Model Error
نویسنده
چکیده
Modern finance would not have been possible without models. Increasingly complex quantitative models drive financial innovation and the growth of derivatives markets. Models are necessary to value financial instruments and to measure the risks of individual positions and portfolios. Yet when used inappropriately, the models themselves can become an important source of risk. Recently, several well-publicized instances occurred of institutions suffering significant losses attributed to model error. This has sharpened the interest in model risk among financial institutions and their regulators. In March of 1997, NatWest Markets, an investment banking subsidiary of National Westminster Bank, announced a loss of £90 million due to mispriced sterling interest rate options. Shortly thereafter, BZW, an investment banking subsidiary of Barclays, sustained a £15 million loss on mispriced currency options and Bank of Tokyo-Mitsubishi announced a loss of $83 million. In April of 1997, Deutsche Morgan Grenfell lost an undisclosed amount. Model errors have been blamed for all these losses.1 This article will describe various models and discuss model errors characteristic of two types—valuation models for individual securities, and models of market risk. The article will discuss the statistical issues that complicate the use of such models, namely the probability distributions of asset returns and estimates of their volatility. It will also discuss a number of practical issues related to model development and describe the approach taken by bank regulators to model risk.
منابع مشابه
An application of Measurement error evaluation using latent class analysis
Latent class analysis (LCA) is a method of evaluating non sampling errors, especially measurement error in categorical data. Biemer (2011) introduced four latent class modeling approaches: probability model parameterization, log linear model, modified path model, and graphical model using path diagrams. These models are interchangeable. Latent class probability models express l...
متن کاملRobust Identification of Smart Foam Using Set Mem-bership Estimation in A Model Error Modeling Frame-work
The aim of this paper is robust identification of smart foam, as an electroacoustic transducer, considering unmodeled dynamics due to nonlinearities in behaviour at low frequencies and measurement noise at high frequencies as existent uncertainties. Set membership estimation combined with model error modelling technique is used where the approach is based on worst case scenario with unknown but...
متن کاملSurvey of Money- Output Causality: Case Study of Iran, Based on Vector Error Correction Model (VECM)
This study investigated the dynamic relationship between money, prices and output in a multivariate structure of casualty analysis in Iran for the two period of 1969 to 2012 (entire period) and 1989 to 2012 (sub-period). This statistical framework has been projected for situations where causal links may have changed over the sample period. Results of a three-variable Vector Error Correction Mod...
متن کاملTHE STUDY OF PACKET LOSS EFFECT ON NETWORK CONTROL SYSTEM ERROR FUNCTION MODEL
Modern control systems widely use network to decrease the implementation cost and also increase the performance. Although they have several advantages, they suffer from some limitations and deficiencies. Packet loss is one of the main limitations which affect the control system in different conditions and finally can lead to system instability. To prevent such problems it is important to model ...
متن کاملA New Ridge Estimator in Linear Measurement Error Model with Stochastic Linear Restrictions
In this paper, we propose a new ridge-type estimator called the new mixed ridge estimator (NMRE) by unifying the sample and prior information in linear measurement error model with additional stochastic linear restrictions. The new estimator is a generalization of the mixed estimator (ME) and ridge estimator (RE). The performances of this new estimator and mixed ridge estimator (MRE) against th...
متن کامل