Mateus Hiro Nagata
M.Res. in Economics and Decision Sciences - HEC Paris
M.A. in Economic Theory - ITAM
B.A. in Economic Sciences - University of Brasilia
View My LinkedIn Profile
View my Google Scholar
Description
Hi! My name is Mateus Hiro Nagata. A first-principles thinker Data Scientist/Economist. My interests revolve around transforming ML and data science analysis into actions using solid economic theory. Specifically, I can translate data-driven analysis into human-interpretable insights.
Currently, I devised new Machine Learning and Reinforcement Learning algorithms and applied them to Game Theory (the analysis of strategic interaction between agents) and Decision Theory (the analysis of individual decision). Those new algorithms were used to create papers that were accepted at JECCO 2025, Science of Decision Making Conference and SSCW 2026 and can be found in my Documents section.
My experiences include 4 papers published into top journals using Time-Series Analysis in Finance (and Extreme-Value Theory) and Machine Learning analysis and Predictions.
Research Projects
Algorithmic Game Theory x Reinforcement Learning
Decision Theory x Statistical Learning
Research
Representativity: A Possibility Theorem
Mateus Hiro Nagata
18th Society for Social Choice and Welfare Conference, Tokyo (2026, forthcoming)
Agents have preferences on policies, not on leaders. Voters want representatives whose future choices match their own. I propose the Representative Dream Problem: given a list of candidates with data on past problems and their preferred actions, can I guarantee a candidate close to society’s preferences both in-sample and out-of-sample? Yes. I solve it by formalizing representative selection as a statistical learning problem and proving that voters can reliably identify representatives who match their future preferences by observing historical agreement. Using Case-Based Decision Theory to model political reasoning, I derive finite-sample convergence bounds showing the number of observations needed grows only logarithmically with the candidate pool size. Under mild structural assumptions on preferences, I prove that the selected representative’s disagreement probability converges to zero asymptotically. This establishes a possibility theorem for representative democracy: political representation is efficiently learnable, bridging behavioral decision theory with statistical learning theory.
On the Comparative Performance of Machine Learning and Economic Models for Risky Decision Making
Mateus Hiro Nagata
SDM 2025 (Chengdu, China) · GAIMSS 2025 (Paris, France)
What constitutes the pinnacle of decision under risk models? This paper addresses this question by comparing a range of models, from popular economic models to black-box machine learning algorithms, as well as hybrid approaches in predicting the choice of certainty equivalents of risky prospects. The findings demonstrate that there is a relevant gain in descriptive prowess in using machine learning techniques. However, this indicates heterogeneity in the population rather than inadequacy of the economic models.
Outcome Selection with Algorithmic Learners
Mateus Hiro Nagata, Francesco Giordano
JECCO 2025 (UK)
What constitutes the pinnacle of decision under risk models? This paper addresses this question by comparing a range of models, from popular economic models to black-box machine learning algorithms, as well as hybrid approaches in predicting the choice of certainty equivalents of risky prospects. The findings demonstrate that there is a relevant gain in descriptive prowess in using machine learning techniques. However, this indicates heterogeneity in the population rather than inadequacy of the economic models.
Published Papers
Retrodicting with the Truncated Lévy Flight
Raul Matsushita, P. Brom, Mateus Hiro Nagata, Sergio da Silva
Communications in Nonlinear Science and Numerical Simulation, 116 (2023)
The Duration of Historical Pandemics
Raul Matsushita, Mateus Hiro Nagata, Sergio da Silva
Communications in Nonlinear Science and Numerical Simulation, 106 (2022)
Bypassing the Truncation Problem of Truncated Lévy Flights
Raul Matsushita, Sergio da Silva, R. da Fonseca, Mateus Hiro Nagata
Physica A, 559 (2020)
An Empirical Overview of Nonlinearity and Overfitting in Machine Learning Using COVID-19 Data
Yuan Peng, Mateus Hiro Nagata
Chaos, Solitons & Fractals, 139 (2020) · 150+ citations
Guides
Academic
Useful links!
- Prorum A forum that has questions in numerous topics related to academia, organized by professor Cajueiro.
- Matheus Facure A website dedicated to causal relations, econometrics and machine learning in an unique way. Matheus Facure is an extremely talented guy working in Nubank.
- Tips4Economists A thread of useful links for econ students aiming to be a good grad researcher, by professor Masayuki Kudamatsu.
- How to Become a Pure Mathematician (or Statistician) For those who have an (unhealthy) deal of interest in mathematics (such as myself).
- quantecon If you believe that Python, Julia and Economics can change the world, this is how.
- https://aplicardesdemexico.github.io If tortillas are your main food source and you dream of doing a Ph.D. in Econ, this will help.
Page template forked from evanca