Valuation of risks on the aggregate level
In project area B, the quantitative projects are summarized, that deal with the estimation of risks at the aggregated level. Regarding content there exists are many connections to separate project areas A and C, like for example on the level of stochastic modeling or the economic approaches. With the help of three CRC 649 seminars (QFS, Schumpeter, WIAS) and further single projects a mutual transfer of knowledge and methods is for the contracted running time guaranteed. A further integration factor is the joint data supply and use within the boundaries of the RDC, project area INF. The applied for workshops and CRC 649 conferences will guarantee the steady exchange with internationally leading scientists.
Project area B1
The project area B1 engages in the dynamic semi-parametric modeling of risk factors. In the foreground is the GPLSIM (generalized partially linear model) class for panel-regression and DSFMs (dynamic semi-parametric factor models). The main direction of this research project is the development and implementation of the statistical theory of his model class and its use on the dynamics of the SPD and on panel data. In doing so extreme computationally extensive procedures will be used (theoretically extensively yet unresearched). The empirical calibration comes from the data sources (e. g. high frequency option price time-series, macro time-series), that are generated in connection with the RDC. The developed procedures will be used by projects B2, B3, B4 and economical modeling of the received VAR time-series together with the projects C1, C2.
Project area B3
Next to humane capital and pension claims, property possession is the most important position in the asset portfolio of many households. The project area B3 examines the actual investment behaviour with property and the risk connected with them and contrasts them with decisions, that, in light of the provided correlations of value more important components of personal assets would be more optimal. The measurement of property prices and the valuation of property assets play a central role in the portfolio context. The property transactions in Berlin from 1995 – 2004 are an important data pool for the CRC 649. The project areas B1, B4 will be working closely together on a methodical level.
Project area B5
The project area B5 focuses on the development of structure adaptive methods for different problems of statistical data analysis with use on risk management. In the context of effective dimension reduction, models will be examined, that permit the high dimensional data projection, without significant data loss, into a low dimensional space. In the modeling of non stationary time series extreme shocks as well as gradual changes of the general economic conditions will be admitted. This leads naturally to local stationary or local homogeneous time series. In this context the project area will work structurally adaptive methods of modelling and estimation in temporally inhomogeneous models. The direct partners of the project are B1, B2, B3 and B4.
::: Projects finished
Project area B2
The project area B2 concentrates on high-risk strategies that deal with the management of product brands and their influence on the shareholder value. It will thereby be concerning itself with the questions of measuring of market value and the relationship between markets and firm values. Examined will be the influence of strategies of the brand elimination and its effect on its business value. Project area B1 will be working on the methodical level and project areas A2 and B5 will be collaborating on the theoretical level.
Project area B4
The project area B4 examines estimation models for businesses, with which risk and capital costs on will be estimated in a factor based way. The existing estimation models will be the starting point. The basic idea is to equate the share value generated from an estimation method with the observed market value and to derive the risk premium from the equation. The project area will work with an extensive database with market and balance data of some thousands of businesses from approximately 30 countries over a period of up to 10 years. The project area plans to let especially the results of the project areas B3 property estimation and project area B2 brand estimation flow into the construction of a suitable business estimation model. On the methodical level we will be collaborating closely with the project area B1 semi-parametric estimation of estimation models.
Project area B6
The project area B6 busies itself with the finance mathematical modeling of markets, on which agents with different information level are active, especially the risk produced by insider trading. In two steps, above all connections between utility theory and information theory, will be devolved: on the mesoscopic level 2 and 3 trade models will be examined with small noise trader and insider and large market makers, analysts, managers or experts; on the micro-economic level the market will be from several types of agents with typical patters of information privileges and trade defined and in the mesoscopic Limes consequences for market behaviour and the statistical identification of insiders will be examined. It gives methodical connections will be with project areas A1, A2, and C6 and regarding content with A1, B1 and B2.
Project area B7
It gives methodical connections will be with project areas B1, B5, B6 and C1.