This project aims to forecast the quarterly growth rate of U.S. Gross Domestic Product (GDP) for the nine quarters following December 2019. The project was designed in the context of pre-pandemic economic uncertainty, with many economists anticipating a potential recession by the end of 2020.
🗞️ "According to a survey by the National Association for Business Economists, two-thirds of economists predicted a U.S. recession by the end of 2020."
🗣️ "The stimulus is going to hit the economy in a big way this year and next year, and then in 2020 Wile E. Coyote is going to go off the cliff." — Ben Bernanke, former Federal Reserve Chairman (June 2018)
However, no one foresaw the impact of a global pandemic, which drastically altered the economic landscape. This project evaluates the predictive power of economic indicators prior to the COVID-19 crisis.
To build an econometric forecasting model that:
- Predicts U.S. GDP growth from Q1 2020 through Q1 2022
- Utilises an Ordinary Least Squares (OLS) autoregressive model
- Incorporates macroeconomic indicators that are historically predictive of GDP growth (e.g., unemployment, interest rates, stock indices, industrial production, consumer sentiment)
-
Data Collection
- U.S. Real GDP growth data (quarterly)
- Macroeconomic indicators from FRED and other sources
- Time series up to Q4 2019
-
Feature Engineering
- Lagged GDP growth terms (autoregressive structure)
- Transformation and normalization of input indicators
- Rolling means and quarterly aggregates as needed
-
Model Development
- Baseline: OLS autoregressive model
- Extended: OLS with additional predictive indicators
- Model evaluated using:
- Root Mean Squared Error (RMSE)
- Mean Absolute Error (MAE)
- Out-of-sample forecasting accuracy
-
Forecasting
- Generate forecasts for 9 quarters beyond Q4 2019
- Visualise prediction intervals and trend changes
- Python 3.x
pandas,numpyfor data manipulationstatsmodelsfor OLS regression modelingmatplotlib,seabornfor data visualization- Data sourced from: FRED, BEA, OECD