Dear Representative Freese:

At the hearing your Committee held October 17th, I submitted written testimony regarding AB 600. Although I was unable to appear in person, my firm was represented by Barry Ashenfelter. As I explained in my testimony, I have doubts about the State's approach to implementing H.A.V.A.

One of the points I tried to make was that the overwhelming majority of new registrations are taken on Election Day, which is very different from the situation in the rest of the country. The fact that "something over 70 percent of all registrations are taken at the polls" has major implications for the pending program to create a statewide voter registration database. In particular, it suggests there will be very little need for data entry workers except after elections, and that there will be very little payoff to the proposed pre-election spending.

I was surprised to hear that Kevin Kennedy, Executive Director of the Elections Board, who had already testified, asked for an opportunity to rebut me, telling the Committee, "I don't know where they get these numbers from. That 75% number they throw around is just absurdly high and wrong... The number of registrations on election day is usually 7% or less..."

My original estimate was an off-the-cuff guess based on five years' experience working with Wisconsin voter files. I didn't expect that it would be the focus of criticism, since I assumed it was common knowledge that the availability of same-day registration has strongly discouraged efforts to register voters before Election Day. Because my assertion has now been drawn into doubt, I have conducted a more serious effort to estimate the fraction of all registrations which are taken on Election Day.

My firm contacted the Elections Board and confirmed that they do not keep any systematic information on the subject. The Board does compile a report which compares the number of Election Day registrations with the total number of votes cast, but apparently the only non-Election Day registrations they keep track of are those that occur within the final ten days before each election. They apparently have no idea how many registrations occur outside the polling place. (Oddly, they compare the number of registrations taken at the polls with the total number of votes in the state - including areas which do not have any system of registration.)

In order to estimate the fraction of registrations which are taken at the polls on Election Day, I contacted Milwaukee, Madison and Green Bay, but apparently none of them keep track of non-election-day voter registration. Instead, I have relied on voter files my firm had previously purchased from several municipalities which happen to include the exact date on which a person registered. We looked for data sets where registration data cover a full four-year cycle in order to smooth out the surge in registration associated with presidential and gubernatorial elections. These are the results of our research from five fairly typical mid-sized municipalities.

Village of Greendale (Milwaukee County) - 75% of all registrations taken from 1999 through 2002 were taken on the day of an election. Fully 45% of all registrations taken over the four year period occurred on the day of the 11/7/2000 Presidential election.

City of Manitowoc - 70% of registrations taken between1999-2002 were taken on an election day. (48% on 11/7/2000.)

City of Wausau - 73% of registrations taken 1999-2002 were taken on an election day. (36% on 11/7/2000.)

City of Appleton - 75% of registrations taken 2000-2002 were taken on an election day. (We were unable to assemble a full four year cycle of data, which we estimate would have reduced the percentage to 74%.)

City of Lake Geneva - 72% of registrations taken 2000-2002 were taken on an election day. (Only three years of data was readily available to us.)

Of course, these do not constitute a random sample and the pattern could conceivably be different in the largest cities, however, I believe this information provides reasonable support for my original estimate. Our raw data is available upon request.

Mr. Kennedy, as the State's chief elections officer, has a duty to ensure the Legislature receives accurate information regarding pending legislation. I have provided this letter to him and invited his response.

I would just like to reiterate my opinion that the creation of a list of "registered" voters, rather than a list of "eligible" or "potential" voters is an expensive waste of time in Wisconsin.

Thanks for your consideration.

Sincerely,
Mark Grebner
Owner, Wisconsin Voter Lists