Models of Corporations and Sovereigns

The sample of corporations included information from different industries (oil and gas, utilities, retail, telecom, etc.) and countries. We considered the rated companies from these industries which also had financial and market indicators. Financial explanatory variables included such group indicators as size of company, capitaliza­tion, assets, management, efficiency, and liquidity. Among the macro indicators it stands out on the corruption perception index by Transparency International. While among market indicators the volatility of the market prices stands out. We also added industry classification dummies, as well as such factors as groups of countries and a company’s affiliation.

We used both the agencies’ and Bloomberg data for this sample. Financial indicators were selected for 30+ countries during 2000-2009 for 211 corporations. Our database included nearly 1,800 estimations (non-balance panel) for three international rating agencies; S&P, Fitch and Moody’s ratings.

Order probit model parameters are presented in Table 3. We do not have the opportunity to use all the explanatory variables. You can see the best models, which differed in profitability indicators (Karminsky 2010).

The signs for all three models are equal, and have a good explanation from a financial point of view. As for its interpretation, a positive sign of coefficient relates

Table 3 Comparison of corporate rating models for international CRA

Variable

S&P

Fitch

Moody’s

LN (market capital)

-0.692***

-0.806***

-0.691***

Sales/Cash

0.00004***

-0.00051

-0.00049

EBIT/interest expenses

-0.0017***

0.0006

-0.0054***

LT debt/capital

0.006***

0.011***

0.019***

Retained earnings/capital

— 1 107***

-0.581**

-1.230***

Volatility (360d)

0.012***

0.013***

0.016***

Corruption perception index

-0.217***

-0.088***

-0.088

Chemicals

-0.235***

0.381***

-0.182

Metal and mining

0.322***

1 317***

0.947***

Pseudo-R2

0.215

0.220

0.276

Number of observations

1,362

423

339

|A|= 0

40.6 %

34.3 %

42.5 %

|Д|< і

87.7%

87.7%

87.0%

Notes: *, **, *** represent 10%, 5%, 1% levels of significant, respectively. Italic texts were connected with statistical summaries of the tables.

image120 Подпись: —S&P Ш Moody’s Fitch

to a negative influence on rating, and vice versa, because of the fact that the scale mapping choice should be taken into account. From this model we can make the following conclusions:

• The size of the company, asset profitability and the EBITDA to interest expenses ratio have a positive influence on the rating level. A ratio such as LT Debt to Capital has a negative influence on the rating grade.

• Industry dummies are significant. We can see that companies from the utility and oil and gas industries have higher ratings.

• Market variables are also important for understanding the behavior of companies, for example, the corruption index has a negative influence.

Time has an important influence as well. We used a system of dummies during the years 2000-2009 to understand the impact of methodology and crisis. Most of the dummies are significant. We can see in Fig. 3 that all the agencies have the same procyclicity connected with the crisis of 1998 and 2008.

The main explanatory variables for sovereign rating models may be classified into 6 groups of quantitative variables such as: bank characteristics, economic growth, international finance, monetary policy, and public finance and stock market characteristics. In our research 30+ parameters from all groups were analyzed.

We also used dummies for regions, financial crisis type and indicators of corruption (CPI index). Our sample included nearly 1,500 estimations for 100+ countries during the 1991-2010 periods. We dealt with Moody’s bank ceiling ratings as a sovereign rating proxy. The models are presented in Karminsky et al. (2011a).

We derived a strong association of sovereign ratings with economic growth, the public sector, monetary policy, the banking sector, the foreign sector, stock market variables and geographical regions. The forecast accuracy of the models is higher for investment-level grades than for speculative-level grades.

The majority of working explanatory variables for higher-investment ratings consists of the financial sector variables and GDP per capita. The majority of working explanatory variables for speculative-grade ratings includes budget deficit, inflation growth rate, export-to-import ratio and GDP per capita.

Leave a reply

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>