Investor Errors and Financial Illiteracy
The literature is rich with examples of how behavioral biases affect
investment decision-making. The behavioral biases include:
( 1) loss aversion, first noted by Kahneman and Tversky (1979)—
or the tendency to gamble losses; ( 2) endowment and disposition
effect—or the tendency to assign a greater value to securities one
holds and not sell losers; and ( 3) short-termism—the tendency to
make decisions based on recent history (TIAA-CREF 2015).
Adding to the challenge, a recent survey of global financial literacy
(Klapper et al. 2015) suggests that close to 40 percent of U.S. adults
are not financially literate (defined as being able to correctly answer
three out of four sets of questions on risk diversification, simple
interest, compound interest, and understanding the impact of
inflation). These basic calculations are paramount to being able to
correctly answer the many questions needed to ensure retirement
safety, even if we ignore the issues with finance theory raised
above. The greater the number of decisions that individuals are
uncertain about, the greater the risk that they will retire poor and
have to be bailed out by government.
Dealing with Multiple Entities
One attractive feature of a DB fund is that accumulation and decumulation are made within the same entity, with little to no input
from the participant; but this is not the case in typical DC plans
because investors need to accumulate assets with the DC plan provider and then purchase an annuity from an insurance provider
(R. Merton, pers. comm. 2016). This poses a significant challenge
to the typical participant (and particularly so for those who are less
14 Recently some pension plan sponsors are trying to
offer in-plan deferred annuities, but this still leaves the basic challenge of understanding annuities (Cornfield 2015).
Complexity/Costs of Annuities
Annuities are complex because the pricing is not transparent (
especially because of the mortality risk calculations), and these characteristics put them beyond the comprehension of basic investors.
For example, in examining why so few Americans buy annuities,
Brown et al. (2012) report that many are deterred by the complexity of the choice, and few have any experience with these instruments during their working lives, making annuities an alien instrument. Further, annuities typically are much more expensive than
market-based financial instruments, are illiquid or expensive to
exit, and engender credit risk (even though insurance companies
are regulated for solvency). Many investors are scared they will
leave money on the table if they die early. These products have surrender fees if the participant wishes to liquidate the annuity position. There is also some concern that seniors have been taken
advantage of disproportionately by mis-selling of annuities.
Finally, given the costs associated with these instruments, some
(Russell 2015) have advocated for the creation of nonprofit annuity
providers, specifically for the state plans that are being designed to
cater to uncovered workers (who tend to be at the lower end of the
Using MPT-based investment approaches on traditional assets leads
to risky (relative to a target retirement income), complex, error-prone, costly, and illiquid approaches, often with multiple entities,
which have the net impact of lowering retirement income as well as
making it highly uncertain. We will explain how the introduction of
a new market instrument, BFFS, can mitigate some of the risks and
help individuals to deal with the challenges. We also discuss one
approach being implemented in the Netherlands (and under consideration in the United States) to try to mitigate the risk.
How a Guaranteed Return Makes DC into DB
Modigliani and Muralidhar (2004) showed analytically that to
achieve a given target guaranteed replacement rate (i.e., ratio of
retirement income to earnings in working life), under certain conditions, simply requires a guaranteed return on all contributions
(fixed as a percentage of income). Modigliani and Muralidhar
(2004) developed this approach in an attempt to simplify the typical social security DB formula and to show that a DB plan was
nothing more than a DC plan with a guaranteed rate of return.
However, in social security (and employer) DB systems, a sponsor
bears residual risk and can potentially smooth investment outcomes over multiple generations (or change contributions and benefits). Some have argued mistakenly for such a guaranteed return
model for all DC plans (Ghilarducci 2009), stating that such guarantees with guaranteed rates above the current risk-free rate can be
purchased from market participants.
16 No financial entity will provide such a high guaranteed return (without charging an exorbitant
price) because it would create arbitrage possibilities. Given the
challenges of accessing such guarantees, rather than guaranteeing a
return over working life (and spanning the savings-retirement time
gap), BFFS focus on a more tractable problem. They are designed
to guarantee payment of a fixed level of income for a particular
time period. With such a cash-flow guarantee, the investor’s problem is reduced solely to how much to contribute.
Collective DC as a Way to Hedge
Retirement Income Risk
An alternative approach being considered in countries such as the
Netherlands and even in the United States, as well as by companies
such as Royal Dutch Shell, is a pension system called “collective
DC.” As pension systems are converting from DB to DC, collective
DC allows individuals to capture the benefit of pooling that is
embedded in DB plans by working collectively to ensure an inter-and intra-generational pooling of risk. In a way, these structures are
trying to overcome the challenge of dealing with multiple entities
and to spread investment risk over multiple lifetimes. But collective
DC structures also create inter- and intra-generational equity issues.
Typically, future generations are at a disadvantage because they are
not participants in decisions on the subsidy provided to the current
working or retirement generation, and hence must bear the risks
(and resulting costs) of decisions made by the current generation.
In effect, this is an attempt to create DB-like profiles within a DC
structure, but it starts to fall apart because of the mismatch in goals