This paper introduces a microsimulation model that simulates the budgetary impact of the 2006 regime change in the Dutch disability scheme. A dynamic population model fits the case of the disability benefits the best. As opposed to macro forecasts, a microsimulation can answer questions about the individual or meso income effects, the exact distribution of expenses among different benefits and the time path of the savings. The introduction of the proposed system change decreases the number of disability benefits by more than 25% from 2020 onwards and reduces total costs by almost €2 billion or 20%. Based on the better incentive structure, participation will increase and boost GDP. Microsimulation can be used to pick the winners and losers of the new system and give the time path of the savings. For almost all partially disabled that are working, the total discounted income after the system change is as large as or larger than before the system change, for the non-working total discounted income is lower. Overall, the Gini-coefficient slightly decreases due to the regime change. © 2006 Elsevier B.V. All rights reserved.