Abstract
This article analyzes new co-innovative sources of firm labor productivity. Using survey data for 120 small and medium-sized travel agencies based in Catalonia (a region in the northeast of Spain) and partial least square–structural equation modeling (PLS-SEM) estimation techniques, three main findings emerged from the study. First, labor productivity is directly explained by those agencies’ capacity to exploit their assets, to use local networks, and to make international transactions. Second, the analysis of indirect effects on labor productivity suggests a circular causality, which is determined by the influence of the workers’ local network use on a firm’s export capacity. In this circular relationship, a firm’s capacity to generate market-leading product innovations and the stock of human capital and training play a decisive role. Third, co-innovation practices exert a negative effect, which may be related to difficulties in terms of securing productivity improvements in the short term.
Keywords
Introduction
Research into productivity is relevant because it is an indicator that gives us an insight into an economy’s long-term growth potential. Economic theory and available empirical evidence (Jorgenson, Ho, and Stiroh 2005) have shown that the level and growth of productivity derive from two basic sources: capital intensification/dependence (capital endowment per worker or hour worked) and the efficiency with which production factors are combined (total factor productivity, TFP). Because of the existence of diminishing returns, the mere accumulation of factors (extensive growth models) ends up weakening the sources of long-term economic expansion. Only when the accumulation of factors is accompanied by TFP improvements can the growth model become intensive and therefore sustainable in the long term (Timmer et al. 2010).
In the transition toward the global knowledge economy, empirical evidence has shown that information and communication technologies (ICT) are emerging as new explanatory sources of productivity (Dolfsma and Soete 2006; Torrent-Sellens and Vilaseca 2008; Torrent-Sellens 2015). The reason for this is twofold: first, their direct contribution to increased productivity and economic growth (Jorgenson and Vu 2007; Jorgenson, Ho, and Stiroh 2008), and second, their indirect contribution resulting from the generation of co-innovation, especially through human capital improvement and organizational change, that enhances an economy’s TFP (Pilat 2006). In this context, one of the reasons explaining the current severity of the economic crisis in Spain is the country’s lack of adaptation to the knowledge economy. The Spanish economy’s sources of productivity are not the best suited to competition in global knowledge-based markets (Mas and Stehrer 2012).
From the perspective of analyzing the impact of ICT on firm productivity, empirical evidence has highlighted two complementary trends. First, that returns on ICT investment and use are relatively much higher than those on investment in other physical components, and second, that ICT investment and use often go hand in hand with other endeavors, generally human capital improvement and organizational structure change (Bresnahan, Brynjolfsson, and Hitt 2002; Arvanitis 2005). Indeed, the transformative impact of ICT investment and use on firm productivity becomes more evident through co-innovation processes (Black and Lynch 2001, 2004; Brynjolfsson and Hitt 2003; Torrent-Sellens and Ficapal 2010).
Empirical evidence of new co-innovative sources of productivity has been obtained mainly from samples of large firms, while research into small and medium enterprises (SMEs) in general, and into small and medium-sized travel agencies in particular, is rather scarce (Wymenga et al. 2012; Hadad et al. 2012; Torrent-Sellens and Díaz-Chao 2014). Therefore, the aim of the study is to analyze new co-innovative sources of firm productivity in small and medium-sized travel agencies and, consequently, to expand the available evidence. To that end, we have used microdata gathered in 2010 from a sample of 120 small and medium-sized travel agencies in Catalonia (a region in the northeast of Spain) in order to propose and contrast, through partial least square–structural equation modeling (PLS-SEM) estimation techniques, a model of causal relationships that includes direct and indirect determinants of firm productivity. This method allows for the analysis of relationships not only between the various factors considered explanatory of productivity, but also between such factors (Nunkoo, Ramkissoon, and Gursoy 2013). Thus, the analysis completes the structural form explaining the productivity of small and medium-sized travel agencies.
This article is organized as follows. The next section presents a review of the literature on relationships between ICT, innovation, and firm productivity, especially in small and medium-sized travel agencies. The third section describes the data and research design. The fourth section reports the model and the research hypothesis. The fifth section describes the empirical findings, and the sixth section provides conclusions, discussion, and policy implications based on those findings.
ICT, Innovation, and Productivity in Small and Medium-Sized Travel Agencies: Literature Review
ICT investment and use do not give rise to generalized productivity improvements until firms and their workers have achieved the required technological, educational/training, organizational, business, labor and cultural competencies. In other words, the role of ICT as a general-purpose technology needs organizational and business process changes to fully exploit its growth opportunities (Ceccobelli, Gitto, and Mancuso 2012). New evidence has demonstrated the existence of these co-innovative sources of productivity among broad samples of firms, first in the United States and then in the rest of the world (for a review of the empirical literature see Matteucci et al. 2005; Draca, Sadun, and Van Reenen 2007; Jiménez-Rodríguez 2012; Cardona, Kretschmer, and Strobel 2013).
Despite this abundant evidence, generally obtained from large firms, there is relatively little available evidence on co-innovative sources of productivity in SMEs in general (Hall, Lotti, and Mairesse 2009; Torrent-Sellens and Díaz-Chao 2014), and in small and medium-sized travel agencies in particular (Barros and Alves 2004; Blake, Sinclair, and Soria 2006; Fuentes and Alvarez 2012; Hadad et al. 2012).
Indeed, although there is considerable evidence in the literature of the impact that ICT (Buhalis 1998; Buhalis and Law 2008) and innovation (Novelli, Schmitz, and Spencer 2006; Hjalager 2010; Camisón and Montfort-Mir 2012; Williams 2014) have on tourism firms, few studies have addressed specific productivity-related problems in small and medium-sized travel agencies (Thomas, Shaw, and Page 2011; Spencer, Buhalis, and Moital 2012). Thus, a validated model of sources of productivity in small and medium-sized travel agencies would be a very practical and useful instrument for evaluating firm efficiency, a task that is not without its difficulties in the context of SMEs, whatever their type (Audretsch 2002, 2006; Hall, Lotti, and Mairesse 2009).
Regarding research into ICT, innovation, and firm productivity in travel agencies, the available evidence suggests that physical capital together with human capital and innovation explain the level of productivity in travel agencies (Blake, Sinclair, and Soria 2006). Other determinants of efficiency in travel agencies are the implementation of e-booking systems, the level of wages, and the age of the firm (Sellers-Rubio and Mas-Ruiz 2009).
In turn, performance in the intermediation industry is explained by technical progress. Hence, investment in organizational and managerial factors in combination with a good balance between inputs and outputs helps firms to achieve positive technical efficiency change. To perform better, travel agencies must adopt new technologies and upgrade managerial skills (Barros, Botti, and Peypoch 2009).
The action of improving productivity in travel agencies usually focuses on capital accumulation and innovation in a push effect that integrates an upward shift to a higher technology change (Barros and Matias 2006; Assaf, Barros, and Machado 2011). In travel agencies, the technology efficiency score is actually related to the dissemination of best technological practices in the business, where assets, human capital, training, management, and organization play an important role (Barros and Dieke 2007). Thus, local networks become critical to improving technological exchange, since efficiency and productivity increase through knowledge transfer between firms (Assaf, Barros, and Dieke 2011).
Although this initial evidence points to some relationships of complementarity in explaining the productivity of travel agencies, the intention of this study is to go one step further. In small and medium-sized travel agencies, the relationships between ICT, innovation, and productivity are not necessarily direct, as they may also be generated indirectly, that is, through the impact on a firm’s other productivity-related results, such as its export capacity or its assets. In this respect, and based on the idea that ICT and co-innovation are levers of change, the aim of this study is to design and test, using PLS-SEM, a more comprehensive model explaining the direct and indirect sources of productivity in small and medium-sized travel agencies (Nunkoo, Ramkissoon, and Gursoy 2013).
This approach is quite consistent with the further progress made by recent research on innovation and productivity in tourism. The first reason is that innovation in tourism is not an isolated event. Much innovative power in tourism does not originate from tourism itself. Tourism innovation is strongly interrelated with other economic and social fields (Hjalager 2015). The second reason is that the tourism innovation analysis should be made with a specific approach, taking into account their different typologies. Usually, innovation term has been used to signify any change undertaken by an organization, business, or individual, without regard to its extent, context, or value contribution to tourism. Terminology suggesting that innovation is either an incremental or radical change to existing conditions has simply been transferred from manufacturing to tourism. However, since radical innovation is rare in the tourism sector, a broader framework based on its characteristics is required to clearly delineate distinctive innovation approaches (Brooker and Joppe 2014). Finally, more recent research has also indicated the existence of specific pathways in tourism productivity. Empirical evidence have addressed what drives productivity improvements and the role of changes in physical capital, innovation, and the competitive environment, and others studies have focused on drivers that include information and communication technology, size of firm, competition versus cooperation, and clustering. However, most research in this area focuses on the critical role played by human capital and has shown that tourism productivity is more likely to come from innovations that result in enhanced product and service labor quality than from cost-cutting (Joppe and Lee 2014).
Data and Research Design
The study used survey data for a sample of 120 small and medium-sized travel agencies (firms with 50 or fewer employees) operating in Catalonia. The sampling universe comprised 1,790 firms with an overall margin of error of ±6.7% in the case of maximum indetermination, p = q = 50, for a confidence level of 95.5%. Catalonia is a region in the northeast of Spain where SMEs account for the bulk of economic activity. Generally, small and medium-sized travel agencies make medium-intensity use of ICT and have high levels of worker and manager education and training, a good record in innovation activities, and important competitiveness problems as a result of the severity of the economic crisis (Torrent-Sellens 2011). Table 1 shows some of the main statistics describing the value process in the sample of small and medium-sized travel agencies in Catalonia.
Descriptive Statistics of Small and Medium-Sized Travel Agencies in Catalonia.
Source: Own elaboration.
A preliminary version of the questionnaire was drafted following the literature review (Joppe and Lee 2014), at the same time, drawing on the research team’s experience in similar studies (Torrent-Sellens and Vilaseca 2008; Torrent-Sellens and Ficapal 2010; Torrent-Sellens and Díaz-Chao 2014). This version was used for the study pilot test. The questionnaire validation process was undertaken in the third week of March 2010 on a sample of 20 respondents. A number of issues were thrown up by the pilot test. The first was the length of the questionnaire and the time taken to complete it. Most of the respondents felt that the initial questionnaire was very long and difficult to answer. To solve this problem, customized response paths were incorporated into it according to the size of the firm being surveyed. The second was associated with difficulties understanding the technical and technological terms, especially those connected with ICT applications and business (e.g., ERP, CRM, B2B, B2C). To solve this problem, simpler definitions of these terms were incorporated into the final version of questionnaire. The third was the lack of response options to certain questions. The number of options was increased based on the respondents’ suggestions. The fourth and final issue was the complexity of response options to questions about newly created firms’ economic and financial information. The options were simplified based on the respondents’ comments. The findings obtained from the pilot test therefore suggested that the research team should make a number of changes to the original questionnaire in order to simplify the technical questions and facilitate the response process. Despite these issues, the pilot test respondents considered the study to be very positive and showed themselves to be very receptive, collaborative, and interested in the final results of it.
The final version of the questionnaire used in the survey contained 42 questions, against which a scoring value had to be assigned. It was answered by business owners or directors with an overall view of the activities of their firms, in telephone interviews using computer-assisted telephone interviewing (CATI), lasting for half an hour each. By gathering data on the value chain, the aim of the study was to analyze new sources of productivity in small and medium-sized travel agencies in Catalonia. A study presentation letter was written to inform potential respondents about the confidentiality of any data provided and the academic aim of the research. The business owners and directors voluntarily answered the questionnaire and did not receive any payment in cash or kind. While the questionnaire was being implemented, an expert was on hand at all times (on the phone and by e-mail) to resolve any queries that the respondents had. The respondent firms were selected by means of probability sampling applied to the small and medium-sized travel agencies contained in the official database of Spain’s Mercantile Register. The response rate was 17% (one respondent firm for every six small and medium-sized travel agencies contacted). The fieldwork was carried out between April and May 2010. The research was conducted by researchers from the interdisciplinary research group on ICT, i2TIC (http://i2TIC.net), and was funded by the Information Society Observatory Foundation (FOBSIC), belonging to the Government of Catalonia.
Hypotheses and Model
In order to identify the presence of relationships of complementarity (co-innovation) in the explanation of the level of productivity in Catalan small and medium-sized travel agencies, PLS-SEM was used to estimate the model and test the proposed hypotheses. In recent years, researchers have become increasingly interested in using PLS-SEM because of its capacity to model latent constructs under conditions of non-normality and small-to-medium sample sizes (Díaz-Casero, Hernandez, and Roldán 2011). For these reasons, PLS-SEM has now gained acceptance in the management and economics field (Hair, Ringle, and Sarstedt 2011; Nunkoo, Ramkissoon, and Gursoy 2013).The use of this technique involves two stages or approaches. The first requires the evaluation of the measurement model, allowing the relationships between the observable variables and the theoretical concepts to be specified. The second assesses the structural model and evaluates the consistency of the relationship proposed with the theory utilized (Henseler, Ringle, and Sinkovics 2009).
The general analysis model in this study establishes eight hypotheses to be tested. The dependent variable is labor productivity (LABPROD) in Catalan small and medium-sized travel agencies, approximated by the logarithm of turnover divided by the number of full-time equivalent workers. The numerator of this ratio was obtained from direct data on firm turnover. The denominator was constructed by taking into account the full-time and part-time jobs in a firm and expressing the number of workers as full-time equivalents.
The direct explanatory factors of labor productivity in the sample of firms are considered to be the logarithm of assets per full-time equivalent worker (ASSETS), the logarithm of the percentage of workers using local networks for their job (LOCNET), and a firm’s capacity to export to international markets (EXPORTS), that is, the logarithm of sales to the European Union and to the rest of the world. Thus, three of the hypotheses identified in international empirical evidence would be valid in small and medium-sized travel agencies in Catalonia: first, the higher the assets per worker, the higher the productivity (hypothesis 1); second, the higher the local network use, the higher the productivity (hypothesis 2); and third, the higher the export intensity, measured as a firm’s capacity to export goods and services, the higher the productivity (hypothesis 3). Hypothesis 1 is related to a firm’s capacity to increase turnover per worker by being bigger and probably better financed. Hypothesis 2 is related to a firm’s capacity to increase productivity through ICT use, especially through local network use. Hypothesis 3 is related to a firm’s capacity to increase turnover per worker through economies of learning, scale, reach, and scope, which can be achieved by growing export intensity.
After establishing the hypotheses related to direct factors of productivity, the analysis model also established a set of hypotheses related to indirect factors and their interrelationships. Specifically, an indirect causal relationship was established between a firm’s export capacity and local network use. Hypothesis 4 argues that a firm’s export capacity is explained by local network use. Similarly, hypothesis 5 argues that a small and medium-sized travel agency’s export capacity also depends on its capacity to generate market-leading product innovations (LEADINNOV). The LEADINNOV latent variable was estimated from two original variables. First, the INNOV variable shows a firm’s innovatory dynamics and takes two values: 0, when a firm has not implemented any innovation in the last two years; and 1, when a firm has implemented some type of innovation in the last two years. Second, the LIDNEWPR variable shows a firm’s leadership capacity in launching new products or services on the market. This variable takes two values: 0, when a firm has not implemented any product innovation that leads the market; and 1, when a firm has implemented some product innovation that leads the market. Thus, this latent variable reflects firm owners’ or directors’ perceptions of a firm’s capacity to innovate and, more specifically, to generate market-leading product or service innovations.
Hypothesis 6 establishes a causal relationship between innovation leadership and human capital and training in a firm. Its capacity to generate market-leading product innovations explains its greater stock of human capital. The human capital and training (HCT) latent variable was estimated from two variables. First, the human capital (HC) variable is formed by the workers’ level of completed studies (primary, secondary, and university education). And second, the training (TRAIN) variable is formed by the actual percentage of workers on training programs. Thus, the human capital and training construct reflects a firm’s educational stock and training of its workers. Hypothesis 7 argues that local network use by a firm’s workers is explained by its human capital and training stock. The higher the level of human capital and training, the more workers use a firm’s local networks.
Finally, hypothesis 8 establishes a reverse causality between a firm’s co-innovation and assets. A greater presence of co-innovation in a firm explains fewer assets and a smaller size. In other words, small and medium-sized travel agencies with fewer assets tend to encourage more use of co-innovation, although this leads to less productivity in the short term. By contrast, larger travel agencies with more assets tend to make less use of co-innovation and, as a result, obtain higher levels of productivity. In short, this hypothesis highlights the fact that more use of co-innovation in small and medium-sized travel agencies only results in labor productivity gains in the medium to long term because of the costs associated with learning and implementation. The COINNOVATION latent variable was estimated from two variables. The work team (WORKTEAM) variable indicates the presence or absence of work teams in a firm (0 = absent; 1 = present). The Internet use (INTERUSE) indicator indicates the intensity of Internet use in a firm’s value process. It takes four values: 1 = very low use, where a firm is not connected to the Internet; 2 = low use, where a firm is connected to the Internet but does not have its own website; 3 = normal use, where a firm is connected to the Internet and has its own website; and 4 = advanced use, where a firm is connected to the Internet and has its own website and e-commerce practices. The co-innovation construct reflects the establishment of relationships of complementarity between a variable that considers new forms of work organization (work teams) and an ICT variable (Internet use). Figure 1 shows the model and the hypotheses postulated in this study.
Modeling the interaction between sources of labor productivity in small and medium-sized travel agencies.
PLS-SEM Estimation Results
The PLS-SEM algorithm was used to estimate the model and the proposed hypotheses. This technique was chosen over others for several reasons (Henseler and Chin 2010). First, it is a highly evolved validated prediction modeling technique in which data multinormality can be relatively relaxed (Henseler, Ringle, and Sinkovics 2009). Second, it allows causal relationships between latent dimensions and measurement variables to be determined. Third, in addition to estimating causal relationships between latent variables, PLS-SEM allows formative latent variables to be calculated by setting the weights of each explanatory (and observable) variable. Estimating models of this type naturally requires appropriate measurements of goodness of fit and robustness of analysis to validate the proposed model.
We shall therefore present the validation of the proposed structural and measurement models before proceeding to report on the results. Regarding the assessment of the measurement model, several tests were conducted while taking account of the fact that the latent variables included in the model were formative constructs. Thus, on the one hand, the analysis of content validity allowed us to check whether the indicators had captured the full scope of the model (Diamantopoulos and Winklhofer 2001). The analysis confirmed that the indicators had been appropriately selected (Straub, Boudreau, and Gefen 2004).
On the other hand, the analysis of construct reliability allowed us to check the internal consistency of the measurement model. The validity and reliability of each indicator and construct was assessed and no multicollinearity problems were found (Table 2). The variance inflation factor was lower than the threshold value of 3.3 (Diamantopoulos and Siguaw 2006). All of the indicators were significant (p < 0.10), as detailed below in the section on the results of estimation using PLS-SEM.
Latent Variable Correlations.
Note: (1) COINNOVATION; (2) Local Network (LOCNET); (3) EXPORTS; (4) ASSETS; (5) Leader innovation (LEADINNOV); (6) Human capital and training (HCT); (7) Labor productivity (LABPROD).
Source: Own elaboration.
Two further validation measurements were made: construct validity and discriminant validity. For construct validity, the correlations between the constructs were lower than 0.5. Pairwise, they were therefore sufficiently different. For discriminant validity, Table 3 shows the cross-loadings obtained from the correlations between each item and each latent variable. The coefficient between each indicator and respective latent was high in all cases.
Cross-loadings.
Note: (1) COINNOVATION; (2) Local Network (LOCNET); (3) EXPORTS; (4) ASSETS; (5) Leader innovation (LEADINNOV); (6) Human capital and training (HCT); (7) Labor productivity (LABPROD).
Source: Own elaboration.
After analyzing the measurement model validation criteria, we dealt with the structural model validation criteria (Table 4). There were two criteria: model validity and predictive power. For model validity, the estimation was performed using 200 bootstrap resamples. Thus, the R2 result obtained for the labor productivity (LABPROD) variable was 0.285. While the result was relatively moderate, it did not undermine the validity of the estimations as a whole, because the causal effects obtained were in accordance with expectations. Moreover, all the variables showed the expected signs, and t statistics were sufficiently high. Indeed, the model’s predictive power was robust because the coefficient obtained for the Q2-statistic communality values were greater than zero for the variables analyzed. The Q2-statistic cross-validated redundancy values were also greater than zero, thus supporting the predictive relevance of the whole model. The Q2-statistic was evaluated by applying the blindfolding procedure with an omission distance of 7. The proposed threshold value was Q2 > 0; a higher Q2 value would mean a higher predictive relevance of the model.
Quality Model Measurements Overview.
Source: Own elaboration.
After analyzing the validation of the estimated model, we addressed the factors explaining labor productivity in Catalan small and medium-sized travel agencies. First, we analyzed the variables that made up the formative constructs and, second, the relationships between the constructs themselves. With respect to the constructs (Table 5), four of them were formed by a single variable and were therefore assigned an estimated weight of 1. The results of latent construct estimation consisting of two variables are also shown. All coefficients except for the TRAIN variable are significant (p < 0.10).
Outer Weights (Mean, Standard Deviation, and t Values).
Source: Own elaboration.
The results of latent variable estimation (path analysis) explaining labor productivity in Catalan small and medium-sized travel agencies are presented in Table 6 and Figure 2. First, it is important to point out that all the coefficients obtained were significant at a maximum 90% confidence level (p < 0.10), and that their values were consistent with the hypotheses postulated. Second, it should be noted that the main direct determinants of labor productivity are a firm’s assets (hypothesis 1: β = 0.340), the workers’ local network use (hypothesis 2: β = 0.231), and a firm’s capacity to export to international markets (hypothesis 3: β = 0.228). In this respect, the results suggest that labor productivity in small and medium-sized travel agencies in Catalonia is directly explained by their capacity to exploit their assets and to make international transactions. Exploiting assets and attracting customers from international markets are complemented by a particular use of ICT: the worker’s local network use in a firm.
Path Coefficients (Mean, Standard Deviation, and t Values).
Source: Own elaboration.
Sources of labor productivity in small and medium-sized travel agencies (path analysis).
After establishing that export intensity and local networks had direct effects on labor productivity in small and medium-sized travel agencies, a set of indirect effects was validated. First, it should be noted that a small and medium-sized travel agency’s export capacity was explained by the workers’ local network use (hypothesis 4: β = 0.157) and by a firm’s capacity to generate market-leading product innovations (hypothesis 5: β = 0.215). Second, the workers’ local network use had a dual basis. A firm’s capacity to generate market-leading product innovations explains its stock of human capital and training (hypothesis 6: β = 0.216), which in turn determine local network use (hypothesis 7: β = 0.839). In summary, the analysis of the indirect effects of labor productivity in small and medium-sized travel agencies suggests a circular causality, which is determined by the influence of the workers’ local network use on a firm’s export capacity. In this circular relationship, a small and medium-sized travel agency’s capacity to generate market-leading product innovations and its stock of human capital and training play a decisive role.
Finally, the analysis of indirect effects confirmed the inverse causal relationship between a firm’s co-innovation and assets (hypothesis 8: β = −0.259). This negative relationship confirms that in small and medium-sized travel agencies, co-innovation, as represented by Internet use and work teams, adversely affects asset exploitation. In this respect, the positive impact on labor productivity would only be reflected in the medium to long term, once the costs associated with learning and implementing co-innovation practices had been amortized.
As a result, the total effects on labor productivity in small and medium-sized travel agencies in Catalonia (Table 7) is explained by assets (β = 0.340), the workers’ local network use (β = 0.266), a firm’s capacity to export to international markets (β = 0.228), the stock of human capital and training (β = 0.223), a firm’s capacity to generate market-leading product innovations (β = 0.097) and co-innovation practices (β = −0.088). These results suggest that labor productivity in Catalan small and medium-sized travel agencies are due to a wide range of determinants that include elements of performance, such as their capacity to exploit their assets and to export, and elements that improve a firm’s internal value process, such as local network use, the capacity to generate market-leading product innovations or the stock of human capital and training. Only co-innovation, secured by Internet use and the presence of work teams, exerts an overall negative effect on labor productivity. This negative effect could be related to the difficulties that co-innovation practices have in terms of securing productivity improvements in the short term.
Total Effects (Mean, Standard Deviation, and t Values).
Source: Own elaboration.
Conclusion, Discussion, and Policy Implications
Using 2010 survey data for a sample of 120 small and medium-sized travel agencies based in Catalonia (a region in the northeast of Spain), this article analyzed the new sources of firm labor productivity based on the establishment of relationships of complementarity (co-innovation) between ICT investment and use, new forms of work organization and labor relations, and human capital and training.
The study’s focus on ICT and innovation had a dual basis. On the one hand, and as noted in the extensive literature, a firm’s ICT use can become a lever of change for introducing new value processes and new sources of productivity. In this respect, the purpose of the study was to test whether, in the field of small and medium-sized travel agencies, ICT use (measured by local network use) and the generation of co-innovation (such as the creation of new products and services, improving human capital, or relationships of complementarity between Internet use and work teams) explain firm productivity. On the other hand, and as highlighted by the latest research, relationships between ICT, innovation, and productivity in SMEs may also be generated indirectly, that is, through the impact on a firm’s other productivity-related results, such as its export capacity or its assets.
In this respect, and based on the idea that ICT and co-innovation are a lever, the aim of this study was to design and test a model explaining the direct and indirect sources
of productivity in small and medium-sized travel agencies. For that purpose, a PLS-SEM model was developed and tested. PLS-SEM allows for the analysis of relationships not only between the various factors considered explanatory of productivity, but also between such factors. Thus, in the case in hand, the analysis completed the structural form explaining the productivity of small and medium-sized travel agencies (94.2% had fewer than 10 workers), producing mostly for internal markets (79.5% of turnover in the Catalan and Spanish markets).
The results obtained confirmed the suitability of the proposed model. In explaining the productivity of small and medium-sized travel agencies, ICT and innovation simultaneously exerted direct and indirect effects. Regarding the path analysis of direct effects, the results suggested that labor productivity in small and medium-sized travel agencies in Catalonia was directly explained by their capacity to exploit their assets and to make international transactions. Exploiting assets and attracting customers from international markets were complemented by a particular use of ICT: the worker’s local network use in a firm. The path analysis of indirect effects on labor productivity suggested a circular causality, which was determined by the influence of the workers’ local network use on a firm’s export capacity. In this circular relationship, a small and medium-sized travel agency’s capacity to generate market-leading product innovations and its stock of human capital and training played a decisive role. Regarding the analysis of total effects, the results suggested that labor productivity in Catalan small and medium-sized travel agencies was due to a wide range of determinants that included elements of performance, such as their capacity to exploit their assets and to export, and elements that improved a firm’s internal value process, such as local network use, the capacity to generate market-leading product innovations, or the stock of human capital and training. Only co-innovation, captured by Internet use and the presence of work teams, exerted an overall negative effect on labor productivity. This negative effect could be related to the difficulties that co-innovation practices have in terms of securing productivity improvements in the short term. In summary, these results add to the available evidence, going beyond the traditional analysis of the relationships between ICT, innovation, and productivity and providing a more comprehensive and realistic view of the productivity path of small and medium-sized travel agencies.
Furthermore, the results obtained suggest that new directions in public policy are required to improve productivity in small and medium-sized travel agencies. First, it is important to emphasize the need to coordinate efforts in the joint promotion of ICT use, organizational change, and training among employers and employees. For example, partial public policies to promote ICT use, without considering other determinants of co-innovative productivity, may not produce the desired effects. Second, it is important to point out the link between productivity and internationalization. In SMEs producing mostly for internal markets, promoting the internationalization of their products and services is something that public policy should address because international competition introduces efficiency enhancement mechanisms. And third, the study results also suggest the need to consider the costs associated with learning and implementing co-innovation practices. To overcome the negative effects on asset exploitation and labor productivity that co-innovation practices have in the short term, public policy should promote sustainable policies to promote change in firms over time.
The study presented in this article has several limitations. Besides the variables and restrictions imposed on the analysis, perhaps the most significant is the unavailability of a time series. However, the availability of survey data for a sample of small and medium-sized travel agencies has provided an excellent opportunity to analyze the determinants of their growth potential. In this respect, and bearing in mind the economic importance of small and medium-sized travel agencies to the regional economy, the availability of data for (1) other territories or business groups, and their possible comparisons; (2) a time series; (3) better indicators; and (4) new criteria for grouping firms would suggest that new approaches could be taken. Such major lines of improvement give this study a preliminary character and suggest that further research needs to be conducted on this issue.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.