8 / 50
8
the existence of potential self-control problems through
the lifecycle, may be able to implement commitment
devices to counteract personal biases and prevent a
premature spending of personal savings, while a naïve
individual would not (Laibson, 1997, Angeletos et al,
2001). Examples of such commitment mechanisms in
financial products are the classic use of illiquid housing
equity as a vehicle for saving, and the partial illiquidity of
retirement savings accounts which imposes tax penalties
to early withdrawals. Lastly, it has been empirically
documented that self-awareness of potential biases
has a positive effect on retirement savings even after
controlling for measures of IQ, financial literacy and
socio-demographic characteristics (Goda et al, 2015).
The relationship between present bias and retirement
savings, driven by the documented existence of
heterogeneous sophisticated and naïve agents, makes
financial regulators and service providers interested
in understanding the benefits of such commitment
technologies in financial products. It became critical
to understand how salient they should be in order
to correctly inform policy and effectively generate
behavioural changes. This question was experimentally
addressed recently by Beshears et al (2015) using a
representative sample of the U.S. adult population. The
authors pointed out that: (i) conventional economic
theory predicts that nothing should be contributed to a
commitment account when it offers the same expected
return as a fully-liquid account. (ii) higher penalties
may reduce premature withdrawals, but they may also
discourage deposits, defeating the goal of raising net
savings; on the other hand, (iii) if savers recognize that
penalties help them overcome self-control problems,
they may welcome higher penalties and make more
deposits in response. In sum, they test whether the
demand for commitment savings accounts is affected by
how illiquid the offered savings accounts are.
Commitment devices have been shown as strongly
appealing. When participants were asked to allocate
money between a liquid account and a commitment
account that randomly varies across participants in
terms of interest rates, prohibitions and penalties for
withdraws prior to a commitment date, they presented
a consistent demand for commitment technologies.
Participants allocated around half of their endowments
to the commitment account when there was no
difference in interest rates between the two vehicles,
and one-quarter of their money even when the interest
rate paid by the commitment account was lower than the
liquid account. These results build on previous evidence
from field experiments on the role of commitment
savings accounts performed in different countries (see
Beshears et al, 2015, p.3). Ultimately, this suggests the
presence of sophisticated present-biased individuals in
the U.S. adult population.
An anxious reader might think that policymakers
should set high withdraw tax penalties on pension
plans because this act would benefit all agents and the
aggregate national level of savings. Although, we need
to consider that sophisticated players are only part
of the market, if not the minority, hence generating
an ambiguous effect of higher withdraw penalties
on total contributions to these accounts that is not
irrespective of how heterogeneous the population in
question is. First of all, financial products like private
and workplace pension plans, given their risk profile and
asset allocations, have expected returns that are higher
than vehicles with immediate liquidity, such as current
account deposits. Beshears et al (2015) show that when
the commitment account presented in their framework
paid an interest rate higher than the liquid account,
as is the real case of pension schemes, the empirical
relationship between illiquidity levels and deposits in
the commitment account was insignificant, suggesting
that the U.S. adult population contains also naïve
present biased individuals and/or individual who have
consistent time-preferences, i.e., without present bias.
The interaction of these two other groups is described in
Table 1. Both make positive commitment deposits that
diminish as the commitment account’s illiquidity rises.
This ends up offsetting the increase in commitment
deposits by the group of sophisticated present-biased
individuals.
Table 1. Relationship between illiquidity levels and de-
posits in commitment accounts
For further references of the lessons learned and to
be learned about the interaction between present
bias, self-control and commitment devices, we suggest
the recent review of O’Donoghue and Rabin (2015).
The study proposes a set of research questions, being
among them: how to assess the impact of present bias
against other phenomena?; and how to assess whether
the demand for commitments are due to present bias?
4
private and workplace pension plans, given their risk profile and asset allocations, have
expected returns that are high r than vehicles with immediate l qui ity, such as current
account deposits. Beshears et al (2015) show that when the commitment account
presented in their framework paid an interest rate higher than the liquid account, as is
the real case of pension schemes, the empirical relationship between illiquidity levels
and deposits in the commitment account was insignificant, suggesting that the U.S.
adult population contains also naïve present biased individuals and/or individual who
have consistent time-preferences, i.e., without present bias. The interaction of these two
other groups is described in Table 1. Both make positive commitment deposits that
diminish as the commitment account’s illiquidity
rises. This ends up offsetting the
increase in commitment deposits by the group of sophisticated present-biased
individuals.
Table 1. Relationship between illiquidity levels and deposits in commitment
accounts
Group characteristic
Without present bias
With present bias
Sophisticated
Deposits decrease as
illiquidity rises because
this group has time-
consistent preferences.
No need for commitment
devices.
Deposits increase as
illiquidity rises because
this group is aware of
their bias and potential
self-control problems.
Naïve
Deposits decrease as illiquidity rises because naïve
agents are unaware of whether biases and potential
self-control problems.
Note: based on Beshears et al (2015).
For further references of the lessons learned and to be learned about the interaction
between present bias, self-control and commitm nt devices, we sugg st the recent
review of O’Donoghue and Rabin (2015). The study proposes a set of research
questions, being among them:
how to assess the impact of present bias against other
phenomena?
; and
how to assess whether the demand for commitments are due to
present bias
? In sum, the recent studies of Delaney and Lades (2015). Goda et al (2015)
and B sh ars et al (2015) provide good initial evaluations to address these questions to
better inform policymaking in the context of retirement savings.
2. Conclusions
Increasing the illiquidity pension schemes may not increase aggregate contributions
because only one segment of the population has the desire for strict commitment in
order to stick with a previous decided plan. This highlights the need for further
developments to deal with innocent unsophistication of the naïve group. Enhancing
financial literacy and the access to professional advice can be potential complementary
mechanisms.