ArticlePDF Available

Figures

Content may be subject to copyright.
Another Look at Dollar Cost Averaging
Gary Smith Heidi Margaret Artigue
Department of Economics Department of Economics
Pomona College Pomona College
425 N. College Avenue 425 N. College Avenue
Claremont CA 91711 Claremont CA 91711
gsmith@pomona.edu hmal2015@MyMail.Pomona.edu
Another Look at Dollar Cost Averaging
Abstract
Dollar cost averaging—spreading an investor’s stock purchases evenly over time—is widely
touted in the popular press because of the mathematical fact that the average cost per share is less
than the average price. The academic press has generally been skeptical, and attributes dollar
cost averaging’s popularity to investor naiveté and cognitive errors. Yet, dollar cost averaging
continues to be recommended by knowledgeable investors as a sensible way to avoid ill-timed
purchases. We argue that dollar cost averaging is, in fact, an imperfect, but helpful strategy for
diversifying investment decisions across time.
keywords: dollar cost averaging!
Another Look at Dollar Cost Averaging
An investor following a dollar cost averaging (DCA) strategy periodically invests a constant
dollar amount in stocks, adjusting the number of shares purchased as stock prices fluctuate.
When stock prices are high, fewer shares are bought; when prices are low, more shares are
purchased. Because the average cost weights the purchase prices by the number of shares
acquired at each price, the average cost is always less than the average price.
This mathematical fact has led many to recommend dollar cost averaging. In several Barron’s
columns and multiple editions of Successful Investing Formulas, first published in 1947, Lucile
Tomlinson persuasively extolled the virtues of cost-averaging plans. In the 2012 edition, with a
preface written by Benjamin Graham, Tomlinson wrote that dollar cost averaging is “the
unbeatable formula,” because, “No one has yet discovered any other formula for investing which
can be used with so much confidence of ultimate success, regardless of what may happen to
security prices, as Dollar Cost Averaging.”
The academic literature has generally been scornful, concluding that DCA is an inferior
strategy. Constantinides (1979) finds it ironic that DCA’s fixed investment rule is considered a
benefit. He argues that DCA’s inflexibility is a fatal flaw because making portfolio decisions
based on the additional information accumulated as time passes must be superior to committing
oneself to a strategy based only on current information. One counterargument is that in an
efficient market, it may be difficult to profit from future information.
Knight and Mandell (1992/1993) argue that DCA is suboptimal because the DCA portfolio’s
riskiness increases as stock purchases are made at regular intervals, but an investor with constant
!2
relative risk aversion would not increase a portfolio’s riskiness over time.
Thorley (1994) and Milevsky and Posner (2003) criticize a comparison of average price with
average cost, since investors cannot sell the stocks they accumulate at the historical average
price; however, this does not address the issue of whether DCA allows investors to acquire
stocks at a relatively low average cost compared to an initial lump-sum (LS) investment. The
answer obviously depends on whether stock prices subsequently increase or decrease. Thus,
Grable and Chatterjee (2015) observe that, while LS does well in the early stages of a bull
market, DCA may be more suitable for risk-averse investors who are fearful of a bear market.
Luskin (2017) reports that DCA does better when initiated during historical periods when
Shiller’s cyclically adjusted price-earnings (CAPE) ratio is abnormally high.
Bierman and Hass (2004) argue that if an investor is considering an investment that has an
expected return that is higher than from not being invested, then delaying purchases must reduce
the investor’s expected return. However, Cho and Kuvvet (2015) use a two-period mean-variance
framework to argue that if the risk-return opportunity locus is concave, some investors may
prefer DCA to LS.
Milevsky and Posner (2003) show that DCA can be financially attractive if investors have a
fixed target portfolio value. In particular, DCA has a positive expected profit if the price of the
stock being purchased is the same at the end of acquisition period as at the beginning, and the
size of this profit increases with the stock’s volatility. More generally, if the terminal stock price
is known in advance, the higher the volatility, the higher the expected value of the return from
DCA, so that there is some sufficiently high level of volatility such that DCA has a higher
expected return than LS. They argue that investors do, in fact, have terminal target prices, and
!3
conclude (presumably tongue in cheek) that
perhaps the investing masses are more sophisticated than we are willing to admit. Indeed,
they may be familiar with the Ito-Markovian structure of asset prices, yet they insist on
dollar-cost averaging because they believe market prices are driven by Brownian bridges as
opposed to Brownian motions. (Milevsky and Posner 2003, p. 191)
Some academics accept the argument that DCA is inferior to LS, and ask why so many
investors continue to use and recommend it. They identify the main benefits as encouraging
regular savings, protecting investors from trend chasing and portfolio churning, and avoiding a
single ill-timed purchase that might scare them away from future stock purchases (Haley 2010).
Others (Statman 1995; Dichtl and Drobetz 2011) argue that dollar cost averaging can be justified
by Kahneman and Tversky’s (1992) prospect theory; however, Leggio and Lien (2001) present
evidence to the contrary and, in addition, argue that DCA is less successful with more volatile
stocks—contradicting Milevsky and Posner.
On the empirical front, Williams and Bacon (1993) look at DCA and LS investments in the
S&P 500 and find that LS generally has a higher return, which is unsurprising since, historically,
being in the stock market has, on average, been more profitable than being out of the market.
Leggio and Lien (2003a, 2003b) consider the average return and various risk measures and
conclude that DCA is inferior to LS. Thorley (1994) reports that LS has a somewhat higher
average return and lower standard deviation than DCA. Similarly, Rozeff (1994) concludes that,
relative to DCA, the LS strategy tends to have a higher return because it is more fully invested in
stocks and the LS strategy tends to have a lower variance because it is more uniformly invested
in stocks, as opposed to being lightly invested early on and heavily invested later. Using
!4
simulations with historically based parameters, Abeysekera and Rosenbloom (2000) conclude
that DCA generally has a lower return and less risk. Using bootstrapped CRSP index returns, Lei
and Li (2007) report that the results from DCA and LS are statistically indistinguishable over
short and long horizons, but that DCA has a higher probability of a negative return over
intermediate-term horizons.
There is some empirical evidence in support of DCA. Balvers and Mitchell (2000) and
Brennan, Li, and Torous (2005) argue that DCA may be advantageous if stock returns are
negatively autocorrelated, as in the mean-reversion literature (Poterba and Summers 1988; Fama
and French 1988). To exploit this mean reversion, Dunham and Friesen (2012) recommend an
“enhanced DCA” strategy in which stock purchases are increased after a market decline, and
reduced after a market advance. Dubil (2005a) and Trainor (2005) argue that DCA reduces
shortfall risk, the chances of wealth dropping below a specified floor during the investor’s
lifetime.
Israelson (1999) uses mutual fund data to argue that DCA is a superior strategy, especially for
low-volatility mutual funds—which contradicts the more widespread argument that DCA is best
suited for volatile investments. Atra and Mann (2001) argue that the January effect and other
seasonal patterns in stock returns affect comparisons of DCA and LS. Paglia and Jiang (2006)
report that the day of the month matters, too. These claims are examples of the more general
point made by Dubil (2005b) that a comparison of DCA with LS depends on the specific stock
price patterns that happen to occur; for example, LS will have an advantage if stock prices head
up after the plan is started, while DCA has the advantage if prices drop. Comparing strategies
that are always launched on January 1 may distort the evidence.
!5
Overall, the academic verdict is that DCA is an inferior strategy both in theory and practice,
but investors use it anyway for not particularly persuasive reasons. The unsettling thing about
this academic conclusion is that many of those who endorse DCA are, in other respects, very
sensible and experienced. Perhaps the academic verdict is too harsh and underestimates the
wisdom of experienced practitioners?
Despite the generally negative conclusions in peer-reviewed academic research, dollar cost
averaging is widely endorsed in the popular press. Damato (1994) wrote in The Wall Street
Journal that,
People who invest in stocks regularly get the benefit of “dollar cost averaging.” Because a
fixed sum is invested every month or quarter buys more shares when prices are down, the
investor’s average cost per share ends up being lower than the average price in the market
over the same period.
Sylvia Porter (1979) describes dollar cost averaging as “how to beat the stock market.” The
Investment Company Institute (1984) says that, “It takes the ups and downs of the market and
turns them to advantage.” The New York Times Financial Planning Guide (1985) calls dollar cost
averaging a “time-tested investing method” with a “seemingly magic result.”
The financial press is one thing, but it is striking that DCA is recommended by several
extremely knowledgeable observers, including Levy and Sarnat (1972), Sharpe (1978), Dreman
(1982), Loeb (1996), Bogle (2015), and Tobias (2016). In the 2015 edition of his classic book, A
Random Walk Down Wall Street, Malkiel writes that,
Dollar cost averaging can reduce the risks of investing in stocks and bonds….Periodic
investments of equal dollar amounts in common stocks can reduce (but not avoid) the
!6
risks of equity investment by ensuring that the entire portfolio of stocks will not be
purchased at temporarily inflated prices. (p. 355)
Many individuals and institutions practice what these advisors preach, ranging from Fidelity’s
Automatic Account Builder, which transfers money each month from a participant’s bank
account to a stock fund, to the MacArthur Foundation, which used dollar cost averaging in
1985-1986 (at the rate of $100 million a month) to invest $1.4 billion realized from the
liquidation of real estate and other tangible assets.
Neither unequivocal praise or scorn is warranted. The core argument in favor of dollar cost
averaging gauges performance by cost, rather than rate of return, and, when this is taken into
account, the alleged simple virtues of cost averaging vanish. On the other hand, many criticisms
of dollar cost averaging neglect the return on funds not invested in stock or don’t compare
equivalent DCA and LS strategies (for example, equally risky strategies). When these factors are
taken into account, dollar cost averaging may be a reasonable approach to investing in volatile
stocks. We first demonstrate that the claimed low-cost virtue is an illusion and then use a novel
approach to analyze how dollar cost averaging can effectively diversify investment risk across
time.
The Illusion
The presumed advantages of dollar cost averaging are invariably based on calculations similar to
those shown in Table 1. A stock currently sells for $60 a share and its price will either rise or fall
by 50 percent. If $900 is invested each period, the average cost is less than the average price, no
matter whether the stock’s price goes up or down. The intriguing implication is that, although the
expected value of the price change is zero, the expected value of the average cost is less than the
!7
expected value of the price in the second period. In Table 1, the expected value of the average
cost is 0.5($72) + 0.5($40) = $56, and the expected value of the price is 0.5($90) + 0.5($30) =
$60. DCA appears to have a positive expected profit even if changes in stock prices are
independent with a mean of zero.
Table 1 Dollar Cost Averaging Reduces Costs
Price Rises to $90 Price Falls to $30
Time period Price Shares Cost Price Shares Cost
1 $60 15 $900 $60 15 $900
2 $90 10 $900 $30 30 $900
Average $75 $72 $45 $40
This conclusion is misleading. The mathematical fact that the expected value of the average
cost is less than the expected value of the price does not imply a positive expected rate of return,
because it neglects the varying amounts invested in the each scenario. The relevant data are
shown in Table 2. DCA gains $18 a share if the price rises and loses $10 a share if it falls. The
average of these two numbers is indeed positive. But it is also irrelevant since the number of
shares is not the same. DCA gains $18 a share on 25 shares or loses $10 a share on 45 shares—a
$450 profit on an $1,800 investment or a $450 loss on an $1,800 investment. Even though the
average cost is less than the average price, if a stock’s expected return is zero, DCAs expected
profit is also zero.
This is reminiscent of the Martingale betting system (Mitzenmacher and Upfal, 2005) in
which bets are doubled after every loss. If the outcomes are independent and each wager has a
zero expected return, the expected return from the strategy must be zero—no matter how one
!8
adjusts the size of the wagers as time passes. This is true of a Martingale betting strategy and it is
true of a DCA investing strategy. What these strategies do change is the shape of the payoff
distribution, including the variance.
Table 2 Dollar Cost Averaging Does Not Increase Expected Return
Price Rises to $90 Price Falls to $30
Average cost $72 $40
Number of shares 25 45
Total investment $1,800 $1,800
Value of portfolio $2,250 $1,350
Dollar gain $450 - $450
Rate of return 25% - 25%
A Mean-Variance Analysis
Malkiel and other supporters of cost averaging emphasize the risk of making a lump sum
investment at an unfortunate time; i.e., shortly before a big drop in prices. Cost averaging
essentially diversifies one’s investments, not across different stocks, but across different purchase
dates. We use mean-variance analysis with Tobin’s separation theorem to demonstrate that time-
diversification can be an important virtue.
An evaluation of dollar cost averaging should take into account the rate of return on funds not
invested in stocks. We assume two assets, safe Treasury bills and a risky stock. We assume that
the return R on Treasury bills is constant and that the stochastic stock return St in period t is
independent of the stock’s return in other periods and has a constant mean 𝜇 and standard
deviation 𝜎, with the stock’s expected return larger than the return on T-bills. DCA does not
!9
benefit from mean reversion in our model, and LS has the built-in advantage of a stock expected
return that is higher than the return on Treasury bills. Nonetheless, DCA has some appealing
features.
The LS strategy involves an immediate investment of all wealth in stocks. The DCA strategy
is to make regular, periodic constant stock purchases over an investment horizon of T periods.
Money not initially invested in stocks is parked in Treasury bills.
Mean-variance analysis is usually applied to a decision on how to allocate wealth among a
portfolio of different stocks or asset classes. It can be applied to dollar cost averaging by
considering each periodic investment in stocks as an initial investment in Treasury bills that is
converted into stocks at the appropriate time.
Specifically, one way to frame the timing question is to identify the delayed purchase of stock
when there are j periods left in the investment horizon as an asset that yields (1 + R) for each of
the first T - j periods and (1 + St) for the remaining j periods. The gross return over the full T-
period horizon is
! j = 0, …, T (1)
Z0 is the gross return on an investment in Treasury bills for all T periods; Z1 is the return on an
investment in Treasury bills for T - 1 periods, followed by a stock investment in the last period;
and so on. A decision to spread stock purchases over a T-period horizon is a portfolio allocation
of wealth among the T assets j = 1, 2, … T. The overall return is
! (2)
where a fraction 𝜆j of wealth is invested in asset j. For each value of 1 j T, 𝜆j is invested now
Zj=(1+R)Tj(1 +St)
t=T+1j
T
Z=
λ
jZj
j=1
T
!10
in the safe asset and, after T - j periods, 𝜆j(1 + R)T-j is invested in stocks. Dollar cost averaging
corresponds to
! j = 1, …, T (3)
where C is set so that the sum of the 𝜆j is 1. For the lump-sum strategy, 𝜆T = 1, and 𝜆j = 0 for 1
j T.
If the stock returns are independent across time, then a standard mean-variance portfolio
analysis can be conducted using
! 1 j T (4)
Since asset 0 is a safe asset with return (1 + R)T, Tobin’s (1957) separation theorem applies, and
we can identify the optimal ratios 𝜆j/𝜆T for j 0, regardless of risk preferences. These can be
rescaled to give the optimal fractions 𝛾j of the risky portfolio:
! 1 j T
The optimal proportions depend on the volatility of stock prices—the more volatility, the
stronger is the case for time diversification. The time-diversification argument is consequently
more persuasive for individual stocks than for stock indexes. We assume a constant 3 percent
annual return on Treasury bills and a 7 percent expected return on stocks, and use different
values for the standard deviation.
EZj
=(1+R)Tj(1 +
µ
)j
Var Zj
=(1+R)2(Tj)(1+
µ
)2+
σ
2
( )
j(1+
µ
)2j
( )
Cov ZiZj
=(1+R)Ti+Tj(1 +
µ
)ji(1+
µ
)2+
σ
2
( )
i(1+
µ
)2i
( )
γ
j=
λ
j/
λ
T
λ
j/
λ
T
( )
j=1
T
!11
The average annual return on Treasury bills since 1926 has been approximately 3 percent and
the average annual return on large-cap stocks has been 10 percent (Ibbotson, Grabowski,
Harrington, and Nunes 2016). However, a study of the twentieth-century performance of the
stock markets in 39 different countries found that the U. S. stock market beat all the rest (Jorian
and Goetzmann 1999). It seems unlikely that investors worldwide knew that the twentieth
century would turn out to be America’s century and the U. S. stock market would turn out to be
the winner. It is more likely that the remarkable performance of the U. S. stock market was a
pleasant surprise to investors who owned U. S. stocks—and we can hardly count on pleasant
surprises indefinitely.
The equity premium seems to have declined over time and the ex ante premium is likely to be
significantly smaller in the future than the ex post premium has been in the past. Fama and
French (2002) estimate a risk premium of 2.55 to 4.32 percent for 1951 to 2000 relative to short-
term risk-free bonds. Siegel (1999) suggests it might be as low as 1 percent to 2 percent going
forward. We use 4 percent as our baseline premium (a 7 percent expected return on stocks
compared to a 3 percent return on safe Treasury securities), but also consider 2 percent and 6
percent premiums.
Table 3 shows the means, standard deviations, and correlation coefficients for annual
investments over a five-year horizon, in accordance with Equation 1, and the optimal allocation
in the risky portfolio using Tobin’s Separation theorem. The fifth asset is an initial investment in
stocks and, because its return is the least correlated with the return on the first asset (four years
of Treasury bills, followed by a stock investment), the first asset has the second largest optimal
allocation. The spread of investments across all five assets is much more dispersed than is the LS
!12
all-in strategy, which sets 𝛾5 = 1 and the other 𝛾j = 0.
Table 3 Mean-Variance Analysis with a 5-Year Horizon and R = 3%, 𝜇 = 7%, and 𝜎 = 50%
correlation coefficients optimal
Asset Mean, % SD, % 1 2 3 4 5 𝛾j
1 20.43 56.27 1.00 .67 .52 .42 .36 .229
2 25.11 87.07 1.00 .77 .63 .54 .174
3 29.96 116.86 1.00 .87 .69 .132
4 35.01 148.11 1.00 .85 .101
5 40.26 182.04 1.00 .365
Figure 1 shows the percentage of the risky portfolio invested initially in stocks for different
investment horizons and 30 percent and 50 percent annual standard deviations of the stock
return. Time diversification is most attractive for less correlated alternatives, and the correlations
among the alternatives drop for riskier stock and longer horizons. In Figure 1, as the investment
horizon lengthens or the standard deviation increases, the initial investment in stocks approaches
that recommended by dollar cost averaging.
Figure 2 shows the Markowitz frontier using the assumptions in Table 3: annual investments
over a five-year horizon with a 50 percent annual standard deviation in the stock return. The
DCA portfolio is very close to the optimal portfolio implied by Tobin’s separation theorem.
!13
!
Figure 1 Fraction Initially Invested in Stocks, annual investments
!
Figure 2 Markowitz Frontier, annual investments over a 5-year horizon with 𝜎 = 50%
Figure 3 is the analogue of Figure 1, but now with monthly investments over horizons up to
60 months and with 20 percent and 30 percent standard deviations of stock returns. Again, a
0
10
20
30
40
50
60
70
80
90
100
012345678910
Initial stock investment, percent
Horizon, years
LS
Optimal
σ = 30
DCA Optimal
σ = 50
0
5
10
15
20
25
30
35
40
45
0 20 40 60 80 100 120 140 160 180 200
Mean, percent
Standard deviation, percent
Optimal
DCA
LS
!14
longer horizon or standard deviation supports the DCA idea of diversifying stock purchases
across time. Figure 4 shows the corresponding Markowitz frontier for monthly investment over a
60-month horizon with a 20 percent standard deviation. Again, the DCA portfolio is close to the
optimal portfolio implied by Tobin’s separation theorem.
!
Figure 3 Fraction Initially Invested in Stocks, monthly investments
One might think that a higher equity premium would pull the optimal portfolio towards a
larger initial investment in stocks, because it increases the opportunity cost of investing in
Treasury bills. However, the equity premium has little effect on the curvature of the Markowitz
frontier and, looking at Figures 2 and 4, an upward shift in the Markowitz frontier slides the
optimal portfolio away from LS. Table 4 confirms this. As the equity premium increases, the
optimal initial investment in stocks moves away from the 100 percent figure used by LS towards
the 21.2 percent figure used by DCA.
0
10
20
30
40
50
60
70
80
90
100
0 10 20 30 40 50 60
Initial stock investment, percent
Horizon, months
LS
DCA
Optimal
= 10
Optimal
= 20
!15
!
Figure 4 Markowitz Frontier, monthly investments over a 5-year horizon with 𝜎 = 20%
Table 4 Initial Stock Investments with a 5-Year Horizon and R = 3%
Optimal Portfolio
LS DCA 𝜎 = 30% 𝜎 = 50%
𝜇 = 5% 1.00 .212 .665 .396
𝜇 = 7% 1.00 .212 .611 .365
𝜇 = 9% 1.00 .212 .558 .335
The time-diversification value of DCA depends on the investor’s assumptions, with a longer
horizon and larger standard deviation of stock returns moving the optional portfolio away from
LS and closer to DCA. Dollar cost averaging is obviously not always a reasonable approximation
to the optimal portfolio implied by Tobin’s separation theorem. Our point is simply that Bogle,
Malkiel, Tobias, and other wise and experienced investors are neither naïve or foolish. When
contemplating an especially risky investment, there is merit in time diversification even if the
0
5
10
15
20
25
30
35
40
45
0 50 100 150 200 250 300 350 400 450
Mean
Standard Deviation
Optimal
DCA
LS
!16
stock returns are independent with a higher expected return than Treasury bills.
Empirical Evidence
We used historical data to compare the DCA and LS strategies. Even though ex post returns may
be an unreliable guide for ex ante decisions, it seems worthwhile to consider how these strategies
would have fared in the past.
We looked at the 84 stocks that have been in the Dow Jones Industrial Average from October
1, 1928, when the Dow was expanded from 20 to 30 stocks, through December 31, 2016, a total
of 23,219 trading days. These are all prominent companies, widely followed by investors, with
reliable data, though the fluctuations in the daily returns are presumably smaller than those for
many stocks that might be DCA candidates. The daily returns on Treasury bills and these 84
stocks were taken from the Center for Research in Security Prices (CRSP) data base.
To avoid biases that might arise because stocks added to the Dow generally did well before
their inclusion, we only looked at purchases of Dow stocks that were made while the stocks were
in the Dow. In order to avoid distortions that might be caused by seasonal or day-of-the-month
patterns, our portfolio simulations used every Dow stock and every starting date; for example,
IBM on April 18, 1978. When occasional gaps in the CRSP data base of returns arise, we assume
that the portfolio temporarily earns the Treasury-bill return. All initial purchases were made on
dates such that the investment horizon ended on or before December 31, 2016.
The LS strategy purchases the selected stock on the selected day and holds it until the end of
the investment horizon. The DCA strategy uses the asset allocation described by Equation 3 to
invest in the T assets that make up the DCA strategy. A specified fraction of the portfolio is
invested in the stock initially, with the remainder parked in Treasury bills. Each successive
!17
period, an additional investment (equal to the initial investment) is made in the stock. After
looking at Figures 1 and 3, we also considered a 50-50 strategy of investing 50 percent of wealth
immediately in stocks and spreading the remainder equally over the horizon.
We considered two horizons consistent with the earlier theoretical analysis—annual purchases
for five years and monthly purchases for five years. There are approximately 250 trading days in
a year, so we assumed that the annual purchases were made every 250 trading days after the
initial purchase and that the monthly purchases consisted of purchases made every 20 trading
days after the initial purchase.
Table 5 shows the results. As expected, the LS strategy had the highest average return and the
highest standard deviation, and the DCA strategy had the lowest values, with the 50-50 strategy
in between. The LS strategy had the lowest Sharpe ratios, while DCA and 50-50 were very
similar. Historically, an investment in Treasury bills and either DCA or 50-50 would have
dominated an equally risky investment in Treasury bills and LS.
Table 5 Simulated of Purchases of Dow Stocks Over a Five-Year Horizon, 1928-2016
Average Annual Return, % Standard Deviation, % Sharpe Ratio
Annual Purchases
LS 73.63 106.54 0.41
DCA 51.76 60.55 0.49
50-50 59.63 74.48 0.50
Monthly Purchases
LS 69.73 101.86 0.40
DCA 44.32 50.36 0.47
50-50 56.55 71.76 0.49
!18
Conclusion
Dollar cost averaging is not as foolish as it is sometimes portrayed. It is well known that its
mechanical nature may encourage saving and reduce the emotional anxiety associated with
making decisions. It is not well known that cost averaging can also be a valuable way of
diversifying investment decisions across time, which is particularly appealing when investing in
volatile stocks over a substantial horizon. Dollar cost averaging is not always optimal but, in
some circumstances, it may be a reasonable approximation.!
!19
References
Abeysekera, Sarath R., and E. S. Rosenbloom. 2000. A Simulation Model for Deciding Between
Lump-Sum and Dollar-Cost Averaging. Journal of Financial Planning 13 (6): 86–92.
Atra, Robert J. and Thomas L. Mann. 2001. Dollar-Cost Averaging and Seasonality: Some
International Evidence. Journal of Financial Planning, 14 (7): 98-105.
Balvers, R. J., & Mitchell, D. W. (2000). Efficient Gradualism in Intertemporal Portfolios.
Journal Of Economic Dynamics And Control, 24(1), 21-38.
Bierman, Jr., Harold, and Jerome E. Hass. 2004. Dollar-Cost Averaging. Journal of Investing, 13
(4): 21–24.
Bogle, J. C. 2015, Bogle on Mutual Funds, New York: Wiley.
Brennan, Michael J., Li, Feifei, and Walter N. Torous. 2005. Dollar Cost Averaging. Review of
Finance 9 (4): 509–35.
Cho, David D., and Emre Kuvvet. 2015. Dollar-Cost Averaging: The Trade-Off Between Risk
and Return. Journal of Financial Planning 28 (10): 52–58.
Constantinides, G.M. 1979. A Note On The Suboptimality Of Dollar-Cost Averaging As An
Investment Policy. Journal of Financial and Quantitative Analysis, 14 (2), 443-450.
Damato, Karen Slater, 1994, Long-term Investors Can Reap Rewards From Buying Stocks
During Bear Markets, The Wall Street Journal, April 20.
Dichtl, H., & Drobetz, W. 2011. Dollar-Cost Averaging and Prospect Theory Investors: An
Explanation for a Popular Investment Strategy. Journal Of Behavioral Finance, 12 (1),
41-52.
Dreman, David. 1982. New Contrarian Investment Strategy. New York: Random House.
!20
Dubil, R. 2005a. Lifetime Dollar-Cost Averaging: Forget Cost Savings, Think Risk Reduction.
Journal of Financial Planning, 18 (10), 86-90.
Dubil, Robert. 2005b. “Investment Averaging: A Risk-Reducing Strategy.” Journal of Wealth
Management 7 (4): 35–42.
Dunham, Lee M., and Geoffrey C. Friesen. 2012. Building a Better Mousetrap: Enhanced Dollar-
Cost Averaging. Journal of Wealth Management 15 (1): 41–50.
Fama, Eugene F. and Kenneth R. French 1988, Permanent and Temporary Components of Stock
Prices, Journal of Political Economy, 96 (2), 246-273.
Fama, Eugene, and Kenneth French, 2002. The Equity Premium, The Journal of Finance, 57 (2)
Grable, John E., and Swarn Chatterjee. 2015. Another Look at Lump-Sum versus Dollar-Cost
Averaging. Journal of Financial Service Professionals, 69 (5): 16–18.
Haley, Simon, 2010. Explaining the Riddle of Dollar Cost Averaging, Cass Business School.
Ibbotson, Roger, Grabowski, Roger J., Harrington, James P., and Carla Nunes, 2016. 2016
Stocks, Bonds, Bills, and Inflation (SBBI) Yearbook, New York: Wiley.
Investment Company Institute. 1984. Discipline: It Can’t really be Good for You, Can It?
Jorian, Philippe, and William N. Goetzmann, 1999, Global Stock Markets in the Twentieth
Century, The Journal of Finance, 54 (3),953-980.
Knight, J.R. and Mandell, L. 1992/1993. Nobody Gains From Dollar Cost Averaging: Analytical,
Numerical And Empirical Results. Financial Services Review, 2 (1), 51-61.
Leggio, Karyl B., and Donald Lien. 2001. Does Loss Aversion Explain Dollar-Cost Averaging?
Financial Services Review 10 (1-4): 117.
Leggio, Karyl B., and Donald Lien, 2003a. Comparing Alternative Investment Strategies Using
!21
Risk-Adjusted Performance Measures. Journal of Financial Planning, 16 (1), 82-86.
Leggio, Karyl B., and Donald Lien, 2003b. An Empirical Examination of the Effectiveness of
Dollar-Cost Averaging Using Downside Risk Performance Measures. Journal Of Economics
And Finance, 27 (2), 211-223.
Lei, Adam Y. C., and Huihua Li. 2007. Automatic Investment Plans: Realized Returns and
Shortfall Probabilities. Financial Services Review 16 (3), 183–95.
Levy, H. and M. Sarnat. 1972, Investment and Portfolio Analysis. New York, NY: John Wiley
and Sons, 244-248.
Loeb, M. Marshall, 1996, Loeb’s Lifetime Financial Strategies. 1996. Boston: Little, Brown, 68–
70.
Luskin, Jon M. 2017. Dollar-Cost Averaging Using the CAPE Ratio: An Identifiable Trend
Influencing Outperformance. Journal of Financial Planning 30 (1): 54–60.
Malkiel, Burton G., 2015, A Random Walk Down Wall Street, New York: W. W. Norton
Milevsky, M.A. and Posner, S.E. 2003. A Continuous-Time Re-examination of the Inefficiency
of Dollar Cost Averaging. International Journal of Theoretical & Applied Finance, 6 (2),
173-194.
Mitzenmacher, Michael and Eli Upfal, 2005, Probability and computing: randomized algorithms
and probabilistic analysis, Cambridge University Press, p. 298.
Paglia, John K., and Xiaoyang Jiang. 2006. Implementing a Dollar-Cost Averaging Investment
Strategy: Does the Date of the Month Matter?, Journal of Wealth Management 9 (2): 54–62.
Porter, Sylvia. 1979. Sylvia Porter’s New Money Book for the 80’s. New York: Doubleday, 1012.
Poterba, James M. and Lawrence H. Summers, 1988, Mean Reversion in Stock Prices, Journal of
!22
Financial Economics, 22, 27-59.
Rozeff, M.S. 1994. Lump-sum Investing Versus Dollar-Averaging. Journal of Portfolio
Management, 20 (2), 45-50.
Sharpe, W.F. 1978. Major Investment Styles, The Journal of Portfolio Management, 5, 68-74.
Siegel, Jeremy J. 1999. The Shrinking Equity Premium, The Journal of Portfolio Management,
26 (1), 10-17. as low as 1% to 2%
Statman, M. 1995, A Behavioral Framework For Dollar-Cost Averaging. Journal of Portfolio
Management, 22 (1), 70-78.
Thorley, S.R. 1994. The Fallacy of Dollar Cost Averaging. Financial Practice and Education, 4
(2), 138-143.
Tobias, Andrew. 2016. The Only Investment Guide You’ll Even Need. New York: Houghton
Mifflin Harcourt.
Tobin, James. 1958. Liquidity preference as behavior towards risk. The Review of Economic
Studies. 25 (2): 65–86.
Tomlinson, L. 1947. Successful Investing Formulas, New York: Barron’s.
Tomlinson, L. 2012. Successful Investing Formulas, Snowball Publishing.
Trainor, William J., Jr. 2005. Within-horizon exposure to loss for dollar cost averaging and lump
sum investing. Financial Services Review, 14 (4), 319-330.
Tversky, Amos; Kahneman, Daniel. 1992. Advances in prospect theory: Cumulative
representation of uncertainty. Journal of Risk and Uncertainty. 5 (4): 297–323.
Williams, R.E. and Bacon, P.W. 1993. Lump-sum Beats Dollar Cost Averaging. Journal of
Financial Planning, 6 (2), 64–67.
... Moreover, using mutual fund data, Israelsen (1999) argued that lump-sum investing does not always yield superior returns over dollar-cost averaging, especially if the volatility of mutual funds is low. For additional literature on DCA and LS investing, the reader is referred to Pye (1971); Dodson (1989); Thorley (1994); Leggio and Lien (2003); Milevsky and Posner (2003); Smith and Artigue (2018) and references therein. ...
Article
Full-text available
In this paper we present new theoretical and practical insights into the method of dollar cost averaging (DCA) and averaging-style investment timing strategies, with a formal analysis of the problem. Firstly, we provide a rigorous mathematical formulation for studying DCA and related strategies. This provides closed form formulae for the expected value and variance of the investor's wealth process, which mathematically proves many properties that have been documented in the literature only by empirical studies. Secondly, we prove a counterintuitive, but important, result that the frequency of DCA investment has a non-monotonic and non-trivial impact on risk, risk-return trade-off and other important performance metrics (such as the Sharpe ratio).Thirdly, we provide a method of valuing the DCA risk for models which incorporate jumps. We also provide a method of hedging DCA risk based on applying Asian options. Finally, using the PROJ method of computation, we obtain a robust and computationally efficient method for calculating standard risk measures of generic and deterministic investment strategies, such as DCA. We provide numerical experiments to illustrate our conclusions, and conduct an empirical study on the S&P500 index (from 1954 to 2019) to substantiate our results.
... Moreover, using mutual fund data, Israelsen (1999) argued that lump-sum investing does not always yield superior returns over dollar-cost averaging, especially if the volatility of mutual funds is low. For additional literature on DCA and LS investing, the reader is referred to Pye (1971); Dodson (1989); Thorley (1994); Leggio and Lien (2003); Milevsky and Posner (2003); Smith and Artigue (2018) and references therein. ...
Preprint
The Dollar-Cost Averaging (DCA) Strategy is an enigma. Proven sub-optimal from a risk-adjusted performance time and again since the late 1970's, it is nevertheless more popular today than ever. Our empirical analysis makes no exception. The DCA Strategy does almost systematically show a lower level of volatility than the so-called Lump Sum Investing (LSI) Strategy, but there is no free lunch. The price to pay is a significantly lower level of return, leading more often than not to lower Sharpe ratios. And, the greater the expected return of the underlying asset, the higher the opportunity cost to adopt the DCA as opposed to the LSI Strategy. Yet, the DCA Strategy has its merits. It does indeed systematically lead to a smaller dispersion of the final outcomes, which may reassure the most risk-averse investors. Another benefit of the DCA Strategy is to introduce discipline in the investment process, at least in the timing and regularity of the investments, which may prevent investors from giving free reins to their "animal spirits". One can even go one step further and argue that the DCA Strategy paves the way of least resistance for the man in the street to save money and build up an estate. This being said, chances are that cognitive and behavioral biases only play a mediating role in the current popularity of the DCA Strategy. Recent market developments shed a new light on the DCA Strategy, and suggest that the liquidity profile of its order flow could very well be the key driving factor of its commercial success. In a market sorely missing depth, the DCA Strategy is indeed a great liquidity provision strategy for the retail brokers and/or the wholesalers. It is therefore promoted aggressively to retail investors. The collateral effect is that retail investors looking for an efficient way to build up positions in risky assets securely, could finally end up with more risks than they think/feel, and possibly than they can really stand.
Book
Burton Malkiel’s 1973 A Random Walk Down Wall Street was an explosive contribution to debates about how to reap a good return on investing in stocks and shares. Reissued and updated many times since, Malkiel’s text remains an indispensable contribution to the world of investment strategy - one that continues to cause controversy among investment professionals today. At the book’s heart lies a simple question of evaluation: just how successful are investment experts? The financial world was, and is, full of people who claim to have the knowledge and expertise to outperform the markets, and produce larger gains for investors as a result of their knowledge. But how successful, Malkiel asked, are they really? Via careful evaluations of performance - looking at those who invested via ‘technical analysis’ and ‘fundamental analysis’ - he was able to challenge the adequacy of many of the claims made for analysts’ success. Malkiel found the major active investment strategies to be significantly flawed. Where actively managed funds posted big gains one year, they seemingly inevitably posted below average gains in succeeding years. By evaluating the figures over the medium and long term, indeed, Malkiel discovered that actively-managed funds did far worse on average than those that passively followed the general market index. Though many investment professionals still argue against Malkiel’s influential findings, his exploration of the strengths and weaknesses of the argument for believing investors’ claims provides strong evidence that his own passive strategy wins out overall.
Article
We estimate the equity premium using dividend and earnings growth rates to measure the expected rate of capital gain. Our estimates for 1951 to 2000, 2.55 percent and4.32 percent, are much lower than the equity premium produced by the average stock return,7.43 percent. Our evidence suggests that the high average return for 1951 to 2000 is due to a decline in discount rates that produces a large unexpected capital gain. Our main conclusion is that the average stock return of the last half-century is a lot higher than expected.
Article
The author argues that the equity premium or the historical spread between expected equity returns and government bond yields is probably far below the approximately 6% figure estimated in much of the finance literature. This, the author contends, is due both to an underestimate of the expected real return on the risk-free asset and an overestimate of the realized returns on equities. Correction of these biases reduces the equity premium to 1% to 2% per year. Furthermore, given the current high level of equity prices relative to earnings and high yield on government price-indexed bonds, it is extremely unlikely that the future premium will exceed the 1% to 2% range without an unprecedented increase in earnings growth.