Model Approach

It takes a multi-step approach to construct economic input-output (IO) tables in the form of two-region, state of interest (SoI) and rest of the United States (RoUS), at the level of summary resolution as defined by the U.S. Bureau of Economic Analysis.

The following table lists the 4 steps needed to generate two-region IO tables. Each step is a prerequisite for the next step. To generate data in each step, manually trigger the designation workflow in GitHub Actions.

Step	Data Group	Data Set	Script (in /data-raw)	GitHub Actions Workflow
1	Raw Data 1	BEA data 2, Census data 3, EIA data 4, FAF data 5, USASpending data 6	BEAData.R, CensusData.R, EIAData.R, FAFData.R, USASpendingData.R	save-raw-data.yaml
2	State Supply Model Data	State Make table and Commodity Output 7	StateSupplyModel.R	save-state-supply-model-data.yaml
3	State Use Model Data	State Use table, Domestic Use table, and Industry Output	StateUseModel.R	save-state-use-model-data.yaml
4	Two-Region Model Data	Two-Region Make, Use, Domestic Use, and Domestic Use with trade tables, Industry and Commodity Output, Value Added	TwoRegionModel.R	save-two-region-model-data.yaml

¹ The BEA employment data is generated and saved by executing BEAData_wAPI.R instead of running the GH-Actions-based workflow due to requirement of a personal BEA API key.

² BEA data includes State Employment Compensation, GOS, GVA, PCE, Tax, Government Consumption and Investment.

³ Census data includes State & Local Government Expenditure, State Trade.

⁴ EIA data includes State Electricity Consumption and Trade.

⁵ FAF data includes State Commodity Flow.

⁶ USASpending data includes Federal Government Expenditure.

⁷ Commodity Output will be validated against Two-Region Commodity Output after Step 4.