Forecasting the sales of combines and tractors in today's ever changing and volatile marketplace is a difficult task; moreover, it is becoming increasingly important to manufacturers whose goal is to maximize profits with just-in-time policies. Farmers are independent business managers who react to a variety of factors and a unique set of circumstances. Predicting how these individuals will respond to changes in the economic atmosphere is an even more perplexing task than forecasting their business environment. Individually, some farmers may choose to postpone machinery purchases when their cashflow is reduced. Collectively, farmers react to certain factors in a more cohesive manner. For example, if farm commodity prices rise and interest rates are relaxed the agricultural sector will undoubtably increase machinery purchases. Therefore, Federal government policies and also fiscal and monetary policies affect both the farmer's income and the associated agricultural industries sales.
Manufacturers attempt to interpret the economic and political situation and plan production schedules accordingly. Forecasting the sales of their products is essential in controlling the inventory of products, parts, and supplies. Manufacturers must access the trade-offs of keeping adequate inventory versus the escalating costs associated with holding this inventory. This business risk is especially acute in such tight financial times as the present. Funds needed to hold inventory for sustained periods of time detract from the increasingly important research funds needed to develop new products. This could allow a competitor to have new products ready when the market turns around. Furthermore, over estimation of inventory needs in a down market becomes a financial burden to manufacturers instead of a necessity. Labor requirements, material orders, production schedules, and distribution requirements all relate to this inventory control. Thus, forecasting sales with greater accuracy should be a primary goal of marketing management.
The use of statistical techniques such as econometric models and time series- autoregressive integrated moving average (ARIMA)-processes produce a forecast that managers can utilize in forming corporate predictions. These statistical procedures are designed to forecast fluctuations more accurately than mere extrapolation of data and straight-line techniques which can only react to changes in the market after the trend has been started. An econometric model is a single equation regression technique that uses independent variables such as commodity prices, indexes, buyer intentions, interest rates, etc. These variables are tested at various lead times to find the ones that correlate the best with past performance, resulting in a statistically significant predictive equation to estimate future performance. Most time series processor programs provide statistical tests that include r2 standard error, Durbin-Watson statistic, student t tests, F tests, etc., that are used to evaluate confidence limits of the forecasting equations. Models can be developed for different predictive time periods by using the leading indicators that have the highest significance levels at that particular lead time. Hence, an econometric model utilizes independent leading indicators in a single equation that can estimate trends that are cyclical and ever changing in nature.