{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,12]],"date-time":"2026-10-12T14:47:09Z","timestamp":1791816429468,"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>Macroeconomic regime models are widely proposed for strategic and tactical asset allocation, yet the literature rarely tests whether regime information retains its value once it must be economically interpretable, estimated out of sample, applied across a global multi-asset universe, and traded after costs. This paper implements an integrated, interpretable machine-learning framework that links (i) growthinflation-stress regime identification; (ii) calibrated supervised regime-probability forecasts from gradient boosting, benchmarked against a Markov chain, a multinomial logit, a Gaussian hidden Markov model (HMM) and k-means; (iii) feature-block attribution of the regime signal; and (iv) regime-conditioned portfolio construction with CVaR optimization and Black-Litterman, organized into a strategic (SAA) and a tactical (TAA) layer with no-trade bands. We evaluate 13 strategies on a ten-ETF universe over 187 walkforward months (May 2010-November 2025) with asset-class transaction costs, a studentized blockbootstrap Sharpe test and the deflated Sharpe ratio. The primary strategy (gradient-boosting regimes, CVaR, bands) earns a net Sharpe ratio of 0.60, against 0.90 for a 60\/40 SPY\/IEF portfolio (difference not statistically significant, p = 0.21) and 0.70 for a static SAA benchmark (lower, p = 0.04, unadjusted). No regime-conditioned variant outperforms the benchmarks before costs, so the break-even cost multiple is zero for all of them. Regime strategies do reduce volatility and tail risk relative to 60\/40 (monthly CVaR 95% of-4.1% vs-5.5%; 2022 loss of-13.8% vs-16.4%), but they do so mainly by holding a bond-and gold-heavy portfolio rather than through regime timing. Notrade bands cut turnover by 42-45% and reduce cost drag, but there is no gross advantage to preserve. Economically anchored regimes perform on par with HMM and k-means regimes, and the stock-bond correlation is positive only in the stagflation and disinflation regimes (+0.19 and +0.05), consistent with the regime-dependence of nominal-bond diversification. Gradient-boosting probabilities are the best calibrated of all forecasts but do not beat a simple Markov chain.<\/jats:p>","DOI":"10.2139\/ssrn.7595598","type":"posted-content","created":{"date-parts":[[2026,10,12]],"date-time":"2026-10-12T14:07:37Z","timestamp":1791814057000},"source":"Crossref","is-referenced-by-count":0,"title":["From Regimes to Portfolios: An Interpretable Machine-Learning Framework for Strategic and Tactical Cross-Asset Allocation"],"prefix":"10.2139","author":[{"given":"Chalam Raja","family":"Ankam","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"78","container-title":[],"original-title":[],"deposited":{"date-parts":[[2026,10,12]],"date-time":"2026-10-12T14:07:37Z","timestamp":1791814057000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.ssrn.com\/abstract=7595598"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026]]},"references-count":0,"URL":"https:\/\/doi.org\/10.2139\/ssrn.7595598","relation":{},"subject":[],"published":{"date-parts":[[2026]]},"subtype":"preprint"}}