WI: Alternate US House Rep Elections - Proportional?

Assuming a slightly different US Constitution, such that the member of the House of Representatives are voted on by all the voters in each state, without the states being broken up into districts that vote for a single seat, what would be the likely consequences of this, and how the voting work? Each voter would be able to vote for more than one Representative, (assuming their state had more than one) but how many, and would this result in a more proportional representation, or would it make it so that individual states would have all or nearly all of their Representatives be of a single party?

And if this doesn't result in a proportional representation system, how could the Constitution be modified so that the elections are more or less proportional, at least for the states that have more than a single Representative? (Either through changing what it was originally, or through amending the above version.)

Also, how would this change the Electoral College, if at all? Would it become more proportional?
 
I'm not sure how the Electoral College could be proportional; there is only one person being elected.
It could have the electors voting based on the proportion of votes in the popular election are, so a single state's electors would be split between the various candidates, instead of all of them going to the overall winner in the state, sort of like how Maine and Nebraska do it. (Though they do so by district, as did a couple states originally, before winner-take-all became the norm.) Also, before the 12th amendment, the Vice President was the runner up, instead of being a separate vote...
 
the problem under the constitution of choosing electors proportionally under the Constitution is that it makes it much more likely that election go to the House on a 1 state 1 vote basis
 
It could have the electors voting based on the proportion of votes in the popular election are, so a single state's electors would be split between the various candidates, instead of all of them going to the overall winner in the state, sort of like how Maine and Nebraska do it. (Though they do so by district, as did a couple states originally, before winner-take-all became the norm.) Also, before the 12th amendment, the Vice President was the runner up, instead of being a separate vote...

OK. That can be done, even OTL; it's up to the states to decide how they wish to award electoral votes. I think a federal law requiring this would be unlikely though.
 
That system doesn't sound proportional, but I admit it's popular in this board. Let's say there's a state with three representatives. It's not inconceivable that the three candidates with the most votes are all from the same party, even if another party's candidates get a similar number of votes. It's also possible that votes for candidates in the most popular party are spread out between enough candidates that the less popular party would gain two or even three seats from that state. If you also have any of the many systems designed to make the votes proportional within the state, then you have the problem of rounding errors. If the state is evenly divided between two parties (as most states are), whichever party has slightly fewer votes would have twice as many representatives in the state. Political parties would focus heavily on elections in states with even numbers of representatives.
The two ways to avoid this are to treat the entire country like one giant district (I wouldn't recommend this for a country this size),or to use mixed-member proportional representation. This means that several candidates are elected on the district level and and several more are added in to make the party balance match the party balance of the electorate. I don't know if either of these systems existed before 1900.
 
As far as I can tell, there wasn't any sort of proportional system at the time the US Constitution was drafted, so whatever system ends up being used, would likely have to be invented by the Framers...
 
Assuming a slightly different US Constitution, such that the member of the House of Representatives are voted on by all the voters in each state, without the states being broken up into districts that vote for a single seat, what would be the likely consequences of this, and how the voting work? Each voter would be able to vote for more than one Representative, (assuming their state had more than one) but how many, and would this result in a more proportional representation, or would it make it so that individual states would have all or nearly all of their Representatives be of a single party?

And if this doesn't result in a proportional representation system, how could the Constitution be modified so that the elections are more or less proportional, at least for the states that have more than a single Representative? (Either through changing what it was originally, or through amending the above version.)

Also, how would this change the Electoral College, if at all? Would it become more proportional?

I had a long reply going, long ago, that was lost in a system crash.

Let's leave the Presidential race aside for the moment; it really deserves a different thread after all.

As far as I can tell, there wasn't any sort of proportional system at the time the US Constitution was drafted, so whatever system ends up being used, would likely have to be invented by the Framers...

Already in the Framer's era there were people thinking of the idea of proportional representation; John Adams made some remark to the effect that such a system would be best; he may have said that late in life after the Constitution was adopted. So it isn't entirely ASB that it might have been, or put in with an Amendment later.

The big stumbling block to such a system is the concept that the House should be refracted through the framework of the States. This is the fatal error of the particular system you suggest, for instance. A certain number of states will be apportioned with just one Congressional District under our current system, and in these states any proportional system boils down to simply selecting the front-runner. If there are more than three candidates running then the front runner, with the plurality, is not even guaranteed to have a simple majority, yet there is only one candidate to be chosen, who better than a front runner holding say 45 percent of the vote? That doesn't sound so bad but what if the front runner only holds say 35 or 25 percent in a very fragmented race? How representative can such a system be? Whereas if there were many seats up in that electorate, a fragmented distribution can be fairly represented with a proportional distribution. Consider that in the more usual case where just two parties dominate and the front runner does hold a majority, say 52 percent, that still means that 48 percent of the voters lose completely, getting no representation at all. This is the situation that currently holds for all voters in Congressional races across the nation, with a winner take all vote district by district. If we take a two party situation to be natural and normal, it doesn't seem so bad, because the losers in one district are likely to be winners in others, and we assume that the national proportions roughly match the number of districts each wins. But this is only approximately true. There is a tendency for the nationally dominant party to wind up with proportionally more seats than they ought to by proportional assignment.

I believe it is not legally necessary to have a Constitutional amendment to mandate the system you propose; Congress could vote it in tomorrow, and require the 2018 election to be conducted as you suggest. But it would not seem fair, and I believe the courts would strike it down, even though Congress does have the power to control the conduct of its own elections, on the grounds of equal protection under the law.

Consider if I live in a single district state like Vermont or Wyoming. If I don't support the single candidate most likely to get the majority vote, I am effectively throwing my vote away, except in a symbolic sense. Making the bigger states proportional does nothing for me, and making a strict proportional rule simply removes the possibility of requiring a front runner with a mere plurality to go through a runoff election if they don't get a majority, since any proportional rule automatically designates the front runner the winner no matter how small their plurality is.

But if I live in California, with 55 House seats assigned, and all of those are assigned proportionally, I can back a candidate who can only hope to win less than 2 percent of the statewide vote, and still expect to elect someone to the House! My individual options, along with every other California voter, are more more flexible. If I live in Texas, they are almost as good--but not quite! The average sized state has 8 or 9 House seats and so voters there (there are 4 8 seat states and 5 9 seat ones) has better options than the single seat state voters, but worse than the bigger states. No one is close to being really equal in their electoral power. The courts will strike it down unless there is an Amendment enforcing this, and no majorities to favor it will come close to passing such an Amendment. Even the Californians won't like it. Say my 2 percent party can form in California, around some issue we all feel passionately about, and reliably elect a House Representative. From our point of view, since there are other people all across the country who feel just as strongly as we do about this issue, we ought to be able to elect 7 or 8 Representatives, not just one! In fact the issue might be something that is even more popular in smaller states, yet we can help our allies in these little states only by electing our one token Rep, when in fact a nationally proportional system might enable ten or twenty to be elected. We would feel robbed, not empowered, and so would voters for very popular parties nationally that happen to be less popular in their one-House Representative state, such as Democrats in Wyoming or Republicans in Vermont.

Making the races proportional state by state is bad, unpopular, and on grounds of equal protection UnConstitutional. As long as the election of the House is shackled tightly to election from separate states, and not conceived as primarily a vote of all the American people for a House representing them as a unified nation, proportionality cannot be introduced and the least unfair system is the current district by district, winner take all system that puts everyone in the same boat. We all suffer the same dilemma minorities face in single representative states--winner take all and if you happen to be in the minority, tough.

Most proportional systems I know of either aggregate the entire populace represented by a legislature to one big pooled race where parties compete for vote share and send delegations chosen nationally to the legislatures. Or sometimes they form multi-member districts of say three or 5 or maybe 10, and proportionality applies only within these aggregated districts.

I have another proposal though, as far as I know not actually done but I think worth a try.

Keep the districts, but have half as many as there are seats. One way to apply that to the USA today, or in say 1868, would be to simply double the size of the legislature while keeping the old districts. Or keep the current size and go through all the hassle of reapportioning with half the number and redistricting districts twice the population.

Let's say we've done this for 2018; the next Congress will have not 435 but 870 seats! (Or 871 to make it an odd number and discourage ties; we can easily sneak in more seats without harming the principle at work). We allow everyone in each district to vote for one candidate as before--if a party winds up getting more than half the total national vote, they can name a list of additional candidates to be brought in. A vote cast for a party candidate counts both as a vote for that person and for their party.

Now then--if a candidate in a given district gets more than half the vote in that district, they have probably got more votes than the number of national votes cast divided by total House seat number, since we now have twice as many seats. (Probably partially because the single-member states include many that have less population than the average district--but it is in the nature of our Constitutional setup to ignore the advantage this gives these small states, so we ignore it here. More significantly is the matter of turnout, which I may or may not get to before posting this). Therefore anyone getting a majority in a district automatically wins as before. Plurality may prove to be a dangerous gray area if the vote winds up split across a great many parties. Anyway instant win for actual majorities is surely workable without compromising proportionality, at least not much, except in cases of gross disparities in voter turnout.

Meanwhile, the party votes across the nation can be simultaneously totaled up, and the 871 seats assigned according to their proportions by any of several widely used methods. These range from the d'Hondt/Jefferson method, which tends to favor the most popular parties and scant the weakest, to the Hamilton method of Largest Remainder which I favor for several reasons, including that it tends to be most generous to the smallest also-ran parties. More common than these across the globe I think is the Lague/Webster method which is similar to d'Hondt's iterative method except it uses a different divisor incorporating the number of seats each faction has gained at each point in the iteration. The American names associated with each by the way highlight the fact that while proportional seating of the legislature was not considered seriously in the Framers' time, the problem of fair apportionment of number of seats of the House to the states came up immediately and was solved in these three ways, with a fourth used today, by such luminaries as Thomas Jefferson, Alexander Hamilton, and Senator Daniel Webster--all of these American usages pre-dating the European sages that they are more commonly named for!

Whatever method we use, definite numbers of House seats can readily be assigned based on national aggregate votes for the various parties. Having done that, if we have assigned seats already according to district majority victories, we subtract these from the quotas each party has collectively earned, and strike the seated individual candidates, now district representatives, from their respective party lists. The parties should each have a remainder of seats left to them that total up to 436, and ranking the yet unseated candidates within each party according to how many votes they individually got, the remaining seats go down the list in that order, so that the delegation for each party shall include the most popular candidates from among them. It may happen, indeed certainly will if one party gets more than 50 percent of the total, that there will be seats left over for a popular party, and to fill these each party will have been called upon to publish, before the election, their system of choosing additional candidates to take these places, so voters may judge them accordingly.

Vice versa, if we were to have large variations in voter turnout between the district, some parties might win more seats by winning majorities in their districts than their total national proportional share. There would be several approaches to dealing with this, but if it is an ongoing and major distortion, it will probably be pointing to some systemic evil such as voter suppression that should be dealt with aggressively outside the electoral system.

On the whole though, if the system develops the dynamic we would expect of favoring more than two parties, the latter problem will not emerge if we don't assign seats for mere plurality victories, voter suppression is combated, and districts are sized properly. For the former issue, especially if we are relaxed about which districts Representatives come from, the parties might run two candidates instead of one in each district; this would tend to automatically create the surplus list for a strong party.

Such a system would not guarantee that a particular district always sends two and only two representatives--indeed if we don't make a rule saying whoever wins the plurality automatically gets a seat (which could lead to gross distortion of the proportionality of the House as a whole) there is no guarantee any district will send anyone to Washington at all! Odds favor their sending someone, but a district that is idiosyncratically very fragmented with its fragments not aligned with strong national parties does risk having no Representative whatsoever. The voters in the district will be fairly represented, on a partisan level if not by a person from their district. But it is possible this means they have no representative at all, if they cannot put at least half their votes to one person and the parties they support are unpopular outside their district--they go unrepresented because they failed to make an alliance with anyone outside, and failed to cohere to a favored candidate among themselves. This would presumably be a rare situation; more often a fragmented electorate is supporting factions that are strong outside their district and therefore help that faction win votes, which might result in elevating one of their own district to office despite getting anemic numbers of votes, if the winning party they are a member of did well nationally. Vice versa every district can guarantee sending someone from there to Washington if over half the electorate there favors one candidate, and even if it is a bit less than half a strong plurality is likely to enable a candidate to win on the basis of the national party vote and being high up on the list in terms of individual votes received.

I would worry that this system, which I think a very good one, would be opposed on the grounds that each district is not guaranteed equal representation. I would counter argue though that under our current system, districts are not all alike; every state has a different number of citizens comprising a CD ideally, and in reality between imperfect district boundaries that can't quite equalize population, and shifts in population since the last census, no two CD even within a state really have the same population anyway. Systematically, the smallest states have far fewer people voting for a CD than the average. Whereas under the system I outline, every voter is exactly equal, for their votes are all thrown into one national pool. And yet, should they desire to organize their campaigns around regional or even single-district issues, they have a framework for doing so, which is not the case with more traditional national pooling proportional voting for legislatures. The system I propose has a maximum of flexibility, combining the virtues of both district and proportional systems, while tending to check the vices of each.

In particular, I like it because if the districting system is abused, to an extent the harm done is automatically negated. This is especially true if we eliminate all guarantees of a district having a representative. For instance, we might have a terribly apportioned system of districts whereby a big city comprising half the state population is assigned just one district out of 5. Then the candidates in the other 4 districts have a very easy bar to become Congressmen, if we allow that majority winners always are deemed to have won a full voting share--4 candidates in this case can win 50 percent plus one vote in their sparse districts and still not have won as many votes put together as someone winning a similarly slim majority in the underrepresented city! Letting the rural faction party take 3 of those 4 seats is a clear injustice that risks denying other parties their fair share of seats. But even if the system is corrupt or mistaken enough for that to happen, the city vote is tremendous. Not just the front runner but the next several rivals are likely to poll absolute numbers putting their parties up nationally just the same as if they had been voting in a properly apportioned district, and the individual candidates, having won from an improperly huge electorate, will dominate their respective party's lists and guarantee a voice for the city in several parties. From this bully pulpit of a multi-party caucus in the Congress, the city dwellers can denounce the felonious misapportionment of their state and demand action to rectify it--or perhaps not, perhaps they will prefer their situation and let it ride?

In general, gerrymandering is going to be frustrated with a system like this, for whether disadvantaged groups are corralled into single districts to try to contain their influence or scattered among many to dilute them below electoral power threshold, either way if they vote, their power will emerge in fair proportion one way or another. It depends on whether they want to organize around opposing systemic discrimination, or ignore the invidious intent of hostile parties and vote according to individual issues. Either way, their power is not blocked.

Guaranteeing a representative from each district, or even from each state, can make gerrymandering more effective to be sure. But even with that risk in place, I think it is a much superior system to the one we have now.

Could it have passed Constitutional muster? Not I think without a bit of Amendment--we'd have to change the system whereby Congress handles elections not resolved in the Electoral college for instance. Would it pass political muster? I'd say, not before the resolution of the Civil War and establishment of equal citizenship rights for former slaves and the adoption of equal protection under the law. Until then, the issue of state's rights would be too entangled with slavery to allow national aggregation of votes.
-----------
Before posting this I want to say also that while to keep the language simple I have been talking about "parties" here, I also have suggestions for how to set the system up to enable end runs around parties. These innovations would also be needed to avoid certain chokepoints in our current system. If there is any interest I would explain further.

Discussion of choosing Presidents is another topic.
 
Let's take a look at what would have happened had the system I suggest had applied to the results of the latest Congressional election, based on the Wikipedia summary page.

The outcome under current rules is 55.4 percent of all seats, 241, being Republican and 44.6, or 194, going to the Democrats. Note that the two big parties together polled 96.6 percent of the votes, which means that 3.4 percent of the voters got nothing, enough votes to have qualified for almost 15 seats!

The dang page does not list a total number of all votes cast. I was going to average together estimates based on dividing totals for each party by the percentage fraction given, but strangely all the minor parties handled this way give totals smaller than those for the two big ones, so just averaging them, I get 128,923,499 votes cast, rounding up to the nearest unit. Dividing by 435, that gets a quota of 296,378 votes per 435 seats. By that criterion, 4 parties--Libertarian, Independent, Green, and Conservative, should have gotten at least one each, and the next two parties down, Working Families and Independence, should have had a possible shot depending on the exact system of proportionality used. Cutting the quota in half to account for my proposal to double the House size and add 1 for an odd 871 total, interestingly only the "Misc write-in" category moves up to a quota and of course they would not count being highly fragmented, but the next party, Constitution, might have a shot.

Looking a bit more, even considering only votes for Democrats and Republicans, the Republicans had only enough for a clear 219, while the Democrats had enough votes for 215, and their remainder is larger so they would get 216 in proportional systems. Thus the Republicans currently have at least 22 more seats than they should have proportionally while the Democrats come up short those same 22--the margin between them in the House should be just two, at most three, not 47! To be fair, some decades ago when Democrats dominated the House the margins were similarly biased in their favor. But clearly if third parties had their fair share of weight, neither big party would control the House and they'd be forced to negotiate some kind of caucusing coalition to assign control to one side or the other. Recall this is true with OTL voting results in which voters went in knowing that votes for any other party but the Ds or Rs would be "thrown away," as indeed over 4.38 million were last year!

Now then--first of all let us apply the Hamilton method of Largest Remainder to the current results, for 435 House seats--that is, we ignore the whole question of Representatives from a district or state, and just used the districts to administer elections mainly by party. Hamilton's method is to divide the votes by the quota and first of all assign the whole number part of the quotient to the parties. Here we have 296378 (and a fraction):

Republicans 211 seats, 251358* votes remainder
Democrats 208 seats, 107295
Libertarian 5 seats, 218214*
Independent 2 seats, 283579 *
Green 1 seat, 218145*
Conservative 1 seat, 44,481
------------------------------
435- 428 leaves 7 remainder votes which go to the largest remainder parties, including those not getting any quota seats
I marked them with asterisks and added below three others that beat the remainders of quota-seat parties:

Working Families 278171 *
Independence Party 199130 *
Constitution Party 129822*
It seems a safe bet none of the misc write-ins beat this last party!

The totals then are

Republicans 212
Democrats 208
Libertarian 6
Independent 3
Green 2
Conservative 1
Working Families 1
Independence 1
Constitution 1

Observe how, by seating at least one person for all parties going down to Constitution Party, the number of votes not counting toward any representation in Congress is thus reduced from 4.38 million to something like 345,000! Out of nearly 129 million votes cast that is just a quarter of a percent, so 3.15 percent of the electorate completely shut out of representation (by their own choice to be sure) now have effective power, which is nearly 93 percent of those shut out by our current system.

Of course, I would expect if we had such a system that a lot more than under 4 percent would vote for a third party, and the resulting subdivisions of votes might raise the unrepresented fraction up considerably. It would be nice to know. Even so, the outcome would be superior I think in that people could vote for someone they judge closely represents their interests, rather than being the lesser of two evils.

Applying the same methods for an 871 seat House,

Republicans 424
Democrats 417
Libertarian 12
Independent 6
Green 4
Conservative 2
Working Families 2
Independence 1
Constitution 1
Legal Marijuana Now 1
Reform 1

This adds two more parties, comprising well over a third of those left unrepresented by a House the current size.

Note how proportionality is applied if we first seat all the candidates who got a plurality in their races--presumably the case with the Republicans and Democrats seated in the OTL house--since only those two parties won any seats, let's consider only them. The number of additional seats each of these gets is

Republicans 183
Democrats 229

This offsets the disproportional results of the OTL Congress.

Note that no one party controls 436 votes, needed to control the Congress unilaterally. The Republican controlling margin is now more than eroded away; they need 12 more votes. Presumably they can count on the Libertarians, alternatively there are 5 members (from Conservative, Independence, Constitution and Reform) I judge will support them before the Democrats, while the Democrats can count on only perhaps 7 from the Greens, Working Families, and the Marijuana Party I guess. With all six Independents on their side, the Democrats still fall short of a majority with those other 7 on their side too. Whereas the Republicans can govern as long as the Independents, the presumptive right wing parties, and at least one Libertarian support them, or with all the Libertarians don't need any other party.

Make no mistake, the vote was pretty right wing last year.
------------
As I remarked before, the earliest I think such a reform could gain traction sufficient for an Amendment necessary to implement it would have been the late 1860s. Then of course the Republican party had strong control over the majority of votes of the nation and would not like to see it eroded. OTOH a lot of the third parties forming in the generation after that were offshoots of the Republicans, whereas implementing this reform would tend to weaken "the Solid South" particularly if representatives from a district were guaranteed only in the case of a full majority of votes winning it. Since the South was repressing a big part of its potential voters, whereas the Census counted the African-Americans as citizens in apportionment, under winner-take all the Southern electorate (often further restricted by literacy or property requirements) could command a larger share of votes than their actual voting population warranted. If the proportional system were adopted, the Democrats would pay a big penalty by excluding a big portion of potential voters from the polls. If of course these voters were sure not to vote Democratic anyway, the party gains more than it loses by continuing Jim Crow restrictions--but it still loses relative to regions that allowed all adult males (after 1920, all adults) to vote. Other voter restriction dodges such as denying the franchise to ex-felons (especially in the context that African-Americans tended to be convicted a lot more than whites, either for the same crimes or by attaching heavier penalties to "black" versions of essentially the same crime, as with draconian penalties applied to "crack" cocaine in the '80s but not to other ways of taking the same drug) would have a similar effect, as would the simple mass terror applied by the Klan and so forth keeping African-American voters away from the polls even if no legal impediment existed. The Jim Crow parties could still control winning their districts, but the low turnout numbers would penalize the party on the national scale. Note that since the surplus, national draft Congress members are chosen in order of number of votes they individually got, Northern Democrats would be favored since Southern ones would labor under the restriction of excluding black numbers and thus have a smaller white electorate to vote for them. This offsets the "Solid South" advantage of being able to count on all whites to vote for that party, especially in states with very high black population like South Carolina, Mississippi, and Louisiana. Thus, the advantage tends to go to the regions and parties that favor full democracy. Perhaps bearing this in mind, sufficiently many states could back it to wind up the law of the land sometime between 1868 and 1916.

Do we really want a Congress of 871 Representatives? Some suggest making it much much bigger still! I am not so sure. It would probably be better to keep the current size, 435, but cut the number of districts in half, to 217, and apportion those. I just chose the bigger House to illustrate the principle most simply. It should be an odd number, and I would favor having one more "at large" seat than the total number of districts.

Such a House, since it is impossible to assign half the seats definitely to being "from" any particular state (or rather, the obvious method of just observing what state the surplus Representatives happened to run in would give very arbitrary results) would require a change in the rules for Congress selecting a President when the Electoral College fails to deliver a majority. I'd assign it to the Senate instead, with the Speaker of the House elected by the new House being a tiebreaker if needed.
 
For the Presidency--it is necessary to focus on the question, do we want to go on selecting one individual with indivisible powers, and maximize the impression that the President is chosen virtually unanimously by all the people, or do we want "all the people" to be able to express nuances of disagreement and somehow try to put the power of the Presidency on a sliding scale to encourage coalition formation?

In the former case, it is simply necessary to pick one, and the current Electoral College is pretty good at that; it hasn't deadlocked since the modern concept of popular vote even became relevant, and it usually selects whoever got the most popular votes too. Since it sometimes fails at the latter task (which I think should be required of any President, that they have more popular vote than anyone else) I favor something like the National Popular Vote Compact being lobbied for in several crucial states at this time, which undertakes to guarantee that whoever gets the most PV gets over 270 EV no matter what, by obliging a set of states holding that many or more EV to cast their EV for that candidate, regardless of outcome within their own states.

Another approach is to simply get rid of the Electoral College and rely on the Popular vote directly, say using transferable votes so voters can indicate second, third and so on choices their votes are transferred to as their earlier choices are eliminated, until someone has a majority.

I've thought of "sliding scales" to encourage the President who is elected to work with other candidates and perhaps give failed candidates certain powers (an office I thought might be called "National Tribune"). In particular I have thought of linking the veto power to the number of popular votes and giving a President the option of recruiting his rivals in order to command an increased veto power. And resolving shortfalls of Electoral votes, instead of going to Congress, by allowing holders of EV to bargain with them and thus the smaller winners serve as kingmakers for the big candidates. I've been thinking of a 9 EV system for instance, whereby the public votes for one as OTL, the 9 EV are assigned proportionally and a virtual Electoral College exists whereby the candidates are the Electoral College in person. Some candidate who wins 5 EV of the 9 wins and is President, but their 5 EV necessary to win are each worth 12 percentage points of veto override--that is, with just 5 EV, the President's veto can be overriden by more than 60 percent of the votes in House and Senate, not 2/3. The remaining 4 EV are good for 5 percent each. The rival candidates who hold them can agree to join with the President or not as they choose, and if the President persuades all of them it takes 80 percent of each house to override the combined veto. If a very popular Presidential candidate can win 6 or 7 or more EV in their own right they get this added veto power, on the same 5 percent per EV basis.

On the other hand, if no one gets 5 EV, it is possible for candidates to combine their votes, under the rule that whoever has the most popular votes in the coalition becomes President and this decision is irrevocable. The kingmaker who puts someone with 3 or 4 EV other the top with their 2 or one does not get to be President but they do get to decide whether to join their vetos to the President; with only 4 EV in their own right, a President has only 48 percent which is to say, essentially no veto at all.

Other options such as privileges to bring a rival candidate who won some EV into the Cabinet and so on could also be legitimized.
 
A certain number of states will be apportioned with just one Congressional District under our current system, and in these states any proportional system boils down to simply selecting the front-runner. If there are more than three candidates running then the front runner, with the plurality, is not even guaranteed to have a simple majority, yet there is only one candidate to be chosen, who better than a front runner holding say 45 percent of the vote? That doesn't sound so bad but what if the front runner only holds say 35 or 25 percent in a very fragmented race?

Well, the solution can be a simple two round race, take the first and second most voted and make a second election two weeks after the first, the winner take the seat, if any candidate win 50%+1 of the votes in the first election you cancel the second poll.

________________________________________

In Brazil we have the proportional system, and there are problems with the system, the two biggest are that candidates with very few votes are elected because other candidates have many more than needed, and the second problem is that the cost of the campaigns skyrocket.
 
Do we really want a Congress of 871 Representatives? Some suggest making it much much bigger still! I am not so sure. It would probably be better to keep the current size, 435, but cut the number of districts in half, to 217, and apportion those. I just chose the bigger House to illustrate the principle most simply. It should be an odd number, and I would favor having one more "at large" seat than the total number of districts.

It seems to me that they that they should make it larger. According to the Constitution, a House district should have only 30,000 residents, and yet there now are over 700,000 residents per district. The number of seats has stayed at 435 for about a century even though the population has tripled since then. It must be difficult for Representatives to keep in touch with their constituents when they have so many. I recognize that they can never get it back down to a 30,000:1 ratio, but they could lower it somewhat.
 
It seems to me that they that they should make it larger. According to the Constitution, a House district should have only 30,000 residents, and yet there now are over 700,000 residents per district. The number of seats has stayed at 435 for about a century even though the population has tripled since then. It must be difficult for Representatives to keep in touch with their constituents when they have so many. I recognize that they can never get it back down to a 30,000:1 ratio, but they could lower it somewhat.

It is a question of balance between "knowing your representative" and "do the Representatives in the House know each other?" Indeed it probably would not make a tremendous difference if the US House were increased in size to say 701 or even 1001 members, in terms of personal relationships formed in the give and take of House politics over two or more years. Clearly it would make some. The bigger the body is, the more bureaucratic it becomes. Also of course the higher the cost of maintaining it--paying not only member salaries and perks but paying for their staff, for the costs of frequent transportation across the nation, junkets overseas and so on and so forth.

If we were to kick it up to say 1001, more than doubling the size, doubling basic administrative costs, and with an imponderable change in the character of House business due to sheer size, at the same time we slash back the size of the districts (recalling that I'm proposing a number of districts just under half the size of the body, in this case 500, or in the case of the current winner take all system, 1001. The latter increase, with the reapportionment of number of districts to each state and the need to draw new smaller districts for all states (except the single district ones, but those will be made far fewer, perhaps just 2 or 3 would be left) will slash, by more than a factor of two, the number of constituents in each district and thus raise the probability of "knowing your legislator."

I'm not sure how important that is though. The major impediment to knowing your Congressman now is lack of will to do so. In my life, I've lived places where the Representative of the district I live in is of a different party than I support, and I have not chosen to regard them as "my friend." Thinking of them as the enemy, I don't even want to get to know them. Vice versa when they happen to be of the party I support, on my own hook I probably would leave them alone to do their work without bothering them. It so happened that the nature of the person I used to live with for 15 years sent us on a course of a certain mode of political activism whereby I did get to meet my Representative--due to post 2000 Census redistricting, suddenly Lynn Woolsey of Marin County, a woman I had admired for some time before the boundary change. Meanwhile another Democrat replaced the Republican Frank Riggs, a particularly despicable individual by the accounts I gave credit to, and the Democrat who took over was also someone I liked--but did not get to meet in person. In fact my charmed life with this lady I lived with got me face to face with US Senator Barbara Boxer whom I had supported for years without meeting.

If it were key to my personal values to "personally know my Representative" the course I know to take would be to get heavily involved as a volunteer in one's party political campaigns, and if it were the Representative alone I cared about, obviously to join and actively support their campaign. Now if everyone wanted to do this in the district it would clearly become impossible, even if the district comprised only 30,000 people. How many people reading this thread can name 30,000 people they know personally? The reason I could get to know my Representative if I wanted to is that relatively few people want to, so it isn't too hard to do them a service such that they will meet with you. Another way is to become involved in some issue with some political electricity in it--then you can, if you are the kind of person who can be friendly with a political enemy, get to know them even while adversaries.

But again, if everyone got politics fever and cared intensely about a personal bond with their representative, even 30,000 is too large a district. In practice, people deal with intermediaries, and if they can trust these to be decent channels as well as filters, with information and power flowing in both directions--then the number of people a single Representative can deal with goes up and up and up, and the size of the legislative body can go down. If one is going to govern a nation of 300 million people, a certain vastness has to enter somehow or other.

Related to the idea of proportional representation, I think the dynamic of communication with "one's own" Representative would change considerably. Right now, the ideal expressed in some civics texts anyway is that once elected, a Representative serves all the residents of their district, regardless of whether they did in fact vote for them or would seem likely to. Therefore if my Representative is a Republican, I still should feel I could go to them if I had some kind of issue or complaint regarding the Federal government.

That might be so if I wanted to resolve an administrative mix-up, or had an issue that I knew was close to the heart of my Representative's interests. But in general I think of it as hogwash. If I have a Representative from a party I don't like, I think of them as the enemy, as the probable source of any problems I feel the government in Washington is giving me, and that taking any help from them would be a betrayal of my own party and its values. What i want is a Representative I can call my own, whose values are aligned with mine, who is fighting for and not against what I think is most important. And a neighbor of mine who is of opposite political views to me ought to be able to have their own Representative, that they trust who is fighting for what they think is important.

Therefore I don't want to share my Representative with that neighbor. Because I don't want their success to result in my going without any champion in my state capital or in Washington at all, I can't morally wish that fate on my neighbor, wrong-headed though I think they are. We should both be able to support someone we like, and let these elected champions of ours do the negotiating and fighting in the halls of government.

Therefore I don't think it is important that "my Representative" be someone who lives within an easy drive of where I do, or that I can drop in on and hang out with. Indeed not; I need them in Carson City or Washington DC, working for me! I don't care if my federal Representative even is from the state I live in now, or any other state in particular; I can easily see my causes being better championed by someone who lives in a different state than someone who lives in the same town with me.

The dynamic of "knowing my Representative" would come about, on a low level, through media reports of what they do and my scrutiny, as much as I care to, of what they've accomplished for me. If I feel moved to help, getting involved in campaigning would be the way to strengthen the bond. With proportional representation, this activity is less likely to be a gamble risking complete failure and more likely to result in a bond to an ongoing, working organization. If I am a well informed member of an open and responsible political organization, I can be well represented by someone I share with a million or more of my fellow citizens--as long as we are a group all on the same side!

Or so I think. It is refreshing and thought provoking to see a response from someone who lives with a real PR system. Apparently the grass is indeed always greener on the other side of the fence!

Well, the solution can be a simple two round race, take the first and second most voted and make a second election two weeks after the first, the winner take the seat, if any candidate win 50%+1 of the votes in the first election you cancel the second poll.
Yes, this could be done. This is a "runoff election." We have them here, in some places, for some offices. I believe it is up to the state to decide whether to require actual majorities as opposed to pluralities. It is I believe never ever done in any state for the US Presidential race. I don't know if any states require it for Congressional district races. I've never voted in a runoff election myself however, not that I recall. If I ever did I took it in stride as normal. But in my experience, which has been only in California from 1984 to 2002, and Nevada from 2004 to now, they are rare. It might be quite different in other states.

We can go one better. The idea of an instant runoff election, aka ranked preference voting, seems pretty good to me. If we had that here in Nevada, I suppose I would either be voting for a third party first, or the Democrats to start with. And here's what I would expect to happen in my own Congressional District:

1) I, and perhaps a fair number of other people, but not near majorities, vote for some third party as my first preference. Neither the Republican nor the Democrat wins with a majority.
2) the way preference voting works is, I cast a ballot with all the choices listed in order. Perhaps I have the option of refusing to vote for some choices at all, which would mean if my preferences were exhausted before the vote is settled, my vote vanishes as though I did not vote at all. Given what the Republican party has mutated into in my lifetime, I can't see myself casting a vote for them, even at the bottom of the list. For state Governor, a lot of people say our current guy Sandoval is a moderate and deserves some support--but I guess I am an extremist; on that race I would never vote for him or anyone else with an R behind their name.

Since I'd be voting for some wacky extreme leftist third party first, one no one expects to come close to 50 percent support, my candidate would presumably be low down on the list of first choices. Besides the Republicans, there are a bunch of other parties I deem just as toxically right wing as the Rs and would never list on my ballot if I had the choice, or anyway would push down below the Democrats.

3)One party, or probably some random write-ins (there is no way to write in votes on Nevada electronic ballots though; if they don't appear on the ballot one can't vote for them though one can vote for "none of the above") will be smallest. It will be eliminated first, and its voter's second choices revealed and moved to their designated candidate. In my case, perhaps there are 2 or 3 other leftist minor parties I list ahead of the Democrats. If, when the sweep of elimination of small candidates reaches mine, my second and third choices are already eliminated, the machinery tosses them aside and moves on to the first one I list that is not eliminated. Let's say that's some 4th leftist party I didn't like much but preferred to the Democrats this year. My vote gets moved to them. If the process has not yet pushed one candidate past the 50 percent plus one line, that left party is probably going to be eliminated in its turn, and my vote at long last lands in favor of the Democrat.

4) But, in my district, before the Democrat could be pushed over the line, I expect the Republican to win. Or perhaps there will be an upset and it is the Libertarian (one of those right wing parties I won't vote for) who crosses the line.

You see, by districting, all but one of Nevada's districts are alloted to subdivisions of Clark County, where Las Vegas is. That's where most of the people of this state live after all! All of Northern Nevada, which is to say the vast majority of the area of the state, is one Congressional District. It includes Washoe County where I live, which has Reno and Sparks, and Carson City which is the state capitol--Democrats are quite competitive in these districts. But it also includes nearly every bit of every other County in Nevada outside Clark/Las Vegas, and the majorities in the small rural counties are strongly conservative. No Democrat has ever won this Congressional District in Nevada history.

So when push comes to shove, the majority of people in my CD are going to vote Republican, or conceivably some other right wing alternative party, but never Democratic or farther left. That is, a whole lot of us will so vote, but we will lose, even if majorities are required and a mechanism exists to allow us to rank our preferences to consolidate our votes onto a leading candidate. Our votes vanish into the ether, in that race, and so it is hard for a Democratic candidate for Congress in our district to get any support. I've met several of the unsuccessful candidates over the years.

Runoffs and the requirement of a majority are good ideas if we accept winner take all. If we have proportional representation they are unnecessary.

In the case of enabling my proposal for a largely and in principle perfectly proportional system based on districts half the total seats to be won in number, I would indeed want to insist on some sort of runoff or ranked voting system (ideally the latter) to determine who is deemed the individual winner of each district. It creates at least the illusion that the single winner enjoys a solid majority mandate of their district, and helps individuals who want to regard their geographical Representative as being "theirs" have confidence the winner of the instant runoff (or regular runoff held some weeks later) will listen to them.

But it has nothing to do with "proportional outcomes!"
________________________________________

In Brazil we have the proportional system, and there are problems with the system, the two biggest are that candidates with very few votes are elected because other candidates have many more than needed, and the second problem is that the cost of the campaigns skyrocket.

Well, indeed it seems you see the grass greener on our winner take all side of the fence than we do here! Just as I envy you the proportional vote opportunity you have there, though it seems a mess to you. Obviously at least one us has some nasty lessons to learn, and I suspect it might be me alone.

Still, having taken a glance at the form of the elections for your Chamber of Deputies, and an article mentioning the outcomes of the latest election, you don't have "the" proportional system, you have a proportional system, one similar to the one the OP suggested for the USA--better in some ways, perhaps worse in others. For instance every state in Brazil is guaranteed to have at least 8 Deputies, and the maximum for the biggest state is capped at 70. We in the USA have no upper limit, and the lower limit is 1. Brazil is a smaller nation than the USA, but has a larger Chamber than our Congress--512 Deputies versus our 435 Representatives. We'd have to expand our House to about 768 Reps to match your average ratio of population to Deputy. If we did that, California would have roughly 27 more House seats, or 80, which would exceed your cap by 10. Meanwhile if we apportioned 768 Reps the way we currently do, the smallest states would still only have 1 Representative each; there would be no way to guarantee a large enough number per state to make proportional voting within the small states really meaningful unless we were to multiply the political power of the residents of small states even more than our current system already does. Meanwhile if we applied the cap on largest delegation your system has, the people of these biggest states would be shortchanged Representatives even more than they already are relative to the small states.

With these considerations in mind, of course you would expect distorted results. I was struck for instance by two parties in the Opposition Coalition "Change Brazil" listed one above the other; Partido Trabalhista Cristão with 338,117 votes gets 2 Deputies, but Partido Trabalhista do Brasil with 812,206 votes gets only 1! Farther up, Partido da Mobilização Nacional has 467,777 votes but gets 3 Deputies, while right below it with only 2 Deputies Partido Ecologic Nacional got 663,108 votes! These oddly disjoint outcomes happen only at the bottom of the electoral barrel to be sure; the bigger parties seem to be more or less in line with the national vote totals.

I suppose these odd discrepancies among the minor parties happen because the "proportional" votes are cast state by state, and also because of the massive distortion that comes from on one hand capping Sao Paulo at 70 deputies when by proportional population it should have 110, shortchanging its voters by some 40 percent, and the rule of a minimum of 8 deputies per state which in the extreme case of Roraima, which ought to have just one deputy, multiplying their voting power by a factor of 10 or so.

All of that would be solved if the nation were to adopt full proportional elections on a national level, every citizen voting the same ballot; then the party allocations ought to closely correspond to actual percentages of votes received. My hybrid district/national allocation system would have a similar result, leveling the votes of all citizens to one level and giving each citizen the same power no matter where they lived. Unless of course the results were distorted by parties that do well in either gaining majorities or winning in instant runoffs or later runoff elections for the district Deputy position, shortchanging other parties that tend not to win majorities district by district. With Brazil's nearly 30 vote-getting parties I don't think there is much danger of that happening though!

However, the remaining discrepancies due to Brazil's system being of only compromised proportionality don't seem to be what bothers you.

Perhaps I don't understand your first big issue at all. How is it that some candidates have "too many" votes and this results in Deputies winning with very few, if this is not simply the result of parties that do well in the overrepresented states?

It is in the nature of any proportional system that some candidates are elevated to win a seat or two with very low votes cast for them. In my proposed system for the USA, parties are alloted a certain number of House seats based on overall national vote numbers, but every candidate (assuming no party gets over 50 percent, which would be a change in the USA though one I would expect with PR opening the door for small parties) who is placed in office is someone who ran for Congress in one of the districts. They would be placed, in the simplest and most proportional system, in their seats based on ranking by number of votes they individually got. This means pretty much by definition that some party candidates will get an excessive glut of votes, and these superstar winners will take the first seats, while enhancing the total numbers of seats their partner party members can take. Pretty much by definition the last Party member to take a seat will be coming in on the coattails of some big winners in their party, and will themselves have won rather few votes. I don't see anything improper about that; they are supposed to be a team and their total numbers, rounded out by the small vote winners, give their faction the weight the voters assigned it in the Legislature. Of course there is the possibility that having won office, some of these might defect; the ultimate remedy of that is the voters who judge the composition of the party/alliances that form for the next election and those after by the past behavior of their members. The law might possibly provide some alternative shorter term remedies too, with legally binding pledges by the candidates to some rules of discipline.

From what I read at the Wikipedia page, the Brazilian system is based on party lists; in a state with say 29 Deputies each party optimistically presents the voters with a list of who they would put in of up to 29 chosen party members in good standing with good histories, if the voters were all, 100 percent, to vote for them. Since there are close to 30 parties I suppose it is rare for any party in any state to get even a third of their chosen members seated, and less than that nationally. But in Brazil, how do you know which persons got too many votes, whatever that means, and which got too few? Are the people in each state voting for individuals as well as party lists? That's pretty much exactly what my proposal would have happen, in that in voting for one individual in one's own district one is also voting for their party.

So perhaps if you gave an example of what you mean, your objection would be clearer. As it stands, it sounds to me like the way a proportional system might be expected to work and that would be good.

-----------

As for campaign finance...well, that is a surprise to me, especially in view of the fact that there doesn't appear to be any superstar frontrunner party getting even close to half the votes, let alone more than half. Looking at the results of the 2016 election it seems the Worker's party, with 13 1/2 million votes and 70 Deputies, is the largest, and that is only 14 percent of the total! I would not be shocked or frightened if the US system evolved that way, with our biggest party being somewhere between 10 and 25 percent of the vote and a few dozen parties circling around to make shifting coalitions. One would think that if money buying votes, by various means from 19th century type machine politics (Boss Tweed and so on) to the power of advertising, were a big thing, that just a few big dominant parties commanding between them well over 60 percent of the vote would tend to result. The outcome I see, with well over two dozen parties scrapping it out, looks to me like what i'd expect if the voters are not voting for large nebulous coalitions of broad and vague ideas masking very focused lobbying efforts, but rather they are picking people who pledge detailed platforms that several dozen categories of voters approve of in detail. They are voting in the people they want, and letting them do the negotiating--exactly what I hope for here.

Now clearly if there are a couple dozen parties vying for seats, the accumulated total of all their campaign budgets would tend to add up to a pretty penny I guess. But consider how expensive campaigns are here in the USA where there are only two parties to speak of. We've cut the number of competing factions by an order of magnitude, but in the age of TV, campaigning has become very expensive for each candidate.

To get a place in the current US Congress, it is necessary to pass the bar of a winner-take-all election, and to keep it, to do that again and again every 2 years. Assuming the population actually eligible to vote (or who would be were it not for peculiar local restrictions) to be say 220,000,000 (out of 300 million) or about 73 percent, that is just over half a million per average district, and even with fantastically high turnout rates, the actual votes cast per election will be somewhat less. It would seem . under current political conditions it can be a lot less; recall only about half the theoretical electorate actually voted for US congress last year!

Something to remember about US Congress races though is that only some of them are very competitive! Typically, if someone is elected once, first of all their own party does not typically contest their re-nomination to run again on their own ticket. Even in a district that is pretty evenly divided, once a Congress member has gotten that seat they tend to be pretty difficult to dislodge. Add to this that the outcome of the political tug of war around drawing district boundaries tends to create "safe" districts for both parties, so that only a minority of them are evenly divided to begin with, and the typical reality is that being re-elected, once elected the first time, is much less of a struggle. In turn this tends to depress turnout since the race is not exciting and the parties tend not to want to spend money on energizing an electorate to no purpose.

Thus, a 50 percent average turnout probably reflects an even lower level of turnout in many nominal "races" offset by pretty high (by US standards which are scandalously low) rates in the competitive races. Note that this past year was one of major and volatile change, in which both major parties' Presidential primaries were characterized by a really drastic restlessness involving bitter accusations by leading candidates against the "swamp" and other charges of inertia and corruption. In these conditions the Democrats were pushing hard to regain control of the House in anticipation of Clinton winning the Presidency, while the Republicans were insurgent and aggressive on the ground too. In other words, 50 percent average turnout with even lower depths of apathy in "safe" races perhaps falling as low as 1/4 voting in Congressional races was a "good" year for US citizen participation!

Under these circumstances, it is necessary, in a highly competitive House race, to galvanize a partisan contingent of perhaps 100,000 hardcore reliable voters to augment themselves enough to neutralize and, by slim margins (for large margins would indicate spending too much money on that race!) just barely overcome the opposition to win one seat.

Now supposing we had some kind of proportional system--would I expect costs to go up or go down?

Up on the whole, I suppose. One of my hopes for a proportional system would be to galvanize interest in the outcomes so that most people allowed to vote would choose to do so, and of course to remove all obstacles that should not be there so that everyone could vote. This probably means that at a stroke the number of people involved is doubled! That in turn doubles the margin of victory needed...or does it?

After all current races are all winner-take all. No matter how we do it, something close to half of those who do gear themselves up to vote at all wake up disappointed with nothing concrete in hand the next morning. With a proportional system however I think one must choose a very marginal long-shot lost cause indeed in order to emerge from the process with nothing at all! The electoral fight is not so much about being in the legislature at all, but about gaining margin, or staving off losing it.

Secondly, much of the energy of US campaigns goes into vilifying the other side. This obviously also makes politics a toxic subject. Either one is blindly and narrowly partisan and tries to structure one's whole life around associating with one's fellow partisans, or else the division is uncomfortable when dealing with people who may or may not be on your side, or whom you know are against it.

Perhaps you can comment on the balance of negativity versus positive advocacy of specific platform items in Brazil's somewhat proportional system. i would think that with 20+ parties there is little to no point in trying to attack and tarnish the brand of a dozen or so that are close to one's own party position. To an extent I guess maybe there might be more petty bickering since it is easy to split, and then necessary to justify the split to the voting public. But in terms of deep values, there would be little to gain even by attacking parties diametrically opposed to one's own on policy issues. With the voters having so many choices, presumably the ones who are likely to vote for something that opposes every important platform of one's own have zero interest in your party, nor do your own voters seem likely to migrate over to it--they'll go sideways, to the most similar one to yours that differs on some subtle issue or other.

Rather than attacking others, it would seem necessary to identify oneself and stress what one's own party is for. There might not have to be a goal to grow membership and votes; it could be that a particular faction judges it has reached all the people who are interested in its core interests and they merely want those people to show up, vote, so they can retain their minor but potentially influential place. Or, a party that is interested in growing must put out feelers to learn what it is that voters are most likely to care about and frame itself as addressing those issues creatively, and somehow or other offset any impression of insincerity this causes.

US politics today tends to rely very heavily on media, on creating and placing political advertisements on TV and nowadays on the internet. These ads in turn tend to be attention grabbing as they must be, for everyone tries to filter out ads since they tend to be obnoxious across the board. And this usually involves vicious negative attacks on the other side.

Do Brazilian politicians rely on media? And if they do, are the ads negative or positive? Can the parties invest in presenting themselves in an attractive light while ignoring the competition, or is my guess that proportional elections would tend to discourage negative attack ads incorrect?

If a lot of money is being spent on campaigns in Brazil, because the many parties each are running in a sense against all others, is more of it than in the USA spent on community organizing events? If candidates propose to get out and meet their potential electorate, if they make themselves public and show up at events where people can ask them questions in detail and propose their own concerns and notions, that I think is actual democracy at work! Especially if the race is not competitive in the sense that only one candidate will wind up in power, with all others shut out, and it instead a matter of whose influence grows and whose shrinks, then the time spent meeting with any candidate is probably not wasted from the public view. I would expect such events cost some money--but it is much better spent than money spent on media advertising. To an extent, such contacts are part of the governance process, not in any sense frivolous at all.

Here in the USA, what I would hope for if we adopted some system of Proportional choosing of Congress is that politicians become focused on learning their electorate and supporting it.

The idea that American races should be more competitive, not less, is common to many reformers. Under our current system, it is not just Congress seats that tend to be "safe" and uncompetitive, it is across the board a problem that few races are meaningfully open. In the Presidential races every 4 years, most voters are largely bypassed by expressions of political concern because they live in "safe" states for one side or the other--it is known in advance that these states will cast their Electoral votes on one side or the other. Attention swoops in on the "swing" states and partisans in the safe states are told to simply write checks for financial contributions. If a Senator is up for re-election, they probably enjoy the advantage of incumbency and opposition might give that race a pass. And so on; races are sometimes hotly competitive, other times deemed safe and boring.

Therefore a number of reforms, such as the National Popular Vote Pact, are meant to inject more competitiveness into the races. By guaranteeing whoever wins the popular vote becomes President for instance, the notion is to spread campaigning effort across the whole nation since any vote anywhere in the Union could be the "swing" vote that puts one or the other over the top.

More competitiveness does not necessarily mean multiplying the sorts of expenses that are current in US campaigns, because I think the role that would be played by expensive media campaigns would saturate pretty fast. At the shallow level of mere image, trashing the leading parties' images with innuendo and slime would at best start a revolving door whereby minor parties rise to take the place of yesterday's major ones, only to be slimed to death in their turn. If that is the target audience, the voting public, doesn't simply tune out and dismiss attack ads. I would expect that with multiple parties and positive choices being offered, campaign expenses will take on new forms--budgets for public meetings, door to door canvassing, that sort of thing. These costs I suspect will be less per contested vote per campaign. Because there will be many campaigns competing for the vote of each voter, we need to multiply that. And if all this means that a higher percentage of American voters do vote, that is another multiplier.

Perhaps then, the cost of campaigns in the USA will rise to even higher levels than they are now. But if the outcome of this is much higher engagement of the citizens with governing, stronger democracy in short, then I would think it is money well spent. I think PR done properly can short-circuit much of the power of money in US campaigns as we conduct them now....


......but if Brazilian experience tells you no, I'd be wrong, that PR just makes money able to buy power by other means, and explain with some examples I would like to learn more about that experience. Maybe it will illustrate pitfalls a different electoral system can avoid. Maybe it will seem still a clear improvement over our experience here in the USA anyway. Maybe it will illustrate fundamental limits of democracy, fundamental realities about wealth owning whatever system it operates in.

One thing I notice in glancing at the Brazilian parties--it seems just about every one of them calls itself "socialist" in some form or other. When I encounter the hard-core propertarian guff that is peddled as conventional wisdom in my own country, it seems that a nation of democratic socialists has got to be more on the right track.

Yet another sad illusion of mine?
 
The OP suggestion is not a alight divergence. Its a radical change from many decisions made by the founders and really seems to be working from modern assumptions rather than 18th century perspectives.
 
The OP suggestion is not a alight divergence. Its a radical change from many decisions made by the founders and really seems to be working from modern assumptions rather than 18th century perspectives.

Indeed. Although I reject the OP idea of proportional state by state, for reasons the Brazilian example seems to underscore, still that would still somewhat approximate the goals of proportional representation, at least in states with with more than one Representative assigned. Where it goes haywire from something likely to be thought of in the 1780s in America is the fundamental idea of "proportion" because this accepts the concept that the electorate is divided into factions and that a functional legislature would better represent and seek consensus among those factions, rather than appealing to a higher ideal that supposedly rises above faction. There was a certain hope in the Framer generation that the electorate would look beyond sordid self-interest and perceive the most worthy of their fellow citizens, who would look beyond factional interests to the greatest good. If the legislature is to embody such paragons of virtue, electing them all by simple majorities in each district seems straightforward and certainly simpler than trying to conduct elections on a statewide or national scale!

Now I have seen that quote by John Adams, asserting by syllogism that proportional representation would be best. I wish I could place it in context. I believe if I did it would show Adams talking about something close enough to our topic to argue the idea was not unthinkable in the era, but I suppose I could be mistaken and Adams was talking about something else, maybe even something like apportionment of representatives to regional populations or something like that. Perhaps.

The knowledge of legislatures that embodied rival factions, not as a temporary state of affairs but in normal operation over decades and lifetimes, would not be rare in this age. Everyone knew how Parliament worked; the colonial governments had their own legislatures for over a century before the Revolution; by the time of the Constitutional Convention the various state governments had been operating for over a decade, as had the Continental Congress. They knew, from over a century of domestic experience as well as study of Parliament's history and their rather selective, rhetorical reading of Classical history of Athens and Rome, and knowledge of yet other republics antique and modern (the Netherlands, perhaps even Corsica) what the nature of a legislative or other deliberative body (the Articles of Confederation executive, a collection of separate magistrates, as well as some experiments in collective executives in some states) of a bunch of nominally equal agents with some democratic mandate would be in action. Faction was in fact inevitable; this may have been optimistically hoped to be overcome by wisdom and patriotism and indeed the often high-minded actions of the Revolutionary generation of leaders gave some grounds to so hope. But the realization that in fact faction was probably something they'd never be rid of must have haunted the most hopeful, and the notion that perhaps the only way to deal with it in a healthy and surviving Republic would be to allow it to show itself openly and with minimal distortion, by recognizing the absolute right of the people to put forth their chosen champions and let them battle it out in public, rather than in diverse private and secret meetings, would not be unthinkable or unspeakable. I would like to pin down the context of Adams's reflections, but I think there is a good chance he did indeed mean devising means of proportional election of representatives reflecting the factional interests of the public in the legislature.

The thing is, for the concept of proportional representation to be applied well, it is necessary first of all to recognize that faction is part of the normal operation of government and not seek to somehow transcend it by some filtering process, but instead let it fight itself out openly in the formal bodies where the voters could observe and judge the process and outcomes, and allow their own trusted representatives to explain their compromises to their constituents so as to enable the minds of the public's members to evolve in understanding and wisdom. This was indeed a hurdle it may have been too much to expect the mentality of the Framer generation to jump. But not impossible I think.

But then secondly, it is necessary to provide a broad field for the voters to work in, to have suitable numbers for each constituency to vote in allowing enough nuance in the deliberative body to embody the distribution of opinion and interest with reasonable fidelity. For that, one needs to coordinate the election process across that broad field. The idea of a single Representative meant to stand for his entire constituency was technically feasible, and provided equal opportunities of representation for all the people regardless of whether their district was carved out of a big state or comprised the whole of a small one. It is certainly true that the Constitution as originally conceived operated as a cellular composite of a national government, with the States forming the essential cellular units. To conduct a proportional election of Congress across the entire nation would first require the conceptual hurdle of enabling factions spanning the entire nation to organize in a formally recognized way, and then for them, and trusted electoral registrars meanwhile, to communicate across its entire sprawl, from Georgia's southern border to the Massachusetts colonists in Maine, from the Atlantic coast over the hurdle of the Appalachian passes to the new states forming west of them in the great Mississippi watershed. Even if there were strong ideological consensus that the composition of the people's Representatives should reflect the factional spectrum of the nation as a whole rather than be a parliament of regional constituencies, to practically implement it would be a tremendous challenge in days before speed of light telecommunications. In contrast, simple winner take all districts, being originally conceived as being smaller than the majority of states and no larger than such small states as Rhode Island, would be much more manageable, and broadly speaking all of a comparable size, so that each one, operating under the authority of their respective states, would tally the totals over a period of a week or so, and the definitive results, district by district, probably filtered through a central state Registrar overseeing and certifying the results, would be locally known and announced, and the outcomes posted rapidly to the national capital. It might take considerably more time for Congressional officials in Washington to learn the outcomes for distant states like Kentucky or Ohio than to get results reported for Delaware or even Massachusetts (considering that seaborne news traveled faster than overland) but that would not matter to the electorates, who would all know about the same time assuming elections were all held on the same day (not true I think in the early days). We see that quite aside from ideological concerns, the winner-take-all District system was just plain more practical at the turn of the 18th to 19th century.

To have a party-based proportional system of the nation at large then would have been a logistical nightmare in 1800 to be sure.

I do think my own suggestion, of having half as many districts as there are offices, would be more practical. Elections would not have to collate across entire states or indeed the whole nation; as under the OTL system they run district by district, and the composition of half of Congress is decided just as rapidly as OTL, as fast as each district can count up the totals. Ballots are written up locally, perhaps not even mentioning the partisan affiliations of each candidate, with that information being published separately and the voters assumed to take the trouble to look it up before they vote. But upon getting the definitive, official results from each district, the information is compiled and sent to Washington, and there, with information about which candidates affiliated with which others in other districts already filed in advance, it is possible to collate the totals so as to arrive at national totals, then determine the proportional allocation of total seats for each alliance of candidates, then deduct the seats already won on a district basis by each, look at the remaining number of seats each alliance gets and select the top vote getters in each alliance to fill out that number, and send back the news of which additional candidates have won the other half of the Congressional seats. The nationally selected other half of Congress must wait for the information to run both ways and time elapsed to compile it. Every one of them, even those in districts neighboring Washington in Virginia and Maryland, must wait until the last state's official report arrives, and then a day or two if that to finish compiling the outcomes officially, and then a variable amount of time for the news to go back to the states. But it will be known in good time before the next Congress meets, who the secondary wave is composed of. All the new Congress members have time to pick up stakes and move to Washington for the next session.

As I said before, I think a major hurdle for this system to be accepted would be the slave/free division of the nation. The radical notion of universal and proportional representation would rub the slave states more tenderly to begin with. But assuming they just shrug and say "slaves aren't people after all," their lower potential turnout versus the possibility of northern states allowing their entire adult male population to vote will tend to weaken the power of factions favored by the South, and shift the identity of second wave national proportion-rectifying Congress members toward northern individuals, since they will be running in districts with larger average turnouts. Southerners, especially in states with rather moderate proportions of slave population, can combat that by being more militant about voting, but it is an inherent differential they have to struggle against. Thus to achieve balance of power between slave and free states it will not be sufficient for both regions to agree to admit equal numbers of slave and free states; the slave states must overturn the demographic balance to equalize representation in Congress, and thus seek more than parity in state numbers, to control the Senate. Either that or accept being consistently outvoted. For these practical reasons the pro-slavery faction is likely to denounce the whole notion of proportional representation and seek to abolish it.


This is why I felt the greater opportunity to implement such a system might be after the Civil War, most likely during Reconstruction.

1) the nature of the war was in part a referendum on the question of whether the USA is a confederation of separate sovereign states, or a unitary nation delegating substantial powers down to the states for reasons of administration and checks and balances. The Union won, underscoring the idea of a unified nation and Congress as properly representing the US people as a whole. With my districting system, it remains practical to keep districts entirely within states and even tolerate considerable discrepancies between district sizes, since the national aggregation makes misapportionment "come out in the wash."
2) All people have exactly equal power under my proposed system, so it would appeal to the idealism of the Radical Republicans.
3) the issue of freedmen voting and practically favoring the Republican party in states dominated by resentful Southern white majorities would tend to be fudged by the system if African-American voting rights are guaranteed. The freedmen would rarely be able to win district or statewide elections, but their votes, if honestly counted, would tend to favor the Republicans on a national scale. Thus African-Americans in the South would tend to be virtually represented by also-ran Northern Republicans taking Congressional seats on their behalf; meanwhile the Republican party retains an interest in guaranteeing Southern black voting rights. OTOH if the Southern states suppress the black vote, they will tend to lose the national seats relative to their census population anyway.
4) During the reign of the Radical Republicans, they pushed trough a number of Constitutional Amendments; the legislation mandating national proportional representation could have been one of these. Given a few years to work the new partisan alignments resulting would tend to entrench themselves against a backlash.
5) the development of the telegraph would have greatly accelerated the time needed to get results of elections from distant states, making the major delay in determining the composition of Congress the time it takes to count up and collate the votes in each state. Within a day or two of that, the composition of the nationally elected non-district House membership would be decided, and by the next day the outcome is known across the nation.
 
I support the idea of having some of the representatives at-large, but disagree that it would need to be as many as half. The United States is close to being proportionate to begin with. I suspect that having 10% of representatives being chosen at-large would be enough to balance things out.
 
Something I thought of, why not have each state function as a multi-member district for the purpose of electing say 75% of the seats while reserving the remaining 25% nationwide (could have different proportions), wouldn't that deal with the biggest issue mentioned earlier about having the states as at-large districts without creating a situation in which overhang seats result from an election like in Germany, and in fact you could divide the big states like California into multiple districts, maybe by capping the size of each district at 10 representatives or so
 
I support the idea of having some of the representatives at-large, but disagree that it would need to be as many as half. The United States is close to being proportionate to begin with. I suspect that having 10% of representatives being chosen at-large would be enough to balance things out.

I don't accept that a system that tends to shift extra seats to the leading party, as ours does, is really "close to being proportionate," nor that one that strongly blocks third party insurgencies is either. In looking at the current composition of Congress after the 2016 election one might agree that substantially the system seems proportionate enough, for if we were to shift over to a strictly proportional system of assigned 435 seats without regard to district by district victories but only by each party's share of the total, the overall balance would still leave the Republicans stronger than the Democrats, and these two parties combined would dwarf a ragtag contingent of third parties--the Libertarians would get 7 reps, there would be 5 or so "Independents," the Greens would have 3 or 4, and about 7 smaller parties would have 1 or two. But note that that would make for one substantial change--the Republicans would no longer control over 50 percent of the House, and would be forced to caucus with some of the smaller party Representatives to lead. That said, they have good options for doing so, while considering the apparent ideology of the third parties, the Democrats would have no realistic prospect of taking control, and trying to do so would involve massive compromises of their core ideology. The Republicans could simply caucus with the Libertarians and rule that way, since the theoretical points on which the parties differ have largely become questions of degree, not direction.

Anyway the current voting pattern is predicated on the assumption that it is futile to back anyone outside the two dominant parties, and even there if you look at the individual House races district by district, a great many of them are unopposed or clearly facing only token competition; again a product of the first past the post rule being the sole criterion of election. Change the rules to be proportional and I would expect the pattern to shift.

Now given the current pattern, it is apparently true that a fairly small "at large" contingent to counterweigh the distorting effects of winner take all would be needed to produce a perfectly proportional Congress! 10 percent would be too small; the Republicans would with 241 seats (won by majorities or in remarkably few cases, pluralities, in their districts) have 7 more seats then they ought to proportionally, and thus to fight for the remaining 44 seats a portion of the total vote considerably larger than 44/(435+44) would be thrown into the ring. My efforts to estimate the outcomes suggest that it is the largest parties among them--the Democrats most of all, followed by the Libertarians whose seats would drop from 7 to 5--who suffer the most from this Republican "squatting" on seats their national share would not entitle them to. The smallest parties, interestingly, would retain their pathetic shares. The Republicans as OTL would rule unilaterally with a solid majority of the 479 total seats.

But raising it by 20 percent, adding 88 or even just 86 seats for a total of 521, would be more than enough--of the 88, the R's would get a handful, the Democrats a whole lot, and again the smaller parties would be left almost unchanged. The R's would still caucus with the L's I think and rule that way.

But now look at Brazil. I linked to a Wikipedia page that gives the outcome of their 2016 election on a post above. The largest party got only 14 percent in their (imperfectly) proportional system. I cannot say how many district seats that would include because of course Brazilian districts are states, capped at 70 seats (which badly shortchanges Sao Paulo) and with a minimum of 8 for even the smallest. If Brazil were apportioned into 435 districts, it would be a major project to estimate where their boundaries would be, then tally up election results district by district to determine who won each with a plurality--but we can be sure the pluralities would be small! Actual majorities would be quite rare without a runoff system of some kind.

I made the assumption though that in a race for district pluralities, we can estimate that the probability of victory would be roughly proportional to the square of the share of total vote each party gets nationally. That is, we can assume that different parties have uneven distributions of popularity, that even small parties may have bastions where their support is strong. But the larger the party, the more likely it is to win a particular race due to its size presumably reflecting some component of broad popularity. On this assumption, I estimated that the largest party would win 120 seats--more than a quarter, less than a third, of 435, but clearly at least doubling their presence in the legislature relative to their share of votes.

In order for these 120 seats to correspond to their 14 percent total vote share, so that they would get zero of the at large allocation, the legislature would have to have some 860 seats! This is very near doubling the size.

As a rule then, the more successful PR is at getting a very large and diverse mix of parties to finely represent the detailed interest groups of the nation, the larger the at large portion must be if we would continue to award seats to each district based on plurality victory. The largest parties would otherwise dominate to an overbearing extent.

Now it might be argued, that that is good. That a legislature composed of 30 parties, the majority by far of which are tiny and collectively represent less than half the votes, is an absurdity and that the most popular parties ought to enjoy an edge to bring the small fry into line so the business of government can go forward decisively. Certainly even in the USA, with people voting as they are, a mere 10 percent, 44 seats, at large is enough to bring in the smallest parties that had a shot at even one seat, and yet leave the dominant party unilaterally dominant.

I disagree of course. I think the legislature ought to be a forum for the whole nation, all the people, to hash out policy collectively. Whether a proportional legislature would have many small parties or be dominated by a handful of major ones depends to a great degree on the subjective judgement of the voters, as it should. There are enough mechanisms in a capitalist society like ours to magnify the influence of the already powerful that we don't need nor should want the democratic bodies of the state to be artificially skewed toward the rule of one of two establishment parties at any given time.

I took a very long time to investigate the matter of smaller than a 1 to 1 ratio of at large seats to district, and I appreciate the eye-opener of just how few it would take to transform our Congress into a proportional system without resorting to purging fairly won plurality district seats nor suffering any distortion of the total outcome. Surely Congress could face being expanded to 521 seats with more aplomb than making it 871!

But in fact this still leaves too much basis for conflict whereby a few parties get grossly more seats than they ought to, by a factor of 2 or more, while others are unfairly shortchanged. It won't happen if voting patterns remain the same, and that is surely a possibility--that despite the freedom to create small third parties, the majority of voters will continue to favor two or three regarded as serious contenders, and only fanatics in similar numbers to stubborn third party voters today would turn elsewhere. I rather doubt the US electorate would scatter their votes quite as widely as Brazilian ones currently do. But I do think our system should allow for the possibility, if that is how Americans would choose to vote, and for that, strictly speaking even a straightforward 50:50 ratio of district to at large seats might not always prove adequate.

I proposed 50:50 on intuition largely. It is simple to explain, simple to illustrate how it works, easy to see how a party that even manages to win every single district would only rule by a small majority, not unanimously, in the working legislature. If 1 to 1 is not quite good enough to prevent parties good at winning blocks of districts from being overrepresented, at least it should not happen unless the votes are very widely scattered indeed, and then the "overbearing" effect would be limited and subject to counteraction by the actual majority.

One might wonder why I would not simply say that we don't even look at district races at all, but just assign the parties proportionally, filling the seats with the largest vote-getters in order, and be done with it. Well, then districts would just be administrative units for conducting the vote. In the US context certainly where we have been accustomed to look to the plurality leader as the winner of a race, and our current Constitution deeply enshrines the idea that each district, and especially each state, should be proportionally represented in Washington, I think such a weak district system would be strongly opposed. Part of the genesis of my wanting to evolve this notion was to enable individual districts to be the organizing center of campaigns, if that is what the voters in that district want. In a big nation like the USA it did not seem entirely reasonable to simply have a race for 435 seats at large decided by a uniform proportional vote across the nation. Nor is subdividing it into districts of 5 or 8 or 20 or whatever number very appealing; the aggregations we base proportional assignment on should be in as large a bin as possible. The notion of retaining districts clearly seems important to you as well.

Rather than expanding Congress to 871, I would favor halving the number of districts and reapportioning. After all as I've argued before it does not matter if the apportionments are mediocre in terms of either having uniform sizes or being demographically gerrymandered; it all comes out in the proportional wash, leaving each voter with exactly the same power no matter which state they live in.

One to one may not always be necessary, but it often will be needed I think.
 
Interesting. I haven't seen it broken down by numbers before. I didn't think the difference from proportionality would be so pronounced if we had 10% of the districts be allotted proportionally. Of course, if we stomped out First Past the Post and politicians choosing the districts, it would help the problem. But it seems that 50% is the best way to do it. The downsides are that now the average district would have 1.5 million people, and the people of the smallest states would enjoy disproportionate direct power to connect to their district representatives. But those are minor. Like you said, it all washes out.


And it goes without saying that when no party comes anywhere a majority, 10% is not enough.
 
Last edited:
Top