The Conversation

By Derek T. Muller, Professor of Law, University of Notre Dame.

The recent FBI search of the Fulton County, Georgia, elections facility and the seizure of election-related materials pursuant to a warrant has attracted concern for what it might mean for future elections.

What if a determined executive branch used federal law enforcement to seize election materials to sow distrust in the results of the 2026 midterm congressional elections?

Courts and states should be wary when an investigation risks commandeering the evidence needed to ascertain election results. That is where a largely forgotten Supreme Court case from the 1970s matters, a case about an Indiana recount that sets important guardrails to prevent post-election chaos in federal elections.

Congress’s constitutionally delegated role

The case known as Roudebush v. Hartke arose from a razor-thin U.S. Senate race in Indiana in 1970. The ballots were cast on Election Day, and the state counted and verified the results, a process known as the “canvass.” The state certified R. Vance Hartke as the winner. Typically, the certified winner presents himself to Congress, which accepts his certificate of election and seats the member to Congress.

The losing candidate, Richard L. Roudebush, invoked Indiana’s recount procedures. Hartke sued to stop the recount. He argued that a state recount would intrude on the power of each chamber, the Senate or the House of Representatives, to judge its own elections under Article I, Section 5 of the U.S. Constitution. That clause gives each chamber the sole right to judge elections. No one else can interfere with that power.

Hartke worried that a recount might result in ballots that could be altered or destroyed, which would diminish the ability of the Senate to engage in a meaningful examination of the ballots if an election contest arose.

But the Supreme Court rejected that argument.

It held that a state recount does not “usurp” the Senate’s authority because the Senate remains free to make the ultimate judgment of who won the election. The recount can be understood as producing new information — in this case, an additional set of tabulated results — without stripping the Senate of its final say.

Furthermore, there was no evidence that a recount board would be “less honest or conscientious in the performance of its duties” than the original precinct boards that tabulated the election results the first time around, the court said.

A state recount, then, is perfectly acceptable, as long as it does not impair the power of Congress.

In the Roudebush decision, the court recognized that states run the mechanics of congressional elections as part of their power under Article I, Section 4 of the U.S. Constitution to set the “Times, Places and Manner of holding Elections for Senators and Representatives,” subject to Congress’s own regulation.

At the same time, each chamber of Congress judges its own elections, and courts and states should not casually interfere with that core constitutional function. They cannot engage in behaviors that usurp Congress’s constitutionally-delegated role in elections.

Evidence can be power

The Fulton County episode is legally and politically fraught not because federal agents executed a warrant — courts authorize warrants all the time — but because of what was seized: ballots, voting machines, tabulation equipment and related records.

Those items are not just evidence. They are also the raw materials for the canvassing of votes and certification of winners. They provide the foundation for audits and recounts. And, importantly, they are necessary for any later inquiry by Congress if a House or Senate race becomes contested.

That overlap creates a structural problem: If a federal investigation seizes, damages, or destroys election materials, it can affect who has the power to assess the election. It can also inject uncertainty into the chain of custody: Because ballots are removed from absentee envelopes or transferred from Election Day precincts to county election storage facilities, states ensure the ballots cast on Election Day are the only ones tabulated, and that ballots are not lost or destroyed in the process.

Disrupting this chain of custody by seizing ballots, however, can increase, rather than decrease, doubts about the reliability of election results.

That is the modern version of “usurpation.”

From my perspective as an election law scholar, Roudebush is a reminder that courts should be skeptical of executive actions that shift decisive control over election proof away from the institutions the Constitution expects to do the judging.

Congress doesn’t just adjudicate contests

There is another institutional reason courts should be cautious about federal actions that seize or compromise election materials: The House already has a long-running capacity to observe state election administration in close congressional races.

The Committee on House Administration maintains an Election Observer Program. That program deploys credentialed House staff to be on-site at local election facilities in “close or difficult” House elections. That staff observes casting, processing, tabulating and canvassing procedures.

The program exists for a straightforward reason: If the House may be called upon to judge a contested election under Article I, Section 5, it has an institutional interest in understanding how the election was administered and how records were handled.

That observation function is not hypothetical. The committee has publicly announced deployments of congressional observers to watch recount processes in tight House races throughout the country.

I saw it take place first-hand in 2020. The House deployed election observers in Iowa’s 2nd Congressional District to oversee a recount of a congressional election that was ultimately certified by a margin of just six votes.

Democratic and Republican observers from the House politely observed, asked questions, and kept records – but never interfered with the state election apparatus or attempted to lay hands on election equipment or ballots.

Congress has not rejected a state’s election results since 1984, and for good reason. States now have meticulous record-keeping, robust chain-of-custody procedures for ballots, and multiple avenues of verifying the accuracy of results. And with Congress watching, state results are even more trustworthy.

When federal investigations collide with election materials

Evidence seizures can adversely affect election administration. So courts and states ought to be vigilant, enforcing guardrails that help respect institutional boundaries.

To start, any executive branch effort to unilaterally inject itself into a state election apparatus should face meaningful scrutiny. Unlike the Fulton County warrant, which targeted an election nearly six years old, warrants that interrupt ongoing state processes in an election threaten to usurp the constitutional role of Congress. And executive action cannot proceed if it impinges upon the ultimate ability of Congress to judge the election of its members.

In the exceedingly unlikely event that a court issues a warrant, a court should not permit seizure of election equipment and ballots during a state’s ordinary post-election canvass. Instead, inspection of items, provision of copies of election materials, or orders to preserve evidence are more tailored means to accomplish the same objectives. And courts should establish clear chain-of-custody procedures in the event that evidence must be preserved for a future seizure in a federal investigation.

The fear driving much public commentary about the danger to midterm elections is not merely that election officials will be investigated or that evidence would be seized. It is that investigations could be used as a pretense to manage or, worse, disrupt elections – chilling administrators, disorganizing record keeping or manufacturing doubt by disrupting custody of ballots and systems.

Roudebush provides a constitutional posture that courts should adopt, a recognition that some acts can usurp the power of Congress to judge elections. That will provide a meaningful constraint on the executive ahead of the 2026 election and reduce the risk of intervention in an ongoing election.