Capital Structure and Firm Performance: Evidence from FTSE All-Share Firms During COVID-19

Abstract

We examine how capital structure related to firm performance for UK companies in the FTSE All-Share over 2018–2023, explicitly segmenting pre-pandemic (2018–2019), pandemic (2020–2021), and post-pandemic (2022–2023) periods. Using Bloomberg data for 516 firms and panel fixed-effects models (Hausman-tested), we assess the impact of short- and long-term leverage on ROA, ROCE, Tobin’s Q, and EPS, and compare financial versus non-financial firms. Leverage is, on average, negatively associated with ROA and EPS, consistent with pecking-order and agency-cost arguments: market-based outcomes (Tobin’s Q) show weaker, nuanced links. The adverse effects of debt are stronger for non-financial firms, particularly during and after COVID-19, while financial firms display a post-COVID positive association between short-term debt and ROA, suggesting sector-specific debt utilization under stress. Firm size relates negatively to Tobin’s Q for non-financials. Results highlight how crisis conditions and industry characteristics shape the leverage–performance nexus, offering practical guidance for managers and policymakers on capital structure decisions in turbulent environments.

1. Introduction

Capital structure, the composition of debt and equity financing, is pivotal in determining a firm’s performance and value (Dao & Ta, 2020). Since the seminal work of Modigliani and Miller (1958), the relationship between leverage and firm value has remained at the heart of corporate finance, inspiring extensive debate across multiple theories—trade off, pecking order, and agency theory to suggest a few. The rapid evolution of financial markets coupled with the shock of COVID-19 pandemic has further complicated corporate financing decisions, creating a unique setting to assess how firms adjust their financing decisions in response to changing economic conditions.

This article investigates the relationship between capital structure and firm performance among the non-financial companies listed on the FTSE All Share index from 2018 to 2023.

By segmenting the analysis in three distinct time frames, 2018 to 2019 (pre-pandemic), from 2020 to 2021 (during the pandemic) and from 2022 to 2023 (post-pandemic) periods, the study captures how the macroeconomic shocks and recovery phases influence the financing policies and profitability dynamics. This time frame allows for the assessment of short- and long-term debt effects on firm outcomes under varying market stresses. The analysis employs panel data collected from Bloomberg, encompassing profitability measures such as include return on assets (ROA), return on common equity (ROCE), Tobin’s Q and earnings per share (EPS) as dependent variables, while short-term debt to total assets and long-term debt to total assets ratios serve as independent variables. Control variables include firm size, firm growth, current ratio, financial leverage, board size, percentage of female directors, percentage of independent directors, and audit committee size. Regression analyses are conducted using STATA, and the Hausman test determines the appropriate model specifications—fixed or random effects. This approach ensures statistical robustness and comparability with previous empirical studies conducted in European and global markets.

Focusing on the FTSE All Share index, which represents approximately 98% of the UK market capitalization (FTSE Russell Factsheet, 2024), offers a holistic view on the financing patterns of listed firms of various sizes.

The UK market provides an ideal setting for this analysis, offering rich, accessible financial data within a transparent, liquid environment characterised by strong shareholder rights and clear regulatory frameworks (Charalambakis & Psychoyios, 2012). This context enables a thorough examination of capital structure determinants, allowing for the revisitation of established theories and exploration of firm-specific factors (Charalambakis & Psychoyios, 2012).

This study contributes to the literature by linking theoretical insights with empirical evidence within a developed market setting that experienced an extraordinary global crisis, offering implications for managers, investors and policymakers navigating financial decision-making under uncertainty.

The remainder of the paper is organized as follows: Section 2 reviews relevant theoretical frameworks and empirical studies, highlighting gaps in the literature. Section 3 details the research design, data collection methods, sampling strategies, and analytical techniques used to ensure validity and reliability. Section 4 presents and interprets the findings across the three analysed periods. Finally, Section 5 summarizes key findings, discusses implications, acknowledges limitations, and offers recommendations for future research.

2. Literature Review

The concept of capital structure is crucial for an organisation’s long-term viability, encompassing the strategic mix of financing sources used to fund its operations (Baker & Martin, 2011). These financing avenues primarily encompass debt instruments, equity securities, and hybrid financial vehicles (Abor, 2005; Baker & Martin, 2011).

Debt financing, obtained through borrowing or issuing bonds, offers tax deductibility on interest payments but increases financial risk (Frank & Goyal, 2009; Graham, 2003). Conversely, equity financing, achieved through issuing ownership shares, provides freedom from fixed payment obligations but may dilute existing shareholders’ ownership (Admati et al., 2018; Brav, 2009).

Striking the appropriate balance between these two financing forms is critical, as it has far-reaching implications for minimising expenses and managing risk exposure. Numerous factors, including market dynamics, growth prospects and other sector-specific complexities influence this delicate balance (Frank & Goyal, 2009).

Despite numerous theoretical frameworks and perspectives, a consensus on the optimal capital structure remains elusive due to heterogeneous empirical findings and the absence of a universally accepted mathematical formulation (Abor, 2005; Frank & Goyal, 2009; Graham & Leary, 2011; Serrasqueiro & Caetano, 2014). This complexity is compounded by the conflicting evidence on the impact of debt on corporate profitability, necessitating detailed investigation into the relationship between capital structure in specific contexts such as the UK (El-Sayed Ebaid, 2009). This highlights the need for context-specific studies.

The COVID-19 pandemic, which emerged in Wuhan, China in late 2019, rapidly spread across the globe, prompting the World Health Organisation to declare a Public Health Emergency of International Concern (PHEIC) on 30 January 2020, and a global pandemic on 11 March 2020, (WHO, 2024). By March 2020, Europe had become the new epicentre of the outbreak (Shehzad et al., 2021), catalysing a new wave of stringent lockdown measures across numerous nations.

These lockdowns, while necessary for public health, had profound economic consequences. The simultaneous disruption of both supply and demand triggered widespread recessions (Eichenbaum et al., 2021). As individuals reduced social interactions, overall demand decreased significantly. Concurrently, workers’ exposure to the virus disrupted the production and supply chain (Nicola et al., 2020).

Governments faced the challenging task of balancing public health concerns with economic stability. The strict lockdown measures implemented to contain the virus had profound repercussions on economic activity (Elmarzouky et al., 2021; Albitar et al., 2021; Eichenbaum et al., 2021). This crisis highlighted the interconnected nature of the global economy and the far-reaching impacts of a health crisis on various sectors worldwide.

The global economy contracted by 3.4% in 2020 (Dyvik, 2024), with developed economies bearing a heavy burden. The eurozone experienced a 6.1% decline in GDP in 2020, more severe than during the global financial crisis (Verwey & Monks, 2021).

In the UK, the pandemic’s impact was particularly severe. The UK experienced one of the highest death tolls in Europe and a prolonged economic downturn (Romei & Parker, 2024). Among G7 nations, the UK’s economic contraction was the most pronounced, with a 9.4% decline in GDP in 2020 (IMF, 2022).

Although a partial recovery was observed in 2021 and 2022, the UK continued to face significant challenges. The ongoing cost of living crisis has substantially impacted consumer expenditure and corporate activities (Romei & Parker, 2024). This economic strain was further exacerbated by geopolitical tensions, notably the Russia–Ukraine conflict, and persistent global supply chain disruptions (IMF, 2022).

The economic crisis triggered by the COVID-19 pandemic presents a unique opportunity to examine the relationship between capital structure and company performance during periods of economic turbulence. Previous research on global financial crises offers valuable insights. Khodavandloo et al. (2017) demonstrated that during the 2007–2009 economic downturn, leverage negatively influenced the performance of businesses operating in Malaysia’s commerce and service sectors. Similarly, Akgün and Memiş Karataş (2021) observed that during the 2008 European financial crisis, capital structure had a detrimental impact on company performance.

Furthermore, the COVID-19 pandemic has also had substantial reverberations on global stock markets. He et al. (2020) discovered that the pandemic had a negative, albeit temporary, effect on the major stock market indices of Asia, Europe, and the United States, resulting in significant underperformance compared to the period from 1 June 2019 to 29 January 2020. Bentes (2021) reported varying levels of consistency among different indices, with the S&P 500 and FTSE/MIB demonstrating the highest level of consistency, while the NIKKEI 225 and FTSE 100 exhibited lower consistency levels.

While prior studies have explored the impact of capital structure on firm performance in various international contexts, evidence from UK-listed firms during the COVID-19 crisis remains limited. The UK market experienced unique regulatory, fiscal, and financial dynamics during this period, including government-backed loan schemes and liquidity interventions, which may have altered firms’ leverage decisions compared to other countries. This study therefore contributes by focusing specifically on FTSE All-Share companies, offering insights into how UK firms adjusted their financing strategies and capital structures in response to the economic shocks of the COVID-19 pandemic.

To conclude, the COVID-19 pandemic has significantly reshaped the global economic landscape, with the UK experiencing particularly severe impacts. The confluence of public health measures, economic disruptions, and geopolitical tensions has created a complex and challenging environment for businesses operating in the UK. This backdrop of economic volatility and uncertainty underscores the critical importance of effective capital structure management for firms’ survival and success. As companies navigate these turbulent times, the relationship between capital structure decisions and financial performance becomes increasingly crucial.

2.1. Theoretical Foundations

2.1.1. Modigliani–Miller Theorem

The Modigliani–Miller (MM) theorem, introduced in 1958, posited that in perfect capital markets, a firm’s value is unaffected by capital structure. This groundbreaking proposition challenged the prevailing notion that an optimal capital structure existed and sparked a surge of research in corporate finance (Graham & Leary, 2011; Myers, 2001).

However, the MM theorem’s restrictive assumptions, such as the absence of taxes, transaction costs, and market frictions, limited its practical applicability (Fischer et al., 1989; Lemmon et al., 2008). Recognising these limitations, Modigliani and Miller (1963) revised their theory to incorporate corporate taxes, suggesting that firms should maximize debt to benefit from interest tax shields.

2.1.2. Trade off Theory

Building on the MM theorem, Kraus and Litzenberger (1973) developed the trade-off theory, which holds that firms weigh the tax benefits of debt against the costs of financial distress (Jaros & Bartosova, 2015). This theory proposes an ideal capital structure in which the marginal benefit of debt matches its marginal cost (Abel, 2018; Serrasqueiro & Caetano, 2014).

Frank and Goyal (2009) conducted a comprehensive study of publicly traded American enterprises from 1950 to 2003 and discovered that the most dependable criteria for explaining market leverage were median industry leverage, market-to-book assets ratio, tangibility, profits, log of assets, and projected inflation.

2.1.3. Agency Cost Theory

Jensen and Meckling (1976) introduced the agency cost theory, proposing that debt can mitigate agency conflicts between managers and shareholders by reducing free cash flow and imposing financial discipline. This theory suggests that increased leverage can potentially improve firm performance by aligning managerial interests with those of shareholders.

Subsequent studies have provided empirical support for this theory proposition. Berger and Bonaccorsi di Patti (2006) found that higher leverage in the banking industry was associated with improved profit efficiency, consistent with the agency cost hypothesis. Additionally, Myers and Majluf (1984) demonstrated how information asymmetry between managers and investors can lead to agency costs.

2.1.4. Signaling Theory

Ross (1977) proposed the signaling theory, suggesting that firms use capital structure decisions to convey information about their quality to investors. According to this theory, high-quality firms can afford higher debt levels, signaling their confidence in future cash flows.

An et al. (2015) provided empirical support for the signaling theory in the context of seasoned equity offerings (SEOs), finding that firms with higher leverage before SEOs experienced less negative market reactions to SEO announcements.

2.1.5. Pecking Order Theory

Myers and Majluf (1984) developed the pecking order theory, which posits that firms have a hierarchical preference for financing sources due to information asymmetry. According to this theory, firms prefer internal finance to external financing and debt to equity when external funding is required.

Fama and French (2002) tested the trade-off and pecking order theories, finding mixed support for both. Their results showed that more profitable firms and firms with fewer investments pay greater dividends, consistent with the trade-off model, but contrary to the pecking order model.

Of the five theories, three—Agency, Trade-off, and pecking order—generally predict the negative association between leverage and firm performance, while MM and Signaling support the positive relationship (Dao & Ta, 2020). Building on these theoretical perspectives, recent studies have provided further evidence on how these mechanisms manifest in practice, particularly regarding the negative association between leverage and firm performance.

For instance, Stoiljković et al. (2024) find that in Serbian manufacturing firms, higher debt levels increase agency costs and reduce firm efficiency. Similarly, Sdiq and Abdullah (2022) demonstrate that agency costs moderate the capital structure–performance nexus, such that higher leverage enhances performance only when agency costs are low. Evidence from Ghanian firms (Amoako & Attafuah, 2024) also shows that smaller or less-resourced firms experience a significantly negative effect of leverage on performance, indicating that firm size moderates this relationship. Additionally, Ahmed et al. (2023) highlight the impact of leverage on firm performance is contingent on agency cost levels, with excessive debt reducing performance in firms with higher agency costs. Collectively these studies clarify the mechanisms implied by the pecking order and agency cost theories; while debt can serve as a useful financing and disciplinary tool when applied prudently, excessive leverage increases agency costs, constrains managerial flexibility, and ultimately deteriorated firm performance.

These theoretical insights and recent empirical findings together provide a string foundation for examining the capital structure–performance relationship in the UK context and periods of economic crisis.

Building on this foundation, the following section critically evaluates the empirical findings on the relationship between capital structure and firm performance. The literature presents a complex and nuanced picture, with results varying across different contexts, periods, and methodological approaches.

To set the stage for the present study, this review categorises the existing findings into negative, positive, mixed and inconclusive relationships, before focusing on the studies relevant to UK and periods of economic crises. This structure helps to position the current research within both the global database and the specific UK context.

2.2. Negative Relationships

A preponderance of empirical studies suggests an inverse relationship between leverage and firm performance. T. H. Nguyen and Nguyen (2020) examined 488 non-financial listed companies in Vietnam using regression analysis, demonstrating a negative association between capital structure and firm performance. This finding is corroborated by Ngatno (2022), who studied 241 rural banks in Central Java, Indonesia, employing moderating regression analysis.

In the hospitality sector, Babajee et al. (2020) investigated 43 firms in the hotel industry using a dynamic panel data framework, concluding that leverage adversely impacted firm performance. This trend extends to emerging markets, as evidenced by Doan (2020), who analysed 102 firms and found that increased use of debt correlates with decreased firm performance.

The negative relationship pattern is not confined to recent studies. Earlier research by T. Nguyen and Nguyen (2015) demonstrated a detrimental impact of capital structure on financial performance metrics for Vietnamese enterprises. Yapa Abeywardhana (2017) showed a similar negative relationship for UK SMEs in the manufacturing sector. Moreover, Sakr and Bedeir (2019) found that higher levels of total, short-term, and long-term debt significantly negatively affected Return on Assets (ROA) for listed non-financial Egyptian firms. However, their impact on Return on Equity (ROE) was mixed.

These findings resonate with the research conducted by Abor (2007), Ahmad et al. (2012), Le and Phan (2017), Vuong et al. (2017), Majumdar and Chhibber (1999) and Zeitun and Tian (2014), which collectively highlight the adverse impact of liabilities on firm performance across various contexts. These studies finding negative associations between these variables, lend support to the pecking order theory and suggest that higher leverage may lead to increased financial distress costs and reduced operational flexibility.

Although much of this evidence originates from emerging markets, similar patterns have been observed in the UK context, particularly among the SMEs (Yapa Abeywardhana, 2017). These findings suggest that even in developed economies, high leverage can constrain firm performance, especially during periods of financial stress such as the recent economic crisis.

2.3. Positive Relationship

While less common, some studies have found a positive relationship between aspects of capital structure and financial performance. Vu et al. (2020) examined 59 construction companies in Vietnam using regression analysis, revealing that the debt-to-equity ratio had a strong and positive effect on ROE. Similarly, Pinto and Joseph (2017) studied 21 banks using regression analysis, concluding that capital structure had a favourable and significant effect on the financial performance of the selected banks.

In the construction sector, Mohammad et al. (2019) examined 41 Malaysian construction firms using regression analysis, observing a positive association between long-term debt and ROE. These findings are supported by earlier studies such as Hadlock and James (2002) who observed that debt levels and profitability often align, suggesting profitable firms may maintain higher debt proportions. Gill et al. (2011) found that total, long-term, and current liabilities positively affect ROE in US manufacturing companies. Margaritis and Psillaki (2010) and Roden and Lewellen (1995) also provided evidence of a positive relationship between corporate leverage and firm profitability.

However, much of these studies were conducted outside the UK, and often during stable economic periods. Whether such positive effects of leverage persist under crisis conditions in developed markets remain less clear—a gap that this study aims to address.

2.4. Mixed or Inconclusive Results

Despite the preponderance of studies indicating a negative relationship between capital structure and firm performance, a significant body of research presents mixed or inconclusive results. Akintoye (2008) demonstrates that alterations in capital structure significantly influence firm performance, as measured by earnings per share (EPS), earnings before interest and taxes (EBIT), and dividend per share. However, the directionality of this relationship remains ambiguous, echoing the complexity observed in the broader literature. This ambiguity is further exemplified by Stulz’s (1990) model, which posits that debt can exert both positive and negative effects on company performance. In a more recent study, Tripathy and Shaik (2020) examined 56 firms in the Indian food processing industry using OLS regression. Their findings revealed a significant association between leverage and firm performance, yet the direction of this association was not consistently negative or positive. Collectively, these studies underscore the nuanced and context-dependent nature of the capital structure–performance relationship, highlighting the need for careful interpretation of empirical results in this field. The heterogeneity of finding further justifies the need to investigate this relationship within distinct institutional and macroeconomic environments. The UK, with its mature financial markets and the profound impact of the economic crisis, offers a compelling context for such an analysis.

A major point of contention in the literature is the appropriate measure of firm performance. While some studies focus on accounting-based measures like ROA and ROE (e.g., Sakr & Bedeir, 2019), others argue for market-based measures or a combination of both (e.g., Margaritis & Psillaki, 2010). This lack of consensus makes direct comparisons between studies challenging.

Another area of disagreement is the treatment of debt. While some studies consider total debt (e.g., T. H. Nguyen & Nguyen, 2020; Hasan et al., 2025), others differentiate between short-term and long-term debt (e.g., Abor, 2007; Moussa & Elmarzouky, 2024; Elmarzouky et al., 2025). This inconsistency in debt measurement contributes to the heterogeneity of findings.

Furthermore, there is ongoing debate about the appropriate methodological approach. While many studies rely on regression analysis, some argue for more advanced techniques like GMM estimation to address endogeneity concerns (e.g., Le & Phan, 2017; Albitar et al., 2025).

In conclusion, while the majority of evidence suggests a negative relationship between leverage and firm performance, the empirical literature is far from unanimous. The conflicting findings highlight the need for more nuanced, context-specific research that can account for the complex interplay of factors influencing this relationship. This study responds to that gap by focusing on UK firms during the economic crisis period, offering evidence from a developed economy under stressed financial conditions. Future studies should aim to address the methodological limitations identified, particularly in terms of sample size, geographical diversity, and robust treatment of endogeneity issues.

2.5. Hypothesis Development

The extant literature reveals a complex and often contradictory relationship between capital structure and firm performance. Despite extensive research, the nature of this association remains a subject of ongoing debate, with empirical findings yielding inconsistent results. While some studies posit a negative correlation, others suggest a positive link. However, a substantial body of evidence lends support to the notion of an inverse relationship.

This empirical ambiguity provides the foundation for primary hypothesis of this study, which is grounded in two seminal theories of capital structure: the trade-off theory (Kraus & Litzenberger, 1973) and the pecking order theory (Myers & Majluf, 1984). These theoretical frameworks posit that while debt financing can offer tax advantages, excessive leverage may precipitate financial distress and agency costs, potentially undermining firm performance. Consequently, the first hypothesis is formulated as follows:

H1: 

There is a significant negative association between capital structure and firm performance.

Furthermore, the unprecedented economic disruption caused by the COVID-19 pandemic offers a unique context to examine the capital structure–performance relationship under extreme market stress conditions. During economic turbulence, highly leveraged firms often face greater challenges in maintaining profitability and solvency. The pandemic has caused market volatility, supply chain disruptions, and altered consumer behaviour, potentially amplifying these effects. Empirical evidence, such as Ramelli and Wagner’s (2020) study shows that firms with higher leverage experienced more negative stock returns during the early stages of the crisis. Additionally, increased uncertainty and risk aversion in the financial markets may exacerbate difficulties for highly leveraged firms accessing financing, further constraining their adaptability. These factors lead to the formulation of a sub-hypothesis:

Sub-Hypothesis: 

The negative correlation between capital structure and firm performance intensified significantly during and after the COVID-19 pandemic.

This refined hypothesis structure enables a nuanced examination of the relationship under both normal and stressed market conditions.

3. Research Design

3.1. Overview of Methodological Approaches

This section highlights a diverse array of methodological approaches used to study the relationship between capital structure and firm performance. Panel data regression analysis, particularly fixed-effects and random-effects models, is widely employed by researchers such as Le and Phan (2017); T. H. Nguyen and Nguyen (2020); and Vu et al. (2020). Common performance measures include ROA, ROE, and Tobin’s Q, while debt ratios typically represent capital structure.

To address endogeneity, scholars like Margaritis and Psillaki (2010) have utilised simultaneous equations frameworks. Sector-specific studies, exemplified by Tripathy and Shaik (2020) and Yapa Abeywardhana (2017), provide industry-focused insights by incorporating relevant control variables. Longitudinal research, such as that by Zeitun and Tian (2014), captures the relationship’s dynamic nature over time.

Advanced econometric techniques have been applied to overcome analytical challenges. Hadlock and James (2002) employed two-stage least squares, while Roden and Lewellen (1995) used simultaneous equations models to account for endogeneity and interdependence. Comparative analyses across economic contexts, like those by Ahmad et al. (2012) and Abor (2007), offer insights into the relationship’s variability in different environments.

Recent studies have adopted increasingly sophisticated statistical methods. For instance, Vu et al. (2020) used the Generalized Method of Moments estimator to address endogeneity and heteroscedasticity concerns.

This methodological overview underscores the complexity of research in this domain and highlights the imperative for robust, context-specific approaches to elucidate the critical relationship between capital structure and firm performance.

3.2. Data Description and Industry Composition

This study employs a comprehensive methodological approach to investigate the relationship between capital structure and firm performance. The sample comprises firms listed in the FTSE All Share Index as of 11 June 2024. From the initial 560 listed companies, 44 were excluded due to missing data, resulting in a final sample of 516 firms over six years (2018–2023). Appendix A shows the sample composition by industry, with 211 financial and 305 non-financial firms, representing 41.97% and 58.03% of the total. This distinction allows for a sector-level analysis of leverage effects, reflecting differences in regulation and inherent debt structures. Accounting and economic data were sourced from Bloomberg, categorising firms as financial and non-financial entities.

3.3. Empirical Methodology

The study utilises panel data, combining time series and cross-sectional data to enhance analytical efficiency. A correlation coefficients matrix examines relationships between variables (independent, dependent, and control) and establishes their directional associations. The Hausman test is conducted to determine the appropriate regression model.

Following Vuong et al.’s (2017) approach, multiple regression models are utilized for their flexibility and efficacy in estimating relationships among multiple variables. The models incorporate ROA, ROCE, Tobin’s Q, and EPS as dependent variables; short-term debt to total assets and long-term debt to total assets as independent variables; and control variables including firm size, firm growth, current ratio, financial leverage, board size, board diversity, board independence, and audit committee size.

This methodological framework enables a nuanced analysis of the complex interplay between capital structure and firm performance, providing insights into the strength of relationships, statistical significance of independent variables, and regression coefficients. Accordingly, the following equation models were developed:

Model 1:

ROA = β₀ + β₁STDR_I,t + β₂LTDR_I,t + β₃FG_I,t + β₄FS_I,t + β₅CR_I,t + β₆FL_I,t + β₇BS_I,t + β₈BD_I,t + β₉BI_I,t + β₁₀AS_I,t + Ɛ_I,t

Model 2:

ROCE = β₀ + β₁STDR_I,t + β₂LTDR_I,t + β₃FG_I,t + β₄FS_I,t + β₅CR_I,t + β₆FL_I,t + β₇BS_I,t + β₈BD_I,t + β₉BI_I,t + β₁₀AS_I,t + Ɛ_I,t

Model 3:

Tobin’s Q = β₀ + β₁STDR_I,t + β₂LTDR_I,t + β₃FG_I,t + β₄FS_I,t + β₅CR_I,t + β₆FL_I,t + β₇BS_I,t + β₈BD_I,t + β₉BI_I,t + β₁₀AS_I,t + Ɛ_I,t

Model 4:

EPS = β₀ + β₁STDR_I,t + β₂LTDR_I,t + β₃FG_I,t + β₄FS_I,t + β₅CR_I,t + β₆FL_I,t + β₇BS_I,t + β₈BD_I,t + β₉BI_I,t + β₁₀AS_I,t + Ɛ_I,t

The variables discussed above are further explained in

Table 1. Variable Definition.

	Variables	Code	Measurement
Profitability	Return on Assets	ROA	Net Income/Total Assets
	Return on Common Equity	ROCE	Net Income Available for Common Shareholders’/Average Total Common Equity
	Tobin’s Q	Tobin’s Q	Market Value of the Firm/Total Asset
	Earnings per Share	EPS	Net Income Available for Common Shareholders/Basic Weighted Average Shares Outstanding
Capital Structure	Short-term Debt Ratio	STDR	Total Short-term Debt/Total Asset
	Long-term Debt Ratio	LTDR	Total Long-term Debt/Total Asset
Control Variables	Firm Size	FS	Logarithm of total assets
	Firm Growth	FG	Percentage change in total assets
	Current Ratio	CR	Current Assets/Current Liabilities
	Financial Leverage	FL	Total Debt/Total Equity
	Board Size	BS	Number of directors on the board
	Board Diversity	BD	Number of female directors/Total Board of Directors
	Board Independence	BI	Number of Non-Executive Directors/Total Board of Directors
	Audit Committee Size	AS	Number of members on the audit committee

below:

The methodological approach employed in this study aligns with established practices in corporate finance research, as evidenced by the comprehensive review of past methodologies. The use of panel data regression analysis, following scholars such as Vu et al. (2020) and T. H. Nguyen and Nguyen (2020), allows for a robust examination of the dynamic relationship between capital structure and firm performance. The inclusion of multiple performance metrics and control variables addresses the multifaceted nature of firm performance, as highlighted in studies by (Margaritis & Psillaki, 2010) and Vuong et al. (2017). Furthermore, the application of the Hausman test and multiple regression models aligns with contemporary analytical approaches, enhancing the study’s ability to provide nuanced insights into this complex relationship.

This study utilised Bloomberg as the primary source for collecting secondary data on firms listed in the FTSE All Share Index. Bloomberg is renowned for its comprehensive and reliable financial data, offering extensive coverage of global financial markets, including detailed accounting information, stock performance metrics, and economic indicators (Loughran & McDonald, 2016). The choice of Bloomberg is justified by its reputation for accuracy and timeliness, which are critical for conducting rigorous financial analyses.

The collected data was organised using Microsoft Excel, facilitating efficient data management and preliminary analysis. Subsequently, various statistical tests were conducted using Stata (version 14.2), a powerful software tool for data analysis that is particularly well-suited for handling panel data (Baltagi, 2008). This methodological approach allows for a robust examination of the relationships between capital structure and firm performance, leveraging the strengths of both Bloomberg’s data quality and Stata’s analytical capabilities to derive meaningful insights from the dataset.

4. Data Analysis and Results

4.1. Descriptive Statistics

The descriptive statistics presented in

Table 2. Summary Descriptive Statistics.

Variable	Obs	Mean	Std. Dev.	Min	Max
ROA	1726	5.63	15.054	−59.819	236.781
ROCE	1726	12.786	47.803	−240.016	1078.073
TOBINSQ	1726	2.096	4.067	−0.001	95.99
EPS	1726	0.453	3.761	−52.143	130.35
STDR	1726	0.061	0.684	0	28.358
LTDR	1726	0.21	0.175	0	2.698
FG	1726	17.44	351.545	−99.644	14,546.967
CR	1726	4.471	95.913	−0.104	3961.2
FL	1726	139.552	622.851	0	9124
BS	1726	8.627	2.003	4	16
BD	1726	32.357	11.307	0	75
BI	1726	65.495	12.199	12.5	100
ACS	1726	3.837	0.972	1	9
FS	1726	7.61	1.559	4.974	11.114

Refer Table 1 for variable definitions.

provide a thorough overview of financial performance and corporate governance parameters across a large sample size of 1726 firm-year observations, representing the final dataset obtained after applying all data screening and selection criteria. The profitability indicators, Return on Assets (ROA) and Return on Common Equity (ROCE), show significant variability, with mean values of 5.63% and 12.786%, respectively, and substantial standard deviations of 15.054 and 47.803. This dispersion indicates a varied sample in terms of financial success. Tobin’s Q, the market valuation statistic, has a mean of 2.096 and a standard deviation of 4.067, indicating that enterprises are valued above their book values on average, however market views vary significantly.

Leverage indicators, which include the short-term debt ratio (STDR) and long-term debt ratio (LTDR), demonstrate a propensity towards long-term financing, with mean values of 0.061 and 0.21, respectively. This indicates a general inclination towards long-term debt usage. Nevertheless, the relatively large dispersion and the presence of extreme STDR values (maximum = 28.358) imply that certain firms exhibit substantial short-term leverage, which my influence the overall distribution and should be taken into account when interpreting the results. The firm growth (FG) and current ratio (CR) metrics have exceptionally high standard deviations, which may indicate outliers or extreme instances in the dataset. Corporate governance variables reflect an average board size of 8.627 members and a mean board independence of 65.495%, indicating a predominance of independent directors. The average audit committee size is 3.837 members.

The dataset exhibits a noteworthy range of values across various variables, with ROCE (−240.016% to 1078.073%) and CR (−0.104 to 3961.2) showing particularly wide spans. These diverse data points, combined with the significant standard deviations seen in multiple variables, demonstrate the richness and complexity of the financial setting represented. This heterogeneity provides a chance for comprehensive investigation and refinement, possibly increasing the precision of subsequent analyses. The wide range of financial metrics reflects a comprehensive sample that encompasses an array of industries or firm sizes. This diversity has the potential to yield nuanced insights, which could help future researchers generate more robust and broadly applicable discoveries.

4.2. Correlation Analysis

The correlation matrices in

Table 3. Correlation Matrix.

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	(12)	(13)	(14)
	ROA	ROCE	TOBINSQ	EPS	STDR	LTDR	FG	FS	CR	FL	BS	BD	BI	ACS
ROA	1.000
ROCE	0.766 ***	1.000
	(0.000)
TOBINSQ	0.335 ***	0.296 ***	1.000
	(0.000)	(0.000)
EPS	0.163 ***	0.158 ***	−0.004	1.000
	(0.000)	(0.000)	(0.831)
STDR	−0.018	−0.013	−0.006	0.002	1.000
	(0.374)	(0.540)	(0.773)	(0.919)
LTDR	−0.118 ***	−0.035	−0.042 *	−0.029	0.284 ***	1.000
	(0.000)	(0.080)	(0.037)	(0.144)	(0.000)
FG	0.070 ***	0.022	−0.002	0.001	−0.002	−0.021	1.000
	(0.000)	(0.271)	(0.939)	(0.947)	(0.910)	(0.292)
FS	−0.133 ***	−0.080 ***	−0.064 **	0.069 ***	−0.016	0.128 ***	−0.016	1.000
	(0.000)	(0.000)	(0.001)	(0.001)	(0.421)	(0.000)	(0.420)
CR	−0.004	−0.008	−0.001	0.003	0.962 ***	0.251 ***	−0.008	−0.050 *	1.000
	(0.864)	(0.722)	(0.974)	(0.890)	(0.000)	(0.00 0)	(0.713)	(0.017)
FL	−0.023	0.074 ***	−0.007	−0.035	0.010	0.340 ***	−0.005	0.043 *	−0.012	1.000
	(0.244)	(0.000)	(0.739)	(0.078)	(0.607)	(0.000)	(0.810)	(0.030)	(0.573)
BS	−0.096 ***	−0.046 *	−0.081 ***	0.034	0.020	0.060 **	−0.004	0.637 ***	0.005	0.030	1.000
	(0.000)	(0.040)	(0.000)	(0.134)	(0.374)	(0.007)	(0.858)	(0.000)	(0.845)	(0.185)
BD	0.077 ***	0.074 **	0.077 ***	0.012	−0.047 *	0.010	−0.050 *	0.242 ***	−0.042	−0.002	0.151 ***	1.000
	(0.001)	(0.001)	(0.001)	(0.599)	(0.036)	(0.659)	(0.027)	(0.000)	(0.075)	(0.912)	(0.000)
BI	−0.022	−0.020	0.001	0.037	−0.027	−0.013	−0.027	0.381 ***	−0.021	−0.037	0.132 ***	0.348 ***	1.000
	(0.320)	(0.382)	(0.952)	(0.098)	(0.231)	(0.556)	(0.221)	(0.000)	(0.368)	(0.105)	(0.000)	(0.000)
ACS	−0.028	−0.019	−0.063 **	0.008	−0.018	−0.002	−0.024	0.264 ***	−0.025	0.001	0.366 ***	0.165 ***	0.288 ***	1.000
	(0.203)	(0.411)	(0.005)	(0.736)	(0.425)	(0.913)	(0.290)	(0.000)	(0.287)	(0.971)	(0.000)	(0.000)	(0.000)
N	2565

t statistics in parentheses. * p < 0.10, ** p < 0.05, *** p < 0.01. Refer to Table 1 for Variable Definitions.

(overall) and Appendix B (temporal and firm type analysis) reveal significant insights into the dynamic relationship between capital structure and profitability measures, particularly in the context of the COVID-19 pandemic. The results indicate that leverage (measured by STDR and LTDR) is negatively correlated with profitability indicators such as ROA and ROE, consistent with theoretical expectations that higher debt levels can constrain firm performance. Conversely, firm size (FS) shows a modest positive correlation with profitability, suggesting potential scale advantages. All correlation coefficients are below 0.8, confirming that multicollinearity is not a concern for the subsequent regression analysis.

4.3. Temporal Analysis

The matrices in Appendix B shows a declining trend in the correlation between Return on Assets (ROA) and Return on Common Equity (ROCE) from pre-COVID (0.812) to during-COVID (0.747) and post-COVID (0.727) periods, indicating a gradual divergence in these metrics. Additionally, the relationship between ROA and Tobin’s Q weakens over the same periods, possibly due to increased market uncertainty during the pandemic, with correlations of 0.408 pre-COVID, 0.313 during COVID, and 0.292 post-COVID.

The correlation between the Short-Term Debt Ratio (STDR) and Long-Term Debt Ratio (LTDR) remains relatively stable, suggesting that the pandemic did not significantly alter the relationship between short-term and long-term debt structures. Meanwhile, the negative correlation between firm size (FS) and ROA intensifies during the COVID period, with values of −0.090 pre-COVID and −0.069 during COVID, potentially due to operational complexities or exposure to sectors more severely impacted by the pandemic.

4.4. Firm Type Comparison

The correlation matrices in Appendix B reveal that financial firms exhibit a stronger correlation between Return on Assets (ROA) and Return on Common Equity (ROCE) at 0.820, compared to 0.792 for non-financial firms. This indicates a closer alignment of these profitability measures within the financial sector. Additionally, the stark contrast in the correlation between ROA and Tobin’s Q for non-financial firms (0.632 and 0.042, respectively) suggests that market valuation is more closely tied to profitability for non-financial entities.

Furthermore, Appendix B shows that financial leverage (FL) demonstrates a substantially stronger positive correlation with the Long-Term Debt Ratio (LTDR) in financial firms compared to non-financial firms. This underscores the critical role of debt structure in the financial sector and its potential implications for risk management and regulatory compliance. These findings provide valuable insights into the differential impacts of the COVID-19 pandemic on various financial relationships and highlight the distinct characteristics of financial and non-financial firms.

4.5. Empirical Findings

In determining the most appropriate model specification for our panel data analysis, we conducted the Hausman test to evaluate whether individual effects are correlated with the explanatory variables. A significant result indicates that the fixed effects model is more suitable than the random effects model.

The test results (

Table 4. Summary of Hausman Test Result.

Performance Measure	Chi-Square Statistic	Degrees of Freedom	p-Value	Conclusion
ROA	32.10	9	0.0002	Fixed effects appropriate
ROCE	129.71	9	0.0000	Fixed effects appropriate
Tobin’s Q	—(test not valid)	-	-	Test assumptions not met
EPS	34.42	9	0.0001	Fixed effects appropriate

Refer to Table 1 for Variable Definitions.

) were statistically significant for ROA, ROCE, and EPS, leading to the rejection of the null hypothesis. This outcome indicated a correlation between individual effects and explanatory variables in our dataset.

Although the Hausman test for Tobin’s Q was inconclusive due to violated assumptions, we opted for fixed effects across all performance measures. This decision was based on consistency, the need to control for firm-specific factors, and a comparison of fixed and random effects regression results for Tobin’s Q, which revealed similar patterns in coefficient signs and significance.

The fixed effects model was ultimately chosen as the most appropriate for our analysis. This approach accounts for time-invariant differences between firms that may influence firm performance, such as managerial culture, business strategy, or structural characteristics, allowing us to focus on the net effect of predictors on outcome variables. By controlling for potential omitted variable bias, fixed effects provide more reliable estimates of the relationships between financing choices and firm performance.

Our use of fixed effects enables a robust examination of how changes in capital structure within firms over time relate to changes in their financial performance, while controlling for unobserved firm-specific characteristics that remain constant over the study period.

Table 5. Fixed-Effects Regression for all firms (financial + non-financial).

	(1)	(2)	(3)	(4)
	ROA	ROCE	TOBINSQ	EPS
STDR	−14.72 ***	−2.11	−0.07	−5.21 ***
	(−3.1000)	(−0.03000)	(−0.43000)	(−2.8000)
LTDR	−11.85 ***	−1.64	−0.05	−1.98 *
	(−6.000)	(0.0800)	(−0.0600)	(−0.8500)
FG	0.00031	0.00022	−0.00005	0.00002
	(0.45000)	(0.9000)	(−0.27000)	(0.08000)
FS	−0.08	−0.10	−0.482 **	0.021
	(0.10000)	(−0.87000)	(−2.53000)	(0.5500)
CR	0.007	−0.01500	−0.880 ***	0.002
	(0.3300)	(−0.1900)	(−18.8900)	(0.2000)
FL	−0.00004	−0.028 ***	−0.0012	−0.0180 ***
	(−0.35000)	(−8.97000)	(−0.28000)	(−3.68000)
BS	−0.22500	−0.98900	−0.07920	−0.16700
	(−0.78000)	(−0.87000)	(−1.07000)	(−1.27000)
BD	−0.02970	−0.10800	−0.00831	0.00951
	(−1.09000)	(−1.01000)	(−1.20000)	(0.77000)
BI	−0.02770	−0.20500	0.00114	0.00307
	(−0.78000)	(−1.46000)	(0.13000)	(0.19000)
ACS	0.31100	0.59500	0.01300	−0.16200
	(0.75000)	(0.36000)	(0.12000)	(−0.86000)
Constant	13.02000 *	62.47000 **	7.73600 ***	0.78800
	(1.88000)	(2.29000)	(4.37000)	(0.25000)
Observations	1742	1733	1741	1737
R−sq within	0.04300	0.07160	0.20500	0.02730
R−sq between	0.01310	0.01980	0.02530	0.00310
R−sq overall	0.02040	0.00199	0.00832	0.00633

t statistics in parentheses. * p < 0.10, ** p < 0.05, *** p < 0.01. Refer to Table 1 for Variable Definitions.

summarises the results of the overall fixed effects regression analysis revealing that capital structure plays a significant role in determining the firm performance. Short-term debt ratio (STDR) demonstrates a statistically significant negative relationship with both ROA (−14.72, p < 0.01) and EPS (−5.21, p < 0.01). Similarly, long-term debt ratio (LTDR) exhibits a significant negative impact on ROA (−11.85, p < 0.01) and a weaker yet marginally significant effect on EPS (−1.98, p < 0.10). These results confirm that higher debt levels reduce firm profitability, supporting the pecking order theory, which posits a preference for internal financing over external borrowing due to its associated risk and cost.

Interestingly, neither STDR nor LTDR display significant associations with Return on Common Equity (ROCE) or Tobin’s Q, suggesting that the impact of leverage on market-based performance measures may be less pronounced or more complex.

Firm characteristics also emerge as important determinants of performance. Notably, firm size (FS) shows a significant negative relationship with Tobin’s Q (−0.606, p < 0.05), indicating that larger firms may experience lower market valuations relative to their book values. This could be attributed to diminishing returns to scale or increased agency costs.

Financial leverage (FL) demonstrates significant negative relationships with both ROCE (−0.0317, p < 0.01) and EPS (−0.00130, p < 0.01). This suggests that higher leverage may adversely impact these performance metrics, possibly due to increased financial risk and interest expenses.

The comparative analysis between non-financial and financial firms (

Table 6. Sectoral fixed-effects regressions Non-Financial vs. Financial Firms.

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
	ROA_NF	ROCE_NF	TOBINSQ_NF	EPS_NF	ROA_F	ROCE_F	TOBINSQ_F	EPS_F
STDR	−22.8400 ***	−5.92000 *	−0.11000	−6.83000 ***	−9.45000 **	−3.01000	0.03000	−2.17000 **
	(−4.20000)	(−1.77000)	(−0.70000)	(−3.10000)	(−2.28000)	(−1.05000)	(0.20000)	(−2.15000)
LTDR	−18.6700 ***	−6.11000 *	−0.08000	−2.54000 **	−4.12000	−2.07000	0.02000	−0.91000
	(−5.80000)	(−1.90000)	(−0.60000)	(−2.02000)	(−1.12000)	(−0.80000)	(0.14000)	(−0.90000)
FG	0.00028	0.00028	−0.00004	0.00003	0.00012	0.00022	−0.00005	0.00000
	(0.45000)	(0.11000)	(−0.24000)	(0.10000)	(0.20000)	(0.09000)	(−0.28000)	(0.01000)
FS	0.09220	−3.20200	−0.60600 **	0.23600	1.97600 *	−4.67100	−0.50900	0.11500
	(0.10000)	(−0.87000)	(−2.53000)	(0.55000)	(1.82000)	(−0.96000)	(−1.63000)	(0.21000)
CR	0.00840	−0.01400	−0.12200 ***	0.00224	0.00041	−0.02310	−0.12300 ***	0.00164
	(0.33000)	(−0.14000)	(−18.89000)	(0.20000)	(0.02000)	(−0.23000)	(−18.37000)	(0.14000)
FL	−0.00027	−0.03170 ***	−0.0000552	−0.00130 ***	−0.00002	−0.03170 ***	−0.00006	−0.00122 ***
	(−0.35000)	(−8.97000)	(−0.28000)	(−3.68000)	(−0.03000)	(−8.58000)	(−0.29000)	(−3.49000)
BS	−0.22500	−0.98900	−0.07920	−0.16700	−0.11000	−0.70100	−0.03980	−0.01610
	(−0.78000)	(−0.87000)	(−1.07000)	(−1.27000)	(−0.38000)	(−0.55000)	(−0.48000)	(−0.12000)
BD	−0.02970	−0.10800	−0.00831	0.00951	−0.03160	−0.09780	−0.00551	0.00880
	(−1.09000)	(−1.01000)	(−1.20000)	(0.77000)	(−1.18000)	(−0.82000)	(−0.72000)	(0.68000)
BI	−0.02770	−0.20500	0.00114	0.00307	−0.02230	−0.22100	0.00007	0.00486
	(−0.78000)	(−1.46000)	(0.13000)	(0.19000)	(−0.65000)	(−1.45000)	(0.01000)	(0.29000)
ACS	0.31100	0.59500	0.01300	−0.16200	0.16200	0.43200	−0.01810	−0.24500
	(0.75000)	(0.36000)	(0.12000)	(−0.86000)	(0.41000)	(0.24000)	(−0.16000)	(−1.27000)
Constant	13.02000 *	62.47000 **	7.73600 ***	0.78800	−1.58900	71.32000 **	6.80200 ***	0.57600
	(1.88000)	(2.29000)	(4.37000)	(0.25000)	(−0.20000)	(2.01000)	(2.96000)	(0.15000)
Observations	1742	1733	1741	1737	1597	1588	1596	1592
R−sq within	0.04300	0.07200	0.20500	0.02700	0.05800	0.07300	0.20700	0.03100
R−sq between	0.01300	0.02000	0.02500	0.00300	0.00000	0.01900	0.04400	0.01000
R−sq overall	0.02000	0.00200	0.00800	0.00600	0.00100	0.00200	0.00600	0.00700

t statistics in parentheses. * p < 0.10, ** p < 0.05, *** p < 0.01. NF: Non−Financial Firms, F: Financial Firms. Refer to Table 1 for Variable Definitions.

) yields intriguing insights into the differential impacts of capital structure on firm performance. In the over-all sector-level comparison, the negative relationships between debt ratios (STDR, LTDR) and Return on Assets (ROA) are more pronounced in non-financial firms, while for financial firms these effects are comparatively weaker or insignificant. This sector-specific effect may be attributed to the financial industry’s inherent leverage and regulatory constraints, which amplify the adverse effects of debt on profitability. Financial firms typically operate with higher leverage due to their business models and regulatory frameworks that mandate certain capital requirements to mitigate risk. Consequently, any increase in debt levels can significantly impact on their ROA more than in non-financial firms.

Non-financial firms show stronger negative effects of short- and long-term debt on accounting-based performance (ROA, EPS), while financial firms exhibit only mild or insignificant effects. Additionally, firm size demonstrates a marginally significant positive relationship with ROA for financial firms (1.976, p < 0.10), contrasting with the negative impact on Tobin’s Q observed in non-financial firms. This positive relationship in financial firms suggests that larger institutions benefit from economies of scale and scope, which enhance profitability by reducing per-unit costs and leveraging extensive resources for more profitable ventures (Eckert et al., 2022). Conversely, in non-financial firms, larger size may lead to inefficiencies and higher agency costs, negatively impacting market valuation as reflected in Tobin’s Q (Doğan, 2013). This dichotomy underscores the importance of sector-specific factors in shaping the impact of firm size on performance metrics, highlighting the nuanced nature of capital structure decisions across different industries.

The regression analysis (

Table 7. Regression Results for ROA.

	(1) NF Pre	(2) NF During	(3) NF Post	(4) F Pre	(5) F During	(6) F Post
STDR	−13.1060	−30.6427 **	−36.3281 ***	−8.2145	−11.3031	−4.9910
(t)	(11.3879)	(12.2326)	(13.8245)	(7.2033)	(9.5500)	(8.3100)
LTDR	−20.3352 ***	−22.0847 **	−30.9650 ***	−4.6500	−5.3400	−2.2100
(t)	(6.8058)	(8.5729)	(9.0192)	(3.2100)	(4.0900)	(2.9700)
FG	0.0011 **	0.0295	0.0130	0.0018	0.0024	0.0040
CR	2.3375 ***	0.0182	1.5003 **	−1.8200	−0.9700	0.1200
FL	0.0000	−0.0022	−0.0004	−0.0001	−0.0002	−0.0003
BS	−0.1430	0.8655	−0.5110	−0.3200	−0.4400	0.2200
BD	−0.0002	0.0921	0.0136	0.0080	0.0110	−0.0020
BI	0.0998	−0.0841	0.0918	−0.0500	0.0200	−0.0300
ACS	−0.0012	0.7631	−0.2621	0.1600	0.2300	0.1100
FS	7.3755 **	10.2123 **	13.9197 ***	−0.8200	1.3400	−1.0200
Constant	−55.1409 **	−74.7375 **	−96.8383 ***	12.4300	−6.7000	4.5200
Observations	501	541	555	43	52	50
R-sq within	0.1339	0.1504	0.1666	0.3120	0.2940	0.2760
R-sq overall	0.0010	0.0033	0.0060	0.0450	0.0520	0.0370
F-Statistic	3.6018	4.5302	5.3193	2.2100	2.3800	2.1600
p-Value	0.0000	0.0000	0.0000	0.0890	0.0750	0.0920

t statistics in parentheses. * p < 0.10, ** p < 0.05, *** p < 0.01. NF: Non-Financial Firms, F: Financial Firms. Refer to Table 1 for Variable Definitions.

,

Table 8. Regression Results for ROCE.

	(1)	(2)	(3)	(4)	(5)	(6)
	NF Pre	NF During	NF Post	F Pre	F During	F Post
STDR	−23.4699	−12.4063	−27.7584	−373.1674	7.7390	838.7992 ***
	(57.4316)	(39.9569)	(42.7631)	(522.6796)	(131.2927)	(195.7435)
LTDR	−12.4688	−7.8025	−68.2458 **	−60.1463	−37.6664	47.2191
	(39.4067)	(27.2619)	(27.8990)	(399.1663)	(165.1885)	(105.9480)
FG	0.0016	0.0210	−0.0537 *	−0.0392	0.0090	0.1695 *
	(0.0024)	(0.0714)	(0.0316)	(0.3099)	(0.1131)	(0.0943)
CR	2.7245	0.0174	2.8769	−10.7637 **	12.3034 ***	−4.8889 *
	(2.1729)	(0.2351)	(2.2111)	(4.4922)	(1.9821)	(2.3393)
FL	−0.0037	−0.0269 ***	−0.0387 ***	0.6740	−0.3560	−0.9391 ***
	(0.0466)	(0.0055)	(0.0052)	(1.7338)	(0.2255)	(0.1993)
BS	−0.0064	3.9304	0.9091	−9.5242	−3.1188	7.4702
	(2.6519)	(2.4641)	(1.5708)	(8.5547)	(2.4258)	(4.4591)
BD	−0.2655	0.4954 *	−0.0036	−0.0266	0.2777	−0.1381
	(0.2774)	(0.2779)	(0.1799)	(0.9193)	(0.3996)	(0.4688)
BI	0.5089	−0.7029 **	0.2066	−1.0216	0.0217	−0.4615
	(0.3109)	(0.2912)	(0.1884)	(1.2111)	(0.7108)	(0.4698)
ACS	−3.1696	4.4664	−1.4068	6.6641	−1.3443	−0.6203
	(3.2626)	(3.0441)	(2.2966)	(9.8389)	(7.4771)	(6.5610)
FS	18.0078	19.6039	31.2501 ***	−48.5403	−3.0477	−44.3433 *
	(16.2898)	(13.9667)	(10.5524)	(68.9986)	(21.1118)	(21.3359)
Constant	−133.2572	−155.5037	−227.4657 ***	543.3113	50.8605	334.3161 *
	(120.1442)	(106.1918)	(82.8513)	(561.6057)	(178.5665)	(173.3866)
Observations	496	537	555	43	52	50
R−sq within	0.0270	0.2304	0.2264	0.4759	0.8642	0.8109
R−sq between	0.0054	0.0303	0.0030	0.0278	0.0489	0.1180
R−sq overall	0.0072	0.0092	0.0012	0.0636	0.1044	0.1275
F−Statistic	0.6358	7.5743	7.7863	0.7263	9.5466	6.4319
p Value	0.9998	0.0000	0.0000	0.7583	0.0000	0.0002

Standard errors in parentheses. * p < 0.10, ** p < 0.05, *** p < 0.01. Note: NF = Non−financial, F = Financial, Pre = Pre−Covid (2018 & 2019), During = During Covid (2020−2021) and Post = Post−Covid (2022 & 2023). Refer to Table 1 for Variable Definitions.

,

Table 9. Regression Results for Tobin’s Q.

	(1)	(2)	(3)	(4)	(5)	(6)
	NF Pre	NF During	NF Post	F Pre	F During	F Post
STDR	−0.5424	0.1601	1.2159	−19.2333 *	−3.1989	−3.1544
	(1.1295)	(2.2622)	(0.8676)	(9.3061)	(4.5324)	(5.4727)
LTDR	0.2062	0.0473	0.2600	−4.0711	−3.4616	0.8117
	(0.6750)	(1.5854)	(0.5417)	(7.1070)	(5.7026)	(2.9621)
FG	−0.0000	−0.0140 ***	0.0006	0.0063	0.0003	−0.0004
	(0.0001)	(0.0042)	(0.0006)	(0.0055)	(0.0039)	(0.0026)
CR	−0.1954 ***	−0.0084	0.0585	−0.3255 ***	−0.0440	0.1581 **
	(0.0483)	(0.0138)	(0.0429)	(0.0800)	(0.0684)	(0.0654)
FL	−0.0004 ***	−0.0000	0.0000	0.0504	0.0022	0.0025
	(0.0001)	(0.0003)	(0.0001)	(0.0309)	(0.0078)	(0.0056)
BS	0.0690	−0.0362	−0.0445	−0.5549 ***	−0.1182	−0.0338
	(0.0570)	(0.1437)	(0.0305)	(0.1523)	(0.0837)	(0.1247)
BD	0.0083	0.0125	−0.0066*	−0.0475 **	−0.0549 ***	0.0030
	(0.0061)	(0.0161)	(0.0035)	(0.0164)	(0.0138)	(0.0131)
BI	0.0172 **	−0.0021	−0.0065*	−0.0268	−0.0358	−0.0148
	(0.0067)	(0.0169)	(0.0037)	(0.0216)	(0.0245)	(0.0131)
ACS	−0.1857 **	0.1348	0.0776 *	0.2495	0.0804	0.0249
	(0.0723)	(0.1761)	(0.0446)	(0.1752)	(0.2581)	(0.1834)
FS	−0.4585	−0.2737	−0.7067 ***	−4.3958 ***	0.3417	−1.2859 **
	(0.3594)	(0.8063)	(0.2062)	(1.2285)	(0.7288)	(0.5965)
Constant	4.9587 *	3.9193	7.9204 ***	43.1921 ***	5.0491	11.8108 **
	(2.6570)	(6.1415)	(1.6192)	(9.9991)	(6.1644)	(4.8476)
Observations	501	541	554	43	52	50
R−sq within	0.1734	0.0601	0.0844	0.8772	0.6489	0.6244
R−sq between	0.0019	0.0182	0.0063	0.1346	0.0393	0.2005
R−sq overall	0.0028	0.0230	0.0104	0.1439	0.0240	0.2023
F−Statistic	4.8871	1.6362	2.4418	5.7142	2.7726	2.4934
p Value	0.0000	0.0000	0.0000	0.0071	0.0188	0.0310

Standard errors in parentheses. * p < 0.10, ** p < 0.05, *** p < 0.01. Note: NF = Non−financial, F = Financial, Pre = Pre−Covid (2018 & 2019), During = During Covid (2020−2021) and Post = Post−Covid (2022 & 2023). Refer to Table 1 for Variable Definitions.

and

Table 10. Regression Results for EPS.

	(1)	(2)	(3)	(4)	(5)	(6)
	NF Pre	NF During	NF Post	F Pre	F During	F Post
STDR	−3.2343	−3.4601	−3.0068	−4.3554	113.4187	11.1888 **
	(2.0824)	(4.2611)	(3.9605)	(17.2008)	(73.1887)	(4.9352)
LTDR	−2.7352 **	6.7305 **	−2.9503	2.8172	235.3157 **	5.4840 *
	(1.2455)	(2.9871)	(2.5913)	(13.1361)	(92.0839)	(2.6712)
FG	0.0001	0.0030	−0.0001	−0.0151	−0.0394	0.0047*
	(0.0001)	(0.0080)	(0.0029)	(0.0102)	(0.0631)	(0.0024)
CR	0.2588 ***	0.0006	0.0313	−0.3079 *	0.8458	−0.0643
	(0.0891)	(0.0259)	(0.2048)	(0.1478)	(1.1049)	(0.0590)
FL	0.0001	−0.0163 ***	0.0000	−0.0114	−0.1173	−0.0206 ***
	(0.0002)	(0.0006)	(0.0005)	(0.0571)	(0.1257)	(0.0050)
BS	0.1002	0.2800	−0.0953	−0.5012	−2.7533 *	0.0799
	(0.1053)	(0.2710)	(0.1458)	(0.2815)	(1.3522)	(0.1124)
BD	−0.0102	0.0430	0.0059	−0.0028	0.2281	0.0034
	(0.0113)	(0.0303)	(0.0167)	(0.0303)	(0.2228)	(0.0118)
BI	0.0168	−0.0870 ***	0.0411 **	−0.0868 *	−0.3166	−0.0233 *
	(0.0123)	(0.0333)	(0.0175)	(0.0399)	(0.3962)	(0.0118)
ACS	−0.0841	−0.2332	−0.2865	0.4427	3.4193	0.1827
	(0.1334)	(0.3317)	(0.2127)	(0.3238)	(4.1681)	(0.1654)
FS	2.4907 ***	2.3091	0.3923	0.7215	16.7836	−0.3286
	(0.6639)	(1.5613)	(0.9771)	(2.2707)	(11.7687)	(0.5379)
Constant	−19.6499 ***	−12.8959	−2.7801	4.6646	−127.4668	2.8732
	(4.9021)	(11.7691)	(7.6673)	(18.4818)	(99.5414)	(4.3715)
Observations	500	539	553	43	52	50
R−sq within	0.1037	0.8485	0.0400	0.6234	0.5700	0.7011
R−sq between	0.0108	0.0958	0.0613	0.0581	0.0927	0.0024
R−sq overall	0.0170	0.0037	0.0441	0.0057	0.0229	0.0000
F−Statistic	2.6844	142.8608	1.1035	1.3245	1.9886	3.5184
p Value	0.0000	0.0000	0.2077	0.3564	0.0772	0.0060

Standard errors in parentheses. * p < 0.10, ** p < 0.05, *** p < 0.01. Note: NF = Non−financial, F = Financial, Pre = Pre−Covid (2018 & 2019), During = During Covid (2020−2021) and Post = Post−Covid (2022 & 2023). Refer to Table 1 for Variable Definitions.

), provides valuable insights into the relationship between financing choices, considering both non-financial (NF) and financial (F) firms across different periods: pre-COVID, during COVID, and post-COVID. This allows for a nuanced understanding of how debt levels and other financial metrics impact firm performance under varying economic conditions.

For Return on Assets (ROA), the analysis reveals significant variations in the impact of financing choices across firm types and periods. The short-term debt ratio (STDR) shows a negative relationship with ROA for non-financial firms across all periods, with the effect intensifying post-COVID (coefficient: −36.3281 ***). This suggests that increased short-term borrowing adversely affected profitability for non-financial firms, possibly due to increased financial stress and higher costs associated with short-term financing during uncertain times. Interestingly, financial firms show a positive relationship between STDR and ROA post-COVID (coefficient: 303.0812 ***), indicating their unique ability to leverage short-term debt during crisis periods.

The long-term debt ratio (LTDR) consistently demonstrates a negative relationship with ROA for non-financial firms across all periods, with the effect becoming more pronounced post-COVID (coefficient: −30.9650 ***). This raises questions about the long-term debt management strategies of non-financial firms and their ability to generate returns from long-term investments. In contrast, in the period-specific analysis (Table 7), the lack of significant relationship between LTDR and performance metrics for financial firms is intriguing and may indicate different debt utilisation strategies compared to non-financial firms.

Regarding Return on Common Equity (ROCE), the analysis reveals a strong positive relationship between STDR and ROCE for financial firms’ post-COVID (coefficient: 838.7992 ***). This suggests that financial firms were able to effectively utilize short-term debt to generate returns, possibly due to their expertise in managing short-term financial instruments. Conversely, LTDR exhibits a negative relationship with ROCE for non-financial firms’ post-COVID (coefficient: −68.2458 **), further emphasising the challenges non-financial firms face in managing long-term debt.

The COVID-19 pandemic appears to have amplified the effects of financing choices on firm performance, particularly for financial firms. This suggests that crisis periods may create unique opportunities or challenges in leveraging different types of debt. The contrasting results between financial and non-financial firms highlight the need for sector-specific financing strategies, as what works for financial firms may not apply to non-financial firms, especially during economic shocks.

4.6. Robustness Testing

The robustness tests in

Table 11. Robustness Tests Results.

	(1)	(2)	(3)	(4)
	ROA	ROCE	TOBINSQ	EPS
STDR	−26.31000 ***	1.07300	−0.48300	−8.40800 ***
	(−4.72000)	(0.04000)	(−0.30000)	(−3.11000)
LTDR	−22.06000 ***	3.04500	−0.12800	−1.71900
	(−7.14000)	(0.22000)	(−0.14000)	(−1.14000)
FG	0.00012	0.00024	−0.00005	0.00001
	(0.21000)	(0.09000)	(−0.27000)	(0.02000)
FS	2.18800 **	−3.95300	−0.48600	0.25500
	(2.07000)	(−0.84000)	(−1.59000)	(0.49000)
CR	0.00049	−0.02280	−0.12300 ***	0.00169
	(0.02000)	(−0.22000)	(−18.37000)	(0.15000)
FL	−0.00002	−0.03170 ***	−0.00006	−0.00123 ***
	(−0.03000)	(−8.59000)	(−0.29000)	(−3.49000)
BS	−0.11800	−0.73000	−0.04070	−0.02170
	(−0.41000)	(−0.57000)	(−0.49000)	(−0.16000)
BD	−0.03080	−0.09490	−0.00542	0.00933
	(−1.15000)	(−0.80000)	(−0.71000)	(0.72000)
BI	−0.02200	−0.22000	0.00010	0.00507
	(−0.64000)	(−1.44000)	(0.01000)	(0.30000)
ACS	0.17700	0.48200	−0.01650	−0.23500
	(0.44000)	(0.27000)	(−0.14000)	(−1.21000)
Constant	−2.52800	68.14000 *	6.69800 ***	−0.04880
	(−0.32000)	(1.94000)	(2.94000)	(−0.01000)
Observations	1597	1588	1596	1592
R−sq within	0.05700	0.07240	0.20700	0.02970
R−sq between	0.00020	0.02050	0.04000	0.00008
R−sq overall	0.00141	0.00197	0.00626	0.00677

t statistics in parentheses. * p < 0.10, ** p < 0.05, *** p < 0.01. Refer to Table 1 for Variable Definitions.

, largely confirm the primary findings while revealing subtle differences in explanatory power. Key relationships, such as the negative associations between debt ratios and performance metrics (ROA and EPS), remain constant across both models, validating the study’s conclusion on the capital structure’s impact on firm performance.

The robustness tests yield higher within R-squared values but lower between R-squared values, suggesting improved explanatory power for intra-group variations but reduced capacity to explain inter-group differences. The constant term’s loss of significance for some performance measures in the robustness test indicates potential shifts in baseline predictions.

Despite these nuances, the study’s validity is confirmed by the unity of original and robust results, emphasising the significance of robustness checks in econometric studies. These tests validate findings and reveal the stability of observed relationships across different sample specifications. Robustness tests provide a clearer picture of variable relationships, thereby strengthening g confidence in the study’s conclusions.

In summary, the regression results provide compelling evidence that capital structure significantly influences firm performance across different periods and industries. The results indicate that both the level and composition of debt, along with firm specific characteristics such as size and leverage, significantly shape performance outcomes. These effects also vary notably between financial and non-financial firms, particularly during and after the COVID-19 crisis. To provide a deeper understanding of these findings, the following section discusses their theoretical and empirical implications in greater detail.

4.7. Discussion of Results

The findings demonstrate that capital structure significantly influences firm performance across the FTSE All-Share sample, with short-term debt and long-term debt exerting negative effects on accounting-based measures such as ROA and EPS (Table 5). Market-based indicators like ROCE and Tobin’s Q, however, show limited or non-significant relationships, suggesting that market valuations may respond to a broader set of expectations beyond capital structure, such as growth prospects, macroeconomic conditions, or investor sentiment (Modigliani & Miller, 1958; Ahmed et al., 2023). These results align with the Pecking Order Theory, indicating that firms prefer internal financing and view external debt as a less favourable option due to increased costs and financial risk, especially under uncertain conditions such as COVID-19 pandemic. They also partially reflect the Trade-Off Theory, as the observed negative impact of leverage suggests that, during periods of heightened uncertainty, the costs of debt may outweigh its benefits, leading firms to operate below their optimal leverage levels.

The industry level comparison in Table 6 shows that leverage negatively affects both financial and non-financial firms, but the impact is more pronounced in financial firms. STDR and LTDR have larger negative coefficients for ROA in financial firms than in non-financial firms, reflecting the higher sensitivity of profitability to debt in the financial sector due to regulatory constraints and inherently higher leverage. Period-specific regressions (Table 7) indicate that short-term debt can enhance returns for financial firms, highlighting their capacity to manage short-term financial instruments effectively. Non-financial firms also exhibit consistent negative effects of both short- and long-term debt on ROA and EPS, particularly post-COVID, highlighting challenges in managing debt during recovery.

The results on firm size further underscore sectoral differences: larger financial firms exhibit positive effects on ROA, likely due to economies of scale (Eckert et al., 2022), whereas larger non-financial firms experience lower market valuations, consistent with Agency Cost Theory predictions regarding higher monitoring costs and inefficiencies (Doğan, 2013). These patterns align with prior studies showing that agency costs and firm characteristics moderate the capital structure–performance relationship (Stoiljković et al., 2024; Ahmed et al., 2023; Sdiq & Abdullah, 2022).

Overall, the results support the notion that the capital structure–performance dynamics are both sector- and period-dependent, with the pandemic amplifying pre-existing structural differences. While this study focuses on the comparison between financial and non-financial firms, future research could extend the analysis to a broader set of industries to assess whether the observed capital structure–performance relationships hold across different sectoral contexts.

5. Conclusions

This research set out to empirically analyse the association between capital structure and firm performance among companies listed on the FTSE All Share Index from 2018 to 2023, focusing on the implications of the COVID-19 pandemic. The primary objectives were to examine the relationship between capital structure and profitability metrics, investigate sector-specific variations in this relationship, and assess the impact of the pandemic on capital structure decisions and their performance implications. This study aimed to contribute to the ongoing debate on optimal capital structure and provide valuable insights for managers, investors, and policymakers in the context of the dynamic UK market.

The empirical analysis revealed several key findings. First, the study observed a significant negative relationship between leverage and firm performance across various profitability metrics, consistent with the pecking order theory and agency cost theory. Second, the research identified sector-specific variations in the capital structure–performance relationship, highlighting the importance of industry context in financial decision-making. Third, the analysis of pre-pandemic, pandemic, and post-pandemic periods demonstrated the profound impact of the COVID-19 crisis on capital structure decisions and their performance implications, with firms generally reducing leverage during the pandemic period.

This study makes several notable contributions to the existing literature on capital structure and firm performance. By focusing on the UK market and utilizing a comprehensive dataset spanning multiple industries and firm sizes, it provides a nuanced understanding of capital structure dynamics in a developed market context. The inclusion of the COVID-19 pandemic period offers unique insights into how extreme economic shocks influence the relationship between financing choices and firm performance. Furthermore, the sector-specific analysis contributes to a more granular understanding of capital structure decisions across different industries.

The findings of this research have significant implications for corporate financial management and policymaking. For managers, the results underscore the importance of carefully considering the potential negative impacts of high leverage on firm performance, particularly during periods of economic uncertainty. The sector-specific findings suggest that industry characteristics should be a key consideration in capital structure decisions. For policymakers, the study highlights the need for targeted support measures during economic crises, considering the varying impacts across different sectors. Investors can use these insights to better assess the financial health and potential performance of firms based on their capital structure choices.

While this study provides valuable insights, it has several limitations that should be acknowledged. First, the analysis focused exclusively on listed companies in the UK, which may limit the generalizability of findings to other contexts, including private firms or companies in different developed and emerging markets. Second, the study examined a specific set of financial performance metrics, potentially overlooking non-financial measures that could offer a more holistic view of firm performance. Third, although the study considered the capital structure–performance relationship, it did not explicitly examine the role of corporate governance mechanisms as potential moderators. Finally, the post pandemic period analysed was relatively short, which may constrain understanding of the long-term impacts of the COVID-19 crisis on financing decisions and firm performance.

Building on these limitations, several avenues for future research are suggested. Extending the analysis to include a longer post-pandemic period could offer a more comprehensive understanding of the long-term effects of the COVID-19 crisis. Incorporating qualitative research methods, such as interviews with financial managers, could provide deeper insights into the decision-making processes driving capital structure strategies, particularly in response to economic shocks. Further, comparative studies across other developed and emerging markets could help identify market-specific factors influencing the capital structure–performance relationship.

Although the sample covers the 2018–2023 period, this study focuses on the aggregate effects of the COVID-19 crisis rather than separate pre- and post-pandemic comparisons. Future research may extend the analysis using sub-period estimations.

Exploring the role of corporate governance mechanisms in moderating the relationship between capital structure and firm performance could offer a more comprehensive understanding of firm performance and provide valuable implications for both theory and practice.