{"paper_id":"2501.12591","title":"Optimal Rebate Design: Incentives, Competition and Efficiency in Auction Markets","authors":["Thibaut Mastrolia","Tianrui Xu"],"paper_date":"2025-01-22","paper_type":"theoretical","paper_url":"https://arxiv.org/pdf/2501.12591v1","page_url":"https://thequant.space/flowcharts/optimal-rebate-design-incentives-competition-and-efficienc/","scores":{"empirical_rigor":2.5,"math_complexity":9,"quadrant":"Lab Rats"},"proposes":{"status":"available","statement":"The paper proposes a continuous‑time principal‑multi‑agent model of auction markets and computes optimal transaction‑fee and rebate policies using a Deep BSDE method to solve the high‑dimensional HJB equation with jumps.","quote":"This study explores the design of an efficient rebate policy in auction markets, focusing on a continuous-time setting with competition among market participants. To numerically compute the optimal rebate and transaction fee policies, we apply the Deep BSDE method.","summary":"This study explores the design of an efficient rebate policy in auction markets, focusing on a continuous-time setting with competition among market participants. In this model, a stock exchange collects transaction fees from auction investors executing block trades to buy or sell a risky asset…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"Time is modeled as a continuous variable and market makers play a Nash‑equilibrium game.","type":"model","stated":"explicit","quote":"In contrast to our earlier work, where we discretize the time span and simplify trader interactions, this paper models time as a continuous variable and considers a Nash equilibrium game between two strategic traders. • Deep learning methods.","support":"full"},{"assumption":"Strategic market makers control both their order‑arrival intensity λ and the quoted spread µ.","type":"market","stated":"explicit","quote":"2.1.1 Strategic trader: auction market makers We consider two strategic traders in this auction: Market Maker p and Market Maker q sending orders according to right-continuous and adapted counting processes Np and Nq with intensity process λp and λq respectively. The strategic traders are called ”strategic” because they control both their arriving intensity λ and the spread µ.","support":"full"},{"assumption":"Non‑strategic traders incur a fixed transaction fee d per order at the end of the auction.","type":"execution","stated":"explicit","quote":"Non-strategic traders, on the other hand, incur a fixed transaction fee d per order at the end of the auction, for some d ≥0.","support":"full"},{"assumption":"The intensity functions λa(µ,d) and λb(µ,d) are monotone (non‑increasing in µ and d) with baseline λ0.","type":"model","stated":"explicit","quote":"To characterize the market impact of the strategic traders and the impact of the fixed incentive d on non-strategic traders, we assume λa t = λa(µt, d) := λ0ia(µp t , µq t)ι(d) and λb t = λb(µt, d) := λ0ib(µp t , µq t)ι(d), for some constant λ0, some monotonically non-increasing function ia, non-decreasing function ib and non-increasing function ι with ia(0, 0) = ib(0, 0) = ι(0) = 1.","support":"full"},{"assumption":"The efficient price follows dP*ₜ = σ dWₜ with constant σ.","type":"model","stated":"explicit","quote":"Efficient price The market trades a risky asset with efficient price P ∗defined as: dP ∗ t = σdWt, t ∈[0, T], where W is a Brownian motion and σ a fixed constant.","support":"full"},{"assumption":"Strategic traders prefer their intensity to stay close to a target λ₀ (benchmark strategy).","type":"market","stated":"explicit","quote":"In addition, we assume that a strategic traders do not want the intensity process to deviate too much from some target intensity λ0 fixed.","support":"full"}]},"datasets":{"status":"available","items":[{"name":"Apple and Alphabet auction data","provider":"YahooFinance","kind":"market-prices","access":"public","period":"Oct-2-2023 to Dec-29-2023","universe":"Apple, Alphabet","quote":"We calibrate P ∗ 0 and σ by using the trading data of ”Apple” and ”Alphabet” extracted from YahooFinance on period Oct-2-2023 to Dec-29-2023. The calibration method is illustrated in [Mastrolia and Xu, 2024, Section 2.4].","support":"full","known_id":"yahoo-finance"}],"known":[{"id":"yahoo-finance","name":"Yahoo Finance","provider":"Yahoo","kind":"market-prices","access":"public","quote":"We calibrate P ∗ 0 and σ by using the trading data of ”Apple” and ”Alphabet” extracted from YahooFinance on period Oct-2-2023 to Dec-29-2023.","mentions":1,"used_in":1}],"sample":{"start":null,"end":null,"market":null,"asset_class":null,"frequency":null},"data_resolution":"daily"},"transaction_costs":{"status":"available","modeled":"yes","headline_net_of_costs":null,"detail":"A fixed transaction fee d per order is charged to non‑strategic traders and enters the exchange's objective function.","quote":"Non-strategic traders, on the other hand, incur a fixed transaction fee d per order at the end of the auction, for some d ≥0.","source":"evidence_card","rule":"R2","candidates":{"front_matter":"unclear","evidence_card":"yes","evidence_card_quoted":true},"card_detail":"A fixed transaction fee d per order is charged to non‑strategic traders and enters the exchange's objective function.","card_quote":"Non-strategic traders, on the other hand, incur a fixed transaction fee d per order at the end of the auction, for some d ≥0.","meaning":"costs are part of the reported headline results"},"results":{"status":"available","evaluation_type":"none","findings":[{"finding":"The optimal transaction fee that minimizes market inefficiency is ˆd = 3.0 for Apple and ˆd = 2.0 for Alphabet.","kind":"performance","quote":"5.2 Illustration and Policy Recommendations In terms of the optimal d which minimizes ρ(d), we find ˆd = 3.0 for Apple and ˆd = 2.0 for Alphabet.","support":"full"},{"finding":"Optimal rebate and fee structures narrow the spread between the auction clearing price and the efficient price, improving market efficiency.","kind":"theoretical","quote":"Our results show that optimal transaction fees and rebate structures improve market efficiency by narrowing the spread between the auction clearing price and the asset's fundamental value, while ensuring a minimal gain for both market makers indexed on the price of the asset on a coexisting limit order book.","support":"full"},{"finding":"The Deep BSDE method successfully computes the optimal incentive (rebate) function ˆξ(d) for the given fee levels.","kind":"theoretical","quote":"To address this, we present a numerical method designed to explicitly find the optimal incentive structure and optimal fee d, using the Deep BSDE method, see [Henry-Labordere, 2017, Han and Long, 2020, Ji et al., 2020, Ji et al., 2022]. In the following, we explain how ˆξ(d) is solved with the Deep BSDE method, assuming d is fixed.","support":"full"}],"reported_performance":null,"out_of_sample":{"method":"none","headline_out_of_sample":null}},"independent_evidence":{"status":"none_found","supports":[],"contradicts":[],"qualifies":[],"compared_with":14,"citations":{"cited_by_in_corpus":2,"by_independent_authors":0,"how_cited":{"used_as_baseline":0,"uses_method_or_data":0,"extends":0,"discusses_findings":0,"mentions":0,"unclassified":2},"independent_papers":[],"scope":"Citations between papers in this corpus only; citations from outside it are not counted."}},"implementation":{"status":"available","public":false,"code_url":null,"note":"No repository presented by the authors as their own was found in the paper. Absence here is not proof that no code exists."},"replication":{"status":"none_found","card":null,"note":"No replication of this paper has been run by thequant.space yet."},"extraction":{"text_source":"fulltext","model":"openai/gpt-oss-120b","version":2,"extracted_at":"2026-11-14T19:52:47Z"},"provenance":{"claims":"Everything here is what the authors state in the paper. Nothing was re-run or reproduced by the platform.","quotes":"Assumptions, datasets, cost detail and findings were extracted by a language model. Every quote is verbatim paper text. Where an item has `support`, \"full\" means every figure in it appears in the quoted sentences and \"partial\" means a figure is stated elsewhere in the paper; items with a figure found nowhere in the paper were discarded.","datasets_known":"datasets.known lists well-known vendors and public datasets by a stable id, found by a dictionary over the full text where the sentence says the data is used; it is independent of the model and covers only names in the dictionary.","flags":"transaction_costs.modeled, results.evaluation_type and results.out_of_sample.method are platform classifications (structured-decision models); see https://thequant.space/score-guide/#penalty-flags","comparisons":"supports / contradicts / qualifies are a language model's reading of two papers' quoted findings, restricted to closely related papers that share no author. They are pointers to read, not verified replications, and coverage is partial.","implementation":"code_url is a repository the paper itself presents as its own (rule-scored, ambiguous cases judged by a model), checked for liveness when it was recorded."},"docs":"https://thequant.space/agents/"}