[EM] The history of preferential voting in the United States

Abd ul-Rahman Lomax abd at lomaxdesign.com
Wed Dec 31 07:15:31 PST 2008

The more I look the more I find. One the FairVote web site, there are 
very sketchy indications regarding the history of preferential 
voting; one would get the impression there of a handful of 
implementations in the past, and that they were dumped because of 
racism and the red scare or the like: this would apply, for the most 
part, only to multiwinner methods, particularly STV.

However, the more I dig into this, the more I find. Books from the 
early twentieth century are gradually being digitized and searchable. 
I just found this:


There is a wealth of information on preferential voting in the U.S.; 
this was  a report prepared for a constitutional convention in 
Massachusetts, 1917.

page 313 contains a list of 55 towns using a preferential ballot, 
with the year of implementation, up to 1917. San Francisco is included (1916).

I don't know what happened to the San Francisco implementation, but, 
of course, ultimately it was replaced with top two runoff, until that 
was replaced with RCV (three-rank IRV). Were they going in circles? I 
saw no mention of the prior use of preferential voting in San 
Francisco in the ballot arguments from proponents or opponents. The 
Santa Clara county ballot arguments called IRV "modern." Hardly. The 
book referenced covers three basic methods of preferential voting: 
the Ware system (IRV), Bucklin, and Nanson's method. Apparently at 
that time all were in use in the U.S.

It *all* disappeared. Only Cambridge was left with proportional 
representation using an STV form.

The book describes and quotes a Maryland statute (Acts of 1912, 
Chapter 2, Section 160K) which implemented three-rank preferential 
voting for Maryland "In case there are more than two candidates for 
any state office." This was a "second choice nomination law," 
governing delegates to state conventions. This was a Bucklin method, 
but like some of them, it incorporated an elimination process as 
well. (I.e., the candidate with the lowest total was dropped before 
proceeding to the next round.) (I've seen some discussion of why the 
elimination was dropped; as I recall, it was realized that it simply 
complicated the canvassing without contributing substantial value.)

Indiana, according to the book, had "one of the most complete and 
comprehensive preferential primary laws at present on the statute books."

What happened? There is now, actually, a wealth of information easily 
accessible on these preferential voting implementations, early 
results, etc. But not on what happened to them? It's like they 
slipped away without notice.

It appears that one of the factors that led to the demise of these 
systems was low usage of additional ranks, causing majority failure. 
FairVote seems to want us to conclude that this was a problem with 
Bucklin, but many of the implementations were not Bucklin, they were 
forms of STV.

There is no evidence that voters under Bucklin are more likely to 
truncate than they are with IRV. That voters under Bucklin will be 
specially concerned about Later No Harm is an obvious concern, and we 
see this argument coming from an STV advocate, back in the 1920s, 
but, just like now, there was no actual evidence, just theory -- and 
there are lots of theoretical reasons to consider that truncation 
would be a phenomenon seen with both American Preferential Voting 
(Bucklin) and the STV method. Truncation is common in Australia where 
STV is used, but full ranking is optional.

Bucklin, unlike pure Approval, does provide some protection for the 
favorite; compromise votes are considered only if there is majority 
failure. Beyond that, many voters have no motivation for additional 
ranked candidates, no matter what the method. Supporters of a 
frontrunner generally have little motivation, and, by definition, 
this is most voters in most elections!

More information about the Election-Methods mailing list