Page 64 - Special Topic Session (STS) - Volume 2
P. 64

STS459 Gan C.P. et al.





























                  Figure  3.1  The  estimated  probability  that  the  future  rating  is  7,  8,  and  9,
                  respectively, for Company 6, given the interest rate for present quarter, and the
                  ratings for previous and present quarters are both equal to 8.

                      We have obtained the plot (not presented here) of the value of the -th
                  (1 ≤  ≤  ,  = 4) latent factor  with the probability of Company 2 staying in
                              ∗
                            ∗
                  rating 10 given that its ratings in the previous and present quarters are both
                  10. The plots when  = 1 and 3 show obvious variation of the probability of
                  staying in rating 10 when the first and third latent factors vary.
                      The similar plots for other companies with given ratings in the previous
                  and  present  quarters  are  expected  to  exhibit  variation  of  transition
                  probabilities  when  the  values  of  some  of  the  latent  factors  vary.  The
                  explanatory power shown by the latent factors is consistent with the previous
                  finding that some of the individual macroeconomic variables have effects on
                  the credit risk.

                  4.  Discussion and Conclusion
                      The  latent  factors  extracted  from  the  Taiwan  macroeconomic  variables
                  exhibit an explanatory power over the credit risk. It might be possible to use
                  the  predicted  probabilities  in  a  number  of  consecutive  quarters  to  get  an
                  indication  of  the  possible  transition  of  rating  in  the  near  future.  It  is  also
                  possible that the latent factors extracted from the macroeconomic variables in
                  other countries will also help to give a better prediction of the credit risk.



                                                                      53 | I S I   W S C   2 0 1 9
   59   60   61   62   63   64   65   66   67   68   69