Combination forecasting for directional accuracy: An application to survey interest rate forecasts

Using published interest rates forecasts issued by professional economists, two combination forecasts designed to improve the directional accuracy of interest rate forecasting are constructed. The first combination forecast takes a weighted average of the individual forecasters' predictions. The more successful the forecaster was in past forecasts at predicting the direction of change in interest rates, the greater is the weight given to his/her current forecast. The second combination forecast is simply the forecast issued by the forecaster who had the greatest success rate at predicting the direction of change in interest rates in previous forecasts. In cases where two or more forecasters tie for best historic directional accuracy track record, the arithmetic mean of these forecasters is used. The study finds that neither combination forecasting method performs better than coin-flipping at predicting the direction of change in interest rates. Nor does either method beat the simple arithmetic mean of the predictions of all the forecasters surveyed at predicting the direction of change in interest rates.     

Rational Expectations and Macroeconomic Forecasts

The article presents extensive results from testing for bias and serially correlated errors in a collection of time series of quarterly multiperiod forecasts for six variables including real GNP growth, inflation, and unemployment. The analysis covers responses by 79 frequent participants in economic outlook surveys conducted regularly since 1968. It shows much greater incidence of apparently systematic errors for inflation than for the other variables. Also, the tests are more favorable to composite group forecasts than to most of the individual forecast sets.     

Combining probability forecasts

Summary.  Linear pooling is by far the most popular method for combining probability forecasts. However, any non-trivial weighted average of two or more distinct, calibrated probability forecasts is necessarily uncalibrated and lacks sharpness. In view of this, linear pooling requires recalibration, even in the ideal case in which the individual forecasts are calibrated. Towards this end, we propose a beta-transformed linear opinion pool for the aggregation of probability forecasts from distinct, calibrated or uncalibrated sources. The method fits an optimal non-linearly recalibrated forecast combination, by compositing a beta transform and the traditional linear opinion pool. The technique is illustrated in a simulation example and in a case-study on statistical and National Weather Service probability of precipitation forecasts.     

Some Problems in Forecasting Population for Government Planning Purposes

Some governments rely on centralized, official sets of population forecasts for planning capital facilities. But the nature of population forecasting, as well as the milieu of government forecasting in general, can lead to the creation of extrapolative forecasts not well suited to long-range planning. This report discusses these matters, and suggests that custom-made forecasts and the use of forecast guidelines and a review process stressing forecast assumption justification may be a more realistic basis for planning individual facilities than general-purpose, official forecasts.     

Using large data sets to forecast sectoral employment

We use several models using classical and Bayesian methods to forecast employment for eight sectors of the US economy. In addition to using standard vector-autoregressive and Bayesian vector autoregressive models, we also augment these models to include the information content of 143 additional monthly series in some models. Several approaches exist for incorporating information from a large number of series. We consider two multivariate approaches—extracting common factors (principal components) and Bayesian shrinkage. After extracting the common factors, we use Bayesian factor-augmented vector autoregressive and vector error-correction models, as well as Bayesian shrinkage in a large-scale Bayesian vector autoregressive models. For an in-sample period of January 1972 to December 1989 and an out-of-sample period of January 1990 to March 2010, we compare the forecast performance of the alternative models. More specifically, we perform ex-post and ex-ante out-of-sample forecasts from January 1990 through March 2009 and from April 2009 through March 2010, respectively. We find that factor augmented models, especially error-correction versions, generally prove the best in out-of-sample forecast performance, implying that in addition to macroeconomic variables, incorporating long-run relationships along with short-run dynamics play an important role in forecasting employment. Forecast combination models, however, based on the simple average forecasts of the various models used, outperform the best performing individual models for six of the eight sectoral employment series.     

PITMAN-CLOSENESS AS A MEASURE TO EVALUATE THE QUALITY OF FORECASTS

ABSTRACT

We consider Pitman-closeness to evaluate the performance of univariate and multivariate forecasting methods. Optimal weights for the combination of forecasts are calculated with respect to this criterion. These weights depend on the assumption of the distribution of the individual forecasts errors. In the normal case they are identical with the optimal weights with respect to the MSE-criterion (univariate case) and with the optimal weights with respect to the MMSE-criterion (multivariate case). Further, we present a simple example to show how the different combination techniques perform. There we can see how much the optimal multivariate combination can outperform different other combinations. In practice, we can find multivariate forecasts e.g., in econometrics. There is often the situation that forecast institutes estimate several economic variables.     

Cointegration and Long-Horizon Forecasting

We consider the forecasting of cointegrated variables, and we show that at long horizons nothing is lost by ignoring cointegration when forecasts are evaluated using standard multivariate forecast accuracy measures. In fact, simple univariate Box–Jenkins forecasts are just as accurate. Our results highlight a potentially important deficiency of standard forecast accuracy measures—they fail to value the maintenance of cointegrating relationships among variables—and we suggest alternatives that explicitly do so.     

基于深度学习LSTM神经网络的全球股票指数预测研究

作为深度学习技术的经典模型之一，长短期记忆（LSTM）神经网络在挖掘序列数据长期依赖关系中极具优势。基于深度神经网络优化技术，本文构造了一个深层LSTM神经网络并将其应用于全球30个股票指数三种不同期限的预测研究，结果发现：①LSTM神经网络具有很强的泛化能力，对全部指数不同期限的预测效果均很稳定；②LSTM神经网络具有优秀的预测精度，相比三种对照模型（SVR，MLP和ARIMA），其对全部指数的平均预测精度在不同期限上均有提升；③LSTM神经网络能够有效控制误差波动，其对全部指数的平均预测稳定度相比三种对照模型在不同期限上亦均有提高。鉴于LSTM神经网络在预测精度和稳定度两方面的优势，其未来在金融预测中将有广阔的应用前景。     

Online learning and forecast combination in unbalanced panels

This article evaluates the performance of a few newly proposed online forecast combination algorithms and compares them with some of the existing ones including the simple average and that of Bates and Granger (1969). We derive asymptotic results for the new algorithms that justify certain established approaches to forecast combination including trimming, clustering, weighting, and shrinkage. We also show that when implemented on unbalanced panels, different combination algorithms implicitly impute missing data differently, so that the performance of the resulting combined forecasts are not comparable. After explicitly imputing the missing observations in the U.S. Survey of Professional Forecasters (SPF) over 1968 IV-2013 I, we find that the equally weighted average continues to be hard to beat, but the new algorithms can potentially deliver superior performance at shorter horizons, especially during periods of volatility clustering and structural breaks.     

The relationship between the length of the base period and population forecast errors

"The base period of a population forecast is the time period from which historical data are collected for the purpose of forecasting future population values. The length of the base period is one of the fundamental decisions made in preparing population forecasts, yet very few studies have investigated the effects of this decision on population forecast errors. In this article the relationship between the length of the base period and population forecast errors is analyzed, using three simple forecasting techniques and data from 1900 to 1980 for states in the United States. It is found that increasing the length of the base period up to 10 years improves forecast accuracy, but that further increases generally have little additional effect. The only exception to this finding is long-range forecasts of rapidly growing states, in which a longer base period substantially improves forecast accuracy for two of the forecasting techniques."     

Real-Time Forecasts of the Real Price of Oil

We construct a monthly real-time dataset consisting of vintages for 1991.1–2010.12 that is suitable for generating forecasts of the real price of oil from a variety of models. We document that revisions of the data typically represent news, and we introduce backcasting and nowcasting techniques to fill gaps in the real-time data. We show that real-time forecasts of the real price of oil can be more accurate than the no-change forecast at horizons up to 1 year. In some cases, real-time mean squared prediction error (MSPE) reductions may be as high as 25% 1 month ahead and 24% 3 months ahead. This result is in striking contrast to related results in the literature for asset prices. In particular, recursive vector autoregressive (VAR) forecasts based on global oil market variables tend to have lower MSPE at short horizons than forecasts based on oil futures prices, forecasts based on autoregressive (AR) and autoregressive moving average (ARMA) models, and the no-change forecast. In addition, these VAR models have consistently higher directional accuracy.     

On the Optimal Use of Suboptimal Forecasts of Explanatory Variables

Ashley (1983) gave a simple condition for determining when a forecast of an explanatory variable (X_t ) is sufficiently inaccurate that direct replacement of X_t by the forecast yields worse forecasts of the dependent variable than does respecification of the equation to omit X_t . Many available macroeconomic forecasts were shown to be of limited usefulness in direct replacement. Direct replacement, however, is not optimal if the forecast's distribution is known. Here optimal linear forms in commercial forecasts of several macroeconomic variables are obtained by using estimates of their distributions. Although they are an improvement on the raw forecasts (direct replacement), these optimal forms are still too inaccurate to be useful in replacing the actual explanatory variables in forecasting models. The results strongly indicate that optimal forms involving several commercial forecasts will not be very useful either. Thus Ashley's (1983) sufficient condition retains its value in gauging the usefulness of a forecast of an explanatory variable in a forecasting model, even though it focuses on direct replacement.     

Composite Forecasting: An Integrated Approach and Optimality Reconsidered

This article shows when the theoretical Lagrange multiplier solution for combining forecasts has a regression representation. This solution is not optimal in general because it imposes a restriction on an otherwise more general linear form. The optimal linear predictor based on N forecasts is presented. This predictor is or is not a regression function depending on whether the latter function is linear. I also show that the Lagrange multiplier solution may often be nearly optimal. Hence, when estimating a composite forecast, the restriction imposed by this solution may prove useful. This observation is supported in an empirical example.     

Forecasting Contemporaneously Aggregated Vector ARMA Processes

Given a multiple time series that is generated by a multivariate ARMA process and assuming the objective is to forecast a weighted sum of the individual variables, then under a mean squared error measure of forecasting precision, it is preferable to forecast the disaggregated multiple time series and aggregate the forecasts, rather than forecast the aggregated series directly, if the involved processes are known. This result fails to hold if the processes used for forecasting are estimated from a given set of time series data. The implications of these results for empirical research are investigated using different sets of economic data.     

Scenarios, Uncertainty and Conditional Forecasts of the World Population

Current official population forecasts differ little from those that Whelpton made 50 years ago either in the cohort–component methodology used or in the arguments used to motivate the assumptions. However, Whelpton produced some of the most erroneous forecasts of this century. This suggests that current forecasters should ensure that they give users an assessment of the uncertainty of their forecasts. We show how simple statistical methods can be combined with expert judgment to arrive at an overall predictive distribution for the future population. We apply the methods to a world population forecast that was made in 1994. Accepting that point forecast, we find that the probability is only about 2% that the world population in the year 2030 will be less than the low scenario of 8317 million. The probability that the world population will exceed the high scenario of 10 736 million is about 13%. Similarly, the probability is only about 51% that the high–low interval of a recent United Nations (UN) forecast will contain the true population in the year 2025. Even if we consider the UN high–low intervals as conditional on the possible future policies of its member states, they appear to have a relatively small probability of encompassing the future population.     

LOCAL LINEAR FORECASTS USING CUBIC SMOOTHING SPLINES

This paper shows how cubic smoothing splines fitted to univariate time series data can be used to obtain local linear forecasts. The approach is based on a stochastic state‐space model which allows the use of likelihoods for estimating the smoothing parameter, and which enables easy construction of prediction intervals. The paper shows that the model is a special case of an ARIMA(0, 2, 2) model; it provides a simple upper bound for the smoothing parameter to ensure an invertible model; and it demonstrates that the spline model is not a special case of Holt's local linear trend method. The paper compares the spline forecasts with Holt's forecasts and those obtained from the full ARIMA(0, 2, 2) model, showing that the restricted parameter space does not impair forecast performance. The advantage of this approach over a full ARIMA(0, 2, 2) model is that it gives a smooth trend estimate as well as a linear forecast function.     

Predicting Turning Points Through the Integration of Multiple Models

A new method for forming composite turning-point (or other qualitative) forecasts is proposed. Rather than forming composite forecasts by the standard Bayesian approach with weights proportional to each model's posterior odds, weights are assigned to the individual models in proportion to the probability of each model's having the correct turning-point prediction. These probabilities are generated by logit models estimated with data on the models' past turning-point forecasts. An empirical application to gross national product/gross domestic product forecasting of 18 Organization for Economic Cooperation and Development countries demonstrates the potential benefits of the procedure     

Multi-step forecasts from threshold ARMA models using asymmetric loss functions

The forecasts generation from nonlinear time series models is investigated under general loss functions. After presenting the main results and some relevant features of these functions, the Linex loss has been used to generate multi-step forecasts from threshold autoregressive moving average models showing their main properties and some results connected to a proper transformation of the forecast errors. A simulation exercise highlights interesting properties of the proposed predictors, both in terms of their bias and their distribution, further clarifying how the Linex predictor can be helpful in empirical applications.     

Insights into the appropriate level of disaggregation for efficient time series model forecasting

This paper provides a potentially valuable insight on how to assess if the forecasts from an autoregressive moving average model based on aggregated data could be substantially improved through disaggregation. It is argued that, theoretically, the absence of moving average (MA) terms indicates that no forecasting efficiency improvements can be achieved through disaggregation. In practice, it is found that there is a strong correlation between the statistical significance of the MA component in the aggregate model and the magnitude of the forecast mean square error (MSE) decreases that can be achieved through disaggregation. That is, if a model includes significant MA terms, the forecast MSE improvements that may be gained from disaggregation could be substantial. Otherwise, they are more likely to be relatively small or non-existent.     

Improved autoregressive forecasts in the presence of non-normal errors

This paper is concerned with obtaining more accurate point forecasts in the presence of non-normal errors. Specifically, we apply the residual augmented least-squares (RALS) estimator to autoregressive models to utilize the additional moment restrictions embodied in non-normal errors. Monte Carlo experiments are performed to compare our RALS forecasts to forecasts based on the ordinary least-squares estimator and the least absolute deviations (LAD) estimator. We find that the RALS approach provides superior forecasts when the data are skewed. Compared to the LAD forecast, the RALS forecast has smaller mean squared prediction errors in the baseline case with normal errors.