Can someone help me replicate this paper? We can do a joint publication on this and similar other topics and I can also reimburse a limited budget for this. Thanks.
MODEL SIMPLIFIED: The model presented int he paper could be simplified into the following steps:
1. Construct a cash flow model
This is shown in equation (1) in the paper, p. 612. This model assumes that cash flow of the firm depends on: expected cash flow + unobserved state factor + the firm's cash flow variation (based on past cash flows).
2. Construct state factor model
This is explained by equation 2 in the paper. The papers assumes that the state factor follows Gaussian distribution and that the data distribution of the state factor comes from the Wiener process. This assumption is problematic---because in construction business, business volume and hence cash flow of the firm does not follow normal distribution curve. In many economies, construction business is used as one of many indicators to measure the health of the economy and vice versa. For instance, in the rainy season or heavy winter, construction slows down regardless of the state of the economy. Yet when the economy slows down, construction business also retracts. Thus, the assumption that the state factor is Gaussian may weaken this second element of the overall structural model.
3. Construct present value model
The paper assumes that the present value of the firm depends on: (i) constant past cash flow growth; (ii) WAAC; and (iii) firm's constant future growth ate. here lies the second problem---constant growth rate. Construction business, unlike other types of businesses, do not have constant growth rate. This part of the model must look into seasonal adjustment. The material on seasonal adjustment may be found in time series literature.
4. Construct default threshold model.
This is explained in equations (4) and (5) in the paper. The paper asserts that default depends on the leverage of the present value of the firm. No default of 0 < leverage < Upper bound of default. Default occurs if leverage > upper bound of default.
5. Construct credit quality score
Using step 4 to construct CQS by answering the query: What is the probability of default of what is P(D > V) where D = default and V = leverage of present value of the firm.
HOW TO REPLICATE: You need to access financial statements of firms. I suggest SEC reports of publicly traded companies. This could be search at Yahoo Finance. The report you are looking for is called SEC 10-Q. This is quarterly operating results and SEC 8-K (Reports of Operations and Financial Condition). Pick an industry, and firms---obtain the reports and get to the cash flow statement of each firm over as many years or quarters as you wish to cover.
MODIFICATION: In order to add to the literature, you need to be aware of the potential weakness of the model: assumptions 2 and 3 above.
Thank You Paul. Its been helpful. I do have the data and wanted to see if we can find some relevant Stata or R codes to do it. Do you have any chance of such codes?
@Muhammad .... no special code no special software. Simple excel would do the job just fine. Let me know if you need help. This is straight forward statistical testing, distribution verification and probability issues.
I have been playing with replicating using Stata. I could not yet got it done. Can I send you material to help me replicate and obtain the results? I will highly acknowledge.