Discount Rates – What is all the fuss about?
In recent months two interesting research papers were issued. The first, published by the Fraser Institute and authored by Malcolm Hamilton and Phillip Cross is Risk and Reward in Public Sector Pension Plans – A Taxpayer’s Perspective. The second, published by the CD Howe Institute and authored by Stuart Landon and Constance Smith is Managing Uncertainty: The Search for a Golden Discount-Rate Rule for Defined-Benefit Pensions. Both papers focus on discount rates and reveal how the assumptions that actuaries make about future interest rates impacts the recognition of the cost of pension promises over time.
The Hamilton-Cross paper presents a clever way of thinking about the difference between the ‘value’ of a pension promise and the ‘cost’ to fund the promise and the paper delivers to readers the truth that the difference between these two amounts is the reward for taking investment risk. Just a warning, after all the math is done the authors expose an estimated $22 billion annual subsidy that taxpayers are providing to government workers in the form of valuable pensions – the cost of which is under-estimated by pretending that investing in equities over the long-term has no risks. This can be upsetting to readers that worry our governments are spending beyond their means.
Where the Hamilton-Cross paper uses pure math to demonstrate its conclusions, the Landon-Smith paper focuses on the question of funding pensions using a different approach. Landon-Smith use Monte Carlo simulations to examine six different approaches to setting a discount rate. These approaches range between the conservative approach of 3% per annum and the aggressive approach of 9% per annum. In between are ‘discount-rate rules’ such as ‘inflation + 3%’. To be honest, I have never been the type of mathematician that sees the elegant answer that Hamilton and Cross have spotted. I have always been partial to the ‘brute force’ method of mathematics employed by Landon and Smith.
So, after 50,000 simulations and the application of a ‘quadratic loss function’ the Landon-Smith paper concludes several findings:
- “We find that none of the discount-rate rules yield both low median excess assets and a high probability that pension plan assets will be adequate to meet future obligations.”
- “To have a relatively high likelihood of meeting future pension obligations, a plan must, on average, use a discount rate that leads to large median excess assets.
- “A notable feature of the discount-rate-setting rules is that they all yield a large number of cases for which there is either a funding shortfall of more than 20 percent or assets exceed obligations by more than 20 percent.”
- “However, even with the best discount-rate rules, it is necessary on average to hold significant assets to achieve a high probability that plan assets will be adequate to meet future obligations.”
The authors conclude by choosing the discount-rate method that they find best balances the risks of too many assets with too few.
But wait, is that what we should be taking from this paper? Should we stop sponsors of pension plans from systematically encouraging actuaries to choose discount rates at the high end of the range and as a result perpetuate a system that is tilted towards underfunded plans?
To me this is not the conclusion at all. The real learning is:
- The more you mismatch bond-like pension liabilities with equity-like assets the more unpredictable your funding requirements will be and the more likely you will overshoot or undershoot the level of required assets. If you want to miss your target by less, then you should better match the assets and the liabilities.
- Pension funding was not historically designed to exactly track the uncertain liability it was chasing. Rather, pension funding was recognized as a self-correcting mechanism where ‘gains and losses’ were amortized, and contributions were smoothed to get the assets and liabilities to line up at some fictional future date in a long time-horizon.
- Entirely absent from the homework here is a discussion of what happens to the excess assets and the shortfall? We know that in traditional target-benefit plans that the excess/shortfall lands in the laps of plan members (although with some pressure on sponsors through collective bargaining to contribute more). When I started in this business we all knew that in a corporate sponsored plan the excess/shortfall landed with the sponsor. Someone didn’t like Conrad Black and unfortunately changed the corporate game to ‘heads I win, tails you lose’.
In the private sector, sponsors looked at the conclusions above and decided that bond-like investments defeated the investment advantage of pooling pension assets for employees and the lose-lose surplus/deficit proposition made chasing a moving target foolish. Thus commenced the exodus of private sector plan sponsors willing to play the DB game.
It is on the final point that I think we have work to do. This is not work for the actuaries. If governments want DB plans in the private sector – the funding regime needs to reflect two truths. First, accumulating pension savings over decades requires prudent risk-taking and this risk taking shouldn’t be discouraged. Second, everyone should be clear on how the risks and the rewards are shared among stakeholders. DC plans became popular in the private sector because they aligned the party taking the investment risk with the party benefiting from doing so – the plan member.
If you go back to the Hamilton-Cross paper, the beauty of their work is that they expose the fact that while targeting public sector pension contributions at a reasonably ‘expected’ level, there is no accounting for who will pay in the 50% of scenarios where the assets fall short. Governments can resolve this challenge easily by clarifying to public sector workers that they have one of two choices:
- Accept that the government is contributing at a reasonable long-term rate and that any shortfall-excess revealed in the future would be borne by plan members and not the taxpayers as is the case in a target benefit plan; or
- Accept that in exchange for an iron-clad guarantee that pensions will be paid ‘come hell or high water’ that the government will reflect a much higher value to this compensation.
In option 2 above, the government should probably also get out of the business of investing in equities and go back to investing in federal and provincial bonds. But that is a discussion for another day.