Government Revenue Forecasting

We develop econometric equations and then construct a model to forecast sales and use taxes by sector, personal or individual income taxes on wages and non-wage income, corporate income taxes, excise taxes, and property taxes. Model equations use our proprietary economic and demographic forecasts as inputs. We then use our econometric expertise to estimate statistically significant equations, unique to each jurisdiction and tax type being consider, that explain tax revenue levels over time based on trends in our forecast variables. Similar approaches are also deployed to forecast receipts from non-tax sources such as license fees, room charges, user fees, and tolls on transportation infrastructure.

To project the local tax revenues, we couple our tax models with our detailed forecasts of the macroeconomic, state, MSA and local economies incorporating such variables such as wage and salary disbursements, personal income, retail sales, consumer expenditures by type of good or service including motor vehicles, unemployment rate, value added by sector, output, population change, and home prices. One major S&P Global competitive advantage in forecasting state and local government revenues is that we already have econometric forecasting models for states and MSA with historical and forecast values for all the major economic and demographic variables that drive government revenues. We then develop and link to them a module that explains the econometric relationships between tax collections and changes in economic activity.  Another advantage is that we use a top-down approach in developing our state and MSA forecasting models. As a result key variables from our energy, industry, US macroeconomic, and global economic forecasts are automatically considered when the state and MSA models are run.

Our state econometric forecasting models already contain a number of variables that are essential to forecasting different types of tax revenues, including: total personal income, wage and salary payments by 2-digit NAICs code, two categories of retail sales, new passenger and light truck vehicle registrations, 21 types of personal consumption expenditures, the CPI, state and local personal tax payments, shipments of manufactured goods and industrial production.  The models also contain, as applicable in an individual state, current tax rates for nine types of major state-level taxes. 

Because of the way our state and MSA forecasting models are designed, and the software used to run them, they can easily be modified to compare the revenue effects of alternative tax policies such different rates, types of goods and services subject to a tax, even alternative economic growth scenarios.  

Our methodology for forecasting government revenues is shown in the flow chart presented below. The figure clearly shows how we leverage all our information and forecasting assets to derive government revenue estimates that are driven by levels of economic activity at the regional, state, US, and global levels.