UTAH v. EVANS
In conducting the 2000 census, the Census Bureau used "hot-deck imputation" to fill in certain gaps in its information and resolved certain conflicts in the data. Under this methodology, the Bureau imputes the relevant information by inferring that the address or unit about which it is uncertain has the same population characteristics as those of its geographically closest neighbor of the same type. Hot-deck Imputation increased North Carolina's population by 0.4% while increasing Utah's population by only 0.2% such that North Carolina will receive one more Representative and Utah one less than if the Bureau had simply filled relevant informational gaps by counting the related number of individuals as zero. Utah brought suit against the officials charged with conducting the census, claiming that the Bureau's use of hot-deck imputation violates 13 USC section 195, which prohibits use of "the statistical method known as 'sampling,'" and is inconsistent with Article 1, section 2, clause 3 of the Constitution, which states that an "actual Enumeration be made." Utah sought an injunction compelling a change of the official census results. The District Court found for the Bureau.
Does the Census Bureau's use of "hot-deck imputation," in the 2000 census, violate the statutory provision forbidding use of the statistical method known as sampling? Is this methodology inconsistent with the Constitution's statement that an "actual Enumeration be made?
Legal provision: 13 U.S.C. 195
No and no. In a 5-4 opinion delivered by Justice Stephen G. Breyer, the Court held that the use of hot-deck imputation violates neither the statute nor the Constitution. The Court reasoned that, whereas sampling seeks to extrapolate the features of a large population from a small one, the Bureau's imputation process simply sought to fill in missing data as part of an effort to count individuals one by one and that these differences placed imputation outside the scope of the prohibitive statute. Furthermore, the Court reasoned that Article 1's wording that the "actual Enumeration" shall take place "in such Manner as" Congress itself "shall by Law direct," suggested a breadth of congressional methodological authority, rather than a limitation.
ORAL ARGUMENT OF THOMAS R. LEE, ESQ. ON BEHALF OF THE APPELLANTS
Chief Justice Rehnquist: We'll hear argument now in number 01-714, Utah vs. Donald L. Evans, Secretary of Commerce.
Mr. Lee: Mr. Chief Justice and may it please the Court: for most of the history of the census of the American population, that census has been conducted by means of an actual count unaltered by any methods of statistical estimation.
At the time of the first census of 1790, James Madison noted that this was the way required by the Constitution and ever since then, that requirement has been implemented by Congress in the Census Act.
The question presented in this case is whether the Census Bureau acted unlawfully in departing from that longstanding historical tradition by using the method of estimation called hot-deck imputation in the 2000 apportionment count.
99.6 percent of the 2000 apportionment count was comprised of actual data compiled by census enumerators in two phases.
First, a mailing sent to all addresses on a master address file, and second, as many as six follow-up visits.
Imputation added an additional.4 percent to that count, 1.2 million statistically generated persons were added to the apportionment count by means of a statistical estimate.
The basic essence of the estimate was to say the 620,000 addresses that the Census Bureau was unable to enumerate are assumed to have estimated to have the same number of occupants as their next door neighbors.
Mr. Dellinger: Mr. Lee, I would like to ask you to address early on whether this injury that Utah alleges to have suffered is redressable in any fashion.
Things have happened since the census was taken.
North Carolina presumably has drawn districts and has gone quite far down the road in reliance on the census.
The President has turned over the numbers and so forth.
How can this be redressed, do you think, now?
Mr. Lee: I suppose, Justice O'Connor, that there are two aspects of your question if I'm understanding it.
One of them would be a redressability standing question.
That's a question as the Solicitor General notes that has already been decided by this Court both in Franklin and in Montana with regard to--
Mr. Dellinger: Well, there is jurisdiction, but I don't know that we have really addressed the redressability issue fully in prior cases.
When could Utah have made its challenge?
When is the earliest time Utah could have challenged this imputation business?
Mr. Lee: --It would have been impossible, Justice O'Connor, for Utah to have brought a challenge prior to the census.
As North Carolina indicated--
Mr. Dellinger: Well, why is that?
I mean, you knew, did you, that the Census Bureau planned to use it?
Mr. Lee: --Yes.
We certainly had constructive knowledge.
Mr. Dellinger: But you wouldn't have had an injury.
Mr. Lee: We wouldn't have had an injury, and in fact as the Census Bureau itself indicates, there would have been no way even to have anticipated that imputation would have had this effect.
Mr. Dellinger: Well, you have no incentive to bring a challenge until you know that you are losing a congressional seat, I suppose.
Mr. Lee: That's exactly right.
And that was one of the problems.
It took us some time to get the numbers from the Census Bureau, specific state by state numbers, in order to be able to determine whether in fact this was a method that impacted Utah, and as soon as we got those numbers, we brought this suit within a few days.
Mr. Dellinger: Well, was there a window of time between completion of the census and the submission of the numbers to the President where you could have sued?
How long was that period of time?
Mr. Lee: That was, the time line is that the Secretary must submit the numbers by the end of the calendar year to the President, and then the President must submit those numbers within seven days after the first day of the Congress.
The problem, Justice Kennedy, was that we didn't have those numbers at that time.
The Census Bureau had not released again state by state imputation numbers until well into March of 2001, and it was within a few days after that that we brought this suit.
Mr. Dellinger: Well, if you prevail, what's supposed to happen?
Mr. Lee: The remedy that we are asking for, Justice Kennedy, is simply a declaration and an injunction to the Secretary of Commerce directing the Secretary to issue a new apportionment count free of the statistical estimates that we believe are unlawful under the Census Act and under the Census Clause of the Constitution.
Mr. Dellinger: Well, what's supposed to happen under the Act?
I mean, I went back and read the Act last night.
And that's geared to certain automatic apportionment, and that in turn is geared to the delivery of the numbers by the date specified.
So let's assume the Secretary and then the President comes with different numbers.
The Act doesn't have any provision in effect for self-adjustment, does it?
Mr. Lee: And I think, Justice Souter, that's effectively North Carolina's argument.
That there are time deadlines set forth in the Census Act that somehow stand in the way of reviewability here.
That is an argument that was raised and rejected in Franklin.
The Franklin opinions, Justice O'Connor's opinion for a four justice plurality on that point, says that if this Court agrees with the Plaintiffs in this case, we may presume that the President of the United States will conform to our decision and will in fact issue a new apportionment count that is consistent with the law as we clarify it.
By the same token--
Mr. Dellinger: I will presume that, too, but that was four justices, and there wasn't a specific discussion of what is supposed to happen under the statute, even assuming the President sort of takes counsel from our opinion.
Mr. Lee: --And the other opinion in that case, Justice Souter, that also addresses this issue was Justice Stevens' concurring opinion, for a separate four-justice concurrence on this issue, and his analysis was in fact that the statute would require the President to conform to a revised count that the Secretary might present in conformity to a decision of this Court.
Mr. Dellinger: Well, that's what he said and as you know, I joined his opinion.
But that didn't have a majority, either.
So isn't it incumbent upon you to tell us whether in fact the position that some of us took that first time is in fact correct?
Why will they, I mean, it seemed so to me at the time but it's a real issue in this case, as to whether the statute which is geared to delivery of information at certain times can in effect reverse itself.
Mr. Lee: The other point that Justice Stevens made in his concurring opinion, Justice Souter, which I think is critical on this issue, is that the time lines under the statute can be understood only to trigger, only with regard to the issue of ripeness.
In other words, it makes no sense to think that a time deadline for executive action under a statute means that you, you must... it makes no sense to think that those deadlines are a bar to judicial review.
In fact, what we usually think of when we look at a statute that has deadlines is that once those deadlines have passed, then executive action is at an end, there is a ripe claim that may be brought and then it's appropriate for the federal courts to step in.
Mr. Dellinger: But those are not cases where after, after the deadline provisions, you have a, a basic change in, in the court's role, that after the deadline in this case, you are going to have to, the court is going to have to issue an opinion that purports to bind the President.
Which is, you know, a significant step.
Mr. Lee: Yes.
I think that the opinion does not need to purport to bind the President, and as the Court indicated in Franklin, there is no need for the order to go to that point.
Mr. Dellinger: Well, but I mean that's, that's even worse.
You want us to decide something on the ground that some good may come out of it.
But I just don't know what your closest case is for the proposition that this satisfies the redressability requirement.
Let's assume that the President, were interested in implementing our orders.
We send something to Congress, and maybe that's the end of it there.
Can we force the Congress to act?
Mr. Lee: Well, the, the order would go to the Secretary.
The Secretary would submit a new apportionment count to the President that would be free of the estimations and the question then is what would the President do.
And the President under two alternative, but not necessarily inconsistent theories, of eight justices in the Franklin case would certainly conform and would present those same numbers, that same revised apportionment count.
Mr. Dellinger: Mr. Lee--
--Let me just... what's your closest case to show that a declaration of this kind satisfies the requirement of redressability?
Mr. Lee: Well, I believe it's Franklin, Justice Kennedy.
Mr. Dellinger: Excuse me.
Other than Franklin.
Because Franklin was a fractured opinion.
Mr. Lee: Well, one understanding consistent with Justice O'Connor's opinion in Franklin would be that of a declaratory judgment.
It does not in and of itself bind the parties.
It rather declares the law.
It is not a, it is not a, a binding order in the sense of ordering parties to do anything.
Mr. Dellinger: Well, it does bind the parties in a certain extent.
It's just, there is no action required as a result of it.
There is no mandate saying you must do this.
Mr. Lee: That's right, Mr. Chief Justice, and I think that that's consistent again with Justice O'Connor's plurality opinion in Franklin.
Mr. Dellinger: We have never held that a declaratory judgment will issue where the Court has no power to compel the action that the declaratory judgment describes.
I mean, you cannot get jurisdiction by saying well, you know, we have no power to, over this person to compel any action, but we are just going to declare what the law is because this person, you know, might follow it.
I don't think that's a basis for jurisdiction.
Mr. Lee: Well, Justice Scalia, but I think there are two alternative theories.
One is Justice O'Connor's approach and the other is Justice Stevens' approach in Franklin, and Justice Stevens' approach in Franklin in fact, is in fact that the President is required to conform with revised numbers.
But the Secretary might represent--
Mr. Dellinger: But you are actually taking a position also that whether or not the President would respond, at least the Secretary would have a duty to do something in response to our order.
Mr. Lee: --That's exactly right.
Mr. Dellinger: Is it quite common that in suits the President is not named, and the Secretary is.
But I wanted to go back to a question... the President has acted, but Congress has not.
Am I correct?
Congress is waiting on what this court does?
Mr. Lee: The final step, Justice Ginsburg, in this process, is for the clerk of the House of Representatives to forward to the states a certificate that shows how many seats in the House of Representatives they are entitled to after clerk receives the revised numbers from the President.
And this Court in Franklin indicated clearly that that duty is a ministerial duty and that we can absolutely expect that the clerk of the House will follow suit.
So the only question is with regard to the President, and either under Justice O'Connor's approach or under Justice Stevens' approach in the Franklin opinions, we can and should expect that the President will in fact conform to this Court's order.
Mr. Dellinger: You only have a certain amount of time, Mr. Lee.
I want to ask you a question about the merits of the case.
Are you saying in effect that there has to be a specific, something more than just a little bit of educated guesswork with respect to every house in the country?
Mr. Lee: Well, the actual enumeration clause of the Constitution would require an actual count, and--
Mr. Dellinger: Okay.
But what, supposing that the census people have, they see a house, and they see a car in the driveway, but they can never find anybody home, and then someone sees lights on late at night.
They do their best.
Do they have to say nobody lives there?
Mr. Lee: --Yes, they do, Mr. Chief Justice, and one of the reasons is that the mere fact that there are lights on and cars there doesn't tell us that this is anyone's usual place of residence.
It may be a seasonal vacation home.
Mr. Dellinger: It doesn't confirm it, but isn't it a permissible inference for the census takers?
Mr. Lee: Well, at some point a census taker may make a conclusion, if it's based on information specific to an individual household, may make a conclusion that a household is in fact occupied.
Our position is that what the Census Act requires, Section 195, is... what the Census Act prohibits, rather, is the use of information specific not to that individual household but specific to an entirely different household for the purpose of drawing an inference.
Mr. Dellinger: Okay.
Would you, on that theory, would you agree that although they may not impute a number beyond one, it would be fair for them at least to count it as one?
Because they start out as I understand it with a, in effect a report from the post office saying, you know, this is a house at which somebody lives.
So isn't that kind of hearsay just as good as, let's say, the next door neighbor's hearsay at least to the extent of saying somebody lives there and they can count one person?
Mr. Lee: Justice Souter, they don't start out with the proposition that this is a house and somebody lives there.
They start out with the proposition that this is an address.
Mr. Dellinger: Well, doesn't an address usually mean somebody lives there, at least if we can establish that it's a residential neighborhood?
Mr. Lee: Not necessarily, for a number of reasons.
Number one, you have got the seasonal home problem.
It may be no one's personal, no one's usual place of residence.
Number two, you have got the problem that some things that look like homes and that are in residential areas actually turn out to be businesses.
Mr. Dellinger: What percentage of homes in the United States when they take the census turn out to be vacation or second homes?
Mr. Lee: I don't have that number handy off the top of my head, Justice O'Connor.
I believe that it is in the record, but I'm not aware of that.
The total number of imputations was with regard to 620,000 unrelated--
Mr. Dellinger: But the imputation techniques apparently treat all the dwellings that are identified as residences to be counted?
They don't discount for a percentage that are vacation homes?
Mr. Lee: --No.
They do not.
And that's part of the problem.
Part of the problem here is that imputation assigns occupancy not only to units, Justice Souter, that are known to be houses.
Mr. Dellinger: That's saying that they are doing it bad.
We have to assume they are doing it good.
I mean, suppose they have learned with experience that an old age home that has two bedrooms in each apartment and that has a sign out in front that says "full", has two people.
One in each bedroom.
And now they learn something about this house.
It's an old age home.
It has two bedrooms.
And the sign in front says full.
Now, can they make the assumption that this too has two people in it?
Mr. Lee: I think not.
Mr. Dellinger: Not.
Now, why not?
Can they make an assumption that if the pizza man delivers pizza to the place and people eat the pizza, i.e., at least it disappears, that there is someone in the house?
Can they make that assumption?
Mr. Lee: They cannot, Justice Breyer.
Mr. Dellinger: They cannot make the assumption about the pizza man.
They deliver it to the door and the food disappears and the lights are on and off.
They can't make that assumption?
Mr. Lee: No.
People in businesses eat pizza and people who are living, staying temporarily in a vacation home eat pizza.
Mr. Dellinger: And by the way, there is a sign in front that says this is not a vacation home.
Mr. Lee: Then I suppose... I suppose under those circumstances, then we'd have information.
Mr. Dellinger: Fine.
Once you are down, once you are down that road, then you are then... then I can easily construct examples where the imputation is absolutely as strong.
I mean, I just did that with my old age home.
So you are not against imputations, you are against weak imputations.
Mr. Lee: --No.
I'm against sampling because that's what the statute prohibits.
Mr. Dellinger: You have to take that position.
Mr. Lee: I think it's important, though, to answer your question, Justice Breyer, to go back to the baseline of the House of Representatives decision.
It holds that a method proposed by the Census Bureau in 1997 is a method of sampling.
We know that much, and so there is really a narrow statutory question presented here and it is, is this particular method meaningfully different?
And the answer is that it is not.
It is not under the statistical understanding of the term sampling.
It also is not because the definition that the appellees are proposing here is a definition that would nullify this Court's decision in the House of Representatives because the two methods simply are not distinguishable.
Let me talk about the appelees' definitions and explain why that's true.
The definitions that they propose say that sampling is limited to methods that are employed as a last resort, only after an initial effort to enumerate the whole population.
Under that definition, in fact, both, neither one of the methods that's at issue here would follow sampling, and so that effectively is an argument that the Court got it wrong in House of Representatives and that that case ought to effectively be nullified.
And here's why.
Mr. Dellinger: Are you suggesting, Mr. Lee, that the Census Bureau goes about this thing in a way, say, supposing you have a lake in northern Wisconsin where the temperatures get down to about 40 below in the winter.
So you know it's not a year-round, you see a bunch of houses around the lake.
The census people come along and do you think if they could find somebody there, they would just automatically say yes, that person counts?
Without inquiring as to whether it's a vacation home or not?
Mr. Lee: Well, if they know that there is somebody who is living there, they have to find out whether the somebody who is living there lives there as his usual place of residence, or that person doesn't count in the apportionment count.
But what's going on--
Mr. Dellinger: But under the imputation method as I understand it, the assumption is made that if a house is there, that it is a residence, and a second assumption is made that the number of occupants will be that in a general geographic area of other homes.
Is that right?
Mr. Lee: --That's right.
And that's exactly the sense in which this particular method of imputation is a method of sampling.
The Census Bureau took a sample of 620,000 addresses.
They called them donor housing units in the 2000 census.
It used that sample of 620,000 addresses or housing units as a representative sample of the 620,000 addresses that they never were able to count on the theory that they were representative, knowing that next door neighbors are not always the same, but that generally across the board, if you take the 620,000 sample or donor units and use them to estimate the 620,000 addresses about which we know nothing, generally across the board it will average out and it will be representative.
And that's sampling.
Sampling is taking information about the parts to make an inference about the whole.
Mr. Dellinger: I'm curious both in your argument here this morning and in the brief, I think both parties seem to assume that the key question is whether or not this is sampling.
Do you take the position that other statistical methods of adjustment are permitted just so long as they are not called sampling?
Mr. Lee: Our position is, Justice Kennedy, that any method that takes information from a part of the population to make an inference about the whole is a method of sampling and is prohibited.
Now, there are methods, statistical methods that the Bureau uses that don't amount to sampling that are permitted.
Let me give you one example.
Mr. Dellinger: Go ahead.
Mr. Lee: It's called the method of quality assurance, and the quality assurance phase, what the Bureau does is to randomly send out enumerators to given census tracts to try to find the tracts where there are a large number of errors, where the enumerators may have made a lot of mistakes.
But the response to that is to send out an additional army of enumerators to those problematic tracts to sort out the problem to get the new information.
Now, that's an actual count.
That's not sampling but it's a statistical method.
And so that's an example.
Mr. Dellinger: What about a house where they have not been able to find anybody to talk to, no returns sent back, so they go to the people next door and they say is there anybody living in that house and the person next door says yes, there are two people.
Now, is that permissible under your theory?
Mr. Lee: That is permissible under our theory.
That's the use of proxy data that the Bureau currently engages in and Mr. Chief Justice, that would be permissible.
Mr. Dellinger: But the Act refers both to sampling and to other methods of statistical adjustment.
It seems to me sampling is a way to gather the data, and what we are talking about here are statistical assumptions made to evaluate the data, and I'm not sure why you have to concede that you lose if we say this isn't sampling.
Mr. Lee: Well, there may be another way to get there, Justice Kennedy, but my understanding under the statute is that Section 195 prohibits the statistical method known as sampling.
That may suggest that there are other statistical methods that are not, that don't amount to sampling, and the quality assurance example that I gave I think is one of those examples that would be permitted under the statute.
But all methods that take information from a part, from these 620,000 addresses that are occupied, to make an inference about the rest of the population, the 620,000 addresses that we don't know anything about, those are methods of sampling and again you really cannot distinguish this method in any meaningful way from the method that was struck down in 1997.
The principal difference is that instead of 620,000 addresses about which we know nothing, under the 1997 plan, the Bureau anticipated that there may be many more.
Mr. Dellinger: Yes.
But there is one significant difference and that is in the technique that was struck down before, a number of inferences are being drawn from the so-called sample.
In this case, only one inference is being drawn from a sample of one, and that is at least a tremendous difference in degree, and I suppose it's a difference in degree that would be likely to have an effect on accuracy.
Mr. Lee: In fact, Justice Souter, that is not a distinction.
The fact that there is an imputation being made one individual housing unit at a time, again, does not distinguish this method from the method struck down in the House of Representatives case.
Mr. Dellinger: Well, it wouldn't in a literal sense if you were simply making one imputation at a time, but you were making a long series of imputations from the one sample, in this case, the one house.
But this is simply one imputation made one time, and it involves not so much a, a principle of selection, or let's put it this way.
You can characterize it not only as based on a principle of selection, but you can, you can characterize it on, on a rule of probability, like birds of a feather flock together.
People who live next door tend to be much alike.
So it seems to me that there is a qualitative difference and a quantitative difference.
Mr. Lee: But Justice Souter, once you pool all of the individual housing units together, the 627... 620,000, rather, of these housing units that were used to estimate the 620,000 addresses that we know nothing about after six follow-up visits, that's your sample.
That's your group.
Mr. Dellinger: No.
Your sample isn't 620 to tell you about just any random 620.
It's, it's a selection of one to tell you about one more which is right next to it.
So it seems to me that it sort of masks the issue to talk about 620 and 620 as opposed to one and one.
Mr. Lee: Mathematically it seems to me, though, Justice Souter, we get to the same end.
This is an intermediate step, even though it happens on a household by household basis.
It's an intermediate step for the clear purpose of making an ultimate inference about the population as a whole.
If, for example, there are three--
Mr. Dellinger: Yes, but that, I mean that would be true about any inferential conclusion.
I mean, strictly it would be true I suppose even if accepting the hearsay conclusion is true.
And there is an awful lot of inference that does not fall within anyone's notion of what is a sample or what is in a technical sense a statistical method.
Mr. Lee: --It is also true, however, Justice Souter, of the method struck down by this Court in the House of Representatives case.
And I probably didn't make this clear enough the first time I mentioned it so let me just make it one more quick time.
Under the 1997 plan, it's clear that the Census Bureau proposed to estimate the roughly 10 percent of the population that it was not going to enumerate one individual housing unit at a time.
As is indicated in the administrative record in this case at page 1647, it was precisely the nearest neighbor assumption that the Bureau had in mind under the 1997 plan.
So that truly is not a distinction between this particular method and the method at issue in the House of Representatives case, and thus the argument that's being proposed here really is an argument that would effect an end run around the House of Representatives decision.
Mr. Dellinger: Mr. Lee, as I understand it, the imputation method that was actually used after the 2000 census has the effect of counting non-households as households in some instances?
Mr. Lee: That's correct, Justice O'Connor.
Mr. Dellinger: And it counts one household several times in some instances.
Mr. Lee: That's correct, because some of these are different--
Mr. Dellinger: And, all right.
And the imputation is not based on the nearest neighbor.
I mean, there is a big block of homes, and the assumption about who lives on it is not based on a nearest neighbor.
Mr. Lee: --Ordinarily, as I understand it, it is the nearest neighbor address that's used to estimate or impute the address about which we know nothing.
Mr. Dellinger: That's not my understanding, but we can explore that with the other side.
But I was concerned because it seems to be a method that does amount to what we said couldn't be done, in House of Representatives, on a much smaller scale.
Mr. Lee: That's exactly right and of course the scale of the practice is beside the point.
Mr. Dellinger: Well, they say it isn't beside the point.
I take it that their argument is basically in House of Representatives, what the census was trying to do was to use a special kind of inferential method to determine the population of an entire area.
Well, here they are using an inferential method to determine the population of an individual house.
Now, if you don't make that distinction, and you say that distinction is irrelevant, we are left with your distinction which seems as Justice Souter just pointed out to be the same as all inference.
In other words, I infer that when you deliver pizza and it disappears, someone is in the house, because of the set of similar to pizza type cases that I have seen in the past.
Or, you know, that's I think what Justice Souter was bringing out, they have a different distinction.
Now how do I deal with the problem?
Mr. Lee: Well, Justice Breyer, let me mention two reasons why that distinction doesn't get them anywhere.
It doesn't get them anywhere number one because it simply isn't borne out in the statistical understanding of methods of sampling, in two senses.
Number one, the statistical understanding of sampling says generally, what statisticians understand to be encompassed within the general category of sampling is taking information from a part to make a statistical definition.
Mr. Dellinger: You have the statistical definition, but my question was directed at a particular problem that, if I take your approach I can't get myself out of.
That is, that your definition applies to all inference.
Their definition distinguishes among kinds of inference and they have their statistical support.
Mr. Lee: The problem with the distinction is, another problem with the distinction is that it's clear that it would permit the Census Bureau to replicate the 1997 plan.
If you buy that distinction, the Bureau can get back exactly to where it wanted to go in 1997 merely by scaling back its nonresponse follow-up efforts.
Mr. Dellinger: I don't understand the pizza man.
Does the pizza man, does that inference consist of imputing something from the part to the whole?
I mean, I would think that's your answer to the pizza man example.
It is an inference.
It goes from the whole to the parts.
But it's not the same kind of inference that is done by what you say is sampling, mainly, imputing information that belongs to a part to the whole.
That's not what's going on in the pizza man case.
Mr. Lee: Maybe, Justice Scalia and Justice Breyer, maybe I missed the pizza man example, but that may well, that may well be a distinction.
Mr. Dellinger: I didn't think the imputation involved any actual things like looking at, when pizza was delivered.
It involves looking at what houses returned, take a big block in the district they are serving.
It involves looking at the houses that returned the form, looking at the houses where they reach people in a follow-up visit, and there are some they may might still be missing and so they impute the map.
They don't go see if there is a car in the garage or look at the pizza delivery.
They arbitrarily say we are going to impute from the data we have that there are X number of additional houses and that X number of houses are occupied at a certain level of occupancy.
That's what's going on.
Mr. Lee: That's exactly right.
That's precisely right, Justice O'Connor.
The pizza man knows how many people live there.
He is a proxy and he can give that information to a census enumerator.
But the end run problem is illustrated by this distinction; if the Census Bureau can use any methods that fall within the definition that's now being proposed, the Bureau can simply scale back slightly its nonresponse follow-up effort and estimate increasing percentages of the population, up to and even exceeding the 10 percent that was proposed in 1997.
Mr. Dellinger: Mr. Lee, didn't it prove out that, after, that in the majority of these cases, something like 75 percent, that they know for sure that these were houses where people lived?
Mr. Lee: Actually not, Justice Ginsburg.
The 75 percent figure is misleading and completely unhelpful.
At page 445 of the joint appendix the memorandum that's at issue there simply says that in 75 percent of the status imputed cases, status imputation was done with regard to enumerator returns or respondent returns.
In other words, these were not mail-back returns.
Mr. Dellinger: But you don't have a negative showing, either, that it wasn't accurate?
And there was something Justice O'Connor said, and I'm not sure whether I fully understand it.
I thought that this imputation was made for people who didn't respond to the mailing, that the like comparison was with the group of people that were eventually counted, but that they were nonrespondents the first time around.
Mr. Lee: That's right.
I think if understand your question, Justice Ginsburg, I think that's right.
They took information from initial nonresponding households, used those as a sample of those who never responded, and some of those may well not have been homes, Justice O'Connor, some of those may well have been duplicates, etcetera.
If there are no further questions, Mr. Chief Justice, I'd like to reserve the remainder of my time for rebuttal.
Mr. Dellinger: Very well, Mr. Lee.
Mr. Dellinger, we'll hear from you.
ORAL ARGUMENT OF WALTER E. DELLINGER ON BEHALF OF THE NORTH CAROLINA APPELLEES.
Mr. Dellinger: May it please the Court, because North Carolina believes very strongly that the right time to challenge the Census Bureau's planned use of a statistical method is before rather than after the census is completed, let me begin by--
Mr. Dellinger: Well, how could the challenge have been brought before Utah knew that the imputation figures caused this problem?
Mr. Dellinger: --Under your decision in House of Representatives, I think on page 332, you note that citizens of Utah, you didn't specify the state, but citizens who anticipate that the use of a method like imputation, as suburbanites might, would dilute their in intrastate districting representation, are specifically said in House of Representatives to have authority to bring that suit.
So I think that there would have been parties able to litigate it.
Mr. Dellinger: Well, you can't bring a suit before you know about what's going on.
Mr. Dellinger: --Yes, you can, Mr. Chief Justice.
It is true, as you noted, that there might not be a special incentive on Utah's part, but they don't know they are going to be the loser.
They might have come out better or worse with or without imputation.
But every governor, attorney general, states and cities are carefully watching this.
Many as you know sued in 1997.
So the suit could have been brought that, by any resident of a suburb could have had an expert allege that they are going to be diluted if you impute households because it's rural areas where the files are often damaged or--
Mr. Dellinger: But Mr. Dellinger, a state can't, you suggest there are suburban people, city people, a state can't, until it knows what the result is going to be, and Congress used the words any aggrieved person.
The state, you are saying is the state can never bring such a suit, because it is not aggrieved before.
It can't predict that it's going to be aggrieved.
It will know only when the returns are in, so in effect, you are saying this is not a question that the state has come too late.
It's not sooner or later, it's a never question for the state.
You're saying Congress simply did not authorize the states to bring this kind of case.
Mr. Dellinger: --Justice Ginsburg, that is correct, if a state is unable to show, does not have the experts about a forthcoming census that is unable to demonstrate that it will be aggrieved.
How much we should worry about a nonaggrieved party not being able to obtain judicial review is--
Mr. Dellinger: They are aggrieved at the end of the line.
Mr. Dellinger: --They are now, but my point is that if Utah had been watching this carefully, and as you know, many states were involved in the '97 litigation as amici and otherwise.
Surely the Governor could have found many citizens of the State of Utah who, and this suit itself includes individual voters, some of whom may very well, the plaintiffs--
Mr. Dellinger: The governor had no incentive to do so.
Mr. Dellinger: --I understand that.
Mr. Dellinger: He could have gotten a stalking horse to, some suburbanite to bring the suit but he had no notion that there was any reason to do that.
Mr. Dellinger: It seems to me that that is exactly correct, but that it's exactly why any governor, major city mayor or others, these citizens, those who brought suit in 1997 and did not include this, suburbanites are going to hurt, would want to sue, could bring the suit, but in any event, they could also participate in the Bureau's process.
Utah never made this objection.
Now, here's why that's so unfair.
It's simply this.
The, it is, we don't know who would have gained this seat if Utah had bought its objections to the Census Bureau before the census was conducted.
Mr. Dellinger: Is there a statutory time period for people to challenge the Census Bureau proposed techniques before the census is taken?
Is there some provision whereby that challenge can be made?
Mr. Dellinger: Yes.
The public law which was the basis for the suit in 1997 is a permanent law, and that suit allows aggrieved parties to bring suit when the census produces its plan.
So it's certainly at least--
Mr. Dellinger: But is it a fact that somebody is aggrieved when you are at the stage of the Census Bureau just saying this is what I plan to do?
Mr. Dellinger: --Yes.
The law involved, the public law, that was a basis of the '97 suit makes it clear that if you could say if they do that plan, that's exactly how it works in House of Representatives, if you do that plan, we expect that our district will be diluted.
Mr. Dellinger: Well, does the law provide for some sort of administrative hearing before the Census Bureau, some sort of an exhaustion requirement?
I thought that's what you were suggesting.
If Utah had just brought this to the attention of the Census Bureau.
Is there some structure for doing that?
Mr. Dellinger: Yes.
The Census Bureau does have a structure for doing this.
I'm not saying that they are foreclosed because they didn't participate in the administrative process, but the Bureau was open to hear these objections and the courts were open to hear them.
Mr. Dellinger: But, you say now the Bureau, is that well publicized?
Is there some person you can go to in the Bureau and say I don't like what you are doing?
Mr. Dellinger: Yes.
And that is the exact process that is followed with every... that's why they publish the plan for the census.
Mr. Dellinger: It seems unfair to Utah, though, to say that they are supposed to bring a suit before they know they have been hurt and why.
How... what are they supposed to do?
Of course they wouldn't come in ahead of time.
For all they know, the system would benefit them.
Mr. Dellinger: --Justice Breyer, those who sued in the House of Representatives litigation had a very thin basis for knowing that they would be adversely affected.
Indiana maybe wrong in thinking that--
Mr. Dellinger: Well, of course, sometimes you could guess in advance, but a lot of times you couldn't, and it's important that there be a fair method that treats states fairly.
So why, why would it be fair, any way, to cut off those states that don't know they have been treated unfairly and hurt, until they find out later?
Mr. Dellinger: --And here's why.
Here's why it's not unfair.
Because the unfairness is so great for the disruption on the other side.
If it had been determined by the Bureau itself or through litigation that imputation could not be used in 2000, the Census Bureau absolutely would have used some means other than imputation to ascertain the enumeration of those established residential addresses whose records were damaged, missing or incomplete.
There are in North Carolina approximately 16,000 households that are established residential addresses on the carefully pruned master address list.
Those houses often were visited by an enumerator in the, often the houses were added where it was status imputation by a field enumerator during the enumeration process.
But the new form, because it wasn't on the master address list before being added, never caught up with the master address list.
So you have no residential addresses, and, and the Bureau absolutely would have had another method.
I mean, even now theoretically they could go back and start again and say what are we going to do about 620,000 established residential addresses for which we don't have input numbers, because at the end of this massive process, now, this is a process that involves 500,000 enumerators, 120 million households, one and a half billion pieces of paper, imputation occurs at the end of that process when all the records are centralized.
It began in 1960, five censuses ago because that's when computers were able to process these cards.
Damaged cards, now forms, can't be read.
The data is discrepant or missing.
It doesn't mean that these are households that were visited six times.
Mr. Dellinger: Mr. Dellinger, you have recognized that your argument is, as stated, that the position of Utah that can't project what the returns will be, does not have, does not qualify as an aggrieved person at the only moment in time when it can say that it's aggrieved.
I think you had another justiciability issue, did you not?
Because time is running and I think we grasp your position that it's too bad they are not aggrieved because they have to come in in the beginning and not at the end.
Mr. Dellinger: Yes.
Justice Ginsburg, we are not suggesting that these issues are immune from judicial review.
I believe someone could challenge imputation and would.
Mr. Dellinger: Are you sure they were not aggrieved at the beginning?
I'm surprised that you... I mean, is it not an aggrievement for the State of Utah or for any state that its districts are distorted?
Even if it doesn't lose, you know, congressmen to another state, isn't the distortation of the districts within that state a grievance of that state?
Mr. Dellinger: That is a very helpful answer, and that is not... a very helpful suggestion.
And that has not been by any means ruled out, nor has it been ruled out in my view that a state could say we believe we are entitled to have a fair process determine our representation.
This process isn't fair.
We don't know how it's going to come out.
But this process is loony.
Now, you would also see there's just a matter on the merits.
Congress has said in Title II, Section 2 that the President's determination shall be final and the states are entitled to that for the next period of time until the next apportionment unless Congress itself acts, which they specifically provide for.
That's surely constitutional because the Constitution itself in providing that the census need be done only every 10 years puts a great stake in permanency.
Now, turning to the merits, it's striking the extent to which the issues in this case are anticipated by the Court's decision in the Wisconsin case.
At page 22 in Wisconsin--
Mr. Dellinger: Before you go on to the merits, you have nothing else to say then about justiciability?
I mean, let's suppose they can file a suit.
What's the remedy going to be at the end of the day at this stage?
Mr. Dellinger: --Oh, at this stage?
Mr. Dellinger: Yes.
Mr. Dellinger: Justice O'Connor, that is a very good question.
Utah seems to assume that if the case went back to the District Court, overturning the District Court's ruling that imputation is satisfactory on the statute and the Constitution, that you would simply take out those occupancy figures for 620,000 households nationwide.
It seems to me that the, that what ought to be done is you return that to the Bureau and say now there may be time to match up the missing forms that were added late in the enumeration process.
There may be other ways to recover that data.
Ways that certainly would have been done if it were brought before the census might still be brought now.
So that we don't, it seems utterly unfair to treat all of those as zero when some of them are houses that, an enumerator comes to 212 Elm Street, it's not on his list, he goes in, interviews Ozzie, Harriet, the kids, sends in the form and it was corrupted or it didn't get matched up at the time, faced with that time deadline.
It may now be available.
We don't know who would prevail on that.
On the statutory issue I think if you look at page 22, the Court itself creates statistical adjustments as done in '70 and '80, and as they were done here--
Mr. Dellinger: Page 22 of what?
Mr. Dellinger: --I'm sorry, Justice Rehnquist.
I referred earlier to the decision of Wisconsin vs. the City of New York, the unanimous opinion.
On the statutory issue, the Court says that the statistical adjustments in 1970 and 1980 which were imputation, hot-deck imputation, were an entirely different type than the adjustment considered here, and they took place on a dramatically smaller scale.
The Court also treats actual enumeration in the Wisconsin case and speaks of actual enumeration at page 6 in the Wisconsin case as having been something that the Bureau has never, or the country has never actually achieved actual enumeration.
It's never been wholly successful.
Treating it clearly as the end result of the process, the right number and not as a method.
Now, let me go right to one point.
We do not believe that there are no constitutional limits on how Congress can conduct the census.
Wisconsin says that there is virtually unlimited deference to Congress, but they also set a standard that the congressional goal must be related to representation according to the respective numbers.
The way I read that is this.
A proposed census as designed is not reasonably calculated to produce distributive accuracy among the states is constitutionally suspect, because it will not produce an apportionment according to the respective numbers.
And that would be a fatal flaw.
Here, every imputed occupant and household is to an established residential address with a precise geographical location.
It was an effort to enumerate every household in the country by using the best information available on known household addresses.
This is historically consistent with what we did when the neighbors are asked for their opinion, the postal worker is asked for his opinion.
You use proxy information.
Here you take--
Mr. Dellinger: Mr. Dellinger, this isn't really a use of proxy information.
This isn't going to a neighbor who lives next door, this is using a statistical method to make assumptions, isn't it?
Mr. Dellinger: --Yes, it is.
Mr. Dellinger: I think you have to be realistic about that.
Mr. Dellinger: Yes, it is a statistical method by which you take, when you have a known established household address, the information from that, basically the next door neighbor, somewhat refined, Justice O'Connor, the next door neighbor that was a nonresponding household of a similar type and they found over the years that that information is more reliable than zero.
Utah's position is that the Constitution requires when you have an address known to be occupied, if you put down that it's occupied by zero people, that's not an actual enumeration, that is literally a counterfactual enumeration from what one knows to be the case.
This is a process where you use one unit's characteristics to supply to another.
Mr. Dellinger: That's just part of it.
It also, it also imputes nonhouseholds as households.
I mean, it does a lot of things.
Mr. Dellinger: It imputes only to known addresses.
Mr. Dellinger: Thank you, Mr. Dellinger.
General Olson, we'll hear from you.
ORAL ARGUMENT OF THEODORE B. OLSON SOLICITOR GENERAL, DEPARTMENT OF JUSTICE ON BEHALF OF THE FEDERAL APPELLEES
Mr. Olson: Mr. Chief Justice and may it please the Court: The Census Bureau has consistently utilized the imputation technique for drawing inferences about a tiny fraction of damaged, discrepant or missing population data for the past five censuses.
Mr. Dellinger: Do you agree that that's a statistical methodology or is it just a method of making deductions from circumstantial evidence?
Mr. Olson: We agree that it is a statistical methodology, Justice Kennedy, and it's very important in that context to focus on the words of the statute.
The words of the statute are that the statistical method known as sampling is the one that's prohibited with respect to the apportionment.
That statement in the statute suggests that other statistical methods are not prohibited by the statuted, and that one particular statistical method, that is the one that is known as sampling, which is in quotes in the statute, which imports that it's a term of art that was known by Congress to be a term of art.
It was a phrase that was suggested by the Secretary himself in 1957, when that statute was added, that exact phrase came from the Secretary.
The Secretary was presumed to know what that phrase meant.
It is a term of art that statisticians know what it is.
Mr. Dellinger: Well, I thought the statute also said that statistical adjustments pose a risk.
Mr. Olson: --It does not say that.
In fact, the import of the statute, Justice Kennedy, is that statistical adjustments, and I will refer to what my colleague Mr. Dellinger just referred to, at page 22 in the Wisconsin decision which was unanimous decision by this Court just six years ago, distinguishing the sampling method that the Court was talking about in that case from statistical adjustments known as imputation, which is described on pages 4 and 5 of the Respondent's brief, referred to on those pages of the Supreme Court, this Court's unanimous decision just six years ago as being an entirely different thing.
We would go back also to the fact that this Court unanimously held in that case that the Constitution vests virtually unlimited discretion in Congress respecting the manner in which the census shall be conducted, and that Congress has delegated its broad authority to the Secretary to take the census in such form and content as he shall direct.
Now, the Congress has exercised that discretion by passing it on to the Secretary with the one limitation with respect, with respect to one statistical method known as sampling.
Now, at the same time, the Secretary was proposing the prohibition of the sampling method; with respect to the apportionment of the census in 1957, the Secretary was planning the 1960 census, which was the first computerized census and the first time that the imputation method which we are talking about today was used.
So it's obvious that the Secretary did not believe that hot-deck imputation was sampling, because the very next census three years later imputation was being used in that census.
And it was used again in 1970, 1980, 1990, and 2000.
And with respect to 1970, and 1980, it was actually considered by this Court in connection with the Wisconsin case.
It is widely understood in the industry that sampling is a collection of, a collection technique whereby a sample, a fraction of the whole population is used to deduct... deduce the actual whole population.
It seemed to me that a metaphor that might be considered outside the context of population gathering would be if the Court asked the library of this Court to ascertain the number of books, to conduct the number of books in the Court's library, and the, but sampling was not permitted to do that, so that the librarian could not go to every third shelf, multiply, count the books, multiply by three and get the census.
But if the librarian went to those shelves and counted every particular volume and found that there was a space here on that shelf, a space this big on the next shelf, and a space this big on another shelf, for example, the imputation would be saying well, all the books or the books right next to this are this size, and therefore that space a book is missing, so we know we have a book, and we will impute one book to that space or two books to this space.
So that the sampling technique is completely discrete from the imputation technique.
We know that because the Secretary has always regarded it that way.
The Congress has given the authority to this expert agency which has been conducting the census for years and years and has drawn various different types of inferences which is what imputations are.
Mr. Dellinger: Why, just out of curiosity.
I mean, I'd like to understand this better.
In the library, you look and see that everything around the book is a history book and so then you impute the characteristic of being a history book to the one that's missing.
That's your analogy of what goes on here, is that right?
Mr. Olson: Well, yes, I think that's an extension of it.
In the example that I was giving you look at the space and impute the size of the space by the book that is immediately next to it.
Mr. Dellinger: Okay.
Why is it called statistical?
Why isn't that just ordinary inference?
Why... you said we do it statistically?
What's statistical about it?
Mr. Olson: Well, I'm not sure.
That's a good question.
I don't know the answer to why it should be called statistics, because in my thinking of it, it is drawing logical inferences from the data available.
Now Justice O'Connor, it is not creating phantom homes.
Most of these cases in, the statistics indicated that with respect to one of the forms of the imputation, and this is in the record at the joint appendix, the information may be found in pages 445 through 448 of the joint appendix, that 98 percent of the household size imputation forms were enumerator forms with the status of occupied homes.
Now, the problem is that with respect to any, and the statistics are different, but in 93 percent with respect to occupancy imputation, 75 percent with respect to status imputation.
Each of these are attempting to find actual people drawing from the closest comparable unit and it's one unit for each inference.
It's not extrapolating from one unit to the whole population.
The problem with the census is that there are billions of pieces of paper as Mr. Dellinger indicated.
Some people refuse to return the forms and their known addresses.
Some people fill out the forms incorrectly.
They may say occupied but zero.
The enumerators might get bad information.
In 1850, a substantial portion of the entire State of California the returns were actually burned, and the Census Bureau in 1850 actually used a process to replace the 70,000 people that were not there.
And one of the questions indicated, neighbors have been used as proxies.
Heads of households have been used as proxies.
Postal service has been used as proxies.
These are all means by which the postal, the Census Bureau attempts to develop the most accurate count it possibly can.
Mr. Dellinger: When you say heads of households have been used as proxies.
Does that mean you go to the head of the household who appears at the door and ask him how many other people live in the house?
Mr. Olson: That's correct, Mr. Chief Justice, and that was the way it was done in 1790.
Mr. Dellinger: That doesn't seem too statistical.
Mr. Olson: Well, it's not necessarily statistical.
And I guess that, the fact is that as far as the statute is concerned, there is only one technique that's prohibited.
The technique of drawing inferences through sampling is prohibited.
It's prohibited probably because Congress feels that it might be subject to manipulation.
Mr. Dellinger: Yes.
On the merits, I think we have to know whether this so-called hot-deck imputation is a form of sampling.
Which it appears that it might well be.
I think that's the--
Mr. Olson: Well, Justice O'Connor, the expert agency to which Congress delegated this broad authority doesn't believe so.
That seems to me that there should be a substantial deference to the expertise of the agency.
Not only that, but Congress with full awareness that hot-deck imputation has been used over the past--
Mr. Dellinger: --Why do we call it hot-deck?
Mr. Olson: --Hot-deck imputation is distinguishable from cold-deck imputation in the sense that information from the most current census and the actual neighborhood, the most current available information for the actual census that's being developed is being used.
Now where that term came from, I don't know.
But that's what it means.
But this methodology has been used with the knowledge of Congress, with the full knowledge of Congress, for the past five censuses.
There was litigation over this matter, it's referred to as the Orr case in the briefs, where a seat may have been allocated to Indiana or Florida, depending upon how the imputation process came out.
So that litigation took place.
Congress actually changed the terminology, Section 141, which provides the Secretary with the authority to conduct the census in a manner that the Secretary thinks appropriate, in 1976.
After imputation had been used in two censuses already.
Mr. Dellinger: Of course, against all of that, and I'm not sure why we should give deference to the agency here.
They didn't conduct a rulemaking.
They didn't have any adjudication on this subject.
That's just what they happened to do, right?
Mr. Olson: Well, as a matter of fact, in the unanimous decision of this Court in the Wisconsin case six years ago, the court said substantial deference should be owed to the agency.
Mr. Dellinger: There's a lot of water over the dam since six years ago.
Mr. Olson: But the reasoning of the Court, I would submit, is--
Mr. Dellinger: I refer to Meade in particular.
And, but wouldn't the deference to the agency, even if there is to be some, potentially be outweighed by a constitutional doubt?
If we thought that even if this isn't sampling, it may well be not enumeration within the meaning of the constitutional requirement, and given, given that constitutional doubt, we think the wise course is to interpret the word sampling as including, including this?
Mr. Olson: --Justice Scalia, not only the Wisconsin case, but the other decisions of this Court which have considered census have suggested that the framers of the Constitution by using the word enumeration didn't mean a particular method by which the census would be conducted, nor did it wish to constrain both the Congress and whomever the Congress may delegate to--
Mr. Dellinger: Even sampling, presumably.
Mr. Olson: --Presumably.
Mr. Dellinger: Pretty accurate sampling, you know?
Mr. Olson: Possibly, Justice Scalia.
But that's nowhere close to that and this is not remotely sampling.
Mr. Dellinger: Is this remotely estimation?
Mr. Olson: This is not remotely estimation.
This is drawing an inference with respect to one particular piece of data.
We would agree that the gross estimation--
Mr. Dellinger: Why isn't it estimation?
Why isn't it estimation?
You estimate that there are so many people in this house because the house next door to it has that many people.
You don't call that estimation?
Mr. Olson: --In each of these cases, Justice Scalia, the words can be changed and added to, but the process by which the framers analyzed this in framing the Constitution, the very first House of Representatives was assigned according to an estimation, and the words actual enumeration were used to compare an actual count, an effort to find the actual number of people by indulging in a process--
Mr. Dellinger: That's why I asked about estimation.
You say there is a difference in an estimation and a deduction, I suppose.
Mr. Olson: --What I think the Constitution and courts with respect to the term actual enumeration, is an effort to go out and find a count.
We, we pointed out that the, the enumeration can mean listing by particular items as used in the Ninth Amendment, the enumerated powers.
It may be a process by which a count might be taken or it might simply refer to a census, find the population.
We submit that the Capitation Clause as we have referred to in our brief, which uses the word census and enumeration indistinguishably as synonyms, and in fact the Appellants in their brief, in their reply brief at page 15 acknowledge that the words enumeration and census are used in the Constitution interchangeably.
Justice Scalia, it didn't seem, it does not seem that the framers of the Constitution actually specified a method.
And in fact, what the Appellants were saying here today and are saying in their briefs, every census conducted by asking people who may have lived next door or drawing inferences from other pieces of information would not have been the individual by individual count that--
Mr. Dellinger: You think sampling--
--Sampling would have been okay as far as the Constitution is concerned?
Real, real sampling?
Mr. Olson: --I think--
Mr. Dellinger: We are going to do two-thirds of the state and just guess that the other third is pretty much like that.
Mr. Olson: --Well, we are not remotely close to that here.
But with respect--
Mr. Dellinger: I understand.
But your argument is that since enumeration doesn't mean anything except census, sampling would be okay.
Mr. Olson: --From the standpoint of the Constitution, I think a reasonably good argument could be made and the Government has in the past made it, that sampling if it is consistent with the process of an attempt to find an actual count utilizing sophisticated accurate and nonmanipulatable techniques--
Mr. Dellinger: Surely the term actual, though, before the word enumeration, narrows the idea that, what might otherwise be an enumeration.
Mr. Olson: --I believe it does and I think, Chief Justice, the District Court in this case distinguished actual enumeration from the conjectural apportionment that actually occurred in the Constitution itself with respect to the first House of Representatives.
What we are talking about here today, though, is an effort to, an effort to produce extremely conscientious and meticulous, to count all of the households in the United States starting with a meticulously prepared master address list.
All of those people were submitted post office forms which has been authorized by Congress for a certain--
Mr. Dellinger: That's not the issue.
Can I ask you a question about the standing issue?
I know that you in your brief do not, do not contest the standing.
You say that Franklin has decided it, because four justices thought that there was standing on one basis, and four thought there was standing on another basis.
Which of those two bases do you agree with?
Mr. Olson: --Well, we--
Mr. Dellinger: Do, do, do, is it your position that the President will have to do whatever, whatever, accept whatever revised census figures are submitted to him by, by the Secretary?
Mr. Olson: --If this Court is to, determines that the process by which the 2000 census was conducted was inconsistent with the statute or inconsistent with the Constitution and orders the Secretary to take out the imputed numbers--
Mr. Dellinger: Right.
Mr. Olson: --and deliver a different piece of information to the President--
Mr. Dellinger: Right.
Mr. Olson: --the President will transmit that certificate or that certified, those certified results to Congress for the process.
In other words, in answer to your question, the President will do what this Court assumed in those cases that the President would do.
Mr. Dellinger: He has told you that that's what he will do?
Mr. Olson: The President will--
Mr. Dellinger: You see, because, I sort of wouldn't want to make the people of North Carolina mad by taking away one of their representatives.
And were I President, I might well say, look at this judgment of the Court, doesn't run against me.
It runs against my Secretary, everybody agrees, you know, that it's not binding upon me.
It's sort of like a declaratory judgment.
And I just think too much time has passed and it would upset things too much and I don't want to take away a representative.
Mr. Olson: --The President is willing to accept the import, not only of the Franklin and the other decision that we referred to, but also if this Court decides that the process was unconstitutional or inconsistent with the statute, the President will accept that, this Court's judgment in that respect.
Mr. Dellinger: So if another, if the next President comes along, we get another case like this, and the next President tells his Solicitor General, I will not accept it, then we come out differently.
Mr. Olson: Well, if, for that reason, if Mr. Chief Justice, the Court decides that that is not the kind of result that this Court can issue, because of that possibility, we'll accept that result as well.
But we think... we think it is very unlikely for that to occur, because it is quite clear that the Constitution intended to give considerable flexibility, did not want to freeze in a system the ability of the Government.
Mr. Dellinger: Mr. Olson, you are saying we can presume the President will obey the law?
Mr. Olson: Yes, Justice Stevens.
Mr. Dellinger: What happens after that?
He transmits to the Congress, I take it the clerk of the Congress says not certified, the House has not certified the results yet?
Mr. Olson: I'm not sure of the answer to that question.
I guess the answer is that is correct.
That it has been certified and I would gather that it would have to be a revised certification if that should occur.
Mr. Dellinger: You say you assume the President will obey the law.
So you are accepting the... you are accepting the position that the President must, even if he didn't want to, that the law requires him to transmit whatever the Secretary gives him, is that right?
Mr. Olson: What we are saying--
Mr. Dellinger: Just answer that question yes or no.
Do you take the position that the President must transmit what the Secretary gives it to him, and he has no, no objection?
Mr. Olson: --I only can answer it this way, Justice Scalia.
If this Court determines that the process before was unconstitutional or in violation of the statute and the Secretary must redo it and if that information is transmitted to the Secretary, he will transmit that.
Mr. Dellinger: That is not the question I asked.
Mr. Olson: Then I misunderstood your question.
Mr. Dellinger: The question I asked is whether, you say the President will obey the law.
I take that to mean that you feel the President is bound by law to transmit whatever revised figures the Secretary takes, is that correct?
Mr. Olson: If it is based upon a decision by this Court that the Court has the power to issue--
Mr. Dellinger: To tell the Secretary.
Let's assume we have the power to tell the Secretary.
Does the President acknowledge that he is bound by law to transmit whatever figures the Secretary gives him?
Mr. Olson: --I think that that would only occur in the context of this Court's decision.
Mr. Dellinger: If not, he is not bound by law and I don't think you are going to give that away.
Mr. Olson: I don't think we need to give anything away, Justice Scalia.
We would be talking about a context in which this Court came to the conclusion it could render a jurisprudentially binding decision in a case in which there was redressability in that context.
Mr. Dellinger: Fine.
But that's a far cry by saying he is bound by law.
You seem to be accepting Marbury and Madison.
Mr. Olson: Let me just summarize because my time is about up.
This Court's words again, that the Constitution gave virtually unlimited discretion to the Congress with respect to the manner in which the census would be--
Mr. Dellinger: Extend your time by two minutes.
I'll extend Mr. Lee's time by two minutes.
Mr. Olson: --Thank you, Mr. Chief Justice.
There is no evidence that the framers of the Constitution wanted to bind themselves to a particular method of counting people.
All of the evidence suggested that what the framers wanted to do was to have a reasonably reliable accurate, reasonably accurate count of the citizens in the manner that the, that Congress would determine.
Congress in the words of this Court has delegated all of that broad authority to the Census Bureau.
The Census Bureau, conscientiously using a technique that they have been using consistently for 50 years with the awareness of Congress, the General Accounting Office, oversight committees and the actual awareness of this Court, as reflected in this 1996 decision, has demonstrated, developed a method that is reasonably accurate, uses statistical methods other than those known as sampling to get an accurate court.
We urge the Court to sustain that outcome.
Mr. Dellinger: Thank you, General Olson.
Mr. Lee, you have five minutes remaining.
REBUTTAL ARGUMENT OF THOMAS R. LEE ON BEHALF OF THE APPELLANTS.
Mr. Lee: Thank you, Mr. Chief Justice.
I have just two or three quick points.
The first is to clarify briefly the record on an important issue.
With all due respect to the Solicitor General, the record does not indicate that the majority of the imputations here were in units known to be occupied.
The Bureau's memo at page 445 of the joint appendix indicates that fully 69 percent of the units subjected to imputation were the kinds of units that Justice O'Connor's questions directed us to be concerned about, units where after as many as six visits, census enumerators were not able to determine whether the unit in question was a valid occupied housing unit and not a duplicate, not a seasonal home, not a home that happens not to be occupied.
Second, I would point the Court to another unanimous decision that this Court has handed down with regard to the census, and it's the Montana decision, and I'd like to read a brief quote from Montana and explain how important I think it is here.
This is from 503 U.S. at 465.
"To the extent that the potentially divisive and complex issues associated with apportionment can be narrowed by the adoption of both procedural and substantive rules that are consistently applied year after year, the public is well served."
That is precisely the goal of both the Census Act and the Census Clause.
To give us rules that can be consistently applied year after year, not rules that will ebb and flow with debates among statisticians, and that's where the Census Bureau is heading us here.
The variable standard that will be produced by a debate as to whether a particular method of sampling is sufficiently premeditated or follows a sufficiently premeditated intent to--
Mr. Dellinger: What about what we said in the Wisconsin vs. New York case by your opponent, Mr. Dellinger, where we referred to this very kind of action, this imputation, and indicated that that was vastly different from the broader statistical?
Mr. Lee: --I don't believe that was an issue in that case, first of all, Justice O'Connor.
Secondly, there isn't any reason to give either deference to the Secretary here or much less to congressional inaction for a very important reason.
Imputation simply has not been on anyone's radar screen.
It's been a nonentity because it's undisputed that it impacted apportionment only once prior to this case, in 1980.
In 1960, in 1970, in 1990, it had no impact.
Congress reenacted the statute, amended the statute in 1976.
At that point in time, imputation had never affected apportionment to any degree whatsoever.
There was no reason for anyone even to be focusing on it and therefore no reason to give any deference here.
I'd like to just close by saying a few brief words about the constitutional question here.
Mr. Chief Justice, to go back to your question about the word actual.
The word actual is important, and it's important because it's not just that the word enumeration is defined to mean an actual count and not an estimate.
It's that this was a term of art.
This phrase actual enumeration was used consistently in the founding era, both in colonial assessments of population when they reported their populations to boards of trade, and also in Great Britain throughout the 18th century.
James Madison himself referred to a distinction between an actual enumeration and a mere estimate.
John Adams similarly said there is a difference between an authentic enumeration and a mere estimate.
Not only that, but the framers of the Constitution thought they were giving us a permanent, fixed standard.
That's what they said they were doing and James Madison said not only is it permanent and fixed, it is the way required by the Constitution and
"which we are obliged to perform."
It was a methodology; they understood it as such; and they understood also that it had its shortcomings.
They knew that it would result in an undercount, that when you require a count, you are going to leave some people out.
George Washington himself said look, we understand at the time of the first census that the real numbers will exceed greatly the official returns.
Thomas Jefferson similarly said we know that the omissions in the census will be great.
If they understood those limitations, then why did they do it?
The answer is clear.
They understood that that was a necessary price of a permanent, fixed, precise standard that would not be subject to manipulation, that would not be subject to the vicissitudes of debates among not only politicians, but statisticians from year to year.
And the impulse, the proper impulse of this Court's unanimous decision in Montana--
Chief Justice Rehnquist: Thank you, Mr. Lee.
The case is submitted.
Argument of Speaker
Mr. Dellinger: The opinion of the Court in No. 01-714 Utah against Evans will be announced by Justice Breyer.
Argument of Justice Breyer
Mr. Breyer: When the Census Bureau counts population, it first sends out forms by mail to each address on a long list of addresses that it has.
It then contacts people who do not respond by sending employees to make personal visits, and they may make up to six such visits, but if they all fail to elicit the information, then the Bureau uses what it calls imputation, which is the subject matter of this case.
The imputation takes an address on its millions of addresses list, an address where there is inadequate information or maybe conflicting information, and it imputes the status say whether the address is a vacant lot or a house, or it imputes the number of occupants, how many people live there, and it does it by assuming that this unknown address has the same characteristics as the closest nearby similar unit.
In the year 2000 the Census imputation accounted for about 0.4% -- it is four tenths of one percent of the total count -- but because the imputation counts were spread in an uneven way across the country, that tiny difference makes a difference.
It gave North Carolina an additional seat in Congress, and if imputation had not been used, the seat would have gone to Utah.
Well, Utah found this out and it brought a suit, and in this suit it asks the Court to set aside imputation as violating both statute and the constitution.
Now, the lower court in this case, which is a three-judge Federal Court in Utah, held that the Bureau could use imputation; it was lawful, and that the seat belongs to North Carolina.
We now review that decision and we affirm.
We first have to decide, as North Carolina argues, whether Utah maybe could not even bring this suit because it would lack standing, but we find that Utah does have standing focusing on the relevant issue because in all likelihood, if Utah had won the case, which it does not, but if it had won, the Census Bureau, the President, and relevant congressional officials would in all likelihood of taken the steps that would have been necessary to provide Utah with that extra seat.
Therefore, this is a real case.
Then second, we turn to the statutory question.
The statute forbids the Bureau to use “the statistical method known as sampling”.
Is imputation that kind of statistical method?
Is it sampling?
The Solicitor General, in argument here, used an example to help explain why he thought it was not.
He said imagine a librarian who wants to find the number of books in a library, suppose the librarian uses a statistically sound way to take a sample, say count the books on every 10th shelf, then suppose the librarian uses a statistically sound method to extrapolate from the sample to the whole, say he multiplies by 10.
That is what is the SG said was the statistical method called sampling.
Now suppose instead that the librarian tries to count each book one by one, that is not sampling, and it does not become sampling just because the librarian finding a few empty spaces on some shelves, say the readers have the books out, imputes to those empty spaces a number of books, not even if he does this imputation in a rather technical way, say by measuring the size of nearby books and dividing the size of the empty space by the average size of a book.
Now, we think that example patches the basic intuitive distinction between sampling and imputation.
Now, we support that intuitive conclusion by elaborating the nature of the differences by referring to technical literature, by looking at statutory history and reduce fill a certain amount of ink along the way, but we end up where we started.
When the statute says statistical sampling, it does not mean to include imputation.
So, the statute does not forbid its use.
Third, we ask whether the constitution forbids imputation.
The relevant part of the constitution says “representatives and direct taxes shall be apportioned among the several States according to their respective numbers counting the whole number of persons in each State.
The actual enumeration shall be made within three years after the first meeting of the Congress in such manner as they shall by law direct.
Now, you may have noticed there were two words in there, actual enumeration.
But we think those two words being there do not require say, an individual contact between a counter and each and every person counted.
Nor do those words otherwise restrict Census taking method so stringently that they would forbid the imputation used here.
In reaching this conclusion, we explore the context, the deligation of authority to Congress, a relevant history, contemporary linguistic usage, and Census practices both in the 18th century and more recent times.
We do not decide precisely what limitations those words, actual enumeration, do impose.
We conclude only that they do not bar the Census Bureau’s use of imputation here where the Bureau has made an effort to reach every household where it involves a tiny percentage of the population and where the alternative would mean a less accurate count.
We consequently affirm the judgment of the District Court.
Justice O’Connor has filed an opinion concurring in part and dissenting in part; Justice Thomas has filed an opinion concurring in part and dissenting in part in which Justice Kennedy has joined; and Justice Scalia has filed a dissenting opinion.