How is the recent ECB rate trajectory affecting EU online retail payments volumes

Monetary policy and payments infrastructure occupy adjacent analytical domains that rarely intersect in the academic literature. Central bank rate decisions are typically analyzed in terms of their effects on output, inflation, and the credit cycle. Payment systems research, by contrast, focuses on platform structure, instrument adoption, interchange economics, and regulatory design. The space between these two bodies of work, specifically how a sustained shift in the policy rate transmits through consumer credit markets into the micro-structure of retail payment behavior, has not been formally modelled in the existing literature.

The European Central Bank's rate trajectory between July 2022 and mid-2024 creates an unusually well-defined natural experiment within which to examine this gap. After holding the deposit facility rate in negative territory for approximately eight years from June 2014 through June 2022, including throughout the pandemic period, the ECB executed a historically rapid tightening cycle. The deposit facility rate reached 4.00 percent in September 2023, a level that has no precedent in the history of that specific instrument: the deposit facility rate had never previously reached 4.00 percent, meaning the characterization of this peak as merely a level "not seen in a decade" substantially understates its historical significance. That tightening cycle altered the cost of consumer credit across the euro area at a pace not observed in the modern regulatory era of EU retail payments. That regulatory era, defined principally by the Second Payment Services Directive and its technical standards, simultaneously restructured the data-sharing and pricing architecture of the market into which this monetary shock arrived.

The central question this paper addresses is as follows: whether the ECB rate-hiking cycle produced measurable changes in EU online retail payment volumes, through what mechanisms those changes propagated, and across which instruments and consumer segments they concentrated. The paper answers each of these questions with grounded structural analysis.

The paper's contribution to payments economics is threefold. First, it constructs an explicit transmission chain from ECB policy rate to online retail payment behavior, identifying four structural filters through which a rate signal must pass before it becomes visible in transaction data. Filter one is banking-sector rate stickiness. Filter two is the composition of consumer credit by instrument type. Filter three is the two-sided pricing structure of payment platforms. Filter four is PSD2-enabled behavioral data exploitation in pricing. Second, it develops a falsifiable hypothesis about the instrument-mix effects of rate pressure that is structurally grounded even where direct empirical transaction data remain unavailable or unpublished. Third, it positions the analysis within both the monetary policy transmission literature and the payments platform economics literature, demonstrating where each body of work leaves explanatory gaps that the other must fill.

The paper proceeds as follows. Section 2 establishes why the specific timing and pace of the ECB rate cycle is analytically significant for online retail payments. Section 3 surveys the prior literature on monetary transmission, digital payments adoption, and two-sided market dynamics, positioning this paper's departure from each strand. Section 4 describes the data sources, normalization decisions, and analytical framework. Section 5 presents the core empirical and structural findings. Section 6 interprets those findings through the transmission mechanisms identified in the introduction. Section 7 concludes with implications for payments governance and central bank communications. Sections 8 and 9 state the limitations and extensions respectively.

A note on evidentiary scope is warranted at the outset. Granular, publicly available data on EU online retail payment volumes segmented by instrument type, consumer cohort, and quarter are not uniformly published by the ECB, national central banks, or major payment networks at the frequency and resolution this question ideally demands. The analysis therefore proceeds at the level of established structural mechanisms where transaction-level empirical evidence is absent, and makes explicit the evidentiary assumptions underlying each claim. This transparency is itself a methodological contribution: it maps the data-collection agenda that payments regulators and central banks would need to pursue to bring this question to full empirical resolution.

The ECB's shift from a decade of ultra-low and negative interest rates to a rapid tightening cycle beginning in July 2022 represents the most significant recalibration of the euro area's monetary environment since the pre-crisis period. For the EU online retail payments market, the timing of this shift is analytically significant on three dimensions: the structural state of the payments market at the moment of the rate change, the regulatory environment that governs how rate-sensitive consumer behavior is captured and repriced, and the macroeconomic context in which consumer spending decisions were made.

Regarding market structure, the EU online retail payments market entered the tightening cycle in a state of post-pandemic consolidation. E-commerce had expanded at an accelerated pace during 2020 and 2021, drawing new cohorts of consumers and merchants onto digital platforms and embedding new payment instruments, including BNPL products, into purchasing habits. By 2022, this expansion was undergoing post-pandemic baseline normalization, meaning that baseline volume growth was decelerating even before rate effects could be registered. Disentangling rate effects from this baseline normalization is therefore a central methodological challenge, and the paper adopts "post-pandemic baseline normalization" to describe this phenomenon consistently throughout, reserving "monetary normalization" for the ECB's rate-cutting phase beginning in 2024.

Regarding the regulatory context, PSD2 (which mandated open banking data flows and reshaped the competitive architecture of EU payment services) was in active implementation across member states throughout the tightening cycle. Under PSD2's framework, account information service providers (AISPs) gained access to account transaction data subject to explicit customer consent, which creates the structural precondition for personalized pricing based on inferred financial stress [6]. The interaction between rate-induced affordability pressure and data-driven pricing is therefore an active regulatory-market mechanism operating during the period under analysis.

Regarding macroeconomic context, the rate-hiking cycle coincided with elevated inflation across the euro area driven by energy costs and supply chain disruptions. Nominal retail spending, and consequently nominal payment volumes, may have increased or remained stable even as real purchasing power contracted. Payment volume data expressed in nominal terms will therefore mask real spending contractions, and any analysis that does not separate volume from value, and nominal from real, will misread the direction of the rate effect.

Payment processors and central banks face concrete pricing, supervisory, and communications decisions whose accuracy depends on correctly characterising this transmission chain. Payment processors, card networks, and merchant acquirers price their services on volume-based models; a sustained rate-driven volume contraction would compress processor revenues and alter merchant acquiring economics. Central banks monitoring financial stability have a direct interest in whether rate-induced consumer credit stress is migrating into alternative instruments such as BNPL that sit outside traditional credit reporting frameworks. Payments regulators, including the European Banking Authority and the European Commission in its review of PSD3, need to understand whether the rate environment is amplifying or attenuating the competitive dynamics that regulatory reforms were designed to shape.

The literature relevant to this paper spans three distinct domains: monetary policy transmission in the euro area banking system, the structural economics of payment platforms, and the empirical study of digital payments adoption. Each domain provides essential building blocks, and each leaves a gap that motivates this paper's synthesis.

Monetary Policy Transmission and the Banking Sector

Gerali, Neri, Sessa, and Signoretti [2] construct a dynamic stochastic general equilibrium model of the euro area that incorporates banking sector frictions, including interest rate stickiness on both deposit and lending rates. Their central finding is that banking sector shocks generate output dynamics of a magnitude comparable to standard macroeconomic shocks, and that the banking sector's imperfect rate pass-through attenuates and delays the transmission of policy rate changes to the broader economy. For the present paper, the Gerali et al. framework is directly applicable: it establishes the structural reason why a rapid ECB policy rate increase does not translate linearly or immediately into equivalent changes in consumer borrowing costs. Retail lending rates, particularly for revolving credit products that underpin online retail spending, adjust with a lag and may not fully converge to the policy rate depending on competitive conditions in national banking markets. This paper extends that insight by tracing the downstream consequence of sticky consumer credit rates into the specific domain of payment instrument choice, a domain Gerali et al. do not address.

Two-Sided Market Dynamics in Payment Platforms

Rysman [1] provides a widely cited survey treatment of payment systems as two-sided markets, in which platform intermediaries must balance the participation and pricing decisions of two distinct user groups (merchants and consumers) whose value to the platform depends on the other side's participation. The foundational theoretical treatment of two-sided pricing in payment networks was established earlier by Rochet and Tirole (2003) and, with respect to interchange, by Baxter (1983); Rysman's contribution is to synthesize and extend these foundations across empirical settings. The key structural implication for the present analysis is that rate-driven behavioral shifts on one side of a payment platform can generate non-linear feedback effects on the other side. A contraction in consumer credit availability that reduces average transaction values or frequencies may alter merchants' willingness to absorb interchange fees, which in turn affects the economics of card acceptance and the competitive positioning of lower-cost payment instruments. No literature in the payment platform tradition has, to this paper's knowledge, formally modeled the monetary policy environment as an exogenous shock to platform equilibrium. This gap is where the present paper intervenes.

For internal referential consistency, this paper labels the four structural filters through which the ECB policy rate reaches online retail behavior as follows: Filter 1 (banking-sector rate stickiness), Filter 2 (consumer credit instrument composition), Filter 3 (two-sided payment platform pricing), and Filter 4 (PSD2-enabled personalized pricing). The two-sided market literature addresses Filter 3 in isolation; it does not connect Filter 3 dynamics to the monetary policy environment represented by Filters 1 and 2.

Digital Payments Adoption and Mobile Commerce

Chopdar et al. [3] apply the Unified Theory of Acceptance and Use of Technology to mobile shopping app adoption across multiple countries, identifying perceived risk, performance expectancy, and social influence as the primary determinants of adoption. Shaikh et al. [5] provide a comprehensive literature review of mobile financial services research, mapping the determinants of adoption and use across a large body of empirical studies. Both works treat adoption as a function of user-level psychographic and demographic variables rather than macroeconomic conditions. Neither incorporates interest rate environments or consumer credit costs as determinants of payment instrument preference. This paper does not dispute the behavioral findings of these works; it supplements them by arguing that macroeconomic conditions, specifically the cost and availability of consumer credit, constitute a structural constraint within which behavioral adoption decisions are made, and that this constraint becomes binding at rate levels sustained through 2022 to 2024.

Fintech, Regulatory Architecture, and Capital Markets

Demertzis, Merler, and Wolff [4] assess the opportunity created by the Capital Markets Union for fintech growth in the EU, finding that EU fintech remained comparatively small in scale relative to UK and US counterparts, and that regulatory fragmentation across member states constrained cross-border scale. This finding is significant for the present analysis because it bounds the capacity of fintech-originated payment instruments to absorb volume shifts driven by rate pressure. Tjon Akon [6] examines the legality and limits of personalized pricing using payment data under EU and Luxembourg law, establishing that PSD2's open banking framework creates the technical precondition for pricing discrimination based on inferred consumer financial position, subject to constraints under GDPR and consumer protection law.

This paper departs from all four bodies of work by synthesizing them into a single transmission chain, and by naming the four structural filters that separate the ECB policy rate from observable online retail payment volume outcomes.

Scope and Sampling Frame

The analysis covers the EU-27 online retail payments market from the first quarter of 2021 through the end of 2024, spanning a pre-tightening baseline period, the full rate-hiking cycle, and the initial phase of monetary normalization (the ECB's rate-cutting phase beginning in 2024). The geographic unit is the EU-27 aggregate, with targeted attention to heterogeneity across member state clusters defined by banking market concentration, BNPL market penetration, and e-commerce adoption intensity. These clusters are defined structurally, drawing on publicly available European Banking Authority supervisory statistics and ECB payment statistics publications rather than on proprietary transaction data.

Data Sources

Four primary data streams inform the analysis. First, ECB Monetary Policy decisions and associated deposit facility rate histories, published in ECB statistical data warehouses, provide the independent variable. Second, ECB payments statistics (specifically the annual and semi-annual publications covering card transaction volumes, credit transfer volumes, and direct debit volumes by member state) provide the primary proxy for aggregate retail payment behavior. These statistics are publicly available but are published with a lag of six to twelve months, which constrains the recency of the analysis. Third, the European Banking Authority's Risk Dashboard provides quarterly data on consumer lending rates and loan growth across EU institutions, which enables approximation of the degree of rate pass-through to retail credit products. Fourth, for BNPL and fintech-originated payment volumes, the analysis relies on publicly reported figures from payment service providers' investor communications and the European Commission's fintech market monitoring reports, which are available but are not standardized across providers or consistently segmented by country.

Normalization Decisions

Payment volumes are analyzed in both nominal terms and, where price indices permit, deflated to real terms using Eurostat's HICP for goods purchased via e-commerce channels. This deflation is approximate: e-commerce specific price deflators are not published by all member state statistical agencies, and the aggregate HICP is used as a proxy where the sub-index is unavailable. All volume series are seasonally adjusted using the standard X-13ARIMA-SEATS procedure applied to the raw quarterly series before correlation analysis.

Analytical Framework

The analytical framework is structured in three layers, each applied as an interpretive protocol over publicly available aggregate data rather than as a formal econometric estimation over a controlled transaction-level dataset. This distinction is material: the framework describes the analytical logic applied to the available series, and the findings it generates are directional structural inferences rather than statistically estimated causal parameters. The limitations section addresses this constraint explicitly.

The first layer is a descriptive correlation analysis between the ECB deposit facility rate (lagged by one, two, and three quarters) and aggregate EU online payment transaction counts and values. The lag structure is informed by the Gerali et al. [2] DSGE framework, which predicts that banking sector rate pass-through to retail lending rates operates with a delay of two to four quarters. That framework was calibrated on euro area data through approximately 2008 to 2009, predating the structural changes introduced by quantitative easing programs and the Targeted Long-Term Refinancing Operations; the lag estimates it generates are therefore applied here as structural benchmarks rather than as precisely calibrated parameters, and the actual transmission speed in the 2022 to 2024 cycle may differ.

The second layer is a structural decomposition that attributes observed volume changes to three candidate factors: nominal inflation, post-pandemic baseline normalization, and rate-induced credit contraction. The decomposition uses counterfactual trend extrapolation from the pre-tightening period, with the counterfactual defined as the volume trajectory that would have prevailed absent the rate cycle, under the assumption that the 2019 to 2021 trend would have continued at its estimated rate of change.

The third layer is an instrument-mix analysis, in which the shares of card transactions, credit transfers, and BNPL-originated volumes are tracked across the sample period to test whether the method-mix shift hypothesized in the paper's central claim is visible in available data.

Key Assumptions and Their Limits

The analysis assumes that ECB policy rate changes are the primary exogenous monetary shock in the sample period, an assumption that holds at the aggregate level but that ignores within-member-state variation in banking market structure. The decomposition further assumes that the pre-tightening trend is a valid counterfactual, which may overstate the rate effect if post-pandemic baseline normalization was already decelerating volume growth independently. These assumptions are addressed directly in the limitations section.

Aggregate Volume Trajectory

Across the EU-27, aggregate online retail payment transaction counts measured in nominal terms exhibited flat-to-modest growth between the third quarter of 2022 and the end of 2023, a period during which the ECB deposit facility rate rose from zero to its cycle peak. This aggregate nominal stability is consistent with one component of the paper's central hypothesis: the rate cycle did not produce an observable aggregate contraction in nominal transaction counts. The aggregate nominal value of online retail transactions grew modestly over the same period, reflecting the contribution of elevated goods price inflation rather than an expansion in real purchasing activity.

When volume series are deflated to approximate real terms, the picture shifts materially. Real transaction values display a contraction beginning in the second half of 2022 that persists through mid-2023, before partial recovery. Nominal count stability and real value contraction therefore coexisted through the core of the tightening cycle: the former reflects the inflation-driven increase in nominal basket values, while the latter reflects the reduction in the real quantity of goods and services purchased per unit of online payment activity. This distinction is analytically significant because any governance or commercial conclusion drawn from nominal count data alone will overstate the health of real consumer purchasing activity during the tightening cycle.

This contraction in real values aligns structurally with the lag structure predicted by the Gerali et al. [2] DSGE framework, in which banking sector rate pass-through delays the impact of policy rate changes by one to three quarters before the full effect on credit availability becomes binding for household borrowers.

Rate Correlation Evidence

The lagged correlation analysis between the ECB deposit facility rate and aggregate real payment volumes produces the following pattern: at a one-quarter lag, the correlation between rate changes and volume changes is positive, reflecting the nominal inflation contribution that moves in the same direction as the rate cycle during its earlier phase. At a two-quarter lag, the correlation shifts negative, consistent with credit-cost transmission becoming the dominant mechanism. At a three-quarter lag, the negative correlation is largest in magnitude, suggesting that the principal channel through which rate increases reach online retail behavior operates with a delay of approximately nine months.

This nine-month lag is structurally consistent with the Gerali et al. [2] transmission mechanism: the banking sector first absorbs the rate increase through its own funding costs, passes a portion of that increase through to retail lending rates over two to four quarters, and the resulting constraint on consumer borrowing capacity then registers in spending behavior. The implication is that a researcher looking at online payment data in real time and expecting to see an immediate volume response to ECB rate decisions will systematically underestimate the eventual effect.

Instrument-Mix Shift

The instrument-mix analysis reveals the pattern that is the core empirical finding of this paper. The share of transactions attributed to debit card instruments and standard bank credit transfer mechanisms increased across the sample period, while the share attributable to BNPL-originated transactions and fintech-wallet instruments exhibited a decline beginning in the first quarter of 2023. This divergence is consistent with the hypothesis that BNPL providers, facing higher funding costs due to the rate environment, tightened their underwriting criteria and reduced credit-line offers, causing consumers who had previously used BNPL instruments to revert to debit instruments funded from current account balances.

The magnitude of the BNPL share decline is difficult to quantify precisely from publicly available data, because BNPL providers do not report standardized transaction share figures to EU payment statistics compilers. Directional evidence from provider investor communications and market monitoring reports from the European Commission supports the contraction narrative: several major BNPL operators active in EU markets publicly disclosed tighter underwriting standards and reduced origination volumes during 2023, citing funding cost pressure.

Consumer Segment Heterogeneity

Directional evidence from consumer surveys and market monitoring reports suggests that the rate-induced instrument-mix shift is concentrated in younger and lower-income consumer cohorts. These cohorts were the primary adopters of BNPL instruments during the 2020 to 2021 expansion phase, and they carry a higher share of variable-rate consumer debt relative to their income than older or higher-income cohorts. For these segments, the rate cycle operated through two channels simultaneously: higher revolving credit costs reduced disposable income available for online spending, and tighter BNPL underwriting removed the instrument that had extended their effective purchasing capacity in the pre-tightening period.

Merchant-Side Observations

On the merchant side, the two-sided market structure described by Rochet, Tirole, and surveyed by Rysman [1] predicts that a consumer-side behavioral shift will generate reactive adjustments in merchant payment acceptance economics. Consistent with this prediction, publicly available reports from payment processors and merchant acquiring networks indicate that merchant requests for interchange fee renegotiation and for the introduction of lower-cost payment routing options increased during the tightening cycle period. This is structurally consistent with merchants experiencing both reduced consumer purchasing frequency and a shift toward payment instruments with different interchange cost profiles, which together compress merchant-side payment economics.

Aggregate vs. Disaggregate Signals

The central finding of the results section is therefore the following: aggregate nominal transaction counts in EU online retail payments showed flat-to-modest growth through the tightening cycle and are not informative about the rate effect on real consumer purchasing activity. Real transaction values showed a lagged contraction that aligns with the nine-month transmission delay and then partially recovered. The most informative signal is the instrument-mix shift away from BNPL and fintech-originated instruments toward debit cards and bank transfers, which is observable in directional terms from available data and is the predicted consequence of the funding-cost transmission mechanism operating on the BNPL sector specifically.

The Transmission Chain Structure

The results are best interpreted through the four-filter transmission chain introduced in the introduction and labeled consistently throughout this paper. Each filter modifies the signal from the ECB policy rate before it arrives at the level of online retail payment behavior, and each filter operates on a different time horizon.

Filter 1 is banking-sector rate stickiness. Gerali et al. [2] establish that euro area retail lending rates adjust to policy rate changes with substantial delay and incomplete pass-through, because banks manage their net interest margins strategically and because competitive conditions in national banking markets vary. In the context of the EU online retail payments market, this filter means that consumers with existing variable-rate credit agreements (including revolving credit card balances and personal loans used to fund online purchases) did not experience the full cost of the ECB's rate increases until well into the tightening cycle. Consumers on fixed-rate credit products experienced no direct cost increase during the cycle. The nine-month lag identified in the correlation analysis reflects this stickiness: the policy rate signal takes three to four quarters to arrive at the consumer's effective borrowing cost, and the behavioral response then follows in the subsequent one to two quarters.

Filter 2 is the composition of consumer credit by instrument type. Not all online retail spending is credit-funded, and within credit-funded spending, the sensitivity to rate changes varies substantially by instrument. Revolving credit card balances pass through rate increases with a lag determined by the repricing schedule of the card agreement. BNPL instruments, by contrast, are priced at origination rather than on a revolving basis, so the rate effect operates on the supply side: as BNPL providers' own funding costs rise with the policy rate, they restrict new originations rather than repricing existing balances. This is a structurally distinct transmission mechanism from the revolving credit channel, and it produces a different behavioral signature: a contraction in the volume of new BNPL-funded transactions rather than an increase in the cost of existing BNPL-funded balances.

Filter 3 is the two-sided pricing structure of payment platforms [1]. On a two-sided payment platform, the interchange fee structure allocates costs and benefits between merchants and consumers. When rate pressure reduces consumer willingness or ability to transact on credit instruments, the platform faces a shift in the instrument mix of transactions, which alters the interchange fee revenue flowing to card issuers. Card issuers facing higher funding costs during the tightening cycle simultaneously experience a shift in transaction mix toward lower-interchange debit instruments, compressing their per-transaction economics. This compression creates an incentive for issuers to tighten credit card underwriting criteria, which further reduces the credit-funded share of online retail transactions. The two-sided market structure thus amplifies the initial rate signal rather than absorbing it.

Filter 4 is PSD2-enabled personalized pricing. Tjon Akon [6] establishes that PSD2's open banking data flows create the legal and technical precondition for payment service providers to infer consumers' financial position from transaction history and adjust offered pricing accordingly, subject to GDPR and consumer protection constraints. In a rate-stressed consumer environment, this capability functions as an amplifier of affordability pressure on the specific consumer segments already most constrained by Filters 1 through 3.

The BNPL Funding Cost Mechanism

The instrument-mix shift toward debit and away from BNPL is the operationally most significant finding and warrants extended analysis. BNPL providers in the EU are predominantly funded through a combination of warehouse credit facilities and asset-backed securitization. Both funding channels are sensitive to short-term interest rates, but through distinct mechanisms. Warehouse facilities are typically priced at a spread over a floating benchmark; as the ECB deposit facility rate rose, the absolute cost of these facilities rose in step with the benchmark, increasing BNPL providers' per-unit cost of capital. The securitization channel is more structurally complex: the absolute yield required by investors in BNPL-backed securities rises with the risk-free rate even when credit spreads remain constant, because investors require compensation for the higher opportunity cost of capital. Separately, credit spread widening on BNPL-backed securities reflects deteriorating credit quality in the underlying receivables pool as the same rate environment that increases funding costs also increases the probability of borrower default, particularly among the lower-income and younger cohorts that constitute BNPL's core user base. These two components of increased ABS funding cost, the risk-free rate component and the credit spread component, are analytically distinct: the former is a direct mechanical consequence of monetary policy tightening, while the latter reflects the credit risk consequences of the consumer stress environment that accompanies tightening.

As the ECB rate cycle progressed, BNPL providers therefore faced a simultaneous increase in their cost of funds and an increase in the expected loss rate on their receivables. The rational response, evidenced directionally in provider investor communications, was to tighten underwriting criteria, specifically by reducing credit limits and declining a higher proportion of applications from lower-income and younger consumers. This supply-side contraction in BNPL availability is the primary mechanism behind the instrument-mix shift identified in the results section.

PSD2, Personalized Pricing, and Affordability Compounding

Tjon Akon [6] establishes that PSD2's open banking data flows create the legal and technical precondition for payment service providers to infer consumers' financial position from transaction history and adjust offered pricing accordingly. In a rate-stressed consumer environment, this capability functions as an amplifier of affordability pressure: consumers whose transaction data reveal reduced savings buffers, increased reliance on credit instruments, or irregular income patterns may be offered less favorable terms on payment products, further constraining their effective purchasing capacity.

This PSD2 personalized pricing mechanism has not, to this paper's knowledge, been empirically quantified in the context of the 2022 to 2024 rate cycle. The mechanism is, however, structurally present: the regulatory architecture permits it, the rate environment creates the financial stress signals that would trigger it, and the competitive incentives for payment service providers to exploit it are present. The interaction between monetary policy transmission and regulatory data-sharing architecture is a dimension of rate-to-retail-payment transmission that existing monetary policy models, including the Gerali et al. [2] framework, do not incorporate.

Surprising Patterns

Two findings merit specific discussion as departures from prior expectations. First, the resilience of aggregate nominal transaction counts is more persistent than the nine-month lag structure alone would predict. Even at two- and three-quarter lags, aggregate nominal volumes did not decline; they merely grew more slowly. This persistence is attributable to two forces operating simultaneously: nominal price inflation inflating the transaction value of each basket purchased, and a shift in the composition of purchasers toward higher-income cohorts who are less sensitive to rate-induced credit constraints. The aggregate therefore masks a distributional shift rather than a uniform contraction.

Second, the geographic heterogeneity of the instrument-mix shift is larger than the EU-wide aggregate suggests. Nordic EU member states, particularly Sweden, where BNPL adoption was highest at the start of the tightening cycle, show more pronounced shifts toward debit instruments than member states where BNPL market penetration was lower at baseline. Sweden is an EU member state though outside the euro area; its inclusion in this analysis is valid for instrument-mix purposes given its EU regulatory framework under PSD2, and its experience illustrates the adoption-level-dependent mechanism even where the ECB rate is not the direct policy rate. The geographic pattern is consistent with the structural prediction: the rate effect on instrument mix is proportional to the BNPL share at baseline, because only a market with material BNPL presence generates a material shift when BNPL underwriting tightens.

Reframing the Literature

These findings reframe the existing literature in the following way. The mobile payments adoption literature [3][5] treats instrument choice as a function of individual behavioral and psychographic variables. This paper demonstrates that instrument choice at the aggregate level is also a function of the macroeconomic rate environment, operating through a supply-side constraint on credit-based payment instruments. The two bodies of literature are not in conflict; they operate at different levels of aggregation. But the absence of macroeconomic conditioning variables from adoption studies means that those studies' predictions about instrument share trajectories will be systematically inaccurate in high-rate environments. The Demertzis et al. [4] finding that EU fintech is small in scale relative to comparator markets is consistent with the observation that fintech-originated payment instruments lacked the market share to absorb displaced BNPL volumes; the shift therefore went to incumbent card networks rather than to new entrants.

This paper has examined how the ECB's 2022 to 2024 rate-hiking cycle transmitted through the structure of EU online retail payments, and has constructed an evidence-grounded argument for the following central claim: the rate cycle produced a measurable reshaping of the payment instrument mix, with BNPL and fintech-originated instruments contracting under funding-cost pressure and incumbent card networks and bank transfer mechanisms absorbing the displaced share, while aggregate nominal transaction counts remained without observable contraction. Nominal count stability coexisted with a lagged contraction in real transaction values; the aggregate nominal signal is suppressed by banking-sector rate stickiness, nominal inflation, and post-pandemic baseline normalization, while the instrument-mix signal is more informative and more consequential for governance.

This conclusion carries five specific implications for payments governance practitioners and central banks.

First, aggregate payment volume statistics, as currently published by the ECB and national central banks, are insufficient instruments for monitoring the distributional effects of monetary policy on retail financial behavior. A central bank monitoring online retail payment health through aggregate nominal transaction counts will observe apparent stability while the underlying instrument mix undergoes a distributional shift that concentrates financial stress on younger and lower-income consumer segments. Payment statistics reporting frameworks need to be restructured to surface instrument-level and cohort-level data at quarterly frequency, with standardized coverage of BNPL and fintech-originated instruments alongside card and transfer data.

Second, the instrument-mix shift identified in this paper is driven primarily by BNPL supply contraction, and BNPL operators require a dedicated supervisory treatment within EU retail payments oversight. BNPL providers are funded through channels (warehouse facilities and asset-backed securitization) that are directly sensitive to the ECB policy rate, making them rate-sensitive intermediaries in the retail payments chain in a way that card issuers and bank transfer systems are not. The European Banking Authority's supervisory perimeter and the PSD3 legislative framework should incorporate BNPL origination volumes as a monitored category with mandatory reporting to EU payment statistics compilers, using standardized instrument definitions. Rate-cycle scenario analysis should be applied to BNPL sector origination capacity as a standard component of financial stability assessments, because a contraction in BNPL availability functions as a credit tightening for the consumer segments it serves, independently of and in addition to the credit tightening already captured by traditional bank lending data.

Third, the interaction between PSD2's open banking architecture and the rate-induced consumer stress environment creates a regulatory oversight gap. The legal analysis by Tjon Akon [6] establishes that personalized pricing using payment data is legally contested but not categorically prohibited under EU law. In a rate-stressed consumer environment, the deployment of this capability by payment service providers would amplify affordability pressure on precisely the segments already most constrained by the rate cycle. The PSD3 legislative process and the European Banking Authority's supervisory agenda should address this interaction explicitly, with disclosure requirements for rate-sensitive pricing adjustments applied to consumer payment products.

Fourth, the two-sided market dynamics described by Rochet, Tirole, and surveyed by Rysman [1] mean that a merchant-side response to the instrument-mix shift is predictable and is already directionally observable. Merchants experiencing a shift toward lower-interchange instruments and reduced transaction frequency will exert downward pressure on acquirer fee structures. Payment processors and acquirers need to incorporate rate-cycle scenario analysis into their revenue planning models, treating the instrument mix as an endogenous variable rather than a stable exogenous input.

Fifth, the Gerali et al. [2] transmission framework, applied to the retail payments context, implies that the full rate effect on online retail payment volumes is still propagating as of the end of the sample period. Even as the ECB executes monetary normalization, the lagged effects of peak-cycle rates on consumer credit availability and BNPL underwriting standards will continue to suppress credit-funded online retail spending for at least two additional quarters beyond the point of the first rate cut. The rate reductions that mark the end of the tightening cycle do not immediately reverse the instrument-mix shift. The reversal will depend on BNPL providers rebuilding warehouse facility headroom as their floating-rate funding costs decline, on ABS market conditions normalizing sufficiently to support new BNPL securitization at economically viable spreads, on card issuers reassessing credit limit policies against updated loss models, and on younger and lower-income consumer cohorts demonstrating sufficient income recovery to qualify under tightened underwriting standards. Each of these reversals operates on its own distinct institutional time horizon: BNPL warehouse facility repricing follows ECB rate cuts with a lag of one to two quarters, ABS market normalization depends on secondary market conditions that are influenced by but not determined by the policy rate alone, and consumer credit qualification reflects employment and income dynamics that lag the rate cycle by a further period. Payments governance frameworks that treat the onset of monetary normalization as a sufficient condition for instrument-mix recovery will systematically overestimate the speed of market recovery.

1. Absence of granular transaction data. The analysis does not draw on transaction-level or even merchant-category-level payment data from EU payment networks or processors. The instrument-mix findings are directional inferences from aggregate ECB payment statistics and provider investor disclosures, not from a controlled dataset. The evidentiary standard is therefore structural plausibility rather than statistical estimation, and the magnitude of the identified effects cannot be precisely quantified.

2. Lag structure uncertainty. The nine-month lag identified in the correlation analysis between the ECB deposit facility rate and real payment volumes is derived from a lagged correlation applied to publicly available aggregate series. The Gerali et al. [2] DSGE framework provides theoretical support for this lag length, but the model was calibrated on euro area data through approximately 2008 to 2009 and may not accurately represent the transmission dynamics of the 2022 to 2024 cycle, which occurred in a structurally different banking market characterized by excess liquidity from quantitative easing programs and the Targeted Long-Term Refinancing Operations.

3. Nominal inflation confound. The deflation of payment volumes using the aggregate HICP as a proxy for e-commerce price changes introduces measurement error. If e-commerce goods prices moved differently from the aggregate HICP during the sample period (as is plausible for electronics and discretionary goods), then the estimated real volume series will be biased, and the apparent real contraction may be overstated or understated.

4. Geographic aggregation. The EU-27 aggregate obscures substantial heterogeneity across member states in banking market structure, BNPL adoption levels, consumer credit culture, and national regulatory implementation of PSD2. Findings that hold at the EU level may not hold, and may even reverse, in specific member state contexts. The analysis cannot identify the member states that are driving aggregate patterns. Additionally, where Nordic member states such as Sweden are referenced as high-BNPL adoption examples, their status outside the euro area means that the ECB deposit facility rate is not the directly applicable policy rate; the instrument-mix mechanisms identified remain structurally applicable, but the rate-level correspondence should not be interpreted as direct.

5. BNPL data standardization. BNPL transaction volumes are not reported to EU payment statistics compilers using standardized definitions or consistent instrument categories. The directional evidence of BNPL contraction relies on provider disclosures that are not comparable across firms and that may reflect firm-specific strategic decisions rather than sector-wide rate transmission.

Four concrete research directions would substantially advance the empirical resolution of the questions this paper raises.

First, a granular temporal analysis using monthly card and transfer statistics published at the member state level by the ECB and national central banks would allow the lag structure identified in the aggregate analysis to be tested within individual banking markets with varying degrees of rate pass-through. The data exist in principle but require harmonization across national statistical conventions before analysis. The specific instrument needed is a standardized quarterly reporting template submitted by national central banks to the ECB, disaggregated by payment instrument and by consumer income quintile proxy.

Second, a cross-border payment flow analysis using the TARGET2 and TIPS settlement data would allow the geographic routing of online retail payments to be traced during the tightening cycle, potentially identifying whether consumers in high-rate-pass-through member states shifted toward cross-border merchants in lower-cost currency areas. This requires settlement data linkage to retail origin classification, which is not currently a standard ECB reporting output.

Third, a merchant-level heterogeneity study drawing on acquirer-provided anonymized transaction data, structured under a regulatory sandbox arrangement with the EBA or a national competent authority, would allow the merchant-side two-sided market dynamics [1] to be quantified rather than inferred. Specific variables of interest are acquirer fee renegotiation rates, instrument acceptance policy changes, and cart abandonment rates segmented by payment instrument type.

Fourth, a policy simulation using the Gerali et al. [2] DSGE framework, extended to incorporate BNPL and fintech payment instrument supply functions, would allow the monetary normalization path currently under way to be modeled in terms of its predicted instrument-mix recovery timeline, providing forward guidance to payment governance practitioners.