 My name is Terri van der Schwan, and I am a fourth year PhD student at Erasmus University Rotterdam. I am introducing my work here, Multiple Impulse Response Functions. Impulse Response Functions are an economic concept to assess the dynamic effect of economic shocks over time. Traditional impulse response functions require assumptions on the underlying structure of the shocks, and the generalized impulse response functions are used when the identifying relations are not clear. However, sometimes multiple shocks trigger each other within a short period of time. So therefore we introduce multiple shock impulse response functions, which take into account the correlation between these shocks. So with a simple VAR example, I show that this concept is necessary in order to estimate the combined effect of these shocks accurately. Hi, my name is Elias Wolf, and I am a PhD student at the Freie Universität in Berlin. Today I am presenting my project Estimating Growth at Risk with Skewed Stochastic Volatility Models. For a long time, economists have observed that economic recessions tend to be more volatile and more severe than economic expansions. And so the aim of my paper is to build a parametric model that can capture these asymmetries and GDP growth rates. A very popular paper amongst academics and policymakers that has been published recently links these asymmetries to the national financial conditions of a country. And so what I do with the model is that I want to estimate the impact that national financial conditions have on the mean, the volatility and the skewness of GDP growth rates conditional of the national financial conditions. What I find is that a deterioration of national financial conditions severely impacts negatively the mean and the skewness of the conditional distribution while the volatility increases. What I also find is that especially the asymmetry in the distribution, so the skewness, matters especially in recessionary periods to accurately predict future negative views GDP growth rates. Hi, my name is Daan Obschor and I am a fourth year PhD student at the Erasmus University in Rotterdam. Today I will be presenting my work on slow expectation maximization convergence in low-noise dynamic factor models. A commonly used method in forecasting macroeconomic time series is the dynamic factor model, which nicely summarizes the covariation in a large number of time series. However, our work shows that the convergence of this algorithm performs rather poorly in the case of low measurement noise, which is a setting that arises in specific now-casting models. To solve this issue, we propose to either use an adaptive version of the expectation maximization algorithm or to artificially increase the noise level. We show in our empirical exercise that we can substantially improve the now-casting performance of euro area GDP growth and therefore we suggest to use this method more commonly in macroeconomic forecasting.