Critical factors influencing information disclosure in public organisations

The study examines why and how local governments disclose information proactively to enhance transparency, accountability, and citizens’ trust. Anchored in agency theory and legitimacy theory, it addresses persistent information asymmetries between citizens and officials and the role of disclosure in reinforcing governmental legitimacy. Portugal provides a relevant context due to its unitary administrative framework, the 2010–2014 financial crisis, and comparatively low trust in government. The research aims to: (1) analyse the evolution of Portuguese municipal information disclosure over time, and (2) identify critical social, economic, and political factors that determine municipal disclosure. Using the Portuguese municipal transparency index (TIAC) as a proxy for disclosure, the paper argues that variation in disclosure arises from municipal actions and contextual factors despite a common legal framework.

Transparency is defined as proactive dissemination and access to intelligible information for citizens and stakeholders. Prior literature links transparency to improved governance outcomes and reduced information asymmetry. Two key theoretical lenses guide the inquiry: agency theory (citizens seek information to monitor agents and reduce agency costs) and legitimacy theory (officials disclose to secure public approval and legitimacy). Cross-national variation exists between proactive and passive transparency regimes. In Portugal, FOI legislation evolved from a passive model (1993, 2007) to an active disclosure mandate in 2016 (Law 26/2016), requiring regularly updated, open, machine-readable information. Municipal disclosure is measured by TIAC’s Municipal Transparency Index (2013–2016), based on 76 indicators across seven dimensions (organisation and functioning; municipal plans; fees/taxes/regulations; interaction with citizens; public procurement; fiscal transparency; urban planning), scored from 0 to 100 based on website content. Hypotheses: H1—disclosure increases over time; H2—greater internet access is associated with more disclosure; H3—wealthier municipalities disclose more; H4—greater financial independence (own-source revenues) increases disclosure; H5—higher debt-to-income ratios lead to more disclosure; H6—lower electoral participation associates with higher disclosure (negative relationship with participation); H7—left-governed municipalities disclose more.

Design: Longitudinal panel data study of Portuguese municipalities (2013–2016). Sample: All 308 municipalities; after removing missing values, an unbalanced panel of 1,098 municipality-year observations. Dependent variable: Information Disclosure Index (ID) from TIAC (0–100), based on presence/absence of 76 website indicators aggregated into seven dimensions. Independent variables: Internet Access (residents with internet access divided by population density); Wealthier Municipalities (regional performance index proxy); Financial Independence (own-source revenue / total actual revenue); Debt-to-Income Ratio (total debt / actual revenues); Electoral Participation (% turnout in municipal elections); Political Ideology (dummy=1 if mayor is leftist). Controls: Electoral cycle dummies—Pre-Election (year before election), Election (election year), Post-Election (year after election). Data sources: TIAC, General Directorate of Local Authorities (Finanças Municipais), National Election Commission, National Statistics Institute of Portugal, Marktest Sales Index; election results from 2009 and 2013. Estimation: Static panel models—pooled OLS, fixed effects (FE), and random effects (RE). Model: ID_it = a + β1 Internet_it + β2 Wealth_it + β3 FinInd_it + β4 DebtInc_it + β5 Turnout_it + β6 Ideology_it + β7 Pre_it + β8 Elect_it + β9 Post_it + μ_i + ε_it. Model selection and diagnostics: Breusch-Pagan LM (p=0.00) favored RE over pooled OLS; F-test (p=0.00) favored FE over pooled OLS; Hausman test χ²=27.21, p=0.0007 favored FE. Serial correlation (Wooldridge F(1,272)=123.329, p=0.0000), heteroskedasticity (modified Wald χ²(278)≈7.3e+31, p=0.0000), and cross-sectional dependence (Pesaran CD=10.996, p=0.000) detected. Final estimation used Prais–Winsten regressions with panel-corrected standard errors (PCSE) accommodating AR(1), heteroskedasticity, and contemporaneous correlation. Preprocessing: Continuous predictors (Internet Access, Wealth, Financial Independence, Debt-to-Income) winsorised at 5% tails to mitigate outliers.

Descriptive: Municipal information disclosure increased markedly: mean ID rose from 34 (2013) to 53 (2016); overall mean 2013–2016 = 42; min=3, max=100. This supports H1 (increasing disclosure over time). Multivariate (PCSE, FE baseline): • Internet Access: positive, significant at 1% (β=0.0840) → H2 accepted. • Wealthier Municipalities: not significant (ns) → H3 rejected. • Financial Independence: not significant → H4 rejected. • Debt-to-Income Ratio: positive, significant at 1% (β=0.0003) → H5 accepted. • Electoral Participation: negative, significant at 1% (β=−0.0029), indicating lower turnout associates with higher disclosure efforts → H6 accepted. • Political Ideology (left): positive, significant at 1% (β=0.0373), left-governed municipalities disclose more → H7 accepted. Electoral cycle controls: • Pre-Election: positive, 1% (β=0.0720). • Election year: negative, 1% (β=−0.0647). • Post-Election: negative, 1% (β=−0.1062). Model fit (PCSE): overall R² ≈ 47.5. Robustness/diagnostics addressed via PCSE with AR(1), heteroskedasticity, and cross-sectional dependence corrections.

Findings substantiate the theoretical expectation that both agency and legitimacy mechanisms shape municipal disclosure. Increased internet access empowers citizens, heightening demand for online information and pressuring municipalities to disclose more, consistent with agency theory. Political variables are pivotal: left-governed municipalities disclose more, and disclosure intensifies pre-election as incumbents seek to legitimise performance, while it declines during and after elections. Lower electoral participation appears to spur municipalities to disseminate more information to regain trust and foster engagement, aligning with legitimacy theory. Contrary to some prior studies, neither municipal wealth nor financial independence significantly affected disclosure in Portugal’s context, suggesting institutional and socio-political specificities may mediate these relationships. The prominence of financial transparency, particularly under higher indebtedness, indicates strategic disclosure to reduce information asymmetry for external stakeholders and signal fiscal credibility.

The study shows a clear upward trend in municipal information disclosure in Portugal and identifies critical determinants. Political factors—ideology and electoral cycle—and civic engagement dynamics (electoral participation) are the most influential drivers. Internet access significantly boosts disclosure, and higher debt levels are associated with greater transparency, likely to signal fiscal discipline to stakeholders. Conversely, municipal wealth and financial independence do not significantly predict disclosure in this context. Contributions include empirical evidence from a unitary country with a common legal framework, highlighting the strong role of political cycles, and reinforcing agency/legitimacy theory linkages. Future research should delve deeper into political-cycle dynamics, explore citizen-side demand and perceptions of disclosed information, test for potential endogeneity, and examine generalisability across varied institutional settings.

- Supply-side focus: analyses organisational disclosure without measuring citizen demand or usage. - No assessment of citizens’ perceptions of accessibility, intelligibility, or quality of disclosed information. - Context-specific to Portugal (unitary, centralised system); external validity to other institutional settings may be limited. - Potential endogeneity and reverse causality (e.g., between disclosure and political/economic variables) not addressed in the empirical strategy.