Professional Documents
Culture Documents
12
247
259
SS
457.604
108.791
566.394
MS
38.134
0.440
F
86.579
Significanc
eF
0.000
Homework 5
Coefficients
1.4727
Standar
d Error
0.6169
t Stat
2.3871
Pvalue
0.0177
Lower 95%
0.2576
Upper
95%
2.6878
rating1
0.1187
0.3575
0.3322
0.7400
-0.5853
0.8228
rating2
rating3
rating4
rating5
0.3613
0.6082
2.3314
3.7291
0.2519
0.2500
0.2761
0.2946
1.4341
2.4331
8.4431
12.6570
0.1528
0.0157
0.0000
0.0000
-0.1349
0.1159
1.7875
3.1488
lnnetpro
-0.2360
0.0775
-3.0450
0.0026
-0.3887
0.8574
1.1005
2.8752
4.3093
0.0834
lnyfmat
default
-0.0837
0.8514
0.0860
0.3600
-0.9735
2.3652
0.3312
0.0188
-0.2532
0.1424
0.0857
1.5603
utildum
calldum
sinkdum
subdum
-0.0822
0.3274
0.7465
0.5382
0.1187
0.1110
0.3132
0.2299
-0.6929
2.9503
2.3837
2.3411
0.4890
0.0035
0.0179
0.0200
-0.3160
0.1088
0.1297
0.0854
0.1515
0.5461
1.3633
0.9910
Intercept
Lower
95.0%
0.2576
0.5853
0.1349
0.1159
1.7875
3.1488
0.3887
0.2532
0.1424
0.3160
0.1088
0.1297
0.0854
It can be seen from the adjusted r-squared value which is 0.799, that this model explains
79.9% the movement of prices of all the non-US bonds. And it can be seen from above
model that most of the parameters except rating1, rating2, innetprp, infymat and utildum
have t-statistics over 2 hence all those parameters are statistically significant and hence
there is 95% probability that the parameters with a t-statistics over 2 are having the yield
spread within 2 standard deviations.
2. Are the signs of the each of the coefficients what you expected from the
U.S. market? (go through each one individually)
There are few parameters which are not statistically significant and most of the
parameters are significant such as
Intercept
Coefficient:- 1.4727 and t-stat:- 2.3871
This is positive and it should be positive since the other markets will take this as
base and hence will increase the yield because of the other constraint or
parameters.
Rating 1
Coefficient:- 0.1187 and t-stat:- 0.3322
The yield spread is dependent on the the yield of US treasury bond so higher the
bond rating the lower would the yield spread be and since this rating is for other
markets, it would have a positive sign and higher the rating coefficient, the higher
the yield spread will be. And the low t-stat indicates that this rating may not be as
accurate in pricing the bonds.
Upper
95.0%
2.6878
0.8228
0.8574
1.1005
2.8752
4.3093
0.0834
0.0857
1.5603
0.1515
0.5461
1.3633
0.9910
Homework 5
Rating 2
Coefficient:- 0.3613 and t-stat:- 1.4341
As explained in rating 1, the same goes for rating 2 and the t-stat is not
significant enough for this rating as well.
Rating 3
Coefficient:- 0.6082 and t-stat:- 2.4331
As explained in rating 1, the same goes for rating 3 and the t-stat is significant
enough for this rating as it is more than 2 and hence it has 95% probability that it
will be in 2 standard error.
Rating 4
Coefficient:- 2.3314 and t-stat:- 8.4431
As explained in rating 1, the same goes for rating 4 and the t-stat is significant
enough for this rating as it is more than 2 and hence it has 95% probability that it
will be in 2 standard error.
Rating 5
Coefficient:- 3.7291 and t-stat:- 12.6570
As explained in rating 1, the same goes for rating 4 and the t-stat is significant
enough for this rating as it is more than 2 and hence it has 95% probability that it
will be in 2 standard error.
Bond Maturity (lnyfmat)
Coefficient:- -0.0837 and t-stat:- -0.9735
The more the bond maturity period the lesser the price of the bond has to be. So
this rating would act as negative parameter and hence it has negative coefficient
and the t-stat for this parameter is not statistically significant.
Issue Size (lnnetpro)
Coefficient:- -0.2360 and t-stat:- -3.0450
The more the size of the bond will be the debt offering will be more stabe and
hence the yield of the bond would be lower. So the coefficient for this has to be
negative and hence it is as shown in the regression analysis.
Market Conditions (default)
Coefficient:- 0.8514 and t-stat:- 0.3600
The price of the bond hugely depends on the market conditions so the better the
market, the more priced the bond will and hence the coefficient of this bond has
to positive and close to 1.
Covenant provisions (calldum,subdum,sinkdum)
Calldum- Coefficient:- 0.3274 and t-stat:- 2.9503
Subdum- Coefficient:- 0.5382 and t-stat:- 2.3411
Homework 5
3. Are there additional factors that the market uses to price foreign
corporate bonds? Specifically, he wants to know if the volatility of the
exchange rate and the firms location are important. To answer this, add to
the U.S. model the variables for Volatility, Rule of Law and Creditor rights.
What do you find, what does it mean?
With the addition of the three new factors in the decision pricing model, the analysis is as
follows:
SUMMARY
OUTPUT
Regression Statistics
Multiple R
0.903139801
R Square
0.815661501
Adjusted R
Square
0.804329216
Standard
Error
0.654142803
Observations
260
ANOVA
df
MS
30.7991
0.4279
F
71.9768
Significanc
eF
0.0000
15
244
259
SS
461.9861
104.4083
566.3944
Coefficients
2.32250
Standard
Error
0.69514
t Stat
3.34107
P-value
0.00097
Lower 95%
0.95327
Upper
95%
3.69174
rating1
0.01168
0.35502
0.03289
0.97379
-0.68761
0.71097
rating2
rating3
rating4
0.33036
0.60781
2.20333
0.25127
0.24729
0.27622
1.31476
2.45792
7.97666
0.18983
0.01467
0.00000
-0.16458
0.12072
1.65925
0.82530
1.09489
2.74742
Regression
Residual
Total
Intercept
Lower
95.0%
0.95327
0.68761
0.16458
0.12072
1.65925
Upper
95.0%
3.69174
0.71097
0.82530
1.09489
2.74742
Homework 5
rating5
3.63522
0.29558
12.29844
0.00000
3.05300
4.21744
lnnetpro
-0.25003
0.07811
-3.20115
0.00155
-0.40388
-0.09618
lnyfmat
default
-0.07005
0.75064
0.08540
0.35836
-0.82022
2.09466
0.41289
0.03723
-0.23826
0.04477
0.09817
1.45651
utildum
calldum
sinkdum
subdum
-0.08789
0.42537
0.72376
0.54695
0.12284
0.11468
0.30953
0.22932
-0.71549
3.70925
2.33823
2.38510
0.47499
0.00026
0.02018
0.01784
-0.32985
0.19949
0.11406
0.09525
0.15407
0.65126
1.33346
0.99865
fxvoliti
40.65339
22.82200
1.78133
0.07610
-4.29987
85.60666
credrgts
-0.08861
0.04636
-1.91113
0.05716
-0.17993
0.00272
ruleofla
-0.08086
0.03055
-2.64682
0.00865
-0.14104
-0.02069
3.05300
0.40388
0.23826
0.04477
0.32985
0.19949
0.11406
0.09525
4.29987
0.17993
0.14104
It can be seen from the adjusted r-squared value which is 0.804, that this model explains
80.4% the movement of prices of all the non-US bonds. And it can be seen from above
model that most of the parameters except rating1, rating2, innetprp, infymat and utildum
have t-statistics over 2 hence all those parameters are statistically significant and hence
there is 95% probability that the parameters with a t-statistics over 2 are having the yield
spread within 2 standard deviations.
Rule of Law (ruleofla)
Coefficient:- -0.08086 and t-stat :- -2.64682
The index of the law and order tradition of the country. It is scaled from 0 to 10, with
higher scores for counties with more tradition for law and order. So more better the laws
would be the more rating a country has so this factor should be negatively correlated with
a negative coefficient.
Creditor Rights (credrgts)
Coefficient:- -0.08861 and t-stat :- -1.91113
This is an index aggregating different creditor rights a particular country provides. This
should be negatively correlated as the higher the chances of default higher would be the
yield spread.
FX Volatility (fxvoliti)
Coefficient:- 40.65339 and t-stat :- 1.78133
The greater the volatility, the greater would be the yield spread of a given bond. But this
factor is statistically not significant since the t-stat value is less than 2.
4.21744
-0.09618
0.09817
1.45651
0.15407
0.65126
1.33346
0.99865
85.60666
0.00272
-0.02069
Homework 5
Homework 5
Rule of Law: The index of the law and order tradition of the country. It is scaled
from 0 to 10, with higher scores for counties with more tradition for law and order.
From LLSV (1998) (variable name in dataset: ruleofla)
Creditor Rights: is an index aggregating different creditor rights a particular
country provides. The index ranges from 0 to 4, with 4 representing the highest
protection. One point is added if there is no automatic stay on assets, secured
creditors get paid first, there are restrictions on reorganizations, and if
management does not stay in reorganizations. From LLSV (1998).(variable name
in dataset: credrgts)
FX Volatility: The 30-day historical volatility of the U.S. to home country currency
exchange rate(variable name in dataset: fxvol)
The Data
The spreadsheet, [bond case data], contains data on 260 corporate bonds issued
by foreign firms in the U.S. over the period 1987-1998.