Time-series model forecasts and structural breaks: evidence from Spanish pre-EMU interest rates

Analysis of the future behaviour of economic variables can be biased if structural breaks are not considered. When these structural breaks are present, the in-sample fit of a model gives us a poor guide to ex ante forecast performance. This problem is true for both univariate and multivariate analysis and can be extremely important when co-integration relationships are analysed. The main goal of this article is to analyse the impact of structural breaks on forecast accuracy evaluation. We focus on forecasting several interest rates from the Spanish interbank money market. In order to carry out the analysis, we perform two forecasting exercises: (a) without structural breaks and (b) when structural breaks are explicitly considered. We use new sequential methods in order to estimate change-points in an endogenous way. This method allows us to detect structural breaks in all four rates in May 1993. However, the effects of these breaks are not very strong, since we found scarce gains in forecasting accuracy when the structural breaks are included in the models.     

Forecasting the US real house price index: Structural and non-structural models with and without fundamentals

We employ a 10-variable dynamic structural general equilibrium model to forecast the US real house price index as well as its downturn in 2006:Q2. We also examine various Bayesian and classical time-series models in our forecasting exercise to compare to the dynamic stochastic general equilibrium model, estimated using Bayesian methods. In addition to standard vector-autoregressive and Bayesian vector autoregressive models, we also include the information content of either 10 or 120 quarterly series in some models to capture the influence of fundamentals. We consider two approaches for including information from large data sets — extracting common factors (principle components) in factor-augmented vector autoregressive or Bayesian factor-augmented vector autoregressive models as well as Bayesian shrinkage in a large-scale Bayesian vector autoregressive model. We compare the out-of-sample forecast performance of the alternative models, using the average root mean squared error for the forecasts. We find that the small-scale Bayesian-shrinkage model (10 variables) outperforms the other models, including the large-scale Bayesian-shrinkage model (120 variables). In addition, when we use simple average forecast combinations, the combination forecast using the 10 best atheoretical models produces the minimum RMSEs compared to each of the individual models, followed closely by the combination forecast using the 10 atheoretical models and the DSGE model. Finally, we use each model to forecast the downturn point in 2006:Q2, using the estimated model through 2005:Q2. Only the dynamic stochastic general equilibrium model actually forecasts a downturn with any accuracy, suggesting that forward-looking microfounded dynamic stochastic general equilibrium models of the housing market may prove crucial in forecasting turning points.     

UK macroeconomic forecasting with many predictors: Which models forecast best and when do they do so?

Block factor methods offer an attractive approach to forecasting with many predictors. These extract the information in these predictors into factors reflecting different blocks of variables (e.g. a price block, a housing block, a financial block, etc.). However, a forecasting model which simply includes all blocks as predictors risks being over-parameterized. Thus, it is desirable to use a methodology which allows for different parsimonious forecasting models to hold at different points in time. In this paper, we use dynamic model averaging and dynamic model selection to achieve this goal. These methods automatically alter the weights attached to different forecasting models as evidence comes in about which has forecast well in the recent past. In an empirical study involving forecasting output growth and inflation using 139 UK monthly time series variables, we find that the set of predictors changes substantially over time. Furthermore, our results show that dynamic model averaging and model selection can greatly improve forecast performance relative to traditional forecasting methods.     

Volatility and VaR forecasting in the Madrid Stock Exchange

This paper provides an empirical study to assess the forecasting performance of a wide range of models for predicting volatility and VaR in the Madrid Stock Exchange. The models performance was measured by using different loss functions and criteria. The results show that FIAPARCH processes capture and forecast more accurately the dynamics of IBEX-35 returns volatility. It is also observed that assuming a heavy-tailed distribution does not improve models ability for predicting volatility. However, when the aim is forecasting VaR, we find evidence of that the Student’s t FIAPARCH outperforms the models it nests the lower the target quantile.      

OPTIMAL WEIGHT FOR REALIZED VARIANCE BASED ON INTERMITTENT HIGH‐FREQUENCY DATA*

Japanese stock markets have two types of breaks, overnight and lunch, during which no trading occurs, causing an inevitable increased variance in estimating daily volatility via a naive realized variance (RV). In order to perform a more stabilized estimation, we modify Hansen and Lunde's weighting technique. As an empirical study, we estimate optimal weights by using a particular approach for Japanese stock data listed on the Tokyo Stock Exchange, and then compare the forecast performance of weighted and non‐weighted RV through an autoregressive fractionally integrated moving average model. The empirical result indicates that the appropriate use of the optimally weighted RV can lead to remarkably smaller estimation variance compared with the naive RV, in many series. Therefore a more accurate forecasting of daily volatility data is obtained. Finally, we perform a Monte Carlo simulation to support the empirical result.     

THE FORECASTING PERFORMANCE OF SETAR MODELS: AN EMPIRICAL APPLICATION

The aim of this paper is to evaluate the forecasting performance of SETAR models with an application to the Industrial Production Index (IPI) of four major European countries over a period which includes the last Great Recession. Both point and interval forecasts are considered at different horizons against those obtained from two linear models. We follow the approach suggested by Teräsvirta et al. (2005) according to which a dynamic specification may improve the forecast performance of the nonlinear models with respect to the linear models. We re‐specify the models every twelve months and we find that the advantages of this procedure are particularly evident in the forecast rounds immediately following the re‐specification.     

Inflation forecasting using the New Keynesian Phillips Curve with a time-varying trend

Does theory aid inflation forecasting? To address this question, we develop a novel forecasting procedure based upon a New Keynesian Phillips Curve that incorporates time-varying trend inflation, to capture shifts in central bank preferences and monetary policy frameworks. We generate theory-implied predictions for both the trend and cyclical components of inflation, and recombine them to obtain an overall inflation forecast. Using quarterly data for the Euro Area and the United States that cover almost half a century, we compare our inflation forecasting procedure against the most popular time series models. We find that our theory-based forecasts outperform these benchmarks that previous studies found difficult to beat. Our results are shown to be robust to structural breaks, geographic areas, and variants of the econometric specification. Our findings suggest that the scepticism concerning the use of theory in forecasting is unwarranted, and theory should continue to play an important role in policymaking.     

Improving forecast accuracy of financial vulnerability: PLS factor model approach

We present a factor augmented forecasting model for assessing the financial vulnerability in Korea. Dynamic factor models often extract latent common factors from a large panel of time series data via the method of the principal components (PC). Instead, we employ the partial least squares (PLS) method that estimates target specific common factors, utilizing covariances between predictors and the target variable. Applying PLS to 198 monthly frequency macroeconomic time series variables and the Bank of Korea's Financial Stress Index (KFSTI), our PLS factor augmented forecasting models consistently outperformed the random walk benchmark model in out-of-sample prediction exercises in all forecast horizons we considered. Our models also outperformed the autoregressive benchmark model in short-term forecast horizons. We expect our models would provide useful early warning signs of the emergence of systemic risks in Korea's financial markets.     

Professional forecasts of interest rates and exchange rates: Evidence from the Wall Street Journal’s panel of economists

We analyze economists’ forecasts of interest rates and exchange rates from the Wall Street Journal. We find that a majority of economists produced unbiased forecasts but that none predicted directions of changes more accurately than chance. Most economists’ forecast accuracy is statistically indistinguishable from a random walk model in forecasting the Treasury bill rate, but many are significantly worse in forecasting the Treasury bond rate and the exchange rate. We also find systematic forecast heterogeneity, support for strategic models predicting the industry employing the economist matters, and evidence that economists deviate less from the consensus as they age.     

FORECASTING US INFLATION USING DYNAMIC GENERAL‐TO‐SPECIFIC MODEL SELECTION

We forecast US inflation using a standard set of macroeconomic predictors and a dynamic model selection and averaging methodology that allows the forecasting model to change over time. Pseudo out‐of‐sample forecasts are generated from models identified from a multipath general‐to‐specific algorithm that is applied dynamically using rolling regressions. Our results indicate that the inflation forecasts that we obtain employing a short rolling window substantially outperform those from a well‐established univariate benchmark, and contrary to previous evidence, are considerably robust to alternative forecast periods.     

Modelling mortality: are we heading in the right direction?

Predicting life expectancy has become of upmost importance in society. Pension providers, insurance companies, government bodies and individuals in the developed world have a vested interest in understanding how long people will live for. This desire to better understand life expectancy has resulted in an explosion of stochastic mortality models many of which identify linear trends in mortality rates by time. In making use of such models for forecasting purposes, we rely on the assumption that the direction of the linear trend (determined from the data used for fitting purposes) will not change in the future, recent literature has started to question this assumption. In this article, we carry out a comprehensive investigation of these types of models using male and female data from 30 countries and using the theory of structural breaks to identify changes in the extracted trends by time. We find that structural breaks are present in a substantial number of cases, that they are more prevalent in male data than in female data, that the introduction of additional period factors into the model reduces their presence, and that allowing for changes in the trend improves the fit and forecast substantially.     

Have structural changes eliminated the out-of-sample ability of financial variables to forecast real activity after the mid-1980s? Evidence from the Canadian economy

This article evaluates how consistently reliable the information content of individual financial variables is for Canada's future output growth. We estimate the timing of structural changes in linear growth models and check robustness to specification changes, multiple breaks, and business cycle asymmetry. Our simulated out-of-sample forecast evaluation strategy, using the Mean Square Error F-type (MSE-F) and the new encompassing (ENC-NEW) tests, shows that the leading information content of most financial variables for Canada's future Gross Domestic Product (GDP) growth has deteriorated substantially after 1984:04, but the 1–3-year term spread exhibits a consistently reliable predictive ability at the 1 and 2 quarter horizons and has significant forecasting ability at the 8 quarter horizon. Also, the real M1 money growth has regained its ability to forecast output growth since 1991:01.     

基于ARIMA与GARCH模型的国际油价预测比较分析

在分析影响油价波动因素的基础上,利用1986年1月至2010年12月的WTI国际原油价格月度数据,分别建立ARIMA和GARCH模型对油价进行预测。并通过对2011年1月至2012年4月WTI原油价格进行外推预测,检验模型的预测效果。比较分析发现,在短期预测中,ARIMA和GARCH模型对油价的预测均比较准确,但当油价由于受到重大事件的影响而有较大波动时,模型的预测精度下降;在长期预测中,GARCH模型的预测效果优于ARIMA模型;整体来看,GARCH模型预测的精度高于ARIMA模型。因此,在国际油价预测中,用GARCH模型是比较合适的。     

A bottom-up approach for forecasting GDP in a data-rich environment

In an increasingly data-rich environment, the use of factor models for forecasting purposes has gained prominence in the literature and among practitioners. Herein, we assess the forecasting behaviour of factor models to predict several GDP components and investigate the performance of a bottom-up approach to forecast GDP growth in Portugal, which was one of the hardest hit economies during the latest economic and financial crisis. We find supporting evidence of the usefulness of factor models and noteworthy forecasting gains when conducting a bottom-approach drawing on the main aggregates of GDP.     

Forecasting Inflation Using Constant Gain Least Squares

This paper assesses the usefulness of constant gain least squares when forecasting inflation. An out‐of‐sample forecast exercise is conducted, in which univariate autoregressive models for inflation in Australia, Sweden, the United Kingdom and the United States are used. The results suggest that it is possible to improve the forecast accuracy by employing constant gain least squares instead of ordinary least squares. In particular, when using a gain of 0.05, constant gain least squares generally outperforms the corresponding autoregressive model estimated with ordinary least squares. In fact, at longer forecast horizons, the root mean square forecast error is reliably lowered for all four countries and for all lag lengths considered in the study.     

A Bayesian method of combining judgmental and model-based density forecasts

This paper introduces a formal method of combining expert and model density forecasts when the sample of past forecasts is unavailable. It works directly with the expert forecast density and endogenously delivers weights for forecast combination, relying on probability rules only. The empirical part of the paper illustrates how the framework can be applied in forecasting US inflation by mixing density forecasts from an autoregressive model and the Survey of Professional Forecasters.     

Comparative accuracy of forecasts of inflation: a Canadian study

An inaccurate forecast of inflation is costlier to economic agents when the inflation rate is high and volatile. In this situation, the use of more sophisticated and information-oriented forecasting models become economically efficient. We test this hypothesis by analysing the forecasting accuracy of vector auto-regressive (VAR), auto-regressive integrated moving average (ARIMA) and static expectation models. We use Canadian data and divide the post-sample forecasting period into four sub-periods, based on high/low and volatile/stable inflation. Prediction errors are compared for both short-term and long-term forecasts. Finally, the paper proposes a portfolio approach for obtaining a more accurate forecast of inflation.     

Forecasting demand for a newly introduced product using reservation price data and Bayesian updating

Forecasting demand during the early stages of a product's life cycle is a difficult but essential task for the purposes of marketing and policymaking. This paper introduces a procedure to derive accurate forecasts for newly introduced products for which limited data are available. We begin with the assumption that the consumer reservation price is related to the timing with which the consumer adopts the product. The model is estimated using reservation price data derived through a consumer survey, and the forecast is updated with sales data as they become available using Bayes's rule. The proposed model's forecasting performance is compared with that of benchmark models (i.e., Bass model, logistic growth model, and a Bayesian model based on analogy) using 23 quarters' worth of data on South Korea's broadband Internet services market. The proposed model outperforms all benchmark models in both prelaunch and postlaunch forecasting tests, supporting the thesis that consumer reservation price can be used to forecast demand for a new product before or shortly after product launch.     

Forecasting economic activity by Bayesian bridge model averaging

This paper proposes the use of Bayesian model averaging (BMA) as an alternative tool to forecast GDP relative to simple bridge models and factor models. BMA is a computationally feasible method that allows us to explore the model space even in the presence of a large set of candidate predictors. We test the performance of BMA in now-casting by means of a recursive experiment for the euro area and the three largest countries. This method allows flexibility in selecting the information set month by month. We find that BMA-based forecasts produce smaller forecast errors than standard bridge model when forecasting GDP in Germany, France and Italy. At the same time, it also performs as well as medium-scale factor models when forecasting Eurozone GDP.     

Dissecting models' forecasting performance

The fact that the predictive performance of models used in forecasting stock returns, exchange rates, and macroeconomic variables is not stable and varies over time has been widely documented in the forecasting literature. Under these circumstances excessive reliance on forecast evaluation metrics that ignores this instability in forecasting accuracy, like squared errors averaged over the whole forecast evaluation sample, masks important information regarding the temporal evolution of relative forecasting performance of competing models. In this paper we suggest an approach based on the combination of the Cumulated Sum of Squared Forecast Error Differential (CSSFED) of Welch and Goyal (2008) and the Bayesian change point analysis of Barry and Hartigan (1993) that tracks the contribution of forecast errors to the aggregate measures of forecast accuracy observation by observation. In doing so, it allows one to track the evolution of the relative forecasting performance over time. We illustrate the suggested approach by using forecasts of the GDP growth rate in Switzerland.