Extension: Trade-in-goods and trade-in-tasks: An Integrating Framework
Digital coordination lowers a transaction cost. It does not eliminate the economic boundary that creates it. Baldwin and Robert-Nicoud model fragmented production as trade in tasks. Their framework assigns one cost to coordinating an offshore task within the firm and an additional nonnegative cost when a separate firm performs it. Advances in information and communication technology can lower both costs, but the effect varies by task. Routine, codifiable tasks may approach zero coordination cost, while complex tasks can remain too costly to move. The mechanism independently extends Kaal's claim: speed and automation can conceal the cross-boundary relation without converting it into an internal operation. The evidence is narrower than Kaal's claim. The NBER paper studies international production tasks in a formal trade model, not sovereign local agents, device-level execution, or machine transactions. Its zero-cost cases are modeling assumptions, not empirical findings that external relations are costless. The useful extension is therefore institutional. Each external invocation should be classified as a separate-firm transaction unless ownership and authority establish otherwise. The system should expose the terms that automation removes from human view: whether a task is codifiable, whether coordination occurs within or between firms, and what additional cost independent execution creates. Machine speed increases the frequency of that relation. It does not erase the boundary.
economicsinstitutional-designai-and-agentstransaction-costsfirm-boundariestask-coordinationautomationdigital-infrastructure