I don’t claim to be an actuary, but how can a simple formula provide an appropriate valuation of a company’s drug plan liability for dynamic post-employment healthcare expenditures?
Why would any plan sponsor want to overstate their estimated liabilities for future healthcare expenditures, and risk having to pull money out of other areas to fund a liability that may not exist to the degree that has been assumed?
It was only two weeks ago that media outlets were reporting how Ontario universities are up against a brick wall because they may have to start making some extremely significant payments to fund pension deficits. If that’s the case, maybe it’s time plan sponsors with significant post-employment benefit obligations dug a bit deeper and put the Excel formulas to bed as it relates to these healthcare liabilities.
Here’s an example from well-known BCE. The following is the company’s 2009 Annual Report.
• The company had total liabilities as of December 31, 2009 of $20 billion, of which $2 billion was related to accrued benefit liabilities.
• Note 1 to the Consolidated Financial Statements states: “Pension and other post-employment benefit costs are determined using management’s best estimate of the plan’s expected investment performance, pay increases, retirement ages of employees and expected healthcare costs.
• Note 23 to the Consolidated Financial Statements states under the heading Significant Assumptions: “We assumed the following trend rates in healthcare costs: an annual rate of increase in the cost of medication for retirees under age 65 of 7% for 2009 with a gradual decline to 4.5% over 10 years, and 4.5% for retirees over age 65.”
Where did they get 7% from? Where did they get a “gradual decline over 10 years” bottoming out at 4.5% from? Where did they get that same 4.5% figure for retirees over 65 in perpetuity?
Is this because an Excel sheet was developed with those inflation figures already populated in those cells? Does that not sound a little too simplistic? If these numbers are nowhere near accurate, then look what BCE themselves said in the same Note 23 to their financials: “Assumed trend rates in healthcare costs have a significant effect on the amounts reported for the healthcare plans. The following table, for example, shows the effect of a 1% change in the assumed trend rates in healthcare costs.”
That table goes on to show that if they overstated the trend rate by 1%, it would decrease their accrued obligations by $104 million!
Really? $104 million and this whole calculation gets left to a sliding scale and some numbers popped out of the sky that are defensible based on many years of what I assume are actuarial best practices. I’m sure nobody ever got fired for using whatever conservative traditions led to the sliding scale approach, but if I’m a senior HR executive, CFO or CEO, I’m not sure I’m too happy with taking the easy way out and possibly leaving hundreds of millions on the table.
Has from BCE looked at their plan data—transaction by transaction—to consider how the following factors will impact their post-employment drug plan costs (by far the biggest component of the total healthcare obligation):
• The current saturation rates of age-related chronic conditions within both the under-65 and over-65 retiree populations? If a given population is fully saturated, will that not materially impact the trend rate and future costs?
• The current saturation rate of spending on expensive specialty (i.e. biological) therapies with the under-65 and over-65 retiree populations? If a population is fully saturated here and there are no markers in the experience for significant increases in specialty spending, won’t that materially impact the trend rate?
• The impact of coordination of benefits with public plans for plan members under-65 who will become eligible for public coverage: the timing of that eligibility, the current therapeutic mix and the geographical mix of employee population. Has anyone checked to see how many more drugs the Ontario Drug Benefit program (for example) covers now compared to three years ago? Is that factored into trend rate calculations?
• The impact of drug pricing legislation on specific populations. Is that not impacted uniquely by specific provincial mix of the current population, legislative impact and the current therapeutic mix?
• The impact of the current and generic substitutes of key brand name pharmaceuticals such as Lipitor, Crestor, Plavix, Diovan, Atacand, Nexium, etc.? Does that depend on the current age and therapeutic mix of the population and saturation rates of current utilization?
• The impact of the current plan designs and tweaks that can make it more cost-effective without at all impacted coverage.
These factors are completely unique to each and every plan sponsor and do not lend themselves to a simple Excel formula. They are items that not just our example plan sponsor should be looking it to, but every plan sponsor.
Who knows, maybe it turns out that the 7% number, sliding scale to 4.5% over 10 years and the current 4.5% number make sense for BCE, and can be fully justified based on a closer look at the plan’s own experience and current/future post-retirement populations. However, what if they have overstated that trend rate by 1% and can justify in tremendous detail why that trend rate is actually lower and they don’t need to make $104 million worth of provisions? I can think of a few key people at the organization who might like to know that.