The Safe Seat Problem: How American Democracy Broke Itself

Here is a number worth sitting with: 97.

That is the percentage of congressional incumbents who won reelection in 2024, according to Ballotpedia. Not 97 percent of the good ones, or the effective ones, or the ones who actually showed up. Ninety-seven percent of all of them. The ones with documented cognitive decline. The ones who missed a third of their votes. The ones who broke public pledges, accepted money from fraudulent crypto exchanges, voted against infrastructure bills and then celebrated the funding. Ninety-seven percent. In 41 states, every single congressional incumbent who sought another term got it.

If you are wondering why Congress looks the way it looks, that number is most of the answer.

The American political system has quietly engineered itself into something the Founders would not recognize and would probably find alarming: a legislature where the primary qualification for holding power is already having it, where the main barrier to entry is the incumbent's continued existence, and where the voters' theoretical authority to remove their representatives has been rendered, in most of the country, almost entirely theoretical.

This is not an accident. It is a design. And it is worth explaining, plainly, how it happened and who benefits from it staying that way.

The 2026 Exception (and Why It Proves the Rule)

There is a caveat worth naming before the argument proceeds.

In 2026, the retirement numbers look different. As of this writing, 56 House members have announced they will not seek reelection — the highest number in over 30 years, and the second-most in any single cycle since 1992, according to Brookings Institution analysis of Vital Statistics on Congress. Thirty-six of them are Republicans. The Hill called it an "unprecedented" exodus. NPR is tracking it in real time. But look at why they are leaving, and the safe seat argument gets stronger, not weaker.

Most of these members are not retiring because accountability mechanisms worked. They are leaving because Republicans read the midterm tea leaves and decided they would rather run for Senate or governor or get out of Washington entirely than face a potential blue wave with a three-seat majority and a president whose approval ratings are underwater. According to Brookings, the average tenure of retiring Republicans is just five terms — these are largely not the 20-year entrenched veterans DNR Congress targets. They are newer members frustrated by institutional dysfunction and declining career prospects in a gridlocked chamber.

The genuine long-timers, the ones who built the machine and benefit most from it, are largely staying. Hal Rogers is running for his 23rd term at 88. Gary Palmer broke his five-term pledge and is running for his seventh. The members most insulated from accountability are, predictably, the least likely to leave. The retirement wave is a political calculation by members who can read polls. It is not accountability. It is self-preservation wearing accountability's clothes.

The Map Is the Message

Start with the geography, because everything else flows from it.

Every ten years, following the census, congressional district lines are redrawn. In most states, that process is controlled by the state legislature. State legislators are, overwhelmingly, members of one of the two major parties. Those parties have an obvious interest in drawing maps that protect their incumbents and maximize their own seat counts. The result, repeated across decades and across party lines with bipartisan enthusiasm, is a Congress whose districts are pre-sorted for partisan safety before a single vote is cast.

According to the Cook Political Report's 2024 House race ratings, only 69 out of 435 House seats were considered competitive in the general election. The remaining 366, or roughly 84 percent, were rated solid or likely for one party. In 20 states, there was not a single competitive congressional district. Not one. Twenty states where the outcome of every federal House race was effectively predetermined by the lines on the map.

The Princeton Gerrymandering Project has documented this in granular detail, measuring the shape, compactness, and partisan fairness of congressional maps across all 50 states. What they find, year after year, is that districts drawn by self-interested legislatures are less compact, more partisan, and more protective of incumbents than districts drawn by independent or court-supervised processes. The maps, in short, are doing exactly what the people who drew them wanted them to do.

Democrats do this. Republicans do it more aggressively in more states at the moment, but Democrats do it wherever they can. Jim Clyburn's own aide, according to a 2023 ProPublica investigation, hand-delivered a map to South Carolina Republicans during the 2021 redistricting process — a map that added Black voters to Clyburn's district and made every other South Carolina congressional seat effectively unwinnable for Democrats. The result: Clyburn gets a safe seat, and South Carolina Democrats lose any realistic shot at representation anywhere else in the state. Both parties sacrificed the interests of their voters for the protection of their incumbent. The voters in question had no say in any of it. 

Mid-Decade Redistricting: Same War, Different Weapons

The decennial process described above is complicated in 2026 by something relatively new: mid-decade redistricting battles that are reshaping maps between census cycles, with consequences that are already landing.

Texas, under Republican legislative control, redrew its congressional map in a move aimed at eliminating five Democratic-leaning districts. The Supreme Court approved the new map in time for the 2026 cycle. The result was immediate: nine Texas incumbents announced they would not seek reelection rather than face reconfigured districts. Among them, 78-year-old Democratic Rep. Lloyd Doggett chose retirement rather than a primary fight against progressive Rep. Greg Casar in a newly drawn seat. Mid-decade redistricting, in Texas's case, accelerated retirements without producing competitive districts. It simply replaced one set of safe Republican seats with a different set of safe Republican seats, minus five Democrats.

California responded with a voter-approved retaliatory redistricting measure, redrawing the state's congressional lines to more heavily favor Democrats. Virginia's General Assembly announced plans to begin its own pre-2026 redistricting process. Colorado, New Jersey, and New York have all been in the conversation, though legal and political challenges make most of these efforts unlikely to materialize before November.

The mid-decade redistricting wars are, in a sense, an argument that the safe seat problem has become so entrenched that both parties are now willing to blow up the norms around the decennial process to try to gain an edge. This is not reform. It is escalation. When Republicans redraw Texas to eliminate Democratic representation and Democrats retaliate by redrawing California, the net effect on the American voter is more partisan cartography, more predetermined outcomes, and fewer competitive races. The specific incumbents in safe seats may change. The safe seats themselves do not.

The Machine That Protects Itself

The gerrymandered map is the foundation. What gets built on top of it is a machine with several moving parts, each one designed to make dislodging an incumbent as difficult as possible.

The first part is money. Congressional incumbents running in 2022 Senate races raised an average of $29.7 million each, according to OpenSecrets. Their challengers averaged $2.1 million. In House races, the top spender wins 92 percent of the time. This is not a coincidence. Donors give to incumbents because they believe incumbents will win, and incumbents win in part because donors give to them. The cycle is self-reinforcing and nearly impenetrable. A credible challenger in a safe district needs to raise enough money to overcome the incumbent's name recognition, their years of constituent services, their franked mail — actually, scratch that, franking is its own rabbit hole — their donor network, and the party infrastructure that will almost certainly line up behind the incumbent rather than the newcomer. Most people who consider running do the math and walk away.

The second part is the seniority system. Congress runs on tenure. Committee chairmanships, ranking member positions, and the ability to steer federal dollars toward a district all accumulate with years served. A freshman member, however talented, however energetic, however better suited to the job than the incumbent they replaced, starts from zero. The colleague who spent 30 years becoming chairman of the Appropriations Committee does not want to give that up. More importantly, the colleague who has spent 15 years working their way up the queue does not want to see an incumbent taken out, because every time an incumbent leaves — voluntarily or otherwise — it disrupts the orderly progression of power that everyone in the building has been counting on. The institution closes ranks around incumbents not out of affection but out of self-interest. Seniority in Congress is still a currency even where formal term limits exist. House Republicans adopted rules in 1995 limiting committee chairmanships to three consecutive terms — a genuine reform that Democrats have never matched. But term-limited chairs do not lose their influence overnight. Senior members accumulate relationships, favors, and informal power that persist long after the gavel passes. Subcommittee assignments, conference standing, and the ability to move legislation all still flow, in part, from time served. More importantly, the norm of incumbent protection operates independently of any committee rule. A primary upset against a sitting member sends a signal that no incumbent is safe, that the party machine will not protect you, that the seniority you have accumulated can be taken away by voters who were not consulted about the queue. Nobody in that building wants that precedent established. So they endorse the incumbent, stay quiet about the decline, and protect the system that protects them.

The third part is the party infrastructure itself. Parties exist to win elections and hold power. An incumbent in a safe seat, whatever their individual failings, is a reliable vote and a reliable seat. Recruiting a challenger to run against your own incumbent in a primary is, from the party's perspective, an investment with uncertain returns and certain costs. You might end up with a weaker general election candidate. You definitely create internal conflict. You signal to other incumbents that the party will not protect them. So the default is always to back the incumbent, regardless of performance, regardless of age, regardless of whether the voters in that district might be better served by someone else. This is how Raul Grijalva missed 97 percent of his votes in his final year while battling cancer and still no senior Democrat publicly called for him to step down. This is how Dianne Feinstein's colleagues watched her cognitive decline for years and said nothing publicly. Saying something costs you. Saying nothing costs you nothing. The incentives are perfectly aligned against honesty.

What the Primary Was Supposed to Fix

In theory, the primary election is the solution. If the general election in a safe district is predetermined by the partisan map, then the real election happens in the primary, where voters within the dominant party choose their representative. A primary challenger can hold an incumbent accountable in a way the general election cannot. This is true. And it is why primary accountability is the only meaningful leverage most voters in safe districts have.

The problem is that the same machine that makes incumbents nearly invincible in general elections makes them nearly invincible in primaries too. Gary Palmer of Alabama signed a pledge to serve no more than five terms. He is now running for his seventh. When the U.S. Term Limits organization put up billboards in his own district calling him out, he acknowledged the pledge and ran again anyway. His 2024 Republican primary gave him 83 percent of the vote against two underfunded challengers. Case Dixon, a 26-year-old physical therapist running against him in the August 11 primary, has two organizational endorsements and a conscience. Palmer has a decade of donor relationships, party infrastructure, and the name recognition that comes from 11 years of congressional mailings and constituent services and press releases and the simple accumulated weight of being the person voters have always seen on the ballot.

Hal Rogers of Kentucky has been in Congress since 1981. Four challengers filed against him for the May 19 primary. Four. In a district that votes R+32. That is how frustrated KY-5 Republicans are with a member who has represented what remains the second most impoverished congressional district in the United States for 45 years. And Rogers will almost certainly survive all four of them, because he has what they do not: the machine.

The primary accountability mechanism exists. It is just surrounded by walls.

The Generational Tax

There is a cost to all of this that does not show up in any budget or any election result, but that is real and accumulating.

When a seat is held for 30 or 40 or 45 years by the same person, an entire generation of potential representatives never gets to try. The qualified 45-year-old who might have run for that seat a decade ago did the math, looked at the incumbent's war chest and name recognition, and chose a different path. The 35-year-old who came of age watching that same incumbent run unopposed cycle after cycle absorbed the lesson that certain seats are simply not available. The pipeline of political talent that might have developed in those districts, that might have brought fresh ideas and different relationships to Washington, simply does not exist, because the incumbent's perpetual presence foreclosed it before it could form.

Meanwhile, the decisions being made in those safe, entrenched offices affect people who had no voice in choosing the people making them. The average age of a House member at the start of the 118th Congress was 58. The average age of a senator was 65. The median age in the United States is 39. The people making decisions about healthcare costs, student debt burdens, climate policy, and housing affordability are, on average, nearly two decades older than the people who will live with those decisions the longest. That is not just a demographic curiosity. It is a structural misalignment between who holds power and who bears the consequences of how it is used.

When Hal Rogers retires or dies, KY-5 will have a candidate field that has been frozen out of the district's political life for 45 years. When Jim Clyburn's tenure ends, South Carolina's 6th will need to build a new political identity from scratch. The longer a safe seat is held, the more atrophied the surrounding political ecosystem becomes. This is the hidden cost of incumbency entrenchment: not just the policies that do not get made, but the leaders who never get developed.

The Only Lever Left

The general election in most congressional districts is not an election. It is a ratification. The outcome is known before it begins. Twenty states had zero competitive House races in 2024. If you live in one of them, and your representative has been in office for two decades, and the map has been drawn to protect them, you have exactly one opportunity to exercise meaningful democratic power: the primary.

the-safe-seat-problem-how-american-democracy-broke-itselfThis is what DNR Congress was built for. Not to replace the political process, but to inject information into the one moment in that process where information can still change the outcome. Before the primary. While challengers are still on the ballot. While the window is still open.

The Congressional Vitality Index scores sitting members on five categories: age and tenure, electoral accountability, health and capacity, legislative effectiveness, and controversy and ethics. It does not tell voters how to vote. It tells voters what they are actually voting on, stripped of the incumbent's press releases and name recognition and 40 years of accumulated goodwill.

The safe seat problem is not going to be solved by redistricting reform, though that would help. It is not going to be solved by term limits legislation, though that argument has merit. It is certainly not going to be solved by the incumbents who benefit from it, who have every structural incentive to leave it exactly as it is.

It might, just barely, be dented by voters who understand what they are up against and decide to act anyway. Primary accountability is the only constitutional mechanism that works. It is surrounded by walls. But walls have doors.