Upload
others
View
3
Download
0
Embed Size (px)
Citation preview
Dollar-Cost Averaging Using the CAPE Ratio: An Identifiable Trend
Influencing Outperformance
By Jon Luskin, MBA CFP®
Define Financial
© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
Financial Planning Association of MinnesotaOctober 16th, 2017, 2:30 p.m.
Abstract
• Previous research: DCA underperformed LSI most of the time• Did not examine the circumstances of this outperformance
• Concurrent with existing research, found LSI outperformed DCA• 15-year periods• roughly two-thirds of the time • on a nominal return basis, when ignoring taxes and transaction costs
• DCA outperformance a function of CAPE • higher CAPE ratios linked to DCA outperformance
• Time Permitting• What I couldn’t fit into the FPA Journal• Successive Research
© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
Why This Paper
© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
Why DCA & CAPE
• Who uses DCA?
• Over what period?
• When/why?
© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
Why DCA & CAPE
• How to Best Advice Clients with• Business Liquidation
• Real Estate Liquidation
• Pension Lump-Sum Distribution
• Inheritance
© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
Investing in Stocks
• History shows:• Grow money
• In the long term
• Can lose money• In the short term
• Because of volatility
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016
© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
Dollar Cost Averaging
• The potential for short-term losses may have inspired the idea behind dollar cost averaging
• Don’t invest your money all at once
• Invest small chunks of your money over time
20
00
20
00
20
01
20
01
20
02
20
02
20
03
20
04
20
04
20
05
20
05
20
06
20
07
20
07
20
08
20
08
20
09
20
09
20
10
20
11
20
11
20
12
20
12
20
13
20
14
20
14
20
15
20
15
20
16
20
16
20
00
20
00
20
01
20
01
20
02
20
02
20
03
20
04
20
04
20
05
20
05
20
06
20
07
20
07
20
08
20
08
20
09
20
09
20
10
20
11
20
11
20
12
20
12
20
13
20
14
20
14
20
15
20
15
20
16
20
16
$ $ $$
$$
$ $$ $
$✓© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
Dollar Cost Averaging
• Proven to be less risky than LSI (as measured by standard deviation)• Williams, R. E., & Bacon, P. W. (1993). Lump Sum Beats Dollar-Cost Averaging.
Journal of Financial Planning, 64-67.
• Dubil, R. (2005). Lifetime Dollar-Cost Averaging: Forget Cost Savings, Think Risk Reduction. The Journal of Financial Planning.
© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
Dollar Cost Averaging
• Proven to be less risky than LSI (as measured by standard deviation)
DCA in progress – less volatile
Stocks
Cash not yet invested
LSI – more volatile
© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
Dollar Cost Averaging
• If less risky, what about a higher investment return?
• Not really
• Less nominal investment return (on average)
20
00
20
00
20
01
20
01
20
02
20
02
20
03
20
04
20
04
20
05
20
05
20
06
20
07
20
07
20
08
20
08
20
09
20
09
20
10
20
11
20
11
20
12
20
12
20
13
20
14
20
14
20
15
20
15
20
16
20
16
20
00
20
00
20
01
20
01
20
02
20
02
20
03
20
04
20
04
20
05
20
05
20
06
20
07
20
07
20
08
20
08
20
09
20
09
20
10
20
11
20
11
20
12
20
12
20
13
20
14
20
14
20
15
20
15
20
16
20
16
$ $ $$
$$
$ $$ $
$✓© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
Dollar Cost Averaging
• Risk and return are positively correlated
© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
$4,744
$27,120
$21
$0
$1
$10
$100
$1,000
$10,000
19
26
19
27
19
28
19
30
19
31
19
33
19
34
19
35
19
37
19
38
19
40
19
41
19
43
19
44
19
45
19
47
19
48
19
50
19
51
19
52
19
54
19
55
19
57
19
58
19
60
19
61
19
62
19
64
19
65
19
67
19
68
19
69
19
71
19
72
19
74
19
75
19
77
19
78
19
79
19
81
19
82
19
84
19
85
19
86
19
88
19
89
19
91
19
92
19
94
19
95
19
96
19
98
19
99
20
01
20
02
20
03
20
05
20
06
20
08
20
09
20
11
20
12
20
13
20
15
$1 Invested in...
All Stocks Small Stocks 1M T-Bills
Indices data sourced from DFA Web Returns, CRSP 1-10, CRSP 10, One Month T-Bills (Ibbotson & Morningstar)
Dollar Cost Averaging
• DCA doesn’t usually beats LSI• Greenhut, J. G. (2006). Mathematical Illusion: Why Dollar-Cost Averaging
Does Not Work. The Journal of Financial Planning.
• LSI beats 2/3 of the time• Shtekhman, A., Tasopoulos , C., & Wimmer, B. (2012, July). Dollar-cost
averaging just means taking more risk later. Retrieved from Vanguard Group• 1926 to 2011; 6, 12, 18, 24, 30, or 36 months; U.S., U.K. & Australia
• Williams, R. E., & Bacon, P. W. (1993). Lump Sum Beats Dollar-Cost Averaging. Journal of Financial Planning, 64-67• 1926 to 1991; 12-month holding periods; SP & 500
© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
Dollar Cost Averaging
• Because on average, the market move up (upward trending)
• Better off lump-sum investing (LSI)
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016
© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
1/3 of the time?
• If LSI outperforms, how does DCA outperform during that 1/3 of time?• Flat, downward trending, or volatile markets
• Shtekhman, et al., 2012• Greenhut, 2006
• Previous literature briefly mentioned high volatility made for DCA success• But stopped short of making a full analysis
• Would it be possible to determine what circumstances make for the success of DCA?
• Is it possible to predict when DCA would the better strategy?
• How would we determine this?
© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
CAPE ratio
• Cyclically Adjusted Price to Earnings ratio• AKA Shiller P/E• Invented by economist & Nobel Laurette Robert Shiller
• A valuation metric• Measures if stocks are “cheap” or “expensive”• Looks at company earnings• Over 10 years• Adjusted for price• Relative to stock price• 10 years earnings, adjusted for inflation ÷ stock share price
• Varies between 5+ and 44+• ~16 being average
© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
CAPE ratio
• Has predictive power for investment return
• Significant negative correlation (-0.41), 1871-2015• Shiller, R. (2016, July 28). Online Data - Robert Shiller. Retrieved from Yale Department of Economics: http://www.econ.yale.edu/~shiller/data.htm
-8%
-3%
3%
8%
13%
18%
4 9 14 19 24 29 34 39 44
An
nu
aliz
ed P
rice
A
pp
reci
atio
n
CAPE at begining of 15-Year Period© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
Using Valuation Metrics as an Investment Strategy• 5-year normalized P/E ratio to tilt equity allocations
• Excess investment returns at highest lowest decile of valuations• Kitces, M., Solow, K., & Locatelli, S. (2011). Improving Risk-Adjusted Returns Using
Market-Valuation-Based Tactical Asset Allocation Strategies. Journal of Financial Planning, 48.
• If a multi-year, inflation adjusted P/E ratio can be used to tilt equity allocation, why not for DCA?• It’s the same thing!
© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
2.50
7.50
12.50
17.50
22.50
27.50
32.50
37.50
42.50
18
81
.01
18
83
.02
18
85
.03
18
87
.04
18
89
.05
18
91
.06
18
93
.07
18
95
.08
18
97
.09
18
99
.1
19
01
.11
19
03
.12
19
06
.01
19
08
.02
19
10
.03
19
12
.04
19
14
.05
19
16
.06
19
18
.07
19
20
.08
19
22
.09
19
24
.1
19
26
.11
19
28
.12
19
31
.01
19
33
.02
19
35
.03
19
37
.04
19
39
.05
19
41
.06
19
43
.07
19
45
.08
19
47
.09
19
49
.1
19
51
.11
19
53
.12
19
56
.01
19
58
.02
19
60
.03
19
62
.04
19
64
.05
19
66
.06
19
68
.07
19
70
.08
19
72
.09
19
74
.1
19
76
.11
19
78
.12
19
81
.01
19
83
.02
19
85
.03
19
87
.04
19
89
.05
19
91
.06
19
93
.07
19
95
.08
19
97
.09
19
99
.1
20
01
.11
20
03
.12
20
06
.01
20
08
.02
20
10
.03
20
12
.04
20
14
.05
Tilting Equity Allocation Relative to CAPE
CAPE
CAPE Average
Underweight equities
Overweight equitiesOverweight equities
Underweight equities
© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
Challenges Using CAPE
• Imperfect valuation metric
• 10 years may be longer than a business cycle
• Technique for measuring inflation has changed with time
• Accounting standards and corporate taxation have changed over time• Wilcox, S. E. (2011, September). A Cautionary Note About Robert Shiller’s
CAPE. Retrieved from AAII: The American Association of Individual Investors
© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
Hypothesis
Can market valuations (CAPE) be used to determine an identifiable trend behind the best opportunities to implement DCA?
5
10
15
20
25
30
35
40
451
88
11
88
31
88
61
88
91
89
11
89
41
89
71
89
91
90
21
90
51
90
71
91
01
91
31
91
51
91
81
92
11
92
31
92
61
92
91
93
11
93
41
93
71
93
91
94
21
94
51
94
71
95
01
95
31
95
51
95
81
96
11
96
31
96
61
96
91
97
11
97
41
97
71
97
91
98
21
98
51
98
71
99
01
99
31
99
51
99
82
00
12
00
32
00
62
00
92
01
12
01
4
CA
PE
Year
1881-2016 1950-2015
© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
Data & Methodology
• S&P 500, Total Return• Yahoo Finance
• 90 Day T-Bills• Board of Governors of the Federal Reserve, 2016
• 1950-2015
• CAPE data from Shiller’s site at Yale
• 15-year rolling time periods• Robust results, vis-à-vis 5 or 10 year• Monthly deposits
• 180 deposits in total• 12 months X 15 years = 180 deposits
• Uninvested cash grew at the risk-free rate© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
Data & Methodology
• No consideration for taxes or fees• Account fees
• Expense Ratios
• Transaction / Trade Fees / Commissions
• IRA, at Vanguard, using Vanguard funds, etc.
© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
36+ 35.99-31 30.99-26 25.99-21 20.99-16 15.99-11 10.99-6 5.99-0
5%
2%1%
14%
33%
25%
20%
0%
2%1%
4%
14%
29%
31%
18%
2%
Freq
uen
cy o
f O
ccu
rren
ce, %
of
Mo
nth
s w
ith
CA
PE
Per
Tim
e Pe
rio
d
CAPE Ratio
1950-2015 1871-2015
© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
Data & Methodology
• Sample Representativeness
• More frequent higher valuations
• Less lower valuations
Results
• DCA outperformance as CAPE increased• With exceptions
• correlation of 0.43
-6%
-4%
-2%
0%
2%
4%
5 10 15 20 25 30 35 40 45
DC
A O
utp
erfo
rman
ce, %
Starting CAPE for the 15-Year Investment Period
© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
Results
• DCA outperformed 1/3 of the time• As per previous literature
• 1/3 of time, CAPE valuations are above ~18.6
• Using DCA at valuations above ~18.6 averaged an excess return of 0.45 BPS per year• Over 15 years
• Increasing valuations made for a greater degree of DCA outperformance – with the tech bubble exception
© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
Results
• Using DCA at higher valuations averaged higher outperformance• But a non-linear relationship, from the tech bubble
-3.0%
-2.0%
-1.0%
0.0%
1.0%
2.0%
0%
20%
40%
60%
80%
100%Frequency of DCAoutperform
Degree of DCAoutperformance
Linear (Frequency ofDCA outperform)
Linear (Degree of DCAoutperformance)
© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
Results
• Higher CAPE usually meant greater DCA outperformance
• With the tech bubble as an exception
-6%
-4%
-2%
0%
2%
4%
5 10 15 20 25 30 35 40 45
DC
A O
utp
erfo
rman
ce
starting CAPE
DCA outperform
mid 90s
© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
Considerations for Applications
• Perspective #1: You Can Use CAPE to Indicate when to Use DCA• CAPE of 31 – as of Friday, October 13th, 2017
• Will valuations peak at ~44, as per the tech bubble?• Or will valuations surpass ~44?
• Is this the valuation peak?
• Perspective #2: You Cannot Use CAPE to Indicate when to Use DCA• Allocate to a Portfolio that is Always Risk Appropriate for your Client
• “Taking Risk Later”
• What Happens During a Drawdown at Year 16?
© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
Future Research
• Closely Examine Additional Time Periods• 12 months, 24 months, 36 months, 60 months
• Incorporate Taxes & Fees
• Examine Time Periods Back to 1926, 1871• Great Depression
• Examine five- & one-year P/E valuation metrics
© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
What I Couldn’t Squeeze into the FPA Journal
• Best Month for LSI? November • foreign and domestic indices; 1970–1998; 12-month periods
• Atra, R. J., & Mann, T. L. (2001). Dollar-Cost Averaging and Seasonality: Some International Evidence. Journal of Financial Planning, 98–105.
• November has the second highest CAPE ratio, on average, behind December (this study)• Returns on short timelines (one year) determined by momentum
• Return on long timelines (15 years) determined by valuation
© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
What I Couldn’t Squeeze into the FPA Journal
Timeline
LongShort
Valuations
High
Low
LSI
LSIDCA
DCA
© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
What I Couldn’t Squeeze into the FPA Journal
• Risk-Adjusted Return• LSI ~0.1+ Sharpe ratios, on average
• Shtekhman, et al. (2012)
• 1926 to 2011; 6, 12, 18, 24, 30, or 36 months; U.S., U.K. & Australia
• LSI ~0.07+ Sharpe ratios• Leggio, K. B., & Lien, D. (2003, January). Comparing Alternative Investment Strategies
Using Risk-Adjusted Performance Measures. The Journal of Financial Planning.
• One-year periods; 1926-1999; S&P 500
© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
What I Couldn’t Squeeze into the FPA Journal
• LSI outperformed roughly 1/2 of the time
(0.10)
(0.05)
-
0.05
0.10
5 10 15 20 25 30 35 40 45DC
A O
utp
erfo
rman
ce, S
har
pe
Starting CAPE
DCA outperformance
Poly. (DCA outperformance)
© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
What I Couldn’t Squeeze into the FPA Journal
• LSI outperformed roughly 1/2 of the time• DCA outperforming when valuations are high and low
• but not median
• DCA strategy outperforms (nominal & risk-adjusted) when valuations are high• Higher Return for Less Risk
• Hence higher Sharpe Ratio
© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
What I Couldn’t Squeeze into the FPA Journal
• DCA outperforms on a risk-adjusted basis when valuations are low• Because of consistently high variations in investment return for LSI
• from upward deviations
• Hence higher Sharpe
• But who really cares about upward volatility?
• Sharpe perhaps not an appropriate metric in this circumstance
© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
Postscript: Shorter timelines?
• Shorter timelines of DCA implementation show no consistent result• 6 months, 9 months, 12 months, 18 months
© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
-80%
-60%
-40%
-20%
0%
20%
40%
5 10 15 20 25 30 35 40 45
ou
tper
form
ance
of
DC
A
CAPE ratio at inception
CAPE, 3 mo look back (real)
Postscript: Shorter timelines?
• Shorter timelines of DCA implementation show no consistent result• 6 months, 9 months, 12 months, 18 months
© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
Postscript: Shorter timelines?
• CAPE most predictive on longer timelines
© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
(0.5)
(0.6)
(0.7)
(0.6)
(0.5)
(0.4)
(0.3)
(0.2)
(0.1)
-
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Co
rre
lati
on
to
An
nu
al R
eal
Ret
urn
Length of Investments, Year
Postscript: CAPE is Broken
• Accounting standards changed, making CAPE today different (Seigel, 2016)• National Income and
Product Accounts (NIPA)
© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
0
0.5
1
1.5
2
2.5
3
18
81
18
83
18
84
18
86
18
88
18
90
18
92
18
94
18
95
18
97
18
99
19
01
19
03
19
05
19
06
19
08
19
10
19
12
19
14
19
16
19
17
19
19
19
21
19
23
19
25
19
27
19
28
19
30
19
32
19
34
19
36
19
38
19
39
19
41
19
43
19
45
19
47
19
49
19
50
19
52
19
54
19
56
19
58
19
60
19
61
19
63
19
65
19
67
19
69
19
71
19
72
19
74
19
76
19
78
19
80
19
82
19
83
19
85
19
87
19
89
19
91
19
93
19
94
19
96
19
98
20
00
20
02
20
04
20
05
20
07
20
09
20
11
20
13
20
15
20
16
rolling CAPE using NIPA, deviation from mean, 3 mos lookback
rolling CAPE using S&P, deviation from mean, 3 mos lookback
Postscript: CAPE is Broken
© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
• w/ NIPA data, a long-enough time period claims below-average valuations
Postscript: CAPE is Broken
© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
• Incorporating real bond yield can improve forecasting (Vanguard, 2017)
© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
Successive Research
© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
• Testing this same strategy• Using NIPA data
• Vanguard’s algorithm incorporating real bond yields
By Jon Luskin, MBA CFP®
Define Financial
750 B Street, Suite 1750, San Diego, CA 92126
619.577.4002
www.JonLuskin.com/DCA
© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary
FPA MN Coordinator: Bonnie Stanley, Financial Planning Association of MN, 3900 Main Street N.E., Columbia Heights, MN 55421, [email protected], 763-781-1212