Ricardian Equivalence and Household Consumption: Evidence from Administrative Banking Microdata

Ricardian Equivalence and Household Consumption: Evidence from Administrative Banking Microdata | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Ricardian Equivalence and Household Consumption: Evidence from Administrative Banking Microdata HELENA CHYTILOVA, PETR FREJLICH, NIKOLAOS D. BELESIS, ANDREAS E. FOUSTERIS This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9246758/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 6 You are reading this latest preprint version Abstract This study tests the Ricardian Equivalence in the Czech Republic using a unique proprietary dataset of real banking data from 2017 to 2023. Applying panel fixed-effects models and generalized method of moments estimation, we find a robust and economically significant negative effect of personal income taxation on household consumption. The results suggest that individuals do not offset tax-induced income changes by increasing savings, providing evidence against Ricardian behaviour. By linking macroeconomic perspectives with high-frequency banking data, the paper makes a novel contribution to fiscal policy research and offers relevant implications for the design of consumption-related tax policies. JEL Classification: H3, E6, H8 Ricardian Equivalence Taxation Real Banking Data Intertemporal Consumption Fixed Effects model Generalized Method of Moments I. Introduction This study examines a central implication of the Ricardian Equivalence framework by testing whether the Barro–Ricardian mechanism operates in practice, i.e., whether reductions in taxation lead consumers to postpone current consumption. Although this question lies at the core of fiscal theory, its empirical identification remains difficult in real-world settings, where observed consumption and saving decisions are shaped by a wide range of macroeconomic, institutional, and behavioural factors. To address this challenge, the analysis draws on a novel proprietary dataset of client-level banking information, capturing detailed patterns of consumption, income, and savings behaviour, and combines it with macroeconomic indicators. To our knowledge, such a granular micro–macro integration has not yet been applied to assess Ricardian behaviour in this context. This unique data environment substantially enhances the precision and credibility of the results, enabling the study to provide new evidence on behavioural responses to fiscal policy that extends beyond conventional survey or aggregate data approaches. The applied panel-data methodology further strengthens the quantitative assessment of the research questions and supports the formulation of analytically sound and policy-relevant insights. Ricardian Equivalence remains one of the central theoretical benchmarks in the analysis of fiscal policy and government debt. In Barro’s ( 1974 ) formulation, government debt is neutral because forward-looking households internalize the intertemporal budget constraint of the state and offset debt-financed tax changes through adjustments in saving. Consumption decisions should therefore be unaffected by whether government expenditure is financed through taxes or borrowing (Barro, 1974 ; Mankiw, 2000 ). Yet the empirical validity of this proposition remains contested. While some studies report findings broadly consistent with Ricardian behaviour, including Seater ( 1993 ), Adji et al. ( 2009 ), Cadsby and Frank ( 1991 ), and Di Laurea and Ricciuti ( 2003 ), many others find evidence to the contrary, such as Bernheim (1987), Slate et al. ( 1995 ), Meissner and Rostam-Afschar ( 2014 , 2017 ), Mertens and Ravn ( 2013 ), Shapiro and Slemrod ( 2003 ), and Souleles ( 1999 ). As Esposito and Mastromatteo ( 2019 ) suggest, the relevance of Ricardian mechanisms may be conditional rather than universal. A major reason for these mixed findings is the difficulty of isolating Ricardian effects from confounding influences such as progressive taxation, political uncertainty, liquidity constraints, preference heterogeneity, and other behavioural or institutional frictions. This makes empirical validation methodologically demanding and increases the value of data that allow consumer behaviour to be observed directly rather than inferred indirectly from aggregate series. Against this backdrop, the present study exploits a unique, non-public banking panel dataset that enables a substantially more granular examination of consumer behaviour than is typically possible with aggregate or survey data. While the empirical setting is the Czech Republic, marked in recent years by rising public debt and the lingering effects of major economic shocks, the mechanisms under study are not country-specific. They speak more broadly to how households adjust consumption and saving decisions in response to fiscal policy across different economic environments. This context is empirically relevant. In recent years, Czech fiscal policy has faced pressures from the pandemic and global economic developments, resulting in a marked increase in public debt and persistent fiscal deficits. Government debt rose from CZK 1.64 trillion in 2020 to CZK 2.89 trillion in 2023, reaching 42.4% of GDP, with further increases projected. Budget deficits also widened sharply during the pandemic and, although partially reduced, remained elevated thereafter. These developments provide a useful macroeconomic setting for analysing how households respond to fiscal imbalances and tax-policy changes, while also reflecting broader fiscal dynamics observable in many advanced and emerging economies. Within this setting, the Barro–Ricardian framework offers a valuable lens for assessing whether households internalize future fiscal burdens. If consumers fully anticipate future taxation implied by current deficits, fiscal policy should have limited effects on aggregate demand. If not, tax and spending measures may exert substantial real effects over the business cycle. Assessing the empirical validity of Ricardian Equivalence is therefore important for understanding the effectiveness of fiscal stabilization policies and the design of countercyclical interventions. Building on this motivation, the present study advances the empirical assessment of Ricardian Equivalence by moving beyond the aggregate macroeconomic indicators typically used in earlier work (e.g., Bernheim, 1987; Ricciuti, 2003 ; Lucke, 1998 ). Instead, it draws on a proprietary micro-level dataset of Czech banking clients that captures detailed consumption, income, and savings behaviour and links these measures to macroeconomic variables. The period 2017–2023 provides a suitable observation window for analysing consumer responses within the banking environment alongside real macroeconomic developments. Importantly, it includes the substantial personal income tax cut implemented in 2021, creating a natural setting in which to examine how fiscal policy changes translate into individual consumption decisions. The panel-data analysis employs fixed effects (FE) and the generalized method of moments (GMM), both of which enhance the robustness and precision of the results. This combination of unique data and advanced methods constitutes the main empirical contribution of the study. The remainder of the paper proceeds as follows. The next section reviews the theoretical foundations and empirical evidence on Ricardian Equivalence, drawing on both experimental and macroeconomic studies. The subsequent sections present the data, model specification, and empirical results, followed by a discussion of the main findings and their policy implications. The paper concludes with a summary of the key results.. II. Literature Review Ricardian Equivalence, much like the Modigliani–Miller theorem in corporate finance, serves as a theoretical benchmark for assessing when fiscal policy becomes neutral (Elmendorf and Mankiw, 1999 ). Under its core logic of debt neutrality, a debt-financed tax cut does not raise consumption if forward-looking households anticipate the implied future tax burden and increase savings accordingly. In that case, aggregate consumption, savings, and investment remain unchanged, contrary to Keynesian predictions. The modern formulation is associated with Barro ( 1974 ), who formalized the idea in an intergenerational framework, while Buchanan ( 1976 ) linked it to Ricardo’s earlier reflections on wartime finance (Ricardo, 1951 ). Related contributions by Tobin ( 1952 ), Bailey ( 1962 ), and Patinkin ( 1965 ) further broadened the analysis of public debt. Its empirical support, however, remains mixed. For example, the 1964 U.S. tax cut increased consumption, whereas the 1968 tax increase produced only a limited response, suggesting heterogeneous household behaviour (Mankiw, 2016 ). Moreover, Ricardian neutrality does not imply that all fiscal policy is irrelevant: if deficits are expected to be offset by lower future spending rather than higher taxes, perceived lifetime income may rise and consumption may increase (Elmendorf and Mankiw, 1999 ). Despite its analytical appeal, the theory rests on strong assumptions that are widely contested. Critics point to limited foresight (Samuelson, 1974 ), behavioural biases (Thaler and Sunstein, 2008 ), market imperfections such as credit constraints and nominal rigidities (Stiglitz, 2000 ; Mankiw, 2000 ), and alternative views on debt sustainability (Kelton, 2020 ). These critiques underscore the restrictive nature of the assumptions underlying Ricardian equivalence and motivate the need for empirical validation. In this context, although still limited, experimental research on Ricardian equivalence is valuable because it isolates key assumptions in controlled settings, making deviations from Ricardian predictions easier to identify. Macroeconomic empirical studies complement this evidence by testing the same mechanisms in real-world fiscal and behavioural environments. Cadsby and Frank ( 1991 ) were among the first to investigate Ricardian equivalence in laboratory conditions and provided early evidence of its partial validity. Using an overlapping-generations framework, they incorporated the welfare of the second generation into the utility of the first, thereby directly testing the intergenerational linkage central to the Ricardian model. Participants in the parent generation received government debt, which either they or their children would ultimately bear through future interest or principal repayment. Under strict Ricardian equivalence, parents should save the full value of the debt as a bequest. The experiment, conducted with economics students at the University of Guelph, showed that when intergenerational transfers were strong, parents saved nearly the full amount of the debt. When such linkages were weaker, however, the offset was incomplete. The results thus suggest that Ricardian behaviour depends critically on the strength of intergenerational altruism. Building on this focus on intergenerational mechanisms, Slate et al. ( 1995 ) examined the role of uncertainty and found that Ricardian behaviour weakens when households are uncertain about debt repayment. Using a similar overlapping-generations design, they varied the probability of government bond repayment across treatments. Intergenerational transfers increased with the likelihood of repayment, as Ricardian theory predicts, but at lower probabilities participants consumed more and saved less. Their findings indicate that uncertainty about future fiscal obligations reduces the likelihood of Ricardian behaviour. Extending this line of inquiry, Di Laurea and Ricciuti ( 2003 ) tested the effects of liquidity constraints and income uncertainty. Their experiment compared a Ricardian control group with groups facing either liquidity constraints or uncertain income. While the control treatment was more consistent with Ricardian predictions, the presence of income uncertainty, and to a lesser extent liquidity constraints, led parents to leave insufficient inheritances to offset government debt. The study therefore rejects Ricardian equivalence under conditions that more closely resemble actual household decision-making. Focusing more specifically on institutional features of fiscal policy, Adji et al. ( 2009 ) examined whether Ricardian equivalence depends on tax structure. Using an overlapping-generations framework with altruistic intergenerational linkages, they compared a baseline treatment with non-distortionary taxes to one with distortionary taxes on savings. Under non-distortionary taxation, increases in government debt were largely offset by higher inheritance transfers, consistent with Ricardian predictions. Under distortionary taxation, however, the offset weakened and consumption became more uneven across periods. Their results show that the form of taxation is an important determinant of whether Ricardian behaviour emerges. While the previous studies emphasize intergenerational transfers and fiscal structure, Meissner ( 2013 ) shifted attention to individual intertemporal decision-making by examining consumption and saving in a laboratory life-cycle model that allowed borrowing. Participants faced either rising or declining income profiles over repeated life cycles. The results revealed strong behavioural asymmetries: participants were more averse to borrowing than to saving and deviated more from optimal decisions when borrowing was required. Individuals moving from a saving to a borrowing environment also encountered persistent difficulties despite repeated learning opportunities. These findings highlight debt aversion as an important behavioural friction in intertemporal choice. Building directly on this behavioural perspective, Meissner and Rostam-Afschar ( 2014 ) tested Ricardian equivalence in a modified life-cycle experiment with random income realizations and varying tax schedules. They found strong evidence against the theory: tax cuts increased consumption by 22%, while tax increases reduced it by 30%, and most participants did not behave in a Ricardian manner. In a follow-up study, Meissner and Rostam-Afschar ( 2017 ) introduced an additional treatment in which future tax schedules were fixed and fully disclosed to participants. Even under this greater fiscal transparency, Ricardian equivalence was still rejected, with a majority of participants deviating from the theory. Together, these studies suggest that tax changes meaningfully affect consumption even under highly controlled and information-rich conditions. Turning from experimental to macroeconomic evidence, a similar pattern of conditional and often non-Ricardian results emerges. Kormendi ( 1983 ), using U.S. data for 1929–1976, argued that government debt influences private consumption, thereby contradicting Ricardian neutrality. In line with this, Bernheim (1987) offered a broad critique of the theory and concluded that most empirical studies reject Ricardian behaviour, with deficits tending to stimulate consumption rather than leave it unchanged. Extending the analysis to a different institutional setting, Leiderman and Blejer ( 1988 ), using Israeli data from 1980 to 1985, found that although some theoretical elements of Ricardian equivalence were consistent with the data, substantial practical deviations remained, shaped by factors such as financial liberalization and economic openness. Further questioning the robustness of earlier findings, Feldstein and Elmendorf ( 1990 ) emphasized the fragility of empirical support for Ricardian equivalence. Reassessing studies including Kormendi ( 1983 ), they showed that results were highly sensitive to sample composition, particularly the inclusion of wartime observations. Consistent with these concerns, Campbell and Mankiw ( 1991 ) demonstrated, using quarterly data for several advanced economies, that a substantial share of households responds directly to current disposable income rather than smoothing consumption fully in line with permanent income. Shapiro and Slemrod (1995) reached a similar conclusion, finding that many consumers increase spending in response to tax cuts. These findings led to the development of macroeconomic models incorporating heterogeneous agents, distinguishing between Ricardian and rule-of-thumb consumers (Mankiw, 2000 ; Galí, López-Salido, and Vallés, 2004). Broadening the perspective to different economic environments, Khalid ( 1996 ) examined 21 developing countries over 1960–1988 and found predominantly non-Ricardian outcomes, attributing them to liquidity constraints, limited credit access, and weak substitutability between public and private consumption. In contrast, Lucke ( 1998 ), using German data from 1960 to 1994, identified a tension between theory and evidence: although the representative-agent model was rejected, aggregate outcomes still showed only modest deviations from Ricardian behaviour. Extending the analysis across countries, Cuaresma and Reitschuler ( 2007 ), in a panel study of EU-15 countries from 1960 to 2002, reported mixed results, with some countries exhibiting negligible fiscal multipliers under stricter fiscal conditions and others continuing to display non-Ricardian responses. Most recently, Georgarakos and Kenny ( 2022 ), using survey-based panel data from six euro area countries during the COVID-19 period, showed that more positive perceptions of fiscal support significantly increased household spending, particularly on durable goods. This finding is clearly inconsistent with Ricardian neutrality and suggests that fiscal policy effects may be especially pronounced in times of crisis. Taken together, the experimental and macroeconomic literature indicates that Ricardian equivalence holds, if at all, only under restrictive conditions and is highly sensitive to intergenerational linkages, uncertainty, liquidity constraints, tax design, and household heterogeneity. While the international literature offers valuable experimental evidence on Ricardian equivalence, especially Meissner and Rostam-Afschar ( 2014 , 2017 ), empirical studies based on high-resolution banking microdata are virtually absent. To our knowledge, this paper is the first to examine Ricardian behaviour using transaction-level data from a financial institution. Such data make it possible to observe actual consumption and saving decisions, including liquidity constraints, income dynamics, and heterogeneity in expectations, in a way experimental settings can only approximate. This helps address concerns about the external validity of laboratory evidence and allows Ricardian behaviour to be tested in a natural environment shaped by real financial constraints. In doing so, the paper bridges the gap between prior experimental research and real-world empirical analysis, providing a more comprehensive assessment of whether households behave in line with Barro’s Ricardian equivalence and when deviations are most likely to arise. III. Methodology and Model Specification We investigate the Barro–Ricardian hypothesis using a quarterly panel dataset spanning 2017–2023 that combines proprietary banking microdata from Raiffeisenbank (2024) with macroeconomic indicators. The dataset contains transaction-level information on household financial behaviour together with standard economic variables relevant for assessing responses to fiscal policy. The use of such granular, real-world data is rare in studies of Ricardian equivalence and allows for a more precise evaluation of household consumption and saving decisions in practice. Accordingly, the study addresses the following central question: Does the Barro–Ricardian equivalence principle holds, implying that lower taxation induces households to postpone consumption? Drawing on theoretical insights and previous empirical work, we construct a linear regression econometric model inspired by Bernheim (1987), Ricciuti ( 2003 ), and Lucke ( 1998 ): $$\:{CONSUMPTION}_{t}\:=\:{\beta\:}_{0}+\:{\beta\:}_{1}{WEALTH}_{t}+{\beta\:}_{2}{INCOME}_{t}-{\beta\:}_{3}{INTEREST\_SA}_{t}\:-\:{\beta\:}_{4}{INCOME\_TAX}_{t}{-\:\beta\:}_{5}{INFLATION}_{t}+{\beta\:}_{6}tGD{P}_{t}-\:{\beta\:}_{7}{DEF\_G}_{t}+{\beta\:}_{8}PUBLIC\_{DEBT}_{t}+dCOVID+dWAR\:{+\:\epsilon\:}_{t}\:$$ 1 In Eq. (46), which specifies the model used to test the hypothesis of this study, the dependent variable \(\:{CONSUMPTION}_{t}\:\) represents individual consumption expenditure at time t. The first explanatory variable, derived from proprietary data, is the individual’s income at time t \(\:{INCOME}_{t}\) . To account for the effect of total wealth, we include the variable \(\:{WEALTH}_{t}\) , defined as the total monetary balances accumulated across all of an individual’s bank accounts at time t. The explanatory variable \(\:{INTEREST\_SA}_{t}\) denotes the interest rate on the individual’s savings account at time t. A key variable for testing the hypothesis is \(\:{INCOME\_TAX}_{t}\) which measures the level of personal income tax at time t. Inflation is represented by the variable \(\:{INFLATION}_{t}\) while \(\:tGD{P}_{t}\:\) reflects GDP growth at time t. Additional macroeconomic explanatory variables include the government expenditure deficit-to-GDP ratio and the public debt-to-GDP ratio, expressed as percentages \(\:{DEF\_G}_{t}\) , \(\:PUBLIC\_{DEBT}_{t}\) . Finally, dummy variables are incorporated to capture the effects of the COVID-19 pandemic and the war in Ukraine \(\:dCOVID,\:dWAR\) . Individual consumption, personal income tax, and the government expenditure deficit-to-GDP ratio are central variables for testing the proposed hypothesis. Consumption serves as the dependent variable, while the direction and magnitude of the effects of personal income tax, the government deficit, and other explanatory variables constitute the core of the analysis. An individual’s income and total wealth, measured as savings across all bank accounts, are expected to influence consumption over time. These variables feature prominently in most consumption models and consistently show positive and significant effects on consumption in empirical research. According to Ricardian equivalence, personal income tax should have no effect on consumption, as tax changes are theoretically neutral. The interest rate on savings accounts is expected to encourage saving and therefore exert a negative influence on consumption. Inflation, measured by the CPI, is included as a control variable capturing precautionary savings behaviour. Variables such as GDP growth, the government expenditure deficit-to-GDP ratio, and the public debt-to-GDP ratio are incorporated to capture the influence of broader fiscal and macroeconomic conditions on individual consumption. GDP growth is expected to have a positive effect on consumption, and a similar positive relationship is anticipated for public debt if it signals increased economic activity or fiscal expansion. Table 1 summarizes all variables included in the model along with the expected signs of their effects. Table 1 Explanatory variables and their anticipated direction of effect on consumption WEALTH + INCOME + INTEREST_SA – INCOME_TAX – INFLATION – tGDP + DEF_G + PUBLIC_DEBT – dCOVID + dWAR + Source : Own elaboration, (2025). Existing empirical research yields mixed results, with studies testing Ricardian equivalence in real-world settings failing to reach a consensus on the theory’s validity. While Seater ( 1993 ) argues that empirical evidence is broadly consistent with Ricardian behaviour, many others report findings that contradict the theory (e.g., Bernheim, 1987; Lucke, 1998 ). These inconsistencies provide a key motivation for the present study. By analysing proprietary banking data, information not publicly accessible, together with macroeconomic indicators, this research offers a novel and more comprehensive perspective on Ricardian equivalence, extending the existing empirical literature in a direction not previously explored. Data To test the Ricardian equivalence hypothesis in the Czech context, we use quarterly panel data covering the period 2017–2023. This timeframe is particularly suitable because it includes the 2021 reduction in personal income tax, which provides a natural setting to examine whether lower taxation leads individuals to adjust their consumption in ways consistent with Ricardian behavior. The dataset combines proprietary banking microdata, capturing individual consumption, income, and savings, with macroeconomic indicators obtained from the Czech Statistical Office. Panel data offer several advantages over purely cross-sectional or time-series analyses, including richer information content, the ability to observe both individual heterogeneity and temporal dynamics, and improved reliability of parameter estimates. A key distinguishing feature of this study is the use of restricted-access banking data, which provide highly detailed information on the financial behaviour of anonymized clients that is not typically available for academic research. These unique proprietary data represent the consumer behavior of Raiffeisenbank clients. Because banking microdata are seldom accessible for research purposes, the dataset itself forms a significant part of the study’s contribution. The final dataset includes 21,885 randomly selected bank clients who meet predefined selection criteria and are observed throughout the sample period. This section describes and defines the variables used in the analysis. Descriptive statistics for all variables are presented in Table 2 . Table 2 Descriptive statistics of the dataset variables CONSUMPTION Mean Median Minimum Maximum 234,0 152,7 0,000 10 610 WEALTH 351,0 119,2 0,000 16 358 INCOME 255,3 173,5 45,00 3 039 INTEREST_SA 1,825 1,000 0,300 5,500 INCOME_TAX 17,29 19,00 15,00 19,00 INFLATION 5,746 2,950 1,900 17,60 tGDP 0,825 1,700 -10,80 4,700 DEF_G -2,303 -2,600 -9,500 3,310 PUBLIC_DEBT 39,16 39,94 30,80 45,20 dCOVID 0,250 0,000 0,000 1,00 dWAR 0,286 0,000 0,000 1,00 Notes : The variables CONSUMPTION, WEALTH, and INCOME are expressed in thousands. INTEREST_SA, INCOME_TAX, INFLATION, tGDP, DEF_GOV, and PUBLIC_DEBT are expressed as percentages. Source Own Elaboration, (2025). Individual consumption at time \(\:\text{t}\) is the dependent variable and represents a bank client’s total consumption expenditures. It aggregates quarterly expenditures made through the client’s current and credit card accounts. To ensure accuracy, transfers between the client’s own accounts, such as movements to savings, term deposits, other current accounts, or investment accounts, are excluded, as are credit card repayments to avoid double-counting. All values are expressed in thousands of CZK (Raiffeisenbank, 2024). Wealth at time \(\:t\) is a proprietary variable capturing the total financial resources available to a client. It is defined as the cumulative end-of-period balance across all accounts held within the institution, including current, savings, and term deposit accounts. This measure can be interpreted as an indicator of savings capacity. All monetary amounts are expressed in thousands of CZK (Raiffeisenbank, 2024). Income at time \(\:\text{t}\) is an explanatory variable representing the individual’s total quarterly income received into their current account. As with consumption, income figures are adjusted to exclude transfers between the client’s own accounts and are expressed in thousands of CZK (Raiffeisenbank, 2024). Personal income tax at time \(\:\text{t}\) captures the statutory income tax rate applied to employees. This variable is central to testing Ricardian equivalence in the Czech setting. Notably, the personal income tax rate fell from 19% to 15% in 2021, providing a natural experiment for assessing the impact of tax reductions on consumption (Czech Statistical Office, 2024). The interest rate on savings accounts at time \(\:\text{t}\) reflects the rate applied to an individual’s savings account and influences consumption and saving decisions. Using client-specific interest rates allows the model to capture the direct incentives faced by households (Raiffeisenbank, 2024). GDP growth rate at time \(\:\text{t}\) measures the year-on-year percentage change in quarterly GDP. This indicator provides insight into macroeconomic conditions, reflecting the pace of economic expansion or contraction and informing expectations about economic policy effectiveness and overall economic activity (Czech Statistical Office, 2024). Inflation rate at time \(\:\text{t}\) is represented by the annual percentage change in the Consumer Price Index (CPI), comparing the price level in the current month to the same month of the previous year. Inflation is included as a control variable linked to precautionary saving behavior (Czech Statistical Office, 2024). The government expenditure deficit-to-GDP ratio at time \(\:\text{t}\) measures the extent to which government expenditure exceeds revenue, expressed as a percentage of quarterly GDP. This indicator reflects fiscal policy stance and the degree of fiscal imbalance, offering insight into the potential long-term sustainability of public finances, (Czech Statistical Office, 2024). The total government debt-to-GDP ratio at time \(\:\text{t}\) captures the size of public debt relative to national economic output. A high ratio suggests increased fiscal pressures and potential macroeconomic risks. Within the context of Ricardian equivalence, this variable is relevant because the theory posits that government debt should not influence real economic activity if households fully anticipate future tax adjustments (Czech Statistical Office, 2024). Regression Analysis of the Main Models In the analysis of panel data, several regression techniques are employed to estimate parameters and test the proposed hypotheses. Before proceeding with the primary analysis of proprietary data to test our hypothesis, it was essential to conduct several preliminary tests and model estimations. These steps were critical in determining the appropriate estimation method and model specification, thereby ensuring the robustness and validity of our results. The correlation matrix confirmed a moderate relationship between key explanatory variables, particularly "wealth" and "income," To mitigate potential issues with autocorrelation and multicollinearity, we opted for the fixed-effects (FE) method and the generalized method of moments (GMM). Zero-order effect regressions highlighted income as the dominant predictor of consumption, while the influence of wealth was comparatively weaker. The Hausman test (χ²=5099.41, p-value < 0.01) rejected the null hypothesis, supporting the use of the fixed-effects model over the random-effects model due to significant differences in parameter estimates.Based on the results of the Hausman test, the fixed effects (FE) estimator was selected as the primary method. The FE approach incorporates individual-specific effects through unit-level intercepts, effectively addressing unobserved heterogeneity and mitigating potential endogeneity arising from omitted time-invariant factors. Its key advantage lies in its ability to control for unobserved characteristics that vary across individuals but remain constant over time, resulting in consistent and reliable parameter estimates. For comparison, the pooled ordinary least squares (pooled OLS) estimator treats the panel dataset as a single cross-sectional sample and applies standard OLS techniques. Although useful for preliminary or simplified analyses where cross-sectional heterogeneity is negligible, pooled OLS does not account for individual-specific effects and may produce biased estimates when heterogeneity or endogeneity is present. To strengthen the robustness of the results, we also apply the generalized method of moments (GMM), an advanced econometric technique capable of addressing challenges such as endogeneity, autocorrelation, and heteroskedasticity. GMM offers a flexible framework that accommodates various model specifications and moment conditions, making it particularly suitable for dynamic or complex panel structures. Its implementation, however, requires careful specification of valid instruments and high-quality data. Despite these demands, GMM produces robust and efficient parameter estimates. Notably, existing empirical literature has not yet applied GMM to tests of Ricardian equivalence, making its use in this study a novel methodological contribution. By integrating these complementary estimation techniques, the analysis achieves a high degree of robustness and precision, which is crucial for evaluating the hypotheses and assessing the influence of both proprietary banking microdata and macroeconomic factors on individual consumption behavior. This multifaceted econometric strategy provides a solid foundation for deriving reliable conclusions from the empirical results. Fixed Effects and Pooled OLS Estimation This section analyses the quarterly panel data using fixed effects (FE) and pooled ordinary least squares (pooled OLS) estimators to empirically test the hypothesis derived from Ricardian equivalence. As noted earlier, the FE approach provides consistent estimates by controlling for unobserved individual-specific characteristics and thereby addressing potential endogeneity. In addition, pooled OLS is employed as a complementary method to assess the magnitude and direction of relationships between the explanatory variables and the dependent variable, enhancing the robustness of the analysis. Both models are estimated using heteroskedasticity- and autocorrelation-consistent (HAC) robust standard errors to ensure reliable inference. The estimated coefficients of the regression models are reported in Table 3 . Table 3 Estimated coefficients of the regression model using the FE method with robust standard errors (HAC) and Pooled OLS with robust standard errors (HAC). Includes 21,885 cross-sectional units, time series length = 28 WEALTH FE Pooled OLS −0,12477*** (− 39,39) −0,02858*** (− 18,19) INCOME 0,67239*** (143,2) 0,75741*** (192,1) INTEREST_SA 0,29839 (0,7174) −1,51849*** (− 3,561) INCOME_TAX −6,12487*** (-16,15) −2,20539*** (− 5,678) INFLATION −0,73531*** (− 5,886) −0,34156*** (− 2,707) tGDP 1,00986*** (6,300) 1,31857*** (7,855) DEF_G 0,04275 (0,3914) −0,13199 (− 1,222) PUBLIC_DEBT 0,04275*** (− 22,88) −1,48437*** (− 15,42) dCOVID 24,7801*** (16,20) 15,3397*** (8,972) dWAR 53,0186*** (19,92) 39,5633*** (14,62) Constant 300,522*** (30,44) 135,083*** (14,84) Adjusted R 2 0,5812 0,5274 Notes : The dependent variable is observed consumption. T-statistics are provided in parentheses. Significance levels. *** p < 0,01; ** p < 0,05; * p < 0,1. Source Own Elaboration, (2025). According to the F-test, both models are statistically significant at the 0.01% level, and most explanatory variables are significant at the 1% level. The adjusted \(\:{R}^{2}\) indicates that the fixed effects (FE) model explains 58.1% of the variance in consumption, whereas the pooled OLS model explains 52.7%. As expected, the FE estimator provides higher explanatory power. The Durbin–Watson statistics 1.9505 for the FE model and 1.9551 for pooled OLS are close to 2, indicating no autocorrelation in the residuals, which is further supported by the very small autocorrelation coefficients (− 0.0184 and − 0.0188). The interpretation that follows focuses primarily on the FE estimates, with pooled OLS results serving as a complementary comparison. The coefficients for WEALTH and INCOME show a strong and significant influence on consumption in both estimation methods. The income coefficient (0.67239, significant at 1%) indicates that higher income leads to higher consumption. In contrast, the wealth coefficient (− 0.12477, significant at 1%) implies that individuals with higher accumulated balances tend to consume less. The effect of income is roughly three times larger than that of wealth, suggesting that individuals prefer to adjust consumption through current income rather than drawing on savings, reflecting precautionary financial behaviour. A crucial variable for testing Ricardian equivalence is personal income tax (INCOME_TAX). This coefficient is statistically significant at the 1% level in both models and negative in the FE estimates (− 6.1249). This indicates that reductions in personal income tax are associated with higher consumption, consistent with Keynesian consumption theory and inconsistent with Ricardian neutrality. According to Ricardian equivalence, tax reductions should have no effect on consumption because households would save rather than spend the additional disposable income. The significant negative coefficient clearly contradicts this prediction and confirms that bank clients increase consumption when the tax burden decreases. The interest rate on savings accounts (INTEREST_SA) is not statistically significant in the FE model but is significant and negative in the pooled OLS estimation. The sign of the coefficient is consistent with substitution effects: higher interest rates encourage saving and reduce consumption by increasing the opportunity cost of spending. Inflation (INFLATION) is statistically significant at the 1% level and has a negative coefficient (− 0.7353). Higher inflation reduces consumption because rising prices lower purchasing power, prompting households to restrain spending or reallocate funds to preserve value. This result supports the view that consumers are aware of inflationary pressures and adjust their behaviour accordingly. GDP growth (tGDP) is positive and significant at the 1% level (coefficient 1.0099). As economic conditions improve, consumers tend to increase their expenditures, consistent with the life-cycle hypothesis, which states that consumption responds to expectations about future income and economic prospects. The coefficients for the government expenditure deficit-to-GDP ratio (DEF_G) and the public debt-to-GDP ratio (PUBLIC_DEBT) show mixed results. DEF_G is statistically insignificant in both models, though its sign differs slightly. 1 PUBLIC_DEBT is positive and significant in the FE model, suggesting that higher public debt is associated with increased consumption, but it is negative and significant in the pooled OLS model (− 1.48437). This inconsistency indicates the need for further analysis using alternative techniques, which is addressed in the following chapter with GMM estimation. Both dummy variables COVID-19 (dCOVID) and the war in Ukraine (dWAR) are statistically significant at the 1% level and have positive coefficients. Average consumption was higher during these periods compared to baseline periods. The COVID-19 dummy has a coefficient of 24.7801, while the war dummy has a coefficient of 53.0186, suggesting that the war had more than twice the impact on consumption compared to the pandemic. These increases in consumption may stem from heightened uncertainty, increased government spending, or broader fiscal stimuli. This interpretation aligns with Keynesian fiscal policy and the multiplier effect, whereby government expenditure boosts aggregate demand during periods of economic instability. Our results for the COVID-19 period correspond with Georgarakos and Kenny ( 2022 ), who show that communication about fiscal support packages increased households’ expectations of future income and stimulated consumption, especially on durable goods. Like their findings, our results offer no evidence that Ricardian equivalence influences consumer behaviour during crisis periods; expected future tax liabilities do not appear to constrain consumption. In conclusion, income, wealth, personal income tax, inflation, GDP growth, and public debt all significantly affect consumption. Higher income and GDP growth increase consumption, whereas higher wealth, inflation, and in some cases public debt reduce it. Additionally, events such as the COVID-19 pandemic and the war in Ukraine exert substantial effects on consumption patterns. These findings highlight the importance of considering both macroeconomic context and household-level behaviour when evaluating Ricardian equivalence and designing fiscal policy. GMM Estimation To enhance the robustness of our conclusions, we re-estimate the main model using the generalized method of moments (GMM). GMM is well suited to complex panel data structures and offers reliable parameter estimates by addressing potential endogeneity, autocorrelation, and heteroskedasticity. Although the Durbin–Watson statistic and autocorrelation coefficient from the fixed effects estimates did not indicate autocorrelation, we employ GMM as an additional validation tool to reinforce the credibility of our findings. Its ability to correct for common econometric complications is particularly valuable when working with proprietary banking microdata, ensuring that our inference rests on a sound and resilient methodological foundation. The GMM framework requires carefully specified moment conditions, appropriate instrumental variables, and high-quality data. Panel data GMM is most implemented as Difference GMM or System GMM. While Difference GMM relies on first differences to remove unobserved fixed effects, System GMM combines moment conditions in first differences and in levels, thereby improving efficiency when the data exhibit persistence. Given the nature of our variables and the objective of achieving consistent and efficient parameter estimates, we adopt the System GMM approach. Both one-step and two-step variants of System GMM are employed. The one-step estimator computes the parameters in a single stage using the initial weighting matrix, which is constructed under the assumption of homoskedasticity. This makes the procedure computationally simpler but less efficient if heteroskedasticity or autocorrelation is present. In contrast, the two-step estimator proceeds sequentially: the first step produces preliminary parameter estimates, which are then used to construct a more robust weighting matrix for the second step. This second-stage estimation typically yields more efficient and accurate results, albeit at greater computational cost. Applying both variants enables us to assess the stability of the results across different specifications and strengthens the robustness of the empirical analysis. The System GMM estimations are conducted on a dataset containing 27,000 observations and 10,000 cross-sectional units. Due to computational constraints, the number of cross-sectional units had to be reduced from the original 21,885 to 10,000 through random sampling, consistent with the sampling strategy used in the initial dataset construction. A sample size of 10,000 cross-sectional units is generally considered sufficient for the GMM method. Several studies confirm that such a sample size provides reliable results when employing GMM. For instance, Han and Kim ( 2023 ) discuss the effectiveness of GMM estimates in their analysis, highlighting that even with a medium-sized sample, valid estimates can be achieved. Furthermore, the study by Hallett and Ma ( 1994 ) suggests that GMM estimates perform well even with smaller sample sizes than those used in our analysis. Hallett and Ma ( 1994 ) emphasize that a sample of just 1,000 cross-sectional units is entirely adequate for robust estimation. This reduction does not compromise the representativeness or validity of the results, as the remaining sample remains large, diverse, and statistically robust. Table 4 reports the estimated coefficients from the System GMM models and the factors influencing observed consumption. Table 4 One-step and two-step dynamic GMM panel model CONSUMPTION (− 1) One-Step GMM Two-Step GMM 0,02725* (1,829) 0,02794* (1,852) WEALTH −0,5139*** (15,65) −0,5165*** (15,72) INCOME 0,8635*** (− 5,487) 0,8647*** (− 5,539) INTEREST_SA −0,22579 (− 0,995) 0,17936* (1,906) INCOME_TAX −2,0595*** (− 5,724) −1,2155*** (− 8,418) INFLATION −0,1803 (− 1,522) −0,1224*** (− 2,908) tGDP 0,4431 (0,7081) 0,3253 (1,327) DEF_G −1,4693*** (− 3,417) 0,0658 (0,3208) PUBLIC_DEBT −0,0866*** (− 5,595) −0,0455*** (− 9,202) dCOVID 55,298*** (4,960) 45,349*** (7,810) dWAR 81,712*** (5,833) 64,617*** (8,433) Constant 630,64*** (6,024) 513,54*** (8,312) Number of Instruments 614 614 Test for AR (1) error z = − 6,229[0,000] z = − 6,23087 [0,000] Test for AR (2) error z = 0,636[0,524] z = 0,557997 [0,577] Sargan test Chi-square (1172) = 74175,3 [0,327] Chi-square (1172) = 3897,73 [0,276] Wald test Chi-square (11) = 2041,65 [0,000] Chi-square (11) = 3765,55 [0,000] Notes : The dependent variable is observed consumption. T-statistics are provided in parentheses. Significance levels. *** p < 0,01; ** p < 0,05; * p < 0,1. Source Own Elaboration, (2025). In both the one-step and two-step System GMM models, the regression results are broadly comparable. The first specification corresponds to the one-step estimator, while the second follows the two-step procedure. Both models explain consumption (CONSUMPTION) using the same set of explanatory variables. The Arellano–Bond autocorrelation tests confirm the absence of second-order serial correlation: the AR(1) test yields a p-value of 0.000, consistent with expectations for first-differenced errors, while the AR(2) test returns a p-value of 0.524, supporting the null hypothesis of no second-order autocorrelation. These diagnostics validate the appropriateness of the GMM framework. The Sargan test of overidentifying restrictions indicates that the instruments are valid, although the relatively high chi-square values suggest the possibility of instrument proliferation, a common issue when the number of instruments is large. The Wald tests confirm that both models are jointly significant. Additional requirements for valid GMM estimation are met: the inclusion of a lagged dependent variable and a number of instruments smaller than the number of cross-sectional units (614 < 10,000). The lagged consumption term (CONSUMPTION(− 1)) is statistically significant at the 10% level in both models, with coefficients of 0.0273 and 0.0279. These positive values indicate a modest degree of consumption persistence, suggesting that past consumption patterns have a small but significant effect on current behavior. Wealth (WEALTH) exhibits a negative and statistically significant effect at the 1% level in both models, with nearly identical coefficients (− 0.5139 and − 0.5165). Income (INCOME) remains positive and significant at the 1% level, with values of 0.8635 and 0.8647. Both results reinforce earlier findings: income increases consumption, whereas greater accumulated wealth is associated with reduced consumption. The coefficients for Personal Income Tax (INCOME_TAX) are negative and statistically significant at the 1% level in both GMM specifications, with magnitudes of − 2.0595 and − 1.2155. These results indicate that higher income taxes reduce consumption through lower disposable income, contradicting Ricardian equivalence and aligning with standard consumption theory. The Interest Rate on Savings (INTEREST_SA) is negative but not statistically significant in the one-step model. In contrast, in the two-step model it is positive and significant at the 10% level. These mixed results suggest that interest rates may influence consumption through both income and substitution channels. In standard consumption theory, the substitution effect typically dominates: higher interest rates raise the return to saving and the opportunity cost of current consumption, thereby discouraging expenditure. The positive coefficient in the two-step model is likely attributable to sensitivity to instrument choice and small-sample properties of the two-step GMM estimator, rather than reflecting a stable behavioural pattern. Inflation (INFLATION) is negative and insignificant in the one-step model but becomes negative and significant at the 1% level in the two-step model. This supports the view that rising inflation dampens consumption by reducing purchasing power and increasing uncertainty. GDP growth (tGDP) is positive but statistically insignificant in both models, implying that short-term fluctuations in GDP growth do not exert a measurable effect on household consumption within this framework. The Government Expenditure Deficit-to-GDP Ratio (DEF_G) is negative and statistically significant at the 1% level in the one-step GMM, but insignificant in the two-step model. The Total Public Debt-to-GDP Ratio (PUBLIC_DEBT), however, is negative and statistically significant at the 1% level in both estimations (− 0.0866 and − 0.0455), suggesting that higher public debt reduces consumption, likely due to increased uncertainty, lower disposable income or expectations of future fiscal adjustment. The dummy variables for COVID-19 (dCOVID) and the war in Ukraine (dWAR) are positive and significant at the 1% level in both models. The COVID-19 coefficients (55.298 and 45.349) and war coefficients (81.712 and 64.617) indicate that consumption rose significantly during these periods. These effects may reflect heightened uncertainty, fiscal support measures, or shifts in household spending patterns during crises. Overall, the one-step and two-step System GMM models provide consistent and meaningful insights into the determinants of consumption. Wealth, income, personal income tax, public debt, and the crisis-related dummy variables consistently exhibit statistically significant effects aligned with economic theory. Although certain variables, such as interest rates and the government deficit ratio, display variability across estimation methods, both GMM specifications appear well-specified and robust. To test our hypothesis, we estimated regression models using three complementary methods: fixed effects (FE), pooled ordinary least squares (pooled OLS), and System GMM. This multi-method approach provides a comprehensive assessment of the determinants of consumption and ensures the robustness of the empirical findings. F-tests confirm that the models are significant at the 1% level. The adjusted \(\:{R}^{2}\) indicates that the fixed effects model explains 82.2% of the variance, while the GMM model explains 79.5%, confirming the strong explanatory power of the specifications. IV. Results and Discussion Our results broadly reject the validity of Ricardian equivalivalence, consistent with the findings of earlier influential studies such as Feldstein (1982), Leiderman and Blejer ( 1988 ), Bernheim (1987), Ricciuti ( 2003 ), and Giorgioni and Holden ( 2003 ). The primary aim of our panel-data analysis was to test the hypothesis derived from Ricardian principles, with particular emphasis on the impact of the 2021 reduction in personal income tax on household consumption. In addition, we examined how other key determinants, including income, wealth, GDP growth, inflation, and fiscal variables, shape consumption behaviour. This approach enables a more comprehensive understanding of how taxation, income dynamics, and broader macroeconomic conditions interact to influence household spending. To facilitate interpretation, Table 5 reports the estimated coefficients from all three estimation methods. A detailed discussion of the effects of each explanatory variable on consumption, based on the fixed effects, pooled ordinary least squares, and generalized method of moments estimations follows the summary table. Table 5 The summary of the estimated coefficients of the main model using the methods applied in our analysis CONSUMPTION (− 1) FE Pooled OLS One-Step GMM Two-Step GMM / / 0,02725* 0,02794* WEALTH −0,12477*** −0,02858*** −0,5139*** −0,5165*** INCOME 0,67239*** 0,75741*** 0,8635*** 0,8647*** INTEREST_SA 0,29839 −1,51849*** −0,22579 0,17936* INCOME_TAX −6,12487*** −2,20539*** −2,0595*** −1,2155*** INFLATION −0,73531*** −0,34156*** −0,1803 −0,1224*** tGDP 1,00986*** 1,31857*** 0,4431 0,3253 DEF_G 0,042745 −0,131994 −1,4693*** 0,0658 dPUBLIC_DEBT 0,04275*** −1,48437*** −0,0866*** −0,0455*** dCOVID 24,7801*** 15,3397*** 55,298*** 45,349*** dWAR 53,0186*** 39,5633*** 81,712*** 64,617*** Constant 300,522*** 135,083*** 630,64*** 513,54*** Notes : The dependent variable is observed consumption. For the sake of clarity, t-statistics from the analysis of proprietary data are not included in this summary table, as was the case in previous instances. Significance levels. *** p < 0,01; ** p < 0,05; * p < 0,1. Income consistently exerts a positive and statistically significant effect on consumption across all models, including the FE and GMM estimations. This result aligns with Keynesian consumption theory, which predicts that higher disposable income increases household spending. Although the proprietary nature of our income variable limits direct comparability with prior studies, most of which rely on macroeconomic proxies, the findings are consistent with Kormendi ( 1983 ), Bernheim (1987), and Khalid ( 1996 ). Unlike these studies, which often rely on aggregate or imputed measures of disposable income, our analysis benefits from direct, high-frequency bank-level income data, representing an empirical contribution of the study. Wealth displays a negative and statistically significant relationship with consumption across the FE and GMM models. This relationship suggests that individuals with higher accumulated assets tend to consume less, potentially reflecting precautionary motives or a preference for maintaining financial buffers. Wealthier individuals may allocate a greater share of resources toward saving or investment rather than current consumption. Lucke ( 1998 ) finds a similar negative association using macroeconomic proxies for wealth, although his estimates lack statistical significance, underscoring the importance of using granular, account-level measures as in our analysis. Personal income tax exhibits a negative and statistically significant impact on consumption in both the FE and GMM models. Reductions in personal income tax increase disposable income and, consequently, consumption, consistent with Keynesian predictions. This pattern stands in contrast to the Ricardian equivalence hypothesis, which posits that tax reductions should be offset by increased saving in anticipation of future tax liabilities. Our findings align with empirical evidence from Leiderman and Blejer ( 1988 ), Bernheim (1987), Khalid ( 1996 ), and Cuaresma and Reitschuler ( 2007 ), which similarly reject the Ricardian neutrality proposition. The effect of interest rates on consumption varies across specifications. The FE model yields a positive and statistically significant coefficient, whereas the pooled OLS estimate is negative and statistically significant. In the one-step GMM model the effect is negative but insignificant, while in the two-step GMM model it becomes positive and significant at the 10% level. These inconsistencies likely reflect differences in sensitivity to interest rate variation across methodologies and potential estimation noise, rather than a stable behavioural relationship. Inflation exerts a negative effect on consumption, with statistical significance in the FE, pooled OLS, and two-step GMM estimates. Higher inflation erodes purchasing power and may induce precautionary behaviour, resulting in reduced consumption. This finding is consistent with theoretical predictions (Samuelson, 2009 ) and with empirical evidence from Dotsey and Sarte ( 2000 ) and Kikuchi and Nakazono ( 2020 ). GDP growth exhibits a positive effect on consumption in the FE and pooled OLS models, consistent with the notion that improved macroeconomic conditions support household expenditure. However, statistical significance is not confirmed in the GMM estimates, suggesting that the contemporaneous relationship between GDP growth and consumption may be weak once endogeneity and dynamic effects are accounted for. Nonetheless, the direction of the coefficients is consistent with prior empirical studies, including Alper ( 2018 ), Lettau and Ludvigson ( 2004 ), and Attanasio and Browning ( 1995 ). Government expenditure deficits display mixed effects. In the FE model the deficit ratio is insignificant, while in the one-step GMM model it is negative and statistically significant, suggesting that higher deficits may reduce consumption, potentially reflecting concerns and uncertainty over fiscal sustainability or future consolidation. This result offers limited support for Ricardian-type behaviour. Public debt exhibits divergent effects across specifications. In the FE model, public debt is positively associated with consumption, consistent with Keynesian views that debt-financed spending may stimulate economic activity. However, in both the one-step and two-step GMM models, the coefficient is negative and statistically significant, indicating that higher public debt reduces consumption, potentially due to heightened uncertainty, expectations of future fiscal tightening, or crowding-out effects. Given the robustness of the GMM estimates, the evidence suggests that elevated public debt dampens household consumption. These findings correspond with Bernheim (1987), Kormendi ( 1983 ), Leiderman and Blejer ( 1988 ), and Lucke ( 1998 ). The dummy variables capturing the COVID-19 pandemic and the war in Ukraine have positive and statistically significant effects across all models. Consumption increased during both periods, holding other factors constant. These patterns likely reflect shifts in expectations and heightened uncertainty, which may prompt households to adjust consumption in anticipation of future conditions. The results suggest that extraordinary macroeconomic events exert substantial influence on consumer behaviour, consistent with theories emphasizing the role of expectations and uncertainty. The findings also carry important implications for fiscal policy design. As emphasized by Esposito and Mastromatteo ( 2019 ), the absence of Ricardian equivalence underscores the need for governments to offset its absence and prioritize public spending that supports productivity, long-term growth, and financial stability. This is particularly relevant in economies where the financial sector has grown in size and complexity and where elevated public debt may heighten perceptions of macroeconomic risk. Public investment strategies should therefore extend beyond traditional infrastructure to include measures that reduce financial fragility and leverage, thereby mitigating risks associated with rising public indebtedness. Our results further highlight the importance of transparent and well-coordinated fiscal and monetary policies. When tax systems are complex or the communication of fiscal measures lacks clarity, uncertainty increases and can constrain effective household decision-making. Strengthening financial literacy and improving public communication about fiscal policy can help reduce this uncertainty, build trust in economic institutions, and enable households and firms to form more accurate expectations. As Esposito and Mastromatteo ( 2019 ) observe, no universal framework exists for managing high public debt; fiscal strategies must be tailored to national institutional and economic conditions. The finding that income tax reductions raise consumption in the Czech Republic illustrates that tax policy has immediate behavioural effects and thus must be assessed within a broader fiscal context. At the same time, persistent fiscal deficits and rising public debt may dampen consumption by increasing perceived economic risk and crowding out private investment. In such an environment, public investment targeted at strengthening financial resilience and reducing leverage becomes particularly important. More broadly, the failure of Ricardian equivalence in the Czech context implies that fiscal policy retains considerable scope for stabilizing economic activity. Government expenditure can support aggregate demand during downturns, while tax measures can be adjusted countercyclically to moderate inflationary or recessionary pressures. Although debt-financed fiscal expansions may stimulate short-run activity when households do not internalize future tax obligations, such policies raise concerns about intergenerational equity and underscore the importance of avoiding excessive debt accumulation. Additionally, behavioural tendencies such as bounded rationality, hyperbolic discounting, and mental accounting may also help explain why Ricardian equivalence breaks down in practice. Although we do not directly test these mechanisms, such behaviours can limit individuals’ capacity to assess future tax liabilities, encourage a preference for immediate consumption, and lead households to compartmentalize financial decisions in ways that hinder intertemporal optimization. These explanations complement more traditional Keynesian views, which emphasize liquidity constraints and a positive marginal propensity to consume as additional reasons why households do not behave as fully forward-looking agents. Taken together, these perspectives reinforce the need for transparent and clearly communicated fiscal measures: improving the salience and clarity of tax policy—as suggested by Chetty, Looney, and Kroft ( 2009 )—can help individuals better understand the long-term implications of fiscal decisions, reduce uncertainty, and support more informed economic behaviour. V. Conclusion This study examined the validity of Barro’s Ricardian equivalence in the Czech Republic by combining unique proprietary banking microdata with macroeconomic indicators in a quarterly panel covering 2017–2023. Using fixed effects (FE), pooled OLS, and generalized method of moments (GMM) estimators, we assessed whether reductions in personal income tax led individuals to defer consumption in anticipation of future tax liabilities. The motivation for this research derives from the longstanding empirical debate, in which evidence on Ricardian behaviour has been inconclusive despite the theory’s central role in fiscal policy analysis. Across all empirical specifications, our findings reject Ricardian equivalence. Consumers do not reduce current consumption in response to tax cuts; instead, lower taxes systematically increase consumption, consistent with Keynesian predictions and inconsistent with Barro’s neutrality proposition. This result is robust to the use of advanced GMM techniques, which address endogeneity, heteroskedasticity, and autocorrelation, and is reinforced by the unique advantage of our dataset: access to non-public, transaction-level banking information. The availability of granular income, wealth, and consumption data allows us to circumvent measurement problems that often limit macroeconomic studies and represents a distinctive contribution to the Ricardian equivalence literature. Our conclusions also align with earlier experimental work (e.g., Slate et al., 1995; Di Laurea and Ricciuti, 2003; Meissner and Rostam-Afschar, 2014, 2017) and empirical findings based on aggregate data (Feldstein, 1982; Bernheim, 1987; Khalid, 1996; Cuaresma and Reitschuler, 2007;Leiderman and Blejer, 1988; Ricciuti, 2003 and Giorgioni and Holden (2003)). The roles of wealth, public deficits, and public debt offer further insights into fiscal transmission. Wealth exerts a negative and significant effect on consumption, consistent with precautionary saving motives. Government expenditure deficits and public debt show negative and significant effects in the GMM estimates, suggesting that fiscal deterioration dampens consumption, possibly reflecting heightened uncertainty or expectations of future consolidation. These results support theories emphasizing debt-induced caution and partial crowding-out, consistent with findings by Bernheim (1987), Kormendi (1983), Leiderman and Bleier (1998), Lucke (1998) and Ricciuti (2003). Taken together, the evidence suggests that consumers respond meaningfully to fiscal variables, contradicting the behavioral invariance assumed under Ricardian equivalence. External shocks such as the COVID-19 pandemic and the war in Ukraine had unexpectedly strong positive effects on consumption. This pattern likely reflects shifts in expectations, increased uncertainty, and the influence of large-scale fiscal interventions. Our findings parallel those of Georgarakos and Kenny (2022), who document that expansive fiscal packages during COVID-19 raised household income expectations and stimulated spending. The evidence challenges Ricardian neutrality and suggests that fiscal policy can meaningfully influence household consumption. Finally, our study highlights the value of integrating micro-level financial data with macroeconomic indicators for empirical fiscal research. Such integration also indicates that fiscal policy should incorporate behavioural patterns that influence how households perceive and react to taxes and public spending, underscoring the importance of transparent, clearly articulated, and context-tailored measures for improving policy effectiveness. Future work could extend the present analysis by incorporating data from additional financial institutions, enabling richer identification strategies and enhancing the generalizability of the results. Such research would further deepen understanding of consumption dynamics and provide strong empirical foundations for fiscal policymaking in small open economies. Declarations Competing Interests: None financial or non-financial interests. The funding sources had no involvement in the study design, data collection, analysis, interpretation, writing, or the decision to submit the article for publication. Contributions All authors contributed to the study conception and design. Funding Declaration: This work was supported by the University of Piraeus Research Center and Program of the Ministry of Education and Religious Affairs of the Hellenic Republic, scientific project No. 16289 (IIS code, TA 5185218), research fellows Helena Chytilova, Andreas Fousteris, “Promoting Quality, Innovation and Extroversion in Universities, Axis 3.2: Strengthening the Digital Capabilities of Education and Modernizing Vocational Education and Training, Pillar 3: Employment, Skills and Social Cohesion,” in cooperation with the University of Piraeus, the Accounting and Auditing Laboratory of the Department of Business Administration. Author Contribution All authors contributed to the study conception and design.Helena Chytilova - Conceptualization; Methodology; Project administration; Writing – original draft; Writing – review and editing; Literature review; VisualizationPetr Frejlich - Conceptualization; Data Curation, Methodology; Formal Analysis; Software; Writing – original draftNikolaos D. Belesis – Conceptualization; Supervision; Project administration; InvestigationAndreas E. Fousteris - Conceptualization; Investigation; Data Curation Acknowledgement Acknowledgements: The authors gratefully acknowledge the academic environment and institutional support provided by their home universities, including the University of Piraeus, Škoda Auto University, and the Prague University of Economics and Business. This work was also supported by the University of Piraeus Research Center and Program of the Ministry of Education and Religious Affairs of the Hellenic Republic, scientific project No. 16289 (IIS code, TA 5185218), research fellows Helena Chytilova and Andreas E. Fousteris, “Promoting Quality, Innovation and Extroversion in Universities, Axis 3.2: Strengthening the Digital Capabilities of Education and Modernizing Vocational Education and Training, Pillar 3: Employment, Skills and Social Cohesion,” in cooperation with the University of Piraeus, the Accounting and Auditing Laboratory of the Department of Business Administration Data Availability The data used in this study consist of confidential banking records that are proprietary and protected under the EU General Data Protection Regulation (GDPR). Due to legal and contractual restrictions, these data cannot be shared publicly or made available to third parties. Although we would gladly provide access, we are not permitted to disclose the dataset in any form. Declaration of generative AI and AI-assisted technologies in the writing process. During the preparation of this work the author(s) used ChatGPT in order to improve language and readability. After using this tool/service, the author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the publication. References Adji, A., Alm, J. and Ferraro, P. J. (2009) ‘Experimental Tests of Ricardian Equivalence with Distortionary Versus Non-Distortionary Taxes’, Economics Bulletin, 29(4), 2556–2572. Alper, E. A. (2018) ‘The Relationship of Economic Growth with Consumption, Investment, Unemployment Rates, Saving Rates and Portfolio Investments in the Developing Countries’, Gaziantep University Journal of Social Sciences, 17(3), 980–987. Attanasio, O. P. and Browning, M. (1995) ‘Consumption over the Life Cycle and over the Business Cycle’, The American Economic Review, 85(5), 1118–1137. Bailey, M. J. (1962) ‘The Welfare Cost of Inflationary Finance’, Journal of Political Economy, 70(3), 244–258. Barro, R. J. (1974) ‘Are Government Bonds Net Wealth?’, Journal of Political Economy, 82(6), 1095–1117. Barro, R. J. (1996) ‘Reflections on Ricardian Equivalence’, Journal of Economic Perspectives, 10(1), 36–49. Buchanan, J. (1976) ‘Barro on the Ricardian Equivalence Theorem’, Journal of Political Economy, 84(2), 337–342. Cadsby, C. B. and Frank, M. (1991) ‘Experimental Tests of Ricardian Equivalence’, Economic Inquiry, 29(4), 645–664. Campbell, Y. J. and Mankiw, N. G. (1991) ‘The Response of Consumption to Income: A Cross-Country Investigation’, European Economic Review, 35(4), 723–756. Cuaresma, C. J. and Reitschuler, G. (2007) ‘Is the Ricardian Equivalence Proposition an “Aerie Fairy” Theory for Europe?’, Economica, 74(296), 682–694. Di Laurea, D. and Ricciuti, R. (2003) ‘An Experimental Analysis of Two Departures from Ricardian Equivalence’, Economics Bulletin, 8(11), 1–11. Dotsey, M. and Sarte, P. (2000) ‘Inflation Uncertainty and Growth in a Cash-in-Advance Economy’, Journal of Monetary Economics, 45(3), 631–655. Esposito, L. and Mastromatteo, G. (2019) Defaultnomics: Making Sense of the Barro-Ricardo Equivalence in a Financialized World. Working paper, No. 933. Feldstein, M. S. and Elmendorf, D. W. (1990) ‘Government Debt, Government Spending, and Private Sector Behavior Revisited: Comment’, The American Economic Review, 80, 589–599. Friedman, M. (1957) A Theory of the Consumption Function. Princeton: Princeton University Press. Georgarakos, D. and Kenny, G. (2022) ‘Household Spending and Fiscal Support during the COVID-19 Pandemic: Insights from a New Consumer Survey’, Journal of Monetary Economics, 129, S1–S14. Giorgioni, G. and Holden, S. (2003) ‘Political Institutions and the Stability of Government Debt’, Journal of Economic Surveys, 17(4), 487–516. Hallett, A. J. H. and Ma, Y. (1994) ‘On the Accuracy and Efficiency of GMM Estimators: A Monte Carlo Study’, in Belsley, D. A. (ed.) Computational Techniques for Econometrics and Economic Analysis. Advances in Computational Economics, 3, 22–44. Han, C. and Kim, H. (2023) ‘Dynamic Panel GMM Estimators with Improved Finite Sample Properties Using Parametric Restrictions for Dimension Reduction’, Empirical Economics, 64(6), 2589–2610. Chetty, R., Looney, A. and Kroft, K. (2009) ‘Salience and Taxation: Theory and Evidence’, The American Economic Review, 99(4), 1145–1177. Kelton, S. (2020) The Deficit Myth: Modern Monetary Theory and the Birth of the People’s Economy. New York: PublicAffairs. Keynes, J. M. (1936) The General Theory of Employment, Interest, and Money. London: Macmillan. Khalid, M. A. (1996) ‘Ricardian Equivalence: Empirical Evidence from Developing Economies’, Journal of Development Economics, 51, 413–432. Kikuchi, J. and Nakazono, Y. (2020) ‘Inflation Expectations and Consumer Spending: Micro-Data Evidence’. ESRI Discussion Paper Series, No. 353. Kormendi, C. R. (1983) ‘Government Debt, Government Spending and Private Sector Behaviour’, American Economic Review, 73, 994–1010. Krugman, P. R. (1988) ‘Financing vs. Forgiving a Debt Overhang’, Journal of Development Economics, 29(3), 253–268. Leiderman, L. and Blejer, M. (1988) ‘Modeling and Testing Ricardian Equivalence: A Survey’, IMF Economic Review, 35, 1–35. Lettau, M. and Ludvigson, C. S. (2004) ‘Understanding Trend and Cycle in Asset Values: Reevaluating the Wealth Effect on Consumption’, American Economic Review, 94(1), 276–299. Lucke, B. (1998) ‘Econometric Tests of Ricardian Equivalence: Results for Germany’, FinanzArchiv / Public Finance Analysis, 55(3), 288–314. Mankiw, N. G. (1999) ‘The Saver’s Paradox’, Journal of Economic Perspectives, 13(3), 13–24. Mankiw, N. G. (2016) Macroeconomics. New York: Worth Publishers. Mankiw, N. G. (2000) ‘The Savers-Spenders Theory of Fiscal Policy’, American Economic Review, 90(2), 120–125. Meissner, T. (2013) Intertemporal Consumption and Debt Aversion: An Experimental Study. SFB 649 Discussion Paper, No. 2013–045. Meissner, T. and Rostam-Afschar, D. (2014) Do Tax Cuts Increase Consumption? An Experimental Test of Ricardian Equivalence. Discussion Paper, No. 2014/16. Meissner, T. and Rostam-Afschar, D. (2017) ‘Learning Ricardian Equivalence’, Journal of Economic Dynamics and Control, 82, 273–288. Mertens, K. and Ravn, M. O. (2013) ‘The Dynamic Effects of Personal and Corporate Income Tax Changes in the United States’, American Economic Review, 103(4), 1212–1247. Modigliani, F. and Brumberg, R. (1954) ‘Utility Analysis and the Consumption Function: An Interpretation of Cross-Section Data’, in Kurihara, K. K. (ed.) Keynesian Economics. New Brunswick: Rutgers University Press. Patinkin, D. (1965) Money, Interest and Prices: An Integration of Monetary and Value Theory. 2nd edn. Evanston and New York: Row, Peterson and Co. and Harper and Row. Ricardo, D. (1951) ‘Funding System’, in Sraffa, P. (ed.) The Works and Correspondence of David Ricardo. Vol. 4. Cambridge: Cambridge University Press. Ricciuti, R. (2003) ‘Assessing Ricardian Equivalence’, Journal of Economic Surveys, 17(1), 55–78. Samuelson, P. A. (1974) ‘Theoretical Notes on Trade Problems’, Review of Economics and Statistics, 46(2), 145–154. Samuelson, P. A. and Nordhaus, W. D. (2009) Economics. New York: McGraw-Hill. Seater, J. J. (1993) ‘Ricardian Equivalence’, Journal of Economic Literature, 31(1), 142–190. Shapiro, M. D. and Slemrod, J. (2003) ‘Consumer Response to Tax Rebates’, American Economic Review, 93(1), 381–396. Slate, S., McKee, M., Beck, W. and Alm, J. (1995) ‘Testing Ricardian Equivalence under Uncertainty’, Public Choice, 85(1–2), 11–29. Souleles, N. S. (1999) ‘The Response of Household Consumption to Income Tax Refunds’, American Economic Review, 89(4), 947–958. Stiglitz, J. E. (2000) Economics of the Public Sector. New York: W. W. Norton & Company. Thaler, R. H. and Sunstein, C. R. (2008) Nudge: Improving Decisions About Health, Wealth, and Happiness. London: Penguin Books. Tobin, J. (1952) ‘A Survey of the Theory of Rationing’, Econometrica, 20, 521–553. Footnotes The negative impact on consumption suggests that a higher government expenditure deficit reduces consumer spending. An increase in the deficit may raise concerns about fiscal sustainability and decrease consumer confidence in the economy, leading to more cautious consumption expenditures. Additional Declarations No competing interests reported. Supplementary Files Appendix.docx Cite Share Download PDF Status: Under Review Version 1 posted Reviewers agreed at journal 01 Apr, 2026 Reviewers agreed at journal 31 Mar, 2026 Reviewers invited by journal 30 Mar, 2026 Editor assigned by journal 28 Mar, 2026 Submission checks completed at journal 28 Mar, 2026 First submitted to journal 27 Mar, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9246758","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":615331983,"identity":"fef6f77c-4d72-4ee1-92de-b438ae09ce52","order_by":0,"name":"HELENA CHYTILOVA","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA/klEQVRIiWNgGAWjYBACAwbmBiiDgeHAAyDBT1gLI0wLM8OBBCBDsoEULQwgLQYHCGgxZ29sfFxQwyBvLpF/8EBCxWF54xvpDx8w1NTh1GLZc7DZeMYxBsOdM5KBDjtz2HDbjRxjA4ZjbLgddiOxTZqHjSHB4AZQS2JbGuO2M2fYJBgbeHBruf+w/TfPP4QW+809x5//YGyQwGMLYxszbxtci03iBvYGM2CYGODWciaxWZq3T8Jww5nHBkC/2CTPON5jLJFwLAG3luOHD37m+WYjb3A88fGHDxUStv3N7A8/fMATYlAAdLkAsrm47UAG/AeIUjYKRsEoGAUjEAAA1wBYhNaRKZAAAAAASUVORK5CYII=","orcid":"","institution":"University of Piraeus","correspondingAuthor":true,"prefix":"","firstName":"HELENA","middleName":"","lastName":"CHYTILOVA","suffix":""},{"id":615331984,"identity":"93d2cc3f-1cf4-4207-8713-8992eb881005","order_by":1,"name":"PETR FREJLICH","email":"","orcid":"","institution":"Raiffeisenbank","correspondingAuthor":false,"prefix":"","firstName":"PETR","middleName":"","lastName":"FREJLICH","suffix":""},{"id":615331985,"identity":"c6b9c6c1-a8e0-4de8-976a-72d71e02115b","order_by":2,"name":"NIKOLAOS D. BELESIS","email":"","orcid":"","institution":"University of Piraeus","correspondingAuthor":false,"prefix":"","firstName":"NIKOLAOS","middleName":"D.","lastName":"BELESIS","suffix":""},{"id":615331986,"identity":"14011e4a-2a72-4435-9866-53d9abbff996","order_by":3,"name":"ANDREAS E. FOUSTERIS","email":"","orcid":"","institution":"University of Piraeus","correspondingAuthor":false,"prefix":"","firstName":"ANDREAS","middleName":"E.","lastName":"FOUSTERIS","suffix":""}],"badges":[],"createdAt":"2026-03-27 16:09:42","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9246758/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9246758/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":106094139,"identity":"f8a36487-a0cd-48ef-9c6a-42f69a7d7d6c","added_by":"auto","created_at":"2026-04-03 11:41:16","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":814968,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9246758/v1/4d6695c9-21c6-4dd5-93f3-abe26c9e7eb9.pdf"},{"id":106045197,"identity":"00522672-5703-41de-b87d-93bc35ea46ac","added_by":"auto","created_at":"2026-04-02 19:10:49","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":36484,"visible":true,"origin":"","legend":"","description":"","filename":"Appendix.docx","url":"https://assets-eu.researchsquare.com/files/rs-9246758/v1/682096fcf3aa8e2c9392869c.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Ricardian Equivalence and Household Consumption: Evidence from Administrative Banking Microdata","fulltext":[{"header":"I. Introduction","content":"\u003cp\u003eThis study examines a central implication of the Ricardian Equivalence framework by testing whether the Barro\u0026ndash;Ricardian mechanism operates in practice, i.e., whether reductions in taxation lead consumers to postpone current consumption. Although this question lies at the core of fiscal theory, its empirical identification remains difficult in real-world settings, where observed consumption and saving decisions are shaped by a wide range of macroeconomic, institutional, and behavioural factors. To address this challenge, the analysis draws on a novel proprietary dataset of client-level banking information, capturing detailed patterns of consumption, income, and savings behaviour, and combines it with macroeconomic indicators. To our knowledge, such a granular micro\u0026ndash;macro integration has not yet been applied to assess Ricardian behaviour in this context. This unique data environment substantially enhances the precision and credibility of the results, enabling the study to provide new evidence on behavioural responses to fiscal policy that extends beyond conventional survey or aggregate data approaches. The applied panel-data methodology further strengthens the quantitative assessment of the research questions and supports the formulation of analytically sound and policy-relevant insights.\u003c/p\u003e \u003cp\u003eRicardian Equivalence remains one of the central theoretical benchmarks in the analysis of fiscal policy and government debt. In Barro\u0026rsquo;s (\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e1974\u003c/span\u003e) formulation, government debt is neutral because forward-looking households internalize the intertemporal budget constraint of the state and offset debt-financed tax changes through adjustments in saving. Consumption decisions should therefore be unaffected by whether government expenditure is financed through taxes or borrowing (Barro, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e1974\u003c/span\u003e; Mankiw, \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2000\u003c/span\u003e). Yet the empirical validity of this proposition remains contested. While some studies report findings broadly consistent with Ricardian behaviour, including Seater (\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e1993\u003c/span\u003e), Adji et al. (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2009\u003c/span\u003e), Cadsby and Frank (\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e1991\u003c/span\u003e), and Di Laurea and Ricciuti (\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2003\u003c/span\u003e), many others find evidence to the contrary, such as Bernheim (1987), Slate et al. (\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e1995\u003c/span\u003e), Meissner and Rostam-Afschar (\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2014\u003c/span\u003e, \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2017\u003c/span\u003e), Mertens and Ravn (\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2013\u003c/span\u003e), Shapiro and Slemrod (\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e2003\u003c/span\u003e), and Souleles (\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e1999\u003c/span\u003e). As Esposito and Mastromatteo (\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2019\u003c/span\u003e) suggest, the relevance of Ricardian mechanisms may be conditional rather than universal.\u003c/p\u003e \u003cp\u003eA major reason for these mixed findings is the difficulty of isolating Ricardian effects from confounding influences such as progressive taxation, political uncertainty, liquidity constraints, preference heterogeneity, and other behavioural or institutional frictions. This makes empirical validation methodologically demanding and increases the value of data that allow consumer behaviour to be observed directly rather than inferred indirectly from aggregate series. Against this backdrop, the present study exploits a unique, non-public banking panel dataset that enables a substantially more granular examination of consumer behaviour than is typically possible with aggregate or survey data. While the empirical setting is the Czech Republic, marked in recent years by rising public debt and the lingering effects of major economic shocks, the mechanisms under study are not country-specific. They speak more broadly to how households adjust consumption and saving decisions in response to fiscal policy across different economic environments.\u003c/p\u003e \u003cp\u003eThis context is empirically relevant. In recent years, Czech fiscal policy has faced pressures from the pandemic and global economic developments, resulting in a marked increase in public debt and persistent fiscal deficits. Government debt rose from CZK 1.64 trillion in 2020 to CZK 2.89 trillion in 2023, reaching 42.4% of GDP, with further increases projected. Budget deficits also widened sharply during the pandemic and, although partially reduced, remained elevated thereafter. These developments provide a useful macroeconomic setting for analysing how households respond to fiscal imbalances and tax-policy changes, while also reflecting broader fiscal dynamics observable in many advanced and emerging economies.\u003c/p\u003e \u003cp\u003eWithin this setting, the Barro\u0026ndash;Ricardian framework offers a valuable lens for assessing whether households internalize future fiscal burdens. If consumers fully anticipate future taxation implied by current deficits, fiscal policy should have limited effects on aggregate demand. If not, tax and spending measures may exert substantial real effects over the business cycle. Assessing the empirical validity of Ricardian Equivalence is therefore important for understanding the effectiveness of fiscal stabilization policies and the design of countercyclical interventions.\u003c/p\u003e \u003cp\u003eBuilding on this motivation, the present study advances the empirical assessment of Ricardian Equivalence by moving beyond the aggregate macroeconomic indicators typically used in earlier work (e.g., Bernheim, 1987; Ricciuti, \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2003\u003c/span\u003e; Lucke, \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e1998\u003c/span\u003e). Instead, it draws on a proprietary micro-level dataset of Czech banking clients that captures detailed consumption, income, and savings behaviour and links these measures to macroeconomic variables. The period 2017\u0026ndash;2023 provides a suitable observation window for analysing consumer responses within the banking environment alongside real macroeconomic developments. Importantly, it includes the substantial personal income tax cut implemented in 2021, creating a natural setting in which to examine how fiscal policy changes translate into individual consumption decisions. The panel-data analysis employs fixed effects (FE) and the generalized method of moments (GMM), both of which enhance the robustness and precision of the results. This combination of unique data and advanced methods constitutes the main empirical contribution of the study.\u003c/p\u003e \u003cp\u003eThe remainder of the paper proceeds as follows.\u003c/p\u003e \u003cp\u003eThe next section reviews the theoretical foundations and empirical evidence on Ricardian Equivalence, drawing on both experimental and macroeconomic studies. The subsequent sections present the data, model specification, and empirical results, followed by a discussion of the main findings and their policy implications. The paper concludes with a summary of the key results..\u003c/p\u003e"},{"header":"II. Literature Review","content":"\u003cp\u003eRicardian Equivalence, much like the Modigliani\u0026ndash;Miller theorem in corporate finance, serves as a theoretical benchmark for assessing when fiscal policy becomes neutral (Elmendorf and Mankiw, \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e1999\u003c/span\u003e). Under its core logic of debt neutrality, a debt-financed tax cut does not raise consumption if forward-looking households anticipate the implied future tax burden and increase savings accordingly. In that case, aggregate consumption, savings, and investment remain unchanged, contrary to Keynesian predictions. The modern formulation is associated with Barro (\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e1974\u003c/span\u003e), who formalized the idea in an intergenerational framework, while Buchanan (\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e1976\u003c/span\u003e) linked it to Ricardo\u0026rsquo;s earlier reflections on wartime finance (Ricardo, \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e1951\u003c/span\u003e). Related contributions by Tobin (\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e1952\u003c/span\u003e), Bailey (\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e1962\u003c/span\u003e), and Patinkin (\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e1965\u003c/span\u003e) further broadened the analysis of public debt.\u003c/p\u003e \u003cp\u003eIts empirical support, however, remains mixed. For example, the 1964 U.S. tax cut increased consumption, whereas the 1968 tax increase produced only a limited response, suggesting heterogeneous household behaviour (Mankiw, \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). Moreover, Ricardian neutrality does not imply that all fiscal policy is irrelevant: if deficits are expected to be offset by lower future spending rather than higher taxes, perceived lifetime income may rise and consumption may increase (Elmendorf and Mankiw, \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e1999\u003c/span\u003e). Despite its analytical appeal, the theory rests on strong assumptions that are widely contested. Critics point to limited foresight (Samuelson, \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e1974\u003c/span\u003e), behavioural biases (Thaler and Sunstein, \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e2008\u003c/span\u003e), market imperfections such as credit constraints and nominal rigidities (Stiglitz, \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2000\u003c/span\u003e; Mankiw, \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2000\u003c/span\u003e), and alternative views on debt sustainability (Kelton, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). These critiques underscore the restrictive nature of the assumptions underlying Ricardian equivalence and motivate the need for empirical validation. In this context, although still limited, experimental research on Ricardian equivalence is valuable because it isolates key assumptions in controlled settings, making deviations from Ricardian predictions easier to identify. Macroeconomic empirical studies complement this evidence by testing the same mechanisms in real-world fiscal and behavioural environments.\u003c/p\u003e \u003cp\u003eCadsby and Frank (\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e1991\u003c/span\u003e) were among the first to investigate Ricardian equivalence in laboratory conditions and provided early evidence of its partial validity. Using an overlapping-generations framework, they incorporated the welfare of the second generation into the utility of the first, thereby directly testing the intergenerational linkage central to the Ricardian model. Participants in the parent generation received government debt, which either they or their children would ultimately bear through future interest or principal repayment. Under strict Ricardian equivalence, parents should save the full value of the debt as a bequest. The experiment, conducted with economics students at the University of Guelph, showed that when intergenerational transfers were strong, parents saved nearly the full amount of the debt. When such linkages were weaker, however, the offset was incomplete. The results thus suggest that Ricardian behaviour depends critically on the strength of intergenerational altruism.\u003c/p\u003e \u003cp\u003eBuilding on this focus on intergenerational mechanisms, Slate et al. (\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e1995\u003c/span\u003e) examined the role of uncertainty and found that Ricardian behaviour weakens when households are uncertain about debt repayment. Using a similar overlapping-generations design, they varied the probability of government bond repayment across treatments. Intergenerational transfers increased with the likelihood of repayment, as Ricardian theory predicts, but at lower probabilities participants consumed more and saved less. Their findings indicate that uncertainty about future fiscal obligations reduces the likelihood of Ricardian behaviour.\u003c/p\u003e \u003cp\u003eExtending this line of inquiry, Di Laurea and Ricciuti (\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2003\u003c/span\u003e) tested the effects of liquidity constraints and income uncertainty. Their experiment compared a Ricardian control group with groups facing either liquidity constraints or uncertain income. While the control treatment was more consistent with Ricardian predictions, the presence of income uncertainty, and to a lesser extent liquidity constraints, led parents to leave insufficient inheritances to offset government debt. The study therefore rejects Ricardian equivalence under conditions that more closely resemble actual household decision-making.\u003c/p\u003e \u003cp\u003eFocusing more specifically on institutional features of fiscal policy, Adji et al. (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2009\u003c/span\u003e) examined whether Ricardian equivalence depends on tax structure. Using an overlapping-generations framework with altruistic intergenerational linkages, they compared a baseline treatment with non-distortionary taxes to one with distortionary taxes on savings. Under non-distortionary taxation, increases in government debt were largely offset by higher inheritance transfers, consistent with Ricardian predictions. Under distortionary taxation, however, the offset weakened and consumption became more uneven across periods. Their results show that the form of taxation is an important determinant of whether Ricardian behaviour emerges.\u003c/p\u003e \u003cp\u003eWhile the previous studies emphasize intergenerational transfers and fiscal structure, Meissner (\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2013\u003c/span\u003e) shifted attention to individual intertemporal decision-making by examining consumption and saving in a laboratory life-cycle model that allowed borrowing. Participants faced either rising or declining income profiles over repeated life cycles. The results revealed strong behavioural asymmetries: participants were more averse to borrowing than to saving and deviated more from optimal decisions when borrowing was required. Individuals moving from a saving to a borrowing environment also encountered persistent difficulties despite repeated learning opportunities. These findings highlight debt aversion as an important behavioural friction in intertemporal choice.\u003c/p\u003e \u003cp\u003eBuilding directly on this behavioural perspective, Meissner and Rostam-Afschar (\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2014\u003c/span\u003e) tested Ricardian equivalence in a modified life-cycle experiment with random income realizations and varying tax schedules. They found strong evidence against the theory: tax cuts increased consumption by 22%, while tax increases reduced it by 30%, and most participants did not behave in a Ricardian manner. In a follow-up study, Meissner and Rostam-Afschar (\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2017\u003c/span\u003e) introduced an additional treatment in which future tax schedules were fixed and fully disclosed to participants. Even under this greater fiscal transparency, Ricardian equivalence was still rejected, with a majority of participants deviating from the theory. Together, these studies suggest that tax changes meaningfully affect consumption even under highly controlled and information-rich conditions.\u003c/p\u003e \u003cp\u003eTurning from experimental to macroeconomic evidence, a similar pattern of conditional and often non-Ricardian results emerges. Kormendi (\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e1983\u003c/span\u003e), using U.S. data for 1929\u0026ndash;1976, argued that government debt influences private consumption, thereby contradicting Ricardian neutrality. In line with this, Bernheim (1987) offered a broad critique of the theory and concluded that most empirical studies reject Ricardian behaviour, with deficits tending to stimulate consumption rather than leave it unchanged. Extending the analysis to a different institutional setting, Leiderman and Blejer (\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e1988\u003c/span\u003e), using Israeli data from 1980 to 1985, found that although some theoretical elements of Ricardian equivalence were consistent with the data, substantial practical deviations remained, shaped by factors such as financial liberalization and economic openness.\u003c/p\u003e \u003cp\u003eFurther questioning the robustness of earlier findings, Feldstein and Elmendorf (\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e1990\u003c/span\u003e) emphasized the fragility of empirical support for Ricardian equivalence. Reassessing studies including Kormendi (\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e1983\u003c/span\u003e), they showed that results were highly sensitive to sample composition, particularly the inclusion of wartime observations. Consistent with these concerns, Campbell and Mankiw (\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e1991\u003c/span\u003e) demonstrated, using quarterly data for several advanced economies, that a substantial share of households responds directly to current disposable income rather than smoothing consumption fully in line with permanent income. Shapiro and Slemrod (1995) reached a similar conclusion, finding that many consumers increase spending in response to tax cuts. These findings led to the development of macroeconomic models incorporating heterogeneous agents, distinguishing between Ricardian and rule-of-thumb consumers (Mankiw, \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2000\u003c/span\u003e; Gal\u0026iacute;, L\u0026oacute;pez-Salido, and Vall\u0026eacute;s, 2004).\u003c/p\u003e \u003cp\u003eBroadening the perspective to different economic environments, Khalid (\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e1996\u003c/span\u003e) examined 21 developing countries over 1960\u0026ndash;1988 and found predominantly non-Ricardian outcomes, attributing them to liquidity constraints, limited credit access, and weak substitutability between public and private consumption. In contrast, Lucke (\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e1998\u003c/span\u003e), using German data from 1960 to 1994, identified a tension between theory and evidence: although the representative-agent model was rejected, aggregate outcomes still showed only modest deviations from Ricardian behaviour. Extending the analysis across countries, Cuaresma and Reitschuler (\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2007\u003c/span\u003e), in a panel study of EU-15 countries from 1960 to 2002, reported mixed results, with some countries exhibiting negligible fiscal multipliers under stricter fiscal conditions and others continuing to display non-Ricardian responses. Most recently, Georgarakos and Kenny (\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2022\u003c/span\u003e), using survey-based panel data from six euro area countries during the COVID-19 period, showed that more positive perceptions of fiscal support significantly increased household spending, particularly on durable goods. This finding is clearly inconsistent with Ricardian neutrality and suggests that fiscal policy effects may be especially pronounced in times of crisis.\u003c/p\u003e \u003cp\u003eTaken together, the experimental and macroeconomic literature indicates that Ricardian equivalence holds, if at all, only under restrictive conditions and is highly sensitive to intergenerational linkages, uncertainty, liquidity constraints, tax design, and household heterogeneity.\u003c/p\u003e \u003cp\u003eWhile the international literature offers valuable experimental evidence on Ricardian equivalence, especially Meissner and Rostam-Afschar (\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2014\u003c/span\u003e, \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2017\u003c/span\u003e), empirical studies based on high-resolution banking microdata are virtually absent. To our knowledge, this paper is the first to examine Ricardian behaviour using transaction-level data from a financial institution. Such data make it possible to observe actual consumption and saving decisions, including liquidity constraints, income dynamics, and heterogeneity in expectations, in a way experimental settings can only approximate. This helps address concerns about the external validity of laboratory evidence and allows Ricardian behaviour to be tested in a natural environment shaped by real financial constraints. In doing so, the paper bridges the gap between prior experimental research and real-world empirical analysis, providing a more comprehensive assessment of whether households behave in line with Barro\u0026rsquo;s Ricardian equivalence and when deviations are most likely to arise.\u003c/p\u003e"},{"header":"III. Methodology and Model Specification","content":"\u003cp\u003eWe investigate the Barro\u0026ndash;Ricardian hypothesis using a quarterly panel dataset spanning 2017\u0026ndash;2023 that combines proprietary banking microdata from Raiffeisenbank (2024) with macroeconomic indicators.\u003c/p\u003e\n\u003cp\u003eThe dataset contains transaction-level information on household financial behaviour together with standard economic variables relevant for assessing responses to fiscal policy. The use of such granular, real-world data is rare in studies of Ricardian equivalence and allows for a more precise evaluation of household consumption and saving decisions in practice.\u003c/p\u003e\n\u003cp\u003eAccordingly, the study addresses the following central question: Does the Barro\u0026ndash;Ricardian equivalence principle holds, implying that lower taxation induces households to postpone consumption? Drawing on theoretical insights and previous empirical work, we construct a linear regression econometric model inspired by Bernheim (1987), Ricciuti (\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2003\u003c/span\u003e), and Lucke (\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e1998\u003c/span\u003e):\u003c/p\u003e\n\u003cdiv id=\"Equ1\" class=\"Equation\"\u003e\n \u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ1\" name=\"EquationSource\"\u003e$$\\:{CONSUMPTION}_{t}\\:=\\:{\\beta\\:}_{0}+\\:{\\beta\\:}_{1}{WEALTH}_{t}+{\\beta\\:}_{2}{INCOME}_{t}-{\\beta\\:}_{3}{INTEREST\\_SA}_{t}\\:-\\:{\\beta\\:}_{4}{INCOME\\_TAX}_{t}{-\\:\\beta\\:}_{5}{INFLATION}_{t}+{\\beta\\:}_{6}tGD{P}_{t}-\\:{\\beta\\:}_{7}{DEF\\_G}_{t}+{\\beta\\:}_{8}PUBLIC\\_{DEBT}_{t}+dCOVID+dWAR\\:{+\\:\\epsilon\\:}_{t}\\:$$\u003c/div\u003e\n \u003cdiv class=\"EquationNumber\"\u003e1\u003c/div\u003e\n\u003c/div\u003e\n\u003cp\u003eIn Eq. (46), which specifies the model used to test the hypothesis of this study, the dependent variable \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{CONSUMPTION}_{t}\\:\\)\u003c/span\u003e\u003c/span\u003erepresents individual consumption expenditure at time t. The first explanatory variable, derived from proprietary data, is the individual\u0026rsquo;s income at time t \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{INCOME}_{t}\\)\u003c/span\u003e\u003c/span\u003e. To account for the effect of total wealth, we include the variable \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{WEALTH}_{t}\\)\u003c/span\u003e\u003c/span\u003e, defined as the total monetary balances accumulated across all of an individual\u0026rsquo;s bank accounts at time t. The explanatory variable \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{INTEREST\\_SA}_{t}\\)\u003c/span\u003e\u003c/span\u003e denotes the interest rate on the individual\u0026rsquo;s savings account at time t. A key variable for testing the hypothesis is \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{INCOME\\_TAX}_{t}\\)\u003c/span\u003e\u003c/span\u003e which measures the level of personal income tax at time t. Inflation is represented by the variable \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{INFLATION}_{t}\\)\u003c/span\u003e\u003c/span\u003e while \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:tGD{P}_{t}\\:\\)\u003c/span\u003e\u003c/span\u003ereflects GDP growth at time t. Additional macroeconomic explanatory variables include the government expenditure deficit-to-GDP ratio and the public debt-to-GDP ratio, expressed as percentages \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{DEF\\_G}_{t}\\)\u003c/span\u003e\u003c/span\u003e, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:PUBLIC\\_{DEBT}_{t}\\)\u003c/span\u003e\u003c/span\u003e. Finally, dummy variables are incorporated to capture the effects of the COVID-19 pandemic and the war in Ukraine \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:dCOVID,\\:dWAR\\)\u003c/span\u003e\u003c/span\u003e.\u003c/p\u003e\n\u003cp\u003eIndividual consumption, personal income tax, and the government expenditure deficit-to-GDP ratio are central variables for testing the proposed hypothesis. Consumption serves as the dependent variable, while the direction and magnitude of the effects of personal income tax, the government deficit, and other explanatory variables constitute the core of the analysis. An individual\u0026rsquo;s income and total wealth, measured as savings across all bank accounts, are expected to influence consumption over time. These variables feature prominently in most consumption models and consistently show positive and significant effects on consumption in empirical research. According to Ricardian equivalence, personal income tax should have no effect on consumption, as tax changes are theoretically neutral. The interest rate on savings accounts is expected to encourage saving and therefore exert a negative influence on consumption. Inflation, measured by the CPI, is included as a control variable capturing precautionary savings behaviour. Variables such as GDP growth, the government expenditure deficit-to-GDP ratio, and the public debt-to-GDP ratio are incorporated to capture the influence of broader fiscal and macroeconomic conditions on individual consumption. GDP growth is expected to have a positive effect on consumption, and a similar positive relationship is anticipated for public debt if it signals increased economic activity or fiscal expansion. Table \u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e summarizes all variables included in the model along with the expected signs of their effects.\u003c/p\u003e\n\u003cp\u003e\u003c/p\u003e\u0026nbsp;\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eExplanatory variables and their anticipated direction of effect on consumption\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eWEALTH\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e+\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eINCOME\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e+\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eINTEREST_SA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e\u0026ndash;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eINCOME_TAX\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e\u0026ndash;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eINFLATION\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e\u0026ndash;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003etGDP\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e+\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eDEF_G\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e+\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003ePUBLIC_DEBT\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e\u0026ndash;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003edCOVID\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e+\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003edWAR\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e+\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eSource\u003c/em\u003e: Own elaboration, (2025).\u003c/p\u003e\n\u003cp\u003eExisting empirical research yields mixed results, with studies testing Ricardian equivalence in real-world settings failing to reach a consensus on the theory\u0026rsquo;s validity. While Seater (\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e1993\u003c/span\u003e) argues that empirical evidence is broadly consistent with Ricardian behaviour, many others report findings that contradict the theory (e.g., Bernheim, 1987; Lucke, \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e1998\u003c/span\u003e). These inconsistencies provide a key motivation for the present study. By analysing proprietary banking data, information not publicly accessible, together with macroeconomic indicators, this research offers a novel and more comprehensive perspective on Ricardian equivalence, extending the existing empirical literature in a direction not previously explored.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eData\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eTo test the Ricardian equivalence hypothesis in the Czech context, we use quarterly panel data covering the period 2017\u0026ndash;2023. This timeframe is particularly suitable because it includes the 2021 reduction in personal income tax, which provides a natural setting to examine whether lower taxation leads individuals to adjust their consumption in ways consistent with Ricardian behavior. The dataset combines proprietary banking microdata, capturing individual consumption, income, and savings, with macroeconomic indicators obtained from the Czech Statistical Office.\u003c/p\u003e\n\u003cp\u003ePanel data offer several advantages over purely cross-sectional or time-series analyses, including richer information content, the ability to observe both individual heterogeneity and temporal dynamics, and improved reliability of parameter estimates. A key distinguishing feature of this study is the use of restricted-access banking data, which provide highly detailed information on the financial behaviour of anonymized clients that is not typically available for academic research. These unique proprietary data represent the consumer behavior of Raiffeisenbank clients. Because banking microdata are seldom accessible for research purposes, the dataset itself forms a significant part of the study\u0026rsquo;s contribution. The final dataset includes 21,885 randomly selected bank clients who meet predefined selection criteria and are observed throughout the sample period.\u003c/p\u003e\n\u003cp\u003eThis section describes and defines the variables used in the analysis. Descriptive statistics for all variables are presented in Table \u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e.\u003c/p\u003e\n\u003cp\u003e\u003c/p\u003e\u0026nbsp;\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eDescriptive statistics of the dataset variables\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\n \u003cp\u003eCONSUMPTION\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003eMean\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003eMedian\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003eMinimum\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003eMaximum\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003cth align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e234,0\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e152,7\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e0,000\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e10 610\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eWEALTH\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e351,0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e119,2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e0,000\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e16 358\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eINCOME\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e255,3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e173,5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e45,00\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e3 039\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eINTEREST_SA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e1,825\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e1,000\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e0,300\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e5,500\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eINCOME_TAX\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e17,29\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e19,00\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e15,00\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e19,00\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eINFLATION\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e5,746\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e2,950\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e1,900\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e17,60\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003etGDP\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e0,825\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e1,700\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e-10,80\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e4,700\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eDEF_G\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e-2,303\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e-2,600\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e-9,500\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e3,310\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003ePUBLIC_DEBT\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e39,16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e39,94\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e30,80\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e45,20\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003edCOVID\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e0,250\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e0,000\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e0,000\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e1,00\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003edWAR\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e0,286\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e0,000\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e0,000\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e1,00\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003ctfoot\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"5\"\u003e\u003cem\u003eNotes\u003c/em\u003e: The variables CONSUMPTION, WEALTH, and INCOME are expressed in thousands. INTEREST_SA, INCOME_TAX, INFLATION, tGDP, DEF_GOV, and PUBLIC_DEBT are expressed as percentages.\u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tfoot\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSource\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eOwn Elaboration, (2025).\u003c/p\u003e\n\u003cp\u003eIndividual consumption at time \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\text{t}\\)\u003c/span\u003e\u003c/span\u003e is the dependent variable and represents a bank client\u0026rsquo;s total consumption expenditures. It aggregates quarterly expenditures made through the client\u0026rsquo;s current and credit card accounts. To ensure accuracy, transfers between the client\u0026rsquo;s own accounts, such as movements to savings, term deposits, other current accounts, or investment accounts, are excluded, as are credit card repayments to avoid double-counting. All values are expressed in thousands of CZK (Raiffeisenbank, 2024).\u003c/p\u003e\n\u003cp\u003eWealth at time \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:t\\)\u003c/span\u003e\u003c/span\u003e is a proprietary variable capturing the total financial resources available to a client. It is defined as the cumulative end-of-period balance across all accounts held within the institution, including current, savings, and term deposit accounts. This measure can be interpreted as an indicator of savings capacity. All monetary amounts are expressed in thousands of CZK (Raiffeisenbank, 2024).\u003c/p\u003e\n\u003cp\u003eIncome at time \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\text{t}\\)\u003c/span\u003e\u003c/span\u003e is an explanatory variable representing the individual\u0026rsquo;s total quarterly income received into their current account. As with consumption, income figures are adjusted to exclude transfers between the client\u0026rsquo;s own accounts and are expressed in thousands of CZK (Raiffeisenbank, 2024).\u003c/p\u003e\n\u003cp\u003ePersonal income tax at time \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\text{t}\\)\u003c/span\u003e\u003c/span\u003e captures the statutory income tax rate applied to employees. This variable is central to testing Ricardian equivalence in the Czech setting. Notably, the personal income tax rate fell from 19% to 15% in 2021, providing a natural experiment for assessing the impact of tax reductions on consumption (Czech Statistical Office, 2024).\u003c/p\u003e\n\u003cp\u003eThe interest rate on savings accounts at time \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\text{t}\\)\u003c/span\u003e\u003c/span\u003e reflects the rate applied to an individual\u0026rsquo;s savings account and influences consumption and saving decisions. Using client-specific interest rates allows the model to capture the direct incentives faced by households (Raiffeisenbank, 2024).\u003c/p\u003e\n\u003cp\u003eGDP growth rate at time \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\text{t}\\)\u003c/span\u003e\u003c/span\u003e measures the year-on-year percentage change in quarterly GDP. This indicator provides insight into macroeconomic conditions, reflecting the pace of economic expansion or contraction and informing expectations about economic policy effectiveness and overall economic activity (Czech Statistical Office, 2024).\u003c/p\u003e\n\u003cp\u003eInflation rate at time \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\text{t}\\)\u003c/span\u003e\u003c/span\u003e is represented by the annual percentage change in the Consumer Price Index (CPI), comparing the price level in the current month to the same month of the previous year. Inflation is included as a control variable linked to precautionary saving behavior (Czech Statistical Office, 2024).\u003c/p\u003e\n\u003cp\u003eThe government expenditure deficit-to-GDP ratio at time \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\text{t}\\)\u003c/span\u003e\u003c/span\u003e measures the extent to which government expenditure exceeds revenue, expressed as a percentage of quarterly GDP. This indicator reflects fiscal policy stance and the degree of fiscal imbalance, offering insight into the potential long-term sustainability of public finances, (Czech Statistical Office, 2024).\u003c/p\u003e\n\u003cp\u003eThe total government debt-to-GDP ratio at time \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\text{t}\\)\u003c/span\u003e\u003c/span\u003e captures the size of public debt relative to national economic output. A high ratio suggests increased fiscal pressures and potential macroeconomic risks. Within the context of Ricardian equivalence, this variable is relevant because the theory posits that government debt should not influence real economic activity if households fully anticipate future tax adjustments (Czech Statistical Office, 2024).\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eRegression Analysis of the Main Models\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eIn the analysis of panel data, several regression techniques are employed to estimate parameters and test the proposed hypotheses. Before proceeding with the primary analysis of proprietary data to test our hypothesis, it was essential to conduct several preliminary tests and model estimations. These steps were critical in determining the appropriate estimation method and model specification, thereby ensuring the robustness and validity of our results. The correlation matrix confirmed a moderate relationship between key explanatory variables, particularly \u0026quot;wealth\u0026quot; and \u0026quot;income,\u0026quot; To mitigate potential issues with autocorrelation and multicollinearity, we opted for the fixed-effects (FE) method and the generalized method of moments (GMM). Zero-order effect regressions highlighted income as the dominant predictor of consumption, while the influence of wealth was comparatively weaker. The Hausman test (\u0026chi;\u0026sup2;=5099.41, p-value\u0026thinsp;\u0026lt;\u0026thinsp;0.01) rejected the null hypothesis, supporting the use of the fixed-effects model over the random-effects model due to significant differences in parameter estimates.Based on the results of the Hausman test, the fixed effects (FE) estimator was selected as the primary method. The FE approach incorporates individual-specific effects through unit-level intercepts, effectively addressing unobserved heterogeneity and mitigating potential endogeneity arising from omitted time-invariant factors. Its key advantage lies in its ability to control for unobserved characteristics that vary across individuals but remain constant over time, resulting in consistent and reliable parameter estimates. For comparison, the pooled ordinary least squares (pooled OLS) estimator treats the panel dataset as a single cross-sectional sample and applies standard OLS techniques. Although useful for preliminary or simplified analyses where cross-sectional heterogeneity is negligible, pooled OLS does not account for individual-specific effects and may produce biased estimates when heterogeneity or endogeneity is present.\u003c/p\u003e\n\u003cp\u003eTo strengthen the robustness of the results, we also apply the generalized method of moments (GMM), an advanced econometric technique capable of addressing challenges such as endogeneity, autocorrelation, and heteroskedasticity. GMM offers a flexible framework that accommodates various model specifications and moment conditions, making it particularly suitable for dynamic or complex panel structures. Its implementation, however, requires careful specification of valid instruments and high-quality data. Despite these demands, GMM produces robust and efficient parameter estimates. Notably, existing empirical literature has not yet applied GMM to tests of Ricardian equivalence, making its use in this study a novel methodological contribution.\u003c/p\u003e\n\u003cp\u003eBy integrating these complementary estimation techniques, the analysis achieves a high degree of robustness and precision, which is crucial for evaluating the hypotheses and assessing the influence of both proprietary banking microdata and macroeconomic factors on individual consumption behavior. This multifaceted econometric strategy provides a solid foundation for deriving reliable conclusions from the empirical results.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eFixed Effects and Pooled OLS Estimation\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThis section analyses the quarterly panel data using fixed effects (FE) and pooled ordinary least squares (pooled OLS) estimators to empirically test the hypothesis derived from Ricardian equivalence. As noted earlier, the FE approach provides consistent estimates by controlling for unobserved individual-specific characteristics and thereby addressing potential endogeneity. In addition, pooled OLS is employed as a complementary method to assess the magnitude and direction of relationships between the explanatory variables and the dependent variable, enhancing the robustness of the analysis.\u003c/p\u003e\n\u003cp\u003eBoth models are estimated using heteroskedasticity- and autocorrelation-consistent (HAC) robust standard errors to ensure reliable inference. The estimated coefficients of the regression models are reported in Table \u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e.\u003c/p\u003e\n\u003cp\u003e\u003c/p\u003e\u0026nbsp;\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eEstimated coefficients of the regression model using the FE method with robust standard errors (HAC) and Pooled OLS with robust standard errors (HAC). Includes 21,885 cross-sectional units, time series length\u0026thinsp;=\u0026thinsp;28\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\n \u003cp\u003eWEALTH\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e\n \u003cp\u003eFE\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e\n \u003cp\u003ePooled OLS\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003cth align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e\u0026minus;0,12477***\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e(\u0026minus;\u0026thinsp;39,39)\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e\u0026minus;0,02858***\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e(\u0026minus;\u0026thinsp;18,19)\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eINCOME\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e0,67239***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e(143,2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e0,75741***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e(192,1)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eINTEREST_SA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e0,29839\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e(0,7174)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e\u0026minus;1,51849***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e(\u0026minus;\u0026thinsp;3,561)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eINCOME_TAX\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e\u0026minus;6,12487***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e(-16,15)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e\u0026minus;2,20539***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e(\u0026minus;\u0026thinsp;5,678)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eINFLATION\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e\u0026minus;0,73531***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e(\u0026minus;\u0026thinsp;5,886)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e\u0026minus;0,34156***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e(\u0026minus;\u0026thinsp;2,707)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003etGDP\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e1,00986***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e(6,300)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e1,31857***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e(7,855)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eDEF_G\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e0,04275\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e(0,3914)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e\u0026minus;0,13199\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e(\u0026minus;\u0026thinsp;1,222)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003ePUBLIC_DEBT\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e0,04275***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e(\u0026minus;\u0026thinsp;22,88)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e\u0026minus;1,48437***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e(\u0026minus;\u0026thinsp;15,42)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003edCOVID\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e24,7801***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e(16,20)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e15,3397***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e(8,972)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003edWAR\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e53,0186***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e(19,92)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e39,5633***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e(14,62)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eConstant\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e300,522***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e(30,44)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e135,083***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e(14,84)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eAdjusted R\u003csup\u003e2\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e0,5812\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e0,5274\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003ctfoot\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"5\"\u003e\u003cem\u003eNotes\u003c/em\u003e: The dependent variable is observed consumption. T-statistics are provided in parentheses.\u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tfoot\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003c/p\u003e\n\u003cp\u003eSignificance levels. *** p\u0026thinsp;\u0026lt;\u0026thinsp;0,01; ** p\u0026thinsp;\u0026lt;\u0026thinsp;0,05; * p\u0026thinsp;\u0026lt;\u0026thinsp;0,1.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSource\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eOwn Elaboration, (2025).\u003c/p\u003e\n\u003cp\u003eAccording to the F-test, both models are statistically significant at the 0.01% level, and most explanatory variables are significant at the 1% level. The adjusted \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{R}^{2}\\)\u003c/span\u003e\u003c/span\u003e indicates that the fixed effects (FE) model explains 58.1% of the variance in consumption, whereas the pooled OLS model explains 52.7%. As expected, the FE estimator provides higher explanatory power. The Durbin\u0026ndash;Watson statistics 1.9505 for the FE model and 1.9551 for pooled OLS are close to 2, indicating no autocorrelation in the residuals, which is further supported by the very small autocorrelation coefficients (\u0026minus;\u0026thinsp;0.0184 and \u0026minus;\u0026thinsp;0.0188). The interpretation that follows focuses primarily on the FE estimates, with pooled OLS results serving as a complementary comparison.\u003c/p\u003e\n\u003cp\u003eThe coefficients for WEALTH and INCOME show a strong and significant influence on consumption in both estimation methods. The income coefficient (0.67239, significant at 1%) indicates that higher income leads to higher consumption. In contrast, the wealth coefficient (\u0026minus;\u0026thinsp;0.12477, significant at 1%) implies that individuals with higher accumulated balances tend to consume less. The effect of income is roughly three times larger than that of wealth, suggesting that individuals prefer to adjust consumption through current income rather than drawing on savings, reflecting precautionary financial behaviour.\u003c/p\u003e\n\u003cp\u003eA crucial variable for testing Ricardian equivalence is personal income tax (INCOME_TAX). This coefficient is statistically significant at the 1% level in both models and negative in the FE estimates (\u0026minus;\u0026thinsp;6.1249). This indicates that reductions in personal income tax are associated with higher consumption, consistent with Keynesian consumption theory and inconsistent with Ricardian neutrality. According to Ricardian equivalence, tax reductions should have no effect on consumption because households would save rather than spend the additional disposable income. The significant negative coefficient clearly contradicts this prediction and confirms that bank clients increase consumption when the tax burden decreases.\u003c/p\u003e\n\u003cp\u003eThe interest rate on savings accounts (INTEREST_SA) is not statistically significant in the FE model but is significant and negative in the pooled OLS estimation. The sign of the coefficient is consistent with substitution effects: higher interest rates encourage saving and reduce consumption by increasing the opportunity cost of spending.\u003c/p\u003e\n\u003cp\u003eInflation (INFLATION) is statistically significant at the 1% level and has a negative coefficient (\u0026minus;\u0026thinsp;0.7353). Higher inflation reduces consumption because rising prices lower purchasing power, prompting households to restrain spending or reallocate funds to preserve value. This result supports the view that consumers are aware of inflationary pressures and adjust their behaviour accordingly.\u003c/p\u003e\n\u003cp\u003eGDP growth (tGDP) is positive and significant at the 1% level (coefficient 1.0099). As economic conditions improve, consumers tend to increase their expenditures, consistent with the life-cycle hypothesis, which states that consumption responds to expectations about future income and economic prospects.\u003c/p\u003e\n\u003cp\u003eThe coefficients for the government expenditure deficit-to-GDP ratio (DEF_G) and the public debt-to-GDP ratio (PUBLIC_DEBT) show mixed results. DEF_G is statistically insignificant in both models, though its sign differs slightly. \u003csup\u003e1\u003c/sup\u003ePUBLIC_DEBT is positive and significant in the FE model, suggesting that higher public debt is associated with increased consumption, but it is negative and significant in the pooled OLS model (\u0026minus;\u0026thinsp;1.48437). This inconsistency indicates the need for further analysis using alternative techniques, which is addressed in the following chapter with GMM estimation.\u003c/p\u003e\n\u003cp\u003eBoth dummy variables COVID-19 (dCOVID) and the war in Ukraine (dWAR) are statistically significant at the 1% level and have positive coefficients. Average consumption was higher during these periods compared to baseline periods. The COVID-19 dummy has a coefficient of 24.7801, while the war dummy has a coefficient of 53.0186, suggesting that the war had more than twice the impact on consumption compared to the pandemic. These increases in consumption may stem from heightened uncertainty, increased government spending, or broader fiscal stimuli. This interpretation aligns with Keynesian fiscal policy and the multiplier effect, whereby government expenditure boosts aggregate demand during periods of economic instability.\u003c/p\u003e\n\u003cp\u003eOur results for the COVID-19 period correspond with Georgarakos and Kenny (\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2022\u003c/span\u003e), who show that communication about fiscal support packages increased households\u0026rsquo; expectations of future income and stimulated consumption, especially on durable goods. Like their findings, our results offer no evidence that Ricardian equivalence influences consumer behaviour during crisis periods; expected future tax liabilities do not appear to constrain consumption.\u003c/p\u003e\n\u003cp\u003eIn conclusion, income, wealth, personal income tax, inflation, GDP growth, and public debt all significantly affect consumption. Higher income and GDP growth increase consumption, whereas higher wealth, inflation, and in some cases public debt reduce it. Additionally, events such as the COVID-19 pandemic and the war in Ukraine exert substantial effects on consumption patterns. These findings highlight the importance of considering both macroeconomic context and household-level behaviour when evaluating Ricardian equivalence and designing fiscal policy.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eGMM Estimation\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eTo enhance the robustness of our conclusions, we re-estimate the main model using the generalized method of moments (GMM). GMM is well suited to complex panel data structures and offers reliable parameter estimates by addressing potential endogeneity, autocorrelation, and heteroskedasticity. Although the Durbin\u0026ndash;Watson statistic and autocorrelation coefficient from the fixed effects estimates did not indicate autocorrelation, we employ GMM as an additional validation tool to reinforce the credibility of our findings. Its ability to correct for common econometric complications is particularly valuable when working with proprietary banking microdata, ensuring that our inference rests on a sound and resilient methodological foundation.\u003c/p\u003e\n\u003cp\u003eThe GMM framework requires carefully specified moment conditions, appropriate instrumental variables, and high-quality data. Panel data GMM is most implemented as Difference GMM or System GMM. While Difference GMM relies on first differences to remove unobserved fixed effects, System GMM combines moment conditions in first differences and in levels, thereby improving efficiency when the data exhibit persistence. Given the nature of our variables and the objective of achieving consistent and efficient parameter estimates, we adopt the System GMM approach.\u003c/p\u003e\n\u003cp\u003eBoth one-step and two-step variants of System GMM are employed. The one-step estimator computes the parameters in a single stage using the initial weighting matrix, which is constructed under the assumption of homoskedasticity. This makes the procedure computationally simpler but less efficient if heteroskedasticity or autocorrelation is present. In contrast, the two-step estimator proceeds sequentially: the first step produces preliminary parameter estimates, which are then used to construct a more robust weighting matrix for the second step. This second-stage estimation typically yields more efficient and accurate results, albeit at greater computational cost. Applying both variants enables us to assess the stability of the results across different specifications and strengthens the robustness of the empirical analysis.\u003c/p\u003e\n\u003cp\u003eThe System GMM estimations are conducted on a dataset containing 27,000 observations and 10,000 cross-sectional units. Due to computational constraints, the number of cross-sectional units had to be reduced from the original 21,885 to 10,000 through random sampling, consistent with the sampling strategy used in the initial dataset construction. A sample size of 10,000 cross-sectional units is generally considered sufficient for the GMM method. Several studies confirm that such a sample size provides reliable results when employing GMM. For instance, Han and Kim (\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) discuss the effectiveness of GMM estimates in their analysis, highlighting that even with a medium-sized sample, valid estimates can be achieved. Furthermore, the study by Hallett and Ma (\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e1994\u003c/span\u003e) suggests that GMM estimates perform well even with smaller sample sizes than those used in our analysis. Hallett and Ma (\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e1994\u003c/span\u003e) emphasize that a sample of just 1,000 cross-sectional units is entirely adequate for robust estimation. This reduction does not compromise the representativeness or validity of the results, as the remaining sample remains large, diverse, and statistically robust. Table \u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e reports the estimated coefficients from the System GMM models and the factors influencing observed consumption.\u003c/p\u003e\n\u003cp\u003e\u003c/p\u003e\u0026nbsp;\u003ctable float=\"Yes\" id=\"Tab4\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003e\u003cem\u003eOne-step and two-step dynamic GMM panel model\u003c/em\u003e\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\n \u003cp\u003eCONSUMPTION (\u0026minus;\u0026thinsp;1)\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e\n \u003cp\u003eOne-Step GMM\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colspan=\"3\" nameend=\"c6\" namest=\"c4\"\u003e\n \u003cp\u003eTwo-Step GMM\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003cth align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e0,02725*\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\n \u003cp\u003e(1,829)\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e0,02794*\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e(1,852)\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eWEALTH\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e\u0026minus;0,5139***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\n \u003cp\u003e(15,65)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e\u0026minus;0,5165***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e(15,72)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eINCOME\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e0,8635***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\n \u003cp\u003e(\u0026minus;\u0026thinsp;5,487)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e0,8647***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e(\u0026minus;\u0026thinsp;5,539)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eINTEREST_SA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e\u0026minus;0,22579\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\n \u003cp\u003e(\u0026minus;\u0026thinsp;0,995)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e0,17936*\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e(1,906)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eINCOME_TAX\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e\u0026minus;2,0595***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\n \u003cp\u003e(\u0026minus;\u0026thinsp;5,724)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e\u0026minus;1,2155***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e(\u0026minus;\u0026thinsp;8,418)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eINFLATION\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e\u0026minus;0,1803\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\n \u003cp\u003e(\u0026minus;\u0026thinsp;1,522)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e\u0026minus;0,1224***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e(\u0026minus;\u0026thinsp;2,908)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003etGDP\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e0,4431\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\n \u003cp\u003e(0,7081)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e0,3253\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e(1,327)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eDEF_G\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e\u0026minus;1,4693***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\n \u003cp\u003e(\u0026minus;\u0026thinsp;3,417)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e0,0658\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e(0,3208)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003ePUBLIC_DEBT\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e\u0026minus;0,0866***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\n \u003cp\u003e(\u0026minus;\u0026thinsp;5,595)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e\u0026minus;0,0455***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e(\u0026minus;\u0026thinsp;9,202)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003edCOVID\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e55,298***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\n \u003cp\u003e(4,960)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e45,349***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e(7,810)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003edWAR\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e81,712***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\n \u003cp\u003e(5,833)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e64,617***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e(8,433)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eConstant\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e630,64***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\n \u003cp\u003e(6,024)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e513,54***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e(8,312)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eNumber of Instruments\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"3\" nameend=\"c4\" namest=\"c2\"\u003e\n \u003cp\u003e614\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e\n \u003cp\u003e614\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eTest for AR (1) error\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"3\" nameend=\"c4\" namest=\"c2\"\u003e\n \u003cp\u003ez\u0026thinsp;=\u0026thinsp;\u0026minus;\u0026thinsp;6,229[0,000]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e\n \u003cp\u003ez = \u0026minus;\u0026thinsp;6,23087 [0,000]\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eTest for AR (2) error\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"3\" nameend=\"c4\" namest=\"c2\"\u003e\n \u003cp\u003ez\u0026thinsp;=\u0026thinsp;0,636[0,524]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e\n \u003cp\u003ez\u0026thinsp;=\u0026thinsp;0,557997 [0,577]\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eSargan test\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"3\" nameend=\"c4\" namest=\"c2\"\u003e\n \u003cp\u003eChi-square (1172)\u0026thinsp;=\u0026thinsp;74175,3 [0,327]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e\n \u003cp\u003eChi-square (1172)\u0026thinsp;=\u0026thinsp;3897,73 [0,276]\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eWald test\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"3\" nameend=\"c4\" namest=\"c2\"\u003e\n \u003cp\u003eChi-square (11)\u0026thinsp;=\u0026thinsp;2041,65 [0,000]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e\n \u003cp\u003eChi-square (11)\u0026thinsp;=\u0026thinsp;3765,55 [0,000]\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003ctfoot\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"6\"\u003e\u003cem\u003eNotes\u003c/em\u003e: The dependent variable is observed consumption. T-statistics are provided in parentheses.\u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tfoot\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003c/p\u003e\n\u003cp\u003eSignificance levels. *** p\u0026thinsp;\u0026lt;\u0026thinsp;0,01; ** p\u0026thinsp;\u0026lt;\u0026thinsp;0,05; * p\u0026thinsp;\u0026lt;\u0026thinsp;0,1.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSource\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eOwn Elaboration, (2025).\u003c/p\u003e\n\u003cp\u003eIn both the one-step and two-step System GMM models, the regression results are broadly comparable. The first specification corresponds to the one-step estimator, while the second follows the two-step procedure. Both models explain consumption (CONSUMPTION) using the same set of explanatory variables. The Arellano\u0026ndash;Bond autocorrelation tests confirm the absence of second-order serial correlation: the AR(1) test yields a p-value of 0.000, consistent with expectations for first-differenced errors, while the AR(2) test returns a p-value of 0.524, supporting the null hypothesis of no second-order autocorrelation. These diagnostics validate the appropriateness of the GMM framework. The Sargan test of overidentifying restrictions indicates that the instruments are valid, although the relatively high chi-square values suggest the possibility of instrument proliferation, a common issue when the number of instruments is large. The Wald tests confirm that both models are jointly significant. Additional requirements for valid GMM estimation are met: the inclusion of a lagged dependent variable and a number of instruments smaller than the number of cross-sectional units (614\u0026thinsp;\u0026lt;\u0026thinsp;10,000).\u003c/p\u003e\n\u003cp\u003eThe lagged consumption term (CONSUMPTION(\u0026minus;\u0026thinsp;1)) is statistically significant at the 10% level in both models, with coefficients of 0.0273 and 0.0279. These positive values indicate a modest degree of consumption persistence, suggesting that past consumption patterns have a small but significant effect on current behavior.\u003c/p\u003e\n\u003cp\u003eWealth (WEALTH) exhibits a negative and statistically significant effect at the 1% level in both models, with nearly identical coefficients (\u0026minus;\u0026thinsp;0.5139 and \u0026minus;\u0026thinsp;0.5165).\u003c/p\u003e\n\u003cp\u003eIncome (INCOME) remains positive and significant at the 1% level, with values of 0.8635 and 0.8647. Both results reinforce earlier findings: income increases consumption, whereas greater accumulated wealth is associated with reduced consumption.\u003c/p\u003e\n\u003cp\u003eThe coefficients for Personal Income Tax (INCOME_TAX) are negative and statistically significant at the 1% level in both GMM specifications, with magnitudes of \u0026minus;\u0026thinsp;2.0595 and \u0026minus;\u0026thinsp;1.2155. These results indicate that higher income taxes reduce consumption through lower disposable income, contradicting Ricardian equivalence and aligning with standard consumption theory.\u003c/p\u003e\n\u003cp\u003eThe Interest Rate on Savings (INTEREST_SA) is negative but not statistically significant in the one-step model. In contrast, in the two-step model it is positive and significant at the 10% level. These mixed results suggest that interest rates may influence consumption through both income and substitution channels. In standard consumption theory, the substitution effect typically dominates: higher interest rates raise the return to saving and the opportunity cost of current consumption, thereby discouraging expenditure. The positive coefficient in the two-step model is likely attributable to sensitivity to instrument choice and small-sample properties of the two-step GMM estimator, rather than reflecting a stable behavioural pattern.\u003c/p\u003e\n\u003cp\u003eInflation (INFLATION) is negative and insignificant in the one-step model but becomes negative and significant at the 1% level in the two-step model. This supports the view that rising inflation dampens consumption by reducing purchasing power and increasing uncertainty.\u003c/p\u003e\n\u003cp\u003eGDP growth (tGDP) is positive but statistically insignificant in both models, implying that short-term fluctuations in GDP growth do not exert a measurable effect on household consumption within this framework.\u003c/p\u003e\n\u003cp\u003eThe Government Expenditure Deficit-to-GDP Ratio (DEF_G) is negative and statistically significant at the 1% level in the one-step GMM, but insignificant in the two-step model. The Total Public Debt-to-GDP Ratio (PUBLIC_DEBT), however, is negative and statistically significant at the 1% level in both estimations (\u0026minus;\u0026thinsp;0.0866 and \u0026minus;\u0026thinsp;0.0455), suggesting that higher public debt reduces consumption, likely due to increased uncertainty, lower disposable income or expectations of future fiscal adjustment.\u003c/p\u003e\n\u003cp\u003eThe dummy variables for COVID-19 (dCOVID) and the war in Ukraine (dWAR) are positive and significant at the 1% level in both models. The COVID-19 coefficients (55.298 and 45.349) and war coefficients (81.712 and 64.617) indicate that consumption rose significantly during these periods. These effects may reflect heightened uncertainty, fiscal support measures, or shifts in household spending patterns during crises.\u003c/p\u003e\n\u003cp\u003eOverall, the one-step and two-step System GMM models provide consistent and meaningful insights into the determinants of consumption. Wealth, income, personal income tax, public debt, and the crisis-related dummy variables consistently exhibit statistically significant effects aligned with economic theory. Although certain variables, such as interest rates and the government deficit ratio, display variability across estimation methods, both GMM specifications appear well-specified and robust.\u003c/p\u003e\n\u003cp\u003eTo test our hypothesis, we estimated regression models using three complementary methods: fixed effects (FE), pooled ordinary least squares (pooled OLS), and System GMM. This multi-method approach provides a comprehensive assessment of the determinants of consumption and ensures the robustness of the empirical findings. F-tests confirm that the models are significant at the 1% level. The adjusted \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{R}^{2}\\)\u003c/span\u003e\u003c/span\u003e indicates that the fixed effects model explains 82.2% of the variance, while the GMM model explains 79.5%, confirming the strong explanatory power of the specifications.\u003c/p\u003e"},{"header":"IV. Results and Discussion","content":"\u003cp\u003eOur results broadly reject the validity of Ricardian equivalivalence, consistent with the findings of earlier influential studies such as Feldstein (1982), Leiderman and Blejer (\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e1988\u003c/span\u003e), Bernheim (1987), Ricciuti (\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2003\u003c/span\u003e), and Giorgioni and Holden (\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2003\u003c/span\u003e). The primary aim of our panel-data analysis was to test the hypothesis derived from Ricardian principles, with particular emphasis on the impact of the 2021 reduction in personal income tax on household consumption. In addition, we examined how other key determinants, including income, wealth, GDP growth, inflation, and fiscal variables, shape consumption behaviour. This approach enables a more comprehensive understanding of how taxation, income dynamics, and broader macroeconomic conditions interact to influence household spending.\u003c/p\u003e \u003cp\u003eTo facilitate interpretation, Table\u0026nbsp;\u003cspan refid=\"Tab5\" class=\"InternalRef\"\u003e5\u003c/span\u003e reports the estimated coefficients from all three estimation methods. A detailed discussion of the effects of each explanatory variable on consumption, based on the fixed effects, pooled ordinary least squares, and generalized method of moments estimations follows the summary table.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab5\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 5\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003e\u003cem\u003eThe summary of the estimated coefficients of the main model using the methods applied in our analysis\u003c/em\u003e\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eCONSUMPTION (\u0026minus;\u0026thinsp;1)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eFE\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003ePooled OLS\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eOne-Step GMM\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eTwo-Step GMM\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003e/\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003e/\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0,02725*\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0,02794*\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eWEALTH\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026minus;0,12477***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u0026minus;0,02858***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u0026minus;0,5139***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u0026minus;0,5165***\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eINCOME\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0,67239***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0,75741***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0,8635***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0,8647***\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eINTEREST_SA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0,29839\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u0026minus;1,51849***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u0026minus;0,22579\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0,17936*\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eINCOME_TAX\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026minus;6,12487***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u0026minus;2,20539***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u0026minus;2,0595***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u0026minus;1,2155***\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eINFLATION\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026minus;0,73531***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u0026minus;0,34156***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u0026minus;0,1803\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u0026minus;0,1224***\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003etGDP\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1,00986***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1,31857***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0,4431\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0,3253\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDEF_G\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0,042745\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u0026minus;0,131994\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u0026minus;1,4693***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0,0658\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003edPUBLIC_DEBT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0,04275***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u0026minus;1,48437***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u0026minus;0,0866***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u0026minus;0,0455***\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003edCOVID\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e24,7801***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e15,3397***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e55,298***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e45,349***\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003edWAR\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e53,0186***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e39,5633***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e81,712***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e64,617***\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eConstant\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e300,522***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e135,083***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e630,64***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e513,54***\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"5\"\u003e\u003cem\u003eNotes\u003c/em\u003e: The dependent variable is observed consumption. For the sake of clarity, t-statistics from the analysis of proprietary data are not included in this summary table, as was the case in previous instances. Significance levels. *** p\u0026thinsp;\u0026lt;\u0026thinsp;0,01; ** p\u0026thinsp;\u0026lt;\u0026thinsp;0,05; * p\u0026thinsp;\u0026lt;\u0026thinsp;0,1.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eIncome consistently exerts a positive and statistically significant effect on consumption across all models, including the FE and GMM estimations. This result aligns with Keynesian consumption theory, which predicts that higher disposable income increases household spending. Although the proprietary nature of our income variable limits direct comparability with prior studies, most of which rely on macroeconomic proxies, the findings are consistent with Kormendi (\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e1983\u003c/span\u003e), Bernheim (1987), and Khalid (\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e1996\u003c/span\u003e). Unlike these studies, which often rely on aggregate or imputed measures of disposable income, our analysis benefits from direct, high-frequency bank-level income data, representing an empirical contribution of the study.\u003c/p\u003e \u003cp\u003eWealth displays a negative and statistically significant relationship with consumption across the FE and GMM models. This relationship suggests that individuals with higher accumulated assets tend to consume less, potentially reflecting precautionary motives or a preference for maintaining financial buffers. Wealthier individuals may allocate a greater share of resources toward saving or investment rather than current consumption. Lucke (\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e1998\u003c/span\u003e) finds a similar negative association using macroeconomic proxies for wealth, although his estimates lack statistical significance, underscoring the importance of using granular, account-level measures as in our analysis.\u003c/p\u003e \u003cp\u003ePersonal income tax exhibits a negative and statistically significant impact on consumption in both the FE and GMM models. Reductions in personal income tax increase disposable income and, consequently, consumption, consistent with Keynesian predictions. This pattern stands in contrast to the Ricardian equivalence hypothesis, which posits that tax reductions should be offset by increased saving in anticipation of future tax liabilities. Our findings align with empirical evidence from Leiderman and Blejer (\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e1988\u003c/span\u003e), Bernheim (1987), Khalid (\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e1996\u003c/span\u003e), and Cuaresma and Reitschuler (\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2007\u003c/span\u003e), which similarly reject the Ricardian neutrality proposition.\u003c/p\u003e \u003cp\u003eThe effect of interest rates on consumption varies across specifications. The FE model yields a positive and statistically significant coefficient, whereas the pooled OLS estimate is negative and statistically significant. In the one-step GMM model the effect is negative but insignificant, while in the two-step GMM model it becomes positive and significant at the 10% level. These inconsistencies likely reflect differences in sensitivity to interest rate variation across methodologies and potential estimation noise, rather than a stable behavioural relationship.\u003c/p\u003e \u003cp\u003eInflation exerts a negative effect on consumption, with statistical significance in the FE, pooled OLS, and two-step GMM estimates. Higher inflation erodes purchasing power and may induce precautionary behaviour, resulting in reduced consumption. This finding is consistent with theoretical predictions (Samuelson, \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2009\u003c/span\u003e) and with empirical evidence from Dotsey and Sarte (\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2000\u003c/span\u003e) and Kikuchi and Nakazono (\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2020\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eGDP growth exhibits a positive effect on consumption in the FE and pooled OLS models, consistent with the notion that improved macroeconomic conditions support household expenditure. However, statistical significance is not confirmed in the GMM estimates, suggesting that the contemporaneous relationship between GDP growth and consumption may be weak once endogeneity and dynamic effects are accounted for. Nonetheless, the direction of the coefficients is consistent with prior empirical studies, including Alper (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2018\u003c/span\u003e), Lettau and Ludvigson (\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2004\u003c/span\u003e), and Attanasio and Browning (\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e1995\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eGovernment expenditure deficits display mixed effects. In the FE model the deficit ratio is insignificant, while in the one-step GMM model it is negative and statistically significant, suggesting that higher deficits may reduce consumption, potentially reflecting concerns and uncertainty over fiscal sustainability or future consolidation. This result offers limited support for Ricardian-type behaviour.\u003c/p\u003e \u003cp\u003ePublic debt exhibits divergent effects across specifications. In the FE model, public debt is positively associated with consumption, consistent with Keynesian views that debt-financed spending may stimulate economic activity. However, in both the one-step and two-step GMM models, the coefficient is negative and statistically significant, indicating that higher public debt reduces consumption, potentially due to heightened uncertainty, expectations of future fiscal tightening, or crowding-out effects. Given the robustness of the GMM estimates, the evidence suggests that elevated public debt dampens household consumption. These findings correspond with Bernheim (1987), Kormendi (\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e1983\u003c/span\u003e), Leiderman and Blejer (\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e1988\u003c/span\u003e), and Lucke (\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e1998\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe dummy variables capturing the COVID-19 pandemic and the war in Ukraine have positive and statistically significant effects across all models. Consumption increased during both periods, holding other factors constant. These patterns likely reflect shifts in expectations and heightened uncertainty, which may prompt households to adjust consumption in anticipation of future conditions. The results suggest that extraordinary macroeconomic events exert substantial influence on consumer behaviour, consistent with theories emphasizing the role of expectations and uncertainty.\u003c/p\u003e \u003cp\u003eThe findings also carry important implications for fiscal policy design. As emphasized by Esposito and Mastromatteo (\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2019\u003c/span\u003e), the absence of Ricardian equivalence underscores the need for governments to offset its absence and prioritize public spending that supports productivity, long-term growth, and financial stability. This is particularly relevant in economies where the financial sector has grown in size and complexity and where elevated public debt may heighten perceptions of macroeconomic risk. Public investment strategies should therefore extend beyond traditional infrastructure to include measures that reduce financial fragility and leverage, thereby mitigating risks associated with rising public indebtedness.\u003c/p\u003e \u003cp\u003eOur results further highlight the importance of transparent and well-coordinated fiscal and monetary policies. When tax systems are complex or the communication of fiscal measures lacks clarity, uncertainty increases and can constrain effective household decision-making. Strengthening financial literacy and improving public communication about fiscal policy can help reduce this uncertainty, build trust in economic institutions, and enable households and firms to form more accurate expectations.\u003c/p\u003e \u003cp\u003eAs Esposito and Mastromatteo (\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2019\u003c/span\u003e) observe, no universal framework exists for managing high public debt; fiscal strategies must be tailored to national institutional and economic conditions. The finding that income tax reductions raise consumption in the Czech Republic illustrates that tax policy has immediate behavioural effects and thus must be assessed within a broader fiscal context. At the same time, persistent fiscal deficits and rising public debt may dampen consumption by increasing perceived economic risk and crowding out private investment. In such an environment, public investment targeted at strengthening financial resilience and reducing leverage becomes particularly important. More broadly, the failure of Ricardian equivalence in the Czech context implies that fiscal policy retains considerable scope for stabilizing economic activity. Government expenditure can support aggregate demand during downturns, while tax measures can be adjusted countercyclically to moderate inflationary or recessionary pressures. Although debt-financed fiscal expansions may stimulate short-run activity when households do not internalize future tax obligations, such policies raise concerns about intergenerational equity and underscore the importance of avoiding excessive debt accumulation.\u003c/p\u003e \u003cp\u003eAdditionally, behavioural tendencies such as bounded rationality, hyperbolic discounting, and mental accounting may also help explain why Ricardian equivalence breaks down in practice. Although we do not directly test these mechanisms, such behaviours can limit individuals\u0026rsquo; capacity to assess future tax liabilities, encourage a preference for immediate consumption, and lead households to compartmentalize financial decisions in ways that hinder intertemporal optimization. These explanations complement more traditional Keynesian views, which emphasize liquidity constraints and a positive marginal propensity to consume as additional reasons why households do not behave as fully forward-looking agents. Taken together, these perspectives reinforce the need for transparent and clearly communicated fiscal measures: improving the salience and clarity of tax policy\u0026mdash;as suggested by Chetty, Looney, and Kroft (\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2009\u003c/span\u003e)\u0026mdash;can help individuals better understand the long-term implications of fiscal decisions, reduce uncertainty, and support more informed economic behaviour.\u003c/p\u003e"},{"header":"V. Conclusion","content":"\u003cp\u003eThis study examined the validity of Barro\u0026rsquo;s Ricardian equivalence in the Czech Republic by combining unique proprietary banking microdata with macroeconomic indicators in a quarterly panel covering 2017\u0026ndash;2023. Using fixed effects (FE), pooled OLS, and generalized method of moments (GMM) estimators, we assessed whether reductions in personal income tax led individuals to defer consumption in anticipation of future tax liabilities. The motivation for this research derives from the longstanding empirical debate, in which evidence on Ricardian behaviour has been inconclusive despite the theory\u0026rsquo;s central role in fiscal policy analysis.\u003c/p\u003e\n\u003cp\u003eAcross all empirical specifications, our findings reject Ricardian equivalence. Consumers do not reduce current consumption in response to tax cuts; instead, lower taxes systematically increase consumption, consistent with Keynesian predictions and inconsistent with Barro\u0026rsquo;s neutrality proposition. This result is robust to the use of advanced GMM techniques, which address endogeneity, heteroskedasticity, and autocorrelation, and is reinforced by the unique advantage of our dataset: access to non-public, transaction-level banking information. The availability of granular income, wealth, and consumption data allows us to circumvent measurement problems that often limit macroeconomic studies and represents a distinctive contribution to the Ricardian equivalence literature. Our conclusions also align with earlier experimental work (e.g., Slate et al., 1995; Di Laurea and Ricciuti, 2003; Meissner and Rostam-Afschar, 2014, 2017) and empirical findings based on aggregate data (Feldstein, 1982; Bernheim, 1987; Khalid, 1996; Cuaresma and Reitschuler, 2007;Leiderman and Blejer, 1988; Ricciuti, 2003 and Giorgioni and Holden (2003)).\u003c/p\u003e\n\u003cp\u003eThe roles of wealth, public deficits, and public debt offer further insights into fiscal transmission. Wealth exerts a negative and significant effect on consumption, consistent with precautionary saving motives. Government expenditure deficits and public debt show negative and significant effects in the GMM estimates, suggesting that fiscal deterioration dampens consumption, possibly reflecting heightened uncertainty or expectations of future consolidation. These results support theories emphasizing debt-induced caution and partial crowding-out, consistent with findings by Bernheim (1987), Kormendi (1983), Leiderman and Bleier (1998), Lucke (1998) and Ricciuti (2003). Taken together, the evidence suggests that consumers respond meaningfully to fiscal variables, contradicting the behavioral invariance assumed under Ricardian equivalence. External shocks such as the COVID-19 pandemic and the war in Ukraine had unexpectedly strong positive effects on consumption. This pattern likely reflects shifts in expectations, increased uncertainty, and the influence of large-scale fiscal interventions. Our findings parallel those of Georgarakos and Kenny (2022), who document that expansive fiscal packages during COVID-19 raised household income expectations and stimulated spending. The evidence challenges Ricardian neutrality and suggests that fiscal policy can meaningfully influence household consumption.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eFinally, our study highlights the value of integrating micro-level financial data with macroeconomic indicators for empirical fiscal research. Such integration also indicates that fiscal policy should incorporate behavioural patterns that influence how households perceive and react to taxes and public spending, underscoring the importance of transparent, clearly articulated, and context-tailored measures for improving policy effectiveness. Future work could extend the present analysis by incorporating data from additional financial institutions, enabling richer identification strategies and enhancing the generalizability of the results. Such research would further deepen understanding of consumption dynamics and provide strong empirical foundations for fiscal policymaking in small open economies.\u003cbr clear=\"all\"\u003e\u0026nbsp;\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eCompeting Interests:\u003c/h2\u003e\n\u003cp\u003eNone financial or non-financial interests. The funding sources had no involvement in the study design, data collection, analysis, interpretation, writing, or the decision to submit the article for publication.\u003c/p\u003e\n\u003ch2\u003eContributions\u003c/h2\u003e\n\u003cp\u003eAll authors contributed to the study conception and design.\u003c/p\u003e\n\u003ch2\u003eFunding\u003c/h2\u003e\n\u003cp\u003eDeclaration: This work was supported by the University of Piraeus Research Center and Program of the Ministry of Education and Religious Affairs of the Hellenic Republic, scientific project No. 16289 (IIS code, TA 5185218), research fellows Helena Chytilova, Andreas Fousteris, \u0026ldquo;Promoting Quality, Innovation and Extroversion in Universities, Axis 3.2: Strengthening the Digital Capabilities of Education and Modernizing Vocational Education and Training, Pillar 3: Employment, Skills and Social Cohesion,\u0026rdquo; in cooperation with the University of Piraeus, the Accounting and Auditing Laboratory of the Department of Business Administration.\u003c/p\u003e\n\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\n\u003cp\u003eAll authors contributed to the study conception and design.Helena Chytilova - Conceptualization; Methodology; Project administration; Writing \u0026ndash; original draft; Writing \u0026ndash; review and editing; Literature review; VisualizationPetr Frejlich - Conceptualization; Data Curation, Methodology; Formal Analysis; Software; Writing \u0026ndash; original draftNikolaos D. Belesis \u0026ndash; Conceptualization; Supervision; Project administration; InvestigationAndreas E. Fousteris - Conceptualization; Investigation; Data Curation\u003c/p\u003e\n\u003ch2\u003eAcknowledgement\u003c/h2\u003e\n\u003cp\u003eAcknowledgements: The authors gratefully acknowledge the academic environment and institutional support provided by their home universities, including the University of Piraeus, \u0026Scaron;koda Auto University, and the Prague University of Economics and Business. This work was also supported by the University of Piraeus Research Center and Program of the Ministry of Education and Religious Affairs of the Hellenic Republic, scientific project No. 16289 (IIS code, TA 5185218), research fellows Helena Chytilova and Andreas E. Fousteris, \u0026ldquo;Promoting Quality, Innovation and Extroversion in Universities, Axis 3.2: Strengthening the Digital Capabilities of Education and Modernizing Vocational Education and Training, Pillar 3: Employment, Skills and Social Cohesion,\u0026rdquo; in cooperation with the University of Piraeus, the Accounting and Auditing Laboratory of the Department of Business Administration\u003c/p\u003e\n\u003ch2\u003eData Availability\u003c/h2\u003e\n\u003cp\u003eThe data used in this study consist of confidential banking records that are proprietary and protected under the EU General Data Protection Regulation (GDPR). Due to legal and contractual restrictions, these data cannot be shared publicly or made available to third parties. Although we would gladly provide access, we are not permitted to disclose the dataset in any form.\u003c/p\u003e\n\u003ch2\u003e\u003cstrong\u003eDeclaration of generative AI and AI-assisted technologies in the writing process.\u0026nbsp;\u003c/strong\u003e\u003c/h2\u003e\n\u003cp\u003eDuring the preparation of this work the author(s) used ChatGPT in order to improve language and readability. After using this tool/service, the author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the publication.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n \u003cli\u003eAdji, A., Alm, J. and Ferraro, P. J. (2009) \u0026lsquo;Experimental Tests of Ricardian Equivalence with Distortionary Versus Non-Distortionary Taxes\u0026rsquo;, Economics Bulletin, 29(4), 2556\u0026ndash;2572.\u003c/li\u003e\n \u003cli\u003eAlper, E. A. (2018) \u0026lsquo;The Relationship of Economic Growth with Consumption, Investment, Unemployment Rates, Saving Rates and Portfolio Investments in the Developing Countries\u0026rsquo;, Gaziantep University Journal of Social Sciences, 17(3), 980\u0026ndash;987.\u003c/li\u003e\n \u003cli\u003eAttanasio, O. P. and Browning, M. (1995) \u0026lsquo;Consumption over the Life Cycle and over the Business Cycle\u0026rsquo;, The American Economic Review, 85(5), 1118\u0026ndash;1137.\u003c/li\u003e\n \u003cli\u003eBailey, M. J. (1962) \u0026lsquo;The Welfare Cost of Inflationary Finance\u0026rsquo;, Journal of Political Economy, 70(3), 244\u0026ndash;258.\u003c/li\u003e\n \u003cli\u003eBarro, R. J. (1974) \u0026lsquo;Are Government Bonds Net Wealth?\u0026rsquo;, Journal of Political Economy, 82(6), 1095\u0026ndash;1117.\u003c/li\u003e\n \u003cli\u003eBarro, R. J. (1996) \u0026lsquo;Reflections on Ricardian Equivalence\u0026rsquo;, Journal of Economic Perspectives, 10(1), 36\u0026ndash;49.\u003c/li\u003e\n \u003cli\u003eBuchanan, J. (1976) \u0026lsquo;Barro on the Ricardian Equivalence Theorem\u0026rsquo;, Journal of Political Economy, 84(2), 337\u0026ndash;342.\u003c/li\u003e\n \u003cli\u003eCadsby, C. B. and Frank, M. (1991) \u0026lsquo;Experimental Tests of Ricardian Equivalence\u0026rsquo;, Economic Inquiry, 29(4), 645\u0026ndash;664.\u003c/li\u003e\n \u003cli\u003eCampbell, Y. J. and Mankiw, N. G. (1991) \u0026lsquo;The Response of Consumption to Income: A Cross-Country Investigation\u0026rsquo;, European Economic Review, 35(4), 723\u0026ndash;756.\u003c/li\u003e\n \u003cli\u003eCuaresma, C. J. and Reitschuler, G. (2007) \u0026lsquo;Is the Ricardian Equivalence Proposition an \u0026ldquo;Aerie Fairy\u0026rdquo; Theory for Europe?\u0026rsquo;, Economica, 74(296), 682\u0026ndash;694.\u003c/li\u003e\n \u003cli\u003eDi Laurea, D. and Ricciuti, R. (2003) \u0026lsquo;An Experimental Analysis of Two Departures from Ricardian Equivalence\u0026rsquo;, Economics Bulletin, 8(11), 1\u0026ndash;11.\u003c/li\u003e\n \u003cli\u003eDotsey, M. and Sarte, P. (2000) \u0026lsquo;Inflation Uncertainty and Growth in a Cash-in-Advance Economy\u0026rsquo;, Journal of Monetary Economics, 45(3), 631\u0026ndash;655.\u003c/li\u003e\n \u003cli\u003eEsposito, L. and Mastromatteo, G. (2019) Defaultnomics: Making Sense of the Barro-Ricardo Equivalence in a Financialized World. Working paper, No. 933.\u003c/li\u003e\n \u003cli\u003eFeldstein, M. S. and Elmendorf, D. W. (1990) \u0026lsquo;Government Debt, Government Spending, and Private Sector Behavior Revisited: Comment\u0026rsquo;, The American Economic Review, 80, 589\u0026ndash;599.\u003c/li\u003e\n \u003cli\u003eFriedman, M. (1957) A Theory of the Consumption Function. Princeton: Princeton University Press.\u003c/li\u003e\n \u003cli\u003eGeorgarakos, D. and Kenny, G. (2022) \u0026lsquo;Household Spending and Fiscal Support during the COVID-19 Pandemic: Insights from a New Consumer Survey\u0026rsquo;, Journal of Monetary Economics, 129, S1\u0026ndash;S14.\u003c/li\u003e\n \u003cli\u003eGiorgioni, G. and Holden, S. (2003) \u0026lsquo;Political Institutions and the Stability of Government Debt\u0026rsquo;, Journal of Economic Surveys, 17(4), 487\u0026ndash;516.\u003c/li\u003e\n \u003cli\u003eHallett, A. J. H. and Ma, Y. (1994) \u0026lsquo;On the Accuracy and Efficiency of GMM Estimators: A Monte Carlo Study\u0026rsquo;, in Belsley, D. A. (ed.) Computational Techniques for Econometrics and Economic Analysis. Advances in Computational Economics, 3, 22\u0026ndash;44.\u003c/li\u003e\n \u003cli\u003eHan, C. and Kim, H. (2023) \u0026lsquo;Dynamic Panel GMM Estimators with Improved Finite Sample Properties Using Parametric Restrictions for Dimension Reduction\u0026rsquo;, Empirical Economics, 64(6), 2589\u0026ndash;2610.\u003c/li\u003e\n \u003cli\u003eChetty, R., Looney, A. and Kroft, K. (2009) \u0026lsquo;Salience and Taxation: Theory and Evidence\u0026rsquo;, The American Economic Review, 99(4), 1145\u0026ndash;1177.\u003c/li\u003e\n \u003cli\u003eKelton, S. (2020) The Deficit Myth: Modern Monetary Theory and the Birth of the People\u0026rsquo;s Economy. New York: PublicAffairs.\u003c/li\u003e\n \u003cli\u003eKeynes, J. M. (1936) The General Theory of Employment, Interest, and Money. London: Macmillan.\u003c/li\u003e\n \u003cli\u003eKhalid, M. A. (1996) \u0026lsquo;Ricardian Equivalence: Empirical Evidence from Developing Economies\u0026rsquo;, Journal of Development Economics, 51, 413\u0026ndash;432.\u003c/li\u003e\n \u003cli\u003eKikuchi, J. and Nakazono, Y. (2020) \u0026lsquo;Inflation Expectations and Consumer Spending: Micro-Data Evidence\u0026rsquo;. ESRI Discussion Paper Series, No. 353.\u003c/li\u003e\n \u003cli\u003eKormendi, C. R. (1983) \u0026lsquo;Government Debt, Government Spending and Private Sector Behaviour\u0026rsquo;, American Economic Review, 73, 994\u0026ndash;1010.\u003c/li\u003e\n \u003cli\u003eKrugman, P. R. (1988) \u0026lsquo;Financing vs. Forgiving a Debt Overhang\u0026rsquo;, Journal of Development Economics, 29(3), 253\u0026ndash;268.\u003c/li\u003e\n \u003cli\u003eLeiderman, L. and Blejer, M. (1988) \u0026lsquo;Modeling and Testing Ricardian Equivalence: A Survey\u0026rsquo;, IMF Economic Review, 35, 1\u0026ndash;35.\u003c/li\u003e\n \u003cli\u003eLettau, M. and Ludvigson, C. S. (2004) \u0026lsquo;Understanding Trend and Cycle in Asset Values: Reevaluating the Wealth Effect on Consumption\u0026rsquo;, American Economic Review, 94(1), 276\u0026ndash;299.\u003c/li\u003e\n \u003cli\u003eLucke, B. (1998) \u0026lsquo;Econometric Tests of Ricardian Equivalence: Results for Germany\u0026rsquo;, FinanzArchiv / Public Finance Analysis, 55(3), 288\u0026ndash;314.\u003c/li\u003e\n \u003cli\u003eMankiw, N. G. (1999) \u0026lsquo;The Saver\u0026rsquo;s Paradox\u0026rsquo;, Journal of Economic Perspectives, 13(3), 13\u0026ndash;24.\u003c/li\u003e\n \u003cli\u003eMankiw, N. G. (2016) Macroeconomics. New York: Worth Publishers.\u003c/li\u003e\n \u003cli\u003eMankiw, N. G. (2000) \u0026lsquo;The Savers-Spenders Theory of Fiscal Policy\u0026rsquo;, American Economic Review, 90(2), 120\u0026ndash;125.\u003c/li\u003e\n \u003cli\u003eMeissner, T. (2013) Intertemporal Consumption and Debt Aversion: An Experimental Study. SFB 649 Discussion Paper, No. 2013\u0026ndash;045.\u003c/li\u003e\n \u003cli\u003eMeissner, T. and Rostam-Afschar, D. (2014) Do Tax Cuts Increase Consumption? An Experimental Test of Ricardian Equivalence. Discussion Paper, No. 2014/16.\u003c/li\u003e\n \u003cli\u003eMeissner, T. and Rostam-Afschar, D. (2017) \u0026lsquo;Learning Ricardian Equivalence\u0026rsquo;, Journal of Economic Dynamics and Control, 82, 273\u0026ndash;288.\u003c/li\u003e\n \u003cli\u003eMertens, K. and Ravn, M. O. (2013) \u0026lsquo;The Dynamic Effects of Personal and Corporate Income Tax Changes in the United States\u0026rsquo;, American Economic Review, 103(4), 1212\u0026ndash;1247.\u003c/li\u003e\n \u003cli\u003eModigliani, F. and Brumberg, R. (1954) \u0026lsquo;Utility Analysis and the Consumption Function: An Interpretation of Cross-Section Data\u0026rsquo;, in Kurihara, K. K. (ed.) Keynesian Economics. New Brunswick: Rutgers University Press.\u003c/li\u003e\n \u003cli\u003ePatinkin, D. (1965) Money, Interest and Prices: An Integration of Monetary and Value Theory. 2nd edn. Evanston and New York: Row, Peterson and Co. and Harper and Row.\u003c/li\u003e\n \u003cli\u003eRicardo, D. (1951) \u0026lsquo;Funding System\u0026rsquo;, in Sraffa, P. (ed.) The Works and Correspondence of David Ricardo. Vol. 4. Cambridge: Cambridge University Press.\u003c/li\u003e\n \u003cli\u003eRicciuti, R. (2003) \u0026lsquo;Assessing Ricardian Equivalence\u0026rsquo;, Journal of Economic Surveys, 17(1), 55\u0026ndash;78.\u003c/li\u003e\n \u003cli\u003eSamuelson, P. A. (1974) \u0026lsquo;Theoretical Notes on Trade Problems\u0026rsquo;, Review of Economics and Statistics, 46(2), 145\u0026ndash;154.\u003c/li\u003e\n \u003cli\u003eSamuelson, P. A. and Nordhaus, W. D. (2009) Economics. New York: McGraw-Hill.\u003c/li\u003e\n \u003cli\u003eSeater, J. J. (1993) \u0026lsquo;Ricardian Equivalence\u0026rsquo;, Journal of Economic Literature, 31(1), 142\u0026ndash;190.\u003c/li\u003e\n \u003cli\u003eShapiro, M. D. and Slemrod, J. (2003) \u0026lsquo;Consumer Response to Tax Rebates\u0026rsquo;, American Economic Review, 93(1), 381\u0026ndash;396.\u003c/li\u003e\n \u003cli\u003eSlate, S., McKee, M., Beck, W. and Alm, J. (1995) \u0026lsquo;Testing Ricardian Equivalence under Uncertainty\u0026rsquo;, Public Choice, 85(1\u0026ndash;2), 11\u0026ndash;29.\u003c/li\u003e\n \u003cli\u003eSouleles, N. S. (1999) \u0026lsquo;The Response of Household Consumption to Income Tax Refunds\u0026rsquo;, American Economic Review, 89(4), 947\u0026ndash;958.\u003c/li\u003e\n \u003cli\u003eStiglitz, J. E. (2000) Economics of the Public Sector. New York: W. W. Norton \u0026amp; Company.\u003c/li\u003e\n \u003cli\u003eThaler, R. H. and Sunstein, C. R. (2008) Nudge: Improving Decisions About Health, Wealth, and Happiness. London: Penguin Books.\u003c/li\u003e\n \u003cli\u003eTobin, J. (1952) \u0026lsquo;A Survey of the Theory of Rationing\u0026rsquo;, Econometrica, 20, 521\u0026ndash;553.\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Footnotes","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003e The negative impact on consumption suggests that a higher government expenditure deficit reduces consumer spending. An increase in the deficit may raise concerns about fiscal sustainability and decrease consumer confidence in the economy, leading to more cautious consumption expenditures.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"review-of-economics-of-the-household","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"reho","sideBox":"Learn more about [Review of Economics of the Household](http://link.springer.com/journal/11150)","snPcode":"11150","submissionUrl":"https://submission.nature.com/new-submission/11150/3","title":"Review of Economics of the Household","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Ricardian Equivalence, Taxation, Real Banking Data, Intertemporal Consumption, Fixed Effects model, Generalized Method of Moments","lastPublishedDoi":"10.21203/rs.3.rs-9246758/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9246758/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThis study tests the Ricardian Equivalence in the Czech Republic using a unique proprietary dataset of real banking data from 2017 to 2023. Applying panel fixed-effects models and generalized method of moments estimation, we find a robust and economically significant negative effect of personal income taxation on household consumption. The results suggest that individuals do not offset tax-induced income changes by increasing savings, providing evidence against Ricardian behaviour. By linking macroeconomic perspectives with high-frequency banking data, the paper makes a novel contribution to fiscal policy research and offers relevant implications for the design of consumption-related tax policies.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eJEL Classification:\u003c/strong\u003e H3, E6, H8\u003c/p\u003e","manuscriptTitle":"Ricardian Equivalence and Household Consumption: Evidence from Administrative Banking Microdata","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-04-02 19:10:46","doi":"10.21203/rs.3.rs-9246758/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"reviewerAgreed","content":"92655407268712034666866302035049877616","date":"2026-04-01T07:50:09+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"19484552516286536513971137740067152253","date":"2026-03-31T13:47:39+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-03-30T06:24:49+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-03-28T14:22:45+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-03-28T14:21:55+00:00","index":"","fulltext":""},{"type":"submitted","content":"Review of Economics of the Household","date":"2026-03-27T16:03:26+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"review-of-economics-of-the-household","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"reho","sideBox":"Learn more about [Review of Economics of the Household](http://link.springer.com/journal/11150)","snPcode":"11150","submissionUrl":"https://submission.nature.com/new-submission/11150/3","title":"Review of Economics of the Household","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"f05d2699-7fc6-447b-8775-7f74e7dc27e6","owner":[],"postedDate":"April 2nd, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-04-02T19:10:46+00:00","versionOfRecord":[],"versionCreatedAt":"2026-04-02 19:10:46","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9246758","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9246758","identity":"rs-9246758","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}