Christos Agiakloglou

Adventures with ARIMA software

Abstract Many software packages are available for fitting autoregressive inegrated moving average models to time-series data. The practitioner is faced with a wide choice of programs for fitting these models. As we illustrate in this paper, the results obtained can depend substantially in important respects on the particular choice that is made. Our examples serve as a basis for exploring the causes of this phenomenon.

The balance between size and power in Dickey-Fuller tests with data-dependent rules for the choice of truncation lag

Abstract The lag in Dickey-Fuller regressions may be chosen through order selection criteria or general-to-specific testing. We examine the trade-offs between size distortion and power that are implied by this approach to testing for unit roots.

Looking for Evolving Growth Rates and Cycles in British Industrial Production, 1700-1913

An alternative analysis of an index of British industrial production for the years 1700–1913 is provided. The focus of attention is the components structure of this time series. Specifically, evidence of evolving growth rates and cycles is sought. The methodology employed is the standard autoregressive integrated moving average model building approach. In contrast with previous findings based on an alternative methodology, we find strong evidence against the hypothesis of evolving growth rates and little support for cycles of any substance.

Evidence of ARCH(1) Errors in the Context of Spurious Regressions

The spurious regression phenomenon is related to first-order serially correlated errors. This study, using a Monte Carlo analysis, finds that this phenomenon is also related to ARCH(1) type errors.

A spatial and economic analysis for telecommunications: Evidence from the European Union

This paper evaluates the role of a number of determinants of telecommunication services in the European Union. We use a logistic model with spatial covariates to estimate the demand function for telecommunications in the Union. Our results show that different types of interconnections generate diverse estimates for country specific demand. The impact on telecommunications from countries with spatial, economic or social similarities differs based on those characteristics. Omitted variable bias from not modeling spatial interdependence is limited in models under spatial connectivity criteria. This satisfies the statistical inference drawn by previous empirical studies regarding determinants of telecommunications.

An alternative approach for testing for linear association for two independent stationary AR(1) processes

Spurious correlations occur when two independent time series are found to be correlated according to the typical statistical procedure for testing the null hypothesis of zero correlation in the population. Using a Monte Carlo analysis, this study examines the spurious correlation phenomenon for two independent stationary AR(1) processes and it finds that if an alternative testing procedure is applied, spurious behaviour is eliminated using the variance of the sample correlation coefficient of these two series, suggested by Bartlett (1935).

Comparing estimates of risk between markets and telecommunications institutions in Europe

We examine and evaluate the concept of risk for the financial market of telecommunications in Europe using the Value-at-Risk (VaR) method. In particular, we compare the estimates of risk between stock market indices and stock prices of telecommunications institutions in Europe. The estimates of risk are obtained as a one-step-ahead forecast using AutoRegressive Integrated Moving Average (ARIMA) analysis with Generalized AutoRegressive Conditional Heteroscedastic (GARCH) errors.

U.S. Common stock prices, 1871–1970: playing with dummies

The received wisdom that the levels of many economic time series are generated by processes with a unit autoregressive root has been called into question by recent work of Perron. When break points, or interventions, in the time series are allowed it emerges that the unit roots hypothesis can often be rejected at quite low significance levels. Taking for illustration a single time series, U.S. common stock prices, we demonstrate that Perrons conclusions are very sensitive to the choice of break point, and that the data contain little support for the particular choice imposed by Perron.

Is spurious behaviour an issue for two independent stationary spatial autoregressive SAR(1) processes

Spurious regression occurs when two independent stationary or nonstationary time series are found to be correlated. Spurious behaviour is also detected in spatial data. Using a Monte Carlo analysis, this study examines the spurious phenomenon for two independent stationary spatial autoregressive processes of order one, that is, SAR(1), and it finds that when spatial econometric models are estimated, as suggested by the LM specification tests, the spurious behaviour is not detected nor the presence of spatially autocorrelated errors.

The balance between size and power in testing for linear association for two stationary AR(1) processes

Abstract The classical statistical procedure in testing the null hypothesis of zero correlation for two independent stationary AR(1) processes produces spurious correlations, contrast to the alternative testing approach that has been proposed by Agiakloglou and Tsimpanos (2012). This study examines the trade-offs between size distortions and power using both testing techniques, including the case where the true values of the autoregressive parameters are replaced by their estimates.