The I.R.S. has been collecting information about the political party identification of people who don't pay their taxes in Colorado. Apparently they have been using voter registration lists to track down dead beat taxpayers and gathered the political party affiliation of registrants in the process.
While I've never seen direct figures, I suspect that Republicans are far more likely to be significant tax cheats than Democrats. Republicans, on average, make more money and the vast majority of taxes are paid by those in the top 20% of household income. More importantly, Republicans make up a grossly disproportionate share of the classes of people most likely to cheat on their taxes -- the self-employed and people with investment income that is not simply from financial instruments issued by large institutions. In contrast, many higher income Democrats are salaried educators, professionals and managers with large organizations, or are, like attorneys, subject to particularly strict third party reporting regimes. The generally anti-tax attitude of the Republican party (particularly the attitude that you owe the government nothing for your own prosperity), probably contributes to a propensity to cheat as well.
Still, since Republicans and Democrats are both so numerous, as are the unaffiliated, it is hard to believe that much significant action arose from those designations. Of greater concern is whether people who were flagged as delinquents with third party identifications, such as the American Taxpayer's Party or the Libertarian Party might be singled out for treatment as tax protesters, against whom the I.R.S. likes to take a hard line, as opposed to run of the mill non-payment cases.