Although the method is verboten at the U.S. Supreme Court, a few states allow their supreme courts to give “advisory opinions” to the state legislature on pending legislation. The idea is, why risk passing important laws that might eventually be ruled unconstitutional if the court can give a “heads up” either way ahead of time? Colorado is one of those states. Under Article VI, Sec. 3 of its constitution the state supreme court can “give its opinion upon important questions upon solemn occasions.” Earlier this month, the court ruled in an interesting “solemn occasion” case, both on questions of the scope of legislative power and judicial independence.

Colorado has recently experimented with how it redistricts its congressional and legislative maps after a new census. Ever since the “one person, one vote” revolution of the 1960s, when the U.S. Supreme Court ruled that the Constitution mandates districts of equal population (including in a Colorado case, Lucas v. Forty-Fourth General Assembly), states have experimented—often unsuccessfully—with how to meet that mandate. In 2018, after the Colorado legislature frequently failed to redraw district lines without the courts getting involved, the voters adopted two ballot initiatives amending the state constitution to give the redistricting task over to two “independent commissions.” One commission is for congressional seats and one for the state legislature, but otherwise the powers and duties of each are fairly identical.

Using “necessary census data,” as the 2018 amendments say, the commissions are charged with applying the data to certain criteria for drawing new congressional and legislative maps, and also to solicit public comment and allow for judicial review of their decisions. Thus, the amendments to the constitution in a sense divested the legislature of certain powers and gave those powers to the commissions. The commissions are composed of a balance of voters of each major party, plus unaffiliated voters, and (supposedly at least) nonpartisan staff.

But the commissions have encountered a problem. The redistricting enterprise depends on the numbers the census comes up with. Of course, the 2020 Census has been greatly delayed because of the pandemic. The initial census numbers of the population of each state (and resulting gain or loss of congressional seats) was delayed by months, and the more detailed data that’s really needed to redraw maps—how many people live in each precinct—is similarly delayed. The amendments refer to “necessary census data” generically, but they also impose deadlines throughout the year after a census, e.g., 2021, for when maps are to be drawn. And those deadlines were based upon when 2020 Census data—what I’ll call the “hard-count numbers” that come out every 10 years—were expected to be available.

So what should the new commissions do in the face of the delay? Well, the legislature stepped in and proposed a couple of things in draft legislation: First, that the commissions be allowed to use other kinds of census data, such as survey data and preliminary data from the census; second, that the commissions’ actions only be reviewable in court if they do not evidence “substantial compliance” with the state constitution, giving them a bit of wiggle room in meeting the various deadlines and other requirements.

Since the state constitutional language is brand new, the pandemic/deadline problem is pretty unprecedented, and the commissions themselves were actually resistant to this legislative assertion of authority, the legislature asked the Colorado Supreme Court for an advisory opinion, asking two separate “interrogatories”: (1) is it constitutional for the legislature to instruct the commissions to use census data other than the “hard count” from the 2020 Census, and (2) is it constitutional for the legislature to dictate that when the state courts review the commissions’ work they apply a “substantial compliance” standard?

In In re Interrogatories on Senate Bill 21-247, the court said the answer on both counts is a pretty firm “no.” The court did say that it’s fine for the commissions to use census data other than the hard count, as “necessary census data” isn’t defined in the Constitution. This is especially true because the hard count data won’t be ready in time to make the commissions’ constitutional deadlines, and they can go back later and backfill the hard count data once it becomes available and make any needed corrections. But even though the court said it’s fine for the commission to do this, it also said it’s not ok for the legislature to instruct the commission to do this.

Here’s where the opinion gets really interesting. State legislatures, unlike Congress, have general powers. Essentially all the normal powers of government are in the hands of the legislature. As I explained earlier this year, under our “Lockean” theory of government, the people of a state delegate the power to enact reasonable (but not unreasonable!) legislation to the state legislature. Contrary to Article I of the U.S. Constitution, which gives Congress certain enumerated powers, state constitutions don’t spell out all those powers (they could—if the people in a constitutional convention said they were only delegating enumerated powers—but that’s never how it’s done). State legislative powers simply are the unenumerated, delegated, powers of sovereignty. Now, again, those powers only go as far as the people have delegated them; that is, they have to be for the public’s health, safety, or welfare. And they can be further limited through the state’s Bill of Rights and other constitutional safeguards such as a bicameral legislature, judicial review, etc. But, in general, if a power isn’t denied to the legislature, it’s reasonable, and if doesn’t conflict with something in the constitution, it’s constitutional. (It has to comply with the U.S. Constitution too, of course, but that’s another story not relevant here.)

However, that’s not true when it comes to powers that affect coordinate branches of the government, such as the executive or the judiciary. And in this case the redistricting commissions are a coordinate “branch.” The people voted to take the powers of redistricting away from the legislature and give them to the commissions in the first place. Thus, the only powers the legislature has left regarding the commissions are those expressly still given to it—the opposite of the usual approach. And since dictating the use of data and timelines is not an express power, that part of the proposed legislation would be unconstitutional. As the court succinctly explained: “any power that the General Assembly asserts over a constitutionally created independent commission that was expressly designed to divest the legislature of authority must derive from the amendment that created the commission, not the constitution’s general grant of legislative authority.”

You might criticize this approach of taking power from the legislature on a variety of grounds—that it itself subverts democratic values, creates a “fourth branch” of government shielded from the voters, or unjustly vests “so-called experts” with too much authority. Similar criticisms are made of federal administrative agencies all the time. But, not to diminish those critiques, in Colorado’s case at least the “fourth branch” is in the state constitution. The 2018 initiatives plainly sought to establish a “fourth branch,” limited in scope as it may be. In denying the legislature’s power the court was recognizing that reality. (It should be noted that, for good or for ill, “fourth” or “fifth” or whatever branches are actually pretty common in state constitutions. They frequently assign powers to boards or commissions or other bodies, although always with much more limited power than the legislature itself.)

Now on to the second question presented to the court, the answer to which is interesting from a judicial engagement perspective. Again, the proposed legislation directs the state supreme court (which would review any challenges) to apply a “substantial compliance” standard to challenges to the commissions’ work. Thus, if someone went to court to oppose the maps the commissions draw, the courts would evaluate whether the commissions “substantially complied” with the dictates of the state constitution, and not throw anything out on a “technicality.” The court said, sure, it might apply that standard, but whether it does or not is its own prerogative. Citing Marbury v. Madison, it said that interpreting the constitution is a core aspect of the “judicial power” and can’t be dictated by the legislature.

With the courts “no” answer on both questions the proposed legislation has, unsurprisingly, died. We’ll now wait and see how the commissions do without the legislation.

Let’s close with a comment on the court’s answer to the second question. It’s great when judges assert their independence to interpret constitutions, whether state or federal, as that places a check on the legislature trying to inoculate itself from judicial review, or, to put it more exactly, inoculate itself from constitutional limits. But, unfortunately, courts often do this without the legislature even asking them to in the first place. The most obvious example is with the rational-basis test. In deferring to the legislature on whether a law violates a constitutional provision—as the rational-basis test demands judges do in a sometimes unrebuttable way—courts are essentially allowing the legislature to dictate their own interpretation of the Constitution, instead of the courts independently making their assessment. This flies in the face of Marbury’s injunction that the judiciary interpret “what the law means” because it abdicates that duty to the legislative branch. That’s a core lesson of Marbury that is so often forgotten. Sometimes, however, courts realize the contradiction. Let’s hope the Colorado Supreme Court does so more often in the future.

Anthony Sanders is the director of the Center for Judicial Engagement at the Institute for Justice.