|| Edinburgh Local group
The 1988 Basel Accord set standards for banks on the minimum amount of capital that they should hold to cover credit risk, effectively limiting the money they can lend. The new Accord has been designed to address a major criticism of “lack of risk-sensitivity” – banks that have very “clean” portfolios are currently being disadvantaged. It also factors “operational risk” into the capital equation.
The new Basel Accord achieves its primary objective of increased risk sensitivity by addressing both “expected losses” and “unexpected losses”, the latter being the statistical variation of the former, and it requires the banks to adopt increased rigour in the statistical calculation of the loss components.
The new Accord also requires the banks to validate their models correctly; to monitor their predictions more effectively; to collect data for modelling over a longer time horizon; to have effective methods for taking a conservative view under uncertainty; to stress test their portfolios; to use the underlying risk models in practice, having appropriate approaches for augmenting them with human expertise if necessary; and to produce written documentation on the model development and procedures, detailing not just those methods selected but also those rejected. In short, to overhaul their risk management practice.
This talk discusses the issues above, picking some examples of current practice and explaining how improvements can be made. Some working calculations will also be shown to help give a feel for the underlying numbers.
Scottish Neighbourhood Statistics is a major project from the Scottish Executive to transform the data held on Scotland’s local areas. The Scottish Executive is working with the Office for National Statistics (ONS), and the administrations in Wales and Northern Ireland to create a neighbourhood statistics service across the UK, with the Scottish system reflecting differences in the education and legal systems.
Scottish Neighbourhood Statistics provides figures on Scotland's local areas for local authorities, electoral wards, and postcodes. It provides information on health, education, poverty, unemployment, housing, population, crime and social / community issues. The ethos of the project is that data is made available for the smallest areas possible and that has involved building partnerships with other information providers and users including local authorities, health boards and the voluntary sector.
Users of the service can obtain a quick profile report for a local area in Scotland or can choose a local area and create standard or custom reports and maps comparing statistical data from multiple sources.
Providing such a diverse range of data across different geographies and different time frames provided many challenges to the project. These covered areas of multi-organisation liaison, data collection and administration, quality assurance, on-going data maintenance, user interface design, statistical validity and technical infrastructure. The three speakers from Communities Scotland, Edina Software and GeoWise will outline these challenges and how they were overcome in delivering a truly joined-up government service based around Scotland's geography.
Angela Dale, Programme Director of the ESRC's Reseach Method's Programme, will outline the issues in achieving the ESRC's aim of improving the standard of research methods across the UK social science community.
Although a variety of sampling methods can be used to select telephone household samples, not all of the methods will provide a representative and unbiased sample of the general population. This presentation will examine the extent to which telephone households represent all households in Scotland and review, in a non-technical manner, different sampling techniques.Download a PowerPoint presentation of this talk
Scotland has higher mortality rates than the rest of Britain, which is not surprising given the higher rates of social deprivation. However, the 'Scottish effect' refers to the fact that Scotland has higher mortality rates than would be expected given the excess deprivation compared to England and Wales. In this paper we compare self-reported illness to mortality in Scotland and find that the rates are lower than we might expect from the mortality rates. We suggest that the 'Scottish effect' might be better termed the 'Scottish mortality effect' and that an interesting 'Scottish self-reported morbidity effect' may also exist.
Generation Scotland (GS) aims to determine the genetic basis of common diseases by creating an ethically sound, population and family based sample collection in Scotland.GS will focus on individuals with a disease diagnosis and their ‘at risk’ relatives. GS seeks to identify genetic contributions to individual risk of disease, disease prognosis and response to treatment. The project raises significant informatics and statistical challenges. This derives from the facts that there will be a need to integrate diverse data sets and interpret soft clinical phenotypic data collected on thousands of patients in relation to hard genotypic data measured at thousands of loci. If these challenges are met, then this information will be valuable for disease screening, surveillance and treatment choice. Importantly, gene discovery through GS will fuel drug target discovery and targeted drug development. GS is thus a model for health and wealth creation. It would most likely operate as a not-for-profit company limited by guarantee. Generation Scotland has been initiated by pump-priming awards to the University of Edinburgh from Scottish Enterprise and the Scottish Higher Education Funding Council Strategic Research Development Grant initiative (�1.8million). Collaborators include the Universities of Edinburgh, Glasgow, Aberdeen, Dundee and St Andrews, the MRC Human Genetics Unit, Edinburgh, the MRC Public Health and Social Sciences Unit, Glasgow, the Scottish Council for Research in Education Centre, the Scottish School of Primary Care and the Information and Statistics Division, NHS Scotland. A detailed proposal for further, major public sector funding is under consideration in response to the SEWHD and SEETLL Genetics Healthcare Initiative.
How will genetics affect insurance underwriting, and what are the actuarial models that will answer questions about genetics and insurance? Will insurance companies create a `genetic underclass'? This talk examines the popular beliefs that genetic tests are strongly predictive, and that insurance underwriting is an exact science, and aims to debunk both.