The Impact of Electronic Money on Demand for Cash: Time Series Analysis

Abstract

The purpose of this chapter is to empirically examine the impact of e-money on the amount of money in circulation by type of money, taking into account as much as possible the structural changes in the financial system, including policy changes such as the increase in the consumption tax rate and the introduction of bank ATM charges.

5.1 Introduction

Recent years have seen the rapid spread of prepaid IC-type electronic payment methods known as electronic money, including Suica, ICOCA, PASMO, SUGOCA, Kitaca, Edy, nanaco, WAON, and others. The inclusion of the word “money” in the name has led to a perception of privately issued “money” being used as an alternative to cash, especially in small denominations. It is sometimes argued that privately-issued “money” will steal the seigniorage that the government and the Bank of Japan earn by issuing cash, or that private companies will be able to earn seigniorage. As things stand, it is clear that e-money is a type of payment method that requires a charge of cash and is not a substitute for cash itself.

It is also true, however, that in recent years there have been reports that the balance of small-denomination money in circulation has declined due to the impact of electronic money’s replacing it. On the other hand, there is a view that the decline in small-denomination money has been affected by structural changes in Japan, such as the introduction of fee-based bank ATMs, in addition to electronic money.

The purpose of this chapter is to empirically examine the impact of e-money on the amount of money in circulation by type of money, taking into account as much as possible the structural changes in the financial system, including policy changes such as the increase in the consumption tax rate and the introduction of bank ATM charges.

We would also like to discuss how widespread e-money actually is, what the characteristics of e-money are nationally, and what the prospects for its future are, as well as what the policy issues surrounding e-money are.

5.2 Overview of Electronic Money and Its Diffusion

With the launch of Edy by BitWallet in November 2001, IC-based e-money in Japan was introduced in earnest. In March 2004, JR East started e-money service for Suica, which had been used only for boarding passes. In October 2005, JR West also inaugurated an e-money service for ICOCA, which had been used only for boarding passes. In 2007, private railway and bus companies in the Tokyo metropolitan area launched PASMO in March, Seven & i Holdings unveiled nanaco in April, and Aeon brought in WAON in April. The year 2007 came to be called the first year of e-money.

As for the macro-statistics of e-money, the Payment and Settlement Organization Bureau of the Bank of Japan has published Recent e-money trends (FY2007 and FY2008) as part of its Survey Report on Payment and Settlement Systems. In the following, we will take a look at the actual situation based on the latest information from this report.

As shown in Table 5.1, e-money can be divided into two categories: (1) IC-type and (2) server-type. Of these, the current expansion of e-money is due to the increase in IC-type e-money. Table 5.1 also summarizes the characteristics of credit cards and debit cards, which are used as similar small-scale payment methods.

Table 5.1 Types of small scale electronic payment methods in Japan

IC-type e-money can be roughly divided into three categories depending on the issuing entity: those specializing in issuing e-money, such as BitWallet, which issues Edy; those issued by distribution companies, such as nanaco and WAON; and those issued by transportation companies, such as Suica, PASMO, and ICOCA. Of these, IC-type e-money and distribution-type e-money, both of which specialize in e-money issuance, can be used throughout Japan in any stores that accept them. In contrast, transportation-related IC-type e-money could initially be used only in the issuing company's respective business area. However, mutual use of transportation-related IC e-money is now expanding, such as mutual use between Suica and ICOCA (mutual use of e-money started in March 2008), and mutual use between Suica and PASMO (mutual use of e-money started in March 2007). In addition, new transportation-related IC e-money is being issued in each region of Japan.

Table 5.2 shows the total number of e-money cards issued, which exceeded 100 million in January 2009, and 105.03 million at the end of March, a year-on-year increase of 30.3%. The number of payment terminals installed in retail stores reached 480,000 at the end of March 2009, an increase of 34.1% year on year, higher than the growth rate of e-money. This is said to be due to the fact that distribution system e-money now allows mutual use across groups, and the installation of e-money-compatible vending machines began in October 2008.

Table 5.2 Number of electronic money issued

Table 5.3 shows that the number of electronic money settlements reached 1,116 million in FY2008 (up 37.8% from the previous year), and the settlement value increased rapidly to 817.2 billion yen (up 45.0% from the previous year). The settlement amount per transaction was 732 yen in fiscal 2008. This per-transaction amount has remained relatively stable at around ¥700 for the past three years. It should be noted that the average e-money transaction is a small transaction of less than 1,000 yen, and this amount has also remained stable at around 700 yen.

Table 5.3 Number and amount of payment by electronic money

The balance of e-money issued (unused balance) was 91.2 billion yen in March 2009 (+18.0% YoY), increasing in line with the increase in the number of cards issued. It should be noted, however, that not all e-money cards issued are being actively used, and the number of dormant cards is also increasing.¹

In addition, when the balance of e-money issued is compared with the amount of cash in circulation, it is clear that it is still very small. As of the end of March 2009, the outstanding balance of e-money was only 2.02% of the outstanding coins in circulation, 0.12% of the outstanding banknotes, and 0.11% of the total cash in circulation (the sum of the outstanding coins in circulation and the outstanding banknotes). It is clear that electronic money is not in a position to influence payment systems or monetary policy.

Electronic money and alternative small payment instruments include credit and debit cards. Table 5.4 shows a comparison of the use of small-payment methods. According to this table, the annual number of payments is by credit card, e-money, and debit card in that order. The amount of payment is also higher for e-money than for debit cards, but the amount per transaction is much smaller for e-money.

Table 5.4 Comparisons of small-scale payments

5.3 E-money Usage Based on Micro-data

So far, an overview of macro-aggregated data has shown that the use of e-money is growing rapidly. However, the empirical feeling is that the use of e-money is dependent on the availability of e-money in areas where it has been actively introduced, the use of transportation, and the availability of e-money at convenience stores and supermarkets. In addition, it can be assumed that it is the younger generation that actively uses these new types of payment methods, and moreover, it is by those who use transportation to commute to work or school. It is important to confirm how these differences in personal attributes affect the ownership and use of e-money through micro-data surveys of individuals and households.

5.3.1 Electronic Money Ownership

The Survey of Household Economy has been conducted by the Statistics Bureau of the Ministry of Internal Affairs and Communications (MIC) every month since October 2001, with the aim of capturing the actual state of personal consumption trends. Included is IT-related consumption, which has been increasing markedly in recent years, and high-priced goods and services, which are purchased infrequently. The “IT-related items” section of the Survey of Household Income and Expenditure began to include a survey on the ownership of electronic money in 2007 (January 2008).

This survey is conducted on approximately 30,000 households selected by the stratified three-stage random sampling method from all households in Japan. A total of 17,843 households with two or more persons and 1,692 single-person households were covered in FY2009, for a total of 19,535 households (65.1% collection rate). The same households are asked to continue the survey for 12 months, with a rotating panel structure in which one-twelfth of the households are replaced every month.

The 2008 report, for which detailed information is available, shows that the national average percentage of households with a household member who owns e-money is 24.4%, with the Kanto region having the highest percentage at 44.3%, followed by the Kinki region at 18.8% (see Fig. 5.1). However, there is a large diffusion gap between the Kanto region and the rest of Japan.

Fig. 5.1

Ownership of electronic money by region in 2008. Source Survey of Household Economy 2008, Statistics Bureau of the Ministry of Internal Affairs and Communications

When asked if any household member used electronic money, 18.0% of respondents nationwide replied that they used electronic money, with the Kanto region having the highest rate at 36.9%. The most common uses were transportation (12.5%) and at convenience stores (3.2%) (see Fig. 5.2). In the Kanto region, a similar trend was observed (30.5% and 3.7%, respectively), but in other regions, with the exception of Kinki and Chugoku, convenience store use was higher than transportation use. This is probably because many stations in these regions have not yet shifted to a system of using e-money in public transportation, and because compatibility of e-money among public transportation systems has not been promoted.

Fig. 5.2

Places and occasions electronic money were most frequently used in 2008. Source Survey of Household Economy 2008, Statistics Bureau of the Ministry of Internal Affairs and Communications

Next, let us look at the status of ownership by age group (Fig. 5.3). The highest percentage of 25–29-year-olds (42.6%), 30–34-year-olds (40.0%), and 35–39-year-olds (41.5%) use a mobile phone, and these age groups are also the most frequent users of electronic money. The most frequent uses are for transportation and at convenience stores. This suggests that workers are beginning to use e-money for transportation to and from work and are also using it at kiosks and local convenience stores. On the other hand, those aged 65 and over who no longer commute regularly are limited in their opportunities to use e-money.

Fig. 5.3

Ownership of electronic money by age-group in 2008. Source Survey of Household Economy 2008, Statistics Bureau of the Ministry of Internal Affairs and Communications

Looking at e-money ownership by income level, it is clear that the higher the income level, the more e-money is owned (see Fig. 5.4). If the majority of people use e-money for transportation to work or school, then differences in income should not be so relevant, but here it is clear that e-money ownership and use is correlated with income level. Of course, income is also correlated with age and occupation, so this may be a spurious correlation, and further analysis is needed on this point.²

Fig. 5.4

Ownership of electronic money by level of income in 2008. Source Survey of Household Economy 2008, Statistics Bureau of the Ministry of Internal Affairs and Communications

Incidentally, when we look at the status of e-money ownership by occupation, it is high (over 30%) among employed people and company executives, and low (under 20%) among the self-employed and unemployed.

5.3.2 Choice of Small-Payment Instruments

The choice of small-settlement instruments has been surveyed since 2007 in the Central Committee for Financial Services Public Relations’ Public Opinion Poll on Household Financial Behavior.

This survey is conducted on 10,500 households selected from among all households in Japan, and as of FY2009, 4,026 households (response rate: 50.3%) were selected from 8,000 households selected by the stratified two-stage random sampling method for households with two or more persons. In addition, 2,500 single-person households were included. It is important to note that this survey was conducted by visiting and mailing only households with two or more persons, using the usual random sampling method, while single-person households were surveyed using Internet monitors who had registered in advance. As we will see later, the significant difference in e-money usage between households with two or more members and single-person households may reflect this difference in survey methodology.³

The 2009 survey asked, “In your household, how do you use different means of payment for daily payments (shopping, etc.) depending on the amount? Please select the payment method you use most often for each amount” [multiple answers up to two means are allowed, Q13(a)]. For households with two or more members, 3.1% choose electronic money for payments of ¥1,000 or less, and 2.2% for payments of ¥5,000 or less, which is not so high, but for single-person households, 25.2% use it for payments of ¥1,000 or less and 14.3% for payments of ¥5,000 or less, which is extremely high.

As we saw earlier in the Survey of Household Economy by the Statistics Bureau of the Ministry of Internal Affairs and Communications, there seemed to be significant differences in ownership by region and age group, so we again tabulated the responses to Q13(a) by region and age group.⁴ The results are summarized in Table 5.5 (by age group) and Table 5.6 (by region) for the 2007 and 2008 surveys.

Table 5.5 Share of payment methods within the range of payment amounts by age group

Table 5.6 Share of payment methods within the renge of payment amounts by region

The points observed here are as follows. (1) Electronic money is chosen as a means of payment for payments of ¥5,000 or less, especially for payments of ¥1,000 or less. (2) In terms of age, working people between the ages of 25 and 49 are the main users. (3) In terms of region, the majority of users are in the Kanto region, but there are also a relatively large number of users in Hokkaido.

As seen earlier, the high use of e-money by single-person households is clearly due to differences in survey methods, so the results here cannot be simply compared, but it can be said that the use of e-money by single-person households is relatively high.⁵

The results of this survey are consistent with the fact that the average settlement amount per transaction is 732 yen, as already seen in Sect. 5.2 of this chapter, and that the majority of e-money use is for payments of 1,000 yen or less.

Incidentally, according to Kitamura (2005), the distinction of settlement methods can be expressed as shown previously in Fig. 4.1. For settlements of tens of thousands of yen or more, credit cards and electronic payments are already being used to make settlements by transferring information using electronic media rather than cash. The introduction of electronic payment methods, such as electronic money, to very small settlements, which used to be limited to cash, suggests that the use of cash may finally be limited.

However, more strictly speaking, the total amount of a small payment does not necessarily determine the means of payment between cash, electronic money, credit cards, and other means. Electronic money may be selected when payment in units of 1 yen or 5 yen is easier than cash transfer. For example, for a payment of 300 yen, it is not so complicated to use cash, but for a payment of 1,376 yen, the payment amount is larger than 300 yen but the amount of change used for exchange is larger and more complicated, so it is quite possible to settle with electronic money. In other words, the choice between electronic money and cash is not simply determined by the total amount of money, but also by the amount of small change that needs to be exchanged. Recall my discussion in Chap. 3 on small change. If this is the case, then it is likely that electronic money will be used primarily for payments at convenience stores and supermarkets where prices are set in units of 1 yen or 5 yen, while cash or credit cards will be used rather than electronic money at department stores and high-end retailers where prices are set in units of 100 yen or 1,000 yen.

5.4 Research on Money Demand

5.4.1 Previous Research

Discussions to date have revealed that e-money payments have expanded fairly rapidly and that the average amount per transaction is small, less than 1,000 yen. It also became clear that the number of users is unevenly distributed by region, age, occupation, and other factors, and that not all consumers in Japan possess or have access to the new means of electronic money. If we assume that the number of e-money cards issued exceeds 100 million and that 24.4% of households own e-money cards, the total number of households in Japan was approximately 50 million in 2009, which means that approximately 12.2 million households own multiple e-money cards (an average of 8 cards per household), and the remaining 37.8 million households have no relationship with e-money.

In econometrics, we should examine whether there is a difference in the demand for money between households that have e-money and households that do not have e-money. In addition, if we can obtain panel data on households that currently hold e-money, it will be possible to see whether there is a difference in money demand between households before and after holding e-money. In general, if micro-data are available, behavioral parameters can be estimated more precisely by modeling and demonstrating the behavior of individuals choosing payment methods.

In fact, empirical studies using micro-data have been actively conducted in other countries. For example, Attanasio et al. (2002) take up the diffusion of ATM cards as a new technology and estimate a Baumol–Tobin type money demand function taking into account the impact of ATM cards using the 1989–1995 Survey of Household Income and Wealth by the Bank of Italy. They find that the interest elasticity of money demand is higher for ATM card holders than for non-card holders. Lippi and Secchi (2007) use the same data as in Attanasio et al. (2002) but extend it to 2004 and consider the interest elasticity of demand for money, taking into account the conditions of ATM card use and the presence of ATM machines within a short distance. They find that the interest elasticity drops as the ATM distance becomes shorter and the elasticity approaches zero in the end. Alvarez and Lippi (2009) also use the Italian Household Income and Assets Survey (1993–2004) to extend the Baumol–Tobin model of cash demand to a dynamic framework and examine how technological progress, such as the proliferation of ATM cards, changes the preliminary demand for cash.

Stix (2004) empirically examines the impact of the widespread use of electronic payment systems (Electronic-Fund-Transfer-at-the-Point-Of-Sale, EFT-POS) and ATM cards on money demand in Austria, based on two surveys conducted by the Austrian National Bank in 2003 with 4,000 people (the actual sample was 2,800). Both EFT-POS and ATM cards have been shown to reduce the demand for money.

Duca and Whitesell (1995) use cross-sectional data from the 1983 Survey of Consumer Finances (SCF) by the Federal Reserve Board (USA) to examine the effect of credit card ownership on money and asset choice using probit estimation. Mulligan and Sala-i-Martin (2000) also use the 1989 SCF to obtain the interest elasticity of household demand for money from cross-sectional differences in asset holdings. They find that the interest elasticity is lower when the interest rate is lower, the holding of interest-bearing financial assets is proportional to the total financial assets, and the participation cost of holding interest-bearing financial assets is lower for those who participate in pension programs.

As we saw in Sect. 5.3, micro-data on e-money in Japan is finally accumulating, but it is still only available for a couple of years and does not contain enough survey items to enable a comprehensive analysis of money demand and asset choice as in the empirical studies in Europe and the United States mentioned above. So far, empirical studies on e-money in Japan have mainly taken the approach of examining the impact of adding a variable indicating the degree of e-money diffusion to the money demand function by money type using the balance of money issued at the national level.⁶

The exception is Fujiki and Tanaka (2009). This chapter examines the impact of e-money on cash holdings using cross-sectional data from the Central Committee for Financial Services Information's Public Opinion Survey on Household Financial Behavior in 2007, which we saw earlier. They report that the introduction of e-money has not had the substitution effect of reducing cash holdings, but rather has increased them in some cases. Nakata (2009) conducted a questionnaire survey on the actual status of e-money diffusion among consumers living in Fukuoka Prefecture and summarized the results. The survey revealed that consumers who use e-money frequently reduce the number of times they make cash payments, and some of them also reduce the amount of cash they hold.

One study of e-money in Japan is The Development of Electronic Money and Financial and Economic Systems by the Financial Research Institute. In it, Saito (2005) estimates the demand for e-money based on the estimation of the real money demand function and describes the impact of e-money through Marshall's k. Specifically, he argues that if e-money is issued that is highly substitutable for small-value payment media and saves the use of coins, which are cumbersome to use, its issuer can establish an issuance base in a zero-interest-rate environment, thereby securing a solid profit opportunity through future rises in nominal interest rates.

Prior studies on the relationship between e-money and money demand include Nakata (2007, 2009). Nakata estimates the money demand function for each money type as in this chapter. Although the functional form differs somewhat from the model in this chapter, it is confirmed that the diffusion of e-money has a significant negative correlation with the growth rate of money circulation. In addition, Nakata estimates two VAR (Vector Auto Regressive) equations, one for the money demand function by money type and the other for the e-money penetration index, and looks at the impulse response function to see the impact of e-money on the money demand by money type. Table 5.7 summarizes the methodology of empirical studies based on money demand functions for e-money in Japan.

Table 5.7 Empirical researches on money demand function taking into account of electronic money in Japan

As already discussed, at present there is no detailed e-money and currency statistical information on individual settlement amounts at the micro level, so we will examine the impact of the increase in e-money payments on the demand for small change by estimating the demand function for each type of money in the aggregate. Our interest here is not in the use of e-money as a means of making small payments, but in examining how it reduces the demand for small coins such as 1 yen, 5 yen and 10 yen coins.

5.4.2 Specification of the Money Demand Function

The demand for money is one of the most extensively studied empirical areas. Historically, there are several models based on the quantity theory of money formulated by Friedman (1956, 1969), the liquidity preference model based on Keynes (1936), and the inventory model based on Baumol (1952) and Tobin (1956) Inventory Model considering transaction costs.⁷ Various theoretical models have been proposed and demonstrated, including Sidrauski's (1967) Shopping Time Model, Clower's (1967) Cash-in-Advance Model, and Tobin's (1958) Portfolio Model, have been proposed and demonstrated.⁸

In this chapter, instead of comparing these models, we consider a general money demand function and add the effects of demographic change and the increase of new payment methods such as electronic money to it.⁹

It is easy to imagine that population increases aggregate money demand: even if per capita money holdings remain constant, an increase in population increases the demand for goods transactions, and the demand for money increases correspondingly. Electronic money, on the contrary, enables small payments of any amount, and the incentive to hold small change decreases. Of course, the infrastructure for e-money payments is not so well developed that all economic transactions can be settled electronically. However, in densely populated metropolitan areas where economic transactions are concentrated, electronic payments tend to increase not only because consumers have a motivation to keep less money, but also because retailers have a motivation to save money by speeding up the settlement process and avoiding errors in change calculation associated with cash payments. In addition to the consumer's motivation to own money, the retailer's motivation to save money is also increasing.

In the following, we consider the following money demand function.

$$\mathrm{ln}\left(\frac{{M}_{t}}{{P}_{t}}\right)=\mathrm{f}({y}_{t},{i}_{t},{em}_{t},{z}_{t})$$

(5.1)

where ${M}_{t}$ is the demand for money, ${P}_{t}$ is the price level (or the face value of each type of money in the case of gold-type money), ${y}_{t}$ is the real economic activity (the real consumption level ${c}_{t}$ may be used depending on the formulation of the model), ${i}_{t}$ is the interest rate, ${em}_{t}$ is the electronic money factor, and ${z}_{t}$ is the population factor.

This Eq. (5.1) is further linearly approximated for the empirical model and organized as follows.

$$ln{m}_{jt}={a}_{0}+{a}_{1}ln{y}_{t}+{a}_{2}{i}_{t}+{a}_{3}\mathrm{ln}\left(e{m}_{t}+1\right)+{a}_{4}ln{z}_{t}+SD{M}_{t}+{\varepsilon }_{t}$$

(5.2)

where ${m}_{jt}={M}_{jt}/{P}_{jt}$ and j means that it is money (denomination) type j. Thus, ${m}_{jt}$ represents the amount of type of money or banknotes in circulation; ${SDM}_{t}$ is a seasonal dummy.

However, Eq. (5.2) does not take into account the structural changes and cyclical factors during the estimation period. Therefore, following the standard method of time series analysis in recent years, we remove the seasonal adjustment and cyclical factors from the original data, and furthermore, we take the first-order difference to ensure stationarity.¹⁰

$$\Delta ln{m}_{jt}={a}_{0}+{a}_{1}\Delta ln{y}_{t}+{a}_{2}\Delta {i}_{t}+{a}_{3}\mathrm{\Delta ln}\left({em}_{t}+1\right)+{a}_{4}\Delta ln{z}_{t}+{\varepsilon }_{t}$$

(5.3)

This model is a dynamic money demand function, and an error correction term is introduced to take into account the long-run equilibrium relationship between the volume of money in circulation and its explanatory variables.¹¹ This is a method of estimation in which the deviation between the equilibrium value and the realized value is regarded as the deviation from the long-run equilibrium. This is an estimation method in which the deviation between the equilibrium value and the realized value is regarded as the deviation from the long-term equilibrium and included in the estimation formula. This estimation method enables us to confirm whether the force to return to equilibrium is working when deviation from the long-term equilibrium occurs. There are two possible ways to create the error correction term. One is to use the residual of the basic estimation Eq. (5.2) as the margin of deviation, and the other is to estimate the amount of money in circulation based on the level of economic activity and the interest rate, and use the residual of that as the margin of deviation. The latter is used here.

$$ \begin{aligned} {\Delta }lnm_{jt} & = a_0 + a_1 {\Delta }lny_t + a_2 {\Delta }i_t + a_3 {\Delta }\ln \left( {em_t + 1} \right) + a_4 {\Delta }lnz_t \\ & \;\;\;\; + a_5 \left( {lnm_{jt} - \widehat{a_1 }lny_t - \widehat{a_2 }i_t } \right) + \varepsilon_t \\ \end{aligned} $$

(5.4)

Here, after a₅ is the error correction term. If the sign of a₅, which is the coefficient of the error correction term, is negative, it indicates that there is a movement back to long-term equilibrium in the following period or later when deviations occur.

Equation (5.4) is the empirical model used in this chapter. The final model selection was made by fitting the actual data to this model.¹² The sample size is 170. The estimation period is from November 1994 to December 2008 (sample size: 170).¹³ The final variables used are: ${m}_{jt}$ is the “volume of money and Bank of Japan notes in circulation,” ${y}_{t}$ is the “industrial production index,” ${c}_{t}$ is the real “commercial sales value (billion yen),” ${i}_{t}$ is the “over-the-counter (OTC) interest rate for less than one year (percent),” ${em}_{t}$ is the “cumulative number of Edy + Suica cards issued (billion)” and ${z}_{t}$ is “population (billion people)”.

5.5 Interpretation of the Empirical Results

Table 5.8 summarizes the estimation results. This result shows that the sign condition required by the theory of money demand function is almost satisfied. That is, the coefficients on the level of real economic activity are mostly positive and significant, and the coefficients on the less-than-one-year term OTC interest rate are mostly negative and significant.

Table 5.8 Estimation results of money type demand function

The coefficients of population growth rate are positive and significant for almost all the money types. This implies that population growth increases the demand for money.

As for the diffusion of electronic money, all coefficients for the 1 yen, 5 yen, 10 yen, and 50 yen coins are negative and significant except for the 1 yen coin. The coefficients for the 100 yen coin and 500 yen coin are negative but no longer significant, while those for the 1,000 yen note are positive and significant, and those for the 5,000 yen note and 10,000 yen note are no longer significant.

The error correction term is negative and significant for the 1 yen coin, 10 yen coin, 50 yen coin and 100 yen coin. The error correction term is negative and significant for the 1 yen coin, the 10 yen coin, the 50 yen coin and the 100 yen coin, while it is positive and significant for the 1,000 yen note and the 5,000 yen note. In other words, when small denomination coins deviate from the long-run equilibrium, there is a tendency to return to equilibrium in the following period.

Since Eq. (5.4) was a model that took the first-order difference of logarithms, the estimated coefficient directly represents the elasticity of demand for money of the variable. In addition to the information in Table 5.8, we also include the estimation results of commercial sales as an explanatory variable, and Table 5.9 summarizes the elasticities.¹⁴

Table 5.9 Various elasticities of money type demand

The coefficient for the elasticity of real economic activity is around 0.15. For interest rate elasticity, the coefficient is around −0.2 to −0.5. For e-money, the order of magnitude is even lower and is around −0.0002 to −0.0007 for significant money types. Comparing the orders of magnitude of these values, it is clear that the impact of e-money on money demand is insignificant compared to the impact of other variables.

If we consider e-money as an additional new means of payment, the current rapid spread of e-money has had a reasonably significant impact on money demand, but the impact is expected to decline further once the spread of e-money has slowed down. It should also be noted that the diffusion of e-money may affect interest elasticity, as Lippi and Secchi (2007) pointed out that interest elasticity is lower for people who have ATM machines close at hand.

In this chapter, the elasticity of the population growth rate is estimated to be 5 to 10, which is extremely elastic. The reason for the extraordinary elasticity of the population growth rate needs to be considered further, including the possibility that this variable is a proxy for some other variable.

Next, let us compare the elasticity between the different money types (denominations). The elasticity of economic activity increases from the 1 yen coin to the 1,000 yen note. In contrast, the elasticity of interest is around −0.2 to −0.5, with little difference between the different denominations. As for electronic money, the elasticity of small denominations is relatively large, and the higher the denomination, the lower the elasticity and the less significant it becomes.

5.6 Policy Implications

5.6.1 Technological Progress and Security Issues in Electronic Money

The current mainstream of electronic money in Japan is in the form of contactless IC cards, the technology for which is called FeliCa, and almost all electronic money uses this technology. The advantage of this card is that it can be used for high-speed authentication and payment processing at automatic gates of transportation systems, building entry, and cash registers of convenience stores and kiosks. Contact-type IC cards are more secure in terms of security, as they require the entry of an ID and password, but they take too much time and are not suitable for authentication and settlement at automatic transportation gates or convenience stores. At present, there is a limit to the amount of balance that can be stored in e-money, and it is unlikely that there have been any cases where someone has deciphered the encryption on a contactless IC card and stolen the balance of e-money without coming into contact with another person, but if e-money issuers raise the balance limit in the future, the incentive to decipher the encryption may increase.

Private companies are urged to make efforts to increase the security of current contactless IC cards.¹⁵ At the same time, however, we must not become overly risk-averse and overemphasize security issues to the point of blocking private-sector initiatives for technological innovation.

If e-money is considered to be a new means of payment, it has been judged that it has had little impact on the payment and financial systems at present, when it accounts for only about 0.1% of total payments. So, will there be any government involvement in the e-money business that is being promoted on a private sector basis? If the government considers this to be a kind of technological innovation in the field of payment methods, and if it functions in a neutral manner, then it would make sense to watch the spread of e-money on a private-sector basis. As long as other payment methods such as cash, bank transfers, postal transfers, and credit cards are widely available, there is no need for the government to correct or subsidize regional differences in the spread of electronic money.

5.6.2 Resource Savings for Small Denominations of Money

It turns out that electronic money is not only used for small payments of 1,000 yen or less, but also for payments with fractions of a yen or five yen. The U.S. penny coin is made of zinc, but due to the recent sharp rise in the price of the metal, it costs about 1.4 penny to make a penny coin, and there have been repeated arguments that the penny coin should be abolished as a supplement to the basic unit of currency, the dollar. In fact, the 1 yen coin is made of aluminum, the 5 yen coin is made of an alloy of copper and zinc (brass), and the 10 yen coin is made of bronze (an alloy of copper and tin containing zinc). For small-denomination coins, at least the 1 yen coin, it is assumed that the raw material costs and production costs exceed the face value.¹⁶

As we have already seen, the distribution of e-money users is skewed toward urban areas, and it cannot be assumed at this point that e-money will be used on a nationwide scale, and in that sense it is likely that small denomination coins will continue to circulate. If the empirical results of this chapter are correct, demand for small denominations such as the 1 yen, 5 yen, and 10 yen coins will certainly decline. We would like to point out that this will also help to conserve precious metal resources.

In addition, if the future consumption tax rate is not a rounded figure, there may be many cases where the amount paid will be a fraction of that amount even if the tax is set within. In this case, too, more and more people will choose to pay with electronic money, which could also lead to savings in the metal resources used for coins. In fact, it is recorded that the demand for 1 yen and 5 yen coins increased when the consumption tax was introduced in 1989 and when the consumption tax rate was raised in 1997. However, as e-money spreads steadily, the impact of a consumption tax rate hike on the small-denomination coins would be negligible.

In addition, settlements of one yen or less can be made on electronic money, which may allow for greater flexibility in pricing. Furthermore, e-money payments are also known to save time, and are being introduced in major stations, kiosks, convenience stores, and cafeterias of large corporations in the Tokyo metropolitan area, where payments are concentrated.

5.6.3 Legal System and Accounting Standards

Until now, the law regulating e-money has been the Prepaid Voucher Regulation Law (Prepaid Card Law), which imposes an obligation to deposit at least half of the balance if the unused balance on the base date exceeds 10 million yen. Until now, this law only applied to IC-type e-money (store value type) and not to server-type e-money. Therefore, in June 2009, the Law on Funds Settlement (Law No. 59 of 2009) was promulgated, and server-based e-money was subject to the same regulations as IC-type e-money, and the Prepaid Card Law was abolished.

Although not the subject of this chapter, points, which are equivalent to premiums or discounts given by companies for the purpose of sales promotion, can be exchanged for e-money, and it is becoming possible to purchase products with a single set of points on a point exchange site or to convert them into cash through online banking. At present, points are discussed in the framework of consumer protection and competition policy under the Act Against Unjustifiable Premiums and Misleading Representations and the Anti-monopoly Act, and the discussion has not converged to the point of inclusion in the Law Concerning Funds Settlement in relation to electronic money, which remains an issue to be considered in the future.

There is also a major debate regarding the accounting treatment of companies that issue points. Under International Financial Reporting Standards (IFRS), the issuance of points should be recorded as a liability and the “fair value” of the points should be excluded from sales, based on the interpretation that points are a “deferred portion of future sales. However, this method would inflate the deferred revenue (liability) and lower the company's profitability index, and there has been much opposition from the corporate side, and no agreement has been reached on the accounting treatment of the points.

Similar to points, there is a type of local currency. It is used as a means of exchanging goods and services only within a region, and is used not only for economic effects but also as a means of stimulating volunteer activities and exchanges in the region. In recent years, it has become possible to record this local currency on electronic money such as Suica and PASMO.¹⁷ At present, this information is stored separately from electronic money, but just as airline mileage and points at mass merchandisers have somehow become exchangeable for electronic money, we should assume that local currencies may become national currencies through private-sector exchange markets.

5.7 Conclusion

In this chapter, after looking at the overview of e-money and its actual diffusion, we examined the actual usage based on more detailed micro-data. The results show that 24.4% of households in Japan have e-money, with the Kanto region having the highest percentage at 44.3%, and that the majority of users are in the 25–49 age group and use e-money for transportation. In addition, e-money is mainly chosen for payments of less than 1,000 yen, and its usage rate is likely to be high among single people.

Furthermore, as a result of estimating the money demand function, it is clear that the demand for small denominations of 50 yen coins and below has declined due to the spread of electronic money. On the other hand, the demand for 1,000 yen bills has increased due to the spread of e-money, which is used for recharging and other purposes, indicating that e-money is not a one-way substitute for money. Although not demonstrated here, the results of Fujiki and Tanaka (2009) and other studies indicate that during the diffusion period of e-money, the demand for money may temporarily increase because the need to hold both money and e-money arises due to uncertainty about the acceptability of e-money. In addition, the money demand elasticity of e-money is extremely low, and the impact of e-money on money demand is found to be limited even during periods of rapidly expanding use such as the current period.

Looking at these results, it can be expected that e-money will be one of the payment methods and that it will be neutral to the real economy in the long run, and that it will have no impact on money demand in the long run. The widespread use of e-money is a kind of infrastructure development, which may provide business opportunities for companies involved in it, but it is unlikely to have an impact that will change the nature of the real economy.

However, electronic money is not limited to time-saving settlements at automatic ticket gates of transportation systems and convenience stores; it also has a high potential to become a highly convenient means of payment for the elderly who find handling small change cumbersome in an aging society. Furthermore, electronic money is also an effective means of saving resources in the form of small denominations of money and expanding the degree of freedom in setting prices. It would be desirable to make active use of these new technologies.

It is necessary to develop a legal system and set accounting standards for this purpose. It will also be important to develop and publish e-money-related statistics in order to understand the actual situation surrounding e-money.