Quadratic shrinkage for large covariance matrices
This paper constructs a new estimator for large covariance matrices by drawing a bridge between the classic Stein (1975) estimator in finite samples and recent progress under large-dimensional asymptotics. The estimator keeps the eigenvectors of the sample covariance matrix and applies shrinkage to...
Saved in:
Published in | IDEAS Working Paper Series from RePEc |
---|---|
Main Authors | , |
Format | Paper |
Language | English |
Published |
St. Louis
Federal Reserve Bank of St. Louis
01.01.2019
|
Subjects | |
Online Access | Get full text |
Cover
Loading…
Abstract | This paper constructs a new estimator for large covariance matrices by drawing a bridge between the classic Stein (1975) estimator in finite samples and recent progress under large-dimensional asymptotics. The estimator keeps the eigenvectors of the sample covariance matrix and applies shrinkage to the inverse sample eigenvalues. The corresponding formula is quadratic: it has two shrinkage targets weighted by quadratic functions of the concentration (that is, matrix dimension divided by sample size). The first target dominates mid-level concentrations and the second one higher levels. This extra degree of freedom enables us to outperform linear shrinkage when optimal shrinkage is not linear (which is the general case). Both of our targets are based on what we term the “Stein shrinker”, a local attraction operator that pulls sample covariance matrix eigenvalues towards their nearest neighbors, but whose force diminishes with distance, like gravitation. We prove that no cubic or higher-order nonlinearities beat quadratic with respect to Frobenius loss under large-dimensional asymptotics. Non-normality and the case where the matrix dimension exceeds the sample size are accommodated. Monte Carlo simulations confirm state-of-the-art performance in terms of accuracy, speed, and scalability. |
---|---|
AbstractList | This paper constructs a new estimator for large covariance matrices by drawing a bridge between the classic Stein (1975) estimator in finite samples and recent progress under large-dimensional asymptotics. The estimator keeps the eigenvectors of the sample covariance matrix and applies shrinkage to the inverse sample eigenvalues. The corresponding formula is quadratic: it has two shrinkage targets weighted by quadratic functions of the concentration (that is, matrix dimension divided by sample size). The first target dominates mid-level concentrations and the second one higher levels. This extra degree of freedom enables us to outperform linear shrinkage when optimal shrinkage is not linear (which is the general case). Both of our targets are based on what we term the “Stein shrinker”, a local attraction operator that pulls sample covariance matrix eigenvalues towards their nearest neighbors, but whose force diminishes with distance, like gravitation. We prove that no cubic or higher-order nonlinearities beat quadratic with respect to Frobenius loss under large-dimensional asymptotics. Non-normality and the case where the matrix dimension exceeds the sample size are accommodated. Monte Carlo simulations confirm state-of-the-art performance in terms of accuracy, speed, and scalability. |
Author | Ledoit, Olivier Wolf, Michael |
Author_xml | – sequence: 1 givenname: Olivier surname: Ledoit fullname: Ledoit, Olivier – sequence: 2 givenname: Michael surname: Wolf fullname: Wolf, Michael |
BookMark | eNrjYmDJy89L5WQwDCxNTClKLMlMVijOKMrMy05MT1VIyy9SyEksArKS88sSizIT85JTFXITS4oyk1OLeRhY0xJzilN5oTQ3g7Kba4izh25BUX5haWpxSXxWfmlRHlAq3sjUwsLYwNTU0tSYOFUAmeQwYw |
ContentType | Paper |
Copyright | 2019. Notwithstanding the ProQuest Terms and conditions, you may use this content in accordance with the associated terms available at https://research.stlouisfed.org/research_terms.html . |
Copyright_xml | – notice: 2019. Notwithstanding the ProQuest Terms and conditions, you may use this content in accordance with the associated terms available at https://research.stlouisfed.org/research_terms.html . |
DBID | 3V. 7WY 7WZ 7XB 87Z 8FK 8FL AAFGM ABLUL ABPUF ABSSA ABUWG ACIOU ADZZV AFKRA AGAJT AGSBL AJNOY AQTIP AZQEC BENPR BEZIV BOUDT CBHQV CCPQU DWQXO FRNLG F~G K60 K6~ L.- M0C PIMPY PQBIZ PQBZA PQCXX PQEST PQQKQ PQUKI PRINS Q9U |
DatabaseName | ProQuest Central (Corporate) ABI/INFORM Collection ABI/INFORM Global (PDF only) ProQuest Central (purchase pre-March 2016) ABI/INFORM Global (Alumni Edition) ProQuest Central (Alumni) (purchase pre-March 2016) ABI/INFORM Collection (Alumni Edition) ProQuest Central Korea - hybrid linking Business Premium Collection - hybrid linking ABI/INFORM Collection (Alumni) - hybrid linking ABI/INFORM Collection - hybrid linking ProQuest Central (Alumni) ABI/INFORM Global - hybrid linking ProQuest Central (Alumni) - hybrid linking ProQuest Central ProQuest Central Essentials - hybrid linking ABI/INFORM Global (Alumni) - hybrid linking Business Premium Collection (Alumni) - hybrid linking ProQuest Women's & Gender Studies - hybrid linking ProQuest Central Essentials ProQuest Central Business Premium Collection ProQuest One Business - hybrid linking ProQuest One Business (Alumni) - hybrid linking ProQuest One Community College ProQuest Central Korea Business Premium Collection (Alumni) ABI/INFORM Global (Corporate) ProQuest Business Collection (Alumni Edition) ProQuest Business Collection ABI/INFORM Professional Advanced ABI/INFORM Global Publicly Available Content (ProQuest) ProQuest One Business ProQuest One Business (Alumni) ProQuest Central - hybrid linking ProQuest One Academic Eastern Edition (DO NOT USE) ProQuest One Academic ProQuest One Academic UKI Edition ProQuest Central China ProQuest Central Basic |
DatabaseTitle | Publicly Available Content Database Business Premium Collection ABI/INFORM Global (Corporate) ProQuest Business Collection (Alumni Edition) ProQuest One Business ABI/INFORM Global ABI/INFORM Global (Alumni Edition) ProQuest Central Basic ProQuest Central Essentials ProQuest One Academic Eastern Edition ProQuest Central (Alumni Edition) ProQuest One Community College ProQuest Business Collection ProQuest Central China ABI/INFORM Complete ProQuest Central ABI/INFORM Professional Advanced ProQuest One Academic UKI Edition ProQuest Central Korea ProQuest One Business (Alumni) ProQuest One Academic ABI/INFORM Complete (Alumni Edition) ProQuest Central (Alumni) Business Premium Collection (Alumni) |
DatabaseTitleList | Publicly Available Content Database |
Database_xml | – sequence: 1 dbid: BENPR name: AUTh Library subscriptions: ProQuest Central url: https://www.proquest.com/central sourceTypes: Aggregation Database |
DeliveryMethod | fulltext_linktorsrc |
Genre | Working Paper/Pre-Print |
GroupedDBID | 3V. 7WY 7XB 8FK 8FL ABUWG AFKRA AZQEC BENPR BEZIV CCPQU DWQXO FRNLG K60 K6~ L.- M0C PIMPY PQBIZ PQBZA PQEST PQQKQ PQUKI PRINS Q9U |
ID | FETCH-proquest_journals_25883055953 |
IEDL.DBID | BENPR |
IngestDate | Thu Oct 10 22:29:28 EDT 2024 |
IsOpenAccess | true |
IsPeerReviewed | false |
IsScholarly | false |
Language | English |
LinkModel | DirectLink |
MergedId | FETCHMERGED-proquest_journals_25883055953 |
OpenAccessLink | https://www.proquest.com/docview/2588305595?pq-origsite=%requestingapplication% |
PQID | 2588305595 |
PQPubID | 2036240 |
ParticipantIDs | proquest_journals_2588305595 |
PublicationCentury | 2000 |
PublicationDate | 20190101 |
PublicationDateYYYYMMDD | 2019-01-01 |
PublicationDate_xml | – month: 01 year: 2019 text: 20190101 day: 01 |
PublicationDecade | 2010 |
PublicationPlace | St. Louis |
PublicationPlace_xml | – name: St. Louis |
PublicationTitle | IDEAS Working Paper Series from RePEc |
PublicationYear | 2019 |
Publisher | Federal Reserve Bank of St. Louis |
Publisher_xml | – name: Federal Reserve Bank of St. Louis |
Score | 3.1880803 |
Snippet | This paper constructs a new estimator for large covariance matrices by drawing a bridge between the classic Stein (1975) estimator in finite samples and recent... |
SourceID | proquest |
SourceType | Aggregation Database |
SubjectTerms | Eigenvalues |
Title | Quadratic shrinkage for large covariance matrices |
URI | https://www.proquest.com/docview/2588305595 |
hasFullText | 1 |
inHoldings | 1 |
isFullTextHit | |
isPrint | |
link | http://utb.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwfV1NSwMxEB20e_GmqLR2KwG9Bs1udrt7EiotxcNSRaG3kp0k9GA_3G37-81sU3oo9BwYwguZNx-PGYBniwolxq8chdBc5pHiOaLhijQXfWu1wkblW6TjH_kxTaa-4FZ7WeXBJzaOWq-QauQvUZJlNJ0qT97Wf5y2RlF31a_QuIQgEpLatMFgWEy-Thxrwxajawgmam2qG7gwy1sQn1ulCWdk9bxyqZ_7wcyFiuyXRNgMVzuXrhL2bNGMyzf1HTyNht_vY36wPvMvXs-O94vvoeVSd9MGFqNAmmqkHVDSploJi5jaNJNJhKosOxCes_Rw_rgLV46-831BIITWptqanqPITfnocfgHjmRy_g |
link.rule.ids | 786,790,21416,33777,43838 |
linkProvider | ProQuest |
linkToHtml | http://utb.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwfV1NS8QwEB20PehNUfFj1YBeg6ZNa3sSlF2qrmWVFfZW0kmCB91d211_v5maxYOw58AQXsi8zOTxBuDSokKJ8TVHITSXeaR4jmi4Is3FjbVaYafyLdPiTT5OkolvuLVeVrnKiV2i1jOkHvlVlGQZuVPlye38i9PUKPpd9SM0NiEky80sgPCuX45e_yXWji0GOxCO1Nw0u7BhpnsgXpZKE87I2vfGlX7uBjP3VGQfJMJmOPt25Sphzz47u3zT7sPFoD--L_gqeuVPvK3-9hcfQOBKd3MILEaB5GqkHVDSploJi5jaNJNJhKquj6C3LtLx-uVz2CrGz8Nq-FA-ncC2o_L8tznQg2DRLM2po8tFfeYx-QEffHX0 |
openUrl | ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=Quadratic+shrinkage+for+large+covariance+matrices&rft.jtitle=IDEAS+Working+Paper+Series+from+RePEc&rft.au=Ledoit%2C+Olivier&rft.au=Wolf%2C+Michael&rft.date=2019-01-01&rft.pub=Federal+Reserve+Bank+of+St.+Louis |