Benjamin Wache

Global data flows sustain the digital economy, from AI and cloud computing to finance, yet they are routed through a network of just a few hundred physical subsea cables. The economic consequences of disrupting this critical infrastructure have never been quantified. We make three contributions to study vulnerability and strategic dependence in the global telecommunications network: (i) GlobeTel, a new dataset covering every international submarine telecom cable laid since 1850; (ii) a traffic-assignment model with endogenous congestion and rerouting; and (iii) sufficient statistics for the welfare costs of disruption requiring only observed flow shares and latency changes, validated against traceroutes, prices, and lit capacity. We find that global data flows are far more geographically concentrated than commonly understood, passing through the same maritime chokepoints that have shaped trade flows for centuries. Severing just 15 cables in the English Channel or off the US East Coast disrupts 20 to 25% of global data flows. Island economies such as Taiwan are particularly exposed, facing large potential welfare losses despite satellite access. Using our 175-year dataset, we also show that hegemons have long exploited network dominance for coercion and surveillance, and that cable investment responds systematically to geopolitical tensions.

How does information impact the international allocation of capital? Using the global rollout of telegraph cables in the 19th century, I show causal evidence that reductions in information frictions had a significant and positive impact on the bilateral international flow of financial capital from the UK. The results indicate that the telegraph lead to more than a doubling of capital flows. For identification I use a geographic instrument, the ruggedness of the seabed. The effect of the telegraph was stronger for capital flows to businesses than to governments, and was significantly positive for flows to most industries. The telegraph had a direct and sizable impact on capital flows that was distinct from the increase in capital flows due to the telegraph's effect on trade. Using data from historical British newspapers, I show that the annual number of articles mentioning connected countries increased around the arrival of the telegraph. I interpret this as evidence that part of the mechanism through which the telegraph affected capital flows went through the news channel.

This paper examines the role of social connectedness in shaping cross-county venture capital (VC) investments. I study yearly county-to-county VC investments, and find a very large effect: a 1% increase in social connectedness is associated with a ~0.5% increase in VC investment. When controlling for social connectedness, physical distance becomes irrelevant as a determinant of VC flows. This result is robust to the addition of several bilateral controls (travel time, trade, commuting, migration). I find that early stage funding is more strongly impacted, and that counties with higher levels of social access to VC host more and better startups.