1: \begin{abstract}

2: We discuss the parameter estimation of the probability of default (PD), the correlation between the obligors, and a phase transition.

3:   In our previous  work,

4:    we studied the problem using the beta-binomial distribution.

5:    A  non-equilibrium phase transition with an order parameter occurs

6:    when the temporal correlation decays by power law. 

7:    In this article, we adopt the Merton model, which uses an asset correlation  as

8:    the default correlation, and find that a phase transition  

9:    occurs 

10:    when the temporal correlation decays  by power law.

11:   When the power  index is less than one, the PD estimator

12:  converges slowly.

13:  Thus, it is difficult to estimate PD with limited historical data.

14:  Conversely, when the power index is greater than one, the convergence speed is inversely proportional to the number of samples.

15: We investigate the empirical default data history of several rating agencies.

16: The estimated  power index is in the slow convergence range when we use long history data.

17: This suggests that PD could have a long memory and 

18: that it is difficult to estimate parameters due to slow convergence.

19: 	

20: 

21: \hspace{0cm}

22: \vspace{1cm}

23: %PACS numbers:  {02.50.Ga, 05.70.Fh, 89.65.Gh, 87.23.Kg} 

24: \end{abstract}

25: