Risk Management in Product Diversification: The Role of Managerial Overconfidence in Cost Stickiness—Evidence from Iran

Abstract

Purpose: This study investigates the relationship between product diversification strategy and cost stickiness, focusing on managerial overconfidence as a moderating factor. It aims to address a critical gap in the literature by providing empirical insights grounded in the Resource-Based View (RBV) theory, specifically examining firms listed on the Tehran Stock Exchange. Methodology: Utilizing a sample of 149 companies from the Tehran Stock Exchange in Iran spanning from 2015 to 2021, this study tests two hypotheses: (1) a positive relationship between product diversification and cost stickiness and (2) the amplification of this relationship by managerial overconfidence. Product diversification is quantified using the Herfindahl Index, while managerial overconfidence is measured through an investment-based index derived from capital expenditures. Cost stickiness is assessed by analysing the asymmetric behaviour of costs in response to changes in sales, focusing on how costs tend to remain high even when sales decrease. Findings: The empirical results substantiate both hypotheses, demonstrating a significant positive relationship between product diversification strategy and cost stickiness. Furthermore, managerial overconfidence amplifies this relationship, highlighting the role of internal resources and managerial perceptions in shaping cost behaviour. Originality: This study contributes substantially to the literature by being among the first to empirically examine the interplay between product diversification strategy, cost stickiness, and managerial overconfidence. Extending the RBV theory to cost behaviour and strategic management provides novel insights for scholars and practitioners in entrepreneurship, corporate strategy, and organizational behaviour. The findings underscore the importance of strategic choices and managerial traits in determining cost stickiness, offering valuable implications for financial analysts, auditors, and stakeholders.

1. Introduction

In an increasingly competitive and volatile business environment, understanding the dynamics of cost behaviour and strategic management is crucial for organizations striving for long-term success (Hitt et al. 2020). Organizations encounter many strategic challenges, among the most critical of which is the need to effectively align their growth strategies with rigorous cost management practices (Grant 2016). Product diversification strategy has received considerable attention in the literature (Bengtsson 2000; Garrido-Prada et al. 2019; Tang et al. 2019; Wang et al. 2014). It represents organizational flexibility by enabling companies to adapt to changing market conditions through an expanded range of products and services. It improves stability by allowing the firm to spread its business risks across different product lines and reduce its vulnerability to market fluctuations while maximising the efficiency of its resource allocation (Dinh et al. 2019; Friberg 2021; Hautz et al. 2014; Mammen et al. 2021; Smeritschnig et al. 2021; Sohl et al. 2022; Solano et al. 2019). The literature supports this notion that product diversification can be considered as one of the competitive strategies (Zúñiga-Vicente et al. 2019). However, given the organisational changes and costs that a product diversification strategy entails, we are unsure if there would be less cost reduction during declining sales, leading to increased cost stickiness. This is the primary question that this study attempts to address.

To examine the relationship between business strategy, product diversification, and cost behaviour, it is crucial to emphasize the importance of cost management in organizations. Cost is one of the most critical factors for any enterprise, and effective cost management is essential for achieving an optimal internal structure, earning a profit, and maintaining continuous development (Cai et al. 2019; Ding et al. 2021; Dormady et al. 2022; Stadtherr and Wouters 2021). Lee et al. (2020) consider cost management a core component of internal structure optimization, risk control, and profitability. Hence, understanding cost behaviour is fundamental to management and cost accounting (Anderson et al. 2003). Managerial overconfidence, a behavioural bias where managers overestimate their abilities and underestimate risks, plays a pivotal role in strategic decision-making (Gurdgiev and Ni 2023). Overconfident managers are more likely to retain resources during sales declines, anticipating future rebounds, which can intensify cost stickiness (Chen et al. 2022). This study explores the relationship between product diversification and cost stickiness, specifically focusing on how managerial overconfidence influences this dynamic.

The theoretical foundation of this study is grounded in the Resource-Based View (RBV) theory, which posits that firms can gain a competitive advantage through unique resources and capabilities (Barney et al. 2011). In this context, product diversification represents a unique resource, while cost stickiness is viewed as resource management and risk mitigation capabilities. This study formulates two hypotheses for testing: First, it proposes a positive relationship between product diversification strategy and cost stickiness. This proposition aligns with RBV, as diversification often involves significant investments, resulting in elevated resource adjustment costs and delayed reduction of underutilized resources during sales declines. Second, this study suggests that managerial overconfidence strengthens the direct link between product diversification and cost stickiness. This hypothesis is consistent with RBV, as overconfident managers are more likely to retain underutilized resources, anticipating future sales rebounds, thus intensifying cost stickiness.

Evolving regulations in Iran’s environment can influence managerial decisions, often leading to cautious resource management during sales declines. This careful approach, driven by overconfidence and regulatory considerations, can contribute to increased cost stickiness, particularly in firms with diversified product portfolios (Nili and Rastad 2007). In addition, external economic conditions have affected the business environment by reducing access to some international markets and capital, leading to challenges in operational stability. In response, managers may adopt more optimistic behaviours as a strategic adjustment, contributing to cost stickiness as they retain resources in anticipation of future opportunities (Katzman 2014). Moreover, the inflation rates add complexity to cost management, as the real value of costs can be distorted. Overconfident managers may underestimate the risks of inflation, leading to increased resource commitments despite declining sales, which heightens cost stickiness (Pesaran 2012). Finally, cultural context influences how risk is perceived and handled. Iranian cultures where risk-taking is en-couraged, managerial overconfidence may be more prevalent, as leaders may feel empow-ered to make bold decisions. This cultural dynamic can lead to decisions favouring maintaining resources during downturns, thereby increasing cost stickiness (Nazarian and Atkinson 2013).

Despite the extensive literature on cost stickiness and its influencing factors in various contexts (Chen et al. 2022; Costa and Habib 2021), the role of managerial overconfidence in moderating the relationship between product diversification and cost stickiness has not been extensively studied, particularly in the context of the Iranian market. This study investigates the intricate relationship between product diversification strategy and cost stickiness, focusing on the moderating influence of managerial overconfidence on decision-making. It aims to fill a notable gap in the existing literature by providing empirical insights into this underexplored area, particularly within the risk management framework in corporate finance.

Using a sample of 149 companies listed on the Tehran Stock Exchange in Iran from 2015 to 2021, this study employs rigorous statistical analysis to assess these hypotheses. The results empirically substantiate both hypotheses, shedding light on the positive relationship between product diversification and cost stickiness and the amplifying effect of managerial overconfidence.

This study makes a significant theoretical contribution by anchoring its research questions in the RBV theory, providing robust empirical evidence, and delivering valuable insights for scholars and practitioners of entrepreneurship, corporate strategy, and organizational behaviour worldwide. By focusing on the intersection of product diversification, cost behaviour, and managerial traits, this study offers new insights into risk management frameworks and financial strategies, contributing to the broader discourse on corporate finance and strategic decision-making in high-risk environments. The remainder of this study is structured as follows. Section 2 reviews the research literature and develops the hypotheses. Section 3 discusses the research method. Section 4 reports the results, and Section 5 and Section 6 provide the discussion and conclusion.

2. Literature Review and Hypotheses Development

Traditional models of cost behaviour divide costs into fixed and variable based on some conventional drivers, such as the sale or production volumes (Bugeja et al. 2015). However, recent research indicates that the relationship between costs and cost drivers is more complex than traditionally assumed, involving multiple factors influencing cost behaviour (Banker and Byzalov 2014; Costa and Habib 2021). Cooper and Kaplan (1998) show that costs increase more when activity rises than they decrease with a proportional decrease in activity. This asymmetric cost behaviour is referred to as cost stickiness (Cooper and Kaplan 1998).

Recent research in cost management adds two crucial concepts to traditional cost analysis. First, cost behaviour is driven by managerial decisions, which depend on managers’ expectations of future demand and their incentives (Jin and Wu 2021). Second, when sales decline, managers’ ability to reduce unutilised resources becomes limited due to resource adjustment costs, such as the cost of employee layoffs (Banker et al. 2018). Managers’ optimism about future demand, managerial incentives driven by self-interest (e.g., empire-building), and adjustment costs are the key factors that motivate managers to retain unutilised resources during declining sales. Managers’ unwillingness to cut resources in such conditions will lead to cost stickiness (Banker et al. 2013; Hartlieb et al. 2020; Xue and Hong 2016).

Asymmetric cost behaviour and cost management have broad implications for owners, investors, creditors, employees (concerning job security), and other stakeholders (Costa and Habib 2021). Cost stickiness affects the authenticity of a firm’s costs and reduces the effectiveness of its resource allocation (Li et al. 2020). Recent studies have investigated the relationship between cost stickiness and various factors (Chen and Xu 2023; Costa and Habib 2021; Haga et al. 2019; Li et al. 2020). However, the relationship between business strategies, particularly growth strategies and cost stickiness, has not been explored. Anderson et al. (2003) argue that a firm’s strategy choice shapes its cost structure. Porter’s generic business-level strategy, namely, cost leadership, differentiation, and focus, have become a central paradigm in the business policy literature (Porter 1980). Each represents a different approach to creating and maintaining a competitive advantage. For example, companies that adopt a cost leadership strategy have a more flexible cost structure. When sales decline, they have to reduce the price of products to maintain their competitive advantage. As a result, they cut costs and experience lower cost stickiness. On the other hand, companies that adopt a differentiation strategy will face high adjustment costs when sales decline. As a result, they retain unutilised resources and experience higher cost stickiness (Zhong et al. 2020).

Diversification is one of the four growth strategies outlined in Ansoff’s product-market matrix (Ansoff 1957). Diversification is a strategy where the company expands its products and services. Diversification has been a common competitive strategy for decades (Jha et al. 2021). A diversification strategy targets new markets with new products. It assumes that a company can adopt and require new resources and capabilities and make necessary changes in organisational structure and processes (Zúñiga-Vicente et al. 2019). So, we can expect that companies adopting a diversification strategy must restructure their business units and change current practices and existing contracts. This strategy requires comprehensive planning to implement diversification processes using internal and external resources and expertise. As a result, recruiting specialists, hiring and training new employees, and new equipment might be needed, which can be very costly (Lin et al. 2020).

As the business environment becomes more competitive, effective cost management has become an essential element of business operations (Chen et al. 2022). As an intuitionistic reflection of the cost decisions made in the accounting information system, cost behaviour shows the current conditions and tendencies of the company’s resource investment. It reflects the executives’ strategies in response to changes in economic conditions (Lee et al. 2020). Resource adjustment decisions are one of these responses that significantly impact firm performance (Chen et al. 2022).

Another factor that may affect organisational strategy and product diversification is managerial behaviour biases (Kak 2004). One of these biases is managerial overconfidence, which refers to managers’ tendency to overestimate their abilities and/or overinvest (Hyun and Seung 2019). Overconfidence is one of the essential concepts in modern behavioural finance. Overconfidence causes people to overestimate their knowledge and skills and underestimate risks, feeling that they have control over events, which may not be the case (Weinstein 1980). Anderson et al. (2003) argue that one reason for asymmetric cost behaviour is the CEO’s optimistic expectations about future sales improvement (Hur et al. 2019; Schrand and Zechman 2012). Hribar and Yang (2016) found that overconfidence increases the degree of optimism in management forecasts. As a result, managers will be more willing to retain slack resources when sales decline to reduce adjustment costs (Hribar and Yang 2016; Lee et al. 2020). Overconfidence can manifest itself by adding more resources when demand increases or retaining unutilised resources when demand decreases, leading to greater cost stickiness (Chen et al. 2022). On the other hand, Chu et al. (2021) found that companies with overconfident managers will have greater cost stickiness if they adopt a differentiation strategy than a cost leadership strategy (Chu et al. 2021).

Companies may adopt different strategies in a competitive market, leading to asymmetric cost behaviour. For example, as noted earlier, cost stickiness can increase with a differentiation strategy increase but decrease with a cost leadership strategy (Zhong et al. 2020). On the other hand, a product diversification strategy aimed at higher profitability and growth can increase sales volume and lead to better financial performance (Dinh et al. 2019; Omosa et al. 2022). Given that companies with a product diversification strategy require new investments, skills, equipment, practices, and changes in organisational structure and processes, it is possible that, when future sales are uncertain, overconfident managers will maintain the business scale and avoid adjusting unutilised resources in the hope of obtaining higher returns in the future, which will lead to more sticky cost behaviour (Chen et al. 2022). Therefore, we expect managerial overconfidence to moderate the relationship between product diversification strategy and cost stickiness. The present research seeks to answer the following questions: Is there a relationship between product diversification strategy and cost stickiness? Does managerial overconfidence moderate the relationship between diversification and cost stickiness?

Following prior research (Lin et al. 2020; Tang et al. 2022; Zúñiga-Vicente et al. 2019), product diversification is measured using the Herfindahl Index. Following Killins et al. (2021), managerial overconfidence is measured using an investment-based index, which is calculated based on the median of capital expenditures. Cost stickiness is measured according to the model proposed by Banker et al. (2013). Our results show a significant positive relationship between product diversification strategy and cost stickiness. In other words, cost stickiness is higher in companies with higher product diversification. In addition, managerial overconfidence increases asymmetric cost behaviour in companies with a product diversification strategy.

In the literature, various methodologies such as scenario analysis, econometric modelling, and stress testing are commonly used to assess the impact of global risks on firms (Schwartz 1997; Wooldridge 2010; Borio et al. 2014). Scenario analysis and stress testing often rely on hypothetical scenarios to predict outcomes, while econometric models analyse historical data to quantify these risks. Our study diverges from these approaches by using empirical data from the Tehran Stock Exchange to examine how economic conditions in Iran affect cost stickiness and managerial decisions. This empirical approach provides a more accurate reflection of real-world impacts, offering specific insights into the Iranian context that might differ from global risk assessments based on hypothetical scenarios.

Product diversification can mitigate risks by spreading them across different markets, which is beneficial in stable or growing economies. However, during economic downturns, the high costs associated with maintaining diverse product lines may outweigh the benefits. The economic environment, including inflation, currency fluctuations, sanctions, and evolving regulations, can further complicate diversification efforts, making them more costly and less effective (Zúñiga-Vicente et al. 2019; Katzman 2014). Cost stickiness, while strategic in some cases, can become a liability during economic downturns, as maintaining unutilized resources can strain a firm’s financial health. Proposed strategies, like cautious resource allocation, may work in stable conditions but could face challenges in volatile economies (Banker et al. 2013). Managerial overconfidence can exacerbate cost stickiness, particularly in volatile economic environments where cost-cutting is essential but may be overlooked (Chen et al. 2022).

The present research will contribute to the literature in two significant ways. First, it sheds light on the relationship between product diversification strategy and cost stickiness, while prior research (Chu et al. 2021; Zhong et al. 2020) has focused on other business strategies. Therefore, the findings of this study can expand the literature on cost behaviour and product diversification strategy. Secondly, by investigating the effect of managerial overconfidence, the present research contributes to the literature on the psychological factors affecting decision-making.

There is an accumulated body of literature on diversification, yet the results are ambiguous and inconclusive (Askarany and Spraakman 2020; Blavatskyy 2023; Denuit et al. 2015; Galavotti et al. 2017; Guidolin and Rinaldi 2013; Wang et al. 2020). Askarany and Spraakman (2020) have investigated the relationship between diversification and financial performance through an excess-capacity theory lens. They found a positive and significant association between diversification and financial performance due to spare capacity (a kind of cost stickiness). In other words, they found that firms with higher excess capacity need to pay less for diversification than those with less or no spare capacity. Therefore, they are more likely to achieve better financial performance than those with less or no excess capacity. This study suggests the possibility of a relationship between diversification and cost behaviour (including cost stickiness).

Other studies explain the diversification through the lens of the Resource-Based View theory (Khurana and Farhat 2021; Situm 2019). The Resource-Based View theory (RBV) posits that a firm’s unique resources and capabilities are the primary drivers of its competitive advantage and performance (Barney et al. 2021). Adopting RBV in this study, we suggest the resource of interest is the diversification strategy, and the capability is the ability to manage cost stickiness.

2.1. Product Diversification Strategy and Cost Stickiness

Traditional models of cost behaviour view cost changes as a direct result of changes in the volume of activity (Noreen 1991). Costs are traditionally classified into fixed and variable costs. A fixed cost remains constant as a company’s sales or production volume changes, while a variable cost changes in proportion to sales and production volume (Anderson et al. 2003). Recent cost stickiness research suggests that the traditional assumption of symmetric cost behaviour does not always hold in practice (Calleja et al. 2006; Lee et al. 2020). A critical review of the findings of Cooper and Kaplan (1998) and Noreen and Soderstrom (1997) introduces a new perspective to cost accounting research. They provided empirical evidence that, in some cases, cost behaviour is asymmetric (Cooper and Kaplan 1998; Noreen and Soderstrom 1997). Later, Anderson et al. (2003) showed that selling, general, and administrative (SG&A) costs decrease less when sales decrease than when sales increase. In other words, these studies state that cost changes are determined by the shift in volume and the direction of activities. This asymmetric cost behaviour is referred to as cost stickiness. The results of Anderson and Lanen (2007) show that SG&A costs increase by an average of 0.55% for a 1% increase in sales but only decrease by 0.35% for a 1% decrease in sales (Anderson and Lanen 2007).

Banker and Byzalov (2014) discussed three factors that cause cost stickiness: agency problems, optimistic managerial expectations, and adjustment costs. Agency problems, where managers prioritize their interests over those of shareholders, can lead to sticky cost behaviour, particularly when managers with empire-building tendencies fail to reduce resources during sales declines (Chen et al. 2012; Zhong et al. 2020). Managerial optimism further intensifies cost stickiness, as optimistic managers are more likely to retain resources even when sales decrease, unlike pessimistic managers who reduce committed resources in response to declining sales. Finally, decisions about resource adjustment are made mainly by the management and significantly impact firm performance (Chen et al. 2022). Hiring/layoff of employees and installation/deployment costs for equipment are examples of adjustment costs. As sales decline, managers tend to retain a certain amount of slack capacity instead of reducing SG&A costs (Banker and Byzalov 2014). Pamplona et al. (2016) found evidence of asymmetric cost behaviour in large South American companies. By analysing the effects of inflation and GDP growth, they also showed that macroeconomic factors explain differences in companies’ cost behaviour. Inflation was negatively associated with cost behaviour, while GDP growth was positively associated with cost behaviour. Zhong et al. (2020) showed that cost stickiness is higher in companies with a differentiation strategy than those with a cost leadership strategy.

Large companies need to make strategic decisions to improve their long-term performance. One of these decisions is product diversification (Azman et al. 2020). The reasons why firms diversify into new businesses and their implications for competitive advantage have been at the core of research on corporate strategy (Becerra et al. 2020). Park and Jang (2013) define diversification as the entry of the company into new markets (sectors, industries) where they previously had no presence (Park and Jang 2013). Zúñiga-Vicente et al. (2019) showed that managers who pursue moderate levels of diversification achieve better profitability during a boom period than those who adopt high or low levels of diversification.

In contrast, moderate and high levels of diversification are equally adequate during an economic downturn. They also found that internationalisation negatively affects profitability, independent of the economic cycle. The effect of product diversification on a firm’s productivity depends on market structure, among other factors. For example, companies operating in a market with little opportunity for growth should diversify by entering different markets. Conversely, diversifying products can be a good strategy for companies operating in a competitive industry. By exploiting and leveraging their resources across multiple markets, companies can gain more competitive advantages and opportunities for further development (Hautz et al. 2013)

The effect of diversification on firm performance has been extensively investigated (Arte and Larimo 2022; Azman et al. 2020). Some studies have reported a positive effect (Kuppuswamy and Villalonga 2016), while others have found a negative impact (Arte and Larimo 2022; Rugman 1976). Lin et al. (2020) examined the effect of operational stickiness on product quality and the moderating effect of product diversification in Chinese exporting firms. The results show an inverted U-shaped relationship between operational stickiness and product quality and that product diversification strategy significantly moderates this relationship. Meanwhile, several studies have reported both positive and negative effects of diversification on performance (Garrido-Prada et al. 2019). For example, Azman et al. (2020) examined the impact of different types of product diversification on firm performance. They found that unrelated diversification reduced company performance in 2003–2016, while related diversification positively affected firm performance during 2010–2016. Similarly, Arte and Larimo (2022) show that the relationship between international diversification and firm performance follows an inverted U-shape and that a firm’s performance is higher with low/related product diversity than with high/unrelated product diversity.

Li and Zheng (2017) show that product market competition increases cost stickiness. They argue that product market competition encourages companies to invest frequently to exploit new investment opportunities under uncertainty (Li and Zheng 2017). Product diversification is also a strategy that requires significant investment in human resources, new machinery and equipment, research and development (R&D), and advertising (Lin et al. 2020). Additionally, this strategy may stem from the agency problem; the top management may try to reduce their employment and reputation risk by diversifying products (Lin et al. 2020). A decrease in sales indicates that the company is losing its competitive position in the market. This can encourage managers to commit more resources to R&D, quality control, marketing, or customer satisfaction to regain the company’s market share and maintain its competitive position in the industry (Li and Zheng 2017).

Moreover, under economic crisis and declining sales, companies may invest less in the related segment and more in unrelated segments to maintain some of their business lines, increasing their internal control, governance, and coordination costs (Chang and Wang 2007). In addition, resource adjustment costs should also be considered. Generally, the cost of increasing resources is lower than the adjustment cost of decreasing resources (Subramaniam and Watson 2016). Due to high adjustment costs, managers are more likely to retain slack resources (Anderson et al. 2003). As a result, costs will not change in proportion to the volume of business activity, leading to cost stickiness (Anderson et al. 2003). Balakrishnan et al. (2004) found that sticky costs are associated with the level of capacity utilisation.

On the other hand, Anderson et al. (2003) showed that companies with more human and physical resources exhibit higher levels of cost stickiness. As Lin et al. (2020) noted, the product diversification strategy requires high human resources and equipment levels. Employment growth increases in the years before diversification, while asset and sales growth increases during and after diversification (Coad and Guenther 2014). Therefore, it can be concluded that resource adjustment costs are higher in companies that adopt a product diversification strategy.

According to Anderson et al. (2003), choosing a business strategy affects cost behaviour. Zhong et al. (2020) documented differences in cost stickiness between companies using differentiation and cost leadership strategies. According to Chen et al. (2022), when sales decline, there is a significant adjustment cost for companies with a product diversification strategy since implementing this strategy requires substantial investment in the company’s structure. Therefore, it can be argued that the product diversification strategy, one of the growth strategies in Ansoff’s matrix (Dawes 2018; Hussain et al. 2013), is associated with cost stickiness. Given the above, we propose our first hypothesis as follows:

H1. 

There is a significant relationship between product diversification strategy and cost stickiness.

2.2. Managerial Overconfidence, Product Diversification and Cost Stickiness

Managerial overconfidence is a behavioural bias that overestimates one’s abilities and underestimates potential risks (Weinstein 1980). Psychological evidence shows that people overestimate the outcomes related to their abilities and underestimate the likelihood of adverse events. This behavioural tendency is called overconfidence (Hyun and Seung 2019). Furthermore, although overconfidence does not equal greater risk-taking, overconfident managers make less conservative decisions by relying too much on positive information (Gervais et al. 2011). Hilary and Hsu (2011) found that managers who accurately forecast earnings in the previous four quarters are less accurate in their subsequent forecasts, suggesting that overconfidence leads managers to be more optimistic. Managers’ cognitive biases affect their decision-making since overconfident managers are optimistic and tend to overestimate returns while underestimating their investment risk (Hilary and Hsu 2011; Killins et al. 2021). As managers become more overconfident, the risk of making wrong decisions increases. This can lead to crises, scandals, and poor performance (Johnson and Fowler 2011; Kunz and Sonnenholzner 2023).

However, some studies have highlighted the benefits of overconfidence. For example, overconfidence motivates managers to pursue promising but risky projects. Overconfident managers tend to pursue innovation and exploit innovative growth opportunities with investment, achieving tremendous innovative success (Hirshleifer et al. 2012). Similarly, Hyun and Seung (2019) found that companies with overconfident managers have higher returns on net operating assets and can better forecast changes in future earnings.

Executives and their personal traits significantly impact organisational capabilities, decision-making, and operations (Kunz and Sonnenholzner 2023). Overconfidence is one of these traits. Previous research shows that managerial overconfidence affects a company’s investment, financing, and financial reporting decisions (Gervais et al. 2011). Overconfident managers are more likely to overestimate returns on investment and underestimate the associated risks (Aabo et al. 2021; Hirshleifer et al. 2012). Empirical studies have provided evidence about the impact of managerial overconfidence on different strategies. Companies with a diversification strategy and overconfident managers experience a decline in value compared to those run by their rational counterparts (Schumacher et al. 2020). Chen et al. (2022) found that managerial overconfidence as an essential factor in resource adjustment decisions is positively associated with cost stickiness.

Similarly, Chu et al. (2021) showed that companies with a differentiation strategy and overconfident managers face higher cost stickiness than those with a cost leadership strategy. Overconfident managers are generally more optimistic and overestimate their abilities to rebound during periods of declining sales (Chen et al. 2022). Therefore, managers’ overconfidence can cause them to ignore cost adjustment, thus increasing cost stickiness (Kuang et al. 2015). Given the above, we propose our second hypothesis as follows:

H2. 

Managerial overconfidence moderates the relationship between product diversification strategy and cost stickiness.

3. Methodology

3.1. Models and Variables

In the cost stickiness literature, the model developed by Anderson et al. (2003) and Banker et al. (2013) is often used to determine the presence or absence of cost stickiness (Haga et al. 2019; Lee et al. 2020). This model is shown in Equation (1).

$$∆ln{COGS}_{it}={\beta }_0+{\beta }_1∆{lnsales}_{it}+{\beta }_2{D}_{it}\ast ∆{lnsales}_{it}+{\epsilon }_{it}$$

(1)

where $∆\ln {COGS}_{it}$ is the dependent variable, equal to the log change in the cost of goods sold for firm $i$ in year $t$; $∆\ln {Sales}_{it}$ is the log change in the sales revenue for firm $i$ in year $t$; $D_{it}$ is a dummy variable that equals one if sales decrease in year $t$, and zero otherwise.

Banker’s model is based on Anderson’s model, which considers resource adjustment decisions under sales increases as a function of managerial discretion and theoretically allows for differences in these decisions under different conditions. In contrast, the traditional model of cost stickiness considers managerial discretion only under declining sales and assumes a mechanical relationship between resource expansion (increase in cost) and increases in sales, which does not account for managerial discretion in decisions to expand resources. Given that our hypothesis considers the possibility of different reactions by managers in various situations (either increase or decrease in sales), our proposed model is adopted from Banker et al. (2013), as shown in Equation (2). Incorporating interaction terms allows us to examine the conditional effects of product diversification based on changes in sales (Δln Sales) and the occurrence of sales decrease (D). In this model, the coefficient ${\beta }_1$ denotes the percentage increase in total cost per 1% increase in sales revenue. The coefficient ${\beta }_2$ captures the degree of asymmetry in cost responses to decreases versus increases in sales. Therefore, ${\beta }_1+{\beta }_2$ show the percentage decrease in total cost per 1% decrease in sales revenue. Cost stickiness is present in a company if ${\beta }_1$ is significant and positive and ${\beta }_2$ is significant and negative. Moreover, if the coefficient of the independent variable (${\beta }_4$) is negative and significant, it suggests that the product diversification strategy reduces the variability of costs concerning changes in revenue and, as a result, increases cost stickiness. Thus, the proposed model is as follows:

$$\begin{gathered} ∆\ln {COGS}_{it}={\beta }_0+{\beta }_1∆\ln {Sales}_{it}+{\beta }_2{D}_{it}\ast ∆\ln {Sales}_{it}+{\beta }_3{PD}_{it}\ast ∆\ln {Sales}_{it}+{\beta }_4{D}_{it}\ast {PD}_{it}\ast ∆\ln {Sales}_{it} \\ +{\beta }_5{ASINT}_{it}\ast ∆\ln {Sales}_{it}+{\beta }_6{D}_{it}\ast {ASINT}_{it}\ast ∆\ln {Sales}_{it}+{\beta }_7{EMPINT}_{it}\ast ∆\ln {Sales}_{it} \\ +{\beta }_8{D}_{it}\ast {EMPINT}_{it}\ast ∆\ln {Sales}_{it}+{\beta }_9{GDP}_t\ast ∆\ln {Sales}_{it}+{\beta }_{10}{D}_{it}\ast {GDP}_t\ast ∆\ln {Sales}_{it} \\ +{\beta }_{11}\sum T_{j}Industry+{\beta }_{12}\sum {\theta }_{j}Year+{\epsilon }_{it} \end{gathered}$$

(2)

where $∆\ln {COGS}_{it}$ is the dependent variable, equal to the log change in the cost of goods sold for firm $i$ in year $t$; $∆\ln {Sales}_{it}$ is the log change in the sales revenue for firm $i$ in year $t$; $D_{it}$ is a dummy variable that equals one if sales decrease in year $t$, and zero otherwise; and ${PD}_{it}$ is product diversification as the independent variable, measured by the Herfindahl index (Tang et al. 2022; Zúñiga-Vicente et al. 2019). The Herfindahl index is calculated as shown in Equation (3):

$$\mathrm{H}\mathrm{e}\mathrm{r}\mathrm{f}\mathrm{i}\mathrm{n}\mathrm{d}\mathrm{a}\mathrm{h}\mathrm{l}=1-\sum _{i=1}^n{p}_i^2$$

(3)

where $p_i^2$ is the ratio of the sales of each segment to the total sales of the company, and $n$ is the number of segments. In Iran, the categorization of industries has been implemented following the standards outlined in the Fourth Edition of the International Standard Industrial Classification (ISIC). The lower the index (i.e., the closer it is to zero), the lower the product diversification.

Following Zhong et al. (2020), Cai et al. (2019), and Anderson et al. (2003), the following control variables are included in the model:

• ${ASINT}_{it}$: Asset intensity, calculated as the ratio of total assets to sales revenue at the end of the fiscal year.
• ${EMPINT}_{it}$: Employee intensity is calculated as the ratio of the number of employees to sales revenue at the end of the fiscal year.
• ${GDP}_t$: Growth rate of gross domestic product.

We expect ${\beta }_6$ and ${\beta }_8$ to be negative since asset- and employee-intensive firms tend to have high-cost stickiness. GDP growth is included to control macroeconomic conditions that might affect the firms’ operations. We expect ${\beta }_{10}$ to be negative since the GDP growth rate increases managers’ optimism regarding future sales, encouraging them to retain slack resources, which could lead to greater cost stickiness (Costa and Habib 2021).

Due to Iran’s inflationary economy, the model is estimated after deflating the primary variables, i.e., sales revenue and costs, based on the inflation rate reported by the Central Bank. Deflation for each year is performed by dividing the sales and expenses of a given year by one plus the inflation rate of that year (MohammadRezaei and Mohd-Saleh 2017). The rationale is to remove the effects of inflation since the proposed cost stickiness model uses sales as a measure of activity level. Therefore, deflated values do not include the increase in sales/costs caused by inflation. In addition, year and industry dummies are included in the model to control for year and industry-fixed effects. The block diagram is presented in Figure 1.

The second hypothesis examines the moderating effect of managerial overconfidence (OCD) on the relationship between product diversification strategy and cost stickiness, as shown in Figure 1. We employ a split-sample method to test this hypothesis, following a similar approach to Costa and Habib (2021). This method divides our statistical sample into companies exhibiting managerial overconfidence and those without. Subsequently, distinct regression models are estimated for each subgroup to examine the differential impact of product diversification on cost stickiness based on the presence or absence of managerial overconfidence. Overconfidence is measured using the index proposed by Killins et al. (2021). This index is a dummy variable equal to one if the capital expenditures scaled by total assets in a given year are greater than the industry median and zero otherwise (Ahmed and Duellman 2013). In the proposed model, a statistically significant difference in coefficients on ${\beta }_4$ indicates the moderating effect of managerial overconfidence.

3.2. Population and Sample

Our statistical population consists of all the companies listed on the Tehran Stock Exchange (TSE) for seven years from 2015 to the end of 2021. Each firm had to meet the following criteria to be included in our sample:

• Financial institutions and intermediaries (e.g., banks, investment companies, etc.) are excluded.
• They were listed on the TSE at the beginning of 2015.
• They had available data and no fiscal year changes during 2015–2021.

TSE is the largest stock exchange in Iran and a reliable source of information (Daryaei et al. 2022; Nassirzadeh et al. 2022; Shandiz et al. 2022).

There were 380 firms listed on TSE by the end of 2021, but only 149 companies met the above criteria that we selected and analysed.

3.3. Descriptive Statistics

The descriptive statistics of the variables are reported in

Table 1. Descriptive statistics of the continuous variables.

Variable	Proxy	Mean	Median	SD	Min.	Max.
Log-change in cost of goods sold.	$∆\ln \mathrm{C}\mathrm{O}\mathrm{G}\mathrm{S}$	0.012	0.013	0.323	−3.298	2.096
Log-change in sales revenue	$∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	0.025	0.022	0.332	−2.709	1.750
Product diversification	$\mathrm{P}\mathrm{D}$	0.471	0.521	0.268	0.000	0.904
Asset intensity	$\mathrm{A}\mathrm{S}\mathrm{I}\mathrm{N}\mathrm{T}$	1.453	1.211	1.084	0.133	9.746
Employee intensity	$\mathrm{E}\mathrm{M}\mathrm{P}\mathrm{I}\mathrm{N}\mathrm{T}$	0.0003	0.0002	0.0003	0.0001	0.003
GDP growth rate	$\mathrm{G}\mathrm{D}\mathrm{P}$	0.018	0.018	0.037	−0.023	0.088

Source: authors’ own work.

and

Table 2. Descriptive statistics of the discreet variables.

Variable	Proxy	Categories	Frequency	Relative Frequency
Sales decrease	$\mathrm{D}$	Decrease	242	0.232
Managerial overconfidence 1	$\mathrm{O}\mathrm{C}\mathrm{D}1$	High OCD	542	0.519
		Low OCD	501	0.481

Source: authors’ own work.

. According to the data, the dispersion of log-change in sales is 0.32. This value is close to the dispersion of log-change in sales revenue (0.33) and indicates low-cost stickiness in the sample companies. According to Herfindahl’s product diversification index (Tang et al. 2022; Zúñiga-Vicente et al. 2019), the findings suggest that manufacturing companies have an average product diversification of 47%. The highest product diversification is 90% (a company with 16 product lines), and some companies have only one product and have not used a diversification strategy. The decrease in the dummy variable for sales shows that 23% of the sample had a reduction in sales, which is close to the result of Zhong et al. (2020). As for the control variables, our results show that the average ratio of assets to sales revenue is more than 1.4, indicating the high asset intensity of the sample companies.

Following Killins et al. (2021), managerial overconfidence is measured using an investment-based index, which is calculated based on the median of capital expenditures. Nearly half of the companies have overconfident managers, similar to Killins et al. (2021) results with the same index. It must be noted that the data are winsorised at the 99% level to reduce the effect of outliers. Also, regression assumptions are tested before estimating the regression models.

4. Results

4.1. Testing the First Hypothesis

The first hypothesis examines the relationship between product diversification strategy and cost stickiness. This hypothesis is tested using panel data after controlling for year and industry-fixed effects. The p-value of the F-statistic is less than 0.01, indicating the significance of the model. This research uses ordinary least squares (OLS) with robust standard errors to solve the problem of autocorrelation of residuals and heteroskedasticity (Petersen 2009). Finally, the variance inflation factor (VIF) is calculated to check for multicollinearity. The VIF values obtained are smaller than ten and indicate the absence of multicollinearity between the independent variables. These results are reported in

Table 3. Estimation results for the first hypothesis.

	Beta	t-Statistic	p-Value	VIF
$\mathrm{C}$	0.06	−2.02	0.045 **	-
${(\beta }_1):∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	0.99	85.56	0.000 ***	1.86
${(\beta }_2):\mathrm{D}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	−0.21	−2.66	0.009 ***	2.59
${(\beta }_3):\mathrm{P}\mathrm{D}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	−0.01	−1.07	0.287	1.29
${(\beta }_4):\mathrm{D}\mathrm{\ast }\mathrm{P}\mathrm{D}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	−0.07	−3.00	0.003 ***	1.35
${(\beta }_5):\mathrm{A}\mathrm{S}\mathrm{I}\mathrm{N}\mathrm{T}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	−0.01	0.20	0.840	9.00
${(\beta }_6):\mathrm{D}\mathrm{\ast }\mathrm{A}\mathrm{S}\mathrm{I}\mathrm{N}\mathrm{T}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	−0.01	−4.50	0.000 ***	8.84
${(\beta }_7):\mathrm{E}\mathrm{M}\mathrm{P}\mathrm{I}\mathrm{N}\mathrm{T}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	−22.97	−1.36	0.177	2.42
${(\beta }_8):\mathrm{D}\mathrm{\ast }\mathrm{E}\mathrm{M}\mathrm{P}\mathrm{I}\mathrm{N}\mathrm{T}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	−2.28	−0.73	0.467	7.66
${(\beta }_9):\mathrm{G}\mathrm{D}\mathrm{P}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	−0.03	−2.96	0.004 ***	1.26
${(\beta }_{10}):\mathrm{D}\mathrm{\ast }\mathrm{G}\mathrm{D}\mathrm{P}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	−0.23	−3.03	0.003 ***	7.62
Year	Controlled
Industry	Controlled
Adjusted R2	0.873
F-statistic	210.84
Sig.	0.000

Source: authors’ own work (**: p < 0.05; ***: p < 0.01).

.

According to these results, the coefficient of $∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$ (${\beta }_1$) is 0.99, which is positive and significant, and the coefficient of $\mathrm{D}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$ (${\beta }_2$) in the model is −0.21, which is negative and significant. This indicates the presence of cost stickiness as per the model adopted by Banker et al. (2013). The economic implication is that costs increase by 0.99% per 1% increase in sales but decrease by 0.78% (0.99% − 0.21%) per 1% decrease in sales. Moreover, given that the coefficient of $\mathrm{D}\mathrm{\ast }\mathrm{P}\mathrm{D}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$ (${\beta }_4$) is negative and significant, the first hypothesis is accepted, indicating that cost stickiness increases with product diversification. Regarding the control variables, the coefficients of $\mathrm{D}\mathrm{\ast }\mathrm{A}\mathrm{S}\mathrm{I}\mathrm{N}\mathrm{T}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$ (${\beta }_6$) and $\mathrm{D}\mathrm{\ast }\mathrm{G}\mathrm{D}\mathrm{P}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$ (${\beta }_{10}$) are negative and significant. This suggests that firms with higher asset intensity and GDP growth rates have greater cost stickiness.

4.2. Testing the Second Hypothesis

The second hypothesis examines the moderating effect of managerial overconfidence on the relationship between product diversification strategy and cost stickiness. The estimation results for the two models related to this hypothesis (two overconfidence states) are reported in

Table 4. Estimation results for the second hypothesis.

	Managerial Overconfidence (OCD1)
	High OCD				Low OCD
Variable	Beta	t-Statistic	p-Value	VIF	Beta	t-Statistic	p-Value	VIF
$\mathrm{C}$	0.001	0.04	0.968	-	0.00	0.00	0.998	-
${(\beta }_1):∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	1.02	8.05	0.000 ***	1.81	0.84	21.87	0.000 ***	1.72
${(\beta }_2):\mathrm{D}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	−0.32	−2.51	0.013 **	2.34	−0.16	−1.99	0.049 **	2.12
${(\beta }_3):\mathrm{P}\mathrm{D}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	−0.14	−0.85	0.398	1.20	0.002	0.37	0.714	1.32
${(\beta }_4):\mathrm{D}\mathrm{\ast }\mathrm{P}\mathrm{D}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	−0.35	−1.89	0.037 **	1.24	−0.04	−2.74	0.007 ***	1.39
${(\beta }_5):\mathrm{A}\mathrm{S}\mathrm{I}\mathrm{N}\mathrm{T}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	−0.02	−0.71	0.481	2.09	−0.00	−0.18	0.856	9.85
${(\beta }_6):\mathrm{D}\mathrm{\ast }\mathrm{A}\mathrm{S}\mathrm{I}\mathrm{N}\mathrm{T}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	−0.05	−1.94	0.054 *	1.23	−0.00	−0.47	0.640	5.86
${(\beta }_7):\mathrm{E}\mathrm{M}\mathrm{P}\mathrm{I}\mathrm{N}\mathrm{T}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	36.26	0.14	0.885	2.65	−15.80	−0.95	0.346	8.61
${(\beta }_8):\mathrm{D}\mathrm{\ast }\mathrm{E}\mathrm{M}\mathrm{P}\mathrm{I}\mathrm{N}\mathrm{T}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	72.24	0.29	0.770	2.19	−53.86	−2.82	0.006 ***	9.64
${(\beta }_9):\mathrm{G}\mathrm{D}\mathrm{P}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	0.25	2.24	0.026 **	1.14	−0.02	−2.09	0.038 **	1.33
${(\beta }_{10}):\mathrm{D}\mathrm{\ast }\mathrm{G}\mathrm{D}\mathrm{P}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	−0.53	−1.99	0.048 **	1.64	−0.22	−2.12	0.036 **	1.07
Year	Controlled				Controlled
Industry	Controlled				Controlled
Adjusted R2	0.868				0.858
F-statistic	919.28				631.46
Sig.	0.000				0.000
z-statistic	−1.904

Source: authors’ own work (*: p < 0.1; **: p < 0.05; ***: p < 0.01).

. The z-test is used to investigate the effect of the moderator variable. This test compares the difference in ${\beta }_4$ between companies with high and low managerial overconfidence. The significant difference in ${\beta }_4$ confirms the moderating effect of managerial overconfidence on the relationship between product diversification and cost stickiness. The z-statistic is obtained using Equation (4):

$$Z={\displaystyle \frac{{\widehat{\beta }}_{4\left(1\right)}-{\widehat{\beta }}_{4\left(2\right)}}{\sqrt{se({\widehat{\beta }}_{4\left(1\right)}{)}^2+se({\widehat{\beta }}_{4\left(2\right)}{)}^2}}}$$

(4)

where ${\widehat{\beta }}_{4\left(1\right)}\mathrm{a}\mathrm{n}\mathrm{d}{\widehat{\beta }}_{4\left(2\right)}$ are the estimated coefficients for high and low overconfidence groups, respectively.

$se{\widehat{\beta }}_{4\left(1\right)}\mathrm{a}\mathrm{n}\mathrm{d}{se\widehat{\beta }}_{4\left(2\right)}$ are the standard errors associated with each coefficient.

The result of the z-test for this measure of managerial overconfidence shows a statistically significant difference in ${\beta }_4$ between companies with high and low managerial overconfidence at the 90% confidence level (p < 0.1). In other words, managerial overconfidence moderates the relationship between product diversification and cost stickiness. This coefficient is more prominent in companies with high managerial overconfidence. Therefore, the second hypothesis is supported.

4.3. Robustness Tests

4.3.1. An Alternative Measure of Managerial Overconfidence

In addition to the indicator variable proposed by Killins et al. (2021), we use the index proposed by Schrand and Zechman (2012) to measure managerial overconfidence. This index is based on excess investment and is the residual from a regression of total asset growth on sales growth, as shown in Equation (5):

$${ASSETGR}_{i,t}={\beta }_0+{\beta }_1{SALEGR}_{i,t}+{\epsilon }_{i,t}$$

(5)

This index takes the value of one if $\epsilon >0$ (i.e., if assets grow at a higher rate than sales, indicating overinvestment due to managerial overconfidence) and zero otherwise. According to the data, about 43% of companies have overconfident managers, nearly similar to the first index. In this alternative model, the magnitude of the difference in ${\beta }_4$ indicates the significant moderating effect of managerial overconfidence. The results are reported in

Table 5. Estimation results using an alternative measure of managerial overconfidence.

	Managerial Overconfidence (OCD2)
	High OCD				Low OCD
Variable	Beta	t-Statistic	p-Value	VIF	Beta	t-Statistic	p-Value	VIF
$\mathrm{C}$	−0.05	−1.09	0.278	-	−0.03	−1.81	0.073 *	-
${(\beta }_1):∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	0.97	45.15	0.000 ***	1.81	0.88	23.34	0.000 ***	2.29
${(\beta }_2):\mathrm{D}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	−0.20	−2.21	0.028 **	1.51	−0.10	−1.80	0.074 *	1.14
${(\beta }_3):\mathrm{P}\mathrm{D}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	0.01	0.71	0.476	1.43	−0.01	−0.71	0.477	1.24
${(\beta }_4):\mathrm{D}\mathrm{\ast }\mathrm{P}\mathrm{D}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	−0.07	−2.90	0.004 ***	1.08	−0.04	−2.38	0.018 **	1.08
${(\beta }_5):\mathrm{A}\mathrm{S}\mathrm{I}\mathrm{N}\mathrm{T}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	−0.00	−0.38	0.705	2.06	−0.00	−1.21	0.229	8.97
${(\beta }_6):\mathrm{D}\mathrm{\ast }\mathrm{A}\mathrm{S}\mathrm{I}\mathrm{N}\mathrm{T}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	−0.02	−2.04	0.043 **	1.51	−0.01	−3.59	0.000 ***	7.58
${(\beta }_7):\mathrm{E}\mathrm{M}\mathrm{P}\mathrm{I}\mathrm{N}\mathrm{T}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	−38.21	−1.79	0.076 *	1.82	6.54	0.41	0.680	4.55
${(\beta }_8):\mathrm{D}\mathrm{\ast }\mathrm{E}\mathrm{M}\mathrm{P}\mathrm{I}\mathrm{N}\mathrm{T}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	−131.95	−8.56	0.000 ***	1.06	−26.54	1.95	0.053 *	1.57
${(\beta }_9):\mathrm{G}\mathrm{D}\mathrm{P}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	−0.03	−2.31	0.022 **	1.27	−0.02	−2.99	0.003 ***	1.15
${(\beta }_{10}):\mathrm{D}\mathrm{\ast }\mathrm{G}\mathrm{D}\mathrm{P}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	−0.06	−0.30	0.768	1.08	−0.01	−0.54	0.592	1.89
Year	Controlled				Controlled
Industry	Controlled				Controlled
Adjusted R2	0.723				0.863
F-statistic	607.27				658.94
Sig.	0.000				0.000
z-statistic	−1.788

Source: authors’ own work (*: p < 0.1; **: p < 0.05; ***: p < 0.01).

.

The result of the z-test for the second measure of managerial overconfidence indicates a significant difference in ${\beta }_4$ between companies with high and low managerial overconfidence at the 90% confidence level (p < 0.1). Again, this coefficient is more prominent in companies with high managerial overconfidence. Therefore, these results support our baseline findings.

4.3.2. An Alternative Measure of Costs

As another sensitivity test, we used SG&A costs instead of the cost of goods sold, as shown in Equation (6). According to the data, log-change dispersion in SG&A costs is 0.37. This value is close to the dispersion of log-change in sales revenue (0.33) and indicates low-cost stickiness in the sample companies. The results are reported in

Table 6. Estimation results using an alternative measure of costs (SG&A costs).

Variable	Beta	t-Statistic	p-Value	VIF
$\mathrm{C}$	0.35	−4.58	0.000 ***	-
${(\beta }_1):∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	0.92	31.60	0.000 ***	1.60
${(\beta }_2):\mathrm{D}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	−0.11	−5.39	0.000 ***	7.34
${(\beta }_3):\mathrm{P}\mathrm{D}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	−0.03	−2.03	0.044 **	1.67
${(\beta }_4):\mathrm{D}\mathrm{\ast }\mathrm{P}\mathrm{D}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	−0.05	−1.78	0.077 *	5.26
${(\beta }_5):\mathrm{A}\mathrm{S}\mathrm{I}\mathrm{N}\mathrm{T}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	0.01	1.01	0.316	9.28
${(\beta }_6):\mathrm{D}\mathrm{\ast }\mathrm{A}\mathrm{S}\mathrm{I}\mathrm{N}\mathrm{T}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	−0.01	−1.68	0.066 *	9.75
${(\beta }_7):\mathrm{E}\mathrm{M}\mathrm{P}\mathrm{I}\mathrm{N}\mathrm{T}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	−66.32	−3.29	0.001 ***	7.04
${(\beta }_8):\mathrm{D}\mathrm{\ast }\mathrm{E}\mathrm{M}\mathrm{P}\mathrm{I}\mathrm{N}\mathrm{T}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	34.96	1.40	0.163	9.53
${(\beta }_9):\mathrm{G}\mathrm{D}\mathrm{P}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	−0.07	−3.86	0.000 ***	1.41
${(\beta }_{10}):\mathrm{D}\mathrm{\ast }\mathrm{G}\mathrm{D}\mathrm{P}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	−0.14	−2.86	0.005 ***	1.56
Year	Controlled
Industry	Controlled
Adjusted R2	0.690
F-statistic	111.79
Sig.	0.000

Source: authors’ own work (*: p < 0.1; **: p < 0.05; ***: p < 0.01).

.

$$\begin{gathered} ∆\ln {SG\&A}_{it}={\beta }_0+{\beta }_1∆\ln {Sales}_{it}+{\beta }_2{D}_{it}\ast ∆\ln {Sales}_{it}+{\beta }_3{PD}_{it}\ast ∆\ln {Sales}_{it}+{\beta }_4{D}_{it}\ast {PD}_{it}\ast ∆\ln {Sales}_{it} \\ +{\beta }_5{ASINT}_{it}\ast ∆\ln {Sales}_{it}+{\beta }_6{D}_{it}\ast {ASINT}_{it}\ast ∆\ln {Sales}_{it}+{\beta }_7{EMPINT}_{it}\ast ∆\ln {Sales}_{it} \\ +{\beta }_8{D}_{it}\ast {EMPINT}_{it}\ast ∆\ln {Sales}_{it}+{\beta }_9{GDP}_t\ast ∆\ln {Sales}_{it}+{\beta }_{10}{D}_{it}\ast {GDP}_t\ast ∆\ln {Sales}_{it} \\ +{\beta }_{11}\sum T_{j}Industry+{\beta }_{12}\sum {\theta }_{j}Year+{\epsilon }_{it} \end{gathered}$$

(6)

According to these results, the coefficient of $∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$ (${\beta }_1$) is 0.92, which is positive and significant, and the coefficient of $\mathrm{D}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$ (${\beta }_2$) in the model is −0.11, which is negative and significant. Consistent with the baseline findings, these results indicate the presence of cost stickiness. Moreover, given that the coefficient of $\mathrm{D}\mathrm{\ast }\mathrm{P}\mathrm{D}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$ (${\beta }_4$) is negative and significant, and the first hypothesis is accepted, this suggests meaningful positive relationship between product diversification and cost stickiness. Similarly, the coefficients of the control variables $\mathrm{D}\mathrm{\ast }\mathrm{A}\mathrm{S}\mathrm{I}\mathrm{N}\mathrm{T}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$ (${\beta }_6$) and $\mathrm{D}\mathrm{\ast }\mathrm{G}\mathrm{D}\mathrm{P}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$ (${\beta }_{10}$) are negative and significant, suggesting that firms with higher asset intensity and GDP growth rates have greater cost stickiness.

4.3.3. Incorporating COVID-19 as a Moderator

To address the unique contextual factors stemming from the COVID-19 pandemic, we included a COVID-19 variable as a moderator in our analysis. The results are reported in

Table 7. COVID-19 and the relationship between product diversification strategy and cost stickiness.

	COVID 19
	Pre COVID 19				COVID 19 Period
Variable	Beta	t-Statistic	p-Value	VIF	Beta	t-Statistic	p-Value	VIF
$\mathrm{C}$	−0.48	−1.98	0.05 *	-	0.003	0.22	0.83	-
${(\beta }_1):∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	0.88	12.23	0.000 ***	1.92	0.87	25.91	0.000 ***	1.12
${(\beta }_2):\mathrm{D}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	−0.25	−2.69	0.008 ***	2.52	−0.05	−1.96	0.052 *	2.82
${(\beta }_3):\mathrm{P}\mathrm{D}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	−0.01	−1.58	0.117	1.31	−0.001	−0.25	0.802	1.72
${(\beta }_4):\mathrm{D}\mathrm{\ast }\mathrm{P}\mathrm{D}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	−0.03	−1.74	0.084 *	1.1	−0.04	−3.22	0.002 ***	1.91
${(\beta }_5):\mathrm{A}\mathrm{S}\mathrm{I}\mathrm{N}\mathrm{T}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	−0.003	−1.26	0.209	8.23	−0.001	−0.64	0.525	8.13
${(\beta }_6):\mathrm{D}\mathrm{\ast }\mathrm{A}\mathrm{S}\mathrm{I}\mathrm{N}\mathrm{T}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	−0.05	−2.32	0.022 **	6.01	−0.02	−2.19	0.03 **	5.81
${(\beta }_7):\mathrm{E}\mathrm{M}\mathrm{P}\mathrm{I}\mathrm{N}\mathrm{T}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	−11.66	−0.31	0.757	8.99	12.11	1.4	0.163	8.01
${(\beta }_8):\mathrm{D}\mathrm{\ast }\mathrm{E}\mathrm{M}\mathrm{P}\mathrm{I}\mathrm{N}\mathrm{T}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	−51.57	−1.22	0.226	7.28	−3.77	−1.89	0.061 *	8.14
${(\beta }_9):\mathrm{G}\mathrm{D}\mathrm{P}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	−0.02	−2.25	0.026 **	1.38	−0.02	−3.61	0.000 ***	1.18
${(\beta }_{10}):\mathrm{D}\mathrm{\ast }\mathrm{G}\mathrm{D}\mathrm{P}\mathrm{\ast }∆\ln \mathrm{S}\mathrm{A}\mathrm{L}\mathrm{E}\mathrm{S}$	−0.048	−1.97	0.051 *	1.92	−0.11	−2.26	0.025 **	2.07
Year	Controlled				Controlled
Industry	Controlled				Controlled
Adjusted R2	0.781				0.855
F-statistic	508.17				680.84
Sig.	0.000				0.000
z-statistic	0.55

Source: authors’ own work (*: p < 0.1; **: p < 0.05; ***: p < 0.01).

. This additional analysis allowed us to explore how the pandemic may have influenced the relationship between our critical variables between 2020 and 2023.

The result of the z-test indicates that there is no significant difference in (${\beta }_4$) between companies affected by the COVID-19 pandemic and those unaffected by it. Our analysis suggests that the COVID-19 pandemic did not significantly impact the relationship between product diversification and cost stickiness.

5. Discussion

This study significantly enhances the understanding of the relationship between product diversification strategy, cost stickiness, and the moderating role of managerial overconfidence, particularly within the context of the Resource-Based View (RBV) theory. The existing literature on diversification and its implications has often yielded inconclusive results, failing to capture this complex relationship comprehensively. Some studies suggest that diversification enhances firm performance by creating synergies and spreading risk (Krivokapić et al. 2017; Bhatia and Thakur 2018), while others argue that diversification can lead to inefficiencies (Dawid and Reimann 2011). These mixed findings highlight the complexity of the relationship between diversification and firm performance, mainly when cost behaviour is considered. Our study addresses this gap by exploring the interplay between diversification and cost stickiness, offering a new perspective and valuable insights.

Our findings reveal a substantial and previously uncharted link between product diversification strategy and cost stickiness, particularly in manufacturing companies. This connection aligns with the RBV theory, which emphasizes the significance of unique resources and capabilities in achieving a competitive advantage (Barney et al. 2011). In this case, product diversification represents a unique resource, while cost stickiness reflects a capacity related to resource management and risk mitigation. Thus, this study underscores the theoretical underpinnings of RBV by empirically validating its principles in the context of corporate strategies. This finding is consistent with Banker and Byzalov (2014), who suggest that firms with greater resource commitments are more likely to exhibit cost stickiness due to the perceived value of maintaining resources in anticipation of future opportunities.

Moreover, this study uncovers the moderating influence of managerial overconfidence on the relationship between diversification and cost stickiness. This aspect further substantiates RBV, emphasizing the role of managerial decisions and perceptions in resource allocation and risk management. Our findings are consistent with those of Chen et al. (2022) as well as Keke (2021), showing that managers’ overconfidence impacts their investment decisions and costs (leading to a more pronounced cost stickiness in companies with product diversification strategies). This empirical validation of managerial overconfidence as a moderating factor reinforces the RBV’s emphasis on the role of internal resources and managerial decisions in shaping firm performance.

6. Conclusions

This study aimed to investigate the relationship between product diversification and cost stickiness, focusing on the moderating role of managerial overconfidence within the context of Iranian manufacturing companies. Through our analysis, we found that product diversification is significantly associated with increased cost stickiness, and this relationship is further amplified by managerial overconfidence. These findings suggest that strategic diversification and resource management decisions are critical in shaping a company’s cost behaviour, especially under economic uncertainty.

This study’s implications are far-reaching. It highlights the critical role of strategic choices in determining cost stickiness, providing valuable insights for shareholders, creditors, and stakeholders in assessing a company’s financial behaviour and risk exposure. Additionally, this study offers auditors a deeper understanding of cost behaviour and its determinants, enhancing their analytical procedures.

For analysts and users of financial statements, this study underscores the importance of considering managerial overconfidence and cost stickiness when evaluating a company’s financial performance and risk profile. Managers are urged to understand external changes and cost control better to improve their strategic positioning and management efficiency. Lastly, this study raises intriguing questions about the optimal level of product diversification, prompting further exploration in this area.

While this research contributes significantly to the literature, it has some limitations. The reliance on financial statements and the absence of face-to-face interactions with respondents pose constraints. Additionally, this study’s focus on Iranian companies may limit its generalizability to other contexts. Iran’s economic and regulatory environment might have influenced the findings, meaning that they should be interpreted with care when considering their applicability to different contexts. Furthermore, the data were modified through deflation to adjust for inflation. This may have influenced the results by providing a more precise reflection of real economic values rather than nominal figures. This adjustment could potentially affect the observed relationships, as it eliminates the distorting effects of inflation, allowing for a more accurate analysis. However, this adjustment assumes a uniform impact of inflation across all firms, potentially overlooking sector-specific differences. Future research can extend these findings by examining different types of product diversification, exploring internationalization strategies, and delving into moderating variables, including other managerial factors, market competition, and intensity of technological innovation, that influence cost stickiness and risk management, thus further enriching the RBV theory’s application in diverse contexts.