Working Notes on Bettencourt Audit

First posted Mar. 5, with two paragraphs added Mar. 7, note added Mar. 8 and two paragraphs deleted when we discovered that the copy of the Bettencourt Audit on file in the archives is not quite complete. Therefore, statements about what is NOT in the audit were corrected.–gm

Here are some working notes from my review of the so-called ‘Bettencourt audit’ submitted on Dec. 20, 2004 to Harris County District Attorney Charles A. Rosenthal, Jr. and containing 167 allegations of illegal voting (not a couple-hundred as reported by one columnist from the Houston Chronicle).

According to the letter of transmittal, the ‘review of election materials’ from Texas House District 149 was conducted pursuant to Texas Election Code 15.028, which reads as follows: “If the registrar determines that a person who is not a registered voter voted in an election, the registrar shall execute and deliver to the county or district attorney having
jurisdiction in the territory covered by the election an affidavit stating the relevant facts.”

Director of Voter Registration for the Harris County Tax Collector’s Office George Hammerlein explained Monday afternoon that ‘poll book audits’ are routinely conducted after every ‘major election’ and that a countywide review is still underway. Hammerlein says collection of the materials for House District 149 was prompted partly by phone calls ‘from both campaigns’ looking for information about particular voters. Also, says Hammerlein, Master of Discovery Will Hartnett (R-Dallas) said he would “appreciate you frontloading the research” for HD 149.

“We probably won’t segment anything else,” says Hammerlein, indicating that voter behaviors will not be broken down according to other House Districts in Harris County. To get an idea about the “routine” nature of the report, The Texas Civil Rights Review has filed an Open Record Request with Hammerlein, asking for copies of cover memos from other “poll book audits” that have been referred to the District Attorney following annual elections since 2000.

As a result of his review of precincts in House District 149, Bettencourt says in the cover memo that he found four voters not registered; two cancelled registrations; five registrations cancelled out of county after the election from provisional ballots; 139 voter registrations cancelled out of county after reviewing statements of residency filed by voters; one voter registration cancelled for lack of citizenship; and 16 voters who went to the wrong precinct. That’s how 167 illegal votes were identified by Bettencourt on Dec. 20.

The Texas Civil Rights Review is interested in the role played by the Bettencourt audit in the Heflin-Vo contest. At first glance, the thick notebook (about as thick as my middle finger is long from tip to knuckle) appears to be Andy Taylor’s play book. For someone who has spent a while going through all these docs, the Bettencourt audit looks very much like the eventual files that were presented into evidence.

In fact, according to my best count (and I would welcome Rick Casey double checking these numbers for me) the Bettencourt audit does point to 90 of the 143 voters listed as deposed in the final Hartnett report, and to 73 of the eventual 110 ballots that were discounted from either the Heflin or Vo column in the Hartnett report of Feb. 7. (Note: all numbers should read ‘at least’ because there are a number of depositions that were taken but not counted in the final report. I am only speaking of depositions reported.)

Had the hearing been confined to cases identified by the Bettencourt audit, then on the basis of Hartnett’s rulings of deposed voters, Vo would have lost 42 votes and Heflin would have lost 31 for a net Heflin gain of eleven votes. Of course, Heflin started out 33 votes behind Vo, so the election would have easily stood the test.

If we think in terms familiar to sales forces, there were a number of leads to illegal votes that were not generated by the Bettencourt audit. At least 53 more depositions were taken outside the database of Bettencourt-identified voters and of these, 23 votes were eventually discounted from Vo’s column, while 14 were deducted from Heflin (a spread of nine votes which still does not threaten the election, even when added to the spread of eleven from before).

For the Texas Civil Rights Review, two questions arise at this point: First, why was the Bettencourt audit produced in the first place? and Second, how did attorneys generate their leads for pursuing voters not found in the Bettencourt book?

The first question, why the review of illegal voters in House District 149? speaks to a concern about due process. Was the list generated because it was a close election? If so, what are the guidelines for determining how close an election has to be? etc. If the Houston Chronicle is not going to just try to sweep these questions under a rug, we’d appreciate some help getting serious answers.

The second question goes to the point of voter harassment entailed in this Republican-led election contest. How were voters identified for pursuit?

For instance, once again, we make reference to the “deportation” of District 149 voters into another district. Two voters who were victim to this scheme filed provisional ballots that clearly explained they were victims of a scheme, and those ballots were accepted by Harris County voting authorities. Yet along comes Andy Taylor who scoops these voters up into his net of alleged illegal voters, along with plainly written evidence that they were victims of unknown others, not perpetrators of voter fraud.

As we have reported, Hartnett called up at least one of these voters at home and ruled they were legal voters. But why did things get that far in the first place?

Or to put the question another way: If Andy Taylor is going to reform Texas election laws, who is going to reform Andy Taylor? He should at least be made to apologize, although we still favor a license hearing on Civil Rights grounds if such a thing is possible in Texas.

How were voters identified for pursuit? I find three examples where one spouse is identified in the Bettencourt audit, but both spouses end up in the final report. Two of the couples in question are identified through depositions as Vo households. The third couple was not deposed.

So this is one way that a couple of subpoenas might be served next time this sort of thing is done. Attorneys can just match spouses to voters previously identified by the local tax administrator and voter registrar.

And let’s be clear how Bettencourt identified 139 illegal voters. They told the truth about where they lived on election day. They went to the polling places in their old neighborhoods, wrote down their new addresses, and signed their names. I don’t know what you call that kind of activity, but fraud is a terrible word for it.

How else were voters identified for pursuit? There is an interesting case of a 24-year-old woman who registered to vote on Sept. 21, 2003, but who forgot to check the citizenship box. According to notes that I took at the archives, she filled out a “new registration” with her name and address, but failed to check the citizenship box. In August 2004 (“why the eleven month gap?”, I ask myself in my notes) she was notified by mail to complete her registration, so she sent it in again, forgetting once more to check the citizenship box. Finally on Oct. 22 she got the box checked properly and she voted on election day 2004. Whew.

But here comes Andy Taylor again and he’s got his copy of the election code which says that because our shero did not correct her registration within ten days of notice in August, the Oct. 22 registration has to be considered brand new. And since you need to register 30 days in advance of the election, and so forth, we have on our hands an illegal voter.

This is one of those interesting cases that Hartnett called a “battle of dueling internet pag

es” because according to the online database on Dec. 7 our shero was listed as a registered voter (“1/1/04 through 12/31/05” note the early 2004 date) but by Jan. 26, her registration had been changed (“11/12/04 through 12/31/05”).

After a couple of phone calls to her house, I can tell you that this voter is not much interested in talking about all this. And I’m probably not going to try again. But it is interesting to see how a voter can submit at least three registrations prior to election day, be listed in good standing in early December, and then by stroke of technicality be de-listed prior to any ruling by the Master of Discovery (who did subsequently rule her vote illegal and deductible from the Vo column.)

The case of our shero is neatly doubled by another voter who turned 18 last Sept. On Oct. 3 she filled out her first-ever voter registration and mailed it the next day, forgetting to check the citizenship box. On Oct. 17, she filled out another form, and again forgot the citizenship box. Finally, on election day, she got the form filled out right and her vote was accepted.

But the good standing of our second shero, like our first, did not survive. Sometime between Jan. 9 and Jan. 26, her online records were changed to reflect new effective dates for her voter registration. A voter registration that on Jan. 9 (three weeks after the Bettencourt review) was listed effective as of Oct. 31, 2004 was by Jan. 26 changed to reflect an effective date of Nov. 24. And her vote, too, was eventually deducted from the Vo column by Master Hartnett following a lengthy presentation by Andy Taylor. Again, with two phone calls I have not been able to reach her, and I doubt I will try again.

Note added March 8: In the hearing of Jan. 27, Taylor claims that one of the above voters was flagged as not registered by the Bettencourt report, but that section of the report was not included in the version he introduced into the archive. He also claims in the hearing that it was his team that demonstrated to Bettencourt’s office why the voter’s registration should be discounted. See updated story of March 8.