Enterprise Risk Management Adoption Practices by US and European Multinationals

Abstract

This study provides the first evidence of the propensity of globally large industrial US and European firms to adopt enterprise risk management (ERM) processes in response to the recent challenges of systematic global risks associated with pandemics (COVID-19), increased geopolitical risks (e.g., the Ukraine–Russia conflict), increased cybersecurity threats and the challenges posed by climate change and biodiversity loss. Consistent with the predictions of standard risk management theory, it is predicted that there is a positive inter-relationship between the propensity to adopt ERM and total firm risk, after controlling for various firm-related financial characteristics, complexity and sources of idiosyncratic risk. The empirical research is based on an industry-matched sample of the 100 largest US and European firms globally. The empirical results are generally consistent with these predictions, but for European firms, total firm risk is not associated with ERM adoption. Furthermore, there is no statistically significant relationship between sample firms’ risk-adjusted performance and their ERM adoption propensity, and there are also significant cultural–institutional variations that explain the differences between the ERM adoption practices between US and European sub-sample firms. The findings raise new questions about the validity of ERM in addressing globally important risk challenges faced by the largest multinational firms.

1. Introduction

Enterprise risk management (ERM) processes have been lauded as a means of helping globally large organizations address overall global challenges such as pandemics, geopolitical risks, cybersecurity threats and imperatives to deal with systemic risks associated with climate and biodiversity loss [1,2]. However, questions have been raised as to why ERM processes have not been more fully implemented, despite the promulgation of both mandatory and voluntary frameworks [3]. Moreover, to date, there has been limited comprehensive empirical evidence on this issue, with prior empirical research mainly confined to US financial sector firms.

This paper makes several contributions to the literature. First, it adds to the findings of existing literature on ERM adoption, which has been to date entirely restricted to studying only financial firms. This study also adds to the literature by going beyond the findings of individual jurisdictional-delimited studies by exploring the role and impact of broader cultural–institutional international considerations affecting multinational firms by examining ERM adoption for a matched sample of 100 globally large European and US multinational non-financial firms (limiting the analysis to the largest US and European multinational firms also avoids the size and institutional ownership issues that limit the findings of prior studies on ERM adoption (e.g., [4,5,6])). This study, therefore, is the first to specifically acknowledge how ERM adoption by multinational firms is influenced by the institutional–cultural factors associated with the different regulatory and accounting rule enforcement requirements faced by European and US firms. The study also contributes to the literature by controlling for other idiosyncratic risk sources of firm-wide risk, such as the number of foreign operations, pension funding risk, and efforts to mitigate market risk by hedging their exposure to interest rate, commodity and foreign exchange risks. Finally, the research updates the prior research findings to incorporate the effects of the implementation of recently upgraded COSO- and ISO-based ERM frameworks that more fully integrate holistic firm-wide risk management processes.

The empirical results are mainly supportive of the predictions that incentives facing globally large multinational firms to adopt ERM practices are inter-related with overall firm risk, after controlling for these factors. However, when results are further decomposed between US and European firms, other firm and cultural–institutional factors appear to be a greater influence in explaining European firms’ total risk other than ERM adoption propensity. There is also no evidence that ERM adoption propensity is associated with sample firms’ risk-adjusted performance. Furthermore, the baseline findings are mitigated when the sample is partitioned by industry and the degree of leverage and derivative usage by firms. These findings raise questions as to whether the traditional risk management theory explanation is appropriate for understanding why multinational firms face incentives to adopt ERM.

The remainder of this paper is organized as follows. Section 2 provides the institutional background and literature review. Section 3 develops the hypotheses. Section 4 outlines the research design. Section 5 discusses the data and sample. Section 6 reports the results of empirical tests. Section 7 provides a conclusion.

2. Institutional Background and Literature Review

This section provides the institutional background required to understand the relevant regulatory and industry developments concerning ERM reporting and disclosures that apply to both European- and/or US-based multinational firms. It then provides an overview of the relevant theoretical antecedents and prior related empirical research.

2.1. Institutional Background

This section comprises a brief overview of the major recent regulatory developments affecting the propensity of multinational firms to adopt ERM. The discussion focuses on the most recent regulatory developments in the US and the European Union, where most of the globally largest multinational firms are based.

The two most widely recognized ERM standards are the COSO (Committee of Sponsoring Organizations of the Treadway Commission) framework [7] and the ISO 31000 (International Organization for Standardization): “Risk Management Guidelines” [8].

There are some specific and subtle differences between these frameworks. COSO [7] is specifically focused on ERM as an ongoing process and is largely aimed at helping US organizations meet their requirements for reporting under the Sarbanes–Oxley Act (SOX) [9]. The latter is meant to be a more voluntary guidance and is a more generic and process-oriented standard focused on the importance of an overall risk management framework (rather than specifically named “ERM”). Unlike COSO [7], ISO [8] more specifically holds organizations, as well as management, to be publicly accountable for their risk management. This standard further proposes a more generic, process-oriented risk management framework to assist the organization in integrating its risk management system into its most significant operational activities and functions. The framework development encompasses integrating, designing, implementing, evaluating, and improving risk management across the organization.

In contrast to the US, there are specified regulatory and statutory obligations for European Union-based companies to publicly disclose information about their overall management strategy and risk management processes [10] For example, the European Union requires all EU-listed consolidated firms to provide a separate narrative management report as a mandatory, additional component of annual financial reporting under the Accounting Directive (EU Accounting Directive (2013/34/EU) [11]. This report should include a generic requirement for a narrative discussion of the “development and position, together with the principal risks and uncertainties “faced (Article 19(1)), as well as a “corporate governance code” (Article 20 [11]).

Moreover, unlike the US, the European Union has subsequently implemented several more bespoke requirements related to how European Union-based firms business strategy and risk management systems incorporate broader sustainability issues and address globally emerging risks related to climate change. Most recently, it issued a Corporate Sustainability Directive (Directive EU 2022/2464) (CSRD) [12], effective 5 January 2023, which requires that European companies produce “… financial and investment plans to ensure that its business model and strategy are compatible with the transition to a sustainable economy and with the limiting of global warming to 1.5 C … and the objective of achieving climate neutrality by 2050 as established in Regulation (EU) 2021/1119”) [12]. EU-listed companies subject to the CSRD must additionally report according to European Sustainability Reporting Standards (ESRS) developed by the European Financial Reporting Advisory Group [13] The International Sustainability Standards Board (ISSB), under the auspices of the International Financial Standards, subsequently issued a standard for publicly listed firms subject to IFRS to also produce climate-related disclosures, which are generally aligned with the TCFD (2017) recommendations [14] (although the UK officially left the European Union after BREXIT in 2020, it has broadly equivalent reporting requirements related to the risk reporting obligations under the Corporate Governance Code [15]).

Another major issue arises over the implementation of these standards and ERM adoption propensity by multinational corporations and their decision to list on various global stock exchanges. First, US-based corporations and those whose cross-list in US stock markets are required to reconcile their accounts with those prepared in accordance with US GAAP (i.e., by submitting either a Form 10K (US domestic firms) or Form 20F (foreign firms) to the Securities and Exchange Commission). Thus, these corporations are required to provide detailed disclosures, and these are enforceable by registration and via SOX (s. 404) internal control enforcement. In contrast, multinational firms that cross-list in non-US exchanges face considerably less stringent requirements. Multinationals based in the European Union (EU) are also required to adopt IFRS as required by the European Commission [16]; however, there is no uniform enforcement of these disclosures as they are subject to disclosure-based regulatory monitoring by national securities regulators.

Therefore, an interesting issue addressed by this study is to examine to what extent ERM adoption practices by multinationals are affected by institutional factors associated with cultural differences in the nature and degree of enforcement of GAAP and ERM frameworks between US- and European-based multinational firms.

2.2. Literature Review

The existing professional and practitioner studies suggest that the adoption of enterprise-wide risk management (ERM) is a major step toward enhancing the quality of firm risk management and corporate governance practices [15]. However, recent survey evidence [17] finds that only 56% of large US corporations fully adopted ERM processes. Furthermore, after reviewing the progress in ERM adoption over the past 20 years, Fraser et al. [3] raise some important questions about the effectiveness of ERM implementation. They identify several common problems that organizations face, including (1) an over-emphasis on reporting, insufficient injection into key decision-making processes, treatment of risks as discrete items and risk aversion, and the misuse of models and lack of role clarity. In an earlier paper, Fraser et al. [18] also identified how global pandemics such as the COVID-19 crisis could enable firms to more effectively consider the longer-term risk implications to their environments. However, Fraser et al. [3] argued that their advice was subsequently “largely ignored”.

Eckles et al. [6] contrast the number of studies examining the determinants of corporate risk management policy with the much fewer studies analyzing the valuation impact of overall firm risk management policies such as ERM. In response to the enactment and implementation of SOX [9], as well as the subsequent global financial crisis in 2007–2008, managing risk from a holistic perspective is becoming an increasingly major consideration for multinational corporations.

Prior US-based studies, mainly focusing on the financial sector (e.g., [5,6,19,20,21,22]) and the information technology services industry [17], find that US insurance firms adopting ERM are likely to lower the marginal cost of adopting risk, which creates incentives for profit-maximizing firms to reduce total risk while increasing firm value. By combining the firm’s risks into a risk portfolio, an ERM-adopting firm is better able to recognize the benefits of natural hedging, prioritize hedging activities towards the risk that most contributes to the total risk of the firm, and optimize the evaluation and selection of available hedging instruments. Thus, by so doing, ERM-adopting firms will realize a greater potential reduction in risk per dollar spent. This reduction in the marginal cost of managing risk is argued to incentivize firms to profit maximize and further reduce risk until the marginal cost of risk management equals the marginal benefits [4]. Grace et al. [23] extend this strand of the literature by examining which specific aspects of ERM create value. They find that US insurers benefit from the use of simple economic capital models (ECMs).

However, the findings of this prior research are contradictory as to the merits of firms adopting ERM in various contexts. For example, whereas Hoyt and Liebenberg [5] find a large valuation premium (as measured by Tobin’s Q) for ERM adopters, Beasley et al. [4] find insignificant negative announcement returns for ERM adoption. Eckles et al. [6] find that, after adopting ERM, firm risk decreases and accounting performance increases for a given unit of risk. Therefore, their results complement the findings of Hoyt and Liebenberg [19], which are based on market valuation of firm performance (This study uses the standard approach used in prior studies in identifying ERM adoption through evidential analysis of keywords in annual reports).

However, other research questions the value-added benefits of ERM adoption. For instance, McShane et al. [20] do not find any positive relationship between insurance firms achieving a higher S&P ERM rating and firm value. More generally, a few papers have criticized the role and relevance of ERM for engaging in risk management issues by various organizational forms [24] or for business continuity management purposes [25]. However, the impact of specific forms of risk, e.g., market and/or idiosyncratic risk, on the financing, accountability and effective management control of organizations affected has not attracted any attention from researchers studying ERM adoption. Further, there is little empirical evidence available on this issue outside the US financial services industry, where ERM adoption is explicitly incorporated in their credit ratings.

For example, in the European context, there is conflicting evidence of the merits of ERM adoption. On the one hand, Florio and Leoni [26] find that Italian firms with advanced levels of ERM implementation produce higher performance. By contrast, Paapke et al. [27] find no evidence that COSO frameworks on risk appetite and tolerance influence ERM adoption by Dutch multinational firms. However, none of the studies have explicitly compared incentives facing US versus European multinationals to adopt ERM or whether these incentives are inter-related with managerial incentives, accounting quality and/or derivative hedge exposure. These findings support the prior empirical [28] and theoretical [29] arguments that the relation between ERM and firm performance is contingent upon the appropriate match between ERM and the following five factors affecting a firm: environmental uncertainty, industry competition, firm size, firm complexity, and board of directors’ monitoring.

Due to the relatively more stringent requirements for risk management strategy disclosure in the European Union than in the US, it may be that societal-wide rather than organizational-specific forms of legitimacy are a primary motivation for ERM adoption by European multinational firms. Neo-institutional theory [30] provides two competing explanations for why companies voluntarily disclose risk information related to their ERM processes [31]. Mimetic or “legitimacy” aspects of the institutional theory argue that ERM risk disclosure is more symbolic than substantive and that risks are not reflected in actual disclosure because managers engage in risk disclosure as a routine activity. Firms are unlikely to disclose ERM-related risk information when other companies fail to do so.

By contrast, a normative aspect of the institutional theory claims that risk disclosure changes over time as a firm’s ex-ante risk changes over time. Based on this argument, managers are more likely to disclose more ERM risk information as their ex-ante risks increase. In the corporate setting, there is a large body of empirical studies that finds that, consistent with this normative aspect prediction, firms with higher risk are more likely to disclose risk. Elshandidy et al. [32] find that voluntary and mandatory risk variations reported both within and across Germany, the UK and the USA are positively associated with firm risk levels, as proxied by market risk measures. However, based on an alternative neo-institutional theory explanation, Tang and Demeritt [33] analyze the carbon reporting practices of a sample of large UK-listed firms following the introduction of mandatory reporting requirements about gas and electricity consumption. They find that firms disclose their emissions in response to social pressure and regulatory compulsion.

To summarize, the overall weight of empirical evidence supports the theory of corporate risk management in justifying the adoption of ERM-based risk management practices. However, there is also evidence that the inter-relation between a firm’s ERM choice and firm risk may be mitigated in institutional environments where cross-country cultural factors, such as regulatory stringency, are important, as suggested by neo-institutional theory.

3. Hypothesis Development

3.1. Costs and Benefits of ERM

This section briefly reviews several arguments for risk management using ERM, each of which justifies the development of hypothesized predictions in the following Section 3.2. Most of the existing theoretical risk management literature [34,35] does not directly discuss ERM adoption incentives but rather focuses on the costs and benefits of hedging market-related financial risk via the use of hedged derivatives.

While the costs of adopting ERM can be high for any firm, for multinational firms these costs are low relative to the benefits of reducing information asymmetries and ameliorating incentive problems associated with firm complexity, such as the number of foreign operations (FORS).

An important attribute of high-reputation firms is their ability to maintain high levels of sustainable or high-quality earnings over time. However, prior empirical research has not so far examined whether the ability of firms to manage their financial risks through ERM adoption, or alternatively, the relationship of firm risk with ERM adoption practices, reporting accounting exposures, and/or other firm, industry or culture-specific factors, or whether it is simply reflecting the underlying economic exposures.

3.2. Hypotheses

This section develops hypotheses concerning the inter-relationship between firm risk and ERM adoption incentives facing multinational firms during the period of the global COVID crisis (2020–2021) and the subsequent geopolitical risks associated with the currently ongoing Russia–Ukraine war (2023–2024). The hypotheses control for firm size (log of total assets, LnSIZE) as well as the timing of adoption (dummy variable Yr1). Additionally, the multivariate tests also specifically control for cross-sectional variation in both market-related factors (book-to-market, BTM) as well as several firm-specific risk factors (e.g., firm complexity (total number of foreign operations (FORS)), financial leverage (LEV), pension funding risk (PFUND), return on assets (ROA) and total notional value of foreign exchange, commodity and interest rate derivatives, scaled by total firm assets (TVH)). The multivariate tests also control for differences in accounting quality of US and European firms (GAAP quality), since the US GAAP is relatively more restrictive than equivalent IAS concerning the disclosure of key financial disclosures such as derivatives, segment reporting and pension costs. All these hypotheses include the assumption that all other factors are held constant.

Propensity to adopt ERM related to firm risk

Following Froot et al.’s [35] arguments concerning the need to reduce incentive problems, it is predicted that the propensity of firms to adopt ERM is primarily related to the desire of multinational firm managers to better manage the overall firm risk exposure of their existing assets, liabilities and internal cash flows. It is therefore predicted that there is a positive association between the propensity of firms to adopt ERM and firm risk, after controlling for various other financial characteristics.

H1. 

Ceteris paribus, the propensity of firms to adopt ERM is positively associated with firm risk.

Vice versa, a positive association is also predicted between total firm risk and the propensity to adopt ERM, after controlling for various firm financial and operating characteristic factors.

H2. 

Ceteris paribus, firm risk is associated with the propensity to adopt ERM.

As noted above, companies based in the European Union as well as the UK are now subject to mandatory public disclosure of their climate change risk reporting alignment with the Task Force for Climate-Related Financial Disclosures [14] climate risk disclosure recommendations, while European firms are additionally required to provide management reports that disclose their risk management strategies. This requires them to disclose whether their ERM-related risk management and strategies are aligned with these requirements. It is therefore predicted that, based on an institutional theory explanation, the mandatory requirements to report both climate and general risk management incentivize European firms to adopt ERM processes related to their reputational risk, which is proxied by political visibility (i.e., firm size).

H3. 

Ceteris paribus, the propensity of European firms to adopt ERM is positively associated with their firm’s reputational risk.

4. Research Design

The research design follows the 2-stage procedure for evaluating ERM retention decisions, as developed in Eckles et al. [6]. Specifically, to test the first hypothesis (H1), a model is specified with the firm’s risks as the dependent variable and ERM adoption and other controls that potentially influence the firm’s risk as the independent variables.

$$firm\_risk=\mathrm{intercept}+\gamma \ast ERM\_adoption+\beta \ast controls.$$

(1)

A finding of $\gamma$ < 0 will be in support of H1. One potential concern in estimating Equation (1) is the self-selection problem.

The first hypothesis (H1) is tested using a multivariate logistic regression model, where the propensity to adopt ERM is the dependent variable and total firm risk and various other financial control variables are outlined in more detail below.

To predict the probability of ERM adoption, the study controls for firm size and operation complexity by using the log of total assets (Lnsize) and the number of overseas business operations (FORS) (Eckles et al. [6] argue that the rationale for these control variables is that (a) the more complex and more myriad risks that a firm faces, the greater benefit a firm can realize by taking a portfolio approach to manage risk). Additionally, there is a lagged measure of firm risk, the log of the annualized standard deviation of monthly stock returns over the previous three years (SDR), to control for the potential relation that riskier firms have a greater incentive to hedge. Since firm earnings are related to the firm’s desire to minimize its operational resilience risk, this study also includes a measure of covariation in firm earnings (COVEARN) (the relation of firm earnings and propensity to use derivatives is a controversial and unresolved issue. Allayannis and Weston [36] argue that derivative use can reduce the volatility of earnings, although these findings are contradicted by Rowntree et al. [37]. Understanding this relation requires further specification of the earnings measure used and whether and how a robust measure and scope of derivative usage are adopted. The relation of earnings to derivative usage and ERM adoption is problematic, and therefore this study does not make any specific prediction on this relation). The study also uses a dummy for GAAP quality for the sample of firms that are non-US and use IFRS (GAAP).

Finally, this study also incorporates institutional-cultural differences between the propensity to adopt ERM for US versus European firms, by including dummy for GAAP quality, for the sample of European multinational firms which are non-US and use IFRS (instead of US GAAP), and are subject to more stringent regulatory requirements for management and risk reporting under the relevant European Union Directives and associated regulations (GAAP quality) (Although the UK BREXIT from the European Union was signaled in 2016, it was not finalized until after the beginning of the study period (i.e., 2020), and therefore the UK firm reporting periods will be assumed to be remain aligned to the relevant EU Directives).

Total firm risk is proxied by the log of the annualized standard deviation of monthly stock returns over the previous three years (SDR). Stock return volatility is the proxy for firm risk, because it is a well-established measure of a firm’s total risk. Mayers and Smith [38] and Smith and Stulz [39] show that, when capital markets are imperfect, firms care about total risk (as opposed to systematic or idiosyncratic risk).

Stock return volatility is also preferred to other alternative measures of firm risk, such as earnings or cash flow volatility, because stock price data are available daily, whereas earnings and cash flow data are only reported quarterly.

To test the second hypothesis (H2), a multivariate OLS regression model is used, with the firm’s risks as the dependent variable and ERM adoption and other controls that potentially influence the firm’s risk as the independent variables.

A finding of $\gamma$ < 0 will be in support of H1.

As mentioned earlier, this study also follows Eckles et al. [6] by including other variables that the existing literature predicts influence firm risk, such as firm size (the log of total assets, LnSIZE), growth opportunities (the log of the book-to-market ratio of assets, BTM), and firm leverage (long-term debt over the sum of debt and equity, LEV), based on the standard argument that debt acts as a lever, magnifying profits and losses, and thus, contributes to higher firm risk [40].

This study also has two additional control variables that are pertinent to firms’ ERM adoption choices but that are not studied by prior research. First, the total of firms’ notional value exposure to interest rate and/or currency risk, scaled by total assets (NVH). This measure of firm risk is relevant to the argument of Froot et al. [35] that derivative usage by firms should be viewed from a broader functional perspective. These arguments imply that derivative usage relates to sources of idiosyncratic or firm-specific risk rather than an institutional perspective, as codified by existing rules to narrowly focus on mitigating specific market risks.

Second, this study includes two specific potential sources of systematic and idiosyncratic risk that are associated with the firm’s total sponsored defined benefit pension fund risk, defined as the relation of market value of pension assets to accrued benefit obligation (PFUND). There are some important reasons why firms’ ERM retention decisions might be associated with pension risk. First, Waring [41] argues that several US-defined benefit pension funds are significantly underfunded, which raises the probability of bankruptcy as firms seek to “put” these obligations to the Federal Government’s Pension Benefit Guarantee Corporation. This issue was illustrated at General Motors, whose significantly underfunded pension fund led to credit rating downgrades in 2008 and subsequent lengthy ongoing legal disputes with employees as to retaining their pension rights, which may cause disruption to a firm’s operational resilience.

Additionally, an important explanation of ERM adoption is firm complexity, which has traditionally been defined as the total number of operating business segments, which is used by prior studies such as Eckles et al. [6] in the second-stage regression to control for the possibility that firms decide to change business mix or other activities in response to a change in firm risk due to ERM adoption. However, the number of operating business segments of a firm is likely to be highly positively correlated with overall firm size. Further, this study sample is restricted to the largest global multinationals, whose business complexity may be complicated by international diversification of business operations. Therefore, this study replaces the standard proxy (number of foreign operating segments) with the number of geographic segments to reflect the agency-related costs of complexity arising from the demand for decentralization of decision-making control facing multinational firms (i.e., FORS).

5. Sample Selection, Data Sources and Sample Description

5.1. Sample Selection and ERM Identification

The sample selection process is based on the population of the globally largest publicly traded companies in the US and Europe, respectively (i.e., the S&P 500 and Euro top 300). This population is then restricted to non-financial companies that are publicly traded and where one can access and utilize stock return data and more easily identify ERM implementation through their public filings (the decision to exclude financial firms is mainly due to the restrictive legal and regulatory obligations imposed on these firms through either mandatory regulatory requirements to establish a “risk committee” and/or appoint a chief risk officer). After controlling for financial firms, non-surviving firms and entries and withdrawals over the period 2021–2023, the study is based on a final SIC three-digit industry-matched balance stratified sample of 100 firms, comprising an equally weighted sub-sample of 50 European and 50 US-listed firms. These firms were also partitioned by industry sub-sector (manufacturing versus non-manufacturing firms) to ensure that there were no intra-industry biases across these two jurisdictions. The decision to focus on only European and US-listed firms is undertaken to identify the impact of cultural differences in GAAP treatment between IFRS and US GAAP enforcement and the relatively more stringent requirements for European companies subject to the requirements of the European Union Accounts and CSRD directives.

Non-financial large firms based in either the US or Europe are not legally required to disclose information about ERM implementation. Therefore, this paper follows the procedure suggested by Hoyt and Liebenberg [19] and Eckles et al. [6] to identify ERM adoption for the above-mentioned 100 firms. Specifically, company annual reports were searched using keywords such as ‘Chief Risk Officer’, ‘Enterprise Risk Management’, ‘Enterprise Risk Officer’, ‘Strategic Risk Management’, ‘Integrated Risk Management’, ‘Holistic Risk Management’ and ‘Consolidated Risk Management’. Furthermore, reference must be made to the relevant ISO [8] and/or COSO [7] standard enterprise risk management frameworks. Additionally, the search is undertaken for each of the three years from 2021 to 2023, and then a dummy variable “year 1” is recorded to document whether the publication of the annual report identified whether the firm adopted in the first year of the study period, i.e., during the COVID crisis. This procedure revealed that 110 (i.e., 63%) of the total sample firms adopted ERM during the study period, i.e., fiscal years from 2021 to 2023. This is a much higher percentage than reported in previous studies and reflects the maturity of ERM adoption practices over time. Furthermore, there was very little change in ERM adoption over the study period. Therefore, this study does not address issues related to the inter-temporal relationship between ERM adoption timing and firm risk issues examined in previous research [6].

5.2. Data Sources

Relevant financial information was collected directly from company annual report financial statements, while stock price data were obtained from the Google Finance website. Soft information such as the number of foreign business segments and derivative reporting data was obtained from the relevant footnote information as disclosed in the “segment reporting” and “risk management” sections of the annual report.

5.3. Descriptive Statistics

Table 1. Descriptive Statistics.

Panel A: Total Sample
Variable	N	Mean	Median	Min	Max	Stdev
SD	300	9.094	8.001	3.520	30.141	3.285
SIZE	300	121,301	85,230	6940	600,340	114,231
FORS	300	3.240	3	0	13	2.887
LEV	300	0.663	0.662	0.278	1.126	0.162
ROA	300	0.065	0.055	−0.130	1.125	0.076
BTM	300	0.562	0.466	−0.024	4.964	0.566
Covearn	300	53.434	38.025	−463.040	518.560	96.036
PFUND	300	0.907	0.981	0.001	1.280	0.219
TVH	300	0.114	0.057	0	3.061	0.231
Panel B: US sub-sample
SD	150	11.91	10.064	4.290	30.14	4.290
SIZE	150	114,284	73,870	6940	564,010	114,603
FORS	150	3.253	3	0	12	2.759
LEV	150	0.652	0.636	0.278	1.126	0.175
ROA	150	0.059	0.050	−0.130	0.440	0.078
BTM	150	0.625	0.531	−0.024	4.565	0.601
Covearn	150	53.051	39.650	−463.040	482.216	118.076
PFUND	150	0.889	0.964	0.007	1.226	0.174
TVH	150	0.116	0.042	0	3.061	0.288
Panel C: European sub-sample
SD	150	8.520	7.054	3.573	26.201	3.490
SIZE	150	121,926	93,780	8680	600,340	112,896
FORS	150	3.194	3	0	13	3.194
LEV	150	0.644	0.656	0.365	1.006	0.114
ROA	150	0.053	0.041	−0.089	0.483	0.065
BTM	150	0.641	0.576	−0.024	4.964	0.516
COVEARN	150	64.671	40.701	−401.998	482.216	97.559
PFUND	150	0.867	0.956	0.007	1.240	0.268
TVH	150	0.131	0.093	0	0.807	0.158

Note: This table provides descriptive statistics on the independent variables for the total sample of firms relating to three fiscal years, 2021–2023. Variable definitions: SD = Standard deviation of firm stock returns, calculated daily over one year. SIZE = Total assets in USD millions on 31 December. BTM = Ratio of book value of common equity to market value of equity. LEV = leverage ratio, which equals long-term debt divided by long-term debt plus common equity. FORS = number of overseas identified operations. ROA = EBIT divided by total assets. Covearn = coefficient of variation for EBIT over the past 3 years. PFUND = ratio of market value of firm’s sponsored defined benefit pension fund assets to projected benefits. TVH = total notional value of hedged foreign exchange and interest rate derivatives, scaled by total assets.

reports the overall descriptive statistics for the main independent variables for the sample firms. These are reported separately for the entire sample firms (Panel A), US sub-samples of firms (Panel B), and EU sub-sample firms (Panel C). The table shows that the average variables are very similar, with US sub-sample firms having slightly higher share price volatility. By contrast, EU sub-sample firms have higher firm size, book-to-market ratio, covariation of earnings and notional value of hedges scaled by total assets.

Table 2. Univariate t-test analysis.

Panel A: ERM Adoption	ERM Firms		Non-ERM Firms		t-Statistic
	N	Mean	N	Mean
SD	190	0.094	110	0.085	1.858 *
SIZE	190	116,857	110	128,976	−0.885
FORS	190	3.610	110	2.600	2.959 ***
BTM	190	0.605	110	0.487	1.752 **
LEV	190	0.668	110	0.655	0.628
ROA	190	0.074	110	0.061	1.473
Covearn	190	55.980	110	49.035	0.603
PFUND	190	0.880	110	0.955	−2.882 ***
TVH	190	0.093	110	0.126	−1.195 ***
Panel B: Volatility of Stock Returns	Low-Volatility Firms		High-Volatility Firms		t-Statistic
	N	Mean	N	Mean
ERM	150	0.587	150	0.680	−1.679 **
SIZE	150	128,318	150	114,284	1.062 *
FORS	150	3.227	150	3.253	−0.080
BTM	150	0.500	150	0.625	−1.941 **
LEV	150	0.675	150	0.652	1.218
ROA	150	0.072	150	0.059	1.409
Covearn	150	53.817	150	53.051	0.069
PFUND	150	0.925	150	0.889	1.449
TVH	150	0.112	150	0.116	−0.115

Note: This table provides univariate two-sample t-tests on the independent variables for the pooled samples of ERM choice (Panel A) and degree of volatility of stock returns (Panel B), relating to three fiscal years, 2021–2023. Where * = 10% level of significance, ** = 5% level of significance and *** = 1% level of significance. Variable definitions (note for reference that these are consistent with those in the following tables): ERM = whether the firm adopted ERM during the fiscal year. SD = standard deviation of firm stock returns, calculated daily over one year. SIZE = market value of equity of stock as of 31 December or total assets in millions on 31 December. BTM = Ratio of book value of common equity to market value of equity. LEV = leverage ratio, which equals long-term debt divided by long-term debt plus common equity. FORS = number of overseas identified operations. ROA = EBIT divided by total assets. Covearn = coefficient of variation for EBIT over the past 3 years. PFUND = Ratio of market value of defined benefit pension fund assets to obligations. TVH = total notional value of hedged foreign exchange and interest rate derivatives, scaled by total assets.

further partitions the sample by whether a firm implemented ERM during the study period (Panel A) and additionally partitions the sample by whether a firm exhibits stock return volatility greater (lower) than the sample mean (Panel B). It also reports two-tailed t-statistics that compare the mean differences for each of the main variables. Table 2 and Panel A show that ERM-adopting firms exhibit relatively higher stock volatility, have a higher number of overseas business operations, and have a higher book-to-market ratio than non-ERM-adopting firms, but have relatively lower pension funding and hedged derivatives. Panel B shows that, relative to low-volatility firms, high-volatility firms are more likely to adopt ERM and have higher book-to-market ratios.

Table 3. Correlations among independent variables.

Variable	LnSIZE	FORS	BTM	LEV	ROA	Covean	Pfund	TVH
LnSIZE	1
FORS	0.001	1
BTM	−0.523	0.020	1
LEV	−0.174	0.044	−0.010	1
ROA	0.080	0.015	0.188	0.039	1
Covearn	−0.135	0.001	0.147	−0.036	0.050	1
PFUND	0.120	0.305	−0.127	−0.090	0.153	−0.010	1
TVH	−0.250	0.097	−0.128	0.168	0.176	−0.010	−0.005	1

reports the correlations among variables. Except for the well-known significantly negative relation between book-to-market and firm size, there are no significant positive or negative correlations, indicating serial correlation and covariation are not an issue.

6. Empirical Tests

This section first reports the results of multivariate logistic and OLS regression models that are used to test hypotheses 1 and 2, respectively. It then reports the results of robustness checks on the baseline results by decomposing the overall findings based on each of the US and EU firm sub-sample firms.

6.1. Impact of the Propensity to Adopt ERM

Table 4. Logistic regression of ERM adoption (full sample).

Variable	Coeff	p Value
Yr1	0.071	0.23
GAAP quality	0.233	0.01
SDR	49.68	0.02
LnSIZE	0.974	0.94
FORS	1.087	0.10
BTM	0.977	0.94
LEV	2.680	0.19
ROA	1.078	0.89
Covearn	1.000	0.85
PFUND	0.332	0.16
TVH	1.610	0.47
Constant	1.813	0.77
Number of observations	300
PseudoR2	0.127
Wald chi-squared	49.92

reports the logistic regression results of the first hypothesis that firms’ propensity to use ERM is related to the predictions of financial risk management theory, after controlling for institutional differences between the US and EU and other firm-based financial and/or broader capital management characteristics (i.e., Equation (1)). The empirical results are generally consistent with both predictions. The relation between ERM adoption and other factors is more equivocal.

6.2. Determinants of Volatility of Stock Return

Table 5. Determinants of volatility of stock return OLS regression (full sample).

Variable	Coefficient	p Value
Yr1	0.028	0.001
GAAP quality	0.025	0.001
ERM	0.012	0.012
LnSIZE	−0.013	0.001
FORS	−0.001	0.481
BTM	0.004	0.322
LEV	−0.054	0.001
ROA	0.012	0.701
Covearn	0.001	0.241
PFUND	−0.007	0.497
TVH	−0.010	0.294
Constant	0.254	0.001
Number of observations	300
F-statistic	9.59
Adj R2	0.223

reports OLS regression results concerning the second hypothesis, i.e., determinants of total firm risk. Consistent with the predictions of H2, we find that the propensity of firms to adopt ERM is positively statistically associated with firm risk. Additionally, US firms have significantly higher volatility exposure than European-based firms. Furthermore, firm risk is positively related to firm size and leverage, indicating that larger and more leveraged firms are subject to greater stock volatility, consistent with the findings of the broader finance literature. Finally, the dummy variable Yr1 indicates that there is significantly higher volatility of firms in the first reporting period, 2021, suggesting that the COVID crisis had a significant impact on overall firm share price volatility.

Table 6. Determinants of volatility of risk to reward ratio OLS regression (full sample).

Variable	Coefficient	p Value
Yr1	−0.038	0.766
GAAP quality	0.075	0.569
ERM	−0.051	0.699
LnSIZE	0.176	0.001
FORS	−0.031	0.153
BTM	−0.251	0.050
LEV	0.122	0.755
Covarearn	−0.001	0.873
PFUND	0.231	0.431
TVH	0.812	0.003
Constant	−1.211	0.003
Number of observations	300
F-statistic	4.79
Adj R2	0.112

reports the OLS multivariate tests of the determinants of the risk-to-reward ratio, defined as ROA/SDR, consistent with the procedure of Eckles et al. [6].

Table 6 shows that there is no statistically significant relationship between the full sample firms’ risk-to-reward ratio and ERM retention decisions. There is also a relatively low F-statistic, indicating that firms’ risk-to-reward ratio may be associated with either missing or misspecified control variables. By contrast, there is a statistically significant positive relationship between risk-to-reward ratio and both firm size (LNSIZE) and total notional value of hedged derivatives (NHV), indicating that legitimacy theory is a potential explanation for this finding.

6.3. Variations Between EU Versus US Firms

To better delineate the standard finance risk management and alternative institutional theory explanations for the baseline findings (H3), the logistical and OLS multivariate empirical tests conducted for the entire sample of firms in Table 4 and Table 5 are repeated for each of the 150 US and European sub-sample firms in

Table 7. Logistic regression of ERM adoption (US and EU sub-samples).

Variable	US Sub-Sample		EU Sub-Sample
	Coeff	p Value	Coeff	p Value
Yr1	0.738	0.42	0.653	0.41
SDR	36.94	0.03	0.001	0.07
LnSIZE	1.09	0.59	0.766	0.30
FORS	1.089	0.25	1.160	0.11
BTM	1.249	0.58	0.869	0.74
LEV	10.68	0.01	0.006	0.03
ROA	59.92	0.10	0.030	0.25
Covearn	0.998	0.32	1.001	0.92
PFUND	0.264	0.30	0.345	0.36
TVH	0.629	0.49	0.239	0.04
Constant	0.051	0.21	601.0	0.15
Number of observations	150		150
PseudoR2	0.117		0.151
Wald chi-squared	15.44		23.50

and

Table 8. Determinants of volatility of stock return OLS regression (US and EU sub-samples).

Variable	US Sub-Sample		EU Sub-Sample
	Coefficient	p Value	Coefficient	p Value
Yr1	0.023	0.001	0.032	0.001
ERM firm	0.015	0.007	0.009	0.154
LnSIZE	−0.014	0.001	−0.008	0.002
FORS	0.001	0.661	−0.001	0.676
BTM	0.012	0.099	−0.001	0.818
LEV	−0.057	0.001	−0.047	0.062
ROA	0.017	0.701	−0.043	0.727
Covearn	0.001	0.561	0.001	0.384
PFUND	0.002	0.911	−0.005	0.621
NHV	−0.006	0.581	−0.022	0.186
Constant	0.272	0.001	0.205	0.001
Number of observations	150		150
F-statistic	8.88		6.31
Adj R2	0.343		0.340

, respectively.

Table 7 shows that stock price volatility is positively associated with ERM adoption propensity for both US () and European sub-sample firms, although only at the 5% and 10% levels of statistical significance, respectively. Additionally, leverage (but not firm size) is statistically significantly associated with ERM adoption propensity for both sub-sample firms. For European firms only, it is also positively associated with the total notional value of hedged derivatives. These results suggest that the incentives facing both US and European firms to adopt ERM are relatively consistent, supporting a financial risk management explanation.

By contrast, Table 8 shows that while ERM adoption is strongly and positively statistically significantly associated with firm risk for both the US sub-sample and European sub-sample firms.

In contrast to the findings of the logistical regression, the OLS regression indicates firm risk is not related to ERM adoption propensity for European sub-sample firms. However, there is a consistent, negative association between firm size and firm risk for both sub-samples. This suggests that there may be alternative institutional reasons why European firms tend to adopt ERM practices, which might be related to their need to adhere to more relatively stringent regulatory requirements for disclosure related to broader climate and systemic risk factors.

Finally, a multivariate OLS regression is also performed on the determinants of risk-to-reward ratio as reported for the entire sample in Table 6, separately for US and European sub-sample firms. The results are reported in

Table 9. Determinants of volatility of risk to reward ratio OLS regression (US and European sub-samples).

Variable	US Firm Sub-Sample		European Sub-Sample
	Coefficient	p Value	Coefficient	p Value
Yr1	0.012	0.950	−0.133	0.438
ERM	0.086	0.631	−0.140	0.495
LnSIZE	0.181	0.032	0.191	0.038
FORS	−0.074	0.047	0.005	0.832
BTM	−0.271	0.163	−0.249	0.158
LEV	−0.286	0.540	1.221	0.136
Covearn	0.001	0.479	−0.001	0.321
PFUND	−0.418	0.511	0.606	0.071
TVH	1.051	0.001	0.045	0.936
Constant	−0.311	0.785	−2.305	0.119
Number of observations	300		300
F-statistic	3.78		2.21
Adj R2	0.144		0.067

.

There is no statistically significant relationship between firm risk-to-reward ratio and ERM adoption propensity for both sub-sample firms. By contrast, there is a positive and statistically significant relationship between both firm size and total notional value of hedging to the risk-to-reward ratios for both US and European sub-samples. Moreover, the overall F and adjusted R-squared statistics suggest that the models do not explain firms’ risk-to-reward ratios. This suggests that ERM adoption is not a major factor in explaining multinational firms’ overall risk-to-reward ratios, which is in contrast to the findings of previous research.

6.4. Robustness Tests

In this section, several robustness tests are conducted on the baseline results reported above (based on the findings above, the robustness checks reported do not incorporate further analysis of the determinants of sample firms’ risk-to-reward ratios). These involve industry-based grouping, incorporating additional country-based macroeconomic variables, changing the dependent variable definition and testing for sub-sample variations between firms with relatively high and low levels of leverage and derivative usage.

Firstly, because ERM adoption practices could vary between industries. Lam [2] argues that manufacturers face relatively greater technology risk and challenges, production-related labor resilience issues, and supply chain issues that are subject to geopolitical- and climate-related systemic risk, being significantly more of an issue facing manufacturing firms than non-manufacturing firms. Exactly 50% of the sample firms are manufacturing-based. The analysis is therefore repeated where the sample firms are decomposed into manufacturing and non-manufacturing industry sub-groups.

Table 10. Logistic regression of ERM adoption (industry sub-samples).

Variable	Manufacturers		Non-Manufacturers
	Coeff	p Value	Coeff	p Value
Yr1	0.988	0.98	0.752	0.51
SDR	0.022	0.60	27.92	0.30
LnSIZE	0.435	0.00	0.827	0.31
FORS	0.968	0.66	1.347	0.00
BTM	0.175	0.03	2.157	0.14
LEV	4.579	0.24	1.863	0.63
ROA	0.003	0.10	0.001	0.00
Covearn	0.998	0.42	0.998	0.43
PFUND	0.001	0.00	3.859	0.17
TVH	1.983	0.34	68.79	0.01
Constant	0.001	0.00	0.172	0.57
Number of observations	150		150
PseudoR2	0.210		0.155
Wald chi-squared	41.50		30.25

and

Table 11. Determinants of volatility of stock return OLS regression (industry sub-samples).

Variable	Manufacturers		Non-Manufacturers
	Coefficient	p Value	Coefficient	p Value
Yr1	0.020	0.001	0.033	0.001
ERM firm	−0.002	0.699	0.004	0.499
LnSIZE	−0.012	0.001	−0.014	0.001
FORS	−0.001	0.268	0.001	0.423
BTM	0.019	0.033	−0.003	0.604
LEV	−0.009	0.001	−0.071	0.001
ROA	−0.068	0.028	0.179	0.009
Covearn	−0.001	0.139	0.001	0.777
PFUND	0.008	0.498	−0.001	0.925
NHV	0.006	0.486	−0.078	0.003
Constant	0.222	0.001	0.293	0.001
Number of observations	150		150
F-statistic	5.73		10.28
Adj R2	0.241		0.383

report the same logistic and OLS regression as reported in Table 4 and Table 5, respectively, but where sample firms are partitioned by whether the firm is manufacturing-based or not.

The ERM incentive logistic multivariate tests (Table 10) results suggest that there is no statistical association between ERM incentive probability and total firm risk (SDR), contrary to the results of the baseline tests and therefore rejecting the prediction of H1. Some control variables also have different significant associations with ERM choice incentives between the two sub-samples. For instance, compared to the manufacturing firm sub-sample, for non-manufacturing sub-sample firms, ERM choice is positively associated with foreign operations (FORS) and total notional value of hedges (TVH). By contrast, there is a significant association between firm size (LNSIZE), book to market (BTM) and pension funding (PFUND) only for the manufacturing firm sub-sample.

Similarly, the OLS regression multivariate tests of the determinants of total firm risk (SDR) reported in Table 11 show that ERM is not associated with total firm risk. However, for both manufacturing and non-manufacturing sub-sample firms, there is a statistically significant negative relationship between total firm risk and both firm size (LNSIZE) and leverage (LEV). For the other control variables, there is only a positive statistical relationship between firm risk (SDR) and book-to-market ratio (BTM) for the manufacturing firms, while there is only a positive statistical relationship between firm risk and notional value of hedging for the non-manufacturing firms.

As a further robustness check, two additional control cross-country-level macroeconomic control variables are included in the multivariate tests, which could explain cross-section variation facing both firms’ incentives to adopt ERM and total firm risk. Country-level cultural factors that are included are (1) the well-known Kaufmann index of institutional development (ID), which is constructed based on a multivariate statistical analysis of six different institutional-level factors associated with “Voice and Accountability”, “Regulatory Quality”, “Control of Corruption” and other factors [42]. Additionally, the country-based stock market development index measure contained in the World Bank Global Development Database is also included as another country-level macroeconomic control variable (SMD).

Table 12. Logistic regression of ERM adoption (including macroeconomic control variables).

Variable	Coeff	p Value
Yr1	−0.743	0.300
SDR	152.21	0.018
LnSIZE	−0.949	0.704
FORS	1.085	0.104
BTM	−0.967	0.904
LEV	2.689	0.237
ROA	1.271	0.901
Covearn	−0.999	0.789
PFUND	−0.446	0.318
TVH	1.625	0.468
ID	5.677	0.005
SMD	−0.991	0.001
Constant	0.699	0.870
Number of observations	300
PseudoR2	0.120
Wald chi-squared	47.50

and

Table 13. Determinants of volatility of stock return OLS regression (including macroeconomic variables).

Variable	Coefficient	p Value
Yr1	0.028	0.001
ERM firm	0.010	0.017
LnSIZE	−0.012	0.001
FORS	−0.001	0.870
BTM	0.007	0.107
LEV	−0.056	0.001
ROA	−0.005	0.853
Covearn	0.001	0.374
PFUND	−0.009	0.347
NHV	−0.012	0.207
ID	−0.012	0.159
SMD	0.001	0.001
Constant	0.240	0.001
Number of observations	300
F-statistic	11.76
Adj R2	0.302

report the logistic multivariate regression and OLS multivariate tests used to examine the determinants of ERM choice and total firm risk, respectively, that include these additional macroeconomic variables.

Table 12 shows that the propensity to adopt ERM is positively associated with firm total risk, even after controlling for other country-wide macroeconomic factors. There is also a statistically significant negative (positive) association with international development (ID) (stock market development (SMD)), indicating that macroeconomic factors are important determinants of firms’ ERM adoption propensity.

Table 13 reports that there is a positive and statistically significant relation between SDR and ERM adoption choice, even after controlling for international-level factors. However, there is only a positive and statistically significant relation between firm risk and stock market development (SMD).

A further robustness check is conducted, which tests the sensitivity of the results to a change in the definition of the dependent variable, ERM choice. Fraser et al. [3] claim that ERM adoption has been diluted by several factors that impede their effective ERM implementation, such as an overemphasis on reporting and insufficient injection into the strategic managerial decision-making processes. Furthermore, Florio and Leoni [26] introduce a new and more complete measure for ERM implementation, which they claim, “concerns not only corporate governance bodies dedicated to risk management, but also the characteristics of the risk assessment process”.

To address these criticisms, the definition of ERM adoption is amended to incorporate a “credible” ERM adoption option, whereby firms have voluntarily reported that they have explicitly incorporated ERM into their overall governance and risk management structures. Therefore, instead of a binary one–zero variable definition as used consistently in prior literature, a categorical “level 2” ERM adoption level was identified, based on a review of the publicly available Management Discussion and Analysis section of the annual report. This procedure identified only 26% (instead of 63%) of the sample firms that had demonstrated via public voluntary disclosures that they had produced “credible” ERM processes, as opposed to “standard” ERM processes produced more generally in compliance with reporting regulations. Therefore, the empirical multivariate analysis reported above is repeated, but instead of a binary logistical model, an ordinal logistical regression model is used. The results are reported in

Table 14. Ordinal logistic regression of ERM adoption (including macroeconomic control variables).

Variable	Coeff	p Value
Yr1	−0.201	0.431
SDR	4.679	0.131
LnSIZE	−0.038	0.741
FORS	0.043	0.292
BTM	0.007	0.973
LEV	1.885	0.013
ROA	0.221	0.893
Covearn	0.001	0.926
PFUND	−1.056	0.068
TVH	0.352	0.543
ID	0.784	0.131
SMD	−0.011	0.001
Number of observations	300
PseudoR2	0.088
Wald chi-squared	57.31

, whereby, reflecting a change in definition of the dependent variable, an ordinal logistical regression is substituted for a binary logistical regression model first-stage specification.

The empirical results show that a change in the definition of the dependent variable ERM from a binary to categorical variable reduces the strength of the statistical association with firm risk. Although it is still positive, it is not statistically significant. By contrast, the relation between the amended “credible” definition of ERM adoption and other control variables is statistically and positively associated with both firm leverage (LEV) and stock market development (SMD) at the 1% level of statistical significance.

The final robustness checks concern the relative importance of both leverage and notional value of hedged derivatives in affecting the relationship between ERM adoption propensity and firm risk. Firstly,

Table 15. Ordinal logistic regression of ERM adoption (high- vs. low-derivative-using sub-sample firms).

Variable	High-Derivative-Using Firms		Low-Derivative-Using Firms
	Coeff	p Value	Coeff	p Value
Yr1	0.501	0.158	0.721	0.452
SDR	1324.48	0.153	2.641	0.001
Size	1.261	0.075	0.681	0.017
FORS	0.960	0.595	1.118	0.178
BTM	0.261	0.186	0.706	0.289
LEV	1.823	0.896	19.355	0.011
ROA	3.695	0.718	3.131	0.687
Cavern	0.997	0.311	1.0001	0.809
PFUND	0.134	0.165	−0.203	0.243
TVH	0.656	0.497	5.001	0.001
ID	20.554	0.001	4.976	0.095
SMD	−0.982	0.001	−0.993	0.177
Number of observations	150		150
PseudoR2	0.089		0.087
Wald chi-squared	36.65		29.90

and

Table 16. Determinants of volatility of stock return (high- vs. low-derivative-using sub-sample firms).

Variable	High-Derivative-Using Firms		Low-Derivative-Using Firms
	Coeff	p Value	Coeff	p Value
Yr1	0.027	0.001	0.028	0.001
ERM	0.016	0.010	0.023	0.001
Size	−0.020	0.001	−0.004	0.125
FORS	0.001	0.878	0.004	0.704
BTM	0.021	0.015	0.010	0.060
LEV	−0.115	0.001	−0.035	0.034
ROA	−0.010	0.801	−0.009	0.824
Cavern	0.001	0.643	0.001	0.220
PFUND	−0.013	0.244	0.004	0.795
TVH	0.002	0.848	−0.663	0.001
ID	−0.041	0.001	4.973	0.641
SMD	0.001	0.001	0.001	0.037
Constant	0.381	0.001	0.128	0.004
Number of observations	150		150
F-statistic	11.12		6.97
Ad R2	0.449		0.379

report the effects on the propensity to adopt ERM and the determinants of firm risk, respectively, when the sample firms are partitioned based on the median (i.e., 5.72% of total assets) of the key control variable notional value of hedged derivatives (TNV).

Table 15 shows that, whereas there is a positive and statistically significant relationship between ERM adoption propensity and firm risk for low derivative hedging firms, it is not statistically significantly associated with high derivative hedging firms. Furthermore, there is a negative association between hedging (TNV) and ERM propensity for low-risk sub-sample firms, but there is no statistical association between these variables for high derivative usage sub-sample firms. Conversely, Table 16 shows that there is a positive and statistically significant relationship between ERM adoption choice and total firm risk for both high- and low-derivative-usage firms. Additionally, there is a consistent and statistically negative association between total firm risk and the notional value of derivatives for both high- and low-derivative-usage sub-sample firms.

Finally,

Table 17. Ordinal logistic regression of ERM adoption (high- vs. low-leverage sub-sample firms).

Variable	High-Leverage Firms		Low-Leverage Firms
	Coeff	p Value	Coeff	p Value
Yr1	0.765	0.529	0.696	0.434
SDR	1.381	0.017	636.718	0.166
Size	0.867	0.381	1.012	0.928
FORS	1.247	0.020	1.007	0.926
BTM	0.689	0.337	1.562	0.399
LEV	0.959	0.982	0.775	0.906
ROA	26.475	0.360	0.755	0.820
Cavern	0.997	0.350	1.002	0.253
PFUND	0.529	0.495	0.245	0.314
TVH	0.617	0.495	987.841	0.013
ID	4.601	0.069	6.076	0.948
SMD	−0.011	0.001	0.989	0.006
Number of observations	150		150
PseudoR2	0.089		0.088
Wald chi-squared	26.30		28.84

and

Table 18. Determinants of volatility of stock return (high- vs. low-leverage sub-sample firms).

Variable	High-Leverage Firms		Low-Leverage Firms
	Coeff	p Value	Coeff	p Value
Yr1	0.016	0.001	0.037	0.001
ERM	0.010	0.027	0.008	0.240
LnSIZE	−0.005	0.016	−0.017	0.001
FORS	0.001	0.260	−0.002	0.068
BTM	0.017	0.001	−0.002	0.714
LEV	0.052	0.020	−0.065	0.127
ROA	−0.065	0.051	−0.009	0.851
Covearn	−0.001	0.590	0.001	0.414
PFUND	0.014	0.140	−0.039	0.124
TVH	−0.001	0.907	0.005	0.879
ID	0.001	0.962	−0.014	0.358
SMD	0.001	0.284	0.001	0.001
Constant	0.063	0.093	0.336	0.001
Number of observations	150		150
F-statistic	6.96		8.48
Adj R2	0.324		0.376

report the effects on the propensity to adopt ERM and the determinants of firm risk, respectively, when the sample firms are partitioned based on the medians of the key control variable firm leverage (LEV).

Table 17 shows that only high-leverage sub-sample firms exhibit a positive and statistically significant relationship between both ERM adoption propensity leverage (LEV) and total firm risk. Equivalently, Table 18 indicates a positive and statistically significant relationship between total firm risk and both ERM adoption and firm leverage for only the high-leverage sub-sample firms.

7. Conclusions

Incentives facing globally large multinational industrial firms to adopt ERM practices have traditionally been regarded as endogenous with financial management theory and management strategic imperatives and/or to adhere to relevant reporting requirements. However, the recent global pandemics related to COVID-19, geopolitical conflicts such as the Ukraine–Russia war, and international prerogatives raise new challenges to whether the globally largest multinational firms that are based in the US and Europe can better deal with the global challenges of systemic risks related to climate change and biodiversity loss by adopting ERM practices. This paper provides the first comprehensive study of how a stratified and industry-matched sample of globally large US and EU multinational industrial firms relate their ERM adoption strategies and total firm risk, controlling for firm complexity and financial characteristics.

Specifically, this study extends prior research concerning the relation of ERM adoption propensity and firm total risk and risk–reward ratios, which has mainly focused on regulated financial firms, to multinational non-financial firms. It also contributes by incorporating new measures of sources of idiosyncratic risk (pension risk), variations in GAAP quality, derivative usage and firm complexity factors that face such firms in adopting ERM practices. It is predicted that there is an inter-relationship between ERM adoption choice and total firm risk. The empirical tests are largely consistent with these predictions; however, they are more equivocal for European firms, which are subject to more stringent regulatory risk reporting requirements but where IFRS is less enforced than US GAAP.

The baseline empirical multivariate test findings suggest that US firms adopting ERM practices are likely to be riskier, supporting an information asymmetry explanation. By contrast, European firms adopting ERM tend to be larger in size, supporting an alternative institutional theory–legitimation explanation. However, there is no statistically significant relationship between firm performance and ERM adoption, perhaps reflecting that many firms are at a relatively mature phase in their risk management frameworks in the two decades since the original explication of well-known COSO and ISO frameworks.

The study’s findings are subject to several caveats. Firstly, the empirical findings are based on a restrictive sample of globally large multinational firms. Furthermore, they are subject to variations in the quality of voluntary disclosure of ERM practices and the notional value (scaled by total assets). Furthermore, the study finds that the relationship between ERM and firm risk is mitigated by several limiting factors, such as high versus low relative leverage and derivative usage, the extent of sophistication of ERM adoption practices that incorporate risk management processes, and possible variable endogeneity issues. Additionally, firms based in different industries (manufacturing versus non-manufacturing) have differing degrees of ERM adoption and may experience different earnings–age profiles of employees that may influence the reported pension funding ratio. Further research is needed to address these and other issues.