Ideological alignment, public sector size and tax morale: empirical evidence from OECD economies

The study investigates how ideological alignment between citizens and their governments shapes citizens’ intrinsic willingness to pay taxes (tax morale), especially under varying sizes of the public sector. The motivation stems from large fiscal losses from tax evasion/avoidance and the role of compliance in enabling redistribution. The authors posit that right-leaning citizens generally prefer smaller government and lower public spending, while left-leaning citizens prefer larger public sectors. Therefore, ideological misalignment combined with perceptions of an oversized or undersized public sector may alter tax morale. The paper sets out to empirically test this mechanism using cross-country survey data and macro indicators and to quantify its relevance for tax morale in developed economies.

The traditional framework for tax evasion (Allingham and Sandmo, 1972) emphasizes deterrence, but more recent work highlights non-pecuniary determinants grouped under tax morale (Torgler, 2007; Luttmer and Singhal, 2014). Two main channels are discussed: (1) personal/social norms (intrinsic motivation, peer effects, culture, religiosity), with evidence from church tax compliance (Dwenger et al., 2016), religiosity (Grasmick et al., 1991; Torgler, 2006), peer influence (Sigala et al., 1999), and cultural/immigration effects (Halla, 2012; Kountouris and Remoundou, 2013); and (2) reciprocity, where tax morale rises with satisfaction in public goods provision, proxied by institutional quality measures such as trust in government (Alm et al., 2006; Torgler, 2004), redistribution/progressivity (Doerrenberg and Peichl, 2013), and ethnic fractionalization (Lago-Peñas and Lago-Peñas, 2010). Prior research often uses broad institutional indicators rather than the specific alignment between citizens’ preferences and government outcomes. Related work on ideological alignment and compliance finds that higher approval or political alignment is associated with higher reported income/lower evasion in the US (Cebula, 2013; Cullen et al., 2021). The present study differs by focusing on tax morale (not observed compliance), using individual-level data across 23 OECD countries, and examining ideological distance interacting with public sector size.

Data: Individual-level data from the combined World Values Survey (WVS) and European Values Study (EVS), yielding approximately 79,189 observations from 23 OECD countries spanning 1995–2018. Country-level variables come from OECD and World Bank sources; government ideology from ParlGov. Eastern European countries are excluded due to differing fairness and evasion patterns. Key variables:

• Dependent variable (Tax morale): Based on the item “Cheating on tax if you have the chance,” originally a 1–10 scale (1 = never justify, 10 = always). Re-coded into a four-category scale for ordered probit: 7–10 → 0 (lowest), 5–6 → 1, 3–4 → 2, 1–2 → 3 (highest). Robustness uses the original 10-point scale and a [0,1] rescaling for GLM.
• Ideology: Self-reported left–right scale (EVS/WVS), used to construct distances to government ideology. Government ideology measured on a 0–10 left–right scale using ParlGov; for coalitions, the unweighted mean of cabinet parties (robustness: weighted by seats, or leading party’s ideology).
• Public sector size: General government expenditure as share of GDP (OECD). Also define mean-adjusted indicators: Public sector oversized (size above OECD average) and Public sector undersized (below average), used in interactions.
• Ideological distance to government: Right to government and Left to government are absolute distances between respondent’s ideology and government ideology when the respondent is to the right or to the left of the government, respectively.
• Interactions: Right to government × Public sector oversized; Right to government × Public sector undersized; Left to government × Public sector oversized; Left to government × Public sector undersized.
• Controls: Income (10-point self-placement), Government effectiveness (WGI, −2.5 to 2.5), Inflation, Unemployment; individual controls include Age, Female, Religious, Patriotic, education (Medium, High vs. Low), occupational status (Unemployed, Self-employed, Retired, Other vs. Employed), marital status (Married/partnership, Divorced/Separated, Widowed vs. Never married). Country and time fixed effects included; standard errors clustered by country.

Empirical models:

• Primary: Ordered probit for the four-category Tax morale with country and year dummies.
• Secondary: OLS on a continuous proxy and GLM on [0,1] scaled outcome for robustness.

Endogeneity handling:

• Ideology (and related terms) may be endogenous. The study applies Lewbel’s (2012) heteroskedasticity-based identification to generate instruments from exogenous regressors exploiting heteroskedasticity, enabling IV estimation via efficient GMM. Instrument validity assessed using Kleibergen-Paap LM and Wald F tests and Hansen J test.

Main ordered-probit results (Table 4, Spec I):

• Ideology and Public sector size main effects are not statistically significant.
• Right to government × Public sector oversized is negative and significant (p<0.01). The marginal effect indicates that when the public sector is 20% above the OECD average, a citizen 5 points to the right of the government has a 2.4 percentage-point lower probability of being in the highest tax morale category; at the maximum observed interaction value, the effect is roughly four times larger. This mean effect is comparable in magnitude to moving from low to high education.
• Left to government × Public sector oversized is positive and marginally significant (p<0.10). Interactions with undersized public sector are directionally consistent but weaker.
• Controls: Higher tax morale among older respondents, women, religious and patriotic individuals; higher education (High) positive; unemployed and self-employed negative; retired and married positive; divorced negative. Macroeconomic controls largely insignificant in Spec I.

IV results using Lewbel instruments (Table 4, Spec III):

• Public sector size becomes significantly negative (−0.325, p<0.01) for tax morale.
• Right to government × Public sector oversized remains negative and significant (−0.072, p<0.01); Right to government × Public sector undersized also negative (−0.031, p<0.05).
• Left to government × Public sector oversized and × undersized are positive and significant (~0.039–0.042, p<0.01).
• Government effectiveness appears negative and significant in IV (−0.092, p<0.01); Inflation small positive; Unemployment small negative; Age, Female, Religious, Patriotic, High education remain positive; Unemployed and Self-employed negative.
• Instrument diagnostics reported as satisfactory (underidentification rejected; weak instrument concerns alleviated; overidentification not rejected according to the paper).

Robustness checks:

• Outcome scale robustness: Results hold using the original 10-point scale and GLM on [0,1] (Table 5).
• Alternative measures of government ideology (weighted coalition mean; leading party ideology) yield similar signs and significance (Table 6).
• Excluding countries with extreme average tax morale (BEL, LUX, IRE; DEN, JAP, TUR) leaves the main interaction results essentially unchanged despite reduced sample size (Table 7).

Descriptive context:

• Sample size: 79,189 individuals across 23 OECD countries (1995–2018).
• Public sector size mean 44.7% of GDP (range: 31.0%—61.4%).

Findings support the hypothesis that ideological alignment interacts with perceived public sector size to shape intrinsic taxpaying motivations. When citizens’ ideology is to the right of the government and the public sector is larger than the OECD average, tax morale declines; conversely, when citizens are to the left of the government and the public sector is smaller than average, tax morale rises. This aligns with broader evidence that reciprocity and perceived congruence between public goods provision and preferences condition tax morale. The effects are nontrivial in magnitude and robust to alternative outcome codings, government ideology measures, and subsample exclusions. Addressing potential endogeneity with Lewbel’s IV approach strengthens the pattern and suggests the relationships are not solely driven by omitted-variable bias. The results imply that political polarization and policy dissatisfaction can erode tax morale, potentially undermining voluntary compliance, even if enforcement moderates actual evasion. These dynamics underscore the importance for policymakers of aligning fiscal policy with citizen preferences or improving perceived fairness/efficiency to maintain tax morale.

The study demonstrates that ideological misalignment with governing coalitions, combined with deviations of public sector size from the OECD average, significantly relates to tax morale: right-leaning citizens relative to the government show lower tax morale when the public sector is oversized, whereas left-leaning citizens show higher tax morale when it is undersized. These insights extend tax morale research by linking individual ideology, policy outcomes (public sector size), and intrinsic compliance motivations across multiple OECD countries using individual-level data. Given rising ideological polarization, the results suggest risks of declining tax morale and, potentially, compliance. Future research should examine specific spending categories and policy mixes, how they interact with ideology to shape tax morale, and the downstream effects on both legal tax avoidance and illegal evasion, ideally with designs that better identify causal effects.

• Causality: Despite IV strategies, the authors note evidence is insufficient to claim a causal relationship; results indicate strong correlations.
• Measurement: Tax morale is self-reported and subject to social desirability or interpretation biases; cross-country evasion data are scarce, limiting validation against behavioral measures.
• Government ideology measurement: Coalition ideology is proxied (unweighted/weighted means or leading party), which may not fully capture policy stance nuances.
• Public sector size benchmark: Oversized/undersized is defined relative to the OECD average, which may not reflect country-specific optimal sizes.
• Sample scope: Analysis limited to 23 OECD countries; Eastern European countries excluded; survey waves end in 2018 due to data availability.
• Potential omitted factors: Although extensive controls are included, unobserved country-time factors or ideological shifts could influence both tax morale and fiscal policy perceptions.