{"paper_id":"2611.20163","title":"Log-optimal portfolios for market models with regime switching","authors":["Tahir Choulli","Xunbai Yin"],"paper_date":"2026-11-15","paper_type":"theoretical","paper_url":"https://arxiv.org/pdf/2611.20163","page_url":"https://thequant.space/flowcharts/log-optimal-portfolios-for-market-models-with-regime-switching/","scores":{"empirical_rigor":2,"math_complexity":9,"quadrant":"Lab Rats"},"proposes":{"status":"available","statement":"The paper introduces a general regime‑switching semimartingale market model (denoted \\(\\widehat{X}\\) or \\(bX\\)) that incorporates both switching coefficients and transition shocks, and quantifies how regime switching impacts the log‑optimal portfolio and its expected logarithmic utility.","quote":"Then we define a general switching model $\\widehat{X}$ that covers the existing cases, and study the impact of the stochasticity of the regime switching and its intrinsic risks on the optimal portfolio in many aspects. (1.4) Our last main contribution lies in measuring the impact of switching on the log-optimal portfolio.","summary":"This paper considers market models under regime switching mechanism with D states, and analyses the log-optimal portfolio. It is the portfolio, with finite expected utility, that maximizes the logarithmic utility from terminal wealth. We consider a family X=(X^(i))_i=1,...,D, of D-semimartingale…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"The underlying assets are modeled as D‑dimensional semimartingales X^{(i)} with predictable characteristics (b(i),c(i),F(i),A) for each regime i.","type":"model","stated":"explicit","quote":"We consider a family $\\mathbb{X}=(X^{(i)})_{i=1,...,D}$, of $D$-semimartingale models, among which our agent can switch to-and-from following a switching regime modelled by a continuous-time Markov Chain $α=(α_t)_{t\\geq 0}$ with $α_0=1$. 2.2 Our switching model and its parametrization Our mathematical model starts with D given market models represented by their return processes denoted by X := (X(i))i=1,...,D. Here","support":"full"},{"assumption":"All regimes share the same increasing process A, i.e., A(i)=A for every i.","type":"model","stated":"explicit","quote":"(2.7) Without loss of generality (it is enough to take A = PD i=1 A(i)+t and rewrite the characteristics accordingly), throughout the paper, we will assume A(i) = A, for any i ∈{1, ..., D} and dt ≪dA. We take A0 = 0.","support":"full"},{"assumption":"The switching process α is a continuous‑time Markov chain with initial state α_0=1.","type":"model","stated":"explicit","quote":"We consider a family $\\mathbb{X}=(X^{(i)})_{i=1,...,D}$, of $D$-semimartingale models, among which our agent can switch to-and-from following a switching regime modelled by a continuous-time Markov Chain $α=(α_t)_{t\\geq 0}$ with $α_0=1$. Recall that, in this paper, we assume that α(0) = 1 (i.e. the first regime is the state 1).","support":"full"},{"assumption":"The log‑optimal portfolio is considered only when the expected logarithmic utility is finite.","type":"other","stated":"explicit","quote":"It is the portfolio, with finite expected utility, that maximizes the logarithmic utility from terminal wealth.","support":"full"},{"assumption":"Existence of the log‑optimal portfolio requires the NUPBR condition for the pair (𝕏,α).","type":"market","stated":"explicit","quote":"On the one hand, we single out the necessary and sufficient conditions on the pair $(\\mathbb{X},α)$, which guarantee the existence of the optimal portfolio for $\\widehat{X}$. This makes NUPBR the minimal viability condition for modelling a frictionless competitive financial market [9, 30].","support":"full"},{"assumption":"Transition shocks γ_α·eJ are assumed to be locally integrable (γ_i·eJ ∈ M_loc).","type":"model","stated":"explicit","quote":"From now on, for each i ∈{1, 2, ..., D}, write qi := dt dAqα−iI{α−̸=i} for the intensity with respect to A, so that eJi = Ji −qi · A. Hence, our switching model is given by b X :=X + γα · eJ = X i I{α−=i} · X(i) | {z } switching coefficients + γα · eJ | {z } transition shock , γα := X i I{α−=i}γi, (2.8) where γi is an Rd×D-valued predictable process satisfying γi · eJ ∈Mloc, i ∈{1, 2, ..., D}.","support":"full"}]},"datasets":{"status":"none_found","items":[],"known":[],"sample":{"start":null,"end":null,"market":null,"asset_class":null,"frequency":null},"data_resolution":"none"},"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":"none","findings":[{"finding":"Necessary and sufficient conditions on (𝕏,α) for the existence of the numéraire portfolio (equivalently NUPBR) of the switching model are derived.","kind":"theoretical","quote":"(1.3) We characterize num´eraire portfolio for b X and its existence (or equivalently NUPBR) in terms of the models X(i) and the parameters of the switching regime.","support":"full"},{"finding":"A complete description of the set of all deflators for the switching model bX is provided.","kind":"theoretical","quote":"As a result, we describe the set of all deflators for b X in various manners.","support":"full"},{"finding":"Necessary and sufficient conditions for the existence of the log‑optimal portfolio of bX are established and an explicit characterization is given.","kind":"theoretical","quote":"When does the log-optimal portfolio of b X exist, and how can it be described?","support":"full"},{"finding":"The impact of regime switching on the optimal utility is quantified by isolating the main factors that perturb the increment Δ_T(bX, X^{(1)}) of expected logarithmic utility.","kind":"theoretical","quote":"5 The interplay between the time-inhomogeneity of (b(i), c(i), F (i)) and regime switching induces an addi-tional effect on the optimal utility, which is one of our principal findings. The non-negligible impact of common jumps on the optimal portfolio is another principal finding of this paper.","support":"full"},{"finding":"The model (2.8) is shown to subsume all existing regime‑switching semimartingale market models as special cases.","kind":"theoretical","quote":"Our switching model (2.8), which is parametrized by the family of quadruplets (b(i), c(i), F (i), γi)i=1,..,D, is the most general regime switching semimartingale market model. In fact, when γα ≡0, then b X = X. 4 Thus, any result for the market b X obtained later automatically applies to the market X by letting γα ≡0.","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":18,"citations":{"cited_by_in_corpus":0,"by_independent_authors":0,"how_cited":{"used_as_baseline":0,"uses_method_or_data":0,"extends":0,"discusses_findings":0,"mentions":0,"unclassified":0},"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":3,"extracted_at":"2026-11-17T15:29:04Z"},"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/"}