The Effects on Forecasting of Autocorrelated Disturbance Terms and Trended Independent Variables

Schieren, George A.; Carr, Richard P.

doi:10.2307/1058658

Cited by 3 publications

(2 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In a recent article on forecasting when disturbances are autocorrelated, Schieren and Carr ( 1982) suggest that it is not necessary to correct for autocorrelation since they found little increase in efficiency over LS. Their results were obtained with the assumption that p was known.…”

Section: Comparison To Previous Workmentioning

confidence: 99%

Regression forecasts when disturbances are autocorrelated

Dielman

1985

Journal of Forecasting

View full text Add to dashboard Cite

A Monte Carlo simulation is used to study the quality of forecasts obtained from regression models with various degrees of autocorrelation present in the disturbances. The methods used to estimate the model parameters include least squares, full maximum likelihood, Prais-Winsten, Cochrane-Orcutt and Bayesian estimation. Results indicate that the Cochrane-Orcutt method should be avoided. The full maximum likelihood, Prais-Winsten and Bayesian methods are relatively more efficient than least squares when the degree of autocorrelation is high (greater than or equal to 0.5) and show little efficiency loss when the degree is low. These results hold for both trended and untrended independent variables. KEY WORDS Cochrane-Orcutt estimation Prais-Winsten estimation Least-squares estimation Bayesian estimation First-order autocorrelation Several estimators have been suggested for regression models with autocorrelated errors. Two step or iterative approaches are intended to provide relatively more efficient estimators of the regression coefficients than least squares. Previous simulation studies suggest that gains in efficiency will occur even in very small samples if the methods are used with discretion. This cautionary note applies t o those estimation methods which omit the first observation from the time series in transforming the data to remove autocorrelation. When the initial observation is dropped, a loss of information occurs which results in significantly inferior parameter estimates under certain conditions. The obvious alternative to prevent such a loss is to transform the data so that the initial observation is retained. Although several studies compared the relative efficiency of estimators in small samples, little work has been done to study the effects on forecasts ofcorrecting for autocorrelation. Goldberger's (1962) seminal paper o n forecasting with autocorrelated errors suggested using the value of the autocorrelation coefficient in order to increase the forecast efficiency. He obtained asymptotic or large sample results assuming that the true value of the autocorrelation coefficient was known.The examination of' possible improvements in forecasting due to corrections for autocorrelation is a n issue that merits further investigation. Although the precision of the forecasts dqpends upon the precision ofthe parameter estimates, other factors also affect forecast accuracy. One of these is 0277-6693/85/030263-09$05 .OO

show abstract

Section: Comparison To Previous Workmentioning

confidence: 99%

Regression forecasts when disturbances are autocorrelated

Dielman

1985

Journal of Forecasting

View full text Add to dashboard Cite

show abstract

“…* The paper is mainly concerned with the effect of exogenous variables on estimation efficiency. For the effect on forecasting accuracy, see Schieren and Carr (1982), which reveals that forecasting accuracy is also affected by the nature of exogenous variables. *My colleague, Shirley Cassing, informs me that the ranking of alternative tests for serial correlation also appears to be altered.…”

Section: Introductionmentioning

confidence: 99%

Fatal shortcomings of stationary AR (1) exogeneous variables in econometric monte carlo tudies of estimators

Maeshiro

1987

Journal of Statistical Computation and Simulation

View full text Add to dashboard Cite

To cite this article: Asatoshi Maeshiro (1987) Fatal shortcomings of stationary AR (1) exogeneous variables in econometric monte carlo tudies of estimators, Journal of Statistical Computation and Simulation, 28:3, The paper first shows that the stationary normal AR(1) process (SNARl), the most frequently used process for generating exogenous variables in econometric Monte Carlo studies, cannot generate realistic exogenous variables, which are generally trended and similar to those generated by ARIMA (p,d,q) process with d B 1 and positive drift (trend). Then, it illustrates that in the context of AR(1) disturbances, trends in exogenous variables can frequently alter the very ranking of two competing estimators, the ordinary least squares estimator (OLS) and the Cochrane-Orcutt estimators (CO). For three common econometric models-a standard regression model, a dynamic model (i.e., a model with a lagged dependent variable), and a seemingly unrelated regression model, OLS becomes superior in many cases. This is so in spite of the fact that the CO estimator in the study utilizes the true value of the first-order autocorrelation coefficient of the disturbances. The message to be derived from these findings should be clear. If one accepts the fact that most if not all economic time series are trended, and endorses a proposition that the fundamental if not sole purpose of Monte Carlo studies in econometrics should be to provide useful guidelines to practicing econometricians, then, he must not employ SNARl (nor any other artificially created nontrended series) as a generator of exogenous variables in a Monte Carlo study, at least in the econometrics of autocorrelated disturbances. A. MAESHIROAlternative methods of generating stochastic exogenous variables that are trended are suggested in the paper.For almost four decades, the principle of the autoregressive transformation of a regression model with first-order autocorrelated disturbances (the CO estimation priciple) has been taken for granted as a method of correcting for the autocorrelation in the disturbances-be it in the two-stage Cochrane-Orcutt estimator, the iterative Cochrane-Orcutt estimator, or an estimator utilizing nonlinear techniques or search procedures. (Omitting the first observation due to transformation is not considered very crucial in general.) The results of the pertinent Monte Carlo studies appear to justify such a procedure only because most studies have employed SNARl exogenous variables, not trended ones. Thus, Monte Carlo experimenters must be blamed, at least partially, for this prevailining malpractice. It is hoped that they will not commit additional sins by not using realistic data in their future experiments.

show abstract

Linear regression forecasting in the presence of ar(1) disturbances

Latif

King

1993

Journal of Forecasting

View full text Add to dashboard Cite

This paper is concerned with time-series forecasting based on the linear regression model in the presence of AR(1) disturbances. The standard approach is to estimate the AR(1) parameter, p, and then construct forecasts assuming the estimated value is the true value. We introduce a new approach which can be viewed as a weighted average of predictions assuming different values of p. The weights are proportional to the marginal likelihood of p. A Monte Carlo experiment was conducted to compare the new method with five more conventional predictors. Its results suggest that the new approach has a distinct edge over existing procedures.

show abstract

The Effects on Forecasting of Autocorrelated Disturbance Terms and Trended Independent Variables

Cited by 3 publications

References 4 publications

Regression forecasts when disturbances are autocorrelated

Regression forecasts when disturbances are autocorrelated

Fatal shortcomings of stationary AR (1) exogeneous variables in econometric monte carlo tudies of estimators

Linear regression forecasting in the presence of ar(1) disturbances

Contact Info

Product

Resources

About