{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,13]],"date-time":"2026-10-13T10:47:50Z","timestamp":1791888470706,"version":"4.3.4"},"reference-count":54,"publisher":"Oxford University Press (OUP)","issue":"2","license":[{"start":{"date-parts":[[2021,1,25]],"date-time":"2021-01-25T00:00:00Z","timestamp":1611532800000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/academic.oup.com\/journals\/pages\/open_access\/funder_policies\/chorus\/standard_publication_model"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2021,4,1]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:p>We consider the problem of predicting a response variable from a set of covariates on a data set that differs in distribution from the training data. Causal parameters are optimal in terms of predictive accuracy if in the new distribution either many variables are affected by interventions or only some variables are affected, but the perturbations are strong. If the training and test distributions differ by a shift, causal parameters might be too conservative to perform well on the above task. This motivates anchor regression, a method that makes use of exogenous variables to solve a relaxation of the \u2018causal\u2019 minimax problem by considering a modification of the least-squares loss. The procedure naturally provides an interpolation between the solutions of ordinary least squares (OLS) and two-stage least squares. We prove that the estimator satisfies predictive guarantees in terms of distributional robustness against shifts in a linear class; these guarantees are valid even if the instrumental variable assumptions are violated. If anchor regression and least squares provide the same answer (\u2018anchor stability\u2019), we establish that OLS parameters are invariant under certain distributional changes. Anchor regression is shown empirically to improve replicability and protect against distributional shifts.<\/jats:p>","DOI":"10.1111\/rssb.12398","type":"journal-article","created":{"date-parts":[[2021,1,25]],"date-time":"2021-01-25T17:04:31Z","timestamp":1611594271000},"page":"215-246","source":"Crossref","is-referenced-by-count":80,"title":["Anchor Regression: Heterogeneous Data Meet Causality"],"prefix":"10.1093","volume":"83","author":[{"given":"Dominik","family":"Rothenh\u00e4usler","sequence":"first","affiliation":[{"name":"Department of Statistics, Stanford University , Stanford, CA , USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Nicolai","family":"Meinshausen","sequence":"additional","affiliation":[{"name":"Seminar f\u00fcr Statistik, ETH Z\u00fcrich , Z\u00fcrich , Switzerland"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Peter","family":"B\u00fchlmann","sequence":"additional","affiliation":[{"name":"Seminar f\u00fcr Statistik, ETH Z\u00fcrich , Z\u00fcrich , Switzerland"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jonas","family":"Peters","sequence":"additional","affiliation":[{"name":"Department of Mathematical Sciences, University of Copenhagen , Copenhagen , Denmark"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"286","published-online":{"date-parts":[[2021,1,25]]},"reference":[{"key":"2023022408371058900_","doi-asserted-by":"crossref","first-page":"15","DOI":"10.1093\/oxfordjournals.oep.a041889","article-title":"Autonomy","volume":"41","author":"Aldrich","year":"1989","journal-title":"Oxford Economic Papers"},{"key":"2023022408371058900_","doi-asserted-by":"crossref","DOI":"10.1002\/9781118619179","volume-title":"Structural Equations with latent variables","author":"Bollen","year":"1989"},{"key":"2023022408371058900_","doi-asserted-by":"crossref","DOI":"10.1093\/acprof:oso\/9780199535255.001.0001","volume-title":"Concentration inequalities: A nonasymptotic theory of independence","author":"Boucheron","year":"2013"},{"key":"2023022408371058900_","volume-title":"Instrumental variables","author":"Bowden","year":"1990"},{"issue":"3","key":"2023022408371058900_","first-page":"404","article-title":"Invariance, causality and robustness","volume":"35","author":"B\u00fchlmann","year":"2020","journal-title":"Statistical Science"},{"key":"2023022408371058900_","doi-asserted-by":"crossref","DOI":"10.1007\/978-3-642-20192-9","volume-title":"Statistics for high-dimensional data: Methods, theory and applications","author":"B\u00fchlmann","year":"2011"},{"issue":"5","key":"2023022408371058900_","doi-asserted-by":"crossref","first-page":"311","DOI":"10.1089\/bio.2015.0032","article-title":"A novel approach to high-quality postmortem tissue procurement: The gtex project","volume":"13","author":"Carithers","year":"2015","journal-title":"Biopreservation and Biobanking"},{"key":"2023022408371058900_","doi-asserted-by":"crossref","first-page":"407","DOI":"10.1080\/01621459.2000.10474210","article-title":"Causal inference without counterfactuals","volume":"95","author":"Dawid","year":"2000","journal-title":"Journal of the American Statistical Association"},{"key":"2023022408371058900_","volume-title":"UCI machine learning repository","author":"Dheeru","year":"2017"},{"key":"2023022408371058900_","doi-asserted-by":"crossref","first-page":"22","DOI":"10.1214\/09-STS316","article-title":"Assumptions of IV methods for observational epidemiology","volume":"25","author":"Didelez","year":"2010","journal-title":"Statistical Science"},{"key":"2023022408371058900_","doi-asserted-by":"crossref","first-page":"981","DOI":"10.1086\/525638","article-title":"Interventions and causal inference","volume":"74","author":"Eberhardt","year":"2007","journal-title":"Philosophy of Science"},{"key":"2023022408371058900_","first-page":"256","volume-title":"Proceedings of the Sixteenth International Conference on Artificial Intelligence and Statistics","author":"Entner","year":"2013"},{"key":"2023022408371058900_","doi-asserted-by":"crossref","first-page":"1491","DOI":"10.1214\/aos\/1017939139","article-title":"Statistical estimation in varying coefficient models","volume":"27","author":"Fan","year":"1999","journal-title":"Annals of Statistics"},{"issue":"2\u20133","key":"2023022408371058900_","doi-asserted-by":"publisher","first-page":"113","DOI":"10.1007\/s13748-013-0040-3","article-title":"Event labeling combining ensemble detectors and background knowledge","volume":"2","author":"Fanaee-T","year":"2013","journal-title":"Progress in Artificial Intelligence"},{"issue":"2","key":"2023022408371058900_","doi-asserted-by":"crossref","first-page":"302","DOI":"10.1214\/07-AOAS131","article-title":"Pathwise coordinate optimization","volume":"1","author":"Friedman","year":"2007","journal-title":"Annals of Applied Statistics"},{"key":"2023022408371058900_","volume-title":"Measurement error models","author":"Fuller","year":"2009"},{"issue":"13","key":"2023022408371058900_","doi-asserted-by":"crossref","first-page":"i288","DOI":"10.1093\/bioinformatics\/btr221","article-title":"Mixed-model coexpression: Calculating gene coexpression while accounting for expression heterogeneity","volume":"27","author":"Furlotte","year":"2011","journal-title":"Bioinformatics"},{"key":"2023022408371058900_","volume-title":"Wasserstein distributional robustness and regularization in statistical learning","author":"Gao","year":"2017"},{"key":"2023022408371058900_","doi-asserted-by":"crossref","DOI":"10.1007\/978-3-319-32774-7","author":"van de Geer","year":"2016","journal-title":"Estimation and testing under sparsity"},{"key":"2023022408371058900_","doi-asserted-by":"crossref","first-page":"37","DOI":"10.1097\/00001648-199901000-00008","article-title":"Causal diagrams for epidemiologic research","volume":"10","author":"Greenland","year":"1999","journal-title":"Epidemiology"},{"issue":"supplement","key":"2023022408371058900_","first-page":"S1","article-title":"The probability approach in econometrics","volume":"12","author":"Haavelmo","year":"1944","journal-title":"Econometrica"},{"key":"2023022408371058900_","doi-asserted-by":"crossref","first-page":"757","DOI":"10.1111\/j.2517-6161.1993.tb01939.x","article-title":"Varying-coefficient models","volume":"55","author":"Hastie","year":"1993","journal-title":"Journal of the Royal Statistical Society, Series B"},{"key":"2023022408371058900_","volume-title":"Conditional variance penalties and domain shift robustness","author":"Heinze-Deml","year":"2018"},{"issue":"1","key":"2023022408371058900_","doi-asserted-by":"crossref","first-page":"73","DOI":"10.1214\/aoms\/1177703732","article-title":"Robust estimation of a location parameter","volume":"35","author":"Huber","year":"1964","journal-title":"Annals of Mathematical Statistics"},{"key":"2023022408371058900_","doi-asserted-by":"crossref","first-page":"799","DOI":"10.1214\/aos\/1176342503","article-title":"Robust regression: Asymptotics, conjectures and Monte Carlo","volume":"1","author":"Huber","year":"1973","journal-title":"Annals of Statistics"},{"key":"2023022408371058900_","doi-asserted-by":"crossref","first-page":"163","DOI":"10.2307\/1911466","article-title":"Consistent sets of estimates for regressions with errors in all variables","volume":"52","author":"Klepper","year":"1984","journal-title":"Econometrica"},{"key":"2023022408371058900_","first-page":"322","volume-title":"Varieties of causal intervention","author":"Korb","year":"2004"},{"key":"2023022408371058900_","doi-asserted-by":"crossref","first-page":"157","DOI":"10.1111\/j.2517-6161.1988.tb01721.x","article-title":"Local computations with probabilities on graphical structures and their application to expert systems","volume":"50","author":"Lauritzen","year":"1988","journal-title":"Journal of the Royal Statistical Society, Series B"},{"key":"2023022408371058900_","doi-asserted-by":"crossref","first-page":"961","DOI":"10.2307\/1909758","article-title":"Least-squares versus instrumental variables estimation in a simple errors in variables model","volume":"46","author":"Leamer","year":"1978","journal-title":"Econometrica"},{"key":"2023022408371058900_","first-page":"10846","volume-title":"Advances in Neural Information Processing Systems 31","author":"Magliacane","year":"2018"},{"key":"2023022408371058900_","first-page":"6","volume-title":"Causality from a distributional robustness point of view","author":"Meinshausen","year":"2018"},{"issue":"4","key":"2023022408371058900_","doi-asserted-by":"crossref","first-page":"1801","DOI":"10.1214\/15-AOS1325","article-title":"Maximin effects in inhomogeneous large-scale data","volume":"43","author":"Meinshausen","year":"2015","journal-title":"Annals of Statistics"},{"key":"2023022408371058900_","doi-asserted-by":"crossref","first-page":"575","DOI":"10.2307\/1909352","article-title":"The bias and moment matrix of the general k-class estimators of the parameters in simultaneous equations","volume":"27","author":"Nagar","year":"1959","journal-title":"Econometrica"},{"issue":"10","key":"2023022408371058900_","doi-asserted-by":"crossref","first-page":"1345","DOI":"10.1109\/TKDE.2009.191","article-title":"A survey on transfer learning","volume":"22","author":"Pan","year":"2010","journal-title":"IEEE Transactions on Knowledge and Data Engineering"},{"key":"2023022408371058900_","doi-asserted-by":"crossref","DOI":"10.1017\/CBO9780511803161","volume-title":"Causality: Models, reasoning, and inference","author":"Pearl","year":"2009","edition":"2nd edition"},{"key":"2023022408371058900_","doi-asserted-by":"crossref","first-page":"579","DOI":"10.1214\/14-STS486","article-title":"External validity: From do-calculus to transportability across populations","volume":"29","author":"Pearl","year":"2014","journal-title":"Statistical Science"},{"issue":"5","key":"2023022408371058900_","doi-asserted-by":"crossref","first-page":"947","DOI":"10.1111\/rssb.12167","article-title":"Causal inference by using invariant prediction: Identification and confidence intervals","volume":"78","author":"Peters","year":"2016","journal-title":"Journal of the Royal Statistical Society, Series B"},{"key":"2023022408371058900_","volume-title":"Elements of causal inference: Foundations and learning algorithms","author":"Peters","year":"2017"},{"issue":"51","key":"2023022408371058900_","doi-asserted-by":"crossref","first-page":"25405","DOI":"10.1073\/pnas.1905688116","article-title":"Learning stable and predictive structures in kinetic systems","volume":"116","author":"Pfister","year":"2019","journal-title":"Proceedings of the National Academy of Sciences"},{"key":"2023022408371058900_","volume-title":"Mixed-effects models in S and S-Plus","author":"Pinheiro","year":"2006"},{"key":"2023022408371058900_","doi-asserted-by":"crossref","first-page":"550","DOI":"10.1097\/00001648-200009000-00011","article-title":"Marginal structural models and causal inference in epidemiology","volume":"11","author":"Robins","year":"2000","journal-title":"Epidemiology"},{"issue":"36","key":"2023022408371058900_","first-page":"1","article-title":"Causal transfer in machine learning","volume":"19","author":"Rojas-Carulla","year":"2018","journal-title":"Journal of Machine Learning Research"},{"key":"2023022408371058900_","doi-asserted-by":"crossref","first-page":"322","DOI":"10.1198\/016214504000001880","article-title":"Causal inference using potential outcomes","volume":"100","author":"Rubin","year":"2005","journal-title":"Journal of the American Statistical Association"},{"key":"2023022408371058900_","first-page":"949","volume-title":"Identification and estimation of causal effects defined by shift interventions","author":"Sani","year":"2020"},{"key":"2023022408371058900_","volume-title":"Certifying some distributional robustness with principled adversarial training","author":"Sinha","year":"2018"},{"key":"2023022408371058900_","volume-title":"Causation, prediction, and search","author":"Spirtes","year":"2000","edition":"2nd edition"},{"key":"2023022408371058900_","first-page":"630","volume-title":"Advances in Neural Information Processing Systems 24","author":"Stegle","year":"2011"},{"issue":"3","key":"2023022408371058900_","doi-asserted-by":"crossref","first-page":"500","DOI":"10.1038\/nprot.2011.457","article-title":"Using probabilistic estimation of expression residuals (PEER) to obtain increased power and interpretability of gene expression analyses","volume":"7","author":"Stegle","year":"2012","journal-title":"Nature Protocols"},{"key":"2023022408371058900_","volume-title":"Economic forecasts and policy","author":"Theil","year":"1958"},{"key":"2023022408371058900_","first-page":"512","volume-title":"Causal discovery from changes","author":"Tian","year":"2001"},{"key":"2023022408371058900_","doi-asserted-by":"crossref","first-page":"267","DOI":"10.1111\/j.2517-6161.1996.tb02080.x","article-title":"Regression shrinkage and selection via the Lasso","volume":"58","author":"Tibshirani","year":"1996","journal-title":"Journal of the Royal Statistical Society, Series B"},{"key":"2023022408371058900_","volume-title":"The tariff on animal and vegetable oils","author":"Wright","year":"1928"},{"key":"2023022408371058900_","first-page":"1801","volume-title":"Robust regression and lasso","author":"Xu","year":"2009"},{"issue":"8","key":"2023022408371058900_","doi-asserted-by":"crossref","first-page":"3920","DOI":"10.1073\/pnas.1901326117","article-title":"Veridical data science","volume":"117","author":"Yu","year":"2020","journal-title":"Proceedings of the National Academy of Sciences"}],"container-title":["Journal of the Royal Statistical Society Series B: Statistical Methodology"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/pdf\/10.1111\/rssb.12398","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/full-xml\/10.1111\/rssb.12398","content-type":"application\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/academic.oup.com\/jrsssb\/article-pdf\/83\/2\/215\/49320927\/jrsssb_83_2_215.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/jrsssb\/article-pdf\/83\/2\/215\/49320927\/jrsssb_83_2_215.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2024,8,22]],"date-time":"2024-08-22T19:57:28Z","timestamp":1724356648000},"score":1,"resource":{"primary":{"URL":"https:\/\/academic.oup.com\/jrsssb\/article\/83\/2\/215\/7056043"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2021,1,25]]},"references-count":54,"journal-issue":{"issue":"2","published-online":{"date-parts":[[2021,1,25]]},"published-print":{"date-parts":[[2021,4,1]]}},"URL":"https:\/\/doi.org\/10.1111\/rssb.12398","relation":{},"ISSN":["1369-7412","1467-9868"],"issn-type":[{"value":"1369-7412","type":"print"},{"value":"1467-9868","type":"electronic"}],"subject":[],"published-other":{"date-parts":[[2021,4]]},"published":{"date-parts":[[2021,1,25]]}}}