California voters decided Tuesday to radically rejigger elections in the nation’s most populous state. Under Proposition 14, a measure that easily passed, traditional party primaries will be replaced in 2011 with wide-open elections. The top two vote-getters – whatever their party, or if they have no party at all – will face off in the general election. Supporters argue that without parties picking candidates for the general election, moderates and independents will move to the fore, and voters will pay more attention to the electoral process. Critics of the measure say it will give a huge advantage to candidates who have the most money or the widest name recognition.
This is a great idea, and just what the sclerotic political system in Massachusetts needs to get new blood into politics, especially the state house. Lest anyone forget, the largest political group in Massachusetts, the 50.1 percent of the electorate which was unenrolled in 2008, compared to 36.95 percent Democrats and 11.62 percent Republicans, is completely unrepresented on Beacon Hill.
Incidentally, I submit that if we had this system there is little likelihood Scott Brown would be a U.S. Senator.
What do you think.
The jury is still out on if this actually changes anything regarding the quality of legislators, their effectiveness, whether it decreases partisanship, raises voter interest and turnout, etc. The Times had an interesting debate amongst different academics and political operatives who had different thoughts about the ideas. The one commentator from WA state where this is in effect said he noticed little change, though personally lamented he could not vote for any Democrats in his conservative district.
<
p>Nate Silver argues it will produce a glut of centrist candidates and ‘a thousand Liebermans’, while others have commented that LA (state) when it had this system tended to elect more extreme candidates including David Duke who made it out of a jungle primary.
<
p>Others have criticized this for marginalizing third parties even more or for diluting the strength of partisan identification which actually is a great way to asses where a candidate stands and whether he agrees with you. I would argue fusion parties, IRV, traditional runoffs, are better electoral reforms. Not to mention non-partisan redistricting coupled with campaign finance reform. This will likely only empower the existing power structure, well financed candidates, and bland candidates.
And the status quo is deeply unsatisfactory to many. A problem with many of the very worthy suggestions you make is that they are complicated. Fusion parties, for example, continue, I believe, to baffle most New Yorkers, where they are allowed. I personally have read many descriptions of IRV, including about an hour researching it online, and I have to stay I still don’t understand how it works. Then again, I freely admit to being obtuse and hard to educate. Proportional representation is another great system that almost no one understands in this country. Even most Cambridge braniacs don’t really understand it. Top two is extremely simple: if millions of Californians can get their heads around it, it should be even easier for infinitely more erudite Bay Staters. đŸ˜‰
Bob: Let’s apply IRV to the Dem Senate primary, since there were four candidates that folks here are all relatively familiar with.
<
p>Capuano
Coakley
Khazei
Pagliuca
<
p>Now, you as a voter go into the booth, and you rank them 1,2,3,4. Let’s say, hypothetically, you ranked them
1. Capuano
2. Khazei
3. Coakley
4. Pagliuca
others will rank them differently.
<
p>At 8pm, polls close, say 600,000 voters for easier math. What happens?
<
p>Step 1: Count how many (1.) votes each candidate has. It could be something like:
Coakley 275k
Capuano 200k
Khazei 75k
Pagliuca 50k
<
p>Step 2: Does a candidate have 50%+1 votes (300,001)? If yes, that candidate wins.
<
p>Step 3: Take the votes received by the candidate in last place. Assign each one of those ballots to the remaining candidates, by looking at each ballot and discovering the highest ranked candidate on the ballot who is still remaining and assigning the ballot to that candidate. Then, go to Step 2. Something like:
<
p>Coakley 275k + 15k = 290k
Capuano 200k + 30k = 230k
Khazei 75k + 5k = 80k
<
p>Step 2: Does a candidate have 50%+1 votes (300,001)? If yes, that candidate wins.
<
p>Step 3: Take the votes received by the candidate in last place. Assign each one of those ballots to the remaining candidates, by looking at each ballot and discovering the highest ranked candidate on the ballot who is still remaining and assigning the ballot to that candidate. Then, go to Step 2. Something like:
<
p>Coakley 275k + 15k + 30k = 320k
Capuano 200k + 30k + 50k = 280k
<
p>Step 2: Does a candidate have 50%+1 votes (300,001)? If yes, that candidate wins.
<
p>Coakley wins
<
p>
<
p>Now, in the actual election, Coakley got 47%. If (nearly) all of the other 53% had ranked Coakley last, then by the time it got down to two (say, Coakley and Capuano) then Capuano would have ended up winning, despite having fewer “#1” votes. Why? Because 53% of the folks voting preferred Capuano to Coakley. This is hypotehtical of course… given that Coakley got 47%, it’s extremely likely that she’d have gotten to 50%+1 by the time it was down to two candidates, perhaps even down to three candidates (since Pags got 12%).
<
p>
<
p>
<
p>A really easy way to understand IRV is to remember what it stands for. Instant Runoff Voting. Imagine if all 600,000 people were in a big room.
<
p>Round 1: Raise your hand if you like C, C, K, P:
<
p>Pagliuca came in last. He’s out. Now, it’s a three person run-off.
<
p>Round 2: Raise your hand if you like C, C, K:
<
p>Khazei came in last. He’s out. Now, it’s a two person run-off:
<
p>Round 3: Raise your hand if you like C, C:
<
p>Whichever C got more votes wins. IRV merely allows all of these steps to take place in one interaction between voter and ballot, not requiring voters to await the results of the results of the first round to register their feelings on the second round.
“Weed from the bottom” might be a better slogan than “IRV.”
<
p>I still submit that it is confusing. It took you several paragraphs to explain it. Contrast that with:
<
p>”The top two vote getters compete in the election.”
<
p>That is simpler.
Someone tried to explain this to me last weekend and I think I get it now. One issue with IRV is that without the results of previous rounds, you might hedge your bets for following rounds.
<
p>Consider this order of preference:
<
p>1. Pagliuca
2. Khazei
3. Coakley
4. Capuano
<
p>Further, assume that I really don’t want Capuano to win (I’ll pick on Mike since it’s known he was my first real choice). If there were no IRV, I’d assume that Pagliuca would get a really small percentage, not enough to boost anyone over 50% for the second round. However, the whole purpose of IRV is to allow people to vote for their dark-horse candidate, so who knows, maybe that percentage might be enough. And maybe enough Pags voters picked Capuano second, possibly giving him the victory. To hedge against that possibility, instead of voting Khazei #2, I might have to give Coakley my #2.
<
p>I hope I’m making sense here. The main point is that while many voters would just vote their preferences (and hopefully they’d get used to voting for more than one), others might still work electability or “anybody but” scenarios into their voting.
Why did it take you 24 inches of text to explain it? Even this disciple of the stomv tribe got lost and distracted.
Just rank your candidates in order of preference. Done. To vote, that’s all you need to know. Not rocket science.
<
p>For those who are interested in the tallying, it works just like a series of runoffs, just calculated at once, without requiring separate trips to the voting booth.
it’s not really essential that each voter understands the mechanics of IRV voter reassign. It simply requires that
a. the voter is confident that the system is fair, and
b. the voter is capable of ranking choices.
<
p>While (a) is non-trivial, (b) is trivial. It’s easy for voters to rank their candidate preference. We rank our preferences all the time, lots of different ways. The concept is so simple that third graders do it all the time — Charley is my bestest friend, David is my second bestest friend, and Bob is my third bestest friend.
I recall my son saying happily, “I have two best friends. Stephen and Eddy are my best friends.”
<
p>I ask, “What about Joey?”
<
p>Without missing a beat he responds, “I have three best friends. Stephen and Eddy and Joey are my best friends.”
<
p>I’m not quite as confident as you that voters are capable of ranking.
<
p>
IRV has a number of flaws to it that show up in simulations. Here’s a bunch of charts. Each chart was generated as follows:
<
p>Each chart represents a voting scenario. Candidates are located on a 2-D plane, and represented by colored circles on the chart.
<
p>For each pixel on the chart, 10,000 “voters” were scattered at random around that pixel by Gaussian distribution. Each voter then “voted” for the candidate they preferred, based on distance from that voter to the candidate. The nature of the vote obviously depends on the voting system. For our current system (“OneVote”), it’s a vote for the closest candidate. For IRV, they are ranked by distance. For approval, there’s a threshold which a candidate must be within to get a vote.
<
p>So, in other words, each pixel on the map represents the winner of the election IF the voting population were centered around that point.
<
p>There’s a handy control in the upper left you can use to reduce the number of charts. But just take a look at the sort of results that IRV returns with 3 or 4 candidates. Do those maps make any sense?
<
p>For the reader’s digest version, you can also take a look at this page which I think accurately sums up the arguments against IRV.
<
p>Do I have a horse in this race? I go back and forth on what I’d consider the ideal voting system, mostly between Range Voting and Approval Voting. Both of which return results that make a lot more sense than IRV. The above site even simulates what Approval Voting would look like if a) voters went in blind (having seen no polls) or b) voters knew what polls were saying and voted strategically. It holds up pretty well.
<
p>And, it’s even simpler than what third graders do. Approval voting asks the questions that most pollsters already ask: Do you approve of so-and-so as your Senator?
<
p>But mostly, if we are going to change from First-Past-The-Post (which is the worst system by almost any measure), I want it to be to something other than IRV (which is the second-worst system by many measures).
<
p>Every system has flaws, both mathematical, sociological, and political. The former was shown (by Condorcet? Don’t remember), the latter two clear through conversation and experimentation. Range voting is really contrived (rate each candidate 0-99?!?!?). Condorcet has more granularity, but requires far more effort when there are more than 3 or 4 candidates. Consider: in a 6 person race, in FPTP I check one box (who I like); in IRV I indicate 1-5 (same effect as 1-6). In condorcet, I have to look at each of the pairs and make a decision: 15 total decisions. If there are 8 candidates, it goes to 1, 7, and 28. Condorcet is really tough if we’ve got more than a handful of candidates, and I suspect likely to have high voter error rates if the ballots aren’t designed really, really well.
<
p>
<
p>They’re pretty, but rather irrelevant. For one thing, it presumes Gaussian and some sort of uniformity on each axis segment (not an axis because it’s bounded). Neither of these seem to match anything close to reality. You could argue 2 axes (economic and social liberal/conservative) but that doesn’t really capture all kinds of different issues, nor does it consider that most voters don’t have have a strong empirical system to measure their own candidates; preferring to use far fuzzier concepts like trustworthiness, tribalism, and so forth. As for the uniformity, it doesn’t make sense because, frankly, people motivated enough to get so involved tin politics that they’re willing to run as a candidate are rarely split-the-difference moderates on most issues, even if such a position exists.
<
p>
<
p>See, I find that much much harder. Do I approve? Compared to what? I would very much approve of Scott Brown to be the senator of South Carolina. Or, the senator of MA in 1844. There are clearly folks who I would approve of and others that I wouldn’t, but those in “the margins” get pretty close. Would I approve of Steven Lynch, for example? Well, I’d rather have him then Mitt Romney as a member of Chogress, but I’d rather have Mac than Lynch. So, do I approve of Lynch? Sometimes, on a case by case basis. Far easier for me to answer “Who do you like better — Lynch of Mac? (Mac) Lynch or Romney (Lynch).”
First off, I never mentioned Condorcet or suggested it as a good idea. And your interpretation of Condorcet is far different from anything I’ve ever read.
<
p>A Condorcet ballot is not cast as a head-to-head ballot, at least not traditionally. The usual format of a Condorcet ballot is a ranked alternative ballot, just as IRV is. The difference comes in how they are tallied.
<
p>A Condorcet ballot is tallied as a series of head-to-head contests. For each pair of candidates, you look at how many ballots A is ranked above B, and vice-versa. And the tallying proceeds from there.
<
p>Second, I agree that the “traditional” (if there is such a thing) Ranged ballot, from 0 to 99, is ridiculous. There’s granular, and then there’s absurd. I think 1 to 5 is probably about the most you can reasonably expect of people. Or just 0 and 1 – aka approval voting.
<
p>Third, I don’t think the graphs on that site are irrelevant at all. Yes, they make assumptions that may or may not apply to the real world (though I think they are more applicable than you seem to think). But just look at some of the shapes IRV produces on those maps. If under simplistic situations it comes out looking like that, what sort of results will it produce in the real world?
<
p>Fourth, I for one find approval questions very easy to answer. Yes, I’d want Coakely for my Senator. I’d also want Capuano. And Khazei. But not Brown or Pagliuca. Doesn’t seem like a difficult question to ask, answer, or tally. Compared to IRV where I’m really supposed to figure out my fourth-place preference between Brown and Pagliuca?
Can you imagine how nasty statewide recount would be involving IRV? No thank you.
Amen. I have grave concerns about any voting system that requires all votes (or a representation thereof) be co-located in a central facility before counting can even begin. Most fully ranked systems (IRV, Condorcet) have this problem.
<
p>First Past The Post (our current system), Borda, Approval, and Range votes do not have this issue.
why would anyone deliberately weaken his or her own ballot vis-a-vis their neighbor’s by giving his or her candidate anything less than a 99 in range voting or the opponents anything other than a zero in approval voting?
I doubt very much that anybody would give “his or her candidate anything less than [10] * ” in range voting. That is, if they truly have a candidate.
<
p>* I said above I think it is ridiculous to have 100 possible values, and that 5 or 10 is much more reasonable. I’m gonna go with 10 for this comment.
<
p>Those people who haven’t fallen in love with a candidate, and truly feel “meh” about several candidates (i.e. they feel like they’re picking the lesser of several evils) may want to vote those candidates a 4, 5, or 6.
<
p>Similarly, those people who are gung-ho about a candidate, but feel that another candidate would do a good job, may give full value to “their” candidate, and a mid-to-high value to that other candidate.
<
p>But you are correct, I think, in that most people would give their top candidate a 10, their bottom candidate a 1, and place the other candidates somewhere in the middle, according to their feelings about those candidates.
<
p>Just as an example, at the convention, I was pretty torn between Bump and Lake. I ended up practically flipping a coin and voting for Lake, but I would have given them both 9 or 10, and Glodis a very low value indeed. I wouldn’t have to think about which one of Bump or Lake I preferred, and if I did prefer one over the other, I could have given him/her a 10 and the other a 9, as a show of a slight preference over Guy Glodious, who I would have given a 1 to.
In most districts, the Rs are fairly non-existent anyway, as most conservative voters (and many moderates and liberals for that matter) register U instead. The open primary lets them vote in the D primary, which is often the main event. The big difference is that it would move that event November, rather than September, which would make this more like our local elections where we have preliminaries.
<
p>The big downside is we wouldn’t have our fun four way governor debates.
I’m skeptical of this, but the truth of the matter is that no one really knows how this will play out in California–whether it will lead to the election of centrists, or whether it will lead to selection of candidates by “back-room insiders”, or what. Why not wait and see how it works before deciding whether it’s a good idea for Massachusetts?
<
p>TedF
It’s classic spoiler nonsense.
<
p>Consider this made up not unreasonable example:
Adam Conservative: 25% support
Betty Moderate: 30% support
Chris Liberal: 20% support
Doug Liberal: 15% support
Ellie Liberal: 10% support
<
p>So, what happens? As a state we’re 25% conservative, 30% moderate, 45% liberal… and we get to choose between a conservative and a moderate? This scenario isn’t that unreasonable — the 50% indie could go 13 to the Con, 30 to the Mod, and 7 to the Libs, for example.
<
p>It’s pure foolishness. All it means is that the politics will be extremely insider — fighting to keep candidates out of the race before the run-off, egging folks in the same party to be especially ugly, dopey conversations about strategic voting, and so forth.
<
p>There’s so much wrong with this. Let’s try it again and consider Bob’s claim. Consider:
Scott Brown: 30% support
Martha Coakley: 35% support
Mike Capuano: 25% support
Alan Khazei: 7%
Steve Pagliuca: 3%
<
p>We get a run-off, with Brown and Coakley. Gee, this is familiar. Why, Bob, do you think that this is an unlikely outcome? The only reason I can think of is that the party would have chased out Pagliuca as quickly as possible, hoping that the 3% would increase the chance that two Dems finish ahead of Brown. Then, they’d have marginalized Khazei and his supporters, insisting that they’re no different than Green-Rainbow Naderites, marginalizing both Khazei and his supporters. Either he gets chased out or he stays, either way the long run goal of a robust democracy is reduced because we either get fewer ideas or we get a better chance of an outcome on the opposite side of moderate as the electorate.
<
p>If both Pags and Khazei get chased out, what happens? Well, the Pags folks could go anywhere, and the Khazei folks may well stay home. It’s not unreasonable to suggest that you get:
Scott Brown: 31%
Martha Coakley: 37%
Mike Capuano: 29%
Stay home: 3%
and again, the Scott Brown v. Martha Coakley run-off. In addition to (still) getting Scott Brown v. Martha Coakley, we also managed to chase out two candidates who were willing to raise (or spend their own) money to bring out new voters. We’ve eliminated outsider candidates, new energy, and plenty of the folks associated with their campaign.
<
p>
<
p>Want to make it wide-open? I’m all for it. Instant Runoff Voting. One election. Watch what happens — the negative campaigning goes way down because you want to get folks’ #2 vote, not their last-vote by alienating them. Candidates will be willing to agree on issues. Candidates will be willing to make clear distinctions on other issues. There won’t be the “toe the line” games, because different candidates within the same party will want to distinguish themselves without forming a circular firing squad.
<
p>As a bonus, you’d attract more candidates without incentive to chase similar-minded candidates out, and you’d be giving those from minor parties or indies who aren’t mega-millionaires a real chance. You could also do it all in a single election, instead of having two elections with different dynamics because the objective is different (first objective: finish in top two, which varies widely in terms of votes or percentages based on field; second objective: get 50%+1 of the votes).
<
p>
<
p>Now, you could do IRV two different ways:
Option 1: Wide open “jungle”, one race, IRV, winner is winner.
Option 2: Primaries, followed by IRV general election
<
p>I prefer Option 2, with IRV in primaries. I think that it helps the parties determine exactly what it is that they stand for, and help establish priorities (Do the Dem candidates who waffle on choice get pasted? Do the GOP candidates who support marriage equality get toasted? etc). It also allows the party members to then get to work together to support the candidate who is far closer to their ideals than the candidates from the other parties… which is also important because of the other races going on, from POTUS down to city councilor.
would be IRV “jungle” primary to select the top two candidates for the final election.
<
p>I believe it satisfies in large measure many of the technical objections to IRV, though I do not have a math proof of that.
I would just add to stomv’s points a problem with this statement:
<
p>
<
p>While unenrolled voters are indeed the largest group, the vast majority of them, like independents elsewhere, clearly lean towards one party or another. So saying that they are “completely unrepresented” on Beacon Hill is not really accurate. Plus, unenrolled voters can already vote in either party’s primary, so they can choose their preferred candidate that way.
can’t be all that bad. The more power taken from political parties, the better.
What if it upsets one party and pleases another?
in California, at least.
<
p>In Massachusetts, there is only on party. In my view, anything that tends to reduce, or one can dream, eliminate, the power of that machine will be to the Commonwealth’s long-term benefit.
…and the Democratic Party in MA is hardly a grand unified coalition that it appears to be on paper.
Finneran. Birmingham. DeMasi. DeLeo.
<
p>These guys have way too much power, and exercise far too much control over members of their party through committee appointments and rules setting. In the absence of sufficient people from some different party, that power is near absolute.
<
p>And then nitiwits act as if switching a GOP seat in the General Court to a Democrat will change things– see he’s a progressive, and once said nice things about Clean Elections!–, rather than entrenching the status quo even more. The status quo in which retiring from the MBTA at 55, in the prime of one’s working life– counts as “reform.”
<
p>Over-identification with a political club, and rampant yay-teamism are pernicious, and though they have metastasized differently in Massachusetts than in California, are no less malignant.
…which wouldn’t necessarily be better under a different party. I’ve said several times here that I think the presiding officers have too much power, but Democrats aren’t always in lockstep behind the leadership. Keep in mind that Finneran was actually elected by a coalition of House Republicans and a minority of conservative Democrats over the objections of a majority of the Democratic caucus.
It would be better if there were more than one, or less than one, party, because the rules are set, how?– by party bosses. More than one isn’t going to happen as long as our Nuke a Gay Terrorist Union Whale for Jesus friends are the choice. Therefore, the best solution is anything that is harmful to the existing, single party.
In fact it has notably been rebellions WITHIN the party in power that lead to rules reform that is needed. Are you suggesting abolishing parties? I suppose it’s an option as that’s how Nebraska does it. I personally like having Democrats in power, though I do have a preference for kind of Democrat. I’d much rather primary entrenched Democrats with better Democrats. I don’t want to hurt (or help) the party in power just for the heck of it; I want to elect whomever will vote my way.
And then your better Democrat shows up and finds that she can’t do a thing unless she does what she is told when told to do it. There is no opposition party to which she could align on leadership issues, which means she is essentially 100% powerless, unless and until she learns how to play, and then she is just another entrenched democrat. That is why I think it would be far healthier for the state if the Republicans were not so thoroughly outnumbered.
<
p>Parties are, on their very best day, a necessary evil, but an evil nonetheless. They are somewhat less awful if they come in pairs.
<
p>
…to form something of a party within a party, and yes it takes work over several cycles to accomplish. Remember again, the example of Finneran elected by a minority of Democrats aligned with Republicans. I must confess that for several years I’ve had this fantasy that the Dems for just one cycle completely run the table, besting the Know-Nothing record of 1854 when they held every seat on the Council, Senate, and CoDel, all the statewide offices and all but three House seats. Once we do that we can have two parties again:) I think we’ve also seen plenty of examples of Democrats being divided; casinos comes to mind.
Let’s take this example into account, for this November’s 9th Congressional District Election. So far, the candidates are as follows:
<
p>Independent:
-Phil Dunkelbarger
<
p>Democrats:
-Mac D’Alessandro
-Stephen Lynch
<
p>Republicans:
-Keith Lepor
-Vernon Harrison
<
p>Now, an “experienced” political operative would tell us that Lynch comes out of all this just a bit poorer after a primary with Mac, and destroys Dunk and Lepor in the General Election after voters forget how hard he’s screwed us for the past 10 years.
<
p>But, take the Californian model into account, and it’s a whole new ball game. It forces the media to focus on the viability of the person, and not the candidate within the party. History shows us that there have been only a handful of candidates on the Federal level who have run as an Independent and emerged victorious. However, we don’t need the Californian rule to be applied in MA for change to occur.
<
p>Joe Moakley, ironically enough, ran in 1972 as an Independent from the 9th Congressional District, and won, making it known he would be sworn in as a Democrat if elected. He ran as an Independent, since he was beaten by Louise Day Hicks in the previous cycle in the Democratic Primary. Moakley knew the only way to beat her was to take her on in the General Election, and leave the foolishness of Party Primaries in the dust. In any case, after he won, he held the seat for 14 straight terms, and never looked back.
<
p>Dunkelbarger, just like Moakely, knows exactly what he’s doing. Running as an Independent affords him more time (no primary in September) and allows him to take on Lynch in the General Election without needing the money to beat him in a Democratic Primary.
<
p>In conclusion, if you want someone who closely represents the ideals of Joe Moakley, the original Independent Voice, VOTE DUNKELBARGER. It’s pretty simple.
<
p>
and would like to see how it plays out in California. I probably would vote for it if this came up in Massachusetts.
<
p>I do think it’s time to change the system in some way. Putting people on the ballot by way of convention is too difficult and creates a needless barrier. As long as candidates can get the signatures they need, I think they have what it takes to be put out to the electorate for votes.
This is a HORRENDOUS idea. It practically ensures a large number of unrepresentative representatives, unless plenty of good candidates decide against running, sacrificing their candidacy for sake of the issues they care about – which will give us less choice and worse candidates on average.
<
p>Look at what happened in the recent Hawaii runoff for House, where a Republican was elected despite >60% Democratic vote, precisely because they used a system like this for special elections.
<
p>Ugh. Please don’t vote for this, anywhere, ever.
<
p>What we need is instant runoff.
This takes the really bad problem we already have in 3+ candidate primaries, and takes it to the general election too. It is “the Nader effect” extra-magnified. Without preference voting / IRV, the result will be a much higher number of people getting elected who are way out of line with their district’s views.
<
p>California governance is already such a disaster that I don’t know if this will really make it that much worse, but wow what an unmitigated idiocy this is.
<
p>Oh really? How?
<
p>If the same candidates had run, we’d have ended up with a race where all Republican votes coalesced on Scott Brown, while Democratic votes would’ve been split between four candidates. This would’ve made Brown an almost sure winner. It would’ve been like the recent special election in Hawaii, which used this system; or like how Allston/Brighton elected a Bush-supporting Democrat as their state rep.
<
p>Or, the Democratic candidates might’ve seen this coming, and some of them dropped out of the race to prevent an easy Brown victory. But Coakley, as the leading Democrat in the polls, would not have dropped out. So either some other Demcrats would’ve stayed, and we’d still split the vote, or all the non-Coakley Democrats would’ve dropped out, and we’d have had a Coakley-Brown matchup, just like we did have.
<
p>There’s no way a system like this could’ve made that election better, it could only make it worse.