Can We Build More Effective Place-Based Policies? Theory and Evidence -- by Matthew Freedman, David Neumark

TEXT START: Amid renewed policy interest, the academic literature on place-based programs has expanded dramatically in recent years.

The Dissection

The essay treats regional decline as an engineering problem: identify design variables, measure employment effects, and improve the subsidy machine. Its implicit objective is to make distressed places economically functional through better-targeted job creation.

Under the Discontinuity Thesis, this is a narrower intervention operating inside a system whose central employment-to-consumption circuit is being severed. Place-based policy may move jobs, firms, and public money geographically. It does not secure human ownership of productive capital or restore the necessity of mass labor.

The Core Fallacy

The central error is confusing local job creation with durable productive participation. A program can generate employment, raise wages temporarily, or attract firms while the underlying production system is rapidly reducing its human labor requirement.

The subsidy changes location economics. AI changes the cost and necessity of human cognition. The second force eventually overwhelms the first. A successful place-based program can therefore become a more efficient way to subsidize an obsolete labor circuit.

Hidden Assumptions

• That newly created jobs remain economically necessary rather than politically maintained or automatable.
• That firms receiving incentives will retain a durable local workforce instead of deploying portable AI capital.
• That local multipliers remain meaningful when income increasingly accrues to owners of automated systems.
• That measured employment, wages, and business formation are adequate proxies for human economic power.
• That policy can coordinate jurisdictions and firms well enough to preserve human-only domains at scale.
• That regional inequality is the main terminal problem, rather than ownership and control of productive AI capital.

These assumptions may hold during the lag period. They fail as a permanent defense under P1, P2, and P3.

Social Function

Primarily transition management, with elements of ideological anesthetic and partial truth.

The partial truth is that design matters. Some place-based programs will outperform others, especially where they secure physical infrastructure, energy, logistics, maintenance, or strategically protected industries. The anesthetic is the implication that better program design can preserve broad-based labor participation after cognitive automation becomes dominant.

The likely result is not systemic rescue but managed differentiation: a few regions become subsidized production or maintenance nodes, while others are abandoned, hollowed out, or converted into consumption zones dependent on transfers.

The Verdict

Place-based policy can manage geography during the decline. It cannot reverse the discontinuity.

At best, it buys time, concentrates scarce activity, and creates temporary servitor niches around sovereign-controlled capital. At worst, it mistakes redistributed employment for restored economic agency. The paper’s policy horizon is a repair manual for the labor market; the underlying machine is losing the need for a labor market at all.