Non-technical innovation Definition, Measurement & Policy Implications Workshop

1. Non-technical innovation Definition, Measurement & Policy Implications Workshop Karlsruhe, 16-17 October 2008 The employment impact of technological and organizational innovations. Firm level evidence Rinaldo Evangelista Antonio Vezzani University of Camerino (IT) University of Rome “Tor Vergata”

2. Structure of the presentation • State of the art and motivation of the empirical study • Issues addressed and methodology • Results (preliminary) • Synthesis of results and conclusions

3. The debate on the innovation-employment relationship: an “never ending story” with no conclusive answers • Very complex issue which can be tackled from different perspectives and levels of theanalysis • previous works and reviews (largely referring to the manufacturing sector)show that it is impossible to come-up with a final answer • results depend very much on the level of aggregation, type of data, time framework • very difficult to grasp the net aggregate effect (i.e. grasp all direct and indirect effects at a macro-economic level) • different contributions and analyses should be used in a complementary way in order to assemble a meaningful interpretative puzzle

4. First fundamental question: Why making a specific analysis on services? • Two possible answers: • a) under-explored area -> lack of evidences and empirical studies (certainly true) • b) specificity of both innovation & the innovation-employment relationships • (open issue….. to be explored empirically). Three broad issues in the research agenda a) What we have learnt from the manufacturing industry does also apply to services? b) Should we change our approach and look at different dimensions, mechanisms, effects? c) Go for a “synthesis approach”?

6. Empirical issues addressed in the paper 1.How do product, process and organizational changes relate to each other (level and type of complementarity)? a) Do emerge distinct patters and strategies made of different combinations of product/process/organizational changes? b) Does organizational innovation represent a stand-alone strategy or it is linked to a product or process innovation strategy?  c) Are there systematic differences between manufacturing and services industries in a) and b) above? 2. The effects on employment a) What are the (direct and indirect) effects of product, process and organizational innovation on employment in Services? b) Is the product/process distinction effective in disentangling the different mechanisms and effects of innovation on employment in the Service sector?

7. Methodology • Data: Italian CIS4 (2002-2004) - Manufacturing & Services (comparative analysis) • Firm level analysis • Main CIS variables used: - Type of innovation (product/process/organizational/marketing) - Effects of tech & non-tech innovation (product vs process related) - Type of innov. input & output (innov. expenditures and innov. sales) - Employment growth (2002-2004) - Sales growth (2002-2004)

8. Main caveats of CIS (and of this study) • vary basic information/types of organizational innovations • cross sectional nature of data and analyses • difficulty of grasping the aggregate-net effects of innovation – firm level evidence cannot be generalized at a macro level • no possibility of investigating qualitative effects on employment • Major strengths • Data representative of all manufacturing and service sectors (across EU) • Possibility of combining non-technological with technological innovations

9. Relevance of (& complementarity between) technological and non-tech. innovations

10. Innovation modes: results from a cluster analysis Sectoral characterization of innovation modes

11. Effects of technological innovation (CIS4) How important were each of the following effects of your product (good or service) and process innovations introduced during the three years 2002 to 2004 (scores from 0 to 3)? • Increased range of good or services • Entered new markets or increased market share • Improved flexibility of production or service provision • Reduced labour costs per unit output • Reduced materials and energy per unit output Product related Process related The importance of product and process related effects has been computed as average values of the scores attached to each set of answers.

12. Effects of non technological innovation (CIS4) If your enterprise introduced an organisational innovation during the three years 2002 to 2004, how important were each of the following effects (scores from 0 to 3)? • Reduced time to respond to customer or supplier needs • Improved quality of your goods or services • Reduced costs per unit output • Improved employee satisfaction • and/or reduced rates of employee turnover Product related Process related Job satisfaction

13. Innovation profiles of different clusters (probit estimates) – Manuf. & Services * significant at 5%; ** significant at 1% - Marginal effects are reported in table

14. Innovation profiles of different clusters (probit estimates) - Services * significant at 5%; ** significant at 1% - Marginal effects are reported in table

15. Product INNOVATION Complex Indirect effects Organisational Process Process Product performances Efficiency gains ECONOMIC PERFORMANCE (Growth of sales) EMPLOYMENT The employment impact of innovation at a firm level Organisational Direct effects

16. The model Labour effects equation: Economic performance equation: ∆Lit = rate of change of employment (2002-2004) ∆Yit = rate of change of turnover (2002-2004) Lt-1 = employment in 2002 πt-1 = productivity in 2002 Proc = dummy for process innovation Org = dummy for organisational innovation Herf = Herfindahl concentration index (for 2-digit level) Dem = Uncertain demand for innovative goods or services

17. Estimation strategy • Three-stage least square estimation: 1th Develops instrumented values for all endogenous variables (In our system: ∆Lit and ∆Yit) 2nd Obtains consistent estimates for the covariance matrix of the equation disturbances 3rd Performs a GLS-type estimation using the covariance matrix estimated in the second stage and with the instrumented values in place of the right-hand-side endogenous variables (In this case: ∆Yit)

18. The employment impact of innovation: estimation results * significant at 10%; ** significant at 5%; *** significant at 1% Industry dummies are included in both ∆L equations, corresponding effects are not shown,but Ind_dummies reports the p-value of a test of the joint significance.

19. Synthesis of results • Relevance (and complementarity between) product, process and organizational innovations • Even when non-technological-innovation are brought into the picture, the product/process distinction is still effective in identifying different innovation strategies and performances. This holds both in the manufacturing and service sectors. • Nonetheless, our analysis has shown the presence of a distinct “pure organizational innovation mode” particularly relevant in Services • Our analysis has also identified two different product-driven innovation modes: - one simply based on the mere introduction of product innovation • - a more “complex” and “innovative” one, made of a combination of • different and complementary types of innovation (product, process and organizational)

20. 2. The employment impact of innovation (synthesis of results) • No negative (labour displacing) direct effects of innovation on employment have been found in both manufacturing and service industries. • In both sectors innovation exerts its impact on employment in an indirect way, that is improving the competitive performances of firms (growth of sales and gain of market shares) • Product related strategies (when complemented with the introduction of both process and organizational innovations) show the strongest employment impact (via improved economic performances) in both Manufacturing and Services.   • The employment impact of process innovation is positive only in the Manufacturing sector. This indicates that only in this sector productivity gains can be obtained through pure process innovation strategies. • -> This implies that in the case of services, pure technology based innovations are less effective and efficiency gains (and employment growth) are likely to be associated to more complex (technological and organizational) innovation mix. • The product/process distinction, although very relevant, does not capture all types of innovation strategies and their different effects on employment. • The introduction of stand-alone organizational changes emerges as an effective innovation strategy for a relatively large section of firms. • Surprisingly enough, the strongest (positive) employment impact of this type of innovation is found in the manufacturing sector.

21. Some implications… All in all the evidence presented provides further support to a “synthesis approach” and suggests that such an approach should be extended also to the analysis of the relationships between innovation (economic performances) and employment. This is not an easy task though, both on a conceptual and on an empirical ground While concepts, definition and data-set on firms’ technological activities are rather settled and widely available, the same does not apply to organizational innovation Future advancements in this research area (as well as any attempt of overcoming a sterile manufacturing/services divide) will crucially depend on the extent to which this type of asymmetry will be overcome, and on the possibility of integrating robust and meaningful data on both technological and organizational innovation