Accuracy in forecasting: A survey

In this study, the author provides a brief and concise summary of empirical investigations pertaining to forecasting with special reference to the accuracy of different forecasting techniques. The study mainly focuses on comparisons of the accuracy of these techniques. The comparisons cover both quantitative and qualitative methods. In addition the summary includes studies seeking to test or improve accuracy by combining forecasting techniques.     

Forecasting future canadian residential heating demand: An illustration of forecasting with aggregated and disaggregated data

This paper examines the sensitivity of forecasts to the level of aggregation of the data. A relative shares regression model and a multinominal logit model are tested with both aggregate and disaggregate survey data from 2109 respondents. The results indicate the appropriate model to use depends on whether the data are disaggregate or aggregate in form. Forecasts of solar heating of dwelling unit demand and market shares are also reported for Canada in terms of the solar price relative to the natural gas price and solar reliability relative to natural gas reliability.     

Feedback forecasting games: An overview

The range and rationales for forecasting schemes involving feedback are surveyed. Classic Delphi schemes involve an iterated exchange of information between a small group of experts in the pursuit of efficiencies arising from the ‘collective mind’. In other applications the event or state to be forecast may itself depend upon the forecast, and feedback forecasting schemes can be derived to allow for such reflexivity. Most situations of informational feedback are at least implicitly informational or predictive games, and a recognition of this character may help in understanding the possible or likely outcomes, including their social desirability. Various informational transfer devices such as real time surveys may speed up the process of feedback.     

The accuracy of extrapolation (time series) methods: Results of a forecasting competition

In the last few decades many methods have become available for forecasting. As always, when alternatives exist, choices need to be made so that an appropriate forecasting method can be selected and used for the specific situation being considered. This paper reports the results of a forecasting competition that provides information to facilitate such choice. Seven experts in each of the 24 methods forecasted up to 1001 series for six up to eighteen time horizons. The results of the competition are presented in this paper whose purpose is to provide empirical evidence about differences found to exist among the various extrapolative (time series) methods used in the competition.     

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 envelope function model for forecasting athletics records

The purpose of this paper is to present a multiplicative model for the fastest time per year in the 400- and 800-metre races. It is composed of an envelope function and a stationary time series. The proposed model is plausible since there is dependency between the fastest times of successive years and because improvements are expected to be smaller now, as compared with several years ago. Data from the years 1860 to 1988 are used for illustration.     

An exploration of some practical issues in the use of quantitative forecasting models

Extrapolative forecasting models have been available for many years and as most organizations have the need to regularly develop forecasts one might anticipate the widespread use of these models. The evidence in Australia indicates that computer based forecasting systems are not being widely used and in fact a number of established systems have been discarded, with the issue of forecast accuracy often being mentioned as a problem area. Two experiments are carried out to examine this issue by comparing judgemental and quantitative forecasts. Other problem areas mentioned as contributing to the abandonment of forecasting systems include the difficulty of manually reviewing the computer forecasts and the effort required to carefully massage the forecast database to remove extraordinary events.     

The teachers/practitioners corner forecasting in business schools: A survey

The study of forecasting techniques has received increased attention in recent years. How to incorporate this topic into the business school curriculum is a frequent subject of discussion. The purpose of this study was to determine whether forecasting is being taught in business schools and how it is incorporated into the curriculum. The survey instrument was sent to 622 member institutions of the American Assembly of Collegiate schools of Business. The importance of teaching forecasting techniques at both the undergraduate and graduate level was investigated.     

An evaluation of bayesian forecasting

‘Bayesian forecasting’ is a time series method of forecasting which (in the United Kingdom) has become synonymous with the state space formulation of Harrison and Stevens (1976). The approach is distinct from other time series methods in that it envisages changes in model structure. A disjoint class of models is chosen to encompass the changes. Each data point is retrospectively evaluated (using Bayes theorem) to judge which of the models held. Forecasts are then derived conditional on an assumed model holding true. The final forecasts are weighted sums of these conditional forecasts. Few empirical evaluations have been carried out. This paper reports a large scale comparison of time series forecasting methods including the Bayesian. The approach is two fold: a simulation study to examine parameter sensitivity and an empirical study which contrasts Bayesian with other time series methods.     

Awareness and use of forecasting techniques in british industry

The paper summarizes results of a mail survey of the use of formal forecasting techniques in British manufacturing companies. It appraises the state of awareness of particular techniques and the extent to which they are used in various functional applications. The extent to which the forecasts generated by the techniques influence company action is assessed; and the reasons for the non-use of particular techniques examined. The paper concludes that although an increasing number of companies appreciate the importance of forecasting, the methods used are predominantly naïve and few companies are taking steps to improve the situation through using alternative techniques or through computerizing established techniques.     

The use of non-time series information in sales forecasting: A case study

The contribution of product and industry knowledge to the accuracy of sales forecasting was investigated by examining the company forecasts of a leading manufacturer and marketer of consumable products. The company forecasts of 18 products produced by a meeting of marketing, sales, and production personnel were compared with those generated by the same company personnel when denied specific product knowledge and with the forecasts of selected judgemental and statistical time series methods. Results indicated that product knowledge contributed significantly to forecast accuracy and that the forecast accuracy of company personnel who possessed industry forecasting knowledge (but not product knowledge) was not significantly different from the time series based methods. Furthermore, the company forecasts were more accurate than averages of the judgemental and statistical time series forecasts. These results point to the importance of specific product information to forecast accuracy and accordingly call into question the continuing strong emphasis on improving extrapolation techniques without consideration of the inclusion of non-time series knowledge.     

An evaluation of forecasting using leading indicators

We consider the use of indices of leading indicators in forecasting and macro-economic modelling. The procedures used to select the components and construct the indices are examined, noting that the composition of indicator systems gets altered frequently. Cointegration within the indices, and between their components and macro-economic variables are considered as well as the role of co-breaking to mitigate regime shifts. Issues of model choice and data-based restrictions are investigated. A framework is proposed for index analysis and selecting indices, and applied to the UK longer-leading indicator. The effects of adding leading indicators to macro models are considered theoretically and for UK data.     

Discretionary databases in forecasting

Information for forecasting databases is often initially under the control of individuals who have no compelling reason to contribute, and who face various significant costs if they do. Such discretionary databases are subject to public goods problems, and are likely to be undersupplied, even when all participants agree that the overall benefits outweigh the overall costs. This paper explores the implications of this incentive structure for the existence, completeness and accuracy of forecasting databases. It also offers some hypotheses as to when the difficulties will be more and less severe, and outlines some directions for possible remedial strategies.     

The long run,causality, and forecasting in the advertising-sales relationship

Co-integration analysis is used in a study of the advertising and sales relationship using the Lydia Pinkham data set. The series are shown to have a valid long-run relationship while Granger-causality runs in both directions. The latter is found by using a causality test involving the co-integration restrictions which seem to constitute a crucial part of such tests in the case of co-integrated variables. A comparison with previous models shows that forecasting co-integrated series is more accurate with error-correction systems, especially in the case of multi-step forecasting.     

Tourism forecasting: A review of empirical research

Many publications on tourism forecasting have appeared during the past twenty years. The purpose of this article is to organize and summarize that scattered literature. General conclusions are also drawn from the studies to help those wishing to develop tourism forecasts of their own. The forecasting techniques discussed include time series models, econometric causal models, the gravity model and expert-opinion techniques. The major conclusions are that time series models are the simplest and least costly (and therefore most appropriate for practitioners); the gravity model is best suited to handle international tourism flows (and will be most useful to governments and tourism agencies); and expert-opinion methods are useful when data are unavailable. Further research is needed on the use of economic indicators in tourism forecasting, on the development of attractivity and emissiveness indexes for use in gravity and econometric models and on empirical comparisons among the different methods.     

On trend extraction models: Interpretation,empirical evidence and forecasting performance

This paper deals with the economic interpretation of the unobserved components model in the light of the apparent problem posed by previous work in that several practiced methodologies seem to lead to very different models of certain economic variables. A detailed empirical analysis is carried out to show how the failure in obtaining quasi-orthogonal components can seriously bias the interpretation of some decomposition procedures. Finally, the forecasting performance (in both the short and long run) of these decomposition models is analyzed in comparison with other alternatives.     

Discretization of stochastic differential equations and econometric forecasting: An application to time-varying autoregressions

Continuous time versions of time varying Vector Autoregressions are stochastic differential equations. Optimal discretization of Stochastic Differential Equations cannot be obtained by replacing all differentials by the corresponding first differences. In this paper we obtain the optimal discretization for time varying VARS. The results are applied to predicting the consumer price index.     

The importance of non-linearities in large forecasting models with stochastic error processes

This paper considers the consequences of the stochastic error process in large non-linear forecasting models. As such models are non-linear, the deterministic forecast is neither the mean nor the mode of the density function of the endogenous variables. Under a specific assumption as to the class of the non-linearity it is shown that the deterministic forecast is actually the vector of marginal medians of the density function. Stochastic simulation techniques are then used to test whether one large forecasting model actually lies within this class.     

Perspectives on forecasting research in accounting and finance

This paper offers some perspectives on forecasting research in accounting and finance. It is maintained that many common areas of forecasting research exist. Yet, most research has focused upon a particular (Box-Jenkins) technique and a particular (reported earnings) variable, virtually neglecting numerous other relevant forecasting research topics. This symposium issue includes papers which address several of these neglected research topics. The eight papers constituting the issue are classified into three categories: (1) univariate time-series modelling; (2) multivariate time-series modelling; and (3) comparison of experts' forecasts with those of statistical models. Following a summary of the papers, some suggestions for future research are offered.