Virtual Migration Could Revive Growth

Mar 16 , 2025
By Carl Benedikt Frey

---

The rise of AI ushers in an “Uber moment” for a wide range of professional services, writes Carl Benedikt Frey, associate professor of AI & Work at the Oxford Internet Institute and director of the Future of Work Program at the Oxford Martin School. In this commentary provided by Project Syndicate (PS), he argued that as virtual migration reduces the need for physical relocation, the world can look forward to a future where skill and talent, not location, determine an individual's earning potential.

---

While technological advances have fueled prosperity for centuries, their benefits remain unevenly distributed, leaving nearly half the world’s population living on less than 6.85 dollars a day. Yet now, emerging tools like generative artificial intelligence (GAI) and remote-work platforms could finally begin to address this historic imbalance, by allowing “virtual migrants” to access high-income opportunities without leaving their home countries.

The disparity in average wages across countries can be striking. In 2022, data from the International Labour Organisation (ILO) revealed that the purchasing power of average monthly earnings in India was 17pc of that in Germany. In Ethiopia, the figure was a mere five percent.

These wage disparities represent the world’s largest arbitrage opportunity. Economist Lant Pritchett estimates that if the 1.1 billion people identified by Gallup as willing to migrate temporarily for work were allowed to do so, their average annual earnings could rise by 15,000 dollars in purchasing-power-parity terms. The resulting economic benefit would total 16.5 trillion dollars, a sum exceeding the impact of all global development aid by over a hundredfold. Even if only a half or a quarter of aspiring migrants actually migrated, the resulting gains would be massive.

As Michael A. Clemens of George Mason University puts it, we are leaving "trillion-dollar bills on the sidewalk."

The primary obstacle to unlocking these benefits is political. A recent survey of wealthy countries found that respondents not only “greatly overestimate the total number of immigrants,” but also perceive immigrants as “culturally and religiously more distant from them, and economically weaker – less educated, more unemployed, and more reliant on and favoured by government transfers – than they actually are.”

Though there is scant evidence that immigrants depress native workers’ wages, the populist right has skillfully exploited the perception that they do, achieving substantial electoral gains across multiple countries in 2024.

At the same time, the combination of longer lifespans and historically low fertility rates has caused a dramatic increase in the ratio of elderly people to those of working age, making existing social-welfare commitments unsustainable in the absence of immigration. According to the OECD, the average fertility rate across the 38 most developed countries has fallen from 3.3 children per woman in 1960 to 1.5 in 2022 – well below the “replacement rate” (2.1) required to maintain a stable population.

One politically feasible solution might be temporary contract work, excluding family reunification and a path to citizenship for migrants in rich countries. But, while this approach may be essential to address the growing demand for eldercare (which must be done in person), virtual migration represents a much easier path for many other sectors. As services become increasingly digitised and tradable, remote work facilitated by technology offers a game-changing solution, especially now that companies like Deel and Remote.com are emerging to handle the various regulatory complexities.

In the wake of the COVID-19 pandemic, the percentage of full days worked from home in the United States soared from almost none to 60pc by the summer of 2020. Although this figure declined to approximately 30pc by 2023, it remains a major departure from the pre-pandemic norm. Companies now have many new opportunities to access affordable talent. In 2021, for example, Google announced plans to adjust employee salaries based on cost-of-living differences, effectively reducing pay for those working remotely or relocating farther from their offices.

Following the same logic, companies may pursue even greater cost savings by offshoring remote jobs to lower-income countries, cutting wage expenses and real estate overhead. One already finds stark wage disparities within multinational firms. At Google, software engineers earn an average of 128,000 dollars annually in New York City compared to 6,000 dollars in Manila. Similarly, Big Four accountants command 76,000 dollars in New York against 3,000 dollars in Cairo.

While these wage differentials reflect underlying productivity gaps across locations, AI technologies quickly become powerful equalisers, eliminating the need for physical migration. In software development, tools like GitHub’s Copilot can help boost productivity, reflected in a study conducted by GitHub and Microsoft which found that developers using the tool completed tasks 56pc faster, on average, with the biggest gains for less experienced programmers. Similarly, AI tools such as ChatGPT have been shown to enhance writing productivity, especially for novices. In customer service, AI assistants have increased productivity by 14pc, benefiting less-skilled workers the most, according to research by Erik Brynjolfsson and his colleagues.

Even language barriers are falling. As my research with Pedro Llanos-Paredes shows, improvements in machine translation are already reducing demand for foreign-language skills, and this trend will only intensify. In a recent demonstration, OpenAI’s GPT-4o model effortlessly translated speech between Italian and English in real time. Given these advances, it is no wonder that low-wage countries like India, Indonesia, and the United Arab Emirates (UAE) have the highest use of generative AI in the world.

Virtual migration is a win-win. It can help drive economic growth, address demographic challenges in developed countries, and reduce worldwide income disparities. However, it does raise legitimate concerns about wage pressures and workforce disruptions in affected industries. A useful analogy is the impact of Uber on the taxi industry. With GPS technology, detailed knowledge of city streets became less valuable, allowing less-skilled drivers to thrive. While Uber’s expansion did not reduce the number of driving jobs, it intensified competition and led to a 10pc drop in hourly earnings for established drivers.

Now, generative AI has ushered in an “Uber moment” for a wide range of professional services.

Such competition poses an even greater challenge in traded industries, where the challengers are often based in other countries. While the rise of China has lifted more than 800 million people out of poverty and reduced costs for consumers worldwide, public debate in the US has overwhelmingly focused on the fate of displaced American manufacturing workers. What usually goes unmentioned is that integration with China provided the solution to US manufacturing firms’ struggle to compete with more productive Japanese companies.

American autoworkers, for example, were three times more productive than their German counterparts in 1950. But by 1980, Japanese autoworkers had achieved a productivity advantage of 17pc over US autoworkers, while Ford and GM reported combined annual losses exceeding 1.3 billion dollars. Similarly, at its peak in 1977, the US semiconductor industry supplied 95pc of the domestic market and 57pc globally. By 1989, however, it had become a net importer, with Japan supplying 25pc of US demand. America’s global share fell to 40pc as Japan’s rose to 50pc.

To counter Japan’s emerging dominance, American corporations adopted a different strategy. Unlike Japan’s vertically integrated "keiretsu" conglomerates, newer companies like Apple developed open, non-proprietary systems that enabled them to leverage global value chains, cutting costs and enhancing flexibility. This shift led to a new era of integration between the US and China, Hong Kong, and Taiwan, providing US innovators with cheaper inputs and sharpening their competitive edge.

Of course, America has since returned to protectionism, even though the trade war with China has not delivered any of the intended benefits to the US heartland. The first Trump administration’s import tariffs on foreign goods had no significant effect on US employment in newly protected sectors, while retaliatory tariffs had negative effects on employment.

Yet, it is worth emphasising that while economists have long pointed to the costs of trade barriers, the implicit “tax” on labour in wealthy countries far exceeds even the most extreme tariffs proposed by the second Trump administration. After all, labour costs are about 400pc higher than they would be if workers from poorer countries were allowed to move freely to fill open jobs. As Clemens notes, the estimated gains from removing barriers to labour mobility represent an economic distortion of staggering scale; removing them would produce estimated gains of 50pc to 150pc of world GDP. By contrast, barriers to trade and investment account for less than a few percent of world GDP.

Unfortunately, once established, protectionist walls can become self-reinforcing by dampening economic growth. One of the most overlooked benefits of growth is that it shifts the economy away from being a zero-sum game (where improving one person’s standard of living comes at the expense of another’s). Perhaps predictably, research shows that recent generations, having experienced slower growth, are more prone to zero-sum thinking and more likely to favour stricter immigration controls.

The most effective antidote to zero-sum thinking is robust growth that drives the creation of new industries and jobs, even as older sectors shrink or migrate. Most jobs held by Americans today did not exist in 1940 – they had to be invented – while most of the jobs that did exist no longer do.

How many telephone operators, ice cutters, and pinsetters are still working in the US?

But this process of creative destruction has slowed significantly over the past decade. From the US to China, the rate of new business formation has been in decline since the 2000s. This trend is deeply concerning, considering that startups play a critical role in translating research into marketable innovations, building new industries, and generating new jobs. The dominance of larger firms in AI development, fueled by scale advantages and defensive patenting, has further curtailed market dynamism.

To unlock AI’s transformative potential and ensure sustained progress, policymakers should stop trying to protect yesterday’s jobs and instead focus on revitalising competition and reducing barriers to entry for new firms to create the jobs of tomorrow.

---

PUBLISHED ON Mar 16, 2025 [ VOL 25 , NO 1298]

---

[ssba-buttons]