Early stage intervention

In David Willett's brilliant book he contests the widely-held view that education investment benefits are highest at an early age. He mentions the benefits depend a lot on, for example, a murder committed by an out-of-scheme male participant of a young person, this cost being substantial relative to an older person (since a young person has more income years to lose).

Here's Heckman's figure 1 of this paper "The Life Cycle Benefits of an Influential Early Childhood Program", https://hceconomics.uchicago.edu/research/working-paper/life-cycle-benefits-influential-early-childhood-program.  It shows that female returns are insignificant and much lower then male returns, much of which seem to be based on crime effects. Later he says
(p.63)

"ABC/CARE has treatment effects on crime for females for a number of
categories (see Appendix G). However, males are much more likely to commit
crimes that are more costly to the victims, to the criminal justice system,
and to society (Cohen and Bowles, 2010; Barak et al., 2015). ABC/CARE
also has treatment effects on crime for females for a number of categories (see
Appendix G). However, males commit crimes that are much more expensive
to society. These two categories are examples of why the magnitudes of the
gains are much higher for males than they are for females."

 I am no expert on this, but these differences seem interesting.