Modelos de frecuencias mixtas :Comparación de alternativas y predicción de variables macroeconómicas

  1. Nicolás Bonino-Gayoso
Supervised by:
  1. Alfredo Garcia-Hiernaux Director

Defence university: Universidad Complutense de Madrid

Fecha de defensa: 01 April 2022

  1. Miguel Jerez Méndez Chair
  2. Sonia Sotoca López Secretary
  3. Elizabeth Bucacos Iguini Committee member
  4. Javier José Peréz García Committee member
  5. Rafael Flores de Frutos Committee member

Type: Thesis


Economic policy-makers, entrepreneurs and investors, among other agents, need to have access to real-time assessments of the state of the economy, along with nowcasts and forecasts of its expected evolution. The sooner they have access to information about the real economic situation, the better prepared they will be to make decisions to update their initial plans. This becomes even more important in non-stable economic and challenging environments like the one generated by COVID-19 outbreak. Unfortunately, data o_ered by the System of National Accounts (SNA) is delivered with considerable delay. On the other hand, there is a signi_cant number of short-term economic indicators available at a much earlier stage that could be used to extract information about the state of the economy. Thus we have access to monthly data from consumer surveys or the industrial production index, daily data from _nancial markets and even more frequently observed variables, such as Google or Twitter trends and mobile phone data. Although not as complete as SNA data, these indicators could anticípate relevant information...