Hillary Clinton, former secretary of state and U.S. presidential candidate, center, speaks during a small business roundtable discussion in Cedar Rapids, Iowa. (Daniel Acker/Bloomberg)

Update: On Friday, Virginia Gov. Terry McAuliffe announced that he would allow an estimated 180,000 to 210,000 ex-felons in his state to vote in the November elections. When Hillary Clinton proposed a national effort to that end in Iowa last year, we assessed how it could reshape American politics. That original story is below.

While in Iowa on Tuesday, Hillary Clinton mentioned a policy reform that could affect the results of presidential races: Allowing ex-felons to vote.

Clinton is not the first 2016 candidate to raise this issue, nor is it the first time that she's done so. Sen. Rand Paul (R-Ky.) has repeatedly advocated for restoring voting rights for felons convicted of certain crimes. At several points while she was in the Senate, including shortly after she announced her 2008 candidacy, Clinton introduced the Count Every Vote Act, which would have restored those rights to anyone not currently incarcerated or not on parole or probation for a felony. We're still early in the 2016 campaign, so it's hard to know if that's still the boundary that Clinton sets.

As it stands, people who are convicted of felonies but are on parole can or cannot vote depending on where they live, since rules on felon voting differ by state. The Sentencing Project has a handy primer on the differences. In 12 states, those convicted of a felony cannot vote even after having repaid their debt to society -- sometimes for certain periods of time, sometimes only for certain felonies. (In two states, Maine and Vermont, there are no restrictions on the voting rights of felons, even if incarcerated.) In total, some 5.8 million people are barred from voting in the United States because of their criminal past, according to the Sentencing Project's data.

University of Florida associate political science professor Michael McDonald tries to tally how many ex-felons are disenfranchised by state at his site United States Election Project. His methodology matches government data on people under correctional control with the Sentencing Project's outline of each state's laws. Moreover, McDonald collects data on voter turnout in elections by matching ballots cast with the size of the voting eligible population.

Assuming that Clinton's advocacy in 2016 matches what she's called for in the past -- namely, restoring voting rights to those permanently disenfranchised -- McDonald's data doesn't help us. He explains why: "Time-series statistics on recidivism, deaths and migration of felons are largely unavailable." In other words, the government tracks people who are in prison or on parole, but once they're free, it's hard to determine where they are or if they're even still alive. In other words, we can't know how the Count Every Vote Act would change an election.

What if a proposal went further, restoring voting rights to those on probation or parole and those in prison? What if the ineligible voters in each state were suddenly eligible? We'll note that this hasn't been proposed by any presidential candidate, and probably wouldn't be popular if it were. But it's still illustrative.

Comparing the number of ineligible felons calculated by McDonald to the voting-age population in each state shows how big a part of the state's population was ineligible to vote due to restrictions. (We're using 2012 data since we're talking presidential politics.) In Georgia, the number of ineligible felons constitutes 4.2 percent of the voting-age population of the state, the highest value nationally.

So if we gave all of those people the right to vote, how many more voters would that be? We took the overall turnout rate (again from McDonald) and applied it to the ineligible population. This is purely speculative, of course; there's a lot of variance among different populations in terms of how heavily they turnout. Younger and poorer people vote less, for example. Painting with a broad brush, then, the biggest increase in the vote would have been in Texas (given its large population), followed by Georgia and Florida.

That Florida example is interesting. If you look at the estimated vote increase versus the margin of difference in the presidential race, Florida sticks out. The number of people who could gain the right to vote and then actually vote is twice the margin of victory for Obama in the state. Florida is also one of the 12 states that bans voting from convicted felons after they're out of institutional control -- in their case, for five years. We don't know how many people that is, but it's safe to assume the figure is in the thousands.

The core question here is not one of politics; it is a question of the boundaries we're willing to put on the responsibilities of citizenship. Clinton and Paul's argument is that we've put overly tight limits on those convicted of crimes. But rolling back those limits more broadly than Clinton proposed in 2007 could have real political implications -- particularly in the always-contested state of Florida.

Knowing politicians as we do, we doubt that this didn't enter into the calculus.