View
225
Download
6
Category
Preview:
Citation preview
International Review of Financial Analysis 14 (2005) 570–586
Exploratory analyses of dividend reinvestment plans
and some comparisons
Kevin Chianga, George M. Frankfurterb,*, Arman Kosedagc
aUniversity of Alaska at Fairbanks, United StatesbEmeritus, Louisiana State University, 60 Sagris Cove, Miramar Beach, FL 32550-3842, United States
cSabanci University, Turkey
Available online 24 November 2004
Abstract
In this paper, we explore salient features of dividend reinvestment plans (DRIPs), analyze their
financial peculiarities and search for the differences between firms that offer DRIPs and those that do
not. As more than 1200 firms currently offer the plan, an understanding of why these plans differ in a
variety of cost/benefit structures and, perhaps more importantly, what separates these firms from No-
DRIP firms is crucial for both investors and adaptors of the plan. Our research suggests that—out of
17 financial and accounting variables—DRIP firms differ from No-DRIP firms in only three
variables. In spite of this, we conclude that there is much to learn about the motivation for DRIPs.
D 2004 Elsevier Inc. All rights reserved.
JEL classification: G30; G32; G35
Keywords: DRIP; Rights offering; Capital structure; Dividend; Growth
1. Introduction
The purpose of this paper and the research it proposes is to expand the corporate,
academic finance literature dealing with the effects of dividend reinvestment plans (DRIPs)
on the dividend policy formulation of firms. In recent years, DRIPs have becomemuchmore
prevalent—both as an investment strategy or tool for the individual investor and as a source
1057-5219/$ -
doi:10.1016/j.i
* Correspon
E-mail add
see front matter D 2004 Elsevier Inc. All rights reserved.
rfa.2004.10.020
ding author. Tel./fax: +1 850 654 5250.
ress: pitypalaty@cox.net (G.M. Frankfurter).
K. Chiang et al. / International Review of Financial Analysis 14 (2005) 570–586 571
of new equity for corporations—since their emergence in the late 1960s and are currently
offered by approximately 1200 firms. According to Baker and Seippel (1981) and others,
firms view these programs as a means of raising equity capital and improving shareholder
goodwill, and investors see them as a means of dollar-cost averaging.
DRIPs are set-up to allow current shareholders to purchase additional shares by buying
directly from the firm and thereby bypass the broker and his associated fees (Carlson,
1992). In addition, approximately 50 companies now allow open enrollment in their
DRIPs. Through these plans, investors can make initial investments and all subsequent
investments without paying broker fees (Burns, 1994). Although the plans appear to be
beneficial to both the shareholders and the firm, there are some problems that could be
considered deterrents. As Schneid (1981) points out, one such problem is that bplanterminations and arbitrage trading may tend to exert a negative impact on market price.Q
Because of the popularity of DRIPs with individual investors and personal investment
analysts, most of the information about DRIPs has been in the laymen finance literature.
These articles focus of course on the relationship of DRIPs to personal finance issues
rather than corporate or academic finance issues and implications. Furthermore, most of
the corporate and academic literature on DRIPs concentrates on their association with
utilities. Consequently, it pays scant attention, if any, to other industries (predominantly the
financial services industry, which now accounts for the largest percentage of DRIPs
offered), much less the entire universe of DRIPs.
Despite their apparent advantages, not all firms offer DRIPs. This seems to suggest that
firm-specific factors might be the driving force for the introduction of such plans, and
perhaps, that firm characteristics may be instrumental in the specifics of the plan offered.
Because of the thinness of the academic literature, there is a void that calls to be filled by
exploring these issues of specificity and characteristics. Accordingly, the analyses of how
firms that offer DRIPs differ from firms that do not is the objective of this paper and our
research. This objective necessarily involves the study of DRIP characteristics as well.
Naturally, our research agenda is exploratory because the lack of well-justified hypotheses
precludes it from being confirmatory.
Section 1 is a brief overview of DRIPs. In Section 2, we review the literature. Section 3
is the statistical part of the paper where we discuss the data, explain the variables of the
study, describe the models we use and the results of our tests. Finally, in Section 4, we
suggest avenues for future research and also offer some brief conclusion.
1.1. DRIPs: an overview
DRIPs are by no means new. Mutual funds and closed-end investment companies have
been providing DRIPs for their shareholders since the early 1940s (Finnerty, 1989). Lehman
Corporation, an investment company, was the first firm to offer market DRIPs.1 DRIPs were
first made available to noninvestment companies through an SEC regulation revision in
1968. Allegheny Power Systems was the first industrial corporation to implement a DRIP.
1 Davey (1976), and Saporoschenko (1996) examine three types of DRIPs: (1) market DRIPs, which are set-up
as a service to shareholders and provide no bnewQ cash to the firm; (2) treasury DRIPs, which are designed to sell
shares from treasury stock; and (3) combination DRIPs (market and treasury).
K. Chiang et al. / International Review of Financial Analysis 14 (2005) 570–586572
Before 1972, all DRIPs were open-market plans. Long Island Lighting offered the first
new-issue DRIP in 1972 (Finnerty, 1989), and AT&T was the first firm to implement a
new-issue DRIP that offered a discount to its investors (Saporoschenko, 1996). As we
discuss in greater detail later in this paper, the majority of DRIPs are offered in the
financial and utility industries.
Cherin and Hanson (1995) identify the following factors that drive the creation of
DRIPs by firms that offer the product:
! Stabilization of the stockholder base. Corporations wishing to avoid excessive
institutional ownership believe that a DRIP will attract and hold small, individual
shareholders, keeping control in the hands of management.
! Greater market price support. Sponsoring firms may believe that price volatility caused
by institutional trading can be moderated through DRIPs, and that quarterly lump sum
purchases to meet plan demand can be beneficial to the firm in that they place upward
pressure on the stock’s market price, providing a degree of price support for the stock.
! New equity capital. By using authorized but previously unissued shares in connection
with the plan, the firm taps a reservoir of new capital, obtaining regular cash infusions,
thereby reducing dependence on other external sources. The boriginal issueQ or bnewcapitalQ DRIP appeals particularly to capital-intensive businesses, like utilities, and the
trend is likely to continue. Modest amounts are raised every year that not only avoid large
periodic offerings and their adverse signaling effects but also allow, as some claim, new
capital to be used more quickly, promoting earnings stability. In addition, by providing
equity inflows, these plans restore balance to debt-burdened capital structures.
! Improved corporate liquidity.
! Lower financing costs. Because investment banker underwriting fees and underpricing
are bypassed and because fewer dividend checks and stock certificates must be
prepared and mailed, the net origination and administrative costs of equity financing are
significantly reduced as the firm uses a bnew-issueQ DRIP. Even rights offering cannot
compete. One estimate of the cost of raising equity capital through a DRIP versus a
normal underwriting is that the plan cost is between 2% and 3%, whereas investment
bankers require a 3% to 5% spread.
! Better shareholder relations. The convenience and other perceived investor benefits of
a DRIP would reap the goodwill of the generally small, individual shareholder.
Presumably, greater shareholder loyalty may reduce the likelihood of an unwanted
takeover. It seems that improved relations with shareholders weigh heavily in planning
the initiation of a DRIP.
! Appeal to shareholders. The plan attracts new stockholders who like DRIPs.
Cherin and Hanson (1995) remark, however, that from the firm’s standpoint, DRIPs
have several disadvantages:
! Dilution. Dilution of earnings per share may be a problem with bnew-issueQ DRIPs.! Wounded investor relations. A DRIP might impair investor relations by causing
shareholders to flirt with the illiquidity of not receiving needed cash dividends. In
addition, a plan may be seen as geared to only those investors with longer term
K. Chiang et al. / International Review of Financial Analysis 14 (2005) 570–586 573
investment, creating conflict between participants and nonparticipants. Moreover, plans
may invite legal headaches with the SEC and others.2 Injuring shareholder relations
through clerical/record-keeping errors is always a possibility as well.
! Costs. The plan start-up and maintenance costs outweigh the advantages. In addition,
nonparticipants are being discriminated against in the form of expenditures on the plan.
The start-up costs for a DRIP can be high. And there are annual expenses associated
with administering the DRIP. Furthermore, there are costs associated with educating
shareholders when initiating a DRIP.
! Participation. Too few stockholders could enroll to make a DRIP worthwhile. And,
perhaps more problematic, there is little control over the extent of participation while an
boriginal issueQ plan is in force without resorting to plan alterations and/or changes in
dollar dividends declared. In other words, a free cash flow problem may exist or new
equity capital may be insufficient, requiring ongoing plan modification.
! Company ownership. To maintain the firm’s shareholder profile, especially if it is
institutionally dominated, a DRIP could be inappropriate.
Because they allow for dollar-cost averaging and the avoidance of brokers’ fees, DRIPs
are touted as a profitable investment strategy for the individual investor. The potential
investor must carefully evaluate the programs, however, as they may differ on several
features, among them are fees that are charged, discounts offered, optional cash payment
availability, enrollment requirements, availability of safekeeping and management of
the plan.
! Fees charged. Although most firms charge no commissions for purchasing stocks
through their DRIPs, many assess the investor one sort of service fee or another. These
fees are classified as dividend reinvestment, optional cash investment, certificate
issuance, sale of shares or termination of account. The fees for dividend reinvestment
are usually set-up either as a certain percentage of the reinvestment amount or as a fixed
dollar amount. The usual percentage ranges from 3% to 10%. The most common dollar
amount is $3 to $5. The optional cash investment fee ranges from 0.025% per share to
$2.50 per share, plus broker’s commission. The most common range for the fees on
certificate issuance is $5 to $15. The sale fee may be either a broker’s commission or a
fixed amount that can range from $0.03 to $15 per share. Charges for the termination of
an account can range from $2.50 to $25 plus broker’s commission.
! Discounts offered. Some firms may offer discounts to investors. These discounts can be
granted either on dividend reinvestment or on optional cash investments. Although
these discounts can vary from 1% to 10%, the most common discount is 3% to 5%.
Strangely enough, quite a few firms do both: offer discounts and charge some fees.
! Optional cash payment (OCP) availability. The availability of OCPs is one of the major
benefits emphasized in case of a DRIP that offers such possibility. Firms allowing for
OCP either will automatically withdraw a fixed amount of money from the investor’s
checking or savings account to make optional cash investments or will accept check
2 DRIPs involve SEC registration.
K. Chiang et al. / International Review of Financial Analysis 14 (2005) 570–586574
payments to the firm. The amounts allowed for an OCP usually range from $10 to
$150,000 per investment period, ranging from weekly to yearly.
! Enrollment requirements. The plan can be set-up to offer shares either to current
investors exclusively or to anyone at all. Some of the firms that allow all investors to
enter their DRIPs include AFLAC, Dial, Exxon, Johnson Controls, Texaco and Mobil.
For these no-load stocks, as they are commonly called, there might be a minimum
investment requirement that can range from $50 to $1000. DRIPs that allow current
investors only may require the investor to hold before enrollment anywhere from 1 to
1000 shares. Whether the firm allows for foreign acceptance simply indicates whether
the firm accepts investments from foreign investors.
! Availability of safekeeping. The safekeeping characteristic signifies whether the firm
accepts certificated shares and holds them in safekeeping for the investors. Sometimes,
the shares are deposited in the account safekeeping for the investor.
! Management of the plan. DRIPs can be administered either by the issuing firm or by an
agent. Only about 12.5% of all firms offering DRIPs administer their own plan. In the rest
of the firms, an agent administers the DRIPs. Some of the most commonly used agents are
The Bank of Boston, First Chicago Trust, RM Trust and The Bank of New York.
Part of our analyses in Section 3 is built on these plan characteristics, which in turn we
translate to observable and quantifiable variables.
2. Literature review
Because of the popularity of DRIPs with individual investors and personal investment
analysts, most of the information published about DRIPs has been in laymen’s finance
literature. Although the popular finance literature on DRIPs is too voluminous to mention,
a few articles merit special notice. Most of these articles are geared of course toward the
individual investor and not toward the firm. Business Week (April 14, 1975) points out that
DRIPs are a method by which firms can try to raise their P/E ratio. Goodman (1993)
advises investors to use DRIPs as a method of bbuilding your assets on autopilot.Q Money
(1994) touts DRIPs as a bsurefire approach to beating the market.Q And according to
Landy (1996), DRIPs provide a simple means of raising capital from existing
shareholders.
Several books pander information on DRIPs. Among them are The moneypaper’s guide
(1996) to dividend reinvestment plans, No-load stocks (Carlson, 1995) and Buying stock
without a broker (Carlson, 1992, 1996).
The academic literature on DRIPs is slight because its thrust, in one way or another, is
to fit the phenomenon into the straight-jacket of one or another cherished or currently in
vogue dividend theory. A recent paper by Bierman (1997), however, lumps DRIPs
together with the bdividend puzzle,Q using the idiom as it was originally intended by Black
(1976).
Most of the DRIPs studies focus on the effects of a DRIPs-related announcement on
shareholder wealth. There is also literature related to DRIPs on equity offerings with
noninsured rights offers being analogous and sometimes offered as a substitute to DRIPs.
K. Chiang et al. / International Review of Financial Analysis 14 (2005) 570–586 575
A convenient taxonomy of the academic literature puts DRIPs in three categories: (A)
event studies; (B) raising new capital/rights offerings; and (3) agency costs/problems.
(A) Event studies
Although most DRIPs research falls into this category, there are a multitude of
problems associated with these studies. One of the most prevalent problems is the
pinpointing of the exact announcement date for the DRIP. In addition, event studies
are very bfragileQ in the sense that it is not uncommon to find contradictory evidence,
given the model from which babnormal returnsQ are derived. Therefore, their resultsshould not be taken lightly or unheeded. Nevertheless, because these studies
constitute such a large portion of the DRIPs literature, we would be remiss not to
report here about their existence and their findings.
Peterson, Peterson and Moore (1987) examine a 1-day window surrounding SEC
filings for new shares being issued by utilities before May 1981 by utilities after July
1981—which received preferential tax treatment (Internal Revenue Code Section
305(e) created tax deferral for DRIPs of qualifying utilities)—and by nonutilities.
They find that utilities adopting DRIPs before the tax law change experienced a
negative reaction on the day after the plan was registered, and that utilities adopting
DRIPs after July 1981 experienced positive abnormal returns significantly greater
than those without preferential treatment. They find that nonutility DRIPs, however,
realized insignificant average abnormal returns.
(B) Chang and Nichols (1992) also examine the effect of the 1981 tax legislation and
find a positive announcement effect for the qualifying utility firms. In comparing the
pre and posttax deferral periods to the 1982–1985 period in which tax deferral was
in place, Chang and Nichols find that there was an increase in DRIP participation for
qualifying utility firms but no increase for the nonqualifying utilities.
Dubofsky and Bierman (1988) examine a 3-day window around the announce-
ment of discount DRIPs. They find significant positive abnormal returns. Dubofsky
and Bierman (ibid.) argue, bthe positive excess returns are consistent with the
hypotheses that the sample firms are moving toward more optimal (less risky) capital
structures, that the internal financing aspects of a DRIP convey positive information
when announced, and/or that on balance transactions costs and brokerage fees are
reduced for the firm and its owners.Q The Dubofsky and Bierman arguments are in
contrast to the braggadocio of the free cash flow hypothesis, according to which
excess cash should be siphoned away from the firm or else management would
squander it on unnecessary perks. Perumpral, Keown and Pinkerton (1991) study the
announcement month of a DRIP and find results similar to those of Dubofsky and
Bierman. In contrast, Allen (1991) reports negative significant abnormal returns for
industrial firms announcing DRIPs.
(C) Raising new capital/rights offerings
Scholes and Wolfson (1989) examine discount DRIPs and find that discount
DRIPs provide an investment banking function in raising new capital for the firm.
They note that these bdiscount dividend–reinvestment and stock–purchase plans
allow shareholders to capture part of the underwriting fees incurred in new stock
offerings and save sponsoring firms some of the usual underwriting costs.Q Scholes
K. Chiang et al. / International Review of Financial Analysis 14 (2005) 570–586576
and Wolfson’s views are supported by Finnerty (1989), who emphasizes the
importance of new-issue DRIPs as a source of equity capital to firms.
Eckbo and Masulis (1992) study a large sample of equity offerings via rights
offerings and standbys. They find that 54% adopted new-issue DRIPs between 1973
and 1981, whereas only 28% of firm commitment issuers used new-issue DRIPs.
Eckbo and Masulis conclude that DRIPs are tantamount to periodic rights issues of
new shares. They also surmise that the wide use of DRIPs by firms that formerly
used rights offering/standbys will lead to the disappearance of equity offerings
through rights offering/standby methods.
(C) Agency costs/problems
Saporoschenko (1996) argues that firms with growth potential are more likely to
use treasury and combination plan DRIPs because of their need for outside funding.
Firms that issue treasury and combination plan DRIPs bypass outside monitoring,
part and parcel of other equity types. With respect to information asymmetries,
Saporoschenko notes that
. . .since treasury and combined DRIP’s are comparable to rights offering, firms
issuing DRIP’s should be confident of high subscription rates (this may be a reason
why some firms offer discounts of shares purchased through DRIP’s) and relative
undervaluation (ibid.).
Saporoschenko’s conjectures are based on the validity of findings by Eckbo and
Masulis (1992) and Smith (1977) that the lower direct flotation costs make rights offerings
a substantially cheaper method of equity issuance than going through an underwriter/
investment banker.
Saporoschenko (1996) also examines the small firm effect. Saporoschenko hypothe-
sizes that smaller firms should benefit more from the alleviation of informational
asymmetries via the use of DRIPs.
Steinbart and Swanson (1998) study the reasons that firms amend their DRIPs to
allow investors to make initial purchases of their stock. Their study attempts to ascertain
whether firms offer bno-loadQ DRIPs for economic or agency purposes. Through
telephone interviews, they asked investor relations’ representatives of firms why they
were selling bno-loadQ DRIPs. The prospectuses of these firms are also studied to learn
about the source of the shares that would be issued through bno-loadQ DRIPs. Their
results show that firms offered bno-loadQ DRIPs to broaden their shareholder base. From
this, Steinbart and Swanson (1998) conclude that their finding is consistent with agency-
related explanations.
3. Data, variables, models and results
As suggested in the introduction, our exploratory analyses of DRIPs focus on detecting
the firm characteristics that may (a) dictate plan features (i.e., a within group analysis of
DRIP firms) and (b) explore differences, if any, between DRIP firms and No-DRIP firms
(i.e., between group analysis).
K. Chiang et al. / International Review of Financial Analysis 14 (2005) 570–586 577
The DRIP plan features, described in Section 1, can be either qualitative or
quantitative. Qualitative variables are coded as binary, while quantitative variables are
in the appropriate units of measurement. In Table 1, we show the possible plan
features as factors, including their definition and the possible values the factors can
take.
From a sample of 1004 firms offering DRIPs on NetStockDirect, we randomly select
just 206 firms. There are two reasons for reducing the sample size to 206:
(1) for these firms, we could identify for certain one or more of the DRIP factors, and
(2) the reduced size allowed us to be sure that a close to full information dataset was
available for other variables of the firms we could to study.
The sample of 206 firms roughly corresponds to the relative frequencies of the total
population as per the three firm categories: financial, utilities and industrial firms. A
complete list of these firms is not included herein but is available on request from
Table 1
DRIP features
Factor name Definition Measure
Covenants
CASHD Cash dividends Yes=1, No=0
IRA IRA option Yes=1, No=0
OCP Optional cash purchase Yes=1, No=0
ACH Direct debit from bank Yes=1, No=0
DISC Discount Yes=1, No=0
LAH Loans against holding Yes=1, No=0
AF Open to/accept foreigners Yes=1, No=0
ADR American deposit receipt Yes=1, No=0
RSA Requires Special Affiliation Yes=1, No=0
MININSH Minimum initial shares required Quantity
MINOCP Minimum optional cash purchase $/year
MAXOCP Maximum optional cash purchase $/year
Fees
OCP fee OCP fee $ Amount
PUCOM Commission—purchase $/Share
SALESCOM Commission—Sales $/Share
SALESFEE Sales fee $/Share
TERMFEE Termination fee $/Share
ENROLFEE Enrollment fee $ Amount
SK Safekeeping/insurance $/Share
NSF Not sufficient funds $/Share
Additional features
PTYPE Plan can be designed either as a registered
DRIP or direct stock plan (DSP)
Registered DRIP=1;
DSP=0
E Enrollment Current shareholders
only=1; anyone=0
MNG Management type Company=1; agent=0
K. Chiang et al. / International Review of Financial Analysis 14 (2005) 570–586578
the authors. Table 2 displays simple statistics relating to the plan feature variables for
the sample firms. Note that not all factors are present for each of the 206 DRIP
firms.
In panel A of the table, the figures for the binary (qualitative) variables are given in
frequency and percent form. The numbers that represent dollars, either total/year or per
share, are presented in panel B with their minimum, maximum, average and standard
deviation.
The data reveal that, for this sample, more than 90% of the DRIP firms
! pay a cash dividend;
! extend the possibility of optional cash purchases of the stock;
! open their plan to foreign investors;
! do not have an IRA;
! do not provide loans against the reinvested dividends;
Table 2
DRIP features—descriptive statistics
(A) Binary variables
Variable Number of Frequency Percentage of 1s Standard
name observations0 1
(mean) [%] deviation [%]
CSHDIV 206 12 194 94.17 23.48
IRA 206 193 13 6.31 24.37
OCP 206 9 197 95.63 20.49
ACH 206 162 46 21.36 41.08
DISC 206 112 94 45.63 49.93
LAH 206 205 1 0.49 6.97
AF 206 5 201 97.57 15.43
ADR 206 196 10 4.85 21.54
RSA 206 203 3 1.46 12.01
PTYPE 206 49 157 76.21 42.68
MNG 195 181 14 7.18 25.88
E 157 1 157 100.00 0.00
(B) Quantitative variables
Variable name Number of
observations
Minimum Maximum Mean Standard
deviation
MININSH 157 1 100 2.88 10.02
MINOCP 205 0 $600.00 $51.53 $91.13
MAXOCP 200 0 $480,000.00 $53,031.00 $67,165.82
OCPFEE 138 0 7.5 0.59 1.56
PUCOM 138 0 0.25 0.01 0.36
SALESCOM 44 0 0.25 0.10 0.05
SALESFEE 62 0 15 $7.15 $5.62
TERMFEE 68 0 15 $2.35 $5.14
ENROLFEE 138 0 15 $0.87 $2.76
SK 106 0 10 $0.33 $1.47
NSF 11 0 25 $18.63 $9.51
K. Chiang et al. / International Review of Financial Analysis 14 (2005) 570–586 579
! have no ADR3 (because they are domestic firms);
! do not require any special affiliations; and
! have an agent who administers the plan.
In addition, almost half the plans offer a discounted share price, and more than 75% are
registered dividend reinvestment plans.
Fees are just averages and are therefore quite misleading because one must look at the
specific combinations of fees with other factors (discounted share price, for instance). The
other two aspects of the sample that justify special scrutiny are the values for the
minimum, maximum and average dollar investment required for OCP. The minimums
range from 0 to $600, with an average of $51.53, whereas the maximums are from 0 to
$480,000, with an average of $53,031. It seems that some of the OCPs are designed for the
large investor who most likely will benefit from bypassing the broker in a deal.
Next, we select 17 financial and stock market variables on which a meaningful within-
group analysis and a comparison between the primary sample and a control sample may be
made. To compare and contrast DRIP firms with No-DRIP firms, we match our primary
samplewith a control sample constructed by stratified random sampling, where thematching
criterion, in addition to the industry stratification, is firm size as measured by total assets.
Although a wide variety of such variables have been used in accounting and financial
research, there is no clear-cut method that would determine the minimum set and its ex-
clusivity. Prior research in this regard was, by and large, arbitrary and depended almost
exclusively onwhat seemed logical and onwhether data were available. Nevertheless, some of
these variables appear almost in every study both as matters of logical selection and as
tradition. The approachwe take here does not differ fromprior practice, and that being the case,
we would be, despite our good intentions, subject to the same criticism as our predecessors.
Data are obtained from COMPUSTAT and represent accounting information and/or
market data as of December 31, 2001. Beta is available directly from COMPUSTAT,
where it is calculated from 60 monthly observations of rate of return on the firm’s stock as
the dependent variable and the S&P 500 Index as the independent variable. Table 3 is a
reference list of the names and the definitions of the 17 accounting/market variables.
Although the selection of these variables is either arbitrary or it is necessitated by data
availability, it represents a reasonable group of variables for which meaningful
comparisons can be made because most are mentioned in conjunction of prior studies
of the DRIP phenomenon.
Naturally, the most critically revealing, stand-alone variable is DPS. Next is a group of
variables of various empirical measures of return and profitability (ROA, ROE, NPM, PE
and EPS). Following return and profitability variables, we have a measure of leverage/
financial risk (LTDEQ) and, after that, variables that traditionally proxy for liquidity (FCF,
CR, NWC), then an additional measure of size, independent of the creation of the control
sample (NSTA), and variables that reflect on growth potential (R&D/S and SAGR). Closing
the set of variables serving for comparison purposes are ownership (IO and INST) and
proxies of risk (B/M and BETA). With regard to this last group, we advise the reader to
peruse the voluminous literature on the CAPM, especially Fama and French (1992).
3 American depository receipt.
Table 3
Accounting variables used in within- and between-group analyses
Variable name Definition
DPS The last dividend paid by the firm from COMPUSTAT
Measures of return or profitability
ROA Return on assets, calculated from financial statements (net income/total assets)
ROE Return on equity, calculated from financial statements (net income/total equity)
NPM Net profit margin (net income/sales), as calculated from financial statements
PE Stock price/earnings, calculated from income statement and COMPUSTAT
EPS Earnings per share, calculated form financial statements
Measure of financial risk
LTDEQ Long-term debt/total equity, calculated from balance sheet
Measures of liquidity
FCF Free cash flowa
CR Current assets/current liabilities (also known as the current ratio),
calculated from financial statements
NWC Net working capital (total current assets less total current liabilities)
Measure of size
NSTA Net sales/total assets (also known as the total asset turnover),
calculated from financial statement
Measures of growth potential
R&D/S R&Db expenses/net sales
SAGR Five-year sales growth ([St/St�4])1/5�1.0, where t=2001
Measures of ownership
IO Insider ownership calculated from baseline and Yahoo finance
INST Institutional ownership as a fraction of total ownership
Market proxies of risk
B/M Book/market ratio calculated from balance sheet and market value
of stock at 2001 year end
BETA As of the ubiquitous market model
a Net cash flow, less dividends, less capital expenditures, COMPUSTAT mnemonic bFREECFLQ.b R&D is all costs that relate to the development of new products or services, COMPUSTAT mnemonic
bXRDQ.
K. Chiang et al. / International Review of Financial Analysis 14 (2005) 570–586580
3.1. Within-group analysis
First, we analyze with a series of probit regressions whether accounting variables
can differentiate/separate within DRIPs. This we call the within-group analysis. In this
analysis, the dependent variable is some factors defined in Table 1 as binary (1,0).
We use in this analysis only DISC, ACH, PTYPE and MNGT. This we do because
the remaining factors fail to partition regarding size, and thus, they would be
meaningless. In addition, we do not consider the fees charged on services as a
differentiating factor because almost all firms charge at least one type of fee
K. Chiang et al. / International Review of Financial Analysis 14 (2005) 570–586 581
(commission on purchase or sales, sales fee, termination fee, enrollment fee,
safekeeping fee etc.).4
A binary probit model can be approached in two ways:
(1) as a regression model or
(2) as a random utility model.
In essence, the two models are quite similar. The reason we opt for the latter is that it
allows for a random vector. Its underlying assumption is that micro units (DRIPs, in this
case) separate/distinguish themselves in the margin by the respective choices they offer.
In its general form, the model is expressed as
y4 ¼ bVXþ e
where y*, bVand e are vectors, and X is a matrix with columns equal to the number of
independent variables. The vector e is a vector of random disturbances distributed standard
normal.
Under the random utility model, we consider DRIP factors as ym if the coding of the
factor is unity and yn, otherwise, corresponding to Ua and Ub as the respective utilities
derived, given the choice.5 Then we have
U a ¼ baVXþ ea and U b ¼ bbVXþ eb:
If we denote Y=0 (choosing the first alternative response), then
Prob Y ¼ 0jX½ � ¼ Prob ½U abU b� ¼ Prob ½baVXþ ea �bbVXþ ebjX�
¼ Prob baV� bbVÞXþ ea � ebÞb0jXð � ¼ Prob b ¼ Xþ eb0jX�:½ð½
For maximum likelihood estimation of the probit model’s parameters, see Greene
(1997, pp. 873–888).
Our interest is in how DRIP firms distinguish one from the other across the 17 financial
variables we defined in Table 3.
We look at the slope coefficients of the individual probit regression run for each
factor. Because our analysis is marginal, in the interest of clarity, we show in Table 4
only the slope coefficients of those variables that are statistically significant at the
level of 0.10 or less. Moreover, we omit those accounting variables for which none of
the factors were distinguishable at the same level of significance of the slope
coefficients.6
Table 4 displays the DRIP factors (dependent variables) as rows and the
accounting/ financial variables (independent variables) as columns with their estimated
5 Please see Table 1 for the binary designations.6 Complete results are available, upon request, from the authors.
4 A subgrouping of the fee structure and hence using dollar amount for each proved to be useless, as in many
cases, this information was not available.
Table 4
Within-group probit analysis
DRIP
features
Accounting
variables
ROA NPM EPS LTDEQ CR NSTA R&D/S SAGR IO INST BETA
DISC 0.041 �0.363 0.407 0.405 122.458 �0.003 0.947 0.532
0.042 0.002 0.052 0.006 0.002 0.001 0.037 0.042
PTYPE 0.293 �0.672 0.001 �0.928
0.014 0.008 0.043 0.004
ACH �0.252 �0.252 0.863 0.767
0.021 0.021 0.003 0.021
MNG 0.052 �0.235
0.072 0.062
K. Chiang et al. / International Review of Financial Analysis 14 (2005) 570–586582
slope coefficient and, in italics, the corresponding probability of error type I. Thus
presented, the reader can verify both the value of the coefficient (specially note its
sign) and the strength of the rejection of the null hypothesis (the probability of a
zero slope).
The strongest bwithin-groupQ separation one can find for the factor DISC (discount).
Eight out of the 12 remaining financial/accounting variables are significantly different than
0 at the 0.05 level of error type I (although for CR, the probability is just a marginal
0.052). The coefficients’ signs are positive with two exceptions, meaning that the higher
the value of a particular accounting measure the higher the probability of offering
discount. The exceptions are SAGR and LTDEQ, implying that lower growth rate and
lower financial risk firms tend to offer discount as part of the plan, other things being
equal.
The next two factors with four significantly nonzero coefficients are ACH and PTYPE
(direct debit from bank and plan type, respectively). Two variables, CR and BETA, are
common to both factors; the other two are not. From the signs of the coefficient for ACH,
it seems that DPS and NPM are negatively correlated with DRIPs that offer a direct debit
and positively correlated with CR and BETA. Other things being equal, these firms have a
higher systematic risk and, perhaps because of that, a higher current ratio. With respect to
plan type, precisely for these two variables, registered plans are less risky (both
coefficients are negative). In addition, they have higher LTDEQ and IO (insider
ownership) than direct stock plans. With regard to the last factor, MNG, the distinguishing
variables are only two (ROA and EPS) and are significant at error type I level higher than
0.05.
For the financial/accounting variables, BETA and CR are the most separating for
four factors, whereas the majority of variables are separating for one or two. Recall
that we omitted already five variables that were not separating for even a single
factor.
In conclusion, we suggest that, as far as the variables we study go, the DRIP plans seem
to be quite homogeneous with the exception of BETA and a liquidity risk measure, where
the signs are consistent with what one would logically expect regarding systematic risk
taking. In the next subsection, we analyze the differences between the primary and the
control samples.
K. Chiang et al. / International Review of Financial Analysis 14 (2005) 570–586 583
3.2. Between-group analysis
Tables 5 and 6 present the results of our tests and contrasts of the primary and the control
samples. Because we match the DRIP firms with No-DRIP firms by a random selection of
size within an industry and because of missing data or other reasons, not all the measures are
present for all the firms initially selected from either the 206 DRIP or the No-DRIP firms.
This is why all the accounting/market variables have less than 206 observations. In fact, the
highest number of pairings is for ROA and NPM (173), and the fewest is for R&D/S (30).
Table 5 is the simple tabulation of sample statistics, first five columns showing the
mean, the number of observations, standard deviation and standard error of the mean,
respectively, for each of the 17 accounting variables for two groups. The last two columns
Table 5
Paired variables of DRIPs and No-DRIPs
Variable Mean N Standard
deviation
Standard error
of mean
Correlation Significance
Pair 1 DPS1 1.00 172 0.908 0.069 0.337 0.000
DPS2 0.78 0.705 0.054
Pair 2 ROA1 2.78 173 4.932 0.375 �0.074 0.330
ROA2 2.04 6.383 0.485
Pair 3 ROE1 9.20 172 18.261 1.392 �0.065 0.397
ROE2 3.12 60.117 4.584
Pair 4 NPM1 9.28 173 14.866 1.130 0.316 0.000
NPM2 15.07 123.178 9.365
Pair 5 PE1 26.30 68 44.527 5.400 0.075 0.546
PE2 30.08 54.655 6.628
Pair 6 EPS1 0.98 82 2.686 0.297 0.056 0.615
EPS2 1.37 3.337 0.369
Pair 7 LTDEQ1 1.03 169 0.874 0.067 0.370 0.000
LTDEQ2 1.13 1.332 0.102
Pair 8 FCF1 0.04 57 0.542 0.072 0.030 0.827
FCF2 �0.87 7.100 0.940
Pair 9 CR1 1.30 103 0.720 0.071 0.294 0.003
CR2 1.46 1.019 0.100
Pair 10 NWC1 196.91 103 1548.897 152.617 0.576 0.000
NWC2 112.96 1103.859 108.766
Pair 11 NSTA1 0.66 171 0.671 0.051 0.688 0.000
NSTA2 0.64 0.691 0.053
Pair 12 R&D1/S1 0.02 30 0.044 0.008 0.532 0.002
R&D2/S2 0.02 0.033 0.006
Pair 13 SAGR1 0.13 147 0.141 0.012 0.317 0.000
SAGR2 0.17 0.181 0.015
Pair 14 IO1 0.11 75 0.120 0.140 0.214 0.065
IO2 0.23 0.217 0.025
Pair 15 INST1 0.51 87 0.261 0.028 0.278 0.009
INST2 0.47 0.280 0.030
Pair 16 B/M1 0.25 71 4.760 0.564 0.000 1.000
B/M2 0.82 0.813 0.096
Pair 17 BETA1 0.52 107 0.450 0.044 0.435 0.000
BETA2 0.58 0.526 0.051
Table 6
Paired difference tests
Variable Paired differences 95% Confidence
interval
t-Value DF Significance
Mean
difference
Standard
deviation
Standard
error
Lower Upper
Pair 1 DPS1–DPS2 0.218 0.943 0.072 0.076 0.360 3.026 171 0.003
Pair 2 ROA1–ROA2 0.746 8.352 0.635 �0.508 1.999 1.174 172 0.242
Pair 3 ROE1–ROE2 6.077 63.955 4.877 �3.549 15.703 1.246 171 0.214
Pair 4 NPM1–NPM2 �5.792 119.321 9.072 �23.698 12.115 �0.638 172 0.524
Pair 5 PE1–PE2 �3.783 67.874 8.231 �20.211 12.646 �0.460 67 0.647
Pair 6 EPS1–EPS2 �0.397 4.165 0.460 �1.312 0.518 �0.863 81 0.391
Pair 7 LTDEQ1–LTDEQ2 �0.101 1.294 0.100 �0.298 0.095 �1.017 168 0.311
Pair 8 FCF1–FCF2 0.907 7.105 0.941 �0.979 2.792 0.963 56 0.340
Pair 9 CR1–CR2 �0.158 1.061 0.105 �0.365 0.050 �1.508 102 0.135
Pair 10 NWC1–NWC2 83.947 1283.767 126.493 �166.952 334.846 0.664 102 0.508
Pair 11 NSTA1–NSTA2 0.022 0.538 0.041 �0.060 0.103 0.523 170 0.602
Pair 12 R&D/S1–R&D/S2 0.002 0.039 0.007 �0.012 0.016 0.287 29 0.776
Pair 13 SAGR1–SAGR2 �0.033 0.191 0.016 �0.065 �0.002 �2.125 146 0.035
Pair 14 IO1–IO2 �0.120 0.225 0.026 �0.171 �0.068 �4.608 74 0.000
Pair 15 INST1–INST2 0.044 0.325 0.035 �0.026 0.113 1.249 86 0.215
Pair 16 B/M1–B/M2 �0.566 4.826 0.573 �1.708 0.577 �0.988 70 0.327
Pair 17 BETA1–BETA2 �0.061 0.523 0.051 �0.161 0.039 �1.209 106 0.229
K. Chiang et al. / International Review of Financial Analysis 14 (2005) 570–586584
present the correlation coefficient of the 17 paired variables and the probability under the
null of correlation. At 0.05 level of error type I, the following 10 variables are significantly
correlated: DPS, NPM, LTDEQ, CR, NWC, NSTA, R&D/S, SAGR, INST and BETA.
With the exception of ROA, all correlations, whether significant or not, are positive.
Therefore, it is safe to conclude that, for the majority of variables, the data cannot show
drastically different group characteristic. Comovement, nevertheless, does not mean that
there are no significant differences between the two types of firms. To test for statistically
significant differences, in Table 6, we show the tests of paired mean differences.
Table 6 is organized in 10 columns. Columns 1 and 2 are the numbered pairs and the
direction of the difference (DRIP minus No-DRIP). Column 3 is the mean difference
(equals to the difference of the means). In columns 4 and 5 are the standard deviation of
the difference and its standard error. Columns 6 and 7 show the 95% confidence interval
of the mean difference, column 8 is the t-value of the difference, and column 9 is the
number of degrees of freedom for a particular pairing. Finally, column 10 is the
probability under the null of error type I of the mean difference coming from the same
population.
It can be seen that, at 0.05 level or less of error type I, just three pairs come from two
different populations: DPS, SAGR and IO. Although significant differences are few, their
signs are nevertheless consistent with what one would expect from the two types of firms.
DRIP firms pay more dividends and, at the same time, entice their shareholders to bleavethe money on the table.Q The growth factor of No-DRIP firms is higher than that of DRIP
firms, and thus, they plow back more internally generated funds because it is a cheap
source of capital (Myers & Majluf, 1984). In addition, insiders’ ownership is higher in No-
K. Chiang et al. / International Review of Financial Analysis 14 (2005) 570–586 585
DRIP firms; therefore, they need less enticement and consistently pay out less in dividends
(DPS1-DPS2 is positive).
4. Future work and conclusions
Perhaps more questions than answers remain at the conclusion of this study. This may
be so because, despite all our good intentions, resources are limited. With more readily
available data, one could carry out more detailed analyses. An obvious avenue of further
research and a different dimension of exploration must follow a route where corporate
decision makers and investors are asked about the good and bad things DRIPs do for them
and why they do or do not use this instrument of financing/investment in the firm. This
issue may obtain an additional edge by considering the 2003 dividend tax reform whereby
dividends are taxed at the fix rate of 15%.
The focus of the survey we propose is the motivation for initiating a DRIP. Such a
survey is badly needed because of the total lack of this line of research on the subject. We
hope that we can also increase the scope of investigation by examining a fuller, more
diverse sample of firms, as well as DRIPs investors. This would be important because,
although they are not a new phenomenon, DRIPs are gaining in popularity as a long-term
investment tool for individual investors. This increase in popularity has led to a dramatic
increase in the number of firms offering DRIPs. This increase is buttressed by a
concomitant increase in the complexity of DRIP characteristics.
The bad news is that such research requires resources that very few researchers have.
But if the role of science is discovery (a search for the truth), then there is no way around
finding both the resources and the know-how to follow this path. Short of that, finance as
an academic profession will never get to the bottom of the dividend puzzle.
Acknowledgements
The authors are indebted to Ms. Kim Alexander, who, as a research assistant of the
second author, did initial exploratory literature search on the subject. All usual disclaimers
of responsibility to others than the authors apply, nevertheless.
References
Allen, Grace Casbum (1991). Dividend reinvestment plans: The effects on shareholder wealth and equilibrium
security returns. Unpublished dissertation, The University of South Carolina, Columbia SC.
Baker, H. Kent, & Seippel, William H. (1981). The use of dividend reinvestment plans by utilities. Akron
Business and Economic Review, 35–41.
Bierman Jr., Harold (1997). The dividend reinvestment plan puzzle. Applied Financial Economics, 7, 267–271.
Black, Fischer (1976). The dividend puzzle. Journal of Portfolio Management, 2, 5–8.
Burns, Greg (1994). Brokers sure won’t plug these DRIPS. Business Week, 3377, 186.
Carlson, Charles B. (1992). Buying stocks without a broker. New York7 McGraw-Hill.
Carlson, Charles B. (1995). No-load stocks–how to buy your first share and every share directly from the
company with no broker’s fee. New York7 McGraw-Hill.
K. Chiang et al. / International Review of Financial Analysis 14 (2005) 570–586586
Carlson, Charles B. (1996). Buying stocks without a broker. New York7 McGraw-Hill.
Chang, Otto H., & Nichols, Donald R. (1992). Tax incentives and capital structures: The case of the dividend
reinvestment plan. Journal of Accounting Research, 30, 109–125.
Cherin, Anthony C., & Hanson, Robert C. (1995). Dividend reinvestment plans: A review of the literature.
Financial Markets, Institutions and Instruments, 4, 59–73.
Davey, Patrick J. (1976). Dividend reinvestment programs. New York7 Conference Board. No. 699.
Dubofsky, David A., & Bierman, Leonard (1988). The effect of discount dividend reinvestment plan
announcements on equity value. Akron Business and Economic Review, 19, 58–68.
Eckbo, B. Espen, & Masulis, Ronald W. (1992). Adverse selection and the rights offer paradox. Journal of
Financial Economics, 32, 293–332.
Fama, Eugene F., & French, Kenneth R. (1992). The cross-section of expected stock returns. Journal of Finance,
47, 427–465.
Finnerty, John D. (1989). New issue dividend reinvestment plans and the cost of equity capital. Journal of
Business Research, 18, 127–139.
Goodman, Jordan E. (1993). Building your assets on autopilot. Money, 22, 82–83.
Greene, William H. (1997). Econometrics. Upper Saddle River, NJ7 Prentice-Hall.
Landy, Eugene W. (1996). Benefits of Dividend Reinvestment Plans and Shareholder Investment Plans for
Shareholders, Real Estate Investment Trusts and the Economy. http://www.mreic.com/landy.html
Myers, Stuart C., & Majluf, N. (1984). Corporate financing and investment decisions when firms have
information that investors do not have. Journal of Financial Economics, 13, 187–221.
Perumpral, S., Keown, Arthur J., & Pinkerton, John (1991). Market reaction to the formulation of automatic
dividend reinvestment plans. Review of Business and Economic Research, 26, 48–58.
Peterson, Pamela P., Peterson, David R., & Moore, Norman H. (1987). The adoption of new issue dividend
reinvestment plans and shareholder wealth. Financial Review, 22, 221–232.
Saporoschenko, Andy (1996). DRIP’s, agency and capital structure issues. Working Paper: University of South
Carolina. June, 19–21.
Scholes, Myron S., & Wolfson, Mark A. (1989). Decentralized investment banking: The case of discount
dividend-reinvestment and stock-purchase plans. Journal of Financial Economics, 17, 7–35.
Schneid, Daniel L. (1981). Packaging the dividend reinvestment plan. Financial Executive.
Smith, Clifford (1977). Alternative methods for raising capital: Rights versus underwritten offerings. Journal of
Financial Economics, 5, 173–207.
Steinbart, Paul John, & Swanson, Jane (1998). dNo-loadT dividend reinvestment plans. Review of Financial
Economics, 7, 121–141.
The Moneypaper. (1996). Guide to dividend reinvestment plans . New York7 Temper of the Times
Communications.
Recommended