{"paper_id":"2305.18136","title":"Exponential Utility Maximization in a Discrete Time Gaussian Framework","authors":["Yan Dolinsky","Or Zuk"],"paper_date":"2023-05-29","paper_type":"methodological","paper_url":"https://arxiv.org/pdf/2305.18136v2","page_url":"https://thequant.space/flowcharts/exponential-utility-maximization-in-a-discrete-time-gaussian/","scores":{"empirical_rigor":1.5,"math_complexity":8.5,"quadrant":"Lab Rats"},"proposes":{"status":"available","statement":"The paper proposes a solution to the discrete‑time exponential utility maximization problem for a multivariate normal risky asset when the investor observes price changes with a delay.","quote":"The aim of this short note is to present a solution to the discrete time exponential utility maximization problem in a case where the underlying asset has a multivariate normal distribution. In addition to the usual setting considered in Mathematical Finance, we also consider an investor who is informed about the risky asset's price changes with a delay.","summary":"The aim of this short note is to present a solution to the discrete time exponential utility maximization problem in a case where the underlying asset has a multivariate normal distribution. In addition to the usual setting considered in Mathematical Finance, we also consider an investor who is…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"The risky asset price vector (S1,…,Sn) follows a non‑degenerate multivariate normal distribution.","type":"model","stated":"explicit","quote":"We assume that the initial stock price S0 is a known constant and the random vector (S1, ..., Sn) has a non-degenerate multivariate normal dis-tribution.","support":"full"},{"assumption":"The increments (S1−S0, S2−S1,…,Sn−Sn‑1) are jointly normal with mean µ and covariance Σ.","type":"model","stated":"explicit","quote":"Hence, (S1 −S0, S2 −S1, ..., Sn −Sn−1) ∼N(µ, Σ) Date: June 27, 2023.","support":"full"},{"assumption":"The investor’s bank account bears no interest.","type":"market","stated":"explicit","quote":"We assume that the investor has a bank account that, for simplicity, bears no interest.","support":"full"},{"assumption":"Absolute risk aversion is set to α=1 without loss of generality.","type":"other","stated":"explicit","quote":"Thus, without loss of generality, we take the risk aversion α = 1.","support":"full"},{"assumption":"The delay D induces banded positive‑definite matrices Q (bandwidth D) and symmetric matrices Γ with zeros for |i−j|≤D.","type":"model","stated":"explicit","quote":"Let SD ⊂Mn(R) be the set of all positive definite matrices Q ∈Mn(R) that satisfy Qij = 0 for |i −j| > D. Namely, SD is the set of all banded positive definite matrices with lower bandwidth and upper bandwidth equal to D. Let TD ⊂Mn(R) be the set of all symmetric matrices Γ ∈Mn(R) that satisfy Γij = 0 for |i −j| ≤D. We arrive at the main result of the paper. Theorem 1.1.","support":"full"},{"assumption":"The time horizon satisfies n≥3.","type":"other","stated":"explicit","quote":"Assume that n ≥3. The matrix Σ has a determinant equal to |Σ| = \u00001 −ρ2\u0001n−1, and a simple tridi-agonal inverse that is given by (see [6]) Λij =            0, |i −j| > 1 − ρ 1−ρ2 , |i −j| = 1 1+ρ2 1−ρ2 , 1 < i = j < n 1 1−ρ2 , otherwise. Case I: D = 0.","support":"full"}]},"datasets":{"status":"available","items":[{"name":"Synthetic Kac‑Murdock‑Szegö covariance matrix Σij:=ρ|i−j| (Example 1.5)","provider":null,"kind":"synthetic","access":"unknown","period":null,"universe":null,"quote":"We first examined the Kac-Murdock-Szeg¨o covariance matrix Σij := ρ|i−j| from Example 1.5.","support":"full"},{"name":"Synthetic discretized fractional Brownian motion covariance matrix (3.3)","provider":null,"kind":"synthetic","access":"unknown","period":null,"universe":null,"quote":"A fractional Brownian motion (fBm) [12] with Hurst parameter H ∈(0, 1) is a centered Gaussian process BH t , t ≥0 with the covariance function (3.2) Cov \u0000BH s , BH t \u0001 = 1 2 \u0000t2H + s2H −|t −s|2H\u0001 . In this case µ = 0 and by (3.2) the covariance matrix is the Toeplitz matrix given by Σij = Cov \u0010 BH (i+1)∆t −BH i∆t, BH (j+1)∆t −BH j∆t \u0011 = ∆t2H 2 \u0000|i −j −1|2H + |i −j + 1|2H −2|i −j|2H\u0001 .","support":"full"}],"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":"For delay D=0 the optimal strategy is γ̂1=0 and γ̂i=ρ/(1−ρ²)(Si‑1‑Si‑2) for i>1, with value −q/(1−ρ²)(1+ρ²)^{n‑2}.","kind":"theoretical","quote":"From Corollary 1.3 we obtain that the optimal strategy is ˆγ1 = 0, and ˆγi = ρ 1−ρ2 (Si−1 −Si−2) for i > 1.","support":"full"},{"finding":"For delay D=1 the optimal strategy is γ̂1=γ̂2=0 and γ̂i given by a weighted sum of past returns (formula in [21]), with value −q(1−ρ²)(1+ρ²)^{n‑2}(1+ρ²+ρ⁴)^{n‑3}.","kind":"theoretical","quote":"Corollary 1.3 gives the optimal strategy ˆγ1 = ˆγ2 = 0 and ˆγi = 1 + ρ2 1 −ρ2 i−2 X j=1 \u0012 − ρ 1 + ρ2 \u0013i−j (Sj −Sj−1) for i > 2.","support":"partial"},{"finding":"Numerical experiments show that the value decreases as the delay D increases for both the Kac‑Murdock‑Szegö and fractional Brownian motion models.","kind":"performance","quote":"The results, displayed in Fig. 1, show the decrease in the value as D in increasing for both models.","support":"full"},{"finding":"For high Hurst parameters (H>0.5) the value increases with D, while for H=0.5 (standard Brownian motion) the value drops to –1 regardless of delay.","kind":"performance","quote":"For high H values (H > 0.5), the increments are positively correlated hence the value increases with D towards the maximum 0. For H = 0.5 we reduce to the independent Brownian motion case, hence the value drops to −1 regardless of the delay.","support":"full"},{"finding":"When scaling D with n (Dn/n fixed), the value converges to a negative number (e.g., –0.283 for n=256, –0.228 for n=1024).","kind":"performance","quote":"while keeping the ratio Dn n fixed amounts to refining the discretization towards the continuous limit with a fixed continuous delay 0 < D < T. The value appears to converge to a negative number for this case (i.e. for Dn n = 1 64), with −0.283 for n = 256 and −0.228 for n = 1024.","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":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":2,"extracted_at":"2026-11-14T20:02:47Z"},"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/"}