A small step today can make a big difference in retirement. “People who save regularly for retirement tend to be happier with their retirement planning than those who do not,” says Kathrin Nies, author of 'Why Saving on a Regular Basis May be Wise', a study conducted by the Munich Center for the Economics of Aging and Allianz Global Investors.
3. Masthead
Publisher
Allianz SE
Koeniginstrasse 28
80802 Munich, Germany
Phone: +49 89 3800-0
Fax: +49 89 3800-3425
www.allianz.com
Author
Dr. Kathrin Nies, Economist Contents
International Pensions
kathrin.nies@allianzam.com
+49 89 12 20 70 03 Executive Summary: Retirement Planning Seems
to be Worth the Effort
Editor
Marilee Williams 04 Retirement Planning Results in Greater Wealth
Layout 05 Data and Descriptive Results
volk:art51 GmbH, Munich
06 Retirement-planning satisfaction according
Closing Date to socio-economic characteristics
20.03.2012
09 Retirement-planning satisfaction according
to savings-related characteristics
These assessments are, as always, subject to the 12 Marginal Impacts on Retirement Planning Satisfaction
disclaimer provided below.
15 Conclusion
Cautionary Note Regarding Forward-Looking Statements
The statements contained herein may include statements 16 References
of future expectations and other forward-looking statements that
are based on management’s current views and assumptions and
involve known and unknown risks and uncertainties that could
17 Recent Publications
cause actual results, performance or events to differ materially
from those expressed or implied in such statements. In addition
to statements which are forward-looking by reason of context,
the words “may”, “will”, “should”, “expects”, “plans”, “intends”,
“anticipates”, “believes”, “estimates”, “predicts”, “potential”,
or “continue” and similar expressions identify forward-looking
statements. Actual results, performance or events may differ
materially from those in such statements due to, without
limitation, (i) general economic conditions, including in particular
economic conditions in the Allianz Group’s core business and core
markets, (ii) performance of financial markets, including emerging
markets, and including market volatility, liquidity and credit events
(iii) the frequency and severity of insured loss events, including
from natural catastrophes and including the development of loss
expenses, (iv) mortality and morbidity levels and trends, (v)
persistency levels, (vi) the extent of credit defaults, (vii) interest
rate levels, (viii) currency exchange rates including the Euro/U.S.
Dollar exchange rate, (ix) changing levels of competition, (x)
changes in laws and regulations, including monetary convergence
and the European Monetary Union, (xi) changes in the policies
of central banks and / or foreign governments, (xii) the impact
of acquisitions, including related integration issues, (xiii)
reorganization measures, and (xiv) general competitive factors,
in each case on a local, regional, national and / or global basis.
Many of these factors may be more likely to occur, or more pro-
nounced, as a result of terrorist activities and their consequences.
The company assumes no obligation to update any forward-
looking statement.
No duty to update
The company assumes no obligation to update any information
contained herein.
2
4. Allianz International Pension Papers No. 2|2012
Executive Summary:
Retirement Planning Seems
to be Worth the Effort
The first baby boomers are now retiring in the United States. Most of them 1 Allianz Global Investors, 2011: Demography
will not be able to live on social security alone. Average earners in the United Series 1-3
States receive about half of their annual lifetime income as a public pension.
Therefore, they are heavily dependent on their private retirement savings.
Due to demographic changes, future generations are likely to face even lower
retirement income from public pensions and will therefore depend more on
private retirement income.1 Consequently, they can now learn from the saving
experiences of the retiring baby-boomer generation.
Previous research has found that people who do some kind of retirement plan-
ning accumulate more wealth than those who do not. Furthermore, planners
who make the effort of calculating their future needs and resulting required
savings rate, and planners who use rules of thumb statistically end up with the
same amount of wealth. Looking at a dataset from retired affluent Americans
who are between 50 and 69 years old, we investigate which type of saver, regular
and/or irregular, is most satisfied with his retirement planning. Our regression
analysis suggests that people who save regularly for retirement are significantly
happier with their retirement planning than those who do not, else equal. A word
of caution: We do not look at people older than 69 who have been living on their
savings a little bit longer and may have a different view.
3
5. Retirement Planning Results
in Greater Wealth
2 Allianz Global Investors, 2011: Putting the When asked in hindsight, many retirees believe they made mistakes in their
Retirement Pieces together: Strategies of the retirement planning. The most often-quoted reason for this is that they started
affluent 50+ Generation in the United States,
International Pension Papers 01/2011 to save too late and invested too riskily.2 The natural follow-up question is:
How can this be avoided? Annamaria Lusardi and Olivia S. Mitchell analyzed the
relationship between retirement planning and total net worth in the United States
and found that households reporting to have done some retirement planning are
already much better off in terms of their net worth than equivalent households
reporting to have done hardly any planning. Furthermore, they found that political
and financial literacy positively affects the probability of planning for retirement.
Based on US citizens, Johannes Binswanger and Katherine G. Carman extended
this analysis. They distinguished three retirement planning types:
• An actual planner: one who tries to calculate how much is needed
in retirement, possibly based on a few future scenarios;
• A rule-of-thumb planner: one who regularly saves an absolute amount
or a percentage of income; and
• Someone who does not have a savings rule.
Of course, the first type of planner requires much more time and effort for his
planning than the second; the last needs the least amount of time and effort.
In their analysis, Binswanger and Carman succeeded in not only showing that the
first two types, the actual and rule-of-thumb planners, accumulate significantly
more retirement wealth than the third type without a savings rule. In addition,
there seems to be little difference between the retirement wealth of an actual
planner and a rule-of-thumb planner. This implies that saving for retirement does
not have to be so expensive (i.e., actual planning); instead, using rules of thumb
statistically lead to the same outcomes.
It is unclear, however, whether a person A who accumulates, say, 100 units of
retirement wealth has saved ‘better’ than a person B with 80 units of retirement
wealth. Person A may have invested less in children’s education, which may
not be optimal. Maximizing retirement wealth therefore does not necessarily
maximize happiness. A complement to the retirement wealth analysis is, then,
the analysis of satisfaction with retirement planning. It gives us a hint at whether
people were able to spread their savings over their life course and are content
with their retirement wealth now. In line with the life-cycle theory and assuming
that people are forward-looking rational agents, it implies that a person manages
to smooth his consumption and savings path over the life course, which makes
him happy. In practice, people are often not forward-looking and may not realize
that they may outlive their assets if they live “too” long. Bearing this fallacy in
mind, we investigate which factors increase retirement-planning satisfaction,
using a relatively young sample of retirees.
4
6. Allianz International Pension Papers No. 2|2012
Data and Descriptive Results
On average, retirees are more satisfied with their retirement planning than people
who are still working or who are part-time retired, 1.59 versus 1.93 and 1.87,
respectively. Workers’ and part-time retirees’ satisfaction with retirement planning
is statistically indistinguishable. There are two immediate reasons for the higher
satisfaction of retirees versus non-retirees, the first being uncertainty. These non-
retirees are not sure whether their accumulated savings will last them through
retirement in the manner they are expecting. As people are typically risk averse,
they may perceive their situation to be worse under uncertainty than under cer-
tainty. Reason number two may be self-selection. Only those people who achieve
a certain planning satisfaction will decide to retire; otherwise, they continue
working (and saving) until they believe they have enough. Our aim is to find the
factors that help to increase retirement-planning satisfaction in order to give
today’s workers some practical guidance about what to do today to be satisfied
during retirement tomorrow. We therefore focus on the retirees in our sample
who can look back at their savings rules and judge their level of satisfaction with
their retirement planning up to their current age.
5
7. Retirement-planning satisfaction
according to socio-economic
characteristics
Table 1 describes this subsample of 580 retirees according to their level of
education and accumulated wealth. It gives us a first indication of which factors
positively affect retirement-planning satisfaction. Satisfaction is measured on
a scale from 1 (very satisfied) to 5 (very dissatisfied).
The better people are educated, the larger their human capital because they can
take more challenging and responsible jobs in the labor market. Consequently,
they have a higher earnings potential. Among the retirees who are beyond their
earnings phase, it gives us a hint of how much they have earned while in the labor
force. The standard life-cycle theory tells us that we should save relative to our
income. People earning less should save less than people who earn more. This is
because the higher earners also consume more and will still do so in retirement.
Especially for this affluent group of people, where everyone should have no
difficulties satisfying basic needs, this should hold true. Consequently, if we treat
education as a proxy for income, there is no reason why retirement-planning
satisfaction should vary across education categories. However, if looking at
education for what it is – an indicator for how much we have learned – we can
expect more-educated individuals to make smarter choices. Whether smarter
choices also result in higher satisfaction with retirement planning is an empirical
Box 1: Description of the dataset
We use a US dataset of 1,506 affluent (more than $ 250,000 in investable assets) individuals who are older than 50.
The survey was conducted in 2010 by Allianz Global Investors AG and TNS Infratest. We know the following socio-
economic characteristics: age, education, employment status (working, retired, part-time retired) and net worth.
Age is recorded in categories 1 (50 – 54) to 4 (65 – 69). Originally, there were 10 education categories, but we recoded
them into categories of the International Standard of Classification for Education (ISCED). Net worth is asked on
a household level and also recorded in categories. We do not know their health status. Furthermore, people indicate
whether they save regularly, irregularly or do not have a savings rule at all, and whether they believe they made errors
in saving for retirement (yes / no). Given their answers to a knowledge question about inflation and annuities, we
can check how financially literate they are. They also indicate how satisfied they are with their retirement planning
on a scale from 1 (very satisfied) to 5 (very dissatisfied).
6
8. Allianz International Pension Papers No. 2|2012
question. Ignoring the small percentage of people in the “middle education”
category, we observe increasing satisfaction levels with education, though only
the highest group is significantly more satisfied than the other ones.
Table 1 also shows that our sample is quite wealthy, with only 3 % of the retirees
having less than $ 250,000 in net worth. For this relatively rich group of people,
private wealth is very important in retirement because their replacement rates
from state pensions are comparatively low. In order to maintain their standard
of living, they have to save a lot on top of the state provision. The satisfaction with
retirement planning generally improves across the wealth categories, although
Table 1: Satisfaction with retirement planning depending on education and wealth
Mean satisfaction Percentage in sample
Education
Middle Education 1.50 > 0 %
Secondary School 1.72 5 %
Vocational Certificate 1.66 54 %
Bachelor / Master 1.54 36 %
Doctorate 1.19 5 %
Wealth
< $ 250,000 2.05 3 %
[$ 250,000; $ 500,000) 1.96 10 %
[$ 500,000; $ 750,000) 1.79 16 %
[$ 750,000; $ 1,000,000) 1.49 17 %
[$ 1,000,000; $ 1,500,000) 1.60 19 %
[$ 1,500,000; $ 2,000,000) 1.50 13 %
[$ 2,000,000; $ 3,000,000) 1.29 10 %
[$ 3,000,000; $ 4,000,000) 1.45 6 %
> $ 4,000,000 1.44 6 %
Source: own calculations
7
9. less clearly across the largest categories. This is in line with our expectations.
A person who has done some retirement planning can allocate his resources
more efficiently, accumulate more wealth and be satisfied with it. Once a certain
wealth level is crossed, however, one unit of additional income does not make
such a big difference anymore. Imagine a frequently unemployed person who
has hardly any wealth – say 100 units – and a permanently employed person
who owns about 1,000 units. If you give 50 extra units both to the frequently
unemployed and the permanently employed person, the former is going to be
happier with the increase than the permanently employed person would be.
Given our sample selection, all individuals in the sample belong to the affluent
group and the effect of an additional unit of wealth should be comparatively low.
8
10. Allianz International Pension Papers No. 2|2012
Retirement planning satisfaction
according to savings-related
characteristics
3 1
1 –
(1 + i / n)nE(y)
P = 1000
Table 2 indicates that 56 % of the people in this sample of retirees are very satisfied i /n
with their retirement planning. It does not come as a surprise to see more people where i is the real interest rate, n the number of
believe they have made mistakes in their retirement planning among the dis- payments per year(12) and E(y) the expected
number of years the payment will be paid out.
satisfied people than among the satisfied people. This percentage monotonically
4 OECD (2011)
increases from the very satisfied to the very dissatisfied. Subjective impressions
are thus in line.
The survey also included two knowledge questions about inflation and annuities.
First, people were asked to indicate how much a life-long annuity that pays $ 1,000
as of age 65 would cost: $ 65,000, $ 115,000, $ 165,000, $ 215,000 or $ 265,000.
They could also say they did not know. The formula to calculate an annuity is quite
complicated and depends on several parameters whose values are not given.3
Assuming a real interest rate of 2.5 % and a life expectancy of 16.8 4 at age 65, the
annuity costs $ 164,480. Thus, option three is correct. Secondly, people were asked
to indicate how much $ 1,000 would be worth in real terms in 20 years, assuming
2 % inflation: $ 1,000, $ 980, $ 820, $ 670 or $ 550. They could also indicate that they
did not know. The correct answer is $ 1,000 x (1–0.02)20 = $ 667.61, thus option 4.
The inflation question is much easier to answer than the annuity question because
it only depends on three given parameters and a relatively easy formula. One can
Table 2: Satisfied with retirement planning and believe to have made errors in planning
Satisfaction with retirement planning
Very satisfied 31 % 58 % 56 %
Somewhat satisfied 52 % 34 % 34 %
Neither satisfied nor dissatisfied 65 % 11 % 5 %
Somewhat dissatisfied 76 % 17 % 4 %
Very dissatisfied 100 % 0 % 1 %
Source: own calculations
9
11. 5 Allianz Global Investors, 2011: Putting the argue about the necessity to be able to calculate the price of an annuity or the
Retirement Pieces together: Strategies of the price of inflation accurately for financial decision making. It may be sufficient to
affluent 50+ Generation in the United States.
International Pension Papers 01/2011 be aware of the influencing factors and their dynamics in order to make smart
financial decisions. Also, not everyone may have had a calculator at hand.
In the upper half of Table 3, we distinguish the retirement-planning satisfaction
of people who answered both questions correctly, who answered the annuity
question correctly, who answered the inflation question correctly and who did
not answer any correctly.
Table 3: Satisfaction with retirement planning depending on financial knowledge and savings behavior
Mean Satisfaction Percentage in Sample
Correct knowledge questions
Among them:
don’t know
0 1.56 68 %
0 1.53 40 %
1 1.59 29 %
2 1.57 31 %
1 (inflation) 1.71 21 %
1 (annuity) 1.63 8 %
2 1.60 3 %
Savings behavior
regular and irregular 1.44 28 %
regular but not irregular 1.63 40 %
irregular but not regular 1.74 13 %
no savings rule 1.66 19 %
Source: own calculations
10
12. Allianz International Pension Papers No. 2|2012
Interestingly, people without any correct answers feel best prepared; and these
are 68 % of the people. They can be split up into those who do not have a correct
answer because they admit they do not know (40 % of the 68 %) and those who
falsely thought they knew (60 % of the 68 %). However, there is no statistical
difference between these subgroups with respect to the degree of satisfaction
with retirement planning. The rest is as expected; among those who do know
something, those who answered both knowledge questions correctly feel best,
followed by those who answered the annuity question correctly and those
who got the inflation question right. Only the difference in retirement-planning
satisfaction between the people with zero correct answers and those with only
the correct inflation answer is statistically significant (at a 10 % significance level).
The results seem to suggest that it is good to either not be able to answer the
financial knowledge questions at all or to be able to answer them perfectly.
However, given the peculiarities of the questions, we should not over-emphasize
the descriptive results.
We can distinguish three saving types: a regular saver, an irregular saver and
someone for whom neither applies. Regular saving can imply putting a certain
dollar amount or a certain fraction of income aside at regular intervals. Regular
and irregular saving are not mutually exclusive. A person can save regularly as
well as irregularly. The bottom half of Table 3 summarizes the satisfaction with
retirement planning depending on people’s savings behavior. Most of the people
save regularly, but not irregularly. This is followed by those who save regularly
as well as irregularly and those that do not have a savings rule. The fewest people
save only irregularly. There is a positive order of satisfaction from saving regularly
and irregularly to irregularly, but not regularly, where the former seem to be
doing best in terms of retirement-planning satisfaction. This regularity is inter-
rupted by those who do not have a savings rule. Only the regular and irregular
savings type’s retirement-planning satisfaction is statistically different from that
of the others (at a 1 % significance level). People who do not have any savings
rule may simply be people who do not care.5
The following regression analysis will tell us whether the separate bivariate
relationships still hold once we correct for all variables.
11
13. Marginal Impacts on
Retirement Planning Satisfaction
We estimate the marginal effects of the factors presented above on satisfaction
with retirement planning. This allows us to understand how important each of
these factors is for retirement planning, while controlling for the other effects.
The regression specification can be found in Box 2.
Table 4 below summarizes the regression results. All variables together explain
about 10 % of the variation in satisfaction with retirement planning and are jointly
significant in determining the level of satisfaction (log-likelihood ratio = 125.01).
Among the socio-demographic characteristics, the age category, gender and
being married do not have a significant impact on satisfaction with retirement
planning (not shown in table); individuals with a doctorate have a 27 % higher
chance of ending up very satisfied with their retirement planning compared to
the reference group of people with secondary education. The low t-ratios indicate
that the effect of the knowledge-related questions is statistically not distinguish-
able from zero. Previous studies around retirement planning found financial
literacy to be an important determinant for wealth accumulation and savings
behavior (cf. Annamaria Lusardi and Olivia S. Mitchell). This discrepancy leads us
to the suspicion that the knowledge questions asked are too specific to capture
an individual’s degree of financial intuition or knowledge.
Box 2: Regression specification
We run an ordered probit regression of retirement planning satisfaction on:
• a constant
• age categories (1 to 4)
• a dummy that equals 1 in case of females
• a dummy that equals 1 in case a person is married
• education dummies ISCED categories 4 to 6
• the natural logarithm of net worth within the household. We use the natural logarithm in order to account
for the decreasing impact of wealth on satisfaction as the wealth level increases.
• a dummy that equals 1 in case a person did his retirement planning without any advisor to control for a possible
systematic difference between people who take and who do not take advice.
• a dummy that equals 1 in case a person believes to have made mistakes in his retirement planning
• the number of correct answers to the knowledge questions
• the number of ‘do not know’ to the knowledge questions
• a dummy that equals 1 in case a person saves regularly and irregularly
• a dummy that equals 1 in case a person saves regularly, but not irregularly
• a dummy that equals 1 in case a person saves irregularly, but not regularly
12
14. Allianz International Pension Papers No. 2|2012
As expected, the net worth affects satisfaction with retirement planning positively.
Here, a 10 % increase in wealth leads to a 0.009 (approximately 0.9 %) increase in
chance of being very satisfied with retirement planning. If we have two almost
identical individuals, the only difference being that one owns the median amount
of net worth ($ 1,250,000) and the other one about $ 125,000 more, the more
affluent one has a 0.9 % higher chance of being very satisfied with his retirement
planning than the other one. Though not very large, this positive impact is the
second-most significant effect in our regression. Thus, even in this relatively
selective sample, there is positive wealth effect.
Our regression results indicate that people who save regularly as well as irregularly
have a 16 % higher chance of being very satisfied with their retirement planning.
Table 4: Marginal effects from ordered probit regression
Variable Marginal effect T-ratio
Education: Vocational certificate (isced = 4) – 0.077 – 0.97
Education: Bachelors / Masters (isced = 5) – 0.028 – 0.33
Education: Doctorate (isced = 6) 0.269 2.09
Nr of correct answers (knowledge) – 0.025 – 1.05
Don’t know answer (knowledge) 0.005 0.22
Saving regularly and irregularly 0.159 3.05
Saving regularly but not irregularly 0.084 1.75
Saving irregularly but not regularly – 0.033 – 0.54
Believe to have made mistakes (subjective) – 0.284 – 9.44
others
N = 580; pseudo R2 = 0.107
Note: The table contains the marginal effects, i.e., the probability of being very satisfied with retirement planning due to each variable holding the other
variables constant as well as the respective t-ratios. The latter indicates the degree of certainty that the marginal effect is not zero (if larger than 2.56, 1 %;
1.96, 5 %; and 1.64, 10 %). It is computed using the Delta Method. Note that the ISCED comparison group is ‘middle and secondary’ education and the savings
type comparison group is ‘no savings rule’.
Source: own calculations
13
15. For those who save regularly, but not irregularly, that probability amounts to about
8 %. Someone, however, who saves only irregularly is not more satisfied than some-
one without a savings rule. The latter provides the reference group. Contrary to
the first descriptive results, the type without a savings rule feels worse about his
retirement preparation once we isolate the effect from the others. Since only the
saver types that include regular savings are significantly positive, it seems essential
to save regularly in order to be satisfied with one’s retirement planning ex post.
This leads to a higher success in planning satisfaction than saving irregularly.
Lastly, we find a statistically significant effect from people believing they have
made mistakes in planning. Someone who believes to have made mistakes in
his retirement planning is 29 % less likely to be very satisfied with his retirement
preparation. It is the largest marginal effect of all variables.
14
16. Allianz International Pension Papers No. 2|2012
Conclusion
The baby-boomers who are now retiring in the United States are one of the first
groups of people who – next to public pension income – depend on their private
savings. Many of them believe they have made mistakes. Since future generations
will depend even more on private savings, it is helpful to analyze which kind of
retiree is now content with his retirement planning ex post.
We built on research from Lusardi and Mitchell, and Binswanger and Carman.
The first authors demonstrated that retirement planning substantially increases
retirement wealth, while the second showed that rule-of-thumb planning and
serious retirement planning (the latter being much more costly) lead to statisti-
cally indistinguishable outcomes in terms of net wealth.
Since it is not necessarily best to maximize retirement wealth, we look at how
content retirees are with their retirement planning ex post. Our sample of retirees
is comparatively rich and young. But even among this group of people, we find
a significant effect of net worth and saving type on the satisfaction with retire-
ment planning. A 10 % increase in net worth leads to an almost 0.9 % increase
in the probability of being very satisfied with retirement planning. Among the
distinguished savings types, only regular, regular and irregular, only irregular and
no savings rule, we find significant differences. While the bivariate descriptive
results imply that the no-savings rule type as similarly content as the other types,
the regression results are able to isolate the effects, clearly showing that this
type does feel worse than the regular saving types. The largest effect here is an
approximate 16 % increase in the chance of being very satisfied with retirement
planning of the regular and irregular savers in comparison to the no-saving rules
type. “Only” saving regularly, but not irregularly, increases this chance by 8 %.
Therefore, it seems to pay off to save regularly if you want to be satisfied with
your retirement planning as a retiree.
15 15
17. References
Allianz Global Investors, 2011: Demographic Turning Point (1). Analysis and Trends
Allianz Global Investors, 2011: Pension Systems in a Demographic Transition (2). Analysis and Trends
Allianz Global Investors, 2011: Putting the Retirement Pieces Together: Strategies of the affluent 50+ Generation
in the United States. International Pension Papers 01/2011
Allianz Global Investors, 2011: The Global Pension Atlas 2011, Project M Publication
Allianz Global Investors, 2011: Paradigm Lost. Project M, 7
Binswanger, Johannes and Carman, Katherine G., 2009: How real people make long-term decisions:
The case of retirement preparation, Center Discussion Paper No 2009-72
Lusardi, Annamaria and Mitchell, Oliva S., 2007: Baby-boomer retirement security: The roles of planning,
financial literacy, and housing wealth. Journal of Monetary Economics, 54
OECD, 2011: Pensions at a Glance 2011: Retirement Income Systems in OECD and G20 Countries
16