Yann Decarie - A New Canadian Retirement Income Microsimulation Model

  • Presenting author: Yann Decarie (HEC Montreal)

  • Authors: David Boisclair, david.boisclair@hec.ca; François Laliberté-Auger, francois.laliberte-auger@hec.ca; Pierre-Carl Michaud, pierre-carl.michaud@hec.ca. Daniel Blanchette, daniel.blanchette@statcan.gc.ca Héloïse Gauvin, heloise.gauvin@statcan.gc.ca Chantal Hicks, chantal.hicks@statcan.gc.ca Jennifer Jones, jennifer.jones@statcan.gc.ca Kevin D. Moore, kevin.moore@statcan.gc.ca Mahbubur Rahman, mahbubur.rahman@statcan.gc.ca Jessie Yeung, jessie.yeung@statcan.gc.ca Zhe Si Yu, zhesi.yu@statcan.gc.ca Li (Grace) Zhuolin, zhuolin.li@statcan.gc.ca

  • Session: A02A - Dynamic / Long term [1] - Monday 16:30-18:00 - Ceremonial Hall

  • Slides: PDF

In partnership with Employment and Social Development Canada (ESDC), Statistics Canada and HEC Montréal teams have developed a large-scale dynamic socio-economic microsimulation population model to meet policy and program needs for high quality information and projections in the retirement income domain. The model will be available to the research community and will be compatible with all operating systems (Windows, Mac and Linux). It is coded in OpenM++, an open-source microsimulation platform derived from Modgen. The initial focus of the project has been to model the detailed provisions and take up of the second-pillar Canada Pension Plan (CPP) and its various benefits, as well as key lifetime economic and demographic attributes of individuals. These include birth, education, unions, migration, mortality, employment, and earnings. In recent years, these types of models have become core analytical tools for many governments, as they are ideal for framing and exploring “what- if” questions. These types of models create a virtual “policy laboratory “ for exploring socio-economic outcomes and detailed implications of policy changes, including complex interactions within and between programs, and for assessing the distributional impacts of policy choices. These models are the only tools capable of effectively evaluating policies with an explicit individual longitudinal dimension, such as pensions, whose outcomes depend on lifetime earnings and coverage histories and links between individuals. Is this paper, we will introduce the model itself but also discuss the effort that was put in creating unique datasets (including a synthetic, public-use dataset) by merging Canadian long form Census and administrative data from various sources (including tax files and CPP administrative files). We will also present how we faced the issues specifically related to the complexity of the CPP.