The Supreme Court and the 2020 Census Citizenship Question: Part B-2

This final post in the series on the census analyzes the Court’s opinion regarding the capricious and arbitrary standard of the Administrative Procedure Act. It also discusses the possible real rationale behind the administration’s attempt to add a citizenship question to the 2020 census.

The Administrative Procedure Act

The Administrative Procedure Act established procedures for federal agency formal and informal rulemaking. The Act defines a rule expansively to include any “agency statement of general or particular applicability and future effect designed to implement, interpret, or prescribe law or policy or describing the organization, procedure, or practice requirements of an agency.” The history of the act suggests that matters of great significance should be accorded more elaborate public procedures. [Congressional Research Service r41546] The APA requires agencies to provide a concise statement of the basis and purpose of the rule so that the public obtains an idea of the purpose of and a statement of the basic justification for the rule. [Italics added for later reference.]

According to the Congressional Research Service report, there is a strong presumption that Congress intends judicial review of administrative action. The Act excludes judicial review in two situations: (1) where other statutes preclude judicial review or (2) where agency action is committed to agency discretion by law.

Courts may hold an agency action unlawful and set aside agency action under several circumstances. The most common circumstance occurs when a court finds any agency action, findings, and conclusions to be “arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law.” In the 1983 State Farm case (463 U.S. 29, 42-44) the Supreme Court focused most specifically on this catch-all standard by applying a narrow standard. The Court said, “a court is not to substitute its judgment for that of the agency.” It further stated that a court should only invalidate an agency determination if it fails to “examine the relevant data and articulate a satisfactory explanation for its action including a ‘rational connection between the facts found and the choice made.'” [Note: this case, technical in nature, involved the setting of motor vehicle standards (air bags) – hardly a case with the kind of broad political/democratic impact of the U.S. census.]

Generally, the arbitrary and capricious standard requires an agency to provide the “essential facts upon which the administrative action was based.” An agency must “explain what justifies the determination with actual evidence beyond a conclusionary statement.” Additionally, an agency’s decision must be the product of logical and consistent reasoning. An agency must consider important factors relevant to its action, such as the policy effects of its decision or vital aspects of the problem in issue before it. And an agency must consider less restrictive yet easily administered regulatory alternatives. If an agency’s action cannot meet these requirements it will fail the capricious and arbitrary test.

Finally, the Court has addressed instances where an agency proposes a change in past practice. In these instances, an agency action that departs from a prior policy must acknowledge the change. If it does not, or if it creates an unexplained inconsistency with prior policy, the change is generally viewed as arbitrary and capricious. The agency must show that there are good reasons for the change. In the Encino Motorcars case (136 S. Ct at 2126-27) the Court determined an agency proposing a change must provide a reasoned explanation for the change.

Analysis

The judicial stance on policy change may have influenced the administration’s rewriting of the nexus between history of the census and the citizenship question. This inaccurate summary of the nexus influenced the District Court’s and the Supreme Court’s finding of no violation of the Enumeration Clause. It also may have kept the administration from needing to address the policy change aspect of the citizenship question.

Justice Breyer’s minority opinion that held the Secretary violated the APA centered on the above paragraphs that focus on what an agency needs to do to avoid violating the arbitrary and capricious standard. A key part of avoiding violation of the arbitrary and capricious standard is that there be a “rational connection between the facts found the choice made.” In other words, an agency’s action cannot be poorly reasoned. This is sometimes called “the reasoned decision-making requirement.” Part of such scrutiny may be the need to “find data rather than simply say it can’t confirm because the data are lacking.” [italics supplied] I supply the italics because the Court found that “when it came time for the Secretary to make a decision, the model [for estimating the 35 million people for whom administrative records were not available to determine citizenship] did not yet exist, and even if it had, there was no way to gauge its accuracy.”

A paradox if not a contradiction in the discussion about the APA is Chief Justice’s opinion on pretext. It is difficult for me to accept how one could determine that the Secretary’s sole rationale for the citizenship question was pretextual (not related to the decision to add a citizenship question) and not determine the decision violated the arbitrary and capricious standard of the APA. [The Chief Justice joined the four other conservative members of the Court to find that the Secretary’s actions were not arbitrary and capricious under the APA.]

Inherently, it seems to me, a finding that the rationale was fully pretextual would in and of itself find the absence of a “rational connection between the facts found and the decision made.” A pretext also hides “the basis and purpose of the rule so that the public obtains an idea of the purpose of and a statement of the basic justification of the rule.” The APA directly speaks to agency action “unsupported by substantial evidence” as falling within the arbitrary and capricious standard.

Arguably, one can fully support the idea that a key objective of the APA is ensuring transparency and openness in the rulemaking process. Other than solely internal management actions, the law exempts only agency actions that require “secrecy in the public interest.”

The Voting Rights Act pretext

In his article for the Columbia Law Review, Justin Levitt, a former Deputy Assistant General in the DOJ’s Civil Rights Division, makes this comment:

“I do not believe the DOJ during my tenure was unduly shy about bringing Voting Rights Act cases when the facts and the law indicated a violation, and I do not believe that we were unduly shy about asking for additional legal or evidentiary authority when that additional authority would enhance our ability to enforce civil right law. Despite a deep commitment to enforcing the Voting Rights Act, we never requested that the decennial enumeration include a question relating to citizenship. Nor had the Civil Rights Division of any Justice Department, under any Administration, for the previous 53 years.” [italics supplied]

Levitt goes on to say, “There are two main reasons for the absence of pre-2017 clamor to place a citizenship question on the decennial enumeration questionnaire in order to enforce the Voting Rights Act. First, since the Voting Rights Act was enacted in 1965, existing survey data on citizenship – originally from the ‘long form’ of the Census and now from the successor American Community Survey – has been largely sufficient to bring and win VRA cases. And second, for any additional data to be incrementally useful as an enforcement tool, it must be not only more precise, but more accurate. The Census Bureau’s action is not likely to meet this basic standard.”

Further, Levitt points out that in all key aspects of VRA enforcement, “Census-based information about citizenship rates has – for the entire history of the Voting Rights Act – come from a survey of a representative American sample.” [italics supplied]

Among the final points he makes: “The very communities most likely to need VRA protection are already chronically undercounted. With the addition of a citizenship question, they will likely be much more likely to go untallied, which means that the enumeration is likely to systematically undercount precisely the people who most need the VRA.”

The Court’s opinion finding no determination of arbitrary and capricious action

The five conservatives on the Court found no violation because they determined the Secretary “examined the relative data” and articulated “a satisfactory explanation for his decision,” including a rational connection between facts found and the choice made.” The five members agreed that the Secretary’s decision remained “within the bounds of reasoned decision making.”

I find these conclusions wide of the mark given the patently false objective provided for adding a citizenship question. The Secretary consciously hid the real objective or purpose of adding the citizenship question (earning him his contempt of Congress citation) but otherwise, his thinking was okay. I must admit that this makes no sense to me. One can perhaps have a rational argument aimed at a wrongly specified or non-specified objective/purpose but that works only if one takes a narrow view of a process, one that makes no connection to an objective or purpose.

Justice Breyer’s minority opinion for justices Ginsburg, Kagan, Sotomayor, and himself focused on several aspects of the APA. One dealt with the absence of any extensive recognition of the negative effect of a citizenship question would have on the accuracy of the census. He noted the absence of any consideration of the risks of a less accurate enumeration, of considering how to minimize those risks, and the failure of the Secretary to link his decision with the available evidence. He particularly stressed the professional estimates of enumeration inaccuracy of the census [about 630,000 to 718,000 fewer households responding] – noting how even a few hundreds of thousands of fewer responses could affect the allocation of House seats. He also raised the absence of any information or data identifying how enforcement of the Voting Rights Act had been adversely affected by the absence of the citizenship question.

The Chief Justice was the swing vote here, moving from finding the only stated rationale as pretextual to concluding no violation of the arbitrary and capricious standard of the APA. But one must understand that four three justices, Alito, Thomas, Kavanaugh, and Gorsuch, saw the APA as wholly irrelevant to the Secretary’s decision to add a citizenship question.

Justice Alito concluded that once the constitutional question was satisfied the Secretary’s discretion made his decision exempt from APA review. Justices Thomas, Gorsuch, and Kavanaugh concluded that once the Court determined the Secretary complied with all laws and gave a reasoned explanation for his decision, the Court’s role was completed. But their opinion contained another line of thought important to the case. They opined that the APA does not permit a court to set aside the Secretary’s decision just because the stated rationale for the citizenship question was pretextual. It made no difference, in their opinion, whether the Secretary had unstated or other reasons, such as “political considerations or other Administrative priorities.” Pretext, in other words, “does not amount to evidence of bad faith or improper behavior.”

In sum, four justices, Alito, Gorsuch, Kavanaugh, and Thomas saw (1) no improper behavior in the Secretary’s decision-making and (2) no need to consider the consequences of a highly probable adverse effect on the quality of enumeration of a census with the citizenship question added. To me, this argues that these justices have a very expansive view of executive powers. That is, unless constrained by explicit and perhaps narrowly targeted constitutional and legislative constraints, the executive branch, defined as the president and all executive-related agencies, boards, and commissions, has very significant discretionary, and perhaps even all-encompassing, powers. It suggests that the Court will produce such decisions with some degree of frequency, at least under a Republican administration. But that is an issue for another day.

Concluding Comments

The rationale for the citizenship question

If the Voting Rights Act was a pretextual reason for the citizenship question, we are still left wondering what the real rationale might have been. Levitt identifies several real rationale alternatives for adding a citizenship question: (1) immigration enforcement and (2) depressing the census’s count of non-citizens and of households related to non-citizen households for one or more of several reasons.

These reasons might include (a) reducing the non-citizenship count as a way of documenting the administration’s success in controlling immigration, (b) reducing such a count to affect the allocation of House seats among the states, (c) reallocating within states federal funds so that they flow to sub-state areas on the basis of citizen counts, or (d) fundamentally altering what determines the allocation of House seats among the states – that is using the population of citizens rather than of people to allocate House seats among states or redistrict House seats within a state.

My own view focuses on the decades-long Republican effort of voter suppression. In other words, adding a citizenship question was designed to reduce the population count for the overall benefit of the Republican party or conservative interests more generally. I recall the words spoken in August 1980 by Paul Weyrich. Weyrich was perhaps the most significant player in combining the religious right and the political right into a widespread, permanent force in American elections. In discussing get-out-the-vote tactics, Weyrich said, “I don’t want everybody to vote.” Voting was not about representational democracy. “Elections,” he said, “are not won by a majority of people. They never have been from the beginning of our country, and they are not now. As a matter of fact, our leverage in the elections, quite candidly, goes up as the voting population goes down.” [See Anne Nelson’s Shadow Network]

Weyrich and his associates were aware of the changing demographics – they strongly believed that by 2030 most of the population would probably lean Democratic. By this time it would be much too late to save America, at least from their perspective. Consequently, they saw a need to develop a multi-faceted and long-range strategy that needed to be effectively implemented well before 2030. Part of that strategy, becoming more apparent and widespread as time passed, was voter suppression.

Per the position taken by Joshua Clark, I use the term voter suppression broadly. The term appropriately refers not only to direct acts of voter suppression, such as photo ID requirements, minimizing early voting, and closing certain voting precincts and so on, but also gerrymandering and other acts, such as excluding ex-felons from voting, that reduce the ability to vote or increase the relative weight of some voters as the expense of others. One of the most consequential of these may be voting on a workday, compared to nearly all other countries voting as a national holiday or on the weekend.

Coming back to the Supreme Court, my nonlegal lay opinion is that when the Court splits in a 5-4 decision in which each side is fully partisan (that is, all Democrat-nominated justices on one side and all Republican-nominated justices on the other side), the opinions reveal that the justices see the same facts differently and/or prioritize the same facts differently. In other words, the opinions are less a matter of reading the law and more a matter of how facts affect the reading of the law.

Thus, arguably a sequence of Court decisions, such as Shelby v. Holder (2013) which invalidated a key part of the Voting Rights Act, Rucho et al. v. Common Cause et al. (2019) and the follow-up Michigan gerrymandering case that determined the federal courts had no jurisdiction over state-based gerrymandering, and the Court’s 5-4 decision in the 2020 census case that found no violation of the APA follow and reinforce the conservative voting suppression initiatives that began in the late ’70s.

One can also suggest the strong possibility that four justices, Alito, Gorsuch, Kavanaugh, and Thomas, take a very expansive view of the power of the executive. In this instance, it becomes difficult to argue against decisions made by a president (and all executive-related agencies under his control according to the strong unitary executive theory). These justices may be strongly inclined from the very start of a case to favor a president in any case that directly involves the president.

Has the objective been accomplished?

This portion of the concluding comment suggests that regardless of the Court’s decision, the objective of reducing the enumeration of non-citizens may already have been accomplished. Part of this suggestion relates simply to the adverse environment regarding non-citizens and immigration generally that now permeates the country. But the focus here is on the nature in which the time and other resources take up regarding the uncertainty about adding a citizenship question may produce results like a census that included a citizenship question.

The extraordinary publicity given to citizenship may have heightened reluctance of non-citizens and families containing non-citizens from responding to the census. But beyond this, the controversy probably added or magnified technical and administrative difficulties in preparing the 2020 census. An earlier post in this series covered some of these, but the remainder of this post highlights some others.

Charlotte Schwartz, Jeffery Zalesin, and Rachel Brown raise in their law review article three such technical issues: address canvassing, nonresponse follow up, and census coverage measurement. An earlier post briefly covered nonresponse follow up, but they provide additional detail that some readers may find informative.

Address canvassing. The Census Bureau determines where people live through a process called address canvassing. The Bureau must confirm addresses for nearly 140 million households. For 2020, the Bureau is using a new, cost-cutting approach. This new approach has caused some concern. The Bureau appears to have miscalculated the amount of effort, testing, and funding required. People of color reside in the types of housing units most likely to be missed.

In the past, address canvassing enumerators walked every block and identified units that had changes of address. This method of canvassing cost about $443 million and required nearly 150,000 field workers. For 2020 the Bureau canvassed addresses in-office rather than in the field, using satellite imagery to evaluate each of the 11.1 million blocks covered by the census. But the Bureau discovered that more addressed canvassing than expected must take place in the field. Consequently, the Bureau discontinued a second phase of imagining. This adversely affected the rollout of other new programs the Bureau had planned to use.

The single end-to-end test that occurred in Rhode Island indicated that the share of addresses that needed in-field canvassing was about 38% rather than 30%. Other glitches occurred due to software issues. Budget problems caused the Bureau to eliminate plans to update address lists in Indian lands in North and South Dakota and Washington.

Census coverage measurement. This refers to how the Bureau determines the net undercount. An inadequate net undercount means those most affected by an undercount cannot tell they have been overlooked and they are powerless to do anything about it. More importantly, knowing how wrong the census count was is essential to making corrections that will reduce future undercounts. Observers express concern that the Bureau coverage measurement actions appear to be no better than the census regarding identifying low visibility housing units. Undocumented immigrants disproportionately live in these kinds of housing units. These units include living spaces above commercial establishments, in basements, in garages, and living in multiple-family households in units identified as single-family. When combined with the address canvassing issues raised above, the net undercount could be problematic in its accuracy.

Additionally, the Bureau’s coverage measurement becomes more problematic as the level of nonresponse follow up occurs. Many see this as likely given the hostile environment immigrants generally face today and to the Bureau’s intention to rely more heavily on proxy interviews and administrative records. Perhaps most significantly, the Bureau canceled its CCM in its only end-to-end test before implementing the 2020 census.

A final note

This long series of posts on the census contains the common thread of voter suppression. The country now faces many problems> These range from a very high degree of income/wealth inequality, to racism and xenophobia, to access to affordable health care, to the deterioration of public facilities, and so on. Readers may assign their priorities to these problems.

To me, however, these problems will never be well addressed unless we minimize voter suppression. Since the mid-’90s, our nation has been more influenced by minority rule than by majority rule because of structural constraints and explicit policy and political actions. We cannot sustain and revitalize our democracy without increasing peoples’ opportunities to vote. We will not be able to address, much less solve, our other problems without increasing access to voting. And we need to define voter access broadly, otherwise, singular improvements will have only marginal benefits.