[Linear and Nonlinear Models for Forecasting the Recent Financial Global Crisis Using Mathematical Software]
I would take this opportunity to thank my research supervisor, family and friends for their support and guidance without which this research would not have been possible.
I, [type your full first names and surname here], declare that the contents of this dissertation/thesis represent my own unaided work, and that the dissertation/thesis has not previously been submitted for academic examination towards any qualification. Furthermore, it represents my own opinions and not necessarily those of the University.
Signed __________________ Date _________________
This paper reviews econometrics and operations research methods used in the empirical literature to describe, predict, and remedy financial crises and mortgage defaults. Such as inter- disciplinary approach is beneficial for future research as many of the methods used in isolation are not capable of accurately predicting financial crises and defaults of financial institutions. Operations research is a complex and interdisciplinary tool that combines mathematical modeling, statistics, and algorithms. This tool is often employed by managers and managerial scientists. It is based on techniques that seek to determine either optimal or near optimal solutions to complex problems and situations.
Table of Contents
CHAPTER 4: RESEARCH METHODOLOGY1
Model parsimony and parameter shrinkage1
Forecast Evaluation: Bias And Cross Sectional Dispersion4
Four More Quarters: Additional Forecast Comparisons13
CHAPTER 5: RESULTS & DISCUSSION18
The Subprime Crisis is not unique25
Selected Analyses Of Bank Failure Prediction27
Low and Falling Rates of Earnings32
The Growth of Inequality33
The US Housing Market Bubble34
CHAPTER 6: CONCLUSION37
Five Key Lessons after Reading Andrew Sorkin's Too Big to Fail42
Do not let your money sleep43
Control your expenses43
Build a capital43
CHAPTER 4: RESEARCH METHODOLOGY
Model parsimony and parameter shrinkage
Perhaps the dominant theme across the set of comments is model parsimony. Clements would like to see a comparison with simpler, single country models, and/or some (or all) of the long-run relationships imposed in the DHPS-GVAR models dropped; Giannone and Reichlin argue for a Bayesian shrinkage approach to reduce the parameter space; and Allen is looking for simplification by dropping foreign variables. The results clearly show that for almost every variable, there are some countries and time periods for which the GVAR-AveAve does not outperform the benchmark models. The benchmarks we use in the paper are, of course, simple single equation models (see below for results on single country VAR benchmarks); and, as Sinclair and Stekler demonstrate, one need not be confined to just statistical models. As an alternative benchmark they suggest the mean forecast of the (US) Survey of Professional Forecasts (SPF), itself (by definition) an average forecast (Bailey 2008, pp. 23).
They report that, for the two variables which are collected by the SPF, namely US inflation and output, the results are split: compared with the GVAR-AveAve forecast, the SPF does better for output but not for inflation. We are in fact heartened by this outcome, considering the acknowledged expertise of the Professional Forecasters in the US. However, given the very short evaluation periods at our disposal, forecast evaluation outcomes for any single variable might not be that informative and the results could have gone ...