2001-02

Archived - Determinants of Retirement: Does Money Really Matter?

Janice Compton

Please note this working paper is available in English only.

• PDF Version - [164 KB]

Abstract

The worker to non-worker ratio in Canada is forecasted to fall dramatically over the next few decades due both to demographic pressures and the recent decline in the average age of retirement. If governments desire to reverse these trends, it is necessary to determine whether altering Canada’s retirement income programs may impact retirement decisions. This analysis focuses on the CPP/QPP program using the internal longitudinal files of the Survey of Labour and Income Dynamics (SLID) and concludes that limited changes to CPP/QPP benefit levels will not have a large impact on the labour force behaviour of older workers. CPP/QPP benefits can affect retirement through two channels – by altering the relative benefits of work (the substitution effect) and by changing the net present value of lifetime wealth (the wealth effect). We first demonstrate that retirement and CPP/QPP benefit take-up are not temporally connected for most Canadians. This means that the substitution effect is unlikely to have an important impact on retirement decisions. We then use hazard regressions and ordered probit analyses to determine which factors influence retirement behaviour. We find no evidence that expected CPP/QPP benefits or current wage levels influence the retirement decision. This result provides further evidence that CPP/QPP benefits are not having a large impact on retirement through the substitution effect. We also find no evidence that wealth affects the retirement decision. Instead, demographic factors such as disability status, labour force status of spouse and class of worker dominate the regressions. This does not necessarily mean that the CPP/QPP program does not influence retirement, only that its influence is likely to be embedded in the structure of the program and the existence of borrowing constraints rather than through the benefit levels. However, the results also indicate that structural changes may not be required in order to raise the retirement age. Variables that do have a strong effect on retirement behaviour are moving in directions that may result in a reversal of the early retirement trend.