Determine whether high-frequency reconfiguration contains unpriced structural information

Determine whether the high-frequency component of the reconfiguration index contains structural information that is simply unpriced, rather than being dominated by re-estimation noise, by directly testing the direction of eigenspace increments rather than only their magnitude.

Background

The paper decomposes the reconfiguration index into a persistent three-month moving average and a high-frequency residual. Only the persistent component is priced in the variance-risk-premium regressions, while the high-frequency component is not statistically associated with the premium.

Because consecutive twelve-month estimation windows overlap by eleven months, the authors argue that the high-frequency component contains substantial re-estimation noise, or “wobble.” However, the reported autocorrelation of the index magnitude is compatible with contamination but cannot determine whether the high-frequency increment also contains genuine structural information that the market leaves unpriced. Resolving this issue requires a direct analysis of the increment’s direction rather than its size.

References

Whether it also carries structural content that simply goes unpriced is not settled here: the autocorrelation of the magnitude series is consistent with contamination but does not establish that the increment is noise-dominated, and separating the two would require a direct test on the direction of the increment rather than its size.

— The Reconfiguration Premium: Co-movement Structure as an Unspanned Dimension of the Variance Risk Premium  (2608.20020 - Carvalho, 20 Aug 2026) in Section 5.3, “Why smoothing helps”