James Allison, August 20, 2015

In Bloomington six City Council members represent six different districts, but each of the other three members, elected at large, represents the city as a whole. Every four years the top vote getter in a district becomes the council member for that district, and the top three at-large candidates become the council members for the city. This venerable system could not be more straightforward. But is it representative democracy? In particular, do the political leanings of the at-large council members reflect those of the voters who put them in office?

Not very well. The seven elections since 1987 gave those 21 at-large seats to 19 Democrats and 2 Republicans, 90% and 10% respectively. Did 90% of the voters vote Democrat, and 10% Republican? No. The Democrats garnered 119,001 votes, the Republicans 71,292–63% and 37% respectively. Had those 21 council seats been allocated proportionally by party vote, the Democrats would have occupied 13, the Republicans 8–63% and 37% of 21, respectively.

This alternative system is proportional representation by party. What difference would it have made in individual elections? No difference at all in 1987, 1991,1995, or 2007. But in 1999, when Democrats swept the at-large seats, the weakest one of the three Democrat winners would have been displaced by the strongest one of the three Republican losers. Likewise in two additional Democratic sweeps, the 2003 and 2011 elections. Applied to the seven individual elections since 1987, a proportional formula would have sent 16 at-large Democrats to City Council (instead of 19), and 5 at-large Republicans (instead of 2).

This closer approach to representative democracy by party would have come at the expense of the three Democrats displaced on City Council by three Republicans with fewer popular votes than they. But before they grumble too much, Bloomington Democrats should consider what would happen if the formula were applied to state government, where Republicans now hold dictatorial supermajority power in both houses despite near parity in the popular vote. How might that work in Indiana? Imagine ten geographic areas with equal populations, each of which elects five state senators and ten state representatives by proportional party representation.

Too many voters feel they have no personal stake in elections. Would voter turnout rise if proportional representation were seen to raise that personal stake? Recent voter turnout in Bloomington gives reason for concern: Between 1987 and 2007 the votes in at-large races averaged nearly 5,000 per candidate; in 2011, the figure was less than 3,000 votes per candidate.

Where did these notions come from? Last Spring I was in Berlin, having a tour of the city and learning something about local politics. I asked the German tour guide about city elections. “Ja, we use proportional representation.” I asked whether they had good turnout in city elections. “Depends what you mean by good!” I rephrased: What percentage of eligible voters turn out for city elections? “Ah! About 80%.”

His answer got me thinking about home.