A research seminar presented by Mr Olivier Scaillet from the University of Geneva was held on 26 of November 2018. The lecture took place at LSF à 5 PM and was presented in English.
The seminar was organised by the Luxembourg School of Finance, the Luxembourg Statistical Society and the Luxembourg Association of Wealth Managers.
We develop a simple approach for estimating the entire distribution of skill across mutual funds. Our approach is non-parametric–as such, it avoids the challenge of correctly specifying the skill distribution. It also allows for a joint analysis of multiple measures– a key requirement for examining skill. Our empirical analysis reveals that most funds are skilled at detecting profitable trades, but unskilled at resisting capacity constraints. These two skill dimensions exhibit strong heterogeneity both within and across fund groups. In addition, they are strongly correlated. Aggregating them using the value added reveals that 75% of the funds earn profits for a total of 7.8 mio. per year on average.
July 10th 2018, a conference on Electoral Surveys was held at 11:45 am at Cercle Cité. The conference examined statistical basis behind electoral surveys, standards and practices, as well as presentation of results. A round table was also held on the effects of electoral surveys.
On march 21st 2018, Claude Lamboray , Head of Unit “Price statistics”, STATEC held a talk on Scanner Data and Indices of Consumer Prices.
16/03/2018: Royal Belgian Statistical Society
Location: Brussels (Belgium)
On March 16, three board members of the LSS met the President of the Royal Belgian Statistical Society, Prof. F. Thomas Bruss, in Brussels. During this meeting, potential future collaborations between the two societies have been discussed, and the LSS has been invited to attend the Annual Meeting of the RBSS in October 2018.)
On March 8th 2018, a lecture was given by Angelo Koudou from the Université de Lorraine.
Talk was hold at LISER at 18:00 in English language.
This talk was divided into two parts. We began by showing some examples of use of the generalized inverse Gaussian (GIG) distribution in statistics. The second part of the talk discussed the (unsolved) problem of a possible use of the so-called Stein's method in deriving rates of convergence to the GIG distribution.