Managing macro-longevity risk
An internal market for macro-longevity risk could provide an alternative to hedging tools
Over recent decades, life expectancy has increased dramatically. In highly developed countries, life expectancy increased by 2.5 years per decade. This has a substantial impact on the sustainability of pension systems. The IMF estimates that adding an extra year to the average lifespan, increases the pension bill by 4%.
A key challenge for pension provisioning is therefore the uncertainty about future life expectancy. This is called macro-longevity risk. Life expectancy may, for example, increase following medical improvements or decrease because of an unhealthier lifestyle. Understandably, pension funds search for ways to manage macro-longevity risk for the benefit of their participants.
Macro-longevity risk differs from micro-longevity risk or the individual uncertainty about the time of death. Micro-longevity risk is an idiosyncratic risk that fully diversifies by pooling enough participants in a pension scheme. Macro-longevity risk, by contrast, is a systematic risk that does not decrease by sharing it within a large pool of participants. Nonetheless, sharing macro-longevity risk across cohorts can be beneficial as the risk does not affect the entire population similarly.
Medical progress or diseases affect cohorts in a different way. Workers can, furthermore, adjust their labour supply in response to changes in longevity, for example by postponing their retirement date. The risk-bearing capacity of workers is larger when they use their labour supply as a hedge against macro-longevity risk. How can pension funds take advantage of these properties?
- Macro-longevity risk can be shared across age cohorts
- Government policy on retirement age can have an impact
- Benefits from sharing macro-longevity risk are large if retirement age is linked to life expectancy
- Macro-longevity risk-sharing between pension funds is an area for future development
Creating a pension fund ‘internal market’ for macro-longevity risk provides an alternative to market-based solutions. The market for macro-longevity risk is not liquid. There are several reasons for the lack of a well-functioning market. First, there are no institutions in the economy that are willing to underwrite this risk. The government, for example, is already exposed to longevity risk through social security payments.
Second, macro-longevity risk is rather specific to a pension fund’s population. This will create a basis risk compared with generic market solutions based on a country’s population.
A pension fund makes decisions on behalf of the participants and can allocate macro-longevity risk to cohorts. To allocate the risk optimally, the pension fund can maximise aggregate expected utility of all participants. Some cohorts absorb macro-longevity risk of other cohorts and receive a fair compensation for this risk transfer. This transaction via the internal market generates welfare benefits.
The intuition behind these benefits is as follows. Take two cohorts that are both exposed to macro-longevity risk but to a different extent. The exposure of cohort A is larger compared with the exposure of cohort B. In the absence of risk sharing, both cohorts absorb a shock in macro-longevity themselves. If the two cohorts pool their risk exposures, however, both cohorts absorb half of the aggregate shock. Furthermore, if cohort A pays a fair compensation to cohort B, the risk-sharing transaction can be attractive for both cohorts from a risk-return perspective. Broeders, Mehlkopf and van Ool (2018) derive the optimal risk-sharing rule such that all cohorts benefit and experience the same welfare gain1.
The example already shows that cohorts can benefit from risk-sharing. These benefits enhance if cohorts adjust their work supply – for example, through advancing or adjourning their retirement. In fact, more and more countries link the retirement age to life expectancy developments. In the UK, for example, the government plans to link the state pension age to the projected longevity of the population in such a way that people receive a state pension during a fixed proportion of adult life. Under this policy both the working and retirement period expand if life expectancy increases.
The Netherlands links the retirement age to the remaining life expectancy of the population at age 65. Under this policy the length of the retirement period is fixed, while the working period increases if life expectancy increases.
To derive the impact of the retirement-age policy on macro-longevity risk sharing, we consider three policies. First, a fixed retirement age: the retirement age does not change after macro-longevity shocks. In this case the working period is constant.
Second, a partial adjustment of the retirement age: the retirement age automatically adjusts to life expectancy developments in a such a way that retirement consumption remains constant. This policy resembles the proposed retirement age policy in the UK.
Third, a full adjustment of the retirement age: the retirement age fully keeps up with life expectancy changes. This means, for example, that if the remaining life expectancy at retirement increases by one month, the retirement age also increases by one month. Under this policy, which is close to the policy in the Netherlands, the length of the retirement period is constant.
To quantify the benefits of risk-sharing we use the concept of welfare benefits. A welfare benefit of 1%, for instance, implies that cohorts find the welfare benefits from risk-sharing equivalent to a permanent consumption increase of 1%. In our calculations, the uncertainty with regard to future mortality rates is based on the well-known Lee-Carter model for macro-longevity risk. We use mortality data of Dutch and UK females and males and consider sharing macro-longevity shocks on a 10-year horizon.
The figure presents the welfare benefits for the three retirement age policies. We observe that for each retirement-age policy collective risk-sharing of macro-longevity risk is welfare improving compared with no risk-sharing. In case of a fixed retirement age, the welfare benefit equals 0.3% for Dutch females. Under this policy a longevity shock impacts the retirement consumption of different cohorts more or less equally. Risk-sharing in this case does not create many additional benefits. If the retirement age is partially adjusted, the welfare benefit is higher. This is a result of the fact that retirement consumption of workers is not affected by macro-longevity shocks.
In case of a full adjustment of the retirement age the aggregate welfare benefit increases significantly. This follows from the large risk-bearing capacity of workers. They adjust their labour supply to hedge against macro-longevity shocks. This increases the workers’ risk appetite to provide insurance to retirees. The figure also shows that the welfare benefits are lower for UK mortality data. This is a result of a smaller macro-longevity risk.
Pension funds face macro-longevity risk, or uncertainty about future mortality rates. As market-based solutions for managing macro-longevity risk are limited, risk-sharing between cohorts in a pension scheme is an interesting risk-management tool.
The design of the retirement age policy has a large impact on both the optimal risk-sharing rule and welfare benefits. If the retirement age is linked to life expectancy, welfare benefits from sharing macro-longevity risk are large and significant.
The risk-bearing capacity of workers is larger, because they can use their labour supply as a hedge against macro-longevity shocks. As a result, workers absorb risk from retirees in the optimal risk-transfer rule because the human capital of workers increases if they work longer.
These findings are relevant for pension policy, especially because of the general trend of transferring risks to pension participants. An interesting area for future development is macro-longevity risk-sharing between pension funds. If two pension funds are heterogeneous in terms of the exposure to macro-longevity risk, they will also benefit from risk-sharing.
1 See Broeders, D, Mehlkopf, R and van Ool, A (2018) The economics of sharing macro-longevity risk, https://ssrn.com/abstract=3311420
Dirk Broeders is a senior strategy adviser at De Nederlandsche Bank and professor at the School of Business and Economics of Maastricht University; Roel Mehlkopf is a pension fund adviser at Cardano Risk Management and is affiliated with Tilburg University; Annick van Ool is policy adviser at De Nederlandsche Bank and a PhD candidate at Maastricht University