Loann Desboulets*, Anna Belianska**
Anushka Chawla : anushka.chawla[at]univ-amu.fr
Laura Sénécal : laura.senecal[at]univ-amu.fr
Carolina Ulloa Suarez : carolina.ulloa-suarez[at]univ-amu.fr
*Sparse Graphical Models are often used in Finance to assess risk. Here I show how to use Non-parametric Variable Selection on manifolds to build Non-parametric Sparse Graphical Models. The advantage in the manifold representation is that it accounts for unobserved non-linear dependences, that can be missed by standard estimation techniques. In that case, the resulting graphs are incomplete and the risk is wrongly assessed. The proposed methodology is then illustrated via an empirical application to the Commodity Futures Market.
**This paper investigates the impact of uncertainty on the skill premium, which is defined as the ratio of the wages of skilled to unskilled workers. First, I show empirically in a structural VAR model that uncertainty shocks are recessionary and lead to a reduction in the skill premium. Then, I propose a dynamic New Keynesian model consistent with these facts. This model highlights the importance of distinguishing different roles of skilled and unskilled labor in production by the means of capital-skill complementarity. The model accounts for the additional transmission channel of uncertainty, which explains its effects on the wage divergence between skilled and unskilled labor. The uncertainty shock is contractionary and pushes the demand for labor and capital inputs down. The reduction in capital investment leads to larger wage decreases for skilled labor than for unskilled labor since skilled workers are complementary to capital.