Eventos

Seminario CEDE - Paul Soto

This paper examines the link between industrial production and the sentiment expressed in natural language survey responses from U.S. manufacturing firms. We compare several natural language processing (NLP) techniques for classifying sentiment, ranging from dictionary-based methods to modern deep learning methods. Using a manually labeled sample as ground truth, we find that deep learning models--partially trained on a human-labeled sample of our data--outperform other methods for classifying the sentiment of survey responses. Further, we capitalize on the panel nature of the data to train models which predict firm-level production using lagged firm-level text. This allows us to leverage a large sample of "naturally occurring" labels with no manual input. We then assess the extent to which each sentiment measure, aggregated to monthly time series, can serve as a useful statistical indicator and forecast industrial production. Our results suggest that the text responses provide information beyond the available numerical data from the same survey and improve out-of-sample forecasting; deep learning methods and the use of naturally occurring labels seem especially useful for forecasting. We also explore what drives the predictions made by the deep learning models, and find that a relatively small number of words--associated with very positive/negative sentiment--account for much of the variation in the aggregate sentiment index.

12:30 pm

Universidad de los Andes

Activo

Seminario CEDE - Guido Alfani

Recent research in economic history has unearthed previously unknown facts about the long-term trends in inequality. We now have, for at least some areas of Europe, continuous time series of key inequality indicators from ca. 1300. Most of these series have resulted from the research conducted by the ERC-funded project EINITE Economic Inequality across Italy and Europe 1300-1800 and by its follow-up project, SMITE Social Mobility and Inequality across Italy and Europe 1300-1800. These new data are changing the way in which we perceive economic inequality not only in the past, but even today as a key lesson from history, is that economic inequality (especially, but not only, of wealth) has a marked tendency for increasing over time, and only catastrophes on the scale of the Black Death or the World Wars managed to bring it down, albeit temporarily. Additionally, the new historical evidence is also relevant to the debate about the long-term determinants of inequality growth. This seems to be independent, to a large degree at least, from economic growth. Other factors seem to have played a crucial role, including institutional factors and in particular (in the early modern period) the rise of the fiscal-military state. The seminar will provide an overview of recent research on inequality in preindustrial Europe, with a particular focus on some areas on which research is currently ongoing, in particular the southern Italian regions of Apulia and Sicily.

12:30 pm

Universidad de los Andes

Activo

Ejercer Economía | Mateo Salazar

Conoce más de nuestro invitado: LinkedIn Mateo Salazar

12:30 pm

Universidad de los Andes

Activo

Seminario PePe - Angela Amaya

¿Sabías que la conformación del capital para financiar las pensiones va más allá de la cotización a pensiones? La oficina de Bonos Pensionales del Ministerio de Hacienda y Crédito Público tiene a su cargo la administración y pago de varios mecanismos que resultan ser aportes para la conformación del capital que posteriormente se traduce en mesadas pensionales.

12:30 pm

Universidad de los Andes

Activo

Seminario CEDE - Kirill Borusyak

We examine how to interpret estimates from a commonly used migration regression relating changes in local population to exogenous local labor demand shocks. Using a simple model of local labor markets with mobility costs, we find that most conclusions drawn from migration regression estimates are likely to be substantially misleading. Intuitively, the conventional migration regression is misspecified due to the bilateral nature of location choices. Workers choose where to live based not only on the shock to their current location, but also on the shocks to potential alternative locations, which are omitted from the regression. Analytical results and simulations based on Brazilian data show that conventional migration regression estimates are inaccurate for the local population effects of either shocks to individual locations or all observed shocks taken together and often substantially understate the amount of worker reallocation driven by observed shocks. These problems are particularly acute when workers face industry switching costs in addition to geographic mobility costs. Simple alternative approaches leveraging the models structure exhibit far better performance.

12:30 pm

Universidad de los Andes

Activo

Seminario CEDE - Diego Ramos

We examine perceptions of the history of race in the United States and its impact on present-day political polarization. Based on survey data from 14,044 US respondents, we examine historical narratives surrounding key racial events among both white and Black individuals. Our analysis unveils notable discrepancies in beliefs regarding the causes of the Civil War, the achievements of the Civil Rights Movement, and the enduring legacy of slavery on Black Americans today. Ideological divisions in historical interpretations, particularly among white respondents, emerge prominently, while differences across race and generation are less pronounced. Moreover, we investigate the political ramifications of these historical viewpoints through two experimental interventions. Participants were randomly prompted to contemplate their own perspectives on these issues and to confront the views of others. The results of our experiments indicate that historical narratives about race can exacerbate polarization in institutional satisfaction. Liberals exhibit heightened pessimism and dissatisfaction with the current institutional framework as a result of the treatments, while conservatives remain largely unaffected. This study underscores the significance of understanding the origins and repercussions of historical accounts concerning intergroup tensions, which may contribute to contemporary political divisions.

12:30 pm

Universidad de los Andes

Activo

Seminario PePe - María Paula Vargas

María Paula Vargas Urquijo es Economista y Profesional de Gobierno y Asuntos Públicos de la Universidad de los Andes y cuenta con una Especialización en Negociación de la misma Universidad. Ha trabajado en la Subdirección de Financiamiento Externo y Relaciones con Inversionistas de la Nación desde marzo del 2019, donde ha mantenido la relación con inversionistas institucionales de deuda y calificadoras de riesgo, ha emitido deuda externa por más de USD 17.000 millones y ha trabajado en el financiamiento sostenible del país mediante la construcción del Marco de Referencia de Bonos Verdes, Sociales y Sostenibles de Colombia, la Taxonomía Verde y la emisión de los bonos inaugurales verdes y sociales del país.

12:30 pm

Universidad de los Andes

Activo

Seminario CEDE - Tomás Rodríguez

In this paper we study the relationship between ideological segregation in Twitter and external events in the context of the "estallido social" of 2021 in Colombia. We analyze the tweeting behavior of approximately 37,000 users active in political discussions during the two most intense months of the 2021 "estallido social" in Colombia, supplemented by event data from ACLED. Our findings reveal a significant increase in political segregation within daily weighted and directed re-tweet networks of right, left, and center groups during spikes in street violence. Notably, around critical days, defined as days with violent activity spikes over two standard deviations above the mean users exhibited a 20 to 30 percentage point increase in attention towards individuals of their own ideology, with a corresponding decrease in attention towards those of differing ideologies. This endogamic shift was most pronounced among right-wing users, followed by the left-wing users. Users from the political center and unclassified users showed a less pronounced shift in behavior. These changes were primarily driven by sporadic participants in Twitter conversations during critical periods, while regular re-tweeters in fact displayed an increase in exogamy during these times.

12:30 pm

Universidad de los Andes