# The degree signal reprices on a phased 20-year schedule: slow erosion 2025–2032, rapid collapse of the non-elite premium 2032–2040, equilibrium of fewer larger universities plus lean outcome-selling teaching institutions after 2040 (Unverified: the phase boundaries)

The signaling facet is the pacing mechanism for the whole university question, and it is phased rather than linear. In 2025–2032, employers strip degree requirements from job postings at the low end — the share of postings demanding a bachelor's has already declined from its late-2010s peak (Burning Glass Institute / Harvard's "Dismissed by Degrees" work) — while AI screening makes competence observable before hiring. The leading indicator is employer-side, not enrollment-side: non-licensure majors at tuition-dependent schools soften first and get misread as price sensitivity, when it is actually the new-graduate wage premium compressing (recent-graduate unemployment ticked above the national rate in 2024–25, BLS; Unverified on exact magnitude).

In 2032–2040, the demographic cliff (2029–2037) collides with a visibly devalued generalist degree, and the two forces compound: this is the window when tuition-dependent institutions close rather than merely shrink, and when even selectivity-protected regionals come under pressure. Post-2040 equilibrium: a smaller number of larger, more selective research universities, lean teaching institutions selling verified outcomes, and net national enrollment modestly down but heavily region-skewed toward the Sun Belt and licensure-heavy program mixes.

The reason to date the phases rather than treat erosion as continuous: the degree's trust/liability function (blame-shifting for employers and licensure boards) decouples it from its screening function, so erosion is abrupt where screening matters most and slow where insurance matters most — producing a cliff shape, not a slope.

What would change this: if the share of postings requiring degrees stabilizes or rebounds and the new-grad premium holds through the late 2020s, the screening erosion is cosmetic and the 2032–2040 rapid repricing collapses into the slow phase; conversely, if AI-verified skills credentials achieve employer-side adoption comparable to the degree before 2030, the rapid phase