How do you boost scientific collaboration by 50%? Clue: it's nothing to do with funding

Scientific papers are increasingly the result of the work of teams of researchers. By 2000, more than 80% of science and engineering articles were coauthored, compared to 50% in 1960 (Wuchty et al. 2007). Scientific collaborations are also more likely to span geographically distant coauthors (Adams et al 2005). Similar patterns have been observed in economics research (Gaspar & Glaeser 1998, Hamermesh & Oster 2004, Rosenblatt & Mobius 2004).

One possible explanation for these findings is the dramatic reduction in communication costs brought about by the internet. Tools such as email, videoconferencing, file sharing and syncing greatly reduce coordination and collaboration costs over distance. Research by Agarwal & Goldarb (2008) shows how the adoption of bitnet, a predecessor of the internet, led to a rise in collaboration in the field of electric engineering between the universities involved. Interestingly, in their data the effect is stronger for co-located pairs, suggesting that communication technology may be a complement rather than a substitute to face-to-face interactions.

In a recent working paper, we investigate an alternative (though complementary) explanation for the rise in scientific collaboration over distance – the reduction in air travel costs (Catalini et al. 2016). Over the last 30 years, the cost per mile flown in the US dropped by more than 50% (Perry 2014). To the extent that face-to-face interactions are important to transmit complex knowledge and foster trust among scientific teams, cheaper air travel fares may have amplified the effect of cheap communication costs.

To explore the effect of air travel costs on scientific collaboration, we leverage the differential timing of entry of Southwest airlines, a low cost carrier, across different US cities. After Southwest Airlines enters a new route, not only do prices drop on average by 20%, but we also observe a 50% increase in scientific collaboration. The effect is present across different fields of science, and is stronger when weighting scientific output by its future impact – that is, the additional collaborations induced by the lower fares are not of marginal quality.

Using a fine-grained dataset on US faculty members in chemistry departments (1991-2012) and a differences-in-differences research design, we find that the timing of the effect is consistent with a causal interpretation.

Figure 1

The availability of low-cost flights had uneven effects across types of scientists – researchers that are younger, have less access to funding, and that are more productive than their department average seem to benefit the most from the lower fares.

Taken together, our results suggest that cheaper flights may have reduced the frictions normally associated with geography by potentially allowing for better matches over distance.

References

Adams, J D, G C Black, J R Clemmons and P E Stephan (2005) “Scientific teams and institutional collaborations: Evidence from US universities, 1981-1999”, Research Policy, 34(3): 259-285.

Agrawal, A and A Goldfarb (2008) “Restructuring research: Communication costs and the democratization of university innovation”, American Economic Review, 98(4): 1578-90.

Catalini, C, C Fons-Rosen and P Gaule (2016) “Did cheaper flights change the direction of science?” CEPR Discussion Paper 11252.

Gaspar, J and E L Glaeser (1998) “Information technology and the future of cities”, Journal of Urban Economics, 43(1): 136-156.

Hamermesh, D and S Oster (2002) "Tools or toys? The impact of high technology on scholarly productivity", Economic Inquiry, 40(4): 539-555.

Perry, M (2014) “The cost of air travel in the US has been remarkably stable for the last decade, and 17% cheaper than 20 years ago" Carpe Diem, Blog of American Entreprise Institute, Accessed 17 March.

Rosenblatt T and M Mobius (2004) “Getting closer or drifting apart?” Quarterly Journal of Economics, 119(3): 971-1009.

Wuchty, S, B F Jones and B Uzzi (2007) “The increasing dominance of teams in production of knowledge”, Science, 316(5827): 1036-1039.