Cookies?
Library Header Image
LSE Theses Online London School of Economics web site

A theoretical and empirical study of asset securitisation: Risk modelling, security design and market pricing.

Jobst, Andreas Alexander (2005) A theoretical and empirical study of asset securitisation: Risk modelling, security design and market pricing. PhD thesis, London School of Economics and Political Science (United Kingdom).

[img]
Preview
PDF
Download (19MB) | Preview

Abstract

Asset securitisation represents an alternative risk management and refinancing method, which allows issues to convert classifiable cash flows from a diversified portfolio of pre-existing assets and receivables (liquidity transformation and asset diversification process) of varying maturity and quality (integration and differentiation process) into negotiable capital market paper, so-called "asset-backed securities" (ABS). Over the recent past ambivalence in the definition of capital adequacy for credit risk has particularly facilitated the development of loan securitisation as a refined "regulatory arbitrage tool". However, as impending regulatory change shifts the prime objective of securitisation to the efficient management of economic capital, procedural and substantive aspects of asset securitisation warrant closer inspection. The dissertation presents a comprehensive examination of the risk modelling, asset selection, optimal security design and competitive market pricing of asset-backed securities. We first provide an overview of the main characteristics of asset securitisation and explain its attendant benefits and drawbacks, especially as they pertain to the refinancing of illiquid asset exposures, such as SME-related payment obligations. Subsequently, we explain the gradual evolution of the regulatory treatment of asset securitisation adopted by the Basle Committee on banking Supervision in the wake of a general revision of the 1988 Basle Accord, which finally led to the adoption of the so-called Basle Securitisation Framework in 2004. We then present a single-factor, loss-based asset pricing model, which estimates the risk-neutral investment return of subordinated debt securities ("tranches") as leveraged contingent claims on a securitised reference portfolio of pooled credit exposures. We challenge common wisdom of robust statistics for the estimation of portfolio credit risk by adopting extreme value analysis, mainly because the leveraged exposure of securitised debt on fundamental asset value changes requires a better parametric specification of extreme quantiles to gauge unexpected loss. Based on the loss sharing between issuers and investors in a common security design, we examine how securitised asset exposure translates into investment risk of asset-backed securities. As a longitudinal extension to this valuation model, we also investigate the price dynamics of securitised debt. A multi-factor GARCH process is applied as an econometric specification of the heteroskedasticity of secondary market spreads of selected types of ABS transactions for valuation and forecasting purposes. In light of the substantial valuation uncertainty in securitisation markets, we conclude with a simple one-shot auction model, in which issuers maximise net payoffs from securitised debt under "winner's curse"-type underpricing as agency cost of adverse selection. In particular, we study how uninformed investment demand at varying degrees of valuation uncertainty affects the utility of endogenous price discovery by informed investors. Overall our synthesis of empirical and theoretical approaches yields instructive findings about important yet unexplored issues concerning the economic rationale of asset securitisation.

Item Type: Thesis (PhD)
Uncontrolled Keywords: Business Administration, Management
Sets: Collections > ProQuest Etheses
URI: http://etheses.lse.ac.uk/id/eprint/1936

Actions (login required)

Record administration - authorised staff only Record administration - authorised staff only

Downloads

Downloads per month over past year

View more statistics