Friday, January 19, 2018

Bob Wilson, Paul Milgrom and Dave Kreps win the Carty Award

Game theory at Stanford:)

Here's the press release from the National Academy of Science

David M. Kreps, Stanford University Graduate School of Business, Paul R. Milgrom, Stanford University Department of Economics, and Robert B. Wilson, Stanford University Graduate School of Business, will receive the 2018 John J. Carty Award for the Advancement of Science.

"Kreps and Wilson provided a framework, known as sequential equilibrium, for modeling dynamic effects in economics. All three of the award winners, together with other collaborators and in particular D. John Roberts, employed these techniques to model and study reputation and collusion, both of which have broad applications in macroeconomics, industrial organization, and labor economics.
Later, the entire modern telecommunications industry arose out of an auction format developed by Milgrom and Wilson, along with Preston McAfee, for the 1994 radio spectrum auctions by the Federal Communications Commission. The simultaneous ascending auction format, in which each bidder can bid for multiple licenses over a series of rounds so long as it remains “sufficiently active,” has since been used around the world to allocate hundreds of billions of dollars’ worth of wireless licenses. Variations of the format have also been applied to numerous other industries, including electricity markets and various commodity markets.  
The three award winners, with collaborators and alone, have contributed broadly to other topics in economics: Kreps has done foundational work in choice theory and financial market theory; Milgrom, in the theories of market microstructure and the principal-agent problem; and Wilson, in nonlinear pricing and utility regulation, as well as the foundations of dynamic equilibria.
The John J. Carty Award for the Advancement of Science is awarded every two years, to recognize noteworthy and distinguished accomplishments. In 2018 the award is presented in the field of economics. The award is presented with a medal and a $25,000 prize."

Thursday, January 18, 2018

Evolution of dating apps: OK Cupid requires mutual consent for messaging

Slate reports (last month):
OkCupid Users Have Long Had to Put Up With Unwanted Messages. That's About to Change
"OkCupid announced a big change to its messaging system in an email to users on Friday afternoon. Starting next week, the email said, “Only the people you like or have responded to will remain in your messages. Messages from people you're not interested in, or people you haven't liked yet, will be moved to their profile.”
"This shift will bring the platform more in line with other online dating platforms, such as Tinder and Bumble, on which users can’t message one another at all until both have shown interest in the other. The new OkCupid way won’t be quite so strict—users can still send messages to whomever they want, though those messages will only appear in the recipients’ mailboxes if they indicate that they like the sender—but engineers hope it will help seed more connections while filtering out some messages that will never get a response."

Wednesday, January 17, 2018

Analytic approaches to allocating organs for transplants

Two recent items:

Adapting a Kidney Exchange Algorithm to Align with Human Values
by Rachel Freedman, Jana Schaich Borg, Walter Sinnott-Armstrong, John P. Dickerson, Vincent Conitzer

The  efficient  allocation  of  limited  resources  is  a  classical problem in economics and computer science. In kidney exchanges,  a  central  market  maker  allocates  living  kidney donors to patients in need of an organ. Patients and donors in kidney exchanges are prioritized using ad-hoc weights decided on by committee and then fed into an allocation algorithm that determines who get what—and who does not. In this paper, we provide an end-to-end methodology for estimating weights of individual participant profiles in a kidney exchange.  We first elicit from human subjects a list of patient attributes they consider acceptable for the purpose of prioritizing patients (e.g., medical characteristics, lifestyle choices,and  so  on).  Then,  we  ask  subjects  comparison  queries between  patient  profiles  and  estimate  weights  in  a  principled way from their responses. We show how to use these weights in kidney exchange market clearing algorithms. We then evaluate the impact of the weights in simulations and find that the precise numerical values of the weights we computed matter little, other than the ordering of profiles that they imply.However, compared to not prioritizing patients at all, there is a significant effect, with certain classes of patients being (de)prioritized based on the human-elicited value judgments.


How analytics and machine learning can aid organ transplant decisions
by Dimitris Bertsimas and Nikolaos Trichakis

"MIT Sloan and Massachusetts General Hospital have developed an analytics tool to help doctors in deceased-kidney acceptance decisions. The model aims to calculate the probability of a patient being offered a deceased-donor kidney of a certain quality level within a specific time frame (three, six, or 12 months), given their individual characteristics. Using machine learning, it looks at 10 years of data and millions of prior decisions to estimate a patient’s waiting time in the context of a current active organ offer until the time to the next offer for a higher quality kidney."

Tuesday, January 16, 2018

Welfare effects of limiting the number of interviews by Beyhaghi and Tardos

Here's a new paper on a subject that is coming up in a number of the markets that I keep an eye on:

Effect of Limited Number of Interviews onMatching Markets
by Hedyeh Beyhaghi and Eva Tardos

Abstract. We study outcome of two-sided matching between prospective medical residents who can only apply to a limited number of positions and hospitals who can interview only a limited number of applicants and show non-intuitive effects in the matching outcomes. We study matching size as our notion of efficiency, and show when the number of interviews is limited, a market with limited number of applications achieves a higher efficiency compared to a market with no limit. Also we find that a system of treating all applicants equally (setting the same limit for their number of applications), is more efficient rather than allowing a small set to apply to one more/less position. This comparison results in a scallop-shape figure 2 that shows expected size of matching with respect to expected number of applications. Finally we show that limiting number of interviews does not always hurt efficiency of matching markets and can improve social welfare in certain cases. 

Monday, January 15, 2018

Spain continues to lead in deceased donation

I'm in Rome to talk today with the 28 EU Competent Authorities on Organ Donation and Transplantation about ways of increasing living donation through kidney exchange, by easing barriers at borders.

In the meantime, Spain remains a model for deceased donation:
Spain breaks organ transplant and donor records again
"The country's National Transplant Organisation (ONT) said 2017 saw a total of 5,259 transplant operations performed beating 2016's record of 4,818 transplants.

The majority were kidney and liver transplants.

Spain also boasts a much higher average of organ donors with 46.9 donors per million people in 2017, compared with 43.9 per million in 2016 and 39.7 in 2015.

The EU average is just 19.6 donors per million and the US average is 26.6 per million."

Sunday, January 14, 2018

Applications and interviews for medical residencies

The computerized clearinghouse for the NRMP medical match solves the congestion problem for new doctors when it comes time to make offers, acceptances and rejections. But electronic applications make it easier to apply for lots of places, and this is coming to seem like a problem in both the resident match and in the later-career fellowship matches.

Residency Placement Fever: Is It Time for a Reevaluation?
Gruppuso, Philip A. MD; Adashi, Eli Y. MD, MS
Academic Medicine
Issue: Volume 92(7), July 2017, p 923–926

Abstract: The transition from undergraduate medical education to graduate medical education (GME) involves a process rooted in the final year of medical school. Students file applications through the Electronic Residency Application Service platform, interview with residency training (i.e., GME) programs from which they have received invitations, and generate a rank-ordered preference list. The National Resident Matching Program reconciles applicant and program rank lists with an eye towards matching students and GME programs. This process has effectively served generations of graduating medical students. However, the past several decades have seen an intensification of the residency placement process that is exemplified by an inexorable increase in the number of applications filed and number of interviews accepted and attended by each student. The authors contend that this trend has untoward effects on both applicants and departments that are home to GME programs. Relevant information in the peer-reviewed literature on the consequences and benefits of the intensification of the residency placement process is scant. The authors address factors that may contribute to the intensity of the residency placement process and the relative paucity of data. They propose approaches to reverse current trends, and conclude that any reevaluation of the process will have to include the generation of outcome data to afford medical educators the opportunity to explore changes in an evidence-based manner.

"In part, the intensification phenomenon is borne out by the aforementioned growth in ERAS-associated traffic. In the eyes of many, this “new normal” draws on the widespread perception that a successful match in highly competitive disciplines is contingent on the filing of applications with a large proportion of the relevant GME programs. For example, in 2015, senior U.S. medical students applied, on average, to 73 of the 163 orthopedic surgery programs and 47 of the 105 neurological surgery programs (based on data extracted from the AAMC 8 and the NRMP 9,10). What is more surprising is that even less competitive disciplines may now be seeing an ever-growing flood of applications. This contention is supported in part by recent observations according to which GME programs in nearly all disciplines have seen a marked increase in their application traffic. For example, the percentage of pediatric GME programs to which graduating U.S. medical students have applied on average increased from 9.8% to 13.7% during the five-year interval from 2010 to 2015.10 For internal medicine GME programs, the corresponding figures are 4.9% to 6.0%.8–10 In making these decisions, students appear to be keeping their own counsel against the advice of medical school advisers and mentors advocating moderation."
"First, consideration should be given to the possibility of coordinating the timing of the interviews and of the Match across all disciplines and GME programs, including the “early match” disciplines of ophthalmology and plastic surgery.32 Consolidation along these lines would address the disruption of fourth-year scheduling, thereby offering educators greater flexibility in designing the fourth-year curriculum. Implementing such changes will not be easy given the longevity, familiarity, and comfort associated with the extant construct. Voluntary action on the part of the relevant professional associations will be required should a realignment of current schedules ever come to pass. Second, reducing if not capping the number of interviews per student would go a long way towards stemming the time and resource drain on both applicants and GME programs. This, too, is not going to be easy given the near-universal presumption that “more is better” and the notion that the times in effect demand such. In this context, consideration might be given to a tiered “screen and schedule” system wherein initial online interviews with many or all eligible applicants would be followed by a limited set of on-site interviews with a select group of “finalists.” As envisioned, this approach, widely used in both the public and the private sectors, stands to rationalize the current residency placement process while maintaining its fundamental premises of excellence and compatibility. Limiting the final on-site interviews to a select number of candidates will also give rise to palpable economies of scale that are likely to be welcomed by applicants, GME programs, and medical schools alike."