Yesterday Virginia voters ratified a constitutional amendment giving the General Assembly temporary authority to redraw the state’s congressional districts before 2031. Most of what you will read about the result over the next few days will concern seats — specifically, whether Virginia Democrats will in fact net four additional House seats in November under the pre-drawn map that House Bill 29 carries into effect. That is a real and consequential question, and I do not have anything useful to say about it.
What does seem worth saying is that the amendment Virginia voters ratified is not a map, and is not — in any particularly interesting sense — a mid-decade redistricting. The General Assembly did not acquire the power to draw congressional districts. It acquired a conditional power to draw congressional districts: a standing authority whose activation depends on the voluntary behavior of other states, with a 2030 sunset and an explicit carve-out for judicially-ordered action. That is a different kind of object than either the map itself or the general question of who gets to draw maps. It is a rule about when a rule applies.
The argument I want to make does not depend on which party benefits from any particular mid-decade redraw, in Virginia or anywhere else. It depends only on what kind of object the amendment is. Partisan consequences are downstream. The structural question is upstream, and structurally distinct.
The operative clause, lifted from the Virginia Department of Elections’ official explanation, grants the General Assembly the power to redraw Virginia’s congressional districts “in the event that another state redraws” its congressional districts before 2031, and only “without being ordered by a court to do so.” The authority expires October 31, 2030, at which point the Virginia Redistricting Commission resumes its regular schedule with the 2031 cycle.
Three features of that sentence do the structural work. The power is conditional: the General Assembly cannot redraw on its own initiative, only in response to another state’s move. The court-order exclusion is doing real categorical work, sorting voluntary moves from compelled ones. And the sunset is doing temporal work — the authority does not transfer redistricting power from the commission to the legislature permanently, but creates a bounded window during which the normal rules do not apply.
Five months ago, California ratified an amendment with superficially similar content. Proposition 50 — officially titled the Election Rigging Response Act — moved redistricting authority from the state’s Citizens Redistricting Commission to the legislature, adopted a specific pre-drawn map under Assembly Bill 604, and provided for the commission to resume its role after the 2030 census. The parallels to Virginia’s arrangement are obvious, and most of the post-passage coverage has treated the two moves as members of the same family: Democratic states responding to Republican-initiated mid-decade redistricting elsewhere by redrawing their own maps in the opposite partisan direction.
They are members of the same family in the surface sense that both are state constitutional amendments that move redistricting authority mid-decade in a partisan direction. They are not members of the same family in the sense that matters if you care about what sort of rule has just been written into a state constitution. Just read the two texts side by side.
California’s text names triggering states in its preamble. It describes the national political context — an administration-led push for mid-decade redistricting — and responds to it with a specific map that was drafted in anticipation of passage. The California amendment is, in structural form, “because X happened, we hereby do Y.” It is backward-looking, specific, and narrowly occasioned. The states whose actions prompted the response are named. The map that executes the response is named. The provocation and the retaliation are identified together, and the amendment executes a single retaliatory move.
Virginia’s text names no state. The conditional is described abstractly, as a class of events: any state voluntarily redrawing its congressional districts mid-decade, excluding judicially-ordered redraws, triggers the General Assembly’s authority. The amendment is, in structural form, “if any state does X, we may do Y, until 2030.” It is forward-looking, general, and operative whenever the conditional is satisfied. No state is named. No particular map is executed by the amendment itself; House Bill 29 is the legislature’s decision about how to exercise the authority, not the authority itself. The amendment writes a rule. It does not execute a move.
California, in other words, wrote a retaliation and ratified it at the constitutional level. Virginia wrote a trigger strategy and ratified it at the constitutional level.[1] The distinction between the two amendments is the distinction between a specific action and a rule about when actions are permitted. I have written before about rules that govern when other rules apply; this is another instance, and a cleaner one than most.
The court-order exclusion deserves its own moment, because it is the signature of careful drafting. Mid-decade redistricting happens for several reasons. A state legislature may move voluntarily to shift the partisan balance — the category of move the Virginia amendment is trying to respond to. Alternatively, a federal or state court may compel a state to redraw because an existing map violates Section 2 of the Voting Rights Act, the state’s own compactness or contiguity requirements, or some other applicable legal constraint. The two categories look the same on the ground: a new map appears mid-decade. They are structurally different: one is a choice, the other is compliance. The Virginia amendment’s trigger counts only the first. That is the correct answer under any coherent retaliation logic. A rule that counted compliance as a trigger would in effect punish lawful redistricting, and would treat a defect and its remedy as equivalent provocations. Whoever drafted this clause was thinking in strategic-form terms, even if they would not describe the work that way.
The rule also has an immediate retrospective dimension. Multiple states have redistricted mid-decade, voluntarily, in the last several months. The conditional in the Virginia amendment was therefore already satisfied on the day of ratification. The authority activates the instant the amendment takes effect. What Virginia wrote yesterday is both a standing trigger strategy and its first execution, in a single legal instrument.
The choice to ratify all of this at the constitutional level, rather than by ordinary statute, is the part of the structural story that political scientists will recognize as a Schelling move. Commitments have value in proportion to their unconditional character. An ordinary legislative majority can undo an ordinary legislative rule at the next session. A constitutional amendment in Virginia requires two separate legislative passages followed by popular referendum — the same process by which this amendment was ratified. The unwinding cost is asymmetric and high. California made the same choice for the same structural reason. Both states committed at the level where commitment is most durable. The difference is only what was committed to. California committed to a map. Virginia committed to a rule.
One more clause is worth dwelling on, because it does work that the rest of the amendment does not. The ballot question presented to Virginia voters asked whether the amendment would allow the General Assembly to “restore fairness in the upcoming elections.” Restore fairness relative to what? The ballot language does not say. It specifies the norm against which the coming elections are to be measured without specifying the baseline that establishes the norm. This is what it means for a classification to be constitutive rather than descriptive, a distinction Maggie and I have developed elsewhere in the context of algorithmic classification, and one that applies with unsettling generality to political rules that specify norms rather than describe states of affairs. The ballot language does not resolve the baseline question. It forecloses it.
Which is to say: the amendment does not eliminate partisan mapmaking, and nobody seriously claims that it does. What it does is shift the equilibrium from unconditional partisan mapmaking to conditional partisan mapmaking, ratified at the constitutional level, and backed by a trigger strategy whose condition is already met. That is a coordination change, not a fairness improvement. The underlying impossibility — every map embodies partisan choices, and the baseline against which those choices are measured is itself a contested object — has not been resolved. It has been relocated. I have suggested before that impossibility in political institutions does not get solved; it gets relocated, often with considerable craftsmanship. Here it has been relocated from a one-shot dominant-action problem (redistrict or don’t, and if you do, own the move) to a conditional-reciprocity problem (redistrict if and only if another state moves first, and if you do, point at the other state). Both are problems. Neither is fair in any baseline-free sense. The conservation law holds.
Florida’s legislature convenes next Tuesday for a special session that is expected to produce a new congressional map. If that happens, Virginia’s trigger will have been pulled again — not a first activation but a further one, another voluntary mid-decade move in a sequence already underway. The rest of the decade will answer a question the Virginia amendment cannot: whether conditional retaliation at the constitutional level functions as an equilibrium stabilizer, where the mutual expectation of reciprocity suppresses escalation, or as an equilibrium accelerator, where each state’s trigger, having been made more credible by its own ratification, induces more moves in response. Trigger strategies are cheap to write and easy to ratify. Whether they are cheap to operate (or, as game theorists would say, “credible”) is another question. I’ll simply say that they must be “no more expensive to write down and proclaim than they are to actually enforce.” Much of game theory essentially boils down to “why parenting is difficult.”
With that, I leave you with this.
1. ↩ The technical term for the Virginia rule is a trigger strategy: an action rule in a repeated game that specifies when punishment follows defection. The classical formal treatment runs through (my undergraduate advisor) James Friedman‘s 1971 paper on noncooperative equilibrium for supergames and the folk theorems that grew out of it. What is distinctive about the Virginia amendment, from a formal-theoretic standpoint, is that the trigger has been ratified at a level of commitment where the cost of abandoning the rule is higher than the cost of being bound by it. Schelling’s original insight — that the value of a commitment is in the burning of the bridge behind it — applies with particular force to state constitutional amendments, which are the most expensive form of rule-ratification in American state government.