{"paper_id":"2511.05463","title":"From sectorial coarse graining to extreme coarse graining of S&P 500 correlation matrices","authors":["Manan Vyas","M. Mijaíl Martínez-Ramos","Parisa Majari","Thomas H. Seligman"],"paper_date":"2025-11-07","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2511.05463v1","page_url":"https://thequant.space/flowcharts/from-sectorial-coarse-graining-to-extreme-coarse-graining-of/","scores":{"empirical_rigor":6.5,"math_complexity":4,"quadrant":"Street Traders"},"proposes":{"status":"available","statement":"The paper proposes to coarse‑grain the Pearson correlation matrix of S&P 500 stock returns into a real symmetric 2 × 2 matrix that retains the average correlation as a parameter.","quote":"Starting from the Pearson Correlation Matrix of stock returns and from the desire to obtain a reduced number of parameters relevant for the dynamics of a financial market, we propose to take the idea of a sectorial matrix, which would have a large number of parameters, to the reduced picture of a real symmetric $2 \\times 2$ matrix, extreme case, that still conserves the desirable feature that the average correlation","summary":"Starting from the Pearson Correlation Matrix of stock returns and from the desire to obtain a reduced number of parameters relevant for the dynamics of a financial market, we propose to take the idea of a sectorial matrix, which would have a large number of parameters, to the reduced picture of a…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"Returns are set to zero on days without a closing quote.","type":"data","stated":"explicit","quote":"(1) For the corresponding returns, we assume zero for the days without closing quote while the return for the active trading day is computed using last active trading day.","support":"full"},{"assumption":"The correlation matrix is computed from logarithmic returns of adjusted closing prices.","type":"model","stated":"explicit","quote":"Pearson correlation matrix elements are computed using logarithmic return time series of adjusted closing prices. We divide the total time horizon T into epochs of 20 trading days, shifted by one trading day, and use logarithmic returns r between these days as the dataset, given the adjusted closing price pi(t) of trading day t for stock i, ri(t) = log \u0014 pi(t) pi(t −1) \u0015 .","support":"full"},{"assumption":"The time horizon is divided into epochs of 20 trading days with a one‑day shift.","type":"data","stated":"explicit","quote":"We divide the total time horizon T into epochs of 20 trading days, shifted by one trading day, and use logarithmic returns r between these days as the dataset, given the adjusted closing price pi(t) of trading day t for stock i, ri(t) = log \u0014 pi(t) pi(t −1) \u0015 . The time horizon divided into epochs of 20 trading days with one trading day shift will define our set of data matrices from which we calculate the","support":"full"},{"assumption":"Two blocks are formed either randomly with equal size or by selecting stocks with strong intra‑ or inter‑sectorial correlations.","type":"model","stated":"explicit","quote":"We will use just two blocks, which we shall choose at random but of equal size, or intuitively depending upon the intra- and inter-sectorial correlations. Note that in Choice 1, we choose sectors with strong intra-sectorial correlations (EG, FN, TC, UT) as the first block and rest of the sectors as second block; in Choice 2, we choose sectors with strong inter-sectorial correlations (CD, FN, ID) as first block and","support":"full"},{"assumption":"The Pearson correlation matrix elements are averaged within each block (excluding self‑correlations) to form the 2 × 2 extreme coarse‑grained matrix.","type":"model","stated":"explicit","quote":"Then, we sum all the correlation matrix elements inside each block (excluding the self-correlations) and divide by the total number of correlation matrix elements in the block resulting in a 2 × 2 dimensional extreme coarse grained (ECG) correlation matrix for each epoch.","support":"full"}]},"datasets":{"status":"available","items":[{"name":"Adjusted end‑of‑day price returns of S&P 500 stocks","provider":"Yahoo Finance","kind":"market-prices","access":"public","period":"2006‑01‑03 to 2023‑08‑10","universe":"322 S&P 500 constituents (after filtering)","quote":"2 Data and Methodology For the present analysis, we use as data the returns for adjusted end of day prices of the stocks of the S&P 500 index. The reason for this is the obvious representativity of the stocks selected as the most relevant components on the New York Stock exchange and NASDAQ.","support":"partial","known_id":"yahoo-finance"}],"known":[],"sample":{"start":null,"end":null,"market":null,"asset_class":null,"frequency":null},"data_resolution":"daily"},"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":"unclear","findings":[{"finding":"Using a 2 × 2 extreme coarse‑grained correlation matrix reduces the number of parameters to three while preserving the average correlation.","kind":"theoretical","quote":"In this paper, we plan to extend the analysis to 2 × 2 dimensional correlation matrices, which generates only three parameters.","support":"full"},{"finding":"The approach yields similar clustering results to sectorial coarse‑graining at lower computational cost.","kind":"statistical","quote":"We shall see that under typical circumstances similar results are obtained at a lower computational cost, reducing the number of parameters to three.","support":"full"},{"finding":"Transition matrices between market states are near‑tridiagonal, satisfying the Markovianity criterion.","kind":"statistical","quote":"The transition matrices are near-tridiagonal. Also, the necessary criterion for Markovianity given in Eq. (2) of [2] is fulfilled for both.","support":"full"},{"finding":"The COVID‑related market state appears only after 2019 and is not captured by the extreme coarse‑grained representation.","kind":"other","quote":"Also, comparison with the time horizon until 2019 showed that the COVID state did not exist previously.","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":13,"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":"2026-11-14T19:46:24Z"},"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/"}