J. Paul Johnston writes:

Guy-Uriel Charles' Findlaw commentary on the constitutionality of single-member districting raises some very basic issues of electoral reform with arguments that many members of the list should find provocative, to say the least.

An important point in his argument that was not overtly stressed is that the solution of moving to multi-member districts only works, to the extent that it does, because he combines it with a switch to using cumulative voting, too. It also depends on voters adopting a partisan, slate-voting approach in making their allocations of cumulated votes. Indeed, the example he offers as an illustration is, however, a "take" on a well-known actual case: that used for elections to the Illinois State House of Representatives prior to 1980. One "knock" on that experience was that the parties reached local arrangements wherein one party's stronghold was not challenged in return for like treatment or one seat was effectively "conceded" to the second strongest party by the dominant one only running two candidates. Personally, I never found the argument that one should change an electoral that had desirable features [which I see this one as having] because of a flaw in the party system wherein party elites chose NOT to compete. The Illinois example is a limited instance, however, and I'm unaware of other instances, except perhaps at the municipal level. Perhaps list members can cite some.
 
A more common combination using multi-member districts, within the majoritarian tradition, is to employ the "block vote" wherein the voter makes as many simple candidate choices as there are positions [seats] to fill. This combination actually worsens the control given to the dominant party and the degree of seats share/votes share disproportionality that occurs. That is certainly NOT the direction one would want to take to improve either minority representation or party competition.

My preference is to use multi-member districts employing three to seven members along with the "single transferable vote" procedure that employ a preferential ballot on which voters can rank any number of candidates up to the number of seats to be filled plus one. It accomplishes much the same outcome as Charles' example would yield, with even closer seats share/votes share proportionality when voters adopt "slate voting" practices in making the rankings, approximating the degree found in largest remainder party-list PR systems.

An advantage of varying the district magnitude from 3 to seven seats is that one can also better approximate the more exact "rep by pop" standards that have evolved in the U.S.A. in the "one person, one vote" tradition. And one can accommodate "local" or traditional geographically-based "communities of interest" more flexibly in drawing district boundaries.

J. Paul Johnston, Associate Professor
Department of Political Science
The University of Alberta, Edmonton, AB  Canada

J. Paul Johnston