A bottom-up hierarchical forecasting of broadband penetration in India using ARIMA & DES approaches

Abstract

Broadband connectivity has been linked with economic growth and is accelerating human progress. On a
global platform, the penetration of broadband in India is still not convincing. A host of initiatives are
aimed at empowering our society digitally, especially the „Digital India‟ programme of Government of
India. While major research thrust is found in areas of policy and technology issues, broadband usage
through handheld devices, and its likes, sparse empirical evidences are found on the future of broadband
penetration in Indian context. This study aims to plug the under explored area uniquely by predicting the
digital connectivity status in India through broadband medium using bottom-up hierarchical forecasting
approach. Here, separate forecasts are made for individual levels of hierarchy, the wireless and the
wireline mediums, and aggregation of them done finally to arrive at the total forecast of the higher level.
To identify the best fitted model that may be used for prediction, two alternative methods; auto regressive
integrated moving average (ARIMA) and double exponential smoothing (DES) techniques have been
used and hierarchical forecast made with it till 2020. The error estimates indicate that the ARIMA model
fits better than the DES model for both wireless and wireline broadband series. From the study results it
may anticipated that wireless medium would drive broadband connectivity in future. Further, by 2020,
wireless broadband coverage in India is expected to reach 800Mn which means a penetration of 58%. The
hierarchical forecast of total broadband reach is expected to touch 821Mn with 59.4% penetration. Even a
conservative estimation of the same is projected at 646Mn i.e. 47% penetration, which is a massive
growth from 85.7Mn in 2014.