Impacto da COVID-19 nos principais indicadores econômicos do Brasil

Carregando...
Imagem de Miniatura

Data

2022-03-17

Autores

Mazuqueli, Juliana Ricci

Título da Revista

ISSN da Revista

Título de Volume

Editor

Universidade Estadual Paulista (Unesp)

Resumo

In the years 2020 and 2021, the world faced the COVID-19 pandemic, reflecting directly on the economy of countries, this change is noticeable by the analysis of economic indicators. In this work, in order to be able to identify the changes present in the Brazilian economy, temporal regression models were applied, composed of dummy and harmonic variables, with periods defined in advance through the analysis of the periodogram. For analysis, monthly values of Gross Domestic Product (GDP), General Market Price Index, Unemployment Rate, National Consumer Price Index and Broad Consumer Price Index, from January 2018 to September 2021. For the model, the dummy variables were defined, Zit , representing whether there was a change between the pre-pandemic period (2018-2019) and during the pandemic (2020- 2021), and Zitt, representing whether there was a change in the slope of the line between the two periods. Residual analysis consisted of the ShapiroWilk test, Runs test and Durbin-Watson test, the metrics for adequacy were the root squared error (RMSE), mean absolute error (MAE) and the coefficient of determination (R2). The models for the National Consumer Price Index, Broad Consumer Price Index, GDP and General Market Price Index passed the tests applied, and all showed changes between the periods under study, in addition to changes in the straight line, however, only the GDP obtained an adequate result in relation to the coefficient of determination, for the Unemployment Rate, we did not obtain satisfactory results in the residual analysis, even applying transformations to the data.
Nos anos de 2020 e 2021, o mundo enfrentou a pandemia da COVID-19, refletindo diretamente na economia dos países, essa mudança é perceptível pela análise dos indicadores econômicos. Neste trabalho, para conseguirmos identificar as alterações presentes na economia do Brasil, foram aplicados os modelos de regressão temporal, compostos por variáveis dummy e harmônicas, com períodos definidos previamente por meio da análise do periodograma. Para análise, foi coletado os valores mensais do Produto Interno Bruto (PIB), Índice Geral de Preços de Mercado (IGP-M), a Taxa de Desemprego, Índice Nacional de Preços ao Consumidor (INPC) e Índice de Preços ao Consumidor Amplo (IPCA), de janeiro de 2018 à setembro de 2021. Para o modelo foram definidas as variáveis dummy, Zit, representando se houve alteração entre o período pré-pandemia (2018-2019) e durante a pandemia (2020-2021), e (Zit)*t, representando se houve alteração na inclinação da reta entre os dois períodos. A análise residual foi composta pelo teste de Shapiro Wilk, teste Runs e teste Durbin-Watson, as métricas para adequação foram a raíz do erro quadrático (RMSE), erro médio absoluto (MAE) e o coeficiente de determinação. Os modelos para o INPC, IPCA, PIB e IGP-M, passaram pelos testes aplicados, e todos apresentaram alteração entre os períodos em estudo, além de alteração na reta, porém, apenas o PIB obteve resultado adequado em relação ao coeficiente de determinação, para a Taxa de Desemprego, não obtivemos resultados satisfatórios na análise residual, mesmo aplicando transformações nos dados.

Descrição

Palavras-chave

Temporal decomposition models, Dummy variables, Periodogram, Economic indicators, COVID-19, Modelos de decomposição temporal, Variáveis dummy, Periodograma, Indicadores econômicos

Como citar