A/
0 / y
income is unity/(thus Y = 0), and the reciprocal rate of
’eturn is therefore given by W. Ult shoulcj/be noted that
a doublepdLded Laplace transform £ould i^dt help us, because
it codld never converge at the empirical values of ©c /•
Ttie density of the rate of return^will thus be truncated.
In~the case of dependence the ./truncation - will be worse,
if \k<1, because the function f* (W) will be stretched
by a factor 1/k.
There is, however,a deeper reason for the restriction:
The validity of the Pareto law cannot be assumed for low
\
values of W (in fact,for negative values, if we put the
unit of wealth at a level which limits the range of
linearity of the distribution). Owing to the irregular
ity of the wealth distribution for low values of wealth
the possibility of very high rates of return might disturb
the regularity of the pattern of incojne distribution. We
mudb, therefore, set a limit to the perkltted rate of
return.
In reality the return rates which have been truncated in
our exercise may, however, exist. The point is, th^n, that
to the extent that they exist - and that will be the x more
likely the smaller k is - the income distribution will \£e
less Regular than the wealth distribution. In practice
this will mean that the range dominated by the Pareto law
will be narrower for income than for wealth.