Baa spreads gave no clear added weekly Nasdaq variance forecast gain after VIX in a 2016–2025 archive test

We tested whether a Baa corporate bond spread improves next-week Nasdaq Composite variation forecasts after past stock variation and VIX are known. Each forecast used a dated ALFRED archive, an earlier observation cutoff and only past fitting labels. Fixed ridge models used 520 common training weeks. The main test has 483 fresh weeks in 2016–2025. Adding spread level and recent change gave a mean QLIKE loss change of -0.00780, with a 95% paired eight-week block interval from -0.03403 to +0.01396. All planned credit checks gave intervals that include zero. Without 2020, the change was +0.00248. Variance mean squared error was 27.3% higher with credit, with wide uncertainty. Thus the test gives no clear added credit forecast gain in this scope. This is a retrospective test on one price index and one fixed model family. It shares a market sample with earlier SCI work and does not establish a trading profit. The package preserves code, source request records, forecasts, fits, scores, figures, uncertainty estimates, negative results and repeat instructions. Source data and measured outcomes keep their provider terms in separate retained records.

stock risk forecasts · credit spreads · VIX · point-in-time data · chronological test · forecast comparison · negative result