Balancing risks and rewards

Value-based asset liability modelling and generational accounting can reveal the hidden value transfers between generations. David White reports

How can the sponsors and boards of pension funds devise fairer occupational pension plans, where the risks and rewards of a pension plan are more equitably distributed between young and old?

This question is particularly relevant in European countries, where employers are moving from traditional defined benefit (DB) pension plans to collective risk sharing schemes.

The net effect is that there is increasing pressure to define more explicit and transparent pension contracts. In particular there is a demand to expose the hidden value transfers between generations, says Roy Hoevenaars, a senior portfolio manager with the Global Tactical Asset Allocation fund at APG Investments, the asset manager of pension fund ABP, in Amsterdam.

Hoevenaars is the author of a recent major study ‘Strategic Asset Allocation and Asset Liability Management’ which considers the role of asset liability modelling (ALM) in the strategic asset allocation decisions of pension funds. The study is a PhD dissertation presented at the faculty of economics and business administration at Maastricht University.

One of the central planks of the study is the importance of understanding the relationship between the various stakeholders in a pension fund by identifying the ‘value transfers’ embedded in the pension deal. In other words, revealing which group in a pension fund benefits most and least in terms of pension risks and rewards.

Hoevenaars’ study draws together three major branches of investment and finance research and extends their application to pension funds.

The first branch is generational accounting, a tool developed by public finance economists to investigate the intergenerational distributional effects of fiscal policy.

However, conventional generational accounting has its limitations; in particular, in evaluating risk, says Hoevenaars. “The problem there was that they didn’t really know how to deal with risks - they couldn’t value them.”

So Hoevenaars has extended the principles of generational accounting to pension funds. A generational account in a pension fund is defined as the difference between the benefits to be received and the contributions to be paid by a specific age group.

The second branch of research is the recognition that a pension contract is effectively a series of ‘embedded options’ on the balance sheet. This perception dates back to work by William Sharpe, of Sharpe ratio fame, who was the first to describe the pension contract as a put option.

Generational accounts can be valued as embedded options, using the same option valuation methodology as is used in financial markets. However, a pension contract differs from a normal contract in a financial market, Hoevenaars points out.

“If you were to buy an equity option the price you would pay would be in line with the market risk. You would pay a simple, single option premium. But with a pension contract, the benefit and the risk are really defined by the terms of the contract.”

As a first step Hoevenaars identifies embedded generational options on the balance sheet as the uncertain cash flows to and from the participating groups in a pension fund, in particular contributions and benefits. He then values these options using option valuation techniques

Hoevenaars in effect rewrites the balance sheet of a pension fund as embedded generational options to be able to explore intergenerational risk sharing: “Understanding the embedded options written by one group of stakeholders to another group is important for a fair and sustainable pension contract,” he says.

The third branch of research is classic asset liability modelling. Hoevenaars points out that the output from a classic asset liability model cannot be used to value the various embedded options, since all the output variables are evaluated in terms of probability distributions rather than risk transfers.

He argues that only a value-based ALM will reveal the hidden value transfers between generations. Value-based ALM uses the same output from scenario analysis - that is, possible future outcomes - that classic ALM uses. However, with value-based ALM the future outcomes are discounted back to the present with an appropriate risk-adjusted discount rate.

Hoevenaars maintains that the value-based ALM, and in particular value-based intergenerational accounting, should be used as an extension of classical ALM to gain extra insights into the relationships between the stakeholders in a pension fund.

Using value-based generational accounting, he shows that changes in the pension contract can lead to sizeable intergenerational value transfers, as the allocation of risk among stakeholder’s changes substantially.

His analysis shows that any policy change due to changes in the investment, contribution rate or indexation policy, or from pension reforms, will inevitably lead to value transfers.

For example, a switch to less risky asset mix is beneficial to older pension plan members at the expense of the younger members who lose value. Conversely, a more risky asset mix is beneficial to younger members at the expense of older members.

Hoevenaars concludes that policy changes in collective pension schemes will inevitably lead to value transfers between generations. In particular the move from a traditional DB plan to a hybrid plan will lead to a considerable redistribution of value from older to younger members.

 

The study breaks new ground by applying the methodology of embedded generational options to real, existing pension funds with intergenerational risk sharing.

His value-based ALM framework is based on a simulation study which projects the development of the pension fund in a number of future scenarios, with a policy horizon of 20 years. To keep the model simple he limits the investment universe to the MSCI world index and nominal bonds with a maturity of 10 years

Hoevenaars considers four variants of pension plan design (see table). The first variant is a ‘no risk allocation’ where current members are not expected to absorb mismatch risk - the gap between assets and liabilities. This risk is implicitly is shifted forward to future members.

The second is a traditional DB plan where the mismatch risk is absorbed by adjustments to the contribution rate. This results in a volatile contribution rate

The third and fourth variants are, respectively, a hybrid DB/DC plan and collective DC plan. Both absorb mismatch risk by adjusting indexation. However, the hybrid plan has a cap and a floor on indexation, while the collective DC plan imposes no restrictions on indexation. As a result, the collective DC variant has more volatility than the hybrid.

To consider what is happening in value terms, Hoevenaars measures ‘pension fund residue’ - that is, the difference between assets and nominal liabilities. He divides the residue option into a surplus option and a deficit option. The surplus option is a European call option on the surplus for participants written by the pension plan. The deficit option is the economic value of a European put option written by the pension plan’s members

The results show that changes in pension plan design lead to value transfers between current and future generations. The first variant ‘no risk allocation’ (NoRM) creates the largest imbalance between value transfers from current to future generations, with the lowest residue option value of -36%, the product of a deficit option value of -44% and a surplus option value of +8%.

This reflects the fact that current generations are not absorbing the mismatch risk by adjustments to contributions and benefits but are transferring this downside risk to future generations (-44%) without a similar transfer of upward potential (+8%).

In contrast, the traditional DB plan’s use of a flexible contribution rate reduces the imbalance in value transfers, producing a residue option value of only -12%

The collective DC scheme offers the lowest residue option value of
-4%. However this achieved with a surplus option value of 97% and a deficit option value of -101%. “There can be a huge range between the deficit option and the surplus option but the residue option may be very small,” Hoevenaars observes.

“The very large deficit option of
-101% for the collective DC - much higher than the other systems - indicates that there are large downside risks for participants because there could be a cut in pension rights during downturns in the global economy.”

Value-based ALM and generational accounting can be applied at two levels, he says. “One is the broad level of pension fund design where one can compare broad categories of pension systems. It can also be applied on the level of a single pension system.

“We could, for example, consider the intergenerational transfers that would result from two or three different changes in the indexation policy.

“The transfers will be smaller than if you changed the whole system, but they would provide insights to enable the pension’s board to evaluate the impact of three or four different indexation policies, and to decide which was the more acceptable.”

As pension funds move from fixed to conditional indexation, this tool will become increasingly useful, Hoevenaars says. “The valuation of embedded options is a useful instrument for good pension governance, to enable policymakers to evaluate pension reforms and other changes in the pension contract in the coming years.”