The Bartlett Test: The Supreme Court of Wisconsin’s New Rule for Assessing the Constitutionality of Partial Vetoes

By: Rick Esenberg, Anthony LoCoco, Luke Berg, and Lucas Vebber

Until recently, it was commonly remarked that Wisconsin’s governor has one of the most powerful veto pens in the country. This is not, or should not be, a source of pride. While most think of a “veto” as the executive’s power to disapprove of a particular legislative policy proposal, the Wisconsin Supreme Court has for decades allowed the governor much broader veto authority, at various times permitting the rejection not only of whole bills or policy proposals, but also individual paragraphs, sentences, words, letters, and digits. The governor’s ability to, for example, strike out the word “not” from a bill and thereby enact a law that provides for the exact opposite of what the Legislature drafted makes a mockery of the entire governmental framework that the Wisconsin Constitution established. Under our constitution the power to legislate is vested in the Legislature alone.

But that all changed on July 10 when the Supreme Court finally said “enough” and imposed serious limits on the governor’s partial veto power in Bartlett v. Evers, striking down three out of the four sets of vetoes challenged by the Petitioners (whom we represented). Because no one of the five opinions spanning 140 pages garnered the support of a majority, commentators were quick to presume that the Court handed down a result without a rule to apply in future cases. Not so fast. The bench and bar have long applied a doctrine — sometimes referred to as the Marks Rule (after Marks v. United States, 430 U.S. 188 (1977)) to extract binding precedent from a mishmash of non-binding writings.

An understanding of preexisting partial veto precedent, combined with a careful read of each of the Bartlett opinions and faithful application of the Marks Rule, reveals that the new test for the constitutionality of partial vetoes going forward is threefold:

First, the law that remains after the exercise of the partial veto must be complete, entire, and workable (see, e.g., State ex rel. Wisconsin Telephone Company v. Henry, 218 Wis. 302, 260 N.W. 486 (1935));

Second, the law that remains after the exercise of the partial veto must involve the same subject matter as the original legislative proposal (see, e.g., State ex rel. Wisconsin Senate v. Thompson, 144 Wis. 2d 429, 424 N.W.2d 385 (1988));

Third, the governor may not exercise the partial veto to create a new policy the Legislature did not itself propose or pass (see Bartlett v. Evers, 2020 WI 68, No. 2019AP1376-OA (2020)).

The first two parts of this test have already been the law for decades; a majority of the justices in Bartlett adopted the third requirement. The new test better preserves the separation of powers, largely restoring the legislative power to the Legislature where it belongs. Under Bartlett, Governor Evers — and all future governors, of whatever party — will need to apply the partial veto much more judiciously. This is a big win for Wisconsin.

I. The Bartlett Decision

The Bartlett litigation began when three taxpayers, the Petitioners, challenged four sets of vetoes Governor Evers had made to the 2019–21 biennial budget. This piece will not set forth in full the facts pertaining to the relevant budget proposals or their veto, which are available elsewhere. But in a nutshell, Governor Evers creatively used the veto — striking paragraphs, sentences, and words — to: (1) transform a grant program for replacing school buses into one that would devote millions to electric vehicle charging stations; (2) remove virtually all conditions on money intended for local road improvement so that the funds could be used elsewhere; (3) disrupt a uniform fee schedule for vehicle registration; and (4) alter the definition of “vapor products” subject to new taxation.

The Petitioners went directly to the Wisconsin Supreme Court and sought original review there, which the Court granted. Their main argument was that the original public meaning of the text of Article V, § 10 (the partial veto power) precluded these broad uses of the veto. This was, in the words of Justice Hagedorn, a “big ask” in that it required overruling a number of Court precedents spanning back to the 1930s.

About one year after the petition was filed, on July 10, 2020, the Supreme Court agreed in a per curiam opinion (i.e., authored “by the Court” as opposed to a specific judge) that three out of the four sets of vetoes challenged violated the Wisconsin Constitution (it upheld the vehicle registration fee vetoes). The Court’s seven justices then set out their individual reasoning in four opinions, none of which enjoyed the support of a four-justice majority.

To understand how Bartlett binds future governors, it’s necessary to take a brief look at each opinion.

Justices Hagedorn and Ziegler: Negation, Not Creation

Justice Hagedorn agreed with the Petitioners that the current reading of the partial veto was inconsistent with the state separation of powers. However, Hagedorn rejected the claim that the Supreme Court’s errors in interpreting the partial veto provision went all the way back to the landmark 1935 case of State ex rel. Wisconsin Telephone Company v. Henry, when the Supreme Court first permitted the governor to veto parts of bills smaller than discrete policy proposals.

Instead acknowledging that the governor could permissibly engage in “some modification of the legislature’s policy choice,” he identified the 1978 case of State ex rel. Kleczka v. Conta, 82 Wis. 2d 679, 264 N.W.2d 539 (1978) as the relevant inflection point, wherein the Supreme Court “allowed the governor to . . . edit the words” of a particular policy proposal to create a new one. In Hagedorn’s view that went too far. His proposed test (as agreed to by Ziegler), therefore, was as follows: “When presented with an appropriation bill containing various legislative proposals, the governor can — as a general matter — negate some proposals and accept others. . . . But what the governor may not do is selectively edit parts of a bill to create a new policy that was not proposed by the legislature. He may negate separable proposals actually made, but he may not create new proposals not presented in the bill.”

One way to view Justice Hagedorn’s “negation, not creation” standard is as one that simply asks whether the “greater” proposal (the original bill) includes the “lesser” proposal (the bill as vetoed). A hypothetical provided by Hagedorn himself illustrates the point. Where a bill appropriates money for a house that may be white, blue, or brown, the governor may veto the word “brown,” leaving only the other two options. Those options were always available and had in fact been passed on by the Legislature. What the governor could not do in that circumstance is strike words to eliminate the condition that the money be used to fund a house: that would be the creation of a brand new policy (an unrestricted pot of money).

Only the veto to the vehicle registration fee schedule survived this test; in that case, Justice Hagedorn explained, the governor had simply chosen from the menu of fee changes those he approved and those he disapproved.

Justices Kelly and Rebecca Grassl Bradley: “Negation, Not Creation” Plus

Justice Kelly, who also saw serious separation of powers problems inherent in the governor’s exercise of the veto, penned the most restrictive view of its scope. Like Justice Hagedorn, Kelly (along with Justice Rebecca Grassl Bradley) believed that, following the governor’s exercise of the partial veto, the remaining bill “must . . . be a law on which the legislature actually voted.” But he layered over this requirement an additional one: “the part of the bill not approved must [also] be one of the proposed laws in the bill’s collection.” This conclusion followed from essentially two propositions: first, that the partial veto “cannot act against any division less than the most elemental part” of an appropriations bill; and second, that the most elemental part of a bill is “a proposal for a complete, entire, and workable law” (what Kelly sometimes referred to as an “idea”).

The practical difference between Justice Hagedorn’s view and Justice Kelly’s view, then, revolves around the smallest element that a governor might permissibly veto from a bill. For Justice Hagedorn, that could be as small as a single word, so long as it leaves behind policy items the Legislature has actually approved; for Kelly, only entire and grammatically separable proposals may be excised. Returning to Hagedorn’s house hypothetical (see above), Kelly would thus apparently find fault with removal of the word “brown” insofar as that word is not, read literally, a discrete proposal. Instead, Kelly’s view of the smallest indivisible part in the example would be the full proposal itself: funding for a white, blue, or brown house. But both Hagedorn and Kelly would object to removal of the requirement that the funding go toward a house.

Chief Justice Roggensack: Germane

Chief Justice Roggensack, writing alone, would have struck down two of the four sets of vetoes based on her interpretation of the preexisting “germaneness” requirement adopted in a 1988 Wisconsin Supreme Court case, State ex rel. Wisconsin Senate v. Thompson. That rule, based not in the constitution but on the “long-standing practical and administrative interpretation or modus vivendi between governors and legislatures,” requires “that the consequences of any partial veto must be a law that is germane to the topic or subject matter of the vetoed provisions.”

There is no dispute that the germaneness requirement is and has been the law since 1988. But its application has never been fully explored until Chief Justice Roggensack’s explication in Bartlett. Moreover, Roggensack and A.W. Bradley’s back-and-forth regarding the germaneness requirement illustrates the significance of Roggensack’s interpretation. She uses the requirement to strike down two partial vetoes that A.W. Bradley would have upheld under her interpretation of the same doctrine. A.W. Bradley views the germaneness requirement as essentially toothless and notes that “no case has rejected a gubernatorial partial veto for defying it.”

Chief Justice Roggensack, however, begs to differ and gives real teeth to the rule. She concludes that the bill being partially vetoed, even after the partial veto, must preserve the “stated legislative idea for which the enrolled bill [i.e. the bill sent to the governor] was passed.” That is, the Chief concludes that the ability to veto a bill in “part” means that to pass the germaneness test the vetoed bill must remain “conceptually” a “part” of the original bill. For this reason, Roggensack struck down Governor Evers’ partial vetoes which turned a school bus modernization bill into a bill for alternative fuels and a bill for local road improvements into a bill for local grants. In each case, the bill as it existed after the partial veto was not “part” of the whole bill as considered and approved by the Legislature and, as a result, the “partial” veto was invalid. Although the Chief claimed that her opinion broke no new ground, she did apply the germaneness requirement in a more substantive way that is different from the tests adopted by Justices Hagedorn/Ziegler and Kelly/R.G. Bradley, but which may lead to results that substantially overlap with the results expected from those tests.

Justices Ann Walsh Bradley and Rebecca Dallet: Complete, Entire, and Workable

Justice Ann Walsh Bradley, joined by Justice Rebecca Dallet, authored the only opinion that would have upheld all four of the vetoes in their entirety. Justice Bradley rejected any calls to change the Court’s established — and permissive — approach to assessing partial vetoes. Under that approach, partial vetoes are generally valid if they leave behind a “complete, entire, and workable law” (subject to certain additional requirements, such as the germaneness requirement). In other words, as long as some kind of law remains, it need not resemble any part of what the legislature passed. Not even the Petitioners thought the challenged vetoes flunked this standard; unsurprisingly, Justice Bradley agreed.

II. The Marks Rule

It is not difficult to see why Bartlett yielded the particular result that it did. A five-justice majority (Roggensack, Ziegler, R. G. Bradley, Kelly, and Hagedorn) agreed to strike down two of the four challenged vetoes and a four-justice majority (Ziegler, R. G. Bradley, Kelly, and Hagedorn) agreed to invalidate a third of the set of four.

What many readers may not know, however, is that the reasoning present in diverse opinions may be stacked in quite the same way as votes to provide a binding rule for future cases. The Supreme Court of the United States has described the so-called “Marks Rule” in this way: “When a fragmented Court decides a case and no single rationale explaining the result enjoys the assent of [a majority], ‘the holding of the Court may be viewed as that position taken by those Members who concurred in the judgments on the narrowest grounds.’” Marks v. United States, 430 U.S. 188, 193 (1977). The Wisconsin Supreme Court, quoting one commentator, has provided a helpful way to consider the rule, explaining that it applies only when “at least two rationales for the majority disposition fit or nest into each other like Russian dolls.”

III. The Bartlett Test

Application of the Marks rule to Bartlett yields a single, clear rule for future cases: the governor may not exercise the partial veto to create a new policy the Legislature did not propose or pass. This is the “negation, not creation” test from Justice Hagedorn’s opinion, which Justice Ziegler joined. The reason that this test is precedential is because Justice Kelly, joined by Justice Rebecca Bradley, agreed that vetoes that transgress this line go too far. True, Justice Kelly would havemadopted an even more stringent rule; but he and Justice Hagedorn both decided that editing laws to produce new proposals never voted on by the legislature is impermissible, producing a four-justice majority for this proposition. For example, all four justices mentioned agreed that the Kleczka Court’s expansive treatment of the partial veto must be overruled (meaning that the case has, in fact, been overruled). And indeed, Justice Hagedorn noted the alignment of views when he said that he “agree[d] with Justice Kelly on the core constitutional limits” and “agree[d] the question is whether the governor vetoed a policy the legislature proposed and passed, which is permissible, or created a new policy the legislature did not propose or pass, which is not.”

Further, as explained, Chief Justice Roggensack’s approach to partial vetoes significantly overlaps with this one, as indicated by her agreement on two of the four vetoes. By conceiving of the germaneness test as one that depends on preservation of the original “stated legislative idea,” Roggensack likewise places the emphasis on investigating whether the end result is legislation actually passed upon by the Legislature. Not coincidentally, her terminology mirrors that used by Justice Kelly, with repeated references to the centrality of the “idea” approved by the Legislature and reflected in the text of a bill.

Together with the preexisting rules discussed — the complete, entire, and workable law test and the germaneness test — the new restriction from Bartlett promises to finally impose true limits on the governor’s use of the partial veto. The jumbled assortment of opinions in Bartlett might not have been the most elegant resolution to this case, but the Court’s decision represents a significant milestone in the restoration of Wisconsin’s separation of powers.