schneider eckl - the difference make a difference
TRANSCRIPT
THE DIFFERENCE MAKES A DIFFERENCE:TEAM DIVERSITY AND INNOVATIVE CAPACITY
Julia Schneider, Verena Eckl
OECD Blue Sky III, Ghent, 21 September 2016: Developing novel approaches to measure human capital and innovation
THE DIFFERENCE MAKES A DIFFERENCE: TEAM DIVERSITY AND INNOVATIVE CAPACITY 2Ghent, 21 September 2016
MOTIVATION
THE DIFFERENCE MAKES A DIFFERENCE: TEAM DIVERSITY AND INNOVATIVE CAPACITY 3
DOES DIVERSITY HELP TO INVENT BREAKTHROUGH INNOVATIONS?
Ghent, 21 September 2016
» Invention as source of innovation: uncertain process of recombining components
» Infinite set of potential combinations individual inventors have only an infinitesimal understanding of all potential combinations
» Technological breakthroughs need new combinations of components well-known to the participating inventors
» Diverse groups of inventors with deep experience in different fields • know more potentially successful combinations of components
higher probability for technological breakthroughs• know more about diverse markets, needs and tastes higher
probability for target group-specific or international innovations
Schumpeter 1939, Fleming 2001, Ahuja 2000, Parotta et al. 2011, Watson et al. 1993, Drach-Zahavi and Somech 2001, Hong and Page 2001, Osborne 2000, Berliant and Fujita 2008, Nelson and Winter 1982
THE DIFFERENCE MAKES A DIFFERENCE: TEAM DIVERSITY AND INNOVATIVE CAPACITY 4Ghent, 21 September 2016
RESEARCH QUESTION
THE DIFFERENCE MAKES A DIFFERENCE: TEAM DIVERSITY AND INNOVATIVE CAPACITY 5
RESEARCH QUESTION: DO THE GAINS OF DIVERSITY OFFSET THE COSTS?
Ghent, 21 September 2016
» Diversity can lead to frictions : • Mixed teams might have greater problems to effectively
communicate and cooperate • Levels of trust may also be lower, due to real and perceived
differences between team members» Costs can be theoretically so high that they offset the gains from
diverse inventor groups
» Can we find empirical evidence that firms that employ researcher teams with a higher degree of diversity – firms with more foreign, female, non-STEM researchers – have a better innovative capacity?
Alesina and Ferrara 2005, Becker 1957, Williams and O'Reilly 1998, Basset-Jones 2005, Zajac et al. 1991, Lazear 1999
THE DIFFERENCE MAKES A DIFFERENCE: TEAM DIVERSITY AND INNOVATIVE CAPACITY 6Ghent, 21 September 2016
PREVIOUS FINDINGS
THE DIFFERENCE MAKES A DIFFERENCE: TEAM DIVERSITY AND INNOVATIVE CAPACITY 7
DO THE GAINS OF DIVERSITY OFFSET THE COSTS? PREVIOUS FINDINGS
Ghent, 21 September 2016
» Innovation-driven firms become more and more diverse in the US: • successful Start-ups more than twice the average percentage of
female employees than failed Start-ups• stock market value of innovation-driven firms increases with the
number of women in top management » But Germany’s firms employ mostly German male engineers » State of research on impact of ethnical and gender diversity:
„Overall, […] the benefits of diversity are more likely to outweigh the costs in high-tech/knowledge intensive sectors than in traditional industries, particularly if the former (latter) are characterized by complex (routine) tasks, negative (positive) complementarities and innovative (functional) output.” (Garnero et al. 2014)
» No findings on impact of educational diversity on innovative capacity
Schneider and Stenke 2016, Dow Jones 2012, Dezsö and Ross 2012
THE DIFFERENCE MAKES A DIFFERENCE: TEAM DIVERSITY AND INNOVATIVE CAPACITY 8Ghent, 21 September 2016
DATA AND METHODS
THE DIFFERENCE MAKES A DIFFERENCE: TEAM DIVERSITY AND INNOVATIVE CAPACITY 9
SAMPLE
» R&D survey 2013 of the German business and enterprise R&D (BERD) collected by the SV Wissenschaftsstatistik (since 1954, full survey every odd-numbered year)
» 2013 = extended version with regard to R&D personnel: nationality, education (subject of study, scientific degree), entry wages, difficulties in recruiting, relevance and success of recruitment strategies
» Observations: 1.873 (~14% of 13.589 R&D active firms in 2013) » Response Bias: overrepresentation of the service/ IT sector and younger
scientist, no significant distortion in terms of firm size, R&D spending, percentage of women
Ghent, 21 September 2016
THE DIFFERENCE MAKES A DIFFERENCE: TEAM DIVERSITY AND INNOVATIVE CAPACITY 10
MEASURES: DEPENDENT VARIABLES
» Input Side: Future innovativeness ~ R&D intensity = R&D spending in 2013 / revenues in 2013 • High correlation of input indicator R&D spending and output indicators, i.e.
number of patents, innovation activity , revenues with new or improved products
Þ Shortcomings: (1) Uncertainty about future innovation development(2) Time lag between R&D spending and innovation output(3) Quantity of R&D covers not necessarily R&D quality
» Output Side: Innovative efficiency ~ R&D efficiency=R&D spending in 2011 / revenues with new or improved products in 2013• Well-functioning R&D teams are able to achieve a higher innovation output
with steady R&D spendingsÞ Shortcomings: Loss of observations
Ghent, 21 September 2016
THE DIFFERENCE MAKES A DIFFERENCE: TEAM DIVERSITY AND INNOVATIVE CAPACITY 11
INDEPENDENT VARIABLES = DIVERSITY MEASURES
» Modified Herfindahl index (HI), which measures not as usual the concentration (with 1 = completely concentrated), but the diversity (1 – HI-concentration).
» Two dimensions of diversity: the “richness”, which refers to the number of defined categories within a firm, and the “evenness”, which informs on how equally populated such categories are
» Range from 0 to ½ for two diversity categories, moving from highly diverse (1/2) to completely homogeneous researchers (0) within a firm
» Three diversity measures of researchers at the firm level: • (1) gender (female/male)• (2) education (subject of study STEM/other) • (3) nationality (German/other)
» Control Variables : size, sector of industry and age of firm
Ghent, 21 September 2016
THE DIFFERENCE MAKES A DIFFERENCE: TEAM DIVERSITY AND INNOVATIVE CAPACITY 13Ghent, 21 September 2016
RESULTS
THE DIFFERENCE MAKES A DIFFERENCE: TEAM DIVERSITY AND INNOVATIVE CAPACITY 14
SUMMARY STATISTICS OF THE STUDY’S VARIABLES
Ghent, 21 September 2016
Variable Number of observations
Mean Min Max
Future innovativeness (R&D Intensity)
1842 0.14 0.00 1
Innovative (R&D) efficiency
1153 -1.97 -7.39 4.14
Share of female researchers over all firms
1515 0.19 0 1
Share of foreign researchers over all firms
1787 0.05 0 1
Share of non-STEM researchers over all firms
1736 0.17 0 1
Gender diversity 1533 0.30 0 0.5Nationality diversity 1758 0.06 0 0.5Subject diversity 1707 0.11 0 0.5
THE DIFFERENCE MAKES A DIFFERENCE: TEAM DIVERSITY AND INNOVATIVE CAPACITY 16
ROBUST REGRESSION COEFFICIENTS: FUTURE INNOVATIVENESS
Diversity Measures Only Educational diversity
Only Gender diversity
Only Nationality diversity
All diversity measures together
Educational diversity 0.045* (0.022) 0.030 (0.025)Gender diversity 0.090***
(0.025) 0.079** (0.027)
Nationality diversity 0.165*** (0.043) 0.173** (0.046) Chemical industry -0.011 -0.027* -0.015 -0.016 Pharmaceutical industry 0.071* 0.053 0.068* 0.061 Electronical industry 0.044*** 0.048*** 0.036*** 0.037***Mechanical Engineering -0.007 0.001 -0.010 -0.009 Automotive industry 0.031 0.045* 0.037* 0.031 ICT sector 0.098*** 0.103*** 0.092*** 0.095***Knowledge intensive Services
0.177*** 0.170*** 0.174*** 0.184***
< 100 Employees 0.069*** 0.071*** 0.071*** 0.099***100-249 Employees 0.010 0.010 0.011 0.010 250-499 Employees 0.010 0.013 0.009 0.008 Firm Age -0.002*** -0.001*** -0.001*** -0.002***Constant 0.069*** 0.048*** 0.061*** -0.003 R-squared 0.219 0.229 0.230 0.246 No. 1600 1724 1645 1346
Ghent, 21 September 2016
» Positive relationship between each diversity measure and the firm’s future innovativeness, independent from the other predictors in the model (robust linear regression)
THE DIFFERENCE MAKES A DIFFERENCE: TEAM DIVERSITY AND INNOVATIVE CAPACITY 17
ROBUST REGRESSION COEFFICIENTS: INNOVATIVE EFFICIENCY
Diversity Measures Only Educational diversity
Only Gender diversity
Only Nationality diversity
All diversity measures together
Educational diversity 0.075 (0.218) -0.067 (0.233) Gender diversity 0.711*(0.314) 0.846** (0.326) Nationality diversity -0.121(0.290) -0.374 (0.325) Chemical industry 0.026 0.011 0.025 0.028 Pharmaceutical industry 1.214*** 1.216*** 1.302*** 1.148***Electronical industry 0.573*** 0.640*** 0.571*** 0.515***Mechanical Engineering -0.350** -0.184 -0.319** -0.422***Automotive industry 0.701** 1.218*** 0.796** 0.753** ICT sector 0.546*** 0.551** 0.572*** 0.637***Knowledge intensive Services
0.835*** 0.791*** 0.862*** 0.841***
< 100 Employees 0.983*** 0.928*** 0.894*** 0.898***100-249 Employees 0.222 0.177 0.154 0.142 250-499 Employees -0.148 -0.192 -0.172 -0.093 Firm Age -0.008** -0.009** -0.009** -0.008* Constant -2.730*** -2.740*** -2.627*** -2.837***R-squared 0.228 0.223 0.220 0.243 No. 1053 906 1074 904
Ghent, 21 September 2016
» Positive relationship between gender diversity and the firm’s innovative efficiency
» No significant relationship between other diversity measures
THE DIFFERENCE MAKES A DIFFERENCE: TEAM DIVERSITY AND INNOVATIVE CAPACITY 20
DISCUSSION
Ghent, 21 September 2016
» We find empirical evidence that firms that employ researcher teams with a higher degree of diversity – firms with more foreign, female, non-STEM researchers – have a better innovative capacity• Firms that employ more female researchers have significantly higher
future innovativeness and innovative efficiency• Firms that employ more non-STEM and more foreign researchers have
significantly higher future innovativeness but not a higher innovative efficiency firms conduct more often basic and applied research instead of
product development - which often does not directly lead to new and innovative, marketable products
» To achieve causality, the quantitative analyses needs the transformation event on the firm level over time, the before and after conditions, to investigate the impact of the event “diversity” and a sufficient time lag to develop influence.