The large cities in the US are the most expensive places to live. Paradoxically, this cost is disproportionately paid by workers who could work remotely, and live anywhere. The greater potential for remote work in large cities is mostly accounted for by their specialization in skilland information-intensive service industries. We highlight that this specialization makes these cities vulnerable to remote work shocks. When high-skill workers begin to work from home or leave the city altogether, they withdraw spending from local consumer service industries that rely heavily on their demand. As a result, low-skill service workers in big cities bore most of the recent pandemic's economic impact.

No abstract

Since 1980, economic growth in the U.S. has been fastest in its largest cities. We show that a group of skill-and information-intensive service industries are responsible for all of this new urban bias in recent growth. We then propose a simple explanation centered around the interaction of three factors: the disproportionate reliance of these services on information and communication technology (ICT), the precipitous price decline for ICT capital since 1980, and the preexisting comparative advantage of cities in skilled services. Quantitatively, our mechanism accounts for most of the urban biased growth of the U.S. economy in recent decades.

We offer causal evidence of higher returns to experience in big cities. Exploiting a natural experiment that settled political refugees across labor markets in Denmark between 1986 and 1998, we find that while refugees initially earn similar wages across locations, those placed in Copenhagen exhibit 35% faster wage growth with each additional year of experience. This gap is driven primarily by differential sorting towards high-wage establishments, occupations, and industries. An estimated spatial model of earnings dynamics attributes an important role to unobserved worker ability: more able refugees transition to more productive establishments faster in Copenhagen than in other cities.

A growing body of empirical research highlights substantial changes in the US economy during the last three decades. Business dynamism -namely job reallocation, firm entry and creative destruction -is declining. Market power, as measured by markups and industry concentration, seems to be on the rise. Aggregate productivity growth is sluggish. We show that declines in the rate of growth of the labor force can qualitatively account for all of these features in a standard model of firm-dynamics. Despite its richness we can characterize the link between population growth and dynamism, markups and growth analytically. When we calibrate the model to the universe of U.S. Census data, the labor force channel can explain a large fraction of the aggregate trends.