Foreign Exchange Rate Forecasting Research Articles

Out-of-sample forecasting of foreign exchange rates: The band spectral regression and LASSO

We propose to utilize the band spectral regression for out-of-sample forecasts of exchange rates. When one period ahead forecast is considered, there is some evidence that the band spectral regression improves its accuracy, especially when the Taylor rule fundamentals model is employed. However, when the forecasting horizon increases, the purchasing power parity (PPP) fundamentals model is found to be powerful, and we can improve the out-of-sample forecast by selecting appropriate frequency bands. Bayesian model averaging shows that placing a large weight on the business cycle frequency improves the accuracy of the out-of-sample forecasting of the PPP model (as well as the monetary fundamentals model) when a longer forecasting horizon is our focus. Without specifying the frequency bands prior to applying the regression, LASSO can provide better out-of-sample exchange rate forecasts for many cases – most patently for the PPP fundamentals model – and provide information about the dynamic relationship between forecasting variables and exchange rates. The frequency domain approach not only improves the accuracy of exchange rate forecast but provides insights for understanding why the PPP fundamentals act as a powerful predictor when the forecasting horizon increases and there is a possible improvement in the time domain regression forecast.

Time series modelling, NARX neural network and hybrid KPCA–SVR approach to forecast the foreign exchange market in Mauritius

PurposeThis study constructs time series model, artificial neural networks (ANNs) and statistical topologies to examine the volatility and forecast foreign exchange rates. The Mauritian forex market has been utilized as a case study, and daily data for nominal spot rate (during a time period of five years spanning from 2014 to 2018) for EUR/MUR, GBP/MUR, CAD/MUR and AUD/MUR have been applied for the predictions.Design/methodology/approachAutoregressive integrated moving average (ARIMA) and generalized autoregressive conditional heteroskedasticity (GARCH) models are used as a basis for time series modelling for the analysis, along with the non-linear autoregressive network with exogenous inputs (NARX) neural network backpropagation algorithm utilizing different training functions, namely, Levenberg–Marquardt (LM), Bayesian regularization and scaled conjugate gradient (SCG) algorithms. The study also features a hybrid kernel principal component analysis (KPCA) using the support vector regression (SVR) algorithm as an additional statistical tool to conduct financial market forecasting modelling. Mean squared error (MSE) and root mean square error (RMSE) are employed as indicators for the performance of the models.FindingsThe results demonstrated that the GARCH model performed better in terms of volatility clustering and prediction compared to the ARIMA model. On the other hand, the NARX model indicated that LM and Bayesian regularization training algorithms are the most appropriate method of forecasting the different currency exchange rates as the MSE and RMSE seemed to be the lowest error compared to the other training functions. Meanwhile, the results reported that NARX and KPCA–SVR topologies outperformed the linear time series models due to the theory based on the structural risk minimization principle. Finally, the comparison between the NARX model and KPCA–SVR illustrated that the NARX model outperformed the statistical prediction model. Overall, the study deduced that the NARX topology achieves better prediction performance results compared to time series and statistical parameters.Research limitations/implicationsThe foreign exchange market is considered to be instable owing to uncertainties in the economic environment of any country and thus, accurate forecasting of foreign exchange rates is crucial for any foreign exchange activity. The study has an important economic implication as it will help researchers, investors, traders, speculators and financial analysts, users of financial news in banking and financial institutions, money changers, non-banking financial companies and stock exchange institutions in Mauritius to take investment decisions in terms of international portfolios. Moreover, currency rates instability might raise transaction costs and diminish the returns in terms of international trade. Exchange rate volatility raises the need to implement a highly organized risk management measures so as to disclose future trend and movement of the foreign currencies which could act as an essential guidance for foreign exchange participants. By this way, they will be more alert before conducting any forex transactions including hedging, asset pricing or any speculation activity, take corrective actions, thus preventing them from making any potential losses in the future and gain more profit.Originality/valueThis is one of the first studies applying artificial intelligence (AI) while making use of time series modelling, the NARX neural network backpropagation algorithm and hybrid KPCA–SVR to predict forex using multiple currencies in the foreign exchange market in Mauritius.

Data decomposition based fast reduced kernel extreme learning machine for currency exchange rate forecasting and trend analysis

In this paper, we propose a hybrid forecasting model that combines Empirical Mode Decomposition (EMD) with fast reduced kernel Extreme Learning Machine (KELM) for day ahead foreign currency exchange rate forecasting. EMD is an efficient method for nonlinear data decomposition in such a noisy environment and the purpose is to find important components in terms of Intrinsic Mode Functions (IMFs) by which the nonlinear time series is converted into stationary time series by making the data smoother and simpler for analysis. The average IMFs decomposed from EMD (AEMD) are hybridized with fast KELM named as AEMD-KELM for producing a more accurate forecast. The experimental results using AEMD-KELM method for seven currency exchange rates like CAD/HKD, CAD/CNY, CAD/USD, CAD/BRL, CAD/JPY, EUR/USD, and GBP/USD provide superior prediction and trend analysis in comparison with EMD based ELM (EMD-ELM) approaches. Further currency exchange rate movement trends are used for generating trading signals like buy, sell or hold.

The information content of retail investors’ order flow

In this paper, we provide evidence that the trading activity of small retail investors carries significant genuine information that can be exploited for the short-term out-of-sample forecasting of foreign exchange rates. Our findings are based on a unique dataset of around 2000 retail investors from the OANDA FXTrade electronic trading platform. Our results are consistent with the view that in the foreign exchange market private information is highly dispersed, but can be extracted by observing customer order flow. Previous studies, however, focused on the information content of costumer order flow of dealers in the interbank market, whose clients are themselves large institutional and professional investors. Our study is the first that analyzes a crowd of small retail investors and shows that even the trading activity of these investors contains, on aggregate, important non-public information that can be exploited for short-term exchange rate forecasting. Our findings lead us to conjecture that retail investors (on aggregate) are not pure noise traders but process dispersed information at least partially in a similar way as large institutional investors and hence place their orders accordingly.

The Information Content of Retail Investors' Order Flow

In this paper we provide evidence that the trading activity of small retail investors carries significant genuine information for the short term out-of-sample forecasting of foreign exchange rates. Our findings are based on a unique dataset of around 2000 retail investors from the OANDA FXTrade electronic trading platform. Our results are consistent with the view that in the foreign exchange market private information is highly dispersed, but can be extracted by observing customer order flow. Previous studies, however, focused on the information content of costumer order flow of dealers in the interbank market, whose clients are themselves large institutional and professional investors. Our study is the first that analyses a crowd of small retail investors and shows that even the trading activity of these investors contains, on aggregate, important non-public information, that can be exploited for short term exchange rate forecasting. Our findings lead us to conjecture that retail investors (on aggregate) are not pure noise traders but process dispersed information at least partially in a similar way as large institutional investors and hence place their orders accordingly.

Foreign exchange rate forecasting with a virtual-expert partial-consensus fuzzy-neural approach for semiconductor manufacturers in Taiwan

Accurately forecasting the foreign exchange rate is very important to export-oriented enterprises, like semiconductor manufacturers in Taiwan. To this end, a virtual-expert partial-consensus fuzzy-neural approach is proposed in this study. In the proposed methodology, instead of calling a number of experts in the field, a committee of virtual experts is formed, and then they are asked to provide views on fuzzy forecasts. For each virtual expert, the corresponding fuzzy linear regression (FLR) equation is constructed to predict the foreign exchange rate. Each FLR equation can be fitted by solving two equivalent non-linear programming problems, based on the virtual experts’ views. To aggregate these fuzzy foreign exchange rate forecasts, a two-step aggregation mechanism is applied. To evaluate the effectiveness of the proposed methodology, the real case of forecasting the foreign exchange rate of NTD for USD is used.

Empirical mode decomposition–based least squares support vector regression for foreign exchange rate forecasting

To address the nonlinear and non-stationary characteristics of financial time series such as foreign exchange rates, this study proposes a hybrid forecasting model using empirical mode decomposition (EMD) and least squares support vector regression (LSSVR) for foreign exchange rate forecasting. EMD is used to decompose the dynamics of foreign exchange rate into several intrinsic mode function (IMF) components and one residual component. LSSVR is constructed to forecast these IMFs and residual value individually, and then all these forecasted values are aggregated to produce the final forecasted value for foreign exchange rates. Empirical results show that the proposed EMD-LSSVR model outperforms the EMD-ARIMA (autoregressive integrated moving average) as well as the LSSVR and ARIMA models without time series decomposition.

Integrating Application of Independent Component Analysis and Support Vector Machine in Foreign Exchange Rate Forecasting

Integrating Application of Independent Component Analysis and Support Vector Machine in Foreign Exchange Rate Forecasting

Applying a Fuzzy and Neural Approach for Forecasting the Foreign Exchange Rate

Accurately forecasting the foreign exchange rate is important for export-oriented enterprises. For this purpose, a fuzzy and neural approach is applied in this study. In the fuzzy and neural approach, multiple experts construct fuzzy linear regression (FLR) equations from various viewpoints to forecast the foreign exchange rate. Each FLR equation can be converted into two equivalent nonlinear programming problems to be solved. To aggregate these fuzzy foreign exchange rate forecasts, a two-step aggregation mechanism is applied. At the first step, fuzzy intersection is applied to aggregate the fuzzy forecasts into a polygon-shaped fuzzy number to improve the precision. A back propagation network is then constructed to defuzzify the polygon-shaped fuzzy number and generate a representative/crisp value to enhance accuracy. To evaluate the effectiveness of the fuzzy and neural approach, a practical case of forecasting the foreign exchange rate in Taiwan is used. According to the experimental results, the fuzzy and neural approach improved both the precision and accuracy of the foreign exchange rate forecasting by 79% and 81%, respectively.

Modeling exchange rates using wavelet decomposed genetic neural networks

Exchange rate forecasting has always been a challenging area of research to applied statisticians and econometricians. In this paper, we propose to use a methodology that combines artificial intelligence modeling techniques with wavelet multiresolution methodology for forecasting of daily spot foreign exchange rates of major internationally traded currencies. The original exchange rate series to be modeled is first decomposed into various frequency resolution related components using wavelet-filtering techniques. The decomposed components represent various components of the original series, like the long-term trend and the periodic components at different periodicities. Artificial intelligence modeling, using genetically optimized neural networks, is applied for modeling components of the decomposed series. The final forecast of the original series is obtained by combining the component series forecasts. In the empirical studies, we apply the proposed modeling technique for forecasting one- and multi-step ahead forecasts of spot exchange rates. Results of the empirical study indicate superior performance of the proposed technique as compared to the traditional exchange rate forecasting models.

Professional vs. amateur judgment accuracy: The case of foreign exchange rates

Highly knowledgeable people often fail to achieve highly accurate judgments, a phenomenon sometimes called the “process-performance paradox.” The present research tested for this paradox in foreign exchange (FX) rate forecasting. Forty professional and 57 sophisticated amateur forecasters made one-day and one-week-ahead FX predictions in deterministic and probabilistic formats. Among the conclusions indicated by the results are: (a) professional accuracy usually surpasses amateur accuracy, although many amateurs outperform many professionals; (b) professionals appear to achieve high proficiency via heavy reliance on predictive information (unlike what has been observed before, e.g., for stock prices); (c) forecast format strongly affects judgment accuracy and processes; and (d) apparent overconfidence can transform itself into underconfidence depending on when and how forecasters must articulate their confidence.

Forecasting foreign exchange rates with an intrinsically nonlinear dynamic speed of adjustment model

Forecasting foreign exchange rates is an important but difficult process; therefore, it is important to use a superior forecasting model. The paper takes up this criterion and proposes to describe and forecast foreign exchange rates by developing an intrinsically nonlinear model with variable and dynamic speeds of adjustment. It is found that the speed of adjusting the random (or expected) to the equilibrium rate is very slow, implying that fiscal policy (statistically insignificat) and monetary policy (statistically significant) may be ineffective to induce changes in the adjustment speed. We also find that the nonlinear dynamic model improves forecasting performance, implying that nonlinearities in the sense of functional forms are exploitable for improved point forecasting of foreign exchange rates.

Combining foreign exchange rate forecasts using neural networks

Combining foreign exchange rate forecasts using neural networks

An assessment of the economic value of non‐linear foreign exchange rate forecasts

AbstractRecent research suggests that non‐linear methods cannot improve the point forecasts of high‐frequency exchange rates. These studies have been using standard forecasting criteria such as smallest mean squared error (MSE) and smallest mean absolute error (MAE). It is, however, premature to conclude from this evidence that non‐linear forecasts of high‐frequency financial returns are economically or statistically insignificant. We prove a proposition which implies that the standard forecasting criteria are not necessarily particularly suited for assessment of the economic value of predictions of non‐linear processes where the predicted value and the prediction error may not be independently distributed. Adopting a simple non‐linear forecasting procedure to 15 daily exchange rate series we find that although, when compared to simple random walk forecasts, all the non‐linear forecasts give a higher MSE and MAE, when applied in a simple trading strategy these forecasts result in a higher mean return. It is also shown that the ranking of portfolio payoffs based on forecasts from a random walk, and linear and non‐linear models, is closely related to a non‐parametric test of market timing.

The performance of UK exchange rate forecasters: David Blake, Michael Beenstock and Valerie Brasse, Economic Journal 96 (1986) 986–999

There have been many tests of the hypothesis that the forward exchange rate is an unbiased predictor of the future spot exchange rate, see e.g. Hansen and Hodrick (I980), Cornell (I977), Levich (I979), Bilson (I98I), Baillie et al. (I983), and Bailey et al. (I984). Their methodology has been largely, but not exclusively, confined to weak form tests of the joint hypotheses of unbiasedness and efficiency in the foreign-exchange markets. In addition there have been limited attempts to test the accuracy and unbiasedness ofvarious exchange rate forecasting services, e.g. Levich (1 980), Goodman (I 979) and Bilik (i 982). The two issues are clearly related, for if the foreign exchange market is efficient and responds promptly and correctly to new, price-sensitive information, then no trading rule, including one based on foreign exchange rate forecasts, can be profitable. That these services have proliferated in recent years would tend to corroborate the findings of Hansen and Hodrick (I980), Geweke and Feige (I979) and Bilson (I98I) that the forward rate has been largely discredited for most major currencies as the best unbiased predictor of future spot rates. Yet Levich (I979), whose sample consisted largely of US-based exchangerate forecasting services, found that the forward rate tends to outperform the forecast services as far as quantitative accuracy is concerned, although the opposite applies for some services with respect to qualitative (i.e. directional) accuracy. Likewise the main conclusion of Goodman's North American-based study was that forecasts derived from 'technical analysis' outperform the forward rate in qualitative tests while judgmental and econometric forecasts are outperformed by the forward rate. These findings are supported by Bilik. In this paper, we carry out quantitative tests of the accuracy of a number of UK-based exchange-rate forecasts. These forecasts are in the main judgemental and econometric but in practice they are no doubt influenced by recent exchange rate movements. Our sample does not contain, therefore, any forecasts derived from technical analysis. The choice of UK forecasters is intentional. In Section I, we discuss our methodology. Results are reported in Section II and our conclusions are drawn in Section III.

The Performance of UK Exchange Rate Forecasters

There have been many tests of the hypothesis that the forward exchange rate is an unbiased predictor of the future spot exchange rate, see e.g. Hansen and Hodrick (I980), Cornell (I977), Levich (I979), Bilson (I98I), Baillie et al. (I983), and Bailey et al. (I984). Their methodology has been largely, but not exclusively, confined to weak form tests of the joint hypotheses of unbiasedness and efficiency in the foreign-exchange markets. In addition there have been limited attempts to test the accuracy and unbiasedness ofvarious exchange rate forecasting services, e.g. Levich (1 980), Goodman (I 979) and Bilik (i 982). The two issues are clearly related, for if the foreign exchange market is efficient and responds promptly and correctly to new, price-sensitive information, then no trading rule, including one based on foreign exchange rate forecasts, can be profitable. That these services have proliferated in recent years would tend to corroborate the findings of Hansen and Hodrick (I980), Geweke and Feige (I979) and Bilson (I98I) that the forward rate has been largely discredited for most major currencies as the best unbiased predictor of future spot rates. Yet Levich (I979), whose sample consisted largely of US-based exchangerate forecasting services, found that the forward rate tends to outperform the forecast services as far as quantitative accuracy is concerned, although the opposite applies for some services with respect to qualitative (i.e. directional) accuracy. Likewise the main conclusion of Goodman's North American-based study was that forecasts derived from 'technical analysis' outperform the forward rate in qualitative tests while judgmental and econometric forecasts are outperformed by the forward rate. These findings are supported by Bilik. In this paper, we carry out quantitative tests of the accuracy of a number of UK-based exchange-rate forecasts. These forecasts are in the main judgemental and econometric but in practice they are no doubt influenced by recent exchange rate movements. Our sample does not contain, therefore, any forecasts derived from technical analysis. The choice of UK forecasters is intentional. In Section I, we discuss our methodology. Results are reported in Section II and our conclusions are drawn in Section III.

Stability of predictors and the forecasting of foreign exchange rates

The forecasting of foreign exchange rates is like reading tea leaves to some and a sophisticated form of economic analysis to others. Several firms currently earn substantial profits selling exchange-rate forecasts based on various models and with different track records. This study critically reviews the pitfalls and the merits of existing approaches to forecasting foreign exchange rates and introduces a new element in the forecasting model: the stability of predictors. The findings are that stability variables do improve the forecasting ability of an econometrically based model. The results of the suggested model are superior to those derived using the forward rate and the current spot rate as forecasting tools for future spot rates. The forecasting model developed here is shown to be transferable from one currency to another. The utilization of complex statistical techniques in the forecasting of foreign exchange rates depends on whether or not the foreign exchange markets are efficient. The foreign exchange market that is efficient in any sense (weak, semistrong, and strong) is signaling foreign exchange managers that regardless of the forecasting technique employed-sophisticated econometric models, simple forecasting tools, or chart reading-the market cannot be beaten. A simple, costless forecasting tool continuously observable in the marketplace, such as the forward rate, should prove to be a sufficient predictor of the future spot rate if markets are functioning efficiently. Empirical tests on the efficiency of the foreign exchange markets suffer the following drawbacks: 1.

Effective Scoring Rules for Probabilistic Forecasts

This paper studies the use of a scoring rule for the elicitation of forecasts in the form of probability distributions and for the subsequent evaluation of such forecasts. Given a metric (distance function) on a space of probability distributions, a scoring rule is said to be effective if the forecaster's expected score is a strictly decreasing function of the distance between the elicited and “true” distributions. Two simple, well-known rules (the spherical and the quadratic) are shown to be effective with respect to suitable metrics. Examples and a practical application (in Foreign Exchange rate forecasting) are also provided.

An econometric model for foreign exchange rate forecasting

(1978). An econometric model for foreign exchange rate forecasting. Investment Analysts Journal: Vol. 7, No. 12, pp. 32-36.