The impact of credit rating announcements on banks: evidence from the United States

(1)

The impact of credit rating announcements on

banks: evidence from the United States

Jarno Herbers S2943778

MSc Finance

Supervisor: dr. E.L. Kramer

9 January 2020 University of Groningen Faculty of Economics and Business

Abstract

This paper examines the informational value of credit rating announcements published by Standard & Poor’s for banks listed on the S&P500 during the period 2000 to 2018, by studying daily stock return data around the announcement date. An event study approach is used to investigate how banks’ share prices respond to credit rating announcements. The results of this study suggest that new information is provided to investors after credit rating changes, where upgrade credit rating announcements result in positive abnormal returns and downgrade credit rating announcements in negative abnormal returns. The results also argue that the impact of credit rating announcements differ per time period and that there is a higher response to upgrade credit rating changes after the financial crisis in 2008.

Keywords: Credit ratings, credit rating announcements, S&P, banks, financial crisis, event

study, abnormal returns

(2)

2

1. Introduction

This study aims to shed light on the effect of credit rating announcements on stock prices of banks listed on the S&P 500, which is the stock market index that measures the stock performance of 500 largest companies listed on stock exchanges in the United States. Corporate credit ratings are independent evaluations of a firm’s ability to make debt payments in the future and can act as a signal of the quality of a particular debt issue, which may indicate the overall credit worthiness of a firm. The ratings are based on two information sources, publicly available information (such as audited financial statements) and non-publicly available information (such as internal reports or budget forecasts). Through this information credit rating agencies (CRAs), such as Standard & Poor’s, Moody’s and Fitch Group, assign credit ratings for issuers of different types of debt obligations based on the above-mentioned ability to repay debt and their probability to default. The agencies are independent actors in the capital market and with the help of credit ratings investors can get an indication of credit risk for a specific firm listed on the index, since the agencies reduce the information asymmetry between a firm’s management and its investors. Investors are expected to react upon a change in credit ratings as a higher credit rating lowers the probability of default of a firm and vice versa (Micu et al., 2007). Good news from a credit rating agency (upgrade in the rating) is expected to result in higher market returns, since it signals that the firm has increased in investment quality. On the other hand, bad news (downgrade in the rating) is expected to result in lower market returns, since it signals a firm’s decrease in investment quality. However, the informational value of credit rating agencies is a controversial and inconclusive issue. It is difficult to establish a clear impact of credit ratings on stock returns as findings on the effects of credit rating announcements differ (Ammer & Clinton, 2004).

(4)

4 reaction to the relevant information. Elayan et al. (2003) found that upgrade credit rating announcements were associated with a positive and statistically significant market reaction from New Zealand share prices. Barron et al. (1997) investigated the UK market and found significant responses toward downgrades, in addition to upgrades. Research on the Australian stock exchange only found effects from downgrades (Choy et al., 2006; Matolcsy & Lianto, 1995), whereas Li et al. (2004) and Amin et al. (2018) suggested that the stock market responds only positively to upgrade credit rating announcements in Sweden and Bangladesh and it does not respond to downgrade credit rating announcements.

Several studies investigated whether there is a dissimilar effect of credit rating changes on banks, because they are high regulated entities compared to corporate firms. Schweitzer et al. (1992) indicated that the regulation of an industry can increase the amount of available information to the market, as the primary concern of the regulatory agencies is maintaining the safety and soundness of the industry. From their results they did not obtain a significant difference in the impact of credit rating changes in comparison to corporate firms for upgrade credit rating announcements. For downgrades, banks seemed to respond significantly more than corporate firms, which supports the belief that bank regulators hold back confidential information from the market. Gropp & Richards (2001) examined credit rating announcements of European banks and they suggested that credit rating agencies perform a useful role in summarizing and obtaining non-public information on banks. Lobão et al. (2019) also suggested that rating agencies provide new information to capital markets through credit rating announcements of EU bank’s debt, with either upgrades or downgrades.

(5)

5 reactive to credit rating changes in the aftermath of a financial crisis by studying Korean rating announcements, where the Korean financial crisis was assumed to have taken place between October 1997 and July 1998 (Choe et al., 1999). Higher market reaction to credit rating announcements is explained by different sensitivity towards these announcements as the informational value and uncertainty became larger during financial instability (Micu et al., 2004). Therefore, stock prices are expected to act differently to credit rating changes after the financial crisis.

(6)

6

2. Literature Review

This section provides a discussion on the fundamental literature relevant to the effect of credit rating announcements on banks in the United States. The theoretical framework from previous studies conducted on the US market and outside of the US market will be discussed and by using these studies, a comprehension of how credit rating changes are affecting bank’s share prices in the United States can be outlined.

Wide-ranging research have examined the effect of stock and bond prices to changes in credit ratings. There are many different results of the responses to up- and downgrades in credit ratings. Weinstein (1977) investigated corporate bond price behaviour during the period around the announcement of a credit rating change. He found no evidence of any reaction six months before the credit rating announcements and little evidence during the change or six months afterwards. He argued that the credit rating changes are not expected to reveal new information to the market. Pinches & Singleton (1978) studied the effect of bond rating changes on stock prices using the monthly return data for twenty years, from 1950 to 1972, and they found no significant up- or downward drift in their cumulative abnormal returns in the month before or after a bond rating change. Additionally, Kaplan & Urwitz (1979) developed a statistical model using a subordinated dummy variable, total assets, one financial ratio, and the common stock market beta coefficient to correctly classify newly issued bonds. They argued that this simple linear model may be predicting the actual risk of a bond better than the credit rating agencies, in their case Moody’s, which raises the question whether rating agencies can outperform the model.

(7)

7 argued that not all rating downgrades are bad news for shareholders, since downgrades indicate shifts of wealth from bond- to shareholders as a result of changes in financial leverage. Kliger & Sarig (2000) tested whether bond ratings contain pricing-relevant information by examining security price reactions to Moody’s refinement of its rating system. They found that rating information does not affect firm value, but that debt value increases and equity value falls when Moody’s announced better than expected ratings. It was the other way around if Moody’s announced worse than expected ratings, which decreases the debt value and increases the equity value.

Previous mentioned studies are all based on the US market, where there also have been many studies that investigated the effect of credit rating announcements outside the United States. Matolcsy & Lianto (1995) examined bond rating changes in the Australian stock market and their results showed that there were only significant abnormal returns for downgraded bonds and non-significant returns for upgraded bonds based on Standard & Poor’s weekly stock returns from 1982 to 1991. Choy et al. (2006) also investigated the Australian stock market and they found the same results in their sample of credit rating changes of all Australian companies rated by Moody’s and Standard & Poor’s for the period 1989 to 2003. Both results from Australia are consistent with the evidence documented for US firms that only downgrades lead to a significant market reaction. A study based on the UK stock market was conducted by Barron et al. (1997) using daily data from 1984 to 1992. They tested the effect for long and short-term debt and found significant excess stock returns associated with bond rating downgrades and positive CreditWatch announcements. Their conclusion was that credit rating announcements provide information to the capital market in the United Kingdom, but having a credit rating could not decrease the cost of equity capital of a firm.

(8)

8 (2003). Amin et al. (2018) concluded from their investigation in the stock market of Bangladesh that upgrade credit rating announcements offer additional information that cannot be obtained by market participants from other sources. Therefore, upgrades in credit ratings are associated with positive market reactions. Despite the overwhelming support in previous mentioned literature, Li et al. (2004) and Amin et al. (2018) both did not find significant evidence of a negative reaction to downgrade credit rating announcements in their samples from Sweden and Bangladesh, indicating that the market had already anticipated to the information provided by the rating agencies.

(9)

(10)

10

3. Hypotheses

In this section the hypotheses based on the literature mentioned in the previous section will be formulated, where two hypotheses are considered as most crucial to explain how banks listed on the S&P500 respond to credit rating announcements.

The mixed results from previous mentioned studies in the literature review raises the question whether credit rating announcements are associated with valuable information to investors such that it impacts bank’s share prices in the United States? If there is an impact, how do bank’s share prices react to different types of credit rating announcements and did this response change after the financial crisis in 2008? This thesis concentrates on the effect of credit rating announcements by studying how banks listed on the S&P500 behave on credit rating changes. The announcements will be divided in three categories: pre-crisis period, crisis period and post-crisis period. The pre-post-crisis period includes credit rating announcements published from January 2000 to August 2008, the crisis period contains credit rating changes published from September 2008 to July 2009 and the post-crisis period includes credit rating announcements published in the years after the financial crisis from August 2009 to December 2018. From the research questions the following hypotheses will be tested:

Hypothesis 1: Credit rating announcements generate significant abnormal returns for banks

listed on the S&P 500, where upgrade (downgrade) credit rating announcements will be associated with statistically significant positive (negative) abnormal returns.

(11)

11 Hypothesis 2: The impact of credit rating announcements on banks listed on the S&P500

became larger for up- and downgrade rating changes published after the financial crisis.

(12)

12

4. Data and Methodology

This section describes the data and methodology used to conduct the research. It amplifies on the restrictions with respect to the dataset and summarizes the descriptive statistics of this dataset. This section also delves deeper into the methodology based on MacKinlay (1997), which explains the framework of an event study and provides the cross-sectional test and multivariate regression analysis.

4.1 Data

(13)

13 Table 1 gives an overview of the credit rating announcements per time period as explained in the previous section. As can be seen from the table, a relatively large part of the downgrade announcements was published by Standard & Poor’s in a short period of time from September 2008 to July 2009. This was due to the global financial crisis at that time which resulted in many downgrade rating changes for banks. Upgrade credit rating announcements, on the other hand, were mostly published in the years before or after the financial crisis, and only one upgrade credit rating announcement was published by Standard & Poor’s during the crisis period.

Table 1. Summary characteristics of the credit rating announcements.

Downgrades Upgrades Time period # % # % Pre-crisis 11 16,2 28 52,8 Crisis 28 41,2 1 1,9 Post-crisis 29 42,6 24 45,3 Total 68 100 53 100

The sample comprises 121 credit rating announcements published by Standard & Poor’s for 32 banks listed on the S&P500 for the period 2000 to 2018. Time periods in the table are denoted as follows: pre-crisis period includes credit rating announcements published from Jan. 2000 to Aug. 2008, crisis period from Sept. 2008 to July 2009, and post-crisis period from Aug. 2009 to Dec. 2018.

4.2 Methodology

4.2.1 Event study framework

An event study is a convenient method to measure the effect of an economic event on the firm’s asset prices (MacKinlay, 1997), and is widely used in studies regarding the impact of credit rating announcements. The events of interest in this study are credit rating announcements of banks listed on the S&P500, where the event day implies the day in which a credit rating change is announced by Standard & Poor’s. The stock price impact of credit rating announcements is estimated using abnormal returns from the single-index market model from Fama (1976), and is presented in the following equation:

𝑅_𝑖𝑡 = 𝛼_𝑖 + 𝛽_𝑖𝑅_𝑚𝑡+ 𝜖_𝑖𝑡 (1)

Where Rit and Rmt are the returns in time t for observation i and the market index m and ϵit is the zero mean disturbance term. The market index on which all banks are trading, the S&P500, will be used as the control group for Rmt. The market model is a statistical model that relates

(14)

14 compared to other models such as the CAPM, mean- and market-adjusted model. Even though these other models perform approximately as well as the market model in numerous studies, the market model remains the most commonly used approach. The mean- and market-adjusted models are, in comparison to the market model, simplified by using assumptions which can be valuable when information is scarce or limited. The capital asset pricing model (CAPM) imposes an additional restriction with intercept αi being equal to the risk-free rate. This

restriction may cause a larger error variance in the formula which could result in a less powerful test than the market model.

The returns for the stock prices and the market index are calculated using the following log function:

𝑅_𝑖𝑡 = 𝑙𝑛 𝑃𝑖𝑡

𝑃𝑖𝑡−1 (2)

With Rit as the normal return of security i at time t, Pit as the closing price on the investigated

day and Pit-1 as the closing price on the prior day. To estimate the impact of credit rating

announcements, a measure of the abnormal returns is required. The abnormal return (AR) for each security i on day t is the deviation of security i’s realized return (Rit) from an expected

return generated by the market model:

𝐴𝑅_𝑖𝑡 = 𝑅_𝑖𝑡− (𝛼_𝑖 + 𝛽_𝑖𝑅_𝑚𝑡) (3)

The coefficients 𝛼_𝑖 and 𝛽_𝑖 are the ordinary least squares (OLS) estimates of security i’s market model parameters obtained from an estimation period of 120 days prior to the announcement date, where the event window itself is not included, and are calculated as follows:

𝛼𝑖 = 𝑅̅ − 𝛽𝑖 𝑖𝑅̅̅̅̅ 𝑚 (4)

𝛽_𝑖 = 𝐶𝑜𝑣 [𝑅𝑖,𝑅𝑚]

𝑉𝑎𝑟 [𝑅𝑚] (5)

(15)

15 the post-announcement event window. The abnormal performance is measured over individual trading days and over specified intervals of trading days in the event windows. To improve the information content of the analysis, the unweighted cross-sectional average of abnormal returns (AAR) with N as the number of events in sample period t is calculated as:

𝐴𝐴𝑅𝑡 = 1 𝑁 ∑ 𝐴𝑅𝑖𝑡 𝑁 𝑖=1 (6) 4.2.2 Testing procedure

To investigate abnormal returns, they must be aggregated in order to draw overall interpretations on the effect of credit rating announcements. For that reason, the cumulative abnormal return (CAR) is calculated, which is the sum of all abnormal returns. The cumulative abnormal return, where the abnormal returns are aggregated from the start of the event window t1 up to time t2, is defined as:

𝐶𝐴𝑅_{𝑖[𝑡1,𝑡2]} = ∑𝑡2 𝐴𝑅_𝑖𝑡

𝑡=𝑡1 (7)

In event studies the CARs are collected over the cross-section of events to get the cumulative average abnormal returns (CAAR), which is a useful statistical tool to better understand the abnormal return evolution:

𝐶𝐴𝐴𝑅_{[𝑡1,𝑡2]} = ∑𝑡2 𝐴𝐴𝑅_𝑡

𝑡=𝑡1 (8)

To test the hypothesis that the mean abnormal return on day t of the event window is equal to zero, a parametric cross-sectional test is used, and the test statistic is given by:

𝑡𝑠𝑡𝑎𝑡𝑖𝑠𝑡𝑖𝑐 = √𝑁 𝐴𝐴𝑅𝑡

𝑆_{𝐴𝐴𝑅𝑡} (9)

Where the standard deviation across the sample is calculated as: 𝑆𝐴𝐴𝑅𝑡 = √ 1 𝑁−1∑ (𝐴𝑅𝑖𝑡 − 𝐴𝐴𝑅𝑡) 2 𝑁 𝑖=1 (10)

The same cross-sectional test is done to examine whether the cumulative average abnormal return is equal to zero in multiple event windows:

𝑡_{𝑠𝑡𝑎𝑡𝑖𝑠𝑡𝑖𝑐} = √𝑁𝐶𝐴𝐴𝑅_𝑆 [𝑡1,𝑡2]

𝐶𝐴𝐴𝑅 (11)

Where the standard deviation is given by: 𝑆_{𝐶𝐴𝐴𝑅} = √ 1

𝑁−1∑ (𝐶𝐴𝑅𝑖[𝑡1,𝑡2]− 𝐶𝐴𝐴𝑅[𝑡1,𝑡2]) 2 𝑁

(16)

16

4.2.3 Multivariate regression

A multivariate regression equation is estimated to explain cross-sectional variation in the window-spanning abnormal returns. The regression is based on firm characteristics and dummy variables. Table 2 explains all definitions of the variables used in the regression analysis. Firm size is the first characteristic in the regression, and it represents the theory that small firms should have larger abnormal returns than large firms. The effect of the change in the stock prices would be stronger for small firms if the abnormal returns arise due to an incomplete market reaction to credit rating announcements (Beard & Sias, 1997; Bernard & Thomas, 1990; Fama, 1998). The size of the firms is calculated by taking the natural logarithm of the market capitalization of the bank. The other variable in the regression is the book-to-market ratio, which is the book value of equity divided by the market value of equity. Book-to-market ratios are used to find a firm’s value by comparing its book value to its market value. If the market value of a firm is trading higher than its book value per share, it is overvalued. Analysts consider a firm to be undervalued if the book value is higher than the market value. Based on the research and accumulating evidence of Barber & Lyon (1997), firm size and book-to-market ratio explain cross-sectional variation in stock returns in an economically meaningful way. Leverage is not used as a variable in the analysis because high leverage is normal for banks and it probably does not have the same meaning as for non-financial firms, where high leverage more likely indicates distress (Fama & French, 1992).

(17)

17 calculated to avoid multicollinearity, where the rule of thumb is used that if the correlation is larger than 0,8 or lower than -0,8 then severe multicollinearity may be present (Brooks, 2014). In the regression analysis no correlations larger than 0,8 or lower than -0,8 between the variables are found.

The dependent variable in the regression analysis will be the absolute value of the abnormal return at the announcement date (t = 0), as not the positive or negative sign of the values but the magnitude of its impact on the share prices are examined in the analysis. The multivariate regression analysis of the abnormal returns for credit rating announcements is done by the following equation:

|𝐴𝑅₀| = 𝛽₀+ 𝛽₁𝑆𝑖𝑧𝑒 + 𝛽₂ 𝐵𝑜𝑜𝑘

𝑀𝑎𝑟𝑘𝑒𝑡+ 𝛽3𝐶ℎ𝑎𝑛𝑔𝑒 + 𝛽4𝐶𝑟𝑖𝑠𝑖𝑠 + 𝛽5𝑃𝑒𝑟𝑖𝑜𝑑 + 𝛽6𝐺𝑟𝑎𝑑𝑒 (13)

Table 2. Definition of the variables in the multivariate regression analysis.

Variable Definition

Size Natural logarithm of the market value of the company.

Book-to-Market Common shareholders’ equity divided by the market capitalization.

Change Dummy with value 1 (0), if the credit rating change was a downgrade

(upgrade).

Crisis Dummy with value 1 (0), if the credit rating change was published during the financial crisis (before/after financial crisis).

Period Dummy with value 1 (0), if the credit rating change was published in the

post-crisis period (pre-crisis period).

Grade Dummy with value 1 (0), if the credit rating is described as investment

grade (speculative grade).

(18)

18

5. Results

This section provides an overview of the empirical analysis conducted on the data and based on the methodology which is presented in section 4.

5.1 Credit rating announcements

The following graphs present the average abnormal returns for downgrade and upgrade credit rating announcements for the event window from t-10 to t+10. Figure 1 shows small positive and negative average abnormal returns on the days prior to the announcement date, a stronger negative average abnormal return on the event day itself, and small positive and negative average abnormal returns on the days after the announcement date. Only the average abnormal return on day t-6 shows a significance at the 5% level, all other days before or after the downgrade credit rating announcement in the event window do not show strong significant average abnormal returns. The announcement date has a negative average abnormal return of 2,53% and is significant at the 1% level. Therefore, the average abnormal return on the announcement date is significantly different from zero.

(19)

19 Figure 2 shows small positive and negative average abnormal returns on the days before and after the upgrade announcement date, where day t-8, t+2 and t+7 are significant at the 5% level. All other average abnormal returns in the event window from t-10 to t+10 are insignificant and thus, they are not significantly different from zero. Also, the average abnormal return on the event day when the upgrade credit rating announcement is published is not significant, which could imply that upgrades have no informational value to investors.

Figure 2. Average abnormal returns for upgrades in the event window [-10 to 10].

(20)

20 Piotroski (2001) who all found a significant negative reaction of the market to downgrades in credit ratings.

Table 3. CAARs and t-statistics for all credit rating announcements.

Downgrades Upgrades

Event window CAAR (%) t-statistic CAAR (%) t-statistic

[-10 to 10] -0,63 -0,22 1,10 1,29

[-5 to 5] -1,52 -0,61 1,18 2,48***

[-1 to 1] -2,43 -2,88*** 0,44 1,16

[-10 to -1] 0,32 0,18 -0,30 -0,44

[0 to 10] -0,96 -0,57 1,39 2,10**

Table 2 shows the cumulative average abnormal return (CAAR) and their t-statistic for multiple event windows. *** p<0.01, ** p<0.05, * p<0.1

5.2 Financial crisis

5.2.1 Pre-crisis period

(21)

21

Table 4. CAARs and t-statistics for the pre-crisis credit rating announcements.

Downgrades Upgrades

[-10 to 10] -4,77 -0,84 -0,10 -0,08

[-5 to 5] -8,75 -1,14 0,48 0,72

[-1 to 1] -6,54 -1,72* 0,64 1,49*

[-10 to -1] -0,22 -0,08 -0,22 -0,21

[0 to 10] -4,54 -1,10 0,12 0,16

Table 3 shows the cumulative average abnormal return (CAAR) and their t-statistic for multiple event windows in the pre-crisis period (Jan. 2000 - Aug. 2008). *** p<0.01, ** p<0.05, * p<0.1

5.2.2 Crisis period

Table 5 presents the cumulative average abnormal returns for downgrades published in the crisis period from September 2008 to July 2009. It shows negative CAARs for several event windows for downgrade credit rating announcements, where the CAAR of -2,96% in the event window [-1 to 1] is statistically different from zero at the 1% significance level. The pre- and post-announcement event windows show positive and negative cumulative average abnormal returns, but these values are insignificant. The other extended event windows also do not show any significance and therefore, they are not statistically different from zero. There is no evidence of an effect of upgrade credit rating announcements during this period, as there was only one observation for upgrades in the crisis period. Thus, the testing procedure was not conducted on upgrades as these would have led to poor estimations (Hogg et al., 2015). The results in the crisis period are consistent with the literature of Holthausen & Leftwich (1986), Hand et al. (1992), and Dichev & Piotroski (2001) who found significant negative abnormal returns for downgrades.

Table 5. CAARs and t-statistics for the downgrade credit rating announcements during the crisis period.

Event window CAAR (%) t-statistic

[-10 to 10] -1,20 -0,18

[-5 to 5] -1,87 -0,37

[-1 to 1] -2,96 -2,56***

[-10 to -1] 1,44 0,34

[0 to 10] -2,64 -0,73

(22)

22

5.2.3 Post-crisis period

In table 6 the CAARs of up- and downgrade credit rating announcements are presented for different event windows in the post-crisis period from August 2009 to December 2018. In contrast to some of the previous literature, but consistent with Lobão et al. (2019), downgrades are not associated with significant negative abnormal returns in the post-crisis period. The event window [-1 to 1] and the pre-announcement event window [-10 to -1] have negative CAARs of 0,38% and 0,55%, but both are not significant at any level. The positive cumulative average abnormal returns for event windows [-5 to 5] and [-10 to 10] are marginally significant at the 10% level. The post-announcement event window [0 to 10] with positive CAAR of 2,03% is significant at 5%. This result suggests that investors anticipated the negative outlooks of the banks before the announcement date, but there was an overreaction to the news which resulted in positive abnormal returns in the days after the announcement to correct the overreaction. Compared to the other time periods, the cumulative average abnormal returns for upgrades do show strong significance in several event windows, consistent with previous literature from Elayan et al. (2003), Li et al. (2004), and Amin et al. (2018). In event window [-5 to 5] the CAAR of 1,87% is highly significant at 1%, whereas the CAAR of 2,52% in event window [-10 to 10] has a 5% significance. Also, the post-announcement event window [0 to 10] is significant at 1% for upgrades, suggesting that the information of an upgrade credit rating announcement is slowly incorporated into the bank’s share prices. The results for upgrade credit rating announcements are in line with the research from Joo & Pruitt (2006), claiming that the market is more reactive to credit rating changes in the aftermath of a financial crisis.

Table 6. CAARs and t-statistics for the post-crisis credit rating announcements.

Downgrades Upgrades

[-10 to 10] 1,48 1,54* 2,52 2,35**

[-5 to 5] 1,56 1,64* 1,87 2,80***

[-1 to 1] -0,38 -0,58 0,53 0,89

[-10 to -1] -0,55 -0,62 -0,38 -0,43

[0 to 10] 2,03 2,24** 2,90 2,54***

(23)

23

5.3 Multivariate regression analysis

Table 7 shows the results from the multivariate regression, with the absolute value of the abnormal returns on the announcement date (t = 0) as the dependent variable in the analysis. Multiple variables are statistically significant in the regression. The first variable ‘’Size’’ with a coefficient of -0,69% shows a marginal 10% significance, which supports the theory that small firms should have larger abnormal returns than large firms due to an incomplete market reaction to credit rating announcements (Beard & Sias, 1997; Bernard & Thomas, 1990; Fama, 1998). The coefficient of 2,77% with a 5% significance for dummy variable ‘’Change’’ supports that there is a larger response of the market to downgrade credit rating announcements in the sample than to upgrade credit rating announcements. The coefficient of -4,82% of dummy variable ‘’Grade’’ is significant at 1% and this indicates that the market is more responsive to credit rating announcements which are graded by Standard & Poor’s as speculative grade compared to investment grade credit rating announcements. Bonds with speculative grade credit ratings are more volatile than investment grade bonds, and regulations occasionally force institutional investors to sell their bonds if credit ratings are below the investment grade. Thus, rating changes into speculative grade are expected to result in larger market responses with higher cumulative average abnormal returns around the announcement date. Table 8 gives an example of a credit rating announcement from investment grade to speculative grade in the sample. It shows large negative cumulative abnormal returns in the event windows [-1 to 1], [-5 to 5] and the pre-announcement event window. The CARs in these event windows are more negative than the cumulative average abnormal returns (CAARs) for downgrade credit rating announcements in the post-crisis period. Also, the post-announcement event window in the example has a large positive cumulative abnormal return, which is higher than the cumulative average abnormal return after post-crisis downgrades in this event window. Unfortunately, due to the small sample of rating changes from investment grade to speculative grade, the statistical evidence is inconclusive.

(24)

24 pre- or post-crisis period. Additionally, the variable ‘’Period’’ does not show significance at any level, meaning that there is no higher market response on the announcement date to credit rating announcements published in the post-crisis period from August 2009 to December 2018 compared to credit rating announcements published in the pre-crisis period from January 2000 to August 2008.

Table 7. Multivariate regression analysis on the absolute values of the abnormal returns of all credit rating announcements in the sample.

Variable Coefficients Coefficient t-statistic Significance

Constant 0,2285 (0,0898) 2,54 0,012** Size -0,0069 (0,0039) -1,76 0,081* Book-to-Market 0,0021 (0,0052) 0,40 0,688 Change 0,0277 (0,0110) 2,53 0,013** Crisis -0,0177 (0,0160) -1,11 0,271 Period -0,0125 (0,0125) -1,00 0,322 Grade -0,0482 (0,0165) -2,92 0,004*** Observations 121 Adjusted R-Squared 0,1752 F-Statistic 5,2493

Table 7 shows the coefficients, t-statistics and significance levels of the variables in the multivariate regression analysis. The sample comprises 121 observations for the period 2000 to 2018. Dependent variable: |AR0| (absolute value of the abnormal return on the announcement date). Independent variables: Size, Book-to-Market, Change, Crisis, Period and Grade (Table 2). Standard error in parentheses.

*** p<0.01, ** p<0.05, * p<0.1

Table 8. Example of a downgrade rating change into speculative grade.

Event window CAR (%)

[-10 to 10] 5,80

[-5 to 5] -6,72

[-1 to 1] -3,66

[-10 to -1] -9,80

[0 to 10] 15,60

(25)

25

6. Discussion

In this section a discussion based on the previous mentioned empirical results is presented, where the hypotheses from section three will be analysed through the help of the theoretical framework from section four. This section will also provide a deeper insight on the obtained results.

6.1 Full sample credit rating announcements

The results for downgrade credit rating announcements show statistically significant negative abnormal returns for the event window [-1 to 1] at the 1% level, suggesting that the market adjusts quickly to the credit rating change. This result corresponds with the results of Holthausen & Leftwich (1986), Hand et al. (1992) and Dichev & Piotroski (2001) who all found a significant negative reaction of the market to downgrades. It also supports the research of Choy et al. (2006) who found that the highest significance in the Australian stock market occurred during the three-day event window. No statistical evidence of an impact is found in the event window [-1 to 1] for upgrade credit rating announcements, but the extended event window [-5 to 5] does show a strong 1% significance. The effect after an upgrade is slowly incorporated into the bank’s share prices on the days after the announcement, as the post-announcement event window [0 to 10] is statistically significant at 5%. This result is consistent with previous literature of Elayan et al. (2003), Li et al. (2004) and Amin et al. (2018) who also observed significant abnormal returns in their post-announcement event window. Therefore, the first hypothesis in this study suggesting that credit rating announcements generate significant abnormal returns for banks listed on the S&P500 is not rejected, as upgrades generate significant positive abnormal returns and downgrades generate significant negative abnormal returns. Although the market takes more time to fully impound upgrade credit rating announcements compared to downgrade credit rating announcements, the overall result finds support for the informational value that credit rating announcements offer to the market which cannot be obtained by investors from other sources.

6.2 Pre- and post-financial crisis

(26)

26 cumulative average abnormal returns are shown in the event window [-10 to 10] and the pre-announcement event window after upgrades. Based on these results, there is a higher market reaction to downgrades than to upgrades in the pre-crisis period. However, only a marginal impact on the bank’s stock prices is found for both up- and downgrade credit rating announcements in the three-day event window [-1 to 1], with a 10% significance. All other event windows do not show any significance, which gives little evidence that the cumulative average abnormal returns for up- and downgrades are statistically different from zero in the pre-crisis period, even though the CAARs in several event windows after downgrades have large negative values. The results correspond with the early literature who also did not find statistical significant abnormal returns for up- and downgrade credit rating announcements (Kaplan & Urwitz, 1979; Pinches & Singleton, 1978; Weinstein, 1977).

Meanwhile, downgrade credit rating announcements in the post-crisis period show small negative CAARs in the event window [-1 to 1] and the pre-announcement event window, both insignificant at any level. The most interesting result from the analysis show positive statistical significance for downgrades in the event windows [-10 to 10] and [-5 to 5] at the 10% level. The post-announcement event window [0 to 10] with a CAAR of 2,03% for downgrades is even significant at the 5% level. This result suggest that the market already anticipated on the negative information of the stocks before the announcement date, but investors overreacted to the news resulting in significantly positive abnormal returns in the days after the announcement to correct this overreaction. The results for upgrade credit rating announcements in the post-crisis period show highly significant cumulative average abnormal returns in the event window [-5 to 5] at the 1% significance level, and a 1% significance in the positive post-announcement event window [0 to 10]. This result indicates that there is a positive impact on bank’s share prices after upgrades, even though the information is slowly incorporated into the prices in the days after the announcement. The results for upgrade credit rating announcements in the post-crisis period are consistent with Joo & Pruitt (2006) who claimed that the market is more reactive to credit rating changes in the aftermath of a financial crisis.

(27)

(28)

28

7. Conclusion

This research examines the impact of credit rating announcements on the stock prices of banks listed on the S&P500. Existing literature on this topic presented conflicting findings, as it appears to be difficult to establish a clear impact of credit rating changes on stock returns. Therefore, the informational value of credit rating announcements is a controversial and inconclusive issue. Throughout the sample of 121 credit rating announcements published by Standard & Poor’s for 32 banks over the period 2000 to 2018, this study provides evidence that credit rating announcements generate statistically significant abnormal returns for both up- and downgrade credit rating announcements. Downgrade credit rating announcements result in negative abnormal returns, where upgrade credit rating announcements are associated with positive abnormal returns in the full sample of credit rating announcements. Significant results for downgrades are the highest on the announcement date, whereas for upgrades the significant results are the highest on the days after the announcement, indicating that the effect after an upgrade is slowly incorporated into the bank’s share prices.

This paper also concludes that the response to credit rating announcements did change after the financial crisis in 2008, after dividing the announcements in three different categories (pre-crisis, crisis and post-crisis period). The pre-crisis credit rating announcements only show marginal significance around the announcement date, which is consistent with early literature (Kaplan & Urwitz, 1979; Pinches & Singleton, 1978; Weinstein, 1977). In the crisis period a significant negative reaction is found to downgrade credit rating announcements and this corresponds with previous literature (Dichev & Piotroski, 2001; Hand et al., 1992; Holthausen & Leftwich, 1986). The post-crisis period shows interesting results with positive statistical significance for downgrade credit rating announcements, suggesting that the market already anticipated the negative returns but overreacted to the negative news. Upgrade credit rating announcements in the post-crisis period show positive and highly significant results on the days after the announcement, consistent with Joo & Pruitt (2006) who claimed that the market is more reactive to credit rating changes in the aftermath of a financial crisis.

(29)

(30)

30

References

Amin, A. S., Jain, P., & Malik, M. (2018). Stock market reaction to credit rating changes: new evidence. Asia-Pacific Journal of Accounting and Economics, 25, 1608–1625.

https://doi.org/10.1080/16081625.2018.1481756

Ammer, J., & Clinton, N. (2004). Good News Is No News? The Impact of Credit Rating Changes on the Pricing of Asset-Backed Securities. International Finance Discussion

Paper, 809, 1–28. https://doi.org/10.17016/ifdp.2004.809

Barber, B. M., & Lyon, J. D. (1997). Firm size, book-to-market ratio, and security returns: A Holdout sample of financial firms. Journal of Finance, 52(2), 875–883.

https://doi.org/10.1111/j.1540-6261.1997.tb04826.x

Barron, M. J., Clare, A. D., & Thomas, S. H. (1997). The effect of bond rating changes and new ratings on UK stock returns. Journal of Business Finance and Accounting, 24(3), 497–509. https://doi.org/10.1111/1468-5957.00117

Basel Committee On Banking Supervision. (2011). Basel III: A global regulatory framework for more resilient banks and banking systems - revised version June 2011. Basel

Committee on Banking Supervision Basel, 1-77.

Beard, C. G., & Sias, R. W. (1997). Is There a Neglected-Firm Effect? Financial Analysts

Journal, 53(5), 19–23.

Bernard, V. L., & Thomas, J. K. (1990). Evidence that stock prices do not fully reflect the implications of current earnings for future earnings. Journal of Accounting and

Economics, 13(4), 305–340. https://doi.org/10.1016/0165-4101(90)90008-R

Binder, J. J. (1998). The event study methodology since 1969. Review of Quantitative

Finance and Accounting, 11(2), 111–137. https://doi.org/10.1023/A:1008295500105

Brooks, C. (2014). Introductory Econometrics for Finance. 3rd edition. Cambridge: Cambridge University Press.

Choe, H., Kho, B. C., & Stulz, R. M. (1999). Do foreign investors destabilize stock markets? The Korean experience in 1997. Journal of Financial Economics, 52(2), 227–264. https://doi.org/10.1016/s0304-405x(99)00037-9

(31)

31 stock returns. Accounting and Finance, 46(5), 755–769. https://doi.org/10.1111/j.1467-629x.2006.00192.x

Dichev, I. D., & Piotroski, J. D. (2001). The long-run stock returns following bond ratings changes. Journal of Finance, 56(1), 173–203. https://doi.org/10.1111/0022-1082.00322 Elayan, F. A., Hsu, W. H., & Meyer, T. O. (2003). The informational content of credit rating announcements for share prices in a small market. Journal of Economics and Finance,

27(3), 337–356. https://doi.org/10.1007/BF02761570

Fama, E. F. (1976). Foundations of finance. New York: Basic Books.

Fama, E. F. (1998). Market efficiency, long-term returns, and behavioral finance. Journal of

Financial Economics, 49(3), 283–306. https://doi.org/10.2469/dig.v30.n1.624

Fama, E. F., & French, K. R. (1992). The Cross-Section of Expected Stock Returns. The

Journal of Finance, 47(2), 427–465. https://doi.org/10.2307/2329112

Financial Crisis Inquiry Commission. (2011). The Financial Crisis Inquiry Report.

Goh, J. C., & Ederington, L. H. (1993). Is a Bond Rating Downgrade Bad News, Good News, or No News for Stockholders? The Journal of Finance, 48, 2001–2008.

https://doi.org/10.1111/j.1540-6261.1993.tb05139.x

Gropp, R., & Richards, A. J. (2001). Rating agency actions and the pricing of debt and equity of European Banks: What can we infer about private sector monitoring of bank

soundness? Economic Notes, 30(3), 373–398. https://doi.org/10.1111/1468-0300.00064 Hand, J. R. M., Holthausen, R. W., & Leftwich, R. W. (1992). The Effect of Bond Rating

Agency Announcements on Bond and Stock Prices. The Journal of Finance, 47(2), 733– 752. https://doi.org/10.1111/j.1540-6261.1992.tb04407.x

Holthausen, R. W., & Leftwich, R. W. (1986). The effect of bond rating changes on common stock prices. Journal of Financial Economics, 17, 57–89. https://doi.org/10.1016/0304-405X(86)90006-1

Hogg, R. V., Tanis, E. & Zimmerman, D. (2015). Probability and Statistical Inference. 9th edition. Boston: Pearson.

(32)

32 https://doi.org/10.1016/j.econlet.2005.04.003

Kaplan, R. S., & Urwitz, G. (1979). Statistical Models of Bond Ratings: A Methodological Inquiry. The Journal of Business, 52(2), 231–261. https://doi.org/10.1086/296045 Kliger, D., & Sarig, O. (2000). The information value of bond ratings. Journal of Finance,

55(6), 2879–2902. https://doi.org/10.1111/0022-1082.00311

Kräussl, R. (2001). Sovereign ratings and their impact on recent financial crises. International

Advances in Economic Research, 7(2), 1–25. https://doi.org/10.1007/bf02296020

Li, H., Visaltanachoti, N., & Kesayan, P. (2004). The Effects of Credit Rating

Announcements on Shares in the Swedish Stock Market. International Journal of

Finance, 16(1), 2872–2894.

Lobão, J., Pacheco, L., & Campos, S. (2019). Stock price effects of bank rating

announcements: An application to European Union countries. International Journal of

Finance and Economics, 24(1), 4–19. https://doi.org/10.1002/ijfe.1645

MacKinlay, A. C. (1997). Event Studies in Economics and Finance. Journal of Economic

Literature, 35(1), 13–39.

Matolcsy, Z. P., & Lianto, T. (1995). The incremental information content of bond rating revisions: The Australian evidence. Journal of Banking and Finance, 19(5), 891–902. https://doi.org/10.1016/0378-4266(94)00082-E

Micu, M., Remolona, E. M., & Wooldridge, P. D. (2004). The price impact of rating announcements: evidence from the credit default swap market. BIS Quarterly Review, 55–65.

Micu, M., Remolona, E. M., & Wooldridge, P. D. (2007). The Price Impact of Rating Announcements: Which Announcements Matter? BIS Working Papers, 207, 1–31. https://doi.org/10.2139/ssrn.911598

Moody's (2008, September 10). Moody's places Lehman's A2 rating on review with direction

uncertain. Retrieved from:

https://www.moodys.com/research/Moodys-places-Lehmans-A2-rating-on-review-with-direction-uncertain--PR_162621

(33)

https://doi.org/10.1111/j.1540-33 6261.1978.tb03387.x

Schweitzer, R., Szewczyk, S. H., & Varma, R. (1992). Bond rating agencies and their role in bank market discipline. Journal of Financial Services Research, 6, 249–263.

https://doi.org/10.1007/BF01047083

Schwert, G. W. (2011). Stock Volatility during the Recent Financial Crisis. European

Financial Management, 17(5), 789–805.

https://doi.org/10.1111/j.1468-036X.2011.00620.x

Sorokina, N., Booth, D., & Thornton, J. (2013). Robust Methods in Event Studies: Empirical Evidence and Theoretical Implications. Journal of Data Science, 11, 575–606.

https://doi.org/10.2139/ssrn.1928942

Standard & Poor’s (2019, September 18). S&P Global Ratings Definitions. Retrieved from: https://www.standardandpoors.com/en_US/web/guest/article/-/view/sourceId/50435 Weinstein, M. I. (1977). The effect of a rating change announcement on bond price. Journal

of Financial Economics, 5, 329–350. https://doi.org/10.1016/0304-405X(77)90042-3

(34)

34

Appendix

Appendix I

Table 9. Standard & Poor’s Long-Term Issuer Credit Ratings* (Standard & Poor’s, 2019).

Category Definition

AAA An obligor rated 'AAA' has extremely strong capacity to meet its financial commitments. 'AAA' is the highest issuer credit rating assigned by S&P Global Ratings.

AA An obligor rated 'AA' has very strong capacity to meet its financial commitments. It differs from the highest-rated obligors only to a small degree.

A An obligor rated 'A' has strong capacity to meet its financial commitments but is somewhat more susceptible to the adverse effects of changes in circumstances and economic conditions than obligors in higher-rated categories.

BBB An obligor rated 'BBB' has adequate capacity to meet its financial commitments. However, adverse economic conditions or changing circumstances are more likely to weaken the obligor's capacity to meet its financial commitments.

BB An obligor rated 'BB' is less vulnerable in the near term than other lower-rated obligors. However, it faces major ongoing uncertainties and exposure to adverse business, financial, or economic conditions that could lead to the obligor's inadequate capacity to meet its financial commitments.

B An obligor rated 'B' is more vulnerable than the obligors rated 'BB', but the obligor currently has the capacity to meet its financial commitments. Adverse business, financial, or economic conditions will likely impair the obligor's capacity or willingness to meet its financial commitments.

CCC An obligor rated 'CCC' is currently vulnerable and is dependent upon favourable business, financial, and economic conditions to meet its financial commitments. CC An obligor rated 'CC' is currently highly vulnerable. The 'CC' rating is used when

a default has not yet occurred, but S&P Global Ratings expects default to be a virtual certainty, regardless of the anticipated time to default.

SD and D

An obligor is rated 'SD' (selective default) or 'D' if S&P Global Ratings considers there to be a default on one or more of its financial obligations, whether long- or short-term, including rated and unrated obligations but excluding hybrid instruments classified as regulatory capital or in non-payment according to terms. A 'D' rating is assigned when S&P Global Ratings believes that the default will be a general default and that the obligor will fail to pay all or substantially all its obligations as they come due. An 'SD' rating is assigned when S&P Global Ratings believes that the obligor has selectively defaulted on a specific issue or class of obligations, but it will continue to meet its payment obligations on other issues or classes of obligations in a timely manner. A rating on an obligor is lowered to 'D' or 'SD' if it is conducting a distressed exchange offer.

*Ratings from 'AA' to 'CCC' may be modified by the addition of a plus (+) or minus (-) sign to show relative standing within the rating categories.

(35)

35

Appendix II

Table 10.S&P500 largest outliers.

Date Market Return

14-4-2000 -6,00% 17-9-2001 -5,05% 24-7-2002 5,57% 29-7-2002 5,27% 29-9-2008 -9,20% 30-9-2008 5,28% 7-10-2008 -5,91% 9-10-2008 -7,92% 13-10-2008 10,96% 15-10-2008 -9,47% 22-10-2008 -6,30% 28-10-2008 10,25% 5-11-2008 -5,41% 6-11-2008 -5,16% 12-11-2008 -5,33% 13-11-2008 6,69% 19-11-2008 -6,31% 20-11-2008 -6,95% 21-11-2008 6,13% 24-11-2008 6,27% 1-12-2008 -9,35% 16-12-2008 5,01% 20-1-2009 -5,43% 10-2-2009 -5,04% 10-3-2009 6,17% 23-3-2009 6,84% 8-8-2011 -6,90%

The impact of credit rating announcements on banks: evidence from the United States