# Signalling Model (How to solve a differential equation?)

I'm currently reading "SAVING RATES AND POVERTY: THE ROLE OF CONSPICUOUS CONSUMPTION AND HUMAN CAPITAL" by Omer Moav and Zvika Neeman, and I encounter some problem deriving one of the equations.

Utility function, where $$y$$ is income, $$\widetilde y$$ is the belief of $$y$$ based on $$x$$ and human capital $$h$$, $$x$$ is the product of interest:

$$u(y,x) = (y-x)^{1-\lambda} \widetilde y(h,x)^\lambda$$

Differentiate with respect to $$x$$:

$$\frac{\lambda}{1-\lambda} \frac{y-x}{\widetilde y(h,x)} = \frac{1}{d \widetilde y (h,x) / d x}$$

This is the confusing part, how do I get this from above knowing that $$y=\widetilde y(h,x)$$?

Given this $$\underline y$$, the lowest possible income an individual of $$h$$ can have ($$\pi$$ is an uncertain component of income that is not observed and has mean zero):

$$\underline y (h) \equiv h + \underline \pi (h)$$

$$\widetilde y (h,x)^{1/(1-\lambda)} - \frac{x}{\lambda} \widetilde y (h,x)^{\lambda / (1-\lambda)} = \underline y (h)^{1/(1-\lambda)}$$

I think it's something to do with utility being indifferent for the marginal person who consume $$x>0$$ and $$x=0$$, but I don't know how to work it out to get the equation shown above.

Thanks!!

• This is not so much a question about signaling, but on how to solve a differential equation. In separating equilibrium, you have $y=\widetilde y$ and then you use $\underline y$ as the boundary. – Bayesian Jul 11 '19 at 9:16
• Thanks for the edit on the question. Yup, I'm not sure how to solve this :( – Estee L Jul 11 '19 at 18:38