A global perspective on tech investment, financing, and ICT on manufacturing and service industry performance

International Journal of Information Management 43 (2018) 130–145

Contents lists available at ScienceDirect

International Journal of Information Management journal homepage: www.elsevier.com/locate/ijinfomgt

A global perspective on tech investment, financing, and ICT on manufacturing and service industry performance Delvin Granta, Benjamin Yeob, a b

T

⁎

Driehaus College of Business, DePaul University, 1 E Jackson Blvd, Chicago, IL 60604, United States School for New Learning, DePaul University, 1 E Jackson Blvd, Chicago, IL 60604, United States

A R T I C LE I N FO

A B S T R A C T

Keywords: Tech investment Decision tree induction Clustering Firm performance ICT

We investigate the influence of manufacturing and service technology investments (i.e., tech investments), Information and Communication Technologies (ICT), and financial factors on global manufacturing and service industry performance, from 2006 to 2014. This is accomplished by employing clustering, and decision tree induction, in conjunction with the Technology, Organization, and Environment framework (TOE) as the theoretical framework of the investigation. ICT and financial factors vary in importance across industries at different levels of technological advancement. Low-tech industries rely on loans and tech investments. As they move to transition industries, tech investments are more important. Once they become highly technologically advanced, bank lending policies become more important.

1. Introduction Manufacturing and service tech investments, cover a wide variety of technologies, referred to as ICT in this study. The former include flexible manufacturing systems (FMS), computer aided design (CAD), computer aided engineering (CAE), computer aided manufacturing (CAM), computer-controlled machines (CNC), bill of materials (BOM), customer relationship management (CRM), supply chain management (SCM), and just-in-time (JIT) systems; the latter include information and communication technologies (ICT) such as computer hardware and software, networking, telecommunications, CRM, ERP, cloud computing, and SCM systems. Companies invest in manufacturing (Meliciani, 2000; Vranakis & Chatzoglou, 2011) or service technologies (Huang, Ou, Chen, & Lin, 2006; Lee, Choi, Lee, Min, & Lee, 2016) to improve performance. Questions persist on how ICT impact performance (Huang et al., 2006; Lee et al., 2016; Vranakis & Chatzoglou, 2011), despite claims that they improve firm performance (Kossaï & Piget, 2014), economic growth and development (OECD, 2008), alter industry structures (Crowston & Myers, 2004), enable globalization (OECD, 2008), and other performance measures (Bloom et al., 2010; Botello & Pedraza Avella, 2014; Draca, Sadun, & Van Reenen, 2006; Vranakis & Chatzoglou, 2011). Tech investments are further motivated to improve performance (Koivunen, Hatonen, & Valimaki, 2008) by reducing transaction costs, inventory and cycle times, product quality, increasing flexibility, efficiency, productivity, and economic growth

⁎

(OECD, 2008; Vranakis & Chatzoglou, 2011). The technology paradox has diminished the effectiveness of ICT and stems from an inability to effectively measure the complexity of tech investment impact on performance (Richard, Devinney, Yip, & Johnson, 2009). ICT do not impact performance (Ho, Wu, & Xu, 2011; Motiwalla, Khan, & Xu, 2005), but several studies show they do (Huang et al., 2006; Kleis, Chwelos, Ramirez, & Cockburn, 2012; Mithas & Rust, 2016; Pakko, 2002). The effect of ICT investments and capability on performance (Piget & Kossaï, 2013; Sein & Harindranath, 2004), and the lack of industry level research (Crowston & Myers, 2004), suggest more research is needed. ICT investment studies exist at the company (Mithas & Rust, 2016), industry (Devaraj & Kohli, 2000), and country levels (Indjikian & Siegel, 2005), but the technology paradox raises concerns about such investments, particularly at the industry level (Abdi, 2008; Sabherwal & Jeyaraj, 2015). Past studies illustrate ICT investment, financial factors, and ICT capability improve performance, but the combined effect has not been extensively studied. High levels of ICT investment and strategies (Mithas & Rust, 2016), as well as financial factors (Park, Kim, & Kim, 2017), and ICT capability, improve performance (Yeo & Grant, 2017a, 2017b). Financial factors such as bank lending affect industry performance and growth (Obamuyi, Edun, & Kayode, 2012; Spring, Hughes, Mason, & McCaffrey, 2017). Hence, Huang et al. (2006) recommend investigating how ICT investment, financial factors, and ICT capability affect performance. Relationships between technological advancement levels and performance exist (cf. Bloom et al., 2010;

Corresponding author. E-mail addresses: [email protected] (D. Grant), [email protected] (B. Yeo).

https://doi.org/10.1016/j.ijinfomgt.2018.06.007 Received 6 February 2018; Received in revised form 28 June 2018; Accepted 28 June 2018 0268-4012/ © 2018 Elsevier Ltd. All rights reserved.

International Journal of Information Management 43 (2018) 130–145

D. Grant, B. Yeo

organizational complementarity of the systems and practices they support. This is consistent with Indjikian and Siegel (2005), who find that complementary tech investments related to labor and the IT-support work environment, increase productivity. Third, tech investments payoffs require companies to first improve their IT-enabled intangible assets, and their human ICT capability (Huang et al., 2006), and performance is contingent on the knowledge characteristics of companies (Liu, Yeung, Lo, & Cheng, 2014). These explain why companies are unable to fully exploit advanced manufacturing and ICT capabilities (Das & Narasimhan, 2001). Kwon (2007) finds a positive relationship between tech investments and five company performance variables: growth, competitiveness, customer relationship, external partnerships, and operational efficiency. Kleis et al. (2012) find that a 10% increase in tech investments is associated with a 1.7% increase in innovation output. Santhanam and Hartono (2003) find that companies with effective ICT capabilities achieve better financial performance than those with ineffective ones. In banking, tech investments are positively related to company performance (Byrd, Lewis, & Bryan, 2006), and Indjikian and Siegel (2005) find a positive relationship between tech investments and economic performance in developing and developed countries. According to Jung (2009), high capability online brokerage companies invest more in ICT and achieve better financial performance by providing better quality customer service. ICT investments influence organizational performance (Ramdani, 2012) and create new business opportunities (Weill & Ross, 2004). Im et al. (2001) use data from 238 companies to investigate the response to price and trading volume on ICT investment announcement and the effect of ICT investments to increase company market value over time. They find that the impact of tech investments is the same between financial and non-financial companies and ICT do not increase market value. Mithas and Rust (2016) investigate information technology strategies, tech investments, and company performances of over 300 U.S. companies. They find that companies with low levels of tech investments need to choose between revenue expansion and cost reduction. However, at higher tech investment levels, dual-emphases on ICT strategy or ICT strategic ambidexterity increasingly pays off.

Botello & Pedraza Avella, 2014; Draca et al., 2006; Kossaï & Piget, 2014), but as this line of research is underdeveloped, these issues motivate our RQ: How do tech investments, ICT use, and financial factors influence global industry performance across industries with different levels of technological advancement? Currently, research findings on the impact of ICT are incomplete (Schryen, 2013). They are better interpreted when ICT contexts are considered (Torero & Von Braun, 2006). We believe that ignoring them contribute to the fragmented findings on technology performance. ICT effectiveness in one context may not be replicated in others (Ko & OseiBryson, 2004). Therefore, Karanasios (2014) recommends rigorous research methods to interpret research findings. We use the Technology, Organization, Environment (TOE) framework (Tornatzky & Fleischer, 1990) to discuss the impact of technology, fixed asset investment, that serves as the organization context, and financial factors, as the environment context on industry performance. Clustering is used to group industries according to their levels of technological advancement. The theoretical research contributions to the extant literature are: (1) Different levels of technological advancement have different levels of ICT use, ICT investment, and financing; (2) Industry performance varies across levels of technological advancement, and influences ICT investment decisions; (3) Tech investments become increasingly important with higher levels of technological advancement, and help identify tradeoffs between tech investments, and levels of technological advancement. The paper is organized as follows, Section 2 is a literature review of tech investments, ICT performance measures, TOE framework, and ICT contexts relevant to the study. Section 3 is the research method, Section 4 covers the results, and we conclude the paper with a discussion of our findings in Section 5. 2. Literature review Section 2.1 discusses tech investments, Section 2.2 ICT performance measures, Section 2.3 the TOE framework, and Section 2.4 the two ICT contexts relevant to the study. 2.1. Technology investments Manufacturing tech investments (Meliciani, 2000; Pakko, 2002; Vranakis & Chatzoglou, 2011) and service tech investments (Huang et al., 2006; Lee et al., 2016) positively influence performance, but companies question their effect on performance (Lee et al., 2016; Vranakis & Chatzoglou, 2011). Technology investments refer to the procurement of IT, which provide long term benefits and can be evaluated based on their costs and benefits (Apostolopoulos & Pramataris, 1997). There are several ways to measure performance, including company market value (Im, Dow, & Grover, 2001), increased sales, revenue, product and service quality, growth, competitiveness, customer relations, partnerships, operational efficiency, profit, return on sales, market share (Campbell, 2012; Kwon, 2007; Liao, Wang, Wang, & Tu, 2015), and product quality, response time, relationship with clients or suppliers (José Tarí, 2005; Kannan & Tan, 2005). Tech investments have been studied at the company (Im et al., 2001; Mithas & Rust, 2016; Pakko, 2002), industry (Abdi, 2008; Devaraj & Kohli, 2000), and country levels (Indjikian & Siegel, 2005; Spring et al., 2017; Vranakis & Chatzoglou, 2011). There are three factors that lead to the technology investment paradox. First, there is a debate on the ability of tech investments to improve performance. Studies show that there is no impact of tech investments on company performance (Ho et al., 2011; Motiwalla et al., 2005), but Huang et al. (2006) and Im et al. (2001) disagree. The second relates to the complexity of measuring the impact of tech investments (Richard et al., 2009), which is often measured as a single factor, but should be measured as multiple components, and broken down into basic infrastructure, wireless, data center, and collaboration (Lee et al., 2016). Brynjolfsson and Hitt (2003) argue that measuring ICT capabilities requires a more accurate view of ICT and

2.2. ICT performance measures Global industry performance can be measured in several ways, such as return on assets (Liu et al., 2014), school performance (Marks & Printy, 2003), sales and profit (Botello & Avella, 2014), turnover and profitability (Koellinger, 2006), firm profitability (Kossaï & Piget, 2014), employee wages (Audretsch, van Leeuwen, Menkveld, & Thurik, 2001), industry sales growth (Yeo & Grant, 2017b), economic growth, jobs, and service (World Bank, 2016), trade (Bankole, Osei-Bryson, & Brown, 2015), capacity utilization (Yeo & Grant, 2017a), quality, operations strategy and flexibility (Arias Aranda, 2003), buyer-supplier relationships (Carr & Pearson, 1999), quality management (Kaynak, 2003; Sila & Ebrahimpour, 2005), and others (Lind, Sepúlveda, & Nuñez, 2000). ICT performance effectiveness depends on contextual factors such as manufacturing infrastructure (Archibugi & Coco, 2004), a capable workforce, country infrastructure (Bankole et al., 2015; Bollou, 2006; Henderson, 2002; Tan, Ng, & Jiang, 2018), ICT usage patterns (Zhang, Li, Qiao, Zhou, & Shen, 2018), and financial institutions’ lending practices (Obamuyi et al., 2012; Yeo & Grant, 2017b). Performance has also been investigated from social, economic, financial, non-financial, micro, and macro perspectives, which are beyond our research scope. Our research scope is limited to how financial factors, ICT, and tech investments affect sales performance. 2.3. The Technology, Organizational, and Environment (TOE) framework There are three reasons for using the TOE framework (Fig. 1). First, the TOE framework is a contextual theory. Given incomplete research 131

International Journal of Information Management 43 (2018) 130–145

D. Grant, B. Yeo

(2005) posit that it should be applied to other domains, thus justifying its use in this investigation. 2.4. ICT contexts relevant to the study The TOE framework is used to frame this investigation, so it is appropriate to discuss the two ICT contexts (environment and organization) of the study. Contexts refer to the settings or situational factors in which ICT are embedded. They play an important role in determining the success or failure of ICT in impacting performance (Piget & Kossaï, 2013; World Bank, 2016). In the ICT4D literature, results on the impact of ICT differ across various levels of analysis. While they are more consistent at the firm level, industry level results are less consistent (Devaraj & Kohli, 2000), prompting Crowston and Myers (2004) to advocate more industry level research. 2.4.1. Fixed asset investments context as organization proxy Fixed assets are long-term resources, such as plants, equipment, ICT, or processes that can generate cash flow, reduce expenses, improve sales in service and manufacturing companies. They are part of the organization context and performance measures such as, net profit, efficiency, and sales (Olatunji & Adegbite, 2014). They include CAD, CAM, CAE, CNC, JIT, BOM, FMS, CRM, SCM, computer software and hardware, telecommunications, networking equipment, industrial and technology related processes, and technology patents. Sayeed and Hogue (2009) investigate the impact of asset and liability management on the profitability of public and private banks in Bangladesh, and find that the profitability of banks is affected by the management of their assets and liabilities. The level of fixed asset investments depends on a company’s line of business. Hence, some industries are more capital intensive than others (Huang et al., 2006). The tech investment paradox (Huang et al., 2006) causes companies to question how their tech investments affect performance. There is evidence that tech investment levels lead to different performance outcomes. Kohli and Devaraj (2003), using meta-analysis, find varying results on ICT investments across sample sizes, data sources, and industries. Bharadwaj (2000) finds that companies with high technology capabilities perform better on a variety of profit and cost-based performance measures with respect to their tech investments. Eriotis, Frangouli, and Ventoura-Neokosmides (2011) study how fixed assets influence company profitability, by looking at the relationships of debt to equity ratios and company profitability. They find companies that self-finance are more profitable than those that borrow capital. Therefore, the need to better understand how fixed assets investments affect performance is important.

Fig. 1. Technology Organization Environment (TOE) Framework (Tornatzky & Fleischer, 1990).

findings on the impact of ICT (Schryen, 2013) and that ICT contexts can enhance interpretations of their impact (Torero & Von Braun, 2006), the TOE framework allows us to frame our investigation within the context of ICT. Second, the TOE framework is flexible and widely used in similar studies (cf. Baker, 2012) that look at how technological, organizational, and environmental factors affect company and industry decisions and performance (cf. Angeles, 2013, 2014; Hackney, Xu, & Ranchhod, 2006; Ryan, Abitia, & Windsor, 2000; Yeo & Grant, 2017a, 2017b, 2018; Zhu & Kraemer, 2005; Zhu, Kraemer, & Dedrick, 2004; Zhu, Kraemer, & Xu, 2006; Zhu, Kraemer, & Xu, 2003). Its constructs – technology, organization, and environment – capture a wide variety of contexts that can be used to study the impact of ICT. The technology construct can represent available or new ones relevant to companies. The organization construct can represent company size, scope, and formal and informal managerial and communication processes and structures; and the environment construct can include competitors, vendors, collaborators, and intermediaries, governmental and regulatory agencies. All three can affect technological innovation, decision making, or performance (Tornatzky & Fleischer, 1990). The framework has been used to investigate the role of RFID in green supply chains at Hewlett Packard’s recycling of ink-jet printers and garbage recycling in Grand Rapids Michigan (Angeles, 2013). It has also been used to study e-business innovation assimilation in more than 1800 companies in 10 countries, so as to develop an integrative model to examine three assimilation stages of initiation, adoption, and routinization (Zhu et al., 2006). Further, it has been used to investigate e-business adoption to identify drivers and inhibitors of e-business decisions in European firms (Zhu et al., 2003), and how TOE factors influence e-business company performance in the financial sector (Zhu et al., 2004). Zhu and Kraemer (2005) examine TOE factors that influence e-business usage in the retail industry. Hackney et al. (2006) use a case study to investigate web services in U.K. companies. Ryan et al. (2000) analyze the adoption of knowledge management technologies in U.S., Japan, and Mexican companies. Among more recent studies, the TOE framework has been used to investigate determinants of cloud computing adoption among firms in Taiwan (Hsu, Ray, & Li-Hsieh, 2014), Portugal (Oliveira, Thomas, & Espadanal, 2014), and India (Gangwar, Date, & Ramaswamy, 2015), as well as Enterprise 2.0 adoption in China (Jia, Guo, & Barnes, 2017). It has been used to study how contextual factors influence web knowledge exchange among small and medium enterprises (SMEs) in Spain (Palacios-Marqués, Soto-Acosta, & Merigó, 2015). Bradford, Earp, and Grabski (2014) use it to study how technological and organizational factors lead to lapses in IT governance and hinder centralized end-to-end identity and access management (CIAM). Third, despite several applications of the TOE framework, it needs more theoretical development (Baker, 2012). Hence, Zhu and Kraemer

2.4.2. Financial context as environment proxy The environment comprises several factors. One of which is lending policies, referred to as the financial context that enables entrepreneurship, technological innovation, and economic performance. A viable financial sector benefits company, industry, and economic performance (Beck, Demirguc-Kunt, & Maksimoic, 2005) is important to firm and industry performance, and economic growth (Levine, 1997). Financial reforms in Latin America increased competitiveness (Libanio & Moro, 2006), and financial liberalization in Ghana improved the economy (Asamoah, 2011). Studies show that financial factors influence industry performance more than ICT (Yeo & Grant, 2017b, 2018). However, there are questions on whether financial factors have varying impact on firms and industries with different levels of technological advancement. The economic importance of financial factors (Bankole et al., 2015; Yeo & Grant, 2017b), the mixed results on the impact of technologies (Piget & Kossaï, 2013), particularly at the industry level (Devaraj & Kohli, 2000), the recommendation for more industry level studies (Crowston & Myers, 2004), and the technology paradox (Huang et al., 2006), motivate us to study their collective impact on the performance of 132

International Journal of Information Management 43 (2018) 130–145

D. Grant, B. Yeo

Percent of firms whose recent loan application was rejected, 6. Proportion of investments financed by banks, and 7. Percent of firms identifying access to finance as a major constraint. Bank lending policies (Obamuyi et al., 2012) and financial factors (Yeo & Grant, 2017b) have been used to investigate ICT and manufacturing performance. Beck et al. (2005) discuss six company performance and growth constraints: 1. The need for special connections with banks; 2. Banks’ lack of money; 3. High interest rates; 4. The adverse impact of having to deal with bank bureaucracies; 5. Collateral requirements; and 6. Lack of access to operations financing. These seven financial variables relate to one or more constraints, justifying their inclusion in the investigation, and representation of the financial, and hence, environment, context. 3.2. Research method overview SPSS Modeler v.15 and R are used for empirical analysis. The method consists of four steps and additional details are discussed in the sub sections and in Section 4.2. In the first step, SPSS Modeler is used to partition the data into distinct clusters, based on levels of technological advancement, using the four ICT variables. Technological advancement serves as a control condition to better understand industry performance. R is used in the second step to perform Shapiro-Wilk Tests to investigate if the data in each cluster are normally distributed, so as to determine the appropriate mean difference tests to perform in step three. In the third step, since we believe each cluster performs differently, R is used to test for mean differences in annual sales growth using Kruskal-Wallis and Dunn’s Tests. Lastly, SPSS Modeler is used to perform predictive analyses on each cluster using decision tree induction, to identify predictors of annual sales growth. The predictors come from the four ICT variables, tech investment variable, and the seven financial variables used in the analysis. This multi-method analysis enriches the findings and discussion more than a single method can.

Fig. 2. Theoretical model for this study.

industries, across different levels of technological advancement. 3. Research method We develop a conceptual research model (Fig. 2) based on the TOE framework and it has four constructs, (i) the technology context represented by ICT variables; (ii) the organization context represented by tech investment; (iii) the environment context represented by financial variables; and (iv) performance, the target variable, measured by annual industry sales growth. The first three constructs constitute the independent variables. 3.1. Data and variables

3.2.1. Clustering Clustering is a data mining technique that places records into homogenous groups. It improves the results by enabling the application of predictive modelling techniques to each cluster at a granular level, rather than to the entire data set (Osei-Bryson & Samoilenko, 2014). Clustering can achieve what control variables do in traditional statistics, and has been used in operations research (Rai, Tang, Brown, & Keil, 2006), ecommerce (Okazaki, 2006), fraud investigation (Aggarwal & Yu, 2001), information systems (Balijepally, Mangalaraj, & Iyengar, 2011; Osei-Bryson, Dong, & Ngwenyama, 2008), and software engineering (Wallace, Keil, & Rai, 2004). No clustering technique is universally applicable in uncovering the variety of structures in multidimensional datasets (Jain, Murty, & Flynn, 1999). Our objective is to partition the dataset into homogenous clusters. Therefore, we use kmeans, a non-hierarchical clustering method available in SPSS Modeler v.15 to group records based on the four ICT variables, that minimize the intra-cluster distance relative to the cluster means. The quality of the clusters measured by their silhouette statistic, represents the validity of the clusters. Validity is the average level of cohesion and separation of the clusters (Osei-Bryson & Ngwenyama, 2014). Cohesion measures the average similarity of data points within a cluster, and separation measures the average dissimilarity between data points from those in the nearest cluster. The k-means algorithm requires users to specify the number of clusters desired, and the highest silhouette measure determines the optimum number of clusters. For purposes of the analysis, we wanted more than two clusters to enrich the analysis, and enough clusters to be manageable and easy to interpret. Using trial and error, we found that three clusters produced the best silhouette measure (s = 0.5). Since the ICT variables are used to define these clusters, they have distinct levels of technological advancement.

The data used in this study come from the World Bank’s World Enterprise Survey (The World Bank, 2015), and were collected from private companies. U.S. and Canadian companies were excluded from the survey. The data are organized into 12 topics: corruption, crime, finance, firm characteristics, gender, informality, infrastructure, innovation and technology (ICT), performance, regulation and taxes, trade, and workforce. We investigate four constructs: ICT, tech investments, financial factors, and industry performance. Inherent data limitations are discussed in Section 5. Data mining techniques help to circumvent these limitations to draw useful insights. Responding to Crowston and Myers (2004) request for more industry research, and the lack of consistent industry findings on the impact of ICT (Devaraj & Kohli, 2000), we use industry data. Each record represents an industry in a specific year and country. For example, in 2006, 70.30% of industries in Argentina’s food industry, use email to communicate with clients and suppliers, hence the value for this variable in this record is 70.30%. The technology context is represented by four ICT variables: 1. Percent of firms with an internationally-recognized quality certification, 2. Percent of firms having their own website, 3. Percent of firms using e-mail to interact with clients or suppliers, and 4. Percent of firms with an annual financial statement reviewed by external auditors. They have been used to investigate technology assimilation and innovation (Zhu et al., 2004, 2006). Our analysis includes manufacturing and service industries, so the fifth technology and innovation variable, the percent of firms using technology licensed from foreign companies, is excluded since it pertains only to manufacturing in the survey. The organization context is represented by tech investments, and is operationalized by the percent of industries buying fixed assets. The environment context, represented by the financial context has seven variables: 1. Percent of firms with a bank loan/line of credit, 2. Proportion of loans requiring collateral, 3. Value of collateral needed for a loan (% of the loan amount), 4. Percent of firms not needing a loan, 5.

3.2.2. Decision tree induction Decision tree induction is a non-parametric data mining method for 133

International Journal of Information Management 43 (2018) 130–145

D. Grant, B. Yeo

financed by banks, and the percent of firms with an internationally recognized quality certification are substantially positively skewed (skewness = 1.80, skewness = 1.23, and skewness = 1.57 respectively). They also have high kurtosis levels (kurtosis = 3.25, kurtosis = 2.11, and kurtosis = 2.90 respectively). The dependent variable, annual sales growth, also has a high kurtosis level (kurtosis = 3.52) and varies from -48.00% to 64.70%. There are several missing data for the percent of firms whose recent loan application was rejected, resulting in 254 records for this variable. There are also other missing values in the data set, but we refrained from imputation for three reasons. First, each record is independent because an industry record in a specific country and specific year, has no bearing on other industry records, be it in the same country and/or year, thus making imputation more difficult. Second, even though imputation is possible with independent records, the process introduces some degree of error. To avoid this, we preserve the dataset and use a decision tree induction method capable of handling missing data. Third, there are many imputation methods, such as mean, regression, and maximum likelihood, but the choice of method depends on factors such as randomness and possible reasons for the missing data. Researchers may not be privy to these factors (Lodder, 2013), which applies in our case, since we are using secondary data. Consequently, we are not in a position to determine the appropriate imputation method.

classification and prediction (Osei-Bryson, 2004), and used to analyze non-linear relationships, requiring no assumptions about the data frequency distribution (Pal & Mather, 2003). Furthermore, the data set has missing values and decision tree induction is not affected by them. For these reasons, decision tree induction is preferred to traditional statistical methods such as regression modelling that has strict requirements for the data distribution and is affected by missing values. It has been used in medicine (Kobayashi, Takahashi, Arioka, Koga, & Fukui, 2013; Murphy & Comiskey, 2013; Rodríguez et al., 2016; Ture, Tokatli, & Kurt, 2009), and marketing (Amir, Osman, Bachok, & Ibrahim, 2015; Díaz-Pérez & Bethencourt-Cejas, 2016; Kim, Timothy, & Hwang, 2011; Legohérel, Hsu, & Daucé, 2015), but comparatively less in mainstream IS research for theory testing and development (Osei-Bryson & Ngwenyama, 2014). It uses a set of decision rules defined by conditional probability that splits data, to identify predictor values that determine the outcome. There are two popular decision tree induction methods, Chi-Squared Automatic Interaction Detection (CHAID) and Classification and Regression (CART), both available in SPSS Modeler v.15. CHAID allows multivariate splits based on each independent predictor (Samoilenko, 2008), whereas CART allows binary splits. The variables in this study are continuous, so we refrained from limiting the induction process to binary splits, thus enabling a richer analysis. There are other decision tree induction methods that are capable of performing multivariate splits (cf. Barros, de Carvalho, & Freitas, 2015), including Entropy and Gini splitting methods, but we chose CHAID given its availability in SPSS Modeler v.15, and its popularity. We acknowledge that researchers may choose to apply other induction methods besides CHAID that are available in various other software packages, and are capable of producing multiple splits.

4.2. Clusters The three clusters are fairly similar in size, with 274, 245, and 241 records, and comprising 36.1%, 32.2%, and 31.7% of the data set respectively (Fig. 4). The ratio of the largest to the smallest cluster is 1.14. All four ICT predictors play varying roles in defining the three clusters. Their relative importance is shown in the Fig. 5. The most important predictor is the percent of firms using email to interact with clients or suppliers (1.00), followed by the percent of firms with an annual financial statement reviewed by external auditors (0.65), the percent of firms having their own websites (0.62), and lastly, the percent of firms with an internationally recognized quality certification (0.21). On average, 35.46% of the 274 industries in Cluster 1 use email to interact with clients and suppliers, 32.83% use external audits, 17.03% have a website, and 9.07% have internationally recognized certifications. The majority of industries in this cluster are in Africa, South Asia, West Asia, and Central America. Africa has 51.46%, South Asia 12.41%, West Asia and Central America with 8.76% each. Service industries make up the majority of Cluster 1, with 26.28% of them in retail, and

4. Findings 4.1. Descriptive analyses The geographical regions of the industries are shown in Fig. 3. Most of the records (30.66%) are from Africa; Western Europe and the Caribbean have the least. Africa has the most industry records followed by Eastern Europe and South America. Western Europe is the only developed region, which is a limitation of the data set discussed in Section 5. Table 1 summarizes the descriptive statistics of the predictors and dependent variable. The sample size N, is less than 760 due to missing data, another limitation discussed in Section 5. The percent of firms whose recent loan application was rejected, proportion of investments

Fig. 3. Breakdown by Geographical Region. 134

International Journal of Information Management 43 (2018) 130–145

D. Grant, B. Yeo

Table 1 Summary of Predictors and Dependent Variable. Variable

Mean

Median

Std. Dev

Skewness

Kurtosis

N

Percent of firms with a bank loan/line of credit Proportion of loans requiring collateral Value of collateral needed for a loan (% of the loan amount) Percent of firms not needing a loan Percent of firms whose recent loan application was rejected Proportion of investments financed by banks Percent of firms identifying access to finance as a major constraint Percent of firms with an internationally-recognized quality certification Percent of firms having their own website Percent of firms using e-mail to interact with clients/suppliers Percent of firms with an annual financial statement reviewed by external auditors Percent of firms buying fixed assets Real annual sales growth

33.27 76.77 170.40 41.55 13.29 15.48 28.85 17.32 40.27 65.42 45.34 43.95 5.63

31.65 82.50 159.30 41.20 7.35 12.30 25.60 13.30 38.55 71.40 43.20 43.60 6.80

20.86 19.63 74.26 18.38 16.33 13.16 17.93 14.34 23.93 27.04 24.16 18.10 12.28

0.35 −1.01 0.94 0.16 1.80 1.23 0.74 1.57 0.30 −0.50 0.22 0.11 −0.66

−0.76 0.48 1.96 −0.32 3.25 2.11 0.13 2.90 −0.86 −0.93 −0.84 −0.58 3.52

742 684 598 739 254 743 751 751 750 745 751 751 682

of 18.24% have internationally recognized certifications. Eastern Europe and South America house most of the industries with 41.08% and 28.63% respectively, followed by Africa with 8.30% and West Asia with 7.05%. Similar to Cluster 1, most industries in Cluster 3 are retail with 23.65%, and other services with 22.82%. Based on the distributions of the four ICT predictors in Cluster 3, we believe they comprise industries in transition that rely heavily on email and websites, but are not sufficiently technologically advanced to achieve internationally recognized certifications. They also do not possess advanced accounting procedures and processes to permit the use of external audits. We define Cluster 3 as the transition cluster. Fig. 6 summarizes the distributions of the ICT predictors in each cluster. The violin charts show how the ICT variables are distributed in each cluster, and the box charts show the respective medians and inter quartile ranges. In terms of internationally recognized quality certifications, we can see that the three clusters – from developing to advanced – have increasingly higher percentages of firms that have international certifications. The mean percentages also increase progressively (9.07%, 18.24% and 25.95% respectively). Regarding the use of websites, the transition cluster has more industries with a higher percentage of industries using websites compared to the advanced cluster. The mean percentage of industries using websites for the transition cluster (55.91%) is also higher than the advanced cluster (51.33%). It is important to note that the transition cluster has industries that do not use websites, with a minimum percentage of 0%, while the advanced cluster has a minimum of 10.50%. The transition cluster has a large proportion of African industries (29.39%), where

18.98% in other services. The distributions of the four ICT predictors, lead us to believe that Cluster 1 represents industries with low technological advancement, because they do not rely heavily on email and websites. They are likely to be low tech service industries in developing countries, which explains their low percentages of internationally recognized certifications and use of external audits. Consequently, we define Cluster 1 as developing, due to its low-tech characteristics. Cluster 2 comprises 245 industries and an average of 79.01% use email to interact with clients and suppliers. An average of 72.16% use external audits. An average of 51.33% has a website and an average of 25.95% has internationally recognized certifications. These industries are more regionally dispersed compared to Cluster 1. Africa, South America, Central America, South Asia, and East Asia account for the majority of the industries, with 29.39%, 16.33%, 14.29%, 11.84%, and 8.98% respectively. The three top industries in Cluster 2 are other services, retail, and food manufacturing, making up 15.51%, 13.88%, and 11.84% respectively. These distributions suggest that Cluster 2 is the most technologically advanced of the three clusters, primarily due to their use of email and websites. They are likely to have advanced business processes, given the high percentage of firms using external audits. This suggests they are larger industries with more established companies. It is likely that the other services industries in this Cluster are technologically focused, providing high-end ICT consulting services. We define Cluster 2 as advanced, due to its high-tech characteristics. Among the 241 industries in Cluster 3, an average of 86.53% use email to interact with clients and suppliers, and an average of 33.29% use external audits. An average of 55.91% use websites and an average

Fig. 4. Cluster Breakdown. 135

International Journal of Information Management 43 (2018) 130–145

D. Grant, B. Yeo

Fig. 5. Predictor Importance for Clustering.

5. Regarding the percentage of firms with an annual statement reviewed by external auditors, we find that the percentages increase progressively from the developing to the advanced cluster. Fig. 7 shows scatter plots of the ICT variables, independent variables, and annual sales growth for each cluster. Pertaining to all four ICT variables, we find that the developing cluster comprises lower percentages and correspondingly lower annual sales growth. We find similar distributions of annual sales growth between the transition and advanced clusters. This suggests that the use of ICT do not always improve annual sales growth. Next, we look at fixed asset investments and annual sales growth. The mean annual sales growth for the developing, transition, and advanced clusters are 3.194%, 8.357%, and 5.453% respectively. ShapiroWilk Tests are performed on the clusters to determine if their annual sales growth are normally distributed. Results from the developing and

having a website can inhibit sales performance (Yeo & Grant, 2018), because poorly developed and maintained websites negatively affect business success, performance, and value (Hahn, 2003; Hoffman, 2011; Schonberg, Cofino, Hoch, Podlaseck, & Spraragen, 2000). Website quality is not captured in the data set, so we are unable to speak with authority on this finding. Pertaining to the use of email, we find very high email use among industries from the transition and advanced clusters. Like websites, the industries in the transition cluster have a higher percentage of firms using email (86.53%) compared to the advanced cluster (79.07%), and the respective minimum percentages are 52.60% and 36.60%. It is plausible that industries from the advanced cluster have moved beyond email to other forms of electronic communication that are used by technologically advanced industries. The data set does not include ICT-related variables to capture these advanced forms of communication. This limitation is discussed in Section

Fig. 6. Cluster Characteristics. 136

International Journal of Information Management 43 (2018) 130–145

D. Grant, B. Yeo

Fig. 7. Annual Sales Growth by Cluster and ICT.

Fig. 8. Correlation of Fixed Asset Investments and Annual Sales Growth by Cluster.

p < 0.001; r = 0.36, p < 0.001, and r = 0.23, p < 0.001 respectively). These bivariate relationships suggest that fixed asset investment, a proxy for tech investments, is relevant to annual sales growth. The transition cluster has the strongest correlation and the highest sales growth, followed by the advanced, and developing cluster.

transition clusters show that their annual sales growth are not normally distributed (W = 0.916, p < 0.001; and W = 0.981, p = 0.004), and for the advanced cluster, it is marginally normal (W = 0.988, p = 0.066). Therefore, rather than using an ANOVA, the Kruskal-Wallis and Dunn’s Tests are used to compare mean annual sales growth differences among the three clusters, to ascertain their performance levels. The mean differences are significant (H = 21.417, p < 0.001). Specifically, the transition cluster has significantly higher annual sales growth than the advanced and developing clusters (z = -3.105, p = 0.003; and z = -4.523, p < 0.001, respectively), and there is no significant difference in annual sales growth between the advanced and developing clusters (z = 1.314, p = 0.283). Fig. 8 shows a scatter plot of both variables, along with the result of a Pearson Correlation. The results show there is a small, positive correlation between the two variables in the developing, transition, and advanced clusters (r = 0.22,

4.3. Predictive analyses We use decision tree induction on each cluster to identify the predictors of sales growth. We re-coded annual sales growth into a categorical variable, comprising positive or negative growth. This simplifies decision trees and yields more useful results that can support theory and management, as recommended by Esposito, Malerba, Semeraro, and Kay, (1997). This approach is also consistent with previous studies (cf. Yeo & Grant, 2017b, 2018). Each decision tree induction comprises 137

International Journal of Information Management 43 (2018) 130–145

D. Grant, B. Yeo

Fig. 9. Important Predictors of Annual Sales Growth in Cluster 1.

loan or line of credit. In the first branch, industries with less than 10.70% of industries with a bank loan or line of credit exhibit negative sales growth (χ2 = 49.495, p < 0.001). The second branch comprises industries with more than 10.70% and less than or equal to 41.30% of industries with a bank loan or line of credit. These exhibit positive growth (χ2 = 49.495, p < 0.001). Among these, those with less than or equal to 7.30% of industries whose loan was rejected, exhibited negative growth. Those with more than 7.30% exhibit positive growth (χ2 = 28.194, p < 0.001). This suggests that poorly performing industries are weeded out by the loan application process. Even though a larger percentage of industries in this branch have loans, those with a higher percentage of rejections include reputable industries with better performance. The third branch includes industries with more than 41.30% of industries with a bank loan or line of credit, exhibit positive growth (χ2 = 49.495, p < 0.001). Industries with access to finance tend to invest more to expand their businesses. However, in this lowtech industry cluster, the percent of industries buying fixed assets is among the important predictors of annual sales growth.

the following settings: maximum tree depth = 3, minimum records in parent branch = 2.0%, and minimum records in child branch = 1.0%. The quality of each decision tree is accessed by its accuracy in correctly predicting the outcomes of the target variable from each corresponding confusion matrix. These matrices are excluded from the paper in view of length considerations. Since there are two possible outcomes – positive and negative growth – the accuracy is computed as the sum of the percentage of records correctly predicted “positive growth” and the percentage of records correctly predicted as “negative growth”. The results and explanations are discussed in the following sub-sections. 4.3.1. Cluster 1 developing Cluster characteristics illustrate that Cluster 1 is made up of lowtech service industries and four variables are significant predictors of fixed asset investments. The decision tree correctly predicts 64.23% of the cases. The four predictors and their relative importance are shown in Fig. 9. The percent of firms with a bank loan or line of credit is the most important predictor, followed by the percent of firms whose recent loan application was rejected, the percent of firms buying fixed assets, and the average value of collaterals needed for a loan. It suggests that these industries are made up of smaller low-tech service industries in developing countries, where access to finance and technology is burdensome, stifling fixed asset investments, which retards annual sales growth. Fig. 10 shows the results of the CHAID decision tree induction that is used on the cluster. The top node (node 0), splits into three branches based on the most important predictor, the percent of firms with a bank

4.3.2. Cluster 2 advanced Cluster 2 comprises industries that are the most technologically advanced of the three clusters. The decision tree correctly predicts 77.14% of the cases. There are two important predictors in Fig. 11, the average value of collateral needed for loans, the most important predictor, followed by the proportion of loans needing a collateral, a distant second. Since these industries are likely to be made up of highertech firms with advanced business processes, financial predictors are

Fig. 10. Decision Tree for Cluster 1. 138

International Journal of Information Management 43 (2018) 130–145

D. Grant, B. Yeo

Fig. 11. Important Predictors of Sales Growth in Cluster 2.

(> 236.00%), resulting in negative growth. The effect of bank lending on performance is evident. There are no ICT predictors in this cluster, implying the business environment is more important than the use of ICT in this cluster.

expected to be more important than ICT variables in improving annual sales growth. Fig. 12 shows the results of the CHAID decision tree induction for this cluster. The top split of the decision tree is based on the average value of collateral needed for loans as a percent of the loan amount. The first branch comprises industries with required collaterals of less than or equal to 147.70% that exhibit positive growth (χ2 = 111.214, p < 0.001). The only exceptions are those that have more than 44.70% and less than or equal to 48.30% of industries not needing a loan, exhibit negative growth. Since there are only five such cases, of which three exhibited negative growth, this may be a prediction anomaly (χ2 = 32.671, p = 0.001). The second branch comprises industries with an average collateral of more than 147.70% and less than or equal to 236.00% of loans. These industries exhibit positive growth (χ2 = 111.214, p < 0.001). The third branch has industries with collaterals of more than 236% of loans, exhibit positive growth. However, those with more than 79.80% of loans requiring a collateral, exhibit negative growth (χ2 = 41.752, p < 0.001). Securing a loan requires a collateral, particularly in developed countries where business practices are well established and transparent. High loan collaterals are burdensome to industries and this is reflected in the high average collateral value

4.3.3. Cluster 3 transition Cluster 3 includes industries in transition. The decision tree predicts 78.01% of the cases correctly. From Fig. 13, the most important predictor is the percent of firms buying fixed assets, followed by the percent of firms using external audits, and finally the percent of firms not needing a loan. Since this cluster comprises industries in transition, we expect the results to exhibit characteristics from the other two clusters. Fig. 14 shows the results of the CHAID decision tree induction. The top split is the percent of firms buying fixed assets. The first branch has industries with less than or equal to 45.90% of industries buying fixed assets, exhibiting positive growth (χ2 = 32.573, p < 0.001). However, those with less than 14.40% of industries having external audits, exhibit negative growth (χ2 = 19.892, p = 0.002). External audits diagnose process and technology limitations and suggest the need for improvements. Some process improvements require software and hardware technology improvements, which are fixed asset investments that may

Fig. 12. Decision Tree for Cluster 2. 139

International Journal of Information Management 43 (2018) 130–145

D. Grant, B. Yeo

Fig. 13. Predictors of Sales Growth in Cluster 3.

be a financial constraint for transition industries. Therefore, the low level of fixed asset investments reduces the benefits of external audits that improve operations and performance such as, increased annual sales growth. This is why external audits have a negative impact on performance on four of the seven cases that exhibit negative growth. The small number of cases may be an anomaly, as predictive models are not 100% accurate. In the other two branches, industries with more than 45.90% of industries buying fixed assets, exhibit positive growth (χ2 = 32.573, p < 0.001), illustrating the importance of fixed asset investments. The findings from this cluster, suggest that fixed asset investments have a positive influence on sales growth. This is corroborated by the highest correlation between fixed asset investments and annual sales growth demonstrated in this cluster (r = 0.36, p < 0.001).

performances across industries. Fourth, studies illustrate the importance of financial factors to ICT innovation, company, and industry performance (Yeo & Grant, 2018). We argue that the influence of financial factors varies across different industries. Fifth, there is insufficient research on the relationship between ICT advancement levels and performance (Bloom et al., 2010; Botello & Pedraza Avella, 2014; Draca et al., 2006; Kossaï & Piget, 2014), so we investigate how ICT, tech investments, and financial factors influence global industry performance across industries at different levels of technological advancement. The TOE framework provided the theoretical research model for the investigation, which has four ICT variables, seven financial variables, and fixed asset investments. The grouping of industries is accomplished by a k-means algorithm to create three industry clusters (developing, transition, and advanced), using the ICT variables. The benefits of data mining techniques and dataset limitations lead us to use decision tree induction to identify predictors for each cluster.

5. Discussion and conclusion 5.1. Summary

5.2. Implications of technology investments on industry practice Five factors motivate this research. First, there is a lack of consensus on the impact of ICT (Piget & Kossaï, 2013; Schryen, 2013; Sein & Harindranath, 2004), particularly at the industry level (Devaraj & Kohli, 2000). Second, Crowston and Myers (2004) recommend more industry ICT research. Third, there is an ongoing debate on the effectiveness of tech investments on company performance (Abdi, 2008; Huang et al., 2006). The impact of technology is contextually dependent (Ko & Osei-Bryson, 2004; Yeo & Grant, 2017a, 2018), resulting in varying and conflicting results across several studies. We posit that tech investments and the innovative use of technology result in varying

The bivariate analysis shows that tech investment is positively correlated with industry performance in the transition cluster. It is an industry performance predictor in the developing and transition clusters, but not the advanced cluster. The positive effect of tech investment is well documented (Huang et al., 2006; Yeo & Grant, 2018), while the inability of tech investments to increase performance may be related to the Law of Diminishing Returns. The advanced industry cluster may be at the point of diminishing returns, so incremental tech investments do not improve performance. Tech investments can also be disruptive and

Fig. 14. Decision Tree for Cluster 3. 140

International Journal of Information Management 43 (2018) 130–145

D. Grant, B. Yeo

Fig. 15. Importance of Tech investments by Industry Type.

Comparatively, we find the percent of firms with external audits is the second most important predictor of industry performance, after tech investments, in the transition cluster. The clustering analysis shows that transition industries are heavier users of ICT compared to the developing industries. Therefore, ICT play a more important role. We find the transition cluster, with higher tech investments and a large percentage of firms with external audits, exhibit positive growth. Yeo and Grant (2017a) posit firms that use external audits are more technologically advanced than those that do not. In the advanced cluster, only the financial variables – the value of collaterals, and the proportion of loans not needing collaterals – are important. We find that high collaterals are unfavorable to industry performance. This explains why technologically advanced industries benefit more from favorable bank lending, particularly collaterals and industry policies that enhance their global competitiveness (Spring et al., 2017). In response to our research question, we find that ICT are less important than financial factors across different levels of technological advancement. From a technological innovation perspective, developing industries initially rely on loans and tech investments. As they move towards becoming transition industries, tech investments become more important. As they become technologically advanced, bank lending policies and financial factors become more important. Our findings illustrate the effect of tech investments across the three industry types (Fig. 15). As industries grow, tech investments become increasingly important with increasing levels of technological advancement. This helps borrowers identify the tradeoff between tech investments, financing, and their levels of technological advancement. As they become highly technologically advanced, the direct impact of technological investments diminishes, as factors like financing become increasingly important. Borrowers armed with this knowledge can improve their lending strategies to maximize returns on investment based on the technological advancement of different industries.

ineffective, due to poor change management procedures. With respect to the developing and transition clusters, tech investments are a moderate and strong predictor respectively. Industries in the developing cluster are low users of ICT, as demonstrated by our descriptive analysis. Most of the industries are in Africa (51.46%), South Asia (12.41%), West Asia (8.76%), and Central America (8.76%). They are mostly retail (26.28%) and other services industries (18.98%) that rely heavily on financing. The moderate importance of tech investments points to their long-term horizon, where returns are not immediately realized. This explains the higher importance of tech investments in the transition cluster, comprising industries from Eastern Europe (41.08%) and South America (28.63%), and mostly in retail (23.65%) and other services (22.82%). Findings from the transition cluster show that higher tech investments lead to stronger industry performance. Incidentally, the transition industries exhibit significantly higher sales growth than the developing and advanced industries. Given the similarities in findings with the developing cluster, and its higher technological advancement, transition industries are better positioned to reap their tech investment benefits, compared to the developing industries. In response to our research question, tech investments are an important predictor of industry performance, which vary with the level of technological advancement. It is moderately important to industries in the early development stages or to low technology industries. In moderately technologically advanced industries, tech investments are important for performance. This is consistent with previous findings (Bloom et al., 2010; Botello & Pedraza Avella, 2014; Draca et al., 2006; Kossaï & Piget, 2014). Among the most technologically advanced industries, tech investments become less critical to industry performance, as financial factors become more important. This knowledge is important to borrowers and lenders alike. Lenders should better understand how industries may benefit from tech investments, given their levels of technological advancement. Borrowers must understand how to best leverage borrowed capital to maximize returns on investment, depending on their stages of technological advancement.

5.4. Policy and management implications ICT capabilities and tech investments have significant economic impact in developed countries and insignificant impact in developing countries (World Bank, 2016) and play a vital role in company, industry, and economic performance (cf. Carr & Pearson, 1999; Crowston & Myers, 2004; Kaynak, 2003; Kossaï & Piget, 2014; Sila & Ebrahimpour, 2005). Performance and competitiveness are influenced by industrial and government policy (Spring et al., 2017) and bank lending policies (Obamuyi et al., 2012). Favorable industrial and lending policies enable technological innovation and economic growth (Spring et al., 2017). Our findings elucidate the role of technology and lending policies at various levels of technological advancement.

5.3. Importance of ICT to industry practice The top two predictors in the developing cluster in descending order are, the percent of firms having a bank loan or line of credit, and the percent of firms whose loan application was rejected, with the former being much more important. Our findings show that industries in this cluster that have a low percent of firms having a bank loan or line of credit generally exhibit negative sales growth. ICT do not appear to be important to these industries. ICT variables are used to define the clusters, so we expect the developing cluster to exhibit low ICT use. 141

International Journal of Information Management 43 (2018) 130–145

D. Grant, B. Yeo

important contribution is we demonstrate the ability of technological advancement levels to moderate ICT use, tech investment, and financing regarding varying industry performance among technological advancement industries and how this knowledge informs ICT investment decisions. Tech investments become more important with increasing levels of technological advancement and aid the evaluation of tradeoffs between tech investment, financing, and ICT use. The level of technological advancement helps to prioritize fixed asset investments. The first industry contribution is that financial variables are important for the performance of industries with varying levels of technological advancement or capability. ICT capabilities affect ICT investments and performance results (Huang et al., 2006). For example, insufficient investments in hardware and human resources, inhibit performance. Therefore, industries benefit by assessing their technological capabilities and other ICT resources, and various financial factors. Second, ICT are relevant to transition industries with moderate levels of technological advancement. An assessment supports better decisions regarding the need for more ICT technology or capability. We find that developing and advanced industries are more dependent on financing, and this is supported by the literature (cf. Obamuyi et al., 2012; Yeo & Grant, 2017b). We find that tech investments are moderately important to industries with low levels of technological advancement, more important to moderately advanced industries, and much less important to advanced technology industries. Therefore, the level of technological advancement influences financing decisions. Third, our findings can influence company policy decision making, which can influence decisions on whether to invest in ICT or other fixed assets. Fourth, the financial industry benefits from assessing the level of industry technological advancement capability and aid policy decision making investment decisions. This knowledge helps to structure lending policies that maximize returns on assets or investments. Lastly, government policies enable industries with different technological advancement levels to develop and prosper, thus fueling economic growth. Our findings support the argument that policies should be tailored to developing, transitional, or advanced technology industries. We recommend that low tech industries, such as low-end retail, textiles, garments, and other service industries, should consider devoting resources towards securing finances via bank loans to support moderate investments in technology assets, which can improve their business processes and productivity levels. These include the use of websites and email, as well as computer-based accounting and bookkeeping. These investments lay the foundation for long term growth. For industries with moderate levels of technological advancement, including but not restricted to electronics and communications equipment manufacturing, fabricated metal products, and transport, storage and communications, heavy investments in technology assets can further propel them towards stronger growth. For technologically advanced industries, such as chemicals and chemical products manufacturing, motor vehicles and transport equipment manufacturing, as well as IT-related industries including software development and computer manufacturing, continued tech investments may result in increased opportunity costs. Instead, they should invest in financial resources and human capital, so as to fully leverage financial factors and their existing technological capabilities, for increased growth. These industry types are generic and used for discussion purposes. They do not suggest that each type has a fixed level of technological advancement. For example, there are both low tech retail industries comprising small retail shops, as well as high tech ones comprising chain stores.

Effective policies are tailored to their local contexts (Akpan, 2003; Yeo & Trauth, 2009), because institutions behave differently in different contexts (Rodrik, 2000). These can help companies compete locally. Policy makers should consider the level of ICT capability – developing, transition, or advanced – in policy decision making, as a one-size-fits-all approach is unlikely to be effective. We find that financing is important to developing industries, to support long-term tech investments and growth. As industries transition to the advanced stage, the use of ICT and tech investments become less important than bank lending policies (cf. Obamuyi et al., 2012; Yeo & Grant, 2017b). Industries in the initial stages of development, or with business models not reliant on ICT, are better off by acquiring bank financing for tech investments. A minimum level of technological advancement, such as email or websites, is required in today’s competitive environment. Transition industries with moderate ICT use, are dependent on tech investments. It is likely their business models are technology dependent, evident by the relevance of ICT in this cluster. For technologically advanced industries, tech investments are unlikely to reap significant rewards because advanced manufacturing technology or IT, may not improve performance (Ho et al., 2011). Therefore, favorable bank lending policies can support future innovation initiatives, particularly for smaller companies that have more potential to exploit technologies (Im et al., 2001). Consequently, financing is more beneficial to technologically advanced industries. 5.5. Theoretical and industry contributions This study makes seven theoretical contributions and five industry practice contributions. The first theoretical contribution is the multimethod approach seldom used in traditional IS research (Osei-Bryson & Ngwenyama, 2014). It provides insights not possible from traditional regression methods that are not suited for non-normal data distributions or analyzing problematic data. Decision tree induction is better suited to accommodate these limitations and clustering provides additional benefits by identifying patterns within each cluster. Osei-Bryson and Ngwenyama (2014) advocate data mining to improve research investigations. Contrary to past ICT4D studies that use qualitative, interpretive methods (Lin, Kuo, & Myers, 2015), we use data mining for the investigation, thus making a contribution to the literature. The second theoretical contribution stems from the lack of consideration to the embedded ICT contexts, a reason for the fragmented and inconclusive ICT impact results in the literature (cf. Bankole et al., 2015; Bollou, 2006; Devaraj & Kohli, 2000; Grant & Yeo, 2017; Henderson, 2002; Ko & Osei-Bryson, 2004). The TOE framework is used to discuss the organization and environment contexts in which ICT are embedded. We identify three distinct clusters with different levels of technological advancement for more in-depth analysis. Third, Crowston and Myers (2004) recommend more industry research because results of ICT impact are inconsistent (Devaraj & Kohli, 2000), so we perform an industry level analysis. Fourth, applying the TOE framework (Zhu & Kraemer, 2005) to other domains improves its theoretical rigor (Baker, 2012) and helps unify the ICT4D industry level literature (Devaraj & Kohli, 2000). Fifth, Gregor (2006) discusses five types of IS theories – Analyzing, Explaining, Predicting, Explaining and Predicting, and Design and Action. The TOE framework is explanatory as it describes how the three constructs influence technology outcomes (Yeo & Grant, 2017b), but this research illustrates how it can be used in a predictive manner. Its predictive capability corroborates existing evidence that the technology, organization, and environment contexts, play a role in industry performance under different conditions. Sixth, the literature on tech investments is inconclusive due to the technology paradox (Huang et al., 2006), the lack of consideration to ICT contexts (Yeo & Grant, 2018), and the complexity of measuring performance (Mithas & Rust, 2016; Richard et al., 2009). We contextualize the level of technological advancement through clustering and decision tree induction to analyze the impact of tech investments in each cluster. Finally, the most

5.6. Limitations and future research The first limitation is that different countries have various industry types, and each type is not equally represented in each year. Furthermore, there are missing data and imputation is not feasible. Future studies can use a better data set, where available, to improve the 142

International Journal of Information Management 43 (2018) 130–145

D. Grant, B. Yeo

analysis. Second, the ICT variables do not fully represent the complexity of today’s wide array of technologies, such as enterprise systems, Big Data, data analytics, cloud computing, CRM, SCM, social network computing, artificial intelligence, and knowledge-based systems. The use of email and websites are still very relevant to developing countries, so this is not a major issue. However, when data on the US and Canada are included, this may change. There is evidence that the use of email and websites are slightly lower in developing industries than transition industries with moderate levels of technological advancement. Future studies may include more advanced technologies and expand the scope to include developed regions, such as North America. Third, fixed asset investment is not broken down into its components, which prevents the ability to tease out specific tech investment components, such as CAM, CAD, and others. The ability to measure the impact of each component on performance will add value to the analysis. Future studies can include these fixed asset components, but we recognize the difficulty to gather global data. Finally, the specific extent of tech investment importance can be ascertained through further research, using better data.

for Economic Performance. Bollou, F. (2006). ICT infrastructure expansion in sub-Saharan Africa: An analysis of six West African countries from 1995 to 2002. The Electronic Journal of Information Systems in Developing Countries, 26. Botello, H. A., & Pedraza Avella, A. (2014). Las Tecnologías De La Información Y La Comunicación Y El Desempeño De Las Firmas: Evidencia De Las Firmas Industriales Del Ecuador (The information technology and communication and firm performance: Evidence from firms industrialists Ecuador). CPR LATAM-Communication Policy Research Conference. Bradford, M., Earp, J. B., & Grabski, S. (2014). Centralized end-to-end identity and access management and ERP systems: A multi-case analysis using the technology organization environment framework. International Journal of Accounting Information Systems, 15(2), 149–165. Brynjolfsson, E., & Hitt, L. M. (2003). Computing productivity: Firm-level evidence. The Review of Economics and Statistics, 85(4), 793–808. https://doi.org/10.1162/ 003465303772815736. Byrd, T. A., Lewis, B. R., & Bryan, R. W. (2006). The leveraging influence of strategic alignment on IT investment: An empirical examination. Information & Management, 43(3), 308–321. Campbell, M. (2012). What a difference a year makes: Time lag effect of information technology investment on firm performance. Journal of Organizational Computing and Electronic Commerce, 22(3), 237–255. Carr, A. S., & Pearson, J. N. (1999). Strategically managed buyer–supplier relationships and performance outcomes. Journal of Operations Management, 17(5), 497–519. Crowston, K., & Myers, M. D. (2004). Information technology and the transformation of industries: Three research perspectives. The Journal of Strategic Information Systems, 13(1), 5–28. Das, A., & Narasimhan, R. (2001). Process-technology fit and its implications for manufacturing performance. Journal of Operations Management, 19, 521–540. https://doi. org/10.1016/S0272-6963(01)00064-X. Devaraj, S., & Kohli, R. (2000). Information technology payoff in the health-care industry: A longitudinal study. Journal of Management Information Systems, 16(4), 41–67. Díaz-Pérez, F. M., & Bethencourt-Cejas, M. (2016). CHAID algorithm as an appropriate analytical method for tourism market segmentation. Journal of Destination Marketing & Management, 5(3), 275–282. https://doi.org/10.1016/j.jdmm.2016.01.006. Draca, M., Sadun, R., & Van Reenen, J. (2006). Productivity and ICT: A review of the evidence (CEP discussion papers). Centre for Economic Performance, LSE. Eriotis, N. P., Frangouli, Z., & Ventoura-Neokosmides, Z. (2011). Profit margin and capital structure: An empirical relationship. Journal of Applied Business Research (JABR), 18(2). Esposito, F., Malerba, D., Semeraro, G., & Kay, J. (1997). A comparative analysis of methods for pruning decision trees. IEEE Transactions on Pattern Analysis and Machine Intelligence, 19(5), 476–491. Gangwar, H., Date, H., & Ramaswamy, R. (2015). Understanding determinants of cloud computing adoption using an integrated TAM-TOE model. Journal of Enterprise Information Management, 28(1), 107–130. Grant, D., & Yeo, B. (2017). How informality and crime affect industry performance in BRICS. Working paper. Gregor, S. (2006). The nature of theory in information systems. MIS Quarterly, 30(3), 611–642. Hackney, R., Xu, H., & Ranchhod, A. (2006). Evaluating web services: Towards a framework for emergent contexts. European Journal of Operational Research, 173(3), 1161–1174. Hahn, J. (2003). Measuring the effectiveness of e-commerce website design and its impact on business value (Ph.D.). Retrieved fromAnn Arbor: University of Minnesota. http:// ezproxy.depaul.edu/login?url=http://search.proquest.com/docview/305331594? accountid=10477. Henderson, V. (2002). Urbanization in developing countries. The World Bank Research Observer, 17(1), 89–112. Ho, J. L., Wu, A., & Xu, S. X. (2011). Corporate governance and returns on information technology investment: Evidence from an emerging market. Strategic Management Journal, 32(6), 595–623. Hoffman, G. (2011). Bad website = lost business: Save the “creativity” for your personal journal. Reeves Journal, 91(9), 10. Hsu, P.-F., Ray, S., & Li-Hsieh, Y.-Y. (2014). Examining cloud computing adoption intention, pricing mechanism, and deployment model. International Journal of Information Management, 34(4), 474–488. Huang, S.-M., Ou, C.-S., Chen, C.-M., & Lin, B. (2006). An empirical study of relationship between IT investment and firm performance: A resource-based perspective. European Journal of Operational Research, 173(3), 984–999. Im, K. S., Dow, K. E., & Grover, V. (2001). A reexamination of IT investment and the market value of the firm—An event study methodology. Information Systems Research, 12(1), 103–117. Indjikian, R., & Siegel, D. S. (2005). The impact of investment in IT on economic performance: Implications for developing countries. World Development, 33, 681–700. https://doi.org/10.1016/j.worlddev.2005.01.004. Jain, A. K., Murty, M. N., & Flynn, P. J. (1999). Data clustering: A review. ACM Computing Surveys (CSUR), 31(3), 264–323. Jia, Q., Guo, Y., & Barnes, S. J. (2017). Enterprise 2.0 post-adoption: Extending the information system continuance model based on the technology-organization-environment framework. Computers in Human Behavior, 67, 95–105. José Tarí, J. (2005). Components of successful total quality management. The TQM Magazine, 17(2), 182–194. Jung, J. (2009). IT investment decision and performance of online brokerage firm under heterogeneity of IT service dimensions (Master’s Thesis). Retrieved fromKorea: Korea Advanced Institute of Science and Technology. http://library.kaist.ac.kr/search/

References Abdi, T. (2008). Machinery & equipment investment and growth: Evidence from the Canadian manufacturing sector. Applied Economics, 40(4), 465–478. Aggarwal, C. C., & Yu, P. S. (2001). Outlier detection for high dimensional data. ACM sigmod record, Vol. 30, ACM37–46. Akpan, P. I. (2003). Basic-needs to globalization: Are ICTs the missing link? Information Technology for Development, 10(4), 261–274. https://doi.org/10.1002/itdj. 1590100405. Amir, S., Osman, M. M., Bachok, S., & Ibrahim, M. (2015). Understanding domestic and international tourists’ expenditure pattern in Melaka, Malaysia: Result of CHAID analysis. Contemporary Issues in Management and Social Science Research, 172, 390–397. https://doi.org/10.1016/j.sbspro.2015.01.386. Angeles, R. (2013). Using the technology-organization-environment framework and zuboff’s concepts for understanding environmental sustainability and RFID: Two case studies. International Journal of Social, Education, Economics and Management Engineering, 7(11), 1599–1608. Angeles, R. (2014). Using the technology-organization-environment framework for analyzing Nike’s “Considered Index” Green Initiative, a decision support system-driven system. Journal of Management and Sustainability, 4(1), 96. Apostolopoulos, T., & Pramataris, K. (1997). Information technology investment evaluation: Investments in telecommunication infrastructure. International Journal of Information Management, 17(4), 287–296. Archibugi, D., & Coco, A. (2004). A new indicator of technological capabilities for developed and developing countries (ArCo). World Development, 32(4), 629–654. Arias Aranda, D. (2003). Service operations strategy, flexibility and performance in engineering consulting firms. International Journal of Operations & Production Management, 23(11), 1401–1421. Asamoah, G. N. (2011). The impact of the financial sector reforms on savings, investments and growth of gross domestic product (gdp) in Ghana. International Business & Economics Research Journal (IBER), 7(No 10 (2008)), https://doi.org/10.19030/Iber. V7i10.3302. Retrieved from http://www.cluteinstitute.com/ojs/index.php/IBER/ article/view/3302. Audretsch, D. B., van Leeuwen, G., Menkveld, B., & Thurik, R. (2001). Market dynamics in the Netherlands: Competition policy and the role of small firms. Competition Policy in Dynamic Markets, 19(5), 795–821. https://doi.org/10.1016/S0167-7187(00) 00094-1. Baker, J. (2012). The technology–organization–environment framework. In Y. K. Dwivedi, M. R. Wade, & S. L. Schneberger (Vol. Eds.), Information systems theory: Explaining and predicting our digital society: Vol. 1, (pp. 231–245). New York, NY: Springer New York. https://doi.org/10.1007/978-1-4419-6108-2_12. Balijepally, V., Mangalaraj, G., & Iyengar, K. (2011). Are we wielding this hammer correctly? A reflective review of the application of cluster analysis in information systems research. Journal of the Association for Information Systems, 12(5), 375. Bankole, F. O., Osei-Bryson, K.-M., & Brown, I. (2015). The impacts of telecommunications infrastructure and institutional quality on trade efficiency in Africa. Information Technology for Development, 21(1), 29–43. https://doi.org/10.1080/02681102.2013. 874324. Barros, R. C., de Carvalho, A. C. P. L. F., & Freitas, A. A. (2015). Decision-tree induction. In R. C. Barros, A. C. P. L. de Carvalho, & A. A. Freitas (Eds.). Automatic design of decision-tree induction algorithms (pp. 7–45). Cham: Springer International Publishing. https://doi.org/10.1007/978-3-319-14231-9_2. Beck, T., Demirguc-Kunt, A., & Maksimoic, V. (2005). Financial and legal constraints to growth: Does firm size matter? The Journal of Finance, 60(1), 137–177. https://doi. org/10.1111/j.1540-6261.2005.00727.x. Bharadwaj, A. S. (2000). A resource-based perspective on information technology capability and firm performance: An empirical investigation. MIS Quarterly, 24, 169–196. https://doi.org/10.2307/3250983. Bloom, N., Draca, M., Kretschmer, T., Sadun, R., Overman, H., & Schankerman, M. (2010). The economic impact of ICT. London, UK: London School of Economics. Centre

143

International Journal of Information Management 43 (2018) 130–145

D. Grant, B. Yeo

Oliveira, T., Thomas, M., & Espadanal, M. (2014). Assessing the determinants of cloud computing adoption: An analysis of the manufacturing and services sectors. Information & Management, 51(5), 497–510. Osei-Bryson, K.-M. (2004). Evaluation of decision trees: A multi-criteria approach. Computers & Operations Research, 31(11), 1933. Osei-Bryson, K.-M., & Ngwenyama, O. (2014). Advances in research methods for information systems research. New York, USA: Springer. Osei-Bryson, K.-M., & Samoilenko, S. (2014). Overview on cluster analysis. Advances in research methods for information systems research. Springer127–138. Osei-Bryson, K., Dong, L., & Ngwenyama, O. (2008). Exploring managerial factors affecting ERP implementation: An investigation of the Klein-Sorra model using regression splines. Information Systems Journal, 18(5), 499–527. Pakko, M. R. (2002). Investment-specific technology growth: Concepts and recent estimates. Federal Reserve Bank of St. Louis Review37–48 November. Pal, M., & Mather, P. M. (2003). An assessment of the effectiveness of decision tree methods for land cover classification. Remote Sensing of Environment, 86(4), 554–565. Palacios-Marqués, D., Soto-Acosta, P., & Merigó, J. M. (2015). Analyzing the effects of technological, organizational and competition factors on web knowledge exchange in SMEs. Telematics and Informatics, 32(1), 23–32. Park, J.-H., Kim, Y. B., & Kim, M.-K. (2017). Investigating factors influencing the market success or failure of IT services in Korea. International Journal of Information Management, 37(1), 1418–1427. Piget, P., & Kossaï, M. (2013). The relationship between information and communication technology use and firm performance in developing countries: A case study of electrical and electronic goods manufacturing SMEs in Tunisia. African Development Review, 25(3), 330–343. Rai, A., Tang, X., Brown, P., & Keil, M. (2006). Assimilation patterns in the use of electronic procurement innovations: A cluster analysis. Information & Management, 43(3), 336–349. Ramdani, B. (2012). Information technology and organisational performance: Reviewing the business value of IT literature. Information systems theory. Springer283–301. Richard, P. J., Devinney, T. M., Yip, G. S., & Johnson, G. (2009). Measuring organizational performance: Towards methodological best practice. Journal of Management, 35(3), 718–804. Rodríguez, A. H., Avilés-Jurado, F. X., Díaz, E., Schuetz, P., Trefler, S. I., Solé-Violán, J., ... Ribas, A. (2016). Procalcitonin (PCT) levels for ruling-out bacterial coinfection in ICU patients with influenza: A CHAID decision-tree analysis. Journal of Infection, 72(2), 143–151. https://doi.org/10.1016/j.jinf.2015.11.007. Rodrik, D. (2000). Institutions for high quality growth: What they are and how to acquire them. National Bureau of Economic Research, Inc. Retrieved from http://www.nber. org/papers/w7540.pdf. Ryan, S. D., Abitia, G. R., & Windsor, J. C. (2000). Factors affecting the adoption of knowledge management technologies: An international perspective. AMCIS 2000 Proceedings197 Retrieved from http://aisel.aisnet.org/amcis2000/197/. Sabherwal, R., & Jeyaraj, A. (2015). Information technology impacts on firm performance: An extension of Kohli and Devaraj. Mis Quarterly, 39(4), 2003. Samoilenko, S. (2008). Contributing factors to information technology investment utilization in transition economies: An empirical investigation. Information Technology for Development, 14(1), 52–75. Santhanam, R., & Hartono, E. (2003). Issues in Linking Information Technology Capability to Firm Performance. MIS Quarterly, 27, 125–153. Sayeed, M. A., & Hogue, M. S. (2009). Impact of assets and liability management on profitability: A study of public vs private commercial bank in Bangladesh. Retrieved fromhttp://www.wbicconpro.com/30%5B/%5D.pdf. Schonberg, E., Cofino, T., Hoch, R., Podlaseck, M., & Spraragen, S. L. (2000). Measuring success. Communications of the ACM, 43(8), 53–57. Schryen, G. (2013). Revisiting IS business value research: What we already know, what we still need to know, and how we can get there. European Journal of Information Systems, 22(2), 139–169. Sein, M. K., & Harindranath, G. (2004). Conceptualizing the ICT artifact: Toward understanding the role of ICT in national development. Information, Communication and Society, 20(1), 15–24. https://doi.org/10.1080/01972240490269942. Sila, I., & Ebrahimpour, M. (2005). Critical linkages among TQM factors and business results. International Journal of Operations & Production Management, 25(11), 1123–1155. Spring, M., Hughes, A., Mason, K., & McCaffrey, P. (2017). Creating the competitive edge: A new relationship between operations management and industrial policy. Journal of Operations Management, 49–51, 6–19. https://doi.org/10.1016/j.jom.2016.12.003. Tan, B., Ng, E., & Jiang, J. (2018). The process of technology leapfrogging: Case analysis of the national ICT infrastructure development journey of Azerbaijan. International Journal of Information Management, 38(1), 311–316. The World Bank (2015). Enterprise surveys. Retrieved from http://www. enterprisesurveys.org. Torero, M., & Von Braun, J. (2006). Information and communication technologies for development and poverty reduction: The potential of telecommunications. Intl Food Policy Res Inst. Tornatzky, L. G., & Fleischer, M. (1990). The processes of technological innovation. 1990. Lexington, MA: Lexington Books. Ture, M., Tokatli, F., & Kurt, I. (2009). Using Kaplan–Meier analysis together with decision tree methods (C&RT, CHAID, QUEST, C4.5 and ID3) in determining recurrence-free survival of breast cancer patients. Expert Systems with Applications, 36(Part 1), 2017–2026. https://doi.org/10.1016/j.eswa.2007.12.002. Vranakis, S., & Chatzoglou, P. (2011). A conceptual model for machinery & equipment investment decisions. International Journal of Business and Management, 7(1), 36. Wallace, L., Keil, M., & Rai, A. (2004). Understanding software project risk: A cluster analysis. Information & Management, 42(1), 115–125.

detail/view.do?bibCtrlNo=309781&flag=t. Kannan, V. R., & Tan, K. C. (2005). Just in time, total quality management, and supply chain management: Understanding their linkages and impact on business performance. Omega, 33(2), 153–162. Karanasios, S. (2014). Framing ICT4D research using activity theory: A match between the ICT4D field and theory? Information Technologies & International Development, 10(2), 1–18. Kaynak, H. (2003). The relationship between total quality management practices and their effects on firm performance. Journal of Operations Management, 21(4), 405–435. Kim, S. S., Timothy, D. J., & Hwang, J. (2011). Understanding Japanese tourists’ shopping preferences using the decision tree analysis method. Tourism Management, 32(3), 544–554. https://doi.org/10.1016/j.tourman.2010.04.008. Kleis, L., Chwelos, P., Ramirez, R. V., & Cockburn, I. (2012). Information technology and intangible output: The impact of IT investment on innovation productivity. Information Systems Research, 23, 42–59. https://doi.org/10.1287/isre.1100.0338. Ko, M., & Osei-Bryson, K.-M. (2004). Using regression splines to assess the impact of information technology investments on productivity in the health care industry. Information Systems Journal, 14(1), 43–63. Kobayashi, D., Takahashi, O., Arioka, H., Koga, S., & Fukui, T. (2013). A prediction rule for the development of delirium among patients in medical wards: Chi-square automatic interaction detector (CHAID) decision tree analysis model. The American Journal of Geriatric Psychiatry, 21(10), 957–962. https://doi.org/10.1016/j.jagp. 2012.08.009. Koellinger, P. (2006). Impact of ICT on corporate performance, productivity, and employment dynamics (European Commission Enterprise and Industry Directorate General No. 01/ 2006). Berlin. Kohli, R., & Devaraj, S. (2003). Measuring information technology payoff: A meta-analysis of structural variables in firm-level empirical research. Information Systems Research, 14(2), 127–145. Koivunen, M., Hatonen, H., & Valimaki, M. (2008). Barriers and facilitators influencing the implementation of an interactive internet-portal application for patient education in psychiatric hospitals. Patient Education and Counseling, 70(3), 412–419. https://doi. org/10.1016/j.pec.2007.11.002. Kossaï, M., & Piget, P. (2014). Adoption of information and communication technology and firm profitability: Empirical evidence from Tunisian SMEs. The Journal of High Technology Management Research, 25(1), 9–20. Kwon, S. O. (2007). Impact of IT investment on firm performance: Focusing on four moderators. (Master’s Thesis). Korea: Korea Advanced Institute of Science and Technology. Lee, H., Choi, H., Lee, J., Min, J., & Lee, H. (2016). Impact of IT investment on firm performance based on technology IT architecture. Procedia Computer Science, 91, 652–661. Legohérel, P., Hsu, C. H. C., & Daucé, B. (2015). Variety-seeking: Using the CHAID segmentation approach in analyzing the international traveler market. Tourism Management, 46, 359–366. https://doi.org/10.1016/j.tourman.2014.07.011. Levine, R. (1997). Financial development and economic growth: Views and agenda. Journal of Economic Literature, 35(2), 688–726. Liao, Y.-W., Wang, Y.-M., Wang, Y.-S., & Tu, Y.-M. (2015). Understanding the dynamics between organizational IT investment strategy and market performance: A system dynamics approach. Computers in Industry, 71, 46–57. Libanio, G., & Moro, S. (2006). Manufacturing industry and economic growth in Latin America: A Kaldorian approach. Lin, C. I., Kuo, F.-Y., & Myers, M. D. (2015). Extending ICT4D studies: The value of critical research. MIS Quarterly, 39(3), 697–712. Lind, P., Sepúlveda, E., & Nuñez, J. (2000). On the applicability of a computer model for business performance analysis in SMEs: A case study from Chile. Information Technology for Development, 9(1), 33–44. https://doi.org/10.1080/02681102.2000. 9525319. Liu, X., Yeung, A. C. L., Lo, C. K. Y., & Cheng, T. C. E. (2014). The moderating effects of knowledge characteristics of firms on the financial value of innovative technology products. Journal of Operations Management, 32, 79–87. https://doi.org/10.1016/j. jom.2013.11.003. Lodder, P. (2013). To impute or not impute: That’s the question. Advising on research methods: Selected topics. Huizen: Johannes van Kessel Publishing. Marks, H. M., & Printy, S. M. (2003). Principal leadership and school performance: An integration of transformational and instructional leadership. Educational Administration Quarterly, 39(3), 370–397. Meliciani, V. (2000). The relationship between R&D, investment and patents: A panel data analysis. Applied Economics, 32(11), 1429–1437. Mithas, S., & Rust, R. T. (2016). How information technology strategy and investments influence firm performance: Conjecture and empirical evidence. MIS Quarterly, 40(1), 223–245. Motiwalla, L., Khan, M. R., & Xu, S. (2005). An intra-and inter-industry analysis of ebusiness effectiveness. Information & Management, 42(5), 651–667. Murphy, E. L., & Comiskey, C. M. (2013). Using chi-squared automatic interaction detection (CHAID) modelling to identify groups of methadone treatment clients experiencing significantly poorer treatment outcomes. Journal of Substance Abuse Treatment, 45(4), 343–349. https://doi.org/10.1016/j.jsat.2013.05.003. Obamuyi, T. M., Edun, A. T., & Kayode, O. F. (2012). Bank lending, economic growth and the performance of the manufacturing sector in Nigeria. European Scientific Journal, 8(3), 19–36. OECD (2008). Measuring the impacts of ICT using official statistics. Paris, France. Okazaki, S. (2006). What do we know about mobile internet adopters? A cluster analysis. Information & Management, 43(2), 127–141. Olatunji, T. E., & Adegbite, T. A. (2014). Investment in fixed assets and firm profitability: Empirical evidence from the Nigerian banking sector. Asian Journal of Social Sciences and Management Studies, 1(3), 78–82.

144

International Journal of Information Management 43 (2018) 130–145

D. Grant, B. Yeo

technology (ICT) use among PLHIV in China: A promising but underutilized venue for HIV prevention and care. International Journal of Information Management, 38(1), 27–33. Zhu, K., & Kraemer, K. L. (2005). Post-adoption variations in usage and value of e-business by organizations: Cross-country evidence from the retail industry. Information Systems Research, 16(1), 61–84. Zhu, K., Kraemer, K. L., & Dedrick, J. (2004). Information technology payoff in e-business environments: An international perspective on value creation of e-business in the financial services industry. Journal of Management Information Systems, 21(1), 17–54. Zhu, K., Kraemer, K. L., & Xu, S. (2006). The process of innovation assimilation by firms in different countries: A technology diffusion perspective on e-business. Management Science, 52(10), 1557–1576. Zhu, K., Kraemer, K., & Xu, S. (2003). Electronic business adoption by European firms: A cross-country assessment of the facilitators and inhibitors. European Journal of Information Systems, 12(4), 251–268.

Weill, P., & Ross, J. W. (2004). IT governance: How top performers manage IT decision rights for superior results. Harvard Business Press. World Bank (2016). World development report 2016: Digital dividendsWashington DC. Yeo, B., & Grant, D. (2018). Predicting service industry performance using decision tree analysis. International Journal of Information Management, 38(1), 288–300. Yeo, B., & Grant, D. (2017a). Exploring the effects of ICTs, workforce, and gender on capacity utilization. Journal of Information Technology for Development. https://doi. org/10.1080/02681102.2017.1383876. Yeo, B., & Grant, D. (2017b). Exploring the factors affecting global manufacturing performance. Journal of Information Technology for Development. https://doi.org/10. 1080/02681102.2017.1315354. Yeo, B., & Trauth, E. (2009). The call for transformational governance in the knowledge economy. In V. Weerakkody, M. Janssen, & Y. DWivedi (Eds.). Handbook of research on ICT-enabled transformational government: A global perspective. (pp. 271–290). U.S.A. U.K: IGI Global. Zhang, Y., Li, X., Qiao, S., Zhou, Y., & Shen, Z. (2018). Information communication

145