The effect of Economic Open Government Data on Bilateral Trade and the
mediating role of Knowledge Creation
Name: Renske Vogel Studentnumbe r: 11138696
Date: 27-01-2017
MSc. Business Administration – International Management Master Thesis
University of Amsterdam
Statement of Originality
This document is written by Student Renske Vogel who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
Abstract
The ability to create a competitive advantage through big data and the increasing use of data in decision-making has led to a rising interest in the matter by both practitioners and scholars (Chen, Chiang & Storey, 2012). Also governments show an increasing acknowledgement of the social and economic value of open data, causing a global movement towards open government data (Bauer & Kaltenböck, 2011; Ubaldi, 2013). In order to make a first step in the process of establishing the role of country level big data in international trade, this research looks at the influence of economic open government data on bilateral trade and the mediating role of knowledge creation in this relationship. This paper uses time-series, cross-country data on the US dollar export value of goods between 30 OECD countries, applying a gravity model. Results show that economic open government data in the importing country positively affect exports towards them. Economic open government data in the exporting country do not influence export levels, however separate analysis on the availability and quality of the datasets in the exporting country shows a negative effect on exports. The ability of businesses and entrepreneurs to gain economic value from open government data has a positive influence on bilateral trade. Furthermore, the availability and quality of datasets in the importing country has a positive influence on exports towards them, which is stronger if firms in the exporting country have a higher ability to gain economic value from open data than firms in the importing country.
Table of Contents
1. INTRODUCTION ... 5
2. LITERATURE REVIEW ... 7
2.1 Big data and analytics ... 7
2.2 Open government data ... 10
2.2.1 Innovation Implementation ... 12
2.3 Knowledge creation ... 13
2.3.1 Business Readiness ... 15
2.3.2 Predictive Analytics ... 18
2.4 Knowledge and export levels ... 19
2.5 Open Government Data and export levels ... 22
2.5.1 Economic Impact ... 22
2.5.2 Economic Open Government Data ... 23
3. CONCEPTUAL FRAMEWORK ... 25
3.1 Influence of open government data on export ... 25
3.2 The mediating role of knowledge ... 30
4. METHODOLOGY ... 33
4.1 Dataset ... 33
4.2 Measures and datacollection... 35
4.1 Data analysis ... 40
4.1 Strengths and limitations of the research method ... 43
5. RESULTS ... 45
5.1 Analytical strategy ... 45
5.2 Hypotheses testing: Business readiness, innovation implementation, economic impact ... 47
5.2.1 Correlations ... 47
5.2.2 Multiple regression analysis ... 47
5.2.3 Moderation analysis ... 51
5.2.3.1 Moderation innovation implementation reporter ... 51
5.2.3.1 Moderation innovation implementation partner ... 53
5.3 Hypotheses testing: Economic open government data ... 55
5.3.1 Correlations ... 55
5.3.2 Multiple regression analysis ... 55
5.3.3 Mediation knowledge creation ... 58
5.3.3.1 Mediation economic open government data reporter ... 58
5.3.3.1 Mediation economic open government data partner ... 59
6. DISCUSSION ... 61
6.1 Discussion business readiness, innovation implementation, economic impact ... 61
6.2 Discussion economic open government data ... 63
7. CONCLUSION ... 66
7.1 Scientific relevance and managerial implications ... 68
7.2 Limitations of research ... 70
7.3 Suggestions for future research ... 70
8. REFERENCES ... 72
Index of Tables and Figures
Table 1. Measurements Open Government Data ... 37
Table 2. Descriptive statistics and correlations 1 ... 48
Table 3. Hierarchical Regression Model of Bilateral Trade 1... 50
Table 4. Analysis of moderation effect Business Readiness Difference ... 52
Table 5. Interaction Variables ... 52
Table 6. Analysis of moderation effect Business Readiness Difference 2 ... 54
Table 7. Interaction Variables 2 ... 54
Table 8. Descriptive statistics and correlations 1 ... 56
Table 9. Hierarchical Regression Model of Bilateral Trade 2... 57
Table 10. Mediation Knowledge Creation OGD partner ... 60
Table 11. Effect Models ... 60
Table 12a. Results on the hypotheses for Business Readiness, Innovation Implementation and Economic Impact ... 62
Table 12b. Results on the hypotheses for Economic Open Government Data ... 65
Figure 1. Conceptual framework split measurement Open Government Data ... 26
1. Introduction
Of all data ever generated, ninety percent has been generated over the last two years (Buchholtz, Bukowski, & Śniegocki, 2016). The increasing amounts and variety of data available has led to an ongoing need for new technologies allowing for big data analytics (Provost & Fawcett, 2013). Firms are more and more starting to acknowledge the opportunities that arise from the use of these large and diverse data sets in decision-making (Chen, Chiang & Storey, 2012). Using big data analytics can provide organizations with high value information on consumer behavior and allows for quick responses to market developments (European Union, 2013). Following, we can identify a shift from intuitive decision-making towards data-driven decision-making (European Union, 2013; Provost & Fawcett, 2013; Ubaldi, 2013; Wu, Li, Lin & Goh, 2015).
An important driver towards data-driven decision-making is the level of open data, more specifically a country’s open government data and its re-use by private bodies (Avital, Bjørn-Andersen & Jetzek, 2013; Ubaldi, 2013). This is because governments are one of the most considerable producers and providers of open data(Charalabidis, Janssen, & Zuiderwijk, 2012).Over the last years, open government data has gained increasing attention. In 2011 the first global initiative for open government data and new technology development was launched, the Open Government Partnership currently counting 75 participating countries (Janssen, 2012; Open Government Partnership, 2016).
The ability for third parties to use open government data allows for the release of the economic value of information, as often argued by for example public bodies (Avital et al., 2013; Ubaldi, 2013). However, very limited studies have been done to actually establish how open government data can generate value (Avital et al., 2013). Therefore the growing influence of open government data on data driven decision-making and economic growth is
still in need of academic research to gain a more thorough understanding of its influence on firm level decision-making.
Using data in decision-making enables firms to more accurately monitor their current business operations and identify new fields for market expansion and revenue creation (Wu et al., 2015). This is thought to lead to increased competitiveness on a global level and a more rapid internationalization of firms (European Union, 2013). An interesting component of this internationalization, however not yet investigated, is the influence of data-driven decision-making on the export levels of firms. Scholars have pointed to the influence of firm level use of open government data on their product development and efficiency gains (World Wide Web Foundation, 2013). Next to that, using open government datasets from foreign countries allows firms to lower the risk on foreign business activities as well as recognize market opportunities outside their national borders (Becerra-Fernandez, Zanakis & Walczak, 2002).
Following the often shown influence of knowledge creation on bilateral trade, the knowledge creation through open government data may influence country level bilateral trade levels trough their aggregated firm behavior (Autio, Sapienza, & Almeida, 2000; Casillas, Moreno, Acedo, Gallego, & Ramos, 2009; Petersen, Welch, & Liesch, 2002). However, current literature has not yet linked firm level use of open government data to knowledge creation and bilateral trade levels.
In this research a country level analysis of open government data on bilateral trade in merchandise will be done. This country level analysis includes the aggregated firm level export behavior as a response to open government datasets. The economic component of open government data will be examined to set out a first step toward establishing its effect on bilateral trade and the mediating role of knowledge creation in this relation.
2. Literature review
In order to justify the expected relationship between open government data and bilateral trade, the literature review will outline all components relative to this research. First the growth of data-driven management and the role of big data and value creation will be touched upon. Next, the literature review will specify on open government data, in order to clarify this concept and explain the focus of this research on economic open government data. This section will explain the first of three components of economic open government data, innovation implementation and its role in value creation. Following, open government data will be linked to bilateral trade, firm-specific knowledge resources, and knowledge creation on foreign markets. Here the second component of economic open government data: business readiness will be explained. The link between open government data and knowledge creation will be further specified though a review on predictive analytics. Strengthening the expected effect of economic open government data on international trade, the previously found role of knowledge and the Internet in bilateral trade will be elaborated upon. In the final section, economic impact, the third component of economic open government data will 2be explained and the possibility that there is no relationship between open government data and bilateral trade will be touched upon.
2.1 Big Data and Analytics
Big data and analytics have been identified as one of four current leading technological trends (Chen et al., 2012). Big data and big data analytics are terms used to describe the current evolvement of datasets and techniques for analysis of data too extensive and complex to be analyzed by humans. The numerous data available can no longer be processed by regular comparative databases (Khan, Yaqoob, Hashem, Inayat, Mahmoud Ali, Alam, Shiraz & Gani, 2014).
Next to the Internet, the world wide use of mobile devices has continuously increased the amounts of generated data (Khan et al., 2014). Furthermore, sensor-generated data are tremendously contributing to the collection of data (Chen et al., 2012, Khan et al., 2014). Following, state-of-the-art technologies are continuously being developed, to enable interpretation and possible value derivation from these large quantities of data (Chen et al., 2012). The ability to use these large datasets through newly developed technologies is, amongst other things, not only changing business processes but influencing management practices as well (Gerbert, Justus & Müller, 2016). Organizations have realized that this data can be used to seek opportunities and create a competitive asset within their firm. Therefore, big data analytics are being adopted in management practices, leading to so-called data-driven management (Chen et al., 2012; Gerbert et al., 2016; Provost & Fawcett, 2013).
The OECD (2014) points to the rise of data-driven innovation, where big data is used to promote new markets, products and processes and to create knowledge and value. Internet firms are the forerunners in the use of big data (OECD, 2014). The rest of the ICT sector has also recognized its new growth opportunities, bringing on large investments (OECD, 2014). Even though the ICT sector remains leading in this department, big data and analytics are currently influencing business along a wide range of sectors (OECD, 2014). Big data and analytics are starting to gain significant interest by, for example, the retail, manufacturing and health care sectors. In the health care sector the analysis of big data can, for example, help create medical support systems and improve management (Wu et al., 2015). Furthermore, being able to systematically analyze data enables practitioners to improve their operating efficiencies and lower the costs of health care (Wu et al., 2015).
Part of big data analytics is a process often referred to as data mining (Provost & Fawcett, 2013). This entails the process of gaining actual knowledge through the use of technologies that allow for information derivation (Provost & Fawcett, 2013). Managers, in their
decision-making processes, can in turn use the knowledge gained from big data analytics. The value of big data is released during the different stages of its use, namely, when the data is transformed into knowledge and when it is used in decision-making (OECD, 2014). In 2012, big data was already used by sixty percent of business leaders as a support for decision-making (OECD, 2014). Looking at the manufacturing sector, there is an increase in the adoption of data-driven decision-making of thirty percent between 2005 and 2010 (Brynjolfsson & McElheran, 2016). Nevertheless, research has pointed out that many organizations still do not include data-driven decision-making in their managerial practices, possibly due to cost barriers or lack of awareness (Brynjolfsson & McElheran, 2016).
Managers that adopt a data-driven decision strategy do not base their decisions solely on personal instinct, perception and insights, but make use of the analysis of data (Provost & Fawcett, 2013). Managerial strategies can be data-driven up to different levels, allowing for personal intuition as well as data-driven decision-making (Provost & Fawcett, 2013). Research has shown the advantages of data-driven decision-making and its positive influence on firm performance, as well as on productivity and a firm’s market value (Hitt & Kim, 2011; Provost & Fawcett, 2013; Brynjolfsson). The growing use of big data analytics and the increase in data-driven decision-making within organizations allows for a wide range of future research on the outcomes of big data analytics and the relating changes in management practices.
Following the growing use of big data in decision-making, it seems interesting to see how these types of data are used in managerial decisions and shape the current business fields. This research will look at a specific example of big data: open government data. Previous research has pointed to the ability of firms and start-ups to create new products, services and generate efficiencies trough the use of open government data (Ubaldi, 2013; World Wide Web Foundation, 2013). Open government data allows for the detraction of economic value
from information, which is now available to private parties (Ubaldi, 2013). Avital et al. (2013) found that open government data creates social and economic value through participation, transparency, efficiency and innovation. Literature has often mentioned the role of knowledge and innovation in international trade. Therefore, there might be an influence of open government data on knowledge creation and bilateral trade.
The following sections will look into the role of open government data on both new idea generation, as well as knowledge creation on foreign market risks. Furthermore, the role of knowledge in bilateral trade and subsequently the possible influence of open government data on bilateral trade will be touched upon.
2.2 Open Government Data
As previously explained, big data refers to the development of datasets and techniques for data analysis (Khan et al., 2014). An important requirement to release the economic value of these datasets is for these data to be open. Open data need to be available for use, re-use and sharing (Molloy, 2011). Open data consist of datasets as made available by governments but also by private bodies within for example the entrepreneurial, scientific or educational field (Bauer & Kaltenböck, 2011). Part of open data is open government data. Open government data consist of government data: “any data and information produced or commissioned by public bodies” and open data: “data that can be freely used, re-used and distributed by anyone, only subject to (at the most) the requirement that users attribute the data and that they make their work available to be shared as well” (Ubaldi, 2013, p.6).
Along with the growth in data generation and its usage by private sector firms, governments are also more and more acknowledging the social and economic value of open data (Ubaldi, 2013). Open government data thus refers to the global movement towards opening up data and information either generated or made available by the government
(Bauer & Kaltenböck, 2011). In a wide variety of subjects, public bodies are one of the biggest producers and providers of open data (Charalabidis et al., 2012). Examples of open government data are datasets on tourist information, traffic and budgets within the public sector (Charalabidis et al., 2012). Data derived from the public sector can be especially useful due to its remarkable quantity and centrality (Ubaldi, 2013). Governments are increasingly contributing to the large quantities of available datasets (Charalabidis, et al., 2012; Ubaldi, 2013). These datasets together with the growing evolvement of technologies allow for the public and private sector to tap into the value of open data, leading towards more data-driven decision-making along a wide range of sectors (OECD, 2014; Ubaldi, n.d.).
The influence of open government data can be seen in its political, social and economic impact (World Wide Web Foundation, 2015). Research has pointed out that the influence of these types of data has been most prominent in the economic field. Especially, one can see a growing influence on entrepreneurial activity (World Wide Web Foundation, 2015). Therefore, this research focuses on the economic aspect of open government data. Due to its influence on product development and efficiency generation and the often mentioned influence of knowledge creation in bilateral trade, it seems highly interesting to focus on the influence of economic open government data on export levels (World Wide Web Foundation, 2013; Ubaldi, 2013). Open government data is seen as a positive force in the creation of transparency of knowledge and value creation (Charalabidis et al., 2012). Establish the influence of open government data on bilateral trade would generate some valuable outcomes on the influence of open data on the international business environment and the importance of governments to actively participate in the development towards a data-driven economy.
Setting out a first measurement on the influence of big data on bilateral trade, this research will look at open government data from a firm perspective. The effect of open government data on bilateral trade will be investigated by examining how the aggregated firm use of these
datasets influences export levels. Thus, the aggregated firm behavior on the use of open government datasets and their export levels will be examined using country level data. In order to account for the different factors that influence the value and outcomes of the use of open data, this study will first look at the role of economic open government data in bilateral trade by splitting up this measurement in three components. This research will look at innovation implementation, business readiness and economic impact of open government data. Second, an economic open government data measurement will be created, including all three components, in order to examine the overall influence on bilateral trade.
2.2.1 Innovation Implementation
Innovation implementation is a measurement on the level of datasets as made available by the government, including a number of different components. Innovation implementation specifically refers to the quality of datasets, relevant to the entrepreneurial and business fields of a country (World Wide Web Foundation, 2014). The quality level of these datasets is based not only on their availability, but includes in total ten components to allow for an actual rating on their quality and usability (World Wide Web Foundation, 2014). The components included in this study measure if the data exist, are online available, machine-readable, bulk, free of charge, openly licensed, up to date, sustainable, easy to find and linked (World Wide Web Foundation, 2014).
Datasets need to be available free of charge and online too (Ubaldi, 2013). Furthermore, the ability to re-use data is often mentioned as an important requirement to unleash the economic value of open government data (Avital et al., 2013). This because the ability to re-use data allows for the combination of different datasets, whereby knowledge and new understandings can be obtained (Ubaldi, 2013). Other important components in establishing the value of open government data are, the frequent and timely updating of the available data
and the incorporation of feedback in the data systems (Charalabidis et al., 2012; Ubaldi, 2013). The open access to public data allows governments to gain feedback from its users. In trying to improve the quality level of their datasets, governments should therefore actively seek and use feedback (Charalabidis et al., 2012).
In order to account for the accessibility and quality of datasets on a firm level, this study will use country level data on the evaluation of the datasets as available to entrepreneurs and businesses in that country. This study will take into account both the level of innovation implementation in the exporting country as well as that of the importing country in order to model their influence on bilateral trade. See the methodology section of this research for a more thorough understanding on the measurements for innovation implementation.
2.3 Knowledge creation
With business practices moving further into the process of data-driven decision-making and the increased understanding of its limitations and possibilities, the ability to gain knowledge through big data analytics grows. Being able to gain knowledge from data allows firms to identify and understand patterns (Becerra-Fernandez et al., 2002). These patterns can, for example, allow a company to identify the chance that a consumer will buy a certain product or enable a company to establish the influence of inventory levels on sales (Becerra-Fernandez et al., 2002). In order to create knowledge through open data, not only the availability and quality of these datasets are of importance. Also the ability of entrepreneurs and businesses to gain value from these datasets should be accounted for. “The value of data depends on the meaning as extracted or interpreted by the receiver” (OECD, 2014, p.26). Therefore, also the value of open government data depends on the readiness of businesses and entrepreneurs in the country to use these data (OECD, 2014).
Many governments and organizations have recognized the value adding ability of big data and analytics. Big data analytics can be used along a wide range of business operations in order to improve a firm’s performance (Wu et al., 2015). Big data can be used to improve customer relationships and enable firms to more accurately monitor their current business operations or identify new fields for market expansion and revenue creation (Wu et al., 2015). Even though not yet established, it might be that these new fields for market expansion and revenue creation are identified overseas, as will be investigated within this research. Well organizing data and being able to integrate internal company databases with external databases allows for great opportunities of knowledge discovery (Hendler, 2014; Peng, Kou, Shi & Chen, 2008).
Within the field of data mining and knowledge discovery, as well as the more specific frameworks for big data analytics, a number of stages in data collection and data usage have been identified. Examples of these stages are data selection, transformation and interpretation (Peng et al., 2008). These frameworks set out a process through which data becomes knowledge. Mandinach, Honey, and Light (2006) include three stages towards knowledge creation in their conceptual model for data-driven decision-making. First, companies have data available, which in itself does not have any meaning yet. This data is transformed into information by connecting it to the environment. However, this information does not give any direction to future behavior (Mandinach et al., 2006). What follows is the process through which information becomes knowledge. Here, one is able to capture and connect the useful components of information in order to shape future behavior (Mandinach et al., 2006).
However, when looking at the role of these new gained insights in management practices, one should be cautious. Multiple studies have attempted to create a theoretical framework for the field of big data analytics, data mining or data-driven decision-making (Khan et al., 2014; Mandinach et al., 2006; Peng et al., 2008). These scholars point to the
ongoing challenges in data-driven decision-making. Big data analytics can create a wrong image on which data-driven manager based their decision. This is due to factors such as data inconsistency and incompleteness (Khan et al., 2014). Next to that, the ability to correctly analyze semi-structured or unstructured data still remains complicated (Khan et al., 2014).
The increasing levels of data availability can increase the difficulty for practitioners to access, or integrate them as well as control their quality (OECD, 2015). Even when using accurate data, it can lead to incorrect output if the data is accessed too late or is misinterpreted (OECD, 2015). As stated by the OECD (2015, p. 193-194): “even though techniques for record linkage are now well developed, and are used by numerous organizations regularly, the capacity with which to carry out successful linkages may be in short supply”. Therefore, scholars have pointed to the need for data scientists who are able to interpret a firm’s challenges from a data perspective (Cavanillas, Curry & Wahlster, 2016; Provost & Fawcett, 2013; Skudiene, Auruskeviciene & Sukeviciute, 2015). Thus research relating to big data analytics should take into account that the knowledge derived from big data analytics might not always be correct.
2.3.1 Business Readiness
In order to account for the ability to gain valuable insights from open data, this study will look at the role of business readiness in bilateral trade levels. Here business readiness accounts for the ability of the private sector to use and obtain benefits from open data (World Wide Web Foundation, 2015). Factors influencing a country’s business readiness are the level of training directed towards the use of open data and the governmental support directed towards innovation in open data technologies (World Wide Web Foundation, 2015). Furthermore, firm-level technology incorporation and Internet penetration are contributing to the level of business readiness (World Wide Web Foundation, 2013).
Being able to critically and correctly assess and use the outcomes from big data allows for extensive knowledge discovery and the ability to identify ne w opportunities and markets (Hendler, 2014; Peng et al., 2008; Wu et al., 2015). An often-used term in literature relating to data-driven decision-making is business intelligence (Chen et al., 2012). This term has evolved into big data analytics, referring to the growing ability to extract information and create a new way of firm-customer interaction (Chen et al., 2012). For example, social media analytics allows for a comprehensive collection of data on customer assessments and behavioral patterns (Chen et al., 2012). Big data analytics thereby allows practitioners to see, hear and interpret new information and can thus derivate knowledge from their consumers as well as all shareholders, such as suppliers and employees (Chen et al., 2012).
Now looking at the influence of business readiness on knowledge creation, one can distinguish between the ability to create knowledge within the exporting country and to gain knowledge on foreign market opportunities. First, looking at the levels of knowledge creation within the exporting country, business readiness together with innovation implementation and economic impact affect the level of new idea and thus new product generation and efficiency gains within a country (Ubaldi, 2013; World Wide Web Foundation, 2013). Another way of knowledge creation through open data is the ability of firms to base for example their product development or market expansion decisions on current market evolvements, possibly leading to first- or early-mover advantages (OECD, 2014).
The quality of government datasets and the ability of firms to gain knowledge from these datasets, thus seem to lead to firm-specific knowledge resources. Resource- and knowledge-based literature has identified intangible resources as the most important source of competitiveness and long-lasting performance superiority (Rodríguez & Rodríguez, 2005; Wang, He & Mahoney, 2009). Firm-specific knowledge is a type of intangible resource and an important facilitator of a competitive advantage (Wang et al., 2009). These firm-specific
knowledge resources are dependent on information inputs as well as the ability of employees to make use of the information inputs in the development of innovations (Wang et al., 2009). Thus human capital determines what kind of firm-specific knowledge is created from the data input, for example government datasets. It is however important to acknowledge that the knowledge resources need to remain firm specific by for example, being inimitable and non-substitutable, in order to generally create a comparative advantage (Rodríguez & Rodríguez, 2005; Wang et al., 2009).
Relating the concept of firm-specific knowledge resources to this study, the information inputs as provided by governments and the readiness of entreprene urs and business determine the creation of firm-specific knowledge resources. Setting out the first measurement on the possible creation of firm-specific knowledge resources through open government data, this study will include country level data on knowledge creation. Knowledge creation in this study is captured by, amongst others, the number of patent and utility model applications in a country (Cornell University, INSEAD & WIPO, 2016). The country level measurement on knowledge creation thus measures the accumulated number of patent and utility model applications by firms and entrepreneurs within a country.
Next to the knowledge creation within the exporting country, one could also look at the creating of knowledge about foreign markets through the use of open government data. In case the datasets from foreign markets are available to entrepreneurs and businesses in the exporting country, it allows them to lower the risks of doing business abroad as well as recognize foreign market opportunities (Becerra-Fernandez et al., 2002). Previous literature has pointed to the negative influence of insufficient information o n foreign market opportunities on international trade (Rauch, 2001). Denis and Depelteau (1985) already identified market intelligence as one of the key conditions for the decision to export or expand export operations. Allen (2014), points to the prominent influence of so-called information
frictions in international trade. Information frictions refer to the costs related to gaining information on foreign market prices (Allen, 2014). Information frictions turn out to be equal or even more important than transaction costs in the decision to export (Allen, 2014). Therefore, it seems that the ability to quickly and correctly gain information, on for example foreign market prices, though open data will positively influence export levels.
As previously mentioned this study will look at the overall country level of business readiness in order to account for the average ability of firms within that country to detract value from open data (World Wide Web Foundation, 2015). This research will include both the business readiness of the exporting as well as the importing country to model its influence on bilateral trade. In order to more specifically measure the level of knowledge creation within the exporting country and its influence on bilateral trade, knowledge creation within the exporting country will be tested as a mediator in the relationship between economic open government data and export.
2.3.2 Predictive analytics
Predictive analytics is identified as one of the key enablers for improved decision-making through big data (Junqué de Fortuny, Martens, & Provost, 2013). Data-driven predictive modelling uses datasets to estimate the levels of a target variable (Junqué de Fortuny et al., 2013). Statistics that have often been used in predictive analytics include variables such as geographic characteristics, consumer characteristics and features relating to prior customer buying behavior (Junqué de Fortuny et al., 2013). The rise of big data allows firms to include more specified data on for example those consumer characteristics. Junqué de Fortuny et al (2013, p.217) refer to these new data as “sparse, fine grained (behavior) data ”. Modern technology leads to an increase in this type of data because it is able to specifically
track individual behavior and monitor consumer characteristics (Junqué de Fortuny et al., 2013).
As data become more specified and up to date, firms should increase their emphasis on data-driven decision-making (Brynjolfsson et al., 2011). By doing so they are able to improve the quality of their decision-making (Brynjolfsson et al., 2011). The research by Junqué de Fortuny et al (2013) found an increased enhancement of a company’s predictive analytics as its data availability grows. This can be either by growth in the number of individuals included in the data or in the amount of data components included (Junqué de Fortuny et al., 2013). This outcome suggests that if firms are able to increase the use of data including specific consumer information in their decision-making, they will be able to make more accurate assumption on for example future consumer behavior and promising foreign market developments. If firms are able to create firm specific knowledge resources through the use of predictive analytics, then open government data would be transformed into an intangible resource and a possible comparative advantage in international trade (Rodríguez & Rodríguez, 2005; Wang et al., 2009).
Thus predictive analytics could possibly lead to knowledge creation in the exporting country, by allowing firms to monitor the development of customer needs and use this information in the creation of new products (Junqué de Fortuny et al, 2013; World Wide Web Foundation, 2013). Furthermore, predictive analytics allows firms to identify promising foreign market opportunities, possibly increasing their likelihood to engage in international trade (Becerra-Fernandez et al., 2002).
2.4 Knowledge and export levels
The influence of knowledge on international trade levels, as previously established in the literature, enables for a better understanding of why open government data and the
growing use of data-driven decision-making might have an impact on bilateral trade (Provost & Fawcett, 2013). Previous research has repetitively pointed to the essential influence of knowledge on internationalization and international trade (Autio et al., 2000; Casillas et al., 2009; Petersen et al., 2002). Here knowledge can be split into experience-based knowledge and planning-based/objective knowledge (Casillas et al., 2009; Petersen et al., 2002). Mostly relevant to open government data and data-driven decision-making is objective knowledge.
Autio et al. (2009), explain the influence of knowledge levels (not obtained by a firm’s activity) on international trade, by stating that firms with a higher focus on knowledge creation are more likely to evolve their learning skills, enabling for better adjustment to international environments. Furthermore, they state that high knowledge-level firms can better capture international opportunities because of the mobility of knowledge as a resource.
Part of knowledge, as measured in various studies, is the level of information and communication technology and innovation (Demirkan, Goul, Kauffman & Weber, 2009; Wakelin, 1998). Extensive research has been done on the influence of innovation and technology on bilateral trade, all finding a significant positive influence on bilateral trade or bilateral trade performance (Demirkan et al., 2009; Eaton, & Kortum, 2002; Wakelin, 1998). This can be explained by relating these concepts to the Transaction Cost Theory (Demirkan et al., 2009). This theory states that when looking at the costs of an economic transaction, one should take into account all costs related to this exchange (Demirkan et al., 2009). ICT lower these costs through a reduction in search costs, management and control costs, shipping costs and time costs (Demirkan et al., 2009). The most interesting component of costs reduction regarding this research is the decrease in search costs. Higher levels of information availability through ICT lower the search costs for sellers and buyers on identifying the availability and price of the seller’s offer (Demirkan et al., 2009).
Furthermore, research by Becerra-Fernandez et al. (2002) has shown that knowledge discovery in databases can be of high value in establishing the risk of investing in a foreign country. Knowledge discovery in big databases can be a decision support system that allows for an online and quick determination of country investment risk (Becerra-Fernandez et al., 2002). Also, as the data input changes, firms can rapidly assess the influence on country investment risk and adjust their strategies accordingly (Becerra-Fernandez et al., 2002).
From a resource-based perspective, a firm’s decision to export and its export intensity is highly dependent on the company’s competitiveness (Rodríguez & Rodríguez, 2005). A firm’s competitive advantage is mostly determined by its intangible resources, more specifically, the previously mentioned firm-specific knowledge resources (Rodríguez & Rodríguez, 2005; Wang et al., 2009). Previous studies have shown the positive effect of intangible resources, such as patents and innovation, on the decision to export and export levels (DiPietro & Anoruo, 2006; Rodríguez & Rodríguez, 2005).
The positive influence of open government data on bilateral trade seems even more likely, looking at previous research on the effect of Internet on internationalization and international trade (Petersen et al., 2002; Skudiene et al., 2015). Here the ability to gain information quickly and at no or low costs is mostly identified as the key driver behind the influence of the Internet on export levels (Petersen et al., 2002; Skudiene et al., 2015). Relating this to the possible influence of open government data, it thus seems likely that the same effect will be observed. Literature has already pointed to the likelihood that market knowledge gained through big data will reduce costs and increase revenues, leading to higher internationalization of European firms (European Union, 2013). The actual effect of open data on international trade is yet to be established. This paper will therefore examine if open government data, similarly to ICT and Internet, have a positive effect on bilateral trade and if this relationship is (partially) driven by knowledge creation in the home country.
2.5 Open Government Data and export levels
As previously mentioned, this research will look at three parts of economic open government data and their influence on bilateral trade in goods. The analysis excludes bilateral trade in services, due to the limited existence of data on service trade (Baier & Bergstrand, 2002). Furthermore, the model will include a gravity equation, mostly applicable to international trade flows in merchandise (Baier & Bergstrand, 2002). First, the individual effect of every component will be examined. Furthermore, innovation implementation and business readiness will be linked to model their joint effect on export levels. Regarding the components innovation implementation and business readiness, there are some possible effects of open data. If two countries both have a high level of innovation implementation, this could enable firms to also tap into the value source of the other country, which might strengthen their willingness to export. However, since open government data will provide the same information to everyone, this does not have to lead to a competitive advantage. Especially not to those firms that already have access to these datasets (Ubaldi, 2013). Nevertheless, as previously established, information is not the same as knowledge and therefore it may very well be that organizations derive different knowledge from the same datasets. If firms would be able to create firm-specific knowledge resource though the use op open government data, this would provide them with a comparative advantage in international trade (Rodríguez & Rodríguez, 2005). Looking at the joint effect of innovation implementation and business readiness on bilateral trade allows for a specification on this effect.
2.5.1 Economic impact
Next to business readiness and innovation implementation, this research includes the economic impact of open government data. This refers to the measured overall impact of open
government data on a country’s economy and its entrepreneurial activity directed towards creating new business (World Wide Web Foundation, 2014). The general impulse of open government to the exporting country’s economy and their new venture creation may positively influence their export levels. Furthermore, the economic impact within the importing country will be taken into account. Global competition is likely to increase due to big data and data-driven decision-making (European Union, 2013; Freund & Weinhold, 2004). It could be that this increased competition will lower the profits from export, decreasing the effect of big data on bilateral trade (Freund & Weinhold, 2004). Freud and Weinhold (2004), state that as the number of firms active in export increases, the amount exported per firm decreases due to a decline in their export equilibrium.
2.5.2 Economic Open Government Data
Following the separate components of economic open government data, an overall measurement including innovation implementation, business readiness and economic impact, will be created. This measurement, including all relevant components of economic open government data will allow the monitoring of its overall effect on bilateral trade. Due to the previously described expected influence of open government data on knowledge creation and the established influence of knowledge levels on exports, this research expects a positive influence of economic open government data on bilateral trade levels.
However, it could be that the influence will not be as expected or not very strong. Previous literature has found that knowledge derived from actual experience is far more valuable and of higher influence on export levels (Denis & Depelteau, 1985). The research by Denis and Depelteau (1985), finds that an increase in information services will not lead to a growth in export activity or in export volumes. According to the process model of internationalization by Johansson and Vahlne the perception of market uncertainty is only
reduced by experimental knowledge (Hadley & Wilson, 2003). Experimental knowledge is market specific knowledge and provides firms with information o n for example consumer characteristics (Hadley & Wilson, 2003). Objective knowledge is knowledge that can be applied on various markets (Hadley & Wilson, 2003). This theory would be in line with Denis and Depelteau (1985), and suggests that the objective knowledge creation obtained through open government data will not affect country export levels. However, as open government data can also provide information on for example country specific consumer characteristics, it seems that these datasets might enable firms to gain knowledge that would normally have to be obtained through experience (Junqué de Fortuny et al., 2013; World Wide Web Foundation, 2013).
Next to that, it can be that the expected relationship between economic open government data and bilateral trade is not mediated by knowledge creation. As previously established knowledge creation will only create a sustainable competitive advantage if it leads to firm-specific knowledge resources (Wang et al., 2009). It might be that firms are not able to gain inimitable, non-substitutable knowledge resources from open government data, therefore not providing them with a comparative advantage in international trade.
3. Conceptual framework
Two conceptual frameworks will be used in order to answer the following research question: What is the effect of economic open government data on bilateral trade and is there a mediating role of knowledge creation?
3.1 Influence of Open Government Data on export
This research will investigate the economic effect of open government data on a country’s export levels. Open government not only captures data about governmental operations but also data that are made publicly available by the government (Ubaldi, 2013; World Wide Web Foundation, 2015). This means that it can also include crowd-sourced data (World Wide Web Foundation, 2015). Therefore this study will take into account open data from the government but also open data from other resources such as crowd sourced data, made available by the government (World Wide Web Foundation, 2015). As previously mentioned, open government data country levels consist of several components, respectively: readiness, implementation and impact (World Wide Web Foundation, 2015). Therefore this research will first look at the influence of business readiness, innovation implementation and economic impact of open government data on export. Within the framework and following hypotheses, reporter refers to the exporting country and partner to the importing country.
The hypotheses to be tested for this part of the research will be elaborated upon based on the following conceptual framework:
Figure 1: Conceptual framework split measurement Open Government Data
The readiness of open data refers to the readiness of open data users to create change, more specifically a valuable outcome, through their technical skills, means and freedom in using these datasets (World Wide Web Foundation, 2013; World Wide Web Foundation, 2015). This research will look into the readiness of business and entrepreneurship to gain advantages from the possible economic gains of open data (World Wide Web Foundation, 2015). Business readiness includes country level measurements on the level of training to make use of open data and the governmental support directed towards open data-driven innovations (World Wide Web Foundation, 2015). As previously explained this research expects a positive influence of business readiness on bilateral trade, through the creation of firm-specific knowledge resources and knowledge creation on foreign markets.
H1: Business Readiness for Open Government Data in the exporting country has a positive influence on its export levels.
Following business readiness, one can look at the role of implementation of open data. Implementation refers to the actual level of government datasets that are open, available for use and updated timely (World Wide Web Foundation, 2015). Again, taking into account the economic influence of open government data, this research will look at the so-called innovation implementation of open government data. This research takes into account all components of open data from the government or made available by the government that can be of significant value for enterprises or entrepreneurs (World Wide Web Foundation, 2013). The quality of the data available influences the ability to gain useful insights and possibly develop firm-specific knowledge resources, thereby influencing export levels.
H2a: Innovation Implementation of Open Government Data in the exporting country has a positive influence on its export levels.
The influence of innovation implementation in the importing country should also be taken into account. A high-level of innovation implementation in the importing country could be a disadvantage to the exporting country because it would mean that businesses and entrepreneurs in the importing country obtain highly valuable data, possibly causing the exporting country to lose their comparative advantage. Nevertheless, if the information available in the exporting country is available to businesses in the exporting country as well, it would enable the exporting country to gain market knowledge, assess country risk and lower international trade costs (Becerra-Fernandez et al., 2002; Demirkan et al., 2009; European Union, 2013). In this case a high level of innovation implementation in the importing country
could positively influence a country’s export levels. Businesses in the exporting country would be able to extract valuable information on markets and opportunities in the importing country, lowering their trade barriers.
H2b Innovation Implementation of Open Government Data in the importing country has a positive influence on the reporter’s export levels.
Allowing comparison between the ability to extract valuable information from open government data between the exporting and the importing country, business readiness of both countries is taken into account. One could expect that a low business readiness in the importing country is an advantage for the exporting country. This is because a low business readiness in the importing country would entail that companies and entrepreneurs in this country have a low readiness to extract value from, and obtain economic gains through, open data (World Wide Web Foundation, 2015). If the exporting country has a higher business readiness, this would mean that they have a higher ability to gain valuable information from open government data. This could lead to a competitive advantage for the firm in the reporting county if they are able to create firm-specific knowledge resources. Thus, if businesses and entrepreneurs in the exporting country have access to the datasets within their own country and the datasets of the partner country, their comparative advantage might increase if their business readiness is higher than that of the importing country.
H2c: The influence of Innovation Implementation in the exporting country on export levels is stronger if the Business Readiness of the exporting country is higher than the Business Readiness of the importing country.
H2d: The influence of Innovation Implementation in the importing country on the reporter’s export levels is stronger if the Business Readiness of the exporting country is higher than the Business Readiness of the importing country.
Next to the implementation of and readiness for open data, the impact of open government data on the economy is an important part of its influence on knowledge creation and export. This study will research how the impact of open government data on, start-up support, already operating companies, entrepreneurship and the eco nomy of a country in general have an influence on exports. One could expect that a positive impact of open government data on these factors leads to a positive influence on export levels, through for example the previously mentioned new business and idea generation (World Wide Web Foundation, 2015).
H3a: Economic Impact of Open Government Data in the exporting country has a positive influence on its export levels.
Taking into consideration the economic impact of open government data in the importing country, a multitude of outcomes could arise. A high level of economic impact in the importing country may increase the reporter’s export levels due to the positive economic impulse to this area of distribution. Nevertheless, it also leads to an impulse to the importing country’s entrepreneurial environment. This could lead to higher competition levels, lowering profits from export (Freund & Weinhold, 2004). Looking at both these effects, it seems that the impulse to the importing country’s distribution market would positive impact on export, whereas the competition levels would have a negative effect. Nevertheless, for this research a
positive effect of the economic impact in the importing country on the reporter’s export levels is expected.
H3b: Economic Impact of Open Government Data in the importing country has a positive influence on the reporter’s export levels.
3.2 The mediating role of knowledge
Following the analysis of the separate components of economic open government data, a second framework will be used to test the overall influence of economic open government data on the export levels of the reporter. This framework is complementary to the first framework. The three components of economic open government data are now weighted according to their importance in value creation and the ability to correctly measure the component (World Wide Web Foundation, 2015). By using this framework it can be examined if knowledge is created in the exporting country through the combination of all three aspects of economic open government data. Hypotheses tested for this part of the research will be elaborated upon based on of the following conceptual framework:
Hypotheses related to this framework will establish if economic open government data influence export levels, through the creation of new knowledge. Knowledge creation will be captured measuring the number of several sorts of patent applications, scientific and technical publications, and citable documents (Cornell University, INSEAD, and WIPO, 2016). This part of the research will take into account both the influence of economic open government data in the exporting and the importing country. First the influence of economic open government data of the reporter on the export levels of the reporter will be tested. Next, the possible mediating role of knowledge creation in this relationship will be analyzed.
H4a: Economic Open Government Data in the exporting country has a positive influence on its export levels.
H4b: Knowledge creation mediates the relationship between Economic Open Government Data in the exporting country and its export levels.
Second, the effect of economic open government data in the importing country on export levels of the reporter will be tested. Also for this relationship, the mediating role of knowledge creation will be evaluated.
H5a: Economic Open Government Data in the importing country has a positive influence on the reporter’s export levels.
H5b: Knowledge creation mediates the relationship between Economic Open Government Data in the importing country and the reporter’s export levels.
All hypotheses belonging the first framework will contribute to answering the research question, by separately examining the effect of every economic open government data component on bilateral trade. Furthermore, they allow a separate link between business readiness and innovation implementation and their combined influence on exports. The hypotheses belonging to the second framework examine the effect of a weighted measurement on economic open government data on export levels. Furthermore, the hypotheses in this framework will clarify the possible mediating role of knowledge creation in the reporting country.
4. Methodology
This research is a quantitative exploratory study, as its aims to generate theory as opposed to testing current theories. All data in this study are secondary data, obtained through database research. The topic of open government data is relatively new and therefore little studies have done research one its economic effect. Even more specifically, very little is known about its effect on international trade. Based on current literature one would expect to see a relation between open data and international trade. This research attempts to show if indeed one can establish an influence of open government data on export. Furthermore, in order to set out a first step towards explaining this relation, it will look into the role of knowledge creation in the different components of open government data. Being able to establish the influence of open government data on export and the role of knowledge creation in this relation would set the groundwork for a wide variety of future research.
4.1 Dataset
To test the hypotheses, this study uses time-series, cross-country data on US dollar export value of goods as collected by the UN COMTRADE (UN Comtrade Database). Previous studies on bilateral trade have also used data obtained from the UN COMTRADE (Carrere, 2006). The sample is specified to merchandise trade flows, and thus excludes trade in services. This is because of the limited existence of data on bilateral trade flows in services (Baier & Bergstrand, 2002). Furthermore, a gravity equation is included in the model for this study, mostly applicable to trade flows in goods (Baier & Bergstrand, 2002). The study uses annual export of goods data between 30 OECD countries, thus 435 country pairs and 2610 observations on export levels in total. As pointed out in the literature review, the OECD set out an open government data project that is directed towards increasing the number of international researches on the impact of open government data (OECD, n.d.). The OECD is
also doing comprehensive research on the influence of open data on for example innovation and growth (OECD, 2015). They indicated data-driven innovation as a key driver of growth, specifically pointing to the role of big data in promoting new industries and the creation of a comparative advantage (OECD, 2015). Therefore, it seems especially interesting to specify the sample for this study to the 30 OECD countries, due to their active role in the development of open government data. Furthermore, it could possibly lead to valuable outcomes for the OECD, stimulating further research on this topic.
The time range of the data is from 2013 until 2015. Measurements start in 2013 due to the independent variables in this study. Data on the independent variables have been obtained through the Open Data Barometer, a large-scale research collaboration between the Open Data Institute and the World Wide Web Foundation (World Wide Web Foundation, 2013). The Open Data Barometer is set out to reveal the use and impact of open data initiatives around the globe (World Wide Web Foundation, 2013). The first report on the Open Data Barometer in 2013 indicates that the concept of open government data has quickly spread around the world. By mid-2013 it is said to be a global idea, initiating the first large-scale measurement of open government readiness, implementation and impact in a set of 77 countries (World Wide Web Foundation, 2013). Following measurements for the Open Data Barometer in 2014 and 2015 included a growing number of countries. However, this study only includes the 30 OECD countries with available measurements from 2013 on. Since the research looks into the economic effect of open government data on export, the sample for the independent variables only includes the measurements related to the economic aspect of open government data. Further explanation on all variables included in the model will follow in the section below.
4.2 Measures and Data collection
This study includes measurements on three economic components of open government data, respectively: business readiness, innovation implementation and economic impact. Taking into account only these components of open government data as measured by the Open Data Barometer allows this study to specifically look at the economic aspect of open data.
Independent variables: Open Data Barometer. The Open Data Barometer sets out open government data levels for already three years. Because re-use of the data both for quantitative and qualitative purposes is highly applauded, the methodology is kept almost identical throughout their research. The measurements of the Open Data Barometer in 2013 and 2014 are gained through three kinds of data, namely: peer-reviewed expert survey responses, detailed data assessment and secondary data (World Wide Web Foundation, 2013; World Wide Web Foundation, 2014). For the Open Data Barometer 2013, the survey was conducted between July and October 2013. In order to gain insights into the open data levels for every country, researchers were asked to respond to a variety of questions (World Wide Web Foundation, 2013). Country expert researchers, who are taking part in the Open Data Barometer research, did the detailed data assessment (World Wide Web Foundation, 2013). The secondary data used in their research was obtained through five indicators, used for the establishment of open data readiness. Four of these indicators are independent expert surveys and the fifth is based on Internet penetration data as collected by the World Bank (World Wide Web Foundation, 2013).
The Open Data Barometer study in 2014 almost exactly replicated that of 2013. For this research the data were obtained between June and September 2014 (World Wide Web Foundation, 2014). The study in 2015 broadly replicated that of 2013 and 2014, however some minor adjustments where made. Only those relative to this research will be mentioned.
Data where obtained between May and July 2015. Next to the same data resources as previous years, this study also included a government self-assessment (World Wide Web Foundation, 2015). Governments where asked to fill-out a self-assessment questionnaire, leading to closer involvement of governments and the gain of an extra source of valuable information (World Wide Web Foundation, 2015). The rest of the data gathering remained constant to allow for comparison of the data over three years (World Wide Web Foundation, 2015).
All scores for the measurements on readiness and impact have been normalized using z-scores and rescaled to a range of 0-100. Here a score of 100 does not refer to a perfect score, but means that it is the highest score of all countries included in the study for that year. This allows for a comparison of the data over the different years (World Wide Web Foundation, 2013). The ratings on innovation implementation as included in the studies by the Word Wide Web Foundation, were not yet normalized (World Wide Web Foundation, 2013, World Wide Web Foundation, 2014, World Wide Web Foundation, 2015). However in order to enable comparison between the countries and to create the measurement on Economic Open Government Data for this study, these scores where also normalized using z-scores and rescaled to a 0-100 range.
All three studies done for the Open Data Barometer measure the different components of open government data under three pillars, namely: the political, the social and the economic pillar (World Wide Web Foundation, 2013, World Wide Web Foundation, 2014, World Wide Web Foundation, 2015). This study only takes into account the economic pillar allowing for a clear distinction on which data from the researches by the World Wide Web Foundation are included in the new dataset. The table below indicates which measures are included in this study and which components from the dataset, as previously described, are used for every measure.
Open government data component:
Economic measurement: Measured via:
Readiness Entrepreneurs & Business Primary & Secondary Data Implementation Innovation Dataset Cluster Dataset assessments
Impact Economic Primary Data
Table 1: Measurements Open Government Data (World Wide Web Foundation, 2013).
Now, the meaning and measurement for all three components of economic open government data will be specified. Business readiness. This variable measures the readiness of entrepreneurs and businesses to gain economic value from open data (World Wide Web Foundation, 2015). The survey questions corresponding to this component establish the level of training to use open data for individual or firms, and the level of direct government support for innovations directed towards increased gains from open data (World Wide Web Foundation, 2015). Furthermore, secondary data are included on firm-level technology absorption and the number of Internet users per 100 people (World Wide Web Foundation, 2013, World Wide Web Foundation, 2015).
Innovation implementation. Innovation implementation refers to; “data commonly used in open data applications by entrepreneurs, or with significant value to enterprise” (World Wide Web Foundation, 2014, p. 55). This measurement contains evaluations for map data, public transport timetables, crime statistics and international trade data (World Wide Web Foundation, 2013). From 2014 on, this measurement also includes ratings on public contract data (World Wide Web Foundation, 2014). All datasets are rated based on a ten-point checklist, in order to establish their value for entrepreneurs and businesses (World Wide Web Foundation, 2014). The checklist measures if the data exist, are online available,
