The backlash against migration around the globe continues to strengthen and, in some cases, move beyond anti-migrant rhetoric to affect actual policy. These policies could have negative medium- and long-term impacts in economies worldwide.
In the US, for example, it has become much harder to obtain a work visa, with data from the US Department of Homeland Security showing that the denial rate for H-1B visa petitions (for skilled immigrant workers) more than doubled in 2018 from 7% to 15% (USCIS 2019). Similar trends are emerging among students too, with foreign enrolment at US universities declining year-on-year as students look to Canada and other alternative destinations.1 The aggregate economic losses resulting from falling rates of skilled migration could be vast.
Innovation – one of the most important engines of economic growth – stands to be strongly negatively affected by these trends. Many noted economists have used quantitative studies to demonstrate the linkages between skilled migration and innovation.
Perhaps the best summary of this research is Bill Kerr’s book, The Gift of Global Talent (Kerr 2018), in which he shows – among other findings – that the share of US patenting by inventors of Chinese and Indian ethnicities combined grew from about 3% in 1975 to nearly 20% in 2015. Recent work by Choudhury and Kim (2019), Doran and Yoon (2019) and Burchardi et al. (2019) also shows the positive impacts of immigration on US innovation, both historically and today.
Immigrants are not only an important determinant of increasing innovation, but also of the type of innovation countries produce. Our most recent work (Bahar et al. forthcoming) presents systematic evidence – based on a 95-country sample spanning several decades – of the role that migrant inventors play in boosting innovation in the specific technologies in which their home countries specialise, as measured by patenting activity.
Our main results suggest that migrants – and in particular, migrant inventors – import knowledge that reshapes patenting activity in the destination country, moving it towards new technologies in which the destination country has no prior patenting activity.
In a sense, our paper generalises the specific historic episode studied by Moser et al. (2014), who find that chemistry-related innovation in the US was positively affected by an inflow of German and Austrian inventors fleeing Nazism before and during WWII.
Specifically, we ask: do migrant inventors boost patent production in their countries of destination (origin) in the same technology classes in which their home (receiving) countries specialise? We focus on what we call ‘technology take-offs’ – episodes where a country has no patent filings in a given technology at the start of a decade, but by the end of the decade, patent filings in that technology (as a share of total patents) outpace global filing rates.
Our main results are based on two data sources:
- First, annual country-level patenting activity data, as compiled by the OECD and reported by the US Patenting Office, for 651 four-character technology subclasses, as defined by the International Patenting Classification.2
- Second, bilateral data on migrant inventors compiled by Miguelez and Fink (2017), which measures the presence of foreign inventors from every country of origin in every country of destination.
Our results are illustrated in Figure 1, which plots – using raw data – the unconditional probability of a technology take-off in any country when the stock of immigrant inventors in that country from other countries that specialise in that same technology is high or low (above and below median, respectively).
To be more precise, the figure shows a country that hosts a larger number of migrant inventors from nations that specialise in a particular technology at the beginning of a decade is more likely to experience a technology take-off in that technology by the end of that decade. For example, a country with a higher stock of Israeli inventors is more likely to start innovating in water technologies (Israel being a country that innovates in water-related technologies).
The figure also shows analogous results for average growth rates in patenting in a technology for which the same pattern as in take-offs does not hold.
Figure 1 Probability of take-off and growth by intensity of migrant inventors
Notes: This figure presents the average probability of a patent technology take-off and the average CAGR (using patent applications) for country-technology pairs with a stock of immigrant inventors from countries that specialize in that same technology (e.g., file patents in that technology subclass with an RTA above 1) below and above the sample median. The stock of immigrants is scaled by population size of the receiving country. The figure is based on simple averages, with no controls whatsoever.
Our more rigorous estimates, based on two different set of instrumental variables to deal with endogeneity concerns, find that countries are 25-60% more likely to gain advantage in patenting in certain technologies within a decade given a twofold increase in the number of foreign inventors from other nations that specialise in those same technologies. For the average country in our sample, this number corresponds to only 25 inventors and a standard deviation of 135. We find no effect of take-offs when a country has emigrants in other nations with technological advantage in a given technology. We also don’t find any results when measuring growth of patenting activity in a given technology (instead of take-offs).
These baseline estimates control for the volume of trade and foreign investment to and from the countries of origin of the migrant inventors (to control for the possibility that these flows could be an alternative pathway for knowledge diffusion).
Our main identification strategy relies on two sets of instrumental variables to deal with the stock of current immigrant inventors. The first is the 30-year lagged presence of immigrants from and emigrants to the same countries of the migrant inventors.
Second, separately, we estimate the specification instrumenting for immigrant inventors using the estimated number of immigrants based on push and pull factors similarly to Burchardi et al. (2019). Our results are shown in Figure 2, which shows our estimators based on the take-off dependent variable are similar in magnitude regardless of the instruments used.
Figure 2 Technology take-offs: OLS versus 2SLS estimators
Notes: This figure plots the point estimates and their corresponding 95% confidence intervals (represented by whiskers) of the OLS and two different 2SLS estimations for both βIM and βIM, based on the results presented in Panel A of Table 2.
We also run a ‘placebo test’ to make sure our results are driven by migration flows and not by other factors that could be at play. To do so, we reconstruct the number of migrant inventors based on random distributions, using two approaches.
First, we reshuffle the bilateral inventor migrants matrix in every year such that it maintains the same mean and distribution (‘Random Model 1’); second, we simply assign a random variable, uniformly distributed between 0 and 1, to every pair of countries and treat that as the number as the stock of migrant inventors (‘Random Model 2’).
Using ‘random’, instead of real, numbers for migrant inventors, and performing 500 iterations of this randomisation, we find that results disappear (in terms of statistical significance) in 45-85% of the cases. These results are summarised in Figure 3.
Figure 3 Summary of 500 estimations using random inventor figures (OLS)
Notes: This figure plots the estimators of βIM (left panel) and βEM (right panel) when substituting the real number of migrant inventors between countries with a random one, for each of 500 iterations. The results are based on an OLS estimation that includes both regressors simultaneously. The upper row is based on a randomisation approach such that the real and the random number of inventors have the same sample mean and distribution. The lower panel is based on a randomisation approach that replaces the actual number of inventors with a random number, with no restrictions whatsoever, distributed uniformly from 0 to 1. The figure also includes, for reference, the estimators using the actual number of migrant inventors, reported in our main results, and marked with a diamond-shaped symbol. Whiskers represent 95% confidence intervals, based on SE clustered at the country level.
Our overall findings show that migrant inventors can play an important role in shaping the patent production of their destination countries. Arguably, these dynamics – driven by migrant inventors – can also affect broader economic outcomes, given the secondary effects of patenting and innovation on productivity and firm performance. Hence, our study serves as yet another piece of evidence of the significant positive impacts of migration on overall economic growth – impacts that persist over both the medium and long term.
Bahar, D, P Choudhury, and H Rapoport (forthcoming), “Migrant Inventors and the Technological Advantage of Nations”, Research Policy (Special Issue on Migration and Innovation).
Buchardi, K, T Chaney, T A Hassan, L Tarquinio and S J Terry (2019), “Immigration, Innovation and Growth”, mimeo.
Choudhury, O and D Y Kim (2019), “The Ethnic Migrant Inventor Effect: Codification and recombination of knowledge across borders”, Strategic Management Journal 40(2): 203-29.
Doran, K and C Yoon (2019), “Immigration and Invention: Evidence from the Quota Acts”, mimeo.
Kerr, W R (2018), The Gift of Global Talent: How Migration Shapes Business, Economy & Society, Stanford University Press.
Miguelez, E and C Fink (2017), “Measuring the International Mobility of Inventors: A new database”, The International Mobility of Talent and Innovation: New Evidence and Policy Implications 8: 114–161.
Moser, P, A Voena, and F Waldinger (2014), “German Jewish Émigrés and US Invention”, American Economic Review 104(10): 3222-55.
USCIS (2019), 2018 USCIS Statistical Annual Report, US Citizen and Immigration Services.
2 One example of a four-character technology subclass in our sample is A63C, which corresponds to patents related to skates, skis, water-shoes, roller skates, courts and rinks.