Discount Rates: Discounting dilemmas
As risk-free rates edge towards zero or below in many regions, questions must be asked about discounting methods, according to Carlo Svaluto Moreolo
At a glance
• Negative yields on fixed income have implications for liability-discounting methods.
• Although technically possible, discounting at negative rates seems illogical.
• There is disagreement on whether discount rates should stop at zero or match lower-yielding assets.
• Experts say better discounting practices should reflect economic factors.
More than $10trn (€9.1trn) worth of negative-yielding sovereign debt is weighing heavily on the world’s financial system, according to Fitch Ratings. When rates are negative, pension funds and insurers stumble upon a difficult question: is it possible to discount liabilities at a negative rate?
Because sovereign and corporate debt are widely used as reference assets in liability-discounting models, falling yields on these assets means growing liabilities. A negative discount rate means that present value of a future liability is higher today than at the future date when that liability will have to be paid. The notion does not seem logical, much as it seems unreasonable to pay for the privilege of holding debt.
Thomas Hagemann, chief actuary at Mercer in Germany, says the issue is being widely discussed in the country. In fact, using models based on the International Accounting Standard Nineteen (IAS 19), the discount rate is low – about 1.5%, depending on how IAS 19 is interpreted – but not negative. IAS 19 is the prevalent set of accounting rules used to calculate employee benefits around the world. In IAS 19, discount rates are based upon high-quality corporate bonds.
But although the headline rate is positive, it does not mean actuaries do not face difficult decisions. Hagemann explains: “The fact the rate is positive does not represent the full truth. IAS 19 says you have to construct a yield curve and calculate discount rates for each duration. The resulting average rate is 1.5%, but that now includes negative discount rates too, for very short durations of up to two years. Even high-quality corporate bonds with very short duration have negative yield.”
This forces actuaries to decide whether for those short durations, where yields are negative, the discount rate should be zero, or reflect the actual yield. “Some believe that a discount rate cannot be negative. The IAS board did not take into account negative rates when they were writing IAS 19,” says Hagemann. “But I am not fully convinced that discounting at 0% instead of at the negative yields on corresponding durations is the right approach.”
The argument supporting his view, Hagemann says, is that it is difficult to find risk-free or low-risk assets that do not offer a negative yield. Negative rates are a reality, and they should be taken into account as such. However, he acknowledges that, from a technical point of view, the models do not need to be amended when discount rates are negative.
The psychological barrier to using negative discount rates seems high. But the consensus seems that the problem of how discounting models are affected has not fully taken hold in the actuarial world. Peter Meier, head of the Centre for Asset Management at Zurich University of Applied Sciences (ZHAW), says negative rates can be looked at as a cyclical phenomenon so far.
“The picture would become truly distressing if markets started to truly price long-term negative rates,” says Meier. At the moment, he points out, equity markets reflect a positive weighted average cost of capital (WACC), which implies that long-term rates must be positive. Meier adds: “Whatever the type of cashflow we are discounting, if rates are negative we should not limit ourselves to using market rates, but add risk premia instead.” In equity valuation, for example, this may consist of adding a premium linked to the dividend level (the higher and riskier the dividend payout, the lower the discount rate).
The question inevitably turns to what level of interest rates we can reasonably expect to see in the long term. Meier says there is a low probability of seeing negative rates five years from now. Others are pondering whether they have reached a floor yet, and whether a floor to interest rates even exists.
Facts show that the existence of an interest rate lower bound is questionable, according to David Lloyd, head of institutional portfolio management for fixed income at M&G Investments. Lloyd argues: “Our understanding has been completely re-written by the events. However, there has recently been a real awakening of the debate as to whether such low interest rates are actually effective. The traditional wisdom is that lower rates would encourage borrowing, consumption and investment, and curb saving.”
“But the overall effectiveness of the current interest rate structure is open to question. The burden of proof is challenging the view that we need lower rates for longer,” adds Lloyd. However, it is not clear how the scenario might change. Lloyd contends that one unintended consequence of lower rates is that many people are compelled to save more for their retirement. This would support the view that yet lower rates may be of limited effectiveness.
Mitul Patel, head of rates at Henderson Global Investors, says the existence of a lower bound to interest rates is a “strong assumption to make”. According to Patel: “Years of falling yields have made it foolhardy to call the lows. However, if a lower bound exists on overnight rates, as central banks now seem to be suggesting, then yields should also be floored by the same bounds.”
It is important to note that in the two European countries where interest rates are most negative – Switzerland and Germany – regulation gives a certain degree of freedom in terms of discounting and funding. In Switzerland, pension funds board decide their own discount rate and discuss recovery plans in cases of underfunding. Furthermore, the rate at which savings are converted into an annuity is politically determined.
In Germany, corporates are under no obligations to fund so-called Direktzusage or direct promise defined benefit liabilities, according to Hagemann. It is not surprising, as he points out, that the German Institute of Public Auditors (Institut der Wirtschaftsprüfer in Deutschland) and the Accounting Standards Committee of Germany (Deutsches Rechnungslegungs Standards Committee) have not officially addressed the question of how negative discount rates should be treated.
Yet IAS 19 rules apply worldwide and dictate not just liability valuations but, ultimately, pension fund’s liability-driven investment (LDI) strategy. Whatever the liability discount rate, pension funds still need to make strategic decisions on how to match those liabilities. Salman Ahmed, chief investment strategist at Lombard Odier Investment Management (LOIM), says pension funds should not fall into the trap of overweighting sovereign fixed income within their matching portfolios.
Ahmed argues: “Due to central bank action, the level of interest rates is not market determined – therefore, sovereign rates are not the right indicator of how you should swap liabilities across time. I think pension funds should focus on corporate bonds to match their liabilities.
“This is for a pure economic reason,” he adds. “If a pension fund makes a promise to stakeholders, it cannot pretend that the promise is as strong as government guarantee. Pension funds cannot print money. A promise a pension fund makes to stakeholders is no different from a promise made by a corporate to repay an investor. Pension funds’ strategies should reflect that dynamic when they think about discounting and matching.”
Patel brings back the question of whether interest rates have a lower bound. He says if the market accepts the existence of a floor, and regulations are adapted to reflect one, pension funds should reduce their fixed-income holdings as yields approach the floor, rather than increasing them. “At the extreme, no hedging of liabilities would be required at all if yields can only go higher”, he notes.