members and Blau Index) and in order to transform them into dummies, the percentile rank scoring methodology was adopted. More specifically, for each continuous variable we introduced a dummy variable that was assigned a value of 1 for values above the third quartile and 0 for values at or below the third quartile. The advantage of the DID approach is that it allows the comparison of abnormal returns between comparative samples of borrowers before and during the COVID-19 pandemic, while controlling for differences in firm-specific characteristics. This approach enables us to make more robust inferences regarding the potential impact of the examined corporate governance mechanisms on the excess returns of borrowers.

Results

Univariate analysis

At the first stage of our analysis, we estimate the CARs upon the announcement of syndicated loans for the entire sample period. Then, we split our sample into two subsets based on the WHO declaration of COVID-19 as a pandemic on 11th March. Table 3 reports the CARs for the European borrowers estimated with the market model for the entire period. Panels A, B and C present the results of the market model employing three different proxies of the market portfolio, namely a domestic market index of each country, the STOXX Europe 600 index and the EURO STOXX 50 index, respectively. The results deriving from all three models show significant positive or insignificant abnormal returns for borrowers in short event windows surrounding the announcement date. This indicates that syndicated loan arrangements are not negative events for European borrowers.

Table 3

Cumulative Abnormal Returns upon syndicated loan announcements (entire period).

	Panel A: Market model (each country’s index)						Panel B: Market model (STOXX Europe 600)						Panel C: Market model (EURO STOXX 50)
Eventwindow	Mean	Median	SD	% Pos	Patell-Z	Corrado	Mean	Median	SD	% Pos	Patell-Z	Corrado	Mean	Median	SD	% Pos	Patell-Z	Corrado
[-10,10]	0.15	0.45	13.80	52	−2.648a	−0.946	−0.06	0.17	13.54	51	−4.096a	−1.639	−0.15	0.33	13.75	52	−4.439a	−1.793c
[-3,3]	0.30	0.36	9.45	53	1.621	0.473	0.21	0.22	9.32	53	0.676	−0.029	0.19	0.12	9.43	51	0.595	−0.199
[-2,2]	0.28	0.23	8.92	54	4.248a	0.800	0.21	0.20	8.83	52	3.436a	0.250	0.17	0.14	8.92	53	3.226a	0.079
[-1,0]	0.24	0.25	6.94	53	2.568b	1.357	0.25	0.23	6.88	53	2.591a	1.319	0.21	0.25	6.94	54	2.232b	1.143
[-1,1]	0.20	0.23	7.91	53	2.514b	1.140	0.22	0.23	7.86	54	2.554b	1.108	0.16	0.26	7.91	53	1.925c	0.751
[-1,10]	0.23	0.31	11.25	53	1.253	−0.703	0.12	0.24	11.16	52	0.100	−1.204	0.06	0.13	11.24	52	−0.108	−1.315
[0,0]	0.00	0.08	5.88	52	−0.659	1.407	0.03	0.09	5.82	52	−0.245	1.529	−0.01	0.08	5.87	53	−0.803	1.355
[0,1]	−0.03	0.09	7.12	52	−0.725	1.039	0.00	0.16	7.08	52	−0.291	1.125	−0.07	0.10	7.14	53	−1.179	0.732

This table reports the cumulative abnormal returns (CARs) surrounding syndicated loan announcements. The sample comprises 637 loan agreements from 434 European borrowers between 01/01/2018 and 31/07/2020. Panel A, B and C present the mean, median, standard deviation and the positive percentage of CARs derived from the market model and estimated using: (a) each country’s benchmark index; (b) STOXX Europe 600 index; and (c) EURO STOXX 50 index, respectively. The statistical significance of CARs is accessed using the Patell−Z test and the Corrado test. The superscripts a, b and c denote significance at 1%, 5% and 10 % levels, respectively.

To further investigate the wealth effects of syndicated loan announcements we split the sample into two subsets with regard to the declaration of COVID-19 pandemic. Panels A and B of Table 4 report the abnormal returns from January 1, 2018 to March 10, 2020 and from March 11, 2020 to July 30, 2020 respectively. The results reveal two different patterns of the abnormal returns prior to and during the COVID-10 pandemic. More specifically, the returns are either negative and statistically significant or statistically insignificant during the first period, suggesting that loan agreements, at best, do not destroy value for borrowers. On the contrary, the returns are positive and statistically significant across all event windows during the pandemic. The gains for borrowers range from around 1% to above 2.5 % and are statistically significant at the 1% level mainly, indicating that loan announcements are positive events for borrowers during the pandemic-induced crisis. Panel C confirms the value discrepancies between the two periods, since the mean differences are statistically significant across all event windows. The results suggest that the market has positive prospects for firms that announce syndicated loans during the economic turmoil driven by the pandemic, supporting hypothesis H and consistent with Li and Ongena (2015).

Table 4

Cumulative Abnormal Returns upon syndicated loan announcements (before and during-the pandemic).

	Panel A: Prior-pandemic(N = 480)						Panel B: During-pandemic(N = 157)						Panel C: Test for differences(2-1)
EventWindow	Mean	Median	SD	% Pos.	Patell-Z	Corrado	Mean	Median	SD	% Pos.	Patell-Z	Corrado	Mean	Median	t-test	MWU
[-20,20]	−1.04	−0.16	19.06	49	−3.566a	−1.219	5.58	4.28	24.17	62	4.822a	1.296	6.61a	4.44a	3.125	−3.846
[-10,10]	−0.52	0.01	12.25	50	−3.589a	−1.281	2.22	2.34	17.61	60	0.942	0.333	2.74c	2.33a	1.812	−2.801
[-3,3]	−0.19	0.08	8.55	51	−0.575	−0.321	1.81	1.26	11.68	61	4.272a	1.431	1.99b	1.18b	1.972	−2.170
[-2,2]	−0.18	0.18	8.38	53	0.879	0.203	1.66	0.74	10.31	58	7.019a	1.192	1.83b	0.56	2.018	−1.516
[-1,0]	−0.24	0.07	5.75	51	−1.193	−0.133	1.71	0.78	9.60	59	7.258a	2.811a	1.95b	0.70b	2.409	−2.411
[-1,1]	−0.24	0.13	7.27	51	−0.889	0.188	1.56	0.73	9.53	57	6.618a	1.871c	1.80b	0.60c	2.167	−1.742
[-1,10]	−0.55	0.02	10.29	50	−2.433b	−1.138	2.62	2.03	13.56	60	6.778a	0.558	3.17a	2.01c	2.689	−2.460
[0,0]	−0.35	0.04	4.89	51	−3.168a	0.083	1.08	0.47	8.11	55	4.213a	2.559a	1.43b	0.43c	2.090	−2.246
[0,1]	−0.35	0.05	6.47	51	−2.796a	0.420	0.93	0.71	8.75	55	3.429a	1.290	1.28c	0.65	1.685	−1.208

This table reports the cumulative abnormal returns (CARs) upon syndicated loan announcements between 01/01/2018 and 31/07/2020. CARs that derived from a multi-country market model were estimated using each country’s benchmark index. Panels A and B present the mean and median CARs, standard deviation, percentage of borrowers with positive CARs for loan announcement prior to the pandemic (N = 480) and for loan announcements during the pandemic (N = 157), respectively. The statistical significance of CARs is accessed using the Patell-Z test and the Corrado test. Panel C reports the mean and median differences of CARs between announcements before and during the pandemic. The statistical significance of the differences between the means and the medians of the two subgroups are tested using the t−test of equality of means and the Mann−Whitney U test, respectively. The superscripts a, b and c denote significance at 1%, 5% and 10 % levels, respectively.

Cross-sectional analysis of borrowers’ gains

We estimate several regression models in order to investigate the effect of the selected determinants on shareholder value surrounding the announcement of syndicated loans. Following prior empirical studies (i.e. Allen et al., 2019; Fotak and Lee, 2020), all models are estimated using robust standard errors and include (unreported) fixed effects for the borrowers’ industry sector based on the two-digit SIC codes and for the borrowers’ country of stock exchange listing. Moreover, to enhance the robustness of our results we winsorize all continuous variables at 1% and 99 % levels in order to alleviate the potential impact of outliers and we report the VIF statistic as an indicator of the severity of potential multicollinearity. Similar to prior literature (i.e. Balafas and Florackis, 2014; Stellner et al., 2015; Chung et al., 2017; Lahlou and Navatte, 2017; Fotak and Lee, 2020), we apply lagged values of all control variables to address potential endogeneity. We also use as dependent variable the borrower 3-day CARs centered on the announcement day (i.e.−1,1), similarly to relevant studies (e.g., Marshall et al., 2019; Fungáčová et al., 2020). Sections 4.2.1 and 4.2.2 present the multivariate results for the entire period and for the two distinct periods (before and during the COVID-19 pandemic), respectively.

Regression results for the entire period

Table 5 reports the results of the multivariate analysis for the entire sample period. All models are estimated with industry and country fixed effects. Using various control factors for borrower-characteristics, loan-characteristics as well ascorporate governance specific variables, we find that COVID-19 pandemic has a positive and statistically significant effect on shareholder value upon the announcement of syndicated loans. This confirms the results of the univariate analysis and further supports hypothesis H, indicating that a loan origination from multiple lenders during the pandemic conveys a signal of creditworthiness and a strong business model. With regard to corporate governance mechanisms, the regression models provide inconclusive results since the estimated coefficients are statistically significant in a few model specifications. Thus, we provide evidence for a negative and statistically significant effect of CEO board member on borrowers’ excess returns, supporting hypothesis H. This is in line with agency theory, suggesting reduced independence and ineffective CEO monitoring by the board with regard to the decision for a syndicated loan. The combination of CEO and chairman roles (i.e. CEO duality) seems to have a positive impact on shareholder value. This finding partially supports hypothesis H and is consistent with stewardship theory. Thus, firms that combine the two roles seem to have better flow of information, effective decision-making and a clearer strategic orientation, which allow them to originate value-enhancing syndicated loans. There is also an indication for a negative effect of board size on returns, which provides partial evidence for hypothesis H. Consistently with agency theory, it seems that firms with large boards are not able to create shareholder value through the announcement of syndicated loans. The remaining governance mechanisms, namely gender diversity, the percentage of non-executive board members and the link between CEO compensation and shareholder return, present insignificant coefficients suggesting that they do not have any explanatory power over the abnormal returns of borrowers. Among the other control variables, only age is positive and statistically significant in all models. Age is an inverse indicator of borrowers’ information asymmetry and therefore the higher the age, the lower the information asymmetry related to the borrower. Hence, the reduced information asymmetry creates a positive signal to the investors and in this way positive market reaction is generated for the “older” borrowers.

Table 5

Determinants of value creation upon syndicated loan announcements (entire period).

	(1)	(2)	(3)	(4)	(5)
Constant	−0.062(-1.18)	−0.072c(-1.89)	−0.097b(-2.25)	0.005(0.12)	−0.056(-1.13)
Age	0.010b(2.26)	0.009b(2.06)	0.009b(2.17)		0.009b(2.07)
Total Assets (Ln)	0.002(0.55)	0.000(0.01)	0.001(0.36)	0.002(0.59)	0.001(0.21)
Total Debt % Common Equity	0.000(-0.74)	0.000(-0.71)	0.000(-0.59)	0.000(-0.73)	0.000(-0.44)
Return on Assets	0.001(1.34)	0.001(1.26)	0.001(1.27)		0.001(1.37)
Fixed Assets % Common Equity	0.000(0.60)		0.000(-0.10)	0.000(0.02)	0.000(-0.27)
Tobin’s Q	−0.003(-0.60)				−0.003(-0.79)
Price Volatility	0.015(0.33)			−0.014(-0.32)	0.002(0.04)
Loan Amount (Ln)	−0.002(-0.33)	−0.003(-0.57)	−0.003(-0.69)	−0.003(-0.64)	−0.002(-0.49)
High Yield/Leveraged LoanDummy	0.009(1.38)	0.008(1.20)	0.010(1.45)	0.006(0.84)	0.008(1.16)
Board Size (Ln)	−0.023c(-1.79)		−0.016(-1.24)	−0.010(-0.80)	−0.014(-1.11)
Non-Executive Board Members (%)	0.000(0.27)		0.000(1.45)
Blau Index	−0.041(-1.44)			0.000(0.00)	−0.002(-0.06)
CEO Board Member	−0.030a(-3.42)
CEO Compensation Link to Shareholder Return	0.009(1.40)
CEO duality	0.022a(3.38)		0.011(1.23)		0.009(1.04)
COVID-19 CRISIS	0.019a(2.57)	0.013c(1.79)	0.014c(1.84)	0.013c(1.77)	0.014c(1.85)
Industry Dummies	Yes	Yes	Yes	Yes	Yes
Country Dummies	Yes	Yes	Yes	Yes	Yes
N	586	625	625	630	624
R2	0.1613	0.1205	0.1283	0.1067	0.1258
VIF	5.00	5.18	5.19	5.27	5.14

This table reports the results of the cross-sectional OLS regression analysis with robust standard errors for announcement period (3-days) excess returns of borrowers estimated using the market model. Coefficients are reported and t-statistics are presented in parentheses. The Huber-White robust standard errors are used to calculate t-statistics in all models. All variables are winsorized at the 1% and 99% levels. For more details about the definition of each variable see Table 2. The subscripts a, b and c denote significance at 1%, 5% and 10% levels, respectively.

Regressions results before and during the COVID-19 pandemic

To further investigate the determinants of value creation upon syndicated loan announcements with regard to the COVID-19 pandemic, we apply separate cross-sectional analysis for the periods before and during the pandemic. Panels A and B of Table 6 present the results of multivariate analysis for the determinants of the borrowers’ excess returns before and during the pandemic. Considering the corporate governance variables, we find significant differences between the two periods. More specifically, the results show a nonlinear relationship between board size and abnormal returns for the period before the pandemic. This implies that the wealth gains of borrowers will decrease as the number of directors increases to a certain point; after this point, the effect of more directors on the abnormal returns will be positive. For the period after the pandemic, the relationship between board size and shareholder value is inverted U-shaped. Thus, increasingly large boards are problematic during an economic meltdown, failing to create more value. The different effect of board size on shareholder value before and during the COVID-19 pandemic indicates that the optimal size of a board depends on the opportunities, risks and challenges of the external environment. The pandemic-induced crisis caused a dramatic shortfall in revenue streams and so firms had to address the unexpected revenue shock. Firms with smaller boards were able to respond more effectively to the impacts of the crisis and get financed through the origination of syndicated loans and, thus, enhance shareholder value. Gender diversity has a negative and statistically significant impact on borrowers’ excess returns before the pandemic, suggesting that the inclusion of more female directors may hamper the financing decisions and therefore destroy value. In contrast, the results provide partial evidence for a positive effect of board gender heterogeneity on shareholder value during the COVID-19 pandemic. The different effect of gender diversity on borrowers’ excess returns before and after the pandemic suggests that in times of increased uncertainty women can contribute positively to the quality of decision-making on board. Consistently with resource dependence and stakeholder theories, women can be particularly useful in building connections with financial intermediaries and meeting their interests as well. They also tend to be focused on moral values, ethics and sustainability issues, building a sense of trust that allows firms to negotiate a syndicated loan origination with banks with better rates and terms. The link between CEO compensation and shareholder return is significant and positive before the pandemic, while after the pandemic the relationship between the two variables becomes insignificant. Considering that the outbreak of the COVID-19 pandemic triggered a plummet in share prices, equity-based compensation had no real worth as a performance incentive. The link of CEO compensation to shareholder value is significantly positive before the crisis, corroborating the notion that CEOs are keener on bank financing when the expected profitability of the underlying projects is high. The decision of borrowers to combine the CEO and chairman roles is significantly positively related to the announcement period excess returns before the pandemic; however, the positive effect of CEO duality on returns disappears during the pandemic. The COVID-19 crisis changed the dynamics of the business environment, causing, amid others, a dramatic increase in economic uncertainty. Under these unprecedented conditions, firms were in need of enhanced CEO monitoring and board independence. Therefore, the separation of the CEO and chairman roles would be preferable in order to ensure that the decision for loan syndication was the product of objective assessment by the board. Finally, the percentage of non-executive board members and the CEO being a board member have a similar (non-significant) effect on borrowers’ wealth gains before and during the COVID-19 pandemic.

Table 6

Determinants of value creation upon syndicated loan announcements (before and during the pandemic).

	Panel ASyndicated loan announcements before the pandemic					Panel BSyndicated loan announcements during the pandemic
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)
Constant	−0.030(-0.50)	−0.044(-0.74)	−0.041(-0.69)	0.167(1.34)	0.197(1.50)	−0.242(-1.22)	−0.329(-1.55)	−0.204(-1.24)	−0.732c(-1.96)	−1.091b(-2.23)
Age	0.007(1.45)	0.007(1.43)	0.008(1.45)	0.007(1.50)	0.007(1.44)	0.023c(1.68)	0.022c(1.72)	0.022c(1.76)	0.024c(1.85)	0.026c(1.87)
Total Assets (Ln)	0.004(0.90)	0.004(1.03)	0.004(0.93)	0.004(0.96)	0.004(0.98)	−0.005(-0.40)	−0.006(-0.57)	−0.007(-0.63)	−0.011(-0.95)	−0.012(-0.98)
Total Debt % Common Equity	0.000(-0.85)	0.000(-0.86)	0.000(-0.80)	0.000(-0.74)	0.000(0.75)	0.000(0.48)	0.000(0.99)	0.000(0.93)	0.000(1.14)	0.000(0.73)
Return on Assets	0.002b(2.09)	0.001c(1.88)	0.002b(2.17)	0.001c(1.84)	0.002a(2.14)	0.002(0.45)	0.002(0.67)	0.002(0.73)	0.002(0.63)	0.001-0.29)
Fixed Assets % Common Equity	0.000(0.99)	0.000(0.71)	0.000(0.81)	0.000(0.51)	0.000(0.69)	0.000(-0.31)	0.000(-1.00)	0.000(-1.06)	0.000(-1.18)	0.000(-0.49)
Tobin’s Q	−0.002(-0.56)	−0.002(-0.41)	−0.002(-0.55)	−0.002(-0.49)	−0.002(-0.56)	0.000(0.02)	−0.006(-0.33)	−0.007(-0.39)	−0.006(-0.34)	0.002(0.13)
Price Volatility	−0.020(-0.37)	−0.023(-0.44)	−0.019(-0.35)	−0.023(-0.45)	−0.019(-0.35)	0.101(1.07)	0.058(0.63)	0.064(0.68)	0.081(0.86)	0.127(1.37)
Loan Amount (Ln)	0.000(0.07)	−0.002(-0.37)	0.000(-0.05)	−0.002(-0.35)	0.000(0.07)	0.001(0.10)	0.000(0.04)	0.002(0.18)	0.007(0.54)	0.007(0.57)
High Yield/Leveraged Loan Dummy	0.018a(2.76)	0.019a(3.12)	0.019a(2.83)	0.018a(3.15)	0.019a(2.93)	−0.009(-0.30)	−0.021(-0.77)	−0.026(-0.92)	−0.022(-0.80)	0.000(0.01)
Board Size (Ln)	−0.025c(-1.85)	−0.022c(-1.69)	−0.023c(-1.72)	−0.225b(-2.36)	−0.236b(-2.22)	−0.019(-0.47)	−0.014(-0.39)	−0.004(-0.12)	0.481(1.62)	0.748b(2.05)
Board Size (Ln) × Board Size (Ln)				0.044b(2.21)	0.046b(2.09)				−0.105c(-1.72)	−0.163b(-2.23)
Non-Executive Board Members (%)	0.000(-0.49)	0.000(-0.02)			0.000(-0.50)	0.000(0.14)	0.001(1.28)			0.000(0.04)
Blau Index	−0.101a(-2.92)	−0.071b(-2.49)	−0.100a(-2.82)	−0.064b(-2.25)	−0.096a(-2.76)	0.160(1.40)	0.233c(1.99)	0.229b(2.03)	0.193(1.60)	0.143(1.23)
CEO Board Member	−0.034a(-3.11)		−0.020c(-2.50)		−0.034a(-3.24)	−0.068a(-2.81)				−0.076a(-3.09)
CEO Compensation Link to Shareholder Return	0.018b(2.50)		0.020a(2.74)		0.018b(2.50)	−0.005(-0.22)				−0.003(-0.16)
CEO duality	0.027a(3.73)			0.022a(3.29)	0.027a(3.70)	0.002(0.09)		−0.040(-1.60)	−0.043c(-1.69)	0.001(0.03)
Industry Dummies	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Country Dummies	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
N	439	467	439	467	439	147	157	157	157	147
R2	0.1998	0.1489	0.1780	0.1757	0.2101	0.4975	0.4541	0.4609	0.4784	0.5369
Mean-VIF	4.33	4.42	4.29	8.31	8.41	3.84	2.95	2.87	10.26	12.15

This table reports the results of the cross-sectional OLS regression analysis with robust standard errors for announcement period (3-days) excess returns of borrowers estimated using the market model before (Panel A) and after the COVID-19 pandemic (Panel B). Coefficients are reported and t-statistics are presented in parentheses. The Huber-White robust standard errors are used to calculate t-statistics in all models. All variables are winsorized at the 1% and 99% levels. For more details about the definition of each variable see Table 2. The subscripts a, b and c denote significance at 1%, 5% and 10% levels, respectively.

We also find significant differences between the two examined periods regarding the other determinants. Specifically, borrowers with long history experience positive value effects during the COVID-19 pandemic only, indicating the significance of low information asymmetry during the economic turmoil. The level of profitability (as measured by the return on assets) has a statistically significant effect on borrowers’ gains before the pandemic, which becomes insignificant during the pandemic. Finally, the results show that high yield/leveraged loans are associated with higher abnormal returns before the pandemic, while during the pandemic loan leveraging is not significant.

Difference-in-Differences results

Table 7 reports the results of the DID regression models. The estimated coefficients of the COVID-19 crisis are mainly positive and statistically significant (at the 1% and 5% levels), suggesting that market participants view syndicated loan announcements as good news during the pandemic. This finding is also consistent with the results of the univariate analysis and in line with hypothesis H. Furthermore, the estimated coefficients of the DID regressions show that, during the COVID-19 pandemic, the combination of the CEO and chairman roles has a negative and statistically significant effect on borrowers’ abnormal returns. Similarly, the effect of large boards on shareholder value is negative and statistically significant since the outbreak of the pandemic. Regarding the other corporate governance mechanisms, namely the percentage of non-executive board members, the link between CEO compensation and shareholder return, the Blau Index and the CEO being a board member, the DID estimators are statistically indistinguishable from zero, suggesting no association between these variables and the excess returns of borrowers during the pandemic. The findings of the DID analysis further support the results of the regression analysis.

Table 7

Differences in Differences analysis: The effect of corporate governance on borrower excess returns with respect to the COVID-19 pandemic.

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
Constant	−0.076(-1.49)	−0.069(-1.34)	−0.076(-1.47)	−0.075(-1.46)	−0.080(-1.58)	−0.075(-1.46)	−0.078(-1.52)	−0.070(-1.36)
Age	0.010b(2.22)	0.010b(2.14)	0.010b(2.22)	0.010b(2.19)	0.010b(2.24)	0.010b(2.19)	0.010b(2.25)	0.009b(2.09)
Total Assets (Ln)	−0.001(-0.33)	−0.002(-0.47)	−0.001(-0.37)	−0.001(-0.33)	−0.001(-0.33)	−0.001(-0.33)	−0.002(-0.43)	−0.002(-0.51)
Total Debt % Common Equity	0.000(-0.66)	0.000(-0.68)	0.000(-0.67)	0.000(-0.67)	0.000(-0.66)	0.000(-0.66)	0.000(-0.59)	0.000(-0.60)
Return on Assets	0.001(1.58)	0.001(1.57)	0.001(1.57)	0.001(1.59)	0.001(1.59)	0.001(1.59)	0.001(1.59)	0.001(1.59)
Fixed Assets % Common Equity	0.000(0.47)	0.000(0.57)	0.000(0.50)	0.000(0.48)	0.000(0.45)	0.000(0.47)	0.000(0.34)	0.000(0.34)
Tobin’s Q	−0.004(-0.93)	−0.004(-0.84)	−0.004(-0.98)	−0.004(-0.90)	−0.004(-0.91)	−0.004(-0.94)	−0.004(-0.96)	−0.004(-0.84)
Price Volatility	0.021(0.46)	0.022(0.50)	0.019(0.43)	0.022(0.48)	0.022(0.48)	0.021(0.47)	0.020(0.44)	0.017(0.37)
Loan Amount (Ln)	−0.001(-0.22)	0.000(-0.04)	−0.001(-0.15)	−0.001(-0.23)	−0.001(-0.23)	−0.001(-0.23)	−0.001(-0.15)	0.000(-0.01)
High Yield/Leveraged LoanDummy	0.011c(1.68)	0.012c(1.82)	0.012c(1.70)	0.011c(1.68)	0.012c(1.72)	0.011c(1.67)	0.011(1.61)	0.012c(1.78)
Large Board Size	−0.006(-1.02)	0.006(0.96)	−0.006(-1.02)	−0.006(-1.02)	−0.006(-1.09)	−0.006(-1.00)	−0.007(-1.22)	0.003(0.50)
COVID-19 CRISIS * Large Board Size		−0.042a(-3.27)						−0.035a(-2.76)
High ratio of Non-Executive Board Members	−0.008(-1.00)	−0.009(-1.16)	−0.003(-0.34)	−0.008(-1.06)	−0.007(-0.89)	−0.008(-1.01)	−0.005(-0.67)	−0.003(-0.39)
COVID-19 CRISIS * High ratio of Non-Executive Board Members			−0.019(-1.51)					−0.014(-0.86)
High Blau Index	0.002(0.35)	0.000(0.03)	0.002(0.34)	0.001(0.12)	0.003(0.37)	0.002(0.34)	0.002(0.26)	−0.006(-0.86)
COVID-19 CRISIS * High Blau Index				0.048(0.63)				0.018(1.29)
CEO Board Member	−0.032a(-3.40)	−0.035a(-3.73)	−0.032a(-3.40)	−0.032a(-3.40)	−0.029a(-2.90)	−0.032a(-3.39)	−0.029a(-3.25)	−0.031a(-2.97)
COVID-19 CRISIS * CEO Board Member					−0.009(-0.64)			−0.009(-0.51)
CEO Compensation Link to Shareholder Return	0.010c(1.69)	0.011c(1.86)	0.011c(1.80)	0.010(1.61)	0.010c(1.67)	0.010(1.48)	0.011c(1.76)	0.012c(1.87)
COVID-19 CRISIS * CEO Compensation Link to Shareholder Return						0.003(0.21)		0.001(0.08)
CEO duality	0.022a(3.26)	0.021a(3.16)	0.022a(3.22)	0.022a(3.27)	0.022a(3.28)	0.022a(3.26)	0.030a(4.22)	0.027a(3.91)
COVID-19 CRISIS * CEO duality							−0.041a(-3.06)	−0.035b(-2.42)
COVID-19 CRISIS	0.017b(2.38)	0.027a(3.20)	0.021b(2.54)	−0.003(-0.08)	0.023b(2.10)	0.015(1.55)	0.024a(3.00)	0.034b(2.32)
Industry Dummies	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Country Dummies	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
N	586	586	586	586	586	586	586	586
R2	0.1618	0.1757	0.1645	0.1626	0.1626	0.1618	0.1725	0.1871
VIF	4.99	4.97	4.97	5.53	5.01	5.01	4.96	5.02

This table reports the results of the Differences-in-Differences regression analysis with robust standard errors for announcement period (3-days) excess returns of borrowers estimated using the market model. Coefficients are reported and t-statistics are presented in parentheses. The Huber-White robust standard errors are used to calculate t-statistics in all models. All variables are winsorized at the 1% and 99% levels. For more details about the definition of each variable see Table 1. The subscripts a, b and c denote significance at 1%, 5% and 10% levels, respectively.

Robustness analysis

Different estimation models

To address any potential bias caused by the use of the market model, we estimate the CARs of borrowers using four alternative models, namely: (i) the four-factor model; (ii) the five-factor model5 ; (iii) the CAPM; and (iv) the two factor CAPM with country and global benchmark. The latter model is a two-factor hybrid CAPM that assumes markets may be rather segmented, and not perfectly integrated and, thus, it simultaneously allows time-varying exposure to global and domestic factors (Bekaert et al., 2009; Warnes and Warnes, 2014). Following Bodnar et al. (2003), we apply a partial segmentation model using a global rate of return and a country rate of return. We use the Global MSCI index as proxy for the world market portfolio and the stock market benchmark index in each country to capture the domestic risk factors.

Table 8 reports the excess returns of European borrowers before the COVID-19 pandemic (Panel A) and after the pandemic (Panel B) estimated with the four applied models. All models calculate negative and mainly not statistically significant abnormal returns for borrowers before the pandemic, while they calculate mainly positive and statistically significant returns for the period during the pandemic. Furthermore, the mean differences between the two subgroups are statistically significant, indicating higher value effects for borrowers during the COVID-19 pandemic. The results estimated with the alternative applied models are similar with those estimated with the market model, which further confirms our main findings. In untabulated results, we also estimate the CARs of borrowers using the market model with the STOXX Europe 600 and the EURO STOXX 50 indices. Once again, the results remain unaltered.

Table 8

Cumulative abnormal returns upon syndicated loan announcements estimated with the four-factor and the five-factor models.

	Panel A: Prior-pandemic announcements(N = 480)						Panel B: During-pandemic announcements(N = 157)						Panel C: Test for differences(2-1)
EventWindow	Mean	Median	Std. Dev	%Pos	Patell-Z	Corrado	Mean	Median	Std. Dev	%Pos	Patell-Z	Corrado	Mean	Median	t-test	MWU
I. Four-factor model
[-10,10]	−0.87	−0.15	12.13	49	−5.374a	−2.214b	−1.26	−1.24	16.53	43	−5.976a	−1.557	−0.39	−1.10	−0.269	−0.620
[-3,3]	−0.24	0.12	8.49	51	−1.251	−0.836	0.92	0.20	10.70	52	2.057b	0.136	1.16	0.09	1.242	−0.470
[-2,2]	−0.21	0.21	8.30	53	0.333	−0.166	1.10	0.09	10.20	50	4.736a	−0.078	1.30	−0.12	1.452	−0.008
[-1,0]	−0.17	0.14	5.72	52	−0.642	0.368	1.59	0.26	9.28	54	6.710a	1.732c	1.76b	0.12	2.242	−1.089
[-1,1]	−0.16	0.05	7.19	51	−0.325	0.502	1.34	0.55	9.18	55	5.417a	0.783	1.50c	0.50	1.867	−0.925
[-1,10]	−0.69	0.04	10.21	50	−3.311a	−1.718c	0.41	−0.11	12.74	50	2.466b	−1.141	1.10	−0.15	0.979	−0.014
[0,0]	−0.32	0.02	4.87	51	−3.084a	−0.007	1.04	0.36	8.06	56	4.157a	1.769c	1.36b	0.34	1.997	−1.555
[0,1]	−0.31	0.02	6.43	51	−2.696a	0.244	0.78	0.16	8.72	52	2.574b	0.479	1.10	0.14	1.452	−0.438
II. Five-factor model
[-10,10]	−0.87	−0.11	11.94	50	−4.863a	−1.930c	−0.48	0.16	16.76	52	−3.761a	−1.020	0.38	0.28	0.266	−0.178
[-3,3]	−0.24	0.06	8.61	51	−1.272	−0.482	1.37	0.68	11.18	57	3.741a	0.557	1.61c	0.62	1.656	−1.047
[-2,2]	−0.22	0.05	8.40	51	0.497	0.142	1.44	0.46	10.16	52	6.275a	0.149	1.66c	0.41	1.848	−0.640
[-1,0]	−0.17	0.12	5.77	53	−0.544	0.655	1.83	0.47	9.26	57	8.159a	2.104	2.00b	0.35c	2.551	−1.782
[-1,1]	−0.16	0.18	7.25	52	−0.214	0.829	1.59	0.56	9.23	55	6.978a	0.860	1.74b	0.38	2.160	−1.212
[-1,10]	−0.67	−0.01	9.93	50	−3.042a	−1.644	0.83	−0.11	12.91	50	3.666a	−1.028	1.50	−0.10	1.329	−0.547
[0,0]	−0.31	0.05	4.88	51	−2.784a	0.296	1.06	0.19	8.02	54	4.294a	1.596	1.37b	0.14	2.020	−1.210
[0,1]	−0.29	0.03	6.45	51	−2.379b	0.571	0.82	0.09	8.72	51	2.847a	0.078	1.11	0.06	1.470	−0.342
III. CAPM
[-10,10]	−0.51	0.20	12.27	51	−4.179a	−1.250	1.13	1.05	17.22	53	−2.981a	−0.757	1.64	0.85	1.104	−1.444
[-3,3]	−0.19	0.08	8.67	50	−0.876	−0.256	1.43	0.31	11.29	51	3.605a	0.868	1.62	0.23	1.650	−1.113
[-2,2]	−0.14	0.23	8.42	53	0.892	0.551	1.34	0.71	9.74	54	5.622a	0.542	1.48c	0.48	1.709	−0.940
[-1,0]	−0.20	0.14	5.71	52	−1.210	0.217	1.69	0.62	9.08	59	7.068a	2.643	1.89b	0.49b	2.453	−1.995
[-1,1]	−0.21	0.13	7.28	52	−1.084	0.119	1.48	0.78	9.04	59	6.183a	1.724	1.69b	0.65	2.133	−1.573
[-1,10]	−0.48	0.06	10.22	51	−2.703a	−1.065	2.04	0.52	13.40	52	4.361a	−0.080	2.52b	0.47	2.159	−1.215
[0,0]	−0.31	0.06	4.86	52	−2.927a	0.669	1.10	0.76	7.73	59	4.570a	2.703	1.41b	0.70b	2.156	−2.290
[0,1]	−0.32	0.11	6.49	53	−2.772a	0.401	0.90	0.42	8.54	54	3.486a	1.380	1.22	0.31	1.639	−0.992
IV. Two-factor CAPM
[-10,10]	−0.48	−0.05	12.00	50	−4.062a	−1.133	0.80	0.65	17.17	54	−2.407b	−0.888	1.28	0.70	0.865	−1.201
[-3,3]	−0.19	−0.08	8.65	50	−1.012	−0.277	1.27	0.31	11.21	52	3.517a	0.704	1.46	0.39	1.493	−0.685
[-2,2]	−0.12	0.16	8.32	53	0.884	0.584	1.26	0.67	9.75	53	5.245a	0.443	1.38	0.51	1.596	−0.598
[-1,0]	−0.20	0.13	5.72	53	−1.185	0.205	1.66	0.84	9.12	59	6.853a	2.749a	1.86b	0.70b	2.412	−2.166
[-1,1]	−0.20	0.19	7.28	53	−1.008	0.213	1.45	0.80	9.18	57	5.879a	1.670c	1.65b	0.62	2.052	−1.406
[-1,10]	−0.45	0.07	10.07	51	−2.562b	−0.970	1.72	0.45	13.44	53	3.313a	−0.598	2.17c	0.39	1.860	−0.926
[0,0]	−0.32	0.05	4.85	52	−3.047a	0.558	1.08	0.53	7.80	59	4.175a	2.599a	1.39b	0.47b	2.109	−2.187
[0,1]	−0.32	0.14	6.47	53	−2.831a	0.450	0.86	0.38	8.59	52	2.982a	1.133	1.18	0.24	1.581	−0.626

This table reports the cumulative abnormal returns (CARs) upon syndicated loan announcements between 01/01/2018 and 31/07/2020. CARs are derived from the four-factor model, the five-factor model, the CAPM and the two factor CAPM with country and global benchmark. Panels A and B present the mean and median CARs, standard deviation, percentage of borrowers with positive CARs for loan announcement prior to the pandemic (N = 480) and for loan announcements during the pandemic (N = 157), respectively. The statistical significance of CARs is accessed using the Patell-Z test and the Corrado test. Panel C reports the mean and median differences of CARs between announcements before and during the pandemic. The statistical significance of the differences between the means and the medians of the two subgroups are tested using the t-test of equality of means and the Mann-Whitney U test, respectively. The superscripts a, b and c denote significance at 1%, 5% and 10% levels, respectively.

Cross-sectional analysis with different estimation models and alternative event windows

The results of the multivariate analysis for the entire period were based on the 3-day CARs (−1,1) estimated with the market model. To test the robustness of the results, we re-run the regression equations using different estimation models as well as alternative event windows. We also include in the regression models national characteristics, such as macroeconomic conditions and institutional factors, in order to examine whether they affect the wealth gains of shareholders of borrowers. More specifically, we control for the effect of GDP growth, GDP per capita, current account balance and inflation, as indicators of the macroeconomic environment. To capture institutional factors, we calculate the arithmetic mean of all dimensions of governance included in the Worldwide Governance Indicator (WGI), which are: (i) voice and accountability; (ii) political stability and absence of violence; (iii) government effectiveness; (iv) regulatory quality; (v) rule of law; and (vi) control of corruption.

Panels A, B, C, D and E from Table 9 report the results using the market model, the four-factor model, the five-factor model, the CAPM and the two factor CAPM, respectively, in three event windows lasting from two to five days. The results confirm the positive and statistically significant effect of the COVID-19 pandemic on borrowers’ gains, irrespective of the period and the estimation model used. With regard tocorporate governance variables, the estimated coefficients of board size and CEO board member are, in many cases, significantly negative, indicating value destruction for shareholders. The decision of borrowers to combine the CEO and chairman roles is significantly positively related to the announcement excess returns in the entire period; however, the relationship between CEO duality and excess returns becomes negative during the COVID-19 pandemic. The estimated coefficients of all macroeconomic variables as well as the WGI are not statistically significant, suggesting no relationship with the abnormal returns of borrowers. From the remaining control variables, the coefficients of age remain positive and statistically significant, indicating the importance of information asymmetry mitigation for borrowers. Overall, our main conclusions were unaffected by the alternative econometric specifications.

Table 9

Cross-sectional analysis with different estimation models and alternative event windows (entire period).

	Panel A: Market Model			Panel B: Four-factor Model			Panel C: Five-factor Model			Panel D: CAPM			Panel E: Two-factor CAPM
	CAR_i(-2,2)	CAR_i(-1,0)	CAR_i(0,1)	CAR_i(-2,2)	CAR_i(-1,0)	CAR_i(0,1)	CAR_i(-2,2)	CAR_i(-1,0)	CAR_i(0,1)	CAR_i(-2,2)	CAR_i(-1,0)	CAR_i(0,1)	CAR_i(-2,2)	CAR_i(-1,0)	CAR_i(0,1)
Constant	−0.007(-0.11)	−0.036(-0.91)	0.009(0.23)	−0.020(-0.23)	−0.062(-1.16)	0.017(0.23)	−0.010(-0.17)	−0.047(-1.24)	0.010(0.24)	−0.007(-0.12)	−0.037(-0.91)	0.009(0.23)	−0.001(-0.02)	−0.032(-0.79)	0.015(0.38)
Age	0.008c(1.68)	0.007b(1.98)	0.006c(1.83)	0.012b(1.97)	0.009c(1.79)	0.009c(1.87)	0.008(1.59)	0.007b(2.10)	0.006c(1.81)	0.008c(1.68)	0.007b(1.98)	0.006c(1.83)	0.008(1.64)	0.006c(1.88)	0.006c(1.85)
Total Assets (Ln)	0.000(-0.02)	0.000(0.06)	0.002(0.91)	−0.002(-0.38)	0.001(0.29)	0.001(0.40)	−0.001(-0.36)	−0.001(-0.34)	0.001(0.56)	0.000(-0.02)	0.000(0.06)	0.002(0.91)	0.000(-0.01)	0.000(0.01)	0.002(0.79)
Total Debt % Common Equity	0.000(0.06)	0.000(-0.78)	0.000(-0.94)	0.000(0.17)	0.000(0.28)	0.000(-0.24)	0.000(-0.13)	0.000(-0.72)	0.000(-0.82)	0.000(0.06)	0.000(-0.78)	0.000(-0.94)	0.000(-0.11)	0.000(-0.81)	0.000(-0.96)
Return on Assets	0.002c(1.87)	0.001(1.11)	0.001(1.26)	0.001(1.21)	0.001(0.80)	0.000(0.33)	0.002c(1.72)	0.000(0.78)	0.001(1.00)	0.002c(1.87)	0.001(1.11)	0.001(1.26)	0.002c(1.89)	0.001(1.10)	0.001(1.20)
Fixed Assets % Common Equity	0.000(-0.75)	0.000(0.96)	0.000(0.10)	0.000(-0.35)	0.000(0.10)	0.000(0.22)	0.000(-0.42)	0.000(0.65)	0.000(-0.09)	0.000(-0.75)	0.000(0.95)	0.000(0.10)	0.000(-0.50)	0.000(1.05)	0.000(0.10)
Tobin’s Q	−0.001(-0.23)	−0.001(-0.20)	−0.003(-0.97)	−0.001(-0.20)	−0.001(-0.17)	−0.003(-0.61)	−0.003(-0.55)	−0.001(-0.32)	−0.004(-1.22)	−0.001(-0.23)	−0.001(-0.20)	−0.003(-0.97)	−0.001(-0.24)	−0.001(-0.18)	−0.003(-0.92)
Price Volatility	−0.001(-0.01)	0.028(0.86)	−0.011(-0.34)	0.022(0.28)	0.048(1.02)	−0.026(-0.41)	0.016(0.35)	0.028(0.92)	−0.008(-0.25)	0.000(-0.01)	0.028(0.86)	−0.011(-0.34)	0.003(0.06)	0.030(0.92)	−0.010(-0.29)
Loan Amount (Ln)	0.001(0.31)	−0.001(-0.48)	−0.002(-0.80)	0.005(0.97)	−0.001(-0.12)	0.000(-0.11)	0.003(0.65)	−0.001(-0.26)	−0.002(-0.63)	0.001(0.30)	−0.001(-0.48)	−0.002(-0.80)	0.002(0.50)	−0.001(-0.32)	−0.002(-0.60)
High Yield/Leveraged LoanDummy	0.014c(1.95)	0.008(1.39)	0.008(1.64)	0.011(1.14)	0.004(0.58)	0.004(0.61)	0.009(1.20)	0.006(1.06)	0.006(1.11)	0.014c(1.95)	0.008(1.39)	0.008(1.64)	0.014b(2.02)	0.007(1.21)	0.008(1.57)
Board Size (Ln)	−0.027b(-2.23)	−0.016c(-1.82)	−0.020b(-2.34)	−0.034b(-2.15)	−0.021c(-1.66)	−0.024c(-1.94)	−0.026b(-2.14)	−0.015c(-1.69)	−0.018b(-2.06)	−0.027b(-2.23)	−0.016c(-1.82)	−0.020b(-2.34)	−0.030b(-2.47)	−0.016c(-1.87)	−0.021b(-2.40)
Non-Executive Board Members (%)	0.000(-0.22)	0.000(0.56)	0.000(-0.14)	0.000(0.09)	0.000(0.85)	0.000(-0.39)	0.000(0.10)	0.000(0.88)	0.000(-0.08)	0.000(-0.23)	0.000(0.56)	0.000(-0.14)	0.000(-0.29)	0.000(0.45)	0.000(-0.42)
Blau Index	−0.035(-1.35)	−0.002(-0.12)	−0.028(-1.50)	−0.042(-1.26)	−0.002(-0.07)	−0.037(-1.39)	−0.022(-0.84)	0.008(0.38)	−0.023(-1.15)	−0.035(-1.35)	−0.002(-0.11)	−0.028(-1.50)	−0.032(-1.24)	−0.001(-0.07)	−0.027(-1.43)
CEO Board Member	−0.014c(-1.66)	−0.008(-1.36)	−0.014b(-2.28)	−0.020c(-1.66)	−0.007(-0.84)	−0.025a(-2.62)	−0.011(-1.23)	−0.002(-0.38)	−0.014b(-2.20)	−0.014c(-1.66)	−0.008(-1.36)	−0.014b(-2.28)	−0.015c(-1.70)	−0.008(-1.31)	−0.014b(-2.34)
CEO Compensation Link to Shareholder Return	0.006(0.93)	0.008c(1.73)	0.010b(2.06)	−0.002(-0.29)	0.002(0.30)	0.002(0.35)	0.004(0.52)	0.006(1.36)	0.008(1.56)	0.006(0.93)	0.008c(1.73)	0.010b(2.06)	0.007(1.04)	0.008c(1.79)	0.010b(2.03)
CEO duality	0.024a(3.26)	0.020a(4.13)	0.017a(3.47)	0.023b(2.34)	0.021a(3.05)	0.019a(2.61)	0.021a(2.81)	0.019a(3.83)	0.016a(3.18)	0.024a(3.26)	0.020a(4.13)	0.017a(3.47)	0.023a(3.26)	0.020a(4.08)	0.017a(3.53)
COVID-19 × CEO duality	−0.059a(-3.66)	−0.033b(-2.38)	−0.037a(-2.84)	−0.045b(-2.21)	−0.026(-1.61)	−0.030b(-1.99)	−0.038c(-1.78)	−0.024(-1.61)	−0.027c(-1.94)	−0.059a(-3.67)	−0.033b(-2.38)	−0.037a(-2.85)	−0.058a(-3.61)	−0.032b(-2.31)	−0.035a(-2.64)
COVID-19 CRISIS	0.028a(2.85)	0.023a(3.11)	0.015b(2.08)	0.037a(2.69)	0.029a(2.89)	0.025b(2.22)	0.029a(2.89)	0.024a(3.17)	0.013c(1.74)	0.028a(2.89)	0.023a(3.13)	0.015b(2.10)	0.027a(2.79)	0.023a(3.08)	0.014b(2.02)
GDP growth	0.002(0.98)	0.001(0.37)	0.000(-0.10)	0.004c(1.68)	0.002(0.80)	0.001(0.45)	0.003(1.46)	0.001(0.77)	0.000(-0.09)	0.002(0.98)	0.001(0.37)	0.000(-0.10)	0.002(1.17)	0.001(0.38)	0.000(-0.04)
GDP per capita	0.000(-0.63)	0.000(-1.22)	0.000(-0.87)	0.000(0.08)	0.000(-0.35)	0.000(0.41)	0.000(-1.03)	0.000(-1.50)	0.000(-0.71)	0.000(-0.63)	0.000(-1.22)	0.000(-0.87)	0.000(-0.91)	0.000(-1.40)	0.000(-1.02)
Current account balance	0.000(0.61)	0.000(-0.12)	0.000(0.65)	0.001(0.72)	0.000(0.11)	0.000(0.45)	0.001(0.97)	0.000(0.44)	0.000(0.65)	0.000(0.62)	0.000(-0.12)	0.000(0.65)	0.001(0.88)	0.000(0.18)	0.001(0.91)
Inflation	−0.001(-0.39)	0.000(0.21)	−0.001(-0.22)	0.003(0.71)	0.002(0.82)	0.001(0.40)	0.001(0.27)	0.001(0.59)	0.000(-0.01)	−0.001(-0.39)	0.000(0.21)	−0.001(-0.22)	−0.002(-0.52)	0.001(0.25)	−0.001(-0.23)
Governance (WGI)	−0.020(-1.60)	−0.007(-0.82)	−0.018c(-1.85)	−0.024(-1.43)	−0.010(-0.88)	−0.024c(-1.70)	−0.010(-0.72)	0.000(0.00)	−0.013(-1.42)	−0.020(-1.60)	−0.007(-0.82)	−0.018c(-1.85)	−0.020(-1.60)	−0.006(-0.71)	−0.017c(-1.77)
Industry Dummies	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
N	586	586	586	586	586	586	586	586	586	586	586	586	586	586	586
R2	0.1414	0.1747	0.1562	0.1130	0.1307	0.1214	0.1378	0.1764	0.1423	0.1418	0.1751	0.1510	0.1423	0.1739	0.1516
VIF	5.25	5.25	5.25	5.25	5.25	5.25	5.25	5.25	5.25	5.25	5.25	5.25	5.25	5.25	5.25

This table reports the results of the cross sectional OLS regression analysis with robust standard errors for various announcement periods (5-days, 2-days and 2-days) excess returns of European borrowers estimated using the market model, the four-factor model, the five-factor model, the CAPM and the two-factor CAPM. Coefficients are reported and t-statistics are presented in parentheses. The Huber-White robust standard errors are used to calculate t-statistics in all models. All variables are winsorized at the 1% and 99% levels. The subscripts a, b and c denote significance at 1%, 5% and 10% levels, respectively.

Propensity score matching

It is possible that the borrowers’ financial characteristics have been affected since the outburst of the COVID-19 pandemic. Thus, an endogenous relationship may exist between the pandemic and the characteristics of borrowers, which may result in sample self-selection bias. Following relevant literature (Leledakis and Pyrgiotakis, 2020; Chintrakarn et al., 2021), we address this issue using propensity score matching (PSM). The PSM method allows us to match syndicated loans announced by firms during the COVID-19 pandemic (treated group) with loans announced by firms before the pandemic (controlled group) with otherwise similar characteristics, and then compare the CARs between the two groups. Panel A of Table 10 reports the propensity scores estimated running the probit regression model. Based on the estimated propensity scores, we used the nearest neighbor matching approach (one-to-one) through which syndicated loans announced by firms during the COVID-19 pandemic are matched with loans announced by firms before the pandemic. Panels B, C, and D of Table 10 report the estimated average treatment effect on the treated (ATT) using the market model with each country’s benchmark index, the STOXX Europe 600 index and the EURO STOXX 50 index, respectively. Panels E and F present the ATT using the four-factor model and the five-factor model, respectively. In all models, the ATT is statistically significant, confirming the results of the univariate and multivariate analyses regarding the higher value effects of syndicated loan announcements during the pandemic.

Table 10

Propensity score matching analysis.

Panel A: Probit estimation results (Loan announcements during COVID-19 pandemic = 1)
Constant	−5.117a(-4.67)
Age	−0.013(-0.16)
Total Assets (Ln)	0.246a(4.65)
Total Debt % Common Equity	0.000(-0.27)
Return on Assets	−0.016(-1.16)
Fixed Assets % Common Equity	0.001(0.80)
Tobin’s Q	0.146(1.62)
Price Volatility	1.166b(2.10)
Loan Amount (Ln)	−0.376a(-5.63)
High Yield/Leveraged LoanDummy	−0.209(-1.33)
Board Size (Ln)	0.473b(2.02)
Non-Executive Board Members (%)	0.006(0.91)
Blau Index	2.123a(3.11)
CEO Board Member	0.103(0.61)
CEO Compensation Link to Shareholder Return	0.387a(2.92)
CEO duality	−0.469a(-2.79)
N	589
LR Chi2	80.52a
Pseudo R2	0.1216
Panel B: ATTs for three-day CARs calculated with market model (each country’s benchmark)
Loan announcements during COVID-19 pandemic	0.0143
Loan announcements before COVID-19 pandemic	−0.0050
Difference	0.0193b
Panel C: ATTs for three-day CARs calculated with market model (STOXX Europe 600)
Loan announcements during COVID-19 pandemic	0.0147
Loan announcements before COVID-19 pandemic	−0.0045
Difference	0.0192b
Panel D: ATTs for three-day CARs calculated with market model (EURO STOXX 50)
Loan announcements during COVID-19 pandemic	0.0132
Loan announcements before COVID-19 pandemic	−0.0049
Difference	0.0181b
Panel E: ATTs for three-day CARs calculated with four-factor model
Loan announcements during COVID-19 pandemic	0.0114
Loan announcements prior to COVID-19 pandemic	−0.0034
Difference	0.0147c
Panel F: ATTs for three-day CARs calculated with five-factor model
Loan announcements during COVID-19 pandemic	0.0153
Loan announcements prior to COVID-19 pandemic	−0.0033
Difference	0.0186b

This table reports the outcome of the propensity score matching (PSM) analysis for European borrowers’ syndicated loan announcements. Panel A presents the results of a probit model that is used to estimate the propensity scores. The dependent variable is a dummy variable that equals 1 if the loan is announced during the COVID-19 pandemic and 0 otherwise. Panels B, C and D present the average treatment effect on the treated (ATT) using CARs estimated with the market model using each country’s benchmark index, STOXX Europe 600 index and EURO STOXX 50 index, respectively. Panels E and F present the ATT using CARs estimated with the four-factor model and the five-factor model, respectively. The PSM results are estimated using the one-to-one matching approach. Standard errors for the ATTs are the heteroskedasticity-consistent standard errors estimated using the method provided by Abadie and Imbens (2006). All variables are winsorized at the 1% and 99% levels. The subscripts a, b and c denote significance at 1%, 5% and 10% levels, respectively.

Conclusions

This study seeks to advance the discussion on the value implication of syndicated loan announcements for the shareholders of borrowing firms. Syndicated loans constitute a major alternative for mid- and long-term financing for firms, for which empirical literature provides inconclusive evidence in terms of value creation. The onset of the COVID-19 pandemic caused severe financial problems and, thus, firms were forced to seek for bank loans to meet the challenges of the new era. Therefore, the pandemic-driven crisis offers a quasi-experimental setting providing the opportunity to comparatively assess the wealth effects of syndicated loan announcements for borrowers. We employ a sample of 637 syndicated loans made by European firms in the period from 01/01/2018 to 31/07/2020 and examine the market reaction upon their announcement before and during the COVID-19 pandemic. We also examine the effect of board characteristics, such as board size, gender diversity, percentage of non-executive directors, CEO duality and compensation, on borrowers’ gains. The board of directors is a crucial corporate governance mechanism, having a significant effect on performance outcomes.

The main finding of our study is that the announcement of syndicated loans during the COVID-19 pandemic is associated with significant value creation for the shareholders of borrowers. This stands in contrast to the period before the pandemic, where the results are at best not distinguishably different from zero. The positive market reaction during the pandemic supports the notion that a loan certification from multiple lenders conveys positive information concerning the creditworthiness of borrowers in times of increased economic and financial uncertainty. It may also signal the potential for the implementation of profitable investment projects during the crisis. Our findings are in line with prior literature (i.e. Li and Ongena, 2015; Gasbarro et al., 2017) suggesting positive excess returns for borrowers during the global financial crisis. Another important finding of our study lies in the investigation of the impact of certain corporate governance mechanisms on borrowers’ gains before and during the crisis. The results yield important conclusions regarding the effect of the examined governance mechanisms on shareholder value during the two periods. Specifically, we find that large boards have a positive effect on value creation before the pandemic; however, this effect turns to be negative during the pandemic. Gender diversity destroys value before the outbreak of COVID-19, but the negative impact disappears during the economic meltdown. The link between CEO compensation and shareholder return is significantly positive before the pandemic but becomes insignificant in the post-pandemic period. CEO duality is associated with value creation during the period before the pandemic, while after the pandemic the positive relationship between the two variables does not hold. Considering the legislative initiatives at national and European level towards the formation of an effective corporate governance framework for firms in Europe, the results of our study highlight the importance of the economic and business environment as far as the effectiveness of each mechanism is concerned. Therefore, the establishment of a unified, tight and strict corporate governance framework may hinder the ability of firms to adapt effectively to the external environment.

The findings of this study have important implications for investors, borrowers and banks. For investors, the market reaction upon the announcement of syndicated loans provides useful signals for the current financial condition and the future prospects of firms. This is particularly important considering the increased information asymmetries and uncertainties amidst a market turmoil triggered by the pandemic. From the borrowers’ viewpoint, the certification of a syndicated loan during a crisis period indicates financial strength and this can be used to lower the cost of other sources of financing may be used as well. It also provides the opportunity to firms to negotiate better terms with other business partners, such as suppliers, contractors and customers. Banks could also use the findings of this study as a basis in the screening process of borrowers when demand for loans soars due to liquidity shortages caused by a financial meltdown. In case a potential borrower has already originated a syndicate loan, the bank has a strong indication of a healthy financial position and, thus, mitigate the credit risk associated with the borrower.

In addition to the above implications, there are also limitations to this timely study. Similarly to others (Ali et al., 2020; Baker et al., 2020; Conlon et al., 2020; Conlon and McGee, 2020; He et al., 2020; Shehzad et al., 2020), we focus on the early stages of the COVID-19 pandemic to examine the value implications of syndicated loan announcements. However, the pandemic is still unfolding, affecting the economic and business environment. Another limitation of this study lies in the fact that it focuses exclusively on syndicated loans, without taking into account that firms may issue a variety of debt and equity securities in order to satisfy their financing needs. Constraints on bank lending since the outbreak of the COVID-19 pandemic may have forced firms to look for a mix of financing instruments in the context of debt and equity capital markets. Finally, our study utilizes a sample of European borrowers and so is geographically bound. That limits the generalizability of the findings. The pandemic caused different impacts on each economy and business sector and, thus, the wealth effects of syndicated loan announcements may vary among businesses and/or countries.

With the limitations of our study as starting point, future research could investigate the wealth effects of syndicated loan announcements in more detail. The examination of the value implications of syndicated loans could be done in a more extended period that would include a larger number of such announcements. This would lead to a better understanding of the effects of syndicated loans on shareholder value. Also, the sample could consist of borrowers not only from Europe but also from North and Latin America, Asia and/or Africa. This would allow the comparative assessment of the results among firms from different regions. Finally, future studies could investigate the effects of other corporate governance mechanisms on borrowers’ gains. An optimal internal governance system may depend on more than a few board attributes, such as age and ethnic diversity, as well as other governance mechanisms.

CRediT authorship contribution statement

Ioannis Tampakoudis: Conceptualization, Methodology, Validation, Formal analysis, Writing - review & editing. Athanasios Noulas: Project administration, Supervision, Writing - original draft. Nikolaos Kiosses: Investigation, Resources, Data curation.

Declaration of Competing Interest

The authors declare no potential conflicts of interests with respect to the authorship and/or publication of this article.