The Security Clearance Salary Premium: About 7% at the Same Employer
The security clearance salary premium is about 7%. At the same employer, for the same role family, seniority and state, a job that requires an active clearance lists 7.2% more pay than one that requires none (95% interval 5.5% to 10.1%; 1,739 jobs at 33 employers). The data is every US job posting with listed pay in our database on November 2, 2026.
On the median comparison pay of $160,200, 7.2% is about $11,500 a year. The median group of comparable jobs shows a smaller gap, $7,500. The two measure different things: the median counts each group once, while the 7.2% gives more weight to groups with more postings.
| Clearance the job requires | Premium | 95% interval | On $160,200 | Jobs | Employers |
|---|---|---|---|---|---|
| TS/SCI | +7.8% | +5.8% to +11.3% | +$12,500 | 489 | 26 |
| Secret | +7.1% | +0.8% to +9.9% | +$11,400 | 730 | 18 |
| Top Secret | +6.5% | +4.6% to +9.4% | +$10,400 | 520 | 20 |
| Sponsored after hire | +2.5% | −1.0% to +5.5% | No measurable gap | 1,370 | 45 |
| Level unnamed | +1.2% | −6.9% to +5.0% | No measurable gap | 45 | 13 |
The sponsored row's label is unreliable; see Checks.
What the data shows
- The premium is about 7%, and the three levels can't be told apart. Their intervals overlap, so this data doesn't show TS/SCI listing more than Secret once the employer and role are held equal.
- A clearance the employer sponsors after hire ("ability to obtain") adds less, if anything. With its label corrected the estimate is +3.8% (0.0% to +6.1%): an interval that reaches zero, around half the premium for already holding one.
- Across all employers the sign flips. Compared by role, seniority and state alone, jobs requiring a held clearance list 3.5% to 7.8% less. Raw medians say the same: $171,500 to $172,100 for cleared jobs against $177,500 for jobs with no clearance. Limit the comparison to space and defense employers and the premium turns positive, so that gap comes largely from AI labs and other higher-paying companies in the comparison group.
How the comparison changes the answer
Every column holds role family and seniority equal, and all but the third hold state equal too. Each compares more like with like than the one before. Each cell is the premium with its 95% interval.
| Clearance | All employers | Space and defense | Same employer | Same employer and state |
|---|---|---|---|---|
| TS/SCI | −7.8% (−14.0 to −0.7) | +3.3% (−2.3 to +7.1) | +6.8% (+3.9 to +10.5) | +7.8% (+5.8 to +11.3) |
| Secret | −3.5% (−19.0 to +3.6) | +8.0% (−8.0 to +13.5) | +5.1% (−4.2 to +8.4) | +7.1% (+0.8 to +9.9) |
| Top Secret | −6.1% (−17.3 to +1.0) | +6.8% (−5.7 to +11.7) | +6.1% (+2.5 to +8.7) | +6.5% (+4.6 to +9.4) |
| Sponsored after hire | −7.7% (−16.1 to −1.6) | +1.7% (−5.6 to +5.6) | +1.0% (−3.4 to +4.7) | +2.5% (−1.0 to +5.5) |
| Level unnamed | −8.7% (−19.2 to −3.0) | +0.6% (−11.3 to +6.3) | +0.1% (−6.1 to +4.4) | +1.2% (−6.9 to +5.0) |
Read the last column: the first compares Anduril with OpenAI as much as it compares cleared jobs with uncleared ones.
Public Trust is left out: only 5 jobs at 2 employers have a comparison in the last column.
By seniority and role family
Any held clearance (Secret, Top Secret, TS/SCI) pooled, at the same employer, role family, seniority and state.
| Group | Premium | 95% interval | Jobs | Employers |
|---|---|---|---|---|
| All | +7.2% | +5.5% to +10.1% | 1,739 | 33 |
| Mid | +8.8% | +6.9% to +13.4% | 745 | 26 |
| Senior | +5.3% | +1.9% to +8.6% | 922 | 26 |
| Aerospace Engineering | +9.0% | +2.5% to +14.4% | 536 | 18 |
| Software | +7.0% | +2.3% to +11.7% | 464 | 19 |
| Manufacturing | +10.1% | +6.3% to +52.2% | 123 | 6 |
| Mechanical | +7.9% | +1.9% to +14.8% | 120 | 8 |
| Electrical | +4.7% | −4.5% to +8.5% | 118 | 5 |
| Operations | +6.9% | +2.1% to +18.4% | 91 | 14 |
| Business & Finance | +14.5% | +6.9% to +23.2% | 84 | 18 |
Too thin to report: Junior (15 jobs), Staff (30), Lead+ (26), Principal (1), and the Security (43), Avionics (34) and Technician (42) role families, whose intervals are too wide to read.
In plain counts, 279 groups of comparable jobs hold both a cleared and an uncleared posting. The cleared one lists more in 165, the same in 23 and less in 91.
Methodology
Data. Every active US job posting with listed pay in Zero G Talent's database on November 2, 2026, ingested directly from employers' applicant-tracking systems: 23,737 postings, of which 21,405 are compared after the exclusions below.
Pay. The midpoint of the posted band, annualized in US dollars, floor $20,000. A posting with one figure uses that figure.
Clearance. An LLM parser reads each posting's text and records the clearance it requires: a named level, or "Ability to obtain…" when the employer sponsors it after hire. TS/SCI includes TS/SCI with polygraph; Secret includes Confidential. The comparison group is postings left blank or marked "Not Required", minus blank postings whose text still mentions clearance (368), which are likely misses. A blank can also mean the parser hasn't read that posting yet, which is why the text check below matters.
Comparable. The same role family (our job category), seniority tier (the structured field, else the job title) and state, with remote as its own location. The headline also requires the same employer.
Excluded. Federal postings from USAJobs (116), whose pay follows government scales rather than an employer's choice; internships (738); postings with no location (238); midpoints above $1 million (3); and repeats of the same employer, title, location, band and clearance (869).
Estimate. The logarithm of pay regressed on the clearance groups with a fixed effect for each comparable group, which amounts to a weighted average of the gaps inside groups. A group with no comparison job adds nothing. Percentages are the exponentiated coefficients.
Uncertainty. 95% intervals from 400 bootstrap resamples of employers, because one employer's postings aren't independent of each other.
Checks
- Labels match the text. Of postings labelled Secret, Top Secret, TS/SCI or TS/SCI with polygraph, 98.7% to 100% name that level in their text.
- Missed labels barely matter. At employers that use the label, 307 of 7,768 blank postings (4.0%) mention clearance words. A sample shows a mix of real misses ("will require an active TS security clearance") and false alarms ("trade secrets", "regulatory clearance"). Keeping them all in the comparison group gives Secret +6.8%, Top Secret +6.3% and TS/SCI +7.6%.
- Either end of the band gives the same answer. Low end: Secret +7.6%, Top Secret +6.9%, TS/SCI +7.9%. High end: +7.0%, +6.6%, +7.9%.
- No single employer drives it, though Secret leans on one. Anduril supplies 1,076 of the 1,739 pooled jobs. Without it: Secret +5.1% (−0.5 to +10.1), Top Secret +6.6%, TS/SCI +8.6%. Without SpaceX: +7.8%, +7.3%, +8.9%. Without Blue Origin: +7.4%, +7.0%, +8.2%.
- The sponsored label is weak. 1,217 of the 2,388 postings labelled "ability to obtain" never mention a clearance, 909 of them at Blue Origin: an older version of our parser over-applied it. Counting those 1,217 as postings with no clearance gives sponsored after hire +3.8% (0.0% to +6.1%; 612 jobs at 42 employers) and any held clearance +7.5% (+5.8% to +9.6%; 1,789 jobs at 33 employers), so the headline stands within its interval. The named levels aren't affected: see the first check.
- Location errors didn't move it. On November 2 our geocoder could misplace a posting that named only a city: a job in Munich could read as Munich, North Dakota, and count as a US posting. That's since fixed. Dropping all 3,028 postings that named only a city, right or wrong, gave +7.3% (+5.3% to +10.2%; 1,709 jobs at 30 employers).
- Pay disclosure isn't selecting the sample. 95% of held-clearance postings list pay in CA, CO, WA and NY, 95% in VA, MD and DC, and 85% elsewhere.
Limits
- Listed bands, not offers. A band can span several levels, and the midpoint isn't what anyone is paid.
- Which employers. The headline can only use employers with both cleared and uncleared postings in the same role, so it leans on new-space and defense-tech companies: after Anduril come SpaceX (165 jobs), The Aerospace Corporation (126), Shield AI (67) and Blue Origin (48). Northrop Grumman and Boeing label nearly every posting, which leaves them almost no comparison. RTX lists pay on 23 of its 151 open postings, and Lockheed Martin, General Dynamics and L3Harris have none open in our data.
- Where. California holds 1,750 of the 3,207 cleared postings; Virginia and DC together hold 498.
- Not causal. A clearance requirement comes with different programs and work. The premium is a difference in listed pay, not what getting cleared would earn someone.
- One day. A snapshot of open postings, which weights employers by how much they're hiring today.
See open jobs that require a clearance or pay by company and role. For the next study by email, get the Zero G Talent newsletter.