2014Asian Journal of Development MattersRequires access

Modeling and Forecasting GDP of India: ARIMA Model

M. Sriram, B. Venkatraja

Open publisher page 0 citations

Abstract

The present study aims at estimating and forecasting of GDP in India. Quarterly time series data on GDP from Q1 of financial year 2004–05 to Q1 of financial year 2014–15 are used for estimation and forecasting. The study begins with Augmented Dickey Fuller unit root test and it is found that the time series data is stationary at second difference but not normally distributed. The series is integrated of order two. Later, GDP is estimated and forecasted by applying ARIMA model. In the present study fourauto regressive models are developed and it is found that ARIMA (5,2,4) is the best fit model in predicting the GDP when compared to ARIMA (5,2,5), ARIMA (5,2,6) and ARIMA (5,2,7). Forecasting GDP using the ARIMA model is carried out for In-sample (Dynamic Forecasting). The study shows that the forecasted values are close to the actual values thus minimising the forecasting error.

About this research paper

What this paper is about

The present study aims at estimating and forecasting of GDP in India. Quarterly time series data on GDP from Q1 of financial year 2004–05 to Q1 of financial year 2014–15 are used for estimation and forecasting. The study begins with Augmented Dickey Fuller unit root test and it is found that the time series data is stationary at second difference but not normally distributed. The series is integrated of order two. Later, GDP is estimated and forecasted by applying ARIMA model. In the present study fourauto regressive models are developed and it is found that ARIMA (5,2,4) is the best fit model in predicting the GDP when compared to ARIMA (5,2,5), ARIMA (5,2,6) and ARIMA (5,2,7). Forecasting GDP using the ARIMA model is carried out for In-sample (Dynamic Forecasting). The study shows that the forecasted values are close to the actual values thus minimising the forecasting error.

Why it matters

A significance statement is not available in the OpenAlex record.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

The present study aims at estimating and forecasting of GDP in India. Quarterly time series data on GDP from Q1 of financial year 2004–05 to Q1 of financial year 2014–15 are used for estimation and forecasting. The study begins with Augmented Dickey Fuller unit root test and it is found that the time series data is stationary at second difference but not normally distributed. The series is integrated of order two. Later, GDP is estimated and forecasted by applying ARIMA model. In the present study fourauto regressive models are developed and it is found that ARIMA (5,2,4) is the best fit model in predicting the GDP when compared to ARIMA (5,2,5), ARIMA (5,2,6) and ARIMA (5,2,7). Forecasting GDP using the ARIMA model is carried out for In-sample (Dynamic Forecasting). The study shows that the forecasted values are close to the actual values thus minimising the forecasting error.

Key concepts: Autoregressive integrated moving average, Box–Jenkins, Econometrics, Unit root, Time series, Statistics, Mathematics, Series (stratigraphy)

Related papers

Back to paper searchBrowse research topicsOriginal source
Modeling and Forecasting GDP of India: ARIMA Model — Research Paper | ScholarLens