IN THE UNITED STATES DISTRICT COURT FOR THE EASTERN DISTRICT OF WISCONSIN

Transcription

1 IN THE UNITED STATES DISTRICT COURT FOR THE EASTERN DISTRICT OF WISCONSIN BETTYE JONES, et al., Plaintiffs, v. Case No. 2:12-cv LA JUDGE DAVID G. DEININGER, et al., (all sued in their official capacity), Defendants. DECLARATION OF M.V. HOOD III I, M.V. Hood III, do hereby declare the following: 1 Case 2:12-cv LA Filed 06/04/12 Page 1 of 33 Document 45

2 I. INTRODUCTION AND BACKGROUND 1. My name is M.V. (Trey) Hood III, and I am a tenured, associate professor at the University of Georgia with an appointment in the Department of Political Science. I also serve as the Director of Graduate Studies for the Department. I have been a faculty member at the University of Georgia since August of I am an expert in American politics, specifically in the area of electoral politics, racial politics, election administration, and Southern politics. I teach courses on American politics, Southern politics, and research methods and have taught a graduate seminar on the topic of election administration. 2. I have received research grants from the National Science Foundation and the Pew Charitable Trust. I have published peer-reviewed journal articles specifically in the area of voter administration and voter identification. My academic publications are detailed in a copy of my vita which is attached to end of this document. 3. During the preceding four years I have offered expert testimony in two cases, State of Florida v. United States (No , D.D.C.) and NAACP v. Walker (11-CV-5492, Dane County Circuit Court). In assisting the defendants in analyzing the potential impact of the Wisconsin voter identification statute, I am receiving $200 an hour for this work and $300 an hour for any testimony associated with this work. In reaching my conclusions, I have drawn on my training, experience, and knowledge as a social scientist who has specifically conducted research in the area under examination in this expert report. 2 Case 2:12-cv LA Filed 06/04/12 Page 2 of 33 Document 45

3 II. OVERVIEW In Section III of this report I analyze data provided to me by the State of Wisconsin in an attempt to estimate the number of registrants who possess a driver s license or state issued identification card. Sections IV through VI provide my response to expert reports produced by the plaintiffs witnesses Mr. Leland Beatty, Professor Barry Burden, and Professor Lorraine Minnite respectively. In Section VIII I provide a synopsis of my overall conclusions for this case. 3 Case 2:12-cv LA Filed 06/04/12 Page 3 of 33 Document 45

4 III. DETERMINING HOW MANY WISCONSIN REGISTRANTS HAVE A LICENSE OR STATE ID CARD A. OBJECTIVE 4. I was asked by attorneys for the State of Wisconsin to make an assessment concerning the number of Wisconsin registrants who lacked either a state driver s license or DMV issued identification card. B. DATA AND METHOD 5. I was provided data by the Wisconsin Attorney General s Office that originated from two sources: The Government Accountability Board and the Department of Transportation, Division of Motor Vehicles. From the Division of Motor Vehicles I received a set of data files that contained a list of Wisconsin residents who had been issued a driver s license and a second file that detailed Wisconsin residents who had been issued a state identification card. Both these data files contained a unique identification number, along with information on the race/ethnicity, gender, date of birth, and name (defined as first name, middle initial, and last name) From the Governmental Accountability Board I received a copy of the current Wisconsin voter registration database which contained a record of all registrants who were classified as Active. Along with a unique voter identification number for each registrant the database contained a state identification number which is analogous to the driver s license or state identification card number I combined the data files that contained information on Wisconsin residents with driver s licenses and those with state identification cards. Once combined, there were a total of 4,825,131 records; however 114,607 of these records were duplicates as determined by the driver s license number. I eliminated these duplicate records leaving a total of 4,710,524 records. Hereafter in this report I will refer to this combined dataset as the driver s license file. 8. I then imported the aforementioned datasets into MS Access, a database program, in order to determine the number of Wisconsin registrants who possessed a driver s license or state identification card. In order to do so I created a join between the state identification number field in the voter registration table and the driver s license number field in the driver s license table. Using the state identification number common between these two files I was able to determine the number of Wisconsin registrants who possessed a driver s license or state ID card. One known issue is that prior to 2006 the State of Wisconsin did not require that citizens applying to register to vote provide their state identification number (from their driver s license or state ID card) on their application form. In light of this issue, the initial match using this field resulted in a heavy undercount of registrants with these types of identification. 1 The driver s license and state ID card files do not list anyone with an expiration date prior to May 16, Prior to January 1, 2006 registrants were not required to provide their driver s license or state identification number when registering to vote (Wisconsin Act 265, Section 49b amended Wisconsin Statute s. 6.33(1)). 4 Case 2:12-cv LA Filed 06/04/12 Page 4 of 33 Document 45

5 9. In order to create the most accurate count possible I executed an additional match using a different set of criteria. The second match criteria involved using several fields in common between the three data files previously described. The second match performed used the first name, last name, and date of birth fields to find records with identical values for these variables between the two files. In this manner, any registrant whose first name, last name, and date of birth matched a holder of a driver s license or state identification card was identified as possessing one of these two types of state-issued identification. 10. After conducting the two matches previously described, I calculated the total number of matches by including any registrant matched by the driver s license number match or any registrant matched by name and DOB. This total is denoted as Match A. After conducting these matches there were a number of registrants in the database who did have a state identification number associated with their record, but who remained unmatched using any of the match criteria previously described. Match B adds this group to the total number of matches produced under Match A under the assumption that these registrants have, or previously possessed, a driver s license or state issued identification card. C. FINDINGS 11. The results of the queries used to match cases across the driver s license and voter registration files are presented in Table 1 below. For each type of match the total number of records are recorded along with the corresponding percentages for the total number of registration records. Using driver s license numbers alone I was only able to match 61.27% of the registration records. The name and DOB match identified 86.87% of registrants who had a matching record in the driver s license file. Combining these two approaches yields a slightly higher match rate. The total percentage of registration records matched using both these methods was 90.57%. In the end, Match A estimates 9.43% of Wisconsin registrants lack a driver s license or state identification card. I should note that these figures would appear inflated if compared to other states that have undertaken similar types of analyses. For example, in Georgia a match run between the driver s license and voter registration databases indicated that 6.04% of registrants lacked proper photo identification Following Match A there were 106,825 registrants with a state identification number who remained unmatched. Match B adds this group to the total number of matches produced following Match A. Adding those unmatched registrants in the database with a state identification number to the group of registrants matched previously leaves 6.15% of registrants possibly lacking a driver s license or state ID card. I think it is certainly reasonable to include this particular group of registrants in the group with a driver s license or state ID card as it is clear that at the time they registered to vote (or made a change to their registration record) that they did have one of these forms of Act 23 identification. Even if expired, these forms of identification can be used for voting as long as the date of expiration occurred after the last general election. In addition, the fact that these registrants at one time were in possession of a driver s license or state ID card is an indication that they could obtain one again (for free if 3 Both the Georgia registration database and the driver s license database contain full Social Security numbers. M.V. Hood III and Charles S. Bullock, III Worth a Thousand Words? An Analysis of Georgia s Voter Identification Statute. American Politics Research 36(4): Case 2:12-cv LA Filed 06/04/12 Page 5 of 33 Document 45

6 necessary) in order to vote. 13. The match rate achieved produces an undercount for two reasons. The first concerns incomplete records in the State ID field, especially for some registrants who registered prior to January 1, 2006 when this information became required for those registering to vote. The second issue involves the fact that the driver s license and state ID file did not include any expired records. According to Act 23, a registrant can use an expired license or state ID card to vote as long as the expiration date occurred after the most recent general election. For this reason the files received from the DMV should have included any record with an expiration date from November 2, 2011 to the present. At least some of the non-matches for registrants with a State ID number may be due to the fact that the driver s license database did not include any records where the state ID card or driver s license had expired. 6 Case 2:12-cv LA Filed 06/04/12 Page 6 of 33 Document 45

7 Table 1. Matching Voter Registration Records to Driver s License Records Number of Records Percent of Total Registrants Driver s License/State ID Match 1,994, % Name/DOB Match 2,827, % Total-Match A 2,948, % Unmatched Registrants -Match A 306, % Unmatched State ID Number 106, % Total-Match B 3,055, % Unmatched Registrants-Match B 200, % Total Registrants 3,255, Table Notes: Driver s License/State ID Match: Includes any registrant matched by Driver s License or State ID Number. Name/DOB Match: Includes any registrant matched by the following fields: first name, last name and date of birth Total-Match A: Includes any registrant matched by Driver s License/State ID number or any registrant matched by name and DOB Total-Match B: Includes any registrant from Match A or any registrant with a State ID present in registration database. 7 Case 2:12-cv LA Filed 06/04/12 Page 7 of 33 Document 45

8 IV. RESPOSNE TO MR. LELAND BEATTY S EXPERT REPORT 14. In this section I will compare my analysis presented above to that rendered in the expert report of Mr. Leland Beatty. In paragraph 8 of his report Mr. Beatty states that 11.1% of Wisconsin registrants lack a driver s license of state ID card. Using Match A, I only found 9.43% of Wisconsin registrants might not possess these types of identification. Using the results of Match B this figure would drop to 6.15%. Mr. Beatty s figure of 356,512 unmatched records is approximately 50,000 records greater than the figure I derived for unmatched registrants using the Match A criteria (see Table 1). Examining his matching methodology it seems apparent that Mr. Beatty failed to make use of the driver s license numbers which are common to both files (though incomplete in the voter registration file) to perform any of the matches he performs. This fact would account for some, but not all, of the discrepancy. 15. I should also note that I performed a similar matching analysis for a state court case, NAACP v. Walker, and derived very similar figures to those presented in Table 1 of this report. My calculations are, in part, slightly different as I was using another set of driver s license/state ID data. Analogous to the Match A total in Table 1 of this report, I found 302,082 registrants lacked a driver s license or state ID card (or 9.28% of the total) in my April 15 th report. 4 Using Match B criteria the figure for unmatched registrants would be 199,552, or 6.13% of the total. I should also note that the plaintiff s expert witness in the NAACP case, Professor Kenneth R. Mayer came up with similar figures. In his expert report of April 14, 2012 Professor Mayer states that there are 301,727 registrants who do not appear in the DOT database. Professor Mayer indicates this is 9.3% of SVRS registrants (p. 3) Mr. Beatty further indicates that he sent these 356,512 unmatched records to a firm, Ethnic Technologies, that identified the race and ethnicity for 91.6% of these cases. Unfortunately, Beatty provides no explanation of the methodology this firm used to produce these estimates which makes it impossible to determine the validity of this technique for estimating the racial/ethnic breakdown for these unmatched records. 17. Given the fact that Mr. Beatty s calculation of unmatched records (with the inference being these registrants lack a driver s license or state ID card) is inordinately high and the fact that he relies on a third-party to produce part of his analysis with no explanation of the methodology used, there is good reason to call the results in his expert report into question. The actual number of Wisconsin registrants without a license or state ID card is conservatively below 10% and I would contend for reasons outlined above that it is closer to 6%. The race and ethnicity of registrants who were not matched to a DMV record is still an open question in my opinion. 4 See Tables 1 and 2 of M.V. Hood III. Supplemental Expert Report for Milwaukee Branch of the NAACP et al. v. Scott Walker et al. (11-CV-5492). April 15, Kenneth R. Mayer. Estimate of Voting Eligible Population Lacking IDs Necessary to Vote Under Act 23. Milwaukee Branch of the NAACP et al. v. Scott Walker et al. (11-CV-5492). April 14, Case 2:12-cv LA Filed 06/04/12 Page 8 of 33 Document 45

9 V. RESPONSE TO PROFESSOR BARRY BURDEN S EXPERT REPORT A. EVIDENCE OF VOTER DISCRIMINATION 18. In regard to Senate Factor 3 in Paragraph 14 of his report Professor Burden s states that Wisconsin has a long history of election practices that facilitate discrimination. He specifically cites a statute in place until 2006 that required voter registration for citizens living in municipalities of 5,000 and an order by the Justice Department that Spanish-language ballots be made available in Milwaukee County. 19. There have been three separate applications of Section 5 since the Voting Rights Act was implemented, in 1965, 1970, and The 1965 and 1970 triggers applied to jurisdictions which had a test or device for registration (i.e. literacy test) and which also had low levels of registration or turnout. In 1975, a third trigger related to single-language minorities was added. If a jurisdiction falls under Section 5 coverage the implication is that discriminatory registration or voting practices were present. Section 5 jurisdictions are required to pre-clear any changes related to the conduct of elections or registration with the federal government prior to implementation in order to prevent new discriminatory measures from taking effect Neither the State of Wisconsin, nor any component therein, has ever fallen under Section 5 coverage of the Voting Rights Act. 7 Had Wisconsin had a long history of discriminatory election practices then certainly the State, or specific jurisdictions within the State, should be covered by Section 5. B. MINORITY TURNOUT 21. In Paragraph 9 of his report Professor Burden cites a number of statistics related to turnout by race and ethnicity in Wisconsin. It should be noted that the State of Wisconsin does not keep official registration and turnout statistics by race and ethnicity. The statistics cited by Professor Burden were derived from the Census Bureau s Current Population Surveys of Voting and Registration. As the name implies, these are surveys and, as such, must be viewed in that light. Political science research has long indicated that self-reports of turnout often result in inflation on this measure. 8 Many surveys that include questions relating to voter turnout go through a validation process where public records are used to correct respondent s answers on the voter turnout question. The voting and registration surveys conducted by the Census Bureau do not undergo any type of vote validation process. Because these estimates are based on survey data and because these types of surveys are prone to specific biases it is important that the CPS turnout estimates be used in conjunction with measures of uncertainty. In other words, a range where the true measure is located should also be provided. 9 6 Charles S. Bullock, III and Ronald Keith Gaddie The Triumph of the Voting Rights Act. Norman, OK: University of Oklahoma Press. 7 A list of covered jurisdictions is located at: 8 For a review of this literature see: Seth C. McKee, M.V. Hood III, and David Hill Achieving Validation: Barack Obama and Black Turnout in State Politics and Policy Quarterly 12: As indicated, Wisconsin does not record the race/ethnicity of registrants and, as such, the CPS Surveys are the best data available for registration and turnout statistics related to race. Even so, the CPS data are based on surveys. 9 Case 2:12-cv LA Filed 06/04/12 Page 9 of 33 Document 45

10 22. For each turnout estimate the Census Bureau provides a margin of error whereby a 90% confidence interval can be calculated. In 2010 the CPS estimated that turnout for blacks in Wisconsin to be 46.3% of their citizen voting age population. The margin of error provided for this estimate is So, in order to create a 90% confidence interval 13.0% should be added and subtracted from 46.3%. From these calculations we can be 90% certain that the true estimate of black turnout in Wisconsin during the 2010 midterm lies between 33.3% and 59.3%. It is also important when we compare turnout figures for various groups that we are working with comparable figures. Professor Burden s report using a combination of figures, some of which are based on the voting age population and some based on the citizen voting age population For the figures presented below I report turnout from the CPS as a percentage of the citizen voting age population for Anglos (non-hispanic whites), blacks, and Hispanics in Wisconsin for 2008 and According to the CPS in 2008 Anglo turnout was 72.3%, black turnout 73.1%, and Hispanic turnout was 44.4%. In 2010 the estimated Anglo turnout was 55.3%, black turnout was 46.3%, and Hispanic turnout 39.4%. These values are plotted in Figure 1 and, in addition, for each point estimate the 90% confidence interval is also presented. When comparing turnout between two or more groups it is important to determine whether the confidence intervals for such estimates overlap. Looking at 2008 turnout in Figure 1, it is easy to see that the confidence intervals for the Anglo and black estimates of turnout overlap. In this case one cannot say with any degree of statistical certainty that the real rate of voter turnout for Anglos differs from that of blacks. For this same election the confidence band for Hispanics does not overlap with either Anglos or Hispanics. In this case then one can say with some statistical certainty that Hispanic turnout in 2008 was lower than Anglos or blacks. In 2010, the confidence intervals for both black and Hispanic turnout overlap with the confidence ban plotted for Anglo turnout. In the most recent general election it is impossible to say with any degree of statistical confidence that black and Hispanic turnout differed from Anglo turnout. Accordingly, estimates based on these data also come with some degree of uncertainty. For states that do record the race and ethnicity of registrants, the official statistics should be used to determine rates of registration and turnout on this factor. 10 U.S. Census Bureau. Voting and Registration. Population Characteristic (P20) Reports and Detailed Tables available at: 11 All data for this section of my report is derived from the Census Bureau s Voting and Registration Reports, See footnote 10 for a more detailed citation. 10 Case 2:12-cv LA Filed 06/04/12 Page 10 of 33 Document 45

11 Figure 1. Wisconsin Voter Turnout (CVAP) by Race/Ethnicity, Percent Turnout Election Anglo Black Hispanic 24. Unfortunately, before 2008 the Census Bureau did not report citizen voting age turnout figures at this level of detail. While most Anglos and blacks in Wisconsin are citizens, a substantial proportion of the Hispanic population is comprised on non-citizens who are not an eligible part of the electorate. As I stated above it is especially important when drawing comparisons among groups to use congruent metrics. For this reason, I present a comparison of Anglo and black turnout based on the total voting age population from 2000 through As with Figure 1, Figure 2 plots turnout estimates along with 90% confidence intervals for each group. For five of the six elections plotted in Figure 2 there is no evidence to conclude that Anglo and black turnout are statistically different from one another. The 2006 general election is the only case where the 90% confidence intervals do not overlap. Anglo turnout in the 2006 election then is higher than black turnout. 11 Case 2:12-cv LA Filed 06/04/12 Page 11 of 33 Document 45

12 Percent Turnout Figure 2. Wisconsin Voter Turnout (VAP) by Race, Election Anglo Black 25. In conclusion, from the data collected and analyzed there is no evidence to conclude that a historical pattern of lower minority turnout in Wisconsin exists. From 2000 through 2010 black and Anglo turnout rates are statistically the same in all but one election cycle. Likewise, in more recent comparisons using the citizen voting age population as a baseline for turnout Hispanic turnout differed from Anglo turnout in 2008, but not the most recent election in C. DESCRIPTIVE REPRESENTATION 26. In response to Senate Factor Seven which deals with the level of descriptive representation for racial or ethnic minorities Professor Burden concludes that [b]lacks are not well represented in Wisconsin public office and that Latinos are vastly underrepresented. According to the 2010 Census blacks comprise 5.4% of Wisconsin s voting age population. 12 At this level it is certainly not possible for the black electorate alone to elect a candidate in a statewide race. Black representation can present itself at lower officeholding levels, however, especially in electoral systems composed of single-member districts. Table 2 below documents the present black representation in the Wisconsin State Legislature U.S. Census Bureau Census National Summary File of Redistricting Data. Table QT-PL. Race, Hispanic or Latino, Age, and Housing Occupancy: Source: Wisconsin Blue Book, Available at: 12 Case 2:12-cv LA Filed 06/04/12 Page 12 of 33 Document 45

13 Table 2. Black Representation in the Wisconsin Legislature, Chamber Total Membership Black Members Percent Black Percent of VAP State Assembly % 5.4% State Senate % 5.4% Source: Wisconsin Blue Book, Black representation within the State Assembly and State Senate at 6.1% of the total seats in each chamber is greater than their share of the voting age population of the state at 5.4%. Likewise, one of Wisconsin s congressional seats of eight is held by a black representative. On this metric black representation in the State s congressional delegation, at 12.5%, is more than double their share of the voting age population. In terms of congressional and legislative representation then blacks are descriptively represented at levels comparable to, or above, their proportion of the voting age population. 28. Unlike Anglos and blacks where there is little difference between the voting age population and the citizen voting age population, for Hispanics in Wisconsin there is a gap. According to the Census Bureau Hispanics comprise 4.6% of the Wisconsin voting age population, but only 2.0% of the citizen voting age population. 14 Currently, there is one Hispanic member of the State Assembly which would equate to 1.0% of total membership. 15 While this figure is slightly below 2.0%, it is an indication that Hispanics in Wisconsin are achieving legislative representation roughly on par with their share of the citizen voting age population. When single-member districts are used in a representational scheme the number of seats occupied by a group is seldom perfectly equated to their share of the electorate. 29. When minorities represent a smaller proportion of the voting age or citizen voting age population of a state, as in the case of Wisconsin, it is important to examine officeholding at the local level in areas where a substantial concentration of these groups may be located. One such area in Wisconsin is the City of Milwaukee, which is home to 66.1% of the total black citizen voting age population in Wisconsin and 30.4% of the total Hispanic citizen voting age population of the State. In Milwaukee itself blacks comprise 37.7% of the citizen voting age population and Hispanics 8.8%. 16 The City is represented by a Common Council which is divided into 15 districts. Table 3 displays the number of black and Hispanic alderman currently on the Common Council. In 2012 there are 5 black alderman and 2 Hispanics which equates to 33.3% and 13.3% of the total council respectively. 17 Comparing the share of seats on the Council to the citizen voting age population figures one can see that black membership on the council is 14 Sources: U.S. Census Bureau Census National Summary File of Redistricting Data. Table QT-PL. Race, Hispanic or Latino, Age, and Housing Occupancy: and U.S. Census Bureau American Community Survey 5-Year Estimates. Table B Sex by Age by Citizenship Status (Hispanic or Latino). 15 Source: Wisconsin Blue Book, Available at: 16 Sources: U.S. Census Bureau American Community Survey 5-Year Estimates. Tables B055003B, B050031, and B Sex by Age by Citizenship Status (Black, Hispanic, Total). 17 Milwaukee Common Council Webpage accessible at: 13 Case 2:12-cv LA Filed 06/04/12 Page 13 of 33 Document 45

14 roughly comparable to their share of the citizen voting age population. Hispanics are actually overrepresented using this metric as their share of council seats is greater than their proportion of the eligible electorate. Table 3. Minority Representation on the Milwaukee Common Council, 2012 Minority Group Total Percent of Council Percent of CVAP Black Alderman % 37.7% Hispanic Alderman % 8.8% Source: Milwaukee Common Council webpage. 30. In summary, blacks and Hispanics in Wisconsin appear to have descriptive representation at levels roughly comparable to their share of the eligible electorate. This seems to especially be the case for minority representation in the State Legislature and on elected bodies in local areas where single-member districts are utilized and substantial geographic concentrations of these groups are present. D. ACADEMIC FINDINGS RELATED TO VOTER ID LAWS 31. Professor Burden s report also indicates in Paragraph 57 that Act 23 will disproportionately inhibit voting by Black and Latino voters. This opinion appears to be based in large part on a number of academic studies which examine the rates of possession of various types of government-issued identification on the part of minorities. These studies, including my own work, do find that blacks and Hispanics are less likely to possess a driver s license than white registrants. 18 Most of these studies, however, cannot fully take into account whether a registrant may possess another type of acceptable identification (i.e. a passport). These academic studies, are also not capable of determining the effects, if any, of voter ID laws on turnout. Further, none of the studies cited can answer the most important question in this case, whether the implementation of a photo ID law produces a disparate racial/ethnic impact in terms of voter turnout. Stated differently, a finding of varying rates of driver s license possession for racial/ethnic minorities does not translate, in any causal sense, into lower rates of turnout. 32. One factor that may mitigate the varying rates of ID possession among minorities is the State s launch of a program where any citizen will be granted a free state ID card for the purpose of voting. As of May 16, 2012 the Wisconsin Division of Motor Vehicles has issued 79,171 free state ID cards since Act 23 was enacted on July 1, DMV form MV3004 clearly states that all ID cards used for voting are FREE by simply checking a box on the form. 19 The inference that can be drawn then is that in just under 11 months approximately 80,000 Wisconsin citizens have taken advantage of the program launched by Act 23 and have applied for a free identification card for the purpose of voting. In addition, by examining the racial breakdown of 18 M.V. Hood III and Charles S. Bullock, III Worth a Thousand Words? An Analysis of Georgia s Voter Identification Statute. American Politics Research 36: Form MV3004 located at: 14 Case 2:12-cv LA Filed 06/04/12 Page 14 of 33 Document 45

15 those Wisconsin residents who have obtained a free ID card for voting it is clear that racial and ethnic minorities comprise a disproportionate share of this group. Table 4. Racial/Ethnic Breakdown for Free State ID Cards Issued by Wisconsin DMV Race/Ethnicity Frequency Percentage White 40, % Black 30, % Hispanic 5, % Asian 1, % Indian 1, % Total 79, % Source: Figures calculated by author from data provided by the Wisconsin DMV. 33. As indicated by Table 4 the percentage of blacks and Hispanics taking advantage of the free ID program far exceeds their share of the voting age population. This is especially the case for blacks who constitute 5.4% of the voting age population in Wisconsin, but who comprise 38.2% of those taking advantage of the free ID program. Likewise, the Hispanic share of free ID s issued, at 7.5%, exceeds their share of the citizen voting age population at 2.0%. 34. Another factor that may mitigate lower driver s license/state ID possession among racial/ethnic minorities is that fact that members of these groups may have other acceptable forms of Act 23 identification such as a passport, tribal ID, military ID, or university ID for example. 35. As to the more specific question of how the implementation of a photo ID law may affect voter turnout, and more specifically turnout on the part of minority registrants, I now turn to a study a conducted on this question in Georgia. Georgia has had a long history of state sponsored disenfranchisement of minority voters. 20 Today, the entire state falls under Section 5 of the Voting Rights Act. Georgia also has a much larger minority population compared to many states including Wisconsin. Currently, blacks comprise 29.5% of active registrants In 2007 a federal court injunction barring the implementation of the Georgia voter ID law was lifted, allowing the law to go into effect prior to the 2008 presidential election. Following the 2008 general election I conducted a study designed to answer a number of questions including: 1.) the extent to which implementation of the new law may have depressed voter turnout and 2.) the possibility that the law may have produced a disparate impact in reference to racial minorities. 20 See for example: V.O. Key, Jr Southern Politics in State and Nation. New York: Knoph, Inc.; Laughlin McDonald A Voting Rights Odyssey, Black Enfranchisement in Georgia. New York: Cambridge University Press; and J. Morgan Kousser The Shaping of Southern Politics: Suffrage Restriction and the Establishment of the One-Party South, New Haven: Yale University Press. 21 Source: Georgia Secretary of State Voter Registration Statistics at: 15 Case 2:12-cv LA Filed 06/04/12 Page 15 of 33 Document 45

16 37. Using an identified group of registrants lacking photo identification by the State in 2007, I compare turnout for this group from 2004 (prior to implementation) to 2008 (following implementation). I am able to also compare turnout for all other registrants during these same election cycles. It is critical in this type of policy impact analysis to be able to compare not only turnout rates for registrants lacking photo identification before and after implementation of the statute, but turnout rates for other registrants over the same time period. This study is the only academic research I am presently aware of that examines the effects of a photo identification statute before and after implementation. Again, the results presented are based on a series of statistical models used to estimate the probability of turnout for Georgia registrants. From these models it was possible to derive probabilities of turnout by various categories including age, identification status, race/ethnicity, and election. 38. From the model results we calculated the turnout rate for registrants lacking photo identification in the 2004 election was.54, compared to.76 for all other registrants, for a difference of.22. In 2008 turnout for registrants without photo ID was.45 and.73 for registrants with photo identification, producing a difference of.28. The difference of these differences is This measure takes into account the effect of identification status on turnout comparing the 2004 election (pre-implementation) to the 2008 election (post-implementation). While turnout for both groups dropped from 2004 to 2008, the decline was greater for those lacking photo ID. I estimate that implementation of the Georgia voter ID law depressed turnout in the 2008 general election by.43%. Stated otherwise, turnout in Georgia would have been about four-tenths of a percentage point higher had the voter identification statute remained enjoined. Most importantly for the present case, in terms of a specific racial or ethnic effect, the law did not appear to disproportionately affect minority registrants. Anglo registrants were actually more affected by the law compared to black or Hispanic registrants. The inter-election difference measure was for Anglo registrants, compared with for blacks I should also note another study I conducted that focused on black turnout in Georgia which found a sizable increase in both registration and turnout rates for this group from 2004 to For example, from 2004 to 2008 black registration increased from 64% to 78% and turnout over the same time period from 72% to 76%. These jumps in registration and turnout occurred across the same presidential election cycle that saw the full implementation of the photo identification law in Georgia. Further, it is important to note that that new registrants accounted for 24% of the total turnout for this racial group. In this study new registrants were defined as those who had registered since January 1, A sizable segment of black turnout in 2008 was attributable then to political newcomers who were able to successfully participate within the confines of the Georgia s new voter identification law M.V. Hood III and Charles S. Bullock, III. Forthcoming Much Ado About Nothing?: An Empirical Assessment of the Georgia Voter Identification Statute. Accepted for publication in State Politics and Policy Quarterly. 23 Seth C. McKee, M.V. Hood III, and David Hill Achieving Validation: Barack Obama and Black Turnout in State Politics and Policy Quarterly 12: Case 2:12-cv LA Filed 06/04/12 Page 16 of 33 Document 45

17 E. CONCLUSIONS 40. In conclusion, it is my opinion that there is little evidence to support a claim of a prolonged history of voting discrimination aimed at racial/ethnic minorities in Wisconsin. There is also no supporting evidence, in terms of turnout rates or descriptive representation, which would sustain such a claim. Finally, claims that the implementation of Act 23 would disproportionately burden minority groups in Wisconsin are based on studies of ID possession. As Act 23 has yet to be fully implemented, these studies are not capable of analyzing the effect on voter turnout. An analysis following implementation of Georgia s voter ID law failed to find any disparate racial effect. 17 Case 2:12-cv LA Filed 06/04/12 Page 17 of 33 Document 45

18 VI. RESPONSE TO PROFESSOR LORRAINE C. MINNITE S EXPERT REPORT 41. In her expert report Professor Lorraine Minnite states: Based on my extensive research on contemporary fraud in U.S. elections, I conclude that stringent photo identification requirements to vote are not justified by claims that such requirements are needed to reduce or prevent voter impersonation forms of election fraud because as the empirical record makes clear, fraud committed by voters wither in registering to vote or a the polls on Election Day is exceedingly rare. (Paragraph 2) In specific reference to Act 23, I take issue with a number of points in Professor Minnite s expert report and her ultimate conclusion offered above. 42. Professor Minnite states in paragraph 13 that she defines voter fraud as the intentional corruption of the voting process by voters. Her search for fraudulent activity then is quite narrow in my opinion as it fails to into account any election-related fraud committed in other contexts. For example, it is certainly possible for individual voters to be involved in electoral fraud in collusion with poll workers. In my own work I argue that a more expansive definition is required to study this subject which can include not only fraud committed on the part of individuals, but also other third-party entities as well such as poll workers, candidates, and political parties. A wider definition of election-related fraud allows one to search for a variety of fraudulent activities and not just fraud which may be perpetrated by a single individual. 43. Second, all of Professor Minnite s conclusions are based on reports of voter fraud. As such, she has simply relied on secondary data sources like federal court indictments or journalistic reports. Relying on reports overlooks fraudulent activity that may have gone unreported. As an illegal activity, those engaged in election fraud do not want to be discovered. I argue in my own work that in order to fully examine election fraud one must go beyond simply relying on reports of voter/election fraud and actively search for the existence of such activities. I suggest a general methodology to scientifically study election fraud through forensic techniques based on KDD (Knowledge Discovery in Databases) Wisconsin is among a handful of states that allow EDR or election-day registration. It is possible then for an unregistered Wisconsin resident to show up at the polls on election-day, register to vote, and then cast a ballot. Given this scenario it seems reasonable for election officials to be able to confidently identify such individuals, given there is no time on election-day for these cases to undergo any type of validity checks. Requiring these individuals to provide a form of Act 23 photo identification could help in this circumstance to reduce the possibility of election-related fraud. 24 M.V. Hood III and William Gillespie They Just Do Not Vote Like They Used To: A Methodology to Empirically Assess Election Fraud. Social Science Quarterly 93: Case 2:12-cv LA Filed 06/04/12 Page 18 of 33 Document 45

19 45. In summary, Professor Minnite s report is not based on any study she conducted that specifically analyzes the incidence of reported and/or unreported voter and/or election fraud in the State of Wisconsin. As such, this report says little about the incidence of election-related fraud in the State of Wisconsin and is therefore not germane to the questions at hand regarding Act In the end, whether or not past election fraud in Wisconsin can be proven is not relevant to the ability of Wisconsin to implement Act 23. Even in the absence of evidence for election fraud the U.S. Supreme Court has concluded in Crawford et al. v. Marion County Election Board et al. that the states should be able to implement reasonable requirements to safeguard against future occurrences of voter fraud. 25 In this case the Court upheld the constitutionality of Indiana s voter ID law, in part, based on this logic U.S. 181, 128 S.Ct Case 2:12-cv LA Filed 06/04/12 Page 19 of 33 Document 45

20 VII. OVERALL CONCLUSIONS 47. In this report I have produced an analysis to determine the number of Wisconsin registrants who may lack a driver s license or state ID card. In addition, I also responded to the expert reports presented by the plaintiffs in this case. My expert opinion in this matter concludes that there is not sufficient evidence to suggest implementation of Act 23 will produce any racially disparate impact on the ability of minorities in the State to fully participate in the election process. 20 Case 2:12-cv LA Filed 06/04/12 Page 20 of 33 Document 45

21 VIII. DECLARATION I declare under penalty of perjury under the laws of the United States that the foregoing is true and correct to the best of my knowledge. Executed on May 31, 2012 M.V. (Trey) Hood III Department of Political Science School of Public and International Affairs The University of Georgia 104 Baldwin Hall Athens, GA Phone: (706) FAX: (706) th@uga.edu 21 Case 2:12-cv LA Filed 06/04/12 Page 21 of 33 Document 45

22 Curriculum Vitae (5/2012) M.V. (Trey) Hood III Contact Information: Department of Political Science Office Phone: (706) School of Public and International Affairs Dept. Phone: (706) Baldwin Hall FAX: (706) The University of Georgia Athens, GA Homepage: Education Ph.D. Political Science Texas Tech University 1997 M.A. Political Science Baylor University 1993 B.S. Political Science Texas A&M University 1991 Peer-Reviewed Books The Rational Southerner: Black Mobilization, Republican Growth, and the Partisan Transformation of the American South. (Forthcoming August 2012). New York: Oxford University Press. (Quentin Kidd and Irwin L. Morris, co-authors). Peer-Reviewed Publications Much Ado About Nothing?: An Empirical Assessment of the Georgia Voter Identification Statute. Forthcoming State Politics and Policy Quarterly. (Charles S. Bullock, III, coauthor). Achieving Validation: Barack Obama and Black Turnout in State Politics and Policy Quarterly 12:3-22. (Seth C. McKee and David Hill, co-authors). They Just Don t Vote Like They Used To: A Methodology to Empirically Assess Election Fraud Social Science Quarterly 96: (William Gillespie, co-author). An Examination of Efforts to Encourage the Incidence of Early In-Person Voting in Georgia, Election Law Journal 10: (Charles S. Bullock, III, coauthor). What Made Carolina Blue? In-migration and the 2008 North Carolina Presidential Vote American Politics Research 38: (Seth C. McKee, co-author). i Case 2:12-cv LA Filed 06/04/12 Page 22 of 33 Document 45

23 Stranger Danger: Redistricting, Incumbent Recognition, and Vote Choice Social Science Quarterly 91: (Seth C. McKee, co-author). Trying to Thread the Needle: The Effects of Redistricting in a Georgia Congressional District PS: Political Science and Politics 42: (Seth C. McKee, co-author). Citizen, Defend Thyself: An Individual-Level Analysis of Concealed-Weapon Permit Holders Criminal Justice Studies 22: (Grant W. Neeley, co-author). Two Sides of the Same Coin?: Employing Granger Causality Tests in a Time Series Cross- Section Framework Political Analysis 16: (Quentin Kidd and Irwin L. Morris, co-authors). Worth a Thousand Words? : An Analysis of Georgia s Voter Identification Statute American Politics Research 36: (Charles S. Bullock, III, co-author). Gerrymandering on Georgia s Mind: The Effects of Redistricting on Vote Choice in the 2006 Midterm Election Social Science Quarterly 89:60-77 (Seth C. McKee, coauthor). Examining Methods for Identifying Latino Voters Election Law Journal 6: (Charles S. Bullock, III, co-author). A Mile-Wide Gap: The Evolution of Hispanic Political Emergence in the Deep South Social Science Quarterly 87: (Charles S. Bullock, III, co-author). Punch Cards, Jim Crow, and Al Gore: Explaining Voter Trust in the Electoral System in Georgia, State Politics and Policy Quarterly 5: (Charles S. Bullock, III and Richard Clark, co-authors). When Southern Symbolism Meets the Pork Barrel: Opportunity for Executive Leadership Social Science Quarterly 86: (Charles S. Bullock, III, co-author). The Reintroduction of the Elephas maximus to the Southern United States: The Rise of Republican State Parties, American Politics Research 31: (Quentin Kidd and Irwin Morris, co-authors). One Person, [No Vote; One Vote; Two Votes ]: Voting Methods, Ballot Types, and Undervote Frequency in the 2000 Presidential Election Social Science Quarterly 83: (Charles S. Bullock, III, co-author). On the Prospect of Linking Religious Right Identification with Political Behavior: Panacea or Snipe Hunt? Journal for the Scientific Study of Religion 41: (Mark C. Smith, co-author). ii Case 2:12-cv LA Filed 06/04/12 Page 23 of 33 Document 45

24 The Key Issue: Constituency Effects and Southern Senators Roll-Call Voting on Civil Rights Legislative Studies Quarterly 26: (Quentin Kidd and Irwin Morris, coauthors). Packin in the Hood?: Examining Assumptions Underlying Concealed-Handgun Research Social Science Quarterly 81: (Grant Neeley, co-author). Brother, Can You Spare a Dime? Racial/Ethnic Context and the Anglo Vote on Proposition Social Science Quarterly 81: (Irwin Morris, co-author). Penny Pinching or Politics? The Line-Item Veto and Military Construction Appropriations Political Research Quarterly 52: (Irwin Morris and Grant Neeley, coauthors). Of Byrds[s] and Bumpers: Using Democratic Senators to Analyze Political Change in the South, American Journal of Political Science 43: (Quentin Kidd and Irwin Morris, co-authors). Bugs in the NRC s Doctoral Program Evaluation Data: From Mites to Hissing Cockroaches PS 31: (Nelson Dometrius, Quentin Kidd, and Kurt Shirkey, co-authors). Boll Weevils and Roll-Call Voting: A Study in Time and Space Legislative Studies Quarterly 23: (Irwin Morris, co-author). Give Us Your Tired, Your Poor,...But Make Sure They Have a Green Card: The Effects of Documented and Undocumented Migrant Context on Anglo Opinion Towards Immigration Political Behavior 20:1-16. (Irwin Morris, co-author). Quedate o Vente!: Uncovering the Determinants of Hispanic Public Opinion Towards Immigration Political Research Quarterly 50: (Irwin Morris and Kurt Shirkey, co-authors). Amigo o Enemigo?: Context, Attitudes, and Anglo Public Opinion toward Immigration Social Science Quarterly 78: (Irwin Morris, co-author). Invited Publications Race and the Ideological Transformation of the Democratic Party: Evidence from the Bayou State American Review of Politics 25: Book Chapters The Republican Party in the South. Forthcoming In Oxford Handbook of Southern Politics, Charles S. Bullock, III and Mark J. Rozell, editors. New York: Oxford University Press. (Quentin Kidd and Irwin Morris, co-authors). iii Case 2:12-cv LA Filed 06/04/12 Page 24 of 33 Document 45

25 The Reintroduction of the Elephas maximus to the Southern United States: The Rise of Republican State Parties, In Controversies in Voting Behavior, 5 th ed., David Kimball, Richard G. Niemi, and Herbert F. Weisberg, editors. Washington, DC: CQ Press. (Quentin Kidd and Irwin Morris, co-authors). [Reprint of 2004 APR article with Epilogue containing updated analysis and other original material.] The Texas Governors In Texas Policy and Politics, Mark Somma, editor. Needham Heights, MA: Simon & Schuster. Other Publications Provisionally Admitted College Students: Do They Belong in a Research University? In Developmental Education: Preparing Successful College Students, Jeanne Higbee and Patricia L. Dwinell, editors. Columbia, SC: National Resource Center for the First-Year Experience & Students in Transition (Don Garnett, co-author). NES Technical Report No The Reliability, Validity, and Scalability of the Indicators of Gender Role Beliefs and Feminism in the 1992 American National Election Study: A Report to the ANES Board of Overseers. (Sue Tolleson-Rinehart, Douglas R. Davenport, Terry L. Gilmour, William R. Moore, Kurt Shirkey, co-authors). Grant-funded Research (UGA) Co-Principal Investigator. An Examination of Non-Precinct Voting in the State of Georgia. Budget: $47,000. October 2008-July (with Charles S. Bullock, III). Funded by the Pew Charitable Trust. Co-Principal Investigator. The Best Judges Money Can Buy?: Campaign Contributions and the Texas Supreme Court. (SES ) Total Budget: $166,576; UGA Share: $69,974. September 2006-August (with Craig F. Emmert). Funded by the National Science Foundation. REU Supplemental Award ( ): $6,300. Principal Investigator. Payola Justice or Just Plain Ole Politics Texas-Style?: Campaign Finance and the Texas Supreme Court. $5,175. January 2000-Januray Funded by the University of Georgia Research Foundation, Inc. Curriculum Grants (UGA) Learning Technology Grant: Converting Ideas Into Effective Action: An Interactive Computer and Classroom Simulation for the Teaching of American Politics. $40,000. January-December (with Loch Johnson). Funded by the Office of Instructional Support and Technology, University of Georgia. iv Case 2:12-cv LA Filed 06/04/12 Page 25 of 33 Document 45