Monopoly Money: The Effect of Payment Coupling and Form on

Spending Behavior
Priya Raghubir
New York University
Joydeep Srivastava
University of Maryland, College Park
This article examines consumer spending as a function of payment mode both when the modes differ in terms
of payment coupling (association between purchase decision and actual parting of money) and physical form
as well as when the modes differ only in terms of form. Study 1 demonstrates that consumers are willing to
spend more when a credit card logo is present versus absent. Study 2 shows that the credit card effect can be
attenuated when people estimate their expenses using a decomposition strategy (vs. a holistic one). Noting that
credit card and cash payments differ in terms of payment coupling and form, Studies 3 and 4 examine
consumer spending when the payment mode differs only in physical form. Study 3 demonstrates that
consumers spend more when they are spending scrip (a form of stored value certificate) versus cash of the
same face value. Study 4 shows that the difference in spending across payment modes (cash and gift
certificates) is attenuated by altering the salience of parting with money through contextual manipulations of
the differences between cash and gift certificates.
Keywords: money illusion, gift certificates, credit cards, subjective value of money, economic psychology
With the proliferation of different payment modes in recent years,
consumers have a wide array of payment options to choose from in
making their purchases. Typically, consumers have the option to
purchase with cash, a check, or a credit or a debit card. However,
other payment modes such as bank drafts, money orders, travelers
checks, gift certificates, gift cards, coupons at fairs, chips at gaming
houses, stored value cards such as those used in mass transit and tolls,
and so forth are quite common in the marketplace. Some types of gift
cards are almost identical to debit cards, such as the American
Express or the Visa gift card. Further, the advent of Internet com-
merce has spurred the growth of new payment modes such as Paypal.
Despite the proliferation of these diverse payment modes, and that the
mode of payment is an important contextual element in any transac-
tion, research on the influence of payment mode on consumer spend-
ing decisions and behavior is relatively sparse (for a review see
Raghubir, 2006). Do consumers spend differently when using one
payment mode relative to another mode? For example, do consumers
spend more when they receive $50 in the form of a gift card than in
the form of cash? If indeed they do, then why? This research ad-
dresses these issues.
Although a growing body of literature demonstrates that the nor-
mative principle of descriptive invariance, which holds that prefer-
ences should not vary when the same objective stimuli are represented
differently, is commonly violated in the domain of money (e.g.,
Gourville, 1998; Raghubir & Srivastava, 2002; Shafir, Diamond, &
Tversky, 1997; Shefrin & Thaler, 1988; 1985), relatively few studies
have investigated differences in spending behavior as a function of the
mode of payment. The few studies that exist demonstrate that con-
sumers tend to spend more when paying with a credit card than when
paying by cash or check, after controlling for other factors (Cole,
1998; Feinberg, 1986; Hirschman, 1979; Prelec & Loewenstein,
1998; Prelec & Simester, 2001; Soman, 2001). Tokunaga (1993)
argued that a credit card is a convenient payment mode that allows
people to defer and spread out payments and, thus, consumers differ
in how they treat credit card and cash purchases. Review of the
previous research on the effect of payment mode on spending behav-
ior suggests that the focus has been primarily on the difference
between credit card and cash payments. Further, the phenomenon of
higher spending when paying with a credit card than with cash has
been attributed to the temporal separation of the purchase decision and
the actual payment in the case of credit card payments (Prelec &
Loewenstein, 1998). We refer to this specific feature of payment
mode as payment coupling.
In this article we argue that payment mode can be differentiated
in at least two ways: payment coupling and payment form. Pay-
ment coupling refers to the extent to which the decision to pur-
chase (or spend) is temporally associated with the actual parting of
money (Loewenstein & Prelec, 1992; Prelec & Loewenstein, 1998;
Thaler, 1999) whereas payment form refers to differences between
Priya Raghubir is a professor at the Stern School of Business, New York
University. Joydeep Srivastava is an associate professor at the Robert H.
Smith School of Business, University of Maryland, College Park.
Order of authorship is alphabetical and reflects equal contribution by
both authors.
The authors acknowledge the helpful comments of seminar participants
at the Hong Kong University of Science and Technology, the London
Business School, the University of Paris at Dauphine, and the University of
Texas at Austin. This research was partially funded by the Hellman family
grant and the undergraduate research apprentice program grant awarded by
the University of California at Berkeley to the first author and the graduate
research board summer award awarded by the University of Maryland to
the second author.
Correspondence concerning this article should be addressed to Priya Ra-
ghubir, Marketing Department, #809 Tisch, Stern School of Business, New
York University, 44 West 4
th
Street, New York, NY 10012-1126. E-mail:
raghubir@[Link], or Joydeep Srivastava, Department of Marketing, Robert
H. Smith School of Business, VMH 3453 Van Munching Hall, University of
Maryland, College Park, MD 20742-1815. E-mail: srivasta@[Link]
Journal of Experimental Psychology: Applied Copyright 2008 by the American Psychological Association
2008, Vol. 14, No. 3, 213225 1076-898X/08/$12.00 DOI: 10.1037/1076-898X.14.3.213
213
two monetary instruments that are identical in coupling and face
value but different in terms of physical appearance (e.g., $50 legal
tender and a $50 gift certificate that can be used anywhere).
Although the emphasis of previous research has been on the effect
of payment mode when the modes differ in coupling, relatively
little research focuses on differences in spending as a function of
payment form. Noting that a credit card differs from cash in
coupling as well as form, this research examines consumer spend-
ing as a function of payment mode both when the mode differs in
coupling and form (e.g., credit card vs. cash) and when the mode
differs only in form, holding constant payment coupling (e.g., gift
certificate and cash of equivalent value).
The conceptual underpinning of our research is that payment
modes differ in transparency or the vividness with which individuals
can feel the outflow of money, with cash being the most transparent
payment mode. We argue that the more transparent the payment
outflow, the greater the aversion to spending or higher the pain of
paying (Prelec & Loewenstein, 1998), leading to less transparent
payment modes such as credit cards and gift cards (vs. cash) being
more easily spent or treated as play or monopoly money. Further, to
the extent that the transparency of paying underlies differences in
spending behavior, altering the salience of parting with money should
attenuate the difference across payment modes.
In a series of four studies, this article examines consumer
spending decisions as a function of payment mode. Studies 1 and
2 examine whether consumers are willing to spend more when the
payment mode is a credit card versus when it is cash (payment
mode differs in coupling and form). Holding payment coupling
constant, Studies 3 and 4 examine spending decisions when an
equivalent amount is given either in the form of a gift certificate or
cash (payment mode differs only in form). Exploring conditions
under which payment modes lead to differences in spending de-
cisions, Studies 2 and 4 examine the extent to which the spending
differences across payment modes can be eliminated or attenuated
by altering the salience of parting with money in the case of credit
cards and gift certificates, respectively.
Conceptual Background
Research on the cognitive psychology of financial behavior sug-
gests that people organize their finances around mental accounts
(Thaler, 1985, 1999). Thaler (1999) described mental accounting as
the set of cognitive operations used by individuals and households to
organize, evaluate, and keep track of financial activities. Analogous
to accounting and budgeting systems used by companies, individuals
are modeled as creating separate source-based expense and income
ledgers in their mind. Shefrin and Thaler (1988) suggested that people
tend to categorize income into different mental accounts that then
affects the propensity to spend (see also Henderson & Peterson,
1992). For example, people group and label expense accounts sepa-
rately such as food and entertainment, and spending is driven by the
available surplus or deficit in each account. Since people treat money
differently depending on how it is labeled, money in one mental
account is not a perfect substitute for money in another account
thereby violating the normative principle of fungibility. In general,
mental accounts serve to simplify financial decision making and aid in
making trade-offs between different types of spending.
At the specific transaction level, a mental account is opened, and
the decision to purchase or not is based on an evaluation of the
perceived benefits of consumption and the costs of payment in this
account (Prelec & Loewenstein, 1998; Thaler, 1999). Prelec and
Loewenstein (1998) suggested that when people make purchases,
there is an immediate pain of paying which can reduce the pleasure
of consumption or even prevent it altogether. In balancing the
pleasure derived from a purchase and the pain of paying for it,
making the cost salient undermines the pleasure one can derive,
whereas making the benefits of the purchase salient may blunt the
pain of paying. One factor that tends to enhance the pleasure of
consumption and reduce the pain of paying is coupling or the
extent to which the consumption and payment are psychologically
linked together due to their temporal proximity.
Previous literature on the effects of payment mode suggests that
consumers tend to spend more when using a credit card than cash
(Feinberg, 1986; Hirschman, 1979). The higher spending with
credit cards relative to cash has been attributed to the temporal
separation of the purchase from the actual payment in the case of
credit cards or the decoupling of the purchase from the payment
(Prelec & Loewenstein, 1998; Thaler, 1999; Tokunaga, 1993). In
the case of cash purchases, there is a tight coupling of the con-
sumption and the payment thereby accentuating the pain of paying.
In the case of credit card purchases, actual parting of the money
occurs after the purchase decision thereby dulling the pain of
paying. Thus, the observed bias in spending across credit card and
cash payments is because the pain of paying is higher when paying
by cash than when paying with a credit card. Further, in balancing
the immediate gratification against the expectation of the pain of
paying in the future, people are likely to underestimate the pain
and thus spend more with a credit card than cash. In line with this
reasoning, Srivastava and Raghubir (2002) reported that people
recall their cash payments better than their credit card expenses,
indicating that the salience of individual payments is lower for
credit card payments. This reasoning also implies that the differ-
ence in spending behavior is likely to attenuate if the salience of
parting with money is increased at the point of purchase.
In addition to the difference in payment coupling across payment
modes, we argue that payment modes may also differ in terms of
form. Consider an individual who receives $50 either in the form of
cash or a gift card (that can be used almost universally). Although
there is no difference in coupling or the face value of the amount of
money given (i.e., $50), the two payment modes are different in
physical appearance and thus differ in terms of payment form. In the
case of credit cards versus cash, the two payment modes differ in
terms of coupling as well as form whereas in the case of gift cards
versus cash, the two modes differ only in terms of form. We argue that
the physical form of the payment mode is likely to affect payment
transparency or the vividness with which the outflow of money is felt.
Thus, extending the reasoning to different payment forms, the more
transparent the payment form, the higher the salience of parting with
money, the greater the aversion to paying, and the lower the likelihood
and level of spending.
In terms of payment transparency, cash is the most transparent
form of money as its status as legal tender makes it salient in both
physical form and amount (Soman, 2003). When paying by cash,
the feeling of parting with money is very vivid and is akin to
having ones meter running (Thaler, 1999). In contrast, a
different payment form (e.g., credit card or gift card) may not feel
or appear as real as legal tender thereby reducing the salience of
parting with real money. In other words, using a different payment
214
RAGHUBIR AND SRIVASTAVA
form (other than legal tender) may seem like play money or
monopoly money, making it easier to spend. Given that the
transparency of payment makes the parting of money more vivid
and real, the pain of paying is likely to be greater when using cash
relative to other less transparent payment forms. In other words,
reducing the salience of parting with real money psychologically
reduces the barrier to spend.
Four studies examine differences in spending behavior as a
function of payment mode. Study 1 begins by demonstrating that
consumers willingness to spend is higher when a credit card logo
is present versus absent. Study 2 shows that the credit card effect
is attenuated when people are asked to estimate credit card pay-
ments using a piecemeal decomposition strategy (vs. a holistic
estimate) that increases the salience of individual payments
(Menon, 1997). Given that a credit card differs from cash in terms
of coupling and form, Studies 3 and 4 examine differences in
spending behavior when the payment mode differs only in form.
Study 3 demonstrates that consumers spend more when $50 is
given in the form of a gift certificate than when an equivalent
amount is given in the form of cash. Study 4 extends the findings
by showing that compared to the baseline where consumers are
more likely to spend when using a gift certificate relative to cash,
the differences in spending behavior can be moderated by contex-
tually altering the salience of parting with money.
Study 1
As a starting point, the objective of Study 1 is to replicate the
previous finding that people tend to spend more when using a
credit card than cash (Feinberg, 1986; Hirschman, 1979). In par-
ticular, we examine whether the mere presence of a credit card
logo increases the price that consumers are willing to pay. Another
objective of Study 1 is to explore whether general attitudes toward
the use of credit cards and cash affect the price that consumers are
wiling to pay in the presence versus absence of a credit card logo.
Method
One hundred and fourteen undergraduate students (68 male and
46 female, median age 21) from an introductory marketing class
participated in the study for partial course credit. Participants were
assigned at random to one of four conditions of a 2 (credit card
logo: present or absent) 2 (replicate: lunch or dinner) between-
subjects design. The cover story was that a restaurant called
Michauls of St. Charles, was thinking of opening in their city.
They were given a brief description of the restaurant:
Known as Nawlins Best Cajun Restaurant, Michauls of St. Charles
is thinking of opening in ________. If you want delicious Cajun food,
Cajun music, and that Cajun atmosphere. . .Michauls of St. Charles is
the place! Right on the busiest street of ________, this restaurant
should be a real find. Spacious areas for parties, music, and drink-
ing. . .and that spicy Cajun food all rolled into one!
Creole and Cajun food has a long history. . .it has been influenced
throughout the years by many cultures French, West Indian, Sicilian
and Indian. When the Spanish began to settle here, they brought us the
pepper and the tomato. . .the beginnings of our Shrimp Creole!
New Orleans has traditions related to foodone of them is eating red
beans and rice on Mondays. Many of us grew up never knowing the
reasons why. . .we just enjoyed it. However, Monday is clothes day,
and red beans and rice is the perfect meal because it cooks slowly
while we are doing the laundry. Our tradition of eating seafood on
Fridays is based on the Catholic practice of fasting from meat on that
day, especially during the Lenten season. (In this city, it is no
sacrifice!).
Participants were asked to estimate the amount they would be
willing to pay for a set of nine menu items, using an open-ended
format. The replicate manipulation of lunch or dinner was introduced
in the question that asked: If you were to go to Michauls for lunch
[dinner], how much would you be willing to pay for. . .. They were
given names of nine dishes that included three appetizers: Red Bean
soup, Caesars salad, and Green salad; three entrees all served with
rice: Cajun roasted chicken (breast and leg), Nawlins Vegetables and
Beans (broccoli, beans, and eggplant), and Creole Seafood (shrimps,
squids, and mussels); and three items in the dessert/beverage section:
ice-cream, soda (Coke, Pepsi, 7-up), and tea/coffee. The price esti-
mates of the nine items were averaged to form an overall food price
index ( .84). Participants were then asked to rate the likelihood
of visiting the restaurant using a 7-point scale, 1 (not at all likely) to
7 (very likely).
To ensure that the conditions did not differ in terms of
participants actual usage and spending, we asked them to estimate
how much they spend, on average, per head for lunch and dinner
(using an open ended format), how often they ate out in the last
week (Did not eat out, 12 times, 34 times, 56 times, and 7 or
more times), and an estimate of the amount they spend on eating
out in a typical week.
The second page of the questionnaire consisted of questions that
related to general usage and attitudes toward credit cards and cash.
Participants were first asked to indicate their usual mode of pay-
ment for lunch (or dinner, depending on the condition to which
they had been assigned). They were given five different payment
modes: credit card, ATM debit card, check, cash, and other, and
responded by circling one of three subjective frequency categories,
1 (never), 2 (sometimes), and 3 (always).
A battery of 20 agree-disagree statements, 1 (strongly disagree) to
7 (strongly agree), was presented next to measure participants atti-
tudes toward credit cards versus cash. The 20-items loaded onto six
factors that represented negative affect (4 items tapping: fear, regret,
guilt, and underestimation of spending); positive affect (3 items tap-
ping: feel good, proud, status), beliefs in overspending (3 items);
control of spending (2 items tapping convenience and control of
spending with credit cards); ease of transaction (4 items tapping the
ease of paying with credit cards), and safety concerns (3 items tapping
misuse of credit cards). One item did not load onto any of the six
factors (with a loading .50). Means of the 20 items and their factor
loading are presented in Table 1. Confirmatory reliability analyses
showed that the internal validity of the negative affect scale ( .84),
positive affect scale ( .78), spending scale ( .70), and ease
scale ( .65) were acceptable, so these were combined into four
separate indices based on a simple average. The two items comprising
the control factor, and the three items comprising the safety factor has
low scale reliability ( .32 and .52, respectively), so they were
treated separately in later analyses, along with the one item that did
not load onto any factor.
Finally, all participants responded to how motivated (M5.16)
and interested (M 5.09) they were in completing the question-
naire using a 7-point scale, 1 (not at all) to 7 (very).
215
MONOPOLY MONEY
Results and Discussion
A multivariate analysis of variance on self reports of frequency of
and amount spent on eating out was conducted. There were no
differences in the reported frequency of eating out (M 3.12) or the
amount spent per week on eating out (M $46.13) across the four
conditions ( ps .39 for main and interaction effects). The reported
amount spent per person on lunch was marginally higher in the
presence (M $7.34) versus absence of a credit card logo (M
$6.47; F(1, 108) 3.43, p .06,
2
.03) as was the amount spent
per person on dinner (Ms $12.70 vs. $10.96; F(1, 108) 3.53,
p .06,
2
.03). No other effects were significant ( ps .45).
A 2 2 ANOVA on the food price index revealed a main effect
of the presence versus absence of the credit card logo (F(1, 110)
4.20; p .05.
2
.04). Means are graphically presented in Figure
1. The main effect of the replicate factor was not significant (F(1,
110) 2.60, p .11,
2
.023) and neither was the interaction (F(1,
110) .06; p .81,
2
.001). On average, participants were
willing to pay more when the credit card logo was present (M
$4.53, SD 1.15) than when it was absent (M$4.11, SD 1.06).
Thus, even though consumers were not explicitly informed which
payment mode they would be using, the mere presence of a credit card
logo increased the price that they were willing to pay.
The same ANOVA incorporating the indices of negative affect,
positive affect, spending and ease, and responses to the remaining
six attitude statements as covariates revealed that the statement I
tend to use my credit card without thinking of the amount I am
charging to it was significant (F(1, 95) 14.90, p .001,
2

.14), whereas the main effect of the presence versus absence of the
credit card logo remained significant (F(1, 95) 5.84, p .05,

2
.06) and the replicate factor of whether the estimates were for
lunch or dinner became significant (F(1, 95) 6.75, p .05,
2

.07). No other effects were significant ( ps .25 for all).

To examine how attitudes to the statement affected judgments,
we conducted a median split on the statement use without think-
ing (Median 3) and incorporated this as an additional measured
variables in the 2 2 ANOVA on the food price index. This 2
2 2 ANOVA revealed that all main effects were significant (F[1,
106] 4.18, 4.69, and 13.04 for the credit card logo, replicate, and
using a credit card without thinking; p .05 for all,
2
s .04,
.04, and .11, respectively). The main effect of the attitude to the
statement I tend to use my credit card without thinking of the
amount I am charging to it revealed that those who agreed with
the statement estimated higher prices (M$4.63, SD 1.19) than
those who did not agree with it (M $4.00, SD .95).
These results suggest that the less thought individuals pay to the
amount they are charging to their card, the more money they are likely
to charge to it. Consistent with the idea that greater transparency of
payment reduces the amount spent, the findings suggest that the less
salient the parting of money, the higher the level of spending. Study
2 tests whether the difference in spending decisions can be attenuated
by manipulating the salience of parting with money by focusing on
many small payments relative to one large payment.
Study 2
Study 2 has two main objectives. First, it attempts to replicate that
people are willing to spend more (or estimate spending more) when
Table 1
Attitude Statements Means, Standard Deviations and Factor Structure in Study 1
M SD
Factor loadings and scale reliability
Neg Pos Spend Control Ease Safety
Scale reliability (Cronbachs alpha) .84 .78 .70 .32 .65 .52
I believe credit cards lead to overspending. 4.12 1.94 .65
Credit cards are convenient. 6.21 1.14 .69
Using credit cards makes me feel good. 3.42 1.44 .79
I prefer combining expenses on a credit card as it helps me control my
monthly budget. 3.33 1.88 .62
Its unsafe to give my credit card number to others. 5.22 1.83 -.73
Paying by credit card makes me feel proud. 2.91 1.41 .89
The type of card I use reflects my status. 3.25 1.80 .78
I fear getting my credit card bill every month. 3.51 1.73 .77
I control my expenses better when I pay by card. 2.93 1.57
I believe companies are trying to get me to spend more by giving me a
high credit limit. 4.89 1.92 .82
I tend to use my credit card without thinking of the amount I am charging
to it. 3.77 1.83 .63
I spend less when I shop with cash. 4.84 1.77 .51
I regret the amount that I charge to my card when I have to finally pay
my bills. 3.82 1.86 .89
I find paying cash inconvenient. 3.81 1.94 .87
I buy unnecessary items when using a credit card. 3.40 1.65 .53
I underestimate the total expense of a shopping trip when I am planning to
use a credit card. 3.50 1.77 .69
Carrying cards is more convenient than carrying cash. 5.32 1.58 .74
When I pay cash, I find it difficult to remember how much I have spent. 4.23 1.90 .52
Credit card bills give a monthly expense summary. 5.46 1.58 .67
Using credit cards makes me feel guilty. 2.95 1.69 0.79
216
RAGHUBIR AND SRIVASTAVA
using a credit card than when using cash. Second, and more important
for the current investigation, Study 2 examines the effect of using a
decomposition strategy to increase the transparency of credit card
payments on attenuating the difference between credit card and cash
payments (Menon, 1997; Srivastava & Raghubir, 2002).
In the context of a shopping basket purchase (numerous product
categories), in the piecemeal decomposition estimate condition,
people were asked to estimate the cost of each item in the basket
individually. In another condition, the holistic estimate condition,
people were just asked to estimate the total cost of the basket when
using either a credit card or cash. It was expected that the salience
of parting with money would be higher in the piecemeal decom-
position estimate condition (Menon, 1997; Srivastava & Raghubir,
2002; Thaler, 1999), thereby attenuating the difference between
credit card and cash in this condition. In contrast, in the holistic
estimate condition, we expected to replicate the previous finding
that people would be willing to spend more when paying with a
credit card than with cash. The rationale is that in the decompo-
sition condition, each of the small costs will loom large thereby
accentuating the pain of paying. In addition, the decomposition
estimation condition creates a tight coupling or makes the link
between the payment and specific consumption act salient, when
the opposite is highly desirable (Thaler, 1999). For example,
Thaler (1999) suggested that for an expensive multicourse dinner,
a prix fixe dinner is preferable to having each of the courses priced
a la carte.
Method
Fifty-seven undergraduate students (33 male and 24 female)
from an introductory marketing class (median age 21) partici-
pated in the experiment for partial course credit. Participants were
assigned at random to one of four conditions of a 2 (payment
mode: cash and credit card) 2 (estimation strategy: holistic total
and piecemeal decomposition) between-subjects design. Sample
size ranged from 9 to 19 in each of the four conditions. The cover
story was a scenario where they had to estimate the budget for a
thanksgiving party. They were told:
Chris is planning a Thanksgiving party. There will be 6 guests. It will
be a nice, sit-down affair. The menu consists of turkey with accom-
paniments, salad, vegetables, and bread. Wine will be served with
dinner. Pie will follow later. Chris has to budget how much this party
will cost. Chris will be paying for the party expenses with cash [credit
card]. Please help Chris estimate how much the food will cost.
The items given were assorted nuts, wine (red and white),
turkey, cranberry sauce, gravy and stuffing, salad dressing and
ingredients, vegetables, breads, pies, whipped cream, cheese, fruit,
and other incidentals. The estimation strategy was manipulated by
asking those in the holistic estimate condition to: Just estimate the
total amount based on the items below, while asking those in the
piece-meal estimate condition to Please estimate how much each
item will cost and then add them together to get the total amount.
After reading the vignette manipulating the payment mechanism
that Chris was going to use, and completing the budget estimation
task that manipulated whether the estimate was arrived at through
a piecemeal decomposition strategy or a holistic total strategy,
participants estimated the overall cost of the party, which served as
the dependent variable. They were also asked to report the average
amount they spent per month on groceries to control for hetero-
geneity in participants spending levels.
Effect of Payment Mode on PWP
$0.00
$0.50
$1.00
$1.50
$2.00
$2.50
$3.00
$3.50
$4.00
$4.50
$5.00
Present Absent
Credit card logo
F
o
o
d

P
r
i
c
e

I
n
d
e
x
:

9

i
t
e
m
s

o
p
e
n
-
e
n
d
e
d
P
W
P
Figure 1. Results of Study 1: Effect of the Presence of a Credit Card Logo on Price-Willing-to-Pay. Error bars
represent standard error of means (standard deviation/ n).
217
MONOPOLY MONEY
Results and Discussion
A 2 2 ANOVA on the estimated budget including the amount
spent per month as a covariate yielded a significant interaction, F(1,
52) 4.27, p .05,
2
.08). Means are graphically presented in
Figure 2. Consistent with our reasoning that the holistic versus piece-
meal decomposition estimation strategy affects the manner in which
credit card (but not cash) judgments are made, the effect of estimation
strategy was only significant in the credit card condition, F(1, 28)
6.22, p .05,
2
.18, and not in the cash condition, F(1, 23) .58,
p .45,
2
.03). Further, replicating previous research and the
results of Study 1, in the holistic estimate condition, participants
estimates of the overall cost of the party were significantly higher
when Chris was using a credit card to pay for the party (M$175.16,
SD 64.01) as compared to when Chris was using cash [M
$145.56, SD 27.87; F(1, 25) 4.25, p .05,
2
.15]. In
contrast, when participants used a piecemeal decomposition strategy
to estimate the overall cost of the party, estimates were directionally
lower in the credit card (M $134.00, SD 39.69) versus cash
condition, [M $163.35, SD 60.36; F(1, 26) 2.75, p .11,

2
.10]. Although not significant, the reversal in the piecemeal
decomposition strategy was not expected and should be replicated
prior to drawing inferences from it.
Study 2 replicates the findings of earlier research on piece-
meal and holistic processing (Menon, 1997) as well as credit
card versus cash estimates (Srivastava & Raghubir, 2002). It
also extends both findings to show that the increased salience of
individual payments in the piecemeal processing strategy atten-
uates observed differences across credit card and cash pur-
chases. The results show that future estimates of spending were
significantly higher in the credit card condition than in the cash
conditions, but only when the total cost of the party was
estimated holistically. The result is consistent with our conten-
tion that due to the differences in coupling of payment as well
as form, the salience of parting with money is lower in the case
of credit card purchases thereby encouraging spending. How-
ever, when the salience of parting with money is increased by
estimating the overall cost using a piecemeal decomposition
strategy, the inclination to spend is curbed and thus the differ-
ence in spending between credit cards and cash is attenuated.
Study 2 thus extends the previous research on spending behav-
ior as a function of payment mode.
Although prior literature has documented the credit card effect,
Studies 1 and 2 explored whether the salience of parting with
money underlies differences across credit card versus cash modes
of payment. However, given that credit cards differ from cash in
terms of payment coupling as well as form, Studies 3 and 4
examine spending behavior as a function of payment form only. In
particular, Study 3 examines spending behavior when an equiva-
lent amount of money is given in the form of a gift certificate
versus cash.
Study 3
Study 3 examines differences in spending behavior when indi-
viduals receive an identical amount of money either in the form of
scrip or cash. Some retailers use scrip, a prepaid amount of
money with the value clearly provided on the face of the instru-
ment. The scrip is usable in a specific store toward purchases and
any of the value that is unspent is returned in cash. In order to
Effect of Payment Mode on Expense
Estimates
$100.00
$110.00
$120.00
$130.00
$140.00
$150.00
$160.00
$170.00
$180.00
$190.00
$200.00
Holistic Decomposition
E
s
t
i
m
a
t
e
d

E
x
p
e
n
s
e

(
o
p
e
n
-
e
n
d
e
d
)
Credit Card Cash
Figure 2. Results of Study 2: Effect of Payment Mode on Expense Estimation Using Holistic and Decompo-
sition Estimation Strategies. Error bars represent standard errors of means (standard deviation/ n).
218
RAGHUBIR AND SRIVASTAVA
control for other factors except payment form, explicit instructions
emphasized that the money was being given for the specific
purpose of purchasing items on a shopping list. To the extent a gift
card (or scrip) is less transparent than cash and is thus treated as
monopoly money, the pain of paying is likely to be lower leading
to more spending when using scrip relative to when using cash.
Method
Twenty-eight students (12 males and 16 females) from an un-
dergraduate introductory marketing class (median age 21) par-
ticipated in the experiment for partial course credit. Participants
were assigned at random to the two conditions: 16 participants
were given a $50 bill and 12 participants were given a $50 scrip
certificate along with a shopping list. Participants were asked to
imagine that they were shopping for groceries from a shopping list
and were clearly informed that the grocery store would give back
change in cash, regardless of whether the payment for the groceries
had been made using scrip or cash. Thus, within the context, the
two payment forms were equally fungible (i.e., mutually substi-
tutable or interchangeable).
The items were chosen so as to be fairly typical of an under-
graduates shopping basket. Expected to cost under $20, the shop-
ping list comprised toothbrushes, canned soup, and ketchup. Par-
ticipants were given a booklet of options for products and brands
available for each of the product categories. These included the
brand name, variety, size, and price. The list within each product
category was sorted by price. Participants were given separate
sheets listing brands in each product category. There were 35
options for toothbrushes ranging from the Colgate Plus Diamond
Compact Head Soft toothbrush for $1.99 to a Sonicare Sonic
Toothbrush Personal Model with 1 Brush head for $89.99. There
were 34 options listed in the canned soup category ranging from
Campbells Cream of Mushroom priced at $0.69 for a 10.7 oz can
to Campbells Broccoli Cheese soup priced for $1.19 also for a
10.7 oz can. There were nine options in the ketchup category. The
lowest priced option was a 14 oz bottle of Heinz ketchup priced at
$1.39 and the most expensive option was the Heinz squeeze
ketchup 64 oz bottle, priced at $4.29. Participants could choose as
many brand units as they wanted within each product category.
Until after they had completed the experiment, they were unaware
that the money was not theirs to keep. This was done to capture
their behavior as realistically as possible. The dependent measure
was the average amount spent per item purchased.
Results and Discussion
There was no difference in the number of items purchased as a
function of payment form for scrip (M 6.08, SD 4.21) and
cash (M 5.50, SD 2.80) conditions, F(1, 26) .19, p .66,

2
.01. A one-way ANOVA on the average amount spent per
item purchased, including the number of items purchased from
each category as a covariate, was significantly different as a
function of payment form, F(1, 25) 5.21, p .05,
2
.17. The
covariate also exerted a significant effect, F(1, 25) 16.62, p
.05,
2
.40. Follow-up analyses to examine the direction of this
effect showed that the correlation between the average amount
spent per item and the number of items purchased was not signif-
icant (r .06, p .74) across the two payment mode condi-
tions. However, examining the correlations separately for the scrip
and cash condition revealed that while the number of items pur-
chased and the average amount spent were unrelated in the scrip
condition (r .13, p .66), they were marginally negatively
correlated in the cash condition (r .46, p .06). This implies
that in the cash condition, the more people purchased in the
product category, the cheaper the price of the individual items
purchased, whereas when they were paying by scrip they did not
appear to make this tradeoff between number of items purchased
and unit cost of the item. Another ANOVA on the total amount
spent, including the number of items purchased from each category
as a covariate, showed no significant effects of the covariate, F(1,
26) 1.16, p .29 or form, F(1, 26) .26, p .50.
In the aggregate, means across all product categories show that
participants spent more per item when they were given scrip (M
1.76, SD .42) than when they were given an equivalent amount
in cash (M 1.53, SD .36). Means are graphically presented in
Figure 3. The total amount spent was higher in the scrip condition
than in the cash condition across all product categories: Soup:
Ms 3.81 (SD 3.86) and 3.02 (SD 3.02), Toothbrush: Ms
3.38 (SD 1.18) and 3.00 (SD 1.85), Ketchup: Ms 2.23
(SD 0.96) and 1.98 (SD 1.11).
Study 3 demonstrates that participants spending behavior var-
ied as a function of payment form. In particular, participants who
were given $50 in the form of scrip spent more than participants
who were given the $50 in cash for the same purpose. Consistent
with the argument that scrip is less transparent than cash and is
treated like play money, participants spent more when paying by
scrip than when paying by cash even for frequently purchased
utilitarian goods. Although Study 3 demonstrates that people are
willing to spend more with scrip than with cash, Study 4 examines
whether the monopoly money effect manifests itself in a real
choice task. Importantly, Study 4 examines whether the difference
in spending behavior when using a gift certificate relative to cash
can be attenuated by altering the difference in transparency and
thereby the pain of paying of the payment mode.
Study 4
Study 4 examines the effect of payment form in a real choice
context where participants were given $1 either in the form of a
gift certificate or in cash. The gift certificate could be exchanged
either for cash ($1) or be used to purchase a Starburst candy
whereas the $1 bill could be either kept or be used to purchase a
Starburst candy. Consistent with the results of Study 3 and the
contention that an equivalent face value gift certificate is more
likely to be treated as play money and thus more readily spent, we
expected participants would be more likely to purchase the Star-
burst candy when they were given the $1 in the form of a gift
certificate versus cash.
Importantly, Study 4 attempts to provide insight into the under-
lying reason for the differential effect of payment form on spend-
ing behavior. The transparency of the payment form was altered by
manipulating the relative similarity between a gift certificate and
cash. In one condition, the difference in physical form was sup-
pressed by asking participants to treat the gift certificate like they
would treat cash. They were asked to place the gift certificate in
their wallet for an hour prior to the choice task. The rationale was
that storing the gift certificate in their wallet would dull the
219
MONOPOLY MONEY
difference in form. The act of storing the gift certificate in the
wallet and then having to take it out at the time of choice is also
likely to give the feeling of parting with something of value, as is
the case with cash. The goal was to explore a condition under
which the spending difference between a gift certificate and cash
would be attenuated.
However, this reasoning is not without limits. In another con-
dition, the difference in physical form was made salient again by
making a $1 bill visually prominent at the time of choice. Thus, as
in the previous condition, participants were asked to store the gift
certificate in their wallets for an hour, but at the time of choice, a
$1 bill was visually prominent. The rationale was that the differ-
ence in transparency of the payment form would be salient again
and given its physical form and appearance relative to legal tender
($1 bill) it would be treated as play money and thus be more likely
to be spent. The goal was to show that the effect would reemerge
in the condition where the gift certificate was held like cash, if the
difference between the two payment modes was highlighted at the
time of spending.
In sum, Study 4 tests the following predictions. First, as in Study
3, it was expected that participants would be more likely to spend
the $1 on purchasing the Starburst candy when they were given the
money in the form of a gift certificate versus cash. Second, we
expected that the difference in the likelihood to spend between the
gift certificate and cash would reduce when the gift certificate (as
well as cash) had been placed in the wallet for an hour prior to
making the spending decision. Third, even when the gift certificate
was placed in the wallet and stored like cash, when the difference
in payment form between the gift certificate and cash was high-
lighted at the time of choice, the difference in the likelihood of
spending between the gift certificate and cash would reemerge:
participants would be more likely to spend the $1 on purchasing
the Starburst candy when they were given the money in the form
of a gift certificate versus cash.
Method
One hundred and thirty students enrolled in an introductory
business class were assigned at random to one of six conditions in
a 2 (payment form) 3 (relative difference in transparency or
salience) between-subjects design. Payment form was manipulated
by whether participants received $1 in the form of cash (n 68)
or in the form of a gift certificate (n 62).
The cover story was that the $1 represented a token of appre-
ciation for participating in the study. All participants were in-
formed that the money was theirs to keep at the time they were
given the money. In the gift certificate condition, the text of the
certificate read
Gift Certificate: This Certificate entitles the holder to $1.00 (One
dollar only). Thank you for participating in the BA106 experiment.
Please remember to redeem your certificate for $1.00.
The gift certificate used in two of the conditions is shown in
Appendix 1. At the time of the choice task, those in the gift
certificate conditions had to choose whether they would exchange
their gift certificate for either a $1.00 bill or a Starburst candy,
whereas those in the cash condition had to decide between keeping
their $1.00 bill or a Starburst candy.
The relative difference in transparency was manipulated in three
ways. In the control condition, participants were given the $1 at
Effect of Payment Mode on Amount Spent
$0.00
$1.00
$2.00
$3.00
$4.00
$5.00
$6.00
Soup Toothbrush Ketchup Amount spent
per item
A
m
o
u
n
t

s
p
e
n
t
Scrip Cash
Figure 3. Results of Study 3: The Effect of Paying Using Scrip versus Cash in a Simulated Purchase Situation.
Error bars represent standard error of means (standard deviation/ n).
220
RAGHUBIR AND SRIVASTAVA
the end of the experimental session (approximately 5 min prior to
making their choice) either in the form of cash or a gift certificate
in an envelope. In the low difference in transparency condition,
participants were given the $1 in an envelope approximately an
hour prior to the choice task. In order to enhance the salience of
parting with money, participants were asked to take out the $1
from the envelope which was either in the form of cash or a gift
certificate and place it inside their wallets. In the high difference in
transparency condition, participants were asked to store the money
in their wallets as in the other condition. However, at the time of
the choice task, to lower the salience of parting with real money,
the difference in form between a gift certificate and cash was
highlighted by placing a $1 bill along with the Starburst candy in
front of the participants.
After completing other unrelated experimental tasks, partici-
pants came to the front of the room one at a time to hand over their
experimental materials and sign out. In all three conditions, par-
ticipants were given a choice task at the end of the experimental
session in which they could either purchase a Starburst with their
$1 or keep their $1 (exchange their gift certificate for cash in the
gift certificate condition or keep their cash in the cash condition).
The dependent measure was the average spending per person in the
six between-subjects conditions, followed by a categorical analysis
of whether or not participants purchased the Starburst with their $1
or retained their money. The Starburst used in the study cost $.95
in the cafeteria at the university where the study was conducted.
Results
A 2 3 ANOVA on the average amount spent per condition
showed a significant effect of payment form, F(1, 124) 14.44, p
.01,
2
.10, as well as a form salience interaction, F(2, 124)
4.51, p .05,
2
.07. The main effect of salience was not
significant, F(2, 124) 1.74, p .18,
2
.03. Means are graph-
ically presented in Figure 4. Replicating the results of Study 3, an
analysis of the means showed that people spent more on average in
the gift certificate condition (M$0.49) than in the $1 cash condition
(M$0.16). Further, extending Study 3 results, the interaction shows
that the difference between the average amount spent in the gift
certificate versus the cash condition was present in the control con-
dition, Ms $0.47 vs. $0.10, F(1, 38) 8.29, p .01,
2
.18,
attenuated when people had been given the money at the beginning of
the experimental hour in the form of a gift certificate versus cash,
Ms $0.24 vs. $0.26 F(1, 34) .04, p .85,
2
.001, but
reappeared when the difference in transparency was increased by
making the $1 bill visually salient at the time of the purchase decision,
Ms $0.67 vs. $0.15, F(1, 52) 20.06, p .01,
2
.28. These
results are consistent with those of Study 3.
To examine whether the effect of payment mode also manifests
on the likelihood to purchase (vs. the amount spent as examined in
Studies 13), we analyzed the likelihood of purchasing or saving
$1 as a function of experimental condition. Means are graphically
presented in Figure 5. A binary logistic regression with decision to
purchase (buy 1) as the dependent variable, and the interaction
between condition context (1 low transparency where money
was given early, 2 high transparency where money was given
early but the $1 was visible, and 3 baseline control where the
money was given late) and form of the money (1 gift certificate)
as the independent variables was significant (Nagelkerke R
2
.06,
B .40, Wald 5.58, p .05).
Effect of Payment Mode on Amount Spent
$0.00
$0.10
$0.20
$0.30
$0.40
$0.50
$0.60
$0.70
$0.80
$0.90
$1.00
$1 given at end $1 given an hour
earlier
$1 given an hour
earlier, but salient
at time of choice
Transparency of Gift Certificate
A
v
e
r
a
g
e

A
m
o
u
n
t

s
p
e
n
t
Gift Certificate Cash
Figure 4. Results of Study 4: The Effect of Spending a gift certificate versus cash as a function of the
transparency of the difference in payment modes. Error bars represent standard error of means (standard
deviation/ n).
221
MONOPOLY MONEY
To understand the nature of the interaction, we cross-tabulated
participants decision to spend or not spend on the Starburst as a
function of whether they had received a gift certificate or cash,
collapsing across the salience conditions. Participants spending
behavior, as measured by the percentage of participants who spent
the $1 on Starburst candy, was significantly different when the $1
was in the form of a gift certificate versus cash (
2
15.34, p
.001). While only 17.6% (12/68) purchased a Starburst in the $1
cash condition, the proportion of participants who purchased a
Starburst increased to 50% (31/62) in the $1 gift certificate con-
dition. In support of the contention that a gift certificate is more
likely to be spent than cash, these data demonstrate that in an
actual spending task, participants were more willing to spend the
$1 when it was in the form of a gift certificate than when it was in
the form of cash. These results are conceptually consistent with
those of Study 3, using likelihood of spending (vs. actual amount
estimated) as the dependent variable.
We further expected that the spending difference between the gift
certificate and cash would attenuate in the low difference in transpar-
ency condition (where the money was given early to enhance the
salience of parting with money), but return in the high difference in
transparency condition (when despite the money being stored in the
wallet for one hour, the context made the difference between the $1
gift certificate and $1 cash transparent to lower the salience of parting
with real money by visually highlighting the $1 cash at the time
people made their decision to save or spend).
In the baseline control condition when the $1 was given at the
time of the purchase decision, 9.5% (2/21) spent their $1 on
Starburst when it was given in cash, whereas as many as 9/19
(47.4%) spent it when it was in the form of a gift certificate (
2

7.17, p .05). However, when participants received their money

early and the $1 bill was not visible there was no difference in
spending behavior as a function of payment form as 23.5% (4/17)
of the participants in the gift certificate condition purchased a
Starburst and 26.3% (5/19) of the participants in the cash condition
purchased a Starburst (
2
0.04, ns). In contrast, when the $1 bill
was visible, despite participants having received their money
early, a higher proportion of participants in the gift certificate
condition (66.7% or 18/27) purchased a Starburst than in the cash
condition (14.8% or 4/23;
2
15.03, p .001). As predicted,
while there was no difference in spending when the gift certificate
was stored like cash in the wallet for an hour prior to the choice
task, the difference in spending reemerged when the context high-
lighted the difference in transparency between the $1 gift certifi-
cate and $1 bill.
There is an alternative explanation favoring greater spending of
gift certificates than cash when both are stored in the wallet. It is
easier for participants in the $1 cash condition to retain their
money as they do not have to take the trouble of taking the money
out of their wallet. However, participants in the $1 gift certificate
condition have to take the certificate out of their wallet regardless
of whether they are exchanging it for cash or the Starburst. Note
that the procedures favor saving cash only in the two conditions
where participants were asked to keep their endowment in their
wallets, but not in the control condition where participants made a
choice almost immediately after receiving the cash or gift certifi-
cate. This alternative explanation is examined by contrasting
Effect of Payment Mode on Likelihood of
Spending
0.0%
20.0%
40.0%
60.0%
80.0%
100.0%
$1 given at end $1 given an hour
earlier
$1 given an hour
earlier, but salient
at time of choice
Transparency of Gift Certificate
P
e
r
c
e
n
t
a
g
e

s
p
e
n
d
i
n
g

$
1

t
o

b
u
y
S
t
a
r
b
u
r
s
t
Gift Certificate Cash
Figure 5. Results of Study 4 on Likelihood of Spending a gift certificate versus cash as a function of the
transparency of the difference in payment modes. Error bars represent the standard error of the proportion
(pq/n).
222
RAGHUBIR AND SRIVASTAVA
spending behavior across the three transparency conditions, after
controlling for payment form.
In the gift certificate condition, the proportion of participants
who purchased a Starburst varied across the three conditions (
2

7.80, p .05). In the control condition, when participants were

given the $1 at the end of the session immediately prior to the
choice task, 47% (9/19) chose to purchase a Starburst. However,
the proportion of participants who purchased a Starburst reduced
to 23.5% (4/17) in the low difference in transparency condition and
this proportion increased even beyond the control condition per-
centage to 67% (18/27, p .12) in the high difference in trans-
parency condition where the $1 bill was conspicuously visible. The
significant difference in spending behavior across the two condi-
tions where the gift certificate had been stored in the wallet (but
where the context either made a $1 bill salient or did not) suggests
that while the difference in form was suppressed in the low
transparency condition (no $1 bill visible), it was made salient
again in the high transparency condition ($1 bill visible at the time
of making a choice) whereby a gift certificate was treated as play
money and was, thus, more likely to be spent than cash.
In contrast, in the cash condition, the proportion of participants
who purchased a Starburst did not vary significantly across the
three transparency conditions (
2
2.13, p .34). In the control
condition, 9.5% (2/21) purchased a Starburst, in the low difference
in transparency condition this increased to 26.3% (5/19), and in the
high difference in transparency condition 14.8% (4/27) spent their
$1. The finding that spending behavior did not vary across the
three conditions when the $1 was given in the form of cash is
reassuring because these manipulations were meant to alter the
relative transparency of the gift certificate relative to cash. Further,
the findings suggest that the $1 given to participants was not
always spent as is the case frequently with a temporary endowment
that must be spent (Arkes et al., 1994; Thaler & Johnson, 1990).
However, the results may be due to small sample sizes and should
be replicated with larger sample sizes prior to drawing any infer-
ences from a null effect.
Discussion
Corroborating the findings of Study 3, Study 4 used an actual
spending task to show that peoples spending decisions are con-
tingent on payment form. In particular, the manner in which even
a $1, a relatively small amount of money, is saved when it is in the
form of cash, but spent when it is in the form of a gift certificate
suggests that a less transparent form of money, such as a gift
certificate is more likely to be treated as play money and thus more
likely to be spent than an equivalent amount of cash. Importantly,
Study 4 provides some insights into the underlying reasons for the
spending differences across gift certificates and cash. The results
show that when the difference in payment form is suppressed, a
gift certificate is less likely to be treated as play or monopoly
money. As such, there was no difference in spending behavior
across the cash and gift certificate conditions when participants
were asked to place their endowment in their wallets. However,
when the difference in payment form is highlighted, a gift certif-
icate is treated like play money and is thus more likely to be spent
(despite the fact that making the $1 bill visible also draws attention
to the fact that the gift certificate can be exchanged for cash). In
other words, when the difference in payment form is suppressed,
the salience of parting with real money is higher for gift certifi-
cates thus curbing the urge to spend whereas when the difference
in payment form is made salient, the salience of parting with real
money is lower thus encouraging spending.
General Discussion
Given the proliferation of a variety of payment modes in the
marketplace, the primary objective of this research was to examine
differences in spending decisions and behavior as a function of
payment mode. This research explicitly distinguishes among pay-
ment modes that differ in terms of payment coupling and in terms
of payment form and argues that relative to paying by cash, other
payment modes are less transparent as one does not feel the
outflow of money as vividly. The vividness of the money outflow
leads to a higher pain of paying with cash than with other less
transparent payment modes. In other words, the pain of paying is
somewhat dulled by less transparent payment modes such as a gift
certificate or credit card thus increasing the likelihood of spending
when using these payment modes.
Four studies examined the effect of payment mode on spending
decisions and behavior. Studies 1 and 2 examined differences in
spending when the payment mode differed in terms of payment
coupling and form. Consistent with previous research, both Studies
1 and 2 demonstrated that people are willing to spend (or pay)
more when they use a credit card than when using cash. Impor-
tantly, the results of both studies suggest that the underlying reason
for the differences in spending is, at least, partly due to differences
in the pain of paying. Study 1 showed that people who anticipate
the future pain of paying were willing to pay less with a credit card
than those who do not. Providing more direct support for the
contention that pain of paying underlies differences in spending
across credit card and cash payment modes, Study 2 showed that
the difference in spending was attenuated when the future pain of
paying was made salient by having participants estimate total
expenses using a piecemeal decomposition strategy.
Studies 3 and 4 examined differences in spending when the
payment mode differed only in form. Study 3 showed that con-
sumers tend to spend more when using a $50 gift certificate than
when using $50 cash. Extending the domain to actual spending,
Study 4 showed that people were more likely to spend $1 which
was in the form of a gift certificate than when it was in the form
of cash. Importantly, providing insight into the underlying process,
Study 4 showed that when individuals were asked to store the gift
certificate like cash, the value of the gift certificate was assimilated
and the difference in transparency was reduced such that the
inclination to spend more with the gift certificate was attenuated.
However, the difference manifested itself again when the decision
context made the difference in transparency between the gift
certificate and cash payment modes salient. Together, the four
studies suggest that less transparent payment forms tend to be
treated like monopoly money and are hence more easily spent (or
parted with).
A possible limitation of Study 4 is that the amount of $1 used
was too low to inflict any pain of paying even in the cash condi-
tion. However, the results contradict this notion as participants
were clearly reluctant to part with the rather nominal sum of $1 to
purchase a Starburst, particularly when they received the money in
the form of cash. Further, the random assignment of participants to
223
MONOPOLY MONEY
the different conditions makes it unlikely that participants affinity
toward Starburst varied systematically across the conditions. In
fact, the results of Study 4 are arguably more compelling because
of the observed differences in spending behavior with the rela-
tively small sum of $1. It is possible that these effects are likely to
be stronger for large sums of money that are associated with a
higher pain of paying.
From a theoretical perspective, our findings lend support to
Prelec and Loewensteins (1998) idea that paying for goods and
services leads to an immediate pain of paying, which is balanced
against the anticipated benefits of the goods and services. While
making the benefits salient may somewhat blunt the pain of
paying, our findings suggest that even less transparent payment
modes dulls the pain of paying thus increasing the likelihood to
spend. The outflow of money is very vivid when individuals use
legal tender such as cash making it painful to part with. In contrast,
any payment mode that makes the outflow of money less vivid,
and thus less painful, reduces the psychological barrier to spend.
Importantly, the results also suggest that contextual variables that
make the pain of paying salient serve to reduce the propensity to
spend more even when using a less transparent payment mode
such as a credit card. Thus, participants in the piecemeal decom-
position estimation strategy were made to feel the full extent of the
future pain of paying which attenuated the tendency to overspend
when using a credit card than when using cash. While the effec-
tiveness of the piecemeal decomposition estimation strategy has
been shown in behavioral frequency estimates (Menon, 1997;
Raghubir & Srivastava, 2002), Study 2 demonstrates its utility in
the context of future expense estimates that have been shown to be
based on contextual cues (Menon, Raghubir, & Schwarz, 1997).
Substantively, this research contributes to our understanding of
individuals propensity to spend as a function of payment mode and
thus the increase in the variety of different payment modes in the
marketplace. For example, American Express in 2004 began focusing
on increasing sales of American Express gift cards, the sales of which
tripled from the previous year. The Incentive Federation Study of
Merchandise and Travel Incentive Users found that gift cards are
commonly offered as incentives for salespeople (78%), resellers
(57%), consumers (77%), and employees (67%).
From a consumer welfare perspective, our results suggest that
individuals are prone to biases in spending when they use nonlegal
tender. Treating nonlegal tender as play money leads to overspend-
ing that authorities can warn consumers about. In the case of credit
cards relative to cash, the actual parting of the money occurs after
the purchase decision thereby dulling the pain that is felt in a cash
purchase. In fact, the immediate gratification is much more salient
relative to the anticipated pain of paying in the future. Srivastava
and Raghubir (2002) reported that people recall their cash pay-
ments better than their credit card expenses, indicating that the full
extent of the pain of paying in the future is also not felt at the time
of purchase. To the extent people can be made to anticipate the
future pain of paying at the point of purchase, the difference in
spending behavior as a function of payment mode is likely to
diminish. In the case of gift cards relative to cash, frivolous
spending is likely to occur more with a gift card than with cash.
Although equivalent in face value, the intrinsic difference in phys-
ical form and appearance serves to anesthetize the pain of paying.
The transparency of payment may also affect how people assign
consumption activities (or spending) to different accounts (Thaler,
1999). People tend to label both resources and consumption
wherein resources may be labeled as regular versus windfall and
consumption may be labeled as necessities versus luxuries (Kivetz,
1999; Thaler, 1999). Further, there is a systematic tendency to
match mental accounts such that people prefer to pay for their
hedonic consumption with unexpected windfall resources (Thaler,
1985). Thaler (1985) noted that hedonically pleasurable luxuries
are generally underconsumed because of self-control reasons.
Since luxuries are not essential by definition, the pain of paying is
likely to be higher when consuming luxuries rather than necessi-
ties. However, different payment modes may help balance the pain
of paying and the consumption of luxuries versus necessities.
If the pain of paying increases with the transparency of payment
mode, cash payments are more likely to be used for justifiable
necessities and less likely to be used for frivolous luxuries which
may accentuate the pain of paying. In contrast, using a less
transparent form of payment such as a credit card or a gift card
lowers the vividness with which one feels that one is parting with
real money, thereby encouraging spending particularly for hedon-
ically pleasurable luxuries. Thus, the effects noted in Study 3 with
utilitarian products may be accentuated with hedonic products.
This is suggested as an area for future research.
In the case of credit cards, there are two additional reasons that
the pain of paying is dulled. First, the payment is temporally
separated from the consumption. Second, credit cards allow mix-
ing of purchases where several purchases are combined into one
payment such that a single payment is not attributable to a specific
consumption. In sum, the extent to which people match more
transparent payment forms to necessities or utilitarian consump-
tion and less transparent payment forms to frivolous luxuries,
using cash discourages spending and using a credit card or a gift
card encourages spending. Disentangling the transparency versus
matching effects of credit card spending for utilitarian and hedonic
goods is also suggested as an area for future research.
The matching of payment modes to spending decisions may lead
to schemas that people develop over time. A schema refers to
cognitive structures of organized prior knowledge that is ab-
stracted from experience over time. Schemas, once developed,
guide the processing of new information and the retrieval of stored
information. For example, Thaler (1999) noted that over time
people come to recognize that consumption of luxuries, within a
reasonable range, can enhance the quality of life and often without
significantly affecting the ability to fulfill the essential needs. In
our context, with experience over time, people may learn and
develop schemas about different payment modes and how these
may aid in balancing consumption pleasure and the pain of paying.
Less transparent payment modes such as a credit card or a gift
card, which somewhat dull the pain of paying, are more likely to
be associated with free spending and hedonic consumption
whereas a transparent payment mode such as cash is more likely to
be associated with thriftiness and utilitarian consumption.
Identifying the schemas associated with different payment modes
is also suggested as an area for future research.
Future research interested in examining these issues could ex-
amine differences in spending behavior with larger sums of money
and nonstudent populations under more naturalistic conditions
than the laboratory experiment environment allows. Further, future
research could investigate whether the differences due to the
transparency of the money also affect differences in the likelihood
224
RAGHUBIR AND SRIVASTAVA
of spending and amount of spending for new payment mechanisms
such as prepaid stored value cards, such as those offered by
American Express, Visa, phone companies, and many retailers.
References
Arkes, H. R., Joyner, C. A., Pezzo, M. V., Nash, J. G., Siegel-Jacobs, K.,
& Stone, E. (1994). The psychology of windfall gains. Organizational
Behavior & Human Decision Processes, 59, 33147.
Cole, C. (1998). Identifying interventions to reduce credit card misuse through
consumer behavior research. In Proceedings of the Marketing and Public
Policy Conference, Washington, DC: Georgetown University.
Feinberg, R. A. (1986). Credit cards as spending facilitating stimuli: A con-
ditioning interpretation. Journal of Consumer Research, 13, 348356.
Gourville, J. T. (1998). Pennies-a-day: The effect of temporal reframing on
transaction evaluation. Journal of Consumer Research, 24, 395408.
Henderson, P., & Peterson, R. (1992). Mental accounting and categorization.
Organizational Behavior and Human Decision Processes, 51, 296312.
Hirschman, E. C. (1979). Differences in consumer purchase behavior by
credit card payment system. Journal of Consumer Research, 6, 5866.
Kivetz, R. (1999). Advances in research on mental accounting and reason-
based choice. Marketing Letters, 10, 249266.
Loewenstein, G. & Prelec, D. (1992). Anomalies in intertemporal choice:
Evidence and an interpretation. The Quarterly Journal of Economics,
May, 573597.
Menon, G. (1997). Are the parts better than the whole? The effects of
decompositional questions on judgments of frequent behaviors. Journal
of Marketing Research, 34, 335346.
Menon, G., Raghubir, P., & Schwarz, N. (1997). How much will I spend?
Factors affecting Consumers estimates of future expense. Journal of
Consumer Psychology, 6, 141164.
Prelec, D., & Loewenstein, G. (1998). The Red and the Black: Mental
accounting of savings and debt. Marketing Science, 17, 428.
Prelec, D., & Simester, D. (2001). Always leave home without it: A further
investigation of the credit-card effect on willingness to pay. Marketing
Letters, 12, 512.
Raghubir, P. (2006). An information processing review of the subjective
value of money and prices. Journal of Business Research, 59, 10531062.
Raghubir, P., & Srivastava, J. (2002). Effect of face value on product valuation
in foreign currencies. Journal of Consumer Research, 29, 335347.
Shafir, E., Diamond, P., & Tversky, A. (1997). Money illusion. Quarterly
Journal of Economics, 112, 341374.
Shefrin, H. M., & Thaler, R. H. (1988). The behavioral life-cycle hypoth-
esis. Economic Inquiry, 26, 609643.
Soman, D. (2001). Effects of payment mechanism on spending behavior:
The role of rehearsal and immediacy of payments. Journal of Consumer
Research, 27, 460474.
Soman, D. (2003). The effect of payment transparency on consumption:
Quasi-experiments from the field. Marketing Letters, 14, 173183.
Srivastava, J., & Raghubir, P. (2002). Debiasing using decomposition: The
case of memory-based credit card expense estimates. Journal of Con-
sumer Psychology, 12, 253264.
Thaler, R. H. (1985). Mental accounting and consumer choice. Marketing
Science, 4, 199214.
Thaler, R. H. & Johnson, E. J. (1990). Gambling with the house money and
trying to break even: The effects of prior outcomes on risky choice.
Management Science, 36, 643660.
Thaler, R. H. (1999). Mental accounting matters. Journal of Behavioral
Decision Making, 12, 183206.
Tokunaga, H. (1993). The use and abuse of consumer credit: Applications
of psychological theory and research. Journal of Economic Psychology,
14, 285316.
Appendix
Gift Certificate Stimulus Used for Study 4
Gi f t Cer t i f i cat e
Thi s Cer t i f i cat e ent i t l es t he
hol der t o
$1. 00 ( One dol l ar
onl y)
Thank you f or par t i ci pat i ng i n
t he BA 106 Exper i ment . Pl ease
pl ace t he Cer t i f i cat e i n your
wal l et f or t he dur at i on of t he
st udy. At t he end of t he st udy,
pl ease r emember t o r edeemyour
cer t i f i cat e f or a $1. 00 not e.
Received June 14, 2007
Revision received March 3, 2008
Accepted March 5, 2008
225
MONOPOLY MONEY