I did a research project on Hispanic Voting in the 2008 election. I thought some people might be interested in the results. Basically, the conclusion is that anyone who said that Barack Obama was going to struggle with Hispanic voters was full of shit. In fact, a glance at the exit polls does not do justice to how well he really did. The results also appear to be correlated with the competitiveness of the state, showing what can happen when you have a ton of money and motivated people to get out the vote. Here are some of the basic ideas of my paper. If the spacing between the tables and the numbers looks weird, that is because HTML code doesn't make sense to me sometimes. There's a lot of tables so if you don't feel like reading the whole diary, just look at Table 6. The numbers will blow your mind.
Hispanics are a rising, hotly contested demographic. They have traditionally supported Democrats; however, George W. Bush made substantial progress in earning Hispanic votes in 2000 and 2004, earning 35% of their vote in 2000 and 40% of their vote in 2004. In 2008, John McCain clearly lost ground with Hispanic votes. However, most demographic groups moved in favor of Barack Obama and the Democratic Party. Thus, to determine whether a distinct change occurred specifically with Hispanic voters, those voters must be compared to the national trend. The key question researched is whether Obama’s success with Hispanic voters in 2008 is strong enough to be distinguished from the national trend to an extent that could jeopardize a long-term political goal of Republicans to reach out to Hispanic voters. The results of an examination of surveys and exit polls indicate that Hispanics voted for Obama based on unfavorable views of the incumbent administration and support for Democratic policies. As a result, the Obama campaign identified them as predisposed supporters, and they waged an exceptionally effective mobilization effort to bring them to the polls.
The states where Hispanic voters were sampled in exit polls reported by CNN in both 2004 and 2008 were Arizona, California, Colorado, Florida, Illinois, Nevada, New Jersey, New Mexico, and Texas. The most competitive of the previous states were Colorado, Florida, Nevada, and New Mexico. In addition to being the presidential election prior to 2008, the 2004 presidential election serves as George W. Bush’s best election with Hispanic voters. Because of Barack Obama’s strong success with black voters, the Hispanic vote results will be compared with the Non-Hispanic White vote.
TABLE 1
HISPANIC VOTE | 2004 | RESULTS | 2008 | RESULTS |
STATE | KERRY | BUSH | OBAMA | McCAIN |
ARIZONA | 56% | 43% | 56% | 41% |
CALIFORNIA | 63% | 42% | 74% | 23% |
COLORADO | 68% | 30% | 61% | 38% |
FLORIDA | 44% | 56% | 57% | 42% |
ILLINOIS | 76% | 23% | 72% | 27% |
NEVADA | 60% | 39% | 76% | 22% |
NEW JERSEY | 56% | 43% | 78% | 21% |
NEW MEXICO | 56% | 44% | 69% | 30% |
TEXAS | 50% | 49% | 63% | 35% |
Right away, we can see how Obama made progress with Hispanic voters. For example in Florida, the numbers have flipped. The exit polls do not distinguish Mexican-Americans from Cuban-Americans, but it is safe to say that Democrats made excellent progress in appealing to Cuban-Americans.
TABLE 2
WHITE VOTE | 2004 | RESULTS | 2008 | RESULTS |
STATE | KERRY | BUSH | OBAMA | McCAIN |
ARIZONA | 41% | 59% | 40% | 59% |
CALIFORNIA | 47% | 51% | 52% | 46% |
COLORADO | 42% | 57% | 50% | 48% |
FLORIDA | 42% | 57% | 42% | 56% |
ILLINOIS | 48% | 51% | 51% | 48% |
NEVADA | 43% | 55% | 47% | 51% |
NEW JERSEY | 46% | 54% | 49% | 50% |
NEW MEXICO | 43% | 56% | 42% | 56% |
TEXAS | 25% | 74% | 26% | 73% |
Table 2 shows that Obama made gains with Non-Hispanic voters, but for the most part they were not as dramatic as the gains with Hispanic voters. One exception is Colorado, where percentage gains with White voters was high, and where the percentage of Hispanics supporting the Democratic presidential candidate fell from 2004 to 2008. As you'll see later this isn't really an exception. Table 2 shows that Obama lost support among White voters in Nevada and New Mexico. Considering that he won those states handily, you can see how important Hispanic turnout was for his victories in those states.
TABLE 3
STATE | % HISPANIC 2004 | % HISPANIC 2008 | NET DIFFERENCE |
NEW MEXICO | 32% | 41% | +9 |
COLORADO | 8% | 13% | +5 |
NEVADA | 10% | 15% | +5 |
ARIZONA | 12% | 16% | +4 |
TEXAS | 20% | 20% | 0 |
FLORIDA | 15% | 14% | - 1 |
ILLINOIS | 8% | 7% | - 1 |
NEW JERSEY | 10% | 9% | - 1 |
CALIFORNIA | 21% | 18% | - 3 |
Table 3 shows that in most states Hispanics made up a greater share of the electorate. This trend won't likely be this dramatic in every election, but you have got to remember that the electorate is going to be less white in the future. The Reagan coalition may cease to be a winning coalition even if the Republicans rebuild it. Somehow, I just don't see them acknowledging this problem. Karl Rove's vision of a permanent Republican majority will never happen if they don't broaden their base to include nonwhite voters. Yet the base will probably reject candidates that support policies like immigration reform. I think it's a delightful catch-22.
TABLE 4
STATE | TOTAL VOTERS 2004 | TOTAL VOTERS 2008 | % CHANGE |
NEVADA | 815,880 | 943,872 | +13.56% |
CALIFORNIA | 12,255,311 | 13,286,254 | +8.41% |
TEXAS | 7,359,621 | 7,988,912 | +7.88% |
NEW MEXICO | 747,872 | 808,278 | +7.47% |
FLORIDA | 7,548,066 | 8,083,337 | +6.62% |
COLORADO | 2,102,987 | 2,236,928 | +5.99% |
ILLINOIS | 5,237,541 | 5,451,200 | +4.08% |
ARIZONA | 1,997,818 | 2,081,208 | +4.01% |
NEW JERSEY | 3,581,433 | 3,630,546 | +1.35% |
Table 4 includes the sum of votes for the McCain-Palin ticket and the Obama-Biden ticket. Sorry Nader voters, you don't matter in exit poll land and thus you weren't included. Plus I'm too lazy to count votes for every nutter that ran for president. As you can see, there was a significant increase in turnout in most of these states. Axelrod's common sense strategy was to target the message at moderate voters and to bolster resources into turning out voters with favorable views of Democrats and Obama. In the 2008 election, Hispanics were clearly such a group. GOTV may even be more important with Hispanics due to language barriers and awkward questions about citizenship. When you look at changes in the size of a voting population you have to think about who voted in much greater numbers. In 2008 it wasn't just blacks and young voters that voted in greater numbers.
TABLE 5
STATE | HISPANIC DEM ESTIMATE 2004 | HISPANIC DEM ESTIMATE 2008 | NET DIFFERENCE |
ARIZONA | 134,253 | 186,476 | +52,223 |
CALIFORNIA | 1,621,378 | 1,769,729 | +148,351 |
COLORADO | 114,402 | 177,388 | +62,896 |
FLORIDA | 498,172 | 645,050 | +146,878 |
ILLINOIS | 318,440 | 267,140 | -51,300 |
NEVADA | 48,953 | 110,433 | +61,480 |
NEW JERSEY | 200,560 | 254,864 | +54,304 |
NEW MEXICO | 134,019 | 228,622 | +94,643 |
TEXAS | 735,962 | 1,006,603 | 270,641 |
Table 5 creates an estimate of the total number of Hispanics that voted for the Democratic presidential candidate in each of the chosen states in 2004 and 2008. This estimate was created by multiplying the total voting population in the state by the percent of the electorate that was Hispanic and the percentage of Hispanics that voted for the presidential candidate. The estimate of the total Hispanic voters in each state in 2004 is then subtracted from the estimate of total voters in 2008. This is meant to illustrate the likely increase or decrease in cumulative votes for Obama in 2008 as compared to the votes for Kerry in 2004.
Now you may be skeptical of this kind of math. I was at first. The margin of error is not pretty. I obviously don't know the exact number of Hispanic voters, but conceptually the math should be valid. Keep in mind that exit polls are among the best polling data that there is minus a census. I'm also making identical types of calculations for both 2004 and 2008.
TABLE 6
STATE | T0TAL ESTIMATED % HISPANIC DEM VOTE INCREASE | TOTAL ESTIMATED % WHITE DEM VOTE INCREASE | DIFFERENCE FROM TREND |
NEVADA | +126% | +6% | +120 |
NEW MEXICO | +71% | - 7% | +78 |
ARIZONA | +39% | - 4% | +43 |
COLORADO | +55% | +19% | +36 |
TEXAS | +37% | +8% | +29 |
FLORIDA | +29% | +8% | +21 |
NEW JERSEY | +27% | 12% | +15 |
CALIFORNIA | +9% | +16% | - 7 |
ILLINOIS | - 16% | +3% | - 19 |
This is the math that explains everything. Consider a hypothetical state of 1000 voters in 2004, 5% of which are Hispanic, meaning that there are 50 Hispanic voters. Let's say 60% of them voted Democrat, meaning that there is a total of 30 Democratic votes from Hispanic voters in 2004. Now, imagine the total size of the electorate grows by 5% to an electorate of 1050 in 2008. However, imagine the composition of the electorate in 2008 is now 10% Hispanic instead of 5% Hispanic. Let's say 60% of Hispanics still voted Democrat in 2008. 10% of 1050 is 105, and 60% of 105 is 63, meaning that the number of cumulative votes held by Hispanics more than doubled in comparison to the cumulative votes in 2004. This is critical to understanding the magnitude of Obama’s success with Hispanic voters in battleground states. As you can see in Table 6, this is exactly what happened in Nevada.
Table 6 shows the estimated percent increases in cumulative votes for Obama. When both increases in turnout and in opinion are taken into account to present changes in cumulative votes, a much more remarkable picture of Obama’s appeal to Hispanics can be determined. The vote for the Democratic presidential candidate increased dramatically among Hispanic voters compared to the previous presidential election in every state measured except Obama’s home state of Illinois. I estimated that Obama increased the number of cumulative Hispanic votes by more than half in Colorado, by more than two-thirds in New Mexico, and by more than double in Nevada (Table 6). Just let that those numbers sink in for a moment. All of those states were heavily contested states that were carried by George W. Bush in 2004. Obama only needed to win those three states and the states won by John Kerry in 2004 in order to win the election, and he won those states handily.
Table 6 shows the importance of considering both vote changes and differences in turnout when assessing how well a candidate mobilized and appealed to a group. For example, in Colorado, support for the Democratic presidential candidate among Hispanics as a percentage of the vote was less in 2008 than in 2004 (Table 1). However, Obama likely increased support in cumulative votes from Hispanics by over 50%, because they made up a much larger portion of a much larger pool of the electorate. (Tables 4,5, and 6).
So one explanation for why Hispanics may have been less represented in exit polls in 2008 than in 2004 is the high mobilization of African-Americans. This does not appear to be the case in Illinois, as it was the anomalous state where the estimated number of Hispanic voters went down in 2008. However, the hypothesis remains possible in Texas. In Texas, the Hispanic vote as a percentage of the vote remained the same as in 2008 (Table 3). However, the estimated cumulative vote change was substantial, because Obama increased the margin of votes by thirteen-points (Table 1). The electorate was also much larger, and thus Obama increased his vote total from Hispanics by roughly by more than one-third in Texas despite Hispanics making up roughly the same share of the electorate (Tables 3 and 6). That's why you don't see the whole story from a glance at the exit polls. I've lived in Chicago television markets almost my whole life and I still struggle to understand politics in Illinois.
While Table 6 shows that Obama made dramatic improvements among Hispanic voters, it also shows additional anomalies. In California, Obama’s improvement among white voters was exceptionally strong, and it is only one of state in which the estimated cumulative Democratic percentage change in the vote was greater (Table 6). The other state in which the percentage increase in cumulative votes from Whites for Obama was greater than the percentage increase in cumulative votes from Hispanics for Obama ironically occurred in his home state of Illinois. There is not enough information available that could explain why this anomaly occurred, but it could be related to reduced funding from the campaign in states that were less competitive.
Republicans have a greater long-term problem than winning back lost voters. They've got to figure out how to build a new base. I don't think the construction process will be pretty.