{"paper_id":"2010.01157","title":"Gold Standard Pairs Trading Rules: Are They Valid?","authors":["Miroslav Fil"],"paper_date":"2020-10-02","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2010.01157","page_url":"https://thequant.space/flowcharts/gold-standard-pairs-trading-rules-are-they-valid/","scores":{"empirical_rigor":8,"math_complexity":5,"quadrant":"Holy Grail"},"proposes":{"status":"available","statement":"The paper applies the two most common pairs‑trading formation approaches, the distance and cointegration methods, to US equities from 1990 to 2020.","quote":"We employ the most common distance and cointegration methods on US equities from 1990 to 2020 including the Covid-19 crisis.","summary":"Pairs trading is a strategy based on exploiting mean reversion in prices of securities. It has been shown to generate significant excess returns, but its profitability has dropped significantly in recent periods. We employ the most common distance and cointegration methods on US equities from 1990…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"The analysis uses daily price data for US stocks listed on the NYSE.","type":"data","stated":"explicit","quote":"Data We use daily data on US stocks traded on NYSE spanning from 1990/01/01 to 2020/06/01.","support":"full"},{"assumption":"Only liquid stocks are retained (bottom decile by liquidity excluded).","type":"data","stated":"explicit","quote":"We further pre-process the dataset by excluding bottom decile stocks in terms of liquidity, and removing any stocks that were not traded during the formation period for at least one trading day.","support":"full"},{"assumption":"Pairs are assumed to exhibit a stable long‑run relationship during the formation period.","type":"statistical","stated":"explicit","quote":"In the forma-tion period, a statistical criterion is utilized to determine which pairs exhibit a stable long-run relationship.","support":"full"},{"assumption":"A one‑period trade execution lag is used to model bid‑ask spread and execution difficulties.","type":"execution","stated":"explicit","quote":"Again following Gatev et al. (1999), we also utilize a one-period trade execution lag in order to estimate the bid-ask spread and other trade exe-cution diﬃculties.","support":"full"},{"assumption":"Trading rules use fixed spread thresholds of ±2 (standard deviations) to open positions.","type":"model","stated":"explicit","quote":"In particular, we short the spread when its normalized value exceeds 2, and close the position once it crosses zero or at the end of trading. We proceed analogically when it goes below -2.","support":"full"},{"assumption":"Transaction costs are assumed to be one‑way costs decreasing from 35 bps in 1990 to 26 bps in 2020, plus a short‑selling fee of 0.6 % p.a.","type":"execution","stated":"explicit","quote":"We further extrapolate the provided analysis and work with one-way transaction costs ranging from 35bps in 1990 to 26bps in 2020. Short-selling costs follow the work of D’Avolio (2002) who found an average fee of 0.6% p.a. for US equities in 2000-2001, which we apply over the whole period.","support":"full"}]},"datasets":{"status":"available","items":[{"name":"Daily US stock prices (NYSE)","provider":"NYSE","kind":"market-prices","access":"public","period":"1990-01-01 to 2020-06-01","universe":"≈2800 US stocks","quote":"Data We use daily data on US stocks traded on NYSE spanning from 1990/01/01 to 2020/06/01. In total, the dataset contains roughly 2800 stocks.","support":"full"}],"known":[],"sample":{"start":null,"end":null,"market":null,"asset_class":null,"frequency":null},"data_resolution":"daily"},"transaction_costs":{"status":"available","modeled":"yes","headline_net_of_costs":null,"detail":"One‑way transaction costs ranging from 35 bps in 1990 to 26 bps in 2020; short‑selling cost 0.6 % p.a.","quote":"We further extrapolate the provided analysis and work with one-way transaction costs ranging from 35bps in 1990 to 26bps in 2020. Short-selling costs follow the work of D’Avolio (2002) who found an average fee of 0.6% p.a. for US equities in 2000-2001, which we apply over the whole period.","source":"front_matter","rule":"R3","candidates":{"front_matter":"yes","evidence_card":"yes","evidence_card_quoted":true},"card_detail":"One‑way transaction costs ranging from 35 bps in 1990 to 26 bps in 2020; short‑selling cost 0.6 % p.a.","card_quote":"We further extrapolate the provided analysis and work with one-way transaction costs ranging from 35bps in 1990 to 26bps in 2020. Short-selling costs follow the work of D’Avolio (2002) who found an average fee of 0.6% p.a. for US equities in 2000-2001, which we apply over the whole period.","meaning":"costs are part of the reported headline results"},"results":{"status":"available","evaluation_type":"trading-pnl","findings":[{"finding":"The overall pairs‑trading strategy fails to outperform the market benchmark even with hyper‑parameter tuning, though it performs very strongly during bear markets.","kind":"performance","quote":"The strategy overall fails to outperform the market benchmark even with hyperparameter tuning, but it performs very strongly during bear markets.","support":"full"},{"finding":"Market factors have a strong relationship with the optimal parametrization, indicating that adjustments are needed for modern market conditions.","kind":"other","quote":"Furthermore, we demonstrate that market factors have a strong relationship with the optimal parametrization for the strategy, and adjustments are appropriate for modern market conditions.","support":"full"},{"finding":"The distance method achieves a maximum of 0.64 % monthly profit, while the cointegration method can reach up to 1.03 % monthly profit, both only slightly above the market’s overall monthly return of 0.47 %.","kind":"performance","quote":"Sharpe 0.64 -1.7 1.1 -1.1 0.94 -0.86 0.38 Table 2: Returns across pairs trading methods and timeframes cointegration method in particular appears to have a higher ceiling for per-formance since it can achieve up to 1.03% monthly proﬁt over the whole history as compared to the distance method’s 0.64%, which is only slightly higher than the market’s overall monthly return of 0.47%.","support":"full"},{"finding":"Static trading rules (even when fine‑tuned) are insufficient to outperform the market benchmark; dynamic or partially dynamic rules appear necessary.","kind":"negative","quote":"In fact, completely static trading rules of any sort are not enough to outperform the market benchmark since the distance and cointegration methods then only achieve 0.23% and 0.26% monthly return, respectively. 3.3. We thus conclude that naive pairs trading with complexity on the level of Gatev et al. (1999) does not work well, regardless of whether hyperparameter tuning is performed or not, and its weak","support":"full"},{"finding":"Pairs trading does not generate high excess returns in recent periods, but it performs very strongly during financial crises such as the Covid‑19 crisis.","kind":"performance","quote":"4. Conclusion We conﬁrm that pairs trading does not have high excess returns in recent times. We further corroborate the ﬁndings of Do and Faﬀ(2010) that pairs trading performs very strongly during ﬁnancial crises including the Covid-19 crisis, but generally lags behind the market benchmark otherwise, both in 8 raw and risk-adjusted returns.","support":"full"}],"reported_performance":null,"out_of_sample":{"method":"unclear","headline_out_of_sample":null}},"independent_evidence":{"status":"none_found","supports":[],"contradicts":[],"qualifies":[],"compared_with":23,"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":"2026-11-15T22:12:21Z"},"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/"}