Embedded Files
"Labor Supply Responses with Lack of Pension Knowledge: Evidence from Linked Administrative-Survey Data" (2024). Submitted. Job Market Paper.
I study the impact of a pension reform that changes the requirements for government transfers. Using a unique database that links individual administrative records with a representative longitudinal survey, I estimate the causal impacts using a difference-in-differences approach. I find that some people do not behave as predicted by a full-information standard model because individuals underestimate pension wealth and overestimate monthly contributions and contribution rates. Thus, people lacking pension knowledge behave according to their perceived incentives, which might be incorrect. A model in the presence of sources of error about the pension rationalizes the results.
"Credit Renegotiations: Characterization and Determinants using Population Administrative Data" (2025)". (joint with Diego Beas).
We analyze consumer, mortgage, and commercial loans, as well as their respective renegotiation rates, offering new insights into their incidence and effectiveness. The main contributions can be summarized in five points. First, we construct a unique dataset using population-level administrative data on individuals and businesses in Chile between 2016 and 2024. Second, we estimate that, on average, 1.3% of consumer loans, 0.3% of commercial loans, and 0.18% of mortgage loans are renegotiated monthly. These renegotiated amounts represent 1.6%, 0.3%, and 0.13% of the total credit for each category, respectively. Third, conditional on renegotiation, the average number of renegotiations per mortgage and consumer loan ranges between 1.1 and 1.4, with a maximum number of renegotiations between 5 and 8, depending on the year analyzed. %There are no significant differences when examining loans of varying amounts and durations. Fourth, focusing on renegotiated loans, we define a "successful renegotiation" as one that restores the borrower's ability to make regular payments, i.e., the loan does not fall into delinquency within twelve months. Under this definition, 50% of consumer loans, 48% of commercial loans, and 35% of mortgage loans are successful. However, this success rate depends on the conditions at the time of renegotiation. If a loan is delinquent or remains in that condition after renegotiation, the success rate changes drastically. For delinquent loans, the success rate is 21% for mortgage loans, 31% for consumer loans, and 28% for commercial loans. For non-delinquent loans, the success rate is 80% for mortgage, 70% for consumer, and 65% for commercial loans. Fifth, we apply artificial intelligence techniques to model the probability of a successful renegotiation and to identify its determinants. Non-linear models such as LightGBM outperform traditional linear models like logistic regression, achieving AUROC values above 70% across all loan types. Using SHAP values (Shapley Additive exPlanations), we identify the most critical factors influencing a successful renegotiation, distinguishing by delinquency status. For non-delinquent loans, the maximum number of days in arrears before renegotiation and the client's tenure with the financial institution are relevant for all three loan types. For mortgage and consumer loans, the ratio of renegotiated debt to total debt is a significant predictor. For commercial loans, the ratio of delinquent debt to total debt helps explain the likelihood of future delinquencies.
"Pension Knowledge and the Incomplete Take-Up of Pension Benefits". (joint with Clement Joubert). Draft coming soon.
This paper examines the role of imperfect pension knowledge on the incomplete take-up of a means-tested pension benefit in Chile, leveraging a longitudinal representative household survey with a rich module on pension knowledge and program awareness combined to administrative data on pension rights and social pension applications. 55% of eligible individuals are unaware of the program and only 48% understand the requirements in the years preceding eligibility. Take-up measures that use self-reported data as in much of the literature significantly overestimate non-take up and generate biased distributional impacts of imperfect take-up on household income among the elderly. Pension literacy pre-retirement is strongly correlated with eventual take-up of the benefits, after controlling for a rich set of individual characteristics.
"Asia's Renewable Energy Future: A Quantitative Analysis of Development" (2025) (joint with Pawan Thiruverkadu). New Draft coming soon.
This paper studies the current state of energy production, usage, and energy demand overall in the Asia Pacific region. Using data from the World Bank and the International Energy Agency (IEA), we analyze the case of the top 9 Asian countries: China, Japan, India, the Republic of Korea, Indonesia, Thailand, Singapore, Vietnam, and the Philippines. We also propose a simple Renewable Index that measures various metrics of the current situation in each country, including GDP, non-renewable energy production mix, whether they are net energy importers, whether they have outlined frameworks to increase renewable usage, and their CO2 emissions in tons. Through this index, this paper enables the ranking of countries within the scope for the most rewarding outcomes, where the focus of further adoption assistance would be most beneficial. We finally make a detailed comparison of the most critical reforms in Japan, Korea, and India. Since Japan and Korea are considering a more nuclear-heavy transition, compared to India, which is prioritizing solar as a more prominent source. With the FYP (China), the Strategic Energy Plan (Japan), and the National Action Plan on Climate Change (India), the largest countries offer a guiding framework for what works and what does not work in terms of policy concerning the adoption of renewables.
"Artificial Intelligence and Automation and its Effect on Foreign Labor" (2024). (joint with George Johnson). New Draft coming soon.
This paper studies artificial intelligence’s (AI) impact on skilled and unskilled immigrant labor in the United States (U.S.). By utilizing other studies, papers, graphs, and data the impact of AI on foreign born workers can be illustrated. Based on the findings and trends, we discussed policy reformations and laws. These policy ideas will help lessen the friction between AI and foreign-born workers.
Other Working Papers
"Labor Cost of Mental Health: Evidence from Chile"(2019) (joint Jaime Ruiz-Tagle) Working Paperswp468, University of Chile, Department of Economics.
“Children of Unilateral Divorce Regime: New Evidence from the HRS” (joint with Suchika Chopra). Preliminary results available upon request.
"Economic growth, natural disasters and climate change: New empirical estimates," (joint Ramón López & Vinod Thomas) Working Papers wp434, University of Chile, Department of Economics.
"Climate Change and Natural Disasters," (joint Ramón López & Vinod Thomas) Working Paperswp414, University of Chile, Department of Economics.
"Where are the Missing Babies: The Role of Higher Education Access on Family Planning", (joint Fabian Duarte & Valentina Paredes)
Page updated
Google Sites
Report abuse