The main goal of the present work is to capture the informational effect in the Brazilian short-term interest rate (Selic rate) using Poisson jumps. This structure is sustained by tests realized by Johannes (2004) and Das (2002), which provide strong evidence of this kind of modeling for FOMC3 announcements. In the above-mentioned articles, the authors show that a big part of the short-term volatility observed in fixed income markets is captured with the introduction of jumps in the stochastic process of the interest rate. The fixed income markets in Brazil, despite of the progressive investment profile changes (in direction of longer maturities), remains highly oriented for short term government bonds, like LTN’s. There are, nevertheless, few academic works estimating the informational effects in the Brazilian short-term interest rate, such as Margueron (2006) and with simplified approaches for this issue4. The capture of the informational effect with stochastic processes might generate improvements not only in understanding the effect per se, but also in explaining the abrupt movements in shorter maturities of the yield curve. As affine models for the yield curve, they are based in the idea of the whole curve as a function of the short-term interest rate, so it is possible to extend this approach, improving the yield curve modeling as a whole. Formalizing this structure potentially generates advances in pricing fixed income instruments and risk management.