Defending globalisation: Isolation would cost us dearly

Klaus Desmet, Dávid Krisztián Nagy, Esteban Rossi-Hansberg

30 November 2016



The process of globalisation, so lauded until recently, is under attack. Many people blame international trade and migration for the stagnation of their incomes and the lack of good job opportunities. Politicians’ views reflect this discontent—neither of the two candidates in the US presidential election supported expanding free trade. The winner, Donald Trump, campaigned on a platform of renegotiating NAFTA and raising tariffs on China. Migration gets even more negative press. Listening to the 1980 primary debate between the Republican contenders, with Reagan and Bush Sr competing on who wanted more open borders, is a far cry from the current discussion on immigration reform. In the UK, Brexit won on the promise of regaining control over Britain’s borders. Today, the idea that migrants are unequivocally good for a country’s development has all but vanished from the political and economic debate. At best, people advance arguments on why it might be beneficial to let in highly educated professionals, such as Indian programmers, or to allow some students to stay after their postgraduate studies. At worst, migration elicits xenophobia and racism.

The dismal view of globalisation

What once sounded like the desperate cries of a few extremist commentators against the indomitable force of globalisation has now become increasingly mainstream. The economics profession has been no stranger to what has perhaps become the most dangerous paradigm shift of our time. Some economists argue that the increasing flow of goods and migrants have contributed to the decline in well-being of less-educated workers in the manufacturing regions of developed countries (e.g. Rodrik 1997, Autor et al. 2013, Borjas 2014). Others say that the evidence is mixed and requires further scrutiny (e.g. Peri 2016). The consensus about the benefits of globalisation has floundered; it has become increasingly hard to find economists who unambiguously favour free trade and free migration.

The misleading short run

We sharply disagree with this dismal view of globalisation. Our recent research indicates that the gains from trade and migration are tremendous and that the world stands to benefit greatly from their further liberalisation (Desmet et al. 2016). The problem with virtually all quantitative and empirical evaluations of trade and migration is their static nature. They completely ignore the dynamic gains from globalisation. As we will later discuss, these dynamic gains quantitatively dwarf any short-run costs.

The dynamic argument in favour of trade is simple – innovation requires scale, and scale requires trade. The incentives of firms and individuals to invest effort, time, and resources in generating new ways of producing valuable goods and services are intrinsically related to their ability to use the resulting knowledge repeatedly. This replicability underlies the scale economies that form the basis of much of modern growth theory. At no point in history have these forces been more evident than now. Ideas generate products that are replicated millions, or even billions, of times in a matter of a few years, and companies can go from just a few programmers in a garage to a multi-billion dollar business in a decade. Clearly, the ability of firms to sell the goods and services they invent and improve depends on their capacity to reach their costumers cheaply and efficiently. The result is an obvious, and often tragically neglected, link between trade and growth. In a recent interview with the Financial Times, Bill Gates, founder of Microsoft, echoes this simple insight: “I wish for a week that we could shut down trade and then Boeing, Microsoft, Hollywood, pharma would resize [downsize] their R&D departments…  people would go ‘Holy smokes, that was not a very good deal’.”

Society’s suspicion towards immigration is likewise distorted by a static, short-run view. When foreigners enter a country, they use up land and other fixed factors, and get access to public goods and services. The result is more congestion, as fixed resources have to be shared between more people. If immigrants do not fully pay for these fixed factors and public services, the autochthonous population loses. Even if the foreign residents are made to foot the bill, the citizens who do not own part of the fixed factors still suffer, as their marginal productivity gets eroded and their income declines. The basic economic logic of this argument is sound, but its premise is flawed – fixed factors are less fixed than they appear, and technology evolves endogenously, partly as a result of immigration.

Many fixed factors, such as land, are not fully utilised. The price of land anywhere in Montana is lower than in virtually every block of Manhattan. There is plenty of space in the US to create many more cities like New York. Hence, the lack of space, at least in the US, cannot be a fundamental factor limiting the productivity of migrants and other workers. As for technology, we know that innovation flourishes in dense locations, partly because of spillovers between workers and partly because of stronger local demand. If people can move to the more productive regions of the world, those areas will become denser, spurring more future innovation and growth. Though there might be some discussion in the short run about the adequate skill level of immigrants, in the long run this is a secondary concern – the human capital, savings, and investment decisions of the descendants will be shaped by the country where they live.

One thing is clear, measuring the static effect that immigrants have on the wages of other similarly skilled agents in the same location misses a tremendous number of effects. What if current residents move to other cities? What if they leave to the suburbs instead? What happens in neighbouring places, and in the immigrants’ regions of origin? How do the incentives to innovate in different cities and locations change? Locations within a country and across the world are linked via trade, migration, and production; when people migrate, the geographic distribution of production changes, affecting both trade flows and innovation in different locations. The static and partial equilibrium measurement and logic that treats locations as isolated islands is flawed, and will result in erroneous answers. In economics, it is unfortunately fairly common to focus on certain details and to forget about the big picture. In this context, such an approach is particularly perilous, as it runs the risk of shaping arguably the most important policy debate of our time.

Quantifying the long-run gains from globalisation

It is of course not enough to say that there are dynamic effects of globalisation that are potentially important. Ultimately, to evaluate whether the dynamic gains compensate any static losses, there is a need to measure the different forces and quantify their effects. Our recent work provides a dynamic spatial growth model that incorporates trade and migration frictions, both within and across countries (Desmet et al. 2016). We split the world up into 64,800 grid cells of 1º by 1º (roughly speaking, 100km by 100km). Using large amounts of data, we calibrate the model and run it forward. The goal is to assess the evolution of the world economy over time and over space, under different assumptions on migration frictions and trade costs.

The results are stark. Completely lifting all migration restrictions would increase real world output by 126% in present discounted value terms. Since such a policy may be unrealistic, consider instead a reform that liberalises migration so that 10% of the world population moves at impact. This would yield a present discounted value increase in real world output of 14%. Such a reform would cause some extra congestion in Europe and the US, implying that average welfare would increase by 9%, a smaller but still impressive figure. It is hard to think about any other policy that could readily be applied at the world level for which estimated benefits are as large. Migration is uniquely powerful in generating positive effects. In economic terms, having an open-door policy is a no brainer, not because of some abstract theoretical arguments, but because the measurement of the relevant forces tells us so.

Turning back the clock on trade would have equally dire consequences. Increasing trade costs by 40% would lower real world output by 30% in present discounted value terms. Although globalisation might create losers in the short run, allowing the free flow of goods and people across regions and countries is still one of the best ways we know to ensure our long-run wealth and well-being. If we worry about secular stagnation, there is no other policy that would be as effective in halting the productivity slowdown. In times of protectionist furore, and with anti-immigrant sentiment running high, we would do well to think about this before plunging ourselves into the abyss.


Autor, D, D Dorn and G Hanson (2013) “The China syndrome: Local labor market effects of import competition in the United States,” American Economic Review, 103(6): 2121-2168.

Borjas, G (2014) Immigration Economics, Harvard University Press.  

Desmet, K, D Nagy and E Rossi-Hansberg (2016) “The geography of development”, Journal of Political Economy, forthcoming, and CEPR Discussion Paper 10544.

Peri, G (2016) “Immigrants, productivity, and labor markets”, Journal of Economic Perspectives, 30(4): 3-30.

Rodrik, D (1997) Has globalization gone too far?, Washington, DC: Institute for International Economics.



Topics:  International trade

Tags:  globalisation, anti-globalisation, immigration, free trade, isolation, innovation

Altshuler Professor of Cities, Regions and Globalisation, Southern Methodist University; CEPR Research Fellow

Junior Researcher, Centre de Recerca en Economia Internacional (CREI)

Theodore A. Wells '29 Professor of Economics, Princeton University