Entrepreneurial orientation has become an hot topic in business research due to its key role as “a driving force behind the organizational pursuit of entrepreneurial activities” (Covin & Wales, 2012, p. 677). The literature has emphasize the multifaceted outcomes of entrepreneurial orientation and how it leads to superior performance (Kollmann & Stöckmann, 2014). Miller (1983) states that an organization with entrepreneurial orientation bets on the innovation of poducts and/or markets with some risk and acts proactively before its competition. For Lumpkin and Dess (1996, p. 136), entrepreneurial orientation “refers to the processes, practices, and decision-making activities that lead to new entry”.

Previous research has suggested that entrepreneurial orientation is a multidimentional construct. According to Miller (1983), entrepreneurial orientation includes three main dimensions: innovation, risk taking and proactiveness. Two more dimensions, competitive aggressiveness and autonomy, were proposed by Lumpkin and Dess (1996). Thus, according to these authors, the main dimensions that characterize an entrepreneurial orientation include a tendency to act autonomously; a willingness to innovate and take risks and a tendency to be aggressive toward competitors and proactive in terms of market opportunities. This is corroborated by Pearce, Fritz, and Davis (2010) who refer that entrepreneurial orientation is regarded as a set of distinct behaviors such as innovation capability, proactivity, competitive aggressiveness, risk assumption, and autonomy. However, the dimensions most commonly used in research are: innovativeness, proactiveness and risk taking (Kropp, Lindsay, & Shoham, 2008). According to Miller (1983), only companies that have a high level in all three dimensions of entrepreneurial orientation would be considered potentially entrepreneurial. The relation between entrepreneurial orientation and export performance has also been studied. Previous research suggests that resources and capabilities are related (Dhanaraj & Beamish, 2003; Morgan et al., 2004) and sustains that entrepreneurs have a unique cognitive ability of to recognize venture opportunities and organize resources (Alvarez & Busenitz, 2001). Nevertheless, further research on the relationship between entrepreneurial orientation and performance in the context of changing markets is needed to enhance theory development in the field (Jantunen, Puumalainen, Saarenketo, & Kyläheiko, 2005). Thus, we intend to test the following hypothesis:H5

Entrepreneurial orientation has a positive effect on export performance.

This study follows a quantitative methodological approach to test the proposed model by using a questionnaire to collect data, which is consistent with the majority of studies in the literature on export performance (Sousa et al., 2008).

After collecting data, we proceed to statistical analysis using structural equation modeling (SEM), which is a statistical technique which specifies that latent variables present a direct or indirect impact in the values of other latent variables included in the model (Byrne, 1998). This analysis involved three phases: (1) preliminary analysis of the data, (2) evaluation of the measurement model, and (3) evaluation of the structural model. In the first phase, we prepared the data for analysis, which consists of the treatment of the missing-values, an analysis of the outliers, central tendency and normality, sample dimension, and, also an analysis of the non-response bias. Subsequently, we proceeded to the characterization of the sample. In the second phase, the measurement model was evaluated in terms of the unidimensionality, reliability and validity of the constructs (Hair, Anderson, Tatham, & Black, 1998; Rubio, Berg-Weger, & Tebb, 2001; Ketchen, Hult, & Kacmar, 2004). Finally, the structural model was assessed to test the hypothesis. The statistical analysis was carried out using the software Statistical Package for the Social Sciences (SPSS) and LISREL.

The results are based on the responses provided by a majority of small and medium companies (78.1%), large companies are 8.7% and micro companies are 13.2% of the sample. Regarding export activity, 49.1% have been exporting for more than 15 years, 33% for 6–15 years, 13.2% for 3–5 years and 4.5% for less than 3 years. Additionally, 52.8% of the firms export more than 40% of the total sales and 44.2% export to 6–15 countries.

To test the proposed framework, we used structural equation modeling as the theoretical model includes complex relationships among latent variables, and measures of different items that are presented simultaneously as independent and dependent variables (Bentler, Bagozzi Cudeck, & Iacobucci, 2001; Hair et al., 1998). We started by assessing the model fit, then, the structural model of path analysis was examined.

The analysis of the measurement model was evaluated in terms of constructs unidimensionality, reliability and validity (convergent and discriminant) and the results are shown in Table 2. In the first-order models, all items relate significantly to factor in terms of loadings and statistically, thus demonstrating the unidimensionality of the single factor and all loadings of the observed variables have values greater than 0.70, demonstrating the existence of convergent validity of the constructs (Garver & Mentzer, 1999). All latent variables have a good level of composite reliability (CR), with values greater than 0.60, which proves the reliability of the scales (Bagozzi & Yi, 1988). The average variance extracted (AVE) is greater than 0.50, providing evidence for discriminating validity of the constructs (Fornell & Larcker, 1981). In the second-order models, the statistical significance of associations between factors of first and second order is confirmed, the coefficients exceed the minimum threshold of 0.40, confirming the convergent validity of the construct (Nunnally & Bernstein, 1994) and the square of the correlation is less than the average variance extracted for each factor evidencing discriminant validity of the construct (Fornell & Larcker, 1981).

In order to test the proposed hypotheses the structural model was estimated.

The evaluation of the structural model was done under a partial aggregation perspective (average of the values corresponding to the items that measure each factor), since the analysis under the total disaggregation forecast becomes, according to Bagozzi and Heatherton (1994) impracticable when the sample is large and when the construct has more than four indicators. According to Baumgartner and Homburg (1996), the creation of composites (the items of each latent variable are organized in a cluster so that they serve as a measure for only one indicator) is a generalized and inevitable procedure when the number of indicators and constructs is relatively high. On the other hand, the use of composites when compared with the use of individual items has associated advantages in terms of reduction of measurement error (greater reliability) and parsimony level (Grapentine, 1995; Hair et al., 1998). Thus, the items corresponding to the same cluster, which demonstrate to form a one-dimensional set, result in a composite that will be used in the analysis of the structural model (Baumgartner & Homburg, 1996; Hair et al., 1998). It should be noted that the measurement error of the dynamic capacities and export performance constructs was set to the value of [(1−Cronbach's alpha)×variance of the indicator], because they are presented with only one indicator or cluster (Bagozzi & Heatherton, 1994). In constructs with more than one indicator (entrepreneurial orientation and intangible resources), only one connection was established. This procedure, according to Diamantopoulos and Siguaw (2000), does not affect the results of the data analysis. Thus, the analysis of the parameters confirm the fitness of the model (χ2(141)=250.56, p<0.05, CFI=0.99, GFI=0.91, NNFI=0.98, RMSEA=0.054). Table 3 presents the parameter estimates of the structural model allowing the testing of the hypotheses and in Table 4 the means, standard deviations, and correlations are presented.

Results indicate that: (1) the presence of intangible resources (financial, informational and relational resources) are important factors in the development of dynamic capabilities; (2) that dynamic capabilities have a positive impact in export performance; and (3) that entrepreneurial orientation positively enhances export performance.

The significance of the mediating effect was tested using Sobel test (Sobel, 1982). This test, considered valid for test the statistical significance of the indirect effects, is widely used in recent research (Parawansa & Anggraece, 2018; Preacher & Hayes, 2008; Yunis, El-Kassar, & Tarhini, 2017). Results confirm that intangible resources have a significant indirect impact in export performance through dynamic capabilities. The indirect effect of financial, informational and relational resources in export performance through dynamic capabilities is 0.05 (0.18×0.29; p<0.05; Z=2.41), 0.11 (0.38×0.29; p<0.01; Z=4.17), and 0.09 (0.32×0.29; p<0.01; Z=3.17), respectively. This means that intangible resources do not directly influence export performance but only through the dynamics capabilities, i.e. the impact of intangible resources is fully mediated by dynamic capabilities.

We conclude also that intangible resources and entrepreneurial orientation justified 85% of export performance (R2=35% and R2=50%, respectively).

This study aims to investigate how intangible resources, dynamic capabilities and entrepreneurial orientation impact export performance. We develop and test a model according to which intangible resources (financial, informational and relational resources) have a positive and direct influence in the development of dynamic capabilities, which in turn impact export performance. In addition, the framework proposes that entrepreneurial orientation positively and directly enhances export performance.

Based on the results to an online questionnaire for top managers of a significant number of exporting firms, we validated all direct relationships of the model. Specifically, H1, H2 and H3 referring to access to financial, informational and relational resources respectively, were supported. This finding suggests that intangible resources contribute to the development of dynamic capabilities. These results are consistent with previous studies in this area which argued for the proposition that resources are antecedents of the development of capabilities (e.g. Lin & Wu, 2014; Morgan et al., 2004; Wu, 2006; Wu & Wang, 2007). These results also empirically validate Morgan et al.’s (2004) argument that resources and capabilities are interrelated.

Our findings also support H4, relative to the impact of dynamic capabilities in performance. In addition, we confirmed the mediating role of dynamic capabilities in the resources-export performance relationship. The influence of dynamic capabilities in export performance is consistent with the results of Wu (2006), Wu and Wang (2007) who studied the internal market performance of technological companies and of Lin and Wu (2014). Specifically, this finding lends credence to the idea that “dynamic capabilities are regarded as a transformer for converting resources into improved performance” (Lin & Wu, 2014, p. 407). This is a central tenet of the DCV, which holds that dynamic capabilities have a transformative role in converting resources into enhanced performance (Lin & Wu, 2014; Wu and Wang, 2007). The explaining power of the DCV, in comparison to the RBV, as been emphasized specially in dynamic and fast changing environments. Our study finds support for the view that resources can be converted into performance through dynamic capabilities in the context of volatile environment of international markets.

Finally, H5, referring to the impact of entrepreneurial orientation on export performance, was also supported. The impact of entrepreneurial orientation in performance is consistent with a large number of studies (e.g. Covin & Slevin, 1989; Covin, Slevin, & Covin, 1990; Kantur, 2016; Kollmann & Stöckmann, 2014; Miller, 1983; Lumpkin & Dess, 1996). In fact, the impact of entrepreneurial on organizational performance has been well-established in the literature, however, this link has not been much studied in the context of internationalization.

Our study makes several theoretical and empirical contributions. First, our results are relevant to the DCV literature as they support the theory that capabilities can boost performance by leveraging the impact of resources. We have focused on the intangible resources of access to financial, informational and relational resources and findings show that these resources contribute to the development of dynamic capabilities. Second, we have found support for the impact of entrepreneurial orientation in export performance. Entrepreneurial orientation may be seen as a specific dynamic capability to identify venture opportunities and deploy resources (Alvarez & Busenitz, 2001). More importantly, our study provides support for these tenets in the context of international markets.

Relevant practical implications are also to be noted, starting forthwith the contribution for the (re)definition of government policies supporting the adoption of strategies leading to enhanced performance through different dedicated programs, including economic and fiscal benefits. In particular, these programs may be more effective it they target the development of capabilities rather than solely access to resources.

This research has a number of limitations. First of all, there are limitations derived from the potential bias caused by the sample size and measurement. In this research, in line with previous studies, we used Likert scales 1–7 to evaluate the constructs, thus the majority of the answers are based on the subjective judgment of the respondents. Although a strong use of subjective measures to assess the performance of exports has been identified in previous studies, some responses may not reflect the actual situation of the studied variables. Additionally, although email is a commonly used tool, we cannot generalize the results to the total population. Is also important to acknowledge that the study ignores the moderating effects of some variables (e.g. hostile external environment) as well the effect of control variables, such as firm size and demographic characteristics of the respondents, which could lead to further insights. It may also be argued that evaluating the different variables in this study based on the opinion of one respondent per firm may not accurately reflect the reality of companies, as especially in large companies, in which more than one person, make decisions, and may have different opinions on the export activity (Leonidou & Katsikeas, 1996). Dealing with each of these limitations provides directions for further research.

In broad terms, the results of our study stress the importance of dynamic capabilities, which play a catalyst role in the relationship between intangible resources and export performance, contributing to filling the gaps identified by Sousa et al. (2008) regarding the need for export performance studies that consider the (direct, indirect and total) effects of mediating variables. This conclusion partly validates criticisms to the limitations of the RBV to explain firm's competitiveness, specially in volatile environments (Lin & Wu, 2014). In fact, without the ability to achieve new resource configurations as environmental conditions shift (Eisenhardt & Martin, 2000), resources are clearly insufficient conditions for competitive advantage. A clear understanding of how companies develop differentiated dynamic capabilities is paramount for managers, public bodies and researchers aiming at contributing for or firms’ competitiveness and performance.