{"paper_id":"2609.38765","title":"Multiperiod bond portfolio optimization with transaction costs using a Markov Decision process","authors":["Balaji Ramachandran","Srikanth Iyer","Shashi Jain"],"paper_date":"2026-09-30","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2609.38765","page_url":"https://thequant.space/flowcharts/multiperiod-bond-portfolio-optimization-with-transaction-costs-using-a-markov/","scores":{"empirical_rigor":6,"math_complexity":8,"quadrant":"Holy Grail"},"proposes":{"status":"available","statement":"The paper proposes a simulation‑based multi‑period bond portfolio optimization framework using a time‑inhomogeneous Markov chain approximation of the Dynamic Nelson‑Siegel model and a finite‑horizon MDP with proportional transaction costs.","quote":"We develop a tractable simulation-based framework for multi-period bond portfolio optimization under interest-rate risk and proportional transaction costs.","summary":"Bank treasury portfolios must balance yield, liquidity, and interest-rate risk across bonds of different maturities. Static allocation rules are ill-suited to this task: portfolios concentrated in long-duration securities with no dynamic adjust- ment mechanism can accumulate large mark-to-market…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"Bonds are zero‑coupon with fixed maturities and the portfolio composition is fixed.","type":"market","stated":"explicit","quote":"Consider a finite horizon optimization problem of maximizing the gains from a portfolio consisting of K bonds over a discrete time horizon of T periods. The bonds are zero‑coupon bonds with maturities τ1 < τ2 < . . . < τK at time t = 0."},{"assumption":"Transaction costs are proportional to trade volume.","type":"execution","stated":"explicit","quote":"We consider different levels of transaction costs, c ∈[0, 0.01]."},{"assumption":"The objective is risk‑neutral expected terminal wealth.","type":"model","stated":"explicit","quote":"the investor maximizes expected terminal wealth by choosing among a discrete set of portfolio allocations at each rebalancing date."},{"assumption":"Yield‑curve dynamics follow the Dynamic Nelson‑Siegel model with VAR parameters from Pericoli and Taboga (2016).","type":"model","stated":"explicit","quote":"The parameters of the VAR process are taken from Pericoli and Taboga (2016)."},{"assumption":"State and action spaces are discretized, introducing approximation error.","type":"other","stated":"explicit","quote":"The discrete action and state spaces introduce approximation error, and the finite set of portfolio weights may preclude some beneficial adjustments available in continuous‑action settings."},{"assumption":"No market impact, bid‑ask spread, or regulatory liquidity constraints are modeled.","type":"execution","stated":"explicit","quote":"... models transaction costs as a simple proportional charge rather than incorporating bid‑ask spreads, market impact, or regulatory liquidity requirements."}]},"datasets":{"status":"available","items":[{"name":"Simulated VAR paths based on Pericoli and Taboga (2016) parameters","provider":null,"kind":"synthetic","access":"unknown","period":null,"universe":null,"quote":"We simulate the VAR process (8) with the parameters from Pericoli and Taboga (2016) and map the yields yt(τ) obtained from (7) to the nearest state St ∈S."}],"known":[],"sample":{"start":null,"end":null,"market":null,"asset_class":null,"frequency":null},"data_resolution":"synthetic"},"transaction_costs":{"status":"available","modeled":"yes","headline_net_of_costs":null,"detail":"Proportional costs c ranging from 0 to 0.01 (0 to 100 basis points) are incorporated in the MDP and in all reported performance tables.","quote":"We consider different levels of transaction costs, c ∈[0, 0.01].","source":"front_matter","rule":"R3","candidates":{"front_matter":"yes","evidence_card":"yes","evidence_card_quoted":true},"card_detail":"Proportional costs c ranging from 0 to 0.01 (0 to 100 basis points) are incorporated in the MDP and in all reported performance tables.","card_quote":"We consider different levels of transaction costs, c ∈[0, 0.01].","meaning":"costs are part of the reported headline results"},"results":{"status":"available","evaluation_type":"trading-pnl","findings":[{"finding":"The DP strategy outperforms equal‑weight buy‑and‑hold by 1.0–4.3 percentage points in mean terminal wealth over six months and by 3.1–9.5 points over twelve months, and outperforms the static mean‑variance portfolio by 0.7–4.0 and 2.7–9.0 points respectively.","kind":"performance","quote":"DP strategy outperforms equal-weight buy-and-hold by 1.0–4.3 percentage points in mean terminal wealth over six months and by 3.1–9.5 percentage points over twelve months, and the static mean‑variance portfolio by 0.7–4.0 and 2.7–9.0 percentage points respectively, with the range spanned by transaction costs from 100 b"},{"finding":"Higher transaction costs substantially reduce trading intensity: cumulative turnover falls from 6.34 at c=0 to 1.10 at c=0.01 (about five‑sixths reduction).","kind":"performance","quote":"cumulative turnover declines monotonically from 6.34 at c = 0 to 1.10 at c = 0.01, a reduction of roughly five‑sixths."},{"finding":"Total transaction‑cost burden peaks at intermediate cost levels (≈0.005) rather than at the highest cost level.","kind":"performance","quote":"rise from 0.291% of initial wealth at c = 0.0005 to a maximum of 1.434% at c = 0.005, and only then decline, to 1.096% at c = 0.01."},{"finding":"The DP strategy’s advantage is a return advantage rather than a risk reduction; fifth‑percentile wealth and maximum drawdown are not systematically improved relative to static benchmarks.","kind":"performance","quote":"DP does not systematically improve fifth‑percentile wealth or maximum drawdown relative to either static benchmark, which follows directly from the risk‑neutral objective."},{"finding":"The ordering of strategies is robust across upward‑sloping, near‑flat, and inverted initial yield curves.","kind":"other","quote":"the ordering of the three strategies is unchanged under upward‑sloping, near‑flat and inverted initial yield curves, so the result does not depend on a particular starting term structure."}],"reported_performance":null,"out_of_sample":{"method":"none","headline_out_of_sample":null}},"independent_evidence":{"status":"available","supports":[{"paper":{"paper_id":"2601.09127","title":"Robo-Advising in Motion: A Model Predictive Control Approach","paper_date":"2026-01-14","authors":["Tomasz R. Bielecki","Igor Cialenco"],"empirical_rigor":6.5,"page_url":"https://thequant.space/flowcharts/robo-advising-in-motion-a-model-predictive-control-approach/"},"question":"How does increasing transaction‑cost scaling affect portfolio turnover?","this_paper_finding":"Higher transaction costs substantially reduce trading intensity: cumulative turnover falls from 6.34 at c=0 to 1.10 at c=0.01 (about five‑sixths reduction).","other_paper_finding":"Increasing the transaction‑cost scaling factor η reduces portfolio turnover, with turnover vanishing for η > 0.05.","other_paper_quote":"With increasing η, the portfolio structure changes less in time, and the total turnover decreases and vanishing for η > 0.05.","note":"Both find turnover falls as costs rise: A sees turnover vanish for η>0.05; B sees turnover drop from 6.34 at c=0 to 1.10 at c=0.01 (≈5‑6× reduction).","confidence":0.901},{"paper":{"paper_id":"2306.08519","title":"A multi-agent targeted trading equilibrium with transaction costs","paper_date":"2023-06-14","authors":["Jin Hyuk Choi","Jetlir Duraj","Kim Weston"],"empirical_rigor":3,"page_url":"https://thequant.space/flowcharts/a-multi-agent-targeted-trading-equilibrium-with-transaction-costs/"},"question":"How do proportional transaction costs impact trading intensity and equilibrium price dynamics?","this_paper_finding":"Higher transaction costs substantially reduce trading intensity: cumulative turnover falls from 6.34 at c=0 to 1.10 at c=0.01 (about five‑sixths reduction).","other_paper_finding":"The level of proportional transaction costs influences the equilibrium stock price drift, with higher costs leading to less trading and a changed drift.","other_paper_quote":"The transaction cost level impacts the equilibrium stock price drift. Larger transaction cost proportions lead to less trading, which leads to changes in the stock price drift.","note":"Both find higher costs reduce trading: A reports less trading and altered drift; B shows turnover falls from 6.34 at c=0 to 1.10 at c=0.01 (~5/6 drop).","confidence":0.617},{"paper":{"paper_id":"2406.09578","title":"Dynamic Asset Allocation with Asset-Specific Regime Forecasts","paper_date":"2024-06-13","authors":["Yizhan Shu","Chenyu Yu","John M. Mulvey"],"empirical_rigor":8,"page_url":"https://thequant.space/flowcharts/dynamic-asset-allocation-with-asset-specific-regime-forecast/"},"question":"How does increasing trade‑aversion or transaction‑cost parameters affect portfolio turnover?","this_paper_finding":"Higher transaction costs substantially reduce trading intensity: cumulative turnover falls from 6.34 at c=0 to 1.10 at c=0.01 (about five‑sixths reduction).","other_paper_finding":"Sensitivity analysis shows that increasing the trade‑aversion parameter reduces turnover from 11.80 to 2.06 while slightly improving risk‑adjusted returns.","other_paper_quote":"The removal of the trading cost term markedly increases turnover, from 2.06 to 11.80. Therefore, using a non-zero trade aversion parameter primarily serves to optimize risk-adjusted returns slightly while substantially reducing turnover to a practical level.","note":"Both findings show turnover drops sharply with higher costs: A from 11.80 to 2.06, B from 6.34 to 1.10, despite different models and asset scopes.","confidence":0.432}],"contradicts":[],"qualifies":[],"compared_with":45,"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":1,"extracted_at":"2026-11-14T17:04:02Z"},"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/"}