This paper fits in the discussion in our course since it shows the importance of taking into account the effects on treated and untreated individuals. Previous research \citep{Kluve_2010} \citep{Card_2010} wrongly thought that the treatment was increasing the pool of jobs offered, and hence matching treated and not treated in the same market was a reasonable estimation of the treatment effect. However, since they did not consider possible externalities, these conclusions could be due to making the untreated worse off. In their findings,  they show that when someone computes the net effect taking the externalities into account, the hypothesis of the policy not having an impact at all cannot be rejected as can be seen in the following graph: