Reduced R&D investments: A potential flip side of immigration

Torje Hegna, Karen Helene Ulltveit-Moe 31 August 2021

a

A

Research and development (R&D) is a key driver of productivity growth. Understanding the impact of economic integration on firms’ incentives to invest in R&D remains a key question in economics. There is an extensive literature investigating the role of trade for innovation, and Melitz and Redding (2021) highlight the key mechanisms through which trade works. But trade is just one aspect of economic integration. Another aspect that may affect firms’ incentives to invest in R&D has received relatively little attention: immigration. The scant literature on the relationship between innovation and immigration has focused on the impact of highly skilled immigrants on innovation (Kerr 2018). Notable exceptions include Lewis (2011) and Gray et al. (2020). 

Theory does not provide an unambiguous answer to the question of how an immigration shock can be expected to affect R&D. Hence, whether and to what extent immigration has an impact on firms’ investments in R&D remains an empirical question. In a recent paper (Hegna and Ulltveit-Moe 2021) we therefore use a natural experiment to examine how firms adjust to a significant immigration shock through their investments in R&D, and the potential implications for productivity growth.

Our empirical analysis is motivated by the fact that in the aftermath of a large and sudden immigration shock, R&D investment in the business sector in Norway declined. Over the same period, R&D investment (measured as share of GDP) for the group of OECD countries increased (OECD 2021). Hence, the development in Norway differs from that in many other advanced countries. Our hypothesis is that the immigration shock contributed to the decline in R&D investments.

Following the Eastern enlargement of the EU in 2004, which extended the common European labour market to include roughly 100 million individuals from the EU accession countries, Norway experienced a large wave of immigration (Bratsberg et al. 2019). With real wages among the highest in Europe, and unemployment among the lowest, Norway became a popular destination for labour migrants. Norway is a member of the European Economic Area (EEA) and therefore part of the EU single market. Before 2004, accession-country citizens had very limited access to the Norwegian labour market. After 2004, Norway had few transitional restrictions on immigration from accession countries compared to most EU countries. Over the next four years, the number of labour migrants quadrupled, and ten years later the share of migrants in employment had risen from 7% to 17%. The policy change was instant, comprehensive, and externally imposed, providing a unique setting to study the impact of immigration.

Empirical approach

We conduct an empirical analysis that relies on two key observations. First, the sudden influx of labour to Norway varied substantially across occupations, with some occupations receiving many immigrants while others were hardly affected. Second, even within industries, firms differ in the occupational mix of their labour force. One may therefore expect them to be differently affected by the immigration shock. Using high-quality and detailed administrative Norwegian employer-employee data, combined with survey data on firms’ R&D investments, we exploit the variation in exposure to the shock across firms to identify the impact of the immigration shock on R&D. An important advantage of our firm-level approach is that we are able to sweep out all industry trends in innovation by fixed effects.

A common concern when analysing the impact of immigration relates to the endogeneity of new immigrants’ choices of occupation and location. Immigrants typically sort into occupations based on current and future demand shocks. To tackle this endogeneity problem, many studies have relied on historical settlements across regions to identify the effects of immigration. Such approaches exploit the ‘ethnic enclave’ design commonly associated with Altonji and Card (1991). However, due to the East-West divide in Europe, the number of Eastern European migrants to Norway was very low before 2004. Hence, past settlements turn out to be a poor predictor. 

Rather than relying on past settlements of migrants to forecast the immigration shock, we build on Bratsberg et al. (2019) and exploit the fact that language requirements constitute significant barriers for foreign workers, and that these requirements typically vary across occupations (see e.g. Peri and Sparber 2009). Combining occupation-specific information on language requirements with firms’ occupational mix in the pre-shock period, we compute a firm-specific language intensity which we use to forecast the immigration shock exposure. These exogenous occupation characteristics turn out to be a powerful predictor of immigration flows across occupations.

Trends in R&D

Aggregate numbers on R&D in the business sector in Norway show that R&D investments, measured as a share of value-added, declined in the period immediately following the Eastern enlargement. We split firms into two groups – high and low exposure – depending on whether the firm has above or below mean exposure to the immigration shock based on the firm-specific measure of language intensity. We proceed by computing firms’ R&D investments for the pre-shock (2003/2004) and post-shock (2006/2007) periods. The micro data confirms the aggregate numbers, and an overall decline in R&D investment after the immigration shock. Moreover, it appears that the median high-exposure firm experienced a decline in R&D investments of 33%, which is almost three times the decline of that of the median low-exposure firm, which experienced a decline of 12%. 

Empirical results

To analyse the causal effect of the immigration shock on firms’ R&D investments, we use a difference-in-difference approach wherein we exploit firms’ heterogeneous exposure to the shock. We find that the immigration shock had a significant negative impact on firms’ investment in R&D. The most exposed firms reduced their R&D investments in the aftermath of the immigration shock. Hence, it appears that the overall decline in R&D investments in the business sector can be partly explained by the immigration shock.

What were the economic magnitudes of the immigration shock? In order to evaluate the economic impact of the shock on R&D, we split firms into percentiles according to their exposure to the increase in immigration penetration and compare the 10th and the 90th percentile of firms. We find that the most exposed firms experienced a decline in R&D growth that was more than six times that of the least exposed firms.

Theoretical and empirical research has documented a strong positive impact of the role of investments in R&D on firms’ productivity. Hence, we would expect that any negative shock to R&D would consequently also have a negative impact on firms’ productivity. As we proceed by investigating the medium to long-term impact on productivity, we find that the immigration shock had a significant negative impact on productivity in the medium to long run. Although our analysis does not allow us to draw conclusions about a causal relationship between the immigration-induced decline in R&D investments and the reduced productivity growth, the results indicate that there is a link.

It is worth noting that a previous study (Gray et al. 2020) finds a positive relationship between the Eastern enlargement immigration shock and firms’ innovation activity. Our results are robust to a set of econometric issues and stand in distinct contrast to their results. Our empirical analysis suggests that the flip side of a vast increase in relatively unskilled labour supply may be a reduced incentive to conduct R&D with long term negative consequences for productivity.

References

Altonji, J and D Card (1991), “The Effect of Immigration on the Labor Market Outcomes of Less-Skilled Natives”, in Immigration, Trade and the Labor Market, University of Chicago Press.

Bratsberg, B, A Moxnes, O Raaum and K H Ulltveit-Moe (2019), “Opening the Floodgates: Industry and Occupation Adjustments to Labour Migration”, CEPR Discussion Paper No. 13670.

Bratsberg, B, A Moxnes, O Raaum and K H Ulltveit-Moe (2019), “Opening the floodgates: Immigration in the aftermath of the eastern enlargement”,  VoxEU.org, 9 May.

Gray, R, G Montresor and G C Wright (2020), “Processing Immigration Shocks: Firm Repsonses on the Innovation Margin”, Journal of International Economics 126.

Hegna, T and K H Ulltveit-Moe (2021), “Reduced R&D Investments: A Flip Side of Immigration?”, CEPR Discussion Paper No. 16325.

Kerr, W (2018), “High-skilled immigration and the growing concentration of US innovation”, VoxEU.org , 26 October.

Lewis, E G (2011), “Immigration, skill mix and capital skill complementarity”, Quarterly Journal of Economics 126: 1029–1069.

Melitz, M and S Redding (2021), “Trade and innovation”, VoxEU.org, 28 July.

OECD (2021), “Main Science and Technology Indicators”, https://stats.oecd.org/.

Peri, G and C Sparber (2009), “Task Specialization, Immigration and Wages”, American Economic Journal: Applied Economics 1(3): 135–169.

a

A

Topics:  Migration Productivity and Innovation

Tags:  immigration, R&D, research and development, R&D investment, Norway

Research Fellow and PhD student, University of Oslo

Professor of International Economics, University of Oslo; and CEPR Research Fellow

Events

CEPR Policy Research