13 
In reality, profits in excess of 1oo % any exist. In 
the case whore k<^1 ve can,however, always ensure 
that the conditions are net and (4) remains applicables 
As before, we rotate the diagonal round a point until 
it is parallel to the regression line (see fig* 3 or 4), 
and along that line the distribution is truncated. If 
we choose the point appropriately, we can reduce the 
number of incomes truncated. In algebraic terns, we 
fiat V 1 •» w-V 0 , and we write (4) in the following 
form* 
q (X) « cdX J ^*(hl/'+ V 0 ~T) e^dV' . 
-|^(|)exp <T*%)J, (3) 
1:1/' > ( x+U b ) 
She resulting incone distributism differs only by a 
constant factor from (4). 3inc© the transformation (5), 
however, is only applicable to wealth not smaller than 
¥' 0 , it can be Intuitively seen that the income distrib 
ution will start only at the level X » . o,to confirm to 
the Pareto low. In other words, the Bare m will 
always be projected from wealth on to incone, but de~ 
pan< >n the shape of the regression line, the incone 
distribution will confirm to linearity only from a more