{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,13]],"date-time":"2026-10-13T06:48:07Z","timestamp":1791874087320,"version":"4.3.4"},"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 paper asks whether machine-learning models can extract persistent cross-sectional return-ranking information from a small, illiquid emerging equity market, and whether that statistical predictability survives the trading frictions of that market. Using a 12-security Nairobi Securities Exchange universe, walk-forward validation over 2024 provides evidence of positive cross-sectional ranking information. Random Forest is the strongest tested estimator, with mean daily Spearman Information Coefficient (IC) of 0.133 and positive monthly IC in 12 of 12 months; the originally selected LightGBM model achieves mean daily IC of 0.085 in the model-comparison experiment. Feature-group experiments show that simple returns and price information contain much of the useful signal: a returns\/price-only ablation achieves mean daily IC of 0.125, compared with 0.111 for the full feature set under the ablation protocol. LLM-derived event variables contribute no measured incremental IC in the evaluated historical window. A retrospective extended out-of-sample evaluation over 2025-2026 remains positive, with mean daily IC of 0.0815 and positive monthly IC in 11 of 15 months, but is not a pristine live holdout. Economically, the evidence is much weaker: an overlapping three-day Top-1 portfolio loses approximately 94.2% cumulatively under a 400 basis point round-trip cost assumption. The main finding is therefore a gap between statistical predictability and implementable profitability in a high-friction market.<\/jats:p>","DOI":"10.2139\/ssrn.7597360","type":"posted-content","created":{"date-parts":[[2026,10,13]],"date-time":"2026-10-13T06:05:06Z","timestamp":1791871506000},"source":"Crossref","is-referenced-by-count":0,"title":["Cross-Sectional Return Prediction in a Small, Illiquid Equity Market: Machine Learning Evidence from the Nairobi Securities Exchange"],"prefix":"10.2139","author":[{"ORCID":"https:\/\/orcid.org\/0009-0001-7087-8586","authenticated-orcid":true,"given":"Felix","family":"Nyabuto","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"78","container-title":[],"original-title":[],"deposited":{"date-parts":[[2026,10,13]],"date-time":"2026-10-13T06:05:06Z","timestamp":1791871506000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.ssrn.com\/abstract=7597360"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026]]},"references-count":0,"URL":"https:\/\/doi.org\/10.2139\/ssrn.7597360","relation":{},"subject":[],"published":{"date-parts":[[2026]]},"subtype":"preprint"}}