It has become a truism that DC plans impose too much risk on employees, but that belief is rarely quantified.
One could point to anecdotal evidence but that isn’t conclusive. A more scientific approach is to run a Monte Carlo simulation to find the distribution of future outcomes, but that is only as good as the input. Finally, one could gauge how DC plans would have done based on historic market performance. This last approach is the approach we have taken, using market data from 1938 to 2012.
Let’s assume that ABC Co. has been maintaining a DC pension plan continuously since 1938. The plan requires a contribution of 8% of pay and is invested 60% in equities, 40% in bonds. At retirement, members purchase a fully indexed annuity with a 10-year guarantee. At the time of launch, ABC Co. decided that the target pension for a 30-year employee should be 30% of final average pay.
Read more by Fred Vettese
With this background, we calculated DC pension results over 30-year periods starting with 1938-1967 and ending with 1983-2012. As expected we found the DC pension varied widely depending on the period. The highest pension generated by the plan occurred in 2001 when retirees in that year would have been able to buy a pension equal to 55% of final average pay; 1975 retirees had the worst result with a pension of just 15%. Many will perceive this as further proof that DC arrangements are unacceptable but we decided to delve a little deeper.
In real-life situations, when an individual is confronted with an unfavourable outcome, he or she will take evasive action if possible.
In this case, we assumed that whenever the ABC Co. plan produced a DC pension of less than 30% of final average pay, the individual would do two things: (a) work two years longer and (b) settle for an annuity that was only 50% indexed rather than 100%.
We ran our projections with these modifications and found the range in pension narrowed to between 24% and 55%. One might say that this is still much too wide a range but notice that most of the volatility is on the upside.
In our modified projection, the pension was 30% or more in 43 out of the 46 30-year periods and never fell below 24%. Suddenly, this DC plan starts to look like a success story!
One of the lessons learned from this is to choose a reasonable target. When it comes to DC pensions, it is better to under-promise and over-deliver. Having said that, a pension of 30% of final average pay seems to be a reasonable target relative to an 8% contribution rate.
By comparison, the Canada Pension Plan has a target pension of 25% with a 9.9% contribution rate and the death benefit is less generous than in the ABC Co. plan.
We then reversed the model to see how the cost to provide a 30% DB pension would have varied over time. For the period 1972-2001, the level annual cost would have been 4.3% of pay which is the lowest it ever got. In the worst period, which was 1946-1975, the level cost would have been 16.1% of pay. It is no surprise that traditional DB plans are being closed down faster than drive-in theatres.
As expected, a 60/40 asset mix produced a lower cost over 30 years than a 40/60 mix. The surprise was that it also produced a more narrow range of cost, in other words, more certainty. The range for the 60/40 mix was 4.3% to 16.1% of pay versus 4.5% to 18.1% for the 40/60 mix. It seems that a higher bond holding reduces short-term volatility but not long-term volatility. This may be another reason to favour target benefit plans which are expected to have the accounting treatment of DC plans and thus have the luxury of taking a longer term approach to investments.
As always, we should point out that past performance is not necessarily an indicator of the future. All we can say for certain is that DC plans—and by extension, target benefit plans—would have withstood the turmoil of the last 75 years.