Geopolitical risk and corporate investment inefficiency: evidence from India

Geopolitical risk (GPR) encompasses uncertainties and potential disturbances triggered by conflicts such as wars, terrorist attacks, and interstate tensions. This risk factor significantly impact global economic stability and international relations (Caldara and Iacoviello, 2022). In recent decades, GPR has intensified worldwide, due to critical events like the escalating US-China trade tensions, and the Russian invasion of Ukraine, among others. GPR has been found to have considerable impact on macroeconomic factors such as oil price volatility (Lee et al., 2021), gold prices (Gkillas et al., 2020), stock markets (Zhang et al., 2023), inflation (Yang et al., 2023), and global supply chains (Asadollah et al., 2024). Furthermore, GPR influences financial policies such as cash holding (Wang et al., 2021), leverage decisions (Khoo and Cheung, 2021; Kotcharin and Maneenop, 2020) and payout policies (Adra et al., 2023). Notably, some research indicates that heightened financial frictions caused by GPR can reduce corporate investments (Dissanayake et al., 2018; Caldara and Iacoviello, 2022; Lai et al., 2023; Rumokoy et al., 2023). However, despite extensive studies on the influence of GPR on corporate finance policies, no empirical studies examined the specific association between GPR and investment inefficiency, which this study addresses, focusing on underinvestment and overinvestment.

Investment is a crucial driver of economic development at both micro and macro levels, influencing firm value creation and national growth and prosperity (Le and Tran, 2021). However, prior literature suggests that various market frictions, such as asymmetry of information and agency conflict, often cause companies to diverge from the optimal level of investment (Lara et al., 2016), leading to investment inefficiencies in the form of overinvestment (excessive investment beyond optimal levels) and underinvestment (less investment than anticipated) (Akron et al., 2022). Agency theory suggests that management in companies with excess free cash flow (FCF) may be involved in non-value-maximizing investments for personal benefit, leading to overinvestment (Jensen and Meckling, 1976). Conversely, information asymmetry increases the cost of external financing, causing underinvestment in financially constrained firms (Chen et al., 2017).

Given the economic significance of investment, understanding the determinants of investment inefficiency is crucial for helping firms make optimal investment decisions. While a number of prior studies primarily focused on firm-level factors affecting investment inefficiency—including financial reporting quality (Biddle et al., 2009), accounting conservatism (Lara et al., 2016), maturity of debt (Cutillas Gomariz and Ballesta, 2014; Yadav and Yadav, 2024a), ownership structure (Chen et al., 2017), corporate social responsibility (CSR) initiatives (Benlemlih and Bitar, 2018) and corporate governance (Yadav and Yadav, 2024b) - less attention has been paid to macro-level factors. Some recent studies examining the role of macro-level determinants of investment inefficiency such as macroeconomic uncertainty (Irawan and Okimoto, 2021), economic policy uncertainty (EPU) in the power and energy sector (Hou et al., 2021), the world uncertainty index (WUI) (Akron et al., 2022), EPU in the small and medium enterprises sector (Hamza et al., 2024), and oil price uncertainty (OPU) (Yang et al., 2024). These studies collectively underscore the significance of macroeconomic forces in shaping efficient investment decisions. Among these uncertainties, GPR stands out as distinct. Unlike economic and financial uncertainties, GPR primarily concerns the adverse effects of geopolitical events on economic stability. Despite GPR's recognized importance, its influence on investment inefficiency and overinvestment remains unknown in the prior literature. Therefore, this study examines how GPR influences investment inefficiency, focusing on both overinvestment and underinvestment.

GPR influences investment inefficiency through several key mechanisms and channels. First, prior research shows that GPR heightens financial constraints and information asymmetry, thereby raising operational risks and increasing the cost of capital, debt, and equity (Nguyen and Thuy, 2023). This restricts firms' access to external financing as investors demand higher returns and stringent lending terms to mitigate default risks (Quagliariello, 2009), leading to underinvestment in profitable projects. Further, agency theory and real options theory explain that under uncertain conditions, managers often adopt risk-averse behavior, delaying or avoiding investments to safeguard their positions (Bertrand and Mullainathan, 2003; Bernanke, 1983; Dixit and Pindyck, 1994). Consequently, GPR-induced uncertainty leads to underinvestment as firms wait for clearer conditions before committing to capital expenditures. Second, GPR can also drive overinvestment, often stemming from managerial overconfidence and agency issues. Managers may allocate resources to low-value projects, resulting in asset misallocation and inefficiencies, especially when firms accumulate excess cash and face reduced scrutiny (Wang et al., 2021). This can result in the misallocation of assets, where increased investments fail to generate proportional growth (Wang et al., 2016). Empirical evidence supports this pattern of overinvestment during times of macroeconomic uncertainty, particularly in sectors like wind power (Liu, 2013), research and development (Ahuja and Novelli, 2017), and resource industries (Irawan and Okimoto, 2021). Thus, while GPR increases financial constraints and leads to underinvestment, it also fosters managerial discretion and overinvestment in uncertain environments. Both dynamics contribute to overall investment inefficiency, with firms either missing opportunities for growth or misallocating resources to projects that fail to deliver value.

We further explore how corporate governance (CG) moderates the effects of GPR on investment inefficiency. Given GPR’s destabilizing effects, it is essential to identify factors that mitigate its adverse effects on firm-level investment decisions—an area that has received limited attention. We propose that robust CG serves as a key mechanism for controlling and monitoring firms’ investment decisions during periods of heightened uncertainty. We believe that robust CG can mitigate GPR’s positive impact on investment inefficiency through several mechanisms. First, high-quality CG improves transparency and disclosure, which allows investors to make informed decisions. This reduces information asymmetry and alleviates financial constraints, facilitating better access to capital and lowering financing costs, even during periods of geopolitical uncertainty (Black et al., 2017). As a result, robust CG helps mitigate underinvestment by enabling more efficient capital allocation. Second, strong CG mitigates agency conflicts by curbing managerial tendencies toward overinvestment, particularly in firms with excess FCF (Yadav and Yadav, 2024b). During periods of heightened GPR, when external oversight may weaken, CG strengthens internal monitoring and curbs opportunistic behaviors, such as empire-building, by ensuring that managerial decisions face greater scrutiny (Rutherford et al., 2007). This reduces the likelihood of overinvestment. Lastly, firms with robust CG maintain a better reputation among investors, signaling accountability, transparency, and stability. This reputational strength helps retain investor confidence during geopolitical shocks, as high-CG firms are better equipped to manage risks and maintain performance (Black et al., 2017). Consequently, effective CG reduces firms' vulnerability to external shocks, including those arising from GPR, by improving capital allocation and risk management. Based on the above arguments, we contend that strong CG moderates the influence of GPR on investment inefficiency by promoting better oversight, reducing risk, and facilitating more informed investment decisions.

This research is focused on Indian firms due to the significant impact of GPR on the Indian economy, which is shaped by various global and domestic factors. First, in an interview with Nikkei Asia, Reserve Bank of India's (RBI) the then Governor Shri. Shaktikanta Das underlined the growing complexity of policy formulation due to geopolitical tensions ^[1]. He emphasized that global geopolitical conflicts fragment Indian economy, disrupt supply chains, and hinder growth. Despite India's projected rise to be the third-largest economy by 2027, it faces external pressures from global economic slowdowns, trade tensions, and capital flow disruptions, all threatening exports and access to capital. This underscores the importance of studying how GPR can destabilize investment flows in India. Second, a KPMG (2023) report reveals that both Indian and global chief executive officers (CEOs) perceive geopolitics as a major threat to business growth, with 14% of Indian CEOs identifying it as their greatest concern, compared to 18% globally. Additionally, 55% (versus 63% globally) of Indian CEOs believe geopolitical conflicts could negatively impact their organizations in the next three years, reflecting growing concerns about the impact of regional and global tensions on business operations. Third, GPR in India has increased by 27% above its 2011–2019 average ^[2], leading to capital outflows, declining equity markets, and rising bond yields. These factors worsen financing conditions, raising firm costs and potentially delaying or postponing investments. GPR also redirects foreign investments from emerging markets like India to developed economies, highlighting its potential to amplify corporate investment inefficiencies (Caldara and Iacoviello, 2022). Fourth, the Russia-Ukraine conflict has exacerbated inflation in India through supply chain disruptions (Maurya et al., 2023), prompting the RBI to raise lending rates to curb inflation. While this is a common central bank response, it inadvertently raises borrowing costs, discouraging new capital investments and negatively impacting firm-level investment efficiency. Fifth, India’s domestic geopolitical tensions, including its strategic disputes with China in the Himalayan region and long-standing tensions with Pakistan over Kashmir and cross-border terrorism, further complicate the country’s economic outlook. Additionally, India faces strategic uncertainty from broader United States (US)-China trade tensions as it balances its traditional alignment with the US and its reliance on China as a major trading partner. Internal and regional geopolitical dynamics, including strained relations with neighboring countries like Nepal and Sri Lanka, contribute to financial market volatility, further exacerbating uncertainty in equity markets and corporate decision-making. Given these factors, the elevated GPR in India significantly impacts economic activity and financing conditions, necessitating a closer examination of its influence on investment inefficiency.

To address gaps in the existing literature, this study utilizes a pooled ordinary least squares (OLS) method on a panel dataset of 3,943 non-financial listed firms in India from 2002 to 2023, comprising 43,182 firm-year observations. We measure GPR by employing the index established by Caldara and Iacoviello (2022), while investment inefficiency is assessed through the residual approach of Biddle et al. (2009). Our findings indicate that heightened GPR significantly exacerbates inefficient investments, both in terms of underinvestment and overinvestment. These results align with theoretical frameworks grounded in information asymmetry, real options theory, and agency theory. Moreover, companies with stronger CG structures demonstrate greater resilience to GPR-induced shocks, demonstrating the moderating effect of CG in lessening the impact of GPR uncertainty on investment inefficiency. Our results remain robust to alternative proxies for GPR, investment inefficiency and quantile regression approach. We address potential endogeneity through firm fixed effects, two-stage least squares (2SLS), two-step system generalized method of moments (2S-SGMM) techniques, Oster (2019) test for omitted variable bias and the incremental effect of GPR. Additionally, we apply entropy balancing to mitigate sample selection bias. Cross-sectional analysis further reveals that the influence of GPR on investment inefficiency is more pronounced for firms with lower degree of cash holdings, more irreversible investments, and that are financially unconstrained, those operating in industries with higher exposure to geopolitical risk and operate in highly competitive industries.

This investigation makes several key contributions. First, while earlier research has largely investigated the influence of GPR on macroeconomic variables, underinvestment and capital investments in terms of scale (Gkillas et al., 2020; Lee et al., 2021; Yang et al., 2023; Dissanayake et al., 2018; Caldara and Iacoviello, 2022), our investigation adopts a novel approach by investigating how GPR affects investment inefficiency, specifically through sub-samples of underinvestment and overinvestment. This investigation contributes to the earlier works on GPR by examining the association between GPR and corporate investment decisions. Second, prior research on investment inefficiency determinants has predominantly concentrated on firm (micro) factors, including debt maturity, financial reporting quality, CG, and accounting conservatism (Biddle et al., 2009; Lara et al., 2016; Yadav and Yadav, 2024a). While recent studies have begun to explore macro-level determinants like macroeconomic uncertainty, EPU, WUI, and OPU (Irawan and Okimoto, 2021; Hou et al., 2021; Akron et al., 2022; Yang et al., 2024), the specific impact of GPR on investment inefficiency and overinvestment remains largely unexplored. Therefore, this investigation contributes to the existing literature on the macroeconomic determinants of investment inefficiency by exploring the association between GPR and investment inefficiency. Finally, this study contributes to the CG literature by demonstrating how strong governance practices can mitigate the positive influences of GPR on investment inefficiency of Indian companies. Despite previous research indicating that robust CG diminishes asymmetry of information and agency concern through effective monitoring and financing functions (Kanagaretnam et al., 2007; Rutherford et al., 2007; Saragih, 2024; Yadav and Yadav, 2024a, b), this area, particularly in emerging economies, remains underexplored. By analyzing the moderating role of CG, this paper enriches the literature on how firms can buffer themselves against external risks, particularly in volatile geopolitical environments.

The structure of the study is ordered as follows: Section 2 formulates the hypotheses of the study. Section 3 presents research design. Section 4 describes the empirical analysis. Section 5 addresses potential endogeneity concerns. Section 6 discusses robustness checks. Section 7 explores cross-sectional analysis, while Section 8 discusses policy implications and concludes the study.

Geopolitical risk (GPR) arises from adverse events or tensions among states, threatening the peaceful course of international relations and contributing to global market instability (Caldara and Iacoviello, 2022). Since 2013, the systemic risk barometer survey conducted by the Depository Trust and Clearing Corporation (DTCC) has consistently identified GPR as one of the top five systemic risks, emphasizing its significant implications for the stability, safety, and resilience of the global financial system (DTCC, 2021). The introduction of the GPR index developed by Caldara and Iacoviello (2022), has facilitated a more rigorous empirical examination of its influence on macro and micro-level factors (Lee et al., 2021; Gkillas et al., 2020; Zhang et al., 2023; Kotcharin and Maneenop, 2020; Adra et al., 2023; Wang et al., 2021). Notably, there is growing evidence that GPR reduces corporate investment due to increased financial frictions (Dissanayake et al., 2018; Le and Tran, 2021; Caldara and Iacoviello, 2022; Lai et al., 2023; Rumokoy et al., 2023).

Investment serves as a vital driver of economic development at both micro and macro levels, affecting not only firm performance and value creation but also national growth and prosperity (Le and Tran, 2021). However, various market frictions, such as asymmetry of information and agency concerns, can lead companies to depart from ideal investment levels, resulting in inefficient investment that includes overinvestment or underinvestment (Lara et al., 2016). GPR significantly causes deviation from optimal investment through several mechanisms. First, financial constraints and information asymmetry are critical channels through which GPR exacerbates investment inefficiency. Rising geopolitical tensions elevate uncertainty, amplifying the asymmetry of information between firms and external investors. This often compels banks to raise interest rates and reduce lending volumes, increasing firms' cost of debt as compensation for the heightened default risk associated with uncertainty (Nguyen and Thuy, 2023; Kotcharin and Maneenop, 2020). Moreover, heightened GPR further increase the equity cost (Pástor and Veronesi, 2013). Thus, firms may find it increasingly challenging to finance profitable projects, raising the likelihood of underinvestment.

Second, agency theory posits that managers often prefer to maintain a “quiet life,” avoiding decisions that could increase the company’s risk exposure and endanger their professional standing or job security (Bertrand and Mullainathan, 2003). In the face of high GPR, managers may exhibit risk-averse behavior, opting to delay or postpone investment decisions. Real options theory supports this perspective, suggesting that investment is often irreversible due to adjustment costs, prompting firms to postpone decisions until uncertainties are resolved (Bernanke, 1983; Gulen and Ion, 2016). Consequently, elevated GPR constrains investment and exacerbates underinvestment as firms defer capital expenditures until uncertainties are resolved. Lastly, GPR can also drive overinvestment due to managerial overconfidence and agency issues (Huang et al., 2011; Malmendier and Tate, 2005). In uncertain times, overconfident managers may overinvest in low-value projects and justify these decisions as a defense against external shocks, especially when external oversight weakens. This results in asset misallocation and inefficiencies, where increased investment fails to generate proportional growth (Wang et al., 2016). Moreover, during heightened geopolitical risks, firms may accumulate excess cash (Wang et al., 2021) while facing reduced scrutiny, allowing managers to pursue self-serving projects misaligned with shareholder interests. Empirical research has shown that uncertain future policies can significantly increase overinvestment behavior (Liu, 2013; Ahuja and Novelli, 2017; Wang et al., 2016; Irawan and Okimoto, 2021). Recently, Nguyen et al. (2025) show that GPR negatively impacts investment efficiency among US firms. Their study further reveals that heightened GPR tends to exacerbate underinvestment while simultaneously reducing instances of overinvestment.

In summary, GPR constrains external financing options, leading to underinvestment while simultaneously creating opportunities for overinvestment through agency conflicts and managerial discretion during periods of reduced oversight. Based on the preceding discussion, we propose the following hypotheses:

While GPR exacerbates investment inefficiency and creates destabilizing effects, we hypothesize that CG is a crucial mechanism for controlling and monitoring companies’ investment behavior during periods of higher geopolitical uncertainty. CG plays a significant role in minimizing asymmetry of information and agency conflicts by enhancing managerial oversight and improving the dissemination of information. These robust governance practices help prevent financial constraints and mitigate opportunistic managerial behavior (Yadav and Yadav, 2024b; Kanagaretnam et al., 2007). Although GPR tends to increase investment inefficiency by introducing uncertainty and financial constraints, firms with strong CG systems may be better equipped to manage the impact of such exogenous shocks, ultimately improving their investment outcomes for several reasons.

First, companies with effective CG practices provide transparent and reliable information to investors, enabling them to make informed decisions about potential investment returns. This enhanced disclosure reduces information asymmetry and alleviates financial constraints by diminishing adverse selection problems (Black et al., 2017). Thus, we expect that robust CG can lessen the positive effects of GPR on investment inefficiency by promoting better capital allocation and reducing the likelihood of underinvestment. Second, high-quality CG mitigates moral hazard by curbing management incentives to be involved in value-destructive investments, including building an empire or overinvestment in companies with excess FCF (Black et al., 2017). CG also restrains managerial opportunistic behavior, mitigating the adverse impact of uncertainties (Chow et al., 2018; Yang and Song, 2023). During periods of geopolitical uncertainty, managers may exploit reduced external oversight to pursue such inefficient behaviors. However, strong CG practices enhance the monitoring of managerial decision-making, thus reducing the scope for opportunistic behavior. By improving investor oversight, CG restricts managers' ability to engage in inefficient investments, lowering the risk of overinvestment. Third, robust governance enhances a firm's reputation among investors, which can help alleviate the negative impacts of current or future crises, including geopolitical shocks. Firms with strong governance signal accountability, transparency, and financial stability, reducing idiosyncratic risk and boosting investor confidence. This reputational advantage makes it more likely that investors will retain their stakes in high-CG firms during periods of heightened GPR, as these firms are perceived to be more adept at managing risks and maintaining market performance. Consequently, CG acts as a buffer against the adverse effects of GPR, maintaining firm’s investment efficiency.

Building on this literature, we posit that companies with strong CG practices are better equipped to lessen the positive impact of GPR on investment inefficiency. Strong CG enhances accountability, transparency, and risk management capabilities, which helps reduce the likelihood of both underinvestment and overinvestment in the face of geopolitical uncertainty. Based on these insights, we propose the following hypotheses:

Our research focuses on a comprehensive sample of Indian listed companies from 2002 to 2023. This period is selected due to the availability of consistent and comprehensive data on Indian firms beginning from 2002. We obtain data from multiple datasets. Specifically, accounting data is obtained from COMPUSTAT Global and Bloomberg, while CG data is hand-collected from firms' CG reports retrieved through Centre for Monitoring Indian Economy's (CMIE) ProwessIQ database. Additionally, macroeconomic variables data are extracted from the world development indicators (World Bank). GPR index data is retrieved from Caldara and Iacoviello (2022) ^[3]. To confirm the relevance of our dataset, rigorous selection criteria were applied. Firms classified under the utility industries (SIC 4900–4999) and financial industries (SIC 6000–6999) are omitted due to their distinct regulatory frameworks and characteristics. Furthermore, observations with missing or incomplete data for key variables are excluded. After filtering, our final sample contains 43,182 firm-year observations from 3,943 non-financial companies. All variables are winsorized at the 1st and 99th percentiles to lessen the influences of outliers.

Investment efficiency is conceptually understood as a firm’s ability to execute all projects with a positive net present value (NPV). While no universal direct measure for investment efficiency exists, we follow the most popular model developed by Biddle et al. (2009), with deviations from expected investment levels as a proxy for inefficiency. In this model, the level of investment is regressed on sales growth, aligning with the accelerator theory, which posits that firms in efficient markets adjust their investments in response to product demand or sales growth. Accordingly, sales growth from the preceding year is a key predictor of the optimum level of investment. Thus, the baseline model is:

where Inves_i,t signifies the total investment of company “i” in year “t”, computed as the sum of expenditures in the capital, acquisitions, and research and development minus the property, plant, and equipment sales, all divided by lagged total assets. $SalesGrowthi,t−1$ represents the prior year’s sales growth rate. The residual $(∈$⁠) represents the deviation from predictable investment level. A positive residual specifies that a company's investment is higher than expected based on its growth opportunities, signaling overinvestment, while a negative residual specifies that the company's actual investment is lower than expected, indicating underinvestment. To quantify investment inefficiency (INV_INEFF), we use the absolute value of the residuals | $∈i,t$ |, where higher values reflect greater deviations from the expected investment levels, signifying greater inefficiency. Overinvestment (OI) is measured by positive residuals, while underinvestment (UI) is quantified as the absolute negative residuals value. In robustness tests, we also employ the Chen et al. (2011) model as an alternative proxy for investment efficiency.

The primary independent variable of interest in this study is the GPR index, established by Caldara and Iacoviello (2022). The GPR index is derived using automated text searches that measure the frequency of terms related to bad GPR events in 10 major international newspapers ^[4]. Specifically, the GPR index encompasses eight categories of adverse events related to geopolitical risk: (1) nuclear tensions, (2) war tensions, (3) military buildups, (4) peace tensions, (5) terrorist acts, (6) the initiation of war, (7) escalation of war, and (8) terror acts. GPR is calculated by counting the total number of articles mentioned in these eight categories in each newspaper monthly and dividing this by the total articles published. Additionally, Caldara and Iacoviello (2022) construct two subindices: the geopolitical risk threats (GPRT) index, which captures threats or expectations of future events (categories 1–5), and geopolitical risk acts (GPRA), which reflects the actual occurrence or escalation of ongoing geopolitical events (categories 6–8). For this study, we convert the monthly GPR, GPRT, and GPRA indices into annual measures by averaging the data over 12 months and then taking the natural logarithm of the annual values to minimize skewness in the data, following previous research (Rumokoy et al., 2023; Fiorillo et al., 2023). Higher values of GPR, GPRT, and GPRA indicate elevated geopolitical risks.

The monthly GPR index for the sample period is shown in Figure 1. As demonstrated from the plot, the index effectively captures major geopolitical events, with the highest spikes corresponding to key incidents such as the onset of the Iraq War in 2003 and the Russia–Ukrainian War in 2022. Additionally, smaller yet notable peaks are observed around events like the Russian annexation of Crimea in 2014, the Paris attacks in 2015, the North Korean crisis during 2017–2018, and the assassination of Qasem Soleimani in 2020. Similarly, Figure 2 presents the monthly Indian GPR index, capturing significant geopolitical developments specific to India. Notable spikes align with events such as the India-Pakistan tensions in 2001–2002, Mumbai terror attacks in 2008, the Uri attack and the Surgical Strikes in 2016, the Pulwama attack and Balakot airstrike in 2019, the abrogation of Article 370 of the Indian Constitution to Jammu and Kashmir in the same year, the India-China border clashes in the Galwan Valley in 2020, the US withdrawal from Afghanistan in 2021, and India's neutral stance during the Russia-Ukraine War in 2022. These visual representations of the GPR index offer clear evidence of its ability to track and reflect geopolitical tensions and their impact over time, both globally and in the Indian context.

We develop a firm-level CGI ^[5] as a moderating variable to evaluate governance practices by considering recent Indian regulations and guidelines related to ownership structure, board structure, nomination and remuneration, disclosure practice, and audit quality. The governance index comprises 65 parameters for good governance (see  Appendix Table A2), drawn from global standards and India's legal framework, including the Indian Companies Act, 2013 and the Organisation for Economic Co-operation and Development (OECD) principles, instead of employing the country-level CG ratings. Following the methodology established by Black et al. (2017), we employed a dichotomous approach to construct a CGI, assigning a score of 1 to firms that comply with the requirements and 0 to those that do not. Thus, a company “i” can attain an extreme value of 65 in any given year “t” if it adheres to all governance parameters. The total scores are normalized and scaled between 0 and 100 to enable firm-level comparisons, where 100 represents full compliance, and 0 indicates no compliance. The governance index is calculated employing the following model:

where in equation (2) $CG_INDICATORS$ are 65 corporate governance elements.

We incorporate a comprehensive set of control variables that may impact inefficient investment based on earlier research (Biddle et al., 2009; Benlemlih and Bitar, 2018; Yadav and Yadav, 2024a). These variables comprise tangibility (TANG), firm size (SIZE), leverage (LEVE), return on assets (ROA), cash (CASH), financial slack (SLACK), operating cycle (Ln_OC), return on equity (ROE), cash flow from operations (CFO), Tobin’s Q (TOBINSQ), firm age (Ln_Age), and GDP growth rate (GDP_GRO). We include year and industry fixed effects in our regression models to account for potential confounding effects arising from time and industry variations. Measurement of variables employed in our analysis is provided in  Appendix Table A1.

To investigate the influences of GPR on investment inefficiency, overinvestment, and underinvestment, we estimate the following baseline regression models:

In these equations, the dependent variable $INV_INEFFi,t$ denotes investment inefficiency, while OI and UI represent overinvestment and underinvestment for company “i” in period “t”, respectively. Specifically, INV_INEFF is quantified as the absolute value of the residuals obtained from equation (1), where overinvestment (OI) corresponds to positive residuals, and underinvestment (UI) refers to the absolute value of negative residuals. The primary independent variable under investigation is GPR. The vector Y includes control variables that impacting inefficient investment, as outlined in the variable’s description section. Additionally, year and industry fixed effects are incorporated, with the error term signified by $∈i,t$ across all models.

To mitigate potential endogeneity concerns, particularly those stemming from reverse causality, and recognizing that investment decisions are typically made in advance, we incorporate a one-year lag for independent and control variables in our analysis (Akron et al., 2022). We estimate all equations using OLS and cluster the standard errors at both the year and firm levels to account for within-firm serial correlation and heteroskedasticity, consistent with the earlier literature (Yadav and Yadav, 2024b; Benlemlih and Bitar, 2018; Biddle et al., 2009).

Further, to analyze the effect of firm-level CG as a mitigating channel, we employ the following model:

In this equation, the term GPR $×$ CGI captures the interaction effect between GPR and CGI. We expect the interaction term coefficient to be negative and significant, as it will support our hypothesis that strong CG weakens the positive association between GPR and inefficiency in investment. Similarly, a negative and significant interaction term for overinvestment (OI) and underinvestment (UI) would indicate that robust governance practices reduce the likelihood of excessive or insufficient investment during periods of geopolitical volatility.

Table 1 shows the descriptive statistics of the main variables used in our analysis. Our sample contains 43,182 firm-year observations, with 30,138 observations (69.79%) classified as underinvesting and the remaining 13,044 observations (30.20%) identified as overinvesting. This highlights that under-investment is more prevalent among Indian listed firms. The average value of investment inefficiency (INV_INEFF) is 0.081, ranging from 0.000 to 0.808. Firms that overinvest exhibit a higher mean investment inefficiency (0.134) than underinvesting firms (0.058). Regarding the geopolitical risk indicators, the mean (and standard deviation) for the GPR, GPRT, and GPRA variables are 4.561 (0.176), 4.601 (0.230), and 4.470 (0.287), respectively. These values suggest varying levels of geopolitical risk perceptions across the sample period. Furthermore, the mean value of the CGI is 69.481, with a standard deviation of 6.205, indicating variations in corporate governance practices across the sample firms. The descriptive statistics for control variables are similar to findings in prior studies (Lara et al., 2016).

Further, Table 2 demonstrates the Pearson correlation matrix for the main variables. As expected, the GPR index shows a positive and significant relationship with investment inefficiency, which supports the hypothesis that heightened GPR leads to increased inefficiencies in corporate investment. Conversely, the CGI demonstrates a negative and significant correlation with inefficient investment, implying that good CG stipulations help to lessen inefficiencies in investment. The correlations among the control variables are consistent with prior studies (Biddle et al., 2009; Cutillas Gomariz and Ballesta, 2014), and the low pairwise correlations among the independent variables suggest an absence of multicollinearity.

Table 3 presents the findings of estimating equations (3) to (5) using pooled OLS. The estimation of equation (3) shows that the coefficient for GPR is positive and significant at the 1% level, indicating that companies facing heightened uncertainty due to increased GPR significantly increases investment inefficiency (models 1 and 2). This result is consistent with H1, indicating that increased geopolitical uncertainty leads to suboptimal capital allocation. These dynamics are supported by the recent study of Nguyen et al. (2025), which finds that GPR increases investment inefficiency among US firms, highlighting the adverse effects of heightened uncertainty on capital allocation decisions.

The analysis further examines the differential impact of GPR on overinvestment and underinvestment. For overinvestment, the estimation results of equation (4) reveal that GPR significantly increases overinvestment (models 3 and 4), supporting H1a. It demonstrates that heightened GPR exacerbates resource misallocation and amplifies agency issues, allowing managers to engage in overinvestment. Prior studies support these dynamics, with studies showing uncertain future policies increase overinvestment (Liu, 2013; Ahuja and Novelli, 2017; Irawan and Okimoto, 2021). With respect to underinvestment, the estimation of equation (5) reveals a similarly positive and significant relationship between GPR and underinvestment (models 5 and 6). These results support H1b, suggesting that increased GPR uncertainty constrains firms from making value-generating investments, thereby heightening underinvestment. The results align with theories of information asymmetry and real options, which suggest that firms facing high uncertainty are likely to delay or avoid investments, leading to underinvestment (Dissanayake et al., 2018; Nguyen and Thuy, 2023). However, it is worth noting that Nguyen et al. (2025) report contrasting evidence in the context of US firms, showing that GPR reduces underinvestment. This divergence may be attributed to differences in institutional environments, market maturity, or risk management practices between developed and emerging markets, highlighting the context-specific nature of firm responses to geopolitical risk. Further, coefficients of the control variables are similar to the earlier literature on investment efficiency, as demonstrated in Table 3 (Benlemlih and Bitar, 2018; Biddle et al., 2009; Lara et al., 2016).

Recent research has explored how other macro-level uncertainties such as EPU, WUI, and OPU impact investment inefficiency (Hou et al., 2021; Akron et al., 2022; Hamza et al., 2024). Interestingly, these studies present findings that are contradictory to our results, indicating that macroeconomic uncertainties may reduce rather than exacerbate investment inefficiency. The divergence in findings between ours and these prior studies could be attributed to the distinct nature of GPR, which is often associated with unpredictable and acute geopolitical events, which usually result in sudden and severe disruptions to firms' operating environments. These events create uncertainty that is far more challenging for firms to anticipate or manage compared to more structured uncertainties like EPU or fluctuations in oil prices. Conversely, macroeconomic uncertainties like EPU, WUI, and oil price fluctuations may be more predictable and follow cyclical patterns that allow firms to mitigate risks through hedging and other financial instruments.

The GPR measure encompasses both the recognition or increase of ongoing adverse geopolitical events (GPRA) and the threats or expectations of future events (GPRT). To further validate our findings, we inspect the separate influences of GPRA and GPRT on inefficiency of investment. In this analysis, we substitute the main regressor in equations (3) to (5) with the GPRA and GPRT indices. The results are summarized in Table 4, with Panel A focusing on GPRT and Panel B on GPRA. Our findings demonstrate that the coefficient of GPRT is positively significant at the 1% level across all models: model 1 (inefficient investment), model 2 (overinvestment), and model 3 (underinvestment). Similarly, we observe a consistent positive coefficient for GPRA in models 4 to 6. These results reinforce our earlier findings in Table 3, confirming that uncertainty stemming from both GPRT and GPRA exacerbates all investment inefficiency variables.

We expand our analysis by examining whether CG moderates the association between GPR and investment inefficiency. Table 5 displays the results from estimating equation (6), where we incorporate interaction effects between GPR and CGI (GPR $×$ CGI) to explore whether strong governance mechanisms can buffer the impact of geopolitical uncertainty. The coefficient for GPR (models 1–3) remains positive and significant, confirming its influence on all forms of investment inefficiency, like the findings in Table 3. In contrast, the coefficient for CGI exhibits a negative and significant impact across all models, aligning with previous research (Yadav and Yadav, 2024b; Kanagaretnam et al., 2007).

The interaction term (GPR × CGI) in model 1 shows a negative and significant coefficient (−0.0025), suggesting that firms with stronger CG can better mitigate the adverse effects of GPR on investment inefficiency. This result supports H2, indicating that effective CG can reduce the adverse effects of GPR on investment behavior by improving access to capital and fostering better oversight, thereby ensuring efficient capital allocation during geopolitical uncertainties. Specifically, firms with strong CG enjoy enhanced reputations, signaling accountability and stability, which helps maintain investor confidence during geopolitical shocks. Thus, effective CG serves as a safeguard mechanism against the negative influences of GPR by facilitating better access to capital, fostering improved oversight, and ensuring more efficient capital allocation even amid geopolitical uncertainty. Further, in model 2, the interaction term coefficient is also negative (−0.0052) at the 5% significance level, demonstrating that robust CG lessens the positive influence of GPR on overinvestment. This finding confirms H2a, showing that strong CG curtails managerial tendencies toward overinvestment. Lastly, model 3 presents similar results for underinvestment, where the interaction effect coefficient (−0.0005) is negative and significant. These findings confirm our hypothesis (H2b), suggesting that companies with high-quality CG are better positioned to enhance their reputation among investors and stakeholders, acting as an “insurance-like” protection against external shocks. This improved reputation helps alleviate financial constraints, as better governance increases investor confidence and facilitates access to capital during uncertain times.

To further validate our findings, we explored the influence of GPRT and GPRA on inefficient investment with CGI. Table 6 shows the results in Panel A (for GPRT) and Panel B (for GPRA), where we observe similar negative coefficients for the interaction terms (GPRT × CGI and GPRA × CGI) across both panels for investment inefficiency, overinvestment, and underinvestment. Overall, the empirical findings reinforce our hypotheses (H2, H2a, and H2b), demonstrating that CG effectively lessens the positive influence of GPR on all forms of investment inefficiency, highlighting the crucial role of CG in enhancing firm resilience against geopolitical risks.

Our primary research design in the baseline model implementing one-period lag addresses reverse causality issues significantly. To mitigate further potential endogeneity concerns between GPR variables and investment inefficiency, we employ several econometric analyses: (1) two-stage least squares estimation; (2) Oster (2019) test for omitted variables; (3) two-step system GMM; (4) analysis of the incremental effect of GPR; and (5) firm fixed-effects estimation. These methods aim to address issues of simultaneity, measurement error, and omitted variable bias. We discuss the first two analysis in the paper, with the rest discussed in detail in the Supplementary Appendix (S.A. 1, 2 and 3) for space conservation.

Endogeneity arises from the correlation between explanatory variables and the error term, leading to inconsistent and biased parameter estimates. To alleviate this, we use a 2SLS instrumental variable (IV) estimation. A valid instrument (IV) must satisfy two key conditions: (1) relevance, meaning it must be correlated with the endogenous regressor; and (2) exclusion, meaning it should not directly affect the dependent variable except through its effect on the endogenous variable. By employing instruments into the model helps remove the correlation between the error term and the endogenous variable. Thus, to address potential endogeneity concerns, this study employs the growth rate of terrorist attacks (GTA) ^[6] as an instrumental variable for GPR. This approach aligns with prior research, including Nguyen and Thuy (2023) and Phan et al. (2022), who similarly utilize GTA as a valid instrument. GTA reflects terrorism's activities in a country, encompassing casualties, attacks, and economic effects factors that contribute directly to geopolitical instability. As terrorism directly escalates security concerns, disrupts political and economic environments, and often triggers international tensions, it serves as a strong predictor of elevated geopolitical risk (Nguyen and Thuy, 2023; Phan et al., 2022). Thus, GTA serves as a relevant proxy for geopolitical risk by providing significant exogenous variation in GPR, satisfying the relevance criterion. Moreover, since GTA influences investment inefficiency only indirectly through GPR and there is no theoretical or empirical evidence of a direct effect it also meets the exclusion restriction. Taken together, these conditions confirm that GTA satisfies both the relevance and exclusion criteria necessary for a valid instrumental variable. In the first stage, GPR is regressed on the GTA alongside control variables, with the equation specified as follows:

To assess the validity of the instrumental variable, we examine the results of the first-stage regressions reported in Table 7 for model 1, model 3, and model 5. As expected, these results demonstrate that GTA is positively and significantly associated with GPR, confirming the relevance of the instrument. In support of this, we also find a positive and significant correlation between GTA and GPR (correlation coefficient = 0.0406) (untabulated). These findings are consistent with Abdullah et al. (2024), Nguyen and Thuy (2023), and Phan et al. (2022), who also document that heightened terrorist activity contributes to increased geopolitical uncertainty. To ensure our instruments' validity, we use several diagnostic tests. The Kleibergen-Paap rk LM statistics are significant across all models, which rules out any under-identification concerns. The value of Cragg-Donald Wald F-statistics is more than the Stock-Yogo critical values, alleviating apprehensions about weak instrument identification. In the second stage, we replace GPR with its predicted value from the first stage and re-estimate the investment inefficiency models using the following specifications:

The findings of the IV estimation are detailed in Table 7. Across all models (model 2, model 4, model 6), confirm the earlier results in Table 3. Specifically, the coefficient for the predicted GPR values (Pred_GPR) remains positive and significant across all investment inefficiency measures.

Despite controlling for an extensive range of firm level and macroeconomic factors, the possibility of unaccounted-for confounders influencing the GPR–investment inefficiency relationship cannot be entirely ruled out. To alleviate this concern, we use the econometric framework given by Oster (2019) to evaluate the stability of our results. The Oster δ statistic measures the relative size of omitted variable bias required to nullify the observed effects (Oster, 2019). By analyzing variations in coefficients and R-squared across model specifications, with and without observable controls, this approach offers insights into the influence of unobserved variables. Oster (2019) suggests that δ values greater than 1 indicate robust results, implying that unobserved confounders would need to have a disproportionately large impact to fully explain away the estimated effects. As shown in Table 8, δ values significantly exceed this conventional threshold of 1. This provides strong evidence that the estimated positive relationship between GPR and investment inefficiency encompassing overinvestment and underinvestment is unlikely to be determined by omitted variable bias.

To provide further robustness to our results, we employ multiple alternative approaches. These include the entropy balancing, alternative measures of investment inefficiency and geopolitical risk and heterogeneity investigation. For space conservation, we provide the discussions on heterogeneity analysis in the  Supplementary Appendix (S. A. 4).

To tackle potential model mis-specification issues and endogeneity related to sample selection bias, we employ the EBAP, which has emerged as a more robust alternative to propensity score matching (PSM). While PSM has been widely used, recent critiques suggest that it may worsen imbalances and introduce bias (King and Nielsen, 2019). In contrast, EBAP adjusts the distribution of covariates by applying weights to the control group, thereby ensuring balanced covariates between treated and control firms (Hainmueller, 2012). In our analysis, we replicate the baseline model using this method to mitigate these concerns. Firm-year observations with GPR values exceeding the median are categorized into the treatment group, while those with GPR values below the median are assigned to the control group. Table 9 (Panels A and B) shows the covariate means and variances for treated and control firms before and after balancing. Initially, there were notable differences between the two groups. However, after implementing EBAP, no statistically significant differences remain between the covariates, confirming successful balancing. We then re-estimated the baseline model using the weighted sample. The findings in Table 9 (Panel C) reveal that the GPR remains a significant driver of investment inefficiency, overinvestment, and underinvestment, further corroborating our primary findings.

To further assess the robustness of our results, we utilize an alternative proxy for investment inefficiency based on the methodology of Chen et al. (2011). This method accounts for the association between sales growth and investment, incorporating the effects of periods of increasing and decreasing sales. A dummy variable, NEGA, is introduced to capture periods of negative sales growth (NEGA = 1 if sales growth is negative; otherwise, 0). We also include the interaction between NEGA and sales growth in our model to compute this measure of investment inefficiency, as shown in the following equation:

The estimated findings using this alternative measure are shown in Table 10, demonstrating consistent results. Untabulated findings also show that the effect remains similar for the alternative GPR proxies (GPRT and GPRA). This confirms that GPR increases investment inefficiencies, irrespective of the measure employed.

We further use alternative measures of GPR to ensure that our findings are not sensitive to the choice of GPR proxy. First, we employ a country-specific GPR index for India (GPR_IND). Second, while the log of average monthly GPR serves as our primary proxy, we also use the log of GPR observed in December of every year (GPR_DECEMBER) as a different measure, following the approach of Le and Tran (2021). Third, to capture the longer-term effects of GPR, we incorporate the log change in GPR over a two-year period (ΔTWO_YEAR_LN_GPR). Lastly, we consider the yearly changes in the GPR index (ΔLN_GPR), as suggested by Abdullah et al. (2024), to account for annual fluctuations in geopolitical risk. The results, presented in Table 11, indicate that, across all alternative measures, the coefficients for GPR_IND, GPR_DECEMBER, ΔTWO_YEAR_LN_GPR, and ΔLN_GPR remain positive and statistically significant. This consistent finding suggests that geopolitical risk continues to contribute to investment inefficiency, overinvestment, and underinvestment, regardless of the specific proxy used.