Stock Market Reaction to Credit Rating Changes: Evidence from a Frontier Market. Abu S Amin 1. Mahfuja Malik. August (Preliminary Draft)

Transcription

1 Stock Market Reaction to Credit Rating Changes: Evidence from a Frontier Market Abu S Amin 1 Mahfuja Malik August 2016 (Preliminary Draft) 1 Corresponding Author. Assistant Professor of Finance, Department of Finance and Law, Central Michigan University, amin1a@cmich.edu. Tel

2 Stock Market Reaction to Credit Rating Changes: Evidence from a Frontier Market Abstract The quality of credit rating announcements by rating agencies in the wake of the financial crisis has been greatly debated. With the passage of Dodd-Frank Act, rating agencies have become more conservative in rating issuances. Although, reaction to rating change announcements has been widely studied in large stock markets, however, in other parts of the world, rating agencies are less regulated than in the USA, which brings up the question whether the rating announcements has information content and how investors react to rating change announcements. We study the case of Bangladesh, a frontier market, where no bond market exists and CRAs just started to rate entities. Although macro-prudential reforms led to the modest start of CRAs, our analysis shows, the market is not fully understanding the information content of announcements. We find a credit downgrade has no stock market reaction while a credit upgrade has a positive reaction to the announcement and more pronounced delayed reaction. Our findings are completely different than what we normally see in the literature, highlighting the institutional difference between frontier and developed markets. 2

3 Stock Market Reaction to Credit Rating Changes: Evidence from a Frontier Market 1. Introduction The financial crisis in the USA triggered a widespread discussion on the quality of rating issued by the credit rating agencies. Credit Rating Agencies (CRA) were blamed for their inability to fully assess credit risk and their practice of inflating rating on structured financial product that might have exacerbated probability of default among many financial institutions. In the wake of this financial meltdown, the legislatures in the USA passed Dodd-Frank Wall Street Reform and Consumer Protection Act (2010) in July Among many reforms, Dodd-Frank Act significantly increases the CRA S liability of issuing inaccurate rating. This law also makes it easier for the SEC to impose a sanction against the CRA. Dimitrov et al. (2015) find that after these reforms, CRAs are now issuing lower rating, giving false warnings, and issue downgrade that is less informative. CRAs conservative approach to rating post Dodd-Frank might have been resulting from protective of their reputation (Morris (2001)). The CRAs in developed market is facing a conundrum between balancing regulation and quality of rating. Although regulating CRAs in the developed market is a recent phenomenon, in less developed markets like Bangladesh CRAs, are not regulated. This brings up an important question of relevance and quality of rating information in less developed markets. We address this question by evaluating the stock market reaction to credit rating change for publicly listed companies in Dhaka Stock Exchange. In contrast to a developed market, there are several institutional differences observed in frontier markets like Bangladesh. Credit rating is a recent phenomenon it all started in 1996 when the Securities Exchange Commission of Bangladesh (SECB) introduced Credit Rating Companies 3

4 Rules (CRC) highlighting the role of credit rating. Following the initiative by the SECB, the Bangladesh Bank (the central bank of Bangladesh) made it mandatory for all the commercial banks to be rated by the end of More recently following Basel II implementation, Bangladesh Bank requires all commercial banks to designate an external credit assessment institute (ECIs) for their own rating and rating of their counterparties. With a modest beginning, CRAs have become key financial institutions in recent years. Another key operational difference is the CRAs in frontier markets receive rating requests the individual clients and they pay for the rating. Moreover, the ratings in Bangladesh is mostly entity rating since there is no well-established bond market. Credit rating offers an additional source of information for the investors; a change in rating explains a material change in the creditworthiness of the firm and can be highly correlated with the likelihood of default. A downgrade in credit rating reflects worsening creditworthiness which is typically followed by higher probability of default causing the bond yield to rise and price of the bond to fall. In the marketplace, especially as equity investors, a downgrade in rating should lead to rebalancing their portfolio. This implies that credit rating announcements are signal to the market and convey information that is important empirically and economically. Therefore, a significant reaction to the announcement of a credit rating change will only occur if the market perceives that the announcement contains information that cannot be obtained from other sources. Hence, our null hypothesis is announcements by rating agencies should not have any incremental information to have any impact on stock prices. Market reaction to credit rating announcements has been studied quite extensively. It seems there is an agreement in the literature that the credit downgrade announcements have a profound stock market impact than credit upgrades. The early work by Katz (1974), Weinstein (1977), Wakeman (1981), and Pinches and Singleton (1978) find no incremental benefit of rating 4

5 information was mainly criticized because of failure to isolate rating announcements with other possible announcements coming at the same time. Holthausen and Leftwich (1986) and Hand et al. (1992) report a significant price reaction to downgrades but not upgrades. The market response to a credit downgrade could be due to unfavorable expectations in earnings, cash flows or financial prospects of the firm. It could also be an outcome of changing leverage condition. Ederington and Goh (1993) find no market reaction to downgrades related to a change in financial leverage while a significant reaction due to deteriorating financial condition. Kliger and Sarig (2000) reports that rating announcements contain inside information that is not publicly available and they also find debt value increases and equity value decreases when Moody s announces better than expected rating. The introduction of Fair Disclosure Regulation in the USA (Reg FD) ensures that any information released by the firm should reach the entire market with the exception of private commercially sensitive information can be released to credit rating agencies. Jorion et al. (2005) find that post-reg FD, the announcement effects of downgrades are more pronounced, and there is a significant positive reaction to upgrade announcements. Similarly, non-us studies by Matolcsy and Lianto (1995) use Australian data, Barron et al. (1997) uses U.K. data, Li et al. (2006) uses Japanese data, Poon and Chan (2008) uses Chinese data. Almost all these prior studies indicate that markets react to negative review or downgrades nonresponsive to rating upgrades. Elayan et al. (2003) find that in New Zealand market responds to rating downgrade and upgrade announcements, and Afik et al. (2014) find in Israel market seems to internalize information before the rating announcement hence at the announcement there is no reaction. Finally, using Spanish corporate bond market data Abad-Romero and Robles-Fernandez (2006) reports negative responses to upgrades and no response to downgrade announcements. Given the considerable 5

6 difference between developed and frontier markets, so far there is not a single research on the effect of rating change announcements in any frontier market. Our analysis shows that downgrade announcement conveys no incremental information while market reacts positively to an upgrade rating change. We also find that there is a pronounced delayed reaction to upgrade rating announcements. Together the results suggest the market might have fully anticipated a possible down rating and the possibility of down rating is fully priced in before the actual announcements. We have only few samples of down rating announcements, so this result needs to be interpreted carefully. On the other hand, an upgrade announcement might have caught the investors as a surprise, and there is an underreaction to initial announcement. Although in the western financial markets there is overwhelming support for credit downgrades, there is only limited evidence of incremental information revealed by credit upgrades. We contribute to the literature in at least two ways: to the best of our knowledge, this is the first study on rating announcement change in the context of the frontier market. Secondly, we show that unlike developed market, in the less developed market information content of rating announcements could be different, especially, in a frontier market, rating downgrade announcements is not significant while rating upgrade announcement could convey incremental information. The paper is organized in the following sections: Section two describes rating agencies and regulations in Bangladesh; Section three discusses the data and methodology; Section four delineates empirical results, and we conclude in section five. 6

7 2. Rating agencies and regulation in Bangladesh Credit Rating and presence of rating agencies is relatively new to the financial system in Bangladesh. In 1996, The Securities Exchange Commission of Bangladesh (SECB), introduced Credit Rating Companies Rules (CRC Rules) to underscore the importance of rating information. This rule suggests that issuance of any debt security (either public or private placement), and IPO including right share offering must be rated by the rating agencies. To maintain the quality of rating, SECB required the local credit rating agencies to have a joint venture or technical collaboration with any established international rating agency. In compliance with the CRC rules, the majority of 178 IPOs between 2004 and 2015 are rated by the CRAs. The financial sector in Bangladesh was among the first sectors to embrace external rating following series of initiatives by Bangladesh Bank, the central bank in Bangladesh. First, on May 29, 2004, the regulators issued a circular requiring all commercial banks going public need to be rated. Following the first circular, to increase transparency, protect depositors and creditors interests, and continuous improvement in overall bank management, on July 5, 2006, Bangladesh Bank issued another circular making annual credit rating mandatory for all banks starting from January It was suggested that all commercial banks must complete the initial credit rating by June 30, As part of continual improvement in the regulatory environment and to comply with the Capital Adequacy Framework for Commercial Banks under Basel II (which is adopted in December 2007 and fully implemented by January 2010), Bangladesh Bank has issued a circular instructing all commercial banks to nominate a recognized External Credit Assessment Institute (ECIs) for their own rating and rating of their counterparties. The assessments of credit risks by CRAs are mapped to the risk weights and the risk-adjusted capital requirements of banks determined against the credit risks. The adoption of Standardized Approach to credit risk evaluation by commercial banks in Bangladesh 7

8 is different than Internal Rating Based approach to credit risk assessment adopted by most of the advanced economies. To streamline operations of insurance agencies, Bangladesh Bank made it mandatory for all general insurance companies to be rated so that a rated insurance company can insure banks clients, not by an insurance company the bank has relation with. With all these regulatory changes, it is evident that credit rating is critical to the operations of financial institutions and capital market access by non-financial institutions in Bangladesh. At present, there are eight licensed credit rating agencies operating in Bangladesh according to financial market regulatory agency Bangladesh Securities and Exchange Commission. Among them, the oldest and first rating agency is Credit Rating Information Services Limited (CRISL) established in 1995 and operating since It is a joint venture of Malaysia Berhad (RAM), JCR-VIS Credit Rating Company of Pakistan, and few other financial institutions. With a license from the Securities Exchange Commission (SEC) under Credit Rating Companies Rules 1996, CRISL now appears as the flagship organization of Bangladesh. It is a public limited company dedicated for credit rating and related services and is being recognized by Bangladesh Bank as the External Credit Assessment Institution (ECAI) to offer its services to the banking community for banking client rating. CRISL provides long-term rating (with validity not more than a year) and short-term rating (with validity not exceeding six months) for corporate entities. All ratings are in local currency, so it doesn t consider foreign currency risk and sovereign risk. In addition to ratings, CRISL also offers a stable, positive and negative outlook for corporate entities. CRISL follows sector-specific rating methodology to reflect different weighting to both qualitative and quantitative factors of each sector. These factors are converted to specific traits with due weighting for highest performance, lowest performance, industry average, etc. to arrive at a meaningful rating of the 8

9 organization. Consistent with rating methodology, CRISL also uses separate rating scales for different sectors. The detailed sample rating scheme is presented in Table 2. CRISL investment grades ratings cover AAA down to BBB, while Speculative grades cover BB down to B. It's Risky Grades fall under the range from CCC to C while D is considered as the Default Grade. The rating requests usually come from the individual clients and they pay for the rating; alternatively, banks working with the clients can also request a credit rating and in that case banks pay for it. In the absence of a well-established corporate bond market, most of these credit rating in Bangladesh is entity rating which reflects entity s debt service capacity, net worth, earnings prospects, and product/service providing a mechanism. This is a unique institutional difference compared to instrument specific rating by credit rating agencies in most of the developed market. 3. Data and Methodology 3.1. Data We obtained the credit rating data from Credit Rating Information and Services Limited (CRISL)-one of the first rating agencies in Bangladesh. Our sample period covers between 2002 and The ratings and announcement dates are supplied by the rating agency. Our initial sample covers 586 announcements. However the final sample consists of rating change announcements that meet the following criteria: (i) Firms that are rated must be publicly listed; (ii) At least 60 days of trading information 30 days prior to actual announcements for estimating market model; and (iii) No other corporate announcements 2 days before and after the date of credit rating announcement. These criteria resulted in 340 rating announcements comprising of 109 up rating change, 15 down rating change and 216 rating remained unchanged. 9

10 Our source of daily stock prices and stock market index data is Dhaka Stock Exchange s library database. We have used two indices, Dhaka Stock Exchange General Index (DGEN) for the period between 1/1/2004 and 07/31/2013 and DSE Board Index (DSEX) between the period 08/01/2013 and 11/3/2015. DSEX is the Broad Index of the Exchange (Benchmark Index) which reflects around 97% of the total equity market capitalization. The main difference between these two indices is instead of using all the shares outstanding as in full market capitalization, which DGEN follows, the free-float method excludes locked-in shares held by promoters and governments, which DSEX follows. 3.2 Methodology Our approach to empirical investigation builds on event study methodology. 2 One of the critical measures of price reaction is the estimation of abnormal return around the event days. We have measured abnormal returns AR i,t using standard market model. For any security i, the market model is estimated by the following equation: R i,t = α i + β i R m,t + ε i,t (1) To determine abnormal returns for any stock i, we subtract the actual returns and the returns predicted by the market model using the parameter. AR i,t = R i,t (α i + β ir m,t ) (2) Based on the estimates of abnormal returns of each stock, we calculated cross-sectional average abnormal returns (AAR t ) and calculated cumulative average abnormal returns (CAAR t ) depending on the event windows [τ 1 and τ 2 ] selected for investigation. 2 See MacKinlay (1997) for more on event-study methodology. 10

11 τ 2 CAAR t = t=τ 1 AAR t (3) Another important element in event study is to estimate the statistical significance of CAAR t. We have used a battery of test statistic to check the significance. These test statistics assume different statistical properties of abnormal return in the event widow to address potential bias. We tested the significance of CAAR t by following Patell s (1976) test statistics. In Patell s test, the event periods abnormal returns are standardized by the standard deviations of the estimation period abnormal returns. This estimation decreases the effects of large return standard deviations on the test. The assumption of this test is that there is no event-induced change in the variance of the event period abnormal return. The test statistic is calculated as follows: t Patell = N i=1 SR 1 N D i 2 i=1 D i 4 Where N is the number of stocks in the portfolio, D i is the number of observations in stock i s estimation period, and SR i is the standardized abnormal returns of the i th stock, calculated by dividing the event period abnormal return on the i th stock on the day t by the standard deviation of the estimation period abnormal returns. By following the literature, we also checked the test statistics with cross-sectional test estimates. The assumption of this test is that there is no cross-sectional dependence in abnormal returns, but it allows the event-induced variance changes. The cross-sectional test is calculated by dividing the mean abnormal return during the event period by its contemporaneous cross-sectional standard error. The test statistics is estimated as follows: 11

12 t CS = 1 N N i=1 AR i 2 i=1 ] N N(N 1) N [AR i 1 N N i=1 AR i Where AR i is the abnormal returns of the i th stock. 4. Empirical Results The original rating data collected from CRISL includes 586 rating announcements between 2002 and Table 3, Panel A reports the year-wise distributions of all rating announcements. The number of ratings conducted by the rating agency increased over the period. One-quarter of the ratings are done in first half of the periods (between 2002 and 2008), whereas the other threequarters are done in the years 2009 to The increase in rating activities reflects changes made through regulatory reforms. Panel B of Table 3 shows the industry-wise distribution of all rating announcements. The data indicates that CRISL conducts rating in all the industrial sectors in Bangladesh economy. Not surprisingly, around 61% of all the ratings are done in financial sectors including bank, insurance and other financial institutions, since these financial institutions are the early adopters of ratings. Insurance is the largest sector with 158 ratings (27%), followed by banking (135 ratings, 23%) and other financial institutions (62 ratings, 11%). Among all nonfinancial industries, the textile is the largest industry where CRISL rated different textile companies 58 times (10%) during between 2002 and On Table 3, Panel C, we have classified all rating announcements into three major groups of ratings, i) investment grade, ii) speculative grade, iii) default grade and reported the year wise distribution of all rating announcements. These tables show that majority of the companies are 12

13 rated as investment grade (568 ratings, 96.93%). Only 2.9% rating (17 ratings) is defined as speculative grade, whereas the default grade is 0.17% (only 1 rating). Bangladesh Bank regularly monitors the consistency of ratings by major rating agencies. The metric used by the regulators is cumulative default rate (CDR) indicates the movement of a rating category into default rating during a particular period. Although it might suggest that CRISL issues a higher percentage of investment grade, throughout these periods CRISL did not exceed either monitoring level or trigger level CDR, which might have caused a downgrade of rating companies ECI status. Although there is measure in place to prevent rating shopping by clients, higher propensity to issue investment grade rating is partially reflective of the client-pays business model following in Bangladesh. The up and down rating changes are defined by a consecutive change in rating from one year to previous. Due to sticky nature of the credit rating data, in cases, if we don t have rating data for two consecutive years, we have compared the ratings of three years before the current year to determine a change in rating. Following the rating change criteria, our original sample of 540 rating announcements reduced to 340 rating change announcements between 2003 and Table 4 shows the breakdown of our sample. Out of 340 rating change announcements, 109 ratings have been upgraded, 15 have been downgraded, and 216 remained unchanged. Further classification shows that out of 109 upgraded ratings, 90 are single-step change, whereas 19 are multiple-step changes. Downgraded rating includes 13 single-step change and two multiple-step change and also the majority of the rating changes happened in the financial, bank and insurance sectors. Table 5 Panel A shows the transitions for the upgraded changes. The majority of the upgraded transition occurs within the investment grade ratings. 17 upgraded changes are from A to A+, another 17 upgraded changes are from A- to A, and 15 upgraded changes are from A+ to 13

14 AA-. 8 upgrade ratings show that companies improved their riskiness from speculative to investment grade. Out of 340 rating announcements, 15 ratings are downgraded, which is only 4.4% of the sample. However, all of the downgrading, except one, are within the investment grade classification. Only one downgrading revisions moved a company from the investment grade to the speculative grade category. Due to non-availability of stock market data between 2000 and 2004, beyond 2015, our final sample file shrank to 82 up announcements, 9 down announcements, and 147 unchanged rating announcements. We defined each rating change announcements as event (t=0) and our event window under investigation is 30 trading days pre-and post-event dates. We have used 90 days daily returns (log-differenced stock prices) 30 days prior to event dates to estimate the market model. According to our results from Table 6, there is a strong evidence of positive abnormal return from an upward change in credit rating at the announcement day. It seems the investors on an average make between 2.0% to 3.5% abnormal returns within 3 days following the announcements. There is no evidence of significant average abnormal returns prior to rating change announcements. Collectively the result supports underreaction hypothesis, following rating change announcements, investors quickly absorb the incremental information disseminated, and the market seems to underreact initially then rebounds with t+3 days. This is also no evidence of information leakage so on an average, and the investors do not make any abnormal returns. The rating downgrade announcements give a mixed result (Table 6, panel b). At the rating downgrade announcements, there is evidence of negative average returns although not statistically significant. Following rating announcements, this is statistically significant evidence of positive average abnormal return according to Patell s t statistic. However, these t statistics could produce bias results in a small 14

15 sample. Consistent with our expectation, when there is no change in rating, there is no incremental information conveyed by the rating agencies and there is also no significant abnormal return observed in the market. 5. Conclusion The financial system in Bangladesh has gone through several reforms over the past three decades. The introduction of credit rating is considered one of the major milestones for the regulators and investors in the capital market. The predominance of unsophisticated or less sophisticated investors sees rating announcements as an additional source of information available to them to gauge the overall health of the companies listed in Dhaka Stock Exchange. In this regard our research finds evidence suggesting that any upward rating revision conveys good information there is significant abnormal return at the announcements. The market seems to underreact to this types of announcements the opportunity to make abnormal return fades off within three days from announcements. Our findings offer a slightly different view of what we know from overall evidence from the western countries which suggest rating upward announcements contains no information content. This striking difference might be arising from the institutional difference between western markets and frontier markets like Bangladesh (see Amin and Orlowski (2014)) and also because of difference is corporate practices persists between these markets. Particularly in western markets, investors can acquire information from different avenues e.g. analyst coverage, media coverage, management s forecast, earnings calls by managements, etc. Further, each company has multiple financial assets traded publicly that also allows the investors to gather information about the company. In Bangladesh, the primary sources of information are through semiannual and annual reports. Due to opaque nature of the market, financial regulators introduced credit rating and made it mandatory for all listed companies to be annually rated. Although a 15

16 downgrade announcement typically conveys negative information in the market, and in developed markets this type of announcements mostly informative, we find only limited evidence of significance down rating announcement for Dhaka Stock Exchange. Small sample bias might have driven this particular result, so we interpret this result cautiously. Finally consistent with our expectation, when there is no change in rating announcement, there is no reaction to such information. Overall rating announcements increase transparency in the capital markets and over time we expect that the in Dhaka Stock Exchange will be able to make even better use of rating announcements. 16

17 References Abad-Romen, P., & Robles-Fernandez, M. (2006). Risk and return around bond rating changes: New evidence from the Spanish stick market. J. Business Finance Account, 33(5/6), Afik, Z., Feinstein, I., & Galil, K., (2014). The (un)informative value of credit rating announcements in small markets. Journal of Financial Stability, 14, Amin, a., Orlowski, L., (2014). Returns, volatilities, and correlations across mature, regional and frontier markets: Evidence from South Asia. Emerging Markets Finance and Trade, 50(3), Barron, M. J., Andrew, D. C., & Stephen, T. H. (1997). The effect of bond rating changes and new ratings on UK store returns. Journal of Business Finance and Accounting, 24, Dimitrov, V., Palia, D., & Tang, L., (2015). Impact of Dodd-Frank act on credit rating. Journal of Financial Economics, 115, Ederington, L. H., & Goh, J. C. (1993). Is a bond rating downgrade bad news, good news, or no news for stockholders? Journal of Finance, 48, Elayan, F. A., Hsu, W., & Meyer, T. O. (2003). The informational content of credit rating announcements for share prices in a small market. Journal of Economics and Finance (forth coming). Hand, R., Holthausen, R., & Leftwich, R., (1992). The effect of bond rating agency announcements on bond and stock prices. Journal of Finance, 47,

18 Holthausen, R., & Leftwich, R., (1986). The effect of bond rating changes on common stock prices. Journal of Financial Economics, 17, Jorion, P., & Zhang, G. (2007). Information effects of bond rating changes: The role of the rating prior to the announcement. The Journal of Fixed Income, 16(4), Jorion, P., Liu, Z., & Shi, C. (2005). Informational effects of regulation FD: Evidence from rating agencies. Journal of Financial Economics, 76, Katz, S. (1974). The price adjustment process of bond to rating reclassifications: A test of bond market efficiency. Journal of Finance, 29, Kliger, D., & Sarig, O. (2000). The information value of bond ratings. Journal of Finance, 55, Li, J., Shin, Y. S., & Moore, W. T. (2006). Reactions of Japanese markets to changes in credit ratings by global and local agencies. J. Bank. Finance, 30(3), MacKinlay, A. C. (1997). Event studies in economics and finance. Journal of Economic Literature, 35, MacKinlay, A., (1997). Event studies in economics and finance. Journal of Economic Literature, 35, Matolcsy, Z. P., & Lianto, T. (1995). The incremental information content of bond rating revisions: The Australian evidence. Journal of Banking and Finance, 19, Morris, S., (2001). Political correctness. Journal of Political Economy, 109, Patell, J., (1976). Corporate forecasts of earnings per share and stock price behavior: Empirical test. Journal of Accounting Research, 14(2),

19 Pinches, G. E., & Singleton, J. C. (1978). The adjustment of stock prices to bond rating changes. Journal of Finance, 33, Poon, W. P. H., & Chan, K. C. (2008). An empirical examination of the informational content of credit ratings in China. J. Business Res., 61(7), Wakeman, M. (1981). The real function of bond rating agencies. Chase Financial Quarterly, 1, Weinstein, M. I. (1977). The effect of change announcement on bond price. Journal of Financial Economics, 6,

20 Table 1 Summary of Market Size Information of Dhaka Stock Exchange Year End Date Total Number of Trades Total Trade Volume Total Traded Value in Taka (million) Total Market Capitalization in Taka (million) Dec-03 8,721 3,226, , Dec-04 9,984 3,540, , Dec-05 10,080 1,675, , Dec-06 18,213 13,257, , Dec-07 36,222 5,775,761 1, , Dec-08 92,350 29,109,210 4, ,043, Dec ,907 30,586,541 9, ,903, Dec , ,464,146 17, ,508, Dec , ,952,716 5, ,616, Dec-12 69,550 69,257,597 2, ,403, Dec-13 91,634 94,221,217 3, ,647, Dec-14 68,764 65,694,198 2, ,259, Dec-15 96,535 97,047,649 4, ,159,

21 Table 2 CRISL Credit Rating Scales and Definition Category Risk Level Definition Ratings Highest Safety Entities rated in this category are adjudged to be of best quality, offer highest safety and have highest credit quality. Risk factors are negligible and risk free, nearest to risk free Government bonds and securities. Changing economic circumstances are unlikely to have any serious impact on this category of companies. AAA Investment Grade High Safety Adequate Safety Entities rated in this category are adjudged to be of high quality, offer higher safety and have high credit quality. This level of rating indicates a corporate entity with a sound credit profile and without significant problems. Risks are modest and may vary slightly from time to time because of economic conditions. Entities rated in this category are adjudged to offer adequate safety for timely repayment of financial obligations. This level of rating indicates a corporate entity with an adequate credit profile. Risk factors are more variable and greater in periods of economic stress than those rated in the higher categories. AA+ AA AA- A+ A A- Moderate Safety Entities rated in this category are adjudged to offer moderate degree of safety for timely repayment of financial obligations. This level of rating indicates that a company is under-performing in some areas. Risk factors are more variable in periods of economic stress than those rated in the higher categories. These entities are however considered to have the capability to overcome the above-mentioned limitations BBB+ BBB BBB- 21

22 Category Risk Level Definition Ratings Inadequate Safety Entities rated in this category are adjudged to lack key protection factors, which results in an inadequate safety. This level of rating indicates a company as below investment grade but deemed likely to meet obligations when due. Overall quality may move up or down frequently within this category. BB+ BB BB- Speculative Grade Risky Vulnerable Entities rated in this category are adjudged to be with high risk. Timely repayment of financial obligations is impaired by serious problems which the entity is faced with. Whilst an entity rated in this category might be currently meeting obligations in time through creating external liabilities. Entities rated in this category are adjudged to be vulnerable and might fail to meet its repayments frequently or it may currently be meeting obligations in time through creating external liabilities. Continuance of this would depend upon favorable economic conditions or on some degree of external support. B+ B B- CCC+ CCC CCC- High Vulnerable Entities rated in this category are adjudged to be very highly vulnerable. Entities might not have required financial flexibility to continue meeting obligations; however, continuance of timely repayment is subject to external support. CC+ CC CC- Extremely Speculative Entities rated in this category are adjudged to be with extremely speculative in timely repayment of financial obligations. This level of rating indicates entities with very serious problems and unless external support is provided, they would be unable to meet financial obligations. C+ C C- Default Grade Default Entities rated in this category are adjudged to be either already in default or expected to be in default. D Source: CRISL Credit Rating Information and Services Limited s website. 22

23 Table 3 Panel A Year-Wise Classification of all Ratings Year Frequency Percent Cumulative Frequency Cumulative Percent % 4 0.7% % % % % % % % % % % % % % % % % % % % % % % % % % % 23

24 Industry Table 3 Panel B Industry-Wise Classification of all Ratings Frequency Percent Cumulative Frequency Cumulative Percent Bank % % Cement % % Ceramics Sector 1 0.2% % Engineering % % Financial Institutions % % Food & Allied % % Fuel & Power % % IT Sector % % Insurance % % Jute 2 0.3% % Miscellaneous % % Paper & Printing 1 0.2% % Pharmaceuticals & Chemicals % % Services & Real Estate 3 0.5% % Tannery Industries % % Telecommunication 3 0.5% % Textile % % Travel & Leisure 7 1.2% % 24

25 Table 3 Panel C Year-Wise Risk Assessments for All Ratings Year Investment Grade Speculative Grade Default Grade Total Total Percentage 96.93% 2.9% 0.17% 100% 25

26 Table 3 Panel D Industry-Wise Risk Assessments for All Ratings Sector Investment Grade Speculative Grade Default Grade Total Bank Cement Ceramics Sector 1 1 Engineering Financial Institutions Food & Allied Fuel & Power IT Sector Insurance Jute 2 2 Miscellaneous Paper & Printing 1 1 Pharma & Chemicals Services & Real Estate 3 3 Tannery Industries Telecommunication 3 3 Textile Travel & Leisure 7 7 Total Percentage 96.93% 2.9% 0.17% 100% 26

27 Table 4 Breakdown of the Sample Sample Upgrades Downgrades Unchanged Final Sample Type of Change Single-Step Change Multiple-Step Change Total Major Industry Breakdown Financial, Bank and Insurance Non-Financial Industries Total

28 Prior Rating AAA Table 5 Panel A Transition Matrix of Upward Rating Change Revised Rating AAA AA+ AA AA- A+ A A- BBB+ BBB BBB- BB+ B+ AA+ 2 2 AA 6 6 AA A A A BBB BBB BBB BB- 1 1 BB B B 1 1 Total Total

29 Prior Rating Table 5 Panel B Transition Matrix of Downward Rating Change Revised Rating AA AA- A+ A A- BBB+ BBB- BB- Total AA+ 1 1 AA 1 1 AA- 3 3 A+ 1 1 A A BB+ 1 1 Total

30 Table 6 Panel A Stock Price Reaction to Upward Rating Change Rating Announcement Windows N Average Abnormal Return % of Positive Abnormal Return Patell s t statistics Cross Sectional t statistics (-3 to 0) % 53.01% (-2 to 0) % 46.99% (-1 to 0) % 50.60% (0 to 0) % 55.42% * *** (0 to +1) % 65.06% ** *** (0 to +2) % 57.83% *** ** (0 to +3) % 55.42% ** Table 6 Panel B Stock Price Reaction to Downward Rating Change Rating Announcement Windows N Average Abnormal Return % of Positive Abnormal Return Patell s t statistics Cross Sectional t statistics (-1 to 0) % 55.56% (-2 to 0) % 66.67% (-3 to 0) % 66.67% (0 to 0) % 33.33% (0 to +1) % 66.67% * (0 to +2) % 66.67% * (0 to +3) % 55.56% **

31 Table 6 Panel C Stock Price Reaction to Rating Unchanged Rating Announcement Windows N Average Abnormal Return % of Positive Abnormal Return Patell s t statistics Cross Sectional t statistics (-3 to 0) % 48.03% (-2 to 0) % 49.34% * (-1 to 0) % 47.37% (0 to 0) % 50.66% (0 to +1) % 46.05% (0 to +2) % 51.97% (0 to +3) % 52.63%

32 1