Example: Case-Shiller house prices (6-month ahead)

National home-price growth from months' supply, the mortgage–Treasury spread, and last 6-month appreciation

This model estimates the six-month log return of the S&P/Case-Shiller U.S. National Home Price Index from months' supply, the 30-year mortgage minus the 10-year Treasury, and the prior six-month log return, all dated six months earlier. The printed index is the last observed Case-Shiller print times exp of that fitted growth. Coefficients are fit through December 2023 and then frozen. This is the collateral series behind 1–4 family and HELOC exposure, not another commodity price.

• MSACSR: months' supply of new houses for the month six months before the month being forecasted
• MORT_SPREAD: 30-year fixed mortgage rate minus the 10-year Treasury yield for that same month
• CS_DLOG6_LAG: six-month log change in Case-Shiller ending in that same month

Our growth regression is

CSUSHPISA_estimated(t) = CSUSHPISA(t−6) * exp(MSACSR * -0.011892210 + MORT_SPREAD * 0.023952034 + CS_DLOG6_LAG * 0.401874048 + 0.043383157)
and yields R^2 = 99.2% as of December 2023 (n = 240 months) in-sample on paired actuals through 2023Q4.
On later paired actuals the same frozen coefficients yield R^2 = 38.2% as of July 2026 (n = 31 months) out of sample.

Coefficients are ordinary least squares of the six-month log return on complete lagged pairs from 2004 through 2023Q4. The printed index is the last observed Case-Shiller print times exp of that fitted growth. Later dates, including the period shown as a forecast, use those frozen coefficients. On the leftover paired actuals that frozen fit beats predicting the holdout mean. The first R² above is in-sample through 2023Q4 on the reconstructed index; the second is out-of-sample on later paired actuals.