Aggregate vs disaggregate data analysis: a paradox in the estimation of money demand function of Japan under the low interest rate policy

We investigate the issue of whether there was a stable money demand function for Japan in 1990's using both aggregate and disaggregate time series data. The aggregate data appears to support the contention that there was no stable money demand function. The disaggregate data shows that there was a stable money demand function. Neither was there any indication of the presence of liquidity trapo Possible sources of discrepancy are explored and the diametrically opposite results between the aggregate and disaggregate analysis are attributed to the neglected heterogeneity among micro units. We also conduct simulation analysis to show that when heterogeneity among micro units is present. The prediction of aggregate outcomes, using aggregate data is less accurate than the prediction based on micro equations. Moreover. policy evaluation based on aggregate data can be grossly misleading.

Bias from ignoring price dispersion in demand estimation

Consumers often pay different prices for the same product bought in the same store at the same time. However, the demand estimation literature has ignored that fact using, instead, aggregate measures such as the “list” or average price. In this paper we show that this will lead to biased price coefficients. Furthermore, we perform simple comparative statics simulation exercises for the logit and random coefficient models. In the “list” price case we find that the bias is larger when discounts are higher, proportion of consumers facing discount prices is higher and when consumers are more unwilling to buy the product so that they almost only do it when facing discount. In the average price case we find that the bias is larger when discounts are higher, proportion of consumers that have access to discount are similar to the ones that do not have access and when consumers willingness to buy is very dependent on idiosyncratic shocks. Also bias is less problematic in the average price case in markets with a lot of bargain deals, so that prices are as good as individual. We conclude by proposing ways that the econometrician can reduce this bias using different information that he may have available.

Inference in differences-in-differences with few treated groups and heteroskedasticity

Differences-in-Differences (DID) is one of the most widely used identification strategies in applied economics. However, how to draw inferences in DID models when there are few treated groups remains an open question. We show that the usual inference methods used in DID models might not perform well when there are few treated groups and errors are heteroskedastic. In particular, we show that when there is variation in the number of observations per group, inference methods designed to work when there are few treated groups tend to (under-) over-reject the null hypothesis when the treated groups are (large) small relative to the control groups. This happens because larger groups tend to have lower variance, generating heteroskedasticity in the group x time aggregate DID model. We provide evidence from Monte Carlo simulations and from placebo DID regressions with the American Community Survey (ACS) and the Current Population Survey (CPS) datasets to show that this problem is relevant even in datasets with large numbers of observations per group. We then derive an alternative inference method that provides accurate hypothesis testing in situations where there are few treated groups (or even just one) and many control groups in the presence of heteroskedasticity. Our method assumes that we can model the heteroskedasticity of a linear combination of the errors. We show that this assumption can be satisfied without imposing strong assumptions on the errors in common DID applications. With many pre-treatment periods, we show that this assumption can be relaxed. Instead, we provide an alternative inference method that relies on strict stationarity and ergodicity of the time series. Finally, we consider two recent alternatives to DID when there are many pre-treatment periods. We extend our inference methods to linear factor models when there are few treated groups. We also derive conditions under which a permutation test for the synthetic control estimator proposed by Abadie et al. (2010) is robust to heteroskedasticity and propose a modification on the test statistic that provided a better heteroskedasticity correction in our simulations.

Inference in differences-in-differences with few treated groups and heteroskedasticity

Differences-in-Differences (DID) is one of the most widely used identification strategies in applied economics. However, how to draw inferences in DID models when there are few treated groups remains an open question. We show that the usual inference methods used in DID models might not perform well when there are few treated groups and errors are heteroskedastic. In particular, we show that when there is variation in the number of observations per group, inference methods designed to work when there are few treated groups tend to (under-) over-reject the null hypothesis when the treated groups are (large) small relative to the control groups. This happens because larger groups tend to have lower variance, generating heteroskedasticity in the group x time aggregate DID model. We provide evidence from Monte Carlo simulations and from placebo DID regressions with the American Community Survey (ACS) and the Current Population Survey (CPS) datasets to show that this problem is relevant even in datasets with large numbers of observations per group. We then derive an alternative inference method that provides accurate hypothesis testing in situations where there are few treated groups (or even just one) and many control groups in the presence of heteroskedasticity. Our method assumes that we know how the heteroskedasticity is generated, which is the case when it is generated by variation in the number of observations per group. With many pre-treatment periods, we show that this assumption can be relaxed. Instead, we provide an alternative application of our method that relies on assumptions about stationarity and convergence of the moments of the time series. Finally, we consider two recent alternatives to DID when there are many pre-treatment groups. We extend our inference method to linear factor models when there are few treated groups. We also propose a permutation test for the synthetic control estimator that provided a better heteroskedasticity correction in our simulations than the test suggested by Abadie et al. (2010).

Consumption smoothing in Latin America: an empirical assessment of present value models

The study aims to assess the empirical adherence of the permanent income theory and the consumption smoothing view in Latin America. Two present value models are considered, one describing household behavior and the other open economy macroeconomics. Following the methodology developed in Campbell and Schiller (1987), Bivariate Vector Autoregressions are estimated for the saving ratio and the real growth rate of income concerning the household behavior model and for the current account and the change in national cash ‡ow regarding the open economy model. The countries in the sample are considered separately in the estimation process (individual system estimation) as well as jointly (joint system estimation). Ordinary Least Squares (OLS) and Seemingly Unrelated Regressions (SURE) estimates of the coe¢cients are generated. Wald Tests are then conducted to verify if the VAR coe¢cient estimates are in conformity with those predicted by the theory. While the empirical results are sensitive to the estimation method and discount factors used, there is only weak evidence in favor of the permanent income theory and consumption smoothing view in the group of countries analyzed.

Restrição ao crédito educacional e investimentos em capital humano no Brasil: evidências através dos retornos à educação

Este artigo procura identificar o papel de restrição ao crédito sobre a decisão de investimento em capital humano no Brasil. Para tanto é usada a hipótese de que custos diretos e de oportunidade de estudar afetam pessoas restritas e não restritas por crédito de forma diferente. Enquanto o custo de oportunidade afeta a todos de forma similar, o custo direto afeta mais fortemente aqueles restritos ao crédito. A partir da estimação de taxas de retorno à educação com o uso de diferentes variáveis instrumentais é possível lançar luz sobre o papel da restrição ao crédito sobre escolaridade. Ao comparar as taxas de retorno utilizando como instrumento a oferta relativa de professores (Proxy dos custos diretos) e salário mediano dos jovens (variável de custo de oportunidade), encontramos uma acentuada diferença no ponto estimado, indicando a existência de efeito da restrição sobre escolaridade. Entretanto, o mau desempenho do segundo instrumento no segundo estágio torna problemática a comparação.

State–society cycles and political pacts in a national–dependent society: Brazil

The history of independent Brazil may be divided into three major state–society cycles, and, after 1930, five political pacts or class coalitions can be identified. These pacts were nationalist; only in the 1990s did the Brazilian elites surrender to the neoliberal hegemony. Yet, since the mid-2000s they have been rediscovering the idea of the nation. The main claim of the essay is that Brazilian elites and Brazilian society are “national–dependent”, that is, they are ambivalent and contradictory, requiring an oxymoron to define them. They are dependent because they often see themselves as “European” and the mass of the people as inferior. But Brazil is big enough, and there are enough common interests around its domestic market, to make the Brazilian nation less ambivalent. Today Brazil is seeking a synthesis between the last two political cycles – between social justice and economic development in the framework of democracy.

The impact of structured teaching methods on the quality of education

This paper estimates the impact of the use of structured methods on the quality of education of the students in primary public school in Brazil. Structure methods encompass a range of pedagogical and managerial instruments applied to the education system. In recent years, several municipalities in the State of São Paulo have contracted out private educational providers to implement these structured methods in their schooling system. Their pedagogical proposal involves structuring curriculum contents, elaboration and use of teachers and students textbooks, and training and supervision of the teachers and instructors. Using a difference in differences estimation strategy, we find that the fourth and eighth grader students in the municipalities with structured methods performed better in Portuguese and Math than students in municipalities not exposed to the methods. We find no differences in approval rates. However, a robustness check is not able to discard the possibility that unobserved municipal characteristics may affect the results.

Análise de resultados intermediários das políticas de bônus em escolas públicas estaduais brasileiras

Programas de bônus para servidores de escolas públicas, visando estimular a melhoria da aprendizagem dos alunos, foram adotados por diferentes estados brasileiros entre 2008 e 2011. A presente pesquisa teve como objetivo avaliar o impacto desses programas em três dimensões: 1. práticas pedagógicas dos professores; 2. absenteísmo e rotatividade dos professores ou visão do diretor ou professor sobre a assiduidade e rotatividade do professor; 3. interlocução com familiares para garantir a frequência dos alunos. A coleta de dados foi feita a partir dos questionários da Prova Brasil, respondidos por diretores, professores e alunos das escolas do grupo de tratamento e do grupo controle. Foram considerados na análise: a variação no ano de implementação dos programas (dose), seus diferentes desenhos (tipo) e a série/ano dos alunos (série) que responderam a Prova Brasil. A abordagem metodológica adotada para a análise proposta foi a de pareamento pelo escore de propensão, seguido do cálculo de diferenças em diferenças, entre os Estados que implementaram a política e os que não adotaram. Os resultados variam de acordo com a série e dose, havendo impacto positivo para algumas categorias de análise na dimensão 1, relacionado à frequência com que professores corrigem o dever de casa. Na dimensão 2, os resultados apontam para uma redução de preocupação dos diretores com a rotatividade dos professores e aumento da percepção de que o absenteísmo dos docentes é um problema na realidade escolar pública brasileira. Quanto à dimensão 3, os dados revelam, de modo geral, ausência de impacto. Permanece a dúvida sobre a eficiência e a eficácia do estímulo financeiro para produzir mudanças no comportamento dos professores e diretores que possam promover resultados intermediários na melhoria da aprendizagem dos alunos.

Current account and capital mobility hypothesis: Evidence from the G-7

This paper investigates an intertemporal optimization model in order to analyze the current account of the G-7 countries, measured as the present value of the future changes in net output. The study compares observed and forecasted series, generated by the model, using Campbell & Shiller’s (1987) methodology. In the estimation process, the countries are considered separately (with OLS technique) as well as jointly (SURE approach), to capture contemporaneous correlations of the shocks in net output. The paper also proposes a note on Granger causality and its implications to the optimal current account. The empirical results are sensitive to the technique adopted in the estimation process and suggest a rejection of the model in the G-7 countries, except for the USA and Japan, according to some papers presented in the literature.

A utilização de modelos de duração para gerenciar risco de crédito

Os modelos hazard, também conhecidos por modelos de tempo até a falência ou duração, são empregados para determinar quais variáveis independentes têm maior poder explicativo na previsão de falência de empresas. Consistem em uma abordagem alternativa aos modelos binários logit e probit, e à análise discriminante. Os modelos de duração deveriam ser mais eficientes que modelos de alternativas discretas, pois levam em consideração o tempo de sobrevivência para estimar a probabilidade instantânea de falência de um conjunto de observações sobre uma variável independente. Os modelos de alternativa discreta tipicamente ignoram a informação de tempo até a falência, e fornecem apenas a estimativa de falhar em um dado intervalo de tempo. A questão discutida neste trabalho é como utilizar modelos hazard para projetar taxas de inadimplência e construir matrizes de migração condicionadas ao estado da economia. Conceitualmente, o modelo é bastante análogo às taxas históricas de inadimplência e mortalidade utilizadas na literatura de crédito. O Modelo Semiparamétrico Proporcional de Cox é testado em empresas brasileiras não pertencentes ao setor financeiro, e observa-se que a probabilidade de inadimplência diminui sensivelmente após o terceiro ano da emissão do empréstimo. Observa-se também que a média e o desvio-padrão das probabilidades de inadimplência são afetados pelos ciclos econômicos. É discutido como o Modelo Proporcional de Cox pode ser incorporado aos quatro modelos mais famosos de gestão de risco .de crédito da atualidade: CreditRisk +, KMV, CreditPortfolio View e CreditMetrics, e as melhorias resultantes dessa incorporação

Municipalização do ensino fundamental da rede pública: os impactos sobre o desempenho escolar

O Fundef, implantado em 1998, pode ser considerado uma das reformas educacionais mais importante nos últimos anos para o País. Com a descentralização fiscal estimulada pelo fundo, um dos efeitos líquidos mais perceptíveis, além do aumento inicial do número de crianças matriculadas, foi a municipalização da rede de ensino público do ensino Fundamental, no período entre 1998 e 2004. Este trabalho mostra o efeito da municipalização sobre a variação das taxas de desempenho escolar dos alunos. Utilizando dados do Censo Escolar para os anos de 1998 e 2004, três modelos são propostos para avaliação deste impacto através das metodologias de primeira diferença e de diferenças em diferenças. Eles sugerem que a municipalização tem impacto sobre o desempenho, porém sua magnitude é pequena. De forma geral, este processo acabou comprometendo a qualidade do ensino. Ademais, através da comparação entre as duas redes, é possível apontar variáveis da infra-estrutura escolar, como tamanho de turma e tamanho de escola, que influenciam o desempenho escolar.

Crescimento econômico secular no Brasil, modelo de thirlwall e termos de troca

Explaining the Brazilian economic growth performance during the 20th is a challenging task to academics. As shown, Brazil had already experienced faster growth and recently it is far from such belle époque. Most explanation lies on conventional frameworks. This paper assesses the Thirlwall’s balance-of-payments constraint model to Brazilian economic growth in the period 1900-2005, highlighting the importance of the terms of trade. The results show, first, that the terms of trade are significant in the estimation of import demand function and change the income elasticity, especially in the first period of the sample (1900-1970), and, second, when the simplest Thirlwall’s model is controlled by terms of trade, it seems that such variable is important through export growth, which cause changes in economic performance.

Decentralization and education performance: a first view to the Brazilian process

This paper analyses the impact of the decentralization in educational system that is taking place in Brazil in the last decade, as a result of several laws that encourage municipalities to invest in fundamental education. The proficiency tests undertaken by the government allows to follow some public schools in two points in time. Therefore we were able to create an experimental group with the schools that were under state system in the SAEB exam and have migrated to the municipality system by the time of Prova Brasil and a control group with the schools that were under the state system between the two exams and compare the difference in their results using a fixed effect panel data analysis. The difference in difference estimator indicates that there is no significant change in the performance of the students.

Estimativas para a elasticidade-preço da demanda por produtos siderúrgicos no Brasil

Essa dissertação apresenta estimativas para a elasticidade-preço da demanda por aço no Brasil, a partir de dados agregados e desagregados da indústria siderúrgica. Os resultados das estimativas a partir do painel com dados desagregados sugerem que existe um viés de agregação nas estimativas já realizadas a partir de dados agregados, e esse viés subestimaria a elasticidade-preço do setor siderúrgico. Com a finalidade de comparar as relações entre as elasticidades-preços de curto e longo prazo foram estimados painéis heterogêneos dinâmicos, através de estimadores Mean Group (MG) e Pooled Mean Group (PMG). É importante ressaltar que, de acordo com o conhecimento do autor, este é o primeiro estudo a usar estimação em painel para estimação da elasticidade-preço da demanda por produtos siderúrgicos no Brasil, dessa forma, controlando a estimativa pela heterogeneidade entre os tipos de aço.