Why do regime‐switching models forecast so badly?

Most non‐linear techniques give good in‐sample fits to exchange rate data but are usually outperformed by random walks or random walks with drift when used for out‐of‐sample forecasting. In the case of regime‐switching models it is possible to understand why forecasts based on the true model can have higher mean squared error than those of a random walk or random walk with drift. In this paper we provide some analytical results for the case of a simple switching model, the segmented trend model. It requires only a small misclassification, when forecasting which regime the world will be in, to lose any advantage from knowing the correct model specification. To illustrate this we discuss some results for the DM/dollar exchange rate. We conjecture that the forecasting result is more general and describes limitations to the use of switching models for forecasting. This result has two implications. First, it questions the leading role of the random walk hypothesis for the spot exchange rate. Second, it suggests that the mean square error is not an appropriate way to evaluate forecast performance for non‐linear models. Copyright © 1999 John Wiley & Sons, Ltd.     

Linear and Non-linear (Non-)Forecastability of High-frequency Exchange Rates

This paper forecasts Daily Sterling exchange rate returns using various naive, linear and non-linear univariate time-series models. The accuracy of the forecasts is evaluated using mean squared error and sign prediction criteria. These show only a very modest improvement over forecasts generated by a random walk model. The Pesaran–Timmerman test and a comparison with forecasts generated artificially shows that even the best models have no evidence of market timing ability.©1997 John Wiley & Sons, Ltd.     

Non‐linear,non‐parametric,non‐fundamental exchange rate forecasting

This paper employs a non‐parametric method to forecast high‐frequency Canadian/US dollar exchange rate. The introduction of a microstructure variable, order flow, substantially improves the predictive power of both linear and non‐linear models. The non‐linear models outperform random walk and linear models based on a number of recursive out‐of‐sample forecasts. Two main criteria that are applied to evaluate model performance are root mean squared error (RMSE) and the ability to predict the direction of exchange rate moves. The artificial neural network (ANN) model is consistently better in RMSE to random walk and linear models for the various out‐of‐sample set sizes. Moreover, ANN performs better than other models in terms of percentage of correctly predicted exchange rate changes. The empirical results suggest that optimal ANN architecture is superior to random walk and any linear competing model for high‐frequency exchange rate forecasting. Copyright © 2006 John Wiley & Sons, Ltd.     

Local to unity,long‐horizon forecasting thresholds for model selection in the AR(1)

This article introduces a novel framework for analysing long‐horizon forecasting of the near non‐stationary AR(1) model. Using the local to unity specification of the autoregressive parameter, I derive the asymptotic distributions of long‐horizon forecast errors both for the unrestricted AR(1), estimated using an ordinary least squares (OLS) regression, and for the random walk (RW). I then identify functions, relating local to unity ‘drift’ to forecast horizon, such that OLS and RW forecasts share the same expected square error. OLS forecasts are preferred on one side of these ‘forecasting thresholds’, while RW forecasts are preferred on the other. In addition to explaining the relative performance of forecasts from these two models, these thresholds prove useful in developing model selection criteria that help a forecaster reduce error. Copyright © 2004 John Wiley & Sons, Ltd.     

Assessing the forecasting accuracy of alternative nominal exchange rate models: the case of long memory

This paper presents an autoregressive fractionally integrated moving‐average (ARFIMA) model of nominal exchange rates and compares its forecasting capability with the monetary structural models and the random walk model. Monthly observations are used for Canada, France, Germany, Italy, Japan and the United Kingdom for the period of April 1973 through December 1998. The estimation method is Sowell's (1992) exact maximum likelihood estimation. The forecasting accuracy of the long‐memory model is formally compared to the random walk and the monetary models, using the recently developed Harvey, Leybourne and Newbold (1997) test statistics. The results show that the long‐memory model is more efficient than the random walk model in steps‐ahead forecasts beyond 1 month for most currencies and more efficient than the monetary models in multi‐step‐ahead forecasts. This new finding strongly suggests that the long‐memory model of nominal exchange rates be studied as a viable alternative to the conventional models. Copyright © 2006 John Wiley & Sons, Ltd.     

Short-run macroeconomic forecasting: The OECD performance

Forecasts for the seven major industrial countries, Canada, France, Germany, Italy, Japan, the United Kingdom and the United States, are published on a regular basis in the OECD's Economic Outlook. This paper analyses the accuracy of the OECD annual forecasts of output and price changes and of the current balance in the balance of payments. As a reference basis, the forecasts are compared with those generated by a naive model, a random walk process. The measures of forecasting accuracy used are the mean-absolute error, the root-mean-square error, the median-absolute error, and Theil's inequality coefficient. The OECD forecasts of real GNP changes are significantly superior to those generated by the random walk process; however, the OECD price and current balance forecasts are not significantly more accurate than those obtained from the naive model. The OECD's forecasting performance has neither improved nor deteriorated over time.     

An outlier robust GARCH model and forecasting volatility of exchange rate returns

Since volatility is perceived as an explicit measure of risk, financial economists have long been concerned with accurate measures and forecasts of future volatility and, undoubtedly, the Generalized Autoregressive Conditional Heteroscedasticity (GARCH) model has been widely used for doing so. It appears, however, from some empirical studies that the GARCH model tends to provide poor volatility forecasts in the presence of additive outliers. To overcome the forecasting limitation, this paper proposes a robust GARCH model (RGARCH) using least absolute deviation estimation and introduces a valuable estimation method from a practical point of view. Extensive Monte Carlo experiments substantiate our conjectures. As the magnitude of the outliers increases, the one‐step‐ahead forecasting performance of the RGARCH model has a more significant improvement in two forecast evaluation criteria over both the standard GARCH and random walk models. Strong evidence in favour of the RGARCH model over other competitive models is based on empirical application. By using a sample of two daily exchange rate series, we find that the out‐of‐sample volatility forecasts of the RGARCH model are apparently superior to those of other competitive models. Copyright © 2002 John Wiley & Sons, Ltd.     

A threshold model for the French franc/deutschmark exchange rate

The behaviour of the French franc/deutschmark exchange rate is examined in this paper. During the time period studied, these currencies were constrained to lie within prescribed bands relative to one another and the usual random walk explanation of the exchange rate may not be appropriate. The data are examined for evidence of non-linear structure and it is shown that a piecewise linear SETAR model provides a better explanation and superior forecasting performance than a random walk.     

The importance of time‐varying volatility and country interactions in forecasting economic activity

This paper examines the relative importance of allowing for time‐varying volatility and country interactions in a forecast model of economic activity. Allowing for these issues is done by augmenting autoregressive models of growth with cross‐country weighted averages of growth and the generalized autoregressive conditional heteroskedasticity framework. The forecasts are evaluated using statistical criteria through point and density forecasts, and an economic criterion based on forecasting recessions. The results show that, compared to an autoregressive model, both components improve forecast ability in terms of point and density forecasts, especially one‐period‐ahead forecasts, but that the forecast ability is not stable over time. The random walk model, however, still dominates in terms of forecasting recessions.     

Forecasting the price of crude oil via convenience yield predictions

The paper develops an oil price forecasting technique which is based on the present value model of rational commodity pricing. The approach suggests shifting the forecasting problem to the marginal convenience yield, which can be derived from the cost‐of‐carry relationship. In a recursive out‐of‐sample analysis, forecast accuracy at horizons within one year is checked by the root mean squared error as well as the mean error and the frequency of a correct direction‐of‐change prediction. For all criteria employed, the proposed forecasting tool outperforms the approach of using futures prices as direct predictors of future spot prices. Vis‐à‐vis the random‐walk model, it does not significantly improve forecast accuracy but provides valuable statements on the direction of change. Copyright © 2007 John Wiley & Sons, Ltd.     

Forecasting the dollar/euro exchange rate: are international parities useful?

In this paper we assess the empirical relevance of an expectations version of purchasing power parity in forecasting the dollar/euro exchange rate. This version is based on the differential of inflation expectations derived from inflation‐indexed bonds for the euro area and the USA. Using the longest daily data for both the dollar/euro exchange rate and for the inflation expectations, our results suggest that, with few exceptions, our predictors behave significantly better than a random walk in forecasts up to five days, both in terms of prediction errors and in directional forecasts. Copyright © 2005 John Wiley & Sons, Ltd.     

Conditional predictability of daily exchange rates

At what forecast horizon is one time series more predictable than another? This paper applies the Diebold–Kilian conditional predictability measure to assess the out‐of‐sample performance of three alternative models of daily GBP/USD and DEM/USD exchange rate returns. Predictability is defined as a non‐linear statistic of a model's relative expected losses at short and long forecast horizons, allowing flexible choice of both the estimation procedure and loss function. The long horizon is set to 2 weeks and one month ahead and forecasts evaluated according to MSE loss. Bootstrap methodology is used to estimate the data's conditional predictability using GARCH models. This is then compared to predictability under a random walk and a model using the prediction bias in uncovered interest parity (UIP). We find that both exchange rates are less predictable using GARCH than using a random walk, but they are more predictable using UIP than a random walk. Predictability using GARCH is relatively higher for the 2‐weeks‐than for the 1‐month long forecast horizon. Comparing the results using a random walk to that using UIP reveals ‘pockets’ of predictability, that is, particular short horizons for which predictability using the random walk exceeds that using UIP, or vice versa. Overall, GBP/USD returns appear more predictable than DEM/USD returns at short horizons. Copyright © 2002 John Wiley & Sons, Ltd.     

Google Trends and the forecasting performance of exchange rate models

In this paper, we use Google Trends data for exchange rate forecasting in the context of a broad literature review that ties the exchange rate movements with macroeconomic fundamentals. The sample covers 11 OECD countries’ exchange rates for the period from January 2004 to June 2014. In out‐of‐sample forecasting of monthly returns on exchange rates, our findings indicate that the Google Trends search query data do a better job than the structural models in predicting the true direction of changes in nominal exchange rates. We also observed that Google Trends‐based forecasts are better at picking up the direction of the changes in the monthly nominal exchange rates after the Great Recession era (2008–2009). Based on the Clark and West inference procedure of equal predictive accuracy testing, we found that the relative performance of Google Trends‐based exchange rate predictions against the null of a random walk model is no worse than the purchasing power parity model. On the other hand, although the monetary model fundamentals could beat the random walk null only in one out of 11 currency pairs, with Google Trends predictors we found evidence of better performance for five currency pairs. We believe that these findings necessitate further research in this area to investigate the extravalue one can get from Google search query data.     

Forecast Evaluation of Nonlinear Models: The Case of Long‐Span Real Exchange Rates

This paper deals with the nonlinear modeling and forecasting of the dollar–sterling and franc–sterling real exchange rates using long spans of data. Our contribution is threefold. First, we provide significant evidence of smooth transition dynamics in the series by employing a battery of recently developed in‐sample statistical tests. Second, we investigate the small‐sample properties of several evaluation measures for comparing recursive forecasts when one of the competing models is nonlinear. Finally, we run a forecasting race for the post‐Bretton Woods era between the nonlinear real exchange rate model, the random walk, and the linear autoregressive model. The nonlinear model outperforms all rival models in the dollar–sterling case but cannot beat the linear autoregressive in the franc–sterling. Copyright © 2011 John Wiley & Sons, Ltd.     

Forecasting house prices in OECD economies

In this paper, we forecast real house price growth of 16 OECD countries using information from domestic macroeconomic indicators and global measures of the housing market. Consistent with the findings for the US housing market, we find that the forecasts from an autoregressive model dominate the forecasts from the random walk model for most of the countries in our sample. More importantly, we find that the forecasts from a bivariate model that includes economically important domestic macroeconomic variables and two global indicators of the housing market significantly improve upon the univariate autoregressive model forecasts. Among all the variables, the mean square forecast error from the model with the country's domestic interest rates has the best performance for most of the countries. The country's income, industrial production, and stock markets are also found to have valuable information about the future movements in real house price growth. There is also some evidence supporting the influence of the global housing price growth in out‐of‐sample forecasting of real house price growth in these OECD countries.     

Structural,VAR and BVAR models of exchange rate determination: A comparison of their forecasting performance

This paper compares the out-of-sample forecasting accuracy of a wide class of structural, BVAR and VAR models for major sterling exchange rates over different forecast horizons. As representative structural models we employ a portfolio balance model and a modified uncovered interest parity model, with the latter producing the more accurate forecasts. Proper attention to the long-run properties and the short-run dynamics of structural models can improve on the forecasting performance of the random walk model. The structural model shows substantial improvement in medium-term forecasting accuracy, whereas the BVAR model is the more accurate in the short term. BVAR and VAR models in levels strongly out predict these models formulated in difference form at all forecast horizons.     

N-step combinations of forecasts

While there is general agreement that a linear combination of forecasts can outperform the individual forecasts, there is controversy about the appropriateness of the combination method to be used in a given situation. Hence, in any given application it may be more beneficial to combine different sets of combined forecasts rather than picking one of them. This paper introduces the concept of N-step combinations of forecasts which involves combining the combined forecasts obtained from different combination procedures used at the preceding step. Using quarterly GNP data, evidence supporting the increase in the accuracy of the one-period-ahead ex-ante forecasts as the combination step increases is provided. The MSE, MAE, MAPE and their corresponding standard deviations are used to evaluate the accuracy of the forecasts obtained.     

Beating the random walk in Central and Eastern Europe

We compare the accuracy of vector autoregressive (VAR), restricted vector autoregressive (RVAR), Bayesian vector autoregressive (BVAR), vector error correction (VEC) and Bayesian error correction (BVEC) models in forecasting the exchange rates of five Central and Eastern European currencies (Czech Koruna, Hungarian Forint, Slovak Koruna, Slovenian Tolar and Polish Zloty) against the US Dollar and the Euro. Although these models tend to outperform the random walk model for long‐term predictions (6 months ahead and beyond), even the best models in terms of average prediction error fail to reject the test of equality of forecasting accuracy against the random walk model in short‐term predictions. Copyright © 2005 John Wiley & Sons, Ltd.     

Shrinkage‐Based Tests of Predictability

In this paper we derive a test of predictability by exploring the possibility that forecasts from a given model, adjusted by a shrinkage factor, will display lower mean squared prediction errors than forecasts from a simple random walk. This generalizes most previous tests which compare forecast errors of a benchmark model with errors of a proposed alternative model, not allowing for shrinkage. We show that our test is a particular extension of a recently developed test of the martingale difference hypothesis. Using simulations we explore the behavior of our test in small and moderate samples. Numerical results indicate that the test has good size and power properties. Finally, we illustrate the use of our test in an empirical application within the exchange rate literature. Copyright © 2012 John Wiley & Sons, Ltd.     

Causal forces: Structuring knowledge for time-series extrapolation

This paper examines a strategy for structuring one type of domain knowledge for use in extrapolation. It does so by representing information about causality and using this domain knowledge to select and combine forecasts. We use five categories to express causal impacts upon trends: growth, decay, supporting, opposing, and regressing. An identification of causal forces aided in the determination of weights for combining extrapolation forecasts. These weights improved average ex ante forecast accuracy when tested on 104 annual economic and demographic time series. Gains in accuracy were greatest when (1) the causal forces were clearly specified and (2) stronger causal effects were expected, as in longer-range forecasts. One rule suggested by this analysis was: ‘Do not extrapolate trends if they are contrary to causal forces.’ We tested this rule by comparing forecasts from a method that implicitly assumes supporting trends (Holt's exponential smoothing) with forecasts from the random walk. Use of the rule improved accuracy for 20 series where the trends were contrary; the MdAPE (Median Absolute Percentage Error) was 18% less for the random walk on 20 one-year ahead forecasts and 40% less for 20 six-year-ahead forecasts. We then applied the rule to four other data sets. Here, the MdAPE for the random walk forecasts was 17% less than Holt's error for 943 short-range forecasts and 43% less for 723 long-range forecasts. Our study suggests that the causal assumptions implicit in traditional extrapolation methods are inappropriate for many applications.