Does anyone know of any open-source U.S. data sets in a format appropriate to whole-economy CGE econometric modeling, i.e. data sets covering the entire economy, divided into at least 20 sectors, with annual price and quantity indexes for each sector? And I think I'd need 20 years or so of reasonably self-consistent data to have any hope of estimating flexible functional forms for production functions, etc.
I know that to really do this right you probably need to combine data from multiple sources and become familiar with their idiosyncrecies. I've done that for certain sectors. But I'd like to see if I can embed my sectors in a whole-economy CGE, and I've also been thinking that I could explore a richer system of own and cross-price elasticities than I have believed in the past to be feasible by using LASSO or some other form of penalized model to prevent excessive dimensionality.
Would BEA IO account data work for this? I think last time I checked the price part of the IO model was only updated every five years. But that was some years ago.