Compare methods for forecasting US public-budget tax and fee collections. A worked monthly example shows how to reconcile forecast receipts, explain shortfalls and separate recurring changes from collection timing.
Government revenue forecasting methods estimate future tax and fee collections from historical receipts, economic drivers, collection schedules and documented adjustments. Use a different method where a source behaves differently, then reconcile the combined forecast to a clearly defined reporting basis. A single growth percentage applied to every source hides the reason the forecast changes.
The practical starting point is a source register: what generates the receipt, what determines its amount, and when the money arrives. Forecast each material source separately before combining funds or departments.
| Method | Useful starting point | Evidence needed | Main limitation |
|---|---|---|---|
| Historical trend | A recurring source with comparable periods | Prior receipts, seasonality and identifiable exceptions | Past growth may not continue |
| Driver-based forecast | Fees or taxes tied to measurable activity | Activity forecast, applicable rate and collection assumptions | A strong activity forecast can still have weak receipt timing |
| Collection schedule | Installments, remittances or known payment dates | Amount due, due date and expected collection pattern | Amounts due are not automatically amounts received |
| Statistical relationship | A source linked to economic conditions | Comparable receipt history and relevant economic series | Relationships can change after policy or economic shifts |
| Documented judgment | A known change missing from the baseline | An explicit adjustment, explanation and responsible owner | An unexplained override cannot be evaluated later |
A statistical relationship is useful when it improves an out-of-period comparison against a simpler baseline. More variables do not establish that improvement. Retain the baseline and test which method would have performed better using information available at the time.
Public agencies also connect economic forecasts with source-specific revenue analysis. For example, the Oregon Department of Transportation describes its economic forecast input and publishes separate forecast-variance reports. That is a useful example of keeping the forecast and its subsequent evaluation visible.
Use receipt records with receipt date, revenue-source code, fund, net amount, applicable service or tax period, and any refund or reversal reference. Add a forecast file with the same source and fund codes, forecast month, issue date, method and assumptions.
Choose the comparison basis before calculating growth. The example below measures monthly cash collections, net of refunds. It does not claim that collections equal revenue recognized under every public-sector accounting basis.
Keep transfers, borrowing proceeds and restricted receipts identifiable rather than treating every deposit as a recurring tax resource. Preserve original source codes even when management wants broader reporting groups. A revised grouping should still reconcile to the original export.
Outside data is a reasonableness check, not a replacement for local records. The Census Bureau's quarterly tax survey provides national estimates and detailed state tax data; its geographic coverage and quarterly frequency differ from a city's monthly collections. Compare like sources and periods before drawing a conclusion from it.
For the distinction between a fixed spending benchmark and changing expectations, use budget vs forecast.
Consider an illustrative US city preparing its October 2026 collections forecast. All assumptions below are hypothetical, not current tax rates or economic predictions. Property-tax receipts refer only to the scheduled installment in this example.
| Source | Starting amount (USD) | Expected permits | Fee per permit (USD) | Assumption | Calculation | Forecast collections (USD) |
|---|---|---|---|---|---|---|
| Sales tax | 480,000 | 4% growth | 480,000 × 1.04 | 499,200 | ||
| Permit fees | 320 | 125 | Volume × fee | 320 × 125 | 40,000 | |
| Property-tax installment | 900,000 | 98% collected in month | 900,000 × 0.98 | 882,000 | ||
| Total | 499,200 + 40,000 + 882,000 | 1,421,200 |
The sales-tax starting amount is prior comparable collections; the property-tax starting amount is the scheduled installment due.
The sales-tax row is a trend baseline. The permit row is activity driven. The installment row estimates collection timing. Combining the methods does not require pretending that the revenue sources share the same behavior.
After close, the receipts export gives these results:
| Source | Forecast collections (USD) | Actual collections (USD) | Actual minus forecast (USD) |
|---|---|---|---|
| Sales tax | 499,200 | 485,000 | -14,200 |
| Permit fees | 40,000 | 38,000 | -2,000 |
| Property-tax installment | 882,000 | 873,000 | -9,000 |
| Total | 1,421,200 | 1,396,000 | -25,200 |
The permit register records 304 paid permits at $125 each: 304 × 125 = $38,000. That explains the fee shortfall through activity, without implying a rate change.
Total error is $1,396,000 − $1,421,200 = −$25,200. Relative to the forecast, the shortfall is 25,200 / 1,421,200 = 1.77%, rounded to two decimal places. Using the forecast denominator answers how far collections missed the issued expectation; label a different denominator explicitly.
The source errors must sum to the overall error: −$14,200 − $2,000 − $9,000 = −$25,200. Actual source collections must equal the same-period ledger control total. Reconcile receipt reversals, deposit cutoffs and unidentified receipts before interpreting the gap.
Then distinguish timing from a changed outlook. A delayed installment may move into the next month while the full-year expectation stays unchanged. Fewer permits can reduce both monthly and full-year collections. Do not automatically annualize either monthly miss.
Keep a sensitivity separate from the issued forecast. If the sales-tax forecast were reduced by 2%, holding the other rows unchanged, that row would become 499,200 × 0.98 = $489,216. Total forecast collections would become $1,411,216, a $9,984 reduction. This is an assumption test, not evidence that the revised amount will occur.
A rolling forecast provides the updating framework. The cash flow forecast handles the broader question of whether receipt dates cover payment dates.
The budget or finance team should maintain the consolidated forecast and its version history. Revenue administrators explain filing patterns, refunds and collection changes; service departments supply fee activity assumptions. The responsible officials evaluate resource and service implications within their established budget process.
A review should end with an explanation and next step for each material source: retain the assumption, revise future periods, or investigate a reconciliation difference. Revenue forecasting automation covers the repeatable comparison work; it does not replace ownership of the assumptions.
Upload a collection-history file and a forecast comparison file with source, fund and period columns. Covirage maps the columns; its tools compute and check the totals and variances, and the external AI model explains the results with every figure traced to its rows. Create a free Covirage account and bring a sample of your own data.
See Covirage for FP&A reporting for the supporting reporting context.
Choose by revenue source. A scheduled property-tax installment needs collection timing assumptions; permit fees need activity volumes and approved fees. A historical trend provides a baseline for a stable source, while a documented economic relationship may explain a more variable tax base.
State the purpose and accounting basis first. A collections forecast estimates when cash arrives. A budget revenue forecast follows the government's adopted reporting basis. Maintain a bridge between them when recognition, remittance delays or year-end adjustments put the same activity in different periods.
Set an update calendar that fits the budget process and collection cycle. Review new receipts when they become available, but keep each previously issued forecast unchanged for evaluation. Update assumptions when evidence changes and label the replacement forecast with its issue date.
Compare actual and forecast collections for the same sources, dates and basis. Report dollar errors by source alongside a total percentage. Inspect absolute errors too, because overestimates and underestimates can cancel in the total while individual revenue sources remain poorly forecast.