{"paper_id":"2503.18609","title":"Asset pre-selection for a cardinality constrained index tracking portfolio with optional enhancement","authors":["N. Meade","C. A. Valle","J. E. Beasley"],"paper_date":"2025-03-24","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2503.18609v1","page_url":"https://thequant.space/flowcharts/asset-pre-selection-for-a-cardinality-constrained-index-trac/","scores":{"empirical_rigor":7,"math_complexity":6.5,"quadrant":"Holy Grail"},"proposes":{"status":"available","statement":"The paper proposes a two‑step approach: first pre‑select assets for a given cardinality, then estimate portfolio weights on the selected assets.","quote":"We divide the problem into two steps - (1) pre-selection of assets; (2) estimation of weights on the assets chosen.","summary":"An index tracker is a passive investment reproducing the return and risk of a market index, an enhanced index tracker offers a return greater than the index. We consider the selection of a portfolio of given cardinality to track an index, both without and with enhancement. We divide the problem…","strategy":"BE-OLS(n) cardinality-constrained index tracking"},"assumptions":{"status":"available","items":[{"assumption":"OLS regression assumes asset returns follow a multivariate normal distribution.","type":"model","stated":"explicit","quote":"Ordinary least squares (OLS) regression assumes a multivariate normal distribution for asset returns, where asset return behaviour is completely described by a mean vector and a covariance matrix"},{"assumption":"LAD regression assumes returns are Laplace distributed with fat tails.","type":"model","stated":"explicit","quote":"Least absolute deviation (LAD) regression assumes Laplace distributed returns, where the distribution of returns has fatter tails than the normal distribution"},{"assumption":"Transaction costs are proportional to the sum of absolute changes in asset holdings (transaction volume).","type":"execution","stated":"explicit","quote":"assume that a component of the transaction costs incurred when rebalancing from period K to period K+1 is proportional to ∑"},{"assumption":"Penalty term forces the portfolio to match the index’s total return over the rebalancing period.","type":"model","stated":"explicit","quote":"To ensure that the tracking portfolio reproduces the same return as the index, we introduce a penalty function"},{"assumption":"Weight on the penalty term is set to 5 for the optimisation.","type":"model","stated":"explicit","quote":"In our analysis limited computational experience indicated that setting weight to 5 is effective."},{"assumption":"Only assets that are index members at the time of each rebalance are considered.","type":"data","stated":"explicit","quote":"we restrict attention at each rebalance only to those assets that are index members at the time of the rebalance"}]},"datasets":{"status":"available","items":[{"name":"S&P 500 daily asset price dataset","provider":null,"kind":"market-prices","access":"public","period":"3/1/2005 to 29/12/2023","universe":"822 assets that have been members of the S&P 500 Composite Price Index","quote":"Our full data set is daily asset prices for 822 assets which have been members of the S&P 500 Composite Price Index (at some time or another) from 3/1/2005 to 29/12/2023"},{"name":"Cardinality Portfolio Data repository","provider":"GitHub (cristianoarbex)","kind":"benchmark","access":"public","period":null,"universe":null,"quote":"the dataset used here is publicly available for use by future researchers on https://github.com/cristianoarbex/cardinalityPortfolioData"}],"known":[],"sample":{"start":"2005","end":"2023","market":"S&P 500 stocks","asset_class":"equities","frequency":"daily"},"data_resolution":"daily"},"transaction_costs":{"status":"available","modeled":"yes","headline_net_of_costs":null,"detail":"A component of transaction costs is assumed proportional to the transaction volume (sum of absolute changes in holdings).","quote":"assume that a component of the transaction costs incurred when rebalancing from period K to period K+1 is proportional to ∑","source":"evidence_card","rule":"R2","candidates":{"front_matter":"unclear","evidence_card":"yes","evidence_card_quoted":true},"card_detail":"A component of transaction costs is assumed proportional to the transaction volume (sum of absolute changes in holdings).","card_quote":"assume that a component of the transaction costs incurred when rebalancing from period K to period K+1 is proportional to ∑","meaning":"costs are part of the reported headline results"},"results":{"status":"available","evaluation_type":"trading-pnl","findings":[{"finding":"Out‑of‑sample tracking error declines roughly with 1/√cardinality.","kind":"performance","quote":"Avoiding these restrictions, we show that out-of-sample tracking errors are roughly proportional to 1/sqrt(cardinality)."},{"finding":"OLS‑based pre‑selection outperforms LAD; backward elimination slightly beats forward selection; omitting a constant term is marginally better.","kind":"statistical","quote":"We find OLS more effective than LAD; BE marginally more effective than FS; (n) marginally more effective than (c)."},{"finding":"For a plain index tracker, out‑of‑sample tracking error, transaction volume and return‑risk ratios improve as cardinality increases.","kind":"performance","quote":"For a S&P 500 index tracker, we find that out-of-sample tracking error, transaction volume and return-risk ratios all improve as cardinality increases."},{"finding":"For enhanced‑return portfolios, cardinalities of about 10‑20 assets give the best results.","kind":"performance","quote":"By contrast for enhanced returns, cardinalities of the order 10 to 20 are most effective."},{"finding":"The proportion of assets common between locally optimal tracking portfolios is less than half, indicating a flat optimum surface.","kind":"statistical","quote":"the proportion of assets common to similarly performing, locally optimal, tracking portfolios"}],"reported_performance":{"strategy":"BE-OLS(n) cardinality-constrained index tracking","sharpe":0.474,"sharpe_benchmark":0.474,"annual_return_pct":null,"total_return_pct":null,"max_drawdown_pct":null},"out_of_sample":{"method":"walk-forward","headline_out_of_sample":true}},"independent_evidence":{"status":"none_found","supports":[],"contradicts":[],"qualifies":[],"compared_with":35,"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":"2026-11-14T17:41:55Z"},"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/"}