Sweden
Sweden
Sweden
August 2019
WORKING PAPERS ARE OBTAINABLE FROM
www.riksbank.se/en/research
Sveriges Riksbank • SE-103 37 Stockholm
Fax international: +46 8 21 05 31
Telephone international: +46 8 787 00 00
The opinions expressed in this article are the sole responsibility of the author(s) and should not be
interpreted as reflecting the views of Sveriges Riksbank.
A Tale of Two Countries: Cash Demand in
Canada and Sweden
Walter Engert, Ben Fung, Björn Segendorfy
Sveriges Riksbank Working Paper Series
No. 376
August 2019
Abstract
Cash is being used less and less for making payments in many countries,
including Canada and Sweden, which might suggest that cash will eventually
disappear. However, cash in circulation in most countries, including Canada,
has been stable for decades, and even rising in recent years. In contrast, ag-
gregate cash demand in Sweden has been falling steadily. This paper explains
these di¤erences between Canada and Sweden by focusing separately on the
transactions demand for cash and on the store-of-value demand. We …nd a
long-term downward trend in small-denomination bank notes relative to gross
domestic product in both Canada and Sweden. This re‡ects similar experi-
ences in decreasing cash use for transactions over time due to the adoption
of payment innovations. This means that payment innovations and di¤usion
are not su¢ cient to explain why aggregate cash demand has been declining
rapidly in Sweden but not in Canada. Instead, the di¤erence in the trends
of cash demand between these two countries is due more to the behaviour of
larger-denomination, store-of-value bank notes. Finally, we identify in‡uences
and frictions that help explain the persistent decline in the demand for larger
bank notes in Sweden relative to Canada.
Keywords: Bank topics: Bank notes; Digital currencies and …ntech; Financial
services; Payment clearing and settlement systems.
JEL codes: E, E4, E41, E42, E5.
We thank Ted Garanzotis, Rod Garratt, Scott Hendry, Annetta Ho, Paul Miller, Steve Thomas,
Maarten van Oordt, and participants in the seminar on ‘Understanding Cash Usage’ in Athens,
Greece, the ‘Shaping a New Reality of Cash’ conference hosted by the Narodowy Bank Polski
in Warsaw, Poland, and seminars at the Bank of Canada and Sveriges Riksbank. The opinions
expressed in this article are the sole responsibility of the authors and should not be interpreted as
re‡ecting the views of the Bank of Canada or Sveriges Riksbank.
y
E-mail: wengert@bankofcanada.ca, bfung@bankofcanada.ca, bjorn.segendorf@riksbank.se.
1. Introduction
Cash is being used less and less for making payments in many countries, including Canada
and Sweden.1 Charts 1a and 1b show that the number and the value of cash transactions have
been declining over the last two decades in both of these countries.2 One of the main reasons
for declining cash use has been the growing reliance on electronic payment methods, such as
credit and debit cards.3 Charts 2a and 2b present the value of cash, debit card and credit card
spending relative to gross domestic product (GDP) in Canada and Sweden. Canadians have
long preferred to use credit cards, and their use of debit cards also overtook cash in the early
2000s (Chart 2a). In comparison, while Swedes were still using cash more than cards in 2001,
they have increasingly preferred debit cards to make purchases since the early 2000s, and in
the last few years credit card use has also overtaken cash (Chart 2b). Corresponding to these
trends, increasing card acceptance by merchants has also contributed to greater card use in
both Canada and Sweden.
Such persistent declines in cash use in numerous countries might suggest that cash will
eventually disappear. Indeed, Chart 3a illustrates that cash demand more generally (the ratio
of cash to GDP) steadily decreased in a number of countries following the end of Second
World War.4 However, the cash-to-GDP ratio stabilized in most countries in the late 1970s
and early 1980s. In Canada, for example, bank notes as a ratio of GDP declined steadily from
10 percent in 1946 but then stabilized in the 1980s and have remained between 3 and 4
percent. There has even been a slight upward trend after the 2008 financial crisis in Canada
and in most other advanced economies (Bech et al. 2018). So, decreasing cash use for
transactions has not resulted in a sustained decline in cash demand in Canada (to this point).
And the story is similar in most other advanced economies (Chart 3a).
The evolution of cash demand in Sweden and Norway, however, has been markedly different
from the experiences of other advanced countries (Chart 3b). In Sweden, for example, the
value of bank notes relative to GDP declined continuously from a peak of 13.5 percent in
1945 to 1.2 percent in 2017. Unlike the experience in most other advanced economies, cash-
to-GDP did not stabilize in the 1980s, nor did it increase after the financial crisis. Moreover,
the absolute value of bank notes outstanding in Sweden started to decline in 2007, from a
peak of SEK 108.5 billion in 2007 to just over SEK 55 billion in 2017. Thus, in Sweden,
declining cash use for payments appears to be more closely associated with the decrease in
cash demand more generally. Indeed, the popular press has been predicting that Sweden will
become the world’s first cashless society when the Swedes stop using cash completely.5
The divergent experiences of Canada (and other advanced countries) on the one hand and
Sweden (and Norway) on the other raise several questions. Why has there been a persistent
decline in overall cash demand in Sweden but not in Canada? Have Swedes been reducing
1 The terms cash and bank notes are used interchangeably in this paper.
2The value and number of cash payments are estimated using cash withdrawal data from automated teller machines (ATMs). For a
discussion of the methodology and limitation of the estimation, see Arango et al. (2012).
3 For a discussion of these developments in Canada, see, e.g., Fung, Huynh and Stuber (2015) and Henry, Huynh and Welte (2018). For
Sweden, see Segendorf and Wretman (2015) and Sveriges Riksbank (2013).
4 The ratio of cash in circulation to GDP is widely used as a measure of the demand for cash and thus the importance of cash in an
economy. The value of cash in circulation, like other goods, is determined by both supply-side factors (e.g. the number of bank branches
and the size of the ATM network) and demand-side factors (e.g. the use of cash for payments and as a store of value). In most countries,
including Canada and Sweden, the central bank typically provides enough cash to meet the demand for cash; thus, the value of cash in
circulation reflects the underlying demand for cash.
5
See, e.g., N. Heller, “Imagining a Cashless World,” The New Yorker, October 3, 2016.
1
cash use more quickly than Canadians? Are there meaningful differences in payment
innovations in the two countries that can explain these experiences? Are there large
differences in the availability and supply of cash? What factors could explain the different
trends in cash demand more generally in the two countries? Does Sweden’s experience of
continuously falling cash demand indicate the future for other advanced economies, such as
Canada?
This paper studies both the use of cash and the demand for cash in Canada and Sweden, and
analyzes the main factors affecting cash use and demand in these two countries. More
specifically, we discuss supply side factors such as the bank note distribution system and the
access of cash through ATMs and bank branches, as well as demand side factors such as
consumers’ use of cash and merchants’ acceptance of cash for transactions, and the holding
of cash as store of value. We also discuss foreign demand and the demand for cash in the
underground economy. In addition, we identify some unique developments in Sweden that
help explain the persistent decline in the demand for bank notes. Such analysis can inform
our understanding of potential turning points for cash demand in Canada, as well as policy
measures that might affect such outcomes.
The Riksbank has been studying the implications of a persistent decline in cash use and
demand in Sweden, highlighting some potential difficulties from a cashless society.6 Engert,
Fung and Hendry (2018) also study the implications of a cashless society and note areas that
could raise some concerns in that context—that is, operational reliability and contestability in
retail payments, and the availability of a safe store of value in an extreme financial crisis.
They also suggest options for policy-makers to deal with these potential problems, including
taking steps to inhibit declining cash demand, regulating retail payment systems and issuing a
central bank digital currency (CBDC) in a cashless society. To assess policy responses, it is
important to understand the factors that could lead to the disappearance of cash in an
economy. In particular, policy-makers might want to know how to avoid or inhibit declining
cash demand. If this is not possible or desirable, then the central bank and other public
authorities would need to consider other policy options to mitigate concerns if a cashless
society were to evolve (as noted in Engert, Fung and Hendry 2018).
Canada and Sweden are similar in many ways. For instance, both of these northern countries
are small, open economies that neighbour on much larger economies (the United States and
those in the euro zone, respectively). Each has its own national currency, while citizens have
easy access to an international reserve currency (the US dollar and the euro). The
macroeconomic environments in Canada and Sweden also share similarities. For example,
both Canada and Sweden have monetary policy frameworks that target inflation, and both
have pursued the same inflation-control objective (2 percent) since the beginning of the
1990s. Their banking systems are broadly similar, each dominated by a handful of large,
universal banks, but also including many smaller institutions. Canada and Sweden also both
rank high in terms of digital infrastructure relevant to cashless payments; both are seen as
among the most significant adopters of cashless payment methods and both are considered
among the most cashless societies in the world.7 Finally, each has experienced a range of
payment innovations over the past few decades (e.g., Swish in Sweden, and Interac e-
Transfer in Canada) that have contributed to the reduced use of cash for payments.
6
See, e.g., Sveriges Riksbank (2017) and Sveriges Riksbank (2018a).
7For rankings of digital infrastructure and network readiness, see the Networked Readiness Index compiled by the World Economic
Forum). For an example of rankings of cashless societies, see, S. Smith, “The 10 Most Cashless Countries in the World: Where Does the UK
Rank?” The Telegraph, October 10, 2017,.
2
And yet, despite the similarities, the evolution of overall cash demand is very different in the
two countries, as shown above. It follows then that studying the factors driving cash demand
in Canada and Sweden could inform our understanding of the importance of various
influences and provide some insight into the evolution of cash demand in Canada: Will cash
demand in Canada follow Sweden’s path?
The rest of the paper is organized as follows. The next section discusses access to cash in
Canada and Sweden, focusing on how bank notes are distributed to the general public. The
following two sections consider transactional and non-transactional demand for cash. More
specifically, the third section discusses the main factors that determine transactional demand
and assesses whether these can explain the differences between cash demand in Canada and
Sweden. Section 4 considers non-transactional demand for cash and related influences.
Section 5 discusses the use of cash in the underground economy. The final section provides
conclusions.
2. Access to cash in Canada and Sweden
Typically, a central bank distributes bank notes to financial institutions; these, in turn, operate
a network of branches and automated teller machines (ATMs) that allow the public to
withdraw or deposit bank notes. This section begins with a brief overview of the banking
systems in Canada and Sweden and then describes how the respective central banks distribute
bank notes to financial institutions. Access to cash by the general public through financial
institutions’ branches and ATMs is also considered. We discuss differences between Canada
and Sweden regarding providing and accessing cash, and whether this affects cash use and
demand.
8These six banks are Bank of Montreal, Banque Nationale, Canadian Imperial Bank of Commerce, Royal Bank of Canada, Scotiabank and
TD Bank.
9These four banks are Handelsbanken, Nordea, SEB (Skandinaviska Enskilda Banken) and Swedbank. In October 2018, Nordea moved its
head office (and legal incorporation) from Sweden to Finland.
3
These banks are also highly interconnected with each other and with other participants in the
financial system. Further, they play important roles in most aspects of the Swedish financial
system. Each of the four major banking groups provides a wide range of financial services,
but they have somewhat different business mixes and geographic concentrations across the
Nordic and Baltic regions and other countries (IMF 2017). Nevertheless, the commercial
banks and their mortgage subsidiaries are the dominant entities for each major bank. Nine
financial institutions have been designated by Swedish authorities as domestic systemically
important (Riksgalden 2018a) and are therefore subject to enhanced regulatory provisions
and a specific, open-bank bail-in resolution regime. One of the major banks operating in
Sweden—Nordea—has also been designated as a global systemically important bank by the
FSB. In addition to the four dominant Swedish banks, there are also many—over 100—
smaller banks, including niche participants and some foreign bank subsidiaries and branches
(Swedish Bankers’ Association 2018).
With low interest rates and interest-spread compression prevailing in the last 10 years, the
major Canadian and Swedish banks have sustained profits with low credit losses, an
expansion of real-estate-based lending (mortgages) and increased fee income, along with
tight control of operating expenses (Finansinspektion 2018; IMF 2017).
In sum, the banking market structures in Canada and Sweden are broadly similar,
characterized by the dominance of a handful of large, universal and systemically important
banks, but with many smaller institutions competing as well. Further, the strategic, business
responses of the major banks in each country to the prevailing economic influences in recent
years have been broadly similar.
A key difference between Canadian and Swedish banks concerns their funding. The major
Canadian banks rely on personal and other deposits, which provide for over half of their
liabilities, almost equally split between demand/notice deposits and fixed-term deposits (IMF
2014). And Canadian-dollar personal deposit funding is about 35 percent of Canadian-dollar
assets. In contrast, the major Swedish banks depend on wholesale, market funding, primarily
through covered bonds. Correspondingly, deposits comprise only about one-third of the
aggregate funding for Swedish deposit-taking institutions (IMF 2017). Similarly, SEK
personal deposits amount to just 17 percent of SEK assets.
4
A member financial institution with surplus notes declares them to the BNDS, receives credit
from the Bank of Canada, and then moves them to the vault of its RDC, segregated from its
own note holdings. A financial institution that needs notes places orders on the BNDS and
arranges for the transportation of notes from other RDCs in the same RDP that have surplus
notes. The Bank then debits its accounts accordingly. Unfit notes are shipped by financial
institution branches to their RDCs, where the notes are registered as deposits to the Bank.
These notes are then shipped by the Bank to one of its AOCs for processing and removal
from circulation. The Bank also supplies new notes (and processed fit notes) as needed by
shipping these notes to the RDCs, where they are held in the vaults of the RDCs until
required.
In Sweden, the Riksbank started issuing bank notes early in the eighteenth century. The
number of Riksbank branches at which banks can obtain cash from the Riksbank has been
declining gradually in recent decades. In the late 1980s, the Riksbank had 20 local branches
that managed note distribution, and by 2006 this number had declined to two. Since 2014,
only one Riksbank regional branch remains in operation. Consequently, the network of cash
distribution depots is now mostly operated by the market.12 Banks bear costs from cash in
terms of foregone interest, which generates an incentive for the banks to deposit cash at the
Riksbank each afternoon and withdraw cash the next morning to cover the day’s needs. To
avoid unnecessary transportation and related costs, banks are credited interest compensation
if they store and report idle cash at the depots in accordance with a specific set of
requirements.
In sum, both the Bank of Canada and the Riksbank have substantially decentralized note
processing, storage and distribution to financial institutions. As a result, financial institutions
bear a significant share of the costs of providing cash to the public. In Canada, for example,
Kosse et al. (2017) estimate that 56 percent of the total resource costs of cash are borne by
financial institutions. Similarly, Segendorf and Jansson (2012) estimate that banks and cash-
in-transit companies bear about half of the total resource costs of cash (and nearly all of the
cash distribution costs specifically). Therefore, financial institutions in both countries are
constantly looking for ways to reduce the costs related to their cash operations.
12Similar to the Canadian experience, the transition in Sweden to rely on market participants to distribute cash was motivated by the
Riksbank’s view that the market is better suited to providing efficient cash distribution than is the central bank itself. See Daltung and
Ericson (2004) for more background on this decision.
13More specifically, over 99 percent of Canadians and 97 percent of Swedes have a bank account according to the latest survey data; for
Canada, see Henry, Huynh and Welte (2018), and for Sweden, Sveriges Riksbank (2018b).
14
The CAD/SEK exchange rate is about 1:7 (July 2019).
5
inhabitants in Canada than in Sweden. Further, the number of ATMs in Canada had been
growing until recently, with much of this increase coming from white-label ATMs. The
number of ATMs in Sweden has been fairly stable over the last 20 years. In Sweden, the
distribution of ATMs has recently started to move away from bank branch locations toward
“cash centres” that congregate several ATMS in high-traffic areas, such as shopping malls.15
Major Canadian institutions also place ATMs in high-traffic areas, as well as in bank
branches.
While there are many more ATMs per capita in Canada than in Sweden, this does not
necessarily mean that there is a friction inhibiting access to transactional cash in Sweden
compared with Canada. The business models underlying the deployment of ATMs in the two
countries appear to be quite different. In Canada, banks apparently see their ATM networks
as part of their branding and a means to attract and retain customers. Major Canadian banks
each have their own branded ATMs to compete for customers, joined in a common network
linking all banks’ ATMs. As a result, customers can withdraw cash from their own-bank
ATMs with no (or minimal) charges. They can also withdraw funds from their account via
another bank’s ATMs or white-label ATMs, but subject to additional fees for most account
holders. And bank customers can deposit cash (and cheques) only at ATMs of their own
bank.
In comparison, Swedish banks appear to have viewed ATMs more as a cost centre than a
means of competition, and thus have collaborated since the late 1960s to work toward shared
and interoperable bank-ATM networks to reduce their costs. By the mid-1990s, bank and
savings bank ATM networks were interoperable to provide for cash withdrawals (but not
deposits) across all ATMs. In 2013, all bank ATMs in Sweden came under the management
of one company, to improve interoperability and service more generally.16 As a result,
customers of any Swedish bank can use any ATM to access their account to withdraw cash
without being subject to any fees. This suggests that a smaller ATM network would be
sufficient to provide for cash distribution (other things equal). In this context, Swedish banks
have also been concerned about possible adverse impact of a common, generic ATM network
on their branding. This has been addressed by having the ATM display the specific card-
issuer’s starting page following recognition of the customer’s inserted access card.17
However, Swedish ATMs do not generally accept cash deposits, and specific cash-deposit
machines have been hard to find. As a result, depositing cash was becoming increasingly
difficult in Sweden (more on this below).18 Recently, however, cash-deposit machines are
being located at cash centres (noted above) along with regular ATMs.
In sum, different competitive and business strategies seem to explain the much smaller
number of ATMs per capita in Sweden than in Canada. But the nature of the interoperability
of the bank ATM networks in Sweden suggests that this is unlikely a material friction in
accessing transactional cash via ATMs. A recent official study (Statens offentliga utredningar
15In 2017 there were about 100 such cash centres, and 30 more were expected to open in 2018. See Bankomat press release, May 17,
2018.
16The Swedish ATM network is currently managed by Bankomat AB, which is jointly owned by the five largest banks in Sweden: Danske
Bank, Handelsbanken, Nordea, SEB, Swedbank (with the Savings Banks) For more information, see Bankomat website. The tradition of co-
operation among Swedish banks in this regard is described in Segendorf and Wretman (2015). See also Bátiz-Lazo (2018), especially Table
6.2.
17
In addition to the dominant ATM network, one bank has kept its own ATM network, and there are some white-label ATMs as well.
Further, ICA (the largest grocery store chain in Sweden) opened ICA Bank, which has an (in-store) ATM network that is interoperable with
the main ATM network. One of the cash-in-transit companies also operates its own ATM network. These two networks each have nearly
20 percent of the market.
18
A search of Bankomat’s website reveals that few of its ATM locations provide cash deposit services.
6
2018) on access to cash in Sweden reached a similar conclusion, but raised questions about
access to deposit services, especially given the development of cashless bank branches in
Sweden (discussed below).
2.5 Summing up
A number of key factors influencing access to cash in Canada and Sweden are similar,
including banking market structures and the bank note distribution systems. While there are
significantly fewer ATMs per capita in Sweden, the nature of the interoperability of the ATM
networks in Sweden suggests that this is unlikely related to any material friction around
accessing transactional cash via ATMs. 21 However, a relative lack of access to deposit
services at ATMs, fewer bank branches per capita and the development of cashless bank
branches in Sweden could represent frictions that inhibit the demand for cash (particularly for
larger value bank notes), compared with in Canada. These aspects are discussed further
below.
3. Small-denomination bank notes and transactional demand for cash
As discussed above, cash-to-GDP in Canada and Sweden trended down for years after the
end of the Second World War. But while the cash-to-GDP ratio stabilized in Canada in the
early 1980s and has even increased somewhat in the last few years, this ratio has continued to
fall steadily in Sweden. As a result, cash-to-GDP in Sweden fell below that of Canada in
2009 (although Sweden has had a higher ratio for most of the post-war period). Considering
small- and large-denomination notes separately illuminates these contrasting trends.
Charts 5a and 5b plot the ratio of small-denomination notes-to-GDP and large-denomination
notes-to-GDP separately. For Canada, notes of less than $50 in value are considered small-
denomination notes, and notes of $50 or more are considered large-denomination bank notes.
For Sweden, notes less than 500 krona are small-denomination notes, and the 500-, 1,000-
19The data for Canada include branches of banks, credit unions and governmental savings institutions; for Sweden, branches of banks and
savings banks are included. All data are from the BIS Redbook.
20A subsidiary of the Swedish Post—Svensk Kassaservice—had provided basic cash and payment services over the counter, but it was
closed in 2008. For more on this, see Statens offentliga utredningar (2018).
21Taken together, the number of cash-withdrawal points in Sweden (ATMs, bank branches and various shops) decreased by 38 percent
between 2011 and 2018 (from 5,232 to 3,212; Länsstyrelserna 2018).
7
and 10,000-krona notes (which were removed from circulation in 1993) are considered large-
denomination bank notes.22
For small denominations, the patterns are similar in Canada and Sweden, with the ratio of
small notes to GDP in both countries following a long-term declining trend.23 This ratio was
generally higher in Sweden than in Canada for most of the immediate post-war period (e.g.,
in 1946, 10.6 percent in Sweden versus 7.4 percent in Canada), suggesting that Sweden was
more cash-intensive than Canada during these decades. However, the value of small notes to
GDP has declined substantially in both countries since 1946, and since 1988 this ratio has
been lower in Sweden than in Canada. In 2018, it reached 0.3 percent in Sweden and 1
percent in Canada.
For large denominations, however, the trends in the two countries have been strikingly
different. In Canada, the value of large notes-to-GDP declined from the late 1940s, stabilized
in the late 1960s, and has been rising since the early 1980s. Further, the rate of increase has
accelerated since the 2008 financial crisis. In Sweden, large notes to GDP increased slightly
from the late 1940s, and this increase accelerated in the1970s. The ratio of large notes to
GDP then stabilized in the early 1990s before going into a persistent downward trend starting
in 2001.
Charts 5a and 5b show that the decline in small denominations was driving the downward
trend in total cash demand in Canada until the 1980s, and in Sweden until the mid-1990s.
After 1980 in Canada, rising demand for large denominations offset declining demand for
small denominations, resulting in fairly steady total cash demand and a slight upward trend
after the 2008 financial crisis. After 1995 in Sweden, total cash demand briefly stabilized (at
around 4 percent) as demand for large denominations remained at a high level while demand
for small denominations continued to decline slowly. In 2001, the demand for large
denominations started to fall, driving the sharp downward trend in total cash demand for the
last two decades in Sweden.
Small-denomination notes are typically used for transactions, especially for day-to-day
purchases, as these notes are widely accepted by merchants and are readily available at
ATMs. While large-denomination notes are also used for transactions, such notes are more
likely held as a store of value.24 The cash-to-GDP ratios in Charts 5a and 5b suggest that cash
held for transactions (small notes) has been consistently declining since the 1940s in both
Canada and Sweden. In contrast, cash held for non-transactional reasons (large notes) has
been following different trends in Canada and Sweden over the last three decades—and so
are the key to explaining the difference in total cash demand over time in Canada and
Sweden.
Indeed, these divergent trends suggest that it is useful to consider separately the factors that
have influenced holdings of small denomination notes and large denomination notes.
Accordingly, the rest of this section focuses on influences affecting the demand for small-
denomination notes in Canada and Sweden. Section 4 then considers the demand for large-
22This classification is consistent with the literature, which typically considers the largest two or three denomination bank notes as “large-
denomination” notes. See, e.g., Amromin and Chkravorti (2009) and Judson (2018). For a long-time series, however, there are some
caveats. Persistent inflation, e.g., could suggest that some notes, especially $50 and 500-krona notes, should not be considered to be
“large denomination” (primarily store-of-value) notes throughout the entire period under consideration. That is, with a rising price level,
such notes become more suitable for making payments over time. This perspective is reinforced by fact that these particular notes ($50
and 500-krona) have become available in ATMs in Canada and in Sweden, respectively.
23It should be noted that small-denomination notes relative to GDP are also declining in many other advanced economies; see, e.g.,
Judson (2018).
24
See Amromin and Chkravorti (2009).
8
denomination notes, where there appear to have been some notable differences between the
two countries.
25See also Henry, Huynh and Welte (2018), Chart 3. Note that contactless credit cards and debit cards in Canada have a limit on the size of
the transaction, set by the issuing bank, typically at CAD 100.
26 Fung, Huynh and Sabetti (2014) and Chen, Felt and Huynh (2017) show that introducing retail payment innovations, such as stored-value
cards and contactless credit cards, leads to a reduction in cash usage for transactions in terms of both value and volume. (Additional work
is underway at the Bank of Canada on the impact of recent innovations such as contactless cards and person-to-person payment methods
on cash use.)
27 For more information on contactless payments in Sweden, see the Contactless Forum website.
28
Canadian Financial Monitor conducted by Ipsos Reid.
9
percent plan to abandon cash in the next five years (Bitcoin survey 2018). Similarly,
according to the Riksbank, the percentage of Swedes using cash in the past month declined
from 94 percent in 2010 to 61 percent in 2018. So, while cash use has been declining in both
Canada and Sweden, it seems likely that some cash use for payments will continue for (at
least) the medium term.
A payments market is a two-sided market, as consumers and merchants influence each other
in the use and acceptance of various payment methods. Fung, Huynh and Kosse (2017) argue
that it is consumers, however, who ultimately determine the prevalence of a payment method.
Therefore, as long as a sufficient number of consumers use cash, merchants are likely to
continue to accept it for transactions. However, if cash use continues to decrease such that
eventually only a small number of consumers use cash, merchants might reconsider their
acceptance of cash given the costs associated with it. Merchant acceptance of cash is
discussed further below.
29 Interac is the Canadian company that operates a debit payment network linking Canadian financial service providers.
30 And at least one of the major banks now offers Interac e-Transfers via voice commands and Touch ID on its banking app.
31 For more on this, see Interac press release, “Use of Interac e-Transfer Service Surges in 2018.” February 19, 2019.
32
See Henry, Huynh and Welte (2018), Table 12.
10
designed initially as mainly an alternative to cash or cheque for P2P payments, in 2017, about
one in six e-Transfer transactions are conducted by a business, and this segment is expected
to grow in the coming years.
Swish is a P2P fast-payment system in Sweden that started operating in 2012. It was created
by a consortium of Swedish financial institutions and is designed for use with smartphones.
Every subscriber to Swish links a mobile phone number to a bank account.33 A payer in
Swish enters the payee’s mobile number and authorizes the payment with the Swish app.
Both the payer and payee receive notification through the Swish app in the course of the
payment process, and only one or two seconds elapse between payment initiation and the
availability of funds to the payee. Financial institutions typically impose a maximum
transaction limit of around SEK 20,000.
According to a 2018 Riksbank payments survey, 62 percent of the Swedish population used
Swish in the month preceding the survey. (This compares with 61 percent of survey
respondents who reported using cash in the past month.) But only a small number of
respondents reported using Swish to pay for their last purchase; this probably reflects the fact
that the expansion of Swish to person-to-business payments and point-of-sale transactions is
still at an early stage. (The point-of-sale service builds on QR-codes and has different levels
of integration with the cashier system to target a broad range of businesses.)
The right-hand panel of Figure 7 shows that use of Swish grew slowly in the first three years
following its introduction in 2012. Starting in 2015, however, the number and value of Swish
transactions increased sharply each year. According to Swish, by December 2018, nearly 6.8
million private individuals were using this system, which is about two-thirds of the Swedish
population.34 Furthermore, during December 2018, individuals made 29 million P2P
transactions worth SEK 17 billion, with an average transaction value of about SEK 572.
Almost 3,500 stores accepted Swish for payments in December 2018, where consumers made
6.3 million Swish payments worth SEK 1.4 billion, with an average transaction value of SEK
227. In terms of total transacted value, Swish overtook cash in 2017.
33 Businesses and charities generally obtain a proxy number to substitute for a mobile phone number.
34 Statistics on Swish were retrieved from the Swish website.
35
Preliminary results from a recent Bank of Canada merchant acceptance survey of cash by Huynh, Nicholls,
and Nicholson (forthcoming) suggest that, among small and medium size businesses (less than 50 employees
and $10 million in revenue), about 4 percent of merchants currently do not accept cash. Only about 8 percent
of merchants plan to stop accepting cash in the next five years, while 79 percent of merchants do not plan to
stop accepting cash and 8 percent are uncertain.
11
by 2025.36 In this regard, the cost of handling cash, given decreasing cash turnover, is an
important driver of potentially lower merchant acceptance of cash. Concern for the working
conditions of employees, including security and allergic reactions to metal coins, is also an
influence according to this survey. But retailers also expressed concern about the potential
impact on customer satisfaction and sales of not accepting cash. (Swedish retailers indicated
that a recommendation to discontinue cash acceptance from a relevant organization, such as
the retailers’ association, or from a government authority, would carry great weight in their
eventual decisions.)
As noted, Swedish merchants report a greater interest than Canadian merchants in moving
toward not accepting cash as a means of payment in the future. This difference might be
partly explained by the different card payment schemes that consumers have been adopting in
these two countries. As discussed above, in Sweden, consumers have been moving away
from cash and toward debit cards for P2B payments, while in Canada, consumers have been
moving toward mostly credit cards for P2B payments (Erlandsson and Guibourg, 2018;
Henry, Huynh and Welte 2018). Given that debit cards tend to be less expensive for
merchants than credit cards are (due to, e.g., lower interchange fees on debit cards), it is
perhaps not surprising that merchants in Sweden report a greater willingness to move away
from cash acceptance than merchants in Canada.37
The development of payment card readers that can be attached to mobile devices, such as
Square in Canada and iZettle in Sweden, are also likely to lead to increasing merchant card
acceptance. In particular, these devices allow merchants without regular store fronts and
associated payment infrastructure, (e.g., pop-up stores, kiosks at festivals, craft and farmers’
markets) to accept payment cards through mobile phones. (There is no comprehensive
payments data currently available on how these devices may have affected cash use.)
3.5 Summing up
The value of small-denomination bank notes in circulation has been declining consistently
relative to GDP for many years in both Canada and Sweden. This decline corresponds with
36
The fact that many merchants indicated they were planning to stop accepting cash in coming years seems consistent with recent
anecdotal evidence that it is becoming more difficult to find stores in Sweden that accept cash.
37For the costs of debit cards and credit cards to merchants in Canada and Sweden, see Kosse et al. (2017) and Segendorff and Jansson
(2012), respectively.
38
For more, see Statistics Canada Digital Economy Survey.
12
the finding that the use of cash for point-of-sale payments has been steadily decreasing in
both countries, in absolute terms and as a share of payments, as Canadians and Swedes have
increasingly relied on debit cards and credit cards. In Canada, the use of contactless payment
cards has also grown quickly in the past few years and is likely to reduce cash use further,
particularly for small-value point-of-sale transactions. In Sweden, contactless cards are still in
the introductory stage, but they are likely to gain popularity rapidly, as has been the case in
Canada. Other recent payment innovations are similar in the two countries, notably Interac e-
Transfer in Canada (since 2001) and Swish in Sweden (since 2012), which have also
contributed to reduced cash use in both countries, especially for P2P payments.
The use of cash versus other payment methods depends on both the consumer and merchant
sides of the market. In practice, merchants tend to offer a range of payment choices that
reflect their perceptions of consumer preferences, in order to maximize sales, conditioned by
the costs of accepting different methods of payment. As a result, consumers drive the
observed utilization of various payment methods over time, with merchant acceptance of
different methods largely reactive in a competitive market. At the same time, as consumer
preferences evolve (e.g., use less cash), merchants are more likely to move away from
accepting cash, further discouraging consumers’ use of cash. (For more on such interplay in
the two-sided payment market, see Fung, Huynh and Kosse 2017, and Arvidsson, Hedman
and Segendorf 2018.)
In sum, the long-term downward trend of small-denomination bank notes relative to GDP in
Canada and in Sweden reflects declining transactional demand for cash in both countries,
driven by the adoption of a range of similar retail payment innovations. And this trend is
expected to continue. At the same time, merchant acceptance of cash has been (to this point)
nearly universal in both Canada and Sweden. Since the end of the Second World War,
transactional demand for cash in Sweden has declined more than that in Canada and is now at
a very low level. However, our analysis suggests that payment innovations and their diffusion
as well as merchant acceptance of cash have been similar in both countries. Accordingly,
considerations related to such factors cannot explain the differences in aggregate cash-to-
GDP trends in these two countries, particularly over the last decade or so. Next, we focus on
non-transactional demand for cash to better understand this cross-country difference.
4. Larger-denomination notes and non-transactional demand for cash
The preceding section is concerned with the demand for cash used for purchases—that is,
transactional cash demand. But bank notes can also provide a way to store value: cash is
essentially free of credit risk, maintains its (nominal) value over time, can be exchanged at a
later date for other assets, or used to purchase goods and services, without penalty. And large-
denomination bank notes are a better (more efficient) store of value than small
denominations. Indeed, holdings of large-denomination notes are usually considered to be
motivated mainly by store-of-value considerations, rather than by payment needs, since the
larger value of the notes makes them generally unsuitable for day-to-day transactions and
comparatively more useful as a store of value. This perspective has probably become
increasingly relevant over time as various technological payment innovations displace cash
used for transactions.
This section argues that understanding the influences driving non-transactional—or store-of-
value—cash demand is important for explaining the different experiences of Canada and
13
Sweden when it comes to the evolution of aggregate cash-to-GDP over recent decades. These
influences are discussed in the rest of this section.39
39
The arguments in this section should be seen as suggestive and not conclusive. More definitive conclusions require more rigorous
empirical examination, also drawing on other countries’ comparative experiences, which is left for future research.
40For a discussion of outside and inside money, and related motivations, see, e.g., Engert, Fung and Hendry (2018), especially Appendix 1
of that paper.
41The marginal tax wedge on labour income includes marginal income taxes, marginal social security contributions and marginal payroll
taxes. According to Du Reitz, Johansson and Stenkula (2015, 36), “tax wedges peaked around 1980 when the top marginal tax wedge and
the marginal tax wedge for the high-income earner could reach 90 percent…The major tax reform in 1990–1991 decreased the marginal
tax wedges to levels that prevailed before [the 1970s].”
14
financial crisis in the early 1990s.42 (For more on these developments, see Carlgren 2015;
Hogan 2010; Honkaphoja 2012; Schon 2008; and Wikfalk 1998).
In response to the banking crisis in the early 1990s, Swedish authorities initially responded in
a piecemeal, ad hoc fashion. As the threat grew, however, policy-makers responded more
aggressively to prevent a financial collapse. This ultimately included various supports for
Swedish banks and the introduction in September 1992 of a blanket guarantee protecting
depositors and other bank counterparties from loss. This was backed by open-ended funding
for the Bank Support Authority created by the Swedish Parliament (Honkaphoja 2009;
Jonung 2010). These formal banking system guarantees lasted until 1996. Sweden’s response
was ultimately widely regarded as a successful model of financial crisis management (see,
e.g., Anderson 2009; Ergungor 2007; and Ingves and Lund 2008). Corresponding to these
events, larger notes-to-GDP held as a store of value stabilized but continued at an elevated
level through this period (Figure 5b).43
During the 2008 global financial crisis, in contrast to the experiences of other countries, the
demand for larger notes in Sweden increased only marginally and briefly. In fact, the value of
1,000-krona notes in circulation continued to decline throughout 2008 and 2009. Notably,
following the lessons of the early-1990s banking crisis, Swedish authorities intervened
promptly and aggressively to protect the financial system and inside-money balances. In
October 2008, Swedish authorities introduced a comprehensive set of measures that provided
state support to financial institutions and their creditors. These included liquidity assistance at
short and longer maturities in SEK and in foreign currencies, guarantees of medium-term
market debt issuances (which reached 10 percent of GDP), a bank-recapitalization scheme,
increased deposit guarantee coverage, and a general stability fund to support other
government interventions in the financial sector (Becker, Bryant and Henderson 2012; IMF
2011; and Jochen 2010). A central part of that program was legislation that gave the Swedish
government unlimited fiscal powers to finance measures needed to ensure financial system
stability, through guarantees, capital injections or other means. Distinguishing features of the
Swedish crisis response were its promptness and thoroughness, as well as public transparency
regarding both objectives and actions.
Senior Swedish officials have been clear that “a crucial lesson from the Nordic experience is
the need for prominent state involvement in crisis resolution” (Ingves and Lund 2008, 21).
Transparency is also seen as critical, including informing the public about official plans and
actions. Finally, “there is also a role for a blanket guarantee to restore confidence and prevent
bank runs and a potential financial meltdown” (Ingves and Lund 2008, 23).
It follows that Swedish authorities have demonstrated twice in the last 25 years
comprehensive and transparent protection of the Swedish banking system and deposits under
severe stress. And since the 2008 financial crisis, a new open-bank resolution framework has
been established in Sweden for systemically important financial institutions (Riksgalden
2018b). Under this (bail-in) arrangement, Swedish authorities will take control of a
systemically important institution that is deemed to be not viable. Furthermore, to avoid the
broader adverse externalities from the closure of a systemically important bank, the
42Along with Sweden, two other Nordic countries, Finland and Norway, also experienced a systemic financial crisis in the early 1990s
(Honkaphoja 2009, 2012). All four of the Nordic countries also provided public support to their banking systems in this environment. In
Finland, Norway and Sweden, public support was significant, while in Denmark this support was small (Honkaphoja 2009).
43The 500-krona note was introduced in 1985, and initially this note appears to have substituted for 1,000-krona notes as a store of value
and the 100-krona notes for transactions. (Given persistent inflation in Sweden from the 1970s into the 1990s, the 500-krona note became
more suitable for transactions during that period.)
15
institution will be kept open and functioning as usual so that depositors will have
uninterrupted access to their accounts and other financial services.
These experiences and associated policy statements may have provided a disincentive for
Swedes to hold cash as a store of value, and, in particular, larger bank notes (500-krona and
1,000-krona notes), as a hedge against crisis. In sum, there may be an expectation that, in
periods of severe crisis, most deposits (inside money) will be secured by the actions of public
authorities, thereby reducing the need for cash (outside money) as a hedge against crisis.44
Canada has had comparatively little experience with banking crises, although Canadian
authorities deployed several measures supporting bank liquidity in response to the 2008
global financial crisis (Zorn, Wilkins and Engert 2009). As well, Canadian authorities have
established a broad range of powers and policies to manage such events, also including an
open-bank (bail-in) resolution framework for the major Canadian banks (Engert, Fung and
Hendry 2018, Appendix 2). But there have not been repeated demonstrations of such
intervention powers as in the Swedish experience.
44A similar proposition arguably might apply for some other advanced economies, but perhaps has been less clearly demonstrated in
practice than has been the case in Sweden.
45
And there is some evidence that bank branches in Canada are becoming (even) more central to major Canadian banks’ retail business
strategies. See, e.g., T. Kiladze, “TD Bank’s Radical New Strategy,” The Globe and Mail, December 2, 2018.
46 See Swedish Bankers Association Bank and Finance Statistics and Statens Offentliga Utredningar (2018).
47According to Bankomat, note denominations up to 500 krona are usually available in Swedish ATMs, subject to withdrawal limits set by
each bank.
16
over SEK 3 billion in 2018. Further, this friction has been exacerbated by increasingly
stringent anti-money laundering (AML) provisions.
As noted, the Swedish Bankomat ATM network (discussed above) provides for access to
most smaller-denomination notes, so cashless branches are not a significant constraint on
access to transactions cash. Instead, they appear to generate a friction inhibiting access to
large-denomination, store-of-value notes in Sweden, particularly since 2010. In addition, as
noted above, ATMs in Sweden are mainly for cash withdrawals, and only specific deposit
machines accept cash deposits. Thus, together with the increasing number of cashless
branches, it has becoming increasingly difficult to deposit cash, which could discourage the
acceptance of cash in general.
Cashless branches could also inhibit merchants’ willingness to accept cash at point-of-sale,
given the increased cost to merchants of cash management arising from the proliferation of
cashless branches. (This might especially affect small and medium-sized merchants.) For
example, merchants that accept cash need to keep a sufficient amount of small denominations
(e.g., 20-krona and 50-krona notes) on hand for change. Since Swedish ATMs do not
dispense such small denominations, merchants must find a branch that provides cash services
or they risk not having sufficient change. And to deposit cash received during the day, in the
absence of a suitable bank branch that accepts cash or has a drop-off box, merchants
increasingly depend on the services of cash-in-transit companies to deliver deposits and to
obtain change—which increases merchants’ cost of accepting cash. Indeed, Swedish
authorities have identified the limited access to deposit services as a problem for businesses
(Statens offentliga utredningar 2018).
As noted above, there are significantly more bank branches per capita in Canada than in
Sweden, and almost all Canadian bank branches provide cash services on demand. In the last
few years, some of the major Canadian banks have introduced branches that focus
exclusively on the provision of financial advice and sales to consumers. This includes
providing mortgages and other lending, investment and wealth management, as well as
business services for small and medium-sized firms. These “financial-advice centres” do not
provide any cash services over the counter (teller) but could provide access to cash through
on-site ATMs. To date, such bank branches without over-the-counter cash services are
relatively unusual in Canada, but they could be under consideration more generally by the
major banks.
For example, it appears that 18 percent of the branches of one of the major Canadian banks
are now cashless or tellerless (i.e., no cash services are provided at the counter), based on
information collected online. ATMs are present at these branches and can dispense a range of
bank notes, including $100 notes, and accept deposits including stacked cash and cheques. As
well, daily ATM withdrawal limits can be increased when needed, including to satisfy
individual customer requests arranged on the spot with branch staff or in advance via
telephone banking. More generally, ATMs in Canada increasingly provide consumers with a
range of choice about the denomination of notes dispensed in ATMs, up to and including $50
notes. Some banks are also considering stocking their ATMs with $100 notes. As a result, , in
contrast to the Swedish case, there do not seem to have been meaningful frictions inhibiting
access to large-denomination notes in Canada due to cashless bank branches. However, some
variant of cashless branches could proliferate in Canada in the future.
The evolution of cashless bank branches seems to reflect interactions characteristic of a two-
sided market, similar to that discussed above concerning payment methods. More
specifically, customer preferences regarding access to cash (e.g., over-the-counter versus
ATMs, choice over denominations, etc.) are the fundamental drivers in a competitive banking
17
market, with banks responding strategically to these evolving customer preferences. At the
same time, banks’ decisions about access to cash, which are also influenced by cost
considerations, can consolidate or reinforce the underlying trends in access to cash driven by
customer preferences. (If, however, the banking market is not competitive or banks collude in
this respect, then consumer preferences would be less dominant.)
48The 20-, 50- and 1,000-krona notes became invalid after June 30, 2016 but were accepted for exchange by banks for two additional
months. Similarly, 100- and 500-krona notes became invalid after June 30, 2017 but were accepted for exchange by banks for an
additional year.
49Additional information and requirements concerning redemption of invalid notes are available at the Sveriges Riksbank website
(“Redeeming invalid banknotes”). These requirements reflect the increasingly stringent AML /CTF regulations in Sweden (e.g., people who
deposit more than SEK 10,000 are required to account for the sources of the cash).
50From 2015 to 2018, the Riksbank received 220,662 note-exchange applications, worth over SEK 1 billion. The number of such
applications has been increasing recently. In 2017, there were 63,723 such applications, and in 2018 there were 67,411. (These
applications can also cover very old bank notes, e.g., notes issued in the early 1900s that were declared invalid in the 1980s.)
18
In sum, the frequency with which bank notes have been deemed invalid in Sweden over the
last 20 years, and the burdensome exchange provisions in some cases, especially coupled
with the decreasing availability of cash services at bank branches, has probably undermined
the demand for cash, particularly larger notes, as a store of value. A notable example is the
decline in 1,000-krona notes in circulation over the last two decades. After peaking in
December 2001 at SEK 48.4 billion, the value of 1,000-krona notes declined gradually to
SEK 21.4 billion in December 2012. Beginning in 2013, this decline accelerated, and the
outstanding value decreased to SEK 9.7 billion by December 2013. Subsequently, this
decline levelled off (see Box 1 for a detailed discussion of the experience regarding legal
tender provisions affecting the 1,000-krona note).
In contrast to these Swedish legal tender frictions, all bank notes issued by the Bank of
Canada remain legal tender. Recently, legislation has been passed to allow the Bank of
Canada to change the legal tender status of a denomination of a series of bank notes. The
current plan is to remove the legal tender status of certain denominations of notes that have
long been long discontinued and are no longer issued: $1, $2, $25, $500 and $1,000 notes.51
There are no plans to remove the legal tender status of any other bank notes. Therefore, there
are no legal tender frictions suppressing demand for larger bank notes in Canada.
51
See upcoming changes to legal tender status for older bank notes at the Bank of Canada.
19
Box 1: Additional information on legal tender provisions affecting the 1,000-krona note
The legal tender frictions associated with the issuance of new 1,000-krona notes in 2006 and again in
2015 appear to have contributed to the rapid decline of 1,000-krona notes, particularly in 2013
(Chart 1-A). With the introduction of the new 1,000-krona notes in 2006 (an improved note with a
security foil strip), the 1,000-krona notes issued from 1989 to 1991 (without a foil strip) were
scheduled to become invalid on December 31, 2013. (Similarly, the 50-krona note without foil strip
was also deemed invalid after December 2013 and replaced.) Thus, holders of the old 1,000-krona
notes had to exchange them for new notes at a bank branch before the end of 2013. In 2012,
however, the Riksbank announced its intention to issue another new series of bank notes in 2015
and 2016, including another new 1,000-krona note, which would make all pre-existing series of bank
notes invalid sometime after 2016. This could have had an impact on note holders’ decisions to
exchange their old 1,000-krona notes (no foil strip) for those issued after 2006 (with foil strip) before
the end of 2013, which would become invalid sometime after 2016—or to simply exchange their old
1,000-krona notes (no foil strip) for bank deposits before the end of 2013.
Further, as discussed above, cashless bank branches started emerging in Sweden in 2010, and many
bank branches (more than 40 percent of them) were cashless by 2012. The number of cashless
branches has continued to grow, making it increasingly difficult to exchange old bank notes for new
notes. As a result, the frequent replacement of 1,000-krona notes, combined with an increasing
number of cashless branches, exacerbated the inconvenience and cost of exchanging outstanding
1,000-krona notes, likely contributing to the sharp decline in the circulation of these notes in 2013.
Chart 1-A: Value of 1000-krona notes in circulation (January 1980 – April 2019)
Source: Riksbank
20
4.4 Other considerations
(i) Domestic demand for large-value foreign notes
Canada and Sweden both have strong ties with much larger, influential neighbours: i.e., the
United States and euro area economies, respectively. Have residents of the former increased
their holdings of large-value foreign notes as a store of value? If so, this could reduce demand
for larger domestic notes in favour of foreign notes. There appears to be little evidence that
Canadians have been swapping large-value Canadian notes for US notes as a store of value.
Have Swedes dropped larger-value SEK notes in favour of euro note holdings as a store of
value? If interpersonal trust or trust in Swedish institutions was very low in Sweden
compared with major countries of the euro area, for example, Swedes might switch their
holdings from large-denomination SEK notes to large euro notes or to deposits of major non-
Swedish banks. The evidence, however, indicates that interpersonal and institutional trust in
Sweden is consistently (and substantially) among the highest in Europe (Ortiz-Ospina and
Roser 2018). And, as discussed above, Swedish authorities have gone to considerable lengths
to provide for the robustness of their banking system under stress. More generally, there
seems to be little evidence that Swedes have been swapping their large-value SEK notes for
large euro notes as a store of value.
52
See Annual Report to Parliament on Immigration by Immigration, Refugees, and Citizenship Canada, various years.
53For details, see Statistics Canada, “International Students in Canadian Universities” and “Canadian Postsecondary Enrolments and
Graduates, 2016/2017.”
54See Statistics Canada, “The Evolution of Canadian Tourism, 1946 to 2015” and “Travel Between Canada and Other Countries, December
2017.”
21
on the purchase of foreign currencies by their nationals. For example, Chinese citizens are
allowed to purchase foreign currencies in amounts equivalent to up to US$50,000 per person,
per year. (While this annual allowance would seem adequate for most peoples’ overseas
travel needs, it might not be sufficient for major investments overseas, such as purchasing a
property.) Also, Chinese visitors to Canada are likely to hold some Canadian bank notes
when entering Canada for convenience and wide acceptance, particularly since Chinese
citizens have had limited access to major global card companies’ products. (However, “China
UnionPay” cards, the main debit and credit cards accessible by Chinese nationals, are
increasingly accepted in Canada.) Also, international students and immigrants from China
might acquire Canadian bank notes before arriving in Canada so that they can make
purchases until they have set up a domestic bank account.
Sweden also has experienced substantial immigration in recent decades, as indicated by the
2.4 million residence permits granted between 1980 and 2018, of which nearly half were
granted after 2008.55 Roughly one-quarter of the latter were granted for asylum reasons, one-
quarter for labour market reasons (i.e., pursuing work opportunities) and one-third were
family unifications.56 Ten percent of the residence permits were issued to students, foremost
from China and India. (European Union citizens moving to Sweden are not included in these
data.) As well, there has been labour immigration from some Eastern European countries
following their accession to the European Union in 2004.57 In sum, most immigration to
Sweden in recent decades has been from cash-intensive countries, and this might have
increased the demand for cash, at least for a period before the immigrants adapt to local
Swedish practices. There is, however, no concrete evidence supporting this perspective.
Indeed, the rapid decline of cash demand in Sweden in recent years seems to suggest that this
influence has been negligible.
Likewise, there are no data on the holdings of Swedish bank notes abroad, for example, in
immigrant countries of origin. Some insight might be gained from the invalid bank note
exchange applications sent to the Riksbank in 2015 and subsequent years: Around 15 percent
of all such applications are made by foreign citizens, often from Eastern Europe. If these
numbers are representative of the distribution of SEK notes generally, then there could be
almost SEK 1 billion held abroad.58 Even if foreign SEK holdings were somewhat higher
than this estimate, they would still be a relatively low amount compared with the total
demand for cash and would not likely have had a significant effect on the demand for high-
or low-denomination notes.
In sum, an increase in foreign demand for Canadian bank notes might have contributed to the
strong demand for large-denomination notes in recent years. While there are some possible
explanations for this increase in foreign demand, as noted above, this is not well understood.
A better understanding of the foreign demand for cash requires better data (including, e.g.,
data on the shipment of Canadian bank notes overseas), along with more rigorous analysis of
foreign sources of demand.59
55 Statistics on granted permits can be found at Swedish Migration Agency. (More detailed statistics are available in Swedish.)
56The most common countries of origin for refugees were Syria, Afghanistan, Eritrea, Somalia and Iran. For family unification, main source
countries were Syria, Iraq, Afghanistan, Eritrea and Somalia. And major source countries for labour market reasons were India, Thailand,
China, Albania and Turkey.
57
Most of this labour came from the Baltic states and Poland.
58 At the end of 2018, SEK 5.5 billion in invalid notes were in circulation. Fifteen percent of this is close to one billion.
59For example, it would be useful to understand whether people from a cash-oriented country are more likely to accumulate and carry
foreign bank notes when travelling overseas or when immigrating to another country. Kosse and Jansen (2013) find that first-generation
migrants from a number of countries that can be seen as cash-oriented are more likely to use cash in the Netherlands.
22
4.5 Summing up
The divergent trends of cash demand in Canada and Sweden since 2000 have been driven
largely by developments affecting larger notes—that is, $50 and $100 notes in Canada, and
500-krona and 1,000-krona notes in Sweden. Several influences together appear to have led
to reduced demand for larger, non-transactional notes in Sweden.
Sweden’s crisis-management experience may have created incentives that discourage reliance
on large notes as a hedge against uncertainty. In addition, the proliferation of cashless bank
branches and the operation of Swedish legal tender rules appear to have created frictions that
inhibit holdings, especially of larger Swedish notes. At the same time, foreign sources could
have generated increased demand for large-denomination Canadian bank notes in recent
years (which could diminish over time). Taken together, these various influences and
frictions help to explain relatively low demand for store-of-value cash in Sweden compared
with Canada. (Recently, there has been a modest uptick in holdings of 500-krona notes; see
Box 2 for more on this.) More definitive conclusions require further empirical examination,
drawing also on other countries’ experiences, which is left for future research.
23
Box 2: Recent developments affecting the 500-krona note
Since late 2017, there has been an increase in the (absolute) value of bank notes in circulation in
Sweden, driven by an increase in 500-krona notes (see Chart 2-A below). (This has had little
discernable impact on the value of 500-krona notes relative to GDP.) The reasons for this modest
reversal are not entirely clear, but it likely reflects a buildup of cash buffers of households. One
possible reason for this is a recent Swedish government initiative to increase reliance on cash as a
hedge against crisis. To this end, an agency of the Swedish government (the Swedish Civil
Contingencies Agency) provided every Swedish household a booklet entitled If War of Crisis Comes to
help Swedish citizens be better prepared for a range of calamities, including serious accidents,
extreme weather, information technology dysfunction and military conflict. The booklet also points
specifically to disruptions in payment card and ATM networks. (See the Swedish Civil Contingencies
Agency website for more information.) Accordingly, the agency recommends, among other things,
that citizens hold a buffer of cash in small denominations at home, and it appears that Swedes have
decided that 500-krona notes (which are available from ATMs) can address this need.
The crisis preparedness booklet was sent to households in May 2018, but news about its circulation
and its key recommendation were publicly available since at least January 2018. (See, for example,
“What If War Comes,” in The Irish Times, January 18, 2018.) The authorities also previously posted
the key recommendation on its webpage. A version of this booklet was issued regularly in Sweden
during the Cold War, from 1943 to 1991. And Swedish authorities have been gradually increasing
their focus on security policy over the last 10 years or so.
The recent build-up of this cash buffer could also reflect an element of involuntary cash hoarding by
households and businesses, given the difficulty of depositing bank notes in Sweden (as discussed in
the text). It is also possible that households did not fully restore their cash buffers to their desired
level during the last bank-note exchange and have been gradually doing so recently in 500-krona
notes.
Chart 2–A: Value of Swedish bank notes in circulation (January 1980 – April 2019)
24
A connection is often made between the underground (or shadow) economy and cash. This
section considers that connection and whether it plays a significant role in explaining the
differences in the evolution of cash-to-GDP in Canada and Sweden.
5.1 Background
The focus here is the underground economy, in which the underlying activities and
transactions are legal but are concealed from public authorities for a variety of reasons.
Following Schneider (2016), we consider the underground economy as including market-
based, legal production of goods and services that are concealed from public authorities
largely to avoid one or more of the following:
payment of taxes, such as income taxes or value-added taxes
payment of social security contributions
compliance with administrative procedures, such as completing statistical
questionnaires
certain labour market standards, such as minimum wage or safety requirements
Given the intent to conceal the benefits of the underlying activity, it seems reasonable to
conjecture that cash, which provides for anonymity, is a preferred means of payment in the
underground economy. As a result, one could expect that changes in the size of the
underground economy might help explain cash demand over time.
While the preceding provides a straightforward conceptual definition, reliably measuring the
size of the underground economy is, of course, challenging. Further, there are typically large
discrepancies in estimates depending on the empirical methodology used, and there does not
appear to be a preferred or dominant approach to measuring the underground economy
(Dunbar and Fu 2015; Medina and Schneider 2018).
In Canada, for example, Dunbar and Fu (2015) note that studies using discrepancies in
national accounts to measure the amount of unreported income estimate the underground
economy at about 3 percent of GDP, while studies using microeconomic consumption data
result in estimates of around 5 percent. And studies based on macroeconomic data find that
the size of the underground economy in Canada could be as much as 15 percent of GDP.
Similarly, Dunbar and Fu (2015) use household-level income data and consumption data to
estimate under-reported income, and they conclude that an upper bound of 14 to 19 percent of
GDP is unreported in Canada.
With respect to Sweden, Guiborg and Segendorf (2007), using an unexplained-demand-for-
cash approach, measure the size of the underground economy in Sweden to be 6.5 percent of
GDP in 2006. In comparison, Medina and Schneider (2018), using macroeconomic and other
data, estimate it at 11.1 percent in 2006.
Similarly large discrepancies across quantitative methods are evident for many other
countries (e.g., Medina and Schneider 2018; Seitz, Reimers and Schneider 2018). In their
review, Medina and Schneider (2018, 28) conclude, “There is no superior method [to
measure the underground economy]. All methodologies, without exception, have their own
advantages as well as weaknesses. If possible, one should use multiple methods [to draw
25
conclusions] … Much more research is needed with respect to the estimation methodology
and the results for different countries and periods.”60
These authors also observe, however, that one stable conclusion from the evidence is that the
size of the underground economy (relative to GDP) appears to have been generally declining
across a wide range of economies (advanced and emerging) from 1991 to 2015 (interrupted
only in 2008, likely due to the global economic crisis).61 Table 2 illustrates this aspect for
Canada and Sweden.
Further, while it is reasonable to expect that the underground economy is relevant for
explaining cash demand, in practice it is difficult to establish a meaningful empirical
relationship between cash and measures of the underground economy. This appears to be the
case across a number of countries and a variety of studies (Seitz, Reimers and Schneider
2018).62
5.2 Swedish policy measures to reduce the size of the underground economy
Given concerns in Sweden about undeclared work and revenues, the Swedish Tax Agency
has established several measures to reduce tax evasion in sectors with high cash turnover. A
focus has been domestic services, such as home repairs, cleaning and home maintenance,
where it is believed that such work traditionally has been undeclared to a significant extent
(Swedish Tax Agency 2012). Accordingly, to reduce the prominence of the underground
economy in Sweden and bring activity into the taxable, formal economy, tax deductions for
50 percent of the labour cost of certain domestic work were introduced in 2007 (for
household services, so-called “RUT” provisions) and in 2008 (for home renovations, the
“ROT” provisions). These measures provide tax incentives to pay for such services in the
declared, formal economy, and therefore to discourage transactions in the underground
economy.
These measures appear to have been effective in bringing some activity into the formal
economy and reducing tax evasion (Swedish Tax Agency 2012). For example, surveys have
indicated that, between 2006 and 2012, the proportion of citizens who know people who have
evaded tax or participated in the underground economy during the preceding 12 months
decreased significantly. More specifically, the proportion of respondents who know people
who evaded tax decreased from 22 to 13 percent, and the proportion of respondents who
know people working in the underground economy decreased from 38 to 22 percent. Swedish
authorities consider these changes likely to be the result of the tax deductions described
above for house repair and maintenance and for domestic work (Swedish Tax Agency 2012,
16)
The commercial sector has also been a focus for Swedish authorities, where there have been
concerns about the manipulation of cash registers and associated under-reporting of revenues,
particularly in businesses with high cash turnover (Eurofound 2013). Accordingly, starting in
January 2010, businesses in Sweden selling goods and services for cash payments must use a
certified cash register that includes a control unit (“black box”), which records transactions
made by the cash register. Only the Swedish Tax Agency can access the sales and payment
26
records in the black box.63 Analyses by the Agency (2013) indicate that this measure also
appears to have reduced the extent of commercial tax evasion in Sweden.
5.3 Summing up
Given the outcomes reported by the Swedish Tax Agency (noted above), it follows that there
would have been corresponding decreases in the size of the underground economy (other
things equal) and cash demand in Sweden. Such outcomes would have contributed to the
overall decline of Swedish cash-to-GDP during the last decade (discussed in earlier sections).
On the other hand, given the considerable uncertainty in the measurement of the underground
economy and its (empirical) relationship to cash demand, it is difficult to draw (strong)
conclusions about the significance of these tax-policy measures in reducing cash demand in
Sweden.
Moreover, as noted, numerous countries in the last 10 to 20 years have experienced a general
trend of declining underground economies, including Canada. If this is correct, other (perhaps
common) underlying factors might also be relevant. From this perspective, it seems unlikely
that a declining underground economy in Sweden driven by tax changes explains the
different experiences of cash-to-GDP in Canada and in Sweden.
6. Conclusions
Cash use for payments has been decreasing in many countries, including Canada and
Sweden. However, notes in circulation relative to GDP in most countries, including Canada,
have been stable for decades and even rising in recent years. In contrast, overall cash-to-GDP
in Sweden has been falling steadily. To understand these outcomes, we focus separately on
transaction (small-denomination) bank notes and store-of-value (large-denomination) bank
notes.
Both Canada and Sweden have seen a long-term downward trend in small-denomination bank
notes relative to GDP, reflecting declining transactional demand for cash in both countries.
These outcomes have been driven by the adoption of a range of similar retail payment
innovations. At the same time, merchant acceptance of cash has been (to this point) nearly
universal in both Canada and Sweden. Therefore, neither payment innovations and their
diffusion nor differences in merchant acceptance of cash are adequate to explain why
aggregate cash demand has been declining rapidly in Sweden but not in Canada.
Instead, divergent trends in the demand for larger bank notes, typically used more as a store
of value, seem to be the key to understanding the different overall cash-to-GDP trends in
Canada and Sweden. More specifically, Sweden’s crisis-management experience may have
created incentives that discourage reliance on large notes as a hedge against uncertainty. In
addition, the recent proliferation of cashless bank branches and the operation of Swedish
legal tender rules appear to have created frictions that inhibit holdings of larger Swedish
notes. These three influences work in the same direction to reduce demand for larger bank
notes, and to some extent have probably reinforced each other to generate larger cumulative
effects on cash demand over time. In Canada, some evidence suggests that foreign sources
could have generated increased demand for large-denomination Canadian bank notes in
recent years (which could diminish over time).
The analysis in this paper points to three broad lessons:
63Businesses bear the costs of installing and maintaining these monitored cash registers, and non-compliance is subject to fines levied by
the Swedish Tax Authority.
27
(i) Policy interventions and bank resolution frameworks (e.g., bail-in) that credibly
protect depositors (“inside money”) in financial crises reduce incentives to hold larger
bank notes as a hedge against crises. Such interventions very likely are unavoidable in
extreme (tail-event) financial crises and indeed are justifiable, particularly if the
resolution framework is well-structured ex ante (as are some bail-in regimes). These
considerations also suggest a reduced need for a central bank digital currency as a safe
store of value in crises (as discussed in Engert, Fung and Hendry 2018).
(ii) Cashless bank branches create a friction inhibiting access to cash, depending on
the ability of ATM networks to satisfy consumer demand for bank notes across a
range of denominations and to provide adequate cash-deposit services, particularly for
merchants. Given the need for merchants to manage relatively large volumes (and
values) of bank notes, cashless bank branches can create disincentives for merchants
to accept cash at the point of sale, which would inhibit the use of, and demand for,
bank notes.
(iii) Legal tender rules, where old bank note series are declared invalid, create
frictions that inhibit the demand for cash. The impact of such measures is stronger the
more frequently these declarations occur and the more burdensome and variable the
provisions are governing the exchange of old (invalid) notes for new notes. In that
case, legal tender frictions can be expected to reduce the demand for bank notes,
particularly as a store of value.
Finally, more work is required to better understand the demand for larger Canadian bank
notes, including the role of foreign demand, which is left for future research.
28
References
Amromin G. and S. Chakravorti. 2009. “Whither Loose Change? The Diminishing Demand
Small-Denomination Currency.” Journal of Money, Credit and Banking 41 (2/3)
(March– April), 315–335.
Anderson, R. G. 2009. “Resolving a Banking Crisis, the Nordic Way.” Economic Syopses,
no. 10. Federal Reserve Bank of St. Louis.
Arango, C., D. Hogg and A. Lee (2012). “Why Is Cash (Still) So Entrenched? Insights from
the Bank of Canada’s 2009 Methods-of-Payment Survey.” Bank of Canada Discussion
Paper 2012-2.
Arvidsson, N., J. Hedman and B. Segendorf. 2018. När slutar svenska handlare att acceptera
kontanter? Handelsrådets rapport. [When do Swedish retailers stop taking cash? Report
for the Swedish Retail and Wholesale Council].
Bátiz-Lazo, B. 2018. Cash and Dash: How ATMs and Computers Changed Banking.
London: Oxford University Press.
Bech, M., U. Faruqui, F. Ougaard and C. Picillo. 2018. “Payments Are A-changin' but Cash
Still Rules.” BIS Quarterly Review (March): 67–80.
Becker, T., R. Bryant and D. W. Henderson. 2012. Maintaining Financial Stability in an
Open Economy: Sweden in the Global Crisis and Beyond. Washington DC: The
Brooking Institution.
Bilkes, G. 1997. “The New Bank Note Distribution System,” Bank of Canada Review
(Summer): 41–54.
Carlgren, F. 2015. “Structural Problems and Reform.” Ekonomifakta (October).
Chen, H., M.-H. Felt and K. P. Huynh. 2017. “Retail Payment Innovations and Cash Usage:
Accounting for Attrition by Using Refreshment Samples.” Journal of the Royal
Statistical Society: Series A (Statistics in Society) 180 (2): 503–530.
Daltung, S. and M. Ericson. 2004. “Cash-Supply Efficiency.” Sveriges Riksbank Economic
Review, no. 3.
Dunbar, G. and C. Fu. 2015. "Sheltered Income: Estimating Income Under-Reporting in
Canada, 1998 and 2004." Bank of Canada Staff Working Paper No. 2015-22.
Du Rietz, G., D. Johansson and M. Stenkula. 2015. "Swedish Labor Income Taxation (1862–
2013).” IFN Workgin Paper No. 977. Research Institute of Industrial Economics.
Engert, W., B. Fung and S. Hendry. 2018. “Is a Cashless Society Problematic.” Bank of
Canada Staff Discussion Paper No. 2018-12.
Ergungor, O. E. 2007. “On the Resolution of Financial Crises: The Swedish Experience.”
Federal Reserve Bank of Cleveland Policy Discussion Paper No. 21.
Erlandsson, F. and G. Guibourg. 2018. “Times Are Changing and So Are Payment Patterns.”
Economic Commentaries, no. 6. Sveriges Riksbank.
Eurofound. 2013. Cash Register Legislation, Sweden. June 2.
Finansinpektionen. 29 May 2018. Stability in the Financial System.
Fung, B., K. P. Huynh and A. Kosse. 2017. “Acceptance and Use of Payments at the Point of
Sale in Canada.” Bank of Canada Review (Autumn): 14–26.
29
Fung, B., K. P. Huynh and L. Sabetti. 2014. “The Impact of Retail Payment Innovations on
Cash Usage.” Journal of Financial Market Infrastructures 3(1).
Fung, B., K. P. Huynh and G. Stuber. 2015. “The Use of Cash in Canada.” Bank of Canada
Review (Spring): 45–56.
Guðmundsson, M. 2016. “Iceland’s recovery: Facts, myths, and the lessons learned.” Speech
at an OMFIF lecture, London, England, 28 January.
Guiborg, G. and B. Segendorf. 2007. “The Use of Cash and the Size of the Shadow Economy
in Sweden.” Sveriges Riksbank Working Paper Series No. 204.
Henry, C. S., K. Huynh and A. Welte. 2018. “2017 Methods-of-Payment Survey Report.”
Bank of Canada Staff Discussion Paper No. 2018-17.
Henry, C. S., K. Huynh, G. Nicholls and M. Nicholson. 2019. “2018 Bitcoin Omnibus
Survey: Awareness and Usage.” Bank of Canada Staff Discussion Paper (forthcoming).
Hogan, J. 2010. “Economic Crises and Policy Change in the Early 1980s: A Four Country
Comparison.” Journal of Australian Political Economy 65 (Winter): 106–37.
Honkapohja, S. 2009. “The 1990s Financial Crises in Nordic Countries.” Bank of Finland
Discussion Papers, 5, 2009.
Honkapohja, S. 2012. “The 1980s Financial Liberalization in the Nordic Countries.” Bank of
Finland Discussion Papers, 36, 2012.
Huynh, K., G. Nicholls, and M. Nicholson. 2019. “2018 Merchant Acceptance Survey.” Bank
of Canada Staff Analytical Note (forthcoming).
30
of Point-of-Sale Payments in Canada.” Bank of Canada Staff Discussion Paper No.
2017-4.
Kosse, A. and D. -J. Jansen. 2013. “Choosing How to Pay: The Influence of Foreign
Backgrounds.” Journal of Banking and Finance 37 (3): 989–998.
Kristinsson, Y. 2012. “Icelandic Banks: Rebuilding a Financial System.” Icelandic Financial
Services Association.
Länsstyrelserna. 2018. Bevakning av grundläggande betaltjänster 2018. Rapport 2018:14
[County Administrative Boards. 2018. Monitoring of basic payment services 2018.
Report 2018:14].
https://www.lansstyrelsen.se/download/18.4e0415ee166afb59324291aa/1543416230776
/Bevakning-betaltj%C3%A4nster%202018.pdf
Medina, L. and F. Schneider. 2018. “Shadow Economies Around the World: What Did We
Learn Over the Last 20 Years?” IMF Working Paper WP/18/17 (January).
Ortiz-Ospina, E. and M. Roser. 2018. “Trust.” Our World in Data.
https://ourworldindata.org/trust
OSFI Capital Adequacy Guideline, Annex 1, 2019.
Riksgalden (Swedish National Debt Office). 2018a. Annual Decisions Taken for Crisis
Management of Banks.
Riksgalden (Swedish National Debt Office). 2018b. Financial Crisis Management (July 6)
Schneider, F. 2016. “Estimating the Size of the Shadow Economies of Highly Developed
Countries: Selected New Results.” CESifo DICE Report 4/2016 (December).
Schneider, F. and A. Buehn. 2018. “Shadow Economy: Estimation Methods, Problems,
Results and Open Questions.” Open Economics 1 (1): 1–18.
Schön, L. 2008. “Sweden: Economic Growth and Structural Change, 1800–2000. ” EH.Net,
edited by R. Whaples. February 10. http://eh.net/encyclopedia/sweden-economic-
growth-and-structural-change-1800-2000/
Segendorf, B. and T. Jansson. 2012. “The Cost of Consumer Payments in Sweden.” Working
Paper Series No. 262. Sveriges Riksbank.
Segendorf, B. and A-L. Wretman. 2015. The Swedish Payment Market in Transformation.
Sveriges Riksbank Economic Review, no. 3.
Seitz, F., H-E. Reimers and F. Schneider. 2018. “Cash in Circulation and the Shadow
Economy: An Empirical Investigation for Euro Area Countries and Beyond.” CESifo
Working Paper 7143 (July).
Statens offentliga utredningar (2018). Tryggad tillgång till kontanter. SOU 2018:42. [The
Swedish Governments Public Inquiries. Ensured Access to Cash. Summary in English].
https://www.regeringen.se/49cf6d/contentassets/79026c9e608946bdbfa60067ddae0c0d/t
ryggad-tillgang-till-kontanter-sou-201842.pdf
Sveriges Riksbank. 2013. “The Swedish Retail Payments Market.”
Sveriges Riksbank. 2016. "Payment Patterns in Sweden 2016."
Sveriges Riksbank. 2017. e-Krona Project Report 1. https://www.riksbank.se/en-
gb/payments–cash/e-krona/e-krona-reports/e-krona-project-report-1/
Sveriges Riksbank. 2018a. e-Krona Project Report 2. Available at
31
https://www.riksbank.se/en-gb/payments–cash/e-krona/e-krona-reports/e-krona-project-
report-2/
Sveriges Riksbank. 2018b. “Payment Patterns in Sweden 2018.”
Swedish Bankers’ Association. 2018. Bank and Finance Statistics 2017, December.
https://www.swedishbankers.se/media/4015/bank-och-finansstatistik-2018-december-
eng.pdf
Swedish Tax Agency. 2012. Taxes in Sweden: An English Summary of Tax Statistical
Yearbook of Sweden.
https://www.skatteverket.se/download/18.3684199413c956649b57c66/1361447643691/
10413.pdf
Swedish Tax Agency. 2013. Requirement of Cash Registers Impact Evaluation. English
Summary of the Swedish Tax Agency Report 2013:2. https://www.skatteverket.se/
Wilkfalk, M. 1998. “Globalization and the Crisis of the Swedish Welfare State.” SAIS
Journal of Global Affairs (April).
Zorn, L., C. Wilkins and W. Engert. 2009. "Bank of Canada Liquidity Actions in Response to
the Financial Market Turmoil." Bank of Canada Review (Autumn): 3–22
32
Chart 1a: Number and value of cash transactions in Canada
Chart 2a: Value share of cash, debit card and credit card payments in Canada
33
Sources: TSI International and authors’ calculations
Chart 2b: Value share of cash, debit card and credit card payments in Sweden
34
Chart 3a: Bank notes as a ratio of GDP for selected countries (1945–2018)
Sources: Bank notes data from central banks and GDP data from national account statistics in
respective countries
Chart 3b: Bank notes as a ratio of GDP in Norway and Sweden (1945–2018)
35
Sources: Bank for International Settlements and authors’ calculations
Chart 5a: Large and small denominations as a ratio of GDP in Canada (1946–2018)
36
Sources: Bank of Canada and Statistics Canada
Chart 5b: Large and small denominations as a ratio of GDP in Sweden (1946–2018)
Chart 6a: Value share of cash, debit cards, credit cards and other payment methods, by
transaction value in Canada
37
Source: Bank of Canada Methods-of-Payment surveys, 2009–2017
Chart 6b: The main payment methods for various transaction values in Sweden
Source: Sveriges Riksbank Payment Patterns in Sweden, based on the questions, “What was your main
payment method for payments below SEK 100 (SEK 100–500 and over SEK 500) in a shop over the
past month?”
38
Note: The average transaction value of Swish in 2017 was about SEK550 (~CAD50) and the average
transaction value of e-Transfer in 2017 was about CAD408
39
Table 1. Bank note series (SEK 20 to SEK 10,000) issued in Sweden from 1894 to 2017
40
1991 1995 2000 2005 2010 2015
41
Earlier Working Papers:
For a complete list of Working Papers published by Sveriges Riksbank, see www.riksbank.se
Website: www.riksbank.se
Telephone: +46 8 787 00 00, Fax: +46 8 21 05 31
E-mail: registratorn@riksbank.se