The Impact of Gender on Tax Compliance in Southern Albania

1. Introduction

Tax compliance-declaring one’s income and paying due taxes on time-stands at the core of public finance and increasingly of behavioral finance. According to the classical deterrence theory, the taxpayer is a rational decision-maker who weighs expected penalties against the gains from evasion. However, many people comply even if audit probabilities are low and detection is virtually impossible. This empirical regularity pushed the field toward a broader perspective that includes psychology and social context: norms, emotions, risk preferences, and identity help explain why taxpayers follow-or do not follow-the rules. Further, recent studies have gone ahead to show how transient emotions and stable personality traits meaningfully shape compliance. Fochmann, Hechtner, Kirchler, & Mohr (2025); Lukovszki, Szabó, & Tóth (2025)

Within this behavioral turn, gender is one of the most obvious dimensions to consider. Different studies show that systematic differences are found between males and females regarding risk tolerance, reputational concerns, and planning horizons (Croson & Gneezy 2009; Barber & Odean 2001). As women, averagely have higher risk aversion and sensitiveness to social norms, they tend as well to display higher tax morale and probability of timely payment. International studies conclude that businesses ran by women in Rwanda display higher compliance (Twesige et al., 2024). Similar patterns appear in Ethiopia (Yiman & Asmare, 2020). But again, there is little direct evidence from Southeastern Europe.

Albania is a particularly good fit for studying these questions. The informal economy is large-often estimated at over one-third of GDP, World Economics-and enforcement varies across region and industry. Southern Albania’s economy is mainly characterized by services and trade, and in the last years, the number of businesswomen has been heavily increasing. However, the domestic literature has focused more on the features of institutional reforms, digitalization, and corruption (Hoxhaj & Kamolli, 2022; Muharremi, Salé & Hoxhaj, 2022) than micro-level behavior. Important as it is, this research leaves open whether risk, norms, and perceived control translate into gender differences in actual payment data. To our best knowledge, no previous study in Albania-or indeed in the wider Western Balkans-has examined gendered tax behavior using administrative microdata rather than surveys, thus leaving a clear empirical gap.

This study fills the gap by integrating behavioral finance with classical deterrence and the Theory of Planned Behavior. The basic premise is that intentions determined by attitudes, subjective norms, and perceived control (Ajzen, 1985), in addition to internal tax morale (Torgler, 2003), should materialize in observable compliance outcomes, especially where monitoring is imperfect. We draw on administrative data from 500 self-employed taxpayers in Fier, Vlorë, Berat, Gjirokastër, and Sarandë between January 2022 and March 2025 to test three hypotheses: H1: Gender predicts the probability of timely payment; H2: Gender predicts the level of unpaid tax debt; H3: The sectoral context moderates these effects through a Gender × Sector interaction.

Methodologically, we estimate logistic and probit models for timely payments and use linear regression for unpaid liabilities, which allows us to capture both the extensive margin-whether a taxpayer pays on time-and the intensive margin-how much arrears accumulate. Anticipating the results, we find that female-led businesses are more likely to meet deadlines and hold lower unpaid debts than male-led firms. These differences persist after controlling for firm size, income, region, and time and remain visible across all sectors. By recognizing gender as a behavioral determinant, revenue administrations can design targeted and gender-sensitive interventions-such as integrating gender into compliance risk models and service tailoring-that raise voluntary payments and reduce arrears without heavier auditing.

I. behavioral finance as applied to tax compliance,

II. gender economics in developing and transition economies, and

III. Evidence-based, EU-aligned tax policy. Beyond deterrence alone, it provides a fuller account of how taxpayers actually make financial decisions in a transitioning economy.

2. Literature Review

Understanding taxpayer compliance and evasion has evolved from a narrow, cost–benefit perspective to one that is broader and interdisciplinary, combining economics with psychology and sociology. We next review the principal theories and evidence on gendered compliance, focusing on insights relevant to Southern Albania.

2.1. From Deterrence to Behavioral Finance

The seminal reference frame of Allingham and Sandmo (1972) considers taxpayers as expected-utility maximizers who compare the gain from evasion with the expected punishment. While being a foundational work, this approach cannot satisfactorily explain the usually high level of compliance that is observed in environments characterized by rather weak audit systems (Torgler & Schneider, 2009). Accordingly, specific aspects of behavioral economics and finance became part of later studies that focused on moral norms, perceived fairness, risk attitudes, and trust in institutions (Alm & Torgler, 2006; Kirchler, 2007).

In this manner, compliance expresses values and expectations instilled within communities, in addition to incentives. The related concept here is one of tax morale-the internalized belief that paying taxes is a civic duty-which is more resilient where institutions are perceived as more legitimate. Ajzen’s 1985 Theory of Planned Behaviour further explains how attitudes, subjective norms, and perceived behavioral control create intentions and actions. In contexts where enforcement is uneven, these social and psychological levers could be more important than raw deterrence in explaining compliance even when risks of detection are low.

2.2. International Evidence on Gender and Tax Compliance

Gender naturally enters this behavioral explanation. Systematic variation in risk tolerance, reputational concern, and long-term planning exists, which has been recorded in finance research (Croson & Gneezy, 2009; Barber & Odean, 2001). Consistently, microlevel studies have found that women tend to display higher tax morale and strong prosocial orientation.

Indeed, D’Attoma, Volintiru, and Steinmo (2017) have found evidence that women in Europe and the United States tend to better comply with tax law. D’Attoma, Volintiru, and Malézieux (2020) relate the social-value orientation of women to voluntary compliance. Then, in developing contexts, there is civic duty, perceived responsibility, and tax knowledge; all these factors have been found to raise women’s compliance with tax laws (Twesige et al., 2024). In the same line of argument, a large-scale field experiment in Argentina shows that women pay more and earlier, especially if the bill value is smaller (López-Luzuriaga & Scartascini, 2023).

A similar pattern is also echoed in the recent World Bank knowledge note on women’s tax morale and fiscal responsibility. The pattern is quite robust from a cross-country perspective, with effects being often strongest where cash use is prevalent and the quality of governance uneven

2.3. Evidence from Albania and Southeastern Europe

Despite the fact that Albania has a very large informal economy, which is often estimated at over one-third of GDP (INSTAT 2023; World Economics 2023), gender has rarely been examined as a primary driver of taxpayer behavior. Domestic research has focused more on institutional factors: regulatory design, digitalization, and corruption (Hoxhaj & Kamolli 2022; Muharremi, Salé, & Hoxhaj 2022). This research is important in understanding the “rules of the game,” but it does shed less light on micro-level decisions by individual taxpayers.

This omission is important because women’s entrepreneurship has grown, especially in services, hospitality, and small trade in Southern Albania. These activities usually imply frequent customer contact and visible reputational effects, conditions under which moral norms and peer expectations can support timely payment.

However, no research has been conducted in Albania using administrative payment data to actually test whether gender-based differences in risk, norms, and perceived control translate into observed compliance. In the post-socialist region in general, women tend to report higher tax morale than men do (Torgler & Valev, 2010), but these results often rely on surveys that are sensitive to social desirability and recall biases. Administrative data provide an objective lens and a cleaner test for the Albanian context.

2.4. Sectoral and Contextual Interactions

The impact of gender, however, is unlikely to be consistent across the different sectors. In most economies, men dominate in construction, transport, and wholesale trade-characterized by higher informality, cash intensity, and evasion opportunities-whereas women are more common in education, health, and personal services, which tend to be more formal and closely monitored (OECD, 2023).

These structural patterns shape the incentives people face and can magnify or mask intrinsic gender differences in risk-taking and ethical behavior. Empirical work supports this view: a female compliance advantage is often most visible in retail and services, where client relationships and reputation loom large, and weakest in construction-a sector consistently flagged for VAT-related risks and targeted enforcement (European Investment Bank, 2022; Junttila, Koivisto, & Nivala, 2024).

Considering disparities by regions and segmentation by sectors, Gender × Sector interactions are of particular importance in interpreting results for Albania.

2.5. Methodological Advances

Prior studies in Albania have used surveys based on perceptions, which provide information, but, risk to overstate civic duty and are vulnerable to recall bias (Alm, McClelland, & Schulze, 1992). By contrast, administrative tax data capture real behavior and enable more thorough econometric designs.

Binary response models-logit and probit-fit the issue of on-time versus delayed payment (Horowitz & Savin, 2001) while linear models suit continuous outcomes like stocks of unpaid debt. By using both, this helps in separating the extensive margin of whether taxpayers comply from the intensive one of how much arrears accumulate, thus shedding light on mechanisms.

Recent econometric contributions also explain under which conditions linear probability models approximate nonlinear indices and how to interpret marginal effects (Chen, Martin, & Wooldridge, 2023). In a nutshell, administrative microdata in conjunction with auxiliary models allow one to connect behavioral theory with the observed compliance in a transparent manner.

2.6. Contribution of the Present Study

Within this landscape, the present study contributes in three ways.

First, to our knowledge, it is the first in Albania to pair gender-disaggregated administrative payment data with econometric models of compliance behavior.

Second, it enriches behavioral finance by linking risk aversion, reputational concern, and prosocial orientation to observed outcomes rather than self-reported intentions.

Third, it provides policy-relevant evidence for a transition economy pursuing EU alignment: if gender reliably predicts timely payment and lower arrears, then targeted and gender-sensitive tools — such as incorporating gender into risk assessment and taxpayer services — can raise voluntary compliance without heavier audits.

Together, these contributions move beyond deterrence alone toward a fuller, behaviorally informed description of tax payment in Southern Albania.

3. Methodology

3.1. Data Source

We use administrative data from the Regional Directorate of Taxes of the Southern Region of Albania covering 500 self-employed taxpayers in Fier, Vlorë, Berat, Gjirokastër, and Sarandë for the period January 2022 – March 2025.

The advantage of the administrative data is immediately apparent: they capture actual payment behavior and not perceptions, avoiding common survey problems such as response and recall bias. For example, the sample of active self-employed individuals has complete payment histories within the observation window, so each observation can be compared to all others.

All data were provided in anonymized form and analyzed under a data-sharing agreement with the authority.

3.2. Description of Variables

We consider two primary outcomes:

• Tax compliance: binary, equal to 1 if the taxpayer paid within the legal deadline, 0 otherwise.

The indicator is built from official payment timestamps related to the statutory due dates.

• Tax debt (continuous): the stock of outstanding arrears for each taxpayer.

The most important covariates are:

• Gender: dummy variable (1 = female, 0 = male).

• Sector of Economy: groups by NACE (trade, services, small industry, transport, etc.).

• Gender × Sector: an interaction term testing whether gender effects vary according to the type of sector.

Control variables:

• Number of employees (firm size)

• Annual income from tax records

• Region (city dummies)

Fiscal year (time controls)

We checked for sensitivity to outliers, looking at the upper tail of the distributions of income and arrears; trimming these variables leaves the conclusions unchanged.

3.3. Econometric Models

We combine binary response and linear models in a way that captures both extensive and intensive margins of compliance.

• Logit model for timely payment:

logit{P(Yi=1)}=β0+β1Genderi+β2Sectori+β3(Gender×Sector)i+βkControlsi+εi

We report odds ratios for interpretability.

• Probit model

Used as a specification check, with average marginal effects reported.

• OLS model for unpaid tax debt:

TaxDebt=α0+α1Gender+α2Sector+α3(Gender×Sector)+αkControls+ui

All models include fixed effects for city and year to absorb any spatial and temporal heterogeneity.

In all specifications, we compute heteroskedasticity-robust standard errors.

3.4. Research Hypotheses

We test three related claims:

• H1: Gender affects the probability of timely payment.
• H2: Gender affects the level of unpaid tax debt.
• H3: The Gender × Sector interaction moderates these relationships.

The three hypotheses access the direction of the gender impact (female advantage) as well as its contextual dependence across sectors.

3.5. Analytical Procedures

The analysis proceeds in two steps.

First, we present descriptive statistics by gender, sector, and region.

Second, we estimate the above models and conduct standard diagnostics:

• VIF for multicollinearity
• Breusch–Pagan test for heteroskedasticity
• Pseudo-R² for logit/probit model fit

We also re-estimate after excluding extreme delays and under alternative sectoral aggregations; the main effects remain robust.

All computations are performed in R.

4. Results

Descriptive views already show a strong gender pattern. Timely fulfillment of obligations is 81.5% for women versus 61.2% for men — a substantial gap that anticipates the regression results. Average unpaid liabilities are also notably lower for women (93,000 lek) than for men (156,000 lek).

Women run slightly smaller firms (2.8 versus 3.2 employees) and report somewhat lower annual income, but the main differences concern deadlines and arrears — two outcomes directly tied to compliance behavior rather than just business scale.

The positive relationship between being female-led and paying on time is confirmed by both the logit (odds ratios) and probit (coefficients/marginal effects). Translated into the probability, the estimated effect represents a roughly 30% higher likelihood of timely payment for female-led businesses.

Firm size is near neutral in these specifications while higher annual income is associated with a modest increase in on-time payment. Regional fixed effects absorb spatial differences in administration and local business conditions, and pseudo-R² values indicate that this is a good fit for a behavioral compliance model at this level of aggregation.

From a comparative perspective, this female advantage lies near the upper bound of effects documented in higher-income settings-consistent with compliance driven by moral norms and reputational concerns where deterrence is less decisive.

OLS estimates mirror the binary results on the intensive margin. Female-led firms have, on average, about 58,000 lek less in unpaid liabilities than male-led firms — a difference that is both economically and statistically significant.

The Gender × Sector term is negative and significant, suggesting that women’s advantage in arrears reduction is strongest in client-facing sectors (e.g., trade and services), where repeated interactions and reputation likely reinforce disciplined payment behavior.

Joint interpretation of Table 2 and Table 3

Taken together, the models show that gender differences emerge on both compliance margins:

• the extensive margin, whether the deadline is met, and

• The intensive margin: how much debt accumulates if not.

This pattern is precisely what one would expect if gender captures behavioral traits-risk attitudes, sensitivity to norms, and perceived control-operating simultaneously in deadlines and in delay management.

Interpretation of the Results

• Descriptive patterns (Table 1).

The descriptive gap-81.5% vs. 61.2% on-time compliance-aligns with a behavioral reading: under imperfect monitoring, internal norms and reputational concerns may dominate, and these appear stronger on average among female-led firms.

The lower means of arrears, 93,000 lek versus 156,000 lek, imply tighter liquidity discipline and a clearer prioritization of payments, not simply a question of differences in business size or income.

• Binary models (Table 2).

About 30% higher odds of timely payment for women persist after controls and across alternative link functions: logit/probit. This persistence strengthens confidence that the effect is not an artifact of sector composition, city differences, or firm scale.

The result also resonates with prior evidence that women display higher tax morale and stronger prosocial orientation (Croson & Gneezy, 2009; Torgler & Valev, 2010), consistent with the Theory of Planned Behaviour’s pathway from attitudes and norms to observed action.

Table 2. Logit and probit models of tax compliance.

Key variables	Logit (OR)	Probit (Coeff.)
Gender (Female = 1)	1.32*	0.29*
Firm size	0.98	–0.04
Annual income	1.05**	0.07**
Region (data dummies)	Yes	Yes
Observations (N)	500	500
Pseudo R²	0.21	0.20

(* p<0.10, ** p<0.05, *** p<0.01).

Table 2. Logit and probit models of tax compliance.

Key variables	Logit (OR)	Probit (Coeff.)
Gender (Female = 1)	1.32*	0.29*
Firm size	0.98	–0.04
Annual income	1.05**	0.07**
Region (data dummies)	Yes	Yes
Observations (N)	500	500
Pseudo R²	0.21	0.20

(* p<0.10, ** p<0.05, *** p<0.01).

• Continuous model (Table 3).

On the intensive margin, female-led firms have lower debt stocks when there are delays; likewise, this interaction of Gender × Sector implies that the advantage is context-sensitive: it is largest where reputational feedback is most immediate (services, retail), and weakens where cash intensity and sub-contracting reduce visibility.

This is precisely the heterogeneity expected from a behavioral mechanism relying on peer expectations and client relationships to sustain rule-following.

Table 3. OLS model of unpaid tax debt.

Key variables	Coefficient (ALL)
Gender (Female = 1)	–58,200 (p<0.05)
Firm size	+5,300 (ns)
Annual income	–12,500 (p<0.10)
Gender × Sector	–24,800 (p<0.05)
Observations (N)	500
R²	0.17

Table 3. OLS model of unpaid tax debt.

Key variables	Coefficient (ALL)
Gender (Female = 1)	–58,200 (p<0.05)
Firm size	+5,300 (ns)
Annual income	–12,500 (p<0.10)
Gender × Sector	–24,800 (p<0.05)
Observations (N)	500
R²	0.17

• Synthesis.

Taken together, the descriptive and multivariate evidence along both margins is consistent with a coherent behavioral story:

In Southern Albania, where monitoring is uneven, gender captures differences in risk attitude, social norms and perceived control that translate into earlier payments and smaller arrears.

5. Discussion and Conclusions

The results document a clear and robust gender gradient in tax compliance: female-led businesses are more likely to meet deadlines and less likely to accumulate significant arrears. The pattern survives controls for firm size, income, region, and time, and remains evident across all sectors.

This fits comfortably within the literature on behavioral finance, which has long emphasized gender differences in risk tolerance, reputational sensitivity, and long-term orientation (Barber & Odean, 2001; Croson & Gneezy, 2009). In sectors where client contact is frequent and visibility is high, reputational mechanisms are likely to raise the perceived cost of non-compliance for female owners. In more cash-intensive and subcontracted environments, these pressures may be weaker, which helps explain the smaller Gender × Sector effect there.

Together, the findings suggest that gender is not merely a demographic correlate, but a behavioral determinant of fiscal conduct in this environment.

6. Limitations and Future Research

Administrative data are a strength because they observe actual behavior rather than self-reported intention. At the same time, they do not capture perceptions — such as fairness, trust, or service quality — that could help unpack mechanisms more directly.

A natural next step is to link surveys or experiments with payment data to test how tax morale and perceived control relate to timing and arrears. Future work could also explore heterogeneity by firm age, liquidity constraints, and local enforcement intensity, as well as track whether the observed gender gaps narrow or widen as digitalization and service improvements advance.

7. Policy Implications

The evidence points to practical and low-cost tools for revenue administration.

1. Incorporating gender indicators into compliance-risk assessment, to identify cases where supportive outreach (rather than heavier audits) may quickly reduce arrears.

2. Tailoring taxpayer services, for example:peer mentoring for new female entrepreneurs, transparent installment plans, proactive digital reminders, and helplines responsive to taxpayers’ needs,all of which leverage existing tax morale and reduce frictions causing delays.

3. Sector-conscious messaging and service design in construction and transport, where the gender advantage is weaker and informality risks are higher.

8. Conclusions

To our knowledge, this is the first Albanian study using administrative microdata to analyze the impact of gender on compliance behavior. We find that female-led firms are more punctual and accumulate smaller arrears, with the advantage strongest in reputation-intensive sectors.

The results reinforce an emerging behavioral view of compliance that integrates deterrence with psychological and social channels, offering a replicable model for transition economies seeking to improve revenues without resorting to heavier audit burdens.

Translating these insights into practice — through risk assessment and targeted taxpayer services — provides a cost-effective path to enhance voluntary payments and reduce arrears, while simultaneously supporting gender-responsive public finance.