The bulk of the analysis in Anderson v. Griswold, which held that Donald Trump can appear on the primary ballot in Colorado, has been a matter I’ve been puzzling this weekend. Most of the opinion is not essential to the holding–that is, most of the legal analysis in the opinion concludes that Trump engaged in insurrection (but nevertheless may appear on the ballot). (Rick H. rightly notes earlier that it’s a reason it has “political implications” as the legal implications remain to be seen.)

There are ordinary reasons (e.g., trial courts include alternative or “non-essential” holdings in judicial cases to allow for a better appellate review that might avoid further litigation in the event of a reversal) and cynical reasons (e.g., trial courts want to reach a particular factual conclusion but not issue a particular remedy in a case and choose to discuss both) for such dicta. I don’t want to psychoanalyze in this post.

Instead, I think a reason I’ve been puzzled, on reflection, is because the court made a mistake in how it approached the jurisdictional component: does state law even authorize this kind of judicial review? Viewed through a lens of cases like U.S. Term Limits, Inc. v. Thornton, I think the error in the court’s framing becomes evident.

The Colorado trial court’s decision in Anderson v. Griswold can be broken down into roughly three parts:

[1] The state legislature empowered the the state court to judge the qualifications of presidential candidates [pp. 60-66]

[2] Trump engaged in insurrection under Section 3 of the Fourteenth Amendment [pp. 66-95]

[3] Section 3 does not apply to Trump because [a] the presidency is not an “office . . . under the United States” and [b] Trump did not taken an oath as “an officer of the United States” “to support the Constitution of the United States” [pp. 95-101]

The dicta in [2] precedes the conclusion in [3]. But the court here erred because [3] is actually a subset of [1].

Here’s what I mean. Suppose–in a deliberately absurd hypothetical–a plaintiff sued in Colorado state court and said, “You cannot print Presidential Candidate Smith’s name on the ballot, because Smith is not qualified–the Constitution limits the office of the presidency only to Presbyterians.”

No court would engage in a factual analysis of whether Candidate Smith was actually Presbyterian, only to conclude, “Ah, but the Constitution does not actually have that requirement for presidential candidates.”

Instead, the court would start with state authority. Could the state election code purport to empower a state court to judge the qualifications of a presidential candidate on the basis of religious affiliation? Of course it could not. That’s because states may not add qualifications to presidential candidates, and only the qualifications enumerated in the Constitution may apply to presidents.

Over the years, I’ve come around on this issue of presidential qualifications. I used to be more ambivalent about whether states might add qualifications to presidential candidates. But after increased examination, I’ve become more confident that state cannot add qualifications. That includes presidential primaries, as I’ve written about. And more courts have come around to this view. Justice Kagan in Chiafalo v. Washington noted, “Checks on a State’s power to appoint electors, or to impose conditions on an appointment, can theoretically come from anywhere in the Constitution. A State, for example, cannot select its electors in a way that violates the Equal Protection Clause. And if a State adopts a condition on its appointments that effectively imposes new requirements on presidential candidates, the condition may conflict with the Presidential Qualifications Clause, see Art. II, § 1, cl. 5.”

This is the long tail of U.S. Term Limits, Inc. v. Thornton, which held in 1995 that states cannot add qualifications to congressional candidates beyond those enumerated in the Constitution. Courts have come around to view an analogous rule for presidential candidates, too. At its core, it’s not really a question of qualifications, but a question of state power.

Back to the court in Anderson. Its [1] analysis is a question of state power, and it explains, “The question before the Court then is does the Election Code incorporate Section Three of the Fourteenth Amendment? The Election Code states that the presidential primary process is intended to ‘conform to the requirements of federal law,’ which includes the U.S. Constitution. C.R.S. § 1-4-1201. Further, C.R.S. § 1-4-1203(2)(a) provides that political parties may participate in a presidential primary only if the party has a ‘qualified candidate.’ . . . While the Court agrees with Intervenors that the Secretary cannot investigate and adjudicate Trump’s eligibility under Section Three of the Fourteenth Amendment, the Election Code gives this Court that authority.”

But, of course, state law could not empower a state court to judge a qualification that was not enumerated in the Constitution. To do so would run afoul of the understanding of law from Term Limits through Chiafalo. So the opinion should have ended here. Or, at least, this holding should be one of state power. So the court erred in concluding the state legislature empowered it to judge a qualification beyond those enumerated in the Constitution for presidential candidates.

We can recreate the proper court’s analysis as follows:

[1][a] The state legislature empowered the the state court to judge the qualifications of presidential candidates

[1][b] But the state legislature cannot add qualifications to presidential candidates, and because Section 3 does not apply to presidential candidates, the state legislature could not empower this court to scrutinize a presidential candidate’s qualifications under Section 3

[2] Nevertheless, if this court were so empowered to judge the qualifications of a candidate because Section 3 does apply to presidential candidates, Trump engaged in insurrection under Section 3 of the Fourteenth Amendment

While this is a subtle shift, it rightly cabins the state legislature’s power only to those things it has the jurisdiction to do. And the court first determines whether it has authority to proceed under the legislature’s mandate.

I’m not sure the holdings that the presidency is excluded from Section 3 are that persuasive (and I have assumed they apply to the president in the past), and I’m not sure the court’s interpretation of Colorado law is right in the first place, but those aren’t the points I’m getting at here. If one is to argue that Section 3 does not apply to presidential candidates, we quickly shift to the mechanics by which someone tries to challenge the qualifications. And if a state is trying to add qualifications to candidates, then the court should evaluate whether the state has the power to do so–much like the Ninth Circuit’s approach in Schaefer v. Townsend on congressional residency statutes. And it ought to do that first as a jurisdictional matter, a matter of state power, before getting into the merits of the qualification. There is surely more to discuss in this opinion and to come on appeal. But as a one who’s written a lot about ballot access, it’s another mistake in a series of mistakes as courts focus on Section 3 issues in isolation without thinking about how they fit with state power in ballot access disputes more generally.