Proportional Representation vs. the 2019 General Election

General elections in the UK are held under a "first-past-the-post" electoral system, so whichever candidate gets the most votes in a seat wins that seat. Under this system, large parties whose votes are fairly geographically concentrated tend to pick up a disproportionate number of seats, whereas smaller parties and those with more geographically dispersed support often collect few or no seats. For example, as we saw in a previous piece, the Scottish National Party, in 2019, achieved a seat for every 25,583 votes received, whereas it took the Green Party gained only one seat for its 865,715 votes.

Most countries, especially in Europe, use some form of proportional representation, an unbrella term for systems that give more weight to proportionality between votes received and seats gained by parties. We can't know what the results would have been had the election been held under some form of proportional representation, because people would almost have voted very differently in that circumstance. First, there are strong incentives to vote tactically under first-past-the-post, and secondly the actual information gathered from voters is usually different under other electoral systems, with voters being asked to rank candidates or parties, or choose multiple options.

We can, however, see what would have happened if seats had been allocated in a more proportional manner. In the maps at the top, (where we have used the hex map designed by @ODI Leeds) we have used the former European Parliamentary Constituencies (also known as "NUTS1 regions") and allocated seats in proportion to the votes cast within each of these regions. The left map shows seats simply arranged geometrically, sorted by party, within each European constituency, while on the right map seats are allocated slightly more sensibly, first trying to give the small parties the most appropriate seats, then working up through to the larger parties. If you compare proportionality between these seat allocations and votes cast, it's pretty good:

(Even the one apparently "bad" point on the x-axis is really "others" which takes in all other parties, and none of those actually gained enough votes for a seats under any form of PR.)

This contrasts strongly with the equivalent graphs for the actual results under first-past-the post, where the dramatic overrepresentation of the Conservatives and the SNP is clear, as is the severe underrepresentation of the Lib Dems and Greens.

[Update 2020-01-15:] Here is the comparison between the votes and seats won in the actual 2019 General Election and the allocation under PR used in this simulation

For more detail, and interaction versions of these and other maps, see this page on our website.

#election2019 #proportionalrepresentation #visualization #datascience

@ODI Leeds @Stochastic Solutions