Abstract
This study applies portfolio balance theory in forecasting exchange rate. The study further argues for the need to account for the role of Global Financial Cycle (GFCy). As such, the first stage of the analysis is estimate a GFCy model and obtain the idiosyncratic shock. Next, we use the results in the first stage as a predictor for exchange rate. The study builds dataset for 20 advanced and emerging countries from 1990Q1-2017Q2. Among other things, there are three important results to note. First, our approach to forecast exchange rate is able to beat the benchmark random walk model. Second, the best prediction is made at short term forecasting horizons, i.e. 1 and 4 quarters forecast ahead. Third, the performance of the early sample size outweighs that of the late sample size.
Original language | English |
---|---|
Pages (from-to) | S81-S92 |
Number of pages | 12 |
Journal | Borsa Istanbul Review |
Volume | 20 |
Early online date | 23 Jun 2020 |
DOIs | |
Publication status | Published - Dec 2020 |
Externally published | Yes |
Keywords
- Exchange rate
- Factor models
- Forecasting
- Global financial cycle