Vocational School Signalling-Screening Benefit

This is a lil bit wonkish.

In a recent Twitter discussion in regards of vocational school trend in Indonesia, with an Indonesian economist @ari_ap, he proposed a standardization mechanism via certification made by government for these vocational school graduates in order to lower the cost of screening for employers. However, I extended his case that it might also lower the cost of signalling for students. Here’s the model:

In status quo, we see that the screening (and subsequent response from students) goes like this:

Here, we have student that has additional qualification Q that is drawn from a pool containing a strictly increasing distribution function F (•) Є {Q, Q} that is continuous. Though only Student Q knows Q, it’s difficult to separate her from the pool since the pool contains all vocational school graduates that passed national exam that’s very general (for instance, she had passed English national exam, but her TOEFL scores will not be known from that,) ergo, additional qualification may not be known. Then we have employer that can choose whether she wants to invest in screening Sc(Q) Є {0, 1} (here, I strictly define screening as an attempt to separate Student Q from Student Q from pool of students, be it via test conducted by that employer, and so forth,) and after that can respond by hiring that student, R(Q) Є {0, 1}. The employer then receives payoff Π_E from profit from employment, P(Q) > 0 with additional human capital benefit H(Q) if student has quality Q, subtracted by cost of screening - C(Sc) < 0.

If the employer chooses not to invest in screening, then it’s responded by whether the Student invest in signalling her additional quality via Sg(Q) Є {0, 1}. If the student choose not to invest in signalling, Sg = 0, then, employer may have probability β of employing Student Q, therefore the payoff is Π_E = P(Q) + βH(Q). If the student invests in signalling, Sg = 1 (assumption: the signal is ALWAYS honest, and ONLY student possessing Q can signal,) she’ll then receive payoff Π_S from benefit of employment B(Q) > 0 subtracted by cost of signalling – C(Sg) < 0.

With a standardization mechanism via certification enacted, the game goes like this:

We can see how a certification allows the pool to separate, therefore employer need not to invest in screening (separating Student Q from the pool), and the student need not to invest in signalling (since the cost of signalling is already paid by government.) We can also see that elimination of – C(Sc) and – C(Sg) from both the employer’s side and student’s side makes the payoff from the second game is more Pareto optimum than the first game/status quo.

Since I’m not yet acquainted with game theory (let alone mastering it) and I don’t have data to do regression analysis, I can’t give the analysis of its multiple equilibria and cost-benefit analysis.

Many thanks to @ari_ap for the insight and Simon Quinn for his brilliant paper (.pdf).