At first glance, unemployed people don't appear to be the likeliest of voters. For one thing, they're often in tough economic straits, and, all else being equal, poor people vote less often than rich people.

But according to new research* by Matthew Incantalupo, unemployed people are an exception to that rule when times are tough. At full employment — that is, when the unemployment rate is at 5 percent, as it was in the 1990s and part of the 2000s — being unemployed either very slightly increases or very slightly decreases one's likelihood of voting.

But when there's a jobs crisis, that all changes. Each additional point increase in the unemployment rate raises unemployed voter turnout by 1.3 percentage points — triple the effect on employed voter turnout during such a crisis. So, starting with unemployment rates of 8 percent, about where we are today (on Friday the Labor Department reported that the nation's jobless rate had risen to 7.9 percent in October), there starts to be a statistically significant positive effect of unemployment on election turnout. That is, in an economy as sluggish as ours unemployed people are more likely to vote, not less. Incantalupo built this chart to illustrate his findings:

With extremely low unemployment rates, people who don't have jobs vote less often than  people who work. But as unemployment rises, many more jobless voters start heading to the polls than do the employed.

*h/t: The Monkey Cage