Hello, I am a new member but have been thinking very seriously about improving US democracy for several decades now. It has long been obvious to me that first past the post single member districts electing multimember bodies has been a dubious and problematic system, for reasons I suppose everyone here well understands. It tends to create a two-party duopoly (or even single party system)–though it is worth noting that Britain’s Parliament has had third parties exist and survive, though rarely prosper, for hundreds of years. Speculation, which I suppose expert people on this site might be able to ground more firmly, is that the US Presidential system tends to tighten US votes to the two party “choice” more firmly. In any case it is rare in US representative bodies to have more than two parties and since the Civil War it has been the same two parties for the most part, with some fairly brief and local exceptions (Hello, People’s Party AKA the Populists! And some other examples, but darn few). But duopoly does not get to the root of the problem as I see it–I see it as a failure to achieve positive representation. That is, every voter should know that their vote counts toward achieving representation they want for themselves, as they see fit, even if they are in a minority in some district where they live. If someone votes for a state legislature they should be able to add their votes to other voters in the state, even those living far away, to combine forces in proportion to their numbers. If we have proportional representation then the legislature represents everyone, while under FPTP–or any system that elects a single winner and does not somehow permit the losers to consolidate and gain some representation, note that this applies to single district winner AV and IRV and ratings votes just the same–many voters are simply losers, failing to get any representation directly dependent on their vote at all. To be sure, insofar as Americans are satisfied to be either Democrats or Republicans, they are “virtually represented” when they lose in their district by other politicians from other districts, but these representatives have no constituency link to the unrepresented. Consider that in 1776, British officials and pundits rejected the notion that American Colonials were not represented in Parliament in London–in those days, very few Britons in the homeland had the vote either, and the Colonials were all “represented” in Westminster just the way voteless poor folk were, by the discretionary wisdom of their betters who took their interests into account justly, as they saw it from above. This answer did not satisfy the Patriots, nor should it have. American democracy is a work in progress, and we should not stop with solutions that seemed advanced and revolutionary in the 18th century but have since become somewhat outmoded.
Of course all around the world, other nations have adopted proportional representation in various forms, but we are often told we cannot have it here because it would not “suit” us. When someone tries to suggest it might work fine for us, we get objections along the lines of “a proportional system can work for Denmark or Belgium, but the United States is far too big; PR requires people to have faith in parties which we despise, and gives party officials power to create “party lists” that determine who is a candidate and in what order they will be elected by your party vote. You can’t vote for an independent if we have PR! What about the interests of people in particular districts; districts would be abolished or be very large. It is just un-American!”
Therefore, aiming at achieving what I call positive representation, my thinking has taken US conditions as a starting point and aimed for a system suitable to the USA. I find when I try to present these ideas, many people assume things must be true of my proposal because they are so of other PR systems, but I came at it with a clean sheet approach.
For years I have thought mainly in terms of bodies where the number of districts is exactly, or nearly exactly, half the number of seats in the body–and this resembles MMP. But more recently I have realized it is not as necessary as I thought to have this doubling, we could do it with smaller numbers.
Note that as in MMP I would retain the existence of districts. Generally when I present this, I am thinking in terms of retaining FPTP single seat districts, not because I love them in themselves, but because if included in a comprehensively proportional system they can work well. This site strongly advocates for AV, and I argue that straight AV is just incompatible fundamentally with PR, because the information about which of several candidates a voter approves is actually the single one they would most like to elect is lost unless the voter bullet-votes, and if everyone did that the system would turn into FPTP plurality. Well, if we were to modify AV so as to have a special provision to indicate which of the many choices approved the voter does favor the most, AV could then serve quite well to elect the district representatives. The important thing is for each voter to be able to say which single choice is their favorite, because that is what we need to know for a positively representative proportional system.
So in the usual form I think of it, voters vote exactly as they currently do in the USA, with certain provisions extending it a bit. I think it is important that any voter anywhere should be able to vote for any party that is running anywhere in any district relevant to the office–that is, anywhere in the city in a city council race, anywhere in the state in a state legislature race, anywhere in the nation in a House of Representatives race. (I expect to post sometime about a specific different approach to House elections to be scrupulously Constitutional, at some sacrifice in perfect proportionality and being forced to take a different tack to justify it–if the Constitutional issues preventing true national consolidation of the House votes hold as I fear they probably do, we cannot have this last extension). So added to the familiar limited candidate ballot we typically see, are some lines offering the option of voting–instead of for a personal candidate as usual–a party that is not running a candidate in one’s own district but is elsewhere. Most states I believe still continue the practice of having a write-in line on the ballot in some form (Nevada however does not; here if someone is not on the ballot you cannot vote for them–one can vote for None of The Above, but that goes nowhere and does nothing, even if a majority were to vote that it would have no actual effect, it is basically just a way of registering “I’m abstaining from voting on this office because everyone sucks.” Such closed ballots could be a fatal stumbling block for this system and so I suggest we simply mandate they not be closed, by providing for write in options which would be much more meaningful in a positive representation framework). I think there would be considerable advantage to be gained by permitting people to vote for registered (or write in) candidates in another district–that kind of vote abstains from voting on who represents one’s own district primarily, but that is a legitimate choice.
So as we are familiar, we can cast a single vote for a candidate actually running, or instead for a party that is not running a candidate here but is elsewhere, or even write in a candidate who is running very far away. In addition to providing the party options and the write in line the state or Federal government should have a central registry of all candidates and parties, and issue unambiguous identification codes for each, so that a voter can research this and write in a code to clarify exactly who and what they are pointing at off ballot.
But they get just the one vote. And whoever wins the plurality wins representing the district.
Here is the difference though. After district ballots are tallied, the state (or Congress, or the city) tabulates the totals for each party. Voting for the person in the race is also a vote for their party–unlike MMP there is no separate party choice (though the list of parties inactive locally could include the active ones, so a voter could support the party while pointedly excluding supporting their local candidate; they could do the same by voting for a party candidate outside their district). My theory here is that if one trusts an individual candidate enough to vote for them, one should trust their partisan affiliation too.
We can then quite straightforwardly calculate the proportional share each party wins of the nominal number of seats in play. I strongly favor using Hamilton’s rule for this rather than Jefferson or other methods, because Hamilton’s greatest remainder method is the most inclusive and will thus result in spanning the electorate most broadly, with the maximum number of factions included with at least one representative. The results would not be catastrophic if a more stringent and limiting method, the worst in that respect being Jefferson’s, were used instead, but Hamilton’s rule has some technical advantages in calculation as well, and is quite simple to explain to skeptical voters.
However we compute it, we then compare the number of seats each party “should” win via PR to those that were won on a plurality basis. Typically, some parties (one or both duopoly parties if people keep voting that way) have more seats won by district plurality than their share would be proportionally. So, on the theory that an extra seat for one party is a loss for another party, we multiply the total excess of seats for the favored parties over their proportional share by 2, and add this number to the nominal number of seats and thus districts in the house, and refigure the proportional share of all parties in this larger house. Typically this single step brings the larger overhanging parties up to match or exceed their plurality wins, while the shortchanged parties pick up the bulk of the added seats.
These make up seats (which will often be the entire delegation of smaller parties) are then chosen from the strongest candidates who failed to win a plurality race for that party. For the larger parties these will be generally the second place winners in their district races, for many though not all small parties, they will be candidates with perhaps ludicrously small vote counts–but they do represent other people from many districts who supported their party, and they are the strongest vote winners in their party, so there is no reason to prefer a different choice. Since they are make up winners, they do not displace any plurality winners and so questions as to their legitimacy would be muted. Such representatives will after all be few in number in the body–though with a closely balanced legislature, they could still play crucial roles.
I have considered having the make up seats chosen from among the strongest performing districts from each party, period, even applying proportional rules to have more than one added from a district where that party already won the plurality seat too. But upon reflection, each party will want to maximize its footprint across the districts.
In the version where we don’t compute the overhang of the overrepresented plurality winning parties but simply double the legislature so there is an average of two per district, it will often happen–always, if the voters overwhelmingly continue to vote for one of two parties–that a party will win more seats than they ran candidates; this must happen if we have a doubled legislature (as say electing 870 US Representatives from the 435 districts) and a party wins the majority; they can’t have run more candidates than there were districts, but now they are entitled to more than that. There would need to be a system for approving supplemental candidates in that case–I recommend developing the practice of deputy candidates, a person named by the actual candidate as their right-hand agent. In a campaign this person might become well known to the voters, much as US Presidential races deploy their Vice Presidential candidates, and normally be legally entitled to serve as the now-elected representative’s proxy in many situations–running the home district office while the representative is in a legislative session or on committee, vice versa returning to the capital while the elected official is back home mending fences and campaigning, to coordinate the capital office, handling constituency issues, perhaps being permitted to vote as the representative’s proxy on the floor or to serve on committees. In the event a party won more seats than it ran for, the top scoring candidates could “clone” themselves by nominating their deputy to also serve as a representative. They would be known to the home constituency, and having been deeply involved in both the campaign and legislative business (in a case where a deputy from a previous session is present when their party wins with large returns in their district) they can credibly be seen as equivalent to the strong candidate the people voted for and a suitable representative of the same constituency in their own right. There are other ways to go of course, such as the party having a pre-declared list the voters can take into account when considering their choices, or simply giving the leading vote winners double votes in the legislature.
These doubling issues are less likely to come up if we have an overhang-correcting top off approach of course. Just as I favor Hamilton’s rule over Jefferson’s (or even say Huntington-Hill) because it maximizes the number of parties and hence the number of voters with some direct representation, it seems plain that when there are party candidates who stood for election and lost the plurality available, they should be the first pick for filling out top off seats, because they spread the party footprint into the most districts. This is in the interest of the voters too; it is not crazy to have two representatives of the same party from one district, but it is not as much a benefit to those who did vote for that party in those districts as it is to the smaller number in the districts where they came in second to have one instead of zero.
At some point some virtual representation is inevitable; not everyone can have a representative of their own choice in their own district. But even the voters who don’t get a representative from their district have direct influence on the delegation that did get elected with the help of the votes of these outliers in “hostile” territory (I do hope for a net rise in civility and open mindedness among the representatives, but competitiveness is inherent in any democratic system when there are serious issues at stake). Therefore these minority supporters will not be completely forgotten and can consider themselves usefully engaged, and the more districts the party they do support has, the closer the nearest one will be.
So consider the simplicity of this system! Each voter can focus on simply supporting the candidate and party they most believe in, and should any bloc of voters who have hitherto supported a given party or candidate have reason to feel they are being taken for granted and shortchanged, they can jump ship and run a rival candidate. As long as their total number is great enough to elect one representative, it is worthwhile for them to volunteer, canvass, speak out, and hold rallying events, so grassroots support will be easier to drum up. The dread chill of “what if we lose?” hanging over competitive FPTP politics will be dispelled and the more festive, communal aspects of political organizing can take the lead with less fear–“we” might be disappointed not to do better, but very likely can expect some heartening wins on some scale or other.
What is the purpose of alternate choices in voting systems, if not simply to cover for the downside of one’s most favored candidate losing? If instead everyone knows they will get something out of their efforts, there is no need for backup votes for fall back alternatives. Instead, everyone shows their hands and the game is played as dealt.
It is often said it would be disastrous to have PR unless some kind of special measures are taken to prevent too much party fragmentation. I think that is bunk though. If voters specialize in dozens of boutique parties, each one narrowly gauged to a narrow range of people–that is toxic if those people were to proceed all alone to hold all power. But they don’t hold all power, just their share of it, and to get anything done in the legislature, someone has to cooperate with someone. Whichever subset of these fragmented parties meet with one another first, and negotiate mutually beneficial deals to agree to support, will get to a working majority first. Furthermore I think in reality people will join or divide on different issues, and majorities passing legislation on one topic will not be the same combinations as those who join to pass another law on another issue. Thus collegial cross-connections between diverse groups will be fostered and grow deep and strong, with representatives understanding that a friend on one issue will be a foe on another, and learning there is no profit in bearing grudges about the latter instances (depending on the nature of the issue of course; there are issues where the wrong action deserves scorn, anger, and even perhaps hatred; we can hope that majorities will tend to avoid electing people capable of really terrible decisions but of course history would want a word with us if we did!) A different model of how a legislature should operate, replacing the centuries old alternating one party dictatorships notion where “a majority is needed to form a government” (as they say in parliamentary systems where the legislature is really also the executive).
I am looking to a new kind of legislature, where the membership is from much more diverse classes and backgrounds than our system has tended to favor, grooming cookie cutter law school students of a certain limited range of backgrounds to believe they are entitled to represent everyone else, and demanding any outliers elected from different backgrounds to conform to their rules and adopt their mentalities and interests. If any group of sufficient size can drum up enough support to send their own representative to the capital, then the legislature might be much closer to a cross section of the whole city, state or nation
Or not perhaps. The representative is not the constituency after all, and the constituency is not the representative. We are familiar with the dark side of this; cooptation and betrayal are cynically believed to be practically synonymous with politics as such, and this lies behind a stance of deep suspicion and contempt for politics as an inherently dirty and cynical game. The stance that the nonpartisan is automatically more virtuous than the partisan, or that people in the middle are automatically superior in both morals and wisdom because to be an outlier is to be dysfunctional has done much damage to our institutions and bids fair to do more. But there is a bright side to representation as well. It is the job of the representative to become expert in esoteric aspects of both law and lawmaking, while networking with other representatives to form communities of interest to get useful things done, and what is most lacking is the strong tie of the representative to their constituency. Let that be strengthened, so they know they cannot throw any of their supporters under any buses and all negotiations proceed with the best interests of their various backers in mind, and perhaps it is not so necessary scientists should be represented by scientists, or schoolteachers by schoolteachers, or factory workers by a factory worker, and a professional class trained and experienced in negotiations as such might be just the thing these people would want to vote for and support. I think the healthy balance will be found with a mixture, various party representatives being a mix of professional negotiators and grass roots delegates.
Meanwhile this is accomplished with a voting system no more complicated, and hardly different than, the one we have now. People cast a single vote for what they want, the dust settles in their districts with the weird fluctuations we are used to and quite probably a lot of the professional politicos might win those races as they do now. But where FPTP fails to elect fair shares, these are taken remedially from the same mix of candidates. The math involved in calculating Hamilton’s method winners is fairly simple and straightforward, much more so than the arcane juggling of quotas and figuring out which votes to transfer from a quota-beating one in STV, and while my math background is pretty good and I am a fairly culturally knowledgeable person, I have to admit a lot of the terms and arguments I find here are complete mysteries to me. But showing that the total votes cast divided by the baseline number of seats to be filled is a quota, and dividing each party’s total by that quota, to get a number with a whole and fractional part, and say "the party gets the whole number, and then of the seats left over, the largest fractions left over get one more seat each, is methodically simple.
Thus, it is less esoteric and much easier to do, yet the fidelity of proportionality is greater. If districts are either a necessity or a convenience, we keep them, and I do think they can play a vital role. But the dark side of districts–that they can be badly apportioned (as was generally the case, invidiously against city dwellers, for most of US history) or gerrymandered, and even when well laid out in terms of fair balance of population must drift out of alignment as differential demographics raise and lower populations, shift neighborhoods, and change in general can render these districts a rather haphazard proposition. But with PR across the board, any imbalances or overconentrations are taken of naturally and automatically. No group can be gerrymandered out of relevance.