{"paper_id":"2409.04903","title":"Semi-analytical pricing of options written on SOFR futures","authors":["Andrey Itkin","Yerkin Kitapbayev"],"paper_date":"2024-09-07","paper_type":"theoretical","paper_url":"https://arxiv.org/pdf/2409.04903v2","page_url":"https://thequant.space/flowcharts/semi-analytical-pricing-of-options-written-on-sofr-futures/","scores":{"empirical_rigor":3,"math_complexity":8.5,"quadrant":"Lab Rats"},"proposes":{"status":"available","statement":"The paper proposes a semi‑analytical approach for pricing options on SOFR futures, including both the American option on the SOFR forward and the arithmetic Asian‑American option when the underlying follows a time‑dependent CEV model.","quote":"In this paper, we propose a semi-analytical approach to pricing options on SOFR futures where the underlying SOFR follows a time-dependent CEV model.","summary":"In this paper, we propose a semi-analytical approach to pricing options on SOFR futures where the underlying SOFR follows a time-dependent CEV model. By definition, these options change their type at the beginning of the reference period: before this time, this is an American option written on a…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"The underlying SOFR rate follows a time‑dependent CEV stochastic process.","type":"model","stated":"explicit","quote":"In this paper, we propose a semi-analytical approach to pricing options on SOFR futures where the underlying SOFR follows a time-dependent CEV model. October 8, 2024 I n this paper, we propose a semi-analytical approach to pricing options on SOFR futures where the underlying SOFR follows a time-dependent CEV model.","support":"full"},{"assumption":"SOFR is assumed to be non‑negative for all times.","type":"market","stated":"explicit","quote":"Page 6 of 38 Semi-analytical pricing of options written on SOFR futures Having this in mind, in this paper we assume that SOFR is always non-negative, and, hence, set yl(t) = [−¯R∗(t)]+.","support":"full"},{"assumption":"The initial value of the auxiliary variable is y0 = 0.","type":"model","stated":"explicit","quote":"It is also assumed that y0 = 0, with t = 0 being the \"as of\" date for which the model is calibrated.","support":"full"},{"assumption":"Deterministic functions (σ, α, γ, R∗) are piecewise continuous and bounded.","type":"model","stated":"explicit","quote":"In [Turfus and Romero-Bermúdez, 2023] they all assumed to be piecewise continuous and bounded, while the first assumption can be relaxed.","support":"full"},{"assumption":"The exercise boundary and option price are obtained by solving a linear Volterra integral equation in the continuation region only.","type":"model","stated":"explicit","quote":"(15) The main idea of pricing American options by considering only the continuation region where the exercise boundary and the option price can be found simultaneously by solving a certain linear integral Volterra equation of the second type (LIVESK) together with a nonlinear algebraic equation, has been proposed and implemented in a series of papers [Carr and Itkin, 2021; Itkin and Muravey, 2024; Itkin, 2024].","support":"full"},{"assumption":"The numerical integration uses Simpson quadrature in time and trapezoid rule in space with specified grid sizes.","type":"execution","stated":"explicit","quote":"The corresponding integrals in the time τ ∈[0, τm] are approximated by using the Simpson quadratures on a 3D grid, which contains Nt = 40 nodes in T. The integrals in x and z are computed by using the trapezoid rule. We take Nx = 200 nodes in X, as it turns out that the further increase of Nx practically doesn’t affect the results.","support":"full"}]},"datasets":{"status":"none_found","items":[],"known":[],"sample":{"start":null,"end":null,"market":null,"asset_class":null,"frequency":null},"data_resolution":"synthetic"},"transaction_costs":{"status":"not_applicable","modeled":"n/a","headline_net_of_costs":null,"detail":null,"quote":null,"source":"front_matter","rule":"R3","candidates":{"front_matter":"n/a","evidence_card":"n/a","evidence_card_quoted":false},"meaning":"the paper reports no trading or execution result"},"results":{"status":"available","evaluation_type":"unclear","findings":[{"finding":"A semi‑analytical expression for the exercise boundary of the Asian‑American put option is derived for any strike.","kind":"theoretical","quote":"The term \"solve\" means the following: i) we provide a semi-analytical expression for the exercise boundary of the Asian-American Put option for any strike; ii) we also provide a semi-analytical expression for the Asian-European and Asian-American Put option prices valid when z ≥K. However, we leave an extension of this approach to the case z < K to be published elsewhere.","support":"full"},{"finding":"Semi‑analytical formulas for Asian‑European and Asian‑American put prices (valid for z ≥ K) are obtained.","kind":"theoretical","quote":"The term \"solve\" means the following: i) we provide a semi-analytical expression for the exercise boundary of the Asian-American Put option for any strike; ii) we also provide a semi-analytical expression for the Asian-European and Asian-American Put option prices valid when z ≥K. However, we leave an extension of this approach to the case z < K to be published elsewhere.","support":"full"},{"finding":"The proposed method reduces computational complexity to O(N²M^{1/2}H) while maintaining accuracy comparable to finite‑difference schemes.","kind":"performance","quote":"Despite M1 ≈M, however, L1 ≫H since the FD scheme needs all points in both continuation and exercise regions, while in our case we deal only with the continuation region. Also, the accuracy of our method can be easily increased by using higher-order quadratures, e.g., the Simpson integration rule with the fourth order of approximation in the grid step, [Itkin, Lipton, and Muravey, 2021].","support":"full"},{"finding":"Numerical experiments show that the method matches CRR tree prices for the log‑normal Black‑Scholes model and is faster than FD methods.","kind":"other","quote":"Put option K 90 92 94 96 98 100 Asian-American 0.5575 0.5189 0.5897 0.6364 0.6635 0.6957 European (Black-Scholes) 0.0431 0.1176 0.2814 0.6014 1.1682 2.0883 Table 2: Comparison of Asian-American Put option prices obtained by using our approach with those obtained by using a CRR tree method for the log-normal (Black-Scholes) model with the same model parameters. Note, that when computing an Asian-American option price","support":"full"},{"finding":"The approach can be combined with FD methods to first compute the exercise boundary (O(N²MI)) and then solve the continuation region (O(NM¹L¹)).","kind":"theoretical","quote":"It is even possible to combine these two methods, meaning you can first find the exercise boundary zB(τ, x) with the complexity O(N2MI), and then in the continuation region solve Eq. (10) with the European payoff by using the FD method with the complexity O(NM1L1).","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":22,"citations":{"cited_by_in_corpus":3,"by_independent_authors":0,"how_cited":{"used_as_baseline":0,"uses_method_or_data":0,"extends":0,"discusses_findings":0,"mentions":0,"unclassified":3},"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":"2026-11-14T19:55:40Z"},"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/"}