Research

Mathematical and Computational Models of Complex Systems 

1. Vallarino, D. (2026). “Stochastic Network Survival Dynamics: A Nonlinear Evolution Problem on Economic Graphs.” Communications in Nonlinear Science and Numerical Simulation, 159, 109904.

2. Vallarino, D. (2026, forthcoming). “Sandpile Dynamics in the Product Space: Curvature, Spectral Criticality, and Diversification Fragility.” Journal of Complex Networks.

3. Vallarino, D. (2026). “Survival and Cascades in Geoeconomic Networks: A Graph-Embedded Hazard Framework.” Complexity.

4. Vallarino, D. (2026). “Spectral Curvature of Stochastic Hazard Operators on Graphs.” Statistics and Probability Letters, 110797.

5. Vallarino, D. (2026). “Narrative Pacing under Fragmented Attention: A Structural Approach to Suspense in Serialized Television.” Poetics.

Institutional Dynamics and Networked Economic Systems 

1. Vallarino, D. (2026). “Successful Investment Projects Are Contagious: Social Learning and Network Effects in Industrial Policy.” Economic Systems, 101420.

2. Vallarino, D. (2026). “Estimating Structural Spillovers in International Trade Networks: An Empirical Assessment with Graph Neural Models.” Empirical Economics, 70(6), 87.

3. Vallarino, D. (2026). “Stability without Transformation: Economic Discourse and Reform Capacity in Small Open Economies.” Review of Political Economy.

4. Vallarino, D. (2026). “Institutional Distance and Governance Performance: A Structural Analysis.” Economics of Governance. 27 (51) 

5. Vallarino, D. (2025). “Augmenting Trade Complexity Analysis with Deep Learning: An AI-Based Framework for Small Open Economies.” Applied Economics Letters, 1–4.

Artificial Intelligence, Data Science, and Decision Systems

1. Vallarino, D. (2025). “Causal-GNN for Ethical AI in Financial Services: Ensuring Fairness, Compliance, and Transparency in Automated Decision-Making.” Artificial Intelligence and Law, 1–16.

2. Vallarino, D. (2026). “Taxing the Invisible: Graph-Based Information Structures for Redistributive Design under Uncertainty.” Journal of Economics, Finance and Administrative Science, 1–22.

3. Vallarino, D. (2026). “Artificial Intelligence, Affective Well-Being and Workplace Happiness in Iberoamerican Companies.” Management Research, 1–32.

4. Vallarino, D. (2025). “Understanding the Market Trends: A Hybrid Approach to Stock Price Prediction Using RNNs and Transformer-Based Sentiment Analysis.” Journal of Applied Economic Sciences, XX(1).

5. Vallarino, D. (2024). “Machine Learning Algorithms for Survival Analysis: Advantages, Disadvantages, and Examples.” International Journal of Artificial Intelligence and Machine Learning, 4(1), 10–21.

1. “Capability-Locked Frontiers: The Feasibility Condition for Leapfrogging Is Accumulation, Not Position.” Research Policy. [Revise & Resubmit]. 

This paper examines whether technological leapfrogging is enabled by accumulated capabilities or by advantageous brokerage positions within the technology space. Using global patent co-classification data for 164 countries, it constructs relatedness density to the artificial-intelligence frontier and a curvature-weighted measure of bridge exposure based on Ollivier–Ricci geometry, and evaluates their ability to predict AI patenting in 2011–2015. The results show that bridge exposure almost entirely collapses onto capability density and that its density-orthogonal component does not predict frontier activity at any level of economic complexity. Instead, entry is governed by a threshold of technological reach: countries possessing at least one capability adjacent to the AI frontier entered at a rate of 31.0%, compared with 3.9% among countries outside the reach set. The findings indicate that leapfrogging depends on broad capability accumulation rather than isolated structural positions, with AI emerging as an unusually capability-locked technological frontier. 

2. “Structural Hazards in Global Production: A Graph Neural Network Approach to Sectoral Survival.” Structural Change and Economic Dynamics. [Revise & Resubmit]. 

This paper models sectoral exit as a network-dependent hazard process by integrating survival analysis with graph-based representations of global production systems. Using a synthetic panel calibrated to empirical distributions and validated with WIOD data for 2000–2014, it compares standard Cox models, network-augmented specifications, embedding-based approaches, and a GNN–Cox model. The results show that sectoral exit risk depends not only on observable characteristics but also on higher-order network position, with peripheral sectors facing higher hazards and highly connected sectors displaying greater structural resilience, especially during periods of network reorganization such as 2008–2014.

3. “Stochastic Network Survival Dynamics: A Computational Extension of the Graph-Embedded Hazard Model.” Computational Economics. [Minor Revision]. 

This paper develops a stochastic extension of the Graph-Embedded Hazard Model to study how uncertainty in network structure affects survival dynamics in interconnected economies. It formulates a coupled GEHM–PDE–SDE framework in which structural complexity evolves stochastically and node-level hazards depend on drift, volatility, and topology-dependent exposure, solved through spectral graph methods, finite differences, and Monte Carlo simulation. Numerical experiments show that structural volatility shifts complexity toward higher-risk states, amplifies exit hazards, produces heavier failure tails, and that these effects are strongest in scale-free network geometries.

4. “Random Attractors for Nonlinear p-Laplacian Diffusion on Weighted Graphs under Stochastic Forcing.” Journal of Mathematical Analysis and Applications. [Revise & Resubmit].

This paper develops a random dynamical systems framework for nonlinear stochastic diffusion on finite weighted graphs governed by a discrete (p)-Laplacian with additive Brownian forcing. It establishes the existence and uniqueness of compact pullback random attractors and derives topology-dependent coercivity bounds linking long-run stochastic dynamics to the spectral and combinatorial structure of the underlying graph. The results show that weaker connectivity and greater structural heterogeneity reduce effective dissipation, enlarge random absorbing sets and attractors, and generate anisotropic stochastic dynamics, extending random-attractor theory from homogeneous domains and lattice systems to general weighted networks.

5. “A Colonial Financial System at a Critical State: The Australian Banking Avalanche of 1893.” Asia-Pacific Economic History Review. [Revise and Resubmit] 

This paper reinterprets the Australian banking crisis of 1893 as a systemic avalanche emerging from the interaction of rapid credit expansion, property-market deterioration, institutional fragility, and a highly interconnected colonial banking system. Combining a synthetic-difference-in-differences counterfactual with bank-level suspension and reconstruction records, bank-by-state branch data, colonial macro-financial indicators, monetary aggregates, and historical asset-price evidence, it estimates that Australian real GDP per capita lay on average about 23 percent below its synthetic comparison path over 1894–1913, while emphasizing the conservative nature of the placebo inference and the presence of competing post-crisis shocks. At the institutional level, the analysis documents a highly compressed April–May 1893 suspension cascade, £57.0 million in reconstructed locked liabilities concentrated among a small number of large banks, geographically broad exposure to later-suspending institutions—accounting for 69.2 percent of the national branch network in 1892—and a contractual overhang that persisted long after banks reopened. The evidence supports a historically grounded systemic interpretation of the crisis characterized by accumulated vulnerability, temporal compression, concentration, geographic reach, and persistence, while distinguishing this interpretation from a formally identified causal network mechanism.

6. "Why Most AI-Based Trading Strategies Fail Backtests: Risk Allocation and Real-World Performance." Journal of Asset Management (Springer). [Revise & Resubmit].

This paper investigates the structural gap between backtested and live performance in AI-driven trading strategies, arguing that the failure is not primarily algorithmic but architectural: standard approaches optimize signal generation while neglecting dynamic risk allocation across strategies. The paper develops a transparent Mixture-of-Experts framework that conditions capital allocation on market volatility regimes, compares it against static and benchmark allocation rules, and evaluates out-of-sample performance across multiple equity markets. The results show that regime-aware allocation materially improves risk-adjusted returns and reduces drawdown exposure relative to signal-only AI approaches, with implications for how practitioners and asset managers should evaluate AI-based systematic strategies.

1. Vallarino, D. (2026). Hall-Like Transversal Stress and Sandpile Criticality on Real Production Networks Available on arXiv

2. Vallarino, D. (2026). Sandpile Economics: Theory, Identification, and Evidence . Available on arXiv.

3. Vallarino, D. (2026). Identification and Inference in Nonlinear Dynamic Network Models. Available on arXiv. 

4. Vallarino, D. (2026). Nonlinear Fiscal Transitions and the Dynamics of Public Expenditure Reform. Available on arXiv.

5. Vallarino, D. (2026). Synthetic Firm-Level Dataset with Realistic Credit Risk and Financial Structures [Data set]. Zenodo. 

6. Vallarino, D. (2025). A Semi-Synthetic Social and Clinical Survival Dataset with Structured Dependence and Informative Censoring [Data set]. Zenodo. 

7. Vallarino, D. (2025) "The Graph-Embedded Hazard Model (GEHM): Stochastic Network Survival Dynamics on Economic Graphs"  Available on arXiv.

8. Vallarino, D. (2025) "Invited to Develop: Institutional Belonging and the Counterfactual Architecture of Developments" Available on arXiv.

9. Vallarino. D. (2025) "Data for Inclusion: The Redistributive Power of Data Economics” Available at ArXiv 

10. Vallarino, D. (2025) “Adaptive Market Intelligence: A Mixture of Experts Framework for Volatility-Sensitive Stock Forecasting” Available at ArXiv

11. Vallarino, D. (2025) “Causal GNNs and the Anthropology of Data: An Ethical Approach to Fairness and Transparency in Financial AI” Available at SSRN

12. Vallarino, D. (2025) “AI-Powered Fraud Detection in Financial Services: GNN, Compliance Challenges, and Risk Mitigation” Available at SSRN

13. Vallarino, D. (2025) “How Do Consumers Really Choose? Exposing Hidden Preferences with the Mixture of Experts Model.” Available at ArXiv

14. Vallarino, D. (2024) “Modeling Adaptive Fraud Patterns: An Agent-Centric Hybrid Framework with MoE and Deep Learning” Available at SSRN

15. Vallarino, D. (2024) “A Dynamic Approach to Stock Price Prediction: Comparing RNN and Mixture of Experts Models Across Different Volatility Profiles” Available at ArXiv

16. Vallarino, D. (2024) “Dynamic Portfolio Rebalancing: A Hybrid new Model Using GNNs and Pathfinding for Cost Efficiency” Available at ArXiv

17. Vallarino, D. (2024) “Machine Learning and Econometric Approaches to Fiscal Policies: Understanding Industrial Investment Dynamics in Uruguay (1974-2010)” Available at ArXiv

18. Vallarino, D. (2024) “Understanding the Market Trends: A Hybrid Approach to Stock Price Prediction Using RNNs and Transformer-Based Sentiment Analysis” Available at SSRN

19. Vallarino, D. (2024) “Decoding the Puzzle of Joblessness: Machine Learning Predicts Unemployment Trends in the Americas” Available at SSRN

20. Vallarino, D. (2024) “Analyzing Economic Convergence Across the Americas: A Survival Analysis Approach to GDP per Capita Trajectories” Available at arXiv

1. Vallarino, D. (2026). “Cross-Market Vulnerability and Central Bank Surveillance: A Geometric Contagion Index for Banking and Production Networks.” 41st Annual Economics Conference, Central Bank of Uruguay. (27-29 July, Montevideo, Uruguay). 

2. Vallarino, D. (2026) "Identification in Duration Models with Network-Correlated Unobservables" Econometric Society - 2026 North American Summer Meeting. (June 4 - 7, 2026, Atlanta, Georgia, US) ppt

3. Vallarino, D. (2025) “Breaking the Product Space: Rethinking Economic Complexity with GNNs and Synthetic Trade Networks in Emerging Economies.” Conference on Economic Complexity 2025 (July 9 - 11, 2025, Toulouse, France). 

4. Vallarino, D. (2025). “Market Power and Tax Incentives for Sustainable Growth: A Network-Based Approach in Latin America”. BALAS 2025 Conference, (San José, Costa Rica, from April 7 to 11, 2025)

5. Vallarino, D. (2024). “Stock Behavior Using RNN Models and Transformer-Based Text Analysis”. LatinR 2024 Conference, Montevideo, Uruguay

6. Vallarino, D. (2023). “Comparative Analysis of Machine Learning Models for Survival Analysis: Empirical Study and Performance Assessment” LatinR 2023 Conference, Montevideo, Uruguay

7. Vallarino, D. (2023). “Buy when? Survival machine learning model comparison for purchase timing.” 4th International Conference on Big Data, Machine Learning and IoT (BMLI 2023) Dubai, UAE.

8. Vallarino, D., Azúa C, Úbeda, R. (2023). “Differentiated Roles and Perceptions of the Controller Function: An Exploratory Study”. Business Association of Latin America Studies (BALAS) Annual Conference, Mexico City, Mexico.