A better forecast informs decisions. It doesn’t make them.
The model produces one thing: a measurably better rent-growth forecast (+17.2% forecast-error reduction vs a fitted AR baseline at 3 months). That forecast is an input to underwriting and development models under a known rate environment — not a buy/sell/build signal. We treat the rate channel as exogenous (a number the user supplies), because we forecast rent growth (an NOI driver), not prices.
A more accurate 3-to-12-month rent-growth forecast tightens NOI and DSCR assumptions on the downside. It informs how conservatively to underwrite or whether to pass — it does not flag a market to exit.
A better forward rent-growth estimate is an input to a build pro-forma alongside cost and the rate environment. It sharpens the demand assumption; it does not tell you to break ground.
The forecast informs the rent-growth assumption in an acquisition model. Valuation still turns on the rate/cap-rate environment, which we treat as a known input the user supplies — the model does not forecast prices, cap rates, or returns.
- Not a trading signal, and not a claim that acting on it “makes money.”
- Not a turn detector and not a timing tool — it is a continuous forecast with a measured error reduction.
- Not a price or cap-rate forecast — the rate environment is a known input, not something the model predicts.
- Not a per-metro score — skill is pooled across the covered metros.