It is conventional wisdom that Democrats favor election laws that tend to encourage voter turnout and Republicans favor election laws that tend to discourage it. That's true, up to a point, argues Yale political scientists Eitan Hersh, in an elegant short article in Nate Silver's fivethirtyeight.com. But, as Hersh, who by his own account partnered with the 2012 Obama campaign on a study of the effect of the presence and absence of party identification data on the conduct of campaigns, reports Democrats systematically favor election laws that tend to discourage turnout in elections for school boards and local government offices.
Drawing (with appropriate attribution) on data compiled by Berkeley political scientist Sarah Anzia, he shows that Democratic state legislators tend to oppose holding such elections on biennial federal election days, on which turnout typically peaks, while Republicans seek to schedule them on such dates. The Democrats' motives are obvious: they want lower turnout in which teacher and other public employee unions can gin up turnout to elect candidates that will vote for higher wages, benefits and pensions for union members (and more dues money for union leaders to spread around to Democrats). "Higher salaries and better benefits for municipal employees can be a good outcome," Hersh writes, giving Democratic legislators credit for good intentions — though by using the verb "can" he leaves room for also crediting with good intentions those who believe that surging pension costs and insulation of public employees from accountability are not good outcomes. (See this American Interest blogpost titled "An Empty Windy City?" on the problems high pensions can cause.)
So Democrats are not consistently acting in accord with their process argument that high turnout is a good thing. They're pleased with low turnout when it benefits their clients/paymasters. Which is yet another illustration of my first law of politics: "All process arguments are insincere, including this one."