Is Our Pension System Really Working?
Further to my earlier post on the Mintz report on pensions http://www.progressive-economics.ca/2009/12/20/the-mintz-report-and-the-pensions-debate/ Statistics Canada have released the major study on income replacement rates in retirement by Yuri Ostrovsky and Grant Schellenberg which was cited at some length by Mintz.
The study looks at the incomes of retirees in their early 70s in 2006 in relation to their earnings some twenty years earlier, set against whether or not the individuals when working belonged to a registered pension plan. The abstract reads:
“Data from the Longitudinal Administrative Data (LAD) base are used to compare the retirement status and earnings replacement rates achieved by individuals who were, and individuals who were not, Registered Pension Plan members in 1991 and/or 1992, when they were in their mid-fifties. Among men in this cohort, the likelihood of being retired at age 70 to 72 was about 4 to 14 percentage points higher among pension plan members than non-members. Data used for the study do not provide information on why RPP non-members tend to retire later than do members. Among retired individuals, earnings replacement rates did not differ significantly between RPP members and non-members.”
This study COULD (and will) be used to argue – based on the last line of the abstract – that there really is no pension system problem, since rates of income replacement in retirement differ very little between those who belonged to pension plans while working, and those who did not and thus had to save through other vehicles such as RRSPs and non registered savings.
However, a number of very important caveats are in order, and can readily be found in a close reading of the study.
First, as noted in my earlier post, the study counts earnings in “retirement” as “replacement income” for previous earnings. This seems pretty dubious when we are talking about the “retirement” income of persons in their early 70s, and it is not a trivial issue
As shown in Table 3 of the paper, the difference in annual earnings and self-employment income combined in 2006 between men with no registered pension in 1991 or 1992 and those with a pension in both earlier years is $500 for the first (lowest) quintile; $2750 for the second quintile; $4500 for the third quintile; $5800 for the fourth quintile; and a huge $16,500 for the top quintile. (The quintiles are based on earnings before retirement.) The numbers have a pretty big impact on “retirement” incomes.
Second, pension and superannuation income in retirement is, not surprisingly, found to be higher for the retirees who were previously in pension plans. This category includes incomes from ALL pensions, and some retirees who did not belong to a pension plan in 1989-91 could and (as is acknowledged) probably did have pension income in retirement from previous or later jobs than those held in 1991 and 1992. This is important in looking at the implications for today’s work force since pension plan coverage over the working lifetimes of these 70 years olds was much higher than for those now entering or, for that matter, those soon to be leaving the workforce.
The category of pension and superannuation income is a composite of pension income and of income from RRSPs converted into RRIFs. Likely, much of the “pension” income of retirees who did not earlier belong to registered pension plans came from RRSP savings.
By quintile, the difference in pension and superannuation income for men in pension plans in 1991 and 1992 compared to those who were not was $2600 for quintile 1; $3450 for quintile 2; $3500 for quintile 3; $6150 for quintile 4; and $13,600 for the top quintile. Overall, those not in pension plans had pension incomes (including from RRSP conversions) about one third to one quarter below the level of those in employer pension plans.
By contrast, as one might expect, investment income from non pension vehicles was higher for those who did not belong to pension plans.
Third, receipt of the Guaranteed Income Supplement – which provides a sub poverty line minimum income for the elderly – was much higher for those who did not belong to pension plans in 1991 and 1992. Of those men who were middle income earners but not in pension plans in the early 1990s, many more collected the GIS in their early 70s compared to those in pension plans (27% vs 16% in quintile 3 and 17% vs 7% in quintile 4.)
In summary, a fair conclusion to be drawn from this report is that those who belonged to pension plans were in far better shape for retirement than those who were not. That is why they retired earlier. (I fail to see any dangers in leaping to the conclusion that people with decent pensions retire earlier for that reason.) It is interesting that the post retirement incomes of retirees by previous earning group were fairly equal as between those who belonged and did not belong to employer pension plans. The biggest reason for this is that “retirees” without decent pensions were much more likely to continue to work. Incomes were also equalized by the income tested portion of the public pension system.