Wavelets em economia: utilizando wavelets para projetar inflação via curva de Phillips

• Ano de defesa: 2017
• Autor(a) principal: Reimermendt, Renan Renie Gevisiez
• Orientador(a): Glasman, Daniela Kubudi
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Palavras-chave em Português: Wavelets; Macroeconomia; Projeção de inflação
• Link de acesso: https://hdl.handle.net/10438/18420

Wavelet analysis allows for a much more flexible approach than spectral analysis, being highly utilized in many fields, it is a natural evolution from the Fourier analysis. Unifying both time domain and frequency domain in its analysis, it gives the researcher the ability to observe relations that were previously inaccessible and its flexibility makes it recommended to analyze series with structural changes. This work contains a chronological and theoretical introduction of the technique, focusing on what will be used, presenting some successful applications in economics. Lastly, it is created a measure of economic activity by denoising, in both global and specific scales, the seasonally adjusted GDP series and utilizing this measure in a Phillips curve, as described by the Brazilian Central Bank in theirs semi-structural aggregate small sized models, to forecast future inflation. This forecast is then compared to forecasts using the traditional HP filter and a measure of output gap elaborated by Areosa (2008), which incorporates some economic structure in the output gap.