From: Larry Summers Sent: 9/12/2018 3:01:20 AM To: [email protected] Subject: Fwd: This morning Importance: — High More than u may want to know Sent from my iPhone Please direct all scheduling inquiries to my office at: Follow me on twitter @lhsummers www.larrysummers.com Begin forwarded message: From: "Maskin, Eric" Date: S eptember 11, 2018 at 10:24:21 PM EDT Subject: RE: This morning | believe there are, but don’t know how careful they are. | will try to get you references. From: LHS Sent: Tuesday, September 11, 2018 10:23 PM To: Maskin, Eric <n Ce: (nso fc rr Subject: Re: This morning Are there empirical studies on turnout. Sent from my iPhone Please direct all scheduling inquiries to my office | i _ Follow me on twitter @lhsummers www.larrysummers.com On Sep 11, 2018, at 10:20 PM, Maskin, Eric <> wrote: Yes, we can expect more extremists to run under RCV. But also more centrists (e.g., Bloomberg). Since the centrists are closer to the median voter, they will defeat the extremists. The evidence | have seen suggests that RCV increases turn-out. [But it’s important that voters be given the choice to rank as many or few candidates as they like, so that a voter always has the option of voting for a single candidate (in effect, he would be ranking all other candidates as tied for second). In practice, most voters choose to rank two or three candidates, but a significant fraction just rank one] HOUSE_OVERSIGHT_026321

In fact the increase in the diversity of candidates under RCV is related to turn-out. If RCV had been used in 2016, Bernie Sanders could have run as an independent in the general election without fear of guaranteeing a Trump victory. Many of the Bernie supporters who stayed home on election day might then have voted---and presumably would have ranked Clinton second. This would have given her a victory over Trump in Michigan, Wisconsin, and Pennsylvania (and possibly elsewhere). Eric From: 4S Sent: Tuesday, September 11, 2018 7:52 PM To: Maskin, Eric Ce: lhsoffice Subject: Re: This morning | get that formal argument What about aspects not quite in model. More folk will run from extremes if they can attract more First run votes. Candidates can position a bit. Separately | pitched this to someone today. He said he had heard that because of its greater complexity African American and lower income turnout was depressed. Is there evidence on turnout impacts? Sent from my iPhone Please direct all scheduling inquiries to my office at ee Follow me on twitter @lhsummers www.larrysummers.com On Sep 11, 2018, at 6:05 PM, Maskin, Eric {i wrote: Hi Larry, Thank you very much for taking the meeting this morning. I’m glad you found it worthwhile, and | agree with you that Adam Friedman’s commitment to the project is impressive. | AM serious about working on this---it’s a nice opportunity to make important practical use of some interesting theory. Voting rules may seem nerdy and dry, but they can make an enormous difference to actual politics The formal argument that RCV promotes centrism better than the current system (plurality rule) is straightforward. Suppose that most voters vote ideologically in the sense that the closer a candidate is to their own position on the left- right spectrum, the more like they are to vote for him. Then under majority rule (my favorite voting system)---in which voters rank candidates and the winner is the candidate who beats all other in pairwise comparisons----the winner will be the median voter’s favorite candidate ----- in other words, the most centrist candidate gets elected (this assumes that there are enough candidates running so that there is one who is reasonably close to the median voter). Now observe that RCV is in between majority rule and plurality rule, and so will promote centrism better than plurality rule. Best wishes, Eric HOUSE_OVERSIGHT_026322

From: LHS Sent: Tuesday, September 11, 2018 4:16 PM To: Maskin, Eric Cc: lhsoffice Subject: This morning It was fun and interesting. | admire your friends determination. Are you serious about working on this? Has anyone done a full analysis of this and polarization. Seems like on one hand it might encourage Ralph Nader cuz he d get more first round votes and not elect George bush. This might be bad. On other hand, gore would not have to move left to take Nader vote. If | like centrism is it clear that this is better. | get that it avoids arbitrary outcomes but assume that a tea party nut is as likely to elect a dem as a lefty is to elect et a republican. Anyway thanks for setting up. Btw | gave them 10,000. Larry Sent from my iPad Please direct all scheduling inquiries to my office me Follow me on twitter @lhsummers www.larrysummers.com HOUSE_OVERSIGHT_026323