Analysts flag data gaps and corporate refund volatility as key risks to state forecast

At the Jan. 16 briefing, Tom Cravat highlighted two areas that could change the current near‑stable forecast: gaps in national statistical series and concentrated corporate refund activity.

On national data, Cravat said recent discontinuities and delays in federal series — notably the CPI shelter component — make some public measures less reliable and can bias state forecasts downward. ‘‘It’s not good statistics,’’ he told members, explaining that missing October shelter data effectively inserts a zero into a six‑month moving average and can depress measured shelter inflation.

On corporate receipts, staff observed an unusual spike in refunds in late 2025 that depressed collections. Cravat said staff are analyzing refund files to determine whether the activity reflects timing and backward‑looking settlements or signals a more persistent drop in corporate income. He described possible technical causes including carry‑forward adjustments, refund settlements of older returns, and large firms’ investment patterns (for example, expensing of AI investment that reduces near‑term profits).

Cravat said the office uses Moody’s Analytics model constructs for some assumptions but is attempting to substitute other indicators or model approaches where national series are unavailable. Members asked staff to return with any new findings and to work with national forecasting firms where possible.