Congressional Analysis

October 10, 2026 • Updated October 14, 2026 • 3 min read• 534 words

Analysis by Anthropic’s Claude.AI of the prompt:

compare, using numbers from 2014 to 2024, the probability
of a US congressional representative or senator voluntarily
leaving office or dying verses the probability that they
are voted out of office.

From 2014 to 2024, members of Congress were about 2.2 times more likely to leave voluntarily or die than to be voted out. That is about 9.8% of seats per two-year cycle against 4.5%. The ratio ranged from 1.5 times (2018) to 3.4 times (2016).

Cycle	House not running	Senate not running	Deaths	Leave or die	House defeated	Senate defeated	Voted out	Leave or die	Voted out	Ratio
2014	41	7	2	50	18	4	22	9.3%	4.1%	2.3×
2016	40	5	2	47	12	2	14	8.8%	2.6%	3.4×
2018	52	3	2	57	34	5	39	10.7%	7.3%	1.5×
2020	36	4	3	43	21	5	26	8.0%	4.9%	1.7×
2022	49	6	6	61	25	0	25	11.4%	4.7%	2.4×
2024	45	7	4	56	15	3	18	10.5%	3.4%	3.1×
Total	263	32	19	314	125	19	144
Per cycle	44	5	3	52 (9.8%)	21	3	24 (4.5%)			2.2×

Percentages are out of 535 members.

Where each number comes from

House retirements: Ballotpedia’s historical table (41, 40, 52, 36, 49, 45). Other sources disagree on a few cycles (for example 2016 at 28, and 2020 at 37 or 48).
House defeats: Ballotpedia’s totals for primaries and general elections combined (18, 12, 34, 21, 25, 15).
Senate defeats: Ballotpedia’s Senate election pages for 2016 to 2024 (2, 5, 5, 0, 3), with 4 for 2014.
Senate retirements: Senators not seeking reelection in each cycle (7, 5, 3, 4, 6, 7). These counts are approximate and could be off by a few.
Deaths: Members who died in office, counted by Congress, such as the 113th (2013 to 2014) for the 2014 row. The list may be missing a few.

Limitations

“Not running” includes members who left to run for the Senate or governor. Those members left the seat but not public life. The 2026 NPR tracker puts that group at roughly 40% of the total. If you count only members leaving public life, the voluntary exits fall to about 6%, and the gap with defeats narrows to roughly 1.3 times.
Some House defeats were incumbents facing each other after redistricting.
A member who announced retirement and then died is counted twice, which is a minor effect.
Deaths are only about 1% of the “leave or die” group, so retirement drives that total.
Nearly all defeats came from the House, and the 2018 wave (34 House defeats) is the main reason the ratio varies by cycle.

Sources: