Title
Generalized beta regression for loss given default
abstract:
In risk analysis of credit portfolios, loss given default (LGD) is the proportion of the exposure that will be lost if a default occurs. It is assumed to follow a beta distribution and a generalized beta regression model can be formulated in a similar way as a generalized linear model. The model can be fitted using the R. Using fitted models for default and LGD, VaR and ES can be calculated.
reference: http://www.risk.net/digital_assets/4564/v7n4a3.pdf
My dissertation structure:
Ch1 introduction
1 concepts
2 default models
3 models for LGD
4 Dissertation structure
Ch2 beta regression
Ch3 joint modelling of mean and dispersion
Ch4 generalized linear mixed models
Ch5 portfolio loss and risk measures
Ch6 empirical results and analysis
Ch7 summary and conclusion
I have finished chapter1 to chapter5, which introduce the theory and how to apply to our model. But I don’t know how to do ch6, which is related to 4.3.1 and 4.3.2 of the reference. I hope you can fit the three model using least-square method (like table2 in the reference) by using the statistical software R, and represent them as a table, then just simply analyze them. The most important is use monte carlo simulation to simulate 200000 portfolio according to the model you fitted (as said in 4.3.2), also use R to do that, calculate VaR(value at risk) and ES by using non parametric method, then analyze and summary. That is, do exactly what 4.3.2 do and calculate ES as an extra .( do the presentation like figure 1 and figure 2). Just put the R code in the appendix.
Do not use MLE method to fit the three model, which I have done. I just want you to use least-square method. The data used to fit these models is simulated according to the table 1, and I have already simulated. I can give the R code to you , then you can use it to get the data.
Last Completed Projects
| topic title | academic level | Writer | delivered |
|---|