The use of simulation models in forecasting tax revenues

This paper provides an overview of the modelling techniques and approaches that Revenue Canada - the federal tax administrator - used to simulate a proposed tax regime, Goods and Services Tax (GST) that was intended as a replacement of an existing tag the Federal Sales Tax (FST). The primary, challenge for Revenue Canada,vas twofold: 1) the orderly wind down of the old lax regime; and 2) the successful launch of the new lax regime; to avoid any double taxation and/or cascading of taxes. The results of the macro forecasts produced by the Department of Finance, who has responsibility for tax policy and fiscal policy, did not provide enough detail suitable for use by Revenue Canada in planning the identification of registrant base or amount of revenues expected by sector for the new im. Revenue Canada, therefore, undertook a modelling approach (consistent with the macro approach of the Department of Finance) to disaggregate the national forecasts to the sectoral level. It was necessary to reconcile our results to the official numbers from the Department of Finance. The paper will touch on the challenge of modelling without any historical framework given that a key design objective was revenue neutrality, i.e. the new tax (GST) yield must equal that from the tax (FST) being replaced The paper will cite examples where other studies and microsimulations were carried out and linked to our macro model.