## Instant Payments — Introduction and Key Findings

## Source details

**Canonical URL:** [Instant Payments — Introduction and Key Findings](https://www.imf.org/-/media/files/publications/wp/2022/english/wpiea2022228-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2022/english/wpiea2022228-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2022/english/wpiea2022228-print-pdf.pdf.json)

---

### Scope and objective
- Examines benefits of instant payments (IP) and their effect on other payment instruments by analyzing a selection of twelve countries where sufficient data are available.
- Countries covered: Australia, China, India, Indonesia, Malaysia, Mexico, Norway, Russia, Singapore, Sweden, Thailand, and United Kingdom.
- Highlights potential additional benefits if IP systems in different countries can be connected to enable lower value cross-border retail payments, travel payments, and remittances.

### Background facts and motivation
- Some 55 countries have implemented instant payments between 2001 and 2019 and at least 10 more are projected to do so.
- Instant payments complete and settle (via a central bank) in seconds or minutes, reducing payment system risk and improving cost efficiency and operating effectiveness.
- Available data on instant payments and other instruments are limited.

### Structure of the analysis
- Section 2: Outlines benefits of IP systems relative to cash, cards, and debit instruments (checks and direct debits).
- Section 3: Illustrates the role of regulatory authorities in catalyzing, overseeing, or operating instant payment systems for the twelve countries.
- Section 4: Illustrates change in payment composition using domestic values of different payment instruments over 2005-2020 (16 years).
- Section 5: Presents an econometric model of payment substitution using payment shares to normalize differences and facilitate cross-country comparisons.
- Section 6: Reports estimates of payment substitution reflecting the underlying distribution of payment costs, revenues, and non-monetary benefits driving use of payment instruments.
- Section 7: Provides conclusions. Appendix contains additional information on payment substitution and data sources.

### Conceptual framing and policy relevance
- Payment substitution analysis is analogous to econometric estimation of elasticity of physical capital substitution for labor; estimates reflect the underlying, but unknown, distribution of payment costs, revenues, and non-monetary benefits at the national level.
- Degree of IP adoption and substitution indicates:
  - Improvement in productive efficiency of a country's payment system by supplying lower cost payment methods.
  - Extent to which risk in the payment system has been reduced.
- Both improved efficiency and reduced risk are important payment policy objectives.

### Changes in Payment Composition Across Twelve Countries — Key findings
- Seven countries experienced statistically significant payment substitution of instant payments for cash, for cards, or for checks: China, Indonesia, Mexico, Norway, Russia, Singapore, and the UK.
- For analysis, ChecksDD combines checks and direct debits.
- Eight countries have seen cash use rise over 2005-2020 while only four find cash use on average constant or falling between 2005 and 2020.
- Cash use often displays an inverted "U-shape": negative annual growth rate of cash over the last three years of the sample (Column 2).
  - Only two countries (Russia and Indonesia) show a positive annual average growth rate in cash over this more recent time period.
- Percentage point change in the share of cash in total payments (Column 3): eight countries show a reduction or no change; four countries (India, Mexico, Malaysia, and Singapore) show a rise.
- Share changes by instrument (2005–2020, Columns referenced):
  - Large rise in the share of instant payments (Column 6).
  - Smaller expansion in the current share of cards (Column 5).
  - Large reduction in the share of debit instruments (ChecksDD) (Column 7).
  - Smaller reduction in the share of cash (Column 3).
- Measurement notes:
  - Cash use measured by average annual compound growth rate of value of cash withdrawn from ATMs (Column 1).
  - Table 1 columns: Column 1: % Change 2005-20 (Cash Payment Values, Annual); Column 2: % Change Last 3-Yrs (Cash Payment Values); Column 3: % Point Change 2005-20 (Cash, All Payment Shares); Column 4: % Point Change Last 3-Yrs (Cash, All Payment Shares); Column 5: % Point Change 2005-20 (Cards, All Payment Shares); Column 6: % Point Change 2005-20 (IP, All Payment Shares); Column 7: % Point Change 2005-20 (ChecksDD, All Payment Shares).
  - Time period covered was 2005-2020 for all countries except China (2006-20) and Indonesia (2009-20).
  - Standard alternative measure: CIC/GDP ratio; reductions using payment shares (Column 3) are more common than CIC/GDP or small denomination ratios.

### Exclusions and data caveats
- Japan and Korea excluded because lower value retail instant payments are comingled with large value instant payments.
- Argentina excluded due to insufficient data.
- Other cash acquisition channels (cash-back, OTC withdrawals) are almost never publicly reported or are inconsistent with consumer-focused measures.

### A Measure of Payment Substitution — Model and estimation
- Objective: Use a production function (translog) framework to determine substitution or complementarity among four payment instruments producing total payments (TP).
- Payment instruments in model:
  - 1. cash
  - 2. debit and credit cards (including e-money)
  - 3. instant payments (IP)
  - 4. Checks and direct debits (ChecksDD)
- Wire transfers excluded as functionally separable.
- Translog quadratic in logs specified:
  - (1) lnTP = lnαo + ∑i αi lnPi + 1/2∑i ∑j βij lnPi lnPj + ε
  - ∂lnTP/∂lnPi gives elasticity; negative cross-coefficients (βij < 0) indicate substitution.
- Practical handling of zeros: insert small positive value to take logs when instrument value is zero.
- Mapping to observed shares: payment elasticities replaced with observed shares Si under revealed-preference analogy.
- Estimation: system of four equations estimated jointly using stacked SURE with coefficient symmetry (βij = βji) imposed.
- Ad hoc linear system (levels) also estimated (equation (4)) with symmetry; βij < 0 indicates substitution, βij > 0 indicates complementarity.
- Robustness:
  - Re-estimated with all RHS variables lagged by one year.
  - Unit root and cointegration tests indicate non-stationarity; first-difference re-estimation conducted for illustration.
  - Degrees of freedom: df = (4 x n) - 14, where n = sum of observations across four equations for 2012-2020 or 2005-2020.
  - Instant payments were zero in 2005 and most implementations occurred after 2011.

### Substitution Results (2012–2020) — Summary of empirical findings
- Six types of payment substitutions possible for four payment classes; with twelve countries there could be seventy-two symmetric substitutions.
- Table 2 reports eighteen statistically significant payment substitutions for the full model (Eq. 3, preferred translog quadratic specification) for 2012–2020.
- Significance levels: p-values .01 (**), .05 (*), and .10 (º).
- Key statistically significant IP substitutions (selected estimates reported exactly as in source):
  - IP replaces Cash:
    - China: Column (3) -.071º
    - Indonesia: Column (1) -.010**, Column (3) -.012**
    - Russia: Column (1) -.040**, Column (3) -.063*
    - Sweden: Column (2) -.006* (Column (3) + reported though not significant)
    - UK: Column (1) -.041**, Column (3) -.041**
  - IP replaces Cards:
    - China: Column (2) +, Column (3) -.069*
    - Indonesia: Column (2) +, Column (3) -.003*
    - Norway: Column (1) -.011**, Column (2) -.032**, Column (3) -.547º
    - UK: Column (1) -.062**, Column (3) -.500*
  - IP replaces ChecksDD:
    - Australia: Column (2) -.066**, Column (3) -.034º
    - China: Column (1) -.058**, Column (2) -
    - Indonesia: Column (1) -.006º
    - Mexico: Column (2) -.038**, Column (3) -.225**, Column (3 alternative?) -.074**
    - Singapore: Column (1) -.013º
    - Thailand: Column (2) -.011*, Column (3) +
    - UK: Column (1) -.120**, Column (3) -.109**
- Cards replacing other instruments (selected):
  - Cards replace Cash:
    - Mexico: Column (3) -.023**
    - Norway: Column (3) -.113*
    - Russia: Column (1) -.170**, Column (3) -.191**
    - Sweden: Column (1) +, Column (2) +, Column (3) -.042**
  - Cards replace ChecksDD:
    - India: Column (1) -.168º, Column (3) +
    - Mexico: Column (1) -.063**, Column (3) -.071**
    - Norway: Column (2) .205**, Column (3) -
    - Russia: Column (1) -.034º, Column (3) -.033º
    - Singapore: Column (2) -.365*
    - Sweden: Column (1) -.301**, Column (3) -.190**
    - UK: Column (1) -.127**, Column (3) -.101**
- ChecksDD replace Cash:
  - Mexico: Column (1) -.101**, Column (3) -.127**

### Model robustness and specification comparisons
- Three specifications reported:
  - (1) Full translog quadratic (preferred).
  - (2) All RHS variables lagged by 1 year.
  - (3) First-differenced variables.
- Counts of statistically significant substitutions by specification:
  - Full model (Column 1): eighteen significant substitutions.
  - Lagged-1-year (Column 2): six significant substitutions.
  - First-differenced (Column 3): twenty significant substitutions; 13 duplicate sign and significance with full model.
- Linear specification (equation (4)) produced only five significant substitutions and none involved instant payments.
- Interpretation: translog captures nonlinear, logistic-like adoption/decline dynamics; linear model fails to capture these features.

### Aggregate share changes and country comparisons (reported totals)
- Country ranking of most to least change over 2005–2020 (as reported): Russia, China, India, UK, Indonesia, Mexico, Malaysia, Singapore, Thailand, Sweden, Australia, Norway.
- Aggregate reported percentage point changes (reported totals):
  - Greatest reduction: checks and direct debits (-390).
  - Next greatest reduction: cash (-105).
  - Largest increases: instant payments (+310) and cards (+143).
  - Overall total percentage point change in shares reported as +501/-502 (with some rounding error).
- Country examples of sum of changes:
  - Russia: sum of positive (negative) changes = +67 (-67) percentage points.
  - Norway: sum of positive (negative) changes = +14 (-14) percentage points.

### Additional empirical notes and data context
- Check use is close to or at zero in Norway, Russia, Sweden, and the UK; India and Mexico still use check/debit instruments extensively though falling.
- Sweden: ATM cash withdrawals have been falling slowly since 2006, seven years before Swish (instant payments) in BiR in 2012; preferred model does not show significant replacement of cash by IP in Sweden over 2012–2020.
- Sample size per country is small; outliers (e.g., 2020 for one country) had little meaningful effect given R2 > .95.
- Appendix results for 2005–2020 show instant payments significantly substitute for cash, cards, or ChecksDD eight times using quadratic model versus ten in the 2012–2020 focus.

### Policy-relevant takeaways and implications for retail CBDC
- Translog quadratic model identifies significant payment substitution consistent with logistic adoption/decline patterns; instant payments are statistically significant substitutes for checks, cash, and cards in multiple countries during 2012–2020.
- Instant payments have experienced the largest increase in payment share (+310 percentage points aggregate), while checks and direct debits have experienced the largest aggregate decline (-390 percentage points).
- The emergence and adoption of instant payments improve payment system cost efficiency and reduce payment system risk; measuring substitution provides a means to monitor these policy objectives.
- Instant payments and retail CBDC share strong similarities when retail CBDC is operated and settled through a central bank:
  - Both provide instantly available good and final funds backed by a central bank.
  - Both could offer zero-user cost (as with Swish) or de minimis fee (as envisioned for FedNow).
  - Both can accommodate P2P payments and can be adjusted to enable wallet-to-wallet transfers without external terminals (CBDC-enabled mobile phones example: China).
  - Both could be modified to transfer funds anonymously within value limits.
- Policy implication: if instant payments are structured as a feasible alternative to retail cash at the retail level, they may limit incentives to adopt retail CBDC. The case for issuing retail CBDC has been found less compelling in some advanced economies (Australia, Singapore, UK) compared to emerging markets and developing economies where motivations include promoting financial inclusion, enhancing payment system efficiency, competition, security, resiliency, and cross-border payments.

*Source: IMF Working Paper — Instant Payments: Regulatory Innovation and Payment Substitution Across Countries (extracted content).*

### 1.  Introduction .......................................................................................................

### 1. Introduction

### Scope and objective
- Examines benefits of instant payments (IP) and their effect on other payment instruments by analyzing a selection of twelve countries where sufficient data are available.
- Countries covered: Australia, China, India, Indonesia, Malaysia, Mexico, Norway, Russia, Singapore, Sweden, Thailand, and United Kingdom.
- Highlights potential additional benefits if IP systems in different countries can be connected to enable lower value cross-border retail payments, travel payments, and remittances.

### Background facts and motivation
- Some 55 countries have implemented instant payments between 2001 and 2019 and at least 10 more are projected to do so.
- Instant payments complete and settle (via a central bank) in seconds or minutes, reducing payment system risk and improving cost efficiency and operating effectiveness.
- Available data on instant payments and other instruments are limited.

### Structure of the analysis
- Section 2: Outlines benefits of IP systems relative to cash, cards, and debit instruments (checks and direct debits).
- Section 3: Illustrates the role of regulatory authorities in catalyzing, overseeing, or operating instant payment systems for the twelve countries.
- Section 4: Illustrates change in payment composition using domestic values of different payment instruments over 2005-2020 (16 years).
- Section 5: Presents an econometric model of payment substitution using payment shares to normalize differences and facilitate cross-country comparisons.
- Section 6: Reports estimates of payment substitution reflecting the underlying distribution of payment costs, revenues, and non-monetary benefits driving use of payment instruments.
- Section 7: Provides conclusions. Appendix contains additional information on payment substitution and data sources.

### Conceptual framing and policy relevance
- Payment substitution analysis is analogous to econometric estimation of elasticity of physical capital substitution for labor; estimates reflect the underlying, but unknown, distribution of payment costs, revenues, and non-monetary benefits at the national level.
- Degree of IP adoption and substitution indicates:
  - Improvement in productive efficiency of a country's payment system by supplying lower cost payment methods.
  - Extent to which risk in the payment system has been reduced.
- Both improved efficiency and reduced risk are important payment policy objectives.

_Italic: IMF Working Papers — Instant Payments (chapter: Introduction)._

### 4.   Changes in Payment Composition Across

### 4.   Changes in Payment Composition Across Twelve Countries

### Overview and key findings
- Seven countries experienced statistically significant payment substitution of instant payments for cash, for cards, or for checks: China, Indonesia, Mexico, Norway, Russia, Singapore, and the UK.
- Countries experiencing significant substitution of a different type—cards replacing cash or checks, or checks replacing cash—are also reported in the figures.
- For the analysis that follows, the two debit instruments—checks and direct debits—are combined as ChecksDD.
- Eight countries have seen cash use rise over 2005-2020 while only four find that cash use is on average constant or falling between 2005 and 2020.
- Cash use over the period often displays an inverted "U-shape": first rising and, toward the end, falling. This is shown by the negative annual growth rate of cash over the last three years of the sample (Column 2).
  - Only two countries (Russia and Indonesia) show a positive annual average growth rate in cash over this more recent time period.
- Looking at percentage point change in the share of cash in total payments (Column 3), eight countries show a reduction or no change in the use of cash, while four countries (India, Mexico, Malaysia, and Singapore) show a rise in their cash share.
- With one exception (Mexico), the share of cash in the last three years of the sample are uniformly negative (Column 4), mirroring the last three years result for payment values.

### Interpretation of changes and substitution patterns
- Two mechanisms can cause the share of cash to fall:
  - Cash falls absolutely (as measured by % change 2005-20 in cash payment values; Column 1).
  - Other instruments grow faster and displace cash (observed in % point change in payment shares; Column 3).
- The percentage point changes across instruments indicate:
  - A large rise in the share of instant payments (Column 6).
  - A smaller expansion in the current share of cards (Column 5).
  - A large reduction in the share of debit instruments (ChecksDD) (Column 7).
  - A smaller reduction in the share of cash (Column 3).
- These patterns suggest that cards and instant payments have likely substituted for both debit instruments and cash over the period.
- Visual assessments (Figures 1 and 2) are suggestive of substitution but require a more formal analysis to confirm negative covariation among instruments while holding constant other covariations.

### Data, measures, and table interpretation
- Cash use is measured here by the average annual compound growth rate of the value of cash withdrawn from ATMs (Column 1).
- Table 1 columns referenced in the analysis:
  - Column 1: % Change 2005-20 (Cash Payment Values, Annual)
  - Column 2: % Change Last 3-Yrs (Cash Payment Values)
  - Column 3: % Point Change 2005-20 (Cash, All Payment Shares)
  - Column 4: % Point Change Last 3-Yrs (Cash, All Payment Shares)
  - Column 5: % Point Change 2005-20 (Cards, All Payment Shares)
  - Column 6: % Point Change 2005-20 (IP, All Payment Shares)
  - Column 7: % Point Change 2005-20 (ChecksDD, All Payment Shares)
- Table Notes: Data are rounded. ChecksDD represents checks and direct debits together. The time period covered was 2005-2020 for all countries except China (2006-20) and Indonesia (2009-20).
- Standard alternative measure: ratio of cash in circulation (CIC) to GDP. The focus on percentage point changes in payment shares is a similar relative measure to assess use and substitution.
- Reductions in cash use shown using payment shares (Column 3) are more common than when the ratio of CIC to GDP, or the ratio of the value of small denomination currency and coin to GDP or household consumption, are used.

### Notes on data coverage and exclusions
- Japan and Korea were excluded because their payment data on lower value retail instant payments are comingled with large value instant payments and are not comparable to the other countries covered.
- Argentina was excluded as the data available was insufficient.
- Other ways of obtaining cash (cash-back at point-of-sale, over-the-counter withdrawals) exist but are almost never publicly reported; OTC withdrawals, when reported, include consumer and business withdrawals together, which is inconsistent with other payment instruments focused on consumer payments.

*Source: BIS, National Central Banks, UK FPS*

### 5.   A Measure of Payment Substitution Across

### 5.   A Measure of Payment Substitution Across Countries

### Overview
- Objective: Use a production function framework to determine substitution or complementary relationships among payment instruments in producing and consuming total output (GDP) of an economy.
- Key concept: Revealed preference in observed payment shares (Pi) over time is used in place of explicit demand determinants; changes are observed only if beneficial attributes (convenience, lower cost, greater safety, faster speed, less risk, or other positive payment attributes) drive users/receivers to switch.

### Payment aggregation and exclusions
- Total payments (TP) are defined as the gross value of all inputs that comprise value added GDP plus final demand and are equated to the sum of payment instrument values: TP = ∑i Pi.
- Payment instruments Pi used in the model:
  - 1. cash
  - 2. debit and credit cards (including e-money)
  - 3. instant payments (IP)
  - 4. Checks and direct debits (ChecksDD)
- Wire transfers are excluded as they are assumed to be functionally separable and have been deleted from both sides of TP = k(∑i Pi).

### Translog production function specification
- Translog quadratic function specified in logarithmic form:
  - (1) lnTP = lnαo + ∑i αi lnPi + 1/2∑i ∑j βij lnPi lnPj + ε
- Interpretation:
  - ∂lnTP/∂lnPi gives the elasticity of total payment output with respect to payment input i; elasticities expected to be overall positive.
  - Example: in the cash equation, the contribution of cash (lnP1) is expected positive (β11 > 0); negative cross coefficients (e.g., β12 < 0 or β13 < 0) indicate substitution of cards or instant payments for cash.
- Practical note on zeros: when a payment value may be zero (e.g., instant payments before issuance or disappearance of checks), a small positive value is inserted to take natural logs.

### Marginal/elasticity link to shares and estimation approach
- Under profit-maximization analogy, payment elasticities ∂lnTP/∂lnPi are replaced with observed shares Si (the observed share of Pi in TP).
- System of four equations estimated jointly using a stacked seemingly unrelated regression estimation (SURE) framework with coefficient symmetry (βij = βji) imposed:
  - S1, S2, S3, S4 each related to lnP1, lnP2, lnP3, lnP4 and their pairwise interactions as per equation system (3) (stacked 4 x 14 matrix post-multiplied by coefficient vector).
- Imposed symmetry implies reciprocal substitution effects are equal (e.g., β12 = β21).

### Linear ad hoc system
- A simpler ad hoc linear equation system (no logs) replaces shares with payment levels Pi:
  - System (4) specifies P1, P2, P3, P4 as dependent on levels of the other Pj (i ≠ j) with symmetry imposed and is estimated in a stacked SURE regression.
- Interpretation of coefficients:
  - βij < 0 indicates payment substitution.
  - βij > 0 indicates complementarity.

### Estimation details and robustness checks
- Estimation method: SURE (stacked SURE) to address contemporaneous correlation of errors and reduce multicollinearity across RHS variables.
- To deal with possible contemporaneous correlation, (3) is re-estimated with all RHS variables lagged by one year.
- Unit root and cointegration tests indicate non-stationarity of variables (variable means and variances are non-constant and vary with time).
- First-difference re-estimation was conducted for illustration; differencing likely lowers the chance of observing payment substitution by removing trending behavior.
- Degrees of freedom: df = (4 x n) - 14, where n = the sum of observations across all four equations together in (3) for 2012-2020 or 2005-2020.
- Implementation timing: instant payments were zero in 2005--the start of our sample--and most instant payment arrangements were implemented after 2011.

### Interpretation and empirical considerations
- The marginal rate of payment substitution between two payment types (holding others constant) is reflected in the sign and value of estimated βij coefficients.
- Coefficient symmetry is commonly applied in translog analyses and satisfies integrability conditions (Young's Theorem); it is testable but rarely violated empirically in this literature.
- Estimating in levels (system (4)) versus logs/quadratic (system (3)) provides complementary views: (3) captures elasticities via shares and log interactions; (4) shows level relationships among payment values directly.

*Source: wpiea2022228-print-pdf - 5.   A Measure of Payment Substitution Across Countries*

### 6.   Substitution Results

### 6.   Substitution Results

### Overview
- Six types of payment substitutions are possible for four classes of payments.  
- With twelve countries, there could be seventy-two (symmetric) substitutions.  
- Table 2 reports eighteen statistically significant payment substitutions for the full model (Eq. 3, preferred translog quadratic specification).  
- If an estimated parameter is negative and significant for at least one of the three models presented, the sign of other estimated parameters is reported even if not significant.  
- Significance levels: p-values .01 (**), .05 (*), and .10 (º).

### Statistically significant substitutions (summary of Table 2 and key table-derived points)
- Total statistically significant substitutions reported in Table 2 (2012–2020, three specifications): eighteen in the full model; alternative specifications show different counts (see below).
- Examples of significant instant payments (IP) substitutions (values are estimated parameter signs and magnitudes as reported):
  - IP replaces Cash:
    - China: Column (3) -.071º
    - Indonesia: Column (1) -.010**, Column (3) -.012**
    - Russia: Column (1) -.040**, Column (3) -.063*
    - Sweden: Column (2) -.006* (Column (3) + reported though not significant)
    - UK: Column (1) -.041**, Column (3) -.041**
  - IP replaces Cards:
    - China: Column (2) +, Column (3) -.069*
    - Indonesia: Column (2) +, Column (3) -.003*
    - Norway: Column (1) -.011**, Column (2) -.032**, Column (3) -.547º
    - UK: Column (1) -.062**, Column (3) -.500*
  - IP replaces ChecksDD:
    - Australia: Column (2) -.066**, Column (3) -.034º
    - China: Column (1) -.058**, Column (2) -
    - Indonesia: Column (1) -.006º
    - Mexico: Column (2) -.038**, Column (3) -.225**, Column (3 alternative?) -.074**
    - Singapore: Column (1) -.013º
    - Thailand: Column (2) -.011*, Column (3) +
    - UK: Column (1) -.120**, Column (3) -.109**
- Examples of cards replacing other instruments:
  - Cards replace Cash:
    - Mexico: Column (3) -.023**
    - Norway: Column (3) -.113*
    - Russia: Column (1) -.170**, Column (3) -.191**
    - Sweden: Column (1) +, Column (2) +, Column (3) -.042**
  - Cards replace ChecksDD:
    - India: Column (1) -.168º, Column (3) +
    - Mexico: Column (1) -.063**, Column (3) -.071**
    - Norway: Column (2) .205**, Column (3) -
    - Russia: Column (1) -.034º, Column (3) -.033º
    - Singapore: Column (2) -.365*
    - Sweden: Column (1) -.301**, Column (3) -.190**
    - UK: Column (1) -.127**, Column (3) -.101**
- ChecksDD replace Cash:
  - Mexico: Column (1) -.101**, Column (3) -.127**

### Model robustness, alternative specifications, and numeric comparisons
- Three specifications examined in Table 2:
  - (1) Full model (preferred translog quadratic).
  - (2) All RHS variables lagged by 1-year (to address contemporaneous correlation).
  - (3) First differencing of all variables (unit root and cointegration tests were rejected).
- Counts of statistically significant substitutions by specification:
  - Full model (Column 1): Table 2 lists eighteen statistically significant substitutions (as reported).
  - Lagged-1-year specification (Column 2): only six possible substitutions are statistically significant.
  - First-differenced specification (Column 3): twenty substitutions are significant; 13 duplicate the sign and statistical significance found with undifferenced data in the full model.
- Linear specification (equation (4)) results:
  - Only five payment substitutions were significant (not shown) and none involved instant payments.
  - Interpretation: the quadratic (translog) specification captures nonlinear substitution dynamics implied by logistic-like adoption and decline patterns; linear model fails to capture those features.

### Payment share changes and country-level comparisons
- Percentage point changes in payment shares (from Table 1 Columns 3 and 5–7) indicate change in each country's level of payment system efficiency and the degree that presumed benefits from payment types have been realized.
- Country examples (sum of positive vs negative changes in value shares reported in text):
  - Russia: sum of positive (negative) changes = +67 (-67) percentage points.
  - Norway: sum of positive (negative) changes = +14 (-14) percentage points.
- Country ranking of most to least change over the period (2005–2020, as reported):
  - Russia, China, India, UK, Indonesia, Mexico, Malaysia, Singapore, Thailand, Sweden, Australia, Norway.
- Aggregate greatest reductions and increases in percentage point payment shares (reported totals):
  - Greatest reduction: checks and direct debits (-390).
  - Next greatest reduction: cash (-105).
  - Largest increases: instant payments (+310) and cards (+143).
  - Overall total percentage point change in shares reported as +501/-502 (with some rounding error).
- Interpretation:
  - Instant payments exhibited the greatest increase in use.
  - Debit instruments (checks and direct debits) experienced the greatest reduction.

### Implications for retail CBDC
- Instant payments and retail CBDC share strong similarities when retail CBDC is operated and settled through a central bank:
  - Both provide instantly available good and final funds backed by a central bank.
  - Both could offer zero-user cost (as with Swish) or de minimis fee (as envisioned for FedNow).
  - Both can accommodate P2P payments and can be adjusted to enable wallet-to-wallet transfers without external terminals (CBDC-enabled mobile phones example: China).
  - Both could be modified to transfer funds anonymously within value limits.
- Policy implication noted:
  - If instant payments are structured as a feasible alternative to retail cash at the retail level, they may limit incentives to adopt retail CBDC, if issued.
  - The case for issuing retail CBDC has been found less compelling in some advanced economies (Australia, Singapore, UK) compared to emerging markets and developing economies where motivations include promoting financial inclusion, enhancing payment system efficiency, competition, security, resiliency, and cross-border payments.

### Additional contextual or data notes from the section
- Check use is close to or at zero (effectively equilibrium) in Norway, Russia, Sweden, and the UK; in India and Mexico, check/debit instruments remain in extensive use though falling.
- Sweden: despite anecdotal use of Swish to replace cash in P2P contexts, instant payments in Sweden over 2012–2020 do not appear to have significantly replaced cash in the preferred model; cash ATM withdrawals have been falling slowly since 2006 in the dataset, seven years before instant payments were implemented (Swish introduced in BiR in 2012).
- Sample and sensitivity:
  - Sample size per country is small; outliers (e.g., 2020 for one country viewed as an outlier by authorities) had little meaningful effect on reported results given R2 > .95.
  - Appendix Table 3 reports translog model results for the longer 2005–2020 period: instant payments would have significantly substituted for cash, cards, or ChecksDD eight times using the quadratic model (rather than ten) when using the longer period; early years include near-zero instant payment usage for many countries, motivating focus on 2012–2020.

### Key takeaways (policy-relevant findings)
- The translog quadratic model identifies significant payment substitution consistent with logistic adoption/decline patterns; instant payments are statistically significant substitutes for checks, cash, and cards in multiple countries during 2012–2020.
- Instant payments have experienced the largest increase in payment share (+310 percentage points aggregate), while checks and direct debits have experienced the largest aggregate decline (-390 percentage points).
- The emergence and adoption of instant payments improve payment system cost efficiency and reduce payment system risk; measuring substitution provides a means to monitor these policy objectives.
- Instant payments may reduce demand for a retail CBDC if they offer similar features (instant finality, low or zero user cost, P2P capability), implying that central banks should account for existing or planned instant payment capabilities when evaluating the case for retail CBDC issuance.

*Source: IMF Working Paper — Section 6. Substitution Results (2012–2020 analysis) — extracted from the supplied content.*

### References

### References (Instant Payments: Regulatory Innovation and Payment Substitution Across Countries — Working Paper No. WP/22/228)

### Central bank digital currencies (CBDC) and digital money
- Arauz, A., R. Garratt., and D. Ramos (2021), “Dinero Electronico: The Rise and Fall of Ecuador’s Central Bank Digital Currency”, Latin American Journal of Central Banking, 2: 1-10.
- Auer, R., H. Banka, N. Yaa Boakye-Adjei, A. Faragallah, J. Frost, H. Natarajan, and J. Prenio (2022), Central Bank Digital Currencies: A New Tool in the Financial Inclusion Toolkit?, World Bank Group and Bank for International Settlement Financial Stability Institute Insights on Policy Implementation, No 41, April.
- Duarte, A., J. Frost., L. Gambacorta., P. W. Wilkens., and H. S. Shin (2022), “Central Banks, the Monetary System and Public Payment Infrastructures: Lessons from Brazil’s Pix”, BIS Bulletin, No. 52, 23 March.
- Economic Affairs Committee (2022), “Central Bank Digital Currencies: A Solution in Search of a Problem?” House of Lords, January 13.
- Grym, A. (2020). “Lessons Learned from the World’s First CBDC”, Bank of Finland Economics Review, 8: 1-20.
- Hayashi, F., and Y. Lei Toh (2022), “Assessing the Case for Retail CBDCs: Central Banks’ Considerations”, Payment Systems Research Briefing, Federal Reserve Bank of Kansas City, May 26.
- Khiaonarong, T., and D. Humphrey (2022), "Falling Use of Cash and Demand for Retail CBDC", Working Paper, International Monetary Fund WP/22/27, February, 2022.
- Monetary Authority of Singapore (MAS) (2021), “A Retail Central Bank Digital Currency: Economic Considerations in the Singapore Context”, Economic Policy Group, November.

### Instant and fast retail payments, interoperability, and operating hours
- Bank of Thailand (BOT) and Monetary Authority of Singapore (MAS) (2022), “Thailand and Singapore Launch World’s First Linkage of Real-Time Payment System”, Joint Press Release, April 29.
- Bank for International Settlements Innovation Hub (BISIH) (2021), Nexus: A Blueprint for Instant Cross-Border Payments, July.
- Bech, M., and J. Hancock (2020), "Innovations in Payments", BIS Quarterly Review, March: 21-36.
- Committee on Payments and Market Infrastructures (CPMI) (2022), Extending and Aligning Payment System Operating Hours for Cross-Border Payments, Bank for International Settlements, May.
- Committee on Payments and Market Infrastructures (CPMI) (2021), Developments in Retail Fast Payments and Implications for RTGS Systems, Bank for International Settlements, December.
- Committee on Payments and Market Infrastructures (CPMI) (2016), Fast Payments--Enhancing the Speed and Availability of Retail Payments, Bank for International Settlements, November.
- Committee on Payment and Settlement Systems (CPSS) (2006), General Guidance for National Payment System Development, Bank for International Settlements, January.
- Hartmann, M., L. Hernandez-van Gijsel, M. Plooij, and Q. Vandeweyer (2019), Are Instant Payments Becoming the New Normal? A Comparative Study. European Central Bank, Occasional Paper Series, No. 229, August: 1-53.
- Lobo, B., and C. Brandt (2022), “Pix: The Brazilian Fast Payments Scheme”, Journal of Payments Strategy & Systems, 15 (4): 367-375.
- Lowe, P. (2021), “Payments: The Future?”, Address to the Australian Payments Network Summit 2021, Reserve Bank of Australia, December 9.
- Menon, R (2022), "Yes to Digital Asset Innovation, No to Cryptocurrency Speculation" Opening Address by Mr. Ravi Menon, Managing Director, Monetary Authority of Singapore, at Green Shoots Seminar on August 29, 2022.

### Payment instrument adoption, substitution, and consumer behavior
- Koulayev, S., M. Rysman, S. Schuh, and J. Stavins (2016), “Explaining Adoption and Use of Payment Instruments by U.S. Consumers”, Rand Journal of Economics, 47 (2): 293-325.
- Snellman, J., J. Vesala, and D. Humphrey (2001), “Substitution of Non-Cash Payment Instruments for Cash in Europe”, Journal of Financial Service Research, 19 (April): 131-145.
- Humphrey, D., and J. Moroney (1975), "Substitution Among Capital, Labor, and Natural Resource Products in American Manufacturing", Journal of Political Economy, 83 (February): 57-82.
- Koulayev, S., M. Rysman, S. Schuh, and J. Stavins (2016), “Explaining Adoption and Use of Payment Instruments by U.S. Consumers”, Rand Journal of Economics, 47 (2): 293-325.

### Methodology, production functions, and elasticities of substitution
- Berndt, E., and L. Christensen (1973), "The Translog Function and the Substitution of Equipment, Structures, and Labor in U.S. Manufacturing, 1929-1968", Journal of Econometrics, 1 (March): 81-114.
- Blackorby, C. and R. Russell (1989), "Will the Real Elasticity of Substitution Please Stand Up? (A Comparison of the Allen/Uzawa and Morishima Elasticities)", American Economic Review, 79(4):882–888.
- Bos, J. and M. Koetter (2011), "Handling Losses in Translog Profit Models", Applied Economics, 43 (3): 307-312.
- McFadden, D. (1999). "Chapter 5. Systems of Regression Equations", https://eml.berkeley.edu/~mcfadden/e240b_f01/ch5.pdf.
- Miller, N., J. Bergtold, and A. Featherstone (2019), "Economic Elasticities of Input Substitution Using Data Envelopment Analysis", PLoS ONE 14(8): e0220478. https://doi.org/10.1371/journal. pone.0220478

*Instant Payments: Regulatory Innovation and Payment Substitution Across Countries — Working Paper No. WP/22/228*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2022/english/wpiea2022228-print-pdf.pdf_