ABSTRACT. No abstract

ABSTRACT. In an era in which artificial intelligence can generate fluent text at unprecedented speed, academic writing must be understood less as a mechanical task and more as a disciplined process of value creation. This keynote examines how scholars can write for high-impact journals by combining analytical clarity, methodological credibility, and persuasive narrative structure. The lecture begins by situating AI as a useful but limited writing companion: powerful for iteration, synthesis, and language refinement, yet unable to replace scholarly judgment, originality, and responsibility. It then develops the central principle that publishable writing creates value for a clearly defined academic conversation. Strong papers do not merely report results; they articulate a problem, establish why it matters, position a contribution, and guide readers through evidence with precision. The lecture also discusses the architecture of effective academic documents, from titles and abstracts to introductions, theory, methods, results, and conclusions. Particular attention is given to coherence, paragraph design, signposting, and the alignment between research question, empirical strategy, and claims. Finally, the lecture presents writing as a rigorous workflow: planning, drafting, revising, seeking feedback, and refining argumentation before submission. The aim is to provide practical principles that help researchers write more clearly, strategically, and convincingly.

ABSTRACT. This study estimates housing supply elasticities across more than one thousand European regions at the NUTS 3 level using a harmonised panel of house prices and housing stock. Employing a three-year dynamic specification and a leave-one-out Bartik instrument to address simultaneity, we identify the medium-run response of housing supply to exogenous price changes. We find that housing supply is highly inelastic: a 10 percent increase in prices raises housing stock by only about 0.2 percent over three years. While some heterogeneity emerges across regions with different development intensity, differences are modest overall. The results provide comparable sub-national elasticity estimates for Europe with direct relevance for housing affordability policy.

ABSTRACT. Rising housing prices in the EU catalyzed significant research and policy interest. Macroeconomic factors are acknowledged as contributors, but the territorial dimension of price propagation remains underexplored in an EU-wide framework. To address this gap, this research integrates official statistics with unconventional data sources for 79,676 municipalities from the EU27 countries. Analyses indicate demand and supply mismatches with territorial regimes. Cities experience limited land, increased population, number of households and the presence of short-term rentals. Rental prices exhibit weaker spatial and socioeconomic associations. The research suggests areas where coordinated EU-wide approaches could be beneficial or subsidiarity principles to be preferred.

ABSTRACT. This study examines the development of retail rents using listing data for 16 major German cities. A hedonic pricing model is employed to estimate quality-adjusted rent dynamics. Listing observations are spatially weighted using retail establishments from OpenStreetMap. The inner-city location classification is constructed using a two-stage approach based on car travel times to the main railway station and walking accessibility to the prime shopping street. First, inner and outer urban areas are distinguished using car travel times to the main railway station. Second, central city-centre locations are delineated based on walking accessibility to the respective prime shopping street. The results show that central city-centre locations command a substantial rent premium relative to the wider urban area. This premium declined between 2018 and 2025, indicating partial convergence between central and more peripheral locations. At the same time, the most central retail locations - prime shopping streets - continue to exhibit the highest rent levels and partly offset the broader decline in inner-city retail rents. Overall, the findings indicate a recovery in bricks-and-mortar retail activity in 2025, accompanied by an increasing spatial concentration of retail demand in the most central locations.

ABSTRACT. This policy review discusses the debate on market failure and government failure in housing markets and provides an overview of key publications on affordable housing released in 2025 by the European Commission, ESPON, and the European Parliament. It reviews how these publications address Europe’s housing affordability challenges and examines the economic perspective on housing provision in a market economy. The paper then discuss the respective roles of markets and government intervention in housing provision, emphasising the need to balance potential market failures with the risks of government failure.

ABSTRACT. This paper revisits the role of systematic skewness (coskewness) in asset pricing by introducing the intelligent Persistent Systematic Skewness (iPSS) factor. Motivated by the fact that coskewness pricing depends critically on its measurement, we employ flexible machine learning models to forecast future coskewness using market information, capturing nonlinear and complex patterns in the data. The resulting factor exhibits strong incremental performance in parsimonious asset pricing models and captures economically meaningful variation in returns. However, similar to existing measures, its contribution weakens in richer models, suggesting partial overlap with standard factors. Overall, nonlinear prediction enhances the identification of coskewness risk and provides new insights into investor preferences for higher moments.

ABSTRACT. Financial stability monitoring is increasingly burdened by an ”information deluge” of dense reports issued by central banks and supervisory authorities. Although Large Language Models offer scalable text analysis, their inherent stochastic nature and ”black-box” reasoning pose significant institutional risks. We construct a novel expert-labeled benchmark dataset of Financial Stability Reports from emerging and developed economies and introduce a taxonomy comprising nine systemic risk categories and one macroprudential policy category, each annotated for directional sentiment: stable, unstable, or uncertain and tightening, loosening, or unchanged. To enable transparent and institutionally robust deployment, we propose a two-stage modular framework that constrains Large Language Models inference to the expert taxonomy, thereby helping to isolate signals from noise. Relative to unconstrained prompting, our approach reduces irrelevant content by 90% and produces fully traceable definition-bound risk labels aligned with expert annotations. The The framework demonstrates how generative artificial intelligence can be disciplined for high-stakes regulatory environments while preserving scalability.

ABSTRACT. Financial time series are subject to frequent structural instability, which complicates forecasting, risk measurement, and real-time monitoring. We propose a unified framework that integrates adaptive state-space diagnostics with global offline segmentation and strictly causal online regime recognition. Offline, a Kalman filter with Dempster–Shafer covariance adaptation is combined with penalized exact segmentation to localize structural breaks across heterogeneous data-generating processes. Online, break detection is formulated as a filtration-respecting probabilistic classification problem and implemented using a LightGBM learner based on strictly causal features derived from both raw returns and model diagnostics. Across synthetic processes with known ground truth, the framework achieves balanced break identification offline and stable short-horizon recognition relative to variance-based and Bayesian benchmarks. Applied to a cross-asset panel of financial series, it identifies economically interpretable regime shifts while maintaining parsimonious segmentation. The results highlight the complementarity of adaptive state-space monitoring and causal machine learning for structural break analysis in financial econometrics.

ABSTRACT. Artificial intelligence (AI) is widely considered a general purpose technology with the potential to reshape production processes, managerial decision making, and competitive dynamics. Despite growing attention to its long term productivity benefits, the short run implications for firm level financial performance remain less clearly understood. This study investigates the relationship between AI exposure, innovation investment, financial structure, and firm profitability. The empirical analysis combines a dynamic panel covering 2015–2024 with a cross sectional examination for 2024 using an industry level AI exposure index from the OECD AI Policy Observatory. Firm level financial and innovation data are obtained from the Orbis database for firms headquartered in G7 economies. The results indicate that higher R&D intensity is associated with lower contemporaneous profitability, reflecting the accounting treatment of research expenditures and short run adjustment costs. Leverage shows a negative relationship with return on assets, whereas Tobin’s Q is positively associated with profitability. Cross sectional estimates suggest a negative relationship between industry level AI exposure and profitability; however, this effect weakens after controlling for industry and country fixed effects, implying that it largely reflects structural technological characteristics rather than firm specific AI adoption. Overall, the findings indicate that AI exposure primarily captures industry level technological intensity, while firm level innovation investment generates short term profitability pressures. These results are consistent with an intertemporal adjustment perspective in which technological transitions impose short run costs before longer term productivity gains emerge.

ABSTRACT. This study examines gender differences in financial literacy in Spain using data from the 2021–2022 Financial Competence Survey (ECF), a nationally representative sample of 7,835 individuals aged 18–79. The analysis combines a descriptive assessment of responses to the widely used “Big Three” questions (inflation, compound interest, and risk diversification) with supervised machine learning techniques to identify the factors most strongly associated with gender profiles. Descriptive results reveal a persistent gender gap. Men show higher proportions of correct answers across all three items, while women exhibit a greater tendency to select “don’t know/no answer,” suggesting that disparities reflect not only differences in objective knowledge but also differences in declared uncertainty. To move beyond bivariate comparisons, the study applies a classification tree (CART) and a Gradient Boosting model (XGBoost). The findings consistently indicate that socioeconomic variables—particularly employment status, education level, and age—account for most of the predictive power in distinguishing gender profiles. Financial literacy variables contribute as secondary refinements within segments defined primarily by structural characteristics. In both modelling approaches, classification accuracy is systematically higher for men than for women, pointing to greater heterogeneity or overlap within the female profile. Overall, the results suggest that the gender gap in financial literacy is embedded in broader socioeconomic structures. Policies aimed at reducing disparities should therefore combine improvements in financial knowledge with context-sensitive strategies that address structural inequalities and confidence-related dimensions of financial decision-making.

ABSTRACT. This study investigates distributional heterogeneity in the relationship between financial literacy and demand-side financial development across Europe. Using microdata from Eurobarometer Flash Survey 525 (N = 24,356), we construct a latent household-level financial development variable via Polychoric Principal Component Analysis and Item Response Theory, then apply conditional quantile regression to estimate how the effect of financial literacy varies across the full distribution of household financial engagement. We document a structural "sticky floor": below the 20th percentile, financial literacy has no statistically discernible effect on financial development, reflecting a non-responsive zone driven by supply-side barriers rather than knowledge deficits. Above this zone, the marginal effect of financial literacy increases monotonically through the upper-middle quantiles before attenuating at the top - consistent with product saturation. A comparative analysis of Eastern and Western European subsamples reveals that this non-responsive zone extends significantly deeper in transition economies, where literacy effects remain statistically dormant until approximately the 40th percentile. These findings challenge the universalist framing of financial education policy and argue for sequenced interventions that prioritize structural access before broad-based literacy campaigns.

ABSTRACT. This study investigates gender disparities in financial literacy by bringing together global scholarship and original evidence from Romanian finance undergraduates. Although women and men now resemble each other far more in terms of educational attainment, labour force participation, and college enrollment, a persistent gender gap in financial literacy endures. International research typically reports an average 12% advantage for men, yet the patterns observed in this Romanian sample complicate that aggregate picture. Using a mixed-methods design that combines standardized financial literacy testing with focus group discussions on financial socialization, I find that female students obtain slightly lower test scores than males (2.73 vs. 2.89), but display more accurate self-assessment. Most participants (59%) underestimated their knowledge, and identified family discussions, household money management, and peer conversations as the main channels through which they learned about money. These findings call into question blanket assertions about male overconfidence and instead support the view that gender differences in financial literacy are socially constructed rather than innate. Both cultural context and educational environment shape what students actually know and how confident they feel about that knowledge. This study contributes to the literature by documenting context-specific patterns in a Romanian sample and by showing how family-based financial socialization jointly influences objective financial knowledge and perceived competence (PC).

ABSTRACT. This study addresses a fundamental measurement challenge in behavioral finance: the absence of a unified analytical framework capable of simultaneously capturing the objective behavioral and subjective perceptual dimensions of individual financial risk profiles. Leveraging the 2023 Flash Eurobarometer 525, a harmonized cross-country dataset of approximately 26,000 observations across the European Union, the study integrates Item Response Theory (IRT), Latent Class Analysis (LCA), Graded Response Models (GRM), and Polychoric Principal Component Analysis to reconstruct latent risk structures from large-scale survey data. The IRT analysis establishes a hierarchical sophistication ladder across seven financial products, while LCA identifies three behaviorally coherent archetypes - Investors (16.7%), Traditionalists (21.2%), and Minimalists (62.1%) - revealing deep structural heterogeneity in European financial participation. A Speculative Leapfrogging effect among Minimalists provides empirical support for risk-seeking behavior under structural exclusion, consistent with Prospect Theory. The categorical rejection of subjective risk unidimensionality, confirmed through GRM diagnostics, antagonistic Polychoric PCA factor loadings, and a statistically significant Wilcoxon shift, establishes that short-term financial fragility and long-term retirement confidence constitute fundamentally distinct psychological dimensions rather than a unified risk trait. Building on these findings, the study introduces a Discrepancy Index, a diagnostic metric quantifying the calibration gap between objective financial positioning and subjective risk perception, with direct implications for financial literacy policy and consumer protection regulation within the EU.