Dating and Forecasting the G7 Business Cycle

The two recent global recessions triggered by the Global Financial Crisis in 2007 and the pandemic in 2020 have put business cycle analysis at the forefront of economic research. An important aspect relates to the identification of turning points. Following the methodology proposed by Stock and Watson (2010) to date turning points in the United States, this thesis uses a disaggregated dataset of economic indicators for the G7 to identify turning points in the global business cycle. A machine learning algorithm XGBoost is used to evaluate the new chronology and compares it to OECD reference chronology. Moreover, the algorithm selects the best indicators of global recessions.