Nico Ravanilla

Working paper · July 2026 draft

Early Signs of the Generative AI Labor Market Shock in the Global South: Evidence from Philippine Business Process Outsourcing

Nico Ravanilla

PDF

In brief

Generative AI can do the work the Philippine outsourcing industry sells. Using the Anthropic Economic Index and Philippine labor force data, this paper shows employment growth is already slowing in the most AI-exposed provinces, industries, and municipalities, while wages and hours have not yet moved.

Two-panel chart. Top: indexed BPO employment from 2021 to 2025, where the most exposed tercile flattens after ChatGPT's November 2022 release while less exposed terciles keep rising past 160. Bottom: share of US firms using AI in production rising from zero to about 12 percent, annotated with major model releases.
BPO employment by tercile of provincial AI exposure (top) against the share of US firms using AI in production (bottom). After ChatGPT's release, employment growth in the most exposed provinces flattens while adoption accelerates.

Abstract

Business process outsourcing is among the developing world's largest employers, and generative AI performs the tasks it sells. I examine the Philippines, the economy most dependent on it. I construct a sub-national exposure index from pre-ChatGPT occupational composition and observed usage in the Anthropic Economic Index, and trace employment as US clients adopt AI and automate what they bought. Employment growth slows in exposed provinces, industries, and municipalities, while wages and hours remain unchanged. Would-be workers remain unemployed, not yet absorbed at home or abroad. Displacement is beginning at the hiring margin — early, measurable, and concentrated where exposure is highest.

Cite

@misc{ravanilla_ai_bpo,
  title = {Early Signs of the Generative AI Labor Market Shock in the Global South: Evidence from Philippine Business Process Outsourcing},
  author = {Nico Ravanilla},
  howpublished = {Working paper},
}