[EM] Re: British Election and Duverger's Law

Bart Ingles bartman at netgate.net
Sun May 8 09:54:44 PDT 2005

> On  6 May 2005 at 00:52 UTC-0700, Alex Small wrote:
>    [...]
>>Now, it may be tempting to explain these results solely in terms of
>>current events: Tony Blair has alienated elements of the left and
>>center, and the Tories are such an abysmal mess that even Gray Davis
>>has lost respect for them.  But the LibDems have persisted despite
>>the fact that they've been the third party in size for 80+ years.
>>I'm more surprised by their persistence over time than I am by their
>>current popularity.
>>Does anybody know why Duverger's Law has been so stubbornly resisted
>>in Britain for 80+ years?  I'd be genuinely curious to know.

If this were a single nationwide election, I'd say it was an example of 
a knife-edge equilibrium (we tend to forget that plurality voting has 
three equilibria-- two Duvergerian and one knife-edge).  But I would 
have to give precedence to whatever is happening in individual regions 
or districts.

I wonder how many districts actually have Liberal Democrats as one of 
the top two parties?  LibDem vs. Conservative doesn't seem very likely, 
but LibDem vs. Labour does in a sufficiently left-wing district (much 
like San Francisco, where the top two parties appear to be the Democrats 
and Greens).  At the district level, a knife-edge equilibrium may be 
transitional between such an ultra-liberal district and a more typical one.

Anyway, with multiple districts and the possibility of three equilibria 
in each, I don't think there's anything unexplainable about the national 

Araucaria Araucana wrote:
> Duverger's "Law" is not absolute, and I think it assumes some party
> stability and regional homogeneity.  Extracting from the top of the
> wikipedia entry (which ought to be imported into electowiki):
>     [...]
> | Additionally, William H. Riker noted that strong regional parties
> | can distort matters, leading to more than two parties nationwide,
> | even if there are only two parties competitive in any single
> | district. He pointed to Canada's regional politics, as well as the
> | U.S. presidential election of 1860, as examples of often temporary
> | regional instability that occurs from time-to-time in otherwise
> | stable two-party systems (Riker, 1982).

Australia seems a good example of this, although there doesn't seem to 
be much instability there.  The regional right-wing parties function as 
a permanent coalition (known as The Coalition) at the national level.

IRV has been shown to have same equilibria as Plurality, at least when 
there are three parties.  Is it possible that IRV's knife-edge is 
"sharper" (less stable) than Plurality's?


More information about the Election-Methods mailing list