Pham, Son DMarshall, Ben RNguyen, Nhut HVisaltanachoti, Nuttawat2026-02-122026-02-122023-01-10Australian Journal of Management, ISSN: 0312-8962 (Print); 1327-2020 (Online), SAGE Publications, 49(3), 478-508. doi: 10.1177/031289622211434940312-89621327-2020http://hdl.handle.net/10292/20630Substantial transaction costs are incurred in exchange-traded fund (ETF) trading each year. This article examines a vector autoregressive (VAR) model’s performance and other trading schedules to time trades in a large sample of 1350 ETFs over the 2011–2017 period. We reject the notion of a one-size-fits-all trading schedule that maximizes spread savings for all ETF traders. ETF traders who want to split their orders could save 7.40% of ETF spread costs, whereas trading at the market closing time would be optimal for ETF traders without motives to split trades. The spread savings for ETF traders are diverse across ETF sectors and depend on the spread volatility.This is the Author's Accepted Manuscript version of an article published in the Australian Journal of Management by Sage Publishing © The Author(s) 2023, the Version of Record available at DOI: 10.1177/031289622211434943502 Banking, Finance and Investment35 Commerce, Management, Tourism and Services15 Commerce, Management, Tourism and ServicesBid-ask spreaddiversificationETFsforecastingliquidityportfolio liquidityJournal ArticleOpenAccess10.1177/03128962221143494