A fixed grade and step in the General Schedule and the Foreign Service draws slightly more in real terms than it did in 1994. Almost none of that came from the statutory schedule, which lost a sixth of its value. Locality pay carried the whole load.
Every line is deflated by CPI-U and indexed to 1994, so the flat line at 100 marks purchasing power held. What an employee is actually paid is the solid line: base pay plus locality pay. It rose steeply until 2010, because locality pay was being built from nothing. FEPCA created it in January 1994 and phased it up over the sixteen years that followed, and it has fallen 10% from that 2010 peak since, through the three-year freeze and the post-2021 price surge. The dashed line is the statutory schedule underneath, which is the part the annual adjustment actually moves. It ends 15% below where it started.
The build-out that lifted the upper line above is largely finished. Locality pay in Washington rose from 24.2% in 2011 to 33.9% in 2026, which is real money, but it no longer outruns inflation the way it did in the sixteen years after FEPCA created it. Both lines below are indexed to 2011, the first year whose OPM tables let base and locality be separated exactly.
Overseas Comparability Pay closes most, but not all, of the gap between an officer posted abroad and the identical class and step in Washington. Across the grades that receive it the overseas rate sits 8.5% below the Washington rate at every step, because both are percentages of the same base. Two limits matter. OCP is paid at FS-01 through FS-09 only, so the Senior Foreign Service receives none of it. And it is temporary authority rather than a standing feature of the schedule: it runs on periodic congressional extension, most recently through 20 December 2026.
Washington GS-15 pay is capped at Executive Schedule level IV. In 2011 the cap bit two of the ten steps; in 2026 it binds five, so half the grade’s steps now pay an identical $197,200. Step 10 would otherwise be $220,065.
A $22,865 haircut at the top step, which is the compression the schedule applies to its own senior ranks.
Two federal lines, because the paycheck and the schedule are different objects. The statutory schedule is what the annual adjustment in 5 U.S.C. 5303 moves, and it is the right subject for a formula-versus-delivery comparison. It is not what anyone is paid: every GS and Foreign Service employee draws base plus locality, or base plus an overseas allowance. Charting only the base against a market benchmark would answer a question nobody asked, so both are drawn.
The federal line is a fixed basket, not an average. It follows one grade and step (GS-13, step 1, and its Foreign Service equivalents) held constant across all 33 years. Average federal salary would not work here: it rises partly because the workforce professionalized as clerical and blue-collar work moved to contract, which is a change in who is employed, not in what the job pays. A fixed anchor is the correct mirror of a fixed-weight benchmark.
The benchmark is fixed-weight too, and it is the statutory one. FEPCA sets the annual adjustment from the Employment Cost Index for private wages and salaries, less half a percentage point (5 U.S.C. 5303). ECI holds occupation and industry constant, so it prices labor rather than counting job mix. The comparison is therefore the one Congress legislated, not one chosen for effect. The statute also contains its own escape: 5 U.S.C. 5303(b) lets the President substitute an alternative adjustment, and Congress legislates the rate directly in most years. Every schedule charted here is lawful. The gap is a choice made inside the statute, not a departure from it.
Why median earnings for degreed workers, and not median household income or an average wage. Household income moves with the number of earners per household. An average wage is pulled by a right tail that the wage surveys measure badly, since they exclude bonuses, profit-sharing and equity. And the all-worker median mismatches the workforce: two thirds of federal employees hold a bachelor’s degree or higher against roughly 43% of private employees, so comparing against all workers measures composition, not pay.
What would move these numbers. ECI excludes equity compensation, so it understates private-sector growth at the top; that biases this chart against the erosion finding. Benefits are excluded on both sides, and the federal package is pension-heavy and worth more than the private average, which also cuts against the finding. Neither adjustment is available as a consistent annual series.
Federal pay. OPM salary tables, 2011–2026, read directly (base schedule and the Washington-Baltimore-Arlington
locality table). 1994–2010 chained backward from the 2011 table using the annual base adjustments actually
authorized, per CRS. The chain reproduces the published 1994 GS-13 step 1 rate of $47,920 exactly, which is the
check that the reconstruction is right.
Authorised and required adjustments. CRS RL34463, Table 1 (1991–2010), cross-checked against CRS RL33158,
Table 1 (1991–2007); the overlapping years are identical.
FEPCA counterfactual. 1994–2010 from the CRS “ECI-Based Annual Adjustment Required by FEPCA” column;
2011–2026 computed from the statute using BLS series CIU2020000000000I.
Benchmarks. CPI-U annual average, BLS CUUR0000SA0. Median usual weekly earnings, full-time wage and salary
workers with a bachelor’s degree or higher, 25 and over, BLS LEU0252918500.
Total pay (base + locality). 1994–2010, the base schedule times the national-average locality
percentage authorised each year, from CRS RL34463 Table 1; 2011–2026, grown at the rate of the OPM
Washington-Baltimore-Arlington table. The two are joined at 2010/2011, when locality pay was frozen at its
2010 rate, so the change of basis introduces no growth at the seam. Washington locality exceeded the national
average throughout, so a Washington-only construction would sit above the line drawn here.
Foreign Service. State Department FS salary schedules for 2024 and 2026. OCP grade coverage and
extension status from the American Foreign Service Association.
Caveats. 2026 CPI is a seven-month average, January to July, and prices have run above that mean
since; a full year would put the base line nearer −15.5%, so the last point understates the loss rather than
overstating it. The schedule and the FEPCA line both start at the same 1994 rate, which leaves out the 2.2%
the formula called for in January 1994 against an authorised 0.0%. Counting it would widen the 2026 gap from
13.6% to 15.4%. The 2025 figure for median weekly earnings is an eleven-month average: October 2025 was not
collected during the government shutdown. The Foreign Service
schedule receives the same statutory annual adjustment as the General Schedule, so the base-schedule line applies
to both; the two diverge only through locality and Overseas Comparability Pay.
| Year | Total $ | Base $ | FEPCA $ | CPI-U | BA+ $/wk | Total idx | Base idx | BA+ idx |
|---|---|---|---|---|---|---|---|---|
| 1994 | 49,813 | 47,920 | 47,920 | 148.2 | 733 | 100.0 | 100.0 | 100.0 |
| 1995 | 51,347 | 48,879 | 49,166 | 152.4 | 747 | 100.2 | 99.2 | 99.1 |
| 1996 | 52,628 | 49,856 | 50,346 | 156.9 | 758 | 99.8 | 98.3 | 97.7 |
| 1997 | 54,252 | 51,003 | 51,504 | 160.5 | 779 | 100.6 | 98.3 | 98.1 |
| 1998 | 55,792 | 52,176 | 52,946 | 163.0 | 821 | 101.8 | 99.0 | 101.8 |
| 1999 | 57,827 | 53,793 | 54,588 | 166.6 | 860 | 103.3 | 99.9 | 104.4 |
| 2000 | 60,651 | 55,838 | 56,662 | 172.2 | 891 | 104.8 | 100.3 | 104.6 |
| 2001 | 62,948 | 57,345 | 58,192 | 177.1 | 921 | 105.7 | 100.1 | 105.1 |
| 2002 | 65,915 | 59,410 | 60,287 | 179.9 | 941 | 109.0 | 102.1 | 105.8 |
| 2003 | 68,675 | 61,251 | 62,156 | 184.0 | 964 | 111.0 | 103.0 | 105.9 |
| 2004 | 71,592 | 62,905 | 63,834 | 188.9 | 986 | 112.8 | 103.0 | 105.5 |
| 2005 | 74,156 | 64,478 | 65,430 | 195.3 | 1,013 | 113.0 | 102.1 | 104.9 |
| 2006 | 76,510 | 65,832 | 66,804 | 201.6 | 1,039 | 112.9 | 101.0 | 104.2 |
| 2007 | 78,199 | 66,951 | 67,939 | 207.3 | 1,072 | 112.2 | 99.9 | 104.5 |
| 2008 | 81,067 | 68,625 | 69,638 | 215.3 | 1,115 | 112.0 | 98.6 | 104.7 |
| 2009 | 84,328 | 70,615 | 71,657 | 214.5 | 1,137 | 116.9 | 101.8 | 107.2 |
| 2010 | 86,023 | 71,674 | 73,377 | 218.1 | 1,144 | 117.4 | 101.7 | 106.1 |
| 2011 | 86,023 | 71,674 | 74,015 | 224.9 | 1,150 | 113.8 | 98.5 | 103.4 |
| 2012 | 86,023 | 71,674 | 74,852 | 229.6 | 1,165 | 111.5 | 96.5 | 102.6 |
| 2013 | 86,023 | 71,674 | 75,742 | 233.0 | 1,194 | 109.9 | 95.2 | 103.6 |
| 2014 | 86,884 | 72,391 | 76,757 | 236.7 | 1,193 | 109.2 | 94.6 | 101.9 |
| 2015 | 87,753 | 73,115 | 77,755 | 237.0 | 1,230 | 110.2 | 95.4 | 104.9 |
| 2016 | 89,030 | 73,846 | 79,139 | 240.0 | 1,259 | 110.4 | 95.2 | 106.1 |
| 2017 | 91,591 | 74,584 | 80,374 | 245.1 | 1,279 | 111.2 | 94.1 | 105.5 |
| 2018 | 93,692 | 75,628 | 81,925 | 251.1 | 1,324 | 111.0 | 93.1 | 106.6 |
| 2019 | 95,819 | 76,687 | 83,646 | 255.7 | 1,367 | 111.5 | 92.8 | 108.1 |
| 2020 | 99,192 | 78,681 | 85,804 | 258.8 | 1,421 | 114.0 | 94.0 | 111.0 |
| 2021 | 100,185 | 79,468 | 87,940 | 271.0 | 1,452 | 110.0 | 90.7 | 108.3 |
| 2022 | 103,212 | 81,216 | 89,857 | 292.7 | 1,544 | 104.9 | 85.8 | 106.7 |
| 2023 | 108,228 | 84,546 | 93,532 | 304.7 | 1,609 | 105.7 | 85.8 | 106.8 |
| 2024 | 113,974 | 88,520 | 97,928 | 313.7 | 1,692 | 108.1 | 87.3 | 109.1 |
| 2025 | 116,503 | 90,025 | 101,836 | 321.9 | 1,740 | 107.7 | 86.5 | 109.3 |
| 2026 | 117,668 | 90,925 | 105,207 | 331.2 | 1,768 | 105.7 | 84.9 | 107.9 |
The Schedule and the Paycheck · a working chart · krucho.ski. Pay tables from the Office of Personnel Management and the Department of State; annual adjustments and the FEPCA counterfactual from the Congressional Research Service; price and earnings series from the Bureau of Labor Statistics. Sources and the full series are in the panel above.