The paper Cruz cited is one in a series of papers in which we have combined public records – from departments of corrections, state courts, and secretaries of state – to directly measure ex-felons’ party registration and turnout history. But our research does not support Cruz’s claim.
Senator Cruz makes a common mistake when characterizing the ex-felon population: he equates ex-felons with violent criminals. But only a small share of ex-felons are convicted of violent crimes. This is both because most felons are convicted of non-violent crimes and because those who are convicted of violent felonies serve longer sentences.
Here are some data we’ve collected from Iowa. We have not yet used these specific data in our research but it is particularly helpful because it includes information about ex-felons’ type of offense. As the table below shows, only about 12 percent of the felons discharged from supervision between 2002 and 2012 were convicted of a violent crime.
We can also examine the partisan affiliation of ex-felons who were convicted of violent or non-violent crimes to more directly test Cruz’s claim. Cruz is correct that people convicted of a violent crime are more likely to identify as Democrats than as Republicans:
But people convicted of violent crimes aren’t more likely to identify as Democratic than non-violent offenders. And even more ex-felons register as independents or with a minor party. As Amy Lerman and Vesla Weaver argue in their book, “Arresting Citizenship,” contact with the criminal justice system demobilizes ex-felons and makes them less trustful of traditional political institutions, including political parties.
Cruz was on stronger footing when, later in the interview, he dropped his focus on violent crimes and claimed instead that “convicted felons tend to vote Democrat.” In the paper that the Cruz campaign referenced, we do find that the majority of ex-felons in New Mexico, New York, and North Carolina register with the Democratic Party.
This fits with what you might expect, given that some Democratic-leaning groups are over-represented among ex-felons, including African-Americans, the young, and the poor. In a seminal book, “Locked Out,” Jeff Manza and Chris Uggen predict that, based on demographics alone, 70 percent of ex-felons would identify as Democrats if eligible to vote.
But in some states, a different pattern holds. In Maine and Rhode Island, as in Iowa, we find that a plurality of ex-felons register as independents or with minor parties. These are states with fewer residents from these Democratic-leaning groups, either in the ex-felon population or overall. For example, in Maine, only 6 percent of the ex-felon population we analyzed was black, while in North Carolina, 55 percent was black.
This leads to a broader problem with Cruz’s claim. He implied that being a felon makes someone more Democratic. But the data suggest that the prevalence of Democrats among ex-felons may have to do with basic demographics, not their criminal record.
Consider Florida, where well over one million citizens are unable to vote due to a criminal conviction, more than in any other state. Although Florida has typically refused to give ex-felons the right to vote, from 2007-2011 it had a different policy and restored the voting rights of approximately 150,000 ex-felons. We matched these ex-felons to publicly available voter records to compare their partisan affiliations to those of the general voting population:
The table below shows that African-Americans overwhelmingly identify as Democrats whether or not they are ex-felons. (For registered voters, we focus on those who registered in 2001 or later.)
Among ex-felons who are black males between the ages of 18 to 29, 72.8 percent identify as Democrats, 4.5 percent identify as Republicans, and 22.7 percent identify with a minor party or as an independent. The same is true among registered voters who are black males age 18-29: 73.6 percent are Democrats, 4.5 percent are Republicans, and 21.9 percent are independent or support a minor party. This similarity between ex-felons and registered voters holds for all of the other combinations of race, age, and gender.
The table also reveals what Cruz also did not mention: although non-black ex-felons are slightly more likely to support the Democratic Party than their demographic counterparts among registered voters, a substantial percentage identify as Republican as well. For example, 33.5 percent of ex-felons who are non-black males over the age of 45 registered with the Republican party, only slightly less than the 37.6 percent who registered as Democrats.
The implication is that denying ex-felons the right to vote disproportionately affects potential Democrats, but it affects a lot of potential Republicans too.
The Sentencing Project estimates that, as of 2010, Florida denied the right to vote to 432,839 black ex-felons, or 23.3 percent of the black voting-age population. But a majority of disenfranchised ex-felons in Florida — 890,521 or 7 percent of the state’s non-black voting-age population — are not black.
This basic fact is why there is actually a debate between Uggen and Manza and political scientist Traci Burch about whether George W. Bush or Al Gore would have netted more votes if ex-felons were allowed to vote in the 2000 presidential election in Florida. In ongoing work, we are investigating the difficulty of trying to administer post-sentence disenfranchisement as there is in Florida. The practice, which is common in the South, often conditions voting rights on the payment of legal financial obligations.
Ultimately, we hope that our response will provide some helpful context for Cruz, whose comments perpetuate a narrative that helps explain Republicans’ reticence to reform laws that disenfranchise ex-felons. Most ex-felons were not convicted of a violent crime, nor is the ex-felon population “overwhelmingly Democratic” in every state. Even more importantly, most felons and many ex-felons — almost 6 million citizens — cannot vote at all.