The Job Mortgage

In the [YOU] on AI Field Guide

The cycle documents a transition moving faster than any previous technological disruption—engineers measuring a twenty-fold productivity multiplier in a week, senior architects whose career-defining expertise commoditizes in months rather than decades. At that speed, the gap between the skill that the market devalues and the skill that the market rewards is bridged not by patient institutional processes but by instruments that can move at something approaching the speed of the disruption. The job mortgage is the financial architecture of that bridge.

In Kaplan’s framework, the job mortgage addresses the same structural problem that Joel Mokyr’s analysis of institutional lag identifies at the macro level: the tendency of institutions to adapt at the speed of politics while technology adapts at the speed of computation. The job mortgage is a micro-level response to the macro-level lag—an instrument that lets individuals navigate the transition without waiting for the educational system or the social safety net to catch up. It does not solve the institutional problem; it buys time while the institutions are being rebuilt.

Jerry Kaplan

Origin

Kaplan introduced the job mortgage in Humans Need Not Apply (2015) as part of a broader argument that the harms of automation are institutionally determined rather than technologically determined, and can therefore be addressed by institutional design. The reasoning begins with a frustrating mismatch that he observed across the economy: workers displaced by automation need new skills, but acquiring those skills costs money and time that the newly unemployed rarely have, and conventional lenders will not finance an education whose payoff is uncertain. Meanwhile, employers in growing fields cannot find workers with the skills they need. The capital that could bridge this gap exists; no instrument connects it to the people who need it on terms that make sense.

The job mortgage borrows its basic structure from income-share agreements, which had been discussed in academic economics since Milton Friedman proposed something similar in the 1950s, but Kaplan gave the instrument a specific design adapted to technological displacement: the repayment is conditioned not just on income level but on the demonstrated value of the retraining itself, creating an incentive for the financing party to ensure the training program actually delivers marketable skills rather than credentials. The instrument is explicitly a market mechanism rather than a government program—a design choice that reflects Kaplan’s conviction that the most durable institutional innovations redirect market incentives rather than replace them.

Key Ideas

Earnings-contingent repayment. The defining feature of the job mortgage is that repayment scales with actual earnings rather than being fixed in advance. This makes the instrument viable for workers whose retraining is genuinely valuable but whose future income is uncertain at the time of borrowing—precisely the population that conventional lending cannot serve. The earnings-contingency is not charity; it is a pricing mechanism that correctly reflects the risk profile of the investment.

Incentive alignment across parties. The job mortgage is designed so that no party can profit from the other’s failure. A lender or employer who finances a training program and holds a claim on subsequent earnings has a direct financial interest in ensuring the training is useful and the job actually materializes. This alignment is structurally different from conventional lending, in which the lender profits regardless of whether the borrower’s investment pays off, and from corporate training programs, in which the employer bears no cost if the training proves irrelevant.

Financing human capital as seriously as physical capital. Kaplan’s deeper point is a structural critique of the existing financial system. We have sophisticated instruments for financing houses, cars, and businesses—assets whose future value can be estimated and collateralized. We have almost nothing for financing the development of human capabilities, which are the most valuable assets most people possess and the ones our financial system does the worst job of helping people develop. The capital-labor split that automation widens is partly a financing problem: capital can always find instruments for its own reproduction, while labor is left to finance its own adaptation through savings it no longer has.

Limits and discomforts. Kaplan did not pretend the job mortgage was perfect. Conditioning repayment on future labor income edges toward arrangements that critics have compared to indenture, and the idea of financiers holding a claim on a person’s earnings raises legitimate concerns about exploitation and control. His response was that these risks are manageable through design and regulation, and that the alternative—a growing population of displaced workers with no path back into the economy—is far worse. The point of the job mortgage was never that it was the final word but that it was a genuine mechanism, something specific enough to argue about and refine in a debate too often content with platitudes.