ICT, innovation, and firm productivity: New evidence

JBR-08287; No of Pages 6 Journal of Business Research xxx (2015) xxx–xxx

Contents lists available at ScienceDirect

Journal of Business Research

ICT, innovation, and firm productivity: New evidence from small local firms☆ Ángel Díaz-Chao a,⁎, Jorge Sainz-González a, Joan Torrent-Sellens b a b

Applied Economics Department, Rey Juan Carlos University, Paseo de los Artilleros, s.n., Madrid 28032, Spain Open University of Catalonia, Rambla Catalunya, 6, Barcelona 08007, Spain

a r t i c l e

i n f o

Article history: Received February 2014 Received in revised form November 2014 Accepted January 2015 Available online xxxx Keywords: Information and communication technologies (ICTs) Co-innovation Firm productivity Small local firms Structural equation modeling (SEM) Total factor productivity (TFP)

a b s t r a c t This study analyzes new co-innovative sources of labor productivity (i.e., ICT use, human capital and training, and new forms of work organization) in small firms that produce for local markets. The study presents an application of structural equation modeling (SEM) to 2009 survey data for a representative sample of 464 SMEs in the province of Girona (Spain). Results show that wage is the main determinant of labor productivity. Furthermore, in contrast to evidence regarding larger firms, co-innovation does not directly affect small local firms' productivity. The study establishes an indirect relationship between co-innovation and productivity in firms that initiate international expansion. The study also identifies guidelines for public policy to improve productivity in small local firms. © 2015 Elsevier Inc. All rights reserved.

1. Introduction: ICT, innovation, and firm productivity in the scientific literature The widespread use of information and communication technologies (ICT) is crucial for economic activity (Jorgenson & Vu, 2007) for two reasons. First, ICTs directly increases productivity and boost economic growth (Jorgenson, Ho, & Stiroh, 2008). Second, ICTs generate complementary innovations that improve economies' total factor productivity (TFP) (Ceccobelli, Gitto, & Mancuso, 2012; Jorgenson, Ho, & Samuels, 2011). Empirical analysis of ICT's effect on firm productivity shows that return rates on digital investment are higher than return rates on physical investment. The reason for this difference is that digital investment and use often occur alongside other endeavors, namely, human capital improvement and changes in organizational structure (Arvanitis, 2005; Bresnahan, Brynjolfsson, & Hitt, 2002; Kunz, Schmitt, & Meyer, 2011). The transformative effect of ICT investment and use on business performance becomes more apparent when firms simultaneously engage in co-innovation processes (Brynjolfsson & Hitt, 2003; Cardona, Kretschmer, & Strobel, 2013; Greenan, L'Horty, & Mairesse, 2002). ☆ The authors are grateful to contributions from participants in the Multidisciplinary Seminar at the IN3 (Barcelona, Spain) and in the Applied Economics Seminar at the URJC and ESIC (Madrid, Spain). The authors give special thanks to Elena Gonzalez from ESIC and Pilar Grau from Rey Juan Carlos University for carefully reading this study and providing valuable suggestions. ⁎ Corresponding author. E-mail addresses: [email protected] (Á. Díaz-Chao), [email protected] (J. Sainz-González), [email protected] (J. Torrent-Sellens).

ICT investment and use improve general productivity only when firms and workers achieve the necessary technological, educational and training, organizational, business, labor, and cultural competencies. In other words, organizational and business process changes enable firms to benefit from ICT's full potential as a general-purpose technology (Arvanitis & Loukis, 2009; Timmer, Inklaar, O'Mahoney, & Van Ark, 2010). New evidence shows the existence of co-innovative productivity sources among broad samples of firms in the United States (Atrostic & Nguyen, 2005; Black & Lynch, 2001, 2004; Bresnahan et al., 2002; Brynjolfsson & Hitt, 2003) and the rest of the world (Cardona et al., 2013; Draca, Sadun, & Van Reenen, 2007; Jiménez-Rodríguez, 2012; Matteucci, O'Mahoney, Robinson, & Zwick, 2005; Torrent & Díaz-Chao, 2014). However, scarce evidence is available on co-innovative productivity sources for small and medium enterprises (SMEs) (Audretsch, 2002, 2006; Hall, Lotti, & Mairesse, 2009; Wymenga, Spanikova, Barker, Konings, & Canton, 2012). Evidence is particularly meager in the case of SMEs that produce primarily for local markets (Díaz-Chao et al., 2013; Torrent & Díaz-Chao, 2014). Such SMEs have low degrees of openness and innovation (Drechsler & Natter, 2012). This study bridges the research gap by examining data from a representative sample of 464 small local firms in the region of Girona (Spain). Structural equation modeling (SEM) exploits these data. SEM is capable of analyzing relationships not only between productivity explanatory factors but also among such factors. Thus, the analysis investigates the structural form explaining firm productivity and provides new findings in co-innovative productivity sources for small local firms.

http://dx.doi.org/10.1016/j.jbusres.2015.01.030 0148-2963/© 2015 Elsevier Inc. All rights reserved.

Please cite this article as: Díaz-Chao, Á., et al., ICT, innovation, and firm productivity: New evidence from small local firms, Journal of Business Research (2015), http://dx.doi.org/10.1016/j.jbusres.2015.01.030

2

Á. Díaz-Chao et al. / Journal of Business Research xxx (2015) xxx–xxx

Section 2 describes the data and the research design. Section 3 presents the model and the results. Section 4 discusses conclusions and implications for innovation policy. 2. Data and research design The study uses survey data from a sample of 464 firms operating in Girona (overall margin of error of ±4.6% in the case of maximum indetermination, p = q = 50, for a confidence level of 95.5%). The sample universe comprised 66,682 firms operating in Girona in 2009. The study took a random sample to achieve a margin of error of less than ±5%. A 47-question pilot questionnaire gathered data from 30 firm managers with an overall view of their companies' activities. Managers responded to pilot questionnaire items in 1-hour face-to-face interviews. By gathering data on the value chain, the study analyzed productivity sources in Girona-based firms. The fieldwork took place between June and October 2009—a period during which the current financial crisis was affecting innovation (Hausman & Johnston, 2014) and the sector in general. The Girona Observatory on ICTs, the Girona Association of New Technology Firms, and the Chamber of Commerce of Girona supported the research. In the region of Girona, small local firms account for the majority of economic activity. These firms' structure is similar in most areas in Spain and may well be representative of Spain's SME sector. The sectors where these firms operate make low-intensity use of technology (food, metal and construction, trade, and tourism). The firms have low levels of worker training, unexploited ICT use, and productivity problems (Torrent & Díaz-Chao, 2014). Table 1 shows descriptive statistics illustrating the value generation process in the sample. 3. Direct and indirect sources of small local firms' productivity 3.1. Modeling small local firms' productivity The research uses structural equation modeling with measurement to test how the presence of co-innovation explains Girona-based firms' productivity. Structural equation systems are formal mathematical models. They consist of a set of linear equations that encompasses Table 1 Descriptive statistics of Girona-based firms. Valid % Business sector: - Manufacturing and construction - Wholesale and retail trade - Hotels, restaurants and tourism - Other market services

Firm innovation: 26.2

- R&D department in firm

20.4 21.6 31.8

- Source of innovation: staff 86.0 - Innovation in last two years 26.5 - Product innovation 50.4

Size of firm: - Fewer than 10 employees - From 11 to 49 employees - 50 or more employees

- No ICT use - Low ICT use - Medium ICT use - High ICT use

21.1 33.1 26.7 19.1

Average turnover (thousands of €): 89.7 10.3

Worker training: - Untrained or primary education - Secondary education - University education - Extended education paid by the firm

8.8

ICT use in value chain: 95.4 4.0 0.6

Firm ownership: - Family firm - Business group

Valid %

- 2008 - 2009

- Girona and rest of Catalonia - Spain - European Union - Rest of the world

η ¼ α þ Вη þ Г ξ þ ζ

183.5 165.0

94.7 2.5 2.5 0.3

ð1Þ

where η (m × 1) and ξ (n × 1) are random vectors of latent dependent and independent variables; α (m × 1) is a vector representing the intersections of axes; В (m × m) is the matrix of coefficients of endogenous latent variables representing the effects of variables η on other variables η; Г (m × n) is the matrix of coefficients of exogenous latent variables representing the direct effects of variables ξ on variables ξ; and ζ is a vector (m × 1), indicating the random perturbations in the equation. According to assumptions in the SEM model, E(η) = 0, E(ξ) = 0, and E(ζ) = 0. The vectors y (p × 1) and x (q × 1) represent the observed (measurable) variables, where p is the number of indicators of η and q is the number of indicators of ξ. The following equations relate x and y to the latent variables: y ¼ τy þ Λy η þ ε

ð2Þ

x ¼ τx þ Λx ξ þ δ

ð3Þ

where ε (p × 1) and δ (q × 1) are the vectors of the error terms. In this model, the assumption is that ε does not correlate with η, ξ, or δ and that δ does not correlate with η, ξ, or ε. Λy (p × m) and Λx (q × n) are matrices containing the structural coefficients λij that relate the latent and observed (measurable) variables; τy (p × 1) and τx (q × 1) are the vectors of constant intersection terms. The fundamental hypothesis of structural equation systems is Σ = Σ(θ), where Σ is the population covariance matrix and Σ(θ) is the model covariance matrix, written as a function of a parameter vector of θ. Minimizing the following function of adjustment obtains the estimation of parameters: F ðθÞ ¼ F ½S; ΣðθÞ

Destination of sales: 32.8 47.3 19.9 8.1

various model types (i.e., regression models, simultaneous equation systems, factor analysis, and path analysis). The equation system's variables can be either directly observable measurable variables or latent (theoretical) variables representing unobservable concepts. While latent variables are continuous, observable dependent variables can be continuous, censored, binary, ordered, categorical (ordinals), or combinations of these variable types. The general SEM model comprises two sub-models: a structural model that relates latent variables to each other and a measurement model that relates each latent variable to the respective variables measuring the model. Scholars generally use the term indicators to refer to these variables measuring the model. In this model, the basic assumption is that a causal structure between latent variables usually exists. SEM has distinctive features that make SEM a suitable analysis tool in the current study: (1) SEM admits the explicit inclusion of measurement error in the estimation process for as many variables as necessary; (2) SEM admits simultaneous estimation of the parameters of a series of dependence relationships, whereby a variable can act as dependent in some equations and independent in others; (3) SEM can show reciprocal causes and recursive and non-recursive models; and (4) SEM is also suitable for prospective analysis with additional out-of-the-sample data. Consistent with the most common notation among scholars (Jöreskog & Sörbom, 2004), the following system of linear structural equations formally defines SEM models:

ð4Þ

After estimating the model's parameters, the next step is to compare the resulting covariance matrix to the data covariance matrix. If the difference between the two matrices is statistically acceptable or zero, the proposed SEM model represents a plausible explanation of the reality. The application of this analysis method to productivity sources in the sample yields (1) a more complete explanatory model using multiple equations and (2) specific measurement errors for each variable. By doing so, the process will eliminate any potential problems that



specification errors may cause. Consequently, the estimated parameters will be unbiased, consistent, and of lower variance. The general analysis model includes 13 hypotheses. The dependent variable is labor productivity (LP) in Girona-based firms, approximated by the logarithm of turnover divided by the number of full-time equivalent workers. This definition of LP as the dependent variable is similar to Hall et al. (2009) definition. Direct data on firm turnover provide the numerator of this ratio. This study considers full-time and part-time jobs in the firms and expresses the number of workers as full-time equivalents to construct the denominator. The direct explanatory factors of labor productivity in Girona-based firms are the logarithm of wage per full-time equivalent worker (WAGE) and the firms' export capacity (sales quota) to extra-EU markets (i.e., exports to the rest of the world, or XRW). Thus, two of the hypotheses in the international empirical literature (Section 1) apply to Girona-based firms: H1. A higher wage per worker means higher productivity. H2. Higher export intensity (i.e., capacity to export goods and services to distant overseas markets) means higher productivity. H1 implies that a firm's capacity to increase turnover per worker relies on finding better-paid and probably higher-quality labor. H2 implies that firms increase their capacity turnover per worker through economies of learning, scale, reach, and scope, which firms achieve by growing their export intensity. In addition to including hypotheses relating to direct factors of productivity, the analysis model includes a set of hypotheses relating to indirect factors and their interrelationships. Specifically, the model includes the indirect causal relationship for the capacity of Gironabased firms to export: H3. The capacity of Girona-based firms to export to extra-EU markets depends on variable forms of worker remuneration (VAREM). H4. The capacity of Girona-based firms to export to extra-EU markets depends on variable forms of worker ICT use (ICTU). VAREM indicates the presence or absence of variable forms of remuneration by objective in a firm (0 = absence; 1 = presence). ICTU indicates the intensity of ICT use in the value generation process in firms. Six value elements exist in firms: (1) accounting, finance, and taxation; (2) administration and human resources; (3) procurement; (4) production; (5) turnover and distribution; and (6) management. Thus, the indicator ICT takes four values: 1 = no ICT use (i.e., when a firm does not use ICT in any value element); 2 = low ICT use (i.e., when a firm uses ICT in one or two of the six value elements); 3 = medium ICT use (i.e., when a firm uses ICT in three or four of the six value elements); and 4 = high ICT use (i.e., when a firm uses ICT in five or six of the value elements). H3 posits that the existence of incentives in worker remuneration affects a firm's capacity to export because workers in that firm probably have greater motivation and commitment to the value generation process. H4 posits a causal relationship between the intensity of ICT use and a greater capacity to export to more distant overseas markets. In the model, the assumption is that intensive ICT use makes a firm's value generation process more competitive, which improves the firm's export capacity. The next set of hypotheses derives from this assumption: H5. The intensity of ICT use affects a firm's capacity to implement variable forms of remuneration. Digital intensity is a lever of change for a firm. Digital intensity leads to new practices of variable remuneration by objective and new practices of human resources management.

3

H6. A dependence relationship exists between human capital and training and the presence of practices of variable remuneration of workers (HCT). The categorical variable HCT takes three values: 1 = when workers' mean training stock falls into the category of untrained or primary education; 2 = when workers' mean training stock falls into the category of secondary education; 3 = when workers' mean training stock falls into the category of university education. Higher training levels among workers favor greater presence of variable remuneration forms. Consistent with the empirical findings, the presence of more workers who are highly qualified does not correspond to more traditional and less flexible organizational and worker-remuneration practices. To improve the value generation process, firms need more workers who are highly qualified, to retain and commit to talent and human capital and to develop new, more flexible remuneration forms that better suit qualified workers' needs. H7. The presence of full-time jobs in a firm negatively affects variable forms of remuneration. In contrast to H6, H7 posits a negative causal relationship between a new and an old work-organization practice. The variable (FTIMEJOB) shows the presence of full time jobs in a firm, taking the value 0 = when absent or 1 = when present. Variable forms of worker remuneration, which full-time jobs do not encourage, presumably affect flexible remuneration practices, which depend on incentives. Finally, and consistent with the previous argument: H8. A negative causal relationship exists between full-time jobs and human capital and training. Once again, the assumption in this model is that the presence of relatively inflexible forms of work inherent in full-time jobs does not promote capturing and retaining human capital in a firm. ICT use also has causal relationships with ICT capital endowment and the presence of innovative practices in a firm. The ICT capital variable ICTK shows the investment of a firm in ICT-related goods and services. The variable INNOV shows a firm's innovative dynamics and takes two values: 0 = when a firm has not implemented any innovation in the last two years; and 1 = when a firm has implemented some type of innovation in the last two years. H9. A causal relationship exists between ICT investment and use. A greater ICT investment leads to a greater ICT use. H10. A causal relationship exists between a firm's innovative dynamics and ICT use. In firms with innovative practices, ICT use needs to increase. For this dimension, the model also includes a causal relationship between ICT capital and innovative practices. H11 posits the synergistic effects of digital investment beyond digital use: H11. ICT investment also explains non-digital innovative practices. Finally, because of the sample's specificity, the model includes two further hypotheses outside the scope of new worker-remuneration practices and ICT. Both hypotheses refer to the explanation of innovation. The earlier review of the descriptive statistics shows that only 26.5% of the firms are innovative. Innovation is informal and involves little collaboration. The basic innovation source in Girona-based firms (86% of cases) is the firm's own workers, who are not explicitly working on research, development, and innovation (R&D and innovation) tasks. H12. a negative causal relationship exists between wage per worker and innovation practices in a firm.


4


This hypothesis posits that innovation is informal and suggests that a firm's mean remuneration negatively affects innovation development because that remuneration does not account for the innovative work. H13. the establishment of variable remuneration-by-objective practices (e.g., innovative work objectives) positively affects a firm's innovation development. The following system of equations with observed variables and measurement errors (Eq. (5)) represents the SEM model, which shows direct and indirect factors that explain labor productivity in Gironabased firms. Eq. (5) formalizes and simplifies the earlier formulation of the SEM model. Eq. (1) relates to the direct explanatory factors of labor productivity (LP) and captures H1 and H2. Eq. (2) relates to the indirect explanatory factors of the capacity to export to extra-EU markets and captures H3 and H4. Eq. (3) relates to the indirect explanatory factors of the presence of variable forms of worker remuneration by objective (VAREM) and captures H5, H6, and H7. Eq. (4) relates to the indirect explanatory factors of human capital and training (HCT) and captures H8. Eq. (5) relates to the indirect explanatory factors of ICT use (ICTU) and captures H9 and H10. Eq. (5) also relates to the indirect explanatory factors of innovation and captures H11, H12, and H13. 0

LP

1

0

β 10

1

0

B XRW C B β C B B C B 20 C B B C B C B B VAREM C B β 30 C B B C B C B B HCT C ¼ B β C þ B B C B 40 C B B C B C B @ ICTU A @ β 50 A @ INNOV

β 11

β12

0

0

0

0

0

0 0

0 0

β 23 0

β24 β34

0 β35

0 β36

0 0

0

0

0

0

0

β46

0

0 β 61 1

0 0

0 β 62

0 0

0 0

0 0

β57 β67

β 60 0 WAGES B XRW C 0 ε 1 1 B C B C B VAREM C B ε2 C C B C B C B ICTU C B ε3 C B C B B B CþB C B HCT C B ε 4 C C B C C B FTIMEJOB C B B C @ ε5 A B C @ ICTK A ε6

0

1

C C C C C 0 C C C β58 A 0 0

0

ð5Þ

INNOV

3.2. Results of SEM estimation Tables 2 and 3 show the results of the SEM model (Eq. (5)) for the productivity of Girona-based firms. The fit indices that illustrate the model's goodness of fit (NIF = 0.94; RFI = 0.87; IFI = 0.97; TLI = 0.94; CFI = 0.97) yield values close to 1. In addition, the value of CMIN/DF equals 1.77 (i.e., b2) and the value of RMSEA equals 0.041, which is less than 0.05. These values suggest that the model has acceptable reliability, validity, and unidimensionality, thereby confirming the constructs' reliability and scale validity.

Table 3 SEM estimation of labor productivity in Girona-based firms⁎. Explained Explanatory Coefficients Standardized Standard variable variable coefficients error LP LP XRM XRM VAREM VAREM VAREM HCT ICTU ICTU INNOV INNOV INNOV

WAGES XRW VAREM ICTU ICTU HCT FTIMEJOB FTIMEJOB ICTK INNOV ICTK WAGES VAREM

0.00 0.82 0.07 0.90 41828.30 3.01 −0.06 −0.01 0.47 0.13 0.21 0.00 0.01

0.92 0.06 0.15 0.13 0.26 0.14 −0.10 −0.12 0.43 0.14 0.25 −0.14 0.10

Critical P-value ratio

0.00 33.87 0.36 2.26 0.02 3.16 0.33 2.76 11364.50 3.68 1.10 2.75 0.03 −2.09 0.00 −2.52 0.05 10.25 0.04 3.03 0.04 5.31 0.00 −2.01 0.00 2.00

0.000 0.024 0.002 0.006 0.000 0.006 0.036 0.012 0.000 0.002 0.000 0.045 0.045

⁎ SEM model with observed variables and measurement errors.

Multiple fit indices check the measurement scale's goodness of fit to the data. These indices evaluate how effectively the data fit the model considering the following: (1) absolute model fit, (2) incremental model comparison, and (3) model parsimony (Hooper, Coughlan, & Mullen, 2009). Overall, evidence implies that the results represent the hypothesized constructs. The resulting standardized coefficients are significant at a maximum 95% confidence level, and the values are consistent with the hypotheses. The main direct determinant of labor productivity in Girona-based firms is mean wage (H1: β = 0.92, p = 0.00). The second, much weaker determinant is exports to extra-EU markets (H2: β = 0.06, p = 0.02). Export intensity directly affects labor productivity in Girona-based firms. Therefore, the set of indirect effects in Fig. 1 are valid. In fact, results reveal a set of indirect determinants of labor productivity in Girona-based firms. Grouping these determinants into two dimensions of co-innovation according to export capacity is possible. The two dimensions are (1) the organizational, labor, and human capital dimension and (2) the ICT and innovation dimension. Regarding dimension (1), the effect of flexible remuneration practices on the capacity to export establishes the indirect effect on productivity (H3: β = 0.15, p = 0.02). ICT use (H5: β = 0.26, p = 0.00), human capital and training (H6: β = 0.14, p = 0.01), and (negatively) the presence of full-time jobs (H7: β = −0.10, p = 0.04) affect these flexible remuneration practices. The negative indirect effect of full-time jobs on human capital and training (H8: β = −0.12, p = 0.01) completes this first organizational, labor, and human capital dimension. Regarding dimension (2), ICT capital (H9: β = 0.43, p = 0.00) and innovation (H10: β = 0.14, p = 0.00) affect ICT use. The effect of ICT use on the capacity to export (H4: β = 0.13, p = 0.01) explains the indirect effect on labor productivity. The indirect effect that ICT capital (H11: β = 0.25, p = 0.00), mean wage (H12: β = − 0.14, p = 0.05), and presence of flexible forms of worker remuneration (H13: β = 0.10, p = 0.05) exert on innovation complete this second ICT and innovation dimension. 4. Conclusions and policy implications

Table 2 Fit indices for the structural model of labor productivity in Girona-based firms. Statistics (goodness of fit)* Index

Value

Good fit

Very good fit

NFI RFI IFI TLI CFI CMIN/DF (χ2/df) RMSEA

0.94 0.87 0.97 0.94 0.97 1.77 0.04

N.8 N.8 N.8 N.8 N.8 [1–2] b.08

N.9 N.9 N.9 N.9 N.9 ~1 b.04

Recent international empirical evidence demonstrates the existence of new co-innovative sources of firm productivity. These productivity sources stem from relationships of complementarity (co-innovation) between ICT investment and use, new forms of work organization and labor relations, and human capital and training. Using 2009 survey data for a representative sample of Girona-based firms, this study analyzes determinants of firms' labor productivity. By developing and testing a structural equations model with observed variables and error measurement, this study yields new evidence on co-innovative sources of productivity in a specific business group. Within this business group, most firms are small (95.4% with fewer



5

Fig. 1. Standardized estimated coefficients and path analysis.

than 10 workers) and sell to local markets (94.7% of turnover from local markets). Two important conclusions emerge from the analysis. First, for small local firms, wage is the main determinant of labor productivity. In this respect, productivity has an association with labor quality. Second, unlike most international evidence, which generally focuses on larger firms, co-innovation does not directly affect small local firms' productivity. The capacity to export to extra-UE markets indirectly affects the causal relationship between co-innovation and productivity. Results suggest that for small local firms to improve their productivity, they require two types of public policy. First, public policies should jointly promote ICT use, organizational change, and training of employers and workers. Incomplete public policies designed to promote ICT use without considering the other determinants of co-innovative productivity may not achieve their targets. Public policies should strengthen the link between productivity and internationalization. Public policies should promote the internationalization of products and services that small firms produce for local markets because international competition introduces efficiency enhancement mechanisms. This study has certain limitations. Besides the variables and restrictions on analysis, perhaps the most significant limitation is the unavailability of a time series. However, the survey data on a representative sample of small local firms permits analysis of the determinants of those firms' growth potential. Hence, considering the economic importance of small local businesses, researchers should consider new approaches regarding data availability for other territories or business groups. Researchers should also explore time series, better indicators, and new criteria for grouping firms. References Arvanitis, S. (2005). Computerization, workplace organization, skilled labour and firm productivity: Evidence for the Swiss business sector. Economics of Innovation and New Technology, 14(4), 225–249.

Arvanitis, S., & Loukis, E.N. (2009). Information and communication technologies, human capital, workplace organization and labour productivity: A comparative study based on firm-level data for Greece and Switzerland. Information Economics and Policy, 21(1), 43–61. Atrostic, B.K., & Nguyen, S.V. (2005). IT and productivity in US manufacturing: Do computers networks matter? Economic Inquiry, 43(3), 493–506. Audretsch, D.B. (2002). The dynamic role of small firms: Evidence from the U.S. Small Business Economics, 18(1–3), 13–40. Audretsch, D.B. (2006). SMEs in the age of globalization. Cheltenham and Northampton, MA: Edward Elgar. Black, S.E., & Lynch, L.M. (2001). How to compete: The impact of workplace practices and information technology on productivity. Review of Economics and Statistics, 83(3), 434–445. Black, S.E., & Lynch, L.M. (2004). What's driving the new economy: The benefits of workplace innovation. Economic Journal, 114(493), 97–116. Bresnahan, T.F., Brynjolfsson, E., & Hitt, L.M. (2002). Information technology, workplace organization and the demand for skilled labor: A firm-level evidence. Quarterly Journal of Economics, 117(1), 339–376. Brynjolfsson, E., & Hitt, L.M. (2003). Computing productivity: Firm-level evidence. Review of Economics and Statistics, 85(4), 793–808. Cardona, M., Kretschmer, T., & Strobel, T. (2013). ICT and productivity: Conclusions from the empirical literature. Information Economics and Policy, 25(3), 109–125. Ceccobelli, M., Gitto, S., & Mancuso, P. (2012). ICT capital and labour productivity growth: A non-parametric analysis of 14 OECD countries. Telecommunications Policy, 36(4), 282–292. Díaz-Chao, A., Ficapal, P., & Torrent, J. (2013). ICT, innovation, wages and labour productivity. New evidence from small local firms. Revista de Estudios Empresariales, 2(2), 29–45. Draca, M., Sadun, R., & Van Reenen, J. (2007). Productivity and ICT: A review of the evidence. In R. Mansell, C. Avgerou, D. Quah, & R. Silverstone (Eds.), The Oxford handbook of information and communication technologies (pp. 100–147). Oxford and New York, NY: Oxford University Press. Drechsler, W., & Natter, M. (2012). Understanding a firm's openness decisions in innovation. Journal of Business Research, 65(3), 438–445. Greenan, N., L'Horty, Y., & Mairesse, J. (2002). Productivity, inequality, and the digital economy. A transatlantic perspective. Cambridge, MA: MIT Press. Hall, B.H., Lotti, F., & Mairesse, J. (2009). Innovation and productivity in SMEs: Empirical evidence for Italy. Small Business Economics, 33(1), 13–33. Hausman, A., & Johnston, W.J. (2014). The role of innovation in driving the economy: Lessons from the global financial crisis. Journal of Business Research, 67(1), 2720–2726. Hooper, D., Coughlan, J., & Mullen, M. (2009). Structural equation modelling: Guidelines for determining model fit. Electronic Journal of Business Research Methods, 6(1), 53–60.


6


Jiménez-Rodríguez, S. (2012). Evaluating the effects of investment in information and communication technology. Economics of Innovation and New Technology, 21(2), 203–221. Jöreskog, K.G., & Sörbom, D. (2004). LISREL 8.8 for Windows. Lincolnwood, IL: Scientific Software International. Jorgenson, D.W., Ho, M.S., & Samuels, J.D. (2011). Information technology and U.S. productivity growth: Evidence from a prototype industry production account. Journal of Productivity Analysis, 36(2), 159–175. Jorgenson, D.W., Ho, M.S., & Stiroh, K.J. (2008). A retrospective look at the US productivity growth resurgence. Journal of Economics Perspectives, 22(1), 3–24. Jorgenson, D.W., & Vu, K. (2007). Information technology and the World growth resurgence. German Economic Review, 8(2), 122–145. Kunz, W., Schmitt, B., & Meyer, A. (2011). How does perceived firm innovativeness affect the consumer. Journal of Business Research, 64(8), 816–823.

Matteucci, N., O'Mahoney, M., Robinson, C., & Zwick, T. (2005). Productivity workplace performance and ICT: Evidence from Europe and the US. Scottish Journal of Political Economy, 52(3), 359–386. Timmer, M.P., Inklaar, R., O'Mahoney, M., & Van Ark, B. (2010). Economic growth in Europe. A comparative industry perspective. Cambridge, MA: Cambridge University Press. Torrent, J., & Díaz-Chao, A. (2014). ICT uses, innovation and SMEs productivity: Modeling direct and indirect effects in small local firms. IN3 Working Paper Series WP14-001 (Retrieved from http://journals.uoc.edu/index.php/in3-working-paper-series/ article/view/n-14-torrent-diaz/n14-torrent-diaz-en). Wymenga, P., Spanikova, V., Barker, A., Konings, J., & Canton, E. (2012). EU SMEs in 2012: At the crossroads. Annual report on small and medium-sized enterprises in the EU, 2011/12. Brussels: European Commission.
