Browse the categories to access the content of academic, scientific and opinion publications of the professors and students of the Department of Economics PUC-Rio.

Forecasting Large Realized Covariance Matrice: The Benefits of Factor Models and Shrinkage

We propose a model to forecast very large realized covariance matrices of returns, applying it to the constituents of the S&P 500 on a daily basis. To deal with the curse of dimensionality, we decompose the return covariance matrix using standard firm-level factors (e.g. size, value, profitability) and use sectoral restrictions in the residual covariance matrix. This restricted model is then estimated using Vector Heterogeneous Autoregressive (VHAR) models estimated with the Least Absolute Shrinkage and Selection Operator (LASSO). Our methodology improves forecasting precision relative to standard benchmarks and leads to better estimates of the minimum variance portfolios.


Diego Siebra de Brito.

Orientador: Marcelo Medeiros.

Co-orientador: Ruy Monteiro Ribeiro.

Banca: Diogo Abry Guillén. Marcelo Fernandes.

Credit Crunches and Inequality Dynamics

I develop an entrepreneurship model with occupational choices in an environment where agents face binding credit restrictions. I show that in economies where financial markets are tighter, the distribution of wealth is characterized by higher levels of inequality. The model is consistent with documented results in the literature concerning losses in TFP and other aggregate outcomes. I also analyze the transition dynamics of the wealth distribution in the aftermath of a once-and-for-all credit crunch shock and show that wealth accumulation might mitigate the misallocation implied by such adverse shocks.


Guilherme Neves Silveira.

Orientador: Eduardo Zilberman.

Banca: Carlos Viana de Carvalho. Pedro Cavalcanti Ferreira.

Does structural change lead to inequality change? A macroeconomic approach.

While structural change literature has been mainly focused on explaining the Kuznets Facts - a set of regularities concerning sectoral dynamics throughout economic growth - important issues were left apart. Inequality was one of them: Kuznets himself, when making some of the first documentation of structural change patterns, repetitively expressed his concern that inequality and sector reallocation were linked. In this regard, we seek to extend the benchmark model of structural change to introduce wealth and income distribution. We allow idiosyncratic risky and incomplete markets in a two-sector environment of growth. In a quantitative exercise, a secular transition from a poor and good's producer economy to a richer and service-based one is conducted. We establish how a time-varying relative price of consumption and investment - yielded by the model's multi-sector structure - plays an important role in the inequality behavior. Watching a rising trend for this relative price, agents allocate more consumption to the beginning of the transition. With the subsequent rise in interest rates and the lower accumulation of capital, inequality soars within this period. We also show that, when workers may be restricted to switch sectors, income inequality jumps during the first years, while wealth inequality actually becomes smaller than in the frictionless case. Finally, we calibrate the model based on the US economy 1950-2000, obtaining a good fit for sectoral shares, and a reasonable fit for income Gini percent variations. Our consumption-investment relative price effect on inequality does not explain the rise in the Gini index after the 1980s.

Bernardo Silva de Carvalho Ribeiro.

Orientador: Eduardo Zilberman.

Co-orientador: Tiago Couto Berriel.

Banca: Carlos Viana de Carvalho. Pedro Cavalcanti Ferreira.

Legal Capacity, Historical Development, and Firm size: Evidence from Colonial Peru

The literature on economic development emphasized that historical events have long-term persistence over current social welfare. Specifically, places with historical state capacity are more developed than places without it. In this paper, we focus on how rule of law presence, one crucial dimension of state capacity, affects the dynamics and structure of firms. Therefore, we focus on one mechanism through which state capacity affects development. This relationship is analyzed for the Peruvian case. We use rich data from the national tax collector, the national economic census and a historical census of the state administration carried in 1793 to analyze the persistent effect of rule of law presence over firms’ development indicators. Our baseline results present evidence of a persistent and significant positive effect of the historical presence of rule of law over formalization, firm’s size and labor productivity; also, we find that firms prefer to relate each other through markets rather than adopt vertical structures in places with historical presence of rule of law. Our evidence is robust to the addition of important historical socioeconomic variables and the exclusion of Lima. Furthermore, the presence of courts appears to be the most important dimension of rule of law presence. Second, we analyze the effect of the historical presence of rule of law over the current presence of rule of law as a potential channel of persistence. We argue that having rule of law institutions in the colonial period affected the relative cost of subsequent investments in state capacity held during the republic period. Our findings are robust to alternative specifications (IV model and network framework). Finally, we present preliminary evidence on mechanisms. Municipalities with historical presence of rule of law have loosened financial frictions, demand less informal credit and have better conditions to obtain a functioning license. These results suggest that informal and formal rules may evolve differently and that the cost of access to rule of law-related services is important.


Alvaro Esteban Cox Lescano.

Orientador: Claudio Ferraz.

Banca: Juliano Assunção. Leonardo Monteiro Monasterio.

Macroprudential Policies at Work: How do Government-Owned Banks affect Credit Markets?

How countercyclical macroprudential credit policies affect the interest rate spread? To answer this question, we propose a banking completion model. We use data from Brazil, where government-owned banks whose outstanding loans accounted for almost half of the credit market, but also have played a strong countercyclical role in the economy. In our model, we build upon Cournot banking competition models in order to consider the differences in behavior between public and private loans, as well as to distinguishing features observed between firms and consumers credit market. Our results show that the presence of counter-cyclical public credit reduces the spread when the risk in the economy is high but, on the other hand, reduces the supply of credit provided by private banks not just through the conventional quantity channel of Cournot competition, but also by lowering the risk premium faced by the banks in credit market, which intensifies the retraction in private credit supply in economic downturns.


Paulo Rodrigo Capeleti.

Orientador: Márcio Gomes Pinto Garcia.

Co-orientador: Fábio Miessi Sanches.

Banca: Klênio de Souza Barbosa. Leonardo Rezende.

Do Neighbors Vote Alike? Evidence from the Brazilian Congress

This work analyses the presence of peer effects in the Brazilian Congress among federal deputy. I test if a deputy is influenced by its nextdoor neighbor when casting a vote for a proposition. Since politicians can select colleagues with similar political position to be their neighbors, I use an office lottery that randomly allocates offices for newcomers and test if office proximity increases the likelihood of agreement. I use data for all 1026 Brazilian federal deputies from 54th and 55th legislature elected in 2010 and 2014 and observe their votes in all propositions held between February 2011 and May 2017. I find that being next-door office neighbors does not increase the probability of agreement. Similar findings are obtained when restricting the sample for different types of proposition, for deputies from the same party, as well as for congressmen from the same state.

Fernando Martins Secco Luce.

Orientador: Claudio Ferraz.

Banca: Fábio Miessi Sanches. Carlos Eduardo Ferreira Pereira Filho.

Wage Inequality, Firms and Informality: Theory and Evidence from Brazil

The labor market in Brazil had significant changes between 2003 and 2012. Wage inequality, informality and unemployment decreased while the real minimum wage rose. Recent empirical evidence suggests that firms had an important role in these processes. This paper has two major features. First, I propose a search and matching model with heterogeneous firms and workers that takes into account several attributes of the Brazilian labor market such as informality, unemployment, minimum wage and wage variance between and within firms. Then, with an estimated model that fits important moments of the labor market in 2003, I make counterfactual exercises to quantify the determinants beneath the reduction of wage inequality.


Roberto Hsu Rocha.

Orientador: Gabriel Ulyssea.

Banca: Claudio Ferraz. Miguel Nathan Foguel.

Why do Brazilian bank-affiliated mutual funds underperform?

This paper investigates financial conglomerates' participation in the Brazilian equity mutual fund industry. Using data from 2002 to 2016, we show that bank-affiliated funds underperform funds managed by stand-alone entities by 1.96\%-2.30\% per year. Moreover, we find that bank-affiliated fund managers have less incentives to take risk than independent funds'. Consistent with incentives, we show that bank-affiliated funds trade less often, try less to time the market and have portfolios more similar to the market's than independent funds. Finally, we show that differences in risk taking can be associated to 7.68-29.6\% of the performance difference between bank-affiliated and independent funds.


Pompeu Hoffmann Junior.

Orientador: Ruy Monteiro Ribeiro.

Co-orientador: Walter Novaes.

Banca: Alexandre Lowerkron. Marco Bonomo.

Aviation Technology and Air Traffic Networks

This paper studies to what extent the development of new aircraft shapes airlines' network structure. I argue that modern aircraft are more efficient and well suited to operate flights between smaller and less central cities, hence favoring the service of more markets in the periphery of the network. Using U.S. air traffic data, I employ a discrete choice framework to model airlines' entry decisions and the subsequent aircraft choice to each market. Counterfactual experiments show that had aircraft technology ceased to improve in 1999, the air traffic network as a whole would be more centralized, airlines would be operating more hub-centered networks, reaching fewer cities, and serving fewer markets


Bruno Henrique Castelo Branco.

Orientador: Pedro Carvalho Loureiro de Souza.

Co-orientador: Leonardo Rezende.

Banca: Fábio Miessi Sanches. João Paulo Cordeiro De Noronha Pessoa.

Labor Market Conditions and Gender Inequality: Evidence from the Brazilian Trade Liberalization

This study investigates the effect of a plausibly exogenous trade shock on gender inequality in labor market outcomes, occupational and schooling choices. In the 1990's, Brazilian government decided to reduce import tariffs, inducing a large, once and for all trade liberalization, with heterogeneous effects across local economies. Using Brazilian Decennial Censuses, I estimate medium (1991-2000) and long (1991-2010) run effects on women's and men's outcomes and gender inequality in labor market. Results point that in harder hit regions women faced an increase in non-employment relative to men in the medium run, inducing a reduction in the share of women among those employed. Besides that, the gender wage gap increased in the medium run, but these effects did not persist in the long term.

M 395

Giovanna Ribeiro Paiva de Souza.

Orientador: Gabriel Ulyssea.

Banca: Cecilia Machado. Miguel Nathan Foguel.

Drug markets and violence: evidence from Brazil

Brazil experienced income growth during the 2000's while at the same time sustaining high homicide rates, especially in the North and Northeast regions where homicide rates soared. Using data from the Brazilian Health Ministry on drug related hospitalization, we provide preliminary evidence on the possible positive impact of income on homicide rates, through the drug markets channel.


Guilherme Coelho Netto Avelar.

Orientador: Claudio Ferraz.

Banca: Gustavo Gonzaga. Joana da Costa Martins Monteiro.

The Mechanisms Behind the Effect of Social Media on Protesting Behavior: Evidence from Brazil

I show that social media activity, through Twitter, positively affected protesting behavior during the Brazilian demonstrations of June 2013. I use variation in internet quality and access across municipalities as a source of exogenous variation in Twitter activity, in order to overcome a problem of reverse causality. I find, for the preferred specifications, that a 10\% increase in Twitter activity led to an increase of 16\% in the number of protestors in the streets and an increase of 10.2\% on the probability of an event happening. Furthermore, results indicate that the mechanism behind the effect of social media in protesting behavior was an intra-day pre-protest coordination, rather than live coordination, or days ahead coordination, and that negative propaganda against the government did not play a role through social media.

M 397

Felipe de Almeida Alvarenga Pereira.

Orientador: Pedro Carvalho Loureiro de Souza.

Banca: Gabriel Ulyssea. Rudi Rocha. Francisco Junqueira Moreira da Costa.

Financial institutions, growth, and inequality: A quantitative exploration of financial development in Brazil

Starting on the early 2000s, financial depth and access to financial services surged in Brazil. The ratio of external finance to GDP increased from just over 50% to 110% from 2003 to 2012. During this period, the Brazilian economy also experienced strong growth with decreasing income inequality. The objective of this work is to gain perspective on the aggregate growth effects and the distributional consequences of the financial development as observed in Brazil from 2003 to 2012 through the lens of a dynamic model with financial frictions, where agents who differ in their ability as workers and entrepreneurs make occupational and productive choices under credit constraints. The model yields predictions of income inequality and wealth mobility, as well as productivity and the size distribution of firms. We calibrate the model to match the Brazilian economy before the financial reforms and use it to quantify the consequences of reducing financial frictions on TFP, GDP, and inequality.


Pedro Martins Pessoa.

Orientador: Juliano Assunção.

Banca: Eduardo Zilberman. Felipe Iachan.

The Fiscal Theory of the Price Level with Nominal Revenues and Expenditures

The usual assumption that fiscal policy is set in real terms is neither realistic nor innocuous. In this article, I propose a model that accounts for the existence of nominal revenues and expenditures. This creates an unexplored channel through which monetary and fiscal policies interact. I show that, in this environment, the price level can be fiscally determinate, even when all government debt is real. Also, the effects of monetary and fiscal policies are sensible to the degree of indexation in the government budget. For instance, a monetary policy tighten can cause a temporarily reduction of inflation, in the short-run. In order to gauge how relevant are these nominal components, using Bayesian techniques, I estimate the model for the US economy.


Marcos Kiehl Sonnervig.

Orientador: Tiago Couto Berriel.

Co-orientador: Carlos Viana de Carvalho.

Banca: Eduardo Henrique de Mello Motta Loyo. Eduardo Zilberman.

Uma proxy para aversão ao risco avaliada no mercado de ações

Diferentes medidas de aversão ao risco têm sido propostas pela literatura, entretanto nenhuma delas é fortemente aceita pelos pesquisadores. A partir do valor das aposta nos cassinos, eu calculo a propensão a apostar de cada período e uso essa medida como uma proxy da aversão ao risco agregada. Dessa forma, eu consigo uma medida de aversão ao risco que pode variar com o tempo. Para avaliar essa medida, eu estudo a ligação que essa proxy tem com o ciclo econômico e a sua capacidade de prever o retorno da carteira de mercado no longo prazo.

M 393

Daniel Lívio Alencar Cordeiro.

Orientador: Ruy Monteiro Ribeiro.

Co-orientador: Eduardo Zilberman.

Banca: João Vitor Issler. Marcelo Medeiros.

Does information on school quality affect voting? Evidence from Brazil

This paper examines if voters act upon information about public service delivery. We explore a natural experiment in Brazil, which provided an objective measure of quality for some public schools, but not for others. To use this variation, we look at polling stations that are located at municipal schools and compare electoral outcomes in mayoral elections in informed and non-informed groups of voters, before and after the information release. We find that, on average, receiving information about school quality does not affect the incumbents’ electoral outcomes. When taking into account the content of the information received, good performance slightly increases the support for the incumbent.

M 391

Marina Villas Boas Dias.

Orientador: Claudio Ferraz.

Banca: Gabriel Ulyssea. Vladimir Pinheiro Ponczek .

Public Sector and the Allocation of Skills in the Labor Market

This paper investigates how the size of public sector employment affects the allocation of skills in the private sector. I develop a Roy model were workers self-select into either public or private sector. Workers are heterogeneous in their skills and risk aversion. There are some distinguishing features in public sector employment that make public sector wages more certain and disproportionately high for some skill levels. Those differences may influence workers' sector decision unevenly across skills, affecting the skill distribution available to the private sector. The private sector is characterized by positive assortative matching between skills and tasks. Changes in the skill distribution available to the private sector have effects on the allocation of skills to tasks which has consequences on wage inequality and productivity in the private sector. I estimate this model for Brazil using worker level data for the years of 2011-4 and use it to perform counter-factual exercises. Results show that removing the public sector from the economy increases private sector average productivity, decreases the college wage premium, but increases wage inequality.


Ana Beatriz Ract Pousada.

Orientador: Gabriel Ulyssea.

Banca: Cecilia Machado. Claudio Ferraz.

The politics of government advertisement: evidence from Brazil

Using an unique data set of central government expenditure on advertising in Brazil, we shed light on the behavior of public advertisers and on the relation between government ads and voting. In particular, we investigate political motivations behind the allocation of the advertisement budget by the federal government and its impacts on voting. Borrowing insights from the literature of distributive politics in Political Science, we first correlate ads money and votes for the government's party on the local level. Next, we exploit plausible exogenous variation on FM radio signal coverage to test if money spent on ads turn into votes for the government's party. Our findings show that although past presidential election outcomes predict where in the territory the government places ads, voters do not seem to be persuaded by those ads to favor the party in power.


Bernardo Barboza Ribeiro.

Orientador: Claudio Ferraz.

Banca: Juliano Assunção. Raphael Bottura Corbi .

Product discovery in the PC games market

This paper investigates the role of product discovery in the demand for video games. We show that the lifetime sales patterns for video games vary widely, with some games selling most of their units in the first months after launch and others having longer tails. To understand these differences we propose a demand model by which consumers are periodically informed about the existence of a game and explore the lifetime sales patterns that this implies. We then take it to the data using price and sales figures from the Steam digital platform and web search figures from google trends. Our results imply that sales 3 months after launch are on average half of what the should be were consumers fully informed.

M 388

Luís Paulo Fernandes Bretanha Jorge.

Orientador: Leonardo Rezende.

Banca: Fábio Miessi Sanches. Andre Garcia de Oliveira Trindade.

Search here