{"type":"rich","version":"1.0","provider_name":"Transistor","provider_url":"https://transistor.fm","author_name":"The Agentic Allocator","title":"Professor Emmanuel Yimfor on Capital Allocation Bias in Private Markets and the Choices That Will Determine Whether AI Fixes or Entrenches Them","html":"","width":"100%","height":180,"duration":1527,"description":"Professor Emmanuel Yimfor, Assistant Professor of Finance at Columbia Business School, joins The Agentic Allocator to share his research that should sit at the centre of every conversation about AI in private markets. His work documents the core friction driving racial and gender disparities in access to capital: not quality, not track record, but networks. Who you can reach, not how good you are.\nThat finding has direct and urgent implications for how AI gets deployed across the LP/GP ecosystem. Used thoughtfully, AI has the potential to widen the top of the funnel dramatically, reducing the cost of due diligence enough that LPs can evaluate managers far beyond their existing networks. Used carelessly, the same tools will automate and entrench the same exclusions, encoding past decisions into future ones in ways that are subtle, hard to detect, and difficult to reverse.\nProfessor Yimfor walks through the mechanics of embedding-based matching and why it is a black box that can pick up on signals of race, gender, and network affiliation even when no one intended it to. He explains what the research on accelerators and structured access programmes shows about what happens when the top of the funnel is genuinely open. He makes a clear, practical case for what LPs, GPs, and technology developers should each be doing differently right now.\nWhat You'll Learn:Why the core friction driving racial and gender disparities in private markets is networks and what the research evidence showsWhy Black and Hispanic founders raise around 40% less capital than peers with identical patent holdings, educational backgrounds, and track recordsWhy the gap in funding disappears entirely when access is structured, as in accelerators and grant programmes, and what that tells us about where the problem liesHow embedding-based matching works, why it is a black box, and how it can encode biases in allocation decisions even when no one intended it toWhy asking AI how similar a new manager is...","thumbnail_url":"https://img.transistorcdn.com/QigjduDJIqFTeeJFynBiyzGolv4eHG5zocOf8vY173o/rs:fill:0:0:1/w:400/h:400/q:60/mb:500000/aHR0cHM6Ly9pbWct/dXBsb2FkLXByb2R1/Y3Rpb24udHJhbnNp/c3Rvci5mbS9hZGY5/NTNkMjIzZGE0NWFj/YWEzYzY0ODU1ZTYx/NzE2Zi5wbmc.webp","thumbnail_width":300,"thumbnail_height":300}