I. Overview

Hispanics accounted for half of the population growth in the United States between the elections of 2000 and 2004 but only one-tenth of the increase in the total votes cast, according to a Pew Hispanic Center analysis of new data from the U.S. Census Bureau.¹

This gap between the very substantial growth of the Hispanic population and much more modest growth in Hispanic electoral clout has been developing for a generation but has widened considerably in recent years.

It is primarily the result of the two key demographic factors that distinguish Latinos from whites and blacks in the electoral arena: a high percentage of Hispanics are either too young to vote or are ineligible because they are not citizens.

As a result, a population increase of 5.7 million Latinos between 2000 and 2004 yielded only 2.1 million new eligible voters. In addition, Hispanic voter participation rates lag those of whites or blacks so that the number of Hispanic voters increased by just 1.4 million.

The combination of demographic factors and participation rates meant that 18 percent of the total Latino population (adults as well as children, citizens and non-citizens) went to the polls in 2004, compared with 51 percent of all whites and 39 percent of all blacks.

Despite these factors, however, the Hispanic population has been growing at such a strong rate that it still has led to an increase—albeit a small one—in the Hispanic share of the overall electorate. In November, 2004, Hispanics accounted for 6.0 percent of all votes cast, up from 5.5 percent four years earlier. During this same period, the Hispanic share of the population rose from 12.8 percent in 2000 to 14.3 percent in 2004.

The Hispanic population is not only much larger than the Hispanic electorate but it also differs in some key characteristics, including language usage. The share of Latino adults living in households where only Spanish is spoken is three times higher in the general population than it is among voters.

This report relies primarily on a supplement of the Current Population Survey (CPS) that is conducted every November of an election year. The CPS is the large monthly survey of U.S. households conducted by the Census Bureau and the Bureau of Labor Statistics which is best known as a source of unemployment rates. The November supplements ask whether individuals were registered to vote and whether they actually voted but do not ask which candidates or political parties they supported. Thus, the CPS data does not directly provide any information on how Hispanics voted in the 2004 election.

However, these data shed some new light on a lingering controversy surrounding the extent of Hispanic support for President George W. Bush last year. An analysis of the 2004 exit poll data in conjunction with this new CPS data suggests that Bush’s share of the Hispanic vote last year was probably closer to 40 percent than to the 44 percent widely reported last year by news organizations that had relied on national exit poll data.

Some of the major findings in this report include:

• Between the 2000 and 2004 elections, the Hispanic population grew by 5.7 million, accounting for half of the increase in the U.S. population of 11.5 million.
• Of those 5.7 million Hispanics added to the U.S. population between the last two presidential elections, 1.7 million persons or 30 percent were less than 18 years old and are thus not eligible to vote. Another 1.9 million or 33 percent of the people added to the Hispanic population between the two elections were adults not eligible to vote because they were not citizens.
• As a result of these factors, only 39 percent of the Latino population was eligible to vote compared to 76 percent of whites and 65 percent of the black population.
• Both the number of Latinos registered to vote (9.3 million) and the number of Latinos who cast ballots (7.6 million) in November 2004 marked increases of political participation over the 2000 election that were larger than for any other ethnic or racial group in percentage terms.
• However, both registration and turnout rates for Latinos were lower than for whites or blacks. As a result, only 47 percent of eligible Hispanics went to the polls compared to 67 percent of whites and 60 percent of blacks. Differences in registration rates explain most of the gaps.
• The combination of demographic factors and participation rates meant that only 18 percent of the Latino population voted in 2004 compared to 51 percent of whites and 39 percent of blacks.
• In November 2004, Hispanics were 14.3 percent of the total population but only 6.0 percent of the votes cast. In the previous election, Hispanics were 12.8 percent of the population and 5.5 percent of the votes cast.
• The gap between the size of the Latino adult population and the number of Hispanic voters has been growing since at least 1972 and is likely to continue growing given current trends.
• The foreign-born account for 56 percent of the Latino adult population but only 28 percent of the 2004 voters. As a result, 27 percent of Latino adults live in households where only Spanish is spoken compared to only 9 percent of voters.
• An analysis of census and exit poll data suggests that President Bush took 40 percent of the Hispanic vote in 2004 rather than the 44 percent originally reported from the major news media exit poll.
• Religion appears to be linked to President Bush’s improved showing among Hispanics in 2004 over 2000, when he took 34 percent of Latino votes. Hispanic Protestants made up a larger share of the Latino vote last year (32% in 2004 compared with 25% in 2000), and 56 percent of these voters supported the president in 2004, compared with 44 percent in 2000. The president’s share of the Hispanic Catholic vote remained essentially unchanged between 2000 and 2004.

The November supplements to the CPS, which are the primary source of data for this report, are generally considered an authoritative source of data on the socio-economic characteristics of the electorate. (DeSipio, 2004; Ramakrishnan and Espenshade, 2001; Cassell, 2002.) It is, for example, the only source of information on whether voters are native or foreign-born, which is a critical variable in examining the Hispanic population. Moreover, the November CPS is by far the largest national survey that provides data on the size and characteristics of the full population, eligible voters, registered voters and actual voters.

All surveys are subject to discrepancies due to margins of error and other factors. This is true of the CPS although it is a very large survey regularly conducted of the American public with an average monthly sample of about 140,000 individuals. The November election year supplements of the CPS routinely show a larger number of persons voting than the actual count. So, the November 2004 CPS reports that 125.74 million persons reported voting in the 2004 national election while the official count of votes for the 2004 Presidential contest is 122.28 million in the Federal Register. The discrepancy is 3.5 million votes or about 3 percent of the official count. The CPS supplement is taken after election day and relies on individuals’ self-reporting of their voting behavior. The difference between the CPS and the official count results from two factors: Some people report having voted when they did not, and some ballots do not get counted because the voter did not mark them properly, a voting machine misread them and other reasons (U.S. Census Bureau, 2002). The CPS covers the civilian, noninstitutional population resident in the country. Almost all active duty military (either in the United States or abroad), as well as persons in institutions (for example, nursing homes and correctional facilities), are not included.

The Pew Hispanic Center was founded in 2001 with support from The Pew Charitable Trusts. The Center conducts non-partisan research that aims at improving understanding of the Hispanic population. It is a project of the Pew Research Center.