Philip D. Zelikow, Ph.D.
Executive Director, National Commission on Federal Election Reform
Director, The Miller Center, U.Va.
National Commission on Federal Election Reform Report
September 4, 2001

Philip Zelikow: First we were struck by the lack of knowledge people have about how election administration really works in this country. We have a lot of knowledge about voting behavior, but we have very little knowledge about how elections are administered. For instance, if we went to the library and got a book on how elections are administered--what machines people use, how much money is spent to run the election, all the different counties in America--there is no such book. Not one that had been written in the last 25 years. The last such book was published by the Brookings Institution in 1974. The previous book before that was published in 1934. So, we would ask fairly basic questions, such as: "how much money do we spend in America on administering elections?" We don't know the answer to that question. To this day we do not know the precise answer to that question. What happened was various people had to do surveys this year, going around calling counties and asking how much they spend on running elections and voter registration and all of that in their county. We could go to subject after subject. We had to do inventory and our task forces had to compile information that has not been published before.

So, the first point is that no one really knows much about election administration. This is very important from a policy perspective because you could go to an election official in Maryland, for instance, and they will talk about the will talk about they agonies that they have faced in trying to set up a state-wide voter registration system. Most are ignorant that the state of Michigan had already solved this problem at a fraction of the cost that Maryland had just spent to solve it. Why? Whose job is it to tell Maryland about the state of Michigan? They did not know. There are no books, no journals out there to publicize this. The first point is how little we understood. How many of you understood what you now understand about punch-card voting, or the fact that we don't have a standard definition of how to count a vote from a punch-card machine so that every county can make up its own definition (which has turned out to be unconstitutional by the Supreme Court). People did not even know what they didn't know.

The second overall observation: The problem is significant. To try to just give a sense of the significance, the premiere scientific study of this done this year was by Caltech (?) and MIT. They didn't really work really hard on the policy or what exactly should be done, but they did do a good job of studying some of the associative engineering because that is their comparative advantage (and we drew heavily from some of their work). Caltech an MIT, partly to get an attention getting headline, said that in the last election, four to six million votes were lost. That is a lot. Just to refresh your memory, about a 100 million votes passed. In the popular vote, Gore had 500,000 more than George W. Bush. And so, 4-6 million, that is a lot. If you asked the question that we asked about how Florida was so close and elections are so close? How often is it that key states in presidential states are decided in elections by vote margins of one percent or less? The election of 1940 was decided by such a margin, the election of 1960, 1968, 1976, and 2000 (and that is just the presidency). In other words, in all those presidential elections that I just cited, if I could re-allocate one percent of the vote, I could change who was president. So, close elections are surprisingly common and so therefore, differences at the margin really do matter. In that sense, Florida is not simply an anomaly that can be discarded.

With those two observations, let me say a little about our substantive recommendations. Our number one recommendation, and the one that is really the centerpiece for everything that follows (but is also the one that got the least publication) is when President Bush comes out and endorses this four squared in the Rose Garden on July 27^th. News about this support was on the front page of every newspaper in America. The editorial boards of every newspaper in America editorialized on the report--some liked it, some didn't like it, some even thought about it before they wrote their editorial for like 60-90 seconds. That was one of the striking things about these editorials--how breezy they are. But, there was very little attention to this. Most of it was about election holiday. But the lead recommendation was statewide voter registration systems. The reason that is so important is because, thinking historically, how we do voter registration in America. Election administration in America was organized around the states. States determined who could vote. In all the jurisdictions of the state, people would just assemble and vote. There was no such thing as voter registration. Voter registration is an innovation that comes into America after the Civil War--the late nineteenth century in the time before WWI. It is a good government innovation. And, like all good government innovations, there are mixed motives behind it. The progressive people wanted voter registration and it was significantly a response to the rise of large cities in which there was much more anonymity. They needed to register who would vote; otherwise there was no way of telling there is voter fraud. So, voter registration was instituted.

The big innovation of voter registration was voter rolls. Voter rolls were maintained by the precinct. One thing this instantly did is it meant that now a lot of authority over voting has moved down from the state to counties and local jurisdictions because they are the ones who maintain and keep the voter rolls which had to be renewed every couple of years. Back then you had to re-register over and over again.

Another big government innovation in the 1940's and 1950's was the widespread adoption of permanent voter registration, so that once you registered in the state of Virginia you didn't have to remember to register over and over again--you stayed on the books.

Now, that is all good, except it does not work well in our modern society. We live in a country where 1/6 of the population moves every year. Think about that for a minute. But, 3/4 of those moves are within the state. And so, if you have statewide voter registration, they are just fluctuating within the state. What happens is that people are on permanent voter registration that keeps them on the books, but they move. Maybe they remember to register in the new place, maybe they don't. So, what happens is that the data system is no longer a system that the counties can keep up with. We have this history of highly decentralized election system in America. Most countries have a national election system--you have a national citizenship cards, national voter I.D.'s which American's wouldn’t want. We have this incredibly decentralized system, but that means that every county decides what election machines it uses, decides how to count votes, decides who is on its voter list, so, how do they share data with each other?

Well, it's very hard. How do they share data with the state so that they know who is in the state prison or so they know who has died in another county? So what happens is that the county voter lists are just ripe with problems--people aren't on them that should be on them and then maybe 15-20% of the people on these voter lists aren't there anymore. You can routinely run newspaper stories that you can prove there are more people on the voter list than live in the county, including children. There are many dead people on the voter lists. There are household pets on the voter lists. So, voter lists are full of, what experts call, dead wood (some cases as high as 25%). This creates opportunities for vote fraud. And, yes Virginia, vote fraud still occurs in America. It occurs frequently and occasionally people get caught and get sent to federal penitentiary. Statewide voter registration systems that are basically keyed above all to getting drivers' licenses help solve a lot of these problems. The Motor Voter Law said that when you get a driver's license you should get an opportunity to register to vote. That works pretty well, but not all that well. There are still a lot of problems with that. That is still two databases. There is the state DMV database and then the county database and then maybe they share data well or maybe they don’t. There are problems as mundane as that.. What if the state DMV list and the state voter list and then supplement that with information from county social service and welfare agencies. Match it up with the state database and their vital statistics list and maybe even create a statewide street index so that you can see the addresses of people living in certain locations. You would solve a lot of this problem. If someone moved from one place to another it would very easy problem to fix and he or she could still be on the statewide voter list--I could either vote or re-register. That model has already been done in the state of Michigan (which is a big state within a very complicated set of election jurisdictions) and it works, and it is cheap. Now they are finding it so wonderful that they use if for all sorts of purposes. The statewide street index is applicable for all kinds of state planning and figuring out where they can put green fields and brown fields and other developments. It works. The software is essentially free (Michigan will share it with whoever wants it). If you solve that problem, other things become possible. And, that brings us to our second recommendation. The first is that you clean up the voter rolls. If you solve that problem, then you can do something called provisional voting.

A big thing with the problem in Florida is that people show up and say that they should be able to vote, but they are not on the voter roll. They are not on the voter roll because either the county messed up, or because they forgot to re-register since they moved wither within the county of from another county. This happens hundreds of thousands of times. And the problem in Florida was that people got in an argument was that people got in arguments with poll workers? How many of you have volunteered to work as poll workers? How many of you have had distressing arguments with voters who said that they should have been on your voter roll? Provisional voting just says that poll workers can't resolve the problem right there but that the voter can go ahead and fill out a ballot and include information that says that they are qualified to vote in this jurisdiction. The county registrar will then check out the information. If they can verify your information as correct, then they will count your vote, and if they cannot verify your information, then they will not count your vote. The poll workers are, therefore, not under pressure. If you then have a statewide voter list, which is networked to local jurisdictions across the state, this becomes much easier to do. This enormous problem in Florida that simply wouldn't go away--no one will ever be turned away from a polling place again and be refused to be give the opportunity to vote. This costs next to nothing to do and it will solve a huge part of the problem.

We also would call for a national holiday. Almost everyone thought that the reason that we would have a national election holiday would be to increase the voter turnout. That is not actually the reason we recommended it. We actually don't think that it increases voter turnout. The reason we recommended a holiday is much more mundane--it is hard to get poll workers, trained good, poll workers, and it is hard to get accessible polling places. It turns out that buildings that are ADA compliant are public places, so you need public places for polling, like schools. Schools are increasingly unavailable as polling places across America. One reason that people don’t want to talk about it very much is for liability reasons. They don't like having a lot of grownups around where there are children, especially girls. And so, insurance companies have raised insurance rates. So, a national holiday opens that all up. It allows get government employees who would be willing to work as poll-workers, and it also allows you to get access to the public buildings. That is a big problem in the country.

The politics of this are terrible, however. We already have a national holiday around the second week of November and it is called Veterans Day. Adding another national holiday, no business man will want to support because of course, you have to pay your employees to take the day off. How about putting it on Veteran's Day and merging it with Election Day, which all the veterans on our commission thought was a splendid idea since they felt that was what they had been fighting for. But, the veterans groups are almost unanimously opposed to this--to leave their day alone for commemorations and such. So, politically this is a very uphill battle.

We made a number of recommendations to simplify and facilitate absentee voting by uniformed and overseas citizens. You have heard a lot about this on t.v. It turns out that if you adopt the statewide voter system that we have discussed, this enormously simplifies the way we can solve the military voting problem. We talked about restoring voting rights to people who have been convicted of felonies, after they have served all of their time (including any period of probation and parole). We think that if it is perfectly legitimate for states to disfranchise people who commit felonies, it is up to them to make that judgment--a moral judgment as of whom you think is qualified for all the rights of citizenship. But, to say that anyone who has committed a felony is disfranchised for life, doesn't pay much attention to the particular and individual circumstances. The only people who see these people as individuals is the judicial process and they take that into account into assigning punishment. So, the commission thought unanimously that once they have served whatever time the judicial process has meted out (including probation and parole), they ought to be entitled to get their voting rights restored. There are a number of states that have automatic lifetime disfranchisement. The effects of this, especially in the African American community, are statistically significant (15% of the adult population in some states). This is not something that civil rights groups like to wave their arms and talk about. But, for various reasons, this is an uncomfortable fact, but the effect is that these provisions do more to disfranchise black people more than any other law in the books in America. And so, the commission thought that this reform would address that concern.

Also, we thought that states should address the voting issue in problem. All of you heard about punch cards and how difficult they were. We decided, though, not to pick winners and losers in the debates over what is the best machine. Technologies have all been fast and there are a lot of different reasons why counties pick out the machines that they pick (which we found out about). So, we said that we would look at best practices around the country and regardless of the machine you use, what can be accomplished with these machines. Some counties say that they don' t have to replace their machines because they can get their error rates down to better voter registration or more investment in training coworkers. This is a managerial point. So, how good is good enough?

What we said was that the bottom line ought to be that your over-votes and your under-votes should never be higher than 2% of the people who voted. We looked around and anything that was over 2% we thought was a problem. Every jurisdiction in America should be able to get their residual vote to 2% or less. We force every county to publish and announce how well they did against that benchmark, and then we think that politics will take care of this. In fact, there are counties around America where residual vote rates are in double digits. We publish the residual vote rates for the forty most populous counties in America and ranked and grade them. So, you could see how people could see what was possible--how Oakland, California can get under 2% and Palm Beach, Florida has more than 6%. So, there is something going on here beyond demographics. Through a managerial perspective, state the bottom line and then hold people accountable for how they get there (but let them have a lot of discretion in how they do it). That benchmark approach was very important to the commission.

We also set up a scheme for federal legislation. It consists, basically, of federal voting system standards, which exist now but we propose various ways those standards can be strengthened (technological innovations which are very promising right now--especially in touch-screen machines--some of you have used them in Albermarle County). So, federal voting system standards and then beneath that, federal support for election administration nationwide. We believe that counties, which now bear almost all the burden for election administration, should not have to spend any more money on it at all. They are already straining as much as they can. The federal and state governments ought to pick up the slack. You can actually increase the amount that we spend on election administration by 30-40% at a very modest cost. We concluded that in the whole country we spend don't spend more than about a billion dollars a year on election administration. That's every state, every county in America put together about a billion dollars on running the infrastructure of democracy. Just to give you a sense of scale, this number is so small that the census of activities of government does not bother to record it. The smallest activity of government that they bother to record is waste disposal on which we spend, nationwide, $16 billion. In other words, sixteen times as much as we do on elections. We think that this problem can be solved by spending $300-$400 million more a year, nationwide, on election administration. If the federal and state governments would capitalize on state-revolving funds administered by the states (judging their own needs) at about a level of a billion dollars each, they could get just that one shot of appropriation which would be more than enough to meet all the needs that I have just described for more than a decade.

We even convinced the Bush White House that was important. Their OMB was very suspicious. In addition to the federal voting system standards, we said that the states that wanted the federal money need to put up the match and need to meet a series of requirements--state-wide voter registration systems, network to local jurisdictions, benchmarks for their residual vote rate, and so on. Then we made it clear that the choice and strategies of how to get those goals should be left to the states. So, the whole emphasis here is to create a national framework, make the federal government a limited partner, and the primary focal point of this administration should be back where it was 200 years ago--with state governments. They run the voter registration systems, they run the grant programs, they set the statewide benchmarks for how their votes should be conducted, and the federal government is supporting them.