Ten Facts I have learned about 401(k) Plans

Document technical information

Format ppt
Size 123.4 kB
First found May 22, 2018

Document content analysis

Category Also themed
Language
English
Type
not defined
Concepts
no text concepts found

Persons

Organizations

Places

Transcript

Ten Facts I Have Learned
About 401(k) Plans
Martin J. Gruber
The Adequacy of Choices Offered by
401(k) Plans
The Performance of Funds Offered by
401(k) Plans
The Impact of Mutual-Fund Family
Membership On Investor Risk
10 Lessons from this Research:
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
Most plans do not offer enough or the appropriate mix of options
to participants
Company stock does not affect the adequacy of options
Plans tend to have increased risk because they tend to select
funds from 1 or 2 families
Given the type of funds offered, administrators tend to pick better
than random funds, but much of the difference is due to lower
expense ratios
Participants don’t do better than the 1/n rule
Funds that were added did better before they were added but not
better after they were added
Funds that were dropped did worse before they were dropped and
no better after they were dropped
Plan administrators who outperformed others in the past have a
tendency to outperform in the future
Participants’ contributions, transfers, and return all have about the
same effect on change in investment proportions
Participant changes in allocation exaggerate the change due to
return
1. Most plans don’t offer enough
or the right mix of options to
participants.
Does adding index funds as suggested by
the literature of financial economics or an
ICDI category index of mutual funds to the
mix of offerings shift the efficient frontier by
an amount which is statistically significant?
A. Data
Moody’s survey of pension plans:
Select 401(k) plans that offer only mutual
funds with or without money market
accounts, GICs, stable value funds and
company stock – 680 plans
417 of these had mutual funds with at least
5 years of data.
Percentages of 680 401(k) Plans Offering
Different Numbers of Investment Choices
(Number of choices and percentages include mutual funds, stable value funds, GICs and company stock.)
Number of Investment Choices
Percentage of Plans
1
2.21%
2
2.35%
3
3.09%
4
4.85%
5
8.97%
6
12.06%
7
12.06%
8
13.82%
9
11.76%
10
9.85%
11
5.59%
12
2.21%
13
2.50%
14
1.91%
15
1.18%
16
1.03%
17 or more
4.56%
The Adequacy of Plan Offerings
To judge adequacy, we need to look beyond
risk to the combination of risk and return.
Do the plans offered to participants span the
8 RB indexes which explain returns?
Excess Return on each RB index =
α + ∑iβi excess return on fund i offered.
Sufficiency of Plan Investment Choices
In Spanning 8 RB Indexes
(Short Sales Not Allowed)
Number of Investment
Choices in Plan
6 or less
7 to 12
Over 12
Total
Number Not Sufficient
223
164
19
74%
50%
10%
406
61%
2. Adding company stock is not
bad per se.
Company Stock
1. Including company stock, assuming 1/n rule:
variance up by 3.17 or 19% (t-value 3.6)
2. Sharpe ratio up from 2.40 to 2.55, but increase
comes from added security. If add random
fund rather than company stock, Sharpe ratio
stays at 2.5.
3. Spanning – no effect. Plans that didn’t span
before still don’t span.
Company stock virtually no effect under 1/n rule.
3. Plans tend to have more risk
because they choose funds
from 1 or 2 families.
Standard devation not higher.
Correlation coefficients are higher.
Correlation between two funds of any type within
families is higher than correlation of two
similar funds across families.
Can make a difference of 52 to 70 bp per year.
For other rules, new sample.
11-K filling, 401(k) Plans 1994-2003
401(k) Plan Sample
Number of 401(k) Plans
43
Number of Plan Years
289
Number of Unique Funds Held
141
Number of Funds Initially Helda
116
Number of Funds Added
215
Number of Funds Deleted
a The
total number of funds held by the 43 sample
plans in the first year each plan enters our sample
45
Methodology
A. Alpha
Rit – Rrt = αi + ∑ βij ∙ Ijt + eit
Stock Funds: S&P 500, Fama French Small-Large
and high minus low, Lehman Gov/Credit, and
MSCI Europe
Bond Funds: Lehman Gov/Credit, Lehman
Mortgage-Backed, Credit Suisse High-Yield
Index, Salomon non-dollar World Gov. Bond
Index
International: S&P 500 and the three MSCI
Indexes (Europe, Pacific, and Emerging Markets
B. Differential Alpha
Mutual funds, in general, have negative
alpha. We took the alpha for each mutual
fund minus the average alpha for funds of
the same general size from the same
ICDI category.
To get alpha on a plan we use two
alternative weightings:
1. Equal weight on each mutual fund
2. Weight by participants’ holdings
4. Given the type of fund offered,
administrators tend to pick better
than random funds, but much of
the difference is due to lower
expense ratios.
5. Participants don’t do better than
the 1/n rule.
Performance 3-Year α
43 Plans, with an average of 6.7 years per plan
Average
P-Value
# Pos.
Equal Wts.
Alpha
Diff. α
-0.026
0.043
0.160
0.010
30
Fee difference .019
Participant Wts.
Alpha
Diff. α
-0.043
0.037
0.030
0.040
32
Performance 1-Year α
Average
P-Value
# Pos.
Equal Wts.
Alpha
Diff. α
-0.080
0.035
0.000
0.040
29
Fee difference .019
Participant Wts.
Alpha
Diff. α
-0.093
0.041
0.000
0.030
33
6. Funds that were added did better
before they were added and not
better after they were added.
7. Funds that were dropped did
worse before they were dropped
and no worse after they were
dropped.
Before Action Diff. Alpha
1-Year
Added (200)
0.000
Dropped (44)
- 0.112
Difference
0.112
P-value
0.020
After Action Diff. Alpha
1-Year
Added (214)
0.004
Dropped (43)
0.087
Difference
- 0.083
P-value
0.207
Added Funds and Past Performance
Of Investment Objectives
Average of Past α
of Objective Added
-0.048
Average Past α
of All Objectives
-0.066
Difference
P-value
0.018
0.000
8. Plan administrators who outperform
in the past have a tendency to
outperform in the future.
Past Performance
Quartiles
1 (lowest)
Average Future
Differential Alpha
-0.024
2
0.040
3
0.063
4 (highest)
0.061
Performance and Plan Characteristics
•
•
•
•
•
Dollar size
Number of choices
Changes in choices
New cash flow
Presence of money market
No relationships.
9. Return, participant contributions,
and transfers all have about the
same effect on change in weights.
What causes change in the percentages
participant place in each choice they are
offered?
1. Return
2. Contribution
3. Transfer
3.8%
3.7%
3.6%
10. Participants’ change in allocation
exaggerate the change in weight
due to return.
Change in weight due to contributions and
transfers equals α + β change due to
return.
β is positive for 36 out of 41 plans.
β = 0.63
R2 = 0.17
10 Lessons from this Research:
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
Most plans do not offer enough or the appropriate mix of options
to participants
Company stock does not affect the adequacy of options
Plans tend to have increased risk because they tend to select
funds from 1 or 2 families
Given the type of funds offered, administrators tend to pick better
than random funds, but much of the difference is due to lower
expense ratios
Participants don’t do better than the 1/n rule
Funds that were added did better before they were added but not
better after they were added
Funds that were dropped did worse before they were dropped and
no better after they were dropped
Plan administrators who outperformed others in the past have a
tendency to outperform in the future
Participants’ contributions, transfers, and return all have about the
same effect on change in investment proportions
Participant changes in allocation exaggerate the change due to
return
×

Report this document