The impact of contactless payment on cash usage at an early stage of diffusion

Abstract

Abstract This paper explores the impact of contactless payment on consumers’ demand for cash at an early stage of diffusion. The specific devices that are investigated are debit and credit cards, in which the feature is embedded. A novel balanced panel dataset drawn from representative surveys on consumer payment behavior in the USA from 2009 to 2013 is analyzed to account for unobserved heterogeneity in cash usage. The results show that contactless credit and debit cards exert no statistically significant effect on cash usage after controlling for unobserved heterogeneity. Consumers’ decision to use contactless payment is an endogenous choice. Card-affined individuals replace conventional card payments with contactless card payments. Hence, the overall effect on cash usage remains unaffected.

Full text

Swiss Journa
l
o
f
Economics and Statistics
Trütsch Swiss Journal of Economics and Statistics (2020) 156:5
https://doi.org/10.1186/s41937-020-00050-0
ORIGINAL ARTICLE Open Access
The impact of contactless payment on
cash usage at an early stage of diffusion
Tobias Trütsch
Abstract
This paper explores the impact of contactless payment on consumers’ demand for cash at an early stage of diffusion.
The specific devices that are investigated are debit and credit cards, in which the feature is embedded. A novel
balanced panel dataset drawn from representative surveys on consumer payment behavior in the USA from 2009 to
2013 is analyzed to account for unobserved heterogeneity in cash usage. The results show that contactless credit and
debit cards exert no statistically significant effect on cash usage after controlling for unobserved heterogeneity.
Consumers’ decision to use contactless payment is an endogenous choice. Card-affined individuals replace
conventional card payments with contactless card payments. Hence, the overall effect on cash usage remains
unaffected.
Keywords: Contactless payment, Money demand, Cash usage, Credit cards, Debit cards
JEL classification: C33, D12, E41, E42
1 Introduction
Cash is still the most prominent payment method at
the point-of-sale (POS) in numerous developed countries,
especially at low transaction values (e.g., von Kalckreuth
et al. (2014); Bouhdaoui and Bounie (2012); Arango et al.
(2015); Bagnall et al. (2016)). However, the promotion of
various technological innovations in retail payment mar-
kets such as credit, debit, and prepaid cards has led to a
decline in cash usage in recent years (e.g., Lippi and Sec-
chi (2009); Amromin and Chakravorti (2009); Stix (2003)).
Recent innovative payment means (e.g., contactless pay-
ment) attempt to mimic the desirable features of cash.
They promise efficient and convenient payment services
that may reduce the transaction costs of payment for con-
sumers. Contactless payment is therefore seen as a more
competitive payment alternative to traditional cash pay-
ments compared to conventional payment cards. Thus,
discussing the prospects of cash usage is high on the
agenda of central banks, which are responsible for cash
distribution.
Correspondence: tobias.truetsch@unisg.ch
University of St. Gallen, Holzstrasse 15, 9010 St. Gallen, Switzerland
This paper explores the effect of contactless credit and
debit cards on cash usage in the early stage of diffusion.
The contactless antenna is usually embedded in con-
ventional payment cards such as debit and credit cards.
Contactless cards include a chip and a simple wireless
sign. The sign is the only distinction to traditional chip
cards. Contactless payment is based on near-field commu-
nication (NFC) technology. This standard radio commu-
nication technology allows paying within a 4-cm range by
waving or tapping the payment card. A signature or PIN
verification is not necessary below a certain transaction
value. Contactless payment therefore offers instantaneous
payment, speed, and convenience compared to traditional
cards. Polasik et al. (2013) found that contactless payment
cards compete with cash payments with respect to speed
and under certain conditions even outperform cash. The
speed of a transaction is key to determining the choice of
payment instruments (e.g., Klee (2006); Jonker (2007)).
Thus, I hypothesize that contactless payment adopters
are more likely to exhibit lower cash usage. The effect on
the number of cash transactions is expected to be larger
comparedtocashexpenses.Thisisbecausecontactless
© The Author(s). 2020 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which
permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit
to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The
images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated
otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended
use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the
copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Trütsch Swiss Journal of Economics and Statistics (2020) 156:5 Page 2 of 35
cards—due to their improved speed and convenience—
are likely to substitute low cash value payments, which are
high in frequency but have low budget impact.
Analyzing the effect of contactless payment on cash
usage is relevant for three reasons. First, one of the main
responsibilities of central banks is to provide efficient pay-
ment services to ensure financial system stability. The
number and transaction size of cash payments affect
the efficiency of payment systems, as expressed in social
welfare costs. van Hove (2008) measured the costs of
cash usage in the Netherlands as being 0.48% of GDP.
Schmiedel et al. (2013) estimated the substantial costs of
cash, which amount to one half percent of GDP for the
EU-27 member states. Thus, understanding the demand
for cash is crucial to evaluating the costs of payment
systems.
Second, central banks are the sole institutions that are
entitled to issue legal tender money. The assessment of
future trends in cash demand is a relevant monetary pol-
icy issue. More contactless payment cards could imply
lower cash in circulation and hence lower seignorage
income.
Third, the literature has shown that money demand
might react less sensitively to interest rates due to techno-
logical improvements in payment processing. This might
result in lower welfare costs of inflation (Alvarez and Lippi
(2009))1.
Three papers have so far examined how contactless
payment impacts cash demand. Fujiki and Tanaka (2014)
found that average cash balances do not decrease with the
adoption of contactless payment and under some specifi-
cations even increase. They used household-level survey
data from Japan. Fung et al. (2014)showedthatcontact-
less credit and stored-value cards reduce average cash
usage for transactions in terms of both value and vol-
ume2. They analyzed consumer-level survey data from
Canada. However, both studies failed to purge unobserved
heterogeneity due to data restrictions. Chen et al. (2017)
used household panel data from Canada to account for
endogeneity. They encountered a high attrition rate of
about 50%. However, they applied refreshment samples to
account for this high attrition rate. They found no sta-
tistically significant impact of contactless credit cards on
cash usage, neither in terms of value nor of volume after
controlling for non-ignorable attrition and unobserved
heterogeneity.
This paper contributes to existing literature in three
respects. First, it is essential to control for unobserved
heterogeneity when examining the effect of contactless
1Welfare costs of inflation are the amount of less seigniorage revenue due to
higher nominal interest rates (real rate plus expected inflation) (Briglevics and
Schuh (2013)).
2Fung et al. (2014) estimated a decline in cash value due to contactless credit
and stored-value cards by roughly –14 and –12% and a reduction in cash
volume by around −13 and −15%, respectively.
payment on cash usage Chen et al. (2017). I draw on a
unique balanced panel dataset from 2009 to 2013. Using
such rich datasets represents a novel approach, which
does not suffer from non-ignorable attrition. Second, I
investigate the effect of contactless debit cards on cash
demand and thereby fill an important gap in the literature.
This is because debit cards are the most popular cashless
payment method. Third, I analyze the impact of contact-
less payment on cash usage in the USA, one of the biggest
payment markets. This is important as there is still miss-
ing empirical evidence of contactless payment in the USA
payment landscape.
I find evidence that contactless credit and debit cards
exert no statistically significant effect on cash usage in
the early stage of diffusion. I account for unobserved het-
erogeneity in cash usage by using the fixed-effects model.
Consumers’ decision to adopt contactless payment is an
endogenous choice. Card-affined individuals replace con-
ventional card payments with contactless card payments.
The overall effect on cash usage therefore remains unaf-
fected.
I proceed as follows. Section 2reviews the relevant liter-
ature. Section 3provides background information on the
theoretical framework of the estimation strategy as well
as the institutional background of contactless payment in
the USA. The data are described in Section 4, followed by
empirical specification in Section 5.Section6discusses
theresultswhileSection7draws conclusions and provides
a research outlook.
2 Literature review
This paper is related to the literature of money demand
and the future use of cash with regard to technological
improvements. Efforts to estimate precise parameters of
the traditional money demand function in light of tech-
nological change have produced an important body of
literature (e.g., Attanasio et al. (2002); Lippi and Secchi
(2009); Alvarez and Lippi (2009); Briglevics and Schuh
(2013)).
Some scholars have estimated the share of cash trans-
actions at the POS and its future usage with respect to
payment enhancement. The effect of payment innovations
on aggregate cash demand is not clear from an empirical
point of view. Columba (2009) studied the effect of ATMs
and POS terminals on the demand for currency and nar-
row money M1. He showed that the impact on cash in
circulation is negative, whereas it positively affects narrow
money.
Others have found that modern payment technolo-
gies have little effect on currency usage, mainly due to
its superior characteristic of anonymity. Amromin and
Chakravorti (2009) showed that demand for low denom-
ination notes and coins decreases as debit card usage
increases. This is because merchants need less purse

Trütsch Swiss Journal of Economics and Statistics (2020) 156:5 Page 3 of 35
money for change. The demand for high denomination
notes is less affected because individuals use them for
non-transactional purposes such as hoarding and illegal
activities. This was highlighted by Drehmann et al. (2004),
who pointed out that POS terminals negatively and ATMs
positively affect demand for small banknotes. Snellman et
al. (2001) argued that debit and credit cards are the main
drivers of substituting away from cash, while the effect of
ATMs remains ambiguous (cf. Humphrey (2004)).
Another strand of the literature has employed house-
hold survey data to more precisely study cash usage. Stix
(2003) found that debit cards negatively affect demand for
purse cash in Austria. von Kalckreuth et al. (2009)argued
that credit cards have no impact on the number of cash
transactions in Germany. However, Huynh et al. (2014)
reported that merchants’ acceptance of payment cards has
a substantial negative impact on the demand for cash in
Austria and Canada.
3 Background information
3.1 Theoretical background
I derive the theoretical background for the estimation
strategy and the empirical methodology used here from
McCallum and Goodfriend (1987) framework. Attana-
sio et al. (2002) presented this framework as an exten-
sion of the traditional Baumol-Tobin model (Baumol
(1952); Tobin (1956)). The extended model takes into
account innovations in transaction technologies. Accord-
ingly, individuals adopt payment innovations if the bene-
fits of adopting the technology exceed the costs. Adoption
costs of contactless payment may include (one-time) oper-
ational learning costs, monetary costs of using and adopt-
ing the payment card (e.g., annual fees, surcharges), and
the availability of contactless terminals.
Benefits of payment innovations increase with improv-
ing transaction efficiency. This makes adopting contact-
less payment more likely since it allows for a fast payment
process. Polasik et al. (2013) showed that contactless pay-
mentcardsarethefirstpaymentmethodtobefasterthan
cash. The transaction speed is one of the most important
factors to determine the choice of a payment instrument
(e.g., Klee (2006); Jonker (2007)). This is because it reduces
queue lines and thus consumers’ payment costs (Brits and-
Winder (2005)). Younger consumers in particular react
more negatively to longer payment processing than older
consumers and are therefore more likely to adopt contact-
less payment (Borzekowski and Kiser (2008)).
Benefits also tend to rise with more consumption
expenditures and higher transaction values because more
spending is subject to longer transaction times. Conse-
quently, the rate of adopting contactless payment varies by
consumers’ demographic characteristics (e.g. income, age,
education), which determine their opportunity costs of
paying. High-income individuals are therefore more likely
to adopt contactless payment to reduce their transaction
costs of paying. This is because their opportunity costs of
paying tend to be higher than for low-income individu-
als. For the same reason, cash demand tends to be lower
for contactless adopters than non-adopters because cash
payments take more time to settle than contactless pay-
ments (cf. Polasik et al. (2013)).
In general, individuals need time to undertake transac-
tions. As a form of exchange and financial innovations,
money reduces the transaction time (Attanasio et al.
(2002)). In the traditional Baumol-Tobin setting, individ-
uals face a trade-off between holding liquidity in form of
money, in order to carry out transactions, and the forgone
interest paid on deposited assets. However, in Attanasio
et al. (2002) extended version of the model, consumers
choose optimal money holdings to trade off transaction
costs against the costs of holding cash. Transaction time
costs originate from the shadow value of time and from
the “shoe-leather” costs of withdrawing cash.
Hence, consumers demand optimal money holdings by
minimizing both the transaction time costs and the for-
gone interest paid on deposited assets subject to their
consumption expenditures. Improvements in transaction
technology (e.g., contactless payment) and lower transac-
tion costs therefore lessen the demand for cash. Contact-
less payment also enables instantly accessing liquid assets
in accounts for making payments. This further reduces
the demand for cash and maximizes the return of inter-
est paid on deposited assets. Thus, higher interest rates on
deposited accounts create more incentives to park money
holdings that in turn reduce the demand for cash. Con-
versely, higher consumption expenditures increase the
demand for cash.
3.2 Institutional background
Contactless payment was first launched in the USA in
2005 by only very few issuing banks. The survey data used
in this study show that the rate of contactless card adop-
tion for credit and debit cards remained relatively stable
(at around 10%) between 2009 and 2013 (see Fig. 1). The
low adoption rate approximately agrees with actual data
about contactless card adoption provided by the Federal
Reserve System for the year 2012 (see Table 1). The actual
rate of contactless credit cards was 7% , that of deb it ca rds 8 %.
At the time, adopting contactless payment was an
endogenous decision in the USA. Only a few banks re-
issued contactless cards by default when traditional pay-
ment cards expired. Some issuers provided contactless
cards only upon request or exclusively to new customers.
These were not required to pay extra for the contactless
feature (cf. Chai (2017)).
Compared to other countries like Canada or Australia,
the adoption of contactless cards in the USA in the 2009–
2013 period failed to take off for various reasons. First, the

Trütsch Swiss Journal of Economics and Statistics (2020) 156:5 Page 4 of 35
10.3%
8.2%
10.5% 10.0%
9.1%
9.9%
11.6 % 11.3 %
10.0%
8.1%
0.02% 0.10% 0.20% 1.00%
2.00%
7.00%
17.00%
0%
2%
4%
6%
8%
10%
12%
14%
16%
18%
2009 2010 2011 2012 2013 2014 2015
Contactless Credit Contactless Debit NFC-Terminals
Fig. 1 The evolution of contactless cards and NFC terminals. Note: The shares refer to the contactless data related to their corresponding total data.
For instance, 9.1% of all credit cards and 2% of all POS terminals were contactless in 2013 in the USA. The Surveys of Consumer Payment Choice
(SCPC) provide the share of contactless cards. They no longer included any information on contactless payment after 2013. The share of contactless
terminals for the year 2015 is derived from LTP (2015), for the year 2013 from SPA (2016). The remaining fractions are computed according to the
logistic regression analysis. Other data points are not available
US banking and merchants sectors were very fragmented.
Both sectors pursued different contactless payment strate-
gies (SPA (2016)). Second, banks were legally obliged
to shift completely to EMV (Europay International,
MasterCard and VISA) standard payment cards by 2015.
These types of cards enable storing data on chips rather
than on magnetic stripes. Banks incurred significant man-
ufacturing costs to overhaul these card portfolios given
the immense US market for payment cards. Therefore,
most banks decided to issue single-interface chip cards
with no contactless antenna for saving money3. Third, and
as a consequence, US retailers had never been eager to
install contactless-enabled POS terminals due to lack of
contactless card adopters. As a result of missing accep-
tance, individuals’ contactless payment adoption lagged
behind and further led banks to slow down issuing con-
tactless payment cards (SPA (2016)).
Figure 1shows the very low level of contactless card
acceptance at the POS during the years 2009–2013, rang-
ing from roughly 0.02 to 2%. This goes hand in hand
with actual usage of contactless cards (see Table 1): Only
around 0.1% of all credit and debit card payments in
the USA were made contactless in 2009 and 2012. Such
payments accounted for approximately 0.1% of total trans-
action value. In other words, an average of less than one
payment per card was made using contactless technology
in 2012 (see Table 1). This made contactless payment a
rare novelty in terms of usage.
3Canada, for instance, skipped first generation single-interface (EMV) chip
cards and deployed contactless cards from the outset.
4Data
4.1 Source
Data are drawn from the Federal Reserve Bank of Boston,
which has conducted the Survey of Consumer Payment
Choice (SCPC) since 2008. The surveys are performed
in autumn (fourth quarter)—primarily in October—by
the RAND Corporation as unique, comprehensive, and
representative online surveys using RAND’s American
Life Panel (ALP). They provide detailed payment infor-
mation about individuals with respect to nine payment
instruments (including cash) used in the USA4.
The ALP’s sampling unit is an individual US consumer
older than 18 years, whose responses to each survey are
weighted to represent all US consumers aged 18 years
and older. The 2008 responses are not comparable due
to major revisions in the questionnaire and methodol-
ogy across years. The survey series aims to provide a
consumer-level longitudinal dataset and forms a valu-
able longitudinal balanced panel from 2009–2013 with
respect to payment choice. The surveys conducted after
2013 no longer include information about contactless pay-
ment. One thousand one hundred thirty-two respondents
completed all five surveys, which included similar and
identical questions (see Table 2)5.
Table 2depicts the number of respondents per survey
and the various panelists. It shows an annual rate of attri-
tion of roughly 10% until 2012, whereas this increased to
4These include cash, checks, money orders, traveler’s checks, debit, credit, and
prepaid cards, online banking bill payments, and bank account number
payments.
5I refer to Foster et al. (2013); Schuh and Stavins (2014) and Schuh and Stavins
(2015) for a comprehensive description of each dataset, a synopsis of the
results and detailed information about the collection process.

Trütsch Swiss Journal of Economics and Statistics (2020) 156:5 Page 5 of 35
Table 1 Actual adoption and usage of contactless payment cards in the USA
2009 in % 2012 in %
Credit cards
Number of contactless cards (m) n/a 23.35 7.0%
Contactless transaction volume (m) 20 0.10% 13 0.07%
Contactless transaction value (m) 1000 0.06% 600 n/a
Debit cards
Number of contactless cards (m) n/a 22.62 8.0%
Contactless transaction volume (m) 30 0.15% 27 0.07%
Contactless transaction value (m) 1000 0.08% 378 n/a
Average number of contactless transactions per
Credit card n/a 0.57
Debit card n/a 1.19
Source: Federal Reserve System (cf. FED (2011;2014). Newer data are not available. Contactless payments are labeled “chip” card payments in the report provided by the
Federal Reserve in 2014. “m” is millions. The shares refer to the contactless data related to their corresponding total data. For instance, 0.1% of all credit card transactions in
2009 were made contactless. In other words, contactless credit card transaction volume is divided by the total credit card transaction volume
around 35% in 2013. This is because the SCPC incorpo-
rated the novel payment diary in 2012, thus more strongly
emphasizing demographic coverage (cf. Angrisani et al.
(2015))6. The retention rate between 2009 and 2012 was
around 70% (1515 individuals). Around 90% of respon-
dents who once participated in the SCPC before 2013 also
participated in 2013 (Angrisani et al. (2015)). Among the
2169 individuals observed in 2009, 52% (1132) partici-
pated throughout (i.e., 2009–2013).
Tables 10 and 11 (see Appendix) provide first-year sum-
mary statistics of stayers participating for five consecutive
years versus attritors, in order to check whether panel
attrition is systematic. The statistics reveal that attrition is
likely to be random, i.e. exhibiting no systematic pattern7.
The SCPC asks consumers what payment instruments
they have and how often they use these instruments. The
survey employs a flexible reporting strategy to enhance
recall and to optimize the accuracy of the number of pay-
ments8. It also collects comprehensive data on consumer
cash holdings and cash withdrawal behavior. Low-value
payments tend to be more easily forgotten due to their
high frequency and low budget impact. They are mostly
effected in cash, which may lead to underreporting. Thus,
cleaning procedures were applied to identify and edit
invalid data entries for the number of monthly payments
of all payment instruments and the typical value of cash
withdrawals. The dataset also provides rich information
6The 2012 SCPC included an additional 1111 new respondents to the 2065
respondents with previous experience due to the novel payment diary. Many of
the new respondents came from demographic strata poorly represented in the
pool of respondents with previous SCPC experience (Angrisani et al. (2015.))
7Statistically, attritors significantly differ with respect to three variables
(among the 45 characteristics): They are more likely to earn between 100,000
and 124,000 USD, to be retired and to withdraw cash more frequently.
However, overall differences are suggested to be unsystematic.
8Typical periods that measure the number of payments are during a week, a
month, or a year. They are quite consistent with the implicit average that
represents consumers’ trend behavior (Schuh and Stavins (2015.))
about consumer demographic characteristics, financial
status, and the rating of payment instrument attributes.
However, there are several limitations. The 2009–2013
estimates are not consistently adjusted for seasonal vari-
ation, inflation, or item non-response (missing values).
The calendar time period of the 2009 survey also dif-
fers slightly from that of the 2010–2013 surveys. The
latest surveys are very similar in terms of size, compo-
sition and timing of the sample. Survey comparability
across years may suffer from different survey timing if cru-
cial monthly seasonal differences occurred in individual
payment behavior. Also, consumers may have underre-
ported the number of payments and withdrawals in the
years 2009–2010 (i.e., during the financial crisis and the
corresponding severe recession). The rationale is that con-
sumers generally relied more on cash payments in those
days. These may be harder to recall due to their high
frequency and low budget impact. Additionally, no longi-
tudinal sample weights are available.
4.2 Description
This section describes the 2009–2013 panel dataset used
here for estimation. The surveys specifically ask respon-
dents if one of their credit and debit cards is equipped
Table 2 Panel data structure
2009 2010 2011 2012 2013
Nr. of respondents 2173 2102 2151 2065 2089
2009–2010 panelists 1913 1913
2010–2011 panelists 1801 1801
2011–2012 panelists 1926 1926
2012–2013 panelists 1330 1330
2009–2012 panelists 1515 1515 1515 1515
2009–2013 panelists 1132 1132 1132 1132 1132
Source: Schuh and Stavins (2014) and Angrisani et al. (2015)

Trütsch Swiss Journal of Economics and Statistics (2020) 156:5 Page 6 of 35
Table 3 Adoption and usage rate of payment cards in the
2009–2013 surveys
Variable Mean SD Obs.
Contactless credit cards 0.095 0.294 5659
Contactless debit cards 0.103 0.304 5657
Credit cards 0.759 0.428 5628
Debit cards 0.78 0.414 5620
Credit card usage 0.613 0.487 5625
Debit card usage 0.63 0.483 5619
Usage describes the fact that respondents make the corresponding type of
payment at least once in a typical month. Survey weights used
with the contactless feature. This estimate is likely to be
fairly robust since the decision to adopt contactless pay-
ment is endogenous. Some consumers actively applied for
contactless cards. Unfortunately, the surveys provide no
information on the specific usage patterns of contactless
payment. Contactless adopters are labeled as innovators,
or as non-innovators, irrespective of having any payment
cards. Non-innovators are a relatively homogenous group
of payment card adopters. Roughly 76% of respondents
owned a conventional credit card and 78% a debit card
within the observed period (see Table 3)9. Credit and debit
cards were used at least once a month by around 61 and
63% of all respondents between 2009 and 2013.
Around 10% of consumers in the overall period reported
that one of their credit cards had the embedded contact-
less feature (see Table 3). Approximately 10% stated that
they possess a contactless debit card.
The surveys also collect data on consumer cash with-
drawal behavior. Consumers were asked about the amount
of cash they most often withdraw and the number of
withdrawals they usually make in a typical period (week,
month, or year). Both questions were asked for two sep-
arate withdrawal locations: the primary one, where con-
sumers most often obtain cash, and all other sources10.
Like Briglevics and Schuh (2013); this study focuses on
the figures for the primary location. These estimates tend
to be more precise. Reporting the usual rather than the
actual withdrawal amount reduces the mental burden to
compute averages of potentially diverse cash withdrawals
(cf. Briglevics and Schuh (2013)).TheSCPCalsostates
the number of cash payments and the total number of all
purchases made in a typical month at the POS. Its ratio
measures the cash share in terms of volume. This is a
robust measure towards outliers.
Table 4describes the summary statistics of the main
cash measure variables in the panel dataset. The average
9These numbers are higher in the estimation sample since only checking
account holders is considered. Additionally, more than half of total payments
in the survey were made by payment cards.
10Cash withdrawal locations include ATMs, bank tellers, check cashing stores,
cash back at retail stores, family or friends and others as well as being paid in
cash.
Table 4 Summary statistics of cash measures
Statistics Usual Nr. of Cash Cash share
withdrawal withdrawals in wallet in volume
Mean 128.845 3.716 72.586 0.355
SD 172.734 6.610 134.691 0.285
Median 80.000 2.000 40.000 0.312
Min. 0.000 0.000 0.000 0.000
P-10% 20.000 0.833 1.000 0.000
P-99% 850.000 26.089 500.000 1.000
Max. 5000.000 434.821 3500.000 1.000
Obs. 5561 5572 5577 5527
Cash management measures are reported in USD except the number of
withdrawals and cash share. The usual cash withdrawal amount and the number of
withdrawals relate to the primary location. Cash share is the ratio of the total
number of cash transactions in a typical month at the POS to the total number of all
purchases in a typical month at the POS. Survey weights used
amount of cash in wallet (73 USD) is roughly half of the
average usual withdrawal amount (130 USD). The average
number of withdrawals at the primary location per month
amounts to around 4. Roughly 36% of all POS payments
are made in cash (cash share in volume). Half of the con-
sumers reported a cash ratio both lower and higher than
28.5%. Median values of the remaining cash measures
were roughly half of the average values. This indicates
that a small number of respondents relied heavily on cash,
resulting in high standard errors. The maximum values of
the cash variables support this finding (see Table 4).
For this reason, I winsorize the usual cash amount
withdrawn, the number of withdrawals and the average
cash value in wallet at the 99% level. This enables prop-
erly analyzing the mean difference between innovators
and non-innovators. Tables 5and 6report (winsorized)
statistics of the relevant cash measures distinguished by
contactless credit and debit card innovators and non-
innovators. I also provide univariate mean comparison
tests and Wilcoxon rank-sum tests between innovators
and non-innovators in order to detect statistically signifi-
cant differences11.
Table 5shows that statistically contactless credit card
adopters significantly make fewer cash withdrawals within
a month than non-adopters (roughly 2.9 vs. 3.5). Another
notable statistical difference is that adopters also have a
9 percentage point lower cash ratio in volume than non-
adopters. Further, while their usual withdrawal amount
tends to be smaller than that of non-innovators (around
8 USD), they carry slightly more cash in wallet (+1 USD).
The Wilcoxon rank-sum test supports these results.
Statistically, contactless debit card innovators make sig-
nificantly more cash withdrawals (+ 0.8) compared to
non-innovators (see Table 6). However, they withdraw
11The Wilcoxon rank-sum test tests if the samples of innovators and
non-innovators come from populations with the same distribution.

Trütsch Swiss Journal of Economics and Statistics (2020) 156:5 Page 7 of 35
Table 5 Cash Measures of Contactless Credit Card Innovators and Non-Innovators
Innovator Non-Innovator t-Test Ranksum-Test
Variable Mean SD Med. Min. Max. Obs. Mean SD Med. Min. Max. Obs. Mean Diff. z-values
Usual Withdrawal 120.855 139.313 60.000 0.000 850.000 550 127.032 155.492 80.000 0.000 850.000 5011 -7.820 0.073*
Nr. of Withdrawals 2.934 2.872 2.000 0.000 21.741 549 3.494 3.505 2.000 0.000 21.741 5023 -0.551*** 0.000***
Cash in Wallet 68.860 92.425 40.000 0.000 500.000 554 67.282 90.598 35.000 0.000 500.000 5023 0.871 0.048**
Cash Share 0.275 0.234 0.229 0.000 1.000 544 0.363 0.289 0.319 0.000 1.000 4983 -0.088*** 0.000***
All variables are winsorized at the 99%-level except cash share. Cash management measures are reported in USD except the number of withdrawals and cash share. The usual cash withdrawal amount and the number of withdrawals
relate to the primary location. Cash share is the ratio of the total number of cash transactions in a typical month at the POS to the total number of all purchases in a typical month at the POS. Survey weights used. T-tests of mean
differences of innovators and non-innovators are displayed. Differences may stray from true values due to rounding and weighting. The Wilcoxon rank-sum test is displayed (z-values). Significance levels are denoted as *** p<0.01, **
p<0.05, * p<0.1.

Trütsch Swiss Journal of Economics and Statistics (2020) 156:5 Page 8 of 35
Table 6 Cash Measures of Contactless Debit Card Innovators and Non-Innovators
Innovator Non-Innovator t-Test Ranksum-Test
Variable Mean SD Med. Min. Max. N Mean SD Med. Min. Max. N Mean Diff. z-values
Usual Withdrawal 126.164 160.103 60.000 0.000 850.000 422 126.575 153.521 80.000 0.000 850.000 5137 -1.280 0.000***
Nr. of Withdrawals 4.114 4.247 3.000 0.000 21.741 424 3.361 3.346 2.000 0.000 21.741 5147 0.754*** 0.008***
Cash in Wallet 59.209 89.023 30.000 0.000 500.000 425 68.512 91.019 40.000 0.000 500.000 5150 -9.611 0.000***
Cash Share 0.329 0.280 0.302 0.000 1.000 424 0.357 0.286 0.312 0.000 1.000 5102 -0.026 0.000***
All variables are winsorized at the 99%-level except cash share. Cash management measures are reported in USD except the number of withdrawals and cash share. The usual cash withdrawal amount and the number of withdrawals
relate to the primary location. Cash share is the ratio of the total number of cash transactions in a typical month at the POS to the total number of all purchases in a typical month at the POS. Survey weights used. T-tests of mean
differences of innovators and non-innovators are displayed. Differences may stray from true values due to rounding and weighting. The Wilcoxon rank-sum test is displayed (z-values). Significance levels are denoted as *** p<0.01, **
p<0.05, * p<0.1.

Trütsch Swiss Journal of Economics and Statistics (2020) 156:5 Page 9 of 35
Table 7 Adoption patterns of contactless payment in the entire sample
Contactless credit cards for t;T Contactless debit cards for t;T
N-I; N-I N-I; I I; N-I I; I Multiple switcher Total
Non-innovator; non-innovator 56.81 3.03 2.03 0.62 9.62 72.11
Non-innovator; innovator 2.83 0.94 0.22 0 1.86 5.85
Innovator; non-innovator 3.21 0 1.59 0 1.24 6.04
Innovator; innovator 0.34 0.24 0 0 0.37 0.95
Multiple switcher 7.71 1.05 0.58 0.1 5.6 15.04
Total 70.92 5.26 4.42 0.72 18.69 100
Numbers are in proportions and correspond to the 2009–2013 year balanced panel. Survey weights used. N-I and Idenote non-innovators and innovators, respectively.
Missings are coded according to the value of their previous year. t=2009, 2010, 2011, 2012, 2013
lower cash amounts (−1.3 USD), have a lower average
cash amount in wallet (−9.6 USD) and a lower aver-
age cash ratio in volume (−3%) than non-innovators.
The Wilcoxon rank-sum test indicates that statistically the
medians of contactless debit card innovators and non-
innovators differ significantly.
To sum up, descriptive evidence shows that contactless
payment may reduce both the volume and the value of
cash transactions, whereas the latter primarily holds for
contactless credit cards.
Five different types of contactless payment adopters can
be defined based on the transition patterns of contactless
payment in all five consecutive years in the entire 2009–
2013 balanced panel:
1. Never-innovators (non-innovator; non-innovator);
2. Stayers (start-adopters and one-time switchers), who
start without contactless payment, eventually adopt it
within the five-year period and hold it to the end
(non-innovator; innovator);
3. Leavers (stop-adopters and one-time switchers), who
start with contactless payment and eventually dismiss
it within the 5-year period (innovator;
non-innovator);
4. Permanent innovators (innovator; innovator);
5. Multiple switchers (the rest), who switch between
adoption and non-adoption of contactless payment
one or several times within the 5-year period.
Table 7provides adoption patterns of contactless pay-
ment for these five types of adopters in the years 2009–
2013. It does so separately for contactless credit and debit
cards. The matrix both displays the total share of each
adoption type and all their possible combinations. This
enables revealing the proportions of consumers who, for
instance, simultaneously have contactless debit and credit
cards.
Overall, penetration rates of contactless credit and debit
cards are very similar. The presence of contactless pay-
ment is quite modest in the sample. Around 72 and
71% of respondents never adopted contactless credit or
debit cards (non-innovator; non-innovator). Roughly 1%
of consumers are permanent innovators of both pay-
ment cards, around 6% (credit) and 5% (debit) are stay-
ers (non-innovator; innovator), roughly 6% (credit) and
4.5% (debit) are leavers (innovator; non-innovator), and
approximately 15% (credit) and 19% (debit) are multiple
switchers12.
Table 7also provides information about multiple pay-
ment innovation adopters. Around 57% of respondents
never adopted any contactless payment card in the entire
period. 41.5% of consumers who adopted contactless
credit cards at some point also adopted contactless debit
cards within the years 2009–2013. 47.5% of one-time con-
tactless debit card adopters had contactless credit cards at
some point13. Furthermore, around 43% of all consumers
in the sample had one of the two innovations at some
stage, but only around 14% possessed both innovations at
the same time14 .
Stayers exhibited lower cash shares in volume in every
year from 2009–2013 compared to never-innovators (see
12The relatively sizable number of multiple switchers could rely on the fact
that US consumers possess on average 6.5 debit and credit cards issued by
different banks.
13The proportions are calculated using the sum of debit card innovators
among credit card innovators (11.56%) in relation to the sum of credit card
innovators (27.89%) and vice versa (13.79 vs. 29.02%).
14Table 7does not provide this information.

Trütsch Swiss Journal of Economics and Statistics (2020) 156:5 Page 10 of 35
20%
22%
24%
26%
28%
30%
32%
34%
36%
38%
40%
2009 2010 2011 2012 2013
Credit Card Stayers Credit Card Never-Innovators
Debit Card Sta
y
ers Debit Card Never-Innovators
Fig. 2 The Evolution of the cash share volume between never-Innovators and stayers. Cash share volume is the ratio of the number of cash
payments to the total number of all purchases made in a typical month at the POS
Fig. 2)15 . Cash shares also decreased over the years. This
indicates that contactless payment adopters may have
already used less cash before adopting contactless pay-
ment compared to never-adopters.
The dataset additionally provides rich information on
demographic and financial characteristics. Table 12 (see
Appendix) tabulates these statistics separately for con-
tactless credit and debit card holders. It also includes
the results of the mean comparison tests. The sample
of contactless credit card adopters is statistically signifi-
cantly more skewed towards higher income and education
brackets (see Table 12). Credit card innovators are more
frequently employed, widowed, between 55 and 64 years
old, Asian, revolvers, and home owners. However, they
are less likely to be single and black compared to non-
innovators. These mean differences are all statistically
significant.
As opposed to credit card innovators, statistically, the
sample of contactless debit card adopters is significantly
more skewed towards lower-income and lower-education
brackets (see Appendix, Table 12). They are mostly
younger, working, Black, Asian, Latino, or of another eth-
nicity, and are less likely to own a home and be retired
compared to non-innovators.
To conclude, the descriptive statistics have offered some
suggestive evidence that contactless payment is related
to reduced cash usage in terms of value and volume.
They have also revealed that assignment to the contactless
feature is likely to be non-random.
15The differences are not statistically significant. Similarly, cash measures in
value overall tend to be smaller for stayers than for never-innovators.
5 Methodology
The panel dimension of the SCPC was used to estimate
the relationship between contactless payment and cash
demand. The standard panel data model with unobserved
heterogeneity ηiis
Mit =αIijt +βXit +γYit +δRit +it +ηi,(1)
where Mit denotes the measurement of cash usage, Iijt
takes the value of one if the individual is an innovator,
i.e., a contactless payment adopter for payment method j,
where jrelates to debit and credit cards, respectively. Xit
are the observed individual characteristics and a vector
of proxies for transaction costs, as evidenced by Connolly
and Stavins (2015). Yit is the household income to proxy
for consumption expenditures16,Rit is the interest rate
for primary checking accounts and the alternative cost
of holding cash, and it is the error term for all i.αis
the parameter of interest, which measures the effect of
contactless payment on cash usage.
The variable Iijt must be strictly exogenous to obtain
an unbiased estimate of the parameter α. However, the
descriptive analysis has shown that adopting contactless
payment is likely to be a non-random decision. Some
unobserved variables may cause individuals to select into
innovation Iijt and simultaneously use less cash (cf. Fung
et al. (2014)). For instance, individuals with an affin-
ity for new technologies may be more prone to hold
lesscashandbemorelikelytousecontactlesspayment.
Thus, the estimate may be biased and inconsistent (selec-
tion bias). Omitted variables related to payment automa-
tism may also confound the estimation results (omitted
16See Appendix for the variable definition.

Trütsch Swiss Journal of Economics and Statistics (2020) 156:5 Page 11 of 35
variable bias). Payment behavior has been found to be
largely habitual (van der Horst and Matthijsen (2013)).
Unobserved individual-specific fixed-effects ηimay thus
correlate with the explanatory variable Iijt, which intro-
duces a bias into the estimation17.
Furthermore, contactless payment and cash usage may
suffer from reverse causality (cf. Fung et al. (2014)). Indi-
viduals who rely less on cash may adopt contactless
payment to meet their personal preferences for frequent
usage of payment cards, as this permits instantaneous
payment. It is thus not evident if innovation drives cash
demand or vice versa (simultaneity bias). The estimation
includes perceived characteristics of cash relative to pay-
ment cards (RCHARkit ) in order to address this issue.
However, appropriate instrumental variables would be
more fruitful.
This study uses the within-group estimator (mean-
difference model) to reduce concerns about unobserved
individual fixed-effects by exploiting the panel dimension
ofthedatatoyield
18
Mit −¯
Mi=αIijt −¯
Iij+βXit −¯
Xi+γYit −¯
Yi
+δRit −¯
Ri+(it −¯i).
(2)
The usual withdrawal amount, the cash kept in wallet,
and the number of withdrawals in a typical month—
winsorized at the 99% level—represent the variable Mit as
the parameter for transactional cash demand. These vari-
ables analyze the effect on cash value (cf. Briglevics and
Schuh (2013))19.Mit also serves as the variable for cash
share in volume.
Equation 2represents the baseline specification to be
estimated. As an additional set of controls, I included the
perceived characteristics k=security, setup, acceptance,
costs, records, and convenience of cash (RCHARkit )into
17Payment markets inherently include the two-sided market structure, where
network effects are predominant. The value of contactless payment for a
consumer depends on the number of others using this instrument. If the
critical level of users is not exceeded, merchants would not invest in
contactless payment terminals due to small economies of scale. Hence, the
adoption and usage of contactless payment may face feedback effects.
Consumers therefore choose contactless payment contingent upon the
number of available contactless terminals.
18If individual effects are random and uncorrelated with the variable Iijt ,it
leads to the random-effects model. However, employing the Hausman test on
the balanced panel for all specifications rejects the null hypothesis of the
random-effects model (test statistics are not provided). Therefore, the
random-effects model does not provide consistent estimates compared to the
fixed-effects model.
19Money holdings
M
conceptually represent cash for transactional purposes
in the classical model of cash demand. Individuals, however, do not only hold
money for spending purposes but also for hoarding and precautionary
reasons. Large cash holdings may also be motivated by anticipating large
purchases and could be related not only to retail payments but also to
in-person payments beyond POS payments. In this study, the measures of
actual cash holdings may thus differ from balances consumers held for actual
cash transactions. It is unfeasible to accurately measure cash demand without
accurate transactional-level data. However, the reported amount of cash
usually withdrawn, the cash kept in wallet, and the number of withdrawals are
closely related to transactional cash balances (cf. Briglevics and Schuh (2013)).
the second specification20.Thisisbecauseasignificant
amount of unobserved heterogeneity in cash usage can
be captured by including individuals’ perceptions on pay-
ment methods characteristics (Jonker (2007); Kim et al.
(2006); Ching and Hayashi (2010)). The third specifica-
tion controls for individuals’ primary cash withdrawal
method WMit as a proxy for the “shoe-leather”costs of
withdrawing cash (cf. Briglevics and Schuh (2013))21.
I assume that consumers must have an interest-
bearing bank account to be eligible for payment cards.
Therefore, I use the subsample of checking account
holders22. This has the advantage of eliminating the
self-selection bias into payment cards since individuals
are likely to open bank accounts to reduce transaction
costs. Also, I limit the analysis to the subsample of never-
innovators and stayers in order to distinctly appraise
the effect of contactless payment on cash demand (see
Table 7).
6 Estimation results
This section presents the estimation results of the model
specification in Eq. 2using fixed-effects (FE). Additionally,
it reports the results of the cross-sectional analysis (OLS)
using the dataset in 2013 (see Appendix). Tables 13, 15,
17, and 19 show the full set of OLS estimates for con-
tactless credit cards. Tables 14, 16, 18, and 20 exhibit
the full set of OLS estimates for contactless debit cards.
Fixed-effects and OLS estimates are compared to com-
prehend the importance of controlling for unobserved
heterogeneity.
Overall, the study reveals two major findings: first,
individual-specific fixed-effects are present and con-
tactless payment adopters positively selected. This is
because FE estimates of contactless payment are mostly
smaller than OLS estimates. The large differences in the
goodness-of-fit (R2) between the OLS and the FE mod-
els also indicate that cash usage differs more between
individuals than over time (within). Second, contactless
payment exerts no impact on individual cash payment
behavior in the early stage of diffusion. Individuals with
an affinity for payment cards are likely to replace tradi-
tional card payments in favor of contactless payments. In
this way, the effect of contactless payment on cash usage is
unaffected.
Below, I first discuss the effect of contactless credit cards
on cash usage before analyzing the impact of contactless
debit cards on cash demand.
20See Appendix for variable definition.
21Cash withdrawal methods include ATM (49%), bank teller (23%), check
casher (2%), cashback (13%), employer (4%), family (6%), and others (3%). The
figures in brackets show the share of respondents using the specific withdrawal
method. Innovators statistically significantly withdraw cash more frequently
from ATMs and less often from check cashers than non-innovators.
22Approximately 97% of individuals in the sample have a checking account.

Trütsch Swiss Journal of Economics and Statistics (2020) 156:5 Page 12 of 35
6.1 Effects of contactless credit cards
I estimate three specifications using different controls for
four types of cash measures. Table 8presents the main
results of the impact of contactless credit cards on cash
usage. The results of the full set of covariates for each
regression are reported in the Appendix23 .
I find evidence that contactless credit cards have no sta-
tistically significant effect on cash expenses and cash share
volumes in all specifications. The results are rather robust
against the inclusion of additional controls. The estimated
coefficients of contactless credit cards regarding the cash
sharevolumeandthenumberofwithdrawalsdisplaythe
expected negative signs. However, the point estimates
on the cash in wallet and usual cash withdrawn exhibit,
against expectation, a positive sign in all regressions.
The estimates of the cross-sectional analysis show that
statistically contactless credit cards significantly reduce
cash share volume by approximately 5 to 6%, holding
everthing else constant. This is around half the size esti-
mated by Fung et al. (2014) and comparable to Chen et al.
(2017), who used cross-sectional and pooled data, respec-
tively24. Unobserved heterogeneity therefore drives the
results using cross-sectional data.
6.2 Effects of contactless debit cards
Table 9displays the main estimates of the regressions
that analyze the impact of contactless debit cards on
cash usage. The full set of estimates is provided in the
Appendix25.
Statistically, contactless debit cards significantly influ-
ence the number of withdrawals (see column [1]). The
coefficient of the contactless feature has the predicted
negative sign, but is sensitive to the inclusion of additional
control variables. The estimated negative impact has a
modest magnitude of −0.9, holding everything else con-
stant. In other words, contactless debit cards induce indi-
viduals to make 0.9 fewer cash withdrawals per month.
This is a sizeable reduction of 28% compared to the aver-
age number of withdrawals per month of 3.2. However,
this effect becomes statistically irrelevant using additional
control variables. It thus proves to be non-robust.
The impact of contactless debit cards on the remaining
cash measure variables is statistically insignificant. Con-
trary to expectation, the estimated coefficients in column
(3) show a positive relationship with cash share volume
and cash in wallet. The point estimate in the usual cash
23The number of observations decreases in the estimations using control sets
one and two because withdrawal methods and some perceived characteristics
were not surveyed in 2009.
24Fung et al. (2014) and Chen et al. (2017) estimated a negative effect of
contactless credit cards on the cash share volume at roughly 13 and 8% using
cross-sectional and pooled data, respectively.
25The number of observations decreases in the estimations using control sets
one and two because withdrawal methods and some perceived characteristics
were not surveyed in 2009.
withdrawal regression has the expected negative sign. In
sum, contactless debit cards exert no impact on individual
cash payment behavior.
In the cross-sectional estimation, contactless debit cards
are associated with a statistically significant decline in
cash share volume by approximately 4 to 6%, holding
everything else constant. The results are rather insensitive
to the inclusion of additional control variables and compa-
rable to the effect of contactless credit cards. Conversely,
contactless debit cards have no statistically significant
effect on cash in value.
Overall, the findings demonstrate that consumers self-
select into contactless payment. This leads to spurious
results if unobserved heterogeneity is ignored (cf. Chen et
al. (2017)).
7Conclusion
This paper has investigated the impact of contactless
payment on cash usage in the early stage of diffusion.
Multiple layers of endogeneity are likely to be present
in this setting, which requires an appropriate estimation
strategy to obtain unbiased and consistent estimates. I
have therefore employed the within estimator using a
unique balanced panel dataset from 2009–2013 in the
USA. This has allowed eliminating individual-specific
fixed-effects.
I have found evidence that contactless credit and debit
cards exert no statistically significant effect on cash usage
in the early stage of diffusion after controlling for unob-
served heterogeneity in cash usage. Consumers non-
randomly choose contactless payment cards. Card-affined
individuals replace conventional card payments with con-
tactless card payments. The overall effect on cash usage
therefore remains unaffected.
The results are in line with previous findings (cf. Chen
et al. (2017)). It is important to control for unobserved
heterogeneity in cash usage: The findings suggest a rela-
tionship between the adoption of contactless payment and
decreasing cash usage if the endogenous choice of con-
tactless payment is not accounted for. Omitted variables
such as payment habitualization may also confound and
bias the estimates. Therefore, the effect of contactless
payment on cash demand is overestimated if unobserved
heterogeneity is ignored.
Several caveats are worth mentioning. First, the dataset
gives no insights into transactional purposes. Hence,
the true values of cash purchases need to be proxied
by the typical amount of cash per withdrawal, kept in
wallet, and the frequency of withdrawals. These mea-
sures are not perfectly equivalent to the theoretical
concept in the Baumol-Tobin framework. The differ-
ence between cash hoarding and cash usage for trans-
actional purposes is likely to result in measurement
errors. Incomplete information on the exact amount of

Trütsch Swiss Journal of Economics and Statistics (2020) 156:5 Page 13 of 35
Table 8 Summary of FE regression results of contactless credit cards
Usual cash withdrawn Number of withdrawals Cash in wallet Cash share volume
Variable Baseline Controls 1 Controls 2 Baseline Controls 1 Controls 2 Baseline Controls 1 Controls 2 Baseline Controls 1 Controls 2
(1) (2) (3) (1) (2) (3) (1) (2) (3) (1) (2) (3)
Contactless 6.766 6.855 8.906 -0.325 -0.326 -0.295 8.371 6.850 6.067 -0.011 -0.013 -0.013
Credit cards
(8.994) (11.044) (9.824) (0.355) (0.435) (0.446) (12.976) (14.477) (14.318) (0.027) (0.030) (0.030)
R20.012 0.016 0.088 0.006 0.009 0.031 0.009 0.014 0.024 0.013 0.033 0.035
Observations 3592 2865 2865 3602 2874 2873 3599 2874 2874 3556 2860 2858
Individuals 853 845 845 853 847 846 851 844 844 852 846 845
Demographics Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Relative characteristics No Yes Yes No Yes Yes No Yes Yes No Yes Yes
Withdrawal method No No Yes No No Yes No No Yes No No Yes
Cluster-robust standard errors are used and given in parentheses. Survey weights are used. Significance levels are denoted as ***p<0.01, **p<0.05, *p<0.1. The full set of FE estimates are reported in the Appendix (see Tables 21, 23, 25
and 27)

Trütsch Swiss Journal of Economics and Statistics (2020) 156:5 Page 14 of 35
Table 9 Summary of FE regression results of contactless debit cards
Usual cash withdrawn Number of withdrawals Cash in wallet Cash share volume
Variable Baseline Controls
1
Controls
2
Baseline Controls
1
Controls
2
Baseline Controls
1
Controls
2
Baseline Controls
1
Controls
2
(1) (2) (3) (1) (2) (3) (1) (2) (3) (1) (2) (3)
Contactless
debit cards
-2.250 -2.966 -5.062 -0.904* -0.942 -0.941 3.789 4.256 3.206 -0.003 0.003 0.004
(7.691) (7.751) (7.725) (0.502) (0.634) (0.625) (5.966) (7.965) (7.945) (0.033) (0.038) (0.038)
R20.012 0.011 0.067 0.008 0.015 0.041 0.011 0.014 0.020 0.013 0.028 0.030
Individuals 3874 3084 3084 3882 3090 3089 3882 3093 3093 3832 3075 3074
Observations 906 899 899 906 900 899 905 898 898 905 899 898
Demographics Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Relative
characteristics
No Yes Yes No Yes Yes No Yes Yes No Yes Yes
Withdrawal
method
No No Yes No No Yes No No Yes No No Yes
Cluster-robust standard errors are used and given in parentheses. Survey weights are used. Significance levels are denoted as ***p<0.01, **p<0.05, *p<0.1. The full set of FE
estimates are reported in the Appendix (see Tables 22, 24, 26, and 28)
checking account interest rates and household income
may also lead to measurement errors. Measuring cash
usage in terms of volume may also suffer from recall
effects. Payment diaries reporting each transaction in
detail, in conjunction with information on exact interest
rates and income, would help to obtain more accurate
results.
Second, the estimation results should be interpreted
with caution since the diffusion of contactless payment
cards and contactless-enabled terminals in the USA was
still in its infancy in the years 2009–2013. It is likely that
the effects are greater and also different if more cards are
contactless and if consumers have readier access to more
contactless-enabled terminals.
Third, appropriate instrumental variables were not
available to properly control for possible reverse causal-
ity in the estimation. Individuals who rely less on cash
may also adopt contactless payment to meet their personal
preferences for frequent usage of payment cards. Plausi-
ble instrumental variables are necessary for future work:
supply-side statistics come to mind (e.g., the number of
contactless-enabled terminals differentiated by geograph-
ical region).
Fourth, the external validity of the results could be
limited. This is because payment composition and pay-
ment infrastructure between the USA and Europe dif-
fer significantly due to their culturally and institution-
ally diverse evolution of payment systems. US Americans
pay more by payment cards while Europeans rely heav-
ily on cash as a payment means Bagnall et al. (2016).
Therefore, specific payment patterns in the two pay-
ment areas may affect the magnitude of the estimated
effects.
Appendix
Variable definition
Household income
Annual household income, as a proxy for consumption,
is surveyed as a categorical variable with 17 categories.
Interest rates of checking accounts are also reported as
categories in the SCPC. I computed the average of each
category’s bounds to convert the categories into contin-
uous variables. This makes interpreting the coefficients
straightforward. Data for the median household income
(over 200,000 USD) are drawn from the 2013 Survey of
Consumer Finances as a proxy for the top income category
(see SCF (2014)). I logarithmically transformed this vari-
able for the estimation since test statistics conclude that
the distribution of household income is skewed.
Perceived characteristics
The absolute ratings of the perceived characteristics of
cash are transformed into relative ones—as in Schuh and
Stavins (2013)—by using the following transformation:
RCHARki (j,h)≡log CHARkij
H
h=1CHARkih ,(3)
where kdescribes the six characteristics such as secu-
rity, setup, acceptance, costs, records, and convenience,
iindexes the consumer, jrelates to cash, and his credit
and debit cards. The construction is applied to every con-
sumer regardless of the adoption stage of the payment
methods. The higher the value of the variable RCHARki,
the more favorable cash is than debit and credit cards with
respect to characteristic k.Thisnormalizestheperception
of a particular attribute.
Descriptives and regression tables

Trütsch Swiss Journal of Economics and Statistics (2020) 156:5 Page 15 of 35
Table 10 First-year demographics of stayers vs. attritors
Stayers Attritors ttest
Variable Mean SD N Mean SD N Mean diff.
Income (in 1000)
< 25 0.185 0.389 1128 0.181 0.386 1039 −0.026
0.3 cm 25–49 0.358 0.480 1128 0.313 0.464 1039 −0.014
50–74 0.219 0.414 1128 0.251 0.434 1039 0.029
75–99 0.132 0.338 1128 0.116 0.321 1039 −0.025
100–124 0.037 0.188 1128 0.061 0.239 1039 0.026**
125–199 0.049 0.216 1128 0.052 0.223 1039 0.009
> 200 0.021 0.142 1128 0.025 0.158 1039 0.001
Education
< High school 0.046 0.209 1132 0.078 0.268 1041 0.023
High school 0.432 0.496 1132 0.343 0.475 1041 −0.072*
Some college 0.265 0.442 1132 0.292 0.455 1041 0.033
College 0.168 0.374 1132 0.193 0.395 1041 0.015
Post graduate 0.089 0.285 1132 0.093 0.291 1041 0.002
Employment
Working 0.789 0.408 1018 0.753 0.431 959 −0.036
Retired 0.119 0.324 1018 0.168 0.374 959 0.050**
Unemployed 0.019 0.135 1018 0.005 0.074 959 −0.013*
Other 0.062 0.242 1132 0.064 0.245 1041 −0.001
Marital status
Single 0.200 0.400 1132 0.195 0.396 1041 0.003
Married 0.622 0.485 1132 0.641 0.480 1041 0.019
Separated 0.134 0.340 1132 0.120 0.325 1041 −0.021
Widowed 0.044 0.206 1132 0.044 0.206 1041 −0.001
Age
< 25 0.099 0.299 1132 0.157 0.364 1040 0.047
25–34 0.178 0.383 1132 0.188 0.391 1040 0.026
35–44 0.186 0.389 1132 0.177 0.382 1040 −0.013
45–54 0.239 0.427 1132 0.154 0.361 1040 −0.095***
55–64 0.135 0.342 1132 0.156 0.363 1040 0.015
> 65 0.161 0.368 1132 0.167 0.373 1040 0.021
Ethnicity
White 0.733 0.443 1132 0.750 0.433 1041 0.035
Black 0.139 0.346 1132 0.098 0.298 1041 −0.064**
Asian 0.036 0.187 1132 0.030 0.171 1041 −0.0001
Latino 0.133 0.34 1132 0.168 0.374 1041 0.037
Other 0.092 0.289 1132 0.122 0.327 1041 0.031
Others
Male 0.485 0.500 1132 0.480 0.500 1041 0.004
HH members 1.316 1.528 1132 1.331 1.584 1041 0.034
Revolver 0.403 0.491 1122 0.403 0.491 1019 −0.015
Home owner 0.671 0.470 1129 0.693 0.462 1018 0.058
HH refers to household. Survey weights used. ttests of mean differences of stayers and attritors are displayed. Differences may stray from true values due to rounding and
weighting. Significance levels are denoted as ***p<0.01, **p<0.05, *p<0.1. Variables are displayed for 2009. Stayers participate five years in a row. Attritors participate in 2009
but not in all 5 years

Trütsch Swiss Journal of Economics and Statistics (2020) 156:5 Page 16 of 35
Table 11 First-year payment card and cash usage characteristics of stayers vs. attritors
Stayers Attritors ttest
Variable Mean SD N Mean SD N Mean diff.
Contactless credit cards 0.103 0.303 1131 0.089 0.285 1027 −0.015
Contactless debit cards 0.099 0.298 1129 0.130 0.336 1029 0.034
Credit cards 0.726 0.446 1131 0.717 0.451 1029 −0.023
Debit cards 0.757 0.429 1129 0.785 0.411 1028 0.029
Credit card usage 0.567 0.496 1120 0.542 0.499 1019 −0.036
Debit card usage 0.641 0.480 1118 0.671 0.470 1019 0.033
Usual withdrawal 122.368 161.119 1122 125.448 193.411 1023 3.974
Nr. of withdrawals 3.258 2.956 1123 3.757 3.710 1023 0.479*
Cash in wallet 72.698 120.903 1116 64.834 104.088 1016 −8.315
Cash share in volume 0.392 0.300 1074 0.36 0.301 968 −0.027
Usage describes the fact that respondents make the corresponding typ of payment at least once in a typical month. Survey weights used. ttests of mean differences of
stayers and attritors are displayed. Differences may stray from true values due to rounding and weighting. Significance levels are denoted as ***p<0.01, **p<0.05, *p<0.1.
Variables are displayed for 2009. Stayers participate five years in a row. Attritors participate in 2009 but not in all 5 years

Trütsch Swiss Journal of Economics and Statistics (2020) 156:5 Page 17 of 35
Table 12 Sample summary statistics
Entire sample Credit cards Debit cards
IN-Ittest I N-I ttest
Variable Mean SD N Mean Mean Mean diff. Mean Mean Mean diff.
Income (in 1000)
< 25 0.202 0.402 5645 0.140 0.209 −0.060*** 0.302 0.189 0.115***
25–49 0.28 0.449 5645 0.253 0.283 −0.029 0.285 0.280 −0.004
50–74 0.199 0.4 5645 0.201 0.199 0.003 0.169 0.203 −0.035
75–99 0.13 0.336 5645 0.133 0.129 0.001 0.097 0.134 −0.038**
100–124 0.084 0.278 5645 0.090 0.084 0.004 0.058 0.088 −0.027*
125–199 0.076 0.264 5645 0.111 0.072 0.037** 0.053 0.078 −0.024*
> 200 0.029 0.169 5645 0.072 0.025 0.044*** 0.037 0.029 −0.012
Education
< High school 0.042 0.2 5660 0.058 0.040 0.025 0.059 0.040 0.019
High school 0.379 0.485 5660 0.250 0.393 −0.132*** 0.461 0.368 0.098**
Some college 0.278 0.448 5660 0.277 0.278 −0.010 0.274 0.279 −0.015
College 0.171 0.377 5660 0.198 0.168 0.022 0.145 0.174 −0.026
Post graduate 0.13 0.336 5660 0.216 0.121 0.095*** 0.060 0.138 −0.076***
Employment
Working 0.648 0.478 5546 0.709 0.641 0.070*** 0.733 0.637 0.099***
Retired 0.208 0.406 5546 0.196 0.209 −0.015 0.126 0.218 −0.088***
Unemployed 0.062 0.242 5546 0.045 0.064 −0.018 0.080 0.060 0.020
Other 0.175 0.38 5660 0.140 0.178 −0.041** 0.158 0.177 −0.022
Marital status
Single 0.142 0.349 5660 0.110 0.145 −0.030* 0.209 0.134 0.047
Married 0.659 0.474 5660 0.674 0.657 0.010 0.610 0.664 −0.027
Separated 0.147 0.354 5660 0.145 0.148 −0.005 0.166 0.145 0.016
Widowed 0.052 0.222 5660 0.072 0.050 0.025* 0.015 0.056 −0.036***
Age
< 25 0.045 0.206 5660 0.052 0.044 0.006 0.085 0.040 0.031
25–34 0.152 0.359 5660 0.119 0.155 −0.031 0.245 0.139 0.095**
35–44 0.169 0.375 5660 0.228 0.163 0.066** 0.165 0.170 −0.004
45–54 0.246 0.431 5660 0.242 0.246 −0.010 0.271 0.243 0.041
55–64 0.182 0.386 5660 0.141 0.186 −0.042** 0.114 0.190 −0.071***
> 65 0.207 0.405 5660 0.219 0.206 0.012 0.120 0.218 −0.092***
Ethnicity
White 0.765 0.424 5660 0.740 0.768 −0.029 0.588 0.787 −0.201***
Black 0.142 0.349 5660 0.076 0.149 −0.070*** 0.198 0.133 0.065**
Asian 0.031 0.174 5660 0.104 0.024 0.074*** 0.058 0.028 0.038**
Latino 0.093 0.29 5660 0.113 0.091 0.023 0.180 0.081 0.107***
Other 0.062 0.241 5660 0.081 0.060 0.025 0.155 0.051 0.098**
Others
Male 0.452 0.498 5660 0.434 0.454 −0.026 0.507 0.445 0.050
HH members 1.276 1.566 5660 0.975 1.307 −0.353*** 1.619 1.237 0.368**
Revolver 0.419 0.493 5602 0.499 0.410 0.086*** 0.339 0.429 −0.078**
Home owner 0.71 0.454 5607 0.752 0.706 0.055* 0.499 0.737 −0.223***
N-I and Idenote non-innovators and innovators, respectively. HH refers to household. The minimum numbers equal zero for every variable. Survey weights used. ttests of
mean differences of innovators and non-innovators are displayed. Differences may stray from true values due to rounding and weighting. Significance levels are denoted as
***p<0.01, **p<0.05, *p<0.1

Trütsch Swiss Journal of Economics and Statistics (2020) 156:5 Page 18 of 35
Table 13 OLS regression results of contactless credit on usual cash withdrawn
(1) (2) (3)
Variable bse bse bse
Contactless credit −5.301 (14.613) −5.507 (14.760) −10.587 (15.396)
log(income)30.885*** (6.308) 35.589*** (6.301) 37.976*** (6.451)
Interest rate −6.433 (11.463) −5.329 (10.787) −1.799 (9.505)
Education
High school −24.227 (31.741) −23.819 (33.539) −16.421 (36.158)
Some college −14.669 (32.814) −13.565 (34.835) −10.931 (37.197)
College −33.381 (33.496) −28.655 (35.686) −20.106 (38.322)
Post graduate −10.333 (35.484) −6.421 (37.492) 1.853 (40.304)
Employment
Working −24.375* (13.123) −23.459* (13.400) −23.946** (12.172)
Retired 3.797 (24.020) 7.888 (23.649) 9.055 (23.476)
Other 31.713** (15.364) 35.325** (15.512) 33.888** (15.257)
Marital status
Single −35.151 (34.478) −32.307 (33.662) −33.348 (30.590)
Married −57.976* (32.407) −58.621* (31.320) −59.704** (29.269)
Separated −27.190 (34.106) −22.297 (33.602) −16.506 (30.831)
Age
25–34 15.620 (19.975) 22.015 (20.370) 36.509** (18.243)
35–44 11.993 (22.396) 15.083 (22.467) 34.465* (19.610)
45–54 38.200* (22.869) 36.187 (22.962) 39.305** (20.018)
55–64 36.924 (25.361) 31.116 (25.617) 34.154 (22.811)
> 65 44.779 (31.808) 30.569 (32.846) 31.962 (30.458)
Ethnicity
White −21.338 (44.445) −22.980 (45.108) −30.039 (45.024)
Black 0.618 (48.577) −5.700 (48.927) −10.741 (48.945)
Latino 23.720 (15.654) 18.551 (14.956) 21.396 (14.878)
Other −26.842 (47.219) −28.048 (48.138) -33.685 (48.045)
Other
Male 25.931*** (9.259) 27.375*** (9.215) 22.688** (8.926)
HH members 0.582 (4.399) −0.194 (4.387) −1.627 (4.491)
CC revolver −41.217*** (9.535) −38.676*** (9.681) −30.695*** (9.370)
Home owner 17.532 (12.353) 16.992 (12.130) 11.995 (11.307)
Rel. characteristics
Security 19.570*** (7.492) 14.066** (6.883)
Setup −18.465 (11.395) −32.118*** (11.052)
Acceptance 22.753 (16.290) 22.957 (16.678)
Costs −1.936 (13.064) −2.073 (12.444)
Records 13.911 (8.519) 6.240 (8.207)
Convenience 37.702*** (11.137) 33.056*** (10.417)
Withdrawal method
Bank teller 83.374*** (13.958)
Check casher 209.064*** (60.061)
Cashback −60.506*** (7.440)
Employer 103.629** (40.740)
Family 8.894 (28.750)
Other 98.102** (45.644)
Constant −154.330* (88.997) −130.304 (90.618) −206.721** (94.081)
R20.083 0.112 0.208
Individuals 1464 1452 1452
bare the point estimates and se the standard errors. Cluster-robust standard errors and survey weights are used. HH and CC refers to household and credit card, respectively.
Base category of categorical variables is lower than high school, unemployed, widowed, lower than 25 years, Asian, and ATM. Significance levels are denoted as ***p<0.01,
**p<0.05, *p<0.1

Trütsch Swiss Journal of Economics and Statistics (2020) 156:5 Page 19 of 35
Table 14 OLS regression results of contactless debit on usual cash withdrawn
(1) (2) (3)
Variable bse bse bse
Contactless debit −0.289 (15.510) 1.223 (15.558) 9.603 (14.638)
log(income)30.668*** (6.243) 35.371*** (6.236) 37.663*** (6.382)
Interest rate −6.450 (11.468) −5.359 (10.789) −1.879 (9.476)
Education
High school −23.307 (32.287) −22.802 (34.366) −14.868 (36.797)
Some college −13.768 (33.322) −12.553 (35.593) −9.257 (37.827)
College −32.630 (33.807) −27.787 (36.219) −18.655 (38.716)
Post graduate −9.745 (35.788) −5.694 (37.963) 3.112 (40.608)
Employment
Working −24.510* (13.185) −23.680* (13.460) −24.508** (12.242)
Retired 3.974 (23.963) 8.098 (23.549) 9.707 (23.396)
Other 31.836** (15.398) 35.356** (15.511) 33.521** (15.309)
Marital status
Single −35.054 (34.424) −32.225 (33.655) −32.793 (30.590)
Married −57.853* (32.409) −58.531* (31.351) −59.349** (29.342)
Separated −27.270 (34.134) −22.438 (33.636) −16.686 (30.905)
Age
25–34 15.708 (19.864) 21.953 (20.344) 36.273** (18.138)
35–44 12.183 (22.272) 15.325 (22.446) 35.682* (19.596)
45–54 38.522* (22.684) 36.575 (22.859) 40.826** (19.913)
55–64 37.198 (25.172) 31.414 (25.512) 35.368 (22.644)
> 65 44.811 (31.841) 30.574 (32.932) 32.553 (30.558)
Ethnicity
White −20.823 (44.223) −22.444 (44.902) −28.547 (44.897)
Black 1.536 (47.930) −4.914 (48.348) −9.532 (48.363)
Latino 23.581 (15.556) 18.363 (14.851) 20.854 (14.828)
Other −26.397 (46.944) −27.731 (47.925) −33.294 (47.947)
Other
Male 25.788*** (9.225) 27.224*** (9.187) 22.361** (8.901)
HH members 0.623 (4.312) −0.164 (4.309) −1.601 (4.375)
CC revolver −41.239*** (9.468) −38.696*** (9.623) −30.571*** (9.314)
Home owner 17.635 (12.448) 17.170 (12.158) 12.646 (11.296)
Rel. characteristics
Security 19.551*** (7.470) 13.900** (6.843)
Setup −18.688* (11.126) −32.062*** (10.874)
Acceptance 22.779 (16.298) 22.936 (16.708)
Costs −1.600 (13.078) −1.383 (12.515)
Records 13.805 (8.518) 5.961 (8.191)
Convenience 37.822*** (11.228) 33.467*** (10.451)
Withdrawal method
Bank teller 83.496*** (14.043)
Check casher 211.711*** (59.847)
Cashback −59.936*** (7.464)
Employer 105.128** (40.819)
Family 10.774 (29.041)
Other 98.630** (45.704)
Constant −154.078* (89.384) −130.316 (90.949) −209.598** (94.679)
R20.083 0.112 0.208
Individuals 1464 1452 1452
bare the point estimates and se the standard errors. Cluster-robust standard errors and survey weights are used. HH and CC refers to household and credit card, respectively.
Base category of categorical variables is lower than high school, unemployed, widowed, lower than 25 years, Asian, and ATM. Significance levels are denoted as ***p<0.01,
**p<0.05, *p<0.1

Trütsch Swiss Journal of Economics and Statistics (2020) 156:5 Page 20 of 35
Table 15 OLS regression results of contactless credit on number of withdrawals
(1) (2) (3)
Variable bse bse bse
Contactless credit −0.723 (0.667) −0.611 (0.619) −0.455 (0.607)
log(income)−0.816 (0.543) −0.880 (0.547) −0.804* (0.456)
Interest rate −0.216 (0.345) −0.210 (0.358) −0.132 (0.360)
Education
High school −4.241 (3.365) −1.537 (1.920) −1.858 (1.803)
Some college −5.073 (3.278) −2.369 (1.795) −2.621 (1.742)
College −4.302 (3.217) −1.714 (1.846) −1.946 (1.754)
Post graduate −4.301 (3.332) −1.636 (1.965) −1.938 (1.811)
Employment
Working 1.685* (0.999) 0.957 (0.691) 0.671 (0.637)
Retired 0.782 (1.170) −0.390 (0.587) −0.386 (0.603)
Other 0.341 (1.032) −0.550 (0.606) −0.591 (0.597)
Marital status
Single −0.181 (2.167) −0.463 (2.130) −0.464 (2.148)
Married −1.721 (1.933) −2.179 (1.859) −2.166 (1.873)
Separated −2.621 (2.038) −2.909 (2.016) −2.781 (2.041)
Age
25–34 −1.942 (2.781) −2.190 (2.776) −1.676 (2.410)
35–44 −0.964 (2.584) −0.989 (2.533) −0.581 (2.246)
45–54 −0.614 (2.681) −1.013 (2.552) −0.614 (2.277)
55–64 −1.012 (2.581) −1.166 (2.455) −0.866 (2.181)
>65 −0.767 (2.692) −1.088 (2.424) −0.561 (2.153)
Ethnicity
White 0.788 (1.041) 0.950 (1.092) 0.587 (0.952)
Black 4.407** (1.840) 3.045** (1.532) 2.715* (1.474)
Latino 0.314 (0.633) 0.484 (0.573) 0.409 (0.587)
Other 5.129** (2.397) 5.319** (2.460) 4.874** (2.155)
Other
Male 0.985* (0.586) 0.728 (0.509) 0.578 (0.463)
HH members −0.103 (0.197) −0.133 (0.203) −0.104 (0.200)
CC revolver −0.636 (0.429) −0.519 (0.433) −0.455 (0.451)
Home owner −1.527** (0.773) −1.125* (0.628) −0.962 (0.670)
Rel. characteristics
Security 0.131 (0.287) 0.051 (0.285)
Setup −0.098 (0.729) −0.098 (0.718)
Acceptance −0.937 (0.777) −0.971 (0.735)
Costs −0.547 (0.661) −0.370 (0.586)
Records −0.029 (0.423) −0.109 (0.406)
Convenience 0.456 (0.702) 0.510 (0.701)
Withdrawal method
Bank teller −0.394 (0.524)
Check casher 14.496 (11.068)
Cashback −0.019 (0.759)
Employer 1.186 (1.587)
Family −0.932 (1.013)
Other 2.537 (1.799)
Constant 18.501*** (6.682) 17.340*** (6.624) 16.594*** (5.579)
R20.092 0.086 0.111
Individuals 1464 1452 1452
bare the point estimates and se the standard errors. Cluster-robust standard errors and survey weights are used. HH and CC refers to household and credit card, respectively.
Base category of categorical variables is lower than high school, unemployed, widowed, lower than 25 years, Asian and ATM. Significance levels are denoted as *** p<0.01, **
p<0.05, * p<0.1

Trütsch Swiss Journal of Economics and Statistics (2020) 156:5 Page 21 of 35
Table 16 OLS regression results of contactless debit on number of withdrawals
(1) (2) (3)
Variable bse bse bse
Contactless debit 1.043 (1.299) 0.028 (0.899) 0.156 (0.863)
log(income)−0.840 (0.546) −0.905 (0.553) −0.820* (0.457)
Interest rate −0.232 (0.354) −0.212 (0.359) −0.134 (0.362)
Education
High school −4.155 (3.334) −1.417 (1.916) −1.775 (1.778)
Some college −4.972 (3.227) −2.250 (1.775) −2.536 (1.705)
College −4.213 (3.171) −1.613 (1.835) −1.873 (1.726)
Post graduate −4.218 (3.283) −1.551 (1.963) −1.876 (1.793)
Employment
Working 1.622* (0.961) 0.935 (0.687) 0.653 (0.630)
Retired 0.796 (1.158) −0.368 (0.589) −0.362 (0.608)
Other 0.279 (0.972) −0.540 (0.606) −0.592 (0.596)
Marital status
Single −0.137 (2.173) −0.458 (2.130) −0.450 (2.147)
Married −1.700 (1.933) −2.170 (1.859) −2.154 (1.873)
Separated −2.652 (2.035) −2.923 (2.017) −2.788 (2.042)
Age
25–34 −2.019 (2.755) −2.189 (2.748) −1.672 (2.391)
35–44 −0.903 (2.610) −0.967 (2.552) −0.546 (2.265)
45–54 −0.518 (2.720) −0.976 (2.575) −0.568 (2.299)
55–64 −0.943 (2.608) −1.138 (2.471) −0.827 (2.195)
>65 −0.731 (2.725) −1.092 (2.448) −0.548 (2.172)
Ethnicity
White 0.900 (1.025) 1.006 (1.084) 0.641 (0.945)
Black 4.444** (1.815) 3.140** (1.553) 2.783* (1.495)
Latino 0.258 (0.645) 0.466 (0.580) 0.393 (0.594)
Other 5.126** (2.426) 5.360** (2.488) 4.904** (2.175)
Other
Male 0.952 (0.580) 0.712 (0.509) 0.567 (0.463)
HH members −0.103 (0.196) −0.130 (0.201) −0.102 (0.197)
CC revolver −0.618 (0.424) −0.523 (0.437) −0.455 (0.455)
Home owner −1.463** (0.742) −1.110* (0.629) −0.943 (0.669)
Rel. characteristics
Security 0.131 (0.284) 0.048 (0.281)
Setup −0.131 (0.745) −0.112 (0.731)
Acceptance −0.934 (0.775) −0.972 (0.730)
Costs −0.509 (0.658) −0.340 (0.586)
Records −0.040 (0.423) −0.118 (0.405)
Convenience 0.467 (0.712) 0.521 (0.710)
Withdrawal method
Bank teller −0.397 (0.521)
Check casher 14.577 (11.053)
Cashback 0.005 (0.762)
Employer 1.233 (1.576)
Family −0.869 (1.015)
Other 2.551 (1.801)
Constant 18.355*** (6.748) 17.354*** (6.673) 16.527*** (5.613)
R20.093 0.085 0.110
Individuals 1464 1452 1452
bare the point estimates and se the standard errors. Cluster-robust standard errors and survey weights are used. HH and CC refers to household and credit card, respectively.
Base category of categorical variables is lower than high school, unemployed, widowed, lower than 25 years, Asian and ATM. Significance levels are denoted as ***p<0.01,
**p<0.05, *p<0.1

Trütsch Swiss Journal of Economics and Statistics (2020) 156:5 Page 22 of 35
Table 17 OLS regression results of contactless credit on cash in wallet
(1) (2) (3)
Variable b se b se b se
Contactless credit 3.606 (9.568) 4.114 (9.588) 3.644 (9.590)
log(income)22.446*** (4.547) 22.969*** (4.569) 23.999*** (4.627)
Interest rate 0.189 (4.865) −0.159 (5.080) 1.013 (5.229)
Education
High school 3.021 (19.040) 5.792 (19.713) 7.755 (20.328)
Some college −1.895 (19.759) 1.181 (20.455) 2.198 (21.104)
College 6.686 (20.460) 10.365 (21.135) 12.963 (21.838)
Post graduate −0.388 (21.309) 3.125 (21.859) 5.460 (22.622)
Employment
Working −0.812 (7.003) −1.902 (6.968) −2.606 (6.863)
Retired 4.829 (12.600) 7.901 (12.779) 8.339 (12.343)
Other 5.847 (7.943) 6.809 (7.767) 6.417 (7.663)
Marital status
Single −23.529 (16.831) −22.292 (16.930) −22.157 (16.365)
Married −37.789** (15.762) −36.988** (15.771) −36.980** (15.222)
Separated −19.454 (15.889) −16.866 (15.985) −14.790 (15.458)
Age
25–34 −8.511 (12.364) −8.941 (12.580) −2.698 (11.661)
35–44 0.027 (13.513) −0.100 (13.677) 7.076 (12.889)
45–54 10.213 (13.987) 9.080 (14.195) 11.900 (13.194)
55–64 25.764* (14.509) 23.771 (14.812) 26.734* (13.945)
> 65 27.009 (17.671) 21.299 (17.972) 23.995 (17.261)
Ethnicity
White −11.911 (15.435) −11.541 (15.553) −13.642 (15.659)
Black −13.987 (16.829) −13.339 (16.991) −14.787 (17.077)
Latino −2.992 (7.632) −2.791 (7.856) −2.092 (7.895)
Other −4.479 (17.283) −5.710 (17.424) −7.681 (17.707)
Other
Male 31.668*** (4.811) 30.825*** (4.872) 29.563*** (4.812)
HH members 1.085 (2.962) 0.912 (2.918) 0.629 (2.942)
CC revolver −15.320*** (5.121) −14.110*** (5.244) −12.023** (5.169)
Home owner 17.379*** (5.441) 15.936*** (5.395) 15.149*** (4.984)
Rel. characteristics
Security −0.793 (3.269) −2.355 (3.236)
Setup −3.409 (6.377) −7.041 (6.319)
Acceptance 1.427 (9.089) 0.830 (9.227)
Costs −8.512 (7.372) −7.981 (7.201)
Records 0.457 (4.830) −1.543 (4.720)
Convenience 12.896* (6.699) 11.800* (6.435)
Withdrawal method
Bank teller 20.917*** (6.983)
Check casher 94.271** (40.477)
Cashback −13.049** (6.096)
Employer 36.990** (15.603)
Family 4.733 (10.735)
Other 28.496 (19.593)
Constant −176.079*** (49.450) −177.989*** (50.897) −206.114*** (52.532)
R20.153 0.155 0.182
Individuals 1465 1453 1453
bare the point estimates and se the standard errors. Cluster-robust standard errors and survey weights are used. HH and CC refers to household and credit card, respectively.
Base category of categorical variables is lower than high school, unemployed, widowed, lower than 25 years, Asian, and ATM. Significance levels are denoted as ***p<0.01,
**p<0.05, *p<0.1

Trütsch Swiss Journal of Economics and Statistics (2020) 156:5 Page 23 of 35
Table 18 OLS regression results of contactless debit on cash in wallet
(1) (2) (3)
Variable b se b se b se
Contactless debit −4.297 (8.632) −2.101 (8.698) 0.426 (8.579)
log(income)22.570*** (4.533) 23.138*** (4.532) 24.163*** (4.586)
Interest rate 0.261 (4.868) −0.116 (5.072) 1.036 (5.223)
Education
High school 2.530 (19.833) 5.129 (20.836) 7.041 (21.495)
Some college −2.438 (20.582) 0.488 (21.607) 1.440 (22.324)
College 6.204 (21.106) 9.803 (22.082) 12.375 (22.850)
Post graduate −0.831 (21.867) 2.645 (22.718) 4.967 (23.545)
Employment
Working −0.532 (7.003) −1.727 (6.989) −2.544 (6.873)
Retired 4.759 (12.582) 7.709 (12.759) 8.146 (12.309)
Other 6.109 (8.026) 6.877 (7.822) 6.326 (7.713)
Marital status
Single −23.694 (16.853) −22.452 (16.982) −22.318 (16.391)
Married −37.883** (15.795) −37.085** (15.815) -37.094** (15.245)
Separated −19.301 (15.917) −16.765 (16.034) −14.785 (15.485)
Age
25–34 −8.175 (12.195) −8.866 (12.419) −2.950 (11.503)
35–44 −0.213 (13.365) −0.347 (13.579) 6.858 (12.795)
45–54 9.806 (13.872) 8.696 (14.121) 11.577 (13.133)
55–64 25.473* (14.316) 23.488 (14.655) 26.461* (13.766)
> 65 26.882 (17.618) 21.233 (17.976) 23.932 (17.258)
Ethnicity
White −12.432 (15.334) −11.990 (15.489) −14.012 (15.560)
Black −14.257 (16.713) −13.823 (16.918) −15.384 (16.975)
Latino −2.734 (7.618) −2.590 (7.836) −1.968 (7.864)
Other −4.511 (17.154) −5.844 (17.351) −7.955 (17.618)
Other
Male 31.826*** (4.800) 30.983*** (4.856) 29.676*** (4.791)
HH members 1.076 (2.949) 0.895 (2.914) 0.601 (2.933)
CC revolver −15.400*** (5.098) −14.107*** (5.223) −11.953** (5.141)
Home owner 17.090*** (5.481) 15.767*** (5.434) 15.089*** (5.014)
Rel. characteristics
Security −0.734 (3.257) −2.321 (3.222)
Setup −3.304 (6.158) −6.769 (6.106)
Acceptance 1.383 (9.081) 0.816 (9.215)
Costs −8.711 (7.340) −8.160 (7.192)
Records 0.583 (4.814) −1.448 (4.716)
Convenience 12.647* (6.708) 11.532* (6.438)
Withdrawal method
Bank teller 20.921*** (7.011)
Check casher 93.853** (40.348)
Cashback −13.221** (6.138)
Employer 36.721** (15.692)
Family 4.369 (10.858)
Other 28.434 (19.534)
Constant −175.522*** (49.091) −177.964*** (50.477) −206.167*** (52.098)
R20.153 0.155 0.182
Individuals 1466 1454 1454
bare the point estimates and se the standard errors. Cluster-robust standard errors and survey weights are used. HH and CC refers to household and credit card, respectively.
Base category of categorical variables is lower than high school, unemployed, widowed, lower than 25 years, Asian, and ATM. Significance levels are denoted as ***p<0.01,
**p<0.05, *p<0.1

Trütsch Swiss Journal of Economics and Statistics (2020) 156:5 Page 24 of 35
Table 19 OLS regression results of contactless credit on cash share volume
(1) (2) (3)
Variable b se b se b se
Contactless credit −0.053** (0.025) −0.052** (0.024) −0.059** (0.023)
log(income)−0.019 (0.012) −0.009 (0.012) −0.008 (0.011)
Interest rate −0.041*** (0.016) −0.032** (0.013) −0.029** (0.012)
Education
High school −0.026 (0.073) −0.062 (0.077) −0.047 (0.071)
Some college −0.026 (0.071) −0.062 (0.075) −0.052 (0.070)
College −0.049 (0.071) −0.076 (0.076) −0.063 (0.070)
Post graduate −0.042 (0.071) −0.080 (0.076) −0.067 (0.072)
Employment
Working −0.012 (0.032) −0.004 (0.031) −0.013 (0.029)
Retired −0.059 (0.038) −0.036 (0.035) −0.047 (0.034)
Other 0.013 (0.030) 0.021 (0.029) 0.017 (0.028)
Marital status
Single 0.097* (0.051) 0.091* (0.051) 0.092** (0.044)
Married 0.059 (0.044) 0.055 (0.044) 0.049 (0.037)
Separated 0.037 (0.047) 0.038 (0.047) 0.040 (0.040)
Age
25–34 0.042 (0.065) 0.063 (0.065) 0.064 (0.060)
35–44 0.093 (0.067) 0.099 (0.067) 0.103* (0.061)
45–54 0.118* (0.065) 0.126* (0.065) 0.123** (0.060)
55–64 0.144** (0.066) 0.144** (0.066) 0.143** (0.061)
> 65 0.185*** (0.072) 0.177** (0.070) 0.183*** (0.065)
Ethnicity
White −0.022 (0.066) −0.046 (0.066) −0.064 (0.064)
Black 0.017 (0.075) −0.005 (0.074) −0.021 (0.073)
Latino 0.008 (0.027) 0.004 (0.027) 0.008 (0.026)
Other 0.034 (0.081) 0.014 (0.078) −0.013 (0.077)
Other
Male 0.066*** (0.018) 0.064*** (0.017) 0.056*** (0.017)
HH members 0.002 (0.007) 0.000 (0.007) −0.001 (0.007)
CC revolver −0.053*** (0.016) −0.053*** (0.015) −0.045*** (0.015)
Home owner −0.058*** (0.020) −0.057*** (0.020) −0.059*** (0.019)
Rel. characteristics
Security −0.005 (0.011) −0.008 (0.010)
Setup 0.043* (0.022) 0.032 (0.021)
Acceptance −0.066** (0.033) −0.064** (0.031)
Costs 0.066** (0.026) 0.061** (0.025)
Records 0.008 (0.016) 0.004 (0.015)
Convenience 0.118*** (0.024) 0.116*** (0.022)
Withdrawal method
Bank teller 0.027 (0.021)
Check casher 0.126 (0.170)
Cashback −0.086*** (0.019)
Employer 0.209*** (0.078)
Family −0.101* (0.060)
Other −0.088* (0.046)
Constant 0.442** (0.189) 0.499*** (0.190) 0.497*** (0.171)
R20.083 0.134 0.173
Individuals 1475 1463 1463
bare the point estimates and se the standard errors. Cluster-robust standard errors and survey weights are used. HH and CC refers to household and credit card, respectively.
Base category of categorical variables is lower than high school, unemployed, widowed, lower than 25 years, Asian, and ATM. Significance levels are denoted as ***p<0.01,
**p<0.05, *p<0.1

Trütsch Swiss Journal of Economics and Statistics (2020) 156:5 Page 25 of 35
Table 20 OLS regression results of contactless debit on cash share volume
(1) (2) (3)
Variable b se b se b se
Contactless debit −0.062** (0.026) −0.047* (0.025) −0.043* (0.024)
log(income)−0.021* (0.012) −0.011 (0.012) −0.011 (0.011)
Interest rate −0.041** (0.016) −0.032** (0.013) −0.029** (0.012)
Education
High school −0.015 (0.071) −0.048 (0.077) −0.033 (0.072)
Some college −0.017 (0.069) −0.050 (0.075) −0.039 (0.070)
College −0.041 (0.070) −0.066 (0.076) −0.051 (0.071)
Post graduate −0.036 (0.070) −0.071 (0.076) −0.057 (0.072)
Employment
Working −0.011 (0.032) −0.005 (0.030) −0.015 (0.029)
Retired −0.057 (0.037) −0.035 (0.035) −0.045 (0.034)
Other 0.019 (0.029) 0.025 (0.028) 0.021 (0.027)
Marital status
Single 0.095* (0.051) 0.087* (0.050) 0.089** (0.044)
Married 0.060 (0.043) 0.054 (0.044) 0.049 (0.037)
Separated 0.037 (0.047) 0.037 (0.047) 0.039 (0.040)
Age
25–34 0.047 (0.064) 0.065 (0.065) 0.066 (0.060)
35–44 0.092 (0.067) 0.097 (0.067) 0.101 (0.062)
45–54 0.117* (0.065) 0.124* (0.065) 0.123** (0.061)
55–64 0.143** (0.066) 0.143** (0.066) 0.143** (0.062)
> 65 0.183** (0.072) 0.173** (0.071) 0.180*** (0.067)
Ethnicity
White −0.020 (0.066) −0.043 (0.066) −0.060 (0.065)
Black 0.030 (0.074) 0.007 (0.074) −0.008 (0.073)
Latino 0.009 (0.027) 0.005 (0.027) 0.008 (0.026)
Other 0.042 (0.080) 0.020 (0.078) −0.006 (0.077)
Other
Male 0.066*** (0.018) 0.064*** (0.017) 0.056*** (0.017)
HH members 0.003 (0.007) 0.001 (0.007) −0.001 (0.007)
CC revolver −0.054*** (0.016) −0.054*** (0.015) −0.046*** (0.015)
Home owner −0.059*** (0.020) −0.057*** (0.020) -0.059*** (0.020)
Rel. characteristics
Security −0.003 (0.011) −0.006 (0.010)
Setup 0.038* (0.022) 0.027 (0.022)
Acceptance −0.065** (0.033) −0.063** (0.031)
Costs 0.071*** (0.027) 0.066*** (0.026)
Records 0.008 (0.016) 0.004 (0.015)
Convenience 0.115*** (0.024) 0.113*** (0.022)
Withdrawal method
Bank teller 0.024 (0.021)
Check casher 0.129 (0.170)
Cashback −0.083*** (0.019)
Employer 0.210*** (0.077)
Family −0.097 (0.060)
Other −0.087* (0.046)
Constant 0.455** (0.191) 0.507*** (0.191) 0.503*** (0.173)
R20.084 0.132 0.170
Individuals 1476 1464 1464
bare the point estimates and se the standard errors. Cluster-robust standard errors and survey weights are used. HH and CC refers to household and credit card, respectively.
Base category of categorical variables is lower than high school, unemployed, widowed, lower than 25 years, Asian, and ATM. Significance levels are denoted as ***p<0.01,
**p<0.05, *p<0.1

Trütsch Swiss Journal of Economics and Statistics (2020) 156:5 Page 26 of 35
Table 21 FE regression results of contactless credit on usual cash withdrawn
(1) (2) (3)
Variable b se b se b se
Contactless credit 6.766 (8.994) 6.855 (11.044) 8.906 (9.824)
log(income)14.122** (5.529) 13.766** (6.516) 13.252** (6.046)
Interest rate −7.159 (4.711) −6.441 (6.039) −7.647 (5.589)
Education
High school 43.406*** (14.167) 122.086*** (14.509) 85.841*** (15.074)
Some college 32.472 (34.641) 89.550** (38.303) 54.721* (33.044)
College 12.856 (40.084) 77.117* (44.403) 48.044 (37.261)
Post graduate 108.020 (96.184) 158.726 (114.876) 129.450 (103.538)
Employment
Working 0.219 (7.063) 1.280 (8.799) −6.703 (8.585)
Retired 16.541* (9.485) 2.175 (15.467) 2.803 (12.956)
Other 11.452 (9.673) −5.453 (11.630) −2.332 (9.974)
Marital status
Single −13.859 (26.602) 23.607 (20.204) 19.462 (16.548)
Married −32.397 (24.074) −4.865 (12.767) −1.174 (10.759)
Separated −64.106** (32.425) −39.145 (27.394) -32.862 (26.563)
Age
25–34 −2.776 (24.980) −54.813 (50.620) −42.631 (31.037)
35–44 7.863 (31.520) −43.913 (56.847) −30.950 (38.648)
45–54 30.710 (33.577) −29.556 (58.464) −20.446 (40.516)
55–64 22.564 (36.222) −54.677 (60.666) −43.059 (43.378)
> 65 22.965 (38.734) −63.749 (63.070) −54.154 (46.441)
Other
HH members 1.347 (4.757) −0.497 (5.104) −0.563 (4.871)
CC revolver −8.449 (5.819) −7.186 (7.256) −4.161 (6.674)
Home owner −0.779 (10.929) −6.971 (13.889) −6.730 (13.816)
Rel. characteristics
Security −4.807 (3.289) −4.943 (3.177)
Setup 11.827** (5.155) 6.380 (5.042)
Acceptance 2.090 (9.564) 2.860 (8.746)
Costs −4.646 (7.541) −6.628 (7.101)
Records −6.953 (5.380) −4.379 (4.876)
Convenience −4.244 (7.456) −3.704 (7.332)
Withdrawal method
Bank teller 59.979*** (11.149)
Check casher 167.391*** (56.741)
Cashback −32.326*** (7.362)
Employer 96.928** (42.046)
Family −19.688 (19.508)
Other 93.385*** (27.951)
Constant −68.904 (73.526) −84.921 (93.207) −74.108 (80.500)
R20.012 0.016 0.088
Observations 3592 2865 2865
Individuals 853 845 845
FE is the fixed-effects estimator obtained on the balanced panel. bare the point estimates and se the standard errors. Cluster-robust standard errors are used. HH and CC
refers to household and credit card, respectively. Base category of categorical variables is lower than high school, unemployed, widowed, lower than 25 years, Asian, and
ATM. Significance levels are denoted as ***p<0.01, **p<0.05, *p<0.1

Trütsch Swiss Journal of Economics and Statistics (2020) 156:5 Page 27 of 35
Table 22 FE regression results of contactless debit on usual cash withdrawn
(1) (2) (3)
Variable b se b se b se
Contactless debit −2.250 (7.691) −2.966 (7.751) −5.062 (7.725)
log(income)13.180*** (5.052) 14.032** (6.080) 14.064** (5.722)
Interest rate −6.115 (4.419) −5.534 (5.620) −5.703 (5.316)
Education
High school 43.297*** (16.342) 110.610*** (14.186) 73.845*** (15.433)
Some college 79.346*** (23.743) 177.691*** (57.411) 89.593* (47.979)
College 50.543 (31.071) 153.333** (60.971) 71.590 (51.456)
Post graduate 134.919* (81.733) 239.194** (120.869) 154.730 (109.015)
Employment
Working −2.741 (7.396) −4.863 (8.944) −9.299 (8.963)
Retired 19.282* (10.956) 5.419 (15.423) 5.857 (13.374)
Other 8.306 (8.753) −6.275 (11.157) −5.310 (9.909)
Marital status
Single −16.405 (25.948) 15.137 (18.551) 15.823 (16.334)
Married −35.387 (23.518) −9.051 (12.606) −3.195 (10.892)
Separated −58.642* (30.508) −30.496 (23.113) −21.275 (22.700)
Age
25–34 27.552* (15.736) −2.481 (9.766) 0.370 (12.142)
35–44 45.886* (23.834) 15.797 (28.556) 21.034 (26.914)
45–54 65.007** (26.469) 26.038 (31.001) 26.697 (29.051)
55–64 63.898** (29.855) 12.884 (35.174) 15.375 (33.524)
> 65 71.227** (32.871) 4.624 (37.857) 5.884 (36.258)
Other
HH members −0.524 (4.680) −0.951 (4.890) −0.696 (4.738)
CC revolver −7.442 (5.676) −3.239 (6.892) −1.042 (6.369)
Home owner −6.778 (10.660) −17.967 (13.238) −17.593 (13.200)
Rel. characteristics
Security -5.028 (3.208) −4.280 (3.126)
Setup 10.431** (4.616) 7.110 (4.588)
Acceptance 3.481 (8.858) 4.735 (8.138)
Costs 1.605 (7.332) 0.797 (6.999)
Records −1.113 (5.140) −0.076 (4.793)
Convenience −4.070 (6.906) −4.625 (6.730)
Withdrawal method
Bank teller 55.641*** (10.210)
Check casher 137.978** (67.915)
Cashback −26.767*** (6.760)
Employer 67.020 (43.680)
Family −11.546 (19.602)
Other 95.660*** (28.368)
Constant −118.698* (67.890) −187.422** (92.827) −134.658 (83.228)
R20.012 0.011 0.067
Observations 3874 3084 3084
Individuals 906 899 899
FE is the fixed-effects estimator obtained on the balanced panel. bare the point estimates and se the standard errors. Cluster-robust standard errors are used. HH and CC
refers to household and credit card, respectively. Base category of categorical variables is lower than high school, unemployed, widowed, lower than 25 years, Asian, and
ATM. Significance levels are denoted as ***p<0.01, **p<0.05, *p<0.1

Trütsch Swiss Journal of Economics and Statistics (2020) 156:5 Page 28 of 35
Table 23 FE regression results of contactless credit on number of withdrawals
(1) (2) (3)
Variable b se b se b se
Contactless credit −0.325 (0.355) −0.326 (0.435) −0.295 (0.446)
log(income)0.290* (0.152) 0.193 (0.198) 0.236 (0.196)
Interest rate 0.203 (0.131) 0.282 (0.184) 0.237 (0.184)
Education
High school 1.449*** (0.391) 1.490*** (0.354) 1.996*** (0.382)
Some college 1.229 (1.358) 1.101 (1.889) 1.803 (2.043)
College 1.800 (1.601) 1.751 (2.139) 2.333 (2.274)
Post graduate 1.793 (1.497) 1.699 (2.123) 2.274 (2.254)
Employment
Working −0.058 (0.292) 0.199 (0.364) 0.236 (0.354)
Retired −0.031 (0.415) −0.040 (0.742) 0.096 (0.742)
Other −0.008 (0.315) 0.237 (0.474) 0.170 (0.469)
Marital status
Single −0.645 (0.711) 0.060 (0.843) 0.035 (0.791)
Married −0.838 (0.577) −0.772 (0.600) −0.780 (0.571)
Separated −0.951 (0.782) −0.952 (0.831) −0.907 (0.812)
Age
25–34 0.826 (0.538) 0.618 (1.199) 0.551 (1.097)
35–44 1.439** (0.653) 1.504 (1.261) 1.458 (1.149)
45–54 1.691** (0.752) 1.615 (1.357) 1.558 (1.242)
55–64 1.688** (0.783) 1.465 (1.394) 1.321 (1.280)
> 65 1.290 (0.862) 0.999 (1.473) 0.736 (1.359)
Other
HH members 0.007 (0.091) 0.090 (0.118) 0.063 (0.115)
CC revolver 0.136 (0.231) −0.117 (0.262) −0.151 (0.255)
Home owner 0.059 (0.418) 0.260 (0.507) 0.259 (0.498)
Rel. characteristics
Security 0.022 (0.088) −0.005 (0.088)
Setup −0.057 (0.220) −0.076 (0.219)
Acceptance −0.322 (0.266) −0.294 (0.257)
Costs 0.074 (0.187) 0.074 (0.186)
Records 0.033 (0.138) −0.010 (0.134)
Convenience −0.034 (0.163) −0.024 (0.162)
Withdrawal method
Bank teller −0.657*** (0.223)
Check casher −1.107 (0.828)
Cashback 0.353 (0.299)
Employer 1.245 (0.937)
Family 0.391 (0.459)
Other 1.723*** (0.576)
Constant −2.080 (2.134) −1.437 (2.954) −2.437 (2.952)
R20.006 0.009 0.031
Observations 3602 2874 2873
Individuals 853 847 846
FE is the fixed-effects estimator obtained on the balanced panel. bare the point estimates and se the standard errors. Cluster-robust standard errors are used. HH and CC
refers to household and credit card, respectively. Base category of categorical variables is lower than high school, unemployed, widowed, lower than 25 years, Asian, and
ATM. Significance levels are denoted as ***p<0.01, **p<0.05, *p<0.1

Trütsch Swiss Journal of Economics and Statistics (2020) 156:5 Page 29 of 35
Table 24 FE regression results of contactless debit on number of withdrawals
(1) (2) (3)
Variable b se b se b se
Contactless debit −0.904* (0.502) −0.942 (0.634) −0.941 (0.625)
log(income)0.186 (0.131) 0.151 (0.186) 0.179 (0.185)
Interest rate 0.180 (0.119) 0.195 (0.164) 0.171 (0.165)
Education
High school 1.397*** (0.281) 1.366*** (0.319) 1.794*** (0.345)
Some college 1.176 (1.763) −5.620 (5.910) −5.640 (6.675)
College 2.032 (1.993) −4.699 (6.010) −4.814 (6.763)
Post graduate 2.141 (1.893) −4.665 (6.004) −4.771 (6.758)
Employment
Working −0.171 (0.258) 0.112 (0.322) 0.096 (0.310)
Retired −0.442 (0.322) −0.378 (0.616) −0.231 (0.605)
Other −0.149 (0.215) −0.178 (0.312) −0.246 (0.310)
Marital status
Single −0.182 (0.864) 0.729 (1.050) 0.685 (1.027)
Married −0.925 (0.578) −0.718 (0.591) −0.725 (0.562)
Separated −1.370 (0.845) −1.189 (0.922) −1.201 (0.909)
Age
25–34 0.377 (0.443) −0.167 (0.892) −0.220 (0.916)
35–44 0.675 (0.567) 0.491 (0.947) 0.490 (0.948)
45–54 0.880 (0.687) 0.472 (1.085) 0.444 (1.075)
55–64 0.779 (0.743) 0.408 (1.155) 0.337 (1.145)
> 65 0.612 (0.841) 0.215 (1.268) 0.033 (1.255)
Other
HH members 0.074 (0.093) 0.125 (0.116) 0.108 (0.113)
CC revolver 0.127 (0.199) −0.033 (0.249) −0.054 (0.244)
Home owner −0.034 (0.391) 0.040 (0.472) 0.048 (0.461)
Rel. characteristics
Security 0.054 (0.084) 0.027 (0.083)
Setup −0.056 (0.191) −0.086 (0.189)
Acceptance −0.168 (0.224) −0.140 (0.212)
Costs 0.098 (0.168) 0.103 (0.166)
Records 0.098 (0.128) 0.060 (0.126)
Convenience 0.118 (0.128) 0.112 (0.126)
Withdrawal method
Bank teller −0.550** (0.215)
Check casher −0.631 (0.392)
Cashback 0.276 (0.302)
Employer 1.804* (0.964)
Family 0.017 (0.448)
Other 1.824*** (0.566)
Constant −0.233 (2.324) 5.909 (5.481) 5.619 (6.049)
R20.008 0.015 0.041
Observations 3882 3090 3089
Individuals 906 900 899
FE is the fixed-effects estimator obtained on the balanced panel. bare the point estimates and se the standard errors. Cluster-robust standard errors are used. HH and CC
refers to household and credit card, respectively. Base category of categorical variables is lower than high school, unemployed, widowed, lower than 25 years, Asian, and
ATM. Significance levels are denoted as ***p<0.01, **p<0.05, *p<0.1

Trütsch Swiss Journal of Economics and Statistics (2020) 156:5 Page 30 of 35
Table 25 FE regression results of contactless credit on cash in wallet
(1) (2) (3)
Variable b se b se b se
Contactless credit 8.371 (12.976) 6.850 (14.477) 6.067 (14.318)
log(income)12.786*** (4.315) 11.055*** (4.237) 10.827** (4.272)
Interest rate −4.682 (4.328) −7.248 (5.624) −6.786 (5.622)
Education
High school 22.749*** (6.643) 21.033** (9.831) 5.685 (10.927)
Some college 34.216 (22.420) 37.017 (28.817) 18.492 (27.996)
College 56.729** (27.883) 57.559* (34.903) 40.469 (34.991)
Post graduate 90.759** (42.140) 69.550 (42.755) 49.777 (41.246)
Employment
Working −5.712 (7.804) −3.206 (8.858) −3.256 (8.653)
Retired 5.639 (9.589) 33.880*** (12.537) 33.505*** (12.363)
Other 8.499 (8.777) 0.718 (10.640) 1.618 (10.546)
Marital status
Single 3.580 (25.940) 31.442 (38.132) 30.960 (37.818)
Married −2.111 (24.427) 10.350 (36.163) 11.797 (35.912)
Separated 0.832 (24.022) 6.940 (35.086) 7.005 (34.823)
Age
25–34 −12.806 (14.095) −47.290* (24.754) −48.077* (26.045)
35–44 −9.874 (18.077) −39.249 (27.810) −40.680 (28.671)
45–54 −11.948 (19.972) −43.834 (30.011) −43.679 (30.794)
55–64 −0.598 (21.883) −31.003 (32.207) −30.349 (32.904)
> 65 2.505 (23.721) −27.453 (33.924) −25.689 (34.512)
Other
HH members 0.613 (2.484) −0.103 (3.061) 0.088 (3.078)
CC revolver −0.053 (4.976) −1.117 (6.277) −1.473 (6.354)
Home owner −2.163 (8.108) −6.272 (10.283) −6.916 (10.301)
Rel. characteristics
Security −2.873 (2.274) −2.415 (2.262)
Setup 2.135 (4.750) 2.088 (4.763)
Acceptance −0.195 (5.424) −0.452 (5.493)
Costs −1.493 (5.273) −1.690 (5.286)
Records 3.188 (3.449) 3.792 (3.452)
Convenience −1.924 (4.760) −1.777 (4.800)
Withdrawal method
Bank teller 21.945*** (7.388)
Check casher 6.857 (13.532)
Cashback −0.751 (4.903)
Employer −18.784 (15.614)
Family 12.484 (9.592)
Other 5.559 (13.758)
Constant −111.279** (55.791) −72.804 (66.349) −58.150 (66.991)
R20.009 0.014 0.024
Observations 3599 2874 2874
Individuals 851 844 844
FE is the fixed-effects estimator obtained on the balanced panel. bare the point estimates and se the standard errors. Cluster-robust standard errors are used. HH and CC
refers to household and credit card, respectively. Base category of categorical variables is lower than high school, unemployed, widowed, lower than 25 years, Asian, and
ATM. Significance levels are denoted as ***p<0.01, **p<0.05, *p<0.1

Trütsch Swiss Journal of Economics and Statistics (2020) 156:5 Page 31 of 35
Table 26 FE regression results of contactless debit on cash in wallet
(1) (2) (3)
Variable b se b se b se
Contactless debit 3.789 (5.966) 4.256 (7.965) 3.206 (7.945)
log(income)14.059*** (5.240) 13.057*** (4.988) 13.154*** (5.006)
Interest rate −3.960 (4.175) −7.424 (5.338) −7.002 (5.358)
Education
High school 22.962*** (5.741) 23.694** (9.272) 13.458 (10.292)
Some college 8.841 (40.354) −42.988 (49.328) −46.158 (35.830)
College 33.112 (43.942) −25.073 (52.826) −27.669 (40.513)
Post graduate 60.546 (52.828) −16.242 (58.945) −20.329 (47.134)
Employment
Working 2.339 (7.593) 4.849 (8.873) 5.420 (8.717)
Retired 6.492 (8.506) 25.442** (12.730) 24.980** (12.509)
Other 7.170 (7.620) 4.548 (9.582) 5.150 (9.593)
Marital status
Single −5.355 (21.662) 3.977 (32.038) 4.851 (31.983)
Married −17.728 (20.232) −11.518 (30.069) −9.448 (29.966)
Separated −25.935 (24.999) −29.421 (36.135) −26.059 (35.745)
Age
25–34 5.047 (13.856) −33.830 (26.245) −31.345 (26.809)
35–44 4.502 (17.233) −32.614 (29.194) −31.246 (29.329)
45–54 2.512 (19.327) −40.682 (31.522) −37.907 (31.710)
55–64 17.968 (21.568) −24.783 (33.801) −21.768 (33.750)
> 65 25.009 (23.794) −17.044 (35.696) −13.001 (35.650)
Other
HH members −0.748 (2.545) −0.673 (3.064) −0.704 (3.049)
CC revolver −3.961 (4.785) −1.643 (5.826) −1.743 (5.844)
Home owner −1.334 (7.684) −7.772 (9.411) −8.442 (9.467)
Rel. characteristics
Security −3.354 (2.297) −3.030 (2.304)
Setup 3.689 (4.183) 3.683 (4.189)
Acceptance −0.064 (4.609) −0.270 (4.640)
Costs −2.416 (4.978) −2.466 (4.982)
Records 1.844 (3.087) 2.089 (3.093)
Convenience −3.301 (4.358) −3.054 (4.361)
Withdrawal method
Bank teller 14.425** (6.723)
Check casher 4.667 (13.153)
Cashback −3.940 (5.414)
Employer −17.593 (14.111)
Family 16.414 (15.787)
Other 5.485 (9.757)
Constant −108.235 (67.233) −13.011 (73.219) −17.689 (67.989)
R20.011 0.014 0.020
Observations 3882 3093 3093
Individuals 905 898 898
FE is the fixed-effects estimator obtained on the balanced panel. bare the point estimates and se the standard errors. Cluster-robust standard errors are used. HH and CC
refers to household and credit card, respectively. Base category of categorical variables is lower than high school, unemployed, widowed, lower than 25 years, Asian, and
ATM. Significance levels are denoted as ***p<0.01, **p<0.05, *p<0.1

Trütsch Swiss Journal of Economics and Statistics (2020) 156:5 Page 32 of 35
Table 27 FE regression results of contactless credit on cash share volume
(1) (2) (3)
Variable b se b se b se
Contactless credit −0.011 (0.027) −0.013 (0.030) −0.013 (0.030)
log(income)−0.007 (0.013) 0.007 (0.015) 0.006 (0.015)
Interest rate 0.013 (0.010) 0.013 (0.011) 0.013 (0.011)
Education
High school 0.268*** (0.069) 0.358*** (0.030) 0.346*** (0.033)
Some college 0.308** (0.127) 0.536*** (0.102) 0.522*** (0.103)
College 0.160 (0.140) 0.288** (0.119) 0.277** (0.119)
Post graduate 0.155 (0.166) 0.281** (0.137) 0.271** (0.135)
Employment
Working −0.011 (0.016) −0.000 (0.018) −0.002 (0.019)
Retired −0.005 (0.019) −0.026 (0.027) −0.027 (0.027)
Other −0.000 (0.020) −0.036 (0.025) −0.035 (0.025)
Marital status
Single 0.040 (0.060) 0.031 (0.065) 0.033 (0.064)
Married 0.008 (0.046) 0.037 (0.050) 0.039 (0.050)
Separated 0.026 (0.052) 0.067 (0.054) 0.068 (0.054)
Age
25–34 −0.073 (0.073) −0.296*** (0.102) −0.298*** (0.102)
35–44 −0.058 (0.078) −0.287*** (0.105) −0.289*** (0.105)
45–54 −0.090 (0.081) −0.320*** (0.108) −0.323*** (0.108)
55–64 −0.130 (0.083) −0.364*** (0.111) −0.365*** (0.111)
>65 −0.099 (0.087) −0.375*** (0.114) −0.375*** (0.114)
Other
HH members −0.004 (0.007) 0.006 (0.008) 0.006 (0.008)
CC revolver −0.006 (0.013) −0.017 (0.015) −0.016 (0.015)
Home owner −0.034 (0.022) −0.032 (0.028) −0.032 (0.028)
Rel. characteristics
Security −0.008 (0.006) −0.008 (0.006)
Setup 0.023* (0.013) 0.023* (0.013)
Acceptance −0.009 (0.016) −0.009 (0.016)
Costs 0.011 (0.016) 0.011 (0.016)
Records 0.010 (0.009) 0.011 (0.009)
Convenience −0.003 (0.013) −0.003 (0.013)
Withdrawal method
Bank teller 0.017 (0.017)
Check casher 0.003 (0.041)
Cashback −0.015 (0.015)
Employer 0.002 (0.039)
Family −0.009 (0.032)
Other −0.004 (0.027)
Constant 0.304 (0.192) 0.229 (0.206) 0.251 (0.207)
R20.013 0.033 0.035
Observations 3556 2860 2858
Individuals 852 846 845
FE is the fixed-effects estimator obtained on the balanced panel. bare the point estimates and se the standard errors. Cluster-robust standard errors are used. HH and CC
refers to household and credit card, respectively. Base category of categorical variables is lower than high school, unemployed, widowed, lower than 25 years, Asian, and
ATM. Significance levels are denoted as ***p<0.01, **p<0.05, *p<0.1

Trütsch Swiss Journal of Economics and Statistics (2020) 156:5 Page 33 of 35
Table 28 FE regression results of contactless debit on cash share volume
(1) (2) (3)
Variable b se b se b se
Contactless debit −0.003 (0.033) 0.003 (0.038) 0.004 (0.038)
log(income)0.015 (0.013) 0.022 (0.014) 0.020 (0.014)
Interest rate 0.012 (0.010) 0.011 (0.010) 0.011 (0.010)
Education
High school 0.269*** (0.067) 0.357*** (0.029) 0.344*** (0.032)
Some college 0.264* (0.158) 0.488*** (0.087) 0.454*** (0.097)
College 0.090 (0.169) 0.226** (0.109) 0.195* (0.116)
Post graduate 0.092 (0.185) 0.211* (0.125) 0.181 (0.129)
Employment
Working −0.009 (0.014) 0.009 (0.017) 0.007 (0.017)
Retired −0.010 (0.017) −0.033 (0.026) −0.035 (0.026)
Other −0.016 (0.018) −0.032 (0.023) −0.031 (0.023)
Marital status
Single 0.013 (0.067) 0.028 (0.072) 0.027 (0.071)
Married 0.026 (0.055) 0.053 (0.058) 0.053 (0.057)
Separated 0.036 (0.058) 0.077 (0.059) 0.076 (0.058)
Age
25–34 −0.101 (0.063) −0.227* (0.124) −0.227* (0.124)
35–44 −0.093 (0.067) −0.222* (0.126) −0.223* (0.126)
45–54 −0.105 (0.070) −0.230* (0.129) −0.233* (0.129)
55–64 −0.133* (0.073) −0.260** (0.131) −0.261** (0.131)
>65 −0.101 (0.076) −0.253* (0.133) −0.254* (0.133)
Other
HH members 0.000 (0.007) 0.004 (0.008) 0.005 (0.008)
CC revolver −0.013 (0.012) −0.024* (0.014) −0.023 (0.014)
Home owner −0.040* (0.021) −0.032 (0.026) −0.031 (0.025)
Rel. characteristics
Security −0.009 (0.006) −0.008 (0.006)
Setup 0.018 (0.012) 0.018 (0.012)
Acceptance −0.006 (0.014) −0.006 (0.014)
Costs 0.014 (0.015) 0.013 (0.015)
Records 0.003 (0.009) 0.003 (0.009)
Convenience 0.003 (0.012) 0.004 (0.012)
Withdrawal Method
Bank teller 0.020 (0.016)
Check casher −0.017 (0.043)
Cashback −0.014 (0.015)
Employer 0.018 (0.040)
Family −0.015 (0.028)
Other −0.014 (0.025)
Constant 0.117 (0.213) 0.006 (0.207) 0.047 (0.210)
R20.013 0.028 0.030
Observations 3832 3075 3074
Individuals 905 899 898
FE is the fixed-effects estimator obtained on the balanced panel. bare the point estimates and se the standard errors. Cluster-robust standard errors are used. HH and CC
refers to household and credit card, respectively. Base category of categorical variables is lower than high school, unemployed, widowed, lower than 25 years, Asian, and
ATM. Significance levels are denoted as ***p<0.01, **p<0.05, *p<0.1

Trütsch Swiss Journal of Economics and Statistics (2020) 156:5 Page 34 of 35
Abbreviations
ALP: American life panel; ATM: Automated teller machine; EMV: Europay
international, MasterCard and VISA; EU: European union; FE: Fixed-effects; FED:
Federal reserve system; GDP: Gross domestic product; LTP: Let’s talk payments;
NFC: Near-field communication; OLS: Ordinary least squares; PIN: Personal
identification number; POS: Point-of-sale; SCPC: Survey of consumer payment
choice; SPA: Smart payment association; USA: United States; USD: United
States dollar
Acknowledgements
I would like to thank Martin Brown for his extremely valuable inputs and
recommendations regarding the analysis in the paper. His comments helped
to substantially improve the work. I am also grateful to the anonymous referees
for their critical review, which has led to significant improvements of the paper.
Authors’ contributions
Not applicable. There is only one author. The author(s) have read and
approved the manuscript.
Funding
Not applicable.
Availability of data and materials
The datasets generated and/or analyzed during the current study are available
from the Federal Reserve Bank of Atlanta, Consumer Payment Research
Center, https://www.frbatlanta.org/banking-and- payments/consumer-
payments/survey-of- consumer-payment- choice.
Competing interests
The author declares that he has no competing interests.
Received: 21 January 2019 Accepted: 22 April 2020
References
Alvarez, F., & Lippi, F. (2009). Financial innovation and the transactions demand
for cash. Econometrica,77(2), 363–402.
Amromin, G., & Chakravorti, S. (2009). Whither loose change? The diminishing
demand for small-denomination currency. Journal of Money, Credit and
Banking,41(2-3), 315–335.
Angrisani, M., Foster, K., Hitczenko, M. (2015). The 2013 survey of consumer
payment choice: Technical appendix, Research Data Reports No. 15-5,
Federal Reserve Bank of Boston. www.bostonfed.org. Accessed 10 May
2017 [Online].
Arango, C., Hogg, D., Lee, A. (2015). Why is cash (still) so entrenched? Insights
from Canadian Shopping Diaries. Contemporary Economic Policy,33(1),
141–158.
Attanasio, O.P., Guiso, L., Jappelli, T. (2002). The demand for money, financial
innovation, and the welfare cost of inflation: An analysis with household
data. Journal of Political Economy,110(21), 317–351.
Bagnall, J., Bounie, D., Huynh, K.P., Kosse, A., Schmidt, T., Schuh, S., Stix, H.
(2016). Consumer cash usage: A cross-country comparison with payment
diary survey data. International Journal of Central Banking,12(4), 1–61.
Baumol, W.J. (1952). The transactions demand for cash: An inventory theoretic
approach. The Quarterly Journal of Economics,66(4), 545–556.
Borzekowski, R., & Kiser, E.K. (2008). The choice at the checkout: Quantifying
demand across payment instruments. International Journal of Industrial
Organization,26, 889–902.
Bouhdaoui, Y., & Bounie, D. (2012). Modeling the share of cash payments in the
economy. International Journal of Central Banking,8(4), 175–195.
Briglevics, T., & Schuh, S. (2013). U.S. consumer demand for cash in the era of
low interest rates and electronic payments Working Paper No. 13-23, The
Federal Reserve Bank of Boston.
Brits, H., & Winder, C. (2005). Payments are no free lunch. DNB Occasional
Studies,3(2).
Chai, C. (2017). Contactless tap-and-go cards finally enter US market.
www.creditcards.com. Accessed 10 Nov 2018 [Online].
Chen, H., Felt, M.H., Huynh, K.P. (2017). Retail payment innovations and cash
usage: Accounting for attrition using refreshment samples. Journal of the
Royal Statistical Society, Series A, Statistics in Society,180(2), 503–530. https://
doi.org/10.1111/rssa.12208.
Ching, A.T., & Hayashi, F. (2010). Payment card rewards programs and consumer
payment choice. Journal of Banking and Finance,34(8), 1773–1787.
Columba, F. (2009). Narrow money and transaction technology: New
disaggregated evidence. Journal of Economics and Business,61, 312–325.
Connolly, J., & Stavins, J. (2015). Payment instrument adoption and use in the
United States 2009–2013 by Consumers’ Demographic Characteristics.
Research Data Reports No. 15-6 Federal Reserve Bank of Boston.
Drehmann, M., Goodhart, C., Krueger, M. (2004). The challenges facing
currency usage: will the traditional transaction medium be able to resist
competition from new technologies? Economic Policy,56, 195–227.
FED (2011). Federal Reserve System, The 2010 Federal Reserve Payments
Study. www.federalreserve.gov. Accessed 11 Aug 2018 [Online].
FED (2014). Federal Reserve System, The 2013 Federal Reserve Payments
Study. www.federalreserve.gov. Accessed 11 Aug 2018 [Online].
Foster, K., Schuh, S., Zhang, H. (2013). The 2010 Survey of Consumer Payment
Choice. Research Data Reports No. 13-2, The Federal Reserve Bank of
Boston.
Fujiki, H., & Tanaka, M. (2014). Currency demand, new technology, and the
adoption of electronic money: Micro evidence from Japan. Economics
Letters,125(1), 5–8.
Fung, B.S., Huynh, K.P., Sabetti, L. (2014). The impact of retail payment
innovations on cash usage. Journal of Financial Market Infrastructures,3(1),
1–29.
van der Horst, F., & Matthijsen, E. (2013). The irrationality of payment
behaviour. DNB Occasional Studies,11(4).
van Hove, L. (2008). On the war on cash and its spoils. International Journal of
Electronic Banking,1(1), 36–45.
Humphrey, D.B. (2004). Replacement of cash by cards in U.S. consumer
payments. Journal of Economics and Business,56, 211–225.
Huynh, K.P., Schmidt-Dengler, P., Stix, H. (2014). The role of card acceptance in
the transaction demand for money. Working Paper No. 44, Bank of Canada.
Jonker, N. (2007). Payment instruments as perveived by consumers: Results
from a household survey. De Economist,155, 21–38.
von Kalckreuth, U., Schmidt, T., Stix, H. (2009). Choosing and using payment
instruments: Evidence from German micro-data. Working Paper Series No.
1144 European Central Bank.
von Kalckreuth, U., Schmidt, T., Stix, H. (2014). Using cash to monitor liquidity –
Implications for payments, currency demand and withdrawal behavior.
Journal of Money, Credit and Banking,46(8), 1753–1785.
Kim, B.M., Yilmazer, T., Widdows, R. (2006). Adoption of internet banking and
consumers’ payment choices. Working Paper, Purdue University.
Klee, E (2006). Paper or Plastic? The effect of time on check and debit card use
at grocery stores Working Paper, Board of Governors of the Federal
Reserve System.
Lippi, F., & Secchi, A. (2009). Technological change and the households’
demand for currency. Journal of Monetary Economics,56, 222–230.
LTP (2015). Let’s Talk Payments Comprehensive 2015. U.S. Market Analysis of
POS Terminals and EMV & NFC Status Review.
McCallum, B.T., & Goodfriend, M.S. (1987). Demand for money: Theoretical
studies. In J. Eatwell, M. Milgate, P. Newman (Eds.), The New Palgrave: A
Dictionary of Economics. Macmillan: Palgrave.
Polasik, M., Gorka, J., Wilczewski, G., Kunkowski, J., Przenajkowska, K., Tetkowska,
N. (2013). Time efficiency of point-of-sale payment methods: Empirical
results for cash, cards and mobile payments. In J. Cordeiro, L.A. Maciaszek,
J. Filipe (Eds.), Enterprise Information Systems, Lecture Notes in Business
Information Processing, Vol. 141 (pp. 306–320). Berlin Heidelberg: Springer.
SCF (2014). Survey of Consumer Finances, 2013 Survey of Consumer Finances.
www.federalreserve.gov/econresdata/scf/scfindex.htm. Accessed 10 Oct
2015 [Online].
Schmiedel, H., Kostova, G.L., Ruttenberg, W. (2013). The social and private costs
of retail payment instruments: A European Perspective. Journal of Financial
Market Infrastructures,2(1), 37–75.
Schuh, S., & Stavins, J. (2013). How consumers pay: Adoption and use of
payments. Accounting and Finance Research,2(2), 1–21.
Schuh, S., & Stavins, J. (2014). The 2011 and 2012 Survey of Consumer Payment
Choice. Research Data Report No. 14-1, The Federal Reserve Bank of Boston.
Schuh, S., & Stavins, J. (2015). The 2013 Survey of Consumer Payment Choice:
Summary Results. Research Data Report No. 15-4, The Federal Reserve
Bank of Boston.

Trütsch Swiss Journal of Economics and Statistics (2020) 156:5 Page 35 of 35
Snellman, J., Vesala, J., Humphrey, D.B. (2001). Substitution of noncash
payment instruments for cash in Europe. Journal of Financial Services
Research,19(2/3), 131–145.
SPA (2016). Smart Payment Association, An Overview of Contactless Payment
Benefits and Worldwide Developments.
www.smartpaymentassociation.com. Accessed 20 Dec 2016 [Online].
Stix, H. (2003). How do debit cards affect cash demand? Survey Data Evidence.
Working Paper 82 Oesterreichische Nationalbank.
Tobin, J. (1956). The interest elasticity of the transactions demand for cash.
Review of Economics and Statistics,38(3), 241–247.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.