Demerits of Least Square Method

(i)  This method is very much rigid in the sense that if any item is added to, or subtracted from the series, it will need a thorough revision of the trend equation to fit a trend line, and find the trend values thereby.

(ii)  In comparision to the other methods of trend determination, the method is bit complicated in as much as it involves many mathematical tabulations, computations, and solutions like those of simultaneous equations.

(iii)  Under this method, we forecast the past and future values basing upon the trend values only, and we do not take note of the seasonal, cyclical and irregular components of the series for the purpose.

(iv)  This method is not suitable for business, and economic data which conform to the growth curves like Gompertz’s curve, Logistic Pearl-Read curve etc.

(v)  It needs great care for the determination of the type of the trend curve to be fitted in viz : linear, parabolic, exponential, or any other more complicated curve. An erratic selection of the type of curve may lead to fallacious conclusions.

(vi)  This method is quite inappropriate for both very short and very long series. It is also unsuitable for a series in which the differences between the successive observations are not found to be constant, or nearly so.