Title
Measuring efficiency of innovation using combined Data Envelopment Analysis and Structural Equation Modeling: empirical study in EU regions
Abstract
The main aim of this paper is to investigate the impact of patent applications, development level, employment level and degree of technological diversity on innovation efficiency. Innovation efficiency is derived by relating innovation inputs and innovation outputs. Expenditures in Research and Development and Human Capital stand for innovation inputs. Technological knowledge diffusion that comes from spatial and technological neighborhood stands for innovation output. We derive innovation efficiency using Data Envelopment Analysis for 192 European regions for a 12-year period (1995–2006). We also examine the impact of patents production, development and employment level and the level of technological diversity on innovation efficiency using Structural Equation Modeling. This paper contributes a method of innovation efficiency estimation in terms of regional knowledge spillovers and causal relationship of efficiency measurement criteria. The study reveals that the regions presenting high innovation activities through patents production have higher innovation efficiency. Additionally, our findings show that the regions characterized by high levels of employment achieve innovation sources exploitation efficiently. Moreover, we find that the level of regional development has both a direct and indirect effect on innovation efficiency. More accurately, transition and less developed regions in terms of per capita GDP present high levels of efficiency if they innovate in specific and limited technological fields. On the other hand, the more developed regions can achieve high innovation efficiency if they follow a more decentralized innovation policy.
Year
DOI
Venue
2020
10.1007/s10479-017-2728-4
Annals of Operations Research
Keywords
DocType
Volume
Technological diversity, R&D, Patents, Data Envelopment Analysis, Structural Equation Modeling
Journal
294
Issue
ISSN
Citations 
1
1572-9338
0
PageRank 
References 
Authors
0.34
0
4