Keynote Speakers

2022-04-26 11:41

Dr. Sebastiano Vitali, Ph.D. (University of Bergamo)

Title of the presentation: Stochastic optimization and stochastic dominance in Asset and Liability Management models.


Asset and Liability Management problems are typically subjected to multiple sources of uncertainty, and they require to take a sequence of decision through multiple stages over a very long horizon. Therefore, multistage stochastic optimization appears a natural tool to tackle them. Pension fund problems belong to this category. Moreover, the pension fund manager must guarantee the long-term sustainability of the pension fund and should be able to challenge some reference benchmark. For this reason, stochastic dominance has been recently added in many pension fund optimization models to ensure that the optimal solution is better than the benchmark for a very large set of economic agents. In this work, we discuss various types of multistage stochastic dominance within a pension fund problem. We show the advantages for a pension fund manager to implement stochastic dominance constraints either on single stages, or on multiple stages disjointly, or on multiple stages jointly. Numerical results show the dissimilarities between the different ways to interpret and apply the multistage stochastic dominance.


Ing. Jan Brůha, Ph.D. (Economic Research Director, Czech National Bank)

Title of the presentation: Labour Market Composite Indexes: Formulation, Estimation, and Implications


Central banks closely monitor labour-market data as indicators of the cyclical position of the economy and of demand-driven inflation pressures. As individual time series may sometime send a conflicting message, a composite index that aggregates the information can be useful.

In this talk, I first overview composite labour market indexes used by various central banks. I concentrate on the problem of the proper formulation of the indexes, and highlight the need of addressing missing data, asynchronous data release and lead-lag relationship among data.

Then I will discuss the possible approaches to the estimation of the parameters of the various formulations of the composite indexes. These approaches include machine learning techniques.

Finally, I will show the usefulness of the index constructed by the Czech National Bank on a real-world issue.