Questioning the fundamentals
When I look at my personal DC plan, should I have to worry about the fact that there is no pension fund working behind the scenes to safeguard my pension rights and which, in order to do so, invests wisely?
Well, it depends. If I have amassed enough wealth at my target retirement age of 65 to buy a life-long income stream that is comparable to the income stream I would have received under a career-average, conditionally indexed, defined benefit scheme, I need not have worried. Should I have amassed even more wealth, I will be even better off and happy that I was in control of the investments and not a pension fund. But should I have amassed less wealth than would be necessary to buy a comparable income stream, or pension, I am in trouble, and I can only retire partially.
How partial my retirement can be depends on the level of interest rates, when I turn 65. If interest rates are high, I can buy more pension income than when interest rates are low. Another question is where to buy the income stream? Is there still a pension fund around that I can turn to or do I have to shop around between several commercial insurers? The first option is, all things being equal, less expensive than the second.
The question as to what happens at retirement age is often overlooked in the design and implementation of DC schemes but is vitally important. And I have just listed a number of questions that at the present time to which I simply do not have a definite answer. The one thing I do know is that more wealth is definitely better than less, or more precisely that more pension income is better than less pension income.
So the question becomes how I invest wisely, that is generate good returns without too much risk, and make sure that I have a policy in place that offers inflation protection and makes my portfolio as invulnerable as possible to interest rates towards the end of my planned investment horizon? This of course is more easily said than done.
These questions have been addressed in the academic literature, and luckily I know a little bit about investment theory and statistics so I understand most of what has been written about optimal funding and asset allocation for DC plans. So in theory my investment odds should be better than those of the average DC plan participant who doesn’t. Probably this is the key issue in discussing the importance of asset allocation for DC plans: the average participant in general has not had access to the investment concepts and literature that the professional investor or advisor has had access to, and therefore remains relatively clueless regarding rational and consistent investment decisions. That is, unless he is helped.
The current standard in individual DC plan offerings is the availability of so-called life cycle funds as a default investment option, which attempts to mimic the ‘optimal’ investment portfolio for the plan participant over the course of the accumulation period. In general this means that the participant moves from one generational cohort, or age bracket, to the next as he gets older. Each cohort has its own benchmark life cycle portfolio. These life cycle funds are geared more towards risky assets like equities and real estate when a cohort is younger, and more towards (inflation-linked) bonds and cash when a cohort is older. Some DC plan providers offer automatic or semi-automatic transitions from one life cycle fund to the other, so the participant that does not wish to structure his portfolio himself has at least a sensible default option that protects him from choosing too conservative, too aggressive or just silly portfolios.
The life cycle concept is not new, in fact a wealth of academic literature is now available on the subject, and the concept is accepted by theorists and practitioners alike. The academic debate one can have is the debate about the assumptions and method of analysis one uses to derive the optimal portfolios, which I will not discuss1. The practical issue that I would like to address is the inherent static nature of the life cycle funds and how this may lead to problems.
Life cycle funds are in many cases carefully constructed and based on asset liability analyses that result in a series of optimal portfolios that move with the age of the participant, and hence move with the remaining investment horizon. These life cycle funds try to represent the portfolios that on average would have worked well using the investment data history that is available.
However, long-term investment strategies are not only exposed to averages, but also to outliers, shocks, tail events or other euphemisms used to describe large negative or positive returns. So suppose that I am about to change from one cohort to the next and that my allocation to risky assets drops to 0% and the allocation to bonds goes to 100% because of my age and a planned date of retirement in five years.
I would be happy with this automatic transfer if I was confident that I had amassed enough wealth to retire in five years time. But what if we had just experienced a stock market crash? First of all my wealth would have dropped and I certainly would not want to move out of equities. I would rather invest more into equities to try and recapture my losses. Second, I would like to be able to extend my investment horizon, a euphemism for postponing retirement. However, my automatic life cycle rebalancing structure directs me to accept my losses and retire poorly in five years time.
This simple example shows that due to circumstances I may want to move away from a pre-defined and thus static asset allocation. In an individual DC world there probably is no way out of this dilemma.
q When I am not a knowledgeable investor I am probably better served with this life cycle product than when left to my own devices. I have at my disposal an asset allocation rule that works well on average, but after the asset allocation rule has been set, my performance is completely determined by the market movements. Blake Cairns, and Dowd (2001)2 have shown that pension ratios (DC to DB) for life-style allocations can vary considerably from 0.5-5, which means that DC pensions are inherently more risky than DB pensions, and that after a static or deterministic asset allocation has been chosen, the outcome is determined by luck;
q When I am a knowledgeable investor I would like to have as much flexibility in setting my asset allocation in response to changing market circumstances, and probably do not need a life cycle default option.
A collective DC world might offer a solution to this dilemma, at least within the Dutch pension fund industry. Recently Teulings and de Vries (2003)3 introduced the idea of collective DC schemes with generational accounts, where there are different asset allocations for different generations of plan participants which vary with the average age and investment horizon of the different generations. Contributions would also differ for the separate generations.
This idea introduces intra-generational solidarity as an alternative to inter-generational solidarity in DB plans. Each generation then has the flexibility to adjust the asset allocation and the level of contributions to changing market circumstances. For example, if a generation is hit by a stock market crash, then the loss can be distributed smoothly over lifetime consumption, contributions should be raised, the allocation to equities can be increased, and retirement possibly postponed.
Off course, the accounts are managed by a professional organisation so that the flexibility of dynamic asset allocation is combined with the comfort for all participants of not having to worry about their investment portfolio. This may be a way forward and a smart compromise in the heated debate about the choice between DB and individual DC. But the real test of this new concept obviously comes when the first pioneers begin to implement it.
And this requires the transformation of an academic concept to a manageable product, which may take a while, but in my opinion it could very well become a success story.
1One could argue that the life cycle concept in itself is debatable, but this holds true for any asset allocation strategy.
2David Blake, Andrew J.G. Cairns, and Kevin Dowd, “Pensionmetrics: stochastic pension plan design and value-at-risk during the accumulation phase”, Insurance: Mathematics and Economics 29 (2001) 187-215.
3Coen N. Teulings and Casper G. de Vries, “Generational Accounting, Solidarity and Pension Losses”, IZA Discussion Paper No. 961, (December 2003).
This article represents the personal views of the author and are not necessarily those of Towers Perrin