{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,29]],"date-time":"2026-09-29T19:59:08Z","timestamp":1790711948243,"version":"4.1.0"},"posted":{"date-parts":[[2026]]},"group-title":"SSRN","reference-count":0,"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>This study compares break-adjusted GARCH and deep-learning approaches to forecasting volatility in Brent crude oil spot prices, the USD\/KES exchange rate and the NSE 20 Share Index. Daily returns are examined for variance shifts using a heteroskedasticity-robust modified ICSS procedure. Expanding-window, rolling-window and break-adjusted GARCH forecasts are evaluated alongside neural-network forecasts at 1-, 20-, 60-and 120-trading-day horizons using rolling-origin validation and multiple loss measures. Two significant variance breaks are identified for the NSE 20 Share Index and six for USD\/KES; none is detected for Brent, despite conspicuous crisis-period volatility. For the equity series, the FIGARCH persistence parameter declines from 0.259 (p = .017) before break adjustment to 0.142 (p = .382) afterwards, indicating that structural change explains part of the apparent long memory. Forecast rankings vary across markets, horizons and loss measures. Break-adjusted GARCH and FIGARCH are competitive at the one-day horizon for equities and USD\/KES, respectively, whereas CNN-LSTM performs strongly at longer equity horizons and LSTM leads the reported longer-horizon USD\/KES comparisons. Conventional GARCH variants remain competitive for Brent at longer horizons. By evaluating variance breaks, persistence and multi-horizon predictive accuracy within a common framework, the study shows why neither parameter stability nor the general superiority of deep learning should be assumed in forecasting volatility is a frontier-market. Its findings support asset-specific model selection and regular reassessment after volatility shifts.<\/jats:p>","DOI":"10.2139\/ssrn.7537899","type":"posted-content","created":{"date-parts":[[2026,9,29]],"date-time":"2026-09-29T18:55:35Z","timestamp":1790708135000},"source":"Crossref","is-referenced-by-count":0,"title":["Break-Adjusted GARCH Versus Deep Learning: Multi-Horizon Volatility Forecasts for Kenyan Equity, Currency and Global Oil Markets"],"prefix":"10.2139","author":[{"given":"Abraham","family":"Wawire","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Christine","family":"Simiyu","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Munene","family":"Laiboni","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Rogers","family":"Ochenge","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"78","container-title":[],"original-title":[],"deposited":{"date-parts":[[2026,9,29]],"date-time":"2026-09-29T18:55:35Z","timestamp":1790708135000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.ssrn.com\/abstract=7537899"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026]]},"references-count":0,"URL":"https:\/\/doi.org\/10.2139\/ssrn.7537899","relation":{},"subject":[],"published":{"date-parts":[[2026]]},"subtype":"preprint"}}