Work organization, innovation, and firm performance

Mariana Sapage Madeira Guerreiro Work organization, innovation, and firm performance Abril 2022 Mariana Sapage Madeira Guerreiro Work organization, innovation, and firm performance Master Dissertation Master in Economics Dissertation realized under the supervision of Professora Doutora Ana Paula Rodrigues Pereira de Faria Abril de 2022 ii DIREITOS DE AUTOR E CONDIES DE UTILIZAO DO TRABALHO POR TERCEIROS Este  um trabalho académico que pode ser utilizado por terceiros desde que respeitadas as regras e boas práticas internacionalmente aceites, no que concerne aos direitos de autor e direitos conexos. Assim, o presente trabalho pode ser utilizado nos termos previstos na licença abaixo indicada. Caso o utilizador necessite de permissão para poder fazer um uso do trabalho em condições não previstas no licenciamento indicado, dever contactar o autor, através do RepositriUM da Universidade do Minho. Licença concedida aos utilizadores deste trabalho Atribuição CC BY https://creativecommons.org/licenses/by/4.0/ iii ACKNOWLEDGEMENTS There are a few things in life that you do alone, this is not one of them. For that reason, I must thank some people for getting here. First, I would like to thank my supervisor, Professor Ana Paula Faria, for her patience, dedication, and availability. Without her help, this dissertation would not be possible. I want to thank my parents for the opportunity they gave me, despite some doubts on my side, their love and trust were always felt. I also thank my sister and my boyfriend, who deal with less good times of pressure and anxiety, the words and courage they give me are inexplicable. Finally, I would like to thank all my family and friends who, probably without noticing, gave me some strength to continue and complete this stage. iv STATEMENT OF INTEGRITY I hereby declare having conducted this academic work with integrity. I confirm that I have not used plagiarism or any form of undue use of information or falsification of results along the process leading to its elaboration. I further declare that I have fully acknowledged the Code of Ethical Conduct of the University of Minho. v Work organization, innovation, and firm performance ABSTRACT This dissertation addresses the role of work organization practices on a firm's propensity to innovate and subsequent firm´s innovation performance. Some work organization practices have been pointed out as an important driver of firms´ innovation. This is because these novel practices enhance employees´ creativity, which in turn increases firms´ propensity to innovate. Despite its importance, the topic is still understudied mostly due to a lack of data. This study aims to fill this gap by making use of recent data on work organization practices, at the firm level and investigating their role in both firms´ propensity to innovate and innovation´s commercial success. The work organization practices under analysis are job rotation, brainstorming sessions, and cross-functional teams. Our data comes from the Portuguese Community Innovation Survey (CIS) 2018 which covers a three-year period from 2016 to 2018 and 15 876 firms. As we look at two dimensions of the innovation process, i.e., the propensity to innovate and innovation performance, the analysis was divided into two steps and relies on two dependent variables. In order to deal with the two linked dependent variables, we applied Heckman's selection model as our econometric tool. Overall, our results show a positive relationship between work organization practices and the propensity to innovate. We also observe a positive relationship between these practices and the firm’s innovation commercial success. Yet, we find that work organization practices seem to be more relevant to the success of radical innovations than to incremental ones. Finally, our results show that, in the case of incremental innovations, the importance of each work organization practice is different whether one is analyzing the outcome or the performance innovation. Our findings provide useful evidence to managers who seek to increase their performance through innovation. Specifically, work organization practices are an effective tool to help firms attain innovation, in particular, with a higher degree of novelty. Keywords: Innovation; Innovation performance; Work organization. vi Organização do trabalho, inovação e desempenho económico da empresa RESUMO Esta dissertação aborda o papel das práticas de organização do trabalho na propensão de uma empresa inovar e no subsequente desempenho inovador da mesma. Algumas práticas de organização do trabalho têm sido apontadas como um importante impulsionador da inovação das empresas. Isso ocorre porque essas novas práticas aumentam a criatividade dos funcionários, o que, por sua vez, aumenta a propensão das empresas a inovar. Apesar da sua importância, o tema ainda é pouco estudado devido à falta de dados. Este estudo pretende colmatar esta lacuna recorrendo a dados recentes sobre as práticas de organização do trabalho, ao nível da empresa, e investigando o seu papel na propensão para inovar e no sucesso comercial da inovação. As práticas de organização do trabalho em análise são rotatividade do trabalho, sessões brainstorming e equipas multifuncionais. Os nossos dados provêm do Inquérito Comunitário à Inovação Portuguesa (CIS) 2018 que abrange um período de três anos de 2016 a 2018 e 15 876 empresas. Ao olharmos para duas dimensões do processo de inovação, ou seja, a propensão a inovar e o desempenho da inovação, a análise foi dividida em duas etapas e conta com duas variáveis dependentes. Para lidar com as duas variáveis dependentes interligadas, aplicamos o modelo de seleção de Heckman como ferramenta econométrica. No geral, os resultados mostram uma relação positiva entre as práticas de organização do trabalho e a propensão a inovar. Também observamos uma relação positiva entre essas práticas e o sucesso comercial de inovação da empresa. No entanto, descobrimos que as práticas de organização do trabalho parecem ser mais relevantes para o sucesso das inovações radicais do que para as incrementais. Por fim, os resultados mostram que, no caso de inovações incrementais, a importância de cada prática de organização do trabalho é diferente, seja analisando os efeitos ou o desempenho de inovação. Este estudo fornece evidências úteis para administradores que procuram aumentar o seu desempenho por meio da inovação. Especificamente, as práticas de organização do trabalho são uma ferramenta eficaz para ajudar as empresas a alcançar a inovação, em particular, com maior grau de novidade. Palavras-chave: Inovação, Organização do trabalho, Performance de inovação. vii LIST OF CONTENS ACKNOWLEDGEMENTS ....................................................................................................................... iii ABSTRACT ................................................................................................................................................. v RESUMO .................................................................................................................................................. vi LIST OF CONTENS................................................................................................................................. vii LIST OF TABLES ....................................................................................................................................... ix LIST OF FIGURES ...................................................................................................................................... x CHAPTER 1 INTRODUCTION ................................................................................................................ 1 1.1. Background and motivation ................................................................................................... 1 1.2. Objectives and research questions ...................................................................................... 2 1.3. Organization ............................................................................................................................... 2 CHAPTER 2 RELEVANT LITERATURE .................................................................................................. 3 2.1. Innovation ................................................................................................................................... 3 2.2. Work organization .................................................................................................................... 4 2.3. The impact of work organization practices on firms´ innovation and performance 6 CHAPTER 3 METHODOLOGY ............................................................................................................... 9 3.1. Data and variables .................................................................................................................... 9 3.2. Econometric model................................................................................................................ 12 CHAPTER 4 RESULTS ........................................................................................................................... 14 4.1. The distribution of work innovation practices among Portuguese firms .............. 14 4.2. Descriptive Statistics ............................................................................................................. 18 4.3. Empirical results ..................................................................................................................... 21 CHAPTER 5 CONCLUSIONS ................................................................................................................ 28 5.1. Synthesis ................................................................................................................................... 28 viii 5.2. Main conclusions .................................................................................................................... 28 5.3. Limitations and future research avenues ....................................................................... 29 REFERENCES .......................................................................................................................................... 30 5 opinions about the main definition of job design. Belias and Sklikas (2013: 85) determine job design as “a combination of job content and the work method which has been adopted in the performance of the job”. Fahr (2011: 30) says that job design “describes the characteristics and working conditions of a workplace in a broader scope”. According to Ali and Aroosiya (2014: 35) job design relates the “functions of arranging task, duties and responsibilities in to an organizational unit of work.” For these authors, this is one of the most important functions in human resource management, that helps to indicate the designing of contents, methods, and functions of a job. According to Kapur (2018) job design is the manner, in which the organizations define and structure jobs. It helps to create a job specification that will motivate the employees promoting their best performance. Therefore, Kapur (2018) sees job design as an important determinant of employee´s motivation and productivity. There are different techniques inside these functions in human resource management, one of the most known is job rotation. Job rotation is a system that gives the possibility to employees to rotate from one job or department to another in the organization (Belias and Sklikas, 2013; Kapur, 2018). According to Belias and Sklikas (2013) job rotation contributes to generating awareness and knowledge among the individuals in terms of various areas and develop their skills and abilities effectually motivating employees and contributing to their inclusion while reduces employee boredom and absenteeism. It is a usual process in large firms (Belias and Sklikas, 2013; Kapur, 2018). Laursen and Foss (2012) says that this type of human resource management can contribute to an innovator performance. Another method of work organization well known is the cross-functional teams. This technique is the combination of groups of people from various areas and backgrounds, with different knowledge and skills, from distinct functions that work together to achieve a common goal (Stipp et al., 2018; Love and Roper, 2009). Love and Roper (2009) say that “Cross-functional teamworking has been strongly advocated as part of leading practice innovation strategy, with attention focused largely on team organization, management and psychology”. The implementation of these teams can promote a better performance in terms of the ability to solve problems, produce quality, increase creativity and a better use of the resources these comparing when done individually, in this way, can increases innovation 6 performance and have a positive impact on different organizational processes (Stipp et al., 2018; Love and Roper, 2009; Zeller, 2002). 2.3. The impact of work organization practices on firms´ innovation and performance Work organization practices are an important procedure for firms since it aims to achieve an objective. As we already saw, work organization is a type of innovation that allows the firm to grow and develop for a better version of itself. In recent years, this kind of innovation has become very popular in the literature. Several authors have been analyzing this theme, although with different study objectives. Some authors focus on the impact of this innovation on firm export performance (for example, Azar and Ciabuschi, 2017; Prange and Pinho, 2017). Others opted to understand the relationship between different types of innovations and their effects on firms (Gunday et al., 2011; Camisón and Villar-López, 2014). On related subjects, Chen and Huang (2009), and Haneda and Ito (2018) analyzed the relationship between human resources practices and innovative performance. Love and Roper (2009) opted to study the profits of cross-functional teams in innovation, and Fonseca et al. (2019) studied the organizational task structure on firm’s innovation processes, they determined the link between tasks and innovation at the firm level. Azar and Ciabuschi (2017) studied the relationship between organizational and technological innovation on 218 Swedish firm export performance. They found that the union of these two different types of innovations can improve export performance. Organizational innovation supports technological innovation performance, which is reflected in exports. Prange and Pinho (2017) analyzed the importance of internal drivers (personal and organizational) on the exporting performance of small and medium-sized enterprises (SMEs) in Portugal. According to the authors, the results show that organizational innovation mediates the effect of drivers on SMEs performance, and “the indirect effect of organizational innovation is particularly strong for the organizational-driver-performance relationship” (Prange and Pinho, 2017: 1115). In sum, both articles agree that organizational innovation has (directly or indirectly) a positive influence on a firm's export performance and simultaneously shows a good 7 correlation with other types of innovations and with the internal drivers (personal and organizational) of SMEs. Gunday et al. (2011) studied how organizational, process, product, and marketing innovations affect diverse firm performance aspects (innovative, production, market, and financial performance). They conclude that all individual innovation types are more or less positively and significantly associated with some aspects of firm performance, in particular, organizational innovation has a fundamental position for innovative capabilities, in a way to promote other innovations. In the same line, Camisón and Villar-López (2014) investigated the relationship between organizational innovation and technological product and process innovation capabilities on the firm's performance. The study finds that organizational innovation favors the technological innovation capabilities and together can lead to better execution. Gunday et al. (2011) and Camisón and Villar-López (2014) found positive effects of innovations on firms, and argue that knowledge management practices, especially in terms of innovation capabilities can lead to superior firm performance. Chen and Huang (2009), in their study about knowledge management capacity and human resources practices on innovative performance, conclude that human resources practices are positively related to a firm's innovation performance, and knowledge management capacity acts as a mediator to attenuate these positive relationships. knowledge management is considered organizational innovation. The authors suggested that a better level of knowledge management capacity can stimulate creative and innovative thoughts that may eventually lead to better innovation performance. Haneda and Ito (2018), when studying the relationship between a firm’s organizational and human resource management practices on the R&D and innovation, observed that when more than one organizational and human resource management practice of a firm is implemented together raises the probability of creating new products. However, when talking about process innovation the same does not necessarily happen, the addition of organizational and work management practices considered in the study does not expand the probability of innovating. Love and Roper (2009) observed the advantages of cross-functional teams for innovation. In the study, we can see that cross-functional teams have a positive impact on innovation 8 outputs, however, if these practices are wrongly implemented in the innovation process can lead to a negative effect on innovation output. Finally, Fonseca et al. (2019) analyzed how a firm task distribution (measured by the degree of abstractionism) affects the propensity to innovate and innovation performance. Concluded that the level of abstractionism of a firm has an impact on the firm's propensity to innovate and on product innovation performance. 9 CHAPTER 3 METHODOLOGY 3.1. Data and variables Data The data used to test our hypotheses is from the Portuguese version of the European Community Innovation Survey (CIS) of 2018, relative to the 2016 – 2018 period. This survey, carried out in all member states of the European Union, follows the methodological recommendations of Eurostat based on the principles defined in the Oslo Manual. CIS 2018 follows the most recent revision of the Oslo Manual (4th edition, 2018). Compared to previous editions of CIS, this one had some changes on the definition of business innovation, reducing four types of innovations (product, process, organizational, and marketing) to two (product innovation and business process innovation). In addition, although CIS has a lot of information on innovation activities, it was only in the last survey, CIS2018, that the work organization topic was included. The key advantage of these data is that we are able to employ direct measures of work practices instead of using proxies as in previous studies. The CIS 2018 sample consists of 15 876 firms. At the end of the data collection period, there were 13 701 responses valid, corresponding to a response rate of 86.3% of the total sample. It should be said that the CIS survey is a representative sample for small (with a minimum of 10 employees) and medium sized firms, and it is the population for large firms (250 or more employees). Dependent variables To answer the research question (the role of work organization practices on the propensity to innovate and innovation performance) it is necessary to use two dependent variables, as we are looking for two dimensions of the innovation process. The first dependent variable measures if the firm introduced a product or service innovation in the period 2016-2018. In particular, we are using two variables to measure innovation. One variable Market, measures if the firm introduced an innovation that is new to the national market, and the variable Global measures if the firm introduced an innovation that is new to the global market. These 10 definitions correspond to the degree of novelty of an innovation concept as in the Oslo Manual (2018). Thus, Market corresponds to an incremental innovation and Global corresponds to radical innovation. Both variables are defined as binary variables (see Table 1). Our second dependent variable measures the commercial success of the innovation introduced by the firm. Specifically, it is measured by the percentage of firm's turnover (in log) originated by the product innovation. Independent variables To empirically test our hypotheses, we used three questions of the CIS survey regarding to methods of work organization in the firm management. The first question is about the planning staff turnover between different functional areas of the firm, the second one, brainstorming sessions (debates) with staff, and the last one working groups or crossfunctional teams. These are the variables corresponding to work organization practices. These questions are measured by the degree of importance (from 0: not important to 3: high), for each question we created a dummy that is equal one if the firm answered 2 or 3 or equal zero if the answer is 1 or 0. Control variables Several factors can influence firms´ decision to innovate. Firms with R&D activities are more likely to invest in innovation performance (Fonseca et al., 2019; Haneda and Ito, 2018). Following the literature (Azar and Ciabuschi, 2017; Fonseca et al.; 2019; Gunday et al., 2011; Haneda and Ito, 2018; Love and Roper, 2009) we decided to include total R&D expenditure, which measures the firm´s capabilities in innovation. Likewise, we also control for the workforce´s level of education, which we measured by the percentage of employees with college degree, since innovation outcomes are expected to be positively affected by highly qualified employees (Fonseca et al., 2019; Love and Roper, 2009). Firm size is a common variable to predict innovation activities and a firms’ propensity to innovate (Azar and Ciabuschi, 2017; Cheng et al., 2014; Prange and Pinho, 2017; Chen and Huang, 2009; Damanpour et al., 2009; Love and Roper, 2009). We measured firm size by three cohorts, small-sized (between 10 and 50 employees), medium-sized (between 50 and 250 employees), and large-sized firm (more than 250). We also control firms´ ownership capital, 11 i.e., if it is national or foreign-owned, and the firms’ degree of internationalization, measured by the percentage of the firm's turnover with foreign customers. According to the literature (Azar and Ciabuschi, 2017; Fonseca et al., 2019; Love and Roper, 2009) the higher the percentage of internationalization, the greater the probability of the firm to innovate. Beyond that, we also include the industry, where it is possible to know if the firm is part of the manufacturing or the services industry. Table 1 presents the variables and their description. Table 1 - Variables description. Source: Author Variables Description Global Binary variable = 1 if the firm introduced product/service innovation new to the world market, 0 otherwise Market Binary variable = 1 if the firm introduced product/service innovation new to the firm, 0 otherwise Turnover Firm's turnover percentage resulting from sales of innovative products (goods or services) in log Job Rotation Binary variable = 1 if the firm answered 2 or 3, 0 otherwise Brainstorming Binary variable = 1 if the firm answered 2 or 3, 0 otherwise Cross-functional Binary variable = 1 if the firm answered 2 or 3, 0 otherwise Size Categorical = 1 small-sized firm (>=10 and <50), 2= medium-sized firm (>=50 and < 250), =3 large-sized firm (>250) employees College/skills Categorical = 0 if 0% of employees with college degree, = 1 if >= 1% and <24% of employees with college degree, and = 2 if >= 25% of employees with college education R&D Total R&D expenditure (intramural and extramural expenditures) in log Exports Sales percentage to foreign markets Foreign Binary variable =1 if the firm is owned by foreign capital, = 0 otherwise Industry Binary variable = 1 if the firm belongs to the manufacturing industry, = 0 if the firm belongs to the services industry 12 3.2. Econometric model In terms of econometric approach, we follow Fonseca et al. (2019) who studied the role of task structure on the propensity of Portuguese firms to innovate and its impact on innovation performance. Since we are looking at two different dimensions of the innovation process, the decision to innovate and innovation performance, we used two intertwined dependent variables to test these hypotheses. This is an important part of the model, because if the interrelation was not considered the estimation probably would be wrong. It is understandable that the decision to innovate determines the innovation performance, so there is no performance to be measured if a firm decides not to innovate (Fonseca et. al, 2019). For that reason, the implementation of Heckman's selection model is necessary. This model is divided in two steps, the first one considers the decision to innovate (the selection equation), and the second step is to estimate the innovation performance model (the outcome equation). Both equations are estimated simultaneously by maximum likelihood (Fonseca et al., 2019; Thornhill, 2006). One needs to specify variable lists for both the selection equation and the outcome equation. Though the same list of variables can be employed, it is common for applied work to look for exclusion restrictions (Cameron and Trivedi, 2010). That is, to seek for variable(s) that can generate nontrivial variation in the selection variable but does not affect the outcome variable directly. This way the independent source of variation in the probability of a positive outcome is identified in a more robust way. To this end, we use an identification variable associated with the decision to innovation but not directly correlated with the outcome variable, i.e., the innovation performance. We chose to use the logarithm of R&D expenditure as our identification variable since this variable manipulates the innovation creation but does not necessarily contribute to the turnover of the firm. Formulating the model, let innovi* be the decision to introduced goods or services (binary), and performi the continuous variable that measures the logarithm of innovation sales performance. 13 The first step of the model, tendency to innovate, can be written as: 𝑖𝑛𝑜𝑣𝑖 ∗= 𝑧𝑖´𝛾 + 𝑢𝑖 (1) Where innovation propensity (innovi*) is a hidden variable, dependent on vector z which includes our work organization practices variables and a set of control variables that influence the propensity to innovate and the logarithm of R&D expenditure variable, is our exclusion restriction variable, and ui is the error term with u ∼ N (0, 1). The observed variable represents the decision to innovate that takes the value 1 when innovi* > 0 and zero otherwise. In the second step, we estimate the firm´s innovation performance, which is conditional on the first step and can be written as: 𝑝𝑒𝑟𝑓𝑜𝑟𝑚𝑖= 𝑥𝑖´𝛽 + 𝑒𝑖 , if innovi=1 (2) Where performi is the innovation performance and is explained by the vector x, that includes only the independent and control variables, and ei is the error term with e ∼ N (0, σ). Equation (2) has a selection of firms that engaged in product innovation activities, this means is only observed if innovi* = 1. The model allows the correlation between the error term in both equations, ρ = corr (u, e), to be different from zero. If we reject the null hypothesis (ρ = 0) then there is a selection effect that would bias the results of the innovation performance equation in case we ignore the first step of the model. 14 CHAPTER 4 RESULTS 4.1. The distribution of work innovation practices among Portuguese firms Table 2 presents the distribution of work organization practices by the degree of its importance among Portuguese firms in 2018. Of the three work practices, Cross-functional is the one with a higher degree of importance, with 18%. Brainstorming is next with 16%, and finally, Job rotation with 13.1%. Also, Cross-functional is seen as medium or high importance by nearly 59% of the firms, and Brainstorming by 57% of the firms, and Job rotation by nearly 54%. Thus, these values suggest that these practices are somehow high valued by Portuguese firms. Table 2 - Degree of importance of work practices, Portugal, 2018 Degree of importance Job rotation Brainstorming Cross-functional Not important 18.55% 17.39% 18.13% Low 27.36% 25.05% 22.64% Medium 40.99% 41.53% 41.20% High 13.09% 16.04% 18.03% Source: Author Figures 1, 2, and 3 we can see the degree of importance of Job Rotation, Brainstorming and Cross-functional, respectively, by firm size. These figures show that large firms tend to value more these practices than medium or small firms, as we observe slightly larger percentage of large firms saying that these practices are very important across the three types of work 21 4.3. Empirical results Table 5 and Table 6 present the estimates of the first and second step of the Heckman model, respectively, in which the dependent variable in the first step is Global – when a firm introduces products that are new to the world. Likewise, Table 7 and Table 8, present the first and second step estimates of the Heckman model, respectively, in which the dependent variable in the first step is Market – when the firm introduces a new product to the national market where operates. Table 5 shows four different Models 1.1, 1.2, 1.3, and 1.4. The first three models show the estimates for the three variables Job rotation (Model 1.1), Brainstorming (Model 1.2), and Cross-functional (Model 1.3), and Model 1.4 shows the estimates with the three work practices variables simultaneously in the regression. In the four Models, it is possible to see that both steps are not independent, this is confirmed by the level of significance of the IMR inverted mills ratio at 5% for all the models. So, the econometric approach is justified. The estimates of the first three models in Table 5 show that all three work practice variables are positive and statistically significant at 1%, therefore these practices increase the likelihood of a firm introducing Global innovations. In the last column of Table 5 (1.4) we find that all three variables are statistically significant either at 1% or 5% level. We can also see that R&D expenditure, the identification variable, has a positive relationship with the dependent variable and a level of significance at 1%. The control variables have the expected signal, in the three models the variable Industry is significant, so the manufacturing industry has a positive relationship with the dependent variable, and it is above the relationship with the omitted class (services industry). The fact that the firm is in the manufacturing industry gives them a higher probability to innovate. Also interesting, the variable Foreign is significant and has a negative effect on the propensity to develop products that are new to the world, if the firm is national has more probability to introduce this type of innovation, that is, radical innovation. This can support that Portuguese firms can be pioneers in the creations of new products and services that are new to the world market. 22 Table 5 - Regression results for the innovation propensity, Global innovation (first step) Model 1.1 Model 1.2 Model 1.3 Model 1.4 Job rotation 0.129*** 0.070** (0.028) (0.030) Brainstorming 0.181*** 0.106*** (0.030) (0.039) Cross-functional 0.173*** 0.079** (0.031) (0.040) Medium size 0.013 0.023 0.001 0.008 (0.064) (0.064) (0.064) (0.064) Large size 0.131* 0.138** 0.112 0.122* (0.070) (0.070) (0.070) (0.070) Medium-skilled 0.060 0.001 0.0141 0.017 (0.121) (0.121) (0.121) (0.122) High-skilled 0.296** 0.186 0.199 0.203 (0.127) (0.128) (0.127) (0.128) Exports 0.001 0.001 0.001 0.001 (0.001) (0.001) (0.001) (0.001) Foreign -0.125 -0.131* -0.130* -0.130* (0.073) (0.073) (0.073) (0.073) log R&D 0.049*** 0.043*** 0.038*** 0.041*** (0.012) (0.012) (0.012) (0.013) Industry 0.100* 0.137** 0.126** 0.131** (0.056) (0.056) (0.056) (0.057) Constant -0.957*** -0.966*** -0.923*** -1.089*** (0.164) (0.162) (0.160) (0.167) Wald chi2 126.44 93.43 70.99 134.15 log likelihood -1841.31 -1834.685 -1836.463 -1828.867 Number of obs. 2.722 2.722 2.722 2.722 Innov. definition (dep. var.) Global Global Global Global Source: Author Notes: Standard errors in parenthesis. level of significance ***, **, * at 1%, 5% and 10% respectively. The dependent variable is a dummy variable equal one when the firm innovates and zero otherwise. Innovation definition is Global when a firm introduces products that are new to the world. Table 6 shows the results for step two. We see some differences. First, we observe that Job rotation and Brainstorming are the only work methods that seem to play a role in the firm's innovation performance. Cross-functional does not appear with statistical significance. These results are the same in the last model (4.3), where only Job Rotation and Brainstorming seem to contribute to the innovation's commercial success. 23 Table 6 - Regression results for the innovation performance, Global innovation (second step) Model 2.1 Model 2.2 Model 2.3 Model 2.4 Job rotation 1.603*** 1.310*** (0.498) (0.426) Brainstorming 2.031** 1.496** (0.854) (0.610) Cross-functional 1.365 -0.424 (0.982) (0.597) Medium size -4.809*** -4.607*** -4.706*** -4.744*** (0.757) (0.891) (0.922) (0.773) Large size -5.533*** -5.177*** -5.301*** -5.416*** (1.015) (1.233) (1.258) (1.058) Medium-skilled -2.263 -2.763 -2.594 -2.414 (1.516) (1.717) (1.810) (1.530) High-skilled -2.071 -2.813 -2.322 -2.426 (1.930) (2.058) (2.243) (1.870) Exports 0.021** 0.020* 0.023* 0.021** (0.010) (0.011) (0.012) (0.010) Foreign -1.452* -1.837* -1.838 -1.650* (0.010) (1.068) (1.154) (0.954) Industry -1.585** -0.969 -1.097 -1.342 (0.749) (0.984) (1.048) (0.869) Constant 1.356 -2.167 -1.807 -0.737 (5.676) (7.723) (9.096) (7.634) IMR 11.466** 14.612** 15.381** 12.283** (4.722) (6.222) (7.358) (5.760) Wald chi2 126.44 93.43 70.99 134.15 log likelihood -1841.31 -1834.685 -1836.463 -1828.867 Number of obs. (step 2) 1.375 1.375 1.375 1.375 Innov. definition (dep. var.) Global Global Global Global Source: Author Notes: IMR inverted mills ratio. Standard errors in parenthesis. level of significance ***, **, * at 1%, 5% and 10% respectively. The dependent variable the log of firm's turnover percentage resulting from sales of innovative products. Innovation definition is Global when a firm introduces products that are new to the world. Table 7 shows the estimates for the propensity to innovate but in the case of Market innovation, that is, new to the market where operates. The table contains the same structure, Models 1.1, 1.2, and 1.3 present the independent variables separately and Model 1.4 has the three work practices variables simultaneously. Again, both equation – selection and outcome, are related as IMR inverted mills ratio as a level of significance at 10% (in Models 1.2, 1.3, and 1.4) and 5% (in Model 1.1). 24 Brainstorming, Cross-functional, and Job rotation variables, individually, have an impact on the likelihood of introducing a Market innovation. However, when included simultaneously in the regression (Model 1.4), we find that only Brainstorming has an impact on the Market innovation sales. R&D and Exports are positive and significant, so both have an impact on the propensity to innovate in products that are new to the national market. The variable Foreign although significant has a negative signal, this means if the firm is national will have more probability to innovate. 25 Table 7 - Regression results for the innovation propensity, Market innovation (first step) Model 1.1 Model 1.2 Model 1.3 Model 1.4 Job rotation 0.057* 0.026 (0.030) (0.032) Brainstorming 0.103*** 0.084** (0.033) (0.043) Cross-functional 0.077** 0.013 (0.034) (0.044) Medium size -0.227*** -0.223*** -0.227*** -0.226*** (0.070) (0.070) (0.120) (0.070) Large size -0.292*** -0.288*** -0.298*** -0.292*** (0.076) (0.076) (0.077) (0.077) Medium-skilled -0.577*** -0.608*** -0.601*** -0.604*** (0.120) (0.120) (0.120) (0.121) High-skilled -0.489*** -0.553*** -0.536*** -0.545*** (0.127) (0.128) (0.128) (0.129) Exports 0.005*** 0.005*** 0.005*** 0.005*** (0.001) (0.001) (0.001) (0.001) Foreign -0.121 -0.133* -0.130 -0.133* (0.080) (0.080) (0.080) (0.080) log R&D 0.035*** 0.031** 0.030** 0.031** (0.013) (0.013) (0.013) (0.013) Industry -0.112* -0.093 -0.102* -0.010 (0.061) (0.061) (0.061) (0.061) Constant -0.636*** -0.668*** -0.620*** -0.706*** (0.170) (0.166) (0.165) (0.172) Wald chi2 52.42 39.74 39.58 49.82 log likelihood -1499.430 -1496.425 -1498.63 -1495.997 Number of obs. 2.807 2.807 2.807 2.807 Innov. definition (dep. var.) Market Market Market Market Source: Author Notes: Standard errors in parenthesis. level of significance ***, **, * at 1%, 5% and 10% respectively. The dependent variable is a dummy variable equal one when the firm innovates and zero otherwise. Innovation definition Market when the firm introduces a new product to the national market where operates. Table 8 shows the results of the second step. We see that only Job rotation plays a role in the innovation´s performance (Models 2.1 and 2.4). Thus, these results suggest that different practices play different roles within the firm. While Brainstorming seems to be more helpful to reach an innovation outcome, Job Rotation contributes to increase the innovation´s commercial success. 26 Table 8 - Regression results for the innovation performance, Market innovation (second step) Model 2.1 Model 2.2 Model 2.3 Model 2.4 Job rotation 1.711** 1.389* (0.806) (0.781) Brainstorming 2.168 1.759 (1.441) (1.303) Cross-functional 1.331 -0.554 (1.293) (1.050) Medium size -9.371*** -9.725*** -9.816*** -9.549*** (2.236) (2.716) (2.832) (2.504) Large size -12.162*** -12.686*** -12.885*** -12.328*** (2.431) (3.004) (3.212) (2.786) Medium-skilled -8.893* -10.933 -10.679 -9.550 (5.067) (6.666) (6.831) (6.087) High-skilled -8.829** -11.388** -10.849* -9.754* (4.231) (5.900) (3.212) (5.389) Exports 0.093** 0.102* 0.104* 0.095* (0.046) (0.057) (0.059) (0.052) Foreign -3.842* -4.292* -4.216* -4.021* (2.011) (2.428) (2.456) (2.217) Industry -4.374*** -4.266** -4.520** -4.193*** (1.525) (1.706) (1.770) (1.560) Constant -6.853 -10.242 -9.018 -9.107 (11.781) (15.748) (15.954) (14.769) IMR 21.487** 24.607* 24.674* 22.344* (11.008) (14.049) (14.597) (12.837) Wald chi2 52.42 39.74 39.58 49.82 log likelihood -1499.430 -1496.425 -1498.63 -1495.997 Number of obs. (step 2) 668 668 668 668 Innov. definition (dep. var.) Market Market Market Market Source: Author Notes: IMR inverted mills ratio. Standard errors in parenthesis. level of significance ***, **, * at 1%, 5% and 10% respectively. The dependent variable the log of firm's turnover percentage resulting from sales of innovative products. Innovation definition is Market when the firm introduces a new product to the national market where operates. Considering our first research question - What is the effect of work organization practices on firms' propensity to innovate? - the results (tables 5 and 7) show that the three variables individually have a positive and significant impact on the likelihood of both types of innovation, which corroborates previous literature (Camisón and Villar-López, 2014; Gunday et al., 2011; Chen and Huang, 2009; Haneda and Ito, 2018; Love and Roper, 2009; Stipp et al., 2018; Zeller, 2002; Laursen and Foss, 2012). The R&D expenditure is significant and positive in all the models, as expected from the literature (Azar and Ciabuschi, 2017; Fonseca et al.; 2019; 27 Gunday et al., 2011; Haneda and Ito, 2018; Love and Roper, 2009), R&D activities are a fundamental part of the innovative process of a firm. When the firm introduces products that are new to the national market (Table 7), not all work organization practices are significant. According to the literature (Haneda and Ito, 2018), when we use more than one organizational and human resource management practice is possible that this combination does not increase the probability of innovating, which is the case for Job rotation and Cross-functional teams. Moving on to our second research question - What is the effect of work organization practices on the innovation´s commercial success? - the estimates in tables 6 and 8 show that when the firm introduces a new product to the world (Table 6), the variables of Job rotation and Brainstorming are significant and positive. These working methods influence positively the sales and the success of the firm, which corroborates the literature (Ali and Aroosiya, 2014; Azar and Ciabuschi, 2017; Prange and Pinho, 2017; Camisón and Villar-López, 2014; Gunday et al., 2011; Chen and Huang, 2009; Haneda and Ito, 2018; Love and Roper, 2009). In Table 8, when the firm introduces a Market innovation, i.e., new to the national market where operates, only the variable Job rotation can contribute to the commercial success of the firm. This goes in agreement with Laursen and Foss (2012) when they say human resource management practices, in which, Job rotation can support a positive influence on innovation performance and lead to the success of the business. The Exports variable has the expected values and confirms the literature (Azar and Ciabuschi, 2017; Prange and Pinho, 2017), innovation has a positive influence on a firm's export performance, which also has a positive effect on sales. 28 CHAPTER 5 CONCLUSIONS 5.1. Synthesis Work organization methods can be considered a type of business process innovation - administration and management. Some of these practices have been pointed out as an important driver of firms´ innovation. Therefore, the topic has received growing interest among scholars. Previous studies conclude that this type of innovation could have a positive impact on firms, and it can be an incentive to achieve other kinds of innovations. Extant evidence has mostly used proxies, in this study we overcame this limitation by using direct measures of novel work organization practices (job rotation, brainstorming sessions, and cross-functional teams). We used the CIS18 data because this is the only CIS that addresses these specific work organization practices. This study main objective was to analyze how work organization methods influence the propensity to innovate and, how they affect firm´s innovation performance, or commercial success. To answer these research objectives, the following questions were formulated: 1. What is the effect of work organization practices on firms´ propensity to innovate? 2. What is the effect of work organization practices on the innovation´s commercial success? Since we want to know the propensity to innovate and innovation performance, we are looking at two dimensions of the innovation process. For that reason, the analysis was divided into two steps and counted on two dependent variables, to solve the two linked dependent variables of our study we applied Heckman's selection model as our econometric tool. 5.2. Main conclusions In the first place, we investigated how these practices and their degree of importance are distributed across firms and industries in Portugal. We concluded that these practices in Portugal are somehow high valued by the firms. Individually, we saw that in terms of firm size 29 the large size firms give more importance to these practices than the medium and small ones. The results in terms of industry type (services or manufacturing) show that the services firms give more importance to these kinds of practices than the manufacturing firms. The first step (tendency to innovate) estimates show a positive influence of the studied practices on the propensity of a firm to innovate, either Global (radical) or Market (incremental) innovations. However, the results vary when the variables enter in the regression individually or together. The work organization practices altogether seem to play a more important role in the propensity to innovate in the Global case than in the Market case. In the latter, it seems that only brainstorming matters in that respect. In relation to the firm´s innovation performance (second step), the results show that only job rotation and brainstorming seem to increase the firm´s innovation performance in the Global innovation case. When innovation is Market, only job rotation appears to contribute to the innovation´s commercial success. Although the three work organization methods contribute to generating innovation, only job rotation and brainstorming seem to have an impact on the firm’s performance, which in turn varies with the type of innovation, i.e., Global (radical) or Market (incremental). 5.3. Limitations and future research avenues The main limitation of this study is that our data is a cross-section and the size of our final sample due to lack of data regarding some variables. Therefore, an important research avenue is to collect data from future CIS waves and enlarge our sample size. In addition, we consider it necessary to continue to investigate this topic so it can be more noticeable the relationship between work organization methods and innovation performance. This detection is essential to conduct, once confirmed, it will allow firms to obtain better results with the correct adoption of these practices. 30 REFERENCES Azar, G., and Ciabuschi , F. (2017). Organizational innovation, technological innovation, and export performance: The effects of innovation radicalness and extensiveness. International Business Review, 26, Issue 2, 324-336. doi:http://dx.doi.org/10.1016/j.ibusrev.2016.09.002 Ali, M., and Aroosiya, M. (2014). Impact of job design on employees’ performance (with special reference to school teachers in the Kalmunai Zone). Journal of Management, 8 (1), 33– 41. Belias, D., and Sklikas, D. (2013). Aspects of Job Design. International Journal of Human Resource Management and Research (IJHRMR), 3, Issue 4, 85-94. Cameron, A. C., and Trivedi, P. K. (2010). Microeconometrics Using Stata, Revised Edition. Stata Press. Camisón, C., and Villar-López, A. (2014). Organizational innovation as an enabler of technological innovation capabilities and firm performance. Journal of Business Research(67), 2891-2902. doi:10.1016/j.jbusres.2012.06.004 Chen, C.-J., and Huang, J.-W. (2009). Strategic human resource practices and innovation performance — The mediating role of knowledge management capacity. Journal of Business Research(62), 104-114. doi:10.1016/j.jbusres.2007.11.016 Damanpour, F., and Evan, W. M. (1984). Organizational Innovation and Performance: The Problem of "Organizational Lag. Administrative Science Quarterly, 29(3), 392-409. doi:10.2307/2393031 Damanpour, F., Walker, R. M., and Avellaneda, C. N. (June de 2009). Combinative Effects of Innovation Types and Organizational Performance: A Longitudinal Study of Service Organizations. Journal of Management Studies(46:4). doi:10.1111/j.14676486.2008.00814.x Damanpour, Fariborz; Rutgers Business School; Newark and New Brunswick. (2017). Organizational Innovation. Oxford Research Encyclopedias, Business and Management. doi:https://doi.org/10.1093/acrefore/9780190224851.013.19