Question

The following OLS regression model has been produced to calculate the Loss Given Default (LGD) of corporate firms.

Model A:

Model A:	Constant	LGD_A	TAN	I_DEF
Coefficient	-1.35	2.05	-1.26	-9.46
Standard error of the coefficient	0.32	1.42	0.38	3.68
p-value	0.00	0.32	0.00	0.01

Where: LGD_A is Average LGD of the industry TAN is Tangibility of the firm i

I_DEF is the industry default rate The R2 for this model is 57%.

1. Explain the overall fit of the model and how well it predicts loss given default. (10% question weight)

2. Describe the impact of each factor of loss given default. (20% question weight)

3. Calculate the t-ratio for each factor in the model. (20% question weight)

4. Discuss the different factors you would consider for inclusion in an LGD model, along with

   their benefit, challenges, and expected relationship with LGD. (50% question weight)

Answer 1

1. Explain the overall fit of the model and how well it predicts loss given default. (10% question weight)

57% of the variability in the model is explained which is not a good fitted model.

1. Describe the impact of each factor of loss given default. (20% question weight)

For LGD_A:

The p-value from the output is 0.32.

Since the p-value (0.32) is greater than the significance level (0.05), we cannot reject the null hypothesis.

Therefore, we cannot conclude that there is an impact of LGD_A on loss given default.

For TAN:

The p-value from the output is 0.00.

Since the p-value (0.00) is less than the significance level (0.05), we can reject the null hypothesis.

Therefore, we can conclude that there is an impact of TAN on loss given default.

For I_DEF:

The p-value from the output is 0.01.

Since the p-value (0.01) is less than the significance level (0.05), we can reject the null hypothesis.

Therefore, we can conclude that there is an impact of I_DEF on loss given default.

1. Calculate the t-ratio for each factor in the model. (20% question weight)

For LGD_A:

t-ratio = Coefficient/Standard error of the coefficient = 2.05/1.42 = 1.44

For TAN:

t-ratio = Coefficient/Standard error of the coefficient = -1.26/0.38 = -3.32

For I_DEF:

t-ratio = Coefficient/Standard error of the coefficient = -9.46/3.68 = -2.57

1. Discuss the different factors you would consider for inclusion in an LGD model, along with

   their benefit, challenges, and expected relationship with LGD. (50% question weight)

Several possibilities exist:
1. Pubic bond issues
2. Secondary markets in defaulted loans and bonds
3. Credit derivatives
4. Studies by commercial banks
5. Evidence from commercial finance

Because of these markets, those who wish to learn more about the LGD of larger corporations and sovereign nations can access price information that provides a free market estimate of the cost should default occur or have taken place.

Market LGD is a recovery rate that can be obtained by observing the post-default prices of bonds and syndicated loans. The rating agencies use market prices with a one month delay to allow the initial exuberance or gloom to settle into a consensus price.

The CECL rules mean that an LGD factor will have to be determined for every loan including new business. This means that there will be loans to businesses where there is precious little default history either in the banking system or the bond market.