Financial and marketing approach to export performance: the mediation role of promotion and research and development

Abstract

The current study investigates the direct and indirect effects of firm-specific and environmental factors on export performance. Specifically, the study investigates the direct and indirect (mediation role of promotion (PR) and research & development (RD) activities) effect of capital structure (CS), availability of cash (CA), competition from informal sector (CI), and political instability in home countries (PI) on export performance (EXP). Partial least square structural equation model (PLS-SEM) is used to analyze a year data (2022) from sample of 161 firms in Ethiopian manufacturing industry (i.e. Leather industry, textile industry, and Food & Beverage industry), to find the relation between independent (CS, CA, CI, and PI), mediating (PR and RD), and dependent (EXP) variables after controlling for factors like firm size, and industry category. The result indicates that CS has robust direct and indirect effects on EXP while PI has only direct effect. However, CA needs the intermediation of other variables (i.e. RD) to affect EXP. This finding supports the argument that assessments of direct relationships should not be the primary concern in marketing studies; rather, the sum of the direct and indirect effects of a particular variable must be evaluated for more interpretation.

Keywords:

• Capital structure
• cash
• promotion
• research & development

REVIEWING EDITOR:

• Huifen (Helen) Cai, Middlesex University Business School, United Kingdom

SUBJECTS:

• Economics; Finance; Business, Management and Accounting

1. Introduction

In recent decades, an increasing number of firms have embraced international expansion through exports in response to globalization, major markets, and trade liberalization initiatives in many countries, resulting in significant changes in the world economy (Buckley & Strange, 2015). Exporting represents a feasible strategic alternative for firms to internationalize, and has remained the most frequently used foreign market entry mode chosen (Zhao & Zou, 2002). It is vital to affirm a firm’s survival or growth while positively influencing current and future EXP by ensuring competitive advantage in international markets (Navarro et al., 2010). While EXP is regarded as one of the key indicators of the success of a firm’s export operations, it reflects firm-specific behavior in leveraging its resources and capabilities in an international context at a given point in time (Beleska-Spasova et al., 2012).

It has been long documented that the investigation on EXP determinants is of vigorous interest to public-policy-makers, practitioners, and researchers (Katsikeas et al., 2000). A number of recent empirical studies have made an effort to investigate the impact of different internal and external factors on EXP (e.g. Bodlaj et al., 2020; Durmaz & Eren, 2018; Ferreras-Méndez et al., 2019; Karami & Tang, 2019; Malca et al., 2020; Osiyevskyy et al., 2020; Racela & Thoumrungroje, 2020; Sousa et al., 2020). This large volume of publications is a strong testimony of not only the importance of the issue but also the legitimacy of inquiry into export marketing. The recognition is well reflected not only by the fact that exporting research has been flourished recently, but also by the sheer number of publications related to exporting (Chen et al., 2016). However, despite these developments, some important issues still require attention in extant research. First, a review of recent publications indicates that, among the studies that investigate internal factors of firm EXP determinants, with a few recent notable exceptions (i.e. Caban-Garcia et al., 2020; Chosiah, Purwanto & Ermawati 2019; Pietrovito & Pozzolo, 2021), most of the studies have failed to consider the role of firms’ financial variables such as CS and CA in determining firms’ EXP. Nonetheless, assets and capabilities are the two essential resources that are imperative to create competitive advantage (Gao, Murray, Kotabe, & Lu, 2010).

As pointed out by Strebulaev (2007), small adjustment costs may cause large variations in CS. Because interest on debt is a tax-deductible expense, the firm effectively reduces its tax bill as it employs more debt. Consequently, the cost of capital will not rise, even if the use of leverage increases to excessive levels. The Trade-off theories of CS envisage that firms choose levels of debt in order to balance the benefits from the interest tax shield with the costs of future financial distress or of current financial inflexibility (Danis et al., 2014). The relationship between sources of financing firms’ business operations and firm performance is well documented. Thus, we theorizes that there is strong relationship between EXP and its CS.

The difference between firm’s cash inflow and outflow indicate the firm’s CA (Ogbeide & Akanji, 2017). Cash policies unquestionably linked to the firm’s operations (Kroes & Manikas, 2014). It is widely linked to firm’s financial performance. The Principal working capital management theory advocates (Brewer & Speh, 2000; Theodore Farris & Hutchison, 2002) contend that firms can improve liquidity, and hence their competitive positioning by handling their cash flows. Thus, we presume that firms should regularly change the extent at which cash is spent to insure CA and enhance their international operations to obtain higher performance.

Second, strategically export success of firms depends not only on its controlled resources and capabilities, but also on how the institutional and industrial environment can configure its behaviors through cognitive, normative, and regulative mechanisms (Welter, 2011; Welter & Smallbone, 2011). However, despite the urge by Leonidou and Katsikeas (2010), only few studies have tried to inject a theoretical perspective in researching institutional and industrial environment effects on export performance, which has been endemic to empirical research in the overall exporting field. The lack of empirical validation of integrative model of both internal (firm-level) and external factors-EXP relationships on a comprehensive pool is also a major limitation in the existing literature (see Chosiah et al., 2019; Pietrovito & Pozzolo, 2021).

This leads the researcher to posit that the Resource based theory (RBT), Dynamic capability view (DCV), and Institution-based view (IBV) can provide a suitable theoretical platform to explain and predict the firm’s performance in international markets (Hoskisson et al., 2000; Peng et al., 2009; Wright et al., 2005).

The RBT has emerged in the past decades as a very influential framework to analyze the sources and sustainability of a firm’s competitive advantage in export research (Nakos et al., 2019). It accentuates how organizations achieve viable competitive advantage by developing resources and capabilities (Hennart, 1991). It has been a central foundation of firms’ internationalization theories in developing the concept of firm-specific advantages (Dunning, 1988). DCV are ‘the organizational and strategic routines by which firms achieve new resource configurations as markets emerge, collide, split, evolve and die’ (Eisenhardt & Martin, 2000). DCV integrates the strategic management with entrepreneurial orientation, which a significant application to diverse and changing contexts as is international markets (Knight & Liesch, 2016). The study also founded on IBV to argue that political instability and competition from informal sector in home countries directly affect firm export performance. IBV highlights the significance of institutional environment, and advocates that institutional forces shape firms’ strategic decisions and determine their performance (Tina Dacin et al., 2002).

Thus, founded on RBV, DCV, and IBV, the researchers contend that CS, CA, CI in home countries, and PI in home countries directly affect firm EXP. Hence, this study reveals the direct and indirect impact of CS, CA, CI, and PI in home countries on EXP. Particularly, this study examines the marketing effort route through which CS and CA influences EXP. The marketing effort routes focus on the mediating role of PR and RD activities in explaining the relationship between CS, CA, and EXP.

The contribution of this study is twofold. First, on the methodological front, this study is focused on advancing and empirically validating a RBT, DCV, and an IBV integrative model of export performance. Particularly, the study focuses on investigating the direct and indirect (marketing mix strategy route) effect of EXP antecedents. This holistic analysis approach could make it possible to estimate the complicated associations among critical internal and external factors, consequently providing more sophisticated results as compared to the previous studies. Second, on the practical and policy fronts, this study used data from firms in developing countries for empirical validation of the theoretical model, which could enhance our understanding of how firms’ specific and institutional factors drive EXP in a developing economy. Specifically, this study provides an important insight for practitioners and/or managers in developing countries when venturing abroad carefully considering the consequences of financial variables, political instability, and competition from informal sector firms’ EXP relationships. Similarly, it helps policymakers identify the important factors to consider in enhancing the EXP of firms and countries. Thus, it is in this way this study contributes to international marketing literature thereby enhancing our understanding of determinants of EXP while facilitating theory development.

2. Theoretical background and research hypotheses

As this study intends to investigate the impact of firm-specific and institutional factors on export performance, it is founded on the confluence of three related theories, RBT, DCV, and IBV. The RBT has emerged in the past decades as a very influential framework to analyze the sources and sustainability of a firm’s competitive advantage in export research (Nakos et al., 2019). RBT accentuates how organizations achieve viable competitive advantage by developing resources and capabilities (Hennart, 1991). It has been a central foundation of firms’ internationalization theories in developing the concept of firm-specific advantages (Dunning, 1988). This study focuses on the role that firms’ CS and CA play in determining export performance. It primarily focuses on examining the direct impact of CS and CA on EXP. Nonetheless, the researcher also argues that a firm’s success in a foreign market depends not only on its given portfolio of resources and capabilities, as per the RBT, but also on its capacity and ability to constantly change and adjust to international uncertainties. Thus, the DCV approach introduced by Teece et al. (1997) appears to offer a more dynamic perspective of the RBV. DCV is defined as ‘the firm’s ability to integrate, build, and reconfigure internal and external competences to address rapidly changing environments’. DCV are ‘the organizational and strategic routines by which firms achieve new resource configurations as markets emerge, collide, split, evolve and die’ (Eisenhardt & Martin, 2000).

Unlike traditional economic development theorists which stress resource accumulation, the DCV framework stresses the importance of enterprise-level entrepreneurship, innovation, learning, and good strategy (Teece, 2014). Consequently, the DCV integrates the strategic management with entrepreneurial orientation, which a significant application to diverse and changing contexts as is international markets (Knight & Liesch, 2016). Furthermore, Theory of strategy-environment co-alignment (Porter, 1980; Venkatraman & Prescott, 1990), states that the ‘'fit’’ between strategy and its context-whether it is the external environment (Anderson & Zeithaml, 1984) or organizational characteristics (Gupta & Govindarajan, 1984) has significant positive implications for firm performance. This study complements the DCV with the RBT and stations within the recent approaches of the RBT and the DCV, and adopts an integrative perspective of resource orientations and dynamic capabilities to explain export performance.

In addition, as export activities are subject to different institutional forces in the host and export markets (Peng et al., 2008), the theoretical model also founded on IBV to argue that political instability and competition from informal sector in home countries directly affect firm export performance. The IBV emphasizes the importance of institutional environment, and suggests that institutional forces shape firms’ strategic decisions and determine their performance (Tina Dacin et al., 2002).

2.1. CS and EXP link

Among the definitions, CS is described as the mix of debt and equity that a company uses to finance its operations. This can be understood as the balance between all of the company’s Liabilities + Equities, and thus concerns the entire ‘Liabilities + Equities’ side of the balance sheet. The concept of CS received much attention after Modigliani and Miller (1958) demonstrated in their paper that the choice between debt and equity does not have any material effects on the firm’s value. However, Strebulaev (2007) points out that small adjustment costs may cause large variations in long-term debt to equity. Modigliani and Miller (1963) subsequently corrected their irrelevance proposition for taxes. Because interest in debt is tax-deductible, a firm effectively reduces its tax bill as it employs more debt. This implies that the cost of capital will not increase even if leverage increases to excessive levels.

However, Solomon (1963) argues that in an extreme leverage position, the cost of capital must increase. This is because excessive debt induces markets to react by demanding higher rates of return. Therefore, to minimize the weighted average cost of capital, firms avoid a pure debt position and seek an optimal mix of debt and equity. In conclusion, the literature on the CS debate has progressed from the irrelevant propositions of Modigliani and Miller (1958) to counter arguments based on more realistic assumptions. The train of thought in CS irrelevance propositions has developed to a general consensus among researchers that there is an optimal CS.

As the Trade-off theories of CS envisage, firms choose levels of debt in order to balance the benefits from the interest tax shield with the costs of future financial distress or of current financial inflexibility (Danis et al., 2014). The relationship between the sources of financing firms’ business operations and firm performance is well-documented. The use of borrowed capital increases the level of investment undertaken by the firm without incurring any additional costs for firm owners other than interest expenses (Eriotis et al., 2011). However, despite the fact that borrowed capital increases the return on invested capital, it also increases the risk for firms as well as for owners because borrowed capital creates fixed expenses (i.e. interest); thus, a minimum profit level is necessary to finance the level of interest. Empirical evidence (Salim & Yadav, 2012) confirms the negative relationship between firm performance and short-term, long-term, and total debt. There is a significant negative relationship between financial leverage, measured by short-term debt to total assets, total debt to total assets, and firm performance (Khan, 2012). There is evidence of a negative, statistically, and economically significant effect of financial constraints on both the probability that a firm exports (the extensive margin) and the share of exports over total sales (the intensive margin) (Pietrovito & Pozzolo, 2021).

Contrary to Khan’s (2012) findings, the results suggest that financial leverage measured by both short- and long-term debt has a positive and significant effect on firm performance (Fosu, 2013). However, there is also evidence that the relationship between CS and corporate performance depends on firm size of the firms (San & Heng, 2011). Similarly, Degryse et al. (2012) argue that small and medium enterprises (SMEs) use profits to reduce their debt level, and growing firms increase their debt position since they need more funds. These contradictory findings can probably be attributed to firm size. The indicators of CS regard the repartition between equity and debt in the company’s resources. The researcher argues that the burden of making regular interest payments to debt holders can pressure managers to cut the short-term investment required to enhance a firm’s EXP. Thus, the researcher theorizes that there is a strong relationship between EXP and CS. This leads to the following hypothesis:

Hypothesis 1:

There is significantly negative relationship between firm capital structure with high debt and export performance

2.2. CA and FXP link

CA is the net amount of cash and cash equivalents transferred into and out of business. Cash in organizations usually takes two directions: cash inflow and outflow. The difference between a firm’s cash inflow and outflow indicates its CA (Ogbeide & Akanji, 2017). This means that cash comes from customers or clients who buy products or services. Cash is going out of business in the form of payments for expenses, such as rent or mortgages, in monthly loan payments, and payments for taxes and other accounts payable.

Firms’ cash policies, which manage working capital in the form of cash receivables from customers, inventory holdings, and cash payments to suppliers, are inevitably linked to their operations (Kroes & Manikas, 2014). Previous studies have investigated how adjustments to a firm’s CA can change its performance. This is linked to a firm’s financial performance. Predominant working capital management theory promoters (Brewer & Speh, 2000; Ii & Hutchison, 2003; Theodore Farris & Hutchison, 2002) argue that firms can improve liquidity and hence their competitive positioning by handling their cash flows. Firm CA is determined by the difference between the working capital and working capital needs. If it is positive, it means that the company has sufficient liquidity to face all its short-term obligations, which will be more important when the company increases and intensifies its export activity (Maurel, 2009). This means that if more money is coming in than is going out, there will be a ‘positive cash position’ situation, which will help firms pay for bills. Thus, based on extant studies, the researcher presumes that firms should regularly change the extent to which cash is spent to ensure CA, enhances their international operations and obtains higher performance. Hence, in this sense the researcher proposes the following hypothesis:

Hypothesis 2:

There is significantly positive relationship between availability of cash and firm export performance

2.3. CI And EXP link

Informal sectors are broadly defined as economic activities not recorded in official GDP statistics (London & Hart, 2004). The size of these informal sectors can be significant even in advanced economies, but are typically more substantial in emerging and developing economies (Schneider et al., 2011). Nonetheless, the fact that a considerable informal sector may generate positive benefits in terms of employment, welfare, and the provision of local services, the competitive effect on ‘formal’ (i.e. officially registered) is likely to be negative and considerable (Schneider & Enste, 2000). The existence of a sizeable informal sector makes macroeconomic policy less effective (Mara, 2011), restricting the domestic prospects of formal growth and impeding economic growth in these markets (Lamanna, 2007). In addition, business activities within the informal sector are typically conducted outside official law, with informal social contracts being used as binding arrangements, and firms in the formal sector may find it difficult to protect their proprietary knowledge and technology through enforceable legal mechanisms (McCann & Bahl, 2017). Given these discriminatory conditions, many firms in the formal sector choose to look for more level playing fields in the presence of strong informal competition. Thus, the researcher’s hypothesis was as follows:

Hypothesis 3:

There will be a negative relationship between the degree of competition from the informal sectors in their home countries and the formal firm export performance.

2.4. PI and EXP link

Firm behavior and strategy both at home and abroad is an ultimate result of institutional configurations (Peng et al., 2008). Hoskisson et al. (2000) argued that well-developed institutions create a favorable business environment with low transaction costs and highly competitive pressures that favor efficiency and innovativeness. The institutional environments in different countries typically differ in many ways, including unstable political environments, pervasive government influence, non-transparent regulatory infrastructure, under-developed capital and labor markets, and greater informality (Rottig, 2016). The collective effects of these institutional deficits mean that doing business in different economies is often beset by opportunistic behavior, unpredictable government policies, inflated transaction costs, and higher levels of uncertainty (Gao et al., 2010). Based on the existing literature (e.g. Schneider et al., 2011), the researcher identifies the attributes of the institutional environments within home countries that need consideration in relation to firm export performance: political instability. Thus, the researcher argues that these specific institutional features are particularly relevant in determining firm EXP because they introduce uncertainty in their domestic markets. PI has several adverse effects on business activities in an economy (Arráiz et al., 2013). To the extreme, PI may give rise to political and/or civil violence, which further worsens uncertainty and leads to even greater operational difficulties (Hiatt & Sine, 2014). Thus, we propose the following hypothesis:

Hypothesis 4:

There will be a negative relationship between political instability in their home countries and the firm export performance.

2.5. The mediation role of PR and RD

As the issue of marketing strategy advances increasing eminence as an orientation that everyone in the organization shares and as a process that all functions participate in deploying, a critical issue that arises is the role of the marketing function. The marketing literature to date has focused on the direct effect of export performance determinants. Evidence from recent studies shows, only few studies (e.g. Boso et al., 2019; Costa et al., 2015; Durmaz & Eren, 2018; Hollender et al., 2017) weigh up the indirect effect of the antecedents on export performance. This study argues that firm PR and RD activities could explain the link between CS, CA and EXP.

Regarding the mediating role of PR, there has been a steady stream of research studying the impact of firm promotion activity on performance. Specifically, prior studies examine the contemporaneous association between advertising expenses and EXP (Gopinath et al., 2013), Sridhar et al., 2014). However, strategic management scholars argue that firms’ marketing strategy influenced by the resource and capability of the firm. Leonidou et al. (2011) indicated the interrelationship of the export-related organizational resources and their impact on EXP. A recent study also argues that resources available for export activities affect export marketing strategy and performance (Imiru, 2018). In addition, Maritan and Lee (2017) encourage more research that examines resource allocation as a central focus of study for achieving deeper and better understandings about firm strategies. Prior research (e.g. Joshi & Hanssens, 2010; Luo & Bhattacharya, 2006; McAlister et al., 2007) has recognized firm promotion activity as key components of a firm’s marketing effort. Advertising creates market-based assets such as brand equity by informing customers, differentiating products from competition, and creating barriers to entry (Narasimhan et al., 2006). Recent empirical evidence shows that leverage leads to lower customer satisfaction, with advertising intensity mediating this effect (Malshe & Agarwal, 2015).

Other mediating factor considered in this study is RD. Firm competitive advantage, long-term growth and technological advancement depends on the amount of investment on RD activities which lead to enhanced performance (Patel et al., 2018; Ruiqi et al., 2017). Grounded on the free cash flow hypothesis, Jensen (1993) established that managers may overspend their free cash flows in projects like RD. The possibility of limited gain from RD investment may come from the higher financing cost associated with RD due to the risky nature of RD (Hillier et al., 2011). The literature has found a positive relationship between RD and firm performance (Eberhart et al., 2004; Yeh et al., 2010). RD investment may not automatically create value for the investing firms. RD investment helps to increase the operating performance of investing firms in the long run (Eberhart et al., 2004). However, the link between RD investment and firm performance may be reinforced or weakened by other firm specific and external factors (Alam et al., 2019, 2020). Here, the notion is the burden of making regular interest payments to debt holders can pressure managers to generate adequate cash flows. Thus, the researcher theorize that this may lead marketers to adopt short-term actions such as cutting PR and RD spending, which can harm export performance by lowering customer’s perceived quality and value of the product or service. From the above discussion, the following hypothesis was postulated:

Hypothesis 5:

There is significantly negative relationship between firm capital structure with high debt and firm export promotion activity

Hypothesis 6:

There is significantly negative relationship between firm capital structure with high debt and firm research & development activity

Hypothesis 7:

There is significantly positive relationship between availability of cash and firm export promotion activity

Hypothesis 8:

There is significantly positive relationship between availability of cash and firm research & development activity

Hypothesis 9:

There is significantly positive relationship between promotion activity and firm export performance

Hypothesis 10:

There is significantly positive relationship between research & development activity and firm export performance

Hypothesis 11:

Firm export promotion activity significantly and negatively mediate the effect of capital structure with high debt on export performance

Hypothesis 12:

Firm research and development activity significantly and negatively mediate the effect of capital structure with high debt on export performance

Hypothesis 13:

Firm export promotion activity significantly and positively mediate the effect of availability cash on export performance

Hypothesis 14:

Firm research and development activity significantly and positively mediate the effect of availability of cash on export performance

3. Methods

3.1. Variables and measures

3.1.1. Dependent variable

We have one dependent variable in the theoretical model. Firm EXP, which refers to all products exported to the foreign market, is used as the unit of analysis. EXP is the degree to which a firm completes its purposes by exporting its products to foreign markets, comprising economic or operational facets (profit, sales, etc.) and strategic aspects (international positioning, increased market share from exporting, achievement of objectives, etc.) (Cavusgil & Zou, 1994). The two principal modes of EXP assessment identified in the general literature are objective (e.g. based mainly on records relating to absolute figures of company profitability and sales level) and subjective (e.g. managers’ perceptions) measures (Katsikeas, 1996). Following our theoretical bearings, we focus on the operational performance encompassing firms’ export sales volume and revenue from export of commodity (objective measure) as EXP measures. Thus, the EXP measurement in this study does not include the companies’ sales or revenue from selling its product in the domestic market. The researchers used total unit a firm sold during specific period as sales volume. We solved sales and price faction to obtain revenue of specific firm. It means we calculated revenue for specific firm by multiplying total unit sold during given period by unit price. The following equations (Equations 1(1) $$\hspace{0.17em}\mathrm{\text{EXP}}\hspace{0.17em}1=\text{sales}$$(1) and 2(2) $$\hspace{0.17em}\mathrm{\text{EXP}}\hspace{0.17em}2=\text{sales*}$$(2) ) were used to measure EXP: (1) $$\hspace{0.17em}\mathrm{\text{EXP}}\hspace{0.17em}1=\text{sales}$$(1) (2) $$\hspace{0.17em}\mathrm{\text{EXP}}\hspace{0.17em}2=\text{sales*}$$(2) where sales represent total unit sold during a given time period, sales*price represent revenue during a given time period.

3.1.2. Predictors

Based on the literature review, we distinguished between internal and external EXP determinants (Chen et al., 2016). Firms CS and CA are used as internal environmental factors. Following other studies in international business, CS and CA were measured as managers’ responses to the five-point Likert scale question, ‘How do you assess the level of long-term and short-term debt financing of your organization relative to the industry?’ and ‘How do you assess liquidity and cash following your organization’s position relative to the industry?’ respectively.

The first factor related to the external environment is informal sector competition (CI), which captures the extent to which competition from the informal sector in home countries affects the EXP of formal firms (‘To what extent are the practices of competitors in the informal sector an obstacle to the current export activity?’). The second factor related to the external environment is political instability (PI), which measures the impact of political instability based on managers’ responses to the question: ‘To what extent is political instability an obstacle to your current to the current export activity?’ For both CI and PI, the responses have five possible values: no obstacle (0), minor (1), moderate (2), major (3), and very severe obstacles (4).

3.1.3. Mediating variables

3.1.3.1. Promotion mix activities (PR)

A specific combination of promotional methods used for one product or a family of products is promotion mix. The main purpose of promotion is to warrant that customers are cognizant of the existence and positioning of products. Promotion mix is one of the most powerful elements in the marketing mix. Promotion activity of a firm can create brand equity by informing customers, and differentiating products from competition. It is marketing manager who decides the level of marketing expenditure on promotion (Singh, 2012). Thus, the promotion mixes activity of a firm during a given time period were measured by the level of its promotion mixes expenditure. Promotion expenses range from giveaways, free samples, or other promotional gimmicks in order to help boost sales and revenue. Promotion expense can promote firm value and create potential intangible asset that may not be measurable by increasing future demand and brand loyalty.

3.1.3.2. Research and development activity (RD)

Research & development activity of a firm represents the activities companies undertake to innovate and introduce new products and services or to improve their existing offerings and allows the company to stay ahead of its competition (Kang et al., 2017). Due to increasing competition, firms should be innovative at an extraordinary pace. Innovativeness is one of the vital gadgets of growth strategies and provides the company with a competitive edge (Gunday et al., 2011). This requires the companies’ put forth large expenditures on RD. Thus, expenditures on RD help companies maintain their competitive advantage and ensure their future viability. Hence, the innovative effort of a firm is peroxide by its RD intensity of which can be measured by the firms’ RD expenditure (Baumann & Kritikos, 2016). Thus, the study used RD expenditure to measure the firms’ RD activity during a given time period.

3.1.3. Control variables

To account for firm heterogeneity, we control for several variables deemed important in the export literature. Moreover, including control variables in the model will provide alternative explanations for the possible results (Ferreras-Méndez et al., 2019). Control variables are the variables (i.e. factors and elements) that researchers seek to keep constant when conducting research to prevent them from influencing the outcome of a study. To properly measure the relationship between dependent and independent variables, other extraneous or confounding variables must be controlled (i.e. neutralized, eliminated, and standardized). If used properly control variables can help the researcher accurately test the value of an independent variable on a dependent variable (Alam et al., 2019). There are evidence for performance difference among strategic types for different industries and firm size (Anwar & Hasnu, 2016). Sousa et al. (2008) note that firm size is the most studied factor. It is considered an indicator of the available internal resources and access to resources. Hence, the larger the firm is, the greater its access to resources and skills that will enable it to compete in international markets (Antoncic & Hisrich, 2001). Unlike larger firms, smaller firms are likely to take the international route gradually or in phase (Karadeniz & Göçer, 2007).

The industry sector to which firms belong is another variable that has a profound effect on firms’ export performance. Thus, based on a systematic literature, firm size and firm industry sector were considered as extraneous variables and included in the model as found to affect firm EXP (Kuivalainen et al., 2007).

3.1.4. Model specification

The model of the relationship between predictors, mediators, and dependent variables takes the following form. To estimate the link between independent and mediating variables the study developed Eqiuations 3 and 4. Similarly Equation 5(5) $$\begin{array}{l}{\text{EXP}}_i={\beta }_0-{\beta }_{1}C{S}_{i,t}+{\beta }_{2}C{A}_{i,t}-{\beta }_{3}C{I}_{i,t}-{\beta }_{4}P{I}_{i,t}+{\beta }_{5}P{R}_{i,t}+{\beta }_{6}R{D}_{i,t}+{\beta }_{7}C{S}_{i,t}\text{*}P{R}_{i,t}+{\beta }_{8}C{S}_{i,t}\text{*}R{D}_{i,t}\\ \hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}+{\beta }_{9}C{A}_{i,t}\text{*}P{R}_{i,t}+{\beta }_{10}C{A}_{i,t}\text{*}R{D}_{i,t}+W+X+{\epsilon }_{\mathit{\text{EXPi}},t}\end{array}$$(5) was used to estimate the direct and indirect effect of predictors on EXP. The direct effect root capture the direct impact of independent and mediating variables on EXP, while the indirect effect root capture the effect of interaction between predictors and mediating variables (mediation effect). (3) $${\mathit{\text{PR}}}_i={\mathrm{\Omega }}_0-{\mathrm{\Omega }}_1{\mathit{\text{CS}}}_{i,t}+{\mathrm{\Omega }}_2{\mathit{\text{CA}}}_{i,t}+W+X+{\epsilon }_{\mathit{\text{PRi}},t}$$(3) where ${\Omega }_1$ and ${\Omega }_1$ are coefficients of the CS and CA capturing their effect on PR respectively. W and X represent the control variable in the model. ${\Omega }_0$ represent the constant term, the subscript i and t donate the sampled companies as well as the year respectively. ${\epsilon }_{\mathit{\text{PRi}},t}$ captures the error term of PR at firm i and year t. (4) $$R{D}_i={\Omega }_0-{\Omega }_{1}C{S}_{i,t}+{\Omega }_{2}C{A}_{i,t}+W+X+{\epsilon }_{\mathit{\text{RDi}},t}$$(4) where ${\Omega }_1$ and ${\Omega }_1$ are coefficients of the CS and CA capturing their effect on RD respectively. W and X represent the control variable in the model. ${\alpha }_0$ represnt the constant term, the subscript i and t donate the sampled companies as well as the year respectively. ${\epsilon }_{\mathit{\text{RDi}},t}$ captures the error term of RD at firm i and year t. (5) $$\begin{array}{l}{\text{EXP}}_i={\beta }_0-{\beta }_{1}C{S}_{i,t}+{\beta }_{2}C{A}_{i,t}-{\beta }_{3}C{I}_{i,t}-{\beta }_{4}P{I}_{i,t}+{\beta }_{5}P{R}_{i,t}+{\beta }_{6}R{D}_{i,t}+{\beta }_{7}C{S}_{i,t}\text{*}P{R}_{i,t}+{\beta }_{8}C{S}_{i,t}\text{*}R{D}_{i,t}\\ \hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}+{\beta }_{9}C{A}_{i,t}\text{*}P{R}_{i,t}+{\beta }_{10}C{A}_{i,t}\text{*}R{D}_{i,t}+W+X+{\epsilon }_{\mathit{\text{EXPi}},t}\end{array}$$(5)

${\beta }_{i,t}$‘s are public coefficients of the corresponding variables of firm i and year t. W and X represent the control variable in the model. ${\beta }_0$ represent the constant term, the subscript i and t donate the sampled companies as well as the year respectively. ${\epsilon }_{\mathit{\text{EXPi}},t}$ captures the error term of EXP at firm i and year t.

3.2. Data source and sample

We test our hypothesis using a sample of firms and firm-level data from one of the sub-Saharan African countries in the Ethiopian manufacturing sector. We also used the Ethiopian Ministry of Industry and Ethiopian Enterprise Development Institute database to measure firm export performance. The Ethiopian Ministry of Industry collects countrywide firm-level data for large firms that cover information about firms’ profiles and performance, while the Ethiopian Enterprise Development Institute collects data for SMEs. Primary data for the independent variables were collected using a self-administered questionnaire.

We initially selected all companies listed in the Ethiopian Ministry of Industry and Ethiopian Enterprise Development Institute database. We obtained financial data from the Ministry of Industry, a database containing annual export information, including sales volume and revenue for 2022, of all manufacturing firms engaged in export trade. In addition to assembling the database, we thoroughly explored each company using company websites, annual reports, archival data, and filings with local regulators. We identified a company’s annual sales volume and revenue, size, industry sector, and whether it had international operations among other variables. We ended with 401 SME’s, 117 large firms, and 518 firm-year observations. After cleaning the dataset by removing firms with missing data, we were left with 191 SMEs and 102 large firms for a total of 293 firms. Given a population size of 293, the sample size was calculated based on the parameters and the total number of firms using Cochran’s (1977) formula.

Applying this formula, we initially obtained a sample of 166 firms. However, after considering the expected 25% non-response rate (adding 42 firms); we obtain a final sample size of 208. Once the total sample size was determined, stratified systematic sampling was used to select the appropriate sample. Cross-sectional data were used for the analysis in this study. A total of 208 SMEs and large firms were invited to participate in the study. The survey questionnaire was then distributed to each company’s CEO and/or founder; and a self-administered survey was used to reduce none response rate. Overall, 161 questionnaires were obtained, for an effective response rate of 77.4% (161/208).

3.3. Analysis approach

A PLS-SEM statistical analysis approach using SmartPLS 4 was used to test the research hypotheses. This approach is suitable for small samples, as is the case in this study, and in general, for studies focused on businesses. PLS-SEM is particularly suitable for early-stage theory development and testing (Hair et al., 2021) and allows investigation of constructs and relationships in complex structural model. PLS-SEM analysis can easily obtain solutions to highly complex models, that is, models with a large number of indicators, constructs, and structural relationships (Hair et al., 2021). Researchers prefer this technique because PLS-SEM does not require a large sample size, works efficiently for complex models, and has no assumptions about the data distributions (Hair et al., 2014). Moreover, this approach is suitable for research on EXP determinants which its theory development process is ongoing.

The condition required in PLS-SEM is that the sample size should be ten times the number of arrows directed at a construct (Hair et al., 2021). As the number of arrows pointing at the constructs in this study was 11 and the sample size was 161, this is well above the required size. This approach proposes that the study be conducted in two stages to analyze and interpret PLS results (Chin et al., 2020): (1) evaluation of the outer model (measurement model), and (2) estimation of the inner model (structural model). Construct reliability and validity assessments were conducted before evaluating the inner model. The detailed analysis and estimated results are presented in the following section.

4. Result and discussion

After descriptive analysis, following Chin et al. (2020), at the beginning the tools were tested for reliability and validity (outer model assessment). Then, SEM analysis was performed to test the hypotheses (inner model assessment).

4.1. Descriptive analysis and correlation

Descriptive and univariate statistics provide a fascinating background. Table 1 report the descriptive statistics and the correlation matrix for the variables used in this study. The average perceived level of debt is 3.32 and 3.58 respectively (midway between high and moderate), whereas the average perceived availability of cash is 2.29 and 2.32 respectively (between low and moderate). The average perceived obstacle presented by informal competitors is 3.09 (above the major obstacle). The average extent to which political instability is an obstacle is 3.17 (between major and very severe obstacles). Regarding the level of promotion expenditure of the sample firms, the result indicate that the average promotion expenditure is less than or equals to 2 which is low. Similarly, the average research & development expenditure is 2. These finding indicate that the firms are allocating insignificant amount of money for their promotion and research & development activities.

Table 1. Descriptive statistics and correlations.

	Descriptive statistics					Correlations
	Variables	Mean	Std. deviation	Minimum	Maximum	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16
1	CS1	3.317	1.0864	1	5
2	CS2	3.584	1.2921	1	5	.923**	1
3	CA1	2.292	1.3069	1	5	−.721**	−.775**	1
4	CA2	2.323	1.3857	1	5	−.733**	−.776**	.897**	1
5	CI	3.087	1.0747	1	4	.645**	.715**	−.810**	−.825**	1
6	PI	3.168	0.9889	1	4	.741**	.784**	−.826**	−.824**	.886**	1
7	PR1	1.932	1.1518	1	5	−.677**	−.733**	.757**	.781**	−.788**	−.818**	1
8	PR2	2.006	1.0868	1	5	−.700**	−.710**	.707**	.729**	−.712**	−.769**	.779**	1
9	PR3	2.043	1.1421	1	5	−.721**	−.737**	.766**	.749**	−.797**	−.814**	.886**	.795**	1
10	PR4	1.969	1.1642	1	5	−.738**	−.777**	.737**	.773**	−.742**	−.761**	.828**	.860**	.838**	1
11	PR5	2.068	1.1891	1	5	−.728**	−.766**	.783**	.798**	−.802**	−.818**	.866**	.841**	.927**	.877**	1
12	RD1	2.062	1.1385	1	5	−.663**	−.671**	.719**	.760**	−.750**	−.798**	.809**	.752**	.777**	.747**	.768**	1
13	RD2	2.068	1.0904	1	5	−.667**	−.672**	.736**	.779**	−.768**	−.828**	.805**	.759**	.791**	.725**	.782**	.847**	1
14	RD3	2.124	1.0883	1	5	−.636**	−.652**	.669**	.723**	−.725**	−.769**	.770**	.713**	.745**	.728**	.732**	.892**	.856**	1
15	EXP1	230924.1	503342.6	3.8	3150016.8	−.618**	−.610**	.605**	.638**	−.615**	−.650**	.688**	.608**	.646**	.615**	.591**	.643**	.648**	.608**	1
16	EXP2	1008960	2235289	3143.4	17701210	−.689**	−.681**	.693**	.708**	−.734**	−.776**	.754**	.679**	.749**	.679**	.732**	.787**	.792**	.735**	.782**	1

**Correlation is significant at the 0.01 level (2-tailed).

4.2. Measurement model assessment

The psychometric properties of the measurement scales (reliability and validity) were assessed before the hypotheses test. Assessment of reflective outer models involves determining the indicator reliabilities (Naala et al., 2017). Coefficients Alpha of Cronbach and of composite reliability is commonly used to evaluate the internal consistency (reliability) of measures and, values equal to or higher than 0.7 are generally considered adequate (Sarstedt et al., 2021). Table 2 shows that Coefficients Alpha of Cronbach and of composite values is above the threshold value of 0.7 for all variables, suggesting that the items used in this study are reliable.

Table 2. Measurement model.

Construct	Item	Weights	Cronbach’s alpha	Composite reliability	AVE
CS	CS1 CS2	0.973 0.975	0.960	0.980	0.961
CA	CA1 CA2	0.980 0.981	0.946	0.973	0.948
IC	Single Item Construct	n.a	n.a	n.a	n.a
PI	Single Item Construct	n.a	n.a	n.a	n.a
PR	PR1	0.931	0.966	0.973	0.880
	PR2	0.909
	PR3	0.949
	PR4	0.938
	PR5	0.962
RD	RD1	0.957	0.951	0.968	0.910
	RD2	0.946
	RD3	0.958
EXP	EXP1 EXP2	0.934 0.953	0.878	0.942	0.891

Note: CS = capital structure, CA = availability of cash, IC = competition from the informal sectors in home countries, PI = political instability in home countries, PR = promotion, RD = research and development, EXP = firm export performance, n.a = Not applicable.

Further, convergent and discriminant validity analyses were conducted to test the validity of constructs. The estimation result of the measurement model in this study support the validity of the measurement scales. As the factor loading of items is above 0.7 and the values of AVE coefficient are above 0.50 (Table 2), convergent validity is not an issue in this study (Hair et al., 2021). In addition, this study follows three approaches to assess discriminant validity of rating scales, Fornell and Larcker (1981), Heterotrait-Monotrait (HT-MT) ration, and cross-loadings criterions (Acosta., 2018). As for this study, because the square root of each AVE coefficient is larger than the correlations between constructs (Table 3), all factors meet the criterion proposed by Fornell and Larcker (1981). Furthermore, the Heterotrait-Monotrait (HT-MT) ratios are, in every case, below the threshold of 0.90 (Henseler et al., 2015) (Table 4). Finally, it is witnessed that factor loadings for each item in the associated construct are in each case greater than the loads on other constructs (Table 3). Thus, these results confirm convergent and discriminant validity is established in this study.

Table 3. Discriminant validity – cross loading.

	CA	CS	EXP	IC	PI	PR	RD
CA1	0.973	−0.763	0.691	−0.810	−0.826	0.800	0.743
CA2	0.975	−0.770	0.716	−0.825	−0.824	0.817	0.791
CS1	−0.747	0.980	−0.695	0.645	0.741	−0.760	−0.688
CS2	−0.797	0.981	−0.687	0.715	0.784	−0.794	−0.697
EXP1	0.638	−0.626	0.934	−0.615	−0.650	0.672	0.664
EXP2	0.720	−0.699	0.953	−0.734	−0.776	0.767	0.809
IC	−0.840	0.694	−0.719	1.000	0.886	−0.820	−0.785
PI	−0.847	0.778	−0.761	0.886	1.000	−0.849	−0.837
PR1	0.790	−0.719	0.766	−0.788	−0.818	0.931	0.833
PR2	0.738	−0.719	0.685	−0.712	−0.769	0.909	0.778
PR3	0.778	−0.744	0.743	−0.797	−0.814	0.949	0.809
PR4	0.776	−0.773	0.688	−0.742	−0.761	0.938	0.769
PR5	0.812	−0.762	0.707	−0.802	−0.818	0.962	0.798
RD1	0.759	−0.680	0.763	−0.750	−0.798	0.822	0.957
RD2	0.778	−0.683	0.768	−0.768	−0.828	0.824	0.946
RD3	0.715	−0.657	0.717	−0.725	−0.769	0.786	0.958

Note: CS = capital structure, CA = availability of cash, IC = competition from the informal sectors in home countries, PI = political instability in home countries, PR = promotion, RD = research & development, EXP = firm export performance.

Table 4. Discriminant validity – Fornell and Larcker Criteria, and Heterotrait-Monotrait (HT-MT).

	CA	CS	EXP	IC	PI	PR	RD
CA	0.974	0.826	0.789	0.863	0.872	0.869	0.830
CS	−0.787	0.980	0.764	0.708	0.794	0.823	0.739
EXP	0.723	−0.705	0.944	0.762	0.806	0.827	0.853
IC	−0.840	0.694	−0.719	n.a	0.886	0.833	0.804
PI	−0.847	0.778	−0.761	0.886	n.a	0.863	0.858
PR	0.830	−0.793	0.766	−0.820	−0.849	0.938	0.886
RD	0.788	−0.706	0.786	−0.785	−0.837	0.850	0.954

Note: Diagonal and italicized are the square roots of AVE. Below the diagonal elements are the correlations between the construct’s values. Above the diagonal elements are the HTMT values. CS = capital structure, CA = availability of cash, IC = competition from the informal sectors in home countries, PI = political instability in home countries, PR = promotion, RD = research & development, EXP = firm export performance, n.a = not applicable.

4.3. Structural model assessment

Following confirmation of the validity and reliability of the constructs, the next step was to assess the general fit of the structure itself; in other words, inner model assessment. However, before assessing the inner model eminence, multi-collinearity factor should be examined. Based on the previous literature, the variance inflation factors (VIFs) values of the constructs were used to assess whether multi-collinearity is present or not (Ke & Zhang, 2010). The presence of a multi-collinearity can be an issue if VIFs are higher than 10 (Mason & Perreault, 1991). More strictly, a VIF thresh-old of 3.3 is recommended by Cenfetelli and Bassellier (2009), and 3 is recommended by Hair et al. (2021). In this study, the results show that VIF values range between 2.630 and 7.372 (Table 5).

Table 5. Model explanatory power.

Predictors	Outcome variable	VIF	R square	f Square	Q square
CS	EXP	3.391	0.678	0.032	0.553
CA		5.138		0.000
CI		5.702		0.004
PI		7.372		0.004
PR		5.871		0.007
RD		4.429		0.110
CS	PR	2.630	0.740	0.195	0.732
CA		2.630		0.432
CS	RD	2.630	0.641	0.054	0.635
CA		2.630		0.395

Note: CS = capital structure, CA = availability of cash, CI = competition from the informal sectors in home countries, PI = political instability in home countries, PR = promotion, RD = research & development, EXP = firm export performance.

The three-stage method proposed by Camps et al. (2016) were followed to assess the structural model. First, coefficient of determination (value R²) for latent variables was assessed which followed by the assessment of the predictive power (Q²) of the model. Finally the assessment of significance of the structural model path coefficients and effect size (bootstrapping) were done.

The study used a bootstrap method with 5000 samples, each of which contains the same number of observations than the original sample to test the accuracy of structural paths in the model (Hair et al., 2013). The study makes an estimate of causal relations between latent variables in the model, through the sign and degree of path coefficients.

R² simply represents how much change in the endogenous variable can be accounted by one or more independent variable(s). It denotes the variance in each of the endogenous constructs which measure the model’s explanatory power, and it can also be called as in-sample predictive power (Sarstedt et al., 2014). Chin (1998) recommended R² values for endogenous latent variables based on: 0.67 (substantial), 0.33 (moderate), 0.19 (weak). However, Hair et al. (2013) suggested in scholarly research that focuses on marketing issues, R² values of 0.75, 0.50 or 0.25 for endogenous latent variables, as a rough rule of thumb can be described as substantial, moderate or weak respectively. In the current study, for all endogenous latent variables R² statistics take values between .641 to .740 (Table 5) indicating that the recommended theoretical model provides a moderate explanation of the variance of dependent variables.

In PLS path model, when an independent variable is omitted from the model, it measures the difference in squared correlation values and establish whether the omitted variable has a robust effect on the value of endogenous variable. Effect size (f²) refers to the change in R² if a given exogenous variables is omitted from the model and used to better estimate the explanatory value of each exogenous variable in the model. The impact of exogenous variable is high at the structural level if f square is 0.35; it is medium if f square is 0.15 and small if f square is 0.02 (Cohen, 1988). Precisely, effect size assesses the magnitude of relationship between the latent variables. The results (Table 5) revealed that f-square effect size ranged from 0.000 (negligent) for CA on EXP to 0.110 for RD on EXP. The dominant effect size of RD and the negligent effect size of CA on export performance draw our attention, eliciting vitality of RD in the proposed model. However, the dominant effect size of CA on both PR and RD is also very striking (Tale 5). Furthermore, using Q² the study establishes the predictive relevance of endogenous constructs. Q square value 0.02, 0.15, 0.35 refers weak, moderate, strong degree of predictive relevance of each effect respectively (Hair et al., 2013). Using the PLSpredict procedure, it is observed that in all case the Q² value is considerably larger than 0.35 (Table 5) backing strong predictive relevance of the model with regard to all endogenous variables.

Following the assessment of the general fit of the structural model, the next step was the assessment of structural path for evaluation of path coefficients (relationship among the study constructs) and their statistical significance.

Table 6 summarizes the results obtained for direct relationships in the structural model, including path coefficients, t values, and the level of significance. H1 evaluates whether firm CS with high debt significantly and negatively affects EXP. The results suggested that CS with high debt to have a negative and significant influence on EXP (B = −0.175, t = 3.416, p < 0.001). Hence, H1 is supported. H2 evaluates whether CA significantly and positively affects EXP of a firm. As shown in Table 6, the CA is suggested to have no significant direct effect on EXP (B = −0.026, t = 0.380, p = 0.352). This result suggests that H2 is not confirmed. H3 assesses whether CI in home countries significantly and negatively affects EXP of formal firms. The results indicated that degree of CI in home countries has no impact on formal firm export performance (B = −0.105, t = 1.277, p = 0.101). Hence, hypothesis 3 is not supported. H4 evaluates whether PI in home countries significantly and negatively affects EXP of a firm. The results revealed that effect of PI in home countries is negatively related to export performance (B = −0.127, t = 2.408, p = 0.008). Hence, this finding supports H4.

Table 6. Direct relationships (with control variables).

Hypothesis	Beta coefficient	Standard deviation	T statistics	p Values
H1: CS -> EXP	−0.175	0.051	3.416	0.000
H2: CS -> PR	−0.373	0.084	4.441	0.000
H3: CS -> RD	−0.231	0.076	3.038	0.001
H4: CA -> EXP	−0.026	0.067	0.380	0.352
H5: CA -> PR	0.415	0.089	4.642	0.000
H6: CA -> RD	0.476	0.075	6.332	0.000
H7: IC -> EXP	−0.105	0.082	1.277	0.101
H8: PI -> EXP	−0.127	0.053	2.408	0.008
H9: PR -> EXP	0.086	0.085	1.015	0.155
H10: RD -> EXP	0.444	0.072	6.174	0.000
SizeL <- PR	0.233	0.128	1.820	0.034
SizeL <- RD	0.233	0.121	1.929	0.027
CategoryFB <- PR	−0.230	0.089	2.589	0.005
CategoryFB <- RD	−0.265	0.09	2.733	0.003

Note: CS = capital structure, CA = availability of cash, CI = competition from the informal sectors in home countries, PI = political instability in home countries, PR = promotion, RD = research & development, EXP = firm export performance, SizeL = firm size large, CategoryFB = industry category food & beverage.

Furthermore, H5 evaluates whether firm CS with high debt financing significantly and negatively affects PR activity of a firm. The results suggested that CS with high debt to have a negative and significant influence on firm PR activity (B = −0.373, t = 4.441, p < 0.001). Hence, H5 is supported. Similarly, H6 evaluates whether firm CS with high debt financing significantly and negatively affects RD activity of a firm. The finding of the study also confirms that CS with high debt have significant influence on firm RD activity (B = −0.231, t = 3.038, p = 0.001). Thus, H6 is supported. CA is suggested to have significant positive effect on PR activity (B = 0.415, t = 4.642, p < 0.001). This result indicated that H7 is confirmed. Similarly, H8 evaluates whether CA significantly and positively affects RD activity of a firm. The results indicated that CA significantly and positively affects RD activity of a firm (B = 0.476, t = 6.332, p < 0.001). This result suggests that H8 is confirmed.

Regarding H9, this hypothesis aims to investigate the impact of firm PR activity on EXP. The finding of the study doesn’t support the presence of positive impact of firm PR activity on EXP (B = 0.086, t = 1.015, p = 0.15). Hence, H9 is not supported. In addition, H10 assess how firm RD activity affect EXP and it has been found that there is significant positive relationship between firm RD activity and EXP (B = 0.444, t = 6.174, p < 0.001). Thus, H10 is supported.

Finally, though is not of interest to the study’s objectives, firm size and firm industry category were included in the model as a control variable (Table 6) because it could influence the outcomes. The results revealed that larger firm has more effect on both firm PR and RD activity compared to SMEs (B = 0.233, t = 1.820, p < 0.034; B = 0.233, t = 1.929, p < 0.027) respectively. Regarding the effect of firm industry category on firm PR and RD activity, the result confirms food & beverage industry has lesser impact on PR and RD activity of a firm compared to leather, and textile & garment industry (B = −0.230, t = 2.589, p = 0.005; B = −0.265, t = 2.733, p = 0.003) respectively.

4.3.1. Mediation analysis

The mediation analysis was performed to assess the mediation role firm PR and RD activities in the linkage between CS and CA, and the endogenous variable, EXP. Regarding the mediation role of firm PR and RD activities in the linkage between CS and EXP, the result (Table 7) revealed there is no significant indirect effect of CS on EXP through PR (B = −0.032, t = 0.889, p = 0.187). Hence, H11 is not supported. However, there is significant indirect effect of CS on EXP through RD (B = −0.102, t = 2.633, p = 0.004). The total effect of CS on EXP was significant (B = −0.309, t = 6.553, p < 0.001), with the inclusion of mediator the effect of CS on EXP was still significant (B = −0.175, t = 3.416, p < 0.001). This shows a competitive partial mediation role of RD in the relationship between CS and EXP. Hence, H12 was supported.

Table 7. Mediation analysis results.

Total effect (CS -> EXP)			Direct effect (CS -> EXP)				Indirect effect of CS on EXP
Coefficient	T value	P-value	Coefficient	T value	p Value	Hypothesis	Coefficient	SE	T value	p Value	Percentile bootstrap 95% confidence interval
											Lower	Upper
−0.309	6.553	0.000	−0.175	3.416	0.000	H11: CS -> PR -> EXP	−0.032	0.036	0.889	0.187	−0.098	0.019
						H12: CS -> RD -> EXP	−0.102	0.039	2.633	0.004	−0.174	−0.047
Total effect (CA -> EXP)			Direct effect (CA -> EXP)				Indirect effect of CA on EXP
Coefficient	T value	p Value	Coefficient	T value	p Value	Hypothesis	Coefficient	SE	T value	p Value	Percentile bootstrap 95% confidence interval
											Lower	Upper
0.222	3.214	0.001	−0.026	0.380	0.352	H13: CA -> PR -> EXP	0.036	0.035	1.017	0.155	−0.023	0.092
						H14: CA -> RD -> EXP	0.211	0.052	4.062	0.000	0.138	0.306

Regarding the mediation role of firm PR and RD activities in the linkage between CA and EXP, the finding (Table 7) confirms no significant indirect effect of CA on EXP through PR (B = 0.036, t = 1.017, p = 0.155). Hence, H13 is not confirmed. However, the result revealed there is significant indirect effect of CA on EXP through RD (B = 0.211, t = 4.062, p < 0.001). The total effect of CA on EXP was significant (B = 0.222, t = 3.214, p = 0.001), with the inclusion of mediator the effect of CA on EXP was not significant (B = −0.026, t = 0.380, p = 0.352). This shows a complementary full mediation role of RD in the relationship between CA and EXP. Hence, H14 is supported.

5. Discussion and implications

5.1. Theoretical implications

The results obtained in this study hold relevant theoretical suggestions with regard to the determinants of firm export performance. Founded on Resource based theory, dynamic capability view, and institutional-base view, the researchers contends that CS, CA, CI and PI in home countries directly affect firm export performance.

The study findings highlight the importance of firm CS, CA, PI in home countries, PR and RD activity for better EXP. Regarding the direct determinants of EXP, the results indicated EXP of firm is negatively influenced by CS with high debt and PI in home countries while CA, CI in home countries, and PR activity found to have no direct impact on EXP. However, the finding of study supports the direct and positive impact of RD activity on EXP. Furthermore, the finding also confirms the direct and negative impact of CS with high debt on both PR and RD activity. Similarly, the finding revealed that CA has significant and positive impact on PR and RD activity, indicating existence of an indirect effect of CS and CA on EXP through PR and RD activity. These findings support the researchers argument that firm success in a foreign market depends not only on its given portfolio of resources and capabilities, as per the resource-based theory (RBT), but also on its capacity and ability to constantly change and adjust to international uncertainties. Similarly, theory of strategy-environment co-alignment (Porter, 1980; Venkatraman & Prescott, 1990), states that the ‘fit’ between strategy and its context, whether it is the external environment (Anderson & Zeithaml, 1984; Hofer, 1990) or organizational characteristics (Chandler, 1962; Gupta & Govindarajan, 1984) has significant positive implications for firm performance.

Furthermore, this study contributes to the body of knowledge on the indirect relationship between CS, CA, and EXP by providing insights into the synergy among firm CS, CA, PR, and RD activity and their effect on the EXP. Decision about whether to standardize or adapt the marketing-mix elements does have great impact on export market performance (Ruzo et al., 2011). Empirical and theoretical literature acknowledged marketing mix strategy as a key driver and mediator of business performance. Although the literature continuously emphasizes the vital role of various firm specific and environmental factors in influencing export performance, most past empirical studies examined the direct effect while ignoring the potentially complementary effects of other factors particularly the mediating role of firm marketing mix strategy. A recent study argues that resources available for export activities affect export marketing mix strategy and performance (Imiru, 2018). Maritan and Lee (2017) urge more research that examines resource allocation as a central focus of study for achieving deeper and better understandings about firm strategies. Although it has been well argued in the resource-performance relationship literature, most previous studies have only focused on the direct effect of firms’ CS, and CA on EXP, neglecting the mediating role of export marketing mix strategy. This gives incomplete view of EXP determinants while hindering the conceptualization of EXP construct. Based on the theoretical foundation proposed by different authors in recent literature on Strategic Management (Deutscher et al., 2016) and internationalization (Hagen et al., 2017; Paul et al., 2017), this study also investigates the mediating role of PR and RD in the link the between CS, CA, and EXP.

Interestingly, this study shows that firm PR activity has no vital role in mediating the link between CS, CA and EXP in the short run. However, the study revealed that RD activity is a key factor in mediating the link between CS, CA and EXP. Meanwhile, the study uncovered that RD activity fully mediate the relationship between CA and EXP while partially mediate between CS and EXP. However, non-significant relationship between PR and EXP is interesting as previous scholars find support for the positive relationship between firm PR activity and its performance (Thabit & Raewf, 2018; Nath et al., 2010). While we know that PR activity of firms generally affect EXP, there may be limit as to its instant and linear nature of the PR-EXP relationship, especially in the short-term. Too great a focus on linear relationship between PR and EXP could be misleading. Thus, an interesting avenue for future research would be to look at non-linear relationship between the constructs specifically using panel data. Together, these findings begin to shine a light as to how resources influence both marketing mix strategy and performance amplifying the Maritan and Lee’s (2017) urge to examine resource allocation as a central focus of study for achieving deeper and better understandings about firm strategies and performance.

5.2. Managerial and policy implications

The study of export market can be a learning paradigm in the field of marketing to enhance the understanding of practitioners, policy makers, and academicians. As this study jointly investigated the impact of internal (CS and CA) and external (CI and PI) factors on EXP, it provides more sophisticated results as compared to the separate examinations in the previous studies. This study provides an important insight for practitioners and/or managers when venturing abroad carefully consider the consequences of resource-marketing strategy interaction to EXP. Moreover, using data from firms in developing economy for empirical validation of the theoretical model, this study help policy makers identify the important factors to consider in enhancing export performance of the firms in developing economy context. Generally, the theoretical and practical contributions of this study are important. The paper developed and validated RBT, DCV and IBV integrative model which will enable academics, managers and governmental bodies may emphasize the importance of both internal (CS, and CA) and external (CI, and PI) factors that may result in higher or lower EXP, and the role PR and RD activity play in the link between these factors.

5.3. Limitations and future research directions

The rise of ever-changing foreign market environmental dynamism allows export performance research domain to foresee a promising future area. There was a call by researcher, managers, and policy-makers for a robust understanding of exporting. This study reveals the direct and indirect impact of CS, CA, CI in home countries, and PI in home countries on EXP. Predominantly, the study examines the marketing effort route through which CS and CA influences EXP.

Though this study provides vital findings, there are issues which deserve further consideration. First, the test of our model was piloted with a diverse cross-section of exporters in Ethiopian manufacturing industry. For cross-national generalizability, the model requires replication, extension, and further evaluation in other export market contexts that may differ in terms of economic development and socio-cultural values. Second, our operationalization of financial variables, while pertinent to the context of this study, may not be comprehensive enough to capture the entire domain of the concept. It is possible that financial variables utilization is a multi-faceted construct, which deserves further scrutiny and development, since this construct is an inevitable part of business operations. Third, subjective (perceived) measures of CS and CA were used because objective measures were difficult to obtain during the pretest stage of this study. While subjective and objective measures have been found to be highly correlated (e.g. Boso et al., 2013), future studies should attempt to include objective measures. Finally, future studies should incorporate the consequence of the interaction between CS and CA on EXP as these two financial variables are highly correlated.

Acknowledgements

The authors would like to thank Arba Minch University for their support in conducting this study. We are also grateful to the sample managers and employees, enumerators, Ethiopian Ministry of Industry, Ethiopian Enterprise Development Institute key informants, and local administrators of the study area for their kindness and cooperation during fieldwork.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request.

Additional information

Notes on contributors

Nesredin Temam Hassen

Nesredin Temam Hassen is an assistant professor of Financial Management and International Business at Dire Dawa University, Ethiopian. He holds a BSc in Business Management from Adama Science Technology University of Ethiopia, an MBA from Andhra University in India, and currently a Ph.D. scholar in Marketing at Arba Minch University, Ethiopia.

Mesfin Lema

Mesfin Lema is an Ethiopian lecturer trained as Business Administrator (MBA) and Economist (MA) at The Free University and University of Technology, Berlin, Germany, respectively and Ph.D. in Business Administration at Bulacan State University, the Philippines. He is now working as Associate professor at International Leadership Institute, Ethiopia.

Gemechu Nemera

Gemechu Nemera is an Assistant Professor of Management at Arba Minch University. He studded MBA and Ph.D. in Business Administration. He is now working as assistant professor at Arba Minch University, Ethiopia.