• assistance
• benefits
• instrumental variable
• propensity score
• unemployment

The assessment of the causal effect is not a straightforward task. It needs a specific framework to model what is to be considered as a causal effect in the most clear and unambiguous way. This evaluation is even more complicated when considering socio-economic policy. This happens because very often people can select themselves into treatment and this generate the so-called selection bias which makes impossible to compute the true effect without taking particular adjustments. In particular, I considered both unemployment insurance and assistance schemes in Spain in 2017 and, after having discussed the differences among them, I applied propensity score matching for the first and instrumental variable for the second, as an exercise to see how one could behave in order to try to reduce the selection bias.