Alternative Forecasting Methods Research Articles

Оценка зависимости финансово-экономических результатов предприятия от макроэкономических индикаторов

Among the main indicators of a company’s financial and economic success are indicators such as revenue and profit. Being separate units, the totality of large, medium and small enterprises is the foundation of the country’s economy. In our case, it is the Russian economy. It is obvious that each sector of the Russian economy will be influenced by macroeconomic indicators with different strengths, despite the external homogeneity of the sectors. This fact determines the need for an individual assessment of the influence of factors on each industry, as well as, in particular, special approaches to assessing their impact. Attempts to use conventional forecasting assessment methods face a number of challenges that affect the reliability of impact estimates and the robustness of long-term model-based forecasts. The article assesses the influence of macroeconomic indicators on the financial and economic results of Russian enterprises from various industries. As part of the work, the selection of key sectors of the domestic economy for the article, for which the procedure for modeling financial results will be carried out, is justified. A hypothesis is formulated about a set of factors that supposedly have a serious impact on both the Russian economy as a whole and on selected industries. The selected factors were used as input variables for the models. Based on processed data on enterprises taken from open sources, a number of models were built that included certain macroeconomic factors in various processing. Also, the rationale for selecting the best models from the resulting set based on information criteria is given. For enterprises in selected sectors of the Russian economy, quantitative estimates of the influence of certain factors that have the greatest impact on the output values of profit and revenue were obtained. The set of models was obtained by searching through many combinations of influencing time series processes and input variables. Some models evaluate the influence of structural elements of the time series that are contained in the dependent variable. Based on the aggregation results, an interpretation was carried out aimed at demonstrating the possibility of obtaining high-quality conclusions when using alternative forecasting methods. It is proposed to use the developed models to predict the financial results of Russian enterprises

Out-of- Sample Stock Return Predictability of Alternative COVID-19 Indices

ABSTRACT We explore the predictive value of the various indices developed to capture COVID-19 pandemic for daily stock return predictability of 24 Emerging Market economies (based on data availability). We identify eight measures of COVID-19 indices, namely, the uncertainty due to pandemics and epidemics (UPE) index, Global Fear Index (GFI), COVID index, vaccine index, medical index, travel index, uncertainty index and aggregate COVID-19 sentiment index. We find that, out of the considered measures, the GFI consistently offers the best out-of-sample forecast gains followed by the aggregate COVID-19 sentiment index while the UPE index offers the least predictability gains. The outcome generally improves after controlling for oil price but the ranking of forecast performance remains the same and robust to multiple forecast horizons and alternative forecast evaluation methods. We infer that the relative predictive powers of the indices are proportional to the extent to which the indices truly measure the pandemic.

Electric Load Forecasting Model Using a Multicolumn Deep Neural Networks

In this article, a new approach to short-term load forecasting is proposed using a multicolumn radial basis function neural network (MCRN). The advantage of this new approach over similar models in speed and accuracy is also discussed, especially in regards to renewable generation forecasting. Because weather and seasonal effects have a direct impact not only on load demand but also on renewable energy production, it follows that as the penetration rate of renewable DG increases, the grid will become even more sensitive to weather impacts in the long term. In our approach, we use a k-d tree algorithm to split our feature-rich dataset into dense specialized subsets. These subsets are then trained in parallel as multiple artificial neural networks using a modified error correction algorithm to form the MCRN. This approach reduces the number of hidden neurons, increases the speed of convergence, and improves generalization over similar alternative forecasting methods.

Optimal selection of expert forecasts with integer programming

Combinations of point forecasts from expert forecasters are known to frequently outperform individual forecasts. It is also well documented that combination by simple averaging very often has performance superior to that of more sophisticated combinations. This empirical fact is referred to as the ‘forecast combination puzzle’ in the literature. In this paper, we propose a combination method that exploits this puzzle. Rather than averaging over all forecasts, our method optimally selects forecasts for averaging. The problem of optimal selection is solved using integer programming, a solution approach that has witnessed astonishing advancements. We apply this new method to forecasts of real GDP growth and unemployment from the European Central Bank Survey of Professional Forecasters. The results show that it is optimal to select only a small number of the available forecasts and that averaging over these small subsets almost always provides performance that is superior to averaging over all forecasts. Importantly, this new method is consistently one of the best performers when evaluated against a wide range of alternative forecast combination methods.

Optimal Selection of Expert Forecasts with Integer Programming

Combinations of point forecasts from expert forecasters are known to frequently outperform individual forecasts. It is also well documented that combination by simple averaging very often has performance superior to that of more sophisticated combinations. This empirical fact is referred to as the ‘forecast combination puzzle’ in the literature. In this paper, we propose a combination method that exploits this puzzle. Rather than averaging over all forecasts, our method optimally selects forecasts for averaging. The problem of optimal selection is solved using integer programming, a solution approach that has witnessed astonishing advancements. We apply this new method to forecasts of real GDP growth and unemployment from the European Central Bank Survey of Professional Forecasters. The results show that it is optimal to select only a small number of the available forecasts and that averaging over these small subsets almost always provides performance that is superior to averaging over all forecasts. Importantly, this new method is consistently one of the best performers when evaluated against a wide range of alternative forecast combination methods.

The Dynamic Newsvendor Model with Correlated Demand

ABSTRACTThe classic newsvendor model was developed under the assumption that period‐to‐period demand is independent over time. In real‐life applications, the notion of independent demand is often challenged. In this article, we examine the newsvendor model in the presence of correlated demands. Specifically under a stationary AR(1) demand, we study the performance of the traditional newsvendor implementation versus a dynamic forecast‐based implementation. We demonstrate theoretically that implementing a minimum mean square error (MSE) forecast model will always have improved performance relative to the traditional implementation in terms of cost savings. In light of the widespread usage of all‐purpose models like the moving‐average method and exponential smoothing method, we compare the performance of these popular alternative forecasting methods against both the MSE‐optimal implementation and the traditional newsvendor implementation. If only alternative forecasting methods are being considered, we find that under certain conditions it is best to ignore the correlation and opt out of forecasting and to simply implement the traditional newsvendor model.

The Dynamic Newsvendor Model with Correlated Demand

The classic newsvendor model was developed under the assumption that period-to-period demand is independent over time. In real-life applications, the notion of independent demand is often challenged. In this paper, we examine the newsvendor model in the presence of correlated demands. Specifically under a stationary AR(1) demand, we study the performance of the traditional newsvendor implementation versus a dynamic forecast-based implementation. We demonstrate theoretically that implementing a minimum mean square error (MSE) forecast model will always have improved performance relative to the traditional implementation in terms of cost savings. In light of the widespread usage of all-purpose models like the moving-average method and exponential smoothing method, we compare the performance of these popular alternative forecasting methods against both the MSE-optimal implementation and the traditional newsvendor implementation. If only alternative forecasting methods are being considered, we find that under certain conditions it is best to ignore the correlation and opt out of forecasting and to simply implement the traditional newsvendor model.

Production planning forecasting method selection in a supply chain: a case study

In modern supply chain, forecasting is critical for managers to ensure the availability of the proper amount of inventory to satisfy customers with minimum deviation. Without appropriate forecasting, it is very difficult to plan effectively and efficiently. After scanning the literature, it is to be noted that there is no recommendation of a specific method that is best for forecasting. This is also due to the availability of many alternative forecasting methods and criteria of preferences. Therefore, this paper proposes a ranking of forecasting methods for production planning in a supply chain. The proposed model is based on the analytical hierarchy process (AHP) since it has been proven useful in multi-criteria decision-making in many industrial and real life applications. It considers seven popular methods (decomposition, winter exponential, double exponential, single exponential, trend analysis, five months moving average and 12 months moving average) and three preference criteria (MAD, MSD and MAPE). The model is tested using a real case study. The sensitivity of the ranking was also tested and the ranges of criteria preferences are determined for robustness.

Modeling the leadership – project performance relation: radial basis function, Gaussian and Kriging methods as alternatives to linear regression

The purpose of this paper is to analyze alternative forecasting methods that produce results at least similar to or better than linear regression (MLR) that can be used in the modeling of social systems. While organizations may be considered as typically non-linear systems, the common feature of most models found in literature continues to be the use of linear regression techniques. From a case study, advanced statistical methods of Gaussian and Kriging are evaluated, as well as an artificial intelligence (AI) tool, the radial basis function (RBF). The results show the best performance of the suggested methods compared to MLR, especially RBF, because of its uniform prediction behavior throughout all ranges of evaluation. These techniques, although somewhat unconventional in social systems modeling, present a potential contribution in increasing the accuracy and precision of the predictions allowing a more accurate assessment of the impact of certain strategies on the project performance to be made before the allocation of material, human and financial resources.

Forecasting for cruise line revenue management

In recent years, the cruise line industry has become an exciting growth category in the leisure travel market. Like airlines and hotels, it reports all characteristics of revenue management (RM). Although RM has attracted widespread research interest in airline and hotel contexts, studies of cruise line revenue management are very limited. Using data from a major North American cruise company, we apply a variety of (24) forecasting methods, which are divided into three categories (non-pickup methods, classical pickup (CP) methods and advanced pickup (AP) methods), to generate forecasts of final bookings for the cruises that have not yet departed at a particular reading point. We use a two-stage framework to test alternative forecasting methods and compare their performance. We found the performance of multiplicative methods to be significantly worse. Among the additive methods, we find that classical methods perform the best, followed by AP and non-pickup methods. All CP methods with the exception of exponential smoothing with trend perform fairly well. Among AP methods, Autoregressive integrated moving average, linear regression and moving average (MA) produce the most accurate forecasts. Within non-pickup methods, MA is the most effective method.

Forecasting monthly industrial production in real-time: from single equations to factor-based models

The aim of this paper is to analyze the performance of alternative forecasting methods to predict the index of industrial production in Italy from 1 to 3 months ahead. We use twelve different models, from simple ARIMA to dynamic factor models exploiting the timely information of up to 110 short-term indicators, both qualitative and quantitative. This allows to assess the relevance for the forecasting practice of alternative combinations of types of data (real-time and latest available), estimation methods and periods. Out-of-sample predictive ability tests stress the relevance of more indicators in disaggregate models over sample periods covering a complete business cycle (about 7 years in Italy). Our findings downgrade the emphasis on both the estimation method and data revision issues. In line with the classical “average puzzle”, the use of simple averages of alternative forecasts often improves the predictive ability of their single components, mainly over short horizons. Finally, selected indicators and factor-based models always perform significantly better than ARIMA models, suggesting that the short-run indicator signal always dominates the noise component. On this regard, selected indicators models can further increase the amount of signal extracted to improve up to 30–40% the short-run predictive ability of factor-based models and to forecast-encompass them.

Lock-in of Extrapolative Expectations in an Asset Pricing Model

This paper examines an agent's choice of forecast method within a standard asset pricing model. To make a conditional forecast, a representative agent may choose one of the following: (1) a rational (or fundamentals-based) forecast that employs knowledge of the stochastic process governing dividends, (2) a constant forecast based on a simple long-run average of the forecast variable, or (3) a time-varying forecast that extrapolates from the last observation of the forecast variable. I show that a representative agent who is concerned about minimizing forecast errors may inadvertently become locked-in to an extrapolative forecast. In particular, the initial use of extrapolation alters the law of motion of the forecast variable so that the agent perceives no accuracy gain from switching to one of the alternative forecast methods. Under extrapolative expectations, the model can generate excess volatility of stock prices, time-varying volatility of returns, long-horizon predictability of returns, bubbles driven by optimism about the future, and sharp downward movements in stock prices that resemble market crashes. All of these features appear to be present in long-run U.S. stock market data.

LOCK-IN OF EXTRAPOLATIVE EXPECTATIONS IN AN ASSET PRICING MODEL

This paper examines an agent's choice of forecast method within a standard asset pricing model. A representative agent may choose: (1) a fundamentals-based forecast that employs knowledge of the dividend process, (2) a constant forecast that is based on a simple long-run average, or (3) a time-varying forecast that extrapolates from the last observation. I show that an agent who is concerned about minimizing forecast errors may inadvertently become “locked-in” to an extrapolative forecast. In particular, the initial use of extrapolation alters the law of motion of the forecast variable so that the agent perceives no accuracy gain from switching to one of the alternative forecast methods. The model can generate excess volatility of stock prices, time-varying volatility of returns, long-horizon predictability of returns, bubbles driven by optimism about the future, and sharp downward movements in stock prices that resemble market crashes.

Lock-in of Extrapolative Expectations in an Asset Pricing Model

This paper examines an agent's choice of forecast method within a standard asset pricing model. To make a conditional forecast, a representative agent may choose one of the following: (1) a rational (or fundamentals-based) forecast that employs knowledge of the stochastic process governing dividends, (2) a constant forecast based on a simple long-run average of the forecast variable, or (3) a time-varying forecast that extrapolates from the last observation of the forecast variable. I show that a representative agent who is concerned about minimizing forecast errors may inadvertently become locked-in to an extrapolative forecast. In particular, the initial use of extrapolation alters the law of motion of the forecast variable so that the agent perceives no accuracy gain from switching to one of the alternative forecast methods. Under extrapolative expectations, the model can generate excess volatility of stock prices, time-varying volatility of returns, long-horizon predictability of returns, bubbles driven by optimism about the future, and sharp downward movements in stock prices that resemble market crashes. All of these features appear to be present in long-run U.S. stock market data.

Forecasting international bandwidth capability

AbstractM‐competition studies provide a set of stylized recommendations to enhance forecast reliability. However, no single method dominates across series, leading to consideration of the relationship between selected data characteristics and the reliability of alternative forecast methods. This study conducts an analysis of predictive accuracy in relation to Internet bandwidth loads. Extrapolation techniques that perform best in M‐competitions perform relatively poorly in predicting Internet bandwidth loads. Such performance is attributed to Internet bandwidth data exhibiting considerably less structure than M‐competition data. Copyright © 2005 John Wiley & Sons, Ltd.

The dynamics of government revenue forecasting from an organizational perspective: a review of the literature

Over a decade ago, Bretschneider and Gorr proposed two directions for future research in government forecasting: one was to extend the research on developing and evaluating alternative forecasting methods and the other, to look at forecasting as a human activity and examine how organizational factors affect forecasting. What has happened since then? To see partially what has been done and what remains to be done, this paper provides a review of the literature on government revenue forecasting with a primary focus on the state level and identifies areas for future research in government revenue forecasting.

Forecast Pooling for European Macroeconomic Variables*

AbstractWe compare alternative forecast pooling methods and 58 forecasts from linear, time‐varying and non‐linear models, using a very large dataset of about 500 macroeconomic variables for the countries in the European Monetary Union. On average, combination methods work well but single non‐linear models can outperform them for several series. The performance of pooled forecasts, and of non‐linear models, improves when focusing on a subset of unstable series, but the gains are minor. Finally, on average over the EMU countries, the pooled forecasts behave well for industrial production growth, unemployment and inflation, but they are often beaten by non‐linear models for each country and variable.

Forecasting Monthly Tourist Departures from Australia

This paper builds time series forecasting models for monthly international tourist departures from Australia, categorized by the major destinations and by the main purposes of travel. Descriptive and statistical analyses reveal that strong trend and seasonality are the major features of all time series considered. The seasonal ARIMA model is used as a forecasting tool and its trend and seasonal components are specified based on inferential outcomes of the HEGY test for seasonal unit roots. Other alternative forecasting methods used include Holt-Winters' exponential smoothing method and the regression models that involve deterministic trend or seasonal dummy variables. It is found that forecasts generated from seasonal ARIMA models are more accurate than other alternative forecasts in most cases.

Predictive performance of the World Bank's commodity price projections

ABSTRACTThe World Bank's commodity price projections are widely used for various planning purposes. Two aspects of the Bank's projections of relative prices are studied in this paper. The first is whether the forecasts make efficient use of the information available at the time the forecast is made. The second is whether the forecasts predict future prices with greater accuracy than alternative forecasting methods. These matters are studied by comparing the World Bank's past price projections with the actual prices that were subsequently observed. The results show that, overall, the World Bank forecasts do not pass either test. First, the World Bank forecasts are informationally inefficient. Prediction error (projection minus actual price) tends to be positively correlated with the projections themselves. Although the direction of future price movements tends to be correctly predicted, the magnitude of these movements tends to be overpredicted. Second, the World Bank forecasts do not perform well even compared with the simplest of alternative forecasting methods ‐ the prediction of no change.

A comparison of the performance of different univariate forecasting methods in a model of capacity acquisition in UK electricity supply

Forecasting competitions have usually compared the accuracy of different forecasting methods across a number of different time series. This paper describes a study of the application of ten forecasting methods to a single time series: that of peak electricity demand in England and Wales. The performance measure used, however, is not one of the usual forecast ones, e.g., MSE or MAPS, but a managerial one in that the impact of different forecast methods on the profitability of the Central Electricity Generating Board for England and Wales is assessed using a financial simulation model. As well as examing the effects of forecast method on profitability the effects of two other factors, namely the use of a temperature corrected data series and the impact of log transformation of the data are considered. All these effects are both statistically and practically significant. The results are then examined from a different standpoint: specifically the extent to which the financial impacts of alternative forecast methods can be explained using a number of conventional forecast accuracy measures. This question is of major importance in applications since accuracy is one of the few easily measured characteristics of a potential forecasting method. It is concluded that much, though not all, of the results can be explained by accuracy considerations.