http://orcid.org/0000-0003-4400-089X
![Sample our Education journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5932/TOC-Subject-Sample-2021-Education.jpg"})
ABSTRACT
In this study, we analyse the relationship between competition and programmatic diversification in 75 Italian universities from the academic years of 2003/2004 to 2011/2012. Results show that local competition, rather than national competition, influences programmatic diversification. The relationship between local competition and programmatic diversification is found to be quadratic such that when competition increases, diversification decreases and specialisation increases, and both relationships are reversed after a certain threshold. We argue that under moderate levels of competition, universities tend to respond to local competition for students by differentiating their programmatic offerings from their competitors. However, when the level of competition is minimum or extreme, universities tend to follow an isomorphic strategy. After the reduction in student demand and the reform of the higher education system in 2008/2009, the relationship was no longer curvilinear because universities operating in extremely competitive environments began to adopt more risk-taking behaviour by engaging in diversification strategies.
Acknowledgements
We wish to thank Andrea Bonaccorsi, Giliberto Capano, Daniele Checchi, Cinzia Daraio, and John Aubrey Douglas. The authors also thank the participants at the Regional Science Association European conference, 2015, in Piacenza, the participants at the 29th CHER Annual Conference, 2016, in Cambridge, and those attending the AiIG conference, 2016, in Bergamo.
Disclosure statement
No potential conflict of interest was reported by the authors.
ORCID
Mattia Cattaneo http://orcid.org/0000-0003-4400-089X
Paolo Malighetti http://orcid.org/0000-0002-2655-9860
Michele Meoli http://orcid.org/0000-0002-9438-0782
Notes
1 Source EUA: http://www.eua.be/activities-services/projects/eua-online-tools/public-funding-observatory-tool.aspx (accessed on 20 December 2016).
2 Source MIUR: http://hubmiur.pubblica.istruzione.it/web/universita/home (accessed on 10 February 2017).
3 This trend is similar to that of other countries (e.g. Farhan Citation2017 for Ontario, Canada; Erhardt and von Kotzebue Citation2016 for Germany).
4 The key issue is to know more about how diversity, diversification, and competition affect one another. On this topic, a consensus exists from the perspective of business and ecological systems concerning the association between diversity and competition. Most economic thinking suggests that moderate rather than intensive competitive dynamics can be more effective in improving the performance of businesses and markets. Assuring moderate competitive dynamics highlight the need to maintain balanced levels of economic diversity (also in terms of the market structure) and underline the key role of niches in stimulating diverse product and service offerings (e.g. Tisdell Citation2013). Ecological systems literature is aligned mostly with this perspective in underlining the key role of ecosystems’ diversity in balancing competition for resources which equilibrates biodiversity and fosters complementary and synergetic inter-species dynamics, thus guaranteeing ecological stability and bio-sustainability in complex evolving ecosystems (Loreau and de Mazancourt Citation2013). Further economic literature shows that while moderately competitive markets promote diversity, overly competitive markets dampen diversity because it constrains the actors to act in a similar fashion, offering products and services that assume similar characteristics, without necessarily adding value (van der Wurff and van Cuilenburg Citation2001).
5 Minelli, Rebora, and Turri (Citation2012) describe the teaching environment in terms of freedoms. The form of freedom mostly affected by the recent reforms of the Italian teaching framework is the ‘freedom to specify the product’, given that all new degree courses have to respect a minimum number of ex-ante requisites, such as the number of teaching staff in each course and the number of seats in lecture halls (see also Rebora and Turri Citation2011, 539).
6 The analysis is limited to 2011/2012 because it represents the last year from which disaggregated Ministerial data on the programmatic offerings of each university are available.
7 Our results are robust when weighting the fields by the number of courses per field of education as in Rossi (Citation2009).
8 Although the coordinates of the official location of universities are used, some universities have campuses located elsewhere. The different campuses are, however, relatively close to the main campus (the average distance is less than 6 km). The only exception is the Catholic University of the Sacred Heart, which has the main campus in Milan and an additional four campuses in other cities in Italy.
9 Two alternative measures for university research productivity were also tested: (1) the ratio between the total number of published articles collected from the Scopus dataset per university and (2) the ratio between the total number of published articles collected from the Scopus dataset by the total number of academic staff. The replacement of the current metric used for research orientation by either of the other measures did not alter the current findings (results are available upon request).
10 Fixed-effects estimations qualitatively support our results. The results are available on request.
11 Due to the similarity of the results, the regressions that only test the linear relationship between competition and differentiation are not included in this study. Tables are available upon request.
12 Overall, Italian universities changed their programmatic offering during 2004–2012 by introducing a total of 85 new fields and withdrawing seven. For example, the University of Reggio Calabria introduced new courses in two different fields, art and social and behavioural science, and discontinued courses in business administration, while the University of Basilicata increased its programmatic offerings by adding courses in four different fields: business and administration, health, personal services, and education.
13 As the linear terms of competition are not significant, we do not report them in the table.
14 To test the roboustness of our specifications, we also run the model excluding the two variables having the highest VIF: value added per capita (5.27) and quality of life (7.94) even if these values are below the cut-off point of 10, a commonly known rule-of-thumb used in the literature (e.g. Craney and Surles Citation2002; Dormann et al. Citation2013). All coefficients keep the same sign, and only a minimum impact on magnitude and significance is found. Therefore, the inclusion of the two variables, even if affecting slightly the model efficiency, does not change any of the outcomes, but contributes to better control for confounding effects related to potentially alternative explanations.
15 When using university diversification as a dependent variable, the local competition variable explains 13% of the variance (R2 of a reduced model with no control variables, which is 55% in the full model). When using local specialisation as a dependent variable, the variance explained is equal to 9% (51% in the full model). By contrast, and considering the insignificance of the competition variables in the regressions, national competition only accounts for 3% of diversification variance and 2% of national specialisation variance.
16 We rely on the Factiva news media database to collect the number of pieces published in each sample year using university name as search criterion. This number is then scaled by the maximum value (obtained by Sapienza University of Rome in 2008). This provides a measure of public endorsement from press citations in local, national, and international newspapers (Desai Citation2008). Since some print articles may also discredit universities by reporting cases of corruption and scandals, we randomly select 10% of the university-year articles to have an idea of the frequency with which this occurs. On average, only 1% of this fraction reports unfavourable news.