Aligning Models and Data for Systemic Risk Analysis
Roger M. Stein · Cambridge University Press eBooks · 2013
The recent financial crisis has brought to the fore issues of quantifying and reducing systemic risk. This focus has precipitated the exploration of various methods for measuring systemic risks and for attributing systemic risk contributions to systemically important financial institutions. Concomitant with this stream of research are efforts to collect, standardize and store data useful to these modeling efforts. While discussions of modeling approaches are pervasive in the literature on systemic risk, issues of data requirements and suitability are often relegated to the status of implementation details. This short chapter is an attempt to deepen this discussion. We provide a 2 × 2 mapping of modeling strategies to key data characteristics and constraints that can help modelers determine which models are feasible given the available data; conversely, the mapping can provide guidance for data collection efforts in cases where specific analytic properties are desired. The framework may also be useful for evaluating, at a conceptual level, the trade-offs for incremental data collection. To provide background for this mapping, we review the analytic benefits and limitations of using aggregate vs. micro-level data, provide background on the role of data linking and discuss some of the practical aspects of data pooling including concerns about confidentiality. Throughout the chapter, we include examples from various domains to make the points we outline concrete. Introduction Modern statisticians are familiar with the notion that any finite body of data contains only a limited amount of information on any point under examination; that this limit is set by the nature of the data themselves, and cannot be increased by any amount of ingenuity expended in their statistical examination; that the statistician's task, in fact, is limited to the extraction of the whole of the available information on any particular issue. R.A. Fisher