2.3.3. CAViaR modelEngle and Manganelli (2004) proposed a conditional autore-gressive specification for VaR. This approach is based on a quantileestimation. Instead of modelling the entire distribution, they pro-posed directly modelling the quantile. The empirical fact that thevolatilities of stock market returns cluster over time may be trans-lated quantitatively in that their distribution is autocorrelated.Consequently, the VaR, which is tightly linked to the standarddeviation of the distribution, must exhibit similar behaviour. Anatural way to formalise this characteristic is to use some typeof autoregressive specification. Thus, they proposed a conditionalautoregressive quantile specification that they called the CAViaRmodel.Let rtbe a vector of time t observable financial return and ˇ˛a p-vector of unknown parameters. Finally, let VaRt(ˇ) ≡ VaRt(rt−1, ˇ˛)be the ˛ quantile of the distribution of the portfolio return formed attime t − 1, where we suppress the ˛ subscript from ˇ˛for notationalconvenience.A generic CAViaR specification might be the following