{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,15]],"date-time":"2026-10-15T17:51:39Z","timestamp":1792086699647,"version":"4.4.0"},"posted":{"date-parts":[[2026]]},"group-title":"SSRN","reference-count":10,"publisher":"Elsevier BV","license":[{"start":{"date-parts":[[2026,1,1]],"date-time":"2026-01-01T00:00:00Z","timestamp":1767225600000},"content-version":"unspecified","delay-in-days":0,"URL":"https:\/\/www.uspto.gov\/ip-policy\/copyright-policy\/copyright-basics"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"abstract":"<jats:p>We ask how quickly the ability to predict a volatility regime decays when the forecaster's information is \u03b4 days out of date. On nine years of daily S&amp;amp;P 500 data, we evaluate four forecasters against a binary target, namely whether realised variance five days hence exceeds its trailing 252-day median, at delays of zero to 55 trading days. A trivial rule, \"is volatility above its median now\", stays above a constant forecast out to three days of delay against the more demanding of the two constants a forecaster could actually have run, and a week and a half against the weaker one, with the crossing itself unresolved at the next delay in each case; beyond that the data cannot separate it from a constant. Neither the Yang-Zhang estimator (+2.0 points at zero delay, 95% CI [-5.0, +8.2]) nor a walk-forward HAR classifier (-4.0 points, [-11.5, +2.5]) is distinguishable from it at any delay; the intervals are wide enough to leave a HAR advantage of up to 12.5 accuracy points unexcluded, so what is shown is a failure to separate and not an equality. We explain that instead of reporting it: for a median-split target, under approximate joint lognormality, the threshold rule is the Bayes rule among functions of the delayed state alone. Measured accuracies sit at 85-106% of that restricted-information Gaussian benchmark, exceeding it at six of ten delays, so Section 5 treats it as an approximation, never a ceiling. The benchmark is the best cut on the delayed state rather than the same-sign rule: the correlation turns negative at the two longest delays, where a same-sign rule is beaten by its own negation and is therefore not a benchmark for anything. The estimated headroom is small relative to sampling uncertainty in this sample, so the negative result is about the target, not the models. Three robustness checks follow, on the loss, the windows and the split. The third breaks the median-split theory: at an upper quantile the Bayes cut becomes z q \/ \u03c1 instead of zero, and at the 90th percentile the Bayes rule itself attains only 90.1% against a majority-class 90.0%.<\/jats:p>","DOI":"10.2139\/ssrn.7504558","type":"posted-content","created":{"date-parts":[[2026,9,24]],"date-time":"2026-09-24T04:47:16Z","timestamp":1790225236000},"source":"Crossref","is-referenced-by-count":0,"title":["Why a Simple Threshold Rule Is Hard to Beat: Volatility Regime Prediction Under Information Delay"],"prefix":"10.2139","author":[{"ORCID":"https:\/\/orcid.org\/0009-0007-3306-7833","authenticated-orcid":true,"given":"Sunny Alex","family":"Vellanikaran","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"78","reference":[{"issue":"4","key":"ref1","doi-asserted-by":"crossref","first-page":"885","DOI":"10.2307\/2527343","article-title":"Answering the skeptics: yes, standard volatility models do provide accurate forecasts","volume":"39","author":"T G Andersen","year":"1998","journal-title":"International Economic Review"},{"issue":"1","key":"ref2","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1016\/j.jeconom.2015.10.007","article-title":"Exploiting the errors: a simple approach for improved volatility forecasting","volume":"192","author":"T Bollerslev","year":"2016","journal-title":"Journal of Econometrics"},{"issue":"2","key":"ref3","doi-asserted-by":"crossref","first-page":"174","DOI":"10.1093\/jjfinec\/nbp001","article-title":"A simple approximate long-memory model of realized volatility","volume":"7","author":"F Corsi","year":"2009","journal-title":"Journal of Financial Econometrics"},{"issue":"1","key":"ref4","doi-asserted-by":"crossref","first-page":"60","DOI":"10.1017\/S0266466613000121","article-title":"Estimating the persistence and the autocorrelation function of a time series that is measured with error","volume":"30","author":"P R Hansen","year":"2014","journal-title":"Econometric Theory"},{"issue":"1","key":"ref5","doi-asserted-by":"crossref","first-page":"246","DOI":"10.1016\/j.jeconom.2010.03.034","article-title":"Volatility forecast comparison using imperfect volatility proxies","volume":"160","author":"A J Patton","year":"2011","journal-title":"Journal of Econometrics"},{"issue":"3","key":"ref6","doi-asserted-by":"crossref","first-page":"1217","DOI":"10.1214\/aos\/1176347265","article-title":"The jackknife and the bootstrap for general stationary observations","volume":"17","author":"H R K\u00fcnsch","year":"1989","journal-title":"Annals of Statistics"},{"issue":"4","key":"ref7","doi-asserted-by":"crossref","first-page":"504","DOI":"10.1214\/aoap\/1177005835","article-title":"Estimating variance from high, low and closing prices","volume":"1","author":"L C G Rogers","year":"1991","journal-title":"Annals of Applied Probability"},{"issue":"3","key":"ref8","doi-asserted-by":"crossref","first-page":"477","DOI":"10.1086\/209650","article-title":"Drift-independent volatility estimation based on high, low, open, and close prices","volume":"73","author":"D Yang","year":"2000","journal-title":"Journal of Business"},{"issue":"2","key":"ref9","doi-asserted-by":"crossref","first-page":"723","DOI":"10.1016\/j.ijforecast.2019.08.010","article-title":"The term structure of volatility predictability","volume":"36","author":"Z Li","year":"2020","journal-title":"International Journal of Forecasting"},{"issue":"4","key":"ref10","doi-asserted-by":"crossref","first-page":"614","DOI":"10.1093\/jjfinec\/nbz025","article-title":"HARK the SHARK: realized volatility modeling with measurement errors and nonlinear dependencies","volume":"19","author":"G Buccheri","year":"2021","journal-title":"Journal of Financial Econometrics"}],"container-title":[],"original-title":[],"deposited":{"date-parts":[[2026,10,15]],"date-time":"2026-10-15T17:34:57Z","timestamp":1792085697000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.ssrn.com\/abstract=7504558"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026]]},"references-count":10,"URL":"https:\/\/doi.org\/10.2139\/ssrn.7504558","relation":{},"subject":[],"published":{"date-parts":[[2026]]},"subtype":"preprint"}}