{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,2,8]],"date-time":"2026-02-08T16:53:35Z","timestamp":1770569615438,"version":"3.49.0"},"reference-count":45,"publisher":"Springer Science and Business Media LLC","issue":"3","license":[{"start":{"date-parts":[[2023,6,24]],"date-time":"2023-06-24T00:00:00Z","timestamp":1687564800000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2023,6,24]],"date-time":"2023-06-24T00:00:00Z","timestamp":1687564800000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["Comput Stat"],"published-print":{"date-parts":[[2024,5]]},"abstract":"<jats:title>Abstract<\/jats:title><jats:p>Many modern statistical models are used for both insight and prediction when applied to data. When models are used for prediction one should optimise parameters through a prediction error loss function. Estimation methods based on multiple steps ahead forecast errors have been shown to lead to more robust and less biased estimates of parameters. However, a plausible explanation of why this is the case is lacking. In this paper, we provide this explanation, showing that the main benefit of these estimators is in a shrinkage effect, happening in univariate models naturally. However, this can introduce a series of limitations, due to overly aggressive shrinkage. We discuss the predictive likelihoods related to the multistep estimators and demonstrate what their usage implies to time series models. To overcome the limitations of the existing multiple steps estimators, we propose the Geometric Trace Mean Squared Error, demonstrating its advantages. We conduct a simulation experiment showing how the estimators behave with different sample sizes and forecast horizons. Finally, we carry out an empirical evaluation on real data, demonstrating the performance and advantages of the estimators. Given that the underlying process to be modelled is often unknown, we conclude that the shrinkage achieved by the GTMSE is a competitive alternative to conventional ones.<\/jats:p>","DOI":"10.1007\/s00180-023-01377-x","type":"journal-article","created":{"date-parts":[[2023,6,24]],"date-time":"2023-06-24T13:06:09Z","timestamp":1687611969000},"page":"1203-1239","update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":6,"title":["Multi-step estimators and shrinkage effect in time series models"],"prefix":"10.1007","volume":"39","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-7826-0281","authenticated-orcid":false,"given":"Ivan","family":"Svetunkov","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Nikolaos","family":"Kourentzes","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Rebecca","family":"Killick","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"297","published-online":{"date-parts":[[2023,6,24]]},"reference":[{"key":"1377_CR1","doi-asserted-by":"publisher","first-page":"9519","DOI":"10.1175\/JCLI-D-17-0863.1","volume":"31","author":"C Beaulieu","year":"2018","unstructured":"Beaulieu C, Killick R (2018) Distinguishing trends and shifts from memory in climate data. J Clim 31:9519\u20139543","journal-title":"J Clim"},{"key":"1377_CR2","doi-asserted-by":"publisher","first-page":"577","DOI":"10.1007\/BF00050857","volume":"48","author":"R Bhansali","year":"1996","unstructured":"Bhansali R (1996) Asymptotically efficient autoregressive model selection for multistep prediction. Ann Inst Stat Math 48:577\u2013602","journal-title":"Ann Inst Stat Math"},{"key":"1377_CR3","first-page":"425","volume":"7","author":"R Bhansali","year":"1997","unstructured":"Bhansali R (1997) Direct autoregressive predictors for multistep prediction: order selection and performance relative to the plug in predictors. Stat Sin 7:425\u2013449","journal-title":"Stat Sin"},{"key":"1377_CR4","volume-title":"Time series analysis: forecasting and control","author":"G Box","year":"1976","unstructured":"Box G, Jenkins G (1976) Time series analysis: forecasting and control. Holden-day, Oakland"},{"key":"1377_CR5","doi-asserted-by":"publisher","DOI":"10.1088\/1748-9326\/10\/8\/084002","volume":"10","author":"N Cahill","year":"2015","unstructured":"Cahill N, Rahmstorf S, Parnell AC (2015) Change points of global temperature. Environ Res Lett 10:084002","journal-title":"Environ Res Lett"},{"key":"1377_CR6","doi-asserted-by":"publisher","first-page":"746","DOI":"10.1111\/j.1467-6419.2007.00518.x","volume":"21","author":"G Chevillon","year":"2007","unstructured":"Chevillon G (2007) Direct multi-step estimation and forecasting. J Econ Surv 21:746\u2013785","journal-title":"J Econ Surv"},{"key":"1377_CR7","doi-asserted-by":"publisher","first-page":"602","DOI":"10.1016\/j.ijforecast.2008.12.004","volume":"25","author":"G Chevillon","year":"2009","unstructured":"Chevillon G (2009) Multi-step forecasting in emerging economies: an investigation of the South African GDP. Int J Forecast 25:602\u2013628","journal-title":"Int J Forecast"},{"key":"1377_CR8","doi-asserted-by":"publisher","first-page":"121","DOI":"10.1016\/j.ijforecast.2015.04.004","volume":"32","author":"G Chevillon","year":"2016","unstructured":"Chevillon G (2016) Multistep forecasting in the presence of location shifts. Int J Forecast 32:121\u2013137","journal-title":"Int J Forecast"},{"key":"1377_CR9","doi-asserted-by":"publisher","first-page":"201","DOI":"10.1016\/j.ijforecast.2004.08.004","volume":"21","author":"G Chevillon","year":"2005","unstructured":"Chevillon G, Hendry DF (2005) Non-parametric direct multi-step estimation for forecasting economic processes. Int J Forecast 21:201\u2013218","journal-title":"Int J Forecast"},{"key":"1377_CR10","doi-asserted-by":"publisher","DOI":"10.1017\/CBO9780511599286","volume-title":"Forecasting economic time series","author":"M Clements","year":"1998","unstructured":"Clements M, Hendry D (1998) Forecasting economic time series. Cambridge University Press, Cambridge"},{"key":"1377_CR11","doi-asserted-by":"crossref","first-page":"414","DOI":"10.1111\/j.2517-6161.1961.tb00424.x","volume":"23","author":"DR Cox","year":"1961","unstructured":"Cox DR (1961) Prediction by exponentially weighted moving averages and related methods. J R Stat Soc B 23:414\u2013422","journal-title":"J R Stat Soc B"},{"key":"1377_CR12","doi-asserted-by":"publisher","first-page":"389","DOI":"10.1111\/j.1467-6419.2009.00581.x","volume":"24","author":"PH Franses","year":"2009","unstructured":"Franses PH, Legerstee R (2009) A unifying view on multi-step forecasting using an autoregression. J Econ Surv 24:389\u2013401","journal-title":"J Econ Surv"},{"key":"1377_CR13","doi-asserted-by":"publisher","first-page":"637","DOI":"10.1016\/j.ijforecast.2006.03.005","volume":"22","author":"ES Gardner","year":"2006","unstructured":"Gardner ES (2006) Exponential smoothing: the state of the art-Part II. Int J Forecast 22:637\u2013666. https:\/\/doi.org\/10.1016\/j.ijforecast.2006.03.005","journal-title":"Int J Forecast"},{"key":"1377_CR14","doi-asserted-by":"publisher","first-page":"253","DOI":"10.1080\/07350015.1983.10509349","volume":"1","author":"W Gersch","year":"1983","unstructured":"Gersch W, Kitagawa G (1983) The prediction of time series with trends and seasonalities. J Bus Econ Stat 1:253\u2013264","journal-title":"J Bus Econ Stat"},{"key":"1377_CR15","first-page":"159","volume-title":"Observations: atmosphere and surface","author":"D Hartmann","year":"2013","unstructured":"Hartmann D, Klein Tank A, Rusticucci M, Alexander L, Br\u00f6nnimann S, Charabi Y, Dentener F, Dlugokencky E, Easterling D, Kaplan A, Soden B, Thorne P, Wild M, Zhai P (2013) Observations: atmosphere and surface. Cambridge University Press, Cambridge, pp 159\u2013254"},{"key":"1377_CR16","doi-asserted-by":"publisher","first-page":"237","DOI":"10.1111\/1467-9868.00066","volume":"59","author":"J Haywood","year":"1997","unstructured":"Haywood J, Tunnicliffe Wilson G (1997) Fitting time series models by minimizing multistep-ahead errors: a frequency domain approach. J R Stat Soc: Ser B (Methodol) 59:237\u2013254","journal-title":"J R Stat Soc: Ser B (Methodol)"},{"key":"1377_CR17","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-540-71918-2","volume-title":"Forecasting with exponential smoothing","author":"RJ Hyndman","year":"2008","unstructured":"Hyndman RJ, Koehler AB, Ord JK, Snyder RD (2008) Forecasting with exponential smoothing. Springer, Berlin"},{"key":"1377_CR18","doi-asserted-by":"publisher","first-page":"254","DOI":"10.1017\/S0266466603192031","volume":"19","author":"CK Ing","year":"2003","unstructured":"Ing CK (2003) Multistep prediction in autoregressive processes. Economet Theor 19:254\u2013279","journal-title":"Economet Theor"},{"key":"1377_CR19","doi-asserted-by":"publisher","first-page":"635","DOI":"10.1002\/jae.1166","volume":"25","author":"\u00d2 Jord\u00e0","year":"2010","unstructured":"Jord\u00e0 \u00d2, Marcellino M (2010) Path forecast evaluation. J Appl Economet 25:635\u2013662","journal-title":"J Appl Economet"},{"key":"1377_CR20","doi-asserted-by":"publisher","first-page":"387","DOI":"10.1016\/S0169-2070(02)00010-9","volume":"19","author":"IB Kang","year":"2003","unstructured":"Kang IB (2003) Multi-period forecasting using different models for different horizons: an application to U.S. economic time series data. Int J Forecast 19:387\u2013400","journal-title":"Int J Forecast"},{"key":"1377_CR21","doi-asserted-by":"publisher","first-page":"788","DOI":"10.1016\/j.ijforecast.2015.12.004","volume":"32","author":"S Kolassa","year":"2016","unstructured":"Kolassa S (2016) Evaluating predictive count data distributions in retail sales forecasting. Int J Forecast 32:788\u2013803","journal-title":"Int J Forecast"},{"key":"1377_CR22","unstructured":"Kourentzes N, Trapero JR (2018) On the use of multi-step cost functions for generating forecasts. Department of Management Science Working Paper Series"},{"key":"1377_CR23","doi-asserted-by":"publisher","first-page":"27","DOI":"10.1057\/s41272-017-0117-x","volume":"18","author":"N Kourentzes","year":"2019","unstructured":"Kourentzes N, Li D, Strauss AK (2019) Unconstraining methods for revenue management systems under small demand. J Revenue Pricing Manag 18:27\u201341","journal-title":"J Revenue Pricing Manag"},{"key":"1377_CR24","doi-asserted-by":"publisher","first-page":"499","DOI":"10.1016\/j.jeconom.2005.07.020","volume":"135","author":"M Marcellino","year":"2006","unstructured":"Marcellino M, Stock JH, Watson MW (2006) A comparison of direct and iterated multistep AR methods for forecasting macroeconomic time series. J Econom 135:499\u2013526","journal-title":"J Econom"},{"key":"1377_CR25","doi-asserted-by":"crossref","unstructured":"Martinez AB (2017) Testing for differences in path forecast accuracy: forecast-error dynamics matter","DOI":"10.2139\/ssrn.3064863"},{"key":"1377_CR26","doi-asserted-by":"publisher","first-page":"315","DOI":"10.1002\/for.2321","volume":"34","author":"T McElroy","year":"2015","unstructured":"McElroy T (2015) When are direct multi-step and iterative forecasts identical? J Forecast 34:315\u2013336","journal-title":"J Forecast"},{"key":"1377_CR27","doi-asserted-by":"publisher","first-page":"378","DOI":"10.1016\/j.ijforecast.2012.08.003","volume":"29","author":"T McElroy","year":"2013","unstructured":"McElroy T, Wildi M (2013) Multi-step-ahead estimation of time series models. Int J Forecast 29:378\u2013394","journal-title":"Int J Forecast"},{"key":"1377_CR28","doi-asserted-by":"publisher","first-page":"1621","DOI":"10.1080\/01621459.1997.10473684","volume":"92","author":"JK Ord","year":"1997","unstructured":"Ord JK, Koehler AB, Snyder RD (1997) Estimation and prediction for a class of dynamic nonlinear statistical models. J Am Stat Assoc 92:1621\u20131629","journal-title":"J Am Stat Assoc"},{"key":"1377_CR29","doi-asserted-by":"crossref","unstructured":"Pesaran MH, Pick A, Timmermann A (2010) Variable selection, estimation and inference for multi-period forecasting problems","DOI":"10.2139\/ssrn.1948480"},{"key":"1377_CR30","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijforecast.2022.07.005","author":"KF Pritularga","year":"2022","unstructured":"Pritularga KF, Svetunkov I, Kourentzes N (2022) Shrinkage estimator for exponential smoothing models. Int J Forecast. https:\/\/doi.org\/10.1016\/j.ijforecast.2022.07.005","journal-title":"Int J Forecast"},{"key":"1377_CR31","doi-asserted-by":"publisher","first-page":"266","DOI":"10.1016\/j.ijforecast.2010.05.014","volume":"27","author":"T Proietti","year":"2011","unstructured":"Proietti T (2011) Direct and iterated multistep AR methods for difference stationary processes. Int J Forecast 27:266\u2013280","journal-title":"Int J Forecast"},{"key":"1377_CR32","unstructured":"R Core Team, 2023. R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna. https:\/\/www.R-project.org\/"},{"key":"1377_CR33","doi-asserted-by":"publisher","DOI":"10.1088\/1748-9326\/aa6825","volume":"12","author":"S Rahmstorf","year":"2017","unstructured":"Rahmstorf S, Foster G, Cahill N (2017) Global temperature evolution: recent trends and some pitfalls. Environ Res Lett 12:054001","journal-title":"Environ Res Lett"},{"key":"1377_CR34","doi-asserted-by":"publisher","DOI":"10.1016\/j.omega.2022.102614","volume":"110","author":"P Saoud","year":"2022","unstructured":"Saoud P, Kourentzes N, Boylan JE (2022) Approximations for the lead time variance: a forecasting and inventory evaluation. Omega 110:102614. https:\/\/doi.org\/10.1016\/j.omega.2022.102614","journal-title":"Omega"},{"key":"1377_CR35","doi-asserted-by":"crossref","first-page":"272","DOI":"10.1111\/j.2517-6161.1985.tb01355.x","volume":"47","author":"RD Snyder","year":"1985","unstructured":"Snyder RD (1985) Recursive estimation of dynamic linear models. J R Stat Soc Ser B (Methodol) 47:272\u2013276","journal-title":"J R Stat Soc Ser B (Methodol)"},{"key":"1377_CR36","doi-asserted-by":"publisher","unstructured":"Svetunkov I (2023a) Forecasting and analytics with the augmented dynamic adaptive model (ADAM). Chapman and Hall\/CRC. https:\/\/openforecast.org\/adam\/. https:\/\/doi.org\/10.1201\/9781003452652","DOI":"10.1201\/9781003452652"},{"key":"1377_CR37","unstructured":"Svetunkov I (2023b) smooth: forecasting using state space models. https:\/\/github.com\/config-i1\/smooth. r package version 3.2.1"},{"key":"1377_CR38","doi-asserted-by":"publisher","first-page":"645","DOI":"10.1016\/j.ijforecast.2008.07.007","volume":"24","author":"JW Taylor","year":"2008","unstructured":"Taylor JW (2008) An evaluation of methods for very short-term load forecasting using minute-by-minute British data. Int J Forecast 24:645\u2013658","journal-title":"Int J Forecast"},{"key":"1377_CR39","doi-asserted-by":"publisher","first-page":"623","DOI":"10.1093\/biomet\/80.3.623","volume":"80","author":"GC Tiao","year":"1993","unstructured":"Tiao GC, Xu D (1993) Robustness of maximum likelihood estimates for multi-step predictions: the exponential smoothing case. Biometrika 80:623\u2013641","journal-title":"Biometrika"},{"key":"1377_CR40","doi-asserted-by":"crossref","first-page":"267","DOI":"10.1111\/j.2517-6161.1996.tb02080.x","volume":"58","author":"R Tibshirani","year":"1996","unstructured":"Tibshirani R (1996) Regression shrinkage and selection via the lasso. J R Stat Soc Ser B (Methodol) 58:267\u2013288","journal-title":"J R Stat Soc Ser B (Methodol)"},{"key":"1377_CR41","doi-asserted-by":"publisher","first-page":"199","DOI":"10.1016\/j.omega.2018.05.004","volume":"84","author":"JR Trapero","year":"2019","unstructured":"Trapero JR, Cardos M, Kourentzes N (2019) Empirical safety stock estimation based on kernel and Garch models. Omega 84:199\u2013211","journal-title":"Omega"},{"key":"1377_CR42","doi-asserted-by":"publisher","first-page":"595","DOI":"10.1214\/aoms\/1177729952","volume":"20","author":"A Wald","year":"1949","unstructured":"Wald A (1949) Note on the consistency of the maximum likelihood estimate. Ann Math Stat 20:595\u2013601","journal-title":"Ann Math Stat"},{"key":"1377_CR43","doi-asserted-by":"publisher","first-page":"135","DOI":"10.1016\/0304-4076(91)90035-C","volume":"48","author":"AA Weiss","year":"1991","unstructured":"Weiss AA (1991) Multi-step estimation and forecasting in dynamic models. J Econom 48:135\u2013149","journal-title":"J Econom"},{"key":"1377_CR44","doi-asserted-by":"publisher","first-page":"484","DOI":"10.2307\/2981579","volume":"147","author":"AA Weiss","year":"1984","unstructured":"Weiss AA, Andersen AP (1984) Estimating time series models using the relevant forecast evaluation criterion. J R Stat Soc Ser A (General) 147:484","journal-title":"J R Stat Soc Ser A (General)"},{"key":"1377_CR45","doi-asserted-by":"publisher","first-page":"21","DOI":"10.1214\/10-STS345","volume":"26","author":"Y Xia","year":"2011","unstructured":"Xia Y, Tong H (2011) Feature matching in time series modeling. Stat Sci 26:21\u201346. https:\/\/doi.org\/10.1214\/10-STS345","journal-title":"Stat Sci"}],"container-title":["Computational Statistics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s00180-023-01377-x.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/article\/10.1007\/s00180-023-01377-x\/fulltext.html","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s00180-023-01377-x.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2024,10,22]],"date-time":"2024-10-22T22:08:23Z","timestamp":1729634903000},"score":1,"resource":{"primary":{"URL":"https:\/\/link.springer.com\/10.1007\/s00180-023-01377-x"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2023,6,24]]},"references-count":45,"journal-issue":{"issue":"3","published-print":{"date-parts":[[2024,5]]}},"alternative-id":["1377"],"URL":"https:\/\/doi.org\/10.1007\/s00180-023-01377-x","relation":{},"ISSN":["0943-4062","1613-9658"],"issn-type":[{"value":"0943-4062","type":"print"},{"value":"1613-9658","type":"electronic"}],"subject":[],"published":{"date-parts":[[2023,6,24]]},"assertion":[{"value":"2 April 2022","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"9 June 2023","order":2,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"24 June 2023","order":3,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}}]}}