Testimony on Redistricting

Transcription

1 Testimony on Redistricting Steven O. Kimbrough, for Team Fred August 31, 2011, 5 p.m. at Esperanza Academy Charter High School, 301 West Hunting Park Avenue. For immediate release. Hello, my name is Steve Kimbrough. I am a Professor at the University of Pennsylvania, in The Wharton School. I am a decision scientist, specializing broadly in computational intelligence. I am also a member of what we call Team Fred. The team consists of four people: Ram Gopalan, Frederic Murphy, Nicholas Quintus, and myself. We have entered a couple of very good redistricting plans in the contest being held by Azavia. I am not here, however, to discuss our particular entries in the contest. Nor am I here to discuss any particular redistricting plan. Instead, I want to discuss with you results from our research and development efforts. These results present something genuinely new and different, something that offers the prospect of a win-win situation for all stakeholders. It will take me just two or three minutes to explain the essentials. Here goes. Suppose that in doing redistricting we could begin with a pool of good plans. Let us call these the contenders. (As in, I coulda been a contender. I coulda been somebody,... 1 ) Each of the contenders think 20 or 30 in all meets the legal requirements for contiguity and population size, and is reasonably compact. Deliberation, both public and private, then focuses on these contenders. Some contenders may be better than others because of neighborhood considerations, historical reasons, and yes even because of political factors. Anyone at any time may, of course, add to the list of contenders, provided the new addition meets the legal requirements and is reasonably compact. Adjusting a contender is a good way to introduce a new, potentially better contender, but any method is in principle acceptable. People will disagree and in the end City Council must and should decide. It is the job of the political process to accommodate differing views and interests, including and even especially those not represented in any given decision model or formula. The job of deliberation, both public and private, is to surface and give hearing to these differing views and interests. The process I have just described is based upon the principle (or even philosophy) of solution pluralism. The core idea is to base decision making not just on one supposedly best solution, but on a collection, or pool, or plurality of reasonable solutions. That is, on the contenders. The principle is to begin deliberation with an ample, comprehensive pool of attractive contenders. 1 Terry: You don t understand. I coulda had class. I coulda been a contender. I coulda been somebody, instead of a bum, which is what I am, let s face it. It was you, Charley. From the 1954 movie On the Waterfront. http: // 1

2 All of this sounds good and commonsensical, but it has not been genuinely possible it has not been practicable because there has been no way of identifying a comprehensive pool of contenders. This is needed because if you bias the start you will bias the conclusion. Here is where our results, Team Fred s results, come in. Using modern optimization techniques, combined with artificial intelligence techniques, and adding in a number of our own innovations, we have produced a replicable method for finding comprehensive pools of good contenders. That is what is fundamentally new here. Our methods can find very many very good plans. These are plans you would be very unlikely to find otherwise. It is this that presents the opportunity for a win-win situation. I am not speaking hypothetically. Limiting ourselves only to plans that did not break up any of the City s 66 wards, we have discovered more than 100 distinct contenders: plans that meet the legal requirements and that are reasonably compact. Many of them look great. No doubt most of them can be improved by taking into account the various views and interests mentioned above. We have posted these plans on the Web. Anyone can view the maps. See for yourself. The URL is sok/phillydistricts. I have five points I want to make in concluding. They each point to much larger issues, but in the interests of time, I shall be quite brief. 1. The methods we invented and used are quite general. We worked at the ward level, but any level is possible. Perhaps you would like to divide the City into neighborhoods that should not be split? Well, we can easily do that too. 2. In consequence of our methods being very general, they will be useful for many districting (aka: zone design) issues beyond voting. Think: police, fire, and other services provided by the City. Here, too, a philosophy of solution pluralism would be very attractive and is now enabled by the innovations we have made. 3. We are acting here in our roles as university researchers, with complete transparency. All of our data are already publicly available. We are happy to make all of our software available as well, and we would be pleased to work with City Council or other groups. 4. This is, to repeat, something new and different. The genie is out of the bottle in the sense that it is now possible to produce for public discussion very many high quality solutions. This places a very public burden of proof on anyone who would insist on a much lower quality that is to say gerrymandered plan. We hope that City Council will embrace this new opportunity. We believe it is in City Council s interests to do so and we offer to work with you in that spirit. 5. Finally, we want to commend and thank Azavea ( and in particular its President Robert Cheetham for spending time and money to organize and conduct the redistricting contest. They have done an excellent job. We note further that this is a very successful example of the neat ideas contests that Fred Murphy and I proposed for Philadelphia in our report for Greater Philadelphia First on stimulating innovation in the 2

3 Philadelphia economy. 2 Contests of this sort have proven their value from the time of Athens in ancient Greece to Netflix today. We urge the City and the Commonwealth, along with relevant foundations and non-profits to follow the example of Azavea to address the nuts and bolts problems of government and local industries. Azavea took the chance of sponsoring the contest. Our new methods and results demonstrate that such contests can stimulate innovation in Philadelphia. Thank-you for listening. I will be happy to answer your questions, now and later. * * * Team Fred: Nicholas Quintus nquintus@gmail.com Steven O. Kimbrough kimbrough@wharton.upenn.edu Frederic H. Murphy fmurphy@temple.edu Ram Gopalan ram.gopalan@gmail.com References Kimbrough, S. O. and Murphy, F. H. (2005). A study of the Philadelphia knowledge economy. Interfaces, 35(3): $Id: testimony tex :41:04Z sok $ 2 Greater Philadelphia: Opportunities for Initiatives in the Knowledge Economy, October 2001, wharton.upenn.edu/ sok/sokpapers/2001-2/gpf/shortreport pdf. See also (Kimbrough and Murphy, 2005). 3

4 Philadelphia City Council Districts Figure 1: Current Philadelphia City Council districting map. Based on 2000 census. 4

5 Aug27_orig70_s13 Figure 2: Example districting found by our algorithms: s13. Contiguous, populations within 5% of each other, and none of the 66 wards broken up. Based on 2010 census. 5