7 publications
Papers and proceedings
Peer-reviewed papers, conference proceedings, and preprints across multi-agent systems, statistical verification, and artificial intelligence.
- 2026 Ranking the wrong places: vulnerability assessment, flood losses and risk sharing in Italy and Europe
Preprint (arXiv:2609.36349); submitted to Premio Giorgio Rota XIV
Abstract
Before a flood, governments decide where to invest in protection; after it, how much of the loss to compensate. The European Union informs the first decision with a composite index that ranks every subnational unit by vulnerability. This paper tests the ranking by comparing Italian provinces struck by the same flood between 2005 and 2024. The published index does not order provinces by the losses they suffer, with or without controls for hazard, exposure and income ($\rho$ between $-0.06$ and $0.09$). The failure lies in what it weights. As published, the index mostly measures economic development, while losses follow water: they arrive in recurring groups of provinces that share a river basin, and more than half of the ranking error lies between those groups. In a pre-registered test that rebuilds it and eight rivals at every forecast origin from 2010 to 2020, the index comes last on both questions an allocator asks. On where damage will land, its top ten provinces capture 16 per cent of later damage, less than a random draw; Italy's public flood hazard map, blind to losses, captures 44. On who loses more from the same water, its rank correlation with later losses is 0.03; its own components, reweighted on past losses, reach 0.42. Neither repair needs new data. The misranked loss is paid locally. Of EUR 12.3 billion recorded over 2013--2022, 85 per cent was not covered by national transfers, and prevention funding, scored the same way, does no better than chance. Because correlated losses cluster within borders, national pools in Europe need 39 per cent more reserve than cross-border pools of the same size. The index fails the same way across 1,142 European units, where the repairs are weaker.
Cite this working paper
@misc{blando2026ranking, title = {Ranking the wrong places: vulnerability assessment, flood losses and risk sharing in Italy and Europe}, author = {Blando, Stefano}, year = {2026}, eprint = {2609.36349}, archivePrefix = {arXiv}, primaryClass = {econ.GN}, doi = {10.48550/arxiv.2609.36349}, url = {https://arxiv.org/abs/2609.36349} } - 2026 Towards Agentic Agent-based Models: Feasibility, Performance, and Statistical Model Checking
AISoLA 2026 (to appear; arXiv:2607.17948)
Abstract
Integrating Large Language Models (LLMs) into traditional Agent-Based Models (ABMs) allows agents to leverage tool calls, semantic reasoning, and adaptive decision-making. In this paper, we evaluate the feasibility, performance, and emergent behavior of agentic ABMs using Mesa and statistical model checking via MultiVeStA. By conducting systematic statistical specification testing across agentic parameter spaces, we quantify the impact of LLM decision-making on macroeconomic and social simulation dynamics.
Cite this conference paper generated from arXiv
@misc{blando2026agentic, title = {Towards Agentic Agent-based Models: Feasibility, Performance, and Statistical Model Checking}, author = {Blando, Stefano and Guerrazzi, Emanuele and Porcedda, Riccardo and Squillace, Giuseppe and Tschaikowski, Max and Vandin, Andrea}, year = {2026}, eprint = {2607.17948}, archivePrefix = {arXiv}, primaryClass = {econ.GN}, note = {AISoLA 2026 (to appear; arXiv:2607.17948)} } - 2026 Statistical Model Checking of the Keynes+Schumpeter Model: A Transient Sensitivity Analysis of a Macroeconomic ABM
arXiv preprint arXiv:2605.10447 (journal version in preparation)
Abstract
Agent-based macroeconomic models, such as the Keynes+Schumpeter (K+S) framework, integrate innovation-driven Schumpeterian growth with Keynesian aggregate demand dynamics. However, formal sensitivity analysis in such high-dimensional stochastic systems remains computationally demanding and methodologically fragmented. In this paper, we apply Statistical Model Checking (SMC) via MultiVeStA to perform a systematic, transient sensitivity analysis of the K+S model. By evaluating transient probabilistic properties across financial, structural, and heuristic parameter spaces, we establish statistical confidence intervals for key macroeconomic observables, including output growth and unemployment trajectories.
Cite this article generated from arXiv
@misc{blando2026statistical, title = {Statistical Model Checking of the Keynes+Schumpeter Model: A Transient Sensitivity Analysis of a Macroeconomic ABM}, author = {Blando, Stefano and Fagiolo, Giorgio and Napoletano, Mauro and Treibich, Tania and Vandin, Andrea}, year = {2026}, eprint = {2605.10447}, archivePrefix = {arXiv}, primaryClass = {econ.GN}, note = {arXiv preprint arXiv:2605.10447 (journal version in preparation)} } - 2026 Statistical model checking of the Island Model: an established economic agent-based model of endogenous growth
MARS @ ETAPS 2026, Turin, Italy (EPTCS vol. 443, pp. 3-22)
Abstract
Agent-based models (ABMs) are increasingly used to study complex economic phenomena such as endogenous growth, but their analysis typically relies on ad-hoc Monte Carlo exercises without formal statistical guarantees. We show how statistical model checking (SMC), and in particular MultiVeStA, can automate and enrich the analysis of a seminal ABM: the Island Model of Fagiolo and Dosi, which captures the exploration-exploitation trade-off in technological search. We reproduce key stylized facts from the original model with formal confidence intervals, confirm the optimality of moderate exploration rates, and perform a counterfactual sensitivity analysis across returns to scale, skill transfer, and knowledge locality. Using MultiVeStA's built-in Welch's t-test, 6 out of 7 pairwise parameter comparisons yield statistically different growth trajectories, while the exception reveals a saturation effect in knowledge locality. Our results demonstrate that SMC offers a principled, reproducible methodology for the quantitative analysis of agent-based economic models.
Cite this conference paper
@inproceedings{blando2026islandsmc, title={Statistical model checking of the Island Model: an established economic agent-based model of endogenous growth}, author={Blando, Stefano and Fagiolo, Giorgio and Giachini, Daniele and Vandin, Andrea and Ivanaj, Ernest}, booktitle={Proceedings Tenth International Workshop on Models for Formal Analysis of Real Systems (MARS 2026)}, series={EPTCS}, volume={443}, pages={3--22}, year={2026}, doi={10.4204/EPTCS.443.2} } - 2026 Network Topology Analysis and Machine Learning Techniques for Systemic Risk Prediction in U.S. Equity Markets
University of Rome Tor Vergata - School of Economics and Finance
Abstract
This paper investigates the predictive power of Graph Neural Networks (GNNs) in forecasting systemic risk events. By modelling financial markets as dynamic complex networks, we extract topological features that serve as early warning signals. The study demonstrates how these signals can be integrated into algorithmic trading strategies to mitigate downside risk during market turmoil, outperforming traditional benchmark models.
Cite this thesis
@mastersthesis{blando2025networktopology, title={Network Topology and Machine Learning for Systemic Risk: From Early Warning to Trading Strategies}, author={Blando, Stefano}, school={University of Rome Tor Vergata}, type={Master thesis, Master in CESMA}, year={2025}, note={Academic year 2024/2025} } - 2026 A Multi-Method Validation Framework for Large-Scale Multilingual Text Analytics
JADT 2026, Palermo, Italy (Proceedings Vol. II, pp. 70–79; ISBN 978-88-5509-883-0)
Abstract
To distinguish genuine findings from methodological artifacts, this paper proposes a validation framework based on method-invariant patterns. Analyzing 999,152 multilingual reviews across 18 independent techniques (from classical clustering to Transformers), we demonstrate that substantive content accounts for 95.4% of variance, while methodological choice explains less than 3%. The study confirms that robust patterns transcend specific algorithms and implementations. Furthermore, while BERT achieves peak accuracy (91.3%), classical approaches like SVM offer comparable performance (89.1%) with a 29-fold reduction in computational cost.
Cite this conference paper
@inproceedings{blando2026multimethod, title={A Multi-Method Validation Framework for Large-Scale Multilingual Text Analytics}, author={Blando, Stefano and Iezzi, Domenica Fioredistella}, booktitle={JADT 2026 Proceedings, Volume II: 18th International Conference on Statistical Analysis of Textual Data}, editor={Plaia, Antonella and Sciandra, Mariangela and Albano, Alessandro}, pages={70--79}, year={2026}, address={Palermo, Italy}, isbn={978-88-5509-883-0} } - 2025 Robust Portfolio Optimization Under Systematic Market Disruptions: A Factor-Analytic Approach
Computational Economics (in review)
Abstract
This thesis addresses the challenge of portfolio optimization in high-dimensional settings (p > n) affected by sparse contamination (outliers). By proposing a robust Factor-Analytic approach, the study demonstrates how to stabilize covariance matrix estimation against data anomalies. The resulting allocation strategies exhibit superior resilience and risk-adjusted returns compared to traditional Mean-Variance frameworks, especially during market shocks.
Cite this preprint / working paper
@unpublished{blando2025robust, title={Robust Portfolio Optimization Under Systematic Market Disruptions: A Factor-Analytic Approach}, author={Blando, Stefano and Farcomeni, Alessio}, year={2025}, note={Under review at Computational Economics} }