Is there a virtuous circle relationship between innovation activities and exports? A comparison of food and agricultural firms

Food Policy 61 (2016) 70–79

Contents lists available at ScienceDirect

Food Policy journal homepage: www.elsevier.com/locate/foodpol

Is there a virtuous circle relationship between innovation activities and exports? A comparison of food and agricultural firms q Silverio Alarcón a,⇑, Mercedes Sánchez b a b

Department of Agricultural Economics, Statistics and Business Management at Universidad Politécnica de Madrid, Spain Department of Business Administration at Universidad Pública de Navarra, Spain

a r t i c l e

i n f o

Article history: Received 27 January 2015 Received in revised form 6 November 2015 Accepted 18 February 2016

Keywords: Agricultural innovation Food innovation Internationalization Bivariate probit model Propensity score matching Spain

a b s t r a c t This study examines the existence of an interrelationship between innovation decisions and exports for food and agricultural firms as such a relationship could be the source of competitive advantages. Thus, taking as a theoretical basis the focus provided by the Resource-Based-View, the innovation and export decisions taken from 2006 to 2011 by 165 agricultural firms and 783 food companies operating in Spain (Europe) are examined here. The results of the bivariate probit and matching models used indicate a bidirectional nature of these decisions in the case of food companies and a positive though not bidirectional one in the case of the agricultural firms. Furthermore, a certain persistence is seen in the use of these decisions in both types of firms. For food companies, capital intensity and size are also determinants of innovation and exports. From the viewpoint of the decisions taken by individual firms, the bidirectional relationship could involve significant pressure in terms of the larger volume of both technological and human resources required. Agricultural and food policy decisions should incentivize these decisions given that in order to operate successfully in the global market it is necessary to acquire these competitive advantages, which also favor the growth of the agriculture and food trades. Ó 2016 Elsevier Ltd. All rights reserved.

Introduction The food and agricultural businesses play a fundamental role in the European production systems, representing about 14% of turnover in 2013 in the European Union (Food Drink, 2014; Hirsch et al., 2014; Eurostat, 2015). The importance of the sector becomes even greater when the contribution it makes to the rural population in disadvantaged areas of rural areas is taken into consideration (Compes and García Alvarez-Coque, 2009; Arnalte and Ortiz, 2011). It is thus of the greatest importance that agri-business firms increase their competiveness in order to continue to make a valuable contribution to economic growth. However, the continuing globalization of the economy in general and the food market in particular constitutes a significant challenge for the various kinds of European agri-firms (Blandford and Hill, 2005; Arnalte et al., 2008). In any case, three factors are seen as crucial by Food Drink (2014) in the development of the sector: the growth of export market share, private (R&D) investment and improvements in labor productivity.

q This study forms part of the results of the research project AGL2012-39793C03-01 of the Ministerio de Economía y Competitividad (Spain). ⇑ Corresponding author. E-mail addresses: [email protected] (S. Alarcón), [email protected] (M. Sánchez).

http://dx.doi.org/10.1016/j.foodpol.2016.02.004 0306-9192/Ó 2016 Elsevier Ltd. All rights reserved.

Thus, the entry of these firms into international markets and the increase of their commercial activities in relation to countries abroad involve considerable effort but they allow them to make progress with their decisions of growth and competitiveness. Furthermore, as has already been mentioned, during this current period of global economic crisis, which has depressed domestic consumption, exports represent one of the few possibilities for growth for these firms, and even for their survival in some cases (Ebersberger and Herstad, 2013; Rama, 2014). Along with their efforts to develop exports, these firms can use other decisions to help improve their competiveness, among these are innovation activities and, indeed, these have become one of their main tools for this purpose (Vega-Jurado et al., 2008; Filippaios et al., 2009; Falk, 2012; Hashi and Stojcic, 2013). Thus the globalized environment and financial crisis in the context of which these firms operate forces them to think about the incorporation of technological advances in order to ensure their survival and growth (Ebersberger and Herstad, 2013). The basic idea is that innovation resources allow for the improvement of certain aspects of productivity which in turn has an effect on the firm’s results, in terms of growth, profitability etc., as well as on the internationalization of the firm. And this in turn opens ways to find new markets and opportunities to increase production. Innovation and exports thus interact and can create a virtuous circle for the

S. Alarcón, M. Sánchez / Food Policy 61 (2016) 70–79

firm, the sector or the country in question (Smith et al., 2002; Damijan et al., 2010). Furthermore, both innovation and market internationalization have their places within the various theoretical focuses that have been applied over time. According to Hirsch et al. (2014) these activities are important from the viewpoint of Industrial Organization (IO), Market-Based-View (MBV) as set down by Porter as well as the more up to date Resource-Based-View (RBV) and the Evolutionary approach (López Rodríguez and García Rodríguez, 2005; Castellacci, 2008; Gallego, 2010; Hashi and Stojcic, 2013; Hervas et al., 2014). All of these, to a greater or lesser extent deal with the competitive advantages which could arise resulting from the efforts made by firms toward innovation and internationalization. However, the greater complexity of the agribusiness market due to the important double influence of both the individual aspects of the firms as well as institutional factors must be taken into account (Triguero et al., 2013). Additionally, the literature highlights the importance of innovation for the agri–food sector as one of the main elements by means of which firms can improve their positions in front of their rivals, both in local and international markets (Rama, 2008, 2014; Grunert et al., 1997; Traill and Meulenberg, 2002; Capitanio et al., 2009). Though it is still considered a low intensity sector with regard to R&D&I (García and Burns, 1999; López Rodríguez and García Rodríguez, 2005; Capitanio et al., 2009, 2010; Triguero et al., 2013) one positive factor is the considerable growth in innovation activities in the agri–food sector in relation to the small amount of money spent on it when compared to other sectors. This paper seeks to provide empirical knowledge regarding these matters so as to contribute to the debate on the innovation measures and policies which should be encouraged. In this context, the objective of this paper is to explore the possible bidirectional relations between innovation and exports in the agri–food sector, that is to say, to examine how the efforts made in innovation by agri-business firms contributes to their internationalization, and how at the same time this strengthens innovation. Two methodological approaches are used, with the study being contextualized within the Resources-Based-View Theory (RBV). This study was carried out in Spain (European Union), a country where the agribusiness sector is very relevant in terms of economic position. Agriculture accounted for 2.3% of Spanish GDP in 2013 while the agri–food business accounted for 18.2% of sales and 16.5% of employment in Spanish industry (INE, 2015). In general terms and with regard to the business activities being examined here, in order to improve their competitiveness the Spanish agribusiness sector has been successfully responding to the challenges of internationalization that it faces. Thus, according to the Exports Report of the FIAB1 (FIAB, 2012) sales outside Spain have increased by more than 60% in the last 10 years, from 19398.63 million Euro in 2001 to 31284.09 million Euro in 2011. Furthermore, the Spanish government is conscious of the importance of these processes and has taken steps to support innovation and the internationalization of these firms. The State Innovation Strategy (E2i) is the policy framework which coordinates measures to achieve higher levels of innovation. Furthermore, Spanish Institute of Foreign Trade and the Spanish Ministry of Food, Agriculture and the Environment] co-finance the internationalization programs of FIAB. The information used was obtained from the PITEC [Innovation Technology Panel] produced by the INE (National Statistical Institute), which provides statistics on the technological activities of firms and is commonly used for analyzing innovation decisions. The bivariate probit models and the matching techniques

1

FIAB: Food and Drink industries Federation (Spain) www.fiab.es.

71

estimated take as dependent variables an indicator of whether or not the firm exports, jointly with variables for product or process innovation. The independent variables include different control variables and various measures of firms’ innovation activities. This paper’s main contributions are a study of both food and agricultural firms, an analysis of innovation both in terms of the inputs that are used (total expenditure on innovation, internal expenditure on R&D, external expenditure on R&D) and the innovation outputs obtained (product innovation and process innovation) and, finally, the relationship between innovation and internationalization decisions. The paper is organized as follows: section ‘Innovation and exports’ presents the most significant features of the study’s conceptual framework, section ‘Methodology’ sets out the methodology used, starting from the data and moving to the econometric models estimated, and section ‘Data’ presents the results obtained. The section ‘Conclusions’ presents the most noteworthy conclusions to be drawn, the study’s limitation and suggests areas for future research. Innovation and exports Considered from the microeconomic viewpoint used in this study, the decisions taken by firms in relation to innovation and exports can be placed in the context of strategic decisions taken with the aim of improving and increasing the firm’s resources and capacities (The Theory of the Growth of the Firm, Penrose, 1959; Wernerfelt, 1984; Barney, 1991 among others) with the final aim of obtaining competitive advantages. Thus, internal and external R&D, technological cooperation with other firms and institutions etc. are innovation inputs which firms can employ in a coordinated manner with the impulse provided by other complementary assets such as human, commercial and financial resources (Christensen, 1995). This innovating effort carried out by the firm, in its different formats (internal, external, collaborations, etc.) conditions the capacity the organization will have to absorb the knowledge so generated. Various authors have referred to this as the capacity to recognize the value of the knowledge generated and incorporate it into the firm in commercial and productive terms (Cohen and Levinthal, 1990; Abecassis-Moedas and Mahmoud-Jouini, 2008). Thus this present study situates itself, as has already been indicated, within the Resource-Based-View theory, which has demonstrated the importance of the connection firms are capable of making between their resources and capacities (Hirsch et al., 2014). Exploring this resources and capabilities perspective in various countries Wakelin (1998), Sterlacchini (1999), Basile (2001), Guan et al. (2006), Pla-Barber and Alegre (2007), Caldera (2010) and Yang and Cheng (2012) find that innovation capabilities encourage exports. López Rodríguez and García Rodríguez (2005), for their part, suggest that technological capabilities have an influence on the decision to export and also find that these capabilities are key factors in its competitiveness. Castellacci (2008) also finds more growth in these circumstances. The model of Clerides et al. (1998) offers a rational explanation for this relationship: companies commit themselves to sunk entry costs in order to enter international markets (i.e. market research, product adaptation, etc.) so that only those that expect to make a gross profit from exports greater than these costs will decide to export. Companies that meet this condition are the most productive and in many cases the most innovative as well. Therefore, companies self-select for the internationalization process i.e. as companies innovate and increase productivity they gain in chances and incentives for access to foreign markets. Some studies differentiate between types of innovation and show that product innovations lead to more export advantages and benefits than process

72

S. Alarcón, M. Sánchez / Food Policy 61 (2016) 70–79

innovations (Caldera, 2010). And Nassimbeni (2001) found that exporters have more technological skills especially for new products in the Small and Medium-Sized Enterprises (SMEs). The reverse causal relationship also seems reasonable, that is, from exports to innovation or learning by exporting (Clerides et al., 1998; Girma et al., 2004; among others). The idea is that exporters learn to operate in foreign markets and obtain information that is not available in the domestic market. Thus they adapt to different consumer tastes, deal with other intermediaries, compete with other companies and manage different business practices and regulations. This international experience may put the company in contact with new technologies and management systems, and this could translate into improvements in productivity and lower variable costs. However, other authors consider that the knowledge acquired in foreign markets impact more directly on promoting innovation (Salomon and Shaver, 2005). The empirical evidence for this causal relationship of learning by exporting is less than that found on the hypothesis that companies with more resources and capabilities are more likely to become exporters (Girma et al., 2004; Greenaway and Kneller, 2007). Furthermore Aw et al. (2007) and Girma et al. (2008) find causal relations in both directions between exports and innovations in some types of firms but not others. Monreal et al. (2012) also find a relationship between innovating and exporting with a moderating effect for productivity, Guan et al. (2006) and Karantininis et al. (2010) draw no firm conclusions regarding the relationship between the two variables. Thus the relevant literature is not completely unanimous in its findings and it is therefore of importance to continue to study the relationship between these two variables in different business contexts. Furthermore, Monreal et al. (2012) hold that studying Spain may be of particular relevance as its productive potential does not correspond to its position in international markets. With the existing literature, based on various theoretical perspectives on the relationship between exports and innovation now examined, this present study will now turn to an examination based on the Resource-Based-View (RBV) of the direction of the relationship between the two in agri–food firms. To this end, various methodologies will be employed with the intention of making a contribution to the discussion on the effect of these decisions and suggesting the best way to proceed at the level of the firm and the business sector as well as at the political and institutional level.

Methodology Various methodological options have been adopted to study the existence or not of bidirectionality relationships. In this study two complementary methodologies are used. Firstly, simultaneous equations with panel data (Girma et al., 2008; Damijan et al., 2010), which allows estimation of the bivariate probit with delayed variables and elaborating on causal relationships. Furthermore, the dependent variables themselves are delayed so that they act as regressor variables for the purpose of studying the inertia and sunk costs caused by the export and innovation processes. Secondly, propensity score matching, which classifies companies according to their propensity to innovate and then compares exporting odds according to their degree of innovation; additionally, businesses are classified according to their propensity to export and it then compares the innovation odds based on their export-oriented nature (Damijan et al., 2010; Becker and Egger, 2013). Propensity score matching improves the homogeneity of the companies that are compared and reduces bias in the estimates for the bivariate probit model.

Bivariate probit model Initially a Probit Bivariate model is used to estimate both an export equation and an innovation equation. The intention in doing this is to identify the dominant relationship, in other words whether innovation leads to exports or vice versa. The following equations are jointly estimated (Aw et al., 2007; Girma et al., 2008; Damijan et al., 2010):

ProbðEXPORT it ¼ 1Þ ¼ f ðEXPORT it1 ; Iit1 ; Z it1; Þ

ð1Þ

ProbðIit ¼ 1Þ ¼ f ðIit1 ; EXPORT it1 ; Z it1; Þ

ð2Þ

Sub-indices i and t indicate company and year respectively. The variables of interest are the innovation ones (I) for the Eq. (1) and the export one ðEXPORTÞ for (2), which have a value of 1 in the case of, respectively, innovating or exporting and zero if this is not the case. With regard to the first, product ðINNPRODÞ and process ðINNPROCÞ were taken as output of innovation variables. EXPORT includes exports both inside and outside the European Union. Control variables (Z) include productive factors such as labor productivity, capital intensity and size as well as other elements which can impact on the decision to export or to innovate, such as internal and external R&D expenditures. All the variables are one year delayed in order to avoid simultaneity problems. Following Aw et al. (2007), Girma et al. (2008) and Damijan et al. (2010) the control variables are the same in both equations. Furthermore, the dependent variables delayed one year ðEXPORT it1 ; Iit1 Þ have been introduced as a regressor to control exporting and innovation persistence. The error terms of both equations may be correlated, given the close relationship between exports and innovation and that both are included as dependent variables and as regressors. For this reason the two equations are estimated jointly using a bivariate probit model. Matching approach The Rubin (1980) model seeks to measure the effect of a particular treatment Tri (with a value of 1 if it occurs and 0 if it does not) on the unit i, defined as the difference between the result of the variable of interest after the treatment Y i1 and what would have been obtained had the treatment not been applied (Y i0 , counterfactual). Although in practice only one of these variables can be observed, the inference is experimentally valid because the individuals are extracted at random from the same population and thus their (Z i ) characteristics are independent of the treatment, as is their assignment to the control or experimental group. When the sample is large the mean Z i values are similar for each group. An average treatment effect (ATE) study can be estimated as follows:

ATE ¼ EðY i jTri ¼ 1Þ  EðY i jTri ¼ 0Þ

ð3Þ

In a study with observational data the mean Z i values in both groups (experimental and control) may not be similar because the individuals may not have been taken from the same population. However, it can be taken that if a balance of the Z i in both groups is achieved, then the assignment to one or other group is independent of the (Y 0 ; Y 1 ; ? TjZ). The causal effect is measured through the mean effect of the treatment on those treated (ATT, average treatment effect for the treated):

ATT ¼ EfEðY i jZ i ; Tri ¼ 1Þ  EðY i jZ i ; Tri ¼ 0ÞjTri ¼ 1g

ð4Þ

Matching is a technique which extracts two groups from an initial population (experimental and control) with similar characteristics (covariates) but with the difference that in the first group individuals receive the treatment while in the second they do

73

S. Alarcón, M. Sánchez / Food Policy 61 (2016) 70–79

not. In the present study the objective was to produce two groups of firms with similar characteristics in general but different with regard to whether they are exporters or not and whether they are innovators or not. Specifically, firstly innovating and noninnovating firms were paired in t  1 on the basis of the Z t1 variables so that their mean values would be the same for both groups. The Z t1 variables used are the same as the control variables for the probit bivariate model. In the second phase the propensity to export at the present moment t is quantified for both groups of forms and the ATT of innovation in t  1 on export in t is calculated as well. In the same way the exporting and non-exporting firms are paired in t  1 with the Z t1 variables and then the propensity to innovate at the present moment in both categories is estimated and the ATT of exports in t  1 on export in t is calculated as well. To optimize the proximity of the two groups a variety of techniques are used:  Mahalanobis distance (Rubin, 1980): the distance between observations is obtained and then each treated unit is matched to the nearest units that have not been treated.  Propensity score (Rosenbaum and Rubin, 1983, among others). This is the probability of each individual receiving the treatment and can be obtained through a logit regression (probit) in which the dependent variable is the treatment and the regressors the selected characteristics; with these probabilities groups (clusters or matches) are established for each observation that has received the treatment (experimental group), there being at least one other observation that has not received the treatment (control group) with a very similar propensity score. Like the previous method it is a procedure which captures the multidimensionality of the variables in a single measurement. As shown by Rosenbaum and Rubin (1985) it is useful to combine these methods because the Mahalanobis distance is useful for minimizing the distance between the X coordinates while the propensity score matching is very good for minimizing the discrepancies in the one-dimensional measurement.  Genetic Matching (Sekhon, 2011; Sekhon and Grieve, 2012) This generalizes the two previous procedures and improves the proximity between the two groups by adding additional properties and maximizing the balance of the X variables between the treated and control units through the use of genetic algorithms. The implementation and computational details of this procedure can be consulted in Sekhon (2011) and Mebane and Sekhon (2011). The goodness of these procedures can be evaluated by checking the average values of the variables in both groups, and checking whether there are significant differences. As already mentioned, in the second phase the difference between both groups in terms of the variable of interest is determined, that is, the mean effect of the treatment on the treated, ATT. The inference is carried out using the standard errors of Abadie and Imbens (2006), which take into account biases produced by the matching. Data This study used the PITEC (Innovation Technology Panel) data base (http://icono.fecyt.es/PITEC) built jointly by the INE (Spanish National Statistics), the Spanish Science and Technology Foundation (FECYT) and the Foundation COTEC following the guidelines in the Oslo Manual (OECD, 1997). Its ultimate goal is to provide statistical information regarding businesses’ technological activities and to analyze their evolution over time so as to identify the different innovation decisions they adopt. PITEC has numerous

Table 1 Definition of variables (per company and year). Ex ExRD InRD INNPROD INNPROC Productivity Capitalintensity Size

Takes value 1 if the company exports and 0 otherwise Logarithm of deflated external R&D total expenditures Logarithm of deflated internal R&D total expenditures Takes value 1 if the company is innovating in products Takes value 1 if the company is innovating in process Logarithm of the ratio turnover over number of employees Logarithm of the ratio accumulated capitala over number of employees Logarithm of number of employees

a

Built based on gross investment in material goods. Capital stock was elaborated applying the perpetual inventory method and 10% annual depreciation rates.

Table 2 Variables statistics. Mean

Standard deviation

Agricultural EXPORT exRD inRD INNPROD INNPROC Productivity Capitalintensity Size

0.5648 3.6067 9.4285 0.4611 0.5691 11.8841 9.5482 2.9811

0.4961 8.8273 9.4872 0.4988 0.4955 1.1864 4.1580 1.3275

Food EXPORT exRD inRD INNPROD INNPROC Productivity Capitalintensity Size

0.7472 2.1751 7.5365 0.5271 0.6605 12.6454 9.9761 3.7458

0.4347 8.2763 9.6357 0.4993 0.4736 0.9585 3.4556 1.4875

advantages: easy access, it can be compared to statistics from other OECD countries, it is not limited to manufacturing firms, its panel structure permits the study of the dynamics of innovation and control for the specific effects of firms, etc. PITEC is made up of 5 data bases including firms of different sizes and R&D activities. The largest one (MID) includes firms of all sizes which implement R&D internal expenditure. Therefore, PITEC is not representative of the company population but it is highly useful when it comes to studying the effects and consequences of technological activities in each sector. We have extracted 2 samples from PITEC, which we have called Agricultural and Food. The Agricultural sample includes agricultural, livestock, forestry and fishing companies (codes NACE-93 01, 02 and 05, and codes NACE-2009 01, 02 and 03) and the Food sample includes information regarding food, beverages and tobacco companies (codes NACE-93 15 and 16, and codes NACE2009 10, 11 and 12). One of the main drawbacks of PITEC is that it does not provide more specific details about the business activities of the firms. Therefore, the level of knowledge in terms of the composition of the two subsamples is fairly basic: the Agricultural sample consists of companies that interact with natural resources while the Food sample is made up of companies with industrial production processes and whose inputs are of agricultural commodities. Both panels are incomplete. The Agricultural subsample includes 165 different companies which present information for some of the 6 years in the period 2006–2011, mainly SMEs,2 91.0%. The Food subsample has 783 companies, 74.2% of which are SMEs and 25.8% large companies.

2 In PITEC the SME-large company classification is carried out by means of the number of employees indicator, the limit between categories being 200.

74

S. Alarcón, M. Sánchez / Food Policy 61 (2016) 70–79

The variables used in this study are defined in Table 1. Innovation has been characterized through input measures such as external R&D expenditure and internal R&D expenditure and output variables such as process innovation and product innovation. The importance of internal and external R&D and other forms of innovation has been revealed by different authors (Cassiman and Table 3 Bivariate probit estimations in the agricultural and food companies. EXPORTt

INNPRODt

EXPORTt

INNPROCt

Agricultural

q = 0.0341 Intercept EXPORTt1 INNPRODt1

3.2852*** (0.6962) 2.3779*** (0.1304) 0.2254* (0.1303)

1.9188*** (0.6822) 0.1109 (0.1286) 2.2640*** (0.1297)

INNPROCt1 Productivityt1 Capitalintensityt1 Sizet1 inRDt1 exRDt1

0.1493*** (0.0552) 0.0036 (0.0160) 0.0666 (0.0483) 0.0231*** (0.0080) 0.0205** (0.0084)

0.0528 (0.0551) 0.0042 (0.0165) 0.0212 (0.0486) 0.0227*** (0.0078) 0.0082 (0.0080)

q = 0.0511

3.1400*** (0.6921) 2.3746*** (0.1300)

0.9119 (0.6739) 0.0037 (0.1282)

0.1456 (0.1366) 0.1422*** (0.0551) 0.0069 (0.0164) 0.0574 (0.0481) 0.0250*** (0.0078) 0.0214** (0.0085)

2.1954*** (0.1327) 0.0453 (0.0546) 0.0256 (0.0173) 0.0817* (0.0488) 0.0229*** (0.0075) 0.0067 (0.0082)

Food

q = 0.0248 Intercept EXPORTt1 INNPRODt1

2.8333*** (0.4237) 2.6973*** (0.0661) 0.0394 (0.0709)

2.1990*** (0.3982) 0.1514** (0.0721) 2.5017*** (0.0622)

INNPROCt1 Productivityt1 Capitalintensityt1 Sizet1 inRDt1 exRDt1

0.1068*** (0.0345) 0.0190** (0.0091) 0.0385* (0.0223) 0.0147*** (0.0040) 0.0004 (0.0045)

0.0080 (0.0324) 0.0235** (0.0100) 0.1634*** (0.0212) 0.0347*** (0.0033) 0.0098** (0.0039)

q = 0.0350

2.8525*** (0.4242) 2.6926*** (0.0661)

1.9096*** (0.3669) 0.1604** (0.0645)

0.1219* (0.0711) 0.1069*** (0.0345) 0.0168* (0.0091) 0.0368* (0.0222) 0.0138*** (0.0039) 0.0003 (0.0045)

2.1410*** (0.0586) 0.0130 (0.0300) 0.0337*** (0.0090) 0.1185*** (0.0193) 0.0236*** (0.0031) 0.0145*** (0.0039)

Veugelers, 2006; Lokshin et al., 2008; Alarcón and Sánchez, 2013; Hashi and Stojcic, 2013; Hervas et al., 2014 among others). The selection of innovation options generally depends on the technological intensity of each company, on the activities they carry out and on their size (Arora and Gambardella, 1990; Audretsch et al., 1996; Veugelers and Cassiman, 1999; Cassiman and Veugelers, 2006; Ghazalian and Furtan, 2007; Schmiedeberg, 2008; VegaJurado et al., 2008; Lazzarotti et al., 2011). Furthermore, the labor productivity, capital intensity and size variables seek to collect information on companies’ heterogeneity in terms of their productive processes. Table 2 collects the main statistics of the selected variables in both subsamples. The mean values of the export variable indicate that the proportion of Food exporting companies (74.7%) is higher than in the Agricultural subsample (56.5%). These percentages indicate that the bulk of the most innovative agricultural and food companies tend to export. This gives an initial idea of the positive relationship between internationalization and innovation activities also in the agri–food sector, as previous studies in diverse economic fields have revealed (Cassiman and Martinez-Ros, 2007; Caldera, 2010, among others). As to innovation variables, the fact that Agricultural companies show higher values in innovation inputs in relation to Food companies is striking, but the latter achieve higher percentages in innovation outputs. Thus, exRD and inRD mean values are higher for Agricultural companies, which reflects higher innovation and external and internal R&D expenditure, but Food companies show higher percentages in product and process innovation. A possible explanation would be that the different productive processes of these two kind of companies and their differences in company structure means that in Agricultural firms successful innovation requires, on average, higher efforts in terms of resources. Moreover, the Productivity, Capitalintensity and Size values show higher mean amounts in Food firms, which would be a possible explanation of the previous point: larger size in Food companies implies economies of scale in innovation and internationalization processes, or transmission along the food chain (Ghazalian and Furtan, 2007).

Results General results The bivariate probit estimations are shown in Table 3 for agricultural and food companies we estimate export and product innovation equations jointly as well as exports and process innovations equations. Although the food panel displays more

Standard deviation between brackets * Significance at 10 per cent level. ** Significance at 5 per cent level. *** Significance at 1 per cent level.

Table 4 ATT, average treatment effect for the treated. Agricultural firms 1 lag

Y ¼ EXPORT t ; Tr ¼ INNPRODts Y ¼ EXPORT t ; Tr ¼ INNPROC ts Y ¼ INNPROD; Tr ¼ EXPORT ts Y ¼ INNPROC t ; Tr ¼ EXPORT ts

Food firms 2 lags

1 lag

2 lags

PS

Genetic

PS

Genetic

PS

Genetic

PS

Genetic

0.1230** (0.0510) 0.1025 (0.0672) 0.0415 (0.0490) 0.0174 (0.0469)

0.0921 (0.0573) 0.0538 (0.0595) 0.0572 (0.0594) 0.0342 (0.0538)

0.1254** (0.0621) 0.1524** (0.0715) 0.0546 (0.0601) 0.0251 (0.0583)

0.0955 (0.0672) 0.0030 (0.0673) 0.0094 (0.0624) 0.0754 (0.0568)

0.0428* (0.0226) 0.0528** (0.0239) 0.0630*** (0.0239) 0.0586** (0.0239)

0.0514** (0.0257) 0.1010*** (0.0285) 0.0808*** (0.0290) 0.0672** (0.0286)

0.0325 (0.0263) 0.0900*** (0.0280) 0.0317 (0.0273) 0.0860*** (0.0279)

0.0511* (0.0291) 0.0884*** (0.0315) 0.0647** (0.0320) 0.0551* (0.0318)

Abadie-Imbens standard errors between brackets. PS: Propensity Score Matching. Genetic: genetic matching. * Significance at 10 per cent level. ** Significance at 5 per cent level. *** Significance at 1 per cent level.

75

S. Alarcón, M. Sánchez / Food Policy 61 (2016) 70–79

Fig. 1. Total exports and innovation.

Table 5 ATT by export destination. Agricultural firms 1 lag

Exports within the EU Y ¼ EXPORT t ; Tr ¼ INNPRODts Y ¼ EXPORT t ; Tr ¼ INNPROC ts Y ¼ INNPROD; Tr ¼ EXPORT ts Y ¼ INNPROC t ; Tr ¼ EXPORT ts Exports beyond the EU Y ¼ EXPORT t ; Tr ¼ INNPRODts Y ¼ EXPORT t ; Tr ¼ INNPROC ts Y ¼ INNPROD; Tr ¼ EXPORT ts Y ¼ INNPROC t ; Tr ¼ EXPORT ts

Food firms 2 lags

1 lag

2 lags

PS

Genetic

PS

Genetic

PS

Genetic

PS

Genetic

0.1407*** (0.0500) 0.0937 (0.0676) 0.0196 (0.0514) 0.0075 (0.0461)

0.1017* (0.0574) 0.0563 (0.0559) 0.0296 (0.0588) 0.0219 (0.0530)

0.1326** (0.0617) 0.1423** (0.0712) 0.1423** (0.0712) 0.0351 (0.0601)

0.0999 (0.0732) 0.0037 (0.0652) 0.0110 (0.0606) 0.0674 (0.0556)

0.0403* (0.0228) 0.0535** (0.0245) 0.0497** (0.0236) 0.0773*** (0.0231)

0.0245 (0.0253) 0.1081*** (0.0294) 0.0776*** (0.0285) 0.0665** (0.0279)

0.0279 (0.0267) 0.0878*** (0.0283) 0.0467* (0.0274) 0.0783*** (0.0271)

0.0504* (0.0297) 0.0883*** (0.0320 0.0528* (0.0296) 0.0546* (0.0292)

0.1675*** (0.0488) 0.0539 (0.0656) 0.0645 (0.0535) 0.0262 (0.0533)

0.1122** (0.0516) 0.0651 (0.0544) 0.0431 (0.0570) 0.0299 (0.0565)

0.1276** (0.0580) 0.1165* (0.0654) 0.1133* (0.0623) 0.0399 (0.0613)

0.0844 (0.0683) 0.0647 (0.0635) 0.1194* (0.0624) 0.0130 (0.0625)

0.0300 (0.0284) 0.0625** (0.0296) 0.0202 (0.0189) 0.0379** (0.0181)

0.0121 (0.0320) 0.1082*** (0.0341) 0.0239 (0.0216) 0.0306 (0.0213)

0.0258 (0.0337) 0.0612* (0.0336) 0.0005 (0.0211) 0.0303 (0.0201)

0.0433 (0.0376) 0.0751** (0.0377) 0.0246 (0.0237) 0.0258 (0.0236)

Abadie-Imbens standard errors between brackets. PS: Propensity Score Matching. Genetic: genetic matching. * Significance at 10 per cent level. ** Significance at 5 per cent level. *** Significance at 1 per cent level.

observations than the agricultural one, 4069 versus 841, the variability explained by the regressors is high in both cases, around 50% in the agricultural panel and 58% in the food one. In all four cases the correlation coefficients are small and close to zero, which indicate that export and innovation decisions are not taken jointly.3 The control variables indicate that food companies with greater capital intensity and with higher internal spending on R&D are most likely to export and innovate in both product and process in subsequent years. Besides, productivity contributes to becoming an exporter, and the proportion of external costs in R&D fosters innovation. 3 The individual probit estimations are not shown in order to save space but the coefficients and significance are very close to those in Table 3.

The situation looks very different in farms, where only the proportion of internal expenditure on R&D is positive and significant in all the estimations and therefore helps increase the probability of exporting and innovating in subsequent years. Therefore, the usual effect of the importance of the absorption capacity for innovation can be initially detected when an internal R&D effort is made. As with the food firms, productivity increases the probability of becoming an exporter. The delayed dependent variable, EXPORT t1 ; INNPRODt1 ; INNPROC t1 , are in all cases positive and highly significant; it shows coefficient values and marginal effects higher than all the other variables. This reveals the inertia generated by exports and innovation: once exporting or innovation activities are launched the odds of continuing in successive years are much higher. In the same way Alfranca et al. (2002) and Triguero et al. (2013) have already

76

S. Alarcón, M. Sánchez / Food Policy 61 (2016) 70–79

suggested that food firms are usually persistent in innovation activities, while Girma et al. (2008) found persistence in exporting among firms in Ireland and Great Britain. This study confirms the persistence of exporting and both product and process innovation not only in food firms but also in agricultural ones. Once the good behavior of the control variables is confirmed we analyze the variables of interest. Significance is high from lagged product innovation to current exports in Agriculture but not vice versa; and both directions between process innovation and export in Food companies. Therefore, there is evidence of a positive relationship between innovation and exports as expected based on previous studies (Basile, 2001; López Rodríguez and García Rodríguez, 2005; Caldera, 2010; Cassiman et al., 2010; Jiménez Castillo, 2013). In a similar vein the work of Love and Ganotakis (2013) holds that exporting helps innovation but does not lead to more intensive innovation, even if there is a positive scale effect of exposure to export markets, though they observed differences depending on whether the sectors in question are high-intensity or low-intensity innovators. Sterlacchini (1999) also detected this relationship for supplier dominated sectors (which is often the case in the agri–food sector). The matching approach (Table 4) was carried out with one or two lags of the treatment variables and using the propensity score (PS) and genetic matching. The results show more differences than

similarities between the two types of firms considered. The similar behavior is that the majority of the coefficients obtained are positive as was expected and this shows once more the narrow positive relationship between exporting and innovation, already demonstrated in other countries and contexts (among others Cassiman et al., 2010; Kumar et al., 2013; Triguero et al., 2013). However, the statistical significance is overwhelming in the case of food and scant in that of agriculture. In the case of the former there is sufficient evidence to affirm that both product and process innovation carried out in previous years (One- and two-year lags) increase the proportion of exporting firms. Furthermore, there is also evidence that goes in the opposite direction, as it were, and that suggests that internationalization stimulates both product and process innovation. In agricultural firms, by contrast, there is only the weakest evidence that product and process innovation has a positive influence on the exporting character of firms and there is no evidence at all that exports lead to an increase in product or process innovation. This result could indicate that both business decisions, that of innovation and that of exporting, have deeper roots in food firms, which have a higher innovation output. Innovation processes differ between industries for technological reasons (Pla-Barber and Alegre, 2007) and it is reasonable to think that there exist intrinsic factors in each of the exported products.

Fig. 2. (a) Exports within EU and innovation; and (b) exports beyond EU and innovation.

S. Alarcón, M. Sánchez / Food Policy 61 (2016) 70–79

Fig. 1 may illustrate what is occurring. There are no big differences among the agricultural exporters with regard to whether they carried out product innovation in the past; in agricultural firms 48.3% had not carried out product innovation while 51.7% had done so while in the case of food firms the figures are 40.2% and 59.8% respectively. With regard to process innovation, among agricultural firms, 36.1% of those which exported did not carry out process innovation while 63.9% did do so. The same figures for food firms were 26.7% and 73.3%, respectively. Turning now to food firms that carried out product innovation we can see that 16.9% did not do so in the previous period while 83.1% did. Similarly, 18% of those who carried out process innovation did not export but 82.0% did. In the case of agricultural firms, the differences between these percentages is lower, 36.3% (did not previously export) and 63.7% (did previously export) for product innovation and 39.3% (did not previously export) and 60.7% (did previously export) for process innovation. One possible explanation for the differences in the results between agricultural and food firms could be the different level of maturity of the firms with regard to exporting. Thus agricultural firms would be less familiar with exports and it would be the more innovative ones which would be more inclined to take the decision to export. In the case of food firms innovation is also crucial for dealing with internationalization. Furthermore, the firms with greater experience of exporting make use of their knowledge of consumer needs in other countries and foreign markets to provide positive feedback for their innovation processes. Empirical studies offer greater evidence that firms that carry out more innovation also have a greater propensity to export and to intensify exports (Basile, 2001; Roper and Love, 2002; López Rodríguez and García Rodríguez, 2005; Caldera, 2010, among others), though there are other studies (Aw et al., 2007; Girma et al., 2008) which show the existence of bi-directional relations and which therefore support the idea that firms strengthen and learn innovation through internationalization (Damijan et al., 2010). By export destination The same analysis was carried out by breaking down the export variable by country of destination, that is to say, whether it was an EU (including EFTA and candidate nations) country or a non-EU one (Table 5). Thus Monreal et al. (2012) suggest that exports are affected by several factors, including the innovative effort, indicating in their study that it is relevant to control for the type of destination of the international exchange. For example, exports to highly competitive markets may involve greater innovation efforts, as indicated by Girma et al. (2008) for Ireland. The results obtained show that the case of EU exports the results are similar to those already described for exports as a whole; in food firms there are significant causal relations in both directions and in both types of innovations but in agricultural firms only the product innovation is significant over exports in the genetic matching with one lag. However, in the case of exports beyond the EU product innovation does play a significant role in agricultural firms as does process innovation in food ones. The causal relationship of exports to innovation loses significance in the case of exports outside the EU in the case of food companies. These results are consistent with the idea that companies have greater market penetration in the countries closest to them, and although the opening to more distant markets has started with a major effort and some time ago, there is still a long way to go before achieving positive innovation–exports feedback relationship similar to that which Spanish food firms have with countries in Europe. Future research should also look at the level of internationalization (Madsen and Servais, 1997) which firms of both types have in the closest and most

77

distant markets. Fig. 2 shows how the proportion of exporters, both inside and outside the EU is greater if the firm innovated in previous years, especially if the firm is a food one and to a lesser degree if it is an agricultural one. Turning now to firms which innovate, in the food category the proportion is always greater if the firm was an exporter independently of the destination but among agricultural firms this is only true for exports with the European Union.

Conclusions The study of the relationship between two important business decisions capable of producing competitive advantages such as innovation and internationalization continues to be relevant given the differences that may exist in various economic contexts. Furthermore, the relationship between the two does not always have a positive significance, and it is not even clear that one influences the other or whether the relationship might be bi-directional. Thus this present study has sought to analyze, with a Resource-BasedView theoretical focus, the relationship between the two decisions in the food and agricultural sectors. The results obtained may be of use in the taking of decisions in the firms themselves, as well as being of assistance in the development of agricultural and food policy. The study was carried out in the European context, specifically in Spain, one of the European countries in which the food and agricultural sectors are of the greatest importance for GDP. Initially, the results show that internal R&D expenditure plays a more decisive role than external R&D expenditure in the innovation–exporting strategy of agri–food companies; this result has an impact on the importance of the absorption capacity generated by internal innovation efforts. However, both types of innovation effort (internal and external) are important in the sector, in line with the suggestion in Hervas et al. (2014). Additionally, both product and process innovation are essential in internationalization. Furthermore, internal and external R&D expenditure is proportionally higher in the agricultural sector, but in terms of innovation outputs the agri–food industry obtains better market results in terms of products and productive processes. A certain level of persistence has also been observed in the carrying out of both activities, that is to say, firms that export or innovate in a given period also tend to do so in the following one, both in the case of agricultural and food firms. Furthermore, this study shows, especially in the case of food firms, that the relationship between innovation and internationalization is bidirectional, with an increased human and economic effort being required from the firms to deal with both decisions. However, this complementary relationship between internationalization and innovation is seen in food firms but not in agricultural ones so that the efforts required in this area both for the firms themselves and at the policy level will be even greater. Furthermore, this production sector continues to maintain important links with the local population and the environmental sustainability that they could produce if supported by appropriate agricultural policies. However, as Tambo and Wünscher (2014) have recently indicated, farmers must not only be seen as adopters of externally driven innovation as some of them also show innovationgenerating behavior. The experience of food firms of PITEC sample is, without a doubt, a good model to follow for other food firms. And for the agricultural ones if they also wish to develop a virtuous relationship between innovation and exports and, given their proximity in the production chain, knowledge transfer between the two sectors is recommended regarding how to approach internationalization and innovation but particularly in how to learn from foreign markets about launching new products and improve their production processes. Furthermore, it remains important to continue

78

S. Alarcón, M. Sánchez / Food Policy 61 (2016) 70–79

studying the relationship between innovation and exporting, as the work of Lachenmaier and Wößmann (2006) has detected that the impact of innovation on exporting is higher in the more traditional sectors (including the agri–food sector) than in the general economy. It must also be recalled, as indicated by recent studies like Boermans and Roelfsema (2015) that there are different options in the area of internationalization with implications for economic and productive effort and assumed risk (international outsourcing, exports or plant development) effort and it could be of interest to study the future implications of the different modes of access to international markets. This in turn might produce different results in terms of the different types of efforts and innovation performance. It must be pointed out, nonetheless, that this positive relationship between innovation and exports occurs more in the case of exports to neighboring countries, as opposed to more distant ones, and that it is with the latter that a generally greater proportion of the internationalization process occurs. In other words, in the case of more distant countries there is not such a significant relationship between exports and innovation. This might be due to a lack of a developed internationalization strategy or because it is not closely linked to innovation. There is, therefore, a requirement for future research to look in greater detail at the differences in internationalization processes between nearby and more distant countries, as well as the economic results obtained from various combinations of innovation and internationalization efforts. The study has shown the importance of the relationship which under specific circumstances exists between innovation and internationalization and has also shown that that there are firms which make significant bidirectional efforts in this regard. These organizations would benefit from the existence of appropriate agricultural and food policies, directed not only toward them as individual enterprises but also toward creating institutional frameworks which would help develop on a collective basis those practices which would produce a better and more competitive global positioning for these sectors. Authors such as Castellacci (2008), Gallego (2010), Hashi and Stojcic (2013) and Matsumura et al. (2013) have already shown the importance for innovation in the agri–food sector of appropriate decisions at the level of the individual firm and the critical role of the institutions and policies, in this case at the European level. In this regard, following Karantininis et al. (2010) and Rogers (2004), the development of successful structures via vertical integration in the food chain or the presence of networks can help improve the results obtained. Furthermore, Yi et al. (2013) propose using instrumental variables as moderators in the relationship between innovation and internationalization variables. Furthermore, Seharmur et al. (2015) suggest that the presence of KIBS can also moderate the effects of these business decisions in the organization. Thus their analysis may be of interest for future research.

References Abadie, A., Imbens, G., 2006. Large sample properties of matching estimators for average treatment effects. Econometrica 74, 235–267. Abecassis-Moedas, C., Mahmoud-Jouini, S.B., 2008. Absorptive capacity and sourcerecipient complementarity in designing new products: an empirically derived framework. J. Prod. Innov. Manage. 25, 473–490. Alarcón, S., Sánchez, M., 2013. External and internal R&D, capital investment and business performance in the spanish agri-food industry. J. Agric. Econ. 64 (3), 654–675. Alfranca, O., Rama, R., von Tunzelmann, N., 2002. A patent analysis of global food and beverage firms: the persistence of innovation. Agribusiness 18 (3), 349– 368. Arnalte, E., Ortiz, D., Moreno, O., 2008. Cambio estructural en la agricultura española. Un nuevo modelo de ajuste en el inicio del siglo XXI. Papeles de Economía Española 117, 59–73. Arnalte, E., Ortiz, D., 2011. Breaking down the growth of family farms: a case study of an intensive Mediterranean agriculture. Agric. Syst. 114, 500–511.

Arora, A., Gambardella, A., 1990. Complementarity and external linkages: the strategies of the large firms in biotechnology. J. Ind. Econ. 38 (4), 361–379. Audretsch, D.B., Menkveld, A.J., Thurik, A.R., 1996. The Decision Between Internal and External R&D. Neuhuys-Research Institute. Aw, B., Roberts, M., Winston, T., 2007. Export market participation, investments in R&D and worker training, and the evolution of firm productivity. World Econ. 30 (1), 83–104. Barney, J.B., 1991. Firm resources and sustained competitive advantage. J. Manage. 17 (1), 99–120. Basile, R., 2001. Export behaviour of Italian manufacturing firms over the nineties: the role of innovation. Res. Policy 30 (8), 1185–1201. Becker, S.O., Egger, P.H., 2013. Endogenous product versus process innovation and a firm’s propensity to export. Empirical Economics 44 (1), 329–354. Blandford, D., Hill, B., 2005. Structural change and public policies in EU agriculture: an overview. In: XIth Congress of the EAAE, Copenhagen, Denmark, August. Boermans, M.A., Roelfsema, H., 2015. The effects of internationalization on innovation: firm-level evidence for Transition Economies. Open Econ. Rev. 26, 333–350. Caldera, A., 2010. Innovation and exporting: evidence from Spanish manufacturing firms. Rev. World Econ. 146 (4), 657–689. Capitanio, F., Coppola, A., Pascucci, S., 2009. Indications for drivers of innovation in the food sector. Br. Food J. 111 (8), 820–838. Capitanio, F., Coppola, A., Pascucci, S., 2010. Product and process innovation in the Italian food industry. Agribusiness 26 (4), 503–518. Cassiman, B., Veugelers, R., 2006. In search of complementarity in innovation strategy: internal R&D and external knowledge acquisition. Manage. Sci. 52 (1), 68–82. Cassiman, B., Martinez-Ros, E., 2007. Product innovation and exports. Evidence from Spanish Manufacturing. Working Paper. IESE. Cassiman, B., Golovko, E., Martínez-Ros, E., 2010. Innovation, exports and productivity. Int. Food Ind. Org. 28, 372–376. Castellacci, F., 2008. Innovation and the competitiveness of industries: comparing the mainstream and the evolutionary approaches. Technol. Forecast. Soc. Chang. 75 (7), 984–1006. Christensen, J.F., 1995. Asset profiles for technological innovation. Res. Policy 24 (5), 727–745. Clerides, S.K., Lach, S., Tybout, J.R., 1998. Is learning by exporting important? Microdynamic evidence from Colombia, Mexico, and Morocco. Quart. J. Econ. 113 (3), 903–947. Cohen, W.M., Levinthal, D.A., 1990. Absorptive capacity: a new perspective on learning and innovation. Adm. Sci. Q. 35, 128–152. Compes, R., García Alvarez-Coque, J.M., 2009. La reforma de la PAC y la agricultura española: alternativas y oportunidades para España. Fundación Alternativas. Documento de Trabajo. 40/2009. ˇ ., Polanec, S., 2010. From innovation to exporting or vice Damijan, J.P., Kostevc, C versa? World Econ. 33 (3), 374–398. Ebersberger, B., Herstad, S.J., 2013. The relationship between international innovation collaboration, intramural R&D and SMEs’ innovation performance: a quantile regression approach. Appl. Econ. Lett. 20 (7), 626–630. Eurostat, 2015. Structural Business Statistics. Falk, M., 2012. Quantile estimates of the impact of R&D intensity on firm performance. Small Bus. Econ. 39 (1), 19–37. FIAB, 2012. Informe de exportaciones, industria alimentación y bebidas 2011. . (Spanish Food and Beverage Association). Filippaios, F., Papanastassiou, M., Pearce, R., Rama, R., 2009. New forms of organisation and R&D internationalisation among the world’s 100 largest food and beverages multinationals. Res. Policy 38, 1032–1043. Food Drink, 2014. Economic Bulletin Q4. Gallego, J.R., 2010. La agricultura a tiempo parcial y la externalización de servicios agrarios como vehículo del cambio estructural. Revista Española de Estudios Agrosociales y Pesqueros 225, 13–45. García, M., Burns, J., 1999. Sources of technological development in the Spanish food and drink industry. A ‘‘supplier dominated” industry? Agribusiness 15 (4), 431– 448. Ghazalian, P.L., Furtan, W.H., 2007. The effect of innovation on agricultural and agrifood exports in OECD countries. J. Agric. Resour. Econ. 32 (3), 448–461. Girma, S., Greenaway, A., Kneller, R., 2004. Does exporting increase productivity? A microeconometric analysis of matched firms. Rev. Inter. Econ. 12 (5), 855–866. Girma, S., Görg, H., Hanley, A., 2008. R&D and exporting: a comparison of British and Irish firms. Rev. World Econ. 144 (4), 750–773. Greenaway, D., Kneller, R., 2007. Firm heterogeneity, exporting and foreign direct investment. Econ. J. 117 (517), 134–161. Grunert, K., Harmser, H., Meulenberg, M., Kuiper, E., Ottowitz, T., Declerck, F., Traill, B., Göransson, G., 1997. A framework for analysing innovation in the food sector. Product and Process Innovation in the Food Industry. Blackie Academic and Professional, London, UK. Guan, J.C., Yam, R., Mok, C.K., Ma, N., 2006. A study of the relationship between competitiveness and technological innovation capability based on DEA models. Eur. J. Oper. Res. 170 (3), 971–986. Hashi, I., Stojcic, N., 2013. The impact of innovation activities on firm performance using a multi-stage model: evidence from the Community Innovation Survey 4. Res. Policy 42, 353–366. Hervas, J.L., Sempere, F., Boronat, C., 2014. Process innovation strategy in SMEs, organizational innovation and performance: a misleading debate? Small Bus. Res. 43, 873–886.

S. Alarcón, M. Sánchez / Food Policy 61 (2016) 70–79 Hirsch, S., Schiefer, J., Gschwandtner, A., Hartmann, M., 2014. The determinants of firm profitability differences in EU food processing. J. Agric. Econ. 65 (3), 1–19. INE, 2015. Encuesta Industrial de Empresas 2013. . Jiménez Castillo, D., 2013. Análisis de los efectos moderadores de la inversión en I+D y la experiencia exportadora sobre la relación entre barreras a la exportación e intensidad exportadora bajo distintos niveles de turbulencia ambiental percibida. Cuadernos de Economía y Dirección de la Empresa 16 (2), 95–107. Karantininis, K., Sauer, J., Furtan, W.H., 2010. Innovation and integration in the agrifood industry. Food Policy 35 (2), 112–120. Kumar, V., Mudambi, R., Gray, S., 2013. Internationalization, innovation and institutions: the 3 I’s underpinning the competitiveness of emerging market firms. J. Int. Manage. 19 (3), 203–206. Lachenmaier, S., Wößmann, L., 2006. Does innovation cause exports? Evidence from exogenous innovation impulses and obstacles using German micro data. Oxford Econ. Papers 58 (2), 317–350. Lazzarotti, V., Manzini, R., Pellegrini, L., 2011. Firm-specific factors and the openness degree: a survey of Italian firms. Eur. J. Innov. Manage. 14 (4), 412–434. López Rodríguez, J., García Rodríguez, R.M., 2005. Technology and export behaviour: a resource-based view approach. Int. Bus. Rev. 14 (5), 539–557. Lokshin, B., Belderbos, R., Carree, M., 2008. The productivity effects of internal and external R&D: evidence from a dynamic panel data model. Oxford Bull. Econ. Stat. 70 (3), 399–413. Love, J.H., Ganotakis, P., 2013. Learning by exporting: lessons from high-technology SMEs. Int. Bus. Rev. 22, 1–17. Madsen, T.K., Servais, P., 1997. The internationalization of born globals: an evolutionary process? Int. Bus. Rev. 6 (6), 561–583. Matsumura, T., Matsushima, N., Cato, S., 2013. Competitiveness and R&D competition revisited. Econ. Model. 31, 541–547. Mebane Jr., W.R., Sekhon, J.S., 2011. Genetic optimization using derivatives: the rgenoud package for R. J. Stat. Softw. 42 (11), 1–26. Monreal, J., Aragón, A., Sánchez, G., 2012. A longitudinal study of the relationship between export activity and innovation in the Spanish firms: the moderating role of the productivity. Int. Bus. Rev. 21, 862–877. Nassimbeni, G., 2001. Technology, innovation capacity, and the export attitude of small manufacturing firms: a logit/tobit model. Res. Policy 30, 245–262. OECD, 1997. Oslo Manual: Proposed Guidelines for Collecting and Interpreting Technological Innovation Data. OECD Publishing, Paris. Penrose, E.T., 1959. The Theory of the Growth of the Firm. John Wiley, New York. Pla-Barber, J., Alegre, J., 2007. Analysing the link between export intensity, innovation and firm size in a science-based industry. Int. Bus. Rev. 16 (3), 275–293. Rama, R., 2008. Handbook of Innovation in the Food and Drink Industry. Haworth Press, New York. Rama, R., 2014. Innovación en la industria agroalimentaria. Características y factores de éxito, 83–100. Ciencia e Innovación en el Sistema Agroalimentario Español. (Coord. Aznar, J. y Calatrava, J.). Rogers, M., 2004. Networks, firm size and innovation. Small Bus. Econ. 22, 141–153.

79

Roper, S., Love, J.H., 2002. Innovation and export performance: evidence from the UK and German manufacturing plants. Res. Policy 31 (7), 1087–1102. Rosenbaum, P.R., Rubin, D.B., 1983. The central role of the propensity score in observational studies for causal EECTS. Biometrika 70 (1), 41–55. Rosenbaum, P.R., Rubin, D.B., 1985. Constructing a control group using multivariate matched sampling methods that incorporate the propensity score. Am. Stat. 39 (1), 33–38. Rubin, D.B., 1980. Bias reduction using mahalanobis-metric matching. Biometrics 36 (2), 293–298. Salomon, R.M., Shaver, J.M., 2005. Learning by exporting: new insights from examining firm innovation. J. Econ. Manage. Strategy 14 (2), 431–460. Schmiedeberg, C., 2008. Complementarities of innovation activities: an empirical analysis of the German manufacturing sector. Res. Policy 37, 492–1503. Seharmur, R., Doloreux, D., Laperrière, A., 2015. Is the degree of internationalization associated with the use of knowledge intensive services or with innovation? Int. Bus. Rev. 24, 457–465. Sekhon, J.S., 2011. Multivariate and propensity score matching software with automated balance optimization: the matching package for R. J. Stat. Softw. 42 (7), 1–52. Sekhon, J.S., Grieve, R.D., 2012. A matching method for improving covariate balance in cost-effectiveness analyses. Health Econ. 21 (6), 695–714. Smith, V., Dilling-Hansen, M., Madsen, E.S., 2002. Do R&D Investments Affect Export Performance? Institute of Economics, University of Copenhagen. Sterlacchini, A., 1999. Do innovative activities matter to small firms in non-R&Dintensive industries? An application to export performance. Res. Policy 28 (8), 819–832. Tambo, J.A., Wünscher, T., 2014. Building farmers’ capacity for innovation generation: what are the determining factors ?. European Congress of Agricultural Economists Congress ‘Agri-Food and Rural Innovations for Healthier Societies’. 26–29 August. Ljubljana, Slovenia. Traill, W.B., Meulenberg, M., 2002. Innovation in the food industry. Agribusiness 18 (1), 1–21. Triguero, Á., Córcoles, D., Cuerva, M.C., 2013. Differences in innovation between food and manufacturing firms: an analysis of persistence. Agribusiness 29 (3), 273–292. Vega-Jurado, J., Gutiérrez-Gracia, A., Fernández-de-Lucio, I., 2008. How do Spanish firms innovate? An empirical evidence. J. Technol. Manage. Innov. 3 (3), 100– 111. Veugelers, R., Cassiman, B., 1999. Make and buy in innovation strategies: evidence from Belgian manufacturing firms. Res. Policy 28 (1), 63–80. Wakelin, K., 1998. Innovation and export behaviour at the firm level. Res. Policy 26 (7), 829–841. Wernerfelt, B., 1984. The resource-based view of the firm. Strateg. Manage. J. 5 (2), 171–180. Yang, C.H., Cheng, Y.H., 2012. R&D, productivity, and exports: plant-level evidence from Indonesia. Econ. Model. 29, 208–216. Yi, J., Wang, C., Kafouros, M., 2013. The effects of innovative capabilities on exporting: do institutional forces matter? Int. Bus. Rev. 22 (2), 392–406.