Exchange Rate Time Series Research Articles

Study on the HMM of Nonlinear Nominal Exchange Rates

The exchange rate model for the study of the exchange rate theory has very important significance. After analyzing the successful nonlinear model of real exchange rate based on the purchasing power parity (PPP) theory, the nonlinear problem of nominal exchange rate is studied in this paper. Through a research on a period of nominal exchange rate with nonlinear characteristics, a nonlinear statistical model of nominal exchange rate based on the hidden Markov model (HMM) is proposed, and the parameters of the model are estimated. Hypothesis testing shows that the model can accurately describe the statistical characteristics of the nominal exchange rate time series. The parameters showed that the nominal exchange rate model proposed in this paper, to some extent, supports that deviations from purchasing power parity (PPP) are nonlinear mean reversions.

Cross-sample entropy of foreign exchange time series

The correlation of foreign exchange rates in currency markets is investigated based on the empirical data of DKK/USD, NOK/USD, CAD/USD, JPY/USD, KRW/USD, SGD/USD, THB/USD and TWD/USD for a period from 1995 to 2002. Cross-SampEn (cross-sample entropy) method is used to compare the returns of every two exchange rate time series to assess their degree of asynchrony. The calculation method of confidence interval of SampEn is extended and applied to cross-SampEn. The cross-SampEn and its confidence interval for every two of the exchange rate time series in periods 1995–1998 (before the Asian currency crisis) and 1999–2002 (after the Asian currency crisis) are calculated. The results show that the cross-SampEn of every two of these exchange rates becomes higher after the Asian currency crisis, indicating a higher asynchrony between the exchange rates. Especially for Singapore, Thailand and Taiwan, the cross-SampEn values after the Asian currency crisis are significantly higher than those before the Asian currency crisis. Comparison with the correlation coefficient shows that cross-SampEn is superior to describe the correlation between time series.

Exchange rate forecasting: comparison of various architectures of neural networks

This paper evaluates the predictive accuracy of neural networks in forecasting exchange rate. The multi-layer perceptron (MLP) and radial basis function (RBF) networks with different architectures are used to forecast five exchange rate time series. The results of each prediction are evaluated and compared according to the networks and architectures used. It is found that neural networks can be effectively used in forecasting exchange rate and hence in designing trading strategies. RBF networks performed better than MLP networks in our simulation experiment. This experiment suggests that it is possible to extract information hidden in the exchange rate and predict it into future.

ESTIMATING BIVARIATE GARCH-JUMP MODEL BASED ON HIGH FREQUENCY DATA: THE CASE OF REVALUATION OF THE CHINESE YUAN IN JULY 2005

This paper analyzes the behavior of one-minute high-frequency time-series data of exchange rates for five currencies (Japanese Yen, Australian Dollar, Canadian Dollar, Euro, and Pound Sterling) against the US Dollar when the Chinese Yuan was revalued on July 21st, 2005. The data show the following distinctive features: (1) There is a large jump in the exchange rates time series at the time of the Yuan revaluation. (2) Large volatility in the returns of exchange rates is observed for a while after the jump. (3) There are many other jumps, possibly correlated, in each exchange rate time series. To capture these features we fit the following models to the data: (i) a univariate GARCH-Jump model with a large jump that is influential on volatility, and (ii) a bivariate GARCH-Jump model with correlated Poisson jumps. For comparison, we also estimate these GARCH models without the associated jumps. The model performance is evaluated based on Value-at-Risk (VaR).

Energy dependence, oil prices and exchange rates: the Dominican economy since 1990

This paper studies the impact that oil prices have had on the floating exchange rate (ER) of the Dominican peso during the 1990–2008 period. The existing empirical literature has documented a link between these two variables for large developed economies and oil-producing countries, always including the 1970s oil crises in their sample periods. Few authors analyze the case of small open economies dependent on oil imports. Since the acceleration of economic growth in the Dominican Republic during the last decade has greatly increased its degree of external energy dependence, this country’s experience presents an ideal case study. We estimate the influence that changes in international gasoline prices have on the real exchange rate between the Dominican peso and the United States dollar. The cointegrated nature of the gasoline price and ER time series are tested and a vector error correction model is developed. Our results indicate that a 10% rise in the price of gas coincides with a 1.2% depreciation of the peso in the long run and that the causality runs from gas prices to the peso.

The dynamics of exchange rate time series and the chaos game

This work presents a novel method of reconstructing some relevant characteristics of exchange rate time series by the superposition of two components: a mostly deterministic one, the chaos game as expressed by the Yuan/USD exchange rate and a purely stochastic one, Gaussian white noise. We analyzed 20 economic systems with the average Index of Economic Freedom above 50. The considered characteristics (the Lempel–Ziv complexity index, the slimness of the distribution and the Iterated Function Systems clumpiness test) are well reproduced by the reconstruction process. Additional confirmation is obtained by an analysis of the exchange rate of the Romanian national currency as an example of an application of the method to a transition economy, and by an analysis of the time series of the Euro-zone as an example of an application to a multinational system using a shorter time series.

Exchange rate prediction using hybrid neural networks and trading indicators

This paper describes a hybrid model formed by a mixture of various regressive neural network models, such as temporal self-organising maps and support vector regressions, for modelling and prediction of foreign exchange rate time series. A selected set of influential trading indicators, including the moving average convergence/divergence and relative strength index, are also utilised in the proposed method. A genetic algorithm is applied to fuse all the information from the mixture regression models and the economical indicators. Experimental results and comparisons show that the proposed method outperforms the global modelling techniques such as generalised autoregressive conditional heteroscedasticity in terms of profit returns. A virtual trading system is built to examine the performance of the methods under study.

Exchange Rate Assessments: Methodologies for Oil Exporting Countries

Are the current account fluctuations in oil-exporting countries excessive? How should their real exchange rate respond to the evolution of external (and domestic) fundamentals? This paper proposes methodologies tailored to the specific features of oil-exporting countries that help address these questions. Price-based methodologies (based on the time series of real effective exchange rates) identify a strong link between the real exchange rate and the terms of trade, but have relatively limited explanatory power. On the other hand, an empirical model of the current account, which fits oil exporting countries' data well, and an intertemporal model that takes into account the stock of oil reserves provide useful benchmarks for oil exporters' external balances.

Exchange Rate Assessments: Methodologies for Oil Exporting Countries

Are the current account fluctuations in oil-exporting countries excessive? How should their real exchange rate respond to the evolution of external (and domestic) fundamentals? This paper proposes methodologies tailored to the specific features of oil-exporting countries that help address these questions. Price-based methodologies (based on the time series of real effective exchange rates) identify a strong link between the real exchange rate and the terms of trade, but have relatively limited explanatory power. On the other hand, an empirical model of the current account, which fits oil exporting countries' data well, and an intertemporal model that takes into account the stock of oil reserves provide useful benchmarks for oil exporters' external balances.

Thick Breaks and Trend Stationarity: The Case of Euro-Dollar Exchange Rate

The taking into account of a period of break (thick break) makes it possible to correctly analyze the time series of the euro-dollar exchange rate. By retaining the posterior period with the Louvre agreements, but by eliminating the first years from existence of the euro, and until today, one can affirm that this rate is stationary and after trend stationary and thus that there is a mechanism of return towards a level (a trend) of equilibrium. This point is shown using a new procedure of test based on the elimination of thick breaks. That makes it possible to propose a forecast based on this deterministic trend.

Minimal Spanning Tree Graphs and Power Like Scaling in FOREX Networks

Correlation matrices of foreign exchange rate time series are investigated for 60 world currencies. Minimal Spanning Tree (MST) graphs for the gold, silver and platinum are presented. Inverse power like scaling is discussed for these graphs as well as for four distinct currency groups (major, liquid, less liquid and non-tradable). The worst scaling has been found for USD and related currencies.

Dynamic Ridge Polynomial Neural Networks for multi-step financial time-series prediction

Concerned with the slow learning problems of multilayer perceptrons, which utilise computationally intensive training algorithms like the backpropagation learning algorithm that can get trapped in local minima, this paper presents a novel type of higher-order recurrent neural networks, which is called the Dynamic Ridge Polynomial Neural Networks (DRPNN). The aim of the proposed network is to improve the performance of the Ridge Polynomial Neural Network by accommodating recurrent links structure. The network is tested for the prediction of non-linear and non-stationary financial signals. Three exchange rate time-series, which are the exchange rate time series between the British Pound and the Euro, the US dollar and the Euro as well as the exchange rate between the Japanese Yen and the Euro, are used in the simulation process. Simulation results showed that DRPNN generate higher profit returns with fast convergence when used to predict noisy financial time series.

An Empirical Analysis of Exchange Rates Cointergrating System

This paper analyzes whether the system of exchange rate time series of the Japanese yen and the South Korean won is cointegrated by using Engle and Granger's two-step approach to modeling cointegrated processes. The first step involves fitting the long-run relationship in levels by least squares. At a second step, the residuals from the static regression are employed as an error correction term in the dynamic, first-difference regression. Because the hypothesis of cointegration cannot be rejected, the estimated parameters can be considered as elements of the cointegrating vector of this system. The conclusion of this paper can provide some insights to investors who are interested in foreign exchange markets.

The Lévy sections theorem: An application to econophysics

We employ the Lévy sections theorem in the analysis of selected dollar exchange rate time series. The theorem is an extension of the classical central limit theorem and offers an alternative to the most usual analysis of the sum variable. We find that the presence of fat tails can be related to the local volatility pattern of the series.

An objective function for simulation based inference on exchange rate data

The assessment of models of financial market behaviour requires evaluation tools. When complexity hinders a direct estimation approach, e.g., for agent based microsimulation models, simulation based estimators might provide an alternative. In order to apply such techniques, an objective function is required, which should be based on robust statistics of the time series under consideration. Based on the identification of robust statistics of foreign exchange rate time series in previous research, an objective function is derived. This function takes into account stylized facts about the unconditional distribution of exchange rate returns and properties of the conditional distribution, in particular, autoregressive conditional heteroscedasticity and long memory. A bootstrap procedure is used to obtain an estimate of the variance-covariance matrix of the different moments included in the objective function, which is used as a base for the weighting matrix. Finally, the properties of the objective function are analyzed for two different agent based models of the foreign exchange market, a simple GARCH-model and a stochastic volatility model using the DM/US-$ exchange rate as a benchmark. It is also discussed how the results might be used for inference purposes.

Are pound and euro the same currency?

Abstract Based on long-range dependence, some analysts claim that the exchange rate time series of the pound sterling and of an artificially extended euro have been locked together for years despite daily changes [M. Ausloos, K. Ivanova, Physica A 286 (2000) 353; K. Ivanova, M. Ausloos, False EUR exchange rates vs DKK, CHF, JPY and USD. What is a strong currency? in: H. Takayasu (Ed.), Empirical Sciences in Financial Fluctuations: The Advent of Econophysics, Springer-Verlag, Berlin, 2002, pp. 62–76]. They conclude that pound and euro are in practice the same currency. We assess the long-range dependence over time through Hurst exponents of pound–dollar and extended euro–dollar exchange rates employing three alternative techniques, namely rescaled range analysis, detrended fluctuation analysis, and detrended moving average. We find the result above (which is based on detrended fluctuation analysis) not to be robust to the changing techniques and parameterizing.

An Objective Function for Simulation Based Inference on Exchange Rate Data

The assessment of models of financial market behavior requires evaluation tools. When complexity hinders a direct estimation approach, e.g., for agent based microsimulation models or complex multifractal models, simulation based estimators might provide an alternative. In order to apply such techniques, an objective function is required, which should be based on robust statistics of the time series under consideration. Based on the identification of robust statistics of foreign exchange rate time series in previous research, an objective function is derived. This function takes into account stylized facts about the unconditional distribution of exchange rate returns and properties of the conditional distribution, in particular, autoregressive conditional heteroscedasticity and long memory. A bootstrap procedure is used to obtain an estimate of the variance-covariancematrix of the different moments included in the objective function, which is used as a base for the weighting matrix. Finally, the properties of the objective function are analyzed for two different agent based models of the foreign exchange market, a simple GARCH-model and a stochastic volatility model using the DM/US-$ exchange rate as a benchmark. It is also discussed how the results might be used for inference purposes.

COMPUTATIONAL INTELLIGENCE METHODS FOR FINANCIAL TIME SERIES MODELING

In this paper, the combination of unsupervised clustering algorithms with feedforward neural networks in exchange rate time series forecasting is studied. Unsupervised clustering algorithms have the desirable property of deciding on the number of partitions required to accurately segment the input space during the clustering process, thus relieving the user from making this ad hoc choice. Combining this input space partitioning methodology with feedforward neural networks acting as local predictors for each identified cluster helps alleviate the problem of nonstationarity frequently encountered in real-life applications. An improvement in the one-step-ahead forecasting accuracy was achieved compared to a global feedforward neural network model for the time series of the exchange rate of the German Mark to the US Dollar.

Statistical fuzzy interval neural networks for currency exchange rate time series prediction

In this paper, the statistical fuzzy interval neural network with statistical interval input and output values is proposed to perform statistical fuzzy knowledge discovery and the currency exchange rate prediction. Time series data sets are grouped into time series data granules with statistical intervals. The statistical interval data sets including week-based averages, maximum errors of estimate and standard deviations are used to train the fuzzy interval neural network to discover fuzzy IF-THEN rules. The output of the fuzzy interval neural network is an interval value with certain percent confidence. Simulations are completed in terms of the exchange rates between US Dollar and other three currencies (Japanese Yen, British Pound and Hong Kong Dollar). The simulation results show that the fuzzy interval neural network can provide more tolerant prediction results.

Average conditional correlation and tree structures for multivariate GARCH models

We propose a simple class of multivariate GARCH models, allowing for time-varying conditional correlations. Estimates for time-varying conditional correlations are constructed by means of a convex combination of averaged correlations (across all series) and dynamic realized (historical) correlations. Our model is very parsimonious. Estimation is computationally feasible in very large dimensions without resorting to any variance reduction technique. We back-test the models on a six-dimensional exchange-rate time series using different goodness-of-fit criteria and statistical tests. We collect empirical evidence of their strong predictive power, also in comparison to alternative benchmark procedures. Copyright © 2006 John Wiley & Sons, Ltd.