Forecast combinations under structural break uncertainty

This paper proposes two new weighting schemes that average forecasts based on different estimation windows in order to account for possible structural change. The first scheme weights the forecasts according to the values of reversed ordered CUSUM (ROC) test statistics, while the second weighting method simply assigns heavier weights to forecasts that use more recent information. Simulation results show that, when structural breaks are present, forecasts based on the first weighting scheme outperform those based on a procedure that simply uses ROC tests to choose and forecast from a single post-break estimation window. Combination forecasts based on our second weighting scheme outperform equally weighted combination forecasts. An empirical application based on a NAIRU Phillips curve model for the G7 countries illustrates these findings, and also shows that combination forecasts can outperform the random walk forecasting model.     

Volatility persistence,long memory and time-varying unconditional mean: Evidence from 10 equity indices

This paper re-examines evidence of volatility persistence and long memory in the light of potential time-variation in the unconditional mean of the volatility series. Daily equity volatility is generally regarded as exhibiting long memory, however, recent evidence has suggested that long memory may be a spurious finding arising from neglected breaks or time-variation in unconditional variance. The results presented here suggested that long memory is apparent when analysed on the assumption that unconditional variance is constant. However, both breakpoint tests and a moving average application suggest that unconditional variance exhibits substantial, although slow moving, time-variation. The apparent long-memory property largely disappears when this time-variation is taken into account. A modification of the GARCH model to allow for mean variation generates improved volatility forecasting performance, but only over long horizon. At the daily level the assumption of a constant unconditional variance does not seem to affect forecasts.     

Optimal forecasts in the presence of structural breaks

This paper considers the problem of forecasting under continuous and discrete structural breaks and proposes weighting observations to obtain optimal forecasts in the MSFE sense. We derive optimal weights for one step ahead forecasts. Under continuous breaks, our approach largely recovers exponential smoothing weights. Under discrete breaks, we provide analytical expressions for optimal weights in models with a single regressor, and asymptotically valid weights for models with more than one regressor. It is shown that in these cases the optimal weight is the same across observations within a given regime and differs only across regimes. In practice, where information on structural breaks is uncertain, a forecasting procedure based on robust optimal weights is proposed. The relative performance of our proposed approach is investigated using Monte Carlo experiments and an empirical application to forecasting real GDP using the yield curve across nine industrial economies.     

Properties of optimal forecasts under asymmetric loss and nonlinearity

Evaluation of forecast optimality in economics and finance has almost exclusively been conducted under the assumption of mean squared error loss. Under this loss function optimal forecasts should be unbiased and forecast errors serially uncorrelated at the single period horizon with increasing variance as the forecast horizon grows. Using analytical results we show that standard properties of optimal forecasts can be invalid under asymmetric loss and nonlinear data generating processes and thus may be very misleading as a benchmark for an optimal forecast. We establish instead that a suitable transformation of the forecast error—known as the generalized forecast error—possesses an equivalent set of properties. The paper also provides empirical examples to illustrate the significance in practice of asymmetric loss and nonlinearities and discusses the effect of parameter estimation error on optimal forecasts.     

Forecast Rationality Tests in the Presence of Instabilities,with Applications to Federal Reserve and Survey Forecasts

This paper proposes a framework to implement regression‐based tests of predictive ability in unstable environments, including, in particular, forecast unbiasedness and efficiency tests, commonly referred to as tests of forecast rationality. Our framework is general: it can be applied to model‐based forecasts obtained either with recursive or rolling window estimation schemes, as well as to forecasts that are model free. The proposed tests provide more evidence against forecast rationality than previously found in the Federal Reserve's Greenbook forecasts as well as survey‐based private forecasts. It confirms, however, that the Federal Reserve has additional information about current and future states of the economy relative to market participants. Copyright © 2015 John Wiley & Sons, Ltd.     

Out-of-sample forecasts of China's economic growth and inflation using rolling weighted least squares

Macroeconomic forecasting in China is essential for the government to take proper policy decisions on government expenditure and money supply, among other matters. The existing literature on forecasting Chinas macroeconomic variables is unclear on the crucial issue of how to choose an optimal window to estimate parameters with rolling out-of-sample forecasts. This study fills this gap in forecasting economic growth and inflation in China, by using the rolling weighted least squares (WLS) with the practically feasible cross-validation (CV) procedure of Hong et al. (2018) to choose an optimal estimation window. We undertake an empirical analysis of monthly data on up to 30 candidate indicators (mainly asset prices) for a span of 17 years (2000–2017). It is documented that the forecasting performance of rolling estimation is sensitive to the selection of rolling windows. The empirical analysis shows that the rolling WLS with the CV-based rolling window outperforms other rolling methods on univariate regressions in most cases. One possible explanation for this is that these macroeconomic variables often suffer from structural changes due to changes in institutional reforms, policies, crises, and other factors. Furthermore, we find that, in most cases, asset prices are key variables for forecasting macroeconomic variables, especially output growth rate.     

Least squares learning and the US Treasury bill rate

Understanding how agents formulate their expectations about Fed behavior is important for market participants because they can potentially use this information to make more accurate estimates of stock and bond prices. Although it is commonly assumed that agents learn over time, there is scant empirical evidence in support of this assumption. Thus, in this paper we test if the forecast of the three month T-bill rate in the Survey of Professional Forecasters (SPF) is consistent with least squares learning when there are discrete shifts in monetary policy. We first derive the mean, variance and autocovariances of the forecast errors from a recursive least squares learning algorithm when there are breaks in the structure of the model. We then apply the Bai and Perron (1998) test for structural change to a forecasting model for the three month T-bill rate in order to identify changes in monetary policy. Having identified the policy regimes, we then estimate the implied biases in the interest rate forecasts within each regime. We find that when the forecast errors from the SPF are corrected for the biases due to shifts in policy, the forecasts are consistent with least squares learning.     

Nonparametric simultaneous testing for structural breaks

In this paper we consider a regression model with errors that are martingale differences. This modeling includes the regression of both independent and time series data. The aim is to study the appearance of structural breaks in both the mean and the variance functions, assuming that such breaks may occur simultaneously in both the functions. We develop nonparametric testing procedures that simultaneously test for structural breaks in the conditional mean and the conditional variance. The asymptotic distribution of an adaptive test statistic is established, as well as its asymptotic consistency and efficiency. Simulations illustrate the performance of the adaptive testing procedure. An application to the analysis of financial time series also demonstrates the usefulness of the proposed adaptive test in practice.     

Testing the assumptions behind importance sampling

Importance sampling is used in many areas of modern econometrics to approximate unsolvable integrals. Its reliable use requires the sampler to possess a variance, for this guarantees a square root speed of convergence and asymptotic normality of the estimator of the integral. However, this assumption is seldom checked. In this paper we use extreme value theory to empirically assess the appropriateness of this assumption. Our main application is the stochastic volatility model, where importance sampling is commonly used for maximum likelihood estimation of the parameters of the model.     

Forecasting the volatility of asset returns: The informational gains from option prices

A new class of forecasting models is proposed that extends the realized GARCH class of models through the inclusion of option prices to forecast the variance of asset returns. The VIX is used to approximate option prices, resulting in a set of cross-equation restrictions on the model’s parameters. The full model is characterized by a nonlinear system of three equations containing asset returns, the realized variance, and the VIX, with estimation of the parameters based on maximum likelihood methods. The forecasting properties of the new class of forecasting models, as well as a number of special cases, are investigated and applied to forecasting the daily S&P500 index realized variance using intra-day and daily data from September 2001 to November 2017. The forecasting results provide strong support for including the realized variance and the VIX to improve variance forecasts, with linear conditional variance models performing well for short-term one-day-ahead forecasts, whereas log-linear conditional variance models tend to perform better for intermediate five-day-ahead forecasts.     

Combining competing forecasts of inflation using a bivariate arch model

This paper sets out the basic structure of the bivariate generalization of Engle's ARCH model. Conditions which guarantee that the conditional covariance matrix is well defined are summarized, as are estimation and hypothesis testing.The process is used to combine forecasts where the weights are allowed to vary over time. Forecast errors from competing models are treated as a bivariate ARCH process so that the conditional covariance matrix adapts over time. At each point in time these conditional estimates of the variances and covariances are used to construct the optimal weights for combining the forecasts. Consequently, when one model is fitting well, its variance will be reduced and its weight will be increased.Two models of US inflation are constructed; one is a stylized monetarist model while the other is a mark-up model. The forecast errors are modeled as a simple bivariate ARCH process. Diagnostic tests reveal that this has overly restricted the parameterization of the covariance matrix. An approximation to the theoretically anticipated factor structure model is then estimated. The results in both cases show the weights varying over the sample period in moderately interpretable fashion.     

Pooling: An experimental study of alternative testing and estimation procedures in a two-way error component model

This paper considers a two-way error component model with no lagged dependent variable and investigates the performance of various testing and estimation procedures applied to this model by means of Monte Carlo experiments. The following results were found: (1) The Chow-test performed poorly in testing the stability of cross-section regressions over time and in testing the stability of time-series regression across regions. (2) The Roy-Zellner test performed well and is recommended for testing the poolability of the data. (3) The Hausman specification test, employed to test the orthogonality assumption, gave a low frequency of Type I errors. (4) The Lagrange multiplier test, employed to test for zero variance components, did well except in cases where it was badly needed. (5) The problem of negative estimates of the variance components was found to be more serious in the two-way model than in the one-way model. However, replacing the negative variance estimates by zero did not have a serious effect on the performance of the second-round GLS estimates of the regression coefficients. (6) As in the one-way model, all the two-stage estimation methods performed reasonably well. (7) Better estimates of the variance components did not necessarily lead to better second-round GLS estimates of the regression coefficients.     

On new variance approximations for linear models with inequality constraints

In this paper, we examine the estimation of linear models subject to inequality constraints with a special focus on new variance approximations for the estimated parameters. For models with one inequality restriction, the proposed variance formulas are exact. The variance approximations proposed in this paper can be used in regression analysis, Kalman filtering, and balancing national accounts, when inequality constraints are to be incorporated in the estimation procedure.     

Complete subset regressions

This paper proposes a new method for combining forecasts based on complete subset regressions. For a given set of potential predictor variables we combine forecasts from all possible linear regression models that keep the number of predictors fixed. We explore how the choice of model complexity, as measured by the number of included predictor variables, can be used to trade off the bias and variance of the forecast errors, generating a setup akin to the efficient frontier known from modern portfolio theory. In an application to predictability of stock returns, we find that combinations of subset regressions can produce more accurate forecasts than conventional approaches based on equal-weighted forecasts (which fail to account for the dimensionality of the underlying models), combinations of univariate forecasts, or forecasts generated by methods such as bagging, ridge regression or Bayesian Model Averaging.     

Volatility forecast comparison using imperfect volatility proxies

The use of a conditionally unbiased, but imperfect, volatility proxy can lead to undesirable outcomes in standard methods for comparing conditional variance forecasts. We motivate our study with analytical results on the distortions caused by some widely used loss functions, when used with standard volatility proxies such as squared returns, the intra-daily range or realised volatility. We then derive necessary and sufficient conditions on the functional form of the loss function for the ranking of competing volatility forecasts to be robust to the presence of noise in the volatility proxy, and derive some useful special cases of this class of “robust” loss functions. The methods are illustrated with an application to the volatility of returns on IBM over the period 1993 to 2003.     

Estimating time series models using the relevant cost function

In many forecasting problems, the forecast cost function is used only in evaluating the forecasts; a second cost function is used in estimating the parameters in the model. In this paper, I explore some of the ways in which the forecast cost function can be used in estimating the parameters and, more generally, in producing the forecasts. I define the optimal forecast and note that it may depend on the entire conditional distribution of the data, which is typically unknown. I then consider three of the steps involved in forming the forecast: approximating the optimal forecast, selecting the model, and estimating any unknown parameters. The forecast cost function forms the basis of the approximation, selection, and estimation. The methods are illustrated using time series models applied to 15 US macroeconomic series and in a small Monte Carlo experiment.     

Identifying Structural Breaks in Cointegrated Vector Autoregressive Models*

This article suggests an alternative formulation of the cointegrated vector autoregressive (VAR) model such that the coefficients for the deterministic terms have straightforward interpretations. These coefficients can be interpreted as growth rates and cointegration mean level coefficients and express long‐run properties of the model. For example, the growth rate coefficients tell us how much to expect (unconditionally) the variables in the system to grow from one period to the next, representing the underlying (steady state) growth in the variables. The estimation of the proposed formulation is made operationally in GRaM, which is a program for Ox Professional. GRaM can be used for analysing structural breaks when the deterministic terms include shift dummies and broken trends. By applying a formulation with interpretable deterministic components, different types of structural breaks can be identified. Shifts in both intercepts and growth rates, or combinations of these, can be tested for. The ability to distinguish between different types of structural breaks makes the procedure superior compared with alternative procedures. Furthermore, the procedure utilizes the information more efficiently than alternative procedures. Finally, interpretable coefficients of different types of structural breaks can be identified.     

Modelling and Forecasting Fiscal Variables for the Euro Area*

In this paper, we assess the possibility of producing unbiased forecasts for fiscal variables in the Euro area by comparing a set of procedures that rely on different information sets and econometric techniques. In particular, we consider autoregressive moving average models, Vector autoregressions, small‐scale semistructural models at the national and Euro area level, institutional forecasts (Organization for Economic Co‐operation and Development), and pooling. Our small‐scale models are characterized by the joint modelling of fiscal and monetary policy using simple rules, combined with equations for the evolution of all the relevant fundamentals for the Maastricht Treaty and the Stability and Growth Pact. We rank models on the basis of their forecasting performance using the mean square and mean absolute error criteria at different horizons. Overall, simple time‐series methods and pooling work well and are able to deliver unbiased forecasts, or slightly upward‐biased forecast for the debt–GDP dynamics. This result is mostly due to the short sample available, the robustness of simple methods to structural breaks, and to the difficulty of modelling the joint behaviour of several variables in a period of substantial institutional and economic changes. A bootstrap experiment highlights that, even when the data are generated using the estimated small‐scale multi‐country model, simple time‐series models can produce more accurate forecasts, because of their parsimonious specification.     

Censored regression quantiles with endogenous regressors

This paper develops a semiparametric method for estimation of the censored regression model when some of the regressors are endogenous (and continuously distributed) and instrumental variables are available for them. A “distributional exclusion” restriction is imposed on the unobservable errors, whose conditional distribution is assumed to depend on the regressors and instruments only through a lower-dimensional “control variable,” here assumed to be the difference between the endogenous regressors and their conditional expectations given the instruments. This assumption, which implies a similar exclusion restriction for the conditional quantiles of the censored dependent variable, is used to motivate a two-stage estimator of the censored regression coefficients. In the first stage, the conditional quantile of the dependent variable given the instruments and the regressors is nonparametrically estimated, as are the first-stage reduced-form residuals to be used as control variables. The second-stage estimator is a weighted least squares regression of pairwise differences in the estimated quantiles on the corresponding differences in regressors, using only pairs of observations for which both estimated quantiles are positive (i.e., in the uncensored region) and the corresponding difference in estimated control variables is small. The paper gives the form of the asymptotic distribution for the proposed estimator, and discusses how it compares to similar estimators for alternative models.     

Forecasting a long memory process subject to structural breaks

We develop an easy-to-implement method for forecasting a stationary autoregressive fractionally integrated moving average (ARFIMA) process subject to structural breaks with unknown break dates. We show that an ARFIMA process subject to a mean shift and a change in the long memory parameter can be well approximated by an autoregressive (AR) model and suggest using an information criterion (AIC or Mallows’ _Cp$C_p$) to choose the order of the approximate AR model. Our method avoids the issue of estimation inaccuracy of the long memory parameter and the issue of spurious breaks in finite sample. Insights from our theoretical analysis are confirmed by Monte Carlo experiments, through which we also find that our method provides a substantial improvement over existing prediction methods. An empirical application to the realized volatility of three exchange rates illustrates the usefulness of our forecasting procedure. The empirical success of the HAR-RV model can be explained, from an econometric perspective, by our theoretical and simulation results.