Multivariate Time Series Approach Research Articles

Estimating the marginal costs of airport operation using multivariate time series models with correlated error terms

This paper explores insights about marginal airport costs that can be gained by analysing labour inputs and their variability. Based on hourly cost and traffic data for the airport of Helsinki, a multivariate time series approach is used incorporating correlated error terms to account for random shocks such as delays. We found for most airport services a linear relationship between labour input and aircraft movements, except for producing passenger services for international departing flights where a cubic cost relationship was estimated. The findings are comparable with earlier studies for US airports.

A Multivariate Time Series Approach to Study the Interdependence among O3, NO x , and VOCs in Ambient Urban Atmosphere

A multivariate time series approach vector autoregression (VAR) along with impulse response function and variance decomposition technique has been employed to look into the interrelationship among O3, NO, NO2, and volatile organic compounds (VOCs, namely, benzene, ethylbenzene, toluene, and xylene in the present study) using 3 months long continuous time series data of 1 h average concentration of these pollutants at one of the traffic sites in Delhi, India. It is found that the VAR of order 2 (i.e., past two lagged values of 1 h interval) is sufficient to represent the observed time series at the station studied. The impulse response function and variance decomposition analysis indicate that O3 concentration shows an immediate rise and persists for a longer duration (typically 8–10 h) once the impulse of NO2, benzene, ethylbenzene, or xylene is given in the ambient environment. However, in case of toluene, the reverse effect has been observed. Since O3 forms in the troposphere due to photolysis of NO2, it is not surprising that its impulse triggers O3 formation in the ambient environment. However, in case of VOCs, this has been attributed to their tendency to show higher inclination toward intermediary reactions leading to the formation of O3 rather than their (VOCs) reaction with O3. Among VOCs, only toluene has been observed to show higher inclination toward its reaction with O3. Apart from this, variance decomposition technique also reveals that the relation of NO with NO2 is more important than the relation of NO with O3 creating a conducive atmosphere for O3 formation in the present scenario. Thus, the multivariate time series approach provides significant insight about the role played by the dominant individual VOCs and NOx in influencing the O3 concentration in ambient urban atmosphere whereas a photochemical modeling approach gives an overall view of NOx and VOCs behavior with respect to O3 by using the O3 isopleth technique.

A multivariate time series approach to modeling and forecasting demand in the emergency department

Study objective The goals of this investigation were to study the temporal relationships between the demands for key resources in the emergency department (ED) and the inpatient hospital, and to develop multivariate forecasting models. Methods Hourly data were collected from three diverse hospitals for the year 2006. Descriptive analysis and model fitting were carried out using graphical and multivariate time series methods. Multivariate models were compared to a univariate benchmark model in terms of their ability to provide out-of-sample forecasts of ED census and the demands for diagnostic resources. Results Descriptive analyses revealed little temporal interaction between the demand for inpatient resources and the demand for ED resources at the facilities considered. Multivariate models provided more accurate forecasts of ED census and of the demands for diagnostic resources. Conclusion Our results suggest that multivariate time series models can be used to reliably forecast ED patient census; however, forecasts of the demands for diagnostic resources were not sufficiently reliable to be useful in the clinical setting.

A systems approach to analysing sub‐state conflicts

PurposeThe purpose of this paper is to provide a more holistic approach to analysing the impact of all the behaviour of a conflict's participants its overall dynamics, using the example of the Northern Irish troubles.Design/methodology/approachA novel multivariate time series approach developed by Professor Paul Smoker is presented which can map the dynamics of this conflict and its causal inferences as a series or “systemograms”.FindingsThe case example reveals high levels of autocorrelation in the variety of techniques used by the state security authorities to suppress terrorism, indicating their strong role in maintaining this conflict. When more than one party exhibits such behaviour, the conflict “locks in”.Research limitations/implicationsThe work remains preliminary and historical. Data was collected on a month‐by‐month basis which suggests associated rather than direct causal influence. It would be useful to further explore these findings using data from similar conflicts.Practical implicationsSuggests that some counter‐terrorism approaches may be dysfunctional especially those adopting sub‐lethal weapons. Provides some insight into behavioural changes required to prevent conflict destabilisation.Originality/valueProvides a novel conflict research methodology which allows the strong structural dynamics of the conflict to be seen – much the way that elapsed time photography enables hidden processes to be revealed. The raw statistics are presented here.

Poverty Reduction: Government Transfer Spending vs. Macroeconomic Change

ABSTRACT The study uses a multivariate time series approach (Vector Autoregression) to compare the relative effects of government transfer spending versus macroeconomic change on poverty reduction. The analysis disaggregates transfer spending among federal budget subcategories and poverty rates among different demographic groups. The analysis also tested for the possibility that spending on government programs is adjusted in response to changes in poverty. The findings indicated that out of five separate transfer spending categories, only Income Security and Social Security spending had significant impacts, affecting White and elderly poverty only. Black, youth, and female head-of-house poverty rates were not significantly influenced by any spending category. Economic growth and lower unemployment rates were both observed to reduce poverty in all categories except the elderly, but the unemployment rate appeared to be the dominant effect. We found only a single instance in which transfer spending (Income Security) appeared to respond to changes in poverty (youth poverty).

Effects of Public and Private R&D on Private-Sector Performance in the United States

This study analyzes the effects of public and private research and development (R&D) on private-sector output, employment, and investment using a multivariate time-series approach. This approach follows the conceptual argument that dynamic feedbacks are essential to understanding the relationship between public-funded R&D and private-sector performance. With this approach, the authors are able to measure the total effects of public and private R&D on private output. The results suggest that in the long term, public R&D does not affect employment but crowds in private investment, whereas private R&D is a substitute for both. As a consequence, the long-term effects of public R&D on output are positive and substantial, whereas the effects of private R&D are at best marginally positive. As a corollary, the authors argue that public R&D should be brought into the limelight as a leading candidate for explaining the growth slowdown in the past decades.

Population Growth and the Dynamics of Canadian Development: A Multivariate Time Series Approach

New long-run GNP estimates for Canada are used to apply time-series methods to the issue of the impact of the rapid settlement of western Canada and the wheat boom from 1896 to 1913 on Canadian development. Using annual data for 1870 to 1939, a vector autoregression is estimated and used to analyze the contributions of innovations to the GNP growth over the wheat boom period. The most striking result is the substantial impact of innovations in population. Exports also played an important role and the investment boom that occurred is seen as arising endogenously from the dynamics of the model. Comparing the results with those of Chambers and Gordon (1966,Journal of Political Economy74,315–332), the estimated impact on per capita income is much larger and highlights the importance of economies of scale, which are neglected in their analysis.

Export Growth and Economic Development in Colonial British America

This paper analyzes the dynamic relationship between tobacco and the rest of the economy in the colonial Chesapeake to test the staple theory version of the export‐led growth hypothesis. The paper adopts a multivariate time‐series approach which accommodates the presence of cointegration and contemporaneous correlations among innovations. The empirical evidence strongly supports the staple theory. British demand for tobacco is statistically exogenous and found to induce cyclical fluctuations in the colonial price of tobacco as well as temporary over‐ or underproduction, as measured by a long‐term cointegration relationship between the British demand for tobacco and the colonial price of land.

Foreign direct investment and exports: a time series approach

The impact of the accelerated internationalization of the last decade on the Austrian economy is a controversial issue. Granger's concept of casuality is used to investigate one aspect of the internationalization of production: the realtionship between foreign outward direct investment and exports using aggregate flow data from the Austrian economy. The stationarity of the time series is examined and cointegration tests for the adequacy of the multivariate time series approach are performed. The estimation results suggest significant causality of Austrian foreign outward direct investment and exports in both directions. Impulse response analysis and varience decomposition show a very slow dynamic response of both variables to exogenous shocks of the other. It furthermnore indicates the possiblity of a positive effect of exogeneously increased foreign direct investment on exports and a negative effect of export shocks on foreign direct investment; however, significant long-run effects are not established.

A Dynamic Analysis of Net Migration and State Employment Change

"The dynamic relationship of net migration and employment change is examined for ten selected states of the U.S. using a multivariate time series approach--a vector autoregression (VAR) model. Granger causality tests and dynamic multipliers provide information on the dynamic process. The results suggest a state-level process in which employment change occurs first, and net migration follows with a lag. The procedure appears promising in investigating the timing of net migration and regional employment change."

A multivariate time series approach to modelling macroeconomic sequences

In this paper we discuss a multivariate generalization of autoregressive integrated moving average models. A methodology for constructing multivariate time series models is developed and the derivation of forecasts from such models is considered. A bivariate model for Austrian macroeconomic sequences is constructed. Furthermore it is discussed whether multivariate time series methods can be expected to lead to a significant increase in prediction accuracy when forecasting macroeconomic series.