Forecasting the Exchange Rate in the Czech Republic Using Non-linear Threshold Models
| Název práce v češtině: | Předpovídání směnného kurzu v České republice s použitím nelinárních prahových modelů |
|---|---|
| Název v anglickém jazyce: | Forecasting the Exchange Rate in the Czech Republic Using Non-linear Threshold Models |
| Klíčová slova: | předpovídání, směnný kurz, časové řady, nelinearita, SETAR, TAR |
| Klíčová slova anglicky: | forecasting, exchange rate, time series, nonlinearity, SETAR, TAR |
| Akademický rok vypsání: | 2016/2017 |
| Typ práce: | diplomová práce |
| Jazyk práce: | angličtina |
| Ústav: | Institut ekonomických studií (23-IES) |
| Vedoucí / školitel: | doc. PhDr. RNDr. Josef Stráský, Ph.D. |
| Řešitel: | skrytý - zadáno vedoucím/školitelem |
| Datum přihlášení: | 09.11.2016 |
| Datum zadání: | 09.11.2016 |
| Datum a čas obhajoby: | 08.02.2017 08:30 |
| Místo konání obhajoby: | Opletalova - Opletalova 26, O206, Opletalova - místn. č. 206 |
| Datum odevzdání elektronické podoby: | 06.01.2017 |
| Datum proběhlé obhajoby: | 08.02.2017 |
| Oponenti: | prof. Ing. Evžen Kočenda, M.A., Ph.D., DSc. |
| Kontrola URKUND: | |
| Zásady pro vypracování |
| In 1983, Meese and Rogoff show that monetary models cannot outperform a random walk in out-of-sample exchange rate forecasting. Since that time, many authors have contributed to this field with more or less controversial results.
The question whether the CZK/EUR exchange rate is predictable has been addressed in a few studies, for example Cuaresma and Hlouskova (2004), Naszodi (2011) using a promising survey forecasts method or Mućk and Skrzypczyński, (2012) using time series models. The thesis shall contribute to this topic by applying threshold models, which have not yet been extensively used for forecasting of Czech koruna exchange rate. Suitable threshold model shall be selected and estimated in order to construct pseudo out-of-sample forecasts of CZK/EUR exchange rate in forecasting horizons of 3, 6 and 12 months. Predictive ability will be evaluated using simple comparison of Mean Square Error and more advanced tests such as Theil statistics or Diebold-Mariano test. The thesis can be then extended by using several different threshold-type models or by considering other exchange rates important for Central European region. |
| Seznam odborné literatury |
| Meese, R. A., and K. Rogoff (1983): “Empirical exchange rate models of the seventies: Do they fit out of sample?,” Journal of International Economics, 14(1-2), 3–24.
Mućk, J., and Skrzypczyński, P., (2012): ”Can we beat the random walk in forecasting CEE exchange rates?”, National Bank of Poland, Working Paper No. 127 Rubaszek, M., Skrzypczyński, P., and Koloch G. (2011): “Forecasting the Polish zloty with non-linear models”, National Bank of Poland, Working Paper No. 81 Hansen, B. E. (2011): “Threshold autoregression in economics”, Statistics and Its Interface, 4, p. 123-127 Hloušková, J., and Cuaresma, J. C. (2004): „Beating the Random Walk in Central and Eastern Europe“, Journal of Forecasting, 24, 189-201 Naszodi, A. (2011): “Beating the Random Walk in Central and Eastern Europe by Survey Forecasts“, Magyar Nemzeti Bank, Working Papers 2011/3 |
| Předběžná náplň práce |
| Based on brief literature review, suitable threshold model(s) to forecast exchange rates will be selected. The forecasting periods will be 3, 6 and 12 months. For comparison of the results and predictive ability the Mean Square Error tests and more advanced tests evaluating model efficiency such as Theil statistics or Diebold-Mariano tests will be used.
The Brock–Dechert–Scheinkman test will be used to find out whether there are non-linearities in the data. The predictive ability and practical usability for exchange rate forecasting of the selected threshold models will be discussed. The research will take into account the interventions on the exchange rate market made by CNB in November 2013 to get rid of possible biases. The thesis shall address following research questions: 1. Threshold models outperform random walk in forecasting exchange rates in the case of the Czech koruna. 2. There are non-linearities in exchange rate time series with regards to the Czech koruna. 3. Threshold models are efficient and accessible tools for exchange rate forecasting for practitioners. Tentative content: 1. Introduction 2. Literature review on exchange rate forecasting and non-linear threshold modelling 3. Data and Methodology 4. Results 5. Discussion, alternative approaches, suggestions for future research 6. Conclusion |
| Předběžná náplň práce v anglickém jazyce |
| Based on brief literature review, suitable threshold model(s) to forecast exchange rates will be selected. The forecasting periods will be 3, 6 and 12 months. For comparison of the results and predictive ability the Mean Square Error tests and more advanced tests evaluating model efficiency such as Theil statistics or Diebold-Mariano tests will be used.
The Brock–Dechert–Scheinkman test will be used to find out whether there are non-linearities in the data. The predictive ability and practical usability for exchange rate forecasting of the selected threshold models will be discussed. The research will take into account the interventions on the exchange rate market made by CNB in November 2013 to get rid of possible biases. The thesis shall address following research questions: 1. Threshold models outperform random walk in forecasting exchange rates in the case of the Czech koruna. 2. There are non-linearities in exchange rate time series with regards to the Czech koruna. 3. Threshold models are efficient and accessible tools for exchange rate forecasting for practitioners. Tentative content: 1. Introduction 2. Literature review on exchange rate forecasting and non-linear threshold modelling 3. Data and Methodology 4. Results 5. Discussion, alternative approaches, suggestions for future research 6. Conclusion |