Research preview

The Indian rupee’s recent slide under the 96 level against the US dollar sparked a wave of trading activity across Asia and North America. While headline news focused on macro fundamentals, the price move also illustrates how modern high‑dimensional models can capture and predict such events more reliably than classic low‑dimensional regressions. From Single‑Factor to High‑Dimensional Regressions Traditional FX models often rely on a handful of macro variables—interest‑rate differentials, inflation gaps, and trade balances. In a low‑dimensional regression, adding another predictor, such as a commodity price, typically reduces the in‑sample residual sum of squares....