The 10-Year Tenure Trap: Why Senior Workers Are 2.5× More Likely to Be Laid Off in 2026

1 · Background

The intuition that long tenure protects against job loss is old. It was approximately correct from the post-war period through the early 2000s: seniority correlated with institutional knowledge, established relationships, and a higher floor on replacement cost. The 2008–2010 layoff cycle was the first major stress on this assumption, but the affected industries (manufacturing, finance ops) recovered without permanent shifts in the tenure–risk relationship.

The 2023–2025 layoff cycle is different. It is concentrated in knowledge-work occupations where AI tooling is mature, and it coincides with a measurable decoupling between tenure and protection. Three independent datasets confirm the new shape.

2 · Evidence

2.1 — Meta layoff multipliers (2024). A class-action complaint filed against Meta in November 2024 documented that during their reductions, employees aged 40 and older were 1.5× more likely to be terminated than employees under 40; employees aged 50 and older were 2.5× more likely [1]. The complaint was supported by internal personnel records covering 11,000+ separations.

2.2 — Age × AI-exposed employment growth. The Anthropic Economic Index report (March 2026) provides cross-tabulated employment and AI exposure data by age cohort [2]. In the period 2024-Q1 through 2026-Q1, employment in AI-exposed occupations among workers aged 22–25 declined 13.0% relative to baseline. Employment in the same occupations among workers aged 35–49 grew 4.6%. The 50+ cohort shows mixed signals — protected within stable firms, but disproportionately exposed during reduction events.

2.3 — Entry-level erosion. Big Tech reduced new-graduate hiring by 25% between 2023 and 2024 [5]. The mechanism is well-documented: tasks Claude and competing models are most fluent at — first drafts, summarization, code stubs, structured data entry — exactly map to the apprenticeship tasks that defined entry-level work. The result is a squeeze at the bottom: the learning curve is being automated faster than firms can hire to replace it.

2.4 — EEOC complaint volume. Age-discrimination complaints to the EEOC grew from 14,247 in 2023 to 16,223 in 2024 — a 13.9% YoY increase [3]. Tech industry filings represent ~20% of total charges (vs the 15% all-industry average), consistent with the Meta-like pattern being more widespread than a single firm[4].

3 · Implications for the Hedj model

The naïve approach would be to treat tenure as monotonically protective: more years → bigger defense bonus. The evidence in §2 invalidates that for non-executive roles. The Hedj model uses a level-aware tenure function:

tenure_bonus(tenure, level) =
  level === "executive"
    ? { "0-2": -2, "3-5": 2,  "6-10": 5, "10+": +7 }  // monotone
    : { "0-2": -4, "3-5": 3,  "6-10": 5, "10+": -2 }  // inverted-U

The executive exception is required because Meta's data and the EEOC pattern reflect reductions of individual contributors and managers, not C-suite. Founders and executives accrue different forms of protection — equity, relationship networks, institutional authority — that the salary-target dynamic doesn't override.

For a non-executive 15-year-tenure senior IC with otherwise typical scores, this swings the final reading by ~9 months compared to a naïve monotone curve. A 32-year-old former-founder now back at a FAANG with 15 years total tenure could argue the model penalizes them inappropriately — see §5.

4 · Worked example

Consider a hypothetical senior IC at a public software company: automation = 60, half-life = 55, wage = 50, AI investment = 60, role demand = 50, ai_fluency = daily, tenure = 10+. A naïve monotone tenure bonus would put this profile at 23 months. With the inverted-U adjustment for level = senior, it reads 14 months — a 9-month shift attributable to the tenure correction.

5 · Limitations

5.1 — Tenure is an imperfect proxy for age.A 32-year-old with 10 years at a single employer faces different layoff dynamics than a 55-year-old at the same tenure. The Meta data is conditional on age; we apply it via tenure. A future revision should ask the user's career start year and model age and tenure jointly.

5.2 — Mostly tech-sector data. The Meta complaint is tech, EEOC tech-skew is documented, AEI exposure is highest in tech and adjacent. The inverted-U almost certainly applies to other knowledge-work sectors, but the 2.5× multiplier specifically is a tech finding.

5.3 — Cyclical vs structural. The 2024 layoff cycle may not be fully representative. If 2025-2026 shows a different distribution (e.g. cuts concentrated mid-tenure as firms try to retain senior IC depth), the model will need re-fitting.

References

1. [1]
   Meta age-discrimination lawsuit (Gizmodo coverage, 2024)

   Primary source for the 1.5× / 2.5× layoff multipliers. The original complaint was filed in N.D. Cal. November 2024.
2. [2]
   Anthropic Economic Index report · March 2026

   Age-cohort × occupation cross-tab on employment changes 2024-Q1 through 2026-Q1 in AI-exposed roles.
3. [3]
   EEOC Charge Statistics FY 2024

   16,223 age-discrimination complaints in FY2024, up 13.9% YoY.
4. [4]
   Fortune · Older tech workers filing age complaints in droves

   Tech industry age-discrimination complaints at ~20% of total charges vs 15% all-industry.
5. [5]
   The Crisis of Entry-Level Labor in the Age of AI (2024–2026)

   25% reduction in Big Tech new-graduate hiring 2023→2024; documents AI automation of onboarding-tier tasks.
6. [6]
   BLS Worker Displacement Survey 2021–2023

   2.6M long-tenured (3+ yrs) displaced workers; tenure brackets show mid-range displacement rates lower than tails.
7. [7]
   AARP · Age Discrimination Holds Steady Among Older Workers 2025

   64% of 50+ workers report seeing or experiencing age discrimination; 22% feel pushed out due to age.