Analysis of Migration of Advertisement Business from Print Media to Electronic Media of Pakistan: A Time Series A.R.I.M.A Model Application for Forecasting Overall Print Industry’s Advertisement Business Curtailment

Abstract

Couple of decades back, it was generally acceptable that print media has secured future, newspapers and magazines have a good market and good readership in Pakistan. But by the end of 90’s, when electronic media introduced and welcomed by peoples, print industry of Pakistan tends to decline and players of the industry was in shocked and started to take steps to survive in the market. Advertisers are curtailing their advertisement business from print media and allocating more budgets to advertiser in electronic media.

Purpose

The purpose of the study is to highlight the future trend and to forecast the future share of publishers in market’s total advertisement business by using autoregressive integrated moving average (ARIMA) models.

Design/Methodology/Approach

This research is divided into two parts; in part one, data about the print media share in total advertisement business is collected from the head office of APNS. In part two is Box-Jenkins approach of ARIMA modeling was applied on the time series data to forecast.

In this study, framework of ARIMA includes the following steps: Data collection, identification of model and order, making data stationary, estimation of different orders of ARIMA models to find the best appropriate model by comparing their AIC and SIC values, diagnostic test of residuals for white noise and forecasting of in-sample and out sample data.

Findings

The traditional Box-Jenkins methodology was used to forecast the time series named ARIMA modeling. Initially the model which was identified was AR(1), by the spikes of the correlogram and used to build up the model. Correlogram of basic data series showed the existence of unit root in the series but on 1^st level of difference, the data series became stationary. Augmented Dickey Fuller Unit Root test was applied to accept or reject the hypothesis of unit root, and on first differentiation the value of t was -4.83 which is significant and the null hypothesis of unit root existence was rejected and series considered as stationary.

ARIMA (2,1,3) was identified as the best appropriate model with lowest AIC and SIC value among 15 combinations of AR and MA which were estimated in this study.

Y_t = α + Φ₁Y_{t- 1} + Φ₂Y_{t- 2} + θ₁e_t-1 + θ₂e_t-2 + θ₃e_t-3

Y_t = -0.0798 - 0.1184Y_{t- 1} + 0.1012Y_{t- 2} - 1.5085e_t-1 - 26.14e_t-2 - 13.46e_t-3

Equation of ARIMA (2,1,3) shows that the today print media’s advertisement revenue share is dependent on the two leg value of past year’s share value and on the three leg values of past years error which is moving average.

The value of R² is 0.9988 which shows that the model is fit and the weight of auto regressive and moving average legs in total sum of square is around 99%.

Correlogram Q Statistics of residuals, shows that the residuals are white noise and model is pass the suggested diagnostic test.

The forecasting of 2013, which is in sample data shows the same result as per actual data so the prediction power of model is considered as reliable and appropriate. The future prediction which is out of sample prediction shows the 6 percent decline in the share of print media in advertisement revenue over the period of four years which is an alarming situation for the publishers. Readers are switching from printed stuff to online stuff and in Pakistan; people are more likely to watch television therefore advertisers of Pakistan are more interested to advertise in electronic media and they are curtailing their advertisement expense from print media and allocating more budget to electronic media.

Practical Implications

This study will help new comers in this industry of publishing in Pakistan to understand the advertisement revenue trend and also to understand the ARIMA model application. It also helps the existing publishers to make vital decision regarding the revenue mix and to take any grand decision like diversification or merger etc.

Originality/Value

This study is unique; it is quantitative in nature. All the data in this paper is original and secondary data was collected from the head office of All Pakistan Newspaper Society which is the main governing body of print media in Pakistan. This study predicts about the future of the industry because the main factor of profitability of print media which is advertisement income is predicted.