Inventor mobility and productivity in Italian regions

ABSTRACT

This paper describes the interregional and international mobility of inventors in Italy and estimates its impact on total factor productivity (TFP) at the regional level for the period 1996–2011. A new database of mobile inventors is constructed and, using a set of geographically based instruments to address endogeneity, it shows that inventor in- and outflows affect regional TFP growth. Moreover, the positive effects of the inventors’ mobility (inflow) between different applicants take more time to materialize (relative to movements within the same company). Finally, the negative effects of inventor outflows are mainly driven by mobility between applicants.

KEYWORDS:

• inventor disambiguation
• inventor mobility
• migration
• economic growth
• regional total factor productivity (TFP)
• Italy

JEL:

• J61
• O30
• O47
• R23

DISCLOSURE STATEMENT

No potential conflict of interest was reported by the authors.

ORCID

Riccardo Cappelli http://orcid.org/0000-0001-7917-5142

Dirk Czarnitzki http://orcid.org/0000-0001-5118-7272

Fabio Montobbio http://orcid.org/0000-0001-9917-2344

Notes

1. Related to this there is the idea that productivity increases because of improved working conditions. For example, Clark, Georgellis, and Sanfey (1998), using the German data of the Socio-Economic Panel, show that workers are more likely to leave when they are not satisfied with their jobs.

2. For an analysis of the impact of cultural diversity on productivity, see, for example, Alesina, Harnoss, and Rapoport (2013).

3. The empirical model of this paper is also in line with the technology-gap approach (e.g., Fagerberg, 1988), which considers regional economic (or productivity) growth as driven primarily by innovation and takes the distinction between the development of new knowledge in a region and the diffusion of knowledge between regions.

4. This paper uses a balanced panel data set. TFP growth data refer to the period 1996–2011, while data on the lagged independent variables refer to the period 1995–2010.

5. Geographical distance (km) is calculated as the great circle distance between regions’ centroids.

6. All the variables measured in euros are expressed at constant prices (reference year: 2005).

7. Data on total hours worked are not available at a regional level.

8. Precision = true-positive/(true-positive + false-positive); Recall = true-positive/(true-positive + false-negative).

9. The constructed database contains 52,696 inventors, of which 33,459 are one-patent inventors. Since this paper considers inventors with a minimum of two patents, the computed inventor mobility measures do not capture potential movements of one-patent inventors. In general, the authors recognize that the computed inventor mobility indices underestimate the real inventor flows, and overall the interregional flows of high skilled people.