Pipeline Churn: Pilots Who Build Time but Never Reach the Airlines — Research Memo

Prepared for FRISA (Front Range Institute for Sustainable Aviation) — June 2026

Executive Summary

Since the 2013 First Officer Qualification rule (the "1,500-hour rule," implementing PL 111-216), every aspiring U.S. airline pilot must bridge roughly 1,250 flight hours between the commercial certificate (~250 hr) and ATP eligibility. The literature measures dropout before the private certificate and flight-instructor turnover, but no government, academic, or industry study quantifies attrition during the 250-to-1,500-hour time-building stage. This memo bounds it from primary FAA data.

Strongest defensible numbers:

• Cohort-flow bound: Over matched 9- and 11-year windows (commercial certificates issued 2013–2023 vs. original ATP certificates issued 2–3 years later), original airplane ATP issuances equal ~60% of original airplane commercial issuances (59.9–61.5% across lag assumptions; computed from FAA U.S. Civil Airmen Statistics Table 17, 2025 workbook, 2021 workbook, 2019 workbook). After adjusting qualitatively for foreign trainees (inflate the denominator) and military-competency ATPs (inflate the numerator), a defensible range for time-building completion is ~55–75%, i.e., roughly 1 in 3 (range ~1 in 4 to ~1 in 2) new commercial pilots never obtains an ATP [ESTIMATED BOUND, not a measurement].
• The divergence is widening: commercial issuances hit a record 20,069 in 2025 while original ATPs fell for a second straight year to 7,714 (down 31% from the 2023 peak of 11,218, which had itself risen from 2022) — a queue or churn signal flagged by the only analysis found that comes close to this question (VSL Aviation, Apr 2026), which still does not quantify attrition.
• Pre-private dropout is the only measured stage: "Between 70 percent and 80 percent of student pilots never earn a pilot certificate" (AOPA Flight Training Experience research, 2010–2011; FlightGlobal coverage).
• Risk bound: Producing 1,000 airline pilots induces ~1.3–1.8 million GA flight hours of exposure (completers' 1,250 hr each plus non-completers' wasted hours). At published instructional-to-overall-GA accident rates (≈1.4–3.9 total, ≈0.15–0.65 fatal per 100,000 hr), that implies roughly 20–60 accidents and 2–10 fatal accidents per 1,000 airline pilots produced, of which the churn-attributable share is ~6–29% [BOUNDING ARITHMETIC].
• Cost/environmental bound: ~3,500–8,000 domestic non-completers per year (the issuance gap net of foreign trainees and non-aspirant commercial pilots) × ~200–600 churn hours each ≈ 0.7–4.8 million zero-output hours/yr, burning ~6–43 million gallons of 100LL → ~50,000–360,000 t CO2/yr and ~13–91 t of lead/yr at the 2.12 g Pb/gal specification maximum (typical blends lower) — a meaningful single-digit-to-double-digit percentage of all U.S. piston-fleet lead emissions [ESTIMATED, assumptions stated in §4].

The gap is real: GAO (2014, 2018, 2023, 2026), DOT OIG (2013), the Pilot Source Studies, and UND's pilot-supply forecasts all count pipeline inputs and outputs or measure hired pilots; none measures who starts time-building and fails to finish (§5).

1. Pipeline Flow Data and Cohort Arithmetic

1.1 Primary data: FAA original certificate issuances (airplane), 2010–2025

Source: FAA U.S. Civil Airmen Statistics, Table 17 ("Original Airmen Certificates Issued by Category"), assembled from the 2025, 2021, and 2019 workbooks (hub: faa.gov civil airmen statistics).

* Student counts before April 2016 are FAA estimates tied to medical issuance (FAA Table 17 note); 2012–2013 values are rounded from FAA's fractional estimates. The 2016 ATP spike (9,520) reflects pilots rushing to complete ATP practical tests before pre-ATP-CTP knowledge-test results expired; the 2013–2014 elevation reflects the rush before the August 2014 ATP-CTP requirement. The FAA does not separate R-ATP from full ATP in Table 17; DPE analyses of FAA data indicate R-ATPs have stabilized at roughly one-third of annual ATP issuances (Jason Blair, "How many ATP certificates in 2024?") [SECONDARY ANALYSIS of FAA data].

1.2 Cohort-flow arithmetic

Time from commercial certificate to ATP is typically 1.5–3 years for full-time time-builders (CFI tenure data, §2). Comparing equal-length windows:

• 3-year lag: ATP 2017–2025 = 4,449+5,795+6,690+4,056+5,020+9,588+11,218+9,513+7,714 = 64,043. Commercial 2014–2022 = 9,803+9,211+10,191+10,506+12,198+14,179+14,442+12,771+13,715 = 107,016. Ratio = 59.9%.
• 2-year lag: ATP 2016–2024 = 65,849 ÷ Commercial 2014–2022 = 107,016 → 61.5%.
• 11-year window: ATP 2015–2025 = 80,107 ÷ Commercial 2013–2023 = 133,130 → 60.2%.

The ratio sits at ~60% across all three lag assumptions — a consistency check, not an independence check, since the windows share most of their underlying years of data. Read naively: about 40% of post-rule commercial-certificate cohorts have not obtained an ATP within the observation window. The 2023–2025 trend will push the forward-looking ratio down sharply: 2023–2025 produced 55,787 commercial certificates against a market in which U.S. majors hired 12,000+ pilots/yr in 2022–2023 but only 4,834 in 2024 (−60% YoY) (AirlineGeeks, Jan 2025; AOPA, Aug 2025).

1.3 Limitations of the arithmetic (each biases the 60% figure)

1. Foreign students train in the U.S. on FAA certificates and return home to convert to foreign licenses; they appear in the commercial denominator but never need an FAA ATP. This makes true domestic non-completion lower than 40%. No FAA statistic isolates foreign-national issuances [UNQUANTIFIED — itself part of the gap].
2. Military pilots can obtain ATP largely on military time (and R-ATP at 750 hr), entering the numerator without a matched civilian time-building cohort entry, biasing measured completion upward.
3. Not all commercial pilots are airline aspirants (ag, banner, corporate, hobbyist career-changers). FAA Table 13 shows the average active commercial pilot is 42.8–46.5 years old across 2013–2025 (FAA 2025 workbook, Table 13) — many are not on the 121 track, so the denominator overstates the at-risk cohort and the aspirant completion rate is higher than 60%.
4. Right-censoring: recent cohorts may still convert (some take 5–10 years), so windowed ratios understate eventual completion.
5. ATP ≠ airline seat, but post-2014 the ATP-CTP prerequisite (a $3,950–$4,800 simulator course, usually airline-paid in new-hire training — LVFA pricing, 2026 price guide) means most ATP/R-ATP practical tests now occur inside airline new-hire programs, making ATP issuance a reasonably tight proxy for reaching a 121/professional seat.

Bottom line: items (1) and (3) push true aspirant completion above 60%; item (2) pushes it below. A defensible bounding statement: time-building completion ≈ 55–75%; churn ≈ 25–45% [ESTIMATED BOUND]. A CFI-based cross-check (CFI initial issuances 2014–2022 = 57,975 vs. ATP 2017–2025 = 64,043, ratio >1) confirms many ATPs arrive via non-CFI routes and that CFI counts cannot tighten the bound alone.

2. Documented Attrition Causes and Rates by Stage

Pre-private: "Between 70 percent and 80 percent of student pilots never earn a pilot certificate" — AOPA's APCO Insight retention research, presented at the 2010 Flight Training Summit (AOPA, 2011; research PDF: download.aopa.org; FlightGlobal; AOPA summit report). Causes: educational quality, customer focus, community, information sharing; ~35% of instructors rated insufficiently professional. Note this 70–80% covers all student starts, not career-track students; collegiate-program attrition is lower but not centrally published.

Private → commercial: Not directly measured anywhere found. Issuance ratios (commercial ≈ 55–65% of private issuances 2–3 years prior, FAA Table 17 above) conflate aspirants with recreational pilots and cannot be read as attrition [GAP].

Commercial → 1,500 hr (time-building): No published measurement (§5). The ~60% cohort-flow bound in §1 is, to our knowledge, the first arithmetic specifically aimed at it. Documented mechanisms:

• CFI-job dependence and turnover: average CFI tenure is 12–18 months, compressing below 12 months in hiring booms (Academy of Aviation; The Flight Brief); 16 of 18 collegiate programs reported CFI recruiting/retention difficulty (Leonard & Christensen, IJAAA). When hiring stalls, time-builders stack up in instructor jobs that don't exist — churn pressure.
• Hiring-window timing: In 2009 U.S. majors hired 30 pilots (FAPA data, Spokesman-Review/Bloomberg); 2010–2013 remained a trough. COVID froze hiring in 2020 (ATP issuances fell to 4,056). The 2021–2022 rebound saw network/LCC carriers hire 15,000+ pilots (GAO-26-107856, Apr 2026), then 2024 collapsed to 4,834 at majors (AirlineGeeks), with modest re-expansion signaled for 2026 (AOPA, Mar 2026). Each trough strands a time-building cohort whose members age out, change careers, or exhaust funds; no one tracks how many return [GAP].
• Medical loss: Third-class denials run ~2,000–2,500/yr on ~135,000–140,000 applications (~1.5–1.9%), with ~5% of all ~450,000 annual applications deferred; final denials for those who persist are <0.2% (FAA FAQ; detail in the Aerospace Medical Certification Statistical Handbook). First-class loss rates specifically among 250–1,500-hr time-builders are not published [GAP]; cumulative career-stage medical loss over a 2–4-year time-building period plausibly removes low single-digit percentages of the cohort [ESTIMATE].
• Funding exhaustion: Collegiate professional-pilot programs cost $85,745–$138,511 for a four-year degree (GAO-26-107856); flight fees often exceed $50,000 on top of tuition (GAO-18-403). Federal student-loan caps don't cover flight costs (the gap motivating the proposed Flight Education Access Act). Private financing is boom-bust: Meritize exited aviation lending; Sallie Mae re-entered flight-school lending in 2025 (AOPA, May 2025). No flight-training loan default or completion-conditional data is published by any lender or agency [GAP].
• Age/career-change: average student pilot age 35.8 and average commercial pilot age ~43–46 (FAA Table 13) imply a large career-changer share with higher opportunity costs and narrower windows before the age-65 retirement cap erodes career payback.

3. Risk Accounting: Induced Exposure per Airline Pilot Produced

Published rates (primary/near-primary):

• Overall GA accident rate 2023: 3.86 total / 0.65 fatal per 100,000 hr (AOPA McSpadden Report, Nov 2025).
• Instructional flying 2023: 118 accidents, 12 fatal, 23 fatalities (McSpadden Report coverage; figure view) against 8.2 million instructional hours in the 2023 FAA GA Survey (AOPA, Apr 2025; survey hub: FAA CY2023) → ≈1.4 total and ≈0.15 fatal per 100,000 hr [COMPUTED from the two published figures]. AOPA's Fatal Flight Training Accident Report 2000–2015 (240 fatal instructional accidents; rate below overall GA; −35% over the period) is consistent.
• Common time-building side jobs (banner tow, survey, pipeline patrol, Part 135 cargo) have no cleanly published per-hour rates; NTSB's Part 135 special investigation (AIR-24-03) analyzed 500+ accidents 2010–2022 and treats on-demand operations as a persistent risk area. Time-building flying that is not dual instruction is better proxied by the overall GA rate than the instructional rate [ASSUMPTION].

The arithmetic. Let p = completion fraction of the time-building stage, h̄ = average hours flown beyond ~250 hr by non-completers before quitting. Induced exposure per successful airline pilot:

E = 1,250 + ((1−p)/p) × h̄

(1,250 = 1,500 − 250; reduce to a ~1,040–1,085-hr average wedge if one-third are R-ATPs at 750–1,000 hr — Jason Blair.)

Applying published per-100,000-hr rates to the central 1.52 M hr per 1,000 airline pilots produced:

• Total accidents: 1.4/100k → ~21; 3.86/100k → ~59. Range ≈ 20–60 accidents per 1,000 airline pilots produced.
• Fatal accidents: 0.15/100k → ~2.3; 0.65/100k → ~9.9. Range ≈ 2–10 fatal accidents per 1,000 pilots produced (≈3–20 deaths at 1.4–1.9 fatalities/fatal accident).
• Churn-attributable share (the (1−p)/p·h̄ term): full scenario envelope ~0.1–3.3 fatal and ~1–19 total accidents per 1,000 airline pilots produced (optimistic-scenario churn hours at the instructional rates through pessimistic-scenario churn hours at the overall-GA rates) — exposure that produces no airline pilot at all.

This is bounding arithmetic, not measurement. It assumes time-building hours carry instructional-to-overall-GA risk (some carry more, e.g., 135 cargo; some less), assumes h̄ without any survey basis (no dataset records where non-completers stop — itself the gap), and ignores risk heterogeneity by pilot experience. The honest claim: any safety evaluation of the 1,500-hour rule that counts only Part 121 outcomes omits a 1.3–1.8-million-GA-hour induced denominator per 1,000 pilots produced, of which roughly 6–29% is churn.

4. Cost and Environmental Accounting of Churn Hours

Volume of churn [ESTIMATED]: Recent issuance gaps (commercial ~18–20k/yr vs. ATP ~8–11k/yr, FAA Table 17) less a foreign-trainee share (unmeasured; assume 15–25% of commercial issuances [ASSUMPTION — flag for verification]) and less non-aspirant commercial pilots (ag, banner, corporate — §1.3 item 3), whose post-commercial hours are productive employment rather than churn, imply ~3,500–8,000 domestic non-completers per year. At h̄ ≈ 200–600 churn hours each: ~0.7–4.8 million churn hours/yr (central ~2.3 M) — plausible against the 8.2 M instructional + balance of 28–29 M total GA hours/yr (FAA GA Survey CY2023, CY2024).

Private sunk cost per non-completer [ESTIMATED]: $85,745–$138,511 collegiate (GAO-26-107856) or roughly $80,000–$110,000 academy/Part 61 to commercial+CFI (Epic Flight Academy cost guide); plus any self-funded churn hours at ~$140–$250/hr wet rental (Leopard Aviation) or ~$149/hr shared time-building blocks. Many churn hours are paid CFI work (a transfer, not a private loss), so the conservative per-pilot sunk private cost is the ~$80–140k of training capital stranded without career payback, plus foregone earnings. Pipeline-wide: 3,500–8,000 non-completers/yr × ~$100k ≈ $0.35–0.8 billion/yr in stranded training investment [ROUGH MAGNITUDE].

Environmental accounting (piston trainer ~8–10 gal/hr; aviation gasoline emits ~18.3 lb ≈ 8.31 kg CO2/gal (EPA emission factors, archived); 100LL contains up to 2.12 g lead/gal (FAA, Leaded Aviation Fuel and the Environment)):

• Per churn-hour: ~9 gal → ~75 kg CO2 and up to ~19 g lead.
• Per non-completer (h̄=400 hr): ~3,600 gal → ~30 t CO2, up to ~7.6 kg lead.
• Pipeline-wide: 0.7–4.8 M churn hr/yr → 6–43 M gal/yr → ~50,000–360,000 t CO2/yr and ~13–91 t lead/yr at the 2.12 g/gal specification maximum (typical blends run somewhat lower [ESTIMATE]). For scale, piston aircraft on leaded avgas are the largest remaining source of airborne lead emissions in the U.S. (EPA endangerment finding, Oct 2023) — churn hours alone plausibly account for a mid-single-digit to ~20% share of that source [BOUNDING ESTIMATE — verify national piston lead tonnage against the EPA NEI before publication].

All figures in this section are transparent multiplications of stated assumptions, suitable for sensitivity tables, not point claims.

5. Who Has Studied This — Proving the Gap

Searched: GAO, DOT OIG, FAA, NTSB, ERAU/UND/Purdue commons (Collegiate Aviation Review, IJAAA, JATE), AOPA/ALPA/industry analyses. No study quantifies attrition between the commercial certificate and the 1,500-hour ATP threshold. The nearest works measure something else:

Conclusion: the 250-to-1,500-hour attrition rate is unmeasured in the public record. The cohort-flow bound in §1 of this memo appears to be the first explicit attempt. The binding data obstacles: FAA does not publish certificate-level longitudinal linkage (the airman registry could support it), does not flag foreign-national issuances in Table 17, and does not separate R-ATP in issuance tables.

Implications for FRISA's White Papers

(a) Risk-denominator correction for the ATP-rule safety/cost analysis. Evaluations of the 1,500-hour rule (e.g., ALPA's) score only Part 121 outcomes. Total time-building exposure (rule-attributable only as an increment over the pre-rule baseline) is flown in general aviation: ~1.3–1.8 million GA hours per 1,000 airline pilots produced (§3), incurring an expected ~2–10 fatal accidents per 1,000 pilots produced at published GA/instructional rates — and 6–29% of that exposure (the churn share) produces no airline pilot at all. Any honest benefit-cost accounting of the rule must add this denominator; FRISA can present it as bounding arithmetic from FAA Table 17 + GA Survey + McSpadden rates, all primary or near-primary sources, while calling for FAA to publish the longitudinal certificate linkage that would turn the bound into a measurement.

(b) Environmental accounting. Churn hours are the cleanest possible target for environmental argument: 0.7–4.8 million flight hours/yr of piston flying — ~50,000–360,000 t CO2 and ~13–91 t of lead annually at the specification maximum, plus training-pattern noise over Front Range communities — that by definition yields zero pipeline output. Unlike training itself, no one defends churn. Policy levers that raise completion p (financing reform per the Flight Education Access Act, R-ATP pathway expansion, hiring-cycle smoothing, structured time-building) reduce lead, CO2, noise, and accident exposure per airline pilot produced without touching the 1,500-hour safety debate directly. FRISA should pair every churn-emissions figure with its assumption table (non-completers/yr, h̄, gal/hr, TEL content) and flag the foreign-trainee share and national piston-lead tonnage as the two numbers most needing verification before publication.

Recommended verification before citation: (1) pull the FAA GA Survey CY2023/CY2024 tables directly for instructional-hour confirmation; (2) confirm the EPA NEI piston-aircraft lead tonnage; (3) FOIA or request FAA registry analysis of commercial-to-ATP conversion by certificate vintage and citizenship — the single dataset that would close this gap.