{"paper_id":"2406.17155","title":"Optimizing Sparse Mean-Reverting Portfolio","authors":["Sung Min Yoon"],"paper_date":"2024-06-24","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2406.17155v1","page_url":"https://thequant.space/flowcharts/optimizing-sparse-mean-reverting-portfolio/","scores":{"empirical_rigor":7,"math_complexity":8.5,"quadrant":"Holy Grail"},"proposes":{"status":"available","statement":"The paper proposes an SDP-based optimization to find optimal stock weights for a fast mean-reverting portfolio with minimum variance and sparsity constraints.","quote":"We further add minimum variance and sparsity constraints to the optimization problem and transform into Semidefinite Programming (SDP) problem to find the optimal weights.","summary":"Mean-reverting behavior of individuals assets is widely known in financial markets. In fact, we can construct a portfolio that has mean-reverting behavior and use it in trading strategies to extract profits. In this paper, we show that we are able to find the optimal weights of stocks to construct…","strategy":"sparse mean-reverting portfolio with minimum variance constraint"},"assumptions":{"status":"available","items":[{"assumption":"The process of portfolio returns is stationary with zero mean.","type":"model","stated":"explicit","quote":"we want the process (y^T x_t) to be a stationary process, which is one of the most popular processes that has mean-reverting behavior."},{"assumption":"The expectation of the process is zero, achieved by subtracting the mean.","type":"statistical","stated":"explicit","quote":"Using this canonical form of the process and assuming that the expectation of the process is 0 (this can be achieved through subtracting the mean for each value x_t and"},{"assumption":"Fixed transaction costs are negligible; costs are proportional to position changes and number of assets.","type":"execution","stated":"explicit","quote":"For transaction costs, we assume that fixed transaction costs are negligible. However, we consider that there exist transaction costs that are incurred each trading date and are proportional to the change in position and number of assets in the portfolio."},{"assumption":"Predictability estimator is appropriate for measuring mean-reversion speed.","type":"model","stated":"explicit","quote":"we focus on predictability estimator, which is the most popular estimator for stationary processes, in our optimization problems."},{"assumption":"Randomly selected 12 stocks from the S&P 500 are sufficient to represent the market universe.","type":"data","stated":"explicit","quote":"In our research, we have randomly selected 12 stocks among 448 stocks in S&P 500 to construct initial portfolio"}]},"datasets":{"status":"available","items":[{"name":"Option implied volatility data of S&P 500","provider":null,"kind":"options","access":"unknown","period":"2010-2020","universe":"S&P 500 constituents","quote":"we have empirically tested how the optimal portfolios perform by using the option implied volatility data of S&P 500 from 2010 to 2020."},{"name":"Randomly selected 12 stocks from S&P 500","provider":null,"kind":"market-prices","access":"unknown","period":null,"universe":"12 stocks out of 448 S&P 500 stocks","quote":"we have randomly selected 12 stocks among 448 stocks in S&P 500 to construct initial portfolio"}],"known":[],"sample":{"start":"2010","end":"2020","market":"S&P 500 stocks","asset_class":"options","frequency":"daily"},"data_resolution":"daily"},"transaction_costs":{"status":"available","modeled":"yes","headline_net_of_costs":true,"detail":"Costs varied as 0.04, 0.08, 0.12, 0.16 cents per contract (tc ∈ {0.0004,0.0008,0.0012,0.0016}) and are applied as tc * k * |N_t - N_{t-1}|.","quote":"In each trading date, we assume the total transaction cost to be tc * k * |N_t − N_{t-1}| where tc is the transaction cost (tc ∈ {0.0004, 0.0008, 0.0012, 0.0016}), k is the number of assets in the portfolio","source":"front_matter","rule":"R3","candidates":{"front_matter":"yes","evidence_card":"yes","evidence_card_quoted":true},"card_detail":"Costs varied as 0.04, 0.08, 0.12, 0.16 cents per contract (tc ∈ {0.0004,0.0008,0.0012,0.0016}) and are applied as tc * k * |N_t - N_{t-1}|.","card_quote":"In each trading date, we assume the total transaction cost to be tc * k * |N_t − N_{t-1}| where tc is the transaction cost (tc ∈ {0.0004, 0.0008, 0.0012, 0.0016}), k is the number of assets in the portfolio","meaning":"costs are part of the reported headline results"},"results":{"status":"available","evaluation_type":"trading-pnl","findings":[{"finding":"Without transaction costs, the non‑sparse portfolio yields a 1.2% return over 14 months, 0.2% higher than the sparse portfolio.","kind":"performance","quote":"The optimal portfolio without sparsity and variance constraint has generated 1.2% return over the 14-month period, which is 0.2% larger than the return from sparse optimal portfolio."},{"finding":"With transaction costs, the sparse portfolio outperforms the non‑sparse portfolio, and its returns remain positive while the non‑sparse returns become negative when costs exceed 0.10 cents per contract.","kind":"performance","quote":"... when transaction cost that are above 0.10 cents per contract, the payoff from the sparse optimal portfolio still remains positive while the payoffs from the non-sparse optimal portfolio is negative."},{"finding":"Overall returns are low: under 1.2% with no transaction costs and under 1% with transaction costs over the 14‑month period.","kind":"performance","quote":"the returns were quite low with under 1.2% with no transaction costs and under 1% with transaction costs over the 14-month period."}],"reported_performance":{"strategy":"sparse mean-reverting portfolio with minimum variance constraint","sharpe":null,"sharpe_benchmark":null,"annual_return_pct":null,"total_return_pct":null,"max_drawdown_pct":null},"out_of_sample":{"method":"holdout","headline_out_of_sample":true}},"independent_evidence":{"status":"none_found","supports":[],"contradicts":[],"qualifies":[],"compared_with":14,"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":"2026-11-14T18:07:20Z"},"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/"}