Buying In and Selling Out: The Dynamic Returns to Investing in Expertise

Felix Zhiyu Feng and Mark M. Westerfield
Working Paper, October 2026

[PDF] [SSRN Version]

Summary

We present a unified framework to study the incentives for young knowledge workers to build hidden, durable, and inalienable expertise under serial employment and medium-term contracts. The division of surplus in future contracts drives current incentives, creating regimes of serial founders and serial salaried employees.

Abstract

We study medium-term optimal contracts for a knowledge worker with hidden, durable, and portable expertise who is matched with a sequence of principals through the labor market. Employment relationships are sufficiently persistent for incentives to evolve but remain temporary, allowing future contracts to shape current incentives. The model therefore characterizes career trajectories within and across jobs in a unified framework. Within a job, the worker may be promoted from salary-like compensation to equity-like pay after sufficient tenure, or begin as a residual claimant and sell part of his stake following good performance. Across jobs, the division of expected match surplus generates distinct career regimes: a serial founder repeatedly begins with full ownership, whereas a serial salaried worker begins with full insurance and waits for promotion. Because the labor market links incentives across successive contracts, allocating all match surplus to either party need not maximize that party’s welfare.

Cite as

Feng, Felix Zhiyu, and Mark M. Westerfield. 2026. “Buying In and Selling Out: The Dynamic Returns to Investing in Expertise.” Working paper, October 2026.

BibTeX

@unpublished{FengWesterfield2026,
  author = {Feng, Felix Zhiyu and Westerfield, Mark M.},
  title  = {Buying In and Selling Out: The Dynamic Returns to Investing in Expertise},
  note   = {Working paper},
  month  = {October},
  year   = {2026},
  url    = {https://markwesterfield.com/papers/FW-Dynamic-Expertise.html}
}

This is a working paper; the posted PDF is the October 2026 draft. Updated October 5, 2026.