Forecasting India's Fertilizer Volumes – Horizon 2030

Consumption volumes in the Indian fertilizer industry are expected to rise at a 2.5% CAGR, from 61m metric tons in 2019 to about 78m metric tons in 2030 (see Figure 1). The fertilizer product portfolio will shift toward complexes and DAP, with the highest growth rates coming from complexes, followed by DAP (see Tables 1 and 2). No growth will be seen for SSP.

The primary premise of time-series forecasting is that the patterns that emerge in the past will also materialize appreciably in the future. In recent years, Indian agriculture has been characterized by stable cropping areas and steady improvements in agriculture infrastructure and cropping practices. Also, the seasonality of agriculture trends in Kharif and Rabi are effectively captured through statistical time-series forecasting models, which have been deployed at monthly intervals. Unexpected events would also be captured by time-series forecasting models. For instance, weather disruptions are a fundamental part of Indian agriculture, and though disruptions are sporadic and in different locations every year, time-series forecasting models will appreciably account for these in the past trends, at an aggregated nationwide level. Also, fertilizer volumes data is captured through government infrastructure as a part of the subsidy regime, assuring its credibility. This makes the fertilizer sector amenable to time-series forecasting.

It is also necessary to establish the circumstances in which the forecasts may become inappropriate for use. This would only happen in the case of a complete disruption of agricultural trends within the forecasting horizon of 2030. The advent of new technology in agriculture could be a potential cause (e.g. biotechnology). Also, a remarkable change in government policies (e.g. subsidies), which have been reasonably consistent since 2013, could disrupt fertilizer consumption patterns. Also, weather or climatic changes that affect the entire Indian agriculture industry might be considered a potential disruptor.

Indian agriculture has been fairly resilient to the situation imposed by Covid-19. Still, it might cause short-term distortions in consumption patterns. For example, precautionary purchases of farm inputs, which would not have been captured in the previous year’s patterns, would not get automatically captured in the time-series forecasting model at a monthly level. However, these forecasts would make considerable sense at an aggregated quarterly, semi-annual, or annual level.

Forecasting Model Details

For Statistically Curious Eyes