About

Saša Zorc is an Assistant Professor of Business Administration in the Data Analytics & Decision Sciences area at the University of Virginia Darden School of Business. He teaches the core course Decision Analysis and the Games, Competition, and Cooperation elective in the MBA program, as well as a PhD course on Dynamic Programming. His research primarily focuses on matching markets and the design of incentives, with applications in healthcare settings. The central theme of his work is how to design rules for complex systems—such as job markets, matching platforms, and healthcare payment models—in a way that ensures good system performance despite individuals pursuing their own often conflicting interests. Methodologically, his research relies on stochastic dynamic games, search theory, mechanism design, contract theory, and data-driven simulations of these systems. His work has been published in Management Science and Operations Research and has received recognition in several best paper competitions, including the Pierskalla Award, IBM Service Science Award, Decision Analysis Publication Award, and the Junior Faculty Interest Group Award. Saša currently serves as an Associate Editor of Service Science. Before joining Darden, he completed his PhD in Decision Sciences at INSEAD and worked as a management consultant for Kearney.

Research topics

• Computer Science
• Political Science
• Business
• Computer Security
• Risk analysis (engineering)
• Microeconomics
• Family medicine
• Mathematics
• Medicine
• Medical education
• Psychology
• Statistics
• Social psychology
• Internal medicine
• Economics
• Programming language
• Law and economics
• Operations research
• Law

Selected publications

• Choosing Outcomes-Based Reimbursement Policies: Should We Worry About Collusion?

  Management Science · 2026-03-03

  article1st authorCorresponding

  Outcomes-based reimbursement rewards health providers with better health outcomes with higher payments. Such reimbursement policies require several design choices, including the type of contract (e.g., capitation or fee-for-service), measure (e.g., population- or provider-level outcomes), and whether to contract with individual providers or larger groups. We explore which outcomes-based reimbursement policies may be vulnerable to potentially illegal collusion, and whether collusion issues can be averted through incentive design. We present a game-theoretic model a chronic care pathway in a two-tier healthcare system. We identify differences in the impact of collusion on health, costs, and system efficiency under different reimbursement policies. Theoretical and numerical results (calibrated to data from two pathways for diabetes) show that whether an outcomes-adjusted reimbursement system is vulnerable to collusion depends critically on one trait: whether the income of physicians who receive referrals scales with volume. In systems where it does, as with fee-for-service models in the United States, there exist financial incentives to collude, underlining the importance of addressing collusion through laws. Systems that lack this trait (e.g., the UK NHS) are more resistant to collusion. Exploring theoretically optimal contracts, we find evidence of strong performance of outcomes-adjusted capitation contracts with individual providers using population-level data. This paper was accepted by Terry Taylor, operations management. Funding: S. E. Chick acknowledges research support through the Novartis Chair for Healthcare Management. Supplemental Material: The online appendix and data files are available at https://doi.org/10.1287/mnsc.2021.02283 .

  PublisherDOI

• From Black to Gray: Improving Access to Antimalarial Drugs in the Presence of Deceptive Counterfeits

  Management Science · 2026-04-21

  articleSenior author

  In malaria-endemic countries, the limited availability of affordable antimalarial medication has contributed to the widespread distribution of counterfeit drugs. This paper examines such markets to determine how philanthropic donors can best allocate limited funds to subsidize the purchase or sales of antimalarial drugs via private-sector distribution channels. We develop a game-theoretic model of the antimalarial supply chain wherein the retailer strategically sources legitimate drugs from a donor-certified supplier, potentially counterfeit drugs from an uncertified supplier, or both. In contrast with the extant literature, we demonstrate that, in the presence of counterfeits, exclusive reliance on purchase subsidies may no longer be optimal. Specifically, under donor budget constraints, offering sales subsidies that incorporate both legitimate and counterfeit drugs may be preferable or, in some cases, abstaining from subsidy provision completely. Additionally, we evaluate the implications of three pertinent nonsubsidy strategies deployed to combat counterfeit drugs: imposing penalties for sourcing counterfeits, eliminating subsidies on counterfeit drugs with traceability technology, and implementing price controls. Our study concludes with extensive numerical analysis calibrated to malaria data from Mozambique. Overall, this paper offers strategic guidance for improving outcomes in the presence of counterfeit drugs. Our results highlight the need for governments and donors to carefully consider market-specific factors, such as retailers’ pricing power and donors’ budget constraints, when designing subsidy schemes and access policies for life-saving medicines. These insights could potentially be extrapolated to address similar challenges in other endemic disease contexts, offering a broader framework for enhancing public health in resource-constrained environments. This paper was accepted by Jeannette Song, operations management. Supplemental Material: The online appendix and data files are available at https://doi.org/10.1287/mnsc.2023.00885 .

  PublisherDOI

• Search without Recall and Gaussian Learning: Structural Properties and Optimal Policies

  SSRN Electronic Journal · 2025-01-01

  preprintOpen accessSenior author
  PublisherDOI

• BEACON: Bayesian Optimal Stopping for Efficient LLM Sampling

  ArXiv.org · 2025-10-09

  preprintOpen access

  Sampling multiple responses is a common way to improve LLM output quality, but it comes at the cost of additional computation. The key challenge is deciding when to stop generating new samples to balance accuracy gains against efficiency. To address this, we introduce BEACON (Bayesian Efficient Adaptive Criterion for Optimal N-stopping), a principled adaptive sampling framework grounded in Sequential Search with Bayesian Learning. BEACON sequentially generates responses from the policy LLM, updates posterior belief over reward distributions in real time without further training, and determines when to stop by weighing expected gains against computational cost. Sampling terminates once the marginal utility of further exploration no longer justifies the expense. We establish both theoretical optimality guarantees and practical tractability, and show empirically that BEACON reduces average sampling by up to 80% while maintaining response quality. We further demonstrate BEACON's utility for cost-efficient preference data generation and outline practical extensions, offering actionable insights for future researchers.

  PublisherOA PDFDOI

• Designing Payment Models for the Poor

  Management Science · 2024-01-01 · 1 citations

  articleOpen accessSenior author
  PublisherOA PDFDOI

• Search in the Dark: The Case with Recall and Gaussian Learning

  Operations Research · 2024-08-16

  articleSenior author

  In “Search in the Dark: The Case with Recall and Gaussian Learning,” Manel Baucells and Saša Zorc address the classic sequential search problem where decision makers sample from a distribution to maximize rewards minus sampling costs. They focus on search with recall, where the reward is the highest sampled value. They solve the long-standing problem of determining optimal stopping rules when the sampling distribution is normal but both its mean and variance are unknown, and hence, they are progressively learned via sampling. The solution reveals that traditional methods, which rely on single reservation prices, are inadequate. The proposed approach offers a practical and efficient way to determine the optimal stopping rule. Relevant applications include job search, the sale of an asset, or technology adoption.

  PublisherDOI

• MP66-11 PREFERENCE SIGNALING IN THE UROLOGY MATCH: IMPACT AND USAGE TRENDS

  The Journal of Urology · 2023-03-23

  articleSenior author

  You have accessJournal of UrologyCME1 Apr 2023MP66-11 PREFERENCE SIGNALING IN THE UROLOGY MATCH: IMPACT AND USAGE TRENDS Ralph Grauer, Daniel Ranti, Kirsten Greene, Michael Gorin, Mani Menon, and Saša Zorc Ralph GrauerRalph Grauer More articles by this author , Daniel RantiDaniel Ranti More articles by this author , Kirsten GreeneKirsten Greene More articles by this author , Michael GorinMichael Gorin More articles by this author , Mani MenonMani Menon More articles by this author , and Saša ZorcSaša Zorc More articles by this author View All Author Informationhttps://doi.org/10.1097/JU.0000000000003329.11AboutPDF ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareFacebookLinked InTwitterEmail Abstract INTRODUCTION AND OBJECTIVE: Analyzing behavior from the implementation of signals during the American Urological Association (AUA) urology match will help clarify applicant/program signal usage and trends. METHODS: Using verified data from the match as well as survey data reported by applicants and programs, a logistic regression was performed on applicant factors associated with obtaining a residency interview: age, gender, degree (MD or DO), distribution of signals, US senior status, minority status, Latino status, IMG status, presence of a home urology program, AUA geographic section, and USMLE Step 1 score. We described signal distribution strategies stratified by program competitiveness and program behavior upon receipt of signals with respect to interviewing/ranking applicants. RESULTS: A total of 2,659 signals were sent by 553 candidates submitting rank lists for 364 positions at 142 programs. Programs received a median of 352 applications and were signaled to a median of 16 times each (IQR: 8–26). In a logistic regression predicting interview status, we found that geographic proximity (OR 3.25, 95% CI, 2.05–5.15; p=0.001), and signal status (OR 6.04, 95% CI, 3.50–10.40; p<0.001) were predictive of receiving an interview. Using multiple imputation to broadening the dataset, male gender (OR: 0.64, 95% CI, 0.45–0.92; p=0.039) and IMG status (OR: 0.35, 95% CI, 0.15–0.81; p=0.036) were negatively predictors, while MD degree (OR 2.36, 95% CI, 1.27–4.36; p=0.023), and US senior status (OR 1.91, 95% CI, 1.13–3.23; p=0.039), were positive predictors. CONCLUSIONS: We analyzed trends of the newly added signals. We reportedstrategies for signal dispersal, factors associated with obtaining interviews, and how signals informed program interviewing/ranking decisions. Source of Funding: none © 2023 by American Urological Association Education and Research, Inc.FiguresReferencesRelatedDetails Volume 209Issue Supplement 4April 2023Page: e936 Advertisement Copyright & Permissions© 2023 by American Urological Association Education and Research, Inc.MetricsAuthor Information Ralph Grauer More articles by this author Daniel Ranti More articles by this author Kirsten Greene More articles by this author Michael Gorin More articles by this author Mani Menon More articles by this author Saša Zorc More articles by this author Expand All Advertisement PDF downloadLoading ...

  PublisherDOI

• The When and How of Delegated Search

  Operations Research · 2023-08-14 · 9 citations

  article1st authorCorresponding

  Contract Design for Outsourcing Search Firms commonly outsource search for new employees, real estate, or technology to external agents. What should the contracts with these agents look like, and under which conditions should companies even hire agents (as opposed to doing the search in house)? These are the questions studied in “The when and how of delegated search” by Zorc et al. The authors find that the optimal contracts pay the agent a per-time fee as well as a bonus for finding an acceptable alternative. The size of this bonus is defined on signing of the contract and decreases over time. The decision of whether to outsource at all hinges on the firm’s trade-off between speed and quality; in-house search becomes optimal for a firm that prioritizes quality, but outsourcing offers better speed.

  PublisherDOI

• Characterization of Applicant Preference Signals, Invitations for Interviews, and Inclusion on Match Lists for Residency Positions in Urology

  JAMA Network Open · 2023 · 33 citations

  Senior authorCorresponding
  • Political Science
  • Family medicine
  • Psychology

  Importance: Preference signals were to be implemented in over 15 specialties during the 2022-2023 residency match. Analyzing results from the implementation of signals during the American Urological Association (AUA) urology match may inform future behavior. Objective: To characterize applicant and program signal usage and results in the Society of Academic Urology and AUA databases with respect to interview invites and rank list creation. Design, Setting, and Participants: This cohort study involved all applicants and residencies in the 2021-2022 AUA match with data analysis conducted in April through July 2022. Exposures: Five signals indicating interest. Main Outcomes and Measures: Using verified match and survey data reported by applicants and programs, a logistic regression was performed on applicant factors associated with obtaining an interview-the main outcome (using inclusion on rank list as a proxy): age, gender, degree (MD or DO), dispersal of signal, US senior status, racial minority group status, Latino ethnicity, international medical graduate status, presence of a home program, AUA geographic section, and US Medical Licensing Examination Step 1 score. Applicant signal dispersal strategies were stratified by applicant and program competitiveness, as well as program behavior upon receipt of signal with respect to extending interviews and rank list ordering of applicants. Results: A total of 2659 signals were sent by 553 candidates (mean [SD] age, 27.4 [2.9] years; 179 female [32.4%], 243 racial minority candidates [61.2%]) submitting rank lists for 364 positions at 143 programs. Programs received a median (IQR) of 352 (295-411) applications and were signaled to a median of 16 (8-26) times each. In a logistic regression estimating interview status, geographic proximity (OR, 3.25; 95% CI, 2.05-5.15; P = .001) and signal status (OR, 6.04; 95% CI, 3.50-10.40; P < .001) were associated with receiving an interview. Using multiple imputation by chained equations to impute missing data and broadening the data set, male gender (OR, 0.64; 95% CI, 0.45-0.92; P = .04) and international medical graduate status (OR, 0.35; 95% CI, 0.15-0.81; P = .04) were negative variables, while MD degree (OR, 2.36; 95% CI, 1.27-4.36; P = .02) and US senior status (OR, 1.91; 95% CI, 1.13-3.23; P = .04) were positive variables. Conclusions and Relevance: This study of the usage and trends of the newly added preference signals reported the most common strategies for signal dispersal; in an analysis of factors involved in obtaining an interview, geographic similarity between applicant and program and preference signal usage were associated with successful applications.

  PublisherOA PDFDOI

• From Black to Grey: Improving Access to Antimalarial Drugs in the Presence of Counterfeits

  SSRN Electronic Journal · 2023-01-01

  articleOpen accessSenior author
  PublisherDOI

Frequent coauthors

• Ilia Tsetlin

  INSEAD

  4 shared

• Manel Baucells

  University of Virginia

  4 shared

• Sameer Hasija

  INSEAD

  3 shared

• Stephen E. Chick

  INSEAD

  3 shared

• Daniel Ranti

  Icahn School of Medicine at Mount Sinai

  2 shared

• Mani Menon

  Icahn School of Medicine at Mount Sinai

  2 shared

• Michael A. Gorin

  2 shared

• Kirsten L. Greene

  University of Virginia

  2 shared

Education

• PhD, Decision Sciences

  INSEAD

  2018

• MBA

  Cotrugli Business School

  2010

• MSc, Mathematics

  University of Zagreb

  2008

Awards & honors

• Pierskalla Award
• IBM Service Science Award
• Decision Analysis Publication Award
• Junior Faculty Interest Group Award