December 6, 2007

Hispanics and the 2008 Election: A Swing Vote?

III. Hispanic Population, Voting and Registration Trends

This section of the report presents a national and state level analysis designed to assess the possible importance of Hispanics in the 2008 presidential election. Toward that end, it provides the most recent available estimates of the size and composition of the Hispanic population eligible to vote. It also provides hypothetical projections of Hispanic registration and turnout in key states in 2008, based on the turnout rates observed in the last presidential election.

Latino Voting Eligibility

As of September 2007, Census data indicate that an estimated 18.2 million Hispanics were eligible to vote. At the time of the 2004 presidential election, an estimated 16.1 million Hispanics were eligible to vote (Table 1). Although the Hispanic electorate has been growing faster (about 13% from November 2004 to September 2007) than the electorate of other major racial and ethnic groups, the share of Hispanics among those eligible to vote continues to lag behind their presence in the overall population. In September 2007, Hispanics were 15.3% of the overall population but only 8.9% of eligible voters. In comparison, blacks were about 12% of eligible voters in September 2007 and whites were 74%. Though Hispanics were the largest minority population, they were not the largest minority group among voting eligible persons.

Since the last presidential election, second-generation Latinos have been the generational segment of the Hispanic electorate experiencing the most growth. The number of second-generation Latinos eligible to vote increased by an estimated 785,000 since November 2004, or nearly 19% (Table 2). In September 2007, 27% of the Latino electorate were second-generation Latinos. This growth pattern differs from that apparent earlier in the decade. Between November 2000 and November 2004, second-generation Hispanics had the smallest growth rate of the generations of the Hispanic electorate. Naturalized Hispanic citizens (or first-generation Hispanic citizens) had the highest growth rate between November 2000 and November 2004, but the growth of the second-generation Hispanic electorate has been more prominent since November 2004 than the growth among naturalized Hispanics. Changes in the generational composition of the Latino electorate may have implications for political behaviors, as Latino attitudes and views on issues vary along generational lines. For example, attitudes about the size of government and social issues (such as the legality of abortion and acceptability of divorce) differ by nativity (Pew Hispanic Center/Kaiser Family Foundation, 2002). Furthermore, as discussed below, the likelihood that Latinos register to vote and that they vote in elections depends upon generational status.

Although the number of naturalized Hispanic adults has increased (by more than 700,000 since November 2004), there are still significant demographic differences between the full Hispanic adult population and the Hispanic population that is eligible to vote. In September 2007, 16.6 million, or 55%, of Hispanic adults were foreign born. But only 4.7 million of them, or 28%, are naturalized U.S. citizens. So of the Hispanic electorate—those who are native born and those who are naturalized citizens—just 26% were foreign born, even though more than half of Hispanic adults are foreign born.

Another factor besides citizenship that diminishes the electoral weight of Hispanics is their relative youthfulness. More than one-third of the estimated 45.5 million Latinos in September 2007 were under 18 years of age—and thus not eligible to vote (Table 3). Given the relatively large numbers of Hispanics who were either too young to be eligible to vote or were not citizens, just 40% of the total Hispanic population was eligible to vote in September 2007. Much higher shares of the other major racial and ethnic groups were eligible to vote. Although Hispanics have been and continue to be among the country’s fastest growing populations, as of September 2007, it continued to be the case that fewer than 10% of the nation’s electorate were Hispanics.

Hispanic Voter Registration and Turnout

Census data indicate that 16.1 million Hispanics were eligible to vote in the November 2004 election and that 7.6 million Hispanics reported voting in that election (Suro, Fry and Passel, 2005), for a Hispanic turnout rate of 47% (compared with a turnout rate of 63.8% for the full population). Because the number of Hispanics eligible to vote has since risen to an estimated 18.2 million, the number of Latinos registering to vote and actually voting in the 2008 election can be expected to increase above the November 2004 levels. How many of the estimated 18.2 million Hispanics eligible to vote will actually register and turn out to vote? Registration and turnout depend on myriad factors, among them interest in the issues and candidates, mobilization and get-out-the-vote efforts, and Election Day weather. Many of these factors are election-specific and hence it is not possible to predict with any certainty the number of Latino registrants and voters in 2008. With that caveat in mind, however, one can project the potential size of Latino registrants and voters based upon past registration and voting patterns.

Table 4 projects the number of Hispanics registering to vote and voting in 2008 on the basis of Hispanic registration and voting rates in the last presidential election. By breaking down the estimated 18.2 million Latinos eligible to vote by age and generation, the projection accounts for the change in the age and generation composition of the Hispanic electorate that has occurred since November 2004.

Based on 2004 behavior, a projected 10.6 million Hispanics would register to vote, up from the 9.3 million registered to vote in November 2004. Hispanic voters would increase to 8.6 million, compared with the estimated 7.6 million Hispanics who reported voting in 2004. The projected growth in Hispanic registrants and voters (about 14%) outstrips the growth in the Hispanic population eligible to vote (about 13%) because the composition of the Hispanic electorate has shifted toward age groups and generations that had higher registration and voting rates in 2004.

The Importance of the Hispanic Vote at the State Level

Given the Electoral College system, U.S. presidential elections are won or lost at the state level, not the national level. However, given the nature of the data available, it is more difficult to gauge the size of the Latino vote at the state level than at the national level.

The 18.2 million estimate of the size of the Latino voting eligible population nationwide is based on the Census Bureau’s September 2007 Current Population Survey. The September 2007 CPS had about 152,000 respondents nationally, and the monthly CPS can not provide reliable estimates of the Latino electorate at the state level.1 For many states, an insufficient number of respondents in a monthly CPS precludes the tabulation of detailed state statistics.

Table 5 presents 2006 estimates of the size of the Hispanic voting eligible population for the 50 states and the District of Columbia. These estimates are derived from the Census Bureau’s 2006 American Community Survey. The ACS is designed to provide annual estimates of population characteristics at the state level and for geographic areas with at least 65,000 residents.

The ACS indicates that an estimated 17.9 million Hispanics were eligible to vote nationwide in 2006.2 In nine states the Hispanic share of the state’s electorate is about 10% or greater. Hispanics’ weight in the electorate is greatest in New Mexico, where an estimated 37.1% of the electorate was Hispanic in 2006.

Table 6 reports the Republican margin of victory in each state in the 2004 election—that is, the difference between all ballots cast for President Bush and Democratic challenger John Kerry (expressed as a percentage of the total ballots cast)—and compares it with the Hispanic share of eligible voters in that state and also with the projected Hispanic share of voters who will hypothetically turn out to vote in that state in 2008. It also shows the percentage of Hispanics who voted for Bush in 2004 (in states where there was a big enough Latino vote in 2004 to permit such estimates from national election day exit polling).

As Table 6 shows, there are only seven states where the estimated Hispanic share of the state electorate exceeds the 2004 Republican margin of victory: New Mexico, Texas, Arizona, Florida, Colorado, Nevada, and Iowa. However, Hispanic turnout rates nationwide have historically been lower than non-Hispanic turnout rates. For example, in the 2004 election it is estimated that nationally Hispanics accounted for 8.2% of eligible voters but only 6.0% of all votes cast (Suro, Fry and Passel, 2005). There are no reliable estimates of the 2004 Hispanic turnout rates in individual states, however. In the absence of such data, Table 6 assumes that voters in a given state turn out to vote at the same rates that voters turned out nationally in the 2004 election. That is, the projected Hispanic share of a state’s voters in column 3 is derived by applying a uniform 47% voting rate for Hispanics and an overall voting rate of 63.8% for the full population of eligible voters.

In six states — New Mexico, Florida, Colorado, Nevada, Arizona and Iowa — the projected Hispanic share of voters who go to the polls, as shown in column 3, exceeds the 2004 Republican margin of victory. In 2004, President Bush received at least 30% of the Hispanic vote in each of these states, according to the 2004 state exit polls (as shown in column 5). Past Hispanic presidential preferences provide some guidance as to how much Hispanic voter party preferences could hypothetically change in 2008. The recent high-water mark of Latino support for a Republican candidate was 40% in 2004 when George Bush defeated John Kerry (Suro, Fry and Passel, 2005). The recent low point of Hispanic support for a Republican nominee was 21% in 1996 when Bill Clinton defeated Bob Dole and Ross Perot (Leal, Barreto, Lee, and de la Garza, 2005). And in the Congressional elections of 2006, GOP candidates received an estimated 30% of the Latino vote nationwide, according to exit polls.

On the other side of the partisan ledger, there are three states—California, New Jersey and Wisconsin—where the projected 2008 Hispanic share of the state vote exceeds the 2004 Democratic margin of victory.

Note on the Appendixes

Appendix A consists of two tables presenting 2006 estimates of the size of the total population and the voting eligible population. Appendix Table 1 presents state level estimates. Appendix Table 2 presents estimates for the 435 congressional districts. Estimates are provided for Hispanics and all persons. Tabulations of the Hispanic share of the voting eligible population are shown. These estimates are obtained from the Census Bureau’s 2006 American Community Survey. The ACS surveys about 3 million households per year and is designed to provide annual estimates for places whose population exceeds 65,000.

  1. numoffset=”3″ The Census Bureau does publish some aggregated voting measures at the state level using the CPS (U.S. Census Bureau, 2006). The state measures are not broken down by race/ethnicity nor in any detail.
  2. The 2006 ACS estimate of 17.9 million Hispanics eligible to vote is consistent with the November 2006 CPS estimate of 17.3 million Hispanics eligible to vote reported in Table 2. The ACS is based on 12 monthly surveys and the estimate is centered on the midpoint of the calendar year or the month of July. The difference between the two estimates largely reflects the difference in the coverage of the two surveys. The CPS covers the civilian noninstitutionalized population. The ACS covers the entire resident population.