Support Bills H238 and S223 for Public Confidence in Elections "People with the most expertise with elections have the most questions about paper ballots. People with the most expertise with computers have the most questions about technology." Rep. Verla Insko, at Jan 25, 2005 hearing of the Joint Select Committee on Electronic Voting Which experts should you believe, and what are the facts on voter-verified paper ballots and the need for independent review of the programming code for all computerized voting technology? A. Views of computer professionals favoring paper ballots: · A poll of 4,600 ACM members of the Association of Computing Machinery, the world's oldest professional computer association showed 95.6% supported the following statement: · "Many electronic voting systems have been evaluated by independent, generally-recognized experts and have been found to be poorly designed; developed using inferior software engineering processes; designed without (or with very limited) external audit capabilities; intended for operation without obvious protective measures; and deployed without rigorous, scientifically-designed testing." "ACM has recommended that e-voting systems enable voters to inspect a physical (e.g., paper) record to verify the accuracy of their vote, and to serve as an independent check on the record produced and stored by the system. In addition, those records should be made permanent, not based solely in computer memory, to allow for an accurate recount." "http://www.acm.org/usacm/weblog/index.php?p=73" · Resolution on Electronic Voting signed by approximately 2000 persons in computer and voting related professions, "http://www.verifiedvoting.org/article.php?id=5028"· Speech by Chuck Herrin (separate paper). Mr. Herrin is Chief Security Officer and Chief Network Security Specialist for consulting firms in the Piedmont Triad area. He has performed vulnerability assessments for many medium-sized and Fortune 500 companies. He is certified as a CISSP, CISA, MCSE , CEH, and CHCP. · Paper presented to the Joint Select Committee on Electronic Voting by Prof. Rebecca Mercuri, Fellow at the Radcliffe Institute for Advanced Study at Harvard University (separate paper). B. Arguments against paper ballots by some elections officials (with rebuttals): 1. Hand counting ballots is no more accurate than electronic counting. The hand recounts in the Washington state gubernatorial race totaled 1,717 more votes than the mandatory machine recount, and 3,145 more than the original announced election result. Which would you assume is more accurate? 2. Paper ballots would establish not a floor but a ceiling for voting systems quality. In 2004, miscounts or malfunctions occurred with paperless electronic voting in the following counties: Burke, Carteret, Craven, Gaston, Guilford, Mecklenburg, Onslow, Pender, and Yadkin. ("http://www.ncvoter.net/news.html") No irregularities were found in the counties that still count paper ballots by hand: Graham, Hoke, and Tyrrell. ___________________________________________________________________________________________________________ "A Florida-style nightmare has unfolded in North Carolina in the days since Election Day, with thousands of votes missing and the outcome o two statewide races still up in the air." AP Newswire, Nov 13, 2004 Responses of the North Carolina State Board of Elections: "Except for the lost votes in Carteret County, Gary Bartlett, executive director of the North Carolina State Board of Elections, called the problems 'easily remedied and lessons learned.'" AP Newswire, Nov 13, 2004 NCBOE Deputy Director Johnnie McLean "Glitches always happen." (New Bern Sun Journal, Nov 4, 2004) And in Maryland, which uses only Diebold direct recording electronic voting systems (DREs) The Montgomery County Election Board's official report, "2004 Presidential General Election Review: 'Lessons Learned,'" on page 11 documents a 12% failure rate ("http://www.truevotemd.org/Resources/Lessons_Learned.pdf"). Specifically, a) 189 units (7%) of units deployed failed on Election Day, and of those, 106 units experienced screen freezes; and b) 122 additional (5%) units . . . suspect because few votes were captured compared to other units at the precincts For a 51-page summary of electronic miscounts and malfunctions in recent elections see: "http://www.verifiedvoting.org/downloads/resources/documents/ElectronicsInRecentElections.pdf"3. Requirement of printers Increases DRE costs. It was irresponsible ever to manufacture DREs that do not produce a voter verified ballot (i.e. seen by the voter and placed in a locked ballot box as a permanent record). The exorbitant prices charged by these irresponsible companies for adding printers should not be tolerated. Companies that the State Board of Elections has refused to allow to sell equipment in North Carolina (Accupoll and Avante) have been manufacturing from the beginning DREs with integrated printing systems for voter verified paper ballots that do not cost more than the paperless machines. The price of the Accupoll DRE recently was reduced to about $2,500 from over $4000. (Other makes also are reducing prices). 4. Paper ballot count is no more verified than electronic - the voter doesn't witness the count. In fact, the counting of votes ought to be transparently honest. It can be so only if witnessed. Internal electronic processes of counting votes are not witnessed. At the very least, the code for calculating vote totals needs to be reviewed by independent experts, as provided in the proposed legislation. More importantly, at least a portion of all machine counts need to be audited manually, by hand counts of paper ballots that can be conducted and witnessed jointly by representatives of all parties. 5. Lost ballots a greater threat than fraud or hacking. There was one case in 2004 in North Carolina of ballots being left unattended and thrown out by a janitor. This simply is a case of disgraceful incompetence of the precinct officers. With or without incompetence of elections officials, however, electronic voting tabulations are vulnerable to being fraudulently manipulated through modems to change thousands of votes in one county without a trace ("http://www.chuckherrin.com/hackthevotedemo.htm"). With incompetence, unauthorized persons may gain access to voting machines and reprogram them to manipulate vote counts, and, if they are quite expert, leave no trace. Hence assertions that elections can not be hacked or that they are not fraudulent can not be proven where an original paper ballot that can be recounted does not exist. Even where paper ballots do exist, the assertion can not be made unless they are counted. ______________________________________________________________________________________________________________ 6. Hand counting, for recounts or mandatory audits, is expensive · This argument implies that honest, accurate elections are not worth the price. But it also is fallacious. Paper ballots counted either by optical scan or by hand, cost, in the long run, only a fraction of what Direct Recording Electronic Systems cost. There are many ancillary costs for DREs, including licensing, maintenance, and operating expenses. Also, with DREs, only as many voters can vote at a time as there are DREs. Hence, while only one optical scanner would be required in each precinct, ten to twenty DREs per precinct are required to allow the same pace of voting and prevent excessively long waits. · A study has been conducted by Dr. Rosemarie Myerson of annual operating costs in neighboring counties in Florida, one of which, Sarasota, purchased ES&S iVotronic DREs at the end of 2001, and the other, Manatee, continued to use paper ballots with optical scan ("http://www.votersunite.org/info/costcomparison.asp"). The initial $4,700,000 capital cost of the DREs is not included in the comparison. Since the population of Sarasota is 25% greater than Manatee, for comparison, expenses for Manatee are increased by 25%. This augmented average annual cost for Manatee for 1999-2001 was $1,388,639, and the average for Sarasota $1,776,735. In the subsequent period, 2002-2004, average cost for Manatee was $1,724,256, and for Sarasota, after the conversion to DREs, the average cost was $2,883,659. "There would have been real savings if the Sarasota County Commissioners had decided to throw away the DREs in April 2004 and had spent $600,000 to buy Optical Scanners for the entire county as (was) suggested to the commissioners. The Supervisor of Elections said that paper ballots were expensive. Manatee bought their paper ballots for 20 cents each. For 100,000 voters this amounts to $20,000, a trivial sum compared to the estimate of $1,100,000 that the DREs annually cost Sarasota over Optical Scanners." 7. Hand counting is too slow; the public will not accept it. Even if hand counts take an extra day or two, costing up to another thousand dollars per precinct, it is a small price to pay for assurance of honest elections. The pressure for the fastest possible calls of elections comes not from the public, but from the media. The motive has zero to do with democracy, good government, or even good journalism - actually it is lousy journalism - and has everything to do with profit. Pressure comes also from election officials who are concerned about the ease of administering elections. Those pressures are in conflict with the requirements of honest and secure elections. Election officials who give higher priority to ease of administration than to being sure that every vote is counted perhaps ought to be replaced by officials with greater respect for the rights of citizens to have their votes counted accurately. 8. Maintaining secure custody of tens of thousands of paper ballot is complex and expensive. Physical security is a straightforward problem that election officials now are competent to implement. Breakdowns occur rarely, and only in cases of extreme incompetence. Computer security is a complex challenge that even hospitals, banks, the Department of Defense, and the FBI have not mastered. It is possible to minimize risks if one is willing to pay billions of dollars per software component, as Boeing has done for the control code of the 777, which has only 5% the length of the Windows operating system which most electronic voting machines use.         _______________________________________________________________________________________________________________   C. Questions about the Impartiality of the NC State Board of Elections · The SBOE has prevented new, possibly improved technology from being used in North Carolina, while allowing equipment of several manufacturers already known to be seriously flawed to continue to be used and to be purchased in 2004. For a limited summary of problems known to exist with equipment of Diebold, ES&S, Sequoia, MicroVote, and Hart Intercivic see "http://www.ncvoter.net/clearinghouse.html". · In the January 7 hearing of the Joint Select Committee on Electronic Voting, SBOE Executive Director Gary Bartlett stated that the reason why Accupoll, which is manufactured in Charlotte and was then the only federally certified DRE system producing a voter verified paper ballot, was denied certification by the SBOE was that it had been federally certified to the 1990 standards, not the 2002 standards. On March 25, 2004, the date that the Accupoll application was rejected and a moratorium imposed by the SBOE on certifying new equipment, no voting equipment had been certified as meeting the 2002 standards. What was the real reason for prohibiting a DRE that produces a voter verified paper ballot? · During the sixteen years that DREs have been used in North Carolina, the SBOE has shown no curiosity about how they are programmed, or their security or reliability, and has asked for no expert assistance in evaluating DRE systems . Viewing SBOE's performance even in the best light, it has been asleep at the switch. · Vendors of paperless DREs appear to have cultivated close relations with members of state and county boards of elections, based on their active involvement with The Election Center ("http://www.ncvoter.net/ElectionCenter.html"). The Election Center receives funding from ES&S, Diebold, and Sequoia. These vendors sponsored parties, awards, and a dinner cruise on the Potomac for election officials at its August 24-28, 2004 national conference for election officials. Mr. Bartlett is on the Board of Directors of the Election Center. D. Other Considerations · To allow blind or disabled voters to vote privately, one touchscreen ballot marking system with audio feedback might be purchased for each precinct. This provides a touchscreen interface, but produces only a marked paper ballot, not an electronic vote. Presently, the only brand available is Automark. It has not yet been federally certified, but legislation can specify that such devices be purchased for handicapped voters instead of DREs after they are certified. E. One section of H238 and S223 ought to be deleted. The North Carolina Coalition on Verified Voting opposes SECTION 3.(e) G.S. 163-132.5G reads as rewritten: "§ 163-132.5G. Voting data maintained by precinct. The only motivation for including this provision, requiring that all votes, including absentee, one-stop, and provisional, be identifiable by precinct, is to facilitate partisan redistricting. The time has come for both Republicans and Democrats to be committed to non-partisan redistricting, possibly by a non-partisan commission. Inserting this requirement will promote the use of electronic voting for the complex record keeping, and, if so, could increase annual operating costs of boards of elections across the state by tens of millions of dollars per year (extrapolating from the cost differential between optical scan and DREs in the Florida study, section B (6). F. Bottom Line: Do not listen just to the "experts." Especially do not Listen to Lobbyists for the Manufacturers. Listen to the people. For the voters, is it more important to obtain rapid election results with little human effort or oversight, or to obtain accurate, publicly verifiable results so that all voters can be confident that their votes are counted and are counted accurately? For election integrity, which is more important? For elected representatives, which is more important?