In recent years, considerable attention has been given to a method of calculating rates-of-return to education that was developed in the USA by Mincer (1974) and which makes use of what has become known as a Mincerian equation. This approach does not include any specific reference to direct educational costs, although it does incorporate earnings forgone which are a high proportion of total costs.
Mincer suggested setting up a multiple regression equation of the form:
ln Y=a+bS+cX1+dX2+eX3+....
where the dependent variable = the natural logarithm (in) of individual earnings (Y) [where a variable increases by progressively larger proportions, using the natural logarithm is simply a device for being able to translate these increases into equal, or nearly equal, steps]. The independent variables are:
S = years of schooling
X1 = training
X2 = experience
X3 = weeks worked etc.
Such an equation can be presented in a number of different forms, including the parabolic where additional terms are included for one or more independent variables squared. The equation can relate to a group of workers for a particular time period, for example, Mincer's original formulation related to 1959 annual earnings of white, nonfarm, men in the USA.
The partial coefficient (b) of years of schooling (S) gives an estimate of the average rate of return to schooling. In the simplest form of the equation, the coefficient gives this return directly (Psacharopoulos and Alam, 1991). In more complex forms, it is arrived at via a mathematical adjustment e.g. Tannen (1991) took "the antilog of the schooling coefficient minus one". Other writers often do not explain the mathematical adjustment they have made (e.g. Al-Qudsi, 1989).
This approach to calculating rates-of-return to education may be contrasted with the full cost-benefit approach outlined previously which is sometimes termed the "elaborate" method; a third approach is the "short-cut" method which "amounts to doing in an explicit way what the earnings function method is doing explicitly, i.e. the returns to education are estimated on the basis of a simple formula" (Psacharopoulos, 1981).
Depending on data availability, the Mincer approach may be relatively quick and easy to compute, with the regression equation being readily produced by a standard computer software package. The equation picks out the effect of S (years of schooling) on Y (incomes) but does not include costs at all and therefore can not be termed a cost-benefit analysis as such. Nevertheless, when researchers have used both the "elaborate" and "short-cut" methods to estimate rates-of-return and compared the results, these are often remarkably close (e.g. Tan and Paqueo, 1989).
The obvious advantage of the Mincer approach is that it is quick and easy to use, assuming only that a suitable computer programme is available. The major disadvantage is that this approach is applied to data for broad aggregates, often for the whole of education, and thus does not provide results that are readily implementable at the micro level.
