Loading…
FX BANK FORECAST · COVERAGE
Institutional FX coverage in your inbox
Aggregated year-end forecasts, scenario shifts, and curated analyst notes from 36 institutional desks. No promotion.
FX BANK FORECAST · COVERAGE
Aggregated year-end forecasts, scenario shifts, and curated analyst notes from 36 institutional desks. No promotion.
Fannie Mae's Economic & Strategic Research group publishes a housing forecast every month. We have scored 140 consecutive monthly releases (2015-01 to 2026-08) against the values that actually realized, sourced from the Federal Reserve.
Only forecasts that were entirely ahead of the firm when published are counted. A figure for a quarter already underway is a nowcast, not a forecast, and is excluded.
Error in basis points against the quarterly average of Freddie Mac's PMMS survey (FRED: MORTGAGE30US). The naive column is what you get by assuming the last published rate simply persists.
| Horizon | Fannie Mae | Naive forecast | Independent quarters |
|---|---|---|---|
| 3–6 months out | 34.2 bps | 45.5 bps | 45 |
| 6–12 months out | 56.4 bps | 66.2 bps | 44 |
| 1–2 years out | 95.9 bps | 104.0 bps | 42 |
| Over 2 years out | 146.9 bps | 154.5 bps | 9 |
Shown separately from quarterly figures: a full-year average is smoother than a single quarter and scores better, so the two are never combined.
| Horizon | Fannie Mae | Naive forecast | Independent years |
|---|---|---|---|
| 1–2 years out | 79.8 bps | 85.3 bps | 10 |
| Over 2 years out | 126.3 bps | 139.5 bps | 8 |
Error as a percentage of the realized quarterly average (FRED: HOUST, seasonally adjusted annual rate).
| Horizon | Fannie Mae | Naive forecast | Independent quarters |
|---|---|---|---|
| 3–6 months out | 6.60% | 6.66% | 44 |
| 6–12 months out | 8.24% | 7.48%worse | 43 |
| 1–2 years out | 10.13% | 8.99%worse | 41 |
| Over 2 years out | 12.23% | 9.76%worse | 9 |
Full-year averages, kept separate from the quarterly figures above.
Compare every row against the naive column before reading anything into it. A naive forecast simply assumes the last published value persists. Fannie Mae's mortgage-rate forecasts beat it in 6 of 6 horizons; the housing-starts forecasts beat it in 1 of 6. Beating a naive forecast is the minimum bar for a forecast to be worth anything, and these figures are recomputed from the data on every run rather than quoted from a past review.
"Independent quarters" is the honest sample size. A firm publishing monthly restates the same quarter many times; counting each restatement would multiply the apparent evidence several-fold. The count shown is distinct outcomes.
Actuals use current published data. Housing starts are revised for about two months after first print, so a forecast is scored against a revision the firm could not have seen when it was made. This is standard practice and applies equally to every horizon.
Horizons under three months are not shown. On a quarterly average, a horizon that short means the quarter is already partly complete.
| Horizon | Fannie Mae | Naive forecast | Independent years |
|---|
| 1–2 years out | 7.72% | 6.58%worse | 10 |
| Over 2 years out | 9.14% | 8.19%worse | 8 |