Content uploaded by David Tercero-Lucas
Author content
All content in this area was uploaded by David Tercero-Lucas on Jul 01, 2021
Content may be subject to copyright.
BIS Working Papers
No 951
Distrust or speculation? The
socioeconomic drivers of US
cryptocurrency investments
by Raphael Auer and David Tercero-Lucas
Monetary and Economic Department
July 2021
JEL classification: D14, D91, E42, G11, G12, G28, O33.
Keywords: digital currencies, cryptocurrencies,
distributed ledger technology, blockchain, payments,
digitalisation, banking, household finance, money,
bitcoin, ether, xrp, bitcoin cash, litecoin, xlm, eos.
BIS Working Papers are written by members of the Monetary and Economic
Department of the Bank for International Settlements, and from time to time by other
economists, and are published by the Bank. The papers are on subjects of topical
interest and are technical in character. The views expressed in them are those of their
authors and not necessarily the views of the BIS.
This publication is available on the BIS website (www.bis.org).
© Bank for International Settlements 2021. All rights reserved. Brief excerpts may be
reproduced or translated provided the source is stated.
ISSN 1020-0959 (print)
ISSN 1682-7678 (online)
Distrust or Speculation? The Socioeconomic
Drivers of U.S. Cryptocurrency Investments∗
Raphael Auer1and David Tercero-Lucas2
1Bank for International Settlements (BIS)
2Universitat Autònoma de Barcelona (UAB)
Abstract
Employing representative data from the U.S. Survey of Consumer Payment
Choice, we disprove the hypothesis that cryptocurrency investors are motivated
by distrust in fiat currencies or regulated finance. Compared with the general
population, investors show no differences in their level of security concerns
with either cash or commercial banking services. We find that cryptocurrency
investors tend to be educated, young and digital natives. In recent years, a gap
in ownership of cryptocurrencies across genders has emerged. We examine how
investor characteristics vary across cryptocurrencies and show that owners of
cryptocurrencies increasingly tend to hold their investment for longer periods.
JEL Classification: D14 D91 E42 G11 G12 G28 O33
Keywords
: digital currencies, cryptocurrencies, distributed ledger technology,
blockchain, payments, digitalisation, banking, household finance, money, bitcoin,
ether, xrp, bitcoin cash, litecoin, stellar, eos
∗
We thank Svetlana Abramova, Douglas Araujo, Stefano Balietti, Lea Bitter, Rainer Böhme,
Stijn Claessens, Sebastian Doerr, Egemen Eren, Jon Frost, Leonardo Gambacorta, Jonas Gross,
Hasu, Destan Kirimhan, Tom Lyons, Tom Minic, Bénédicte Nolens, Nandan Rao, Nicolas Reigl, and
an anonymous referee at the BIS working paper series for their useful comments and suggestions.
The views expressed in this paper are those of the authors and do not necessarily reflect those of
the Bank for International Settlements. This research is part of the PhD dissertation of David
Tercero-Lucas, who would like to acknowledge the support of a fellowship from “la Caixa” Foundation
(ID 100010434 - fellowship code LCF/BQ/ES18/11670005).
1 Introduction
The rise and fall of bitcoin, ether, and related cryptocurrencies – with market
capitalisations (see Figure 1a) at times rivalling that of silver, the world’s major
financial companies, and even the stock markets of large advanced economies
1
–
warrants a close examination of investor motivations and levels of sophistication.
One aspect of particular relevance is that the purported motivation for the
creation of these cryptocurrencies has been to design an alternative to fiat money
and commercial banking, with the goal of creating a new form of exchange that is
resistant to debasement and censorship by governments and financial institutions.
As put by Nakamoto (2008, p.1):2
"What is needed is an electronic payment system based on cryptographic proof
instead of trust, allowing any two willing parties to transact directly with each other
without the need for a trusted third party”.
This narrative is also relied upon frequently by the proponents of this asset
class. One noteworthy episode is the trading suspension of GameStop shares on
the app Robinhood that occurred in early 2021. The suspension was – incorrectly –
interpreted as censorship and market manipulation in favour of large hedge funds
(see Ossinger and Hunter (2021) and Appendix A). Due to ensuing media coverage,
a substantial inflow of funds into cryptocurrencies resulted.
Figure 1: Market valuations have reached new records.
(a) Market capitalisations* (b) Bitcoin and dogecoin prices
Source: CoinMarketCap (Panel (a)) and Coindesk.com (Panel (b)). *Other cryptocurrencies include
the sum of the market capitalization of the biggest cryptocurrencies (excluding stablecoins) after
bitcoin and ether.
1
As of 26 May 2021, silver had a market capitalization of around $1.51 trillion, gold’s market
capitalization was $12.04, JPMorgan Chase was $497.38 billion and Bank of America was $365.75
billion (CompaniesMarketCap,2021). In comparison, at that day, total market capitalisation of
cryptocurrencies was $1.72 trillion. As of 8 May 2021, total market capitalisation of cryptocurrencies
(including stablecoins) reached its peak at $2.42 trillion, doubling the combined market capitalisation
of all companies included in the German DAX 30 index (CoinMarketCap,2021).
2
For example, Vitalik Buterin, considered “non-discrimination and non-censorship” one of the
key principles behind the design of Ethereum (Buterin,2013)
2
In the subsequent hype, the price of bitcoin continued surging, exceeding $50,000
soon after the trading suspension, and peaking at over $63,000 several weeks later (see
Figure 1b). Brought to the attention of many via widely read tweets by Elon Musk,
the cryptocurrency Dogecoin saw an almost ten-fold price increase during this episode
(Ossinger and Hunter,2021). However, cryptocurrency prices collapsed during mid-
May 2021, after renewed statements by Mr Musk and the announcement of a ban for
financial institutions and payment companies from providing cryptocurrency services
in China.3
As the end of the GameStop episode exemplifies, the narrative of
cryptocurrencies as a censorship-resistant asset class does not always square with
reality.
4
Cryptocurrencies are rife with fraud and theft (Auer and Claessens,2019;
Foley et al.,2019;Twomey and Mann,2020),
5
mostly due to the fact that coins are
held in the custody of unregulated middlemen (Kharif,2020). Cryptocurrencies such
as bitcoin that are sustained by costly computing (“proof-of-work”) tend to be
centralised (Huang,2020) and their basic security model might not be sustainable
(Auer,2019a).
6
There is ample debate on the censorship-resistance, decentralisation,
and legal nature of other cryptocurrencies, as well (Fröwis and Böhme,2017;Walch,
2019;SEC,2020).7
Amidst a discrepancy between sociological narrative and factual evidence, it is
important to understand who the retail investors in cryptocurrencies are, what their
level of trust and knowledge is, and how they interact with the mainstream financial
system. The objective of this paper is hence threefold. We start by examining
the hypothesis that cryptocurrencies are sought out of distrust in fiat currencies or
regulated finance. Second, we study the broader socioeconomic characteristics of U.S.
retail consumers and disentangle the role of knowledge acquisition and investment
decisions conditional on knowledge. Third, we examine the evolution of patterns of
cryptocurrency investments across time and cryptocurrencies.
3
See BBC (2021). Dogecoin has been subject to even higher volatility during May. Its price
dropped by around 30% on May 9, the day of the appearance of Elon Musk on the TV show
“Saturday Night Live”.
4
The emergence of cryptocurrencies has also led to the development of stablecoins, such as the
Facebook’s Diem pro ject. For an analysis of this and other global stablecoins, see Arner et al.
(2020), Frost et al. (2020), Melachrinos and Pfister (2020) and Tercero-Lucas (2020).
5
There have been many cases of fraud in the industry. One example is the project PlusToken,
which turned out to be a cryptocurrency Ponzi scheme that attracted millions of people with
promises of high returns on investment. The operators were taken to court and found guilty of
defrauding investors of almost $2.3 billion (Akhtar,2021). Investors are also subject to cyber
attacks that have affected both open source distributed ledgers (e.g., in February, 2020, the IOTA
Foundation had to temporarily shut down the IOTA network after suffering an attack on its wallet
app (Pan,2020)) and cryptocurrencies exchange markets (e.g., in the first quarter of 2020, Altsbit,
an Italian cryptocurrency exchange, had to close because the majority of user funds were stolen in
a cyber attack (Partz,2020)).
6
Also, the environmental impact must not be forgotten. Bitcoin’s energy consumption is
exceeding that of entire countries (see Carstens (2018a) and De Vries (2018)).
7
They are further used for illegal activities, including money laundering and the financing of
terrorism. See e.g., Fanusie and Robinson (2018). Foley et al. (2019) estimate that around one
quarter of bitcoin users were involved in illegal activity in the pre-2018 period, which translated to
$76 billion per year.
3
We employ the Survey of Consumer Payment Choice (SCPC), a representative
micro-level dataset provided by the Federal Reserve Bank of Atlanta. The survey
covers the 2014-19 period and is representative of the US population. Using a variety
of econometric specifications, we first disprove the hypothesis that cryptocurrencies
are sought as an alternative to fiat currencies or regulated finance in the US. Compared
with the general population, cryptocurrency investors show no differences in their
level of security concerns with either cash or commercial banking services. We do,
however, find that those who are concerned with the security of cash or bank accounts
tend to acquire information about cryptocurrencies.8
Our second set of results regards the broader sociodemographics of cryptocurrency
investors. Higher educational attainment is associated with more knowledge about
and likelihood of owning a cryptocurrency. Male gender is associated with a 2 to
2.2 percentage points higher likelihood of owning at least one cryptocurrency. The
probability of knowing at least one cryptocurrency is higher for men and for those
individuals with higher levels of both income and an education.
We provide some evidence that the impact of gender and age on cryptocurrency
investment is unrelated to differences in knowledge about the underlying technology.
Despite converging knowledge levels about cryptocurrencies, a gender gap in terms of
ownership has emerged.
9
In the same vein, although age has no effect on knowledge
about cryptocurrencies, it does have a strong effect on investment decisions.
Our last set of results regards the evolution of these patterns of cryptocurrency
investments across time and cryptocurrencies. Owners of ether and xrp are the
most educated in our sample, followed by bitcoin cash and bitcoin users. Conversely,
those owning litecoin are the least educated. We document trends in knowledge
and ownership, and develop an empirical test for so called “hodling”- a term in the
cryptocurrency community that refers to buying and holding a cryptocurrency over
the long-term.
10
Specifically, we estimate that owning a cryptocurrency in one year
increases the probability, on average, of owning a cryptocurrency in the following
year by 50%.
Overall, our paper contributes to several literatures spanning the fields of sociology,
economics, financial stability, and computer science. Our focus on the sophistication
of retail cryptocurrency investors and on whether the demand for cryptocurrencies is
driven by distrust in fiat money and/or the commercial financial industry contributes
to the literature of the sociology of financial markets (Pixley,2004;Knorr Cetina
and Preda,2005;Preda,2007;Knorr Cetina and Preda,2012). Lack of trust has
been shown to be a main driver of investment decisions and limited stock market
participation (see Guiso et al. (2008); Georgarakos and Pasini (2011); Balloch
et al. (2015)).
11
Given the paramount importance of trust for the monetary and
8
Our findings also suggest that experience with digital finance, captured through having a debit
card, and using a mobile payment app, increases the probability of investing in cryptocurrencies.
9
This is in contrast to the finding of a survey discussed in Baker (2019), which found that
“67 percent of women felt their lack of familiarity with bitcoin stopped them from investing in it,
compared to 48 percent of men”.
10For a further explanation of the origin of the term, see StormGain (2020).
11
Employing data from all Bitnodes operating worldwide between 2014 and 2018, (Saiedi
4
financial system (see Carstens (2018b,2019); Borio (2019)), the persistent rise of
cryptocurrencies could potentially evidence rising distrust in today’s arrangements.
12
In that light, our key finding alleviates these concerns: cryptocurrency investors do
not present differences in their level of security concerns with mainstream payment
options, i.e., trust in cash or the banking deposits.
From a policy angle, one of the main takeaways is that as the goals of investors
are the same as those for other asset classes, so should be the regulation. A clarifying
regulatory and supervisory framework for cryptocurrency markets may be useful
for the industry. In fact, regulatory announcements have had a strong impact on
cryptocurrency prices and transaction volumes (Auer and Claessens,2019,2020), and
those pointing to the establishment of specific regulations tailored to cryptocurrencies
and initial coin offerings are strongly correlated with relevant market gains. Here, one
important consideration regards how one could apply technology-neutral regulation,
while at the same time harnessing the potential of the technology itself in the
supervision process. One option for such a framework is “embedded supervision”,
developed in Auer (2019b). This means implementing a supervisory framework for
cryptocurrencies that allows for compliance to be automatically monitored by reading
the market’s ledger. The goal is low-cost supervision of decentralised markets, which
may be particularly relevant amidst recent deliberations of the need for adequate
prudential oversight of the cryptocurrency industry (Basel Committee,2019,2021).
We are also interested in the prevention of consumer fraud in the cryptocurrency
industry, and in particular, whether cryptocurrencies in general, or specific projects,
target poor and uneducated or rather the sophisticated and wealthy investor class
who can afford to experiment. Several consumer agencies have warned against
cryptocurrency scams. Policymakers have also shown their concern about the
increasing adoption of cryptocurrencies.
13
Our results – which show that
cryptocurrency investors tend to be educated – to some extent imply that a
majority of cryptocurrency investors are aware of the inherent risks. Our findings
also suggest that being young increases the likelihood of owning cryptocurrencies.
Since older people are at greater risk of both consumer and financial fraud (Temple,
2007;DeLiema et al.,2020), young people stand a better chance of avoiding them.
This part of the analysis is closely related to recent literature analysing the profile
and behaviour of cryptocurrency users.14
et al., 2020) argue that low trust in the financial system has contributed to the spread of
Bitcoin infrastructure. In addition, they show that Bitcoin’s support is higher in cities with
well-developed banking services. Beyond trust, (De Bondt,2005) finds that self-confidence and
financial sophistication are important determinants of the perceived attractiveness of different asset
classes and investment strategies. Analysing a survey of more than 3,100 European investors, he also
documents differences in values and beliefs by age, health status, religious affiliation, and gender.
As he highlights, people’s values and beliefs are highly correlated with investment behaviour.
12
The level of trust is also determined by other factors such as differences in educational attainment
(Guiso et al.,2004) or in religious upbringing (Guiso et al.,2003).
13
For instance, a member of the US Federal Reserve Board, Lael Brainard, stated that
“cryptocurrencies may raise important investor and consumer protection issues. The lack of
strong governance and questions about the applicable legal framework for some cryptocurrencies
may make consumers vulnerable to mistakes, thefts, and security breaches without much, or any,
recourse” (Brainard,2018).
14
Almost all previous studies focus solely on Bitcoin and do not tackle the security and convenience
5
Our analysis is also related to the literature examining gender gaps in finance.
15
Women tend to be more risk-averse than men when it comes to holding risky assets
(Jianakoplos and Bernasek,1998;Borghans et al.,2009;Arano et al.,2010), and
there are significant differences across genders in the use of FinTech (Chen et al.,
2021).
Further, understanding the concerns and sociodemographic characteristics of
cryptocurrency owners is crucial to those wanting to gauge the potential of
cryptocurrency markets and estimate how large this asset class could eventually
become. We find that cryptocurrencies, at least at current, remain niche markets
dominated by young male investors, while other parts of the population acquire
information about this asset class, but ultimately do not invest in it. However, a
feature pertinent to analysing the potential of the cryptocurrency market is the
phenomenon of hodling. Our results suggest that it is a persistent trait among
cryptocurrency investors.
In addition, our findings may be informative for the likely user sophistication of
future digital currencies,
16
including stablecoins or central bank digital currencies.
Understanding the socioeconomic characteristics of cryptocurrency investors can be
a first step in forecasting who the initial adopters of such future digital currencies
may be.17
The remainder of this paper is structured as follows. Section 2 describes the data
used. Section 3 outlines the empirical strategy used to identify the effects of interest.
Section 4 provides an overview of the main results and presents some robustness
checks. Section 5 documents some trends. Section 6 concludes.
aspects of cryptocurrencies. In this respect, Bohr and Bashir (2014) show that age, mining status,
time of initial use, engaging in online discourse, geographical location, and political orientation
are relevant factors to use bitcoin. Henry et al. (2018) estimate that being a man and having
a higher level of education increases the probability of knowing bitcoin in Canada. Moreover,
bitcoin awareness is more common among unemployed individuals. Using a survey among Austrian
households, Stix (2019) argues that potential adopters of cryptocurrencies are younger and are
more willing to accept financial risk. Fujiki (2020) finds that Japanese crypto-asset owners are
more likely to be men, young, have a high pre-tax income, hold graduate degrees, and have a high
level of financial literacy.
15
Employing U.S. data, Bannier et al. (2019) find that men have greater knowledge regarding
the features of bitcoin than women.
16
Voskobojnikov et al. (2020) find that the type of cryptocurrency and its area of application are
critical to determining which risks and mitigation strategies the user employs.
17
In particular, the degree of user sophistication may be key to the adoption of a token-based
CBDC that requires the handling of digital signatures and allows for anonymity. Auer and Böhme
(2020) analyse the technical design choices for retail CBDCs and their trade-offs, while Auer et al.
(2020) take stock of approaches around the world.
6
2 Data
In this paper, we use data from the Survey of Consumer Payment Choice (SCPC).
The SCPC is a representative micro-level data set provided by the Federal Reserve
Bank of Atlanta since 2009. It provides an overview of the payment behaviour of
US consumers. The SCPC
18
comprises information from the Diary of Consumer
Payment Choice (DCPC). In the DCPC, consumers record details of specific payment
transactions and choices (Foster et al.,2009). The SCPC does not have a perfect
longitudinal design, meaning that if a person is selected in a given wave, that person
does not always continue as a sample member in the subsequent editions.19
Each annual wave contains information about US consumers’ payment behaviour
regarding the use of cash, electronic payments and cryptocurrencies, as well as the
number of transactions made via these means of payment. A crucial feature of
the SCPC for this analysis is that it makes information for every respondent on
whether that person is aware of one or more cryptocurrencies and whether that
person owns any cryptocurrencies. In addition, the data include such socioeconomic
characteristics as age, gender, race, region, education level, marital status, household
income level and the number of people living in the same household.
2.1 The Estimation sample
The main analysis uses the 2019 wave, completed by a total of 3,372 individuals.
20
As
an additional analysis, we use the 2014-19 waves. In Figure 2a, it is shown that the
2014, 2015, 2016, 2017 and 2018 samples contain 1,238, 1,429, 3,404, 3,099 and 3,153
individuals respectively. From the data, 1,264 individuals participated in just one
wave and 3,876 participated in more than one wave. From them, 470 participated in
the six waves of the survey (Figure 2b).
2.2 Descriptive statistics
Table 1presents the main descriptive statistics for the 2019 wave.
21
We use four
main outcome variables: ownership; ownership number; knowledge; and knowledge
number. Ownership (knowledge) captures whether an individual owns (recognises
or knows of) at least one of the following cryptocurrencies: bitcoin, xrp, litecoin,
ether, bitcoin cash, stellar, eos, or any other cryptocurrency. Ownership number
(knowledge number) stands for the number of different cryptocurrencies that a person
owns (recognises or knows of). Of the people who responded to the survey, 73%
knew of at least one cryptocurrency. However, only 1.4% owned a cryptocurrency in
2019.22
18
For detailed analysis on each version of the SCPC, see Foster et al. (2009,2011,2013), Schuh
and Stavins (2014,2015), Greene et al. (2017), Greene and Schuh (2017), and Foster et al. (2020).
19
For instance, since 2014 (N=1238), only 470 individuals have remained in the sample in the
subsequent editions until 2019 (Figure 2b).
20
In the 2019 survey, respondents participate in September (SCPC) and October (DCPC) (Foster
et al.,2020).
21Table B1, in the Appendix, presents the main descriptive statistics with weights.
22
There are no cryptocurrency owners who state that they do not know about cryptocurrencies.
7
Figure 2: Observations and repeated individuals.
(a) Observations per year (b) Number of repeated individuals
Source: 2014-19 SCPC.
It is likely that those individuals with a higher level of digital skills know and own
more cryptocurrencies rather than those with a lower level. The level of digitalisation
is captured through three different variables: i) whether a person has a debit card,
ii) whether a person has used a mobile app to pay in the past 12 months, iii) and
whether a person has used PayPal to make a purchase or pay another person.
23
In
2019, more than 81% of the survey respondents have a debit card, 25.4% have used
a mobile app to pay and almost 40% have used at least once PayPal to make an
online purchase in the past 12 months.
Digitalisation has led to a proliferation of cashless or contactless payment methods
driven by consumers and companies, both of whom want to be able to make payments
quickly and safely and transfer funds around the world at no cost. Hence, the decision
to invest in cryptocurrencies may be driven by consumer perception of other means of
payments. These variables allow us to test whether the demand for cryptocurrencies is
indeed driven by distrust in cash or the financial industry. In the SCPC, respondents
classify the security and convenience of cash, bank account number payments and
online banking bill payments into five categories: 1 – very inconvenient/risky, 2 –
inconvenient/risky, 3 – neither inconvenient nor convenient/risky nor secure, 4 –
convenient/secure, 5 – very convenient/secure. In all cases, the average ranges from
around 2.7 to 4.
Regarding the socioeconomic indicators included in the analysis, the average
income category is 11.3. Household income is divided into sixteen categories (see
Table C1 in the Appendix). Category 11 corresponds to a level of household income
between $40,000 and $49,999. The average educational attainment of the 2019 sample
is 3.4. Educational attainment was divided into 16 categories in the 2018 and 2019
survey waves. However, since it was divided into just five categories in the 2014–17
waves, we transform the 16 categories into five (see Table C2). The variable “married”
23
PayPal launched its own cryptocurrency service, allowing people to buy, hold and sell
cryptocurrencies on its site and applications (BBC,2020).
8
Table 1: Descriptive statistics (2019 wave)
Variable Mean Std. Dev. Min. Max.
Main outcome variables
Ownership 0.014 0.118 0 1
Ownership-number 0.031 0.315 0 7
Knowledge 0.730 0.444 0 1
Knowledge-number 1.281 1.250 0 8
Digitalisation variables
Having a debit card 0.815 0.388 0 1
Mobile app for payments 0.254 0.435 0 1
Usage of PayPal 0.391 0.488 0 1
Secur. and conv. variables
Convenience of cash 3.994 1.140 1 5
Security of cash 2.685 1.551 1 5
Conv. of bank acc. number paym.
3.234 1.183 1 5
Sec. of bank acc. number paym. 2.828 1.283 1 5
Conv. of on. bank. bill payments
3.926 1.074 1 5
Security of on. bank. bill paym. 3.260 1.211 1 5
Sociodemographic variables
Income 11.320 3.929 1 16
Education 3.407 1.119 1 5
Married 0.695 0.461 0 1
Age 52.981 15.298 18 109
Retired 0.255 0.436 0 1
Male 0.436 0.496 0 1
White 0.839 0.368 0 1
The final sample includes 3235 observations. Descriptive statistics are computed without using
weights. Ownership (knowledge) captures whether an individual owns (recognises or knows) at
least one of the following cryptocurrencies: Bitcoin, xrp, litecoin, ether, bitcoin cash, stellar, eos, or
any other different cryptocurrency. Ownership-number (knowledge-number) stands for the number
of different cryptocurrencies that a person owns (recognises or knows).
represents current marital status. It takes a value of 1 if the respondent is married
and a value of 0 if the respondent is separated, divorced or widowed or never married.
In the sample, 43.6% of individuals are men; the average age is close to 53; 25.6%
are retired; and 83.8% are white.
9
3 Methodology
In order to corroborate or disprove the hypothesis that cryptocurrencies are sought
after as an alternative to fiat currencies or regulated finance, as well as to study the
socioeconomic characteristics of US retail cryptocurrency investors, we employ the
standard linear probability model (LPM):
Yi,t =β0+β1Di,t +β2Si,t +β3Xi,t +i,t (1)
where
Yi,t
is a categorical variable that takes the value 1 if individual i owns
(recognises) at least one cryptocurrency in the year t, and 0 otherwise.
Di,t
is a vector
of digitalisation variables at individual level in year t.
Si,t
is a vector of security and
convenience variables at individual level in year t. Xi,t is a vector of socioeconomic
variables at the individual level that includes gender, age, level of education, income,
race and marital status in year t. Finally, standard errors are clustered by individual.
In the cases in which the outcome of interest is a count variable that takes
nonnegative integer values 0, 1, 2, 3, . . . , i.e., the number of known or owned
cryptocurrencies, a standard count model is applied.
Since the outcome of interest is nonnegative, we need to specify our object of
interest,
E
(
Yi,t|Wi,t
), by means of a function that guarantees nonnegative values.
24
The simplest model in this context is the Poisson model, which models the conditional
mean through the exponential function. However, once the goodness-of-fit chi-
squared test is performed
25
, we conclude that the data do not fit the model well.
Therefore, a negative binomial model is performed. The main difference between
both models is that the negative binomial model relaxes the assumption of the
equality of the conditional mean and the conditional variance. Let us consider the
following specification for the conditional mean:
E(Yi,t|Wi,t ) = exp(β0+β1Di,t +β2Si,t +β3Xi,t) = exp(W0
iβ)(2)
The conditional variance is modelled as follows:
V(Yi,t|Wi,t ) = exp(W0
iβ)∗(1 + α·exp(W0
iβ)) (3)
where
α
is the overdispersion parameter. The negative binomial model is estimated
by maximum likelihood (ML) and average marginal effects are computed.
4 Empirical results
The main purpose of this paper is to analyse whether the demand for cryptocurrencies
is indeed driven by distrust in cash or the financial industry, and to develop a
socioeconomic profile of US retail cryptocurrency investors. The next sections aim
at answering these research questions
24Assume for simplicity that Wi,t =β0+β1Di,t +β2Si,t +β3Xi,t.
25The results of this test are available upon request.
10
4.1 LPM results
4.1.1 Initial results
Tables 2and 3show the econometric results of estimating Equation (1). In each
column, we estimate the relationship between each independent variable and the
main outcome variable (i.e., owning or recognising at least one cryptocurrency).
Weights are included in all regressions.26
The first set of initial results (Table 2, upper rows) show that all the
digitalisation variables have a positive impact on the likelihood of owning at least
one cryptocurrency. Having a debit card, using a mobile app for payments and using
PayPal increase the probability of investing into cryptocurrencies by 1.9, 3.5 and 2
percentage points, respectively. The findings also show that the demand for
cryptocurrencies is not driven by distrust in cash or the financial industry, given
that there are no differences in the perceived security of cash and offline and online
banking. We can thus preliminarily disprove the hypothesis that cryptocurrencies
are sought as an alternative to fiat currencies or regulated finance. However,
compared with non-owners, cryptocurrency owners tend to find both cash and
traditional banking services less convenient, although this is not the case for online
banking.27
We also show the correlation between payment experience and level of knowledge
about cryptocurrencies. We find that the level of digitalisation increases knowledge
about cryptocurrencies by around 16–17 percentage points. Those who recognise
at least one cryptocurrency find cash less secure. In other words, ceteris paribus,
if a consumer considers cash to be one step lower in terms of the security scale,
the probability of that consumer knowing about at least one cryptocurrency is 3.5
percentage points lower. That same consumer is also likely to consider paying bills
via online banking more secure and paying bills via online banking more convenient.
28
These patterns are consistent with those observed among people who have security
concerns around fiat money and so acquire information about cryptocurrencies, but
ultimately decide not to invest in them.
Of course, broader socioeconomic characteristics matter as well when it comes
to knowledge acquisition and investment decisions around cryptocurrencies. From
Table 3, it can be derived that education, income, being a man and being married
positively influence both knowing about and owning a cryptocurrency. However,
being one year older (or being retired) has a negative significant effect only on owning
a cryptocurrency. Race is uncorrelated with cryptocurrency ownership.
26Survey weights are provided by the SCPC to generate population estimates.
27
In order to purchase cryptocurrencies, investors need a mean of payment such as a debit or a
credit card. In Appendix D, we restrict the sample to those individuals who are: (i) debit card
adopters, and (ii) credit card adopters. Replicating columns 4-9 of Table 3, results are completely
robust.
28
Krombholz et al. (2017) present one of the first studies analysing how bitcoin users assess the
bitcoin ecosystem in terms of privacy, anonymity and security.
11
Table 2: Payment behaviour, cryptocurrency ownership and knowledge
Debit
Mobile PayPal
Cash Trad. Bank. Online Bank.
Conv. Sec. Conv. Sec. Conv. Sec.
(1) (2) (3) (4) (5) (6) (7) (8) (9)
Owner. 0.019***0.035***0.019**
-
0.007*
-0.001
-
0.008***
-0.002 -0.001 -0.001
(0.005) (0.011)
(0.009) (0.004)
(0.003)
(0.003) (0.004)
(0.004) (0.004)
R20.003 0.015 0.005 0.004 0.000 0.006 0.000 0.000 0.000
Know. 0.158***0.171***0.171***
-0.007
-
0.035***
-
0.042***
-0.010
0.035***0.024**
(0.033) (0.023)
(0.022) (0.010)
(0.007)
(0.009) (0.009)
(0.011) (0.010)
R20.017 0.030 0.034 0.000 0.014 0.012 0.001 0.007 0.004
Weights
Yes Yes Yes Yes Yes Yes Yes Yes Yes
Obs. 3,235 3,235 3,235 3,235 3,235 3,235 3,235 3,235 3,235
Notes: ***, ** and * indicate 1%, 5% and 10% significance levels respectively. In parentheses are
presented robust standard errors clustered by individual. Constant included but not reported.
Owner. (know.) captures whether an individual owns (knows or recognises) at least one
cryptocurrency. Debit stands for having a debit card; Mobile: Using of mobile app for payments.
Trad. Bank.: bank account number payments; Online Bank.: online banking bill payments. Conv.
and Sec. stand for convenience and security respectively.
Table 3: Sociodemographics and cryptocurrency ownership and knowledge
Education
Income Age Retired
Married
Male White
(1) (2) (3) (4) (5) (6) (7)
Owner.
0.009*** 0.002**
-
0.001***
-
0.020***
0.017*** 0.023***
0.002
(0.003) (0.001) (0.000) (0.005) (0.007) (0.008) (0.009)
R20.007 0.004 0.010 0.004 0.004 0.008 0.000
Know. 0.110***
0.031***
-0.001 0.007 0.056**
0.114***
0.086**
(0.010) (0.003) (0.001) (0.027) (0.025) (0.023) (0.029)
R20.086 0.080 0.001 0.000 0.003 0.016 0.007
Weights
Yes Yes Yes Yes Yes Yes Yes
Obs. 3,235 3,235 3,235 3,235 3,235 3,235 3,235
Notes: ***, ** and * indicate 1%, 5% and 10% significance levels respectively. In parentheses are
presented robust standard errors clustered by individual. Constant included but not reported.
Owner. (know.) captures whether an individual owns (knows or recognises) at least one
cryptocurrency.
4.1.2 Payment behaviour and sociodemographics: joint regressions
Table 4presents a new set of regressions that examine the partial effect of payment
behaviour and broader sociodemographics. It shows the econometric results of
12
estimating Equation (1) when the main outcome variable is owning at least one
cryptocurrency. Since education and income are jointly influenced by an unmeasured
third variable (latent variable), they are included separately in each regression. This
is a way of dealing with the possible endogeneity problem.
29
In addition, weights are
included. Columns 1 and 2 present the main results excluding the payment behaviour
indicators. Columns 3-4 and 5-6 present the outcomes including the digitalisation
variables (i.e., having a debit card, usage of mobile app payments, and usage of
PayPal) and the convenience and security variables respectively. Finally, columns 7
and 8 present the results with all variables included at the same time.30
Security concerns have no impact on cryptocurrency investments also conditioned
on sociodemographic indicators. This highlights that the demand for cryptocurrencies
is not driven by distrust in cash, traditional banking payments or online banking
payments. Both having a debit card and having used a mobile app to make a payment
in the last 12 months increase the probability of owning a cryptocurrency by 0.9 and
2.4-2.7 percentage points, respectively. Compared with non-owners, cryptocurrency
investors tend to find traditional banking services less convenient. There are no
differences with respect to cash and online banking payments.
Income level does not affect the main outcome variable of this specification. Each
additional year of age reduces, on average, the probability of owning a cryptocurrency
by 0.1 percentage points. In order to corroborate this result, Table F1, in Appendix
F, reproduces Table 4, but instead of using the variable “age” it uses the variable
“retired”. “Retired” captures whether a person has withdrawn from active working
life or not. Results show that being retired reduces the probability of owning a
cryptocurrency, on average, by between 1 and 1.7 percentage points. Albert and
Duffy (2012) show that older adults are more risk-averse than younger adults, having
a lower expected utility from future income. The study of Stix (2019) also supports
this finding. He argues that potential adopters of cryptocurrencies are younger
and are more willing to accept financial risks. Analysing mortality beliefs, Heimer
et al. (2019) estimate that older individuals place more weight on natural ageing,
overestimating long-run survival rates. This reduces consumption and investment
during retirement, in line with our results.
Being a man in the US increases, on average, the likelihood of owning at least one
cryptocurrency by 2 to 2.2 percentage points.
31
As Jianakoplos and Bernasek (1998)
document, there are gender differences in financial risk-taking. When it comes to
holding risky assets, women tend to be more risk-averse than men and therefore are
unlikely to trust digital currencies that are volatile. Related research by Borghans
et al. (2009) and Arano et al. (2010) also support this finding. Gender differences
also extend to a sector closely related to the cryptocurrency industry: the FinTech
industry. Chen et al. (2021) find that men are more likely to use Fintech products
and services than women. Being married seems to increase the likelihood of investing
29
Table E1 presents the results of estimating Equation (1) with both income and education
included at the same time. Results do not differ from those presented in Table 4.
30
In order to test multicollinearity, we compute the Variable Inflation Factor (VIF) for the
regressors in columns 7 and 8. The values of all variables lie in the (1, 2) interval.
31
This can be interpreted as the difference between 0.7% of female respondents and 2.9% of male
respondents in the survey (Figure 8a).
13
in cryptocurrencies as well. On the contrary, race does not affect the probability of
owning a cryptocurrency.
In contrast, the higher the educational attainment achieved by the individual, the
more likely it is that that individual owns a cryptocurrency. This result is consistent
with the findings of Black et al. (2018). They show that an extra year of education
increases participation in financial markets, and therefore risk-taking, by 2 percentage
points.
4.1.3 Drivers of knowledge acquisition: joint regressions
For the purpose of this study, it is essential to understand the impact of payment
behaviour variables and socioeconomic characteristics on information and knowledge
about cryptocurrencies. Table 5shows the results of estimating Equation (1) when
the main outcome variable is knowing about at least one cryptocurrency.
In this case, we find that all the digitalisation indicators are statistically significant,
i.e. the digitalisation level increases the knowledge about cryptocurrencies. Those
who recognise at least one cryptocurrency also find traditional banking payments
less convenient and online banking bill payments more convenient than those who
do not. While they also find cash less secure, there is no discernible effect found
regarding their opinion of the safety of either traditional or online banking.
Moving from a lower category of education to a higher one increases the probability,
on average, of recognising at least one cryptocurrency by around 8.7 to 11.1 percentage
points.
32
Similarly, the higher the income level, the higher is the probability (from
2.3 percentage points to 3.1) of knowing about at least one cryptocurrency. On the
other hand, current marital status (being married, separated, divorced or widowed
or never having married) is not statistically significantly related to the dependent
variable. Being a man in the US increases, on average, the probability of knowing
about at least one cryptocurrency by between 9.6 and 12.1 percentage points. This
result is in line with the study of Bannier et al. (2019). They highlight that men
have a higher degree of knowledge regarding bitcoin than women. Age, however,
is not a relevant factor in terms of knowing about at least one cryptocurrency. If
“retired” is incorporated into the model instead of age, results coincide with the ones
in the main regression (Table F2 in Appendix F).
32
The magnitude of the coefficients is generally larger in this section because of the percentage of
individuals who know about cryptocurrencies is much larger than the percentage of owners.
14
Table 4: Ownership – payment behavior and sociodemographics
(1) (2) (3) (4) (5) (6) (7) (8)
Digitalisation variables
Debit
card
0.009*** 0.009** 0.010** 0.009**
(0.003) (0.004) (0.004) (0.004)
Mobile
app
0.029*** 0.026** 0.027*** 0.024**
(0.010) (0.010) (0.010) (0.010)
PayPal 0.007 0.009 0.007 0.008
(0.008) (0.008) (0.008) (0.008)
Convenience variables
Cash -0.006 -0.006 -0.005 -0.005
(0.004) (0.004) (0.004) (0.004)
Trad.
Banking
-
0.007**
-0.006* -0.006* -0.006*
(0.004) (0.004) (0.004) (0.004)
Online
Banking
0.002 0.002 0.000 0.001
(0.004) (0.004) (0.004) (0.004)
Security variables
Cash -0.000 0.000 0.000 0.000
(0.003) (0.003) (0.003) (0.003)
Trad.
banking
0.001 0.000 0.001 0.001
(0.005) (0.005) (0.005) (0.005)
Online
Banking
-0.002 -0.000 -0.002 -0.001
(0.005) (0.005) (0.005) (0.005)
Sociodemographic variables
Educ.
0.009***
0.006*
0.008**
0.005*
(0.004) (0.003) (0.004) (0.003)
Income 0.002* 0.000 0.001 0.001
(0.001) (0.001) (0.001) (0.001)
Age
-
0.001***
-
0.001***
-
0.001***
-
0.001**
(0.000) (0.000) (0.000) (0.000)
Married 0.014**
0.011
0.013**
0.012*
0.013**
0.010
0.012**
0.011*
(0.006) (0.007) (0.006) (0.007) (0.006) (0.007) (0.006) (0.007)
Male
0.022*** 0.020*** 0.022*** 0.021*** 0.022*** 0.020*** 0.022*** 0.021***
(0.008) (0.007) (0.008) (0.007) (0.008) (0.007) (0.008) (0.007)
White -0.000 -0.002 -0.001 -0.001 -0.001 -0.002 -0.001 -0.002
(0.009) (0.009) (0.009) (0.009) (0.009) (0.009) (0.009) (0.009)
Weights
Yes Yes Yes Yes Yes Yes Yes Yes
R20.018 0.022 0.031 0.032 0.025 0.028 0.036 0.037
Obs. 3,235 3,235 3,235 3,235 3,235 3,235 3,235 3,235
Notes: ***, ** and * indicate 1%, 5% and 10% significance levels respectively. In parentheses are
presented robust standard errors clustered by individual. Constant included but not reported.
Trad. Banking: bank account number payments. Online Banking: online banking bill payments.
15
Table 5: Knowledge – payment behavior and sociodemographics
(1) (2) (3) (4) (5) (6) (7) (8)
Digitalisation variables
Debit
card
0.079**
0.058*
0.075**
0.056*
(0.032) (0.032) (0.031) (0.032)
Mobile
app
0.086*** 0.095*** 0.078*** 0.087***
(0.023) (0.025) (0.023) (0.024)
PayPal
0.085*** 0.098*** 0.074*** 0.084***
(0.021) (0.022) (0.021) (0.021)
Convenience variables
Cash 0.010 0.003 0.016* 0.008
(0.009) (0.009) (0.009) (0.009)
Trad.
banking
-
0.044***
-
0.052***
-
0.040***
-
0.049***
(0.009) (0.009) (0.009) (0.009)
Online
Banking
0.040*** 0.041*** 0.032*** 0.034***
(0.011) (0.011) (0.011) (0.011)
Security variables
Cash
-
0.026***
-
0.024***
-
0.023***
-
0.022***
(0.007) (0.007) (0.007) (0.007)
Trad.
banking
-0.008 -0.015 -0.006 -0.011
(0.010) (0.010) (0.010) (0.010)
Online
Banking
0.017 0.019* 0.013 0.014
(0.011) (0.011) (0.011) (0.011)
Sociodemographic variables
Educ.
0.111*** 0.092*** 0.101*** 0.087***
(0.009) (0.010) (0.009) (0.010)
Income
0.031*** 0.025*** 0.028*** 0.023***
(0.003) (0.003) (0.003) (0.003)
Age -0.001 0.000 -0.000 0.001
(0.001) (0.001) (0.001) (0.001)
Married
0.022 -0.041 0.013 -0.036 0.016 -0.044* 0.009 -0.039*
(0.024) (0.025) (0.023) (0.024) (0.023) (0.024) (0.023) (0.023)
Male
0.117*** 0.096*** 0.121*** 0.104*** 0.118*** 0.099*** 0.120*** 0.105***
(0.022) (0.022) (0.022) (0.022) (0.022) (0.022) (0.021) (0.022)
White
0.080***
0.038
0.074***
0.039
0.074***
0.034
0.068**
0.034
(0.028) (0.028) (0.027) (0.027) (0.027) (0.027) (0.026) (0.026)
Weights
Yes Yes Yes Yes Yes Yes Yes Yes
R20.111 0.094 0.136 0.120 0.136 0.124 0.156 0.143
Obs. 3,235 3,235 3,235 3,235 3,235 3,235 3,235 3,235
Notes: ***, ** and * indicate 1%, 5% and 10% significance levels respectively. In parentheses are
presented robust standard errors clustered by individual. Constant included but not reported.
Trad. Banking: bank account number payments. Online Banking: online banking bill payments.
16
4.1.4 Ownership conditioned on knowing about cryptocurrencies
Table 6presents the results when we restrict the sample to those individuals that
know about at least one cryptocurrency. The hypothesis that the demand for
cryptocurrencies is driven by distrust in cash or the financial sector is rejected.
Security concerns have no impact on cryptocurrency investment conditioned on
knowing about cryptocurrencies. Education increases the likelihood of owning a
cryptocurrency conditioned on knowing about at least one cryptocurrency.
Nonetheless, becoming one year older decreases, on average, the likelihood of owning
a cryptocurrency by 0.1 percentage points. These results are consistent with those
presented in Table 4.
4.2 Negative binomial model results
Tables 7and 8show the results of estimating Equation (2), when the main outcome
variable is the number of cryptocurrencies owned and the number of cryptocurrencies
known.
Table 7reveals that, compared with non-owners, cryptocurrency investors show no
differences in their level of security concerns with either cash or commercial banking
services. These investors do find traditional banking services less convenient, results
that are consistent through the whole study. Nevertheless, educational attainment
does not play a role here and income level is not statistically significant when the
digitalisation variables are included in the model. With each increasing year of
age, the difference in the logs of expected counts would be expected to decrease by
0.002–0.003 units, while holding the other variables in the model constant. This
result is in line with the estimates of Stix (2019). The difference in the logs of
expected counts is expected to be 0.068–0.106 units higher for males compared with
females, while holding the other variables constant in the model. Gender differences
in financial risk-taking (Jianakoplos and Bernasek,1998;Borghans et al.,2009;Arano
et al.,2010;Chen et al.,2021) are also present in this case. Being married compared
with being single, divorced or widowed also positively affects the main variable of
interest.
Table 8exhibits results in line with the outcomes presented in Table 5. Its
interpretation is as follows: if an individual had achieved one extra level of education,
the difference in the logs of expected counts would be expected to increase by
around 0.22–0.29 units, while holding the other variables in the model constant.
Moreover, being in a higher category of income increases the difference in the logs of
expected counts by 0.05–0.07 units. The relationship between the number of known
cryptocurrencies and age is negative. The difference in the logs of expected counts is
expected to be 0.49–0.56 units higher for males compared with females, while holding
the other variables constant in the model. Marital status and race are not relevant
factors.
17
Table 6: Ownership conditional on knowing at least one cryptocurrency
(1) (2) (3) (4) (5) (6) (7) (8)
Digitalisation variables
Debit
card
0.015*** 0.011** 0.015*** 0.011**
(0.005) (0.005) (0.006) (0.005)
Mobile
app
0.034*** 0.029** 0.032*** 0.027**
(0.013) (0.013) (0.012) (0.012)
PayPal 0.010 0.012 0.010 0.011
(0.010) (0.010) (0.010) (0.010)
Convenience variables
Cash -0.008 -0.008 -0.006 -0.006
(0.006) (0.006) (0.005) (0.005)
Trad.
Banking
-0.009* -0.008* -0.008* -0.007
(0.005) (0.005) (0.005) (0.005)
Online
Banking
0.002 0.002 0.000 0.001
(0.005) (0.005) (0.005) (0.005)
Security variables
Cash 0.000 0.000 0.000 0.000
(0.003) (0.003) (0.003) (0.003)
Trad.
banking
0.002 0.001 0.002 0.001
(0.007) (0.007) (0.007) (0.007)
Online
Banking
-0.003 -0.001 -0.003 -0.002
(0.006) (0.006) (0.007) (0.006)
Sociodemographic variables
Educ.
0.011**
0.007* 0.010* 0.006
(0.005) (0.004) (0.005) (0.005)
Income 0.001 0.000 0.001 0.000
(0.002) (0.002) (0.002) (0.002)
Age
-
0.001***
-
0.001***
-
0.001***
-
0.001**
(0.000) (0.000) (0.000) (0.000)
Married 0.020**
0.020*
0.020** 0.022** 0.019**
0.019*
0.019** 0.020**
(0.009) (0.011) (0.009) (0.011) (0.008) (0.010) (0.008) (0.010)
Male
0.028*** 0.025*** 0.029*** 0.028*** 0.028*** 0.026*** 0.029*** 0.027***
(0.011) (0.010) (0.011) (0.010) (0.010) (0.010) (0.010) (0.010)
White -0.002 -0.002 -0.001 -0.001 -0.001 -0.001 -0.000 -0.000
(0.013) (0.013) (0.013) (0.013) (0.013) (0.013) (0.013) (0.013)
Weights
Yes Yes Yes Yes Yes Yes Yes Yes
R20.018 0.022 0.031 0.032 0.025 0.028 0.036 0.037
Obs. 3,235 3,235 3,235 3,235 3,235 3,235 3,235 3,235
Notes: ***, ** and * indicate 1%, 5% and 10% significance levels respectively. In parentheses are
presented robust standard errors clustered by individual. Constant included but not reported.
Trad. Banking: bank account number payments. Online Banking: online banking bill payments.
18
Table 7: Number of owned cryptocurrencies (negative binomial model)
(1) (2) (3) (4) (5) (6) (7) (8)
Digitalisation variables
Debit
card
0.106* 0.078 0.100 0.070
(0.057) (0.051) (0.062) (0.057)
Mobile
app
0.066**
0.050*
0.065** 0.052**
(0.029) (0.028) (0.028) (0.023)
PayPal 0.054 0.050* 0.057 0.047*
(0.035) (0.028) (0.038) (0.028)
Convenience variables
Cash
-
0.017**
-0.017* -0.012 -0.013
(0.008) (0.008) (0.008) (0.008)
Trad.
banking
-
0.021***
-
0.016**
-
0.022**
-
0.017**
(0.008) (0.008) (0.010) (0.009)
Online
Banking
0.006 0.007 0.003 0.005
(0.012) (0.010) (0.012) (0.011)
Security variables
Cash 0.010 0.010 0.011 0.011
(0.008) (0.009) (0.010) (0.010)
Trad.
banking
0.010 0.012 0.008 0.010
(0.010) (0.010) (0.010) (0.010)
Online
Banking
-0.004 -0.006 -0.008 -0.008
(0.008) (0.009) (0.011) (0.010)
Sociodemographic variables
Educ. 0.021* 0.005
0.027**
0.013*
(0.011) (0.012) (0.013) (0.009)
Income
0.007**
0.002 0.009* 0.005
(0.003) (0.003) (0.005) (0.004)
Age
-
0.003***
-
0.002**
-
0.003***
-
0.002**
(0.001) (0.001) (0.001) (0.001)
Married 0.058** 0.067** 0.066** 0.072** 0.052*** 0.060** 0.062** 0.066**
(0.023) (0.027) (0.030) (0.031) (0.020) (0.024) (0.028) (0.027)
Male
0.068** 0.072** 0.098** 0.090** 0.079** 0.085*** 0.106** 0.100***
(0.031) (0.031) (0.041) (0.035) (0.038) (0.033) (0.043) (0.038)
White 0.007 0.009 0.022 0.025 0.003 0.001 0.018 0.016
(0.018) (0.020) (0.022) (0.023) (0.020) (0.021) (0.022) (0.021)
Weights
Yes Yes Yes Yes Yes Yes Yes Yes
Pseudo-
R2
0.053 0.086 0.101 0.116 0.077 0.106 0.117 0.131
Obs. 3,235 3,235 3,235 3,235 3,235 3,235 3,235 3,235
Notes: ***, ** and * indicate 1%, 5% and 10% significance levels respectively. In parentheses are
presented robust standard errors clustered by individual. Constant included but not reported.
Trad. Banking: bank account number payments. Online Banking: online banking bill payments.
19
Table 8: Number of known cryptocurrencies (negative binomial model)
(1) (2) (3) (4) (5) (6) (7) (8)
Digitalisation variables
Debit
card
0.245**
0.191*
0.243**
0.189*
(0.116) (0.111) (0.112) (0.108)
Mobile
app
0.388*** 0.352*** 0.374*** 0.341***
(0.077) (0.079) (0.077) (0.079)
PayPal
0.229*** 0.260*** 0.205*** 0.233***
(0.070) (0.071) (0.071) (0.073)
Convenience variables
Cash -0.007 -0.015 0.018 0.005
(0.029) (0.029) (0.028) (0.029)
Trad.
banking
-
0.113***
-
0.114***
-
0.097***
-
0.104***
(0.031) (0.031) (0.032) (0.032)
Online
Banking
0.075** 0.077**
0.051 0.057*
(0.034) (0.034) (0.034) (0.034)
Security variables
Cash
-
0.083***
-
0.075***
-
0.077***
-
0.073***
(0.021) (0.022) (0.021) (0.021)
Trad.
banking
0.005 -0.020 0.011 -0.011
(0.037) (0.036) (0.036) (0.036)
Online
Banking
0.020 0.038 0.010 0.023
(0.040) (0.039) (0.040) (0.040)
Sociodemographic variables
Educ.
0.293*** 0.225*** 0.271*** 0.215***
(0.031) (0.030) (0.033) (0.031)
Income
0.073*** 0.051*** 0.067*** 0.049***
(0.014) (0.013) (0.014) (0.013)
Age
-
0.012***
-
0.008***
-
0.011***
-
0.006***
(0.002) (0.002) (0.002) (0.002)
Married
-0.004 -0.119 -0.022 -0.101 -0.024 -0.136 -0.033 -0.115
(0.077) (0.088) (0.074) (0.086) (0.075) (0.087) (0.073) (0.085)
Male
0.555*** 0.494*** 0.556*** 0.513*** 0.559*** 0.503*** 0.557*** 0.518***
(0.076) (0.073) (0.073) (0.071) (0.075) (0.072) (0.072) (0.071)
White 0.066 -0.016 0.056 -0.008 0.053 -0.024 0.042 -0.020
(0.088) (0.087) (0.086) (0.086) (0.087) (0.086) (0.086) (0.086)
Weights
Yes Yes Yes Yes Yes Yes Yes Yes
Pseudo-
R2
0.038 0.035 0.052 0.047 0.045 0.042 0.058 0.053
Obs. 3,235 3,235 3,235 3,235 3,235 3,235 3,235 3,235
Notes: ***, ** and * indicate 1%, 5% and 10% significance levels respectively. In parentheses are
presented robust standard errors clustered by individual. Constant included but not reported.
Trad. Banking: bank account number payments. Online Banking: online banking bill payments.
20
4.3 Robustness checks
In the baseline specification (Equation (1)), a LPM was employed. One limitation of
the LPM is that it does not impose any restriction on the probability. The inability to
impose restrictions on the values of the regressors or the parameters means we could
obtain estimated probabilities outside the unit interval. The logit model overcomes
this limitation by modelling the probability as a cumulative distribution function
that always lies in the [0,1] interval. Tables 9and 10 reports the estimations using a
logit model. As marginal effects are not constant, the average marginal effects are
computed.
The main results are completely robust to the use of a logit model.
33
Overall, the
findings in Table 9suggest that we can disprove the hypothesis that cryptocurrencies
are sought as an alternative to fiat currencies or regulated banking. Cryptocurrency
investors show no differences in their level of security concerns with either cash
or commercial banking services. They also consider traditional banking services
less convenient. Results highlight that being a digital native boosts the usage of
cryptocurrencies. Education and being a man increase, on average, the probability
of owning at least one cryptocurrency by 0.5 and 1 percentage points, respectively.
On the contrary, becoming one year older decreases, on average, the likelihood of
owning at least one cryptocurrency by 0.1 percentage points. Results in Table 10
are in line with those in Table 5.
4.4 Differences across cryptocurrencies
Are there differences in investor characteristics across groups of cryptocurrencies?
This section starts by introducing the best-known cryptocurrencies. As Figure 3a
shows, the most widely known cryptocurrency is bitcoin, followed by bitcoin cash
and ether. Figure 3b depicts the percentage of owners classified by each kind of
cryptocurrency in 2019. Bitcoin is the most widely owned cryptocurrency, with twice
as many investors as is the case for ether. After ether, litecoin is the third most
owned cryptocurrency.
So far, we have treated the owners of cryptocurrencies as if they were a uniform
and homogeneous group of investors who share the same sociodemographic profile.
Investor characteristics may, however, vary across groups of cryptocurrencies. Figure
3c shows that cryptocurrency owners are generally more educated than the average.
Among the various cryptocurrencies, owners of xrp and ether are the most educated,
while those owning litecoin are the least educated, with bitcoin owners ranking in
the middle. Cryptocurrency owners have a household income level higher than the
average, with owners of xrp, ether and stellar being the wealthiest (Figure 3d).
33
In Appendix G, we control for potential rare event bias in the logit model. The sign and
significance of the main variables of the logistic rare event regression à la King and Zeng (2001) are
completely robust.
21
Table 9: Ownership - payment behaviour and sociodemographics (logit model)
(1) (2) (3) (4) (5) (6) (7) (8)
Digitalisation variables
Debit
card
0.031**
0.026*
0.031**
0.026*
(0.015) (0.015) (0.015) (0.015)
Mobile
app
0.023*** 0.019** 0.021** 0.017**
(0.009) (0.008) (0.008) (0.007)
PayPal 0.008 0.009 0.009 0.009
(0.007) (0.007) (0.008) (0.007)
Convenience variables
Cash -0.005* -0.005 -0.004 -0.003
(0.003) (0.003) (0.003) (0.003)
Trad.
banking
-
0.007**
-0.006*
-
0.006**
-0.006*
(0.003) (0.003) (0.003) (0.004)
Online
Banking
0.002 0.002 0.000 0.000
(0.003) (0.003) (0.003) (0.003)
Security variables
Cash -0.001 -0.000 -0.001 -0.001
(0.003) (0.002) (0.002) (0.002)
Trad.
banking
0.001 0.001 0.001 0.001
(0.004) (0.004) (0.004) (0.004)
Online
Banking
-0.002 -0.002 -0.002 -0.001
(0.004) (0.004) (0.004) (0.004)
Sociodemographic variables
Educ.
0.010**
0.006*
0.009**
0.005*
(0.004) (0.004) (0.004) (0.004)
Income 0.002 0.001 0.002 0.002
(0.002) (0.002) (0.002) (0.001)
Age
-
0.001***
-
0.001**
-
0.001**
-
0.001**
(0.000) (0.000) (0.000) (0.000)
Married
0.020* 0.024* 0.020*
0.024**
0.020* 0.023* 0.020*
0.023**
(0.012) (0.013) (0.011) (0.012) (0.012) (0.013) (0.011) (0.012)
Male
0.025** 0.025** 0.025** 0.025** 0.025** 0.024** 0.024*** 0.024***
(0.010) (0.010) (0.010) (0.010) (0.010) (0.010) (0.009) (0.009)
White -0.000 -0.003 0.001 -0.003 -0.000 -0.003 0.001 -0.002
(0.009) (0.009) (0.009) (0.009) (0.009) (0.009) (0.009) (0.009)
Weights
Yes Yes Yes Yes Yes Yes Yes Yes
Pseudo-
R2
0.110 0.168 0.184 0.217 0.151 0.197 0.213 0.238
Obs. 3,235 3,235 3,235 3,235 3,235 3,235 3,235 3,235
Notes: ***, ** and * indicate 1%, 5% and 10% significance levels respectively. In parentheses are
presented robust standard errors clustered by individual. Constant included but not reported.
Trad. Banking: bank account number payments. Online Banking: online banking bill payments.
22
Table 10: Knowledge - payment behaviour and sociodemographics (logit model)
(1) (2) (3) (4) (5) (6) (7) (8)
Digitalisation variables
Debit
card
0.066**
0.050*
0.064**
0.048*
(0.027) (0.028) (0.026) (0.027)
Mobile
app
0.098*** 0.107*** 0.086*** 0.094***
(0.026) (0.028) (0.025) (0.027)
PayPal
0.089*** 0.102*** 0.076*** 0.087***
(0.023) (0.023) (0.023) (0.023)
Convenience variables
Cash 0.010 0.002 0.015 0.007
(0.009) (0.009) (0.009) (0.009)
Trad.
banking
-
0.045***
-
0.053***
-
0.041***
-
0.049***
(0.010) (0.010) (0.009) (0.010)
Online
Banking
0.039*** 0.040*** 0.028*** 0.032***
(0.011) (0.011) (0.011) (0.011)
Security variables
Cash
-
0.024***
-
0.024***
-
0.022***
-
0.021***
(0.007) (0.007) (0.007) (0.007)
Trad.
banking
-0.011 -0.017 -0.008 -0.013
(0.011) (0.010) (0.011) (0.010)
Online
Banking
0.020* 0.022* 0.015 0.016
(0.011) (0.012) (0.011) (0.011)
Sociodemographic variables
Educ.
0.110*** 0.090*** 0.098*** 0.084***
(0.009) (0.009) (0.009) (0.009)
Income
0.028*** 0.021*** 0.025*** 0.020***
(0.002) (0.003) (0.002) (0.003)
Age -0.001 0.000 -0.000 0.001
(0.001) (0.001) (0.001) (0.001)
Married
0.023 -0.037 0.011 -0.035 0.016 -0.040* 0.006 -0.038*
(0.023) (0.024) (0.022) (0.023) (0.022) (0.023) (0.022) (0.023)
Male
0.118*** 0.095*** 0.122*** 0.104*** 0.118*** 0.100*** 0.120*** 0.105***
(0.022) (0.022) (0.021) (0.022) (0.021) (0.022) (0.021) (0.021)
White
0.076***
0.035
0.066***
0.034
0.069***
0.030
0.060**
0.029
(0.026) (0.026) (0.025) (0.026) (0.025) (0.025) (0.025) (0.025)
Weights
Yes Yes Yes Yes Yes Yes Yes Yes
Pseudo-
R2
0.097 0.077 0.121 0.102 0.121 0.104 0.139 0.124
Obs. 3,235 3,235 3,235 3,235 3,235 3,235 3,235 3,235
Notes: ***, ** and * indicate 1%, 5% and 10% significance levels respectively. In parentheses are
presented robust standard errors clustered by individual. Constant included but not reported.
Trad. Banking: bank account number payments. Online Banking: online banking bill payments.
23
Figure 3: Knowledge and owners by group of crypto (2019).
(a) Recognising crypto (in %). (b) Cryptocurrency owners (in %).
(c) Education average by crypto owner. (d) Income average by crypto owner.
Source: 2019 SCPC.
Note. In panels (c) and (d), the brown line represents the average income and education for
all individuals (see Appendix C, Tables C1 and C2 for the household income and educational
attainment classifications respectively). Survey weights are included.
5 Trends in and outlook for the cryptocurrency
industry
What do the findings of this study imply for the future of the cryptocurrency industry?
We discuss whether cryptocurrencies are likely to attract new investors in the future
and whether they may retain their existing ones. We also discuss trends in the
socioeconomic characteristics of cryptocurrency investors.
5.1 Attracting new investors
While knowledge about cryptocurrencies is becoming pervasive, ownership remains
limited to a niche population. In 2014 only some 40% of US citizens were aware of
at least one cryptocurrency (mainly bitcoin). This percentage increased to almost
70% in 2019 (see Figure 4). If the trend continues, in one or two years the entire US
population will recognise at least one cryptocurrency. The acceptance and usage of
cryptocurrencies are nonetheless not high. Only 1.4% of the US population owned
24
at least one cryptocurrency in 2019.
34
The fraction of crypto-owners experienced a
positive trend since 2016 with a slight decline in 2019.
Figure 4: Cryptocurrency facts
Source: SCPC.
Note. Survey weights are included.
Since knowledge about cryptocurrencies is already pervasive, it is not likely that
significant numbers of new investors will be won over to the asset class via the route
of new people learning about the topic. Figure 5shows the amount of new weekly
Twitter followers of some of the major cryptocurrency exchanges, such as Binance or
Coinbase. There was a significant increase of Twitter followers at the end of the first
bitcoin rush.
35
By the end of 2020, cryptocurrency exchanges were gaining some new
followers, but not as much as during the last quarter of 2017 or the first quarter of
2018. It is likely that once a person follows one of these accounts, she is much more
likely to show interest in a cryptocurrency and to invest in the future. Shen et al.
(2019) show that the number of tweets concerning bitcoin can predict the next day’s
trading volume and volatility.
At the end of 2017, LendEDU launched two surveys
36
asking about bitcoin as
an investment option. At the time, over 80% of respondents believed that bitcoin
would be the largest cryptocurrency in terms of market capitalisation in five years
(Figure 6a). Asked whether they would like to own bitcoin through futures contracts
if that were possible, more than 40% of the respondents answered affirmatively, while
34
This percentage is in line with that of other countries. For instance, Stix (2019) shows that
about 1.5% of Austrians own crypto-assets.
35Note that it is possible to game this statistic by buying Twitter followers.
36
Both surveys were answered by the same people, i.e., 564 Americans. See Gitlen (2017) for
further details about the methodology of the survey.
25
Figure 5: New Weekly Twitter Followers of Cryptocurrency Exchanges
Source: Socialblade (retrieved from The Block).
around 34% were not sure (Figure 6b). As a result of the high volatility of bitcoin
and other cryptocurrencies, the market for cryptocurrency futures contracts grew
during 2020 as cryptocurrency investors started to make agreements to buy or sell
a cryptocurrency at a later date for a fixed price. Concurrently, some regulatory
agencies banned the sale of cryptocurrency derivatives and exchange-traded notes,
arguing that these products pose harm and so are ill-suited for the average investor.
The UK Financial Conduct Authority (FCA), for example, was the first agency to
take such a step, banning the sale of cryptocurrency derivatives and exchange-traded
notes.37
Figure 6: Bitcoin’s prospects
(a) In 5 years, will Bitcoin be the largest crypto
in terms of market capitalization? (in %).
(b) If possible, would you rather own Bitcoin
through futures contracts? (in %).
Source: Gitlen (2017).
Note: These questions belong to Part #2 of the survey conducted in November 2017, and correspond
to questions 4 and 3, respectively.
37For further information, see FCA (2020).
26
5.2 Retaining existing investors
One of the main features of the cryptocurrency market is its volatility. The price
of cryptocurrencies can rise and fall dramatically in the course of a single day.
Recognising that the market is prone to such swings, as well as manipulation, the
“holding strategy” has evolved. Hodling means to buy a cryptocurrency and hold
onto it for a prolonged time without any selling or trading activity. But what is
the probability of hodling in the population we examine? To answer this question,
we compute the likelihood that an individual that has a cryptocurrency one year
continues being an owner the following year.
38
To do so, we estimate Equation (1)
maintaining just those individuals that are repeated each two years in the SCPC.
39
Figure 7presents the estimated coefficients for the variable of owning at least
one cryptocurrency in the previous year.
40
Five regressions are performed. Results
are as follows. The likelihood of owning a cryptocurrency in 2015 and 2016 is not
affected by owning a cryptocurrency in 2014 and 2015 respectively.
Figure 7: Estimated coefficients (LPM)
Source: Authors’ elaboration.
Note: The regression of the 2014-2015 waves contains 900 repeated individuals. The 2015-2016,
2016-2017, 2017-2018 and 2018-2019 waves have 1013, 2575, 2526, 2652 repeated individuals
respectively.
Notwithstanding, owning a cryptocurrency in 2016, 2017 and 2018 increases the
probability, on average, of owning a cryptocurrency in 2017, 2018 and 2019 by 61,
38This measure only captures one dimension of hodling. Investors might also increase or reduce
the size of their position.
39
For instance, if an individual responded to the 2014 survey but did not answer the 2015 wave,
this observation is deleted.
40
Figure H1 in the Appendix H shows the estimated coefficients and standard errors presented in
Figure 7.
27
56 and 55 percentage points, respectively. In other words, those who invested in
cryptocurrencies in the past are likely to remain invested.
If this finding – that hodling has become more pervasive, remains constant in
years to come, it may indicate a certain stabilisation in cryptocurrency markets
since investors may be using this asset as a store of value rather than a speculative
asset. This confirms concurrent work in the field of computer science Abramova et al.
(2021), which identifies that the population of cryptocurrency users has grown out of
the original group of tech-savvy “cypherpunks” into a heterogeneous community of
individuals, including both professional and amateur investors (called “hodlers” and
“rookies”, respectively).
5.3 Trends in the gender and age gap
The impact of characteristics such as gender and age on cryptocurrency investments
may be driven by preferences rather than differences in knowledge about the
underlying technology. In terms of gender, the knowledge gap has decreased over
time. In 2014 only 30% of women had knowledge of at least one cryptocurrency (the
percentage was 50% for men), while in 2019 almost 65% of women (and 75% of men)
recognised at least one cryptocurrency (Figure 8b). However, during that time, an
ownership gap actually emerged, evidencing that preferences matter. As Figure 8a
shows, whilst the percentage of male and female owners was pretty similar from
2014 to 2017, in 2018 a gender gap emerged.41
At the same time, Figure 8c presents the age profile of cryptocurrency users.
Owners are younger, on average, than non-owners of cryptocurrencies. This has a
strong effect on investment decisions. On the other hand, the age profile of those who
recognise at least one cryptocurrency is similar to those who have not (Figure 8d).
This evidences that differences in ownership across age are not driven by knowledge,
but rather by preferences
41
The gender gap is present not only in the cryptocurrency industry but also in the FinTech
industry. Chen et al. (2021) find that only 21% of women use fintech products and services while
29% of men do.
28
Figure 8: Cryptocurrency trends (2014-19).
(a) % who own cryptocurrency
(b) % with knowledge about cryptocurrencies
(c) Age of owners vs age of non-owners
(d) Age of people with knowledge about crypto
vs. age of people without it
Source: 2019 SCPC.
Note. Survey weights are included.
6 Conclusion and policy implications
Providing an in-depth exploration of representative data on cryptocurrency owners,
we disprove the hypothesis that cryptocurrencies are sought as an alternative to
fiat currencies or regulated finance: compared with the general population, US
cryptocurrency investors show no differences in their level of security concerns with
either cash or commercial banking services.
We also examine the underlying socioeconomic correlates of cryptocurrency
investing. We find that men tend to invest more in cryptocurrencies than women.
Furthermore, higher levels of income and education, and having digital financial
experience (captured through having a debit card and using a mobile app to pay
for products and/or services) increase the likelihood of recognising at least one
cryptocurrency.
29
We show that these patterns are driven by the impact of socioeconomic features
on knowledge about cryptocurrencies, but also on investment decisions conditional
on knowledge about this asset class. Among the various cryptocurrencies, owners of
ether and xrp have the highest income and educational levels, while those owning
litecoin are the least educated. Last, we document that owning a cryptocurrency
increases the probability, on average, of owning a cryptocurrency in one’s portfolio
the following year by more than 50%.
Finally, we provide some evidence that the impact of gender and age on
cryptocurrency investment is unrelated to differences in knowledge about the
underlying technology. For example, while knowledge levels have converged over the
sample we observe, a gender gap in terms of ownership has emerged.
From a policy perspective, the overall takeaway of our analysis is that as the
objectives of investors are the same as those for other asset classes, so should be the
regulation.
42
Cryptocurrencies are not sought as an alternative to fiat currencies or
regulated finance, but instead are a niche digital speculation object. A clarifying
regulatory and supervisory framework for cryptocurrency markets may be beneficial
for the industry. In fact, regulatory announcements have had a strong impact on
cryptocurrency prices and transaction volumes (Auer and Claessens, 2019, 2020), and
those pointing to the establishment of specific regulations tailored to cryptocurrencies
and initial coin offerings are strongly correlated with relevant market gains.
Better regulation may also be beneficial – quintessential in fact – for the industry
when it comes to the basic security model of many cryptocurrencies. This is so as the
long-term viability of cryptocurrencies based on proof-of-work is questionable. Auer
(2019a) shows that proof-of-work can only achieve payment security (i.e., finality) if
the income of miners is high,
43
and it is questionable whether transaction fees will
always be high enough to generate an adequate level of income to guarantee save
transactions and rule out majority attacks. In the particular for the case of Bitcoin,
the security of payments will decrease each time the “block subsidy” declines (Auer,
2020). Potential solutions44 often involve some degree of institutionalisation, which
in the long-run may require regulation or supervision.45
In the light of these considerations, an important point regards how one could
42
Bouri et al. (2017) stress that although bitcoin was a poor hedge in the 2011-2015 period, it
may be suitable for diversification purposes. In the same spirit, Corbet et al. (2018) find that
cryptocurrencies may serve as diversification assets for investors in the short-term. Bonaparte
(2021) argues that cryptocurrency investors consider cryptocurrencies as a portfolio diversification
vehicle.
43
See also Chiu and Koeppl (2017) and Budish (2018) for related arguments of the cost of
decentralised trust, and Leshno and Strack (2020) for a generalization.
44
Hasu et al. (2019) and Moroz et al. (2020) propose protocol-level changes, among others by
increasing miner income or implementing double-spend counter-attacks. Other important platforms
in the crypto sphere have already moved or are planning to move to proof-of-stake (Kim,2021).
One of the drawbacks of proof-of-stake is however that so called “long-run” attacks may occur.
45
Also the enforcement of anti-money laundering and countering the financing of terrorism
regulations are paramount for the industry (FATF,2014,2020).Coelho et al. (2021) map out
regulatory approaches followed around the world in this respect.
30
apply technology-neutral regulation to this asset class, while at the same time
harnessing the potential of the technology itself in the supervision process. In this
regard, one promising option that supervisory and regulatory agencies could pursue
is “embedded supervision” (Auer,2019b). By this, we understand implementing
a supervisory framework for cryptocurrencies that allows for compliance to be
automatically monitored by reading the market’s ledger. The main aim is low-cost
supervision of decentralised markets, which may be particularly relevant amidst recent
deliberations of the need for adequate prudential oversight of the cryptocurrency
industry (Basel Committee,2019,2021).
31
References
Abramova, S., Voskobojinikov, A., Beznosov, K., and Böhme, R. (2021). Bits under
the Mattress: Understanding different Risk Perceptions and Security Behaviors
of Crypto-asset users. Proceeding of the ACM Conference on Human Factors in
Computing Systems, in press.
Akhtar, T. (2021). Ringleaders of PlusToken Scam Jailed for Up to 11 Years.
Coindesk, https://www.coindesk.com/ringleaders-of-plustoken-scam-jailed-for-up-
to-11-years (site accessed on 12.9.2020).
Albert, S. M. and Duffy, J. (2012). Differences in Risk Aversion Between Young and
Older Adults. Neuroscience and Neuroeconomics, 2012(1).
Arano, K., Parker, C., and Terry, R. (2010). Gender-based Risk Aversion and
Retirement Asset Allocation. Economic Inquiry, 48(1):147–155.
Arner, D., Auer, R., and Frost, J. (2020). Stablecoins: Risks, Potential and Regulation.
Revista de Estabilidad Financiera, Banco de España, (39):99–127.
Auer, R. (2019a). Beyond the Doomsday Economics of ’Proof-of-work’ in
cryptocurrencies. BIS Working Papers, (765).
Auer, R. (2019b). Embedded Supervision: How to Build Regulation into Blockchain
Finance. BIS Working Papers, (811).
Auer, R. (2020). The Security Trilemma and the Future of Bitcoin.
Coindesk, https://www.coindesk.com/the-halving-exposes-bitcoin-to-51-attacks-
heres-what-we-can-do (site accessed on 04.03.2021).
Auer, R. and Böhme, R. (2020). The Technology of Retail Central Bank Digital
Currency. BIS Quarterly Review, March, pages 85–100.
Auer, R. and Claessens, S. (2019). Cryptocurrencies: Why not (to) regulate? In
Fatás, A., editor, The Economics of Fintech and Digital Currencies, pages 83–90.
VoxEU Books.
Auer, R. and Claessens, S. (2020). Cryptocurrency Market Reactions to Regulatory
News. Federal Reserve Bank of Dallas, Globalization Institute Working Papers,
(381).
Auer, R., Cornelli, G., and Frost, J. (2020). Rise of the Central Bank Digital
Currencies: Drivers, Approaches and Technologies. BIS Working Papers, (880).
Baker, P. (2019). What Do Women Want? More Educational Materials Before
Investing in Bitcoin. Coindesk. https://www.coindesk.com/what-do-women-want-
more-educational-materials-before-investing-in-bitcoin (site accessed 06.06.2021).
Balloch, A., Nicolae, A., and Philip, D. (2015). Stock Market Literacy, Trust, and
Participation. Review of Finance, 19(5):1925–1963.
Bannier, C., Meyll, T., Röder, F., and Walter, A. (2019). The Gender Gap in ‘Bitcoin
Literacy’. Journal of Behavioral and Experimental Finance, 22:129–134.
32
Basel Committee (2019). Designing a Prudential Treatment for Cryptoassets. Basel
Committee on Banking Supervision, BIS Discussion paper.
Basel Committee (2021). Prudential treatment of cryptoasset exposures. Basel
Committee on Banking Supervision, BIS Consultative Document.
BBC (2020). PayPal allows Bitcoin and Crypto spending. BBC,
https://www.bbc.com/news/technology-54630283 (site accessed on 1.14.2021).
BBC (2021). Bitcoin falls further as China cracks down on Crypto-currencies. BBC,
https://www.bbc.com/news/business-57169726 (site accessed on 05.28.2021).
Black, S. E., Devereux, P. J., Lundborg, P., and Majlesi, K. (2018). Learning to
Take Risks? The Effect of Education on Risk-taking in Financial Markets. Review
of Finance, 22(3):951–975.
Bohr, J. and Bashir, M. (2014). Who uses Bitcoin? An Exploration of the Bitcoin
Community. Twelfth Annual International Conference on Privacy, Security and
Trust, pages 94–101.
Bonaparte, Y. (2021). On the Portfolio Choice of Crypto Asset Class: Meet the
Gentlemen Investors. SSRN Electronic Journal.
Borghans, L., Heckman, J. J., Golsteyn, B. H., and Meijers, H. (2009). Gender
Differences in Risk Aversion and Ambiguity Aversion. Journal of the European
Economic Association, 7(2-3):649–658.
Borio, C. (2019). On Money, Debt,Trust and Central Banking. BIS Working Papers,
(763).
Bouri, E., Molnár, P., Azzi, G., Roubaud, D., and Hagfors, L. I. (2017). On the
Hedge and Safe Haven Properties of Bitcoin: Is it really more than a Diversifier?
Finance Research Letters, 20:192–198.
Brainard, L. (2018). Cryptocurrencies, Digital Currencies, and Distributed Ledger
Technologies: What Are We Learning? Decoding Digital Currency Conference
sponsored by the Federal Reserve Bank of San Francisco.
Budish, E. (2018). The Economic Limits of Bitcoin and the Blockchain. NBER
Working Paper, (24717).
Buterin, V. (2013). A Next-Generation Smart Contract and Decentralized Application
Platform. Ethereum White Paper, https://ethereum.org/en/whitepaper/ (site
accessed on 05.29.2021).
Carstens, A. (2018a). Money in the Digital Age: What Role for Central Banks?
Lecture at House of Finance, Goethe University, Frankfurt, 6 February.
Carstens, A. (2018b). Technology is no Substitute for Trust. Op-ed in the Börsen-
Zeitung, 23 May 2018.
Carstens, A. (2019). The Future of Money and the Payment System: What Role for
Central Banks? Lecture at the Princeton University, Princeton, New Jersey, 5
December 2019.
33
Chen, S., Doerr, S., Frost, J., Gambacorta, L., and Shin, H. S. (2021). The Fintech
Gender Gap. BIS Working Papers, forthcoming.
Chiu, J. and Koeppl, T. (2017). The Economics of Cryptocurrencies - Bitcoin and
Beyond. Queen’s University, Department of Economics Working Papers, (1389).
Chung, J. (2021). This Hedge Fund Made $700 Million on GameStop. Wall
Street Journal, https://www.wsj.com/articles/this-hedge-fund-made-700-million-
on-gamestop-11612390687 (site accessed on 02.04.2021).
Coelho, R., Fishman, J., and Garcia Ocampo, D. (2021). Supervising Cryptoassets
for Anti-money Laundering. FSI Insights, (31).
CoinMarketCap (2021). Historical Data for Bitcoin.
https://coinmarketcap.com/currencies/bitcoin/historical-data/ (site accessed on
05.31.2021).
CompaniesMarketCap (2021). Gold’s Market Cap.
https://companiesmarketcap.com/gold/marketcap/ (site accessed on 05.30.2021).
Corbet, S., Meegan, A., Larkin, C., Lucey, B., and Yarovaya, L. (2018). Exploring
the Dynamic Relationships between Cryptocurrencies and other Financial Assets.
Economics Letters, 165:28–34.
De Bondt, W. (2005). The values and beliefs of european investors. In Knorr Cetina,
K. and Preda, A., editors, The Sociology of Financial Markets, pages 163–186.
Oxford University Press Oxford.
De Vries, A. (2018). Bitcoin’s Growing Energy Problem. Joule, 2(5):801–805.
DeLiema, M., Deevy, M., Lusardi, A., and Mitchell, O. S. (2020). Financial Fraud
among Older Americans: Evidence and Implications. The Journals of Gerontology:
Series B, 75(4):861–868.
Fanusie, Y. and Robinson, T. (2018). Bitcoin Laundering: An Analysis of Illicit
Flows into Digital Currency Services. Center on Sanctions and Illicit Finance
Memorandum, January.
FATF (2014). Virtual Currencies Key Definitions and Potential AML/CFT Risks.
The Financial Action Task Force Report, June.
FATF (2020). 12-Month Review of the Revised FATF Standards on Virtual Assets
and Virtual Asset Service Providers. The Financial Action Task Force Review,
June.
FCA (2020). FCA bans the sale of crypto-derivatives to retail consumers.
Financial Conduct Authority, https://www.fca.org.uk/news/press-releases/fca-bans-
sale-crypto-derivatives-retail-consumers (site accessed on 1.14.2021).
Foley, S., Karlsen, J. R., and Putni
n
,
š, T. J. (2019). Sex, Drugs, and Bitcoin: How
much Illegal Activity is financed through Cryptocurrencies? The Review of
Financial Studies, 32(5):1798–1853.
34
Foster, K., Greene, C., and Stavins, J. (2020). The 2019 Survey of Consumer Payment
Choice: Summary Results. Federal Reserve Bank of Atlanta, Research Data Report,
(20-3).
Foster, K., Meijer, E., Schuh, S., and Zabek, M. A. (2009). The 2008 Survey of
Consumer Payment Choice. Federal Reserve Bank of Boston Research Department
Public Policy Discussion Papers, (9-10).
Foster, K., Meijer, E., Schuh, S., and Zabek, M. A. (2011). The 2009 Survey of
Consumer Payment Choice. Federal Reserve Bank of Boston Research Department
Public Policy Discussion Papers, (11-1).
Foster, K., Schuh, S., and Zhang, H. (2013). The 2010 Survey of Consumer Payment
Choice. Federal Reserve Bank of Boston Research Department Public Policy
Discussion Papers, (13-2).
Frost, J., Shin, H. S., and Wierts, P. (2020). An early Stablecoin? The Bank of
Amsterdam and the Governance of Money. BIS Working Papers, (902).
Fröwis, M. and Böhme, R. (2017). In Code We Trust? Measuring the Control Flow
Immutability of All Smart Contracts Deployed on Ethereum. In Garcia-Alfaro, J.,
Navarro-Arribas, G., Hartenstein, H., and Herrera-Joancomartí, J., editors, Data
Privacy Management, Cryptocurrencies and Blockchain Technology, pages 357–372.
Springer.
Fujiki, H. (2020). Who Adopts Crypto Assets in Japan? Evidence From the 2019
Financial Literacy Survey. Journal of the Japanese and International Economies,
58:101107.
Georgarakos, D. and Pasini, G. (2011). Trust, Sociability, and Stock Market
Participation. Review of Finance, 15(4):693–725.
Gitlen, J. (2017). Investing in Bitcoin: Survey & Report. Lendedu Research.
Greene, C., O’Brien, S., and Schuh, S. (2017). U.S. Consumer Cash Use, 2012 and
2015: An Introduction to the Diary of Consumer Payment Choice. Federal Reserve
Bank of Atlanta, Research Data Report, (17-6).
Greene, C. and Schuh, S. (2017). The 2016 Diary of Consumer Payment Choice.
Federal Reserve Bank of Atlanta, Research Data Report, (17-7).
Guiso, L., Sapienza, P., and Zingales, L. (2003). People’s Opium? Religion and
Economic Attitudes. Journal of Monetary Economics, 50(1):225–282.
Guiso, L., Sapienza, P., and Zingales, L. (2004). The Role of Social Capital in
Financial Development. American Economic Review, 94(3):526–556.
Guiso, L., Sapienza, P., and Zingales, L. (2008). Trusting the Stock Market. Journal
of Finance, 63(6):2557–2600.
Hasu, Prestwich, J., and Curtis, B. (2019). A Model for Bitcoin’s Security and the
Declining Block Subsidy. https://medium.com/@hasufly/research-paper-a-model-
for-bitcoins-security-and-the-declining-block-subsidy-11a21f600e33.
35
Heimer, R. Z., Myrseth, K. O. R., and Schoenle, R. S. (2019). YOLO: Mortality
Beliefs and Household Finance Puzzles. The Journal of Finance, 74(6):2957–2996.
Henry, C. S., Huynh, K. P., and Nicholls, G. (2018). Bitcoin Awareness and Usage
in Canada. Journal of Digital Banking, 2(4):311–337.
Huang, R. (2020). The ‘Chinese Mining Centralization’ Of Bitcoin and
Ethereum. Forbes, https://www.forbes.com/sites/rogerhuang/2021/12/29/the-
chinese-mining-centralization-of-bitcoin-and-ethereum/?sh=4497e9d62f66 (site
accessed on 02.03.2021).
Jasinski, N. (2021). Why Did Robinhood Stop GameStop Trading? Everything
to Know. Barron’s, https://www.barrons.com/articles/why-did-robinhood-stop-
gamestop-trading-51611967696 (site accessed on 02.03.2021).
Jianakoplos, N. A. and Bernasek, A. (1998). Are Women More Risk Averse? Economic
Inquiry, 36(4):620–630.
Kelly, K., Phillips, M., and Friedman, G. (2021). Trading Curbs Reverse
GameStop Rally, Angering Upstart Traders. The New York Times,
https://www.nytimes.com/2021/01/28/business/gamestop-robinhood.html (site
accessed on 02.03.2021).
Kharif, O. (2020). Bitcoin Whales’ Ownership Concentration Is Rising During Rally.
Bloomberg, https://www.bloomberg.com/news/articles/2020-11-18/bitcoin-whales-
ownership-concentration-is-rising-during-rally (site accessed on 02.01.2021).
Kim, C. (2021). Valid Points: More Ethereum Upgrades to Come After Proof of
Stake, Buterin Says. CoinDesk, https://www.coindesk.com/after-proof-of-stake-
pos-more-upgrades-ethereum-2-0-vitalik-buterin (site accessed on 05.05.2021).
King, G. and Zeng, L. (2001). Logistic regression in rare events data. Political
analysis, 9(2):137–163.
Knorr Cetina, K. and Preda, A. (2005). The Sociology of Financial Markets. Oxford
University Press.
Knorr Cetina, K. and Preda, A. (2012). The Oxford Handbook of the Sociology of
Finance. Oxford University Press.
Krombholz, K., Judmayer, A., Gusenbauer, M., and Weippl, E. (2017). The Other
Side of the Coin: User Experiences with Bitcoin Security and Privacy. In Grossklags,
J. and Preneel, B., editors, Financial Cryptography and Data Security, pages 555–
580. Springer Berlin Heidelberg.
Leshno, J. D. and Strack, P. (2020). Bitcoin: An Axiomatic Approach and an
Impossibility Theorem. American Economic Review: Insights, 2(3):269–86.
Melachrinos, A. and Pfister, C. (2020). Stablecoins: A Brave New World? Banque
de France WP, (757).
Moroz, D. J., Aronoff, D. J., Narula, N., and Parkes, D. C. (2020). Double-Spend
Counter-Attacks: Threat of Retaliation in Proof-of-Work Systems. Cryptoeconomic
Systems, (1).
36
Nakamoto, S. (2008). Bitcoin: A Peer-to-peer Electronic Cash System. White paper.
Ossinger, J. and Hunter, G. S. (2021). Dogecoin Surges 77% as Retail Fever Pitch
Stretches Into Crypto. Bloomberg, https://www.bloomberg.com/news/articles/2021-
01-28/dogecoin-surges-77-as-retail-fever-pitch-stretches-into-crypto (site accessed
on 02.03.2021).
Pan, D. (2020). IOTA Foundation Suspends Network, Probes Fund Theft in Trinity
Wallet. CoinDesk, https://www.coindesk.com/iota-foundation-suspends-network-
probes-fund-theft-in-trinity-wallet (site accessed on 02.02.2021).
Partz, H. (2020). Hacked Italian Exchange Altsbit to Shut Down in May 2020.
CoinTelegraph, https://cointelegraph.com/news/hacked-italian-exchange-altsbit-to-
shut-down-in-may-2020 (site accessed on 01.12.2020).
Pixley, J. (2004). Emotions in Finance: Distrust and Uncertainty in Global Markets.
Cambridge University Press.
Preda, A. (2007). The Sociological Approach to Financial Markets. Journal of
Economic Surveys, 21(3):506–533.
Saiedi, E., Broström, A., and Ruiz, F. (2020). Global Drivers of Cryptocurrency
Infrastructure Adoption. Small Business Economics, pages 1–54.
Schuh, S. and Stavins, J. (2014). The 2011 and 2012 Surveys of Consumer Payment
Choice: Summary Results. Federal Reserve Bank of Boston Research Data Reports,
(14-1).
Schuh, S. and Stavins, J. (2015). The 2013 Survey of Consumer Payment Choice:
Summary Results. Federal Reserve Bank of Boston Research Data Reports, (15-4).
SEC (2020). SEC Charges Ripple and Two Executives with Conducting $1.3 Billion
Unregistered Securities Offering. U.S. Securities and Exchange Commission,
https://www.sec.gov/news/press-release/2020-338 (site accessed on 1.14.2021).
Shen, D., Urquhart, A., and Wang, P. (2019). Does Twitter predict Bitcoin?
Economics Letters, 174:118–122.
Stix, H. (2019). Ownership and Purchase Intention of Crypto-assets – Survey Results.
Oesterreichische Nationalbank Working Papers, (226):1–42.
StormGain (2020). HODL’s meaning. StormGain, https://stormgain.com/blog/hodls-
meaning (site accessed on 18.1.2021).
Temple, J. (2007). Older People and Credit Card Fraud. Trends & Issues in Crime
& Criminal Justice, (343).
Tercero-Lucas, D. (2020). A Global Digital Currency to rule them all? A Monetary-
financial view of the Facebook’s LIBRA for the Euro Area. GEAR WP, (6).
Twomey, D. and Mann, A. (2020). Fraud and Manipulation within Cryptocurrency
Markets. In Alexander, C. and Cumming, D., editors, Corruption and Fraud in
Financial Markets: Malpractice, Misconduct and Manipulation, pages 205–250.
John Wiley & Sons.
37
Voskobojnikov, A., Obada-Obieh, B., Huang, Y., and Beznosov, K. (2020). Surviving
the Cryptojungle: Perception and Management of Risk Among North American
Cryptocurrency (Non)Users. In Bonneau, J. and Heninger, N., editors, Financial
Cryptography and Data Security, pages 595–614. Springer International Publishing.
Walch, A. (2019). Deconstructing ’Decentralization’: Exploring the Core Claim of
Crypto Systems. In Brummer, C., editor, Cryptoassets: Legal, Regulatory, and
Monetary Perspectives, pages 39–68. Oxford University Press.
38
Appendix
This Appendix provides additional explanations, tables and figures that are also
discussed in the paper.
Appendix A. Social networks move markets
A rumour that professional money managers and some hedge funds were shorting
GameStop’s shares spread on the forum Reddit in the mid of January 2021. Boosted
by comments on WallStreetBets, a subreddit page where users discuss stock trading,
a large number of online traders – and some hedge funds
46
– started to buy shares
and stock options, increasing the price of GameStop’s shares. GameStop’s market
capitalisation increased to over $22.6 billion from $5 billion, with the stock opening
on 27 January at $354.83 a share. Many retail investors were using the American
financial services firm Robinhood, which is used to trade stocks and exchange-traded
funds through a mobile app. Robinhood, however, had to cease trading of GameStop’s
shares, as it was facing a $3 billion security demand by its clearing house NSCC
amid a massive spike of trading activity and heightened price volatility (Kelly et al.,
2021).47
46Some hedge funds have made profits of more than $700 million. See Chung (2021).
47
See Jasinski (2021) for further details about all of the factors that led to Robinhood having to
cease trading GameStop shares.
39
Appendix B. Descriptive statistics
Table B1 presents the main descriptive statistics with weights. In the SCPC,
respondents were assigned post-stratified survey weights designed to align as much
as possible the composition of the SCPC sample with that of the Current
Population Study (CPS) (Foster et al.,2020).
Table B1: Weighted Descriptive statistics
Variable Mean Std. Dev. Min. Max.
Main outcome variables
Ownership 0.017 0.131 0 1
Ownership-number 0.04 0.356 0 7
Knowledge 0.709 0.454 0 1
Knowledge-number 1.327 1.382 0 8
Digitalisation variables
Having a debit card 0.826 0.379 0 1
Mobile app for payments 0.301 0.459 0 1
Usage of PayPal 0.379 0.485 0 1
Secur. and conv. variables
Convenience of cash 3.945 1.172 1 5
Security of cash 2.733 1.554 1 5
Conv. of bank acc. number paym.
3.213 1.198 1 5
Sec. of bank acc. number paym. 2.836 1.304 1 5
Conv. of on. bank. bill payments
3.909 1.084 1 5
Security of on. bank. bill paym. 3.244 1.214 1 5
Sociodemographic variables
Income 11.386 4.083 1 16
Education 3.111 1.209 1 5
Married 0.666 0.472 0 1
Age 48.218 16.824 18 109
Retired 0.195 0.396 0 1
Male 0.482 0.5 0 1
White 0.732 0.443 0 1
The final sample includes 3235 observations. Ownership (knowledge) captures whether an
individual owns (recognises or knows) at least one of the following cryptocurrencies: Bitcoin, xrp,
litecoin, ether, bitcoin cash, stellar, eos, or any other different cryptocurrency. Ownership-number
(knowledge-number) stands for the number of different cryptocurrencies that a person owns
(recognises or knows).
40
Appendix C. Income and education
Table C1 shows the household income classification. Table C2 depicts the educational
attainment classification. Education was divided into 16 categories in the 2018 and
2019 waves. However, since it was divided into just five categories in the 2014–17
waves, we reduced it to five.
Table C1: Household income classification
Category Interval Category Interval
1 Less than 5,000. 9 30,000 to 34,999.
2 5,000 to 7,499. 10 35,000 to 39,999.
3 7,500 to 9,999. 11 40,000 to 49,999.
4 10,000 to 12,499. 12 50,000 to 59,999.
5 12,500 to 14,999. 13 60,000 to 74,999.
6 15,000 to 19,999. 14 75,000 to 99,999.
7 20,000 to 24,999. 15 100,000 to 149,999.
8 25,000 to 29,999. 16 150,000 or more.
Source: 2014-19 SCPC.
Table C2: Educational attainment classification
Category Education level
1 12th grade (no diploma) or less.
2 High school graduate - high school diploma or the equivalent.
3 Some college but no degree.
4
Associate degree in college (occupational/vocational program or)
academic program or bachelors degree.
5 Master’s degree, professional school degree or Doctorate degree.
Source: 2014-19 SCPC.
41
Appendix D. Ownership conditioned on having a debit - credit
card
The most common and accepted payment methods to buy cryptocurrencies include
debit cards, credit cards and bank transfers. As the SCPC allows us to restrict
the sample to those individuals that are (i) debit card adopters and (ii) credit card
adopters, we replicate columns 4–9 of Table 3to assess whether our results are
robust.
As Table D1 shows, compared with the general public, cryptocurrency owners
show no differences in their level of security concerns with either cash or commercial
banking services. Cryptocurrency investors find cash and traditional banking services
less convenient. These results are consistent with those presented in Table 3although
the coefficients of cash convenience and traditional banking payments convenience
are higher in magnitude.
Table D1: Ownership conditional on being a debit/credit card adopter
Cash Trad. Banking Online Banking
Conv. Sec. Conv. Sec. Conv. Sec.
(1) (2) (3) (4) (5) (6)
Debit card adopter
Owner. -0.008* -0.001 -0.010*** -0.003 -0.003 -0.001
(0.005) (0.003) (0.004) (0.004) (0.005) (0.005)
R20.005 0.000 0.007 0.001 0.000 0.000
Credit card adopter
Owner. -0.009* -0.002 -0.011*** -0.002 -0.004 -0.001
(0.005) (0.003) (0.004) (0.005) (0.005) (0.005)
R20.005 0.001 0.008 0.000 0.001 0.000
Weights Yes Yes Yes Yes Yes Yes
Obs. 2,636 2,636 2,636 2,636 2,636 2,636
Notes: ***, ** and * indicate 1%, 5% and 10% significance levels respectively. In parentheses are
presented robust standard errors clustered by individual. Constant included but not reported.
Owner. captures whether an individual owns at least one cryptocurrency. Trad. Banking: bank
account number payments; Online Banking: online banking bill payments. Conv. and Sec. stand
for convenience and security respectively.
Appendix E. LPM with income and education
Table E1 and E2 present the results of the LPM with both income and education
included at the same time. If these results are compared with those in Tables 4and
5, they do not differ.
42
Table E1: LPM (ownership) with income and education
(1) (2) (3) (4)
Digitalisation variables
Debit card 0.009** 0.009**
(0.004) (0.004)
Mobile app 0.024** 0.023**
(0.010) (0.009)
PayPal 0.007 0.006
(0.008) (0.008)
Convenience variables
Cash -0.005 -0.004
(0.004) (0.004)
Trad. Banking -0.006* -0.005*
(0.004) (0.004)
Online Banking 0.001 0.000
(0.004) (0.004)
Security variables
Cash 0.000 0.000
(0.003) (0.003)
Trad. banking 0.001 0.001
(0.005) (0.005)
Online Banking -0.001 -0.002
(0.005) (0.005)
Sociodemographic variables
Education 0.008*** 0.007** 0.007** 0.006*
(0.003) (0.003) (0.003) (0.003)
Income 0.000 -0.001 0.000 -0.000
(0.001) (0.001) (0.001) (0.001)
Age -0.001*** -0.001*** -0.001*** -0.000**
(0.000) (0.000) (0.000) (0.000)
Married 0.014* 0.014** 0.012* 0.013*
(0.007) (0.007) (0.007) (0.007)
Male 0.021*** 0.022*** 0.021*** 0.021***
(0.008) (0.008) (0.008) (0.008)
White -0.000 0.000 -0.000 -0.000
(0.009) (0.008) (0.009) (0.008)
Weights Yes Yes Yes Yes
R20.015 0.029 0.023 0.034
Observations 3,235 3,235 3,235 3,235
Notes: ***, ** and * indicate 1%, 5% and 10% significance levels respectively. In parentheses are
presented robust standard errors clustered by individual. Trad. Banking: bank account number
payments. Online Banking: online banking bill payments.
43
Table E2: LPM (knowledge) with income and education
(1) (2) (3) (4)
Digitalisation variables
Debit card 0.054* 0.053*
(0.032) (0.031)
Mobile app 0.080*** 0.075***
(0.025) (0.024)
PayPal 0.077*** 0.068***
(0.022) (0.021)
Convenience variables
Cash 0.010 0.014
(0.009) (0.009)
Trad. Banking -0.046*** -0.044***
(0.009) (0.009)
Online Banking 0.035*** 0.029***
(0.011) (0.011)
Security variables
Cash -0.023*** -0.021***
(0.007) (0.007)
Trad. banking -0.008 -0.005
(0.010) (0.010)
Online Banking 0.014 0.010
(0.011) (0.011)
Sociodemographic variables
Education 0.082*** 0.073*** 0.075*** 0.068***
(0.011) (0.011) (0.011) (0.011)
Income 0.019*** 0.015*** 0.018*** 0.014***
(0.003) (0.003) (0.003) (0.003)
Age -0.001 0.000 -0.000 0.001
(0.001) (0.001) (0.001) (0.001)
Married -0.019 -0.018 -0.023 -0.021
(0.024) (0.024) (0.024) (0.023)
Male 0.107*** 0.113*** 0.108*** 0.112***
(0.022) (0.022) (0.022) (0.021)
White 0.056** 0.055** 0.050* 0.049*
(0.027) (0.027) (0.027) (0.027)
Weights Yes Yes Yes Yes
R20.130 0.147 0.153 0.167
Observations 3,235 3,235 3,235 3,235
Notes: ***, ** and * indicate 1%, 5% and 10% significance levels respectively. In parentheses are
presented robust standard errors clustered by individual. Trad. Banking: bank account number
payments. Online Banking: online banking bill payments.
44
Appendix F. Retired population
Table F1 estimates Equation (1). It reproduces Table 4but instead of using the
variable "age", it uses the variable "retired". "Retired" captures whether a person
has withdrawn from active working life or not.
Results show that being retired reduces the likelihood of owning at least one
cryptocurrency, on average, between 1 and 1.7 percentage points. The rest of the
results are completely in line with those in Section 4.1.2.
Table F2 estimates Equation (1). As in the previous case, it reproduces Table 5
but instead of using the variable "age", it uses the variable "retired".
Results show that being retired does not affect the probability of recognising at
least one cryptocurrency. The rest of the results are consistent with those presented
in Section 4.1.3.
Appendix G. Logistic rare event regression.
Table G1 presents the outcomes of the logistic regression controlling for rare events
à la King and Zeng (2001). The sign and significance of the main variables are
completely consistent with those of the logistic regressions without applying the rare
events correction.
45
Table F1: Owning at least one cryptocurrency
(1) (2) (3) (4)
Digitalisation variables
Debit card 0.010** 0.010**
(0.004) (0.004)
Mobile app 0.029*** 0.027***
(0.011) (0.010)
PayPal 0.009 0.008
(0.008) (0.008)
Convenience variables
Cash -0.006 -0.005
(0.004) (0.004)
Trad. Banking -0.008** -0.006*
(0.004) (0.003)
Online Banking 0.002 0.001
(0.004) (0.004)
Security variables
Cash -0.000 0.000
(0.003) (0.003)
Trad. banking 0.000 0.001
(0.005) (0.005)
Online Banking -0.001 -0.002
(0.005) (0.005)
Sociodemographic variables
Income 0.001 0.000 0.001 0.000
(0.001) (0.001) (0.001) (0.001)
Being retired -0.018*** -0.010*** -0.015*** -0.008***
(0.004) (0.003) (0.004) (0.003)
Married 0.011 0.012* 0.010 0.011
(0.007) (0.007) (0.007) (0.007)
Male 0.020*** 0.022*** 0.020*** 0.021***
(0.007) (0.007) (0.008) (0.008)
White -0.002 -0.001 -0.002 -0.001
(0.009) (0.009) (0.009) (0.009)
Weights Yes Yes Yes Yes
R20.015 0.029 0.023 0.034
Observations 3,235 3,235 3,235 3,235
Notes: ***, ** and * indicate 1%, 5% and 10% significance levels respectively. In parentheses are
presented robust standard errors clustered by individual. Trad. Banking: bank account number
payments. Online Banking: online banking bill payments.
46
Table F2: Knowing at least one cryptocurrency
(1) (2) (3) (4)
Digitalisation variables
Debit card 0.060* 0.058*
(0.032) (0.031)
Mobile app 0.100*** 0.089***
(0.024) (0.023)
PayPal 0.099*** 0.085***
(0.022) (0.021)
Convenience variables
Cash 0.002 0.009
(0.009) (0.009)
Trad. Banking -0.054*** -0.049***
(0.009) (0.009)
Online Banking 0.041*** 0.033***
(0.011) (0.011)
Security variables
Cash -0.024*** -0.023***
(0.007) (0.007)
Trad. banking -0.015 -0.011
(0.010) (0.010)
Online Banking 0.018 0.013
(0.011) (0.011)
Sociodemographic variables
Income 0.031*** 0.025*** 0.028*** 0.023***
(0.003) (0.003) (0.003) (0.003)
Being retired 0.009 0.045* 0.024 0.055**
(0.026) (0.026) (0.025) (0.026)
Married -0.041 -0.035 -0.043* -0.038
(0.025) (0.024) (0.024) (0.023)
Male 0.097*** 0.105*** 0.100*** 0.106***
(0.022) (0.022) (0.022) (0.022)
White 0.037 0.037 0.032 0.031
(0.028) (0.027) (0.027) (0.026)
Weights Yes Yes Yes Yes
R20.093 0.121 0.124 0.145
Observations 3,235 3,235 3,235 3,235
Notes: ***, ** and * indicate 1%, 5% and 10% significance levels respectively. In parentheses are
presented robust standard errors clustered by individual. Trad. Banking: bank account number
payments. Online Banking: online banking bill payments.
47
Table G1: Ownership - Logistic rare events regression
(1) (2) (3) (4) (5) (6) (7) (8)
Digitalisation variables
Debit
card
1.361* 1.079 1.376* 1.110
(0.761) (0.828) (0.779) (0.824)
Mobile
app
1.409*** 1.167*** 1.262*** 1.031**
(0.424) (0.439) (0.401) (0.402)
PayPal 0.489 0.549 0.524 0.567
(0.444) (0.427) (0.471) (0.460)
Convenience variables
Cash -0.319* -0.284 -0.220 -0.198
(0.178) (0.187) (0.170) (0.173)
Trad.
Banking
-
0.449**
-
0.396**
-
0.401**
-0.351
(0.188) (0.201) (0.194) (0.217)
Online
Banking
0.089 0.099 0.008 0.009
(0.192) (0.189) (0.189) (0.193)
Security variables
Cash -0.049 -0.024 -0.040 -0.034
(0.153) (0.154) (0.146) (0.146)
Trad.
banking
0.081 0.055 0.082 0.063
(0.260) (0.239) (0.236) (0.226)
Online
Banking
-0.139 -0.111 -0.129 -0.092
(0.266) (0.256) (0.251) (0.239)
Sociodemographic variables
Educ.
0.584***
0.375*
0.516**
0.330*
(0.195) (0.202) (0.204) (0.208)
Income 0.134 0.068 0.144* 0.087
(0.093) (0.095) (0.087) (0.085)
Age
-
0.079***
-
0.064***
-
0.070***
-
0.056***
(0.020) (0.021) (0.020) (0.021)
Married
1.106* 1.397* 1.153*
1.438**
1.149 1.316* 1.182*
1.366**
(0.669) (0.736) (0.658) (0.698) (0.707) (0.732) (0.684) (0.686)
Male
1.455*** 1.486*** 1.491*** 1.531*** 1.493*** 1.459*** 1.478*** 1.466***
(0.539) (0.549) (0.523) (0.537) (0.503) (0.520) (0.481) (0.495)
White -0.047 -0.221 0.028 -0.180 -0.028 -0.188 0.018 -0.148
(0.553) (0.561) (0.552) (0.554) (0.535) (0.536) (0.534) (0.547)
Weights
Yes Yes Yes Yes Yes Yes Yes Yes
Notes: ***, ** and * indicate 1%, 5% and 10% significance levels respectively. In parentheses are
presented robust standard errors clustered by individual. Constant included but not reported.
Trad. Banking: bank account number payments. Online Banking: online banking bill payments.
48
Appendix H. Owners over time
Table H1 shows the estimated coefficients and standard errors presented in Figure 7.
Table H1: Owners over time
2015 2016 2017 2018 2019
Estimated coefficient 0.132 -0.002 0.612*** 0.556*** 0.549***
(0.128) (0.003) (0.170) (0.119) (0.075)
R20.039 0.014 0.169 0.142 0.418
Observations 900 1,013 2,575 2,526 2,652
Notes: ***, ** and * indicate 1%, 5% and 10% significance levels respectively. In parentheses are
presented robust standard errors clustered by individual.
49
All volumes are available on our website www.bis.org.
Previous volumes in this series
950
June 2021
Fiscal regimes and the exchange rate Enrique Alberola, Carlos Cantú,
Paolo Cavallino and Nikola Mirkov
949
June 2021
The natural interest rate in China Sun Guofeng and Daniel M Rees
948
June 2021
Central bank digital currency: the quest for
minimally invasive technology
Raphael Auer and Rainer Böhme
947
June 2021
Money, technology and banking: what
lessons can China teach the rest of the world?
Michael Chui
946
June 2021
The pricing of carbon risk in syndicated loans:
which risks are priced and why?
Torsten Ehlers, Frank Packer and
Kathrin de Greiff
945
May 2021
US monetary policy and the financial channel
of the exchange rate: evidence from India
Shesadri Banerjee and M S
Mohanty
944
May 2021
Income inequality, financial intermediation,
and small firms
Sebastian Doerr, Thomas Drechsel
and Donggyu Lee
943
May 2021
Income inequality and the depth of economic
downturns
Emanuel Kohlscheen, Marco
Lombardi and Egon Zakrajšek
942
May 2021
FX policy when financial markets are
imperfect
Matteo Maggiori
941
May 2021
The digitalisation of money Markus K Brunnermeier, Harold
James and Jean-Pierre Landau
940
May 2021
Monetary-fiscal crosswinds in the European
Monetary Union
Lucrezia Reichlin, Giovanni Ricco
and Matthieu Tarbé
939
May 2021
The constraint on public debt when
r < g but g < m
Ricardo Reis
938
May 2021
CCP Auction Design Wenqian Huang and Haoxiang Zhu
937
April 2021
Growth, coal and carbon emissions:
economic overheating and climate change
Emanuel Kohlscheen, Richhild
Moessner and Előd Takáts
