The causal model is so called because it employs the cause-effect relationship between fertilizer demand and the factors affecting it. The model does not depict fertilizer demand over time or for a particular point of time but presents demand in relation to a set of circumstances. While the trend extension method assumes that time reflects all factors, the causal method (also called the regression method) seeks to establish direct relationships between fertilizer demand and factors influencing it. Factors affecting demand, as we have seen earlier, include crop prices, fertilizer prices, credit availability, irrigated area, rainfall, area under high-yielding varieties, crop pattern and distribution arrangements. By analysing past data, two or three critical factors that have the most profound effect can be selected and the effect of the selected factors quantified and expressed in the form of mathematical equations. To project demand for future years, the likely state of each selected critical factor at that point of time has to be first assessed.
This method is extremely complex, involving mathematical equations to express the relationships and inter-relationships between variables. Moreover, reliability is not guaranteed because it depends on forecasts of values of the selected critical factors. At best, it tells us what the demand is likely to be in a given set of circumstances but there is no certainty that those set of circumstances will prevail in the years under forecast.
An indispensable prerequisite for using this method is the availability of data, by area and season, on various critical factors (amount of credit, fertilizer prices, crop prices, area under high-yielding varieties, rainfall, etc.) in addition to data on fertilizer demand for several years. As mentioned, a serious limitation of the causal method is that the forecast is, in turn, dependent on indicators that must themselves be forecast. Isolating the influence of each factor on demand is a complex statistical process involving expert intervention and use of computers. While the underlying principle of using causative factors to predict demand is logical, its application is not only complex but requires elaborate data for the past which is rarely available. For these various reasons, this method is not of practical use either to governments or to marketing organisations.
