This is one of my first hands-on projects using machine learning for economic forecasting, so I’m especially keen to hear feedback and suggestions for improvement from those with more experience in this area.

#MachineLearning #EconomicForecasting #PublicPolicy #DataScience #ScottishFiscalCommission

2. Applying machine learning methods: I benchmarked traditional econometric models (like ARIMA) against modern machine learning approaches (LSTM and hybrid ARIMA-LSTM models) using simulated macroeconomic data. The results show that machine learning can significantly reduce forecasting errors.

1. Distinguishing new hire wages from average wages: New hire wages are much more responsive to economic cycles and can provide earlier signals of turning points than average wages, which are often dampened by contract rigidities.

I focused on the wage forecasting component of their model and proposed two key enhancements:

The SFC plays an important role in shaping Scotland’s fiscal policy through its independent economic forecasts. Their current methodology, while robust, could be further refined to better capture real-time labor market dynamics and improve short-term forecasting accuracy.