The effects of credit rating changes on US stock prices

(1)

Luk Jongsma

10772871

The effects of credit rating changes on US stock prices

Programme:

Economie en Bedrijfskunde: Financiering en Organisatie

Supervisor:

Jeroen Ligterink

Date of Submission:

15-02-2017

This thesis investigates the relationship between a credit rating change and a change in stock prices. The focus is on US companies that have experienced a credit rating change, performed by Moody’s, between 2014 and 2015. An event study methodology is used to discover potential abnormal returns in the short-run which are related to the credit rating change. The results of this study display a significant negative response to credit rating downgrades, but no significant response to credit rating upgrades. Because there is a response to rating downgrades, it can be concluded that credit rating agencies possess private company information and that rating changes convey new information to the market.

(2)

1

Statement of Originality

This document is written by Student Luk Jongsma who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

(3)

2

In the present-day economic environment, there’s a lot of company specific information available to the public. This is useful to investors, as they need the information to assess the profitability and the creditworthiness of an investment. One important measure for investors could be the credit rating of a company. Credit ratings are performed by credit rating agencies and are formed to give an independent analysis of a company’s creditworthiness (Becker & Milbourn, 2011). This way companies can be ranked against each other by their credit ratings. These ratings are of great importance, as an inaccurate assessment could lead to significant financial losses. Inaccurate ratings were partly the cause of the financial crisis of 2008, as too many good ratings were given to unsafe asset backed securities (Crawford & Wolfson, 2010). Because of its importance in the economy, the credit ratings are included in this analyses. In this paper, the effect of credit rating changes on stock prices is examined.

Credit rating agencies(CRAs) are processors of information to the public. Considering the financial crisis of 2008 and the risk of a potential new crisis, an accurate risk assessment is an essential part of the financial market. Ideally, CRAs should base their ratings on publicly available information. In this paper, it is tested whether CRAs also have access to private information, which they can use in the rating process. If this were to be true, CRAs have more information than the market and information asymmetry exists. Hence, a credit rating change should have no impact on the stock price, unless it conveys private information in which case information asymmetry is present between the CRAs and the market.

Stock prices are used in this research as they display whether private information is being disclosed to the market. Gropp and Richards (2001) argue that rating changes are useful to investors as they disclose and summarize non-public information. By investigating stock prices around the time of a credit rating change, it can be concluded whether CRAs convey private information through rating changes. The main research question is hereby:

Does a change in a US company’s credit rating indicate a change in the company’s stock price within three days?

In this research, data from62US companies is used of which 32 have experienced an upgrade and thirty have experienced a downgrade, all of them between 2014 and 2015. Only credit rating changes performed by Moody’s are included. The effect of the rating change is examined by analyzing the short-term average cumulative abnormal returns(ACAR) using an event study approach. The results of this research imply a significant negative reaction on stock prices after a credit rating downgrade,

(5)

4

but no response regarding a rating upgrade. They also show, to some extent, an additional effect regarding a downgrade for companies with leveraged securities.

This study contributes to the existing literature by analyzing the effect of very recent credit rating changes. Most studies about this topic are performed with data from the twentieth century, for example Hand and Holthausen (1992), Holthausen and Leftwich (1986) and Stickel (1986). These studies could lose their relevance as time passes. In the present-day market, much more company specific information is available to investors, because of the improved technology and the increasing transparency through the internet (Boone & White, 2015). This could have implications for the efficient market and could lead to different effects regarding a credit rating change. That’s why this study contributes to existing literature, as this effect is being analyzed in a modern-day market. In the next chapter, the theory behind the credit rating process and the studies to this topic will be discussed. In chapter 3, the data collection process and the methodology will be explained. This will be followed by a review of the results in chapter 4 and the conclusion in chapter 5. The references and the appendix can be found in chapter 6 and 7, respectively.

(6)

5

2. Literature Review

In this section the previous researches to this topic will be discussed. The common results will become visible and a possible explanation for these results will be given. But first, the definitions and the reasoning behind credit ratings will be explained.

2.1 Theoretical Background

Credit ratings help investors analyze the credit quality and the credit risk of a fixed-income security. This can contribute to the efficiency in the market by providing an independent and credible

assessment of the investments risk (Moody`s Investors Service, 2016). Graham and Harvey (2001) found out that credit ratings are of the second highest priority to CFOs as they determine their capital structure, which implies a great importance of credit ratings in the financial market. White (2010) argues that Moody’s, Standard and Poor’s and Fitch are the dominant entities in the credit rating industry. In this paper, only Moody’s database is used, as the focus must be on the effects of a rating change and not on the differences between CRAs. Bolton, Freixas and Shapiro (2012) argue that there is competition present between the CRAs, which might influence the outcome of the credit rating process. Hence, by solely focusing on Moody’s, these differences between the CRAs are excluded. Moody’s distinguishes between the following nine long-term rankings: Aaa Aa A Baa Ba B Caa Ca C. Hereby, Aaa represents the highest quality ranking with the lowest level of credit risk and C represents the lowest possible ranking with the highest level of risk. Besides the long-term rankings, Moody’s also provides the following scale of short-term rankings: P-1 P-2 P-3 NP. Hereby, P-1 represents a superior ability to repay its short-term debt obligations, P-2 a strong ability and P-3 an acceptable ability. NP, also Not Prime, means that the issuer doesn’t fall within any prime rating category (Moody’s Investor Service, 2016). Both the long- and short-term rating changes are included in the analyses as either an upgrade or a downgrade.

There can also be a distinction made between investment- grade and leveraged securities. Investment-grade securities being of the relatively higher credit quality and leveraged securities being of the relatively lower credit quality. Investment-grade securities can be defined as securities with a rating between Aaa and Baa and leveraged securities between Ba and C (Moody’s Investor Service, 2016). The different implications of a credit rating change on investment-grade securities and leveraged securities will be measured to get a better understanding of the effect of a rating change.

Elayan, Maris and Young (1996) state that there are two views regarding the role of CRAs as

providers of new information. The first view implies that CRAs base their ratings on publicly available information only. This would demonstrate the CRAs role of an information processor, instead of a

(7)

6

provider of new information. This assumption should lead to results reflecting zero abnormal returns, thus no response in the stock price from a credit rating change. Elayan, Maris and Young (1996) also declare a second view, stating that credit rating agencies are a true source of information to

investors. They argue that credit ratings reveal the true credit quality of the company. This implies that new and private information is being disclosed by CRAs in the form of rating companies. Information asymmetry between the CRAs and the market would be reduced by every credit rating change. If this view were to be true, research to this effect should result in abnormal returns.

Hsueh and Liu (1992) argue that there is a difference between companies regarding the transparency of firm-specific information. For some firms, more information is made available to the public, due to the difference in disclosure of financial information by the press and financial analysts. Because of this difference between companies, information provided by CRAs through credit rating changes can’t be treated the same for all companies. The extent to which CRAs provide new information is dependent on the firm-specific information that is already available to the market. Hence, the information content of the rating changes is more significant for companies of which relatively less information is available to the market.

Gropp and Richards (2001) state that the impact of a rating change is dependent on the underlying reason of the rating change. Instead of investigating companies, they investigated European banks between 1989 and 2000. They found significant positive abnormal returns following from rating changes that is caused by an increase in volatility. Rating changes caused by negative changes in earnings outlooks appeared to have a negative impact on abnormal returns. The authors also claim that rating agencies are valuable to stockholders, as they disclose and summarize non-public information. Whether CRAs actually provide the public with company specific information is examined in this paper.

2.2 Literature review

The general findings of previous researches show evidence for negative abnormal returns caused by a downgrade, but no evidence for positive abnormal returns. Altman and Rijken (2005) argue that one reason for these results might be that CRAs are more responsive to negative news, as the consequences of a too high rating are more serious than a too low rating. When a rating appears to be lower than initially defined by a CRA, investors could make significant losses in the event of default and the CRA could be blamed for it. When a rating appears to be higher than initially anticipated, the investor can only gain from this situation, as the chances of default are smaller. Therefore, the responsibility for CRAs is bigger for a downgrade than for an upgrade. This leads to higher responsiveness to negative news, which leads to more accurate timing for a downgrade.

(8)

7

Hence, more accurate timing implies that downgrades contain more new information than upgrades, resulting in stocks being affected only after a downgrade.

An early study from Pinches and Singelton (1978) was performed on monthly data and concentrated on the long-term effect. Although this research is focused on daily data and the short-term effect, it might be useful to discuss the results and implications form these researches. Pinches and Singelton (1978) found no significant stock price reaction from credit rating changes. These results imply that investors have already determined the credit quality of the companies themselves. Other studies performed on monthly data by Griffin and Sanvicente (1982) and Dichev and Piotroski (2001) did find abnormal returns, but only regarding a downgrade. This implies that CRAs possess private

information and information asymmetry is present between the market and the CRAs. In the following paragraphs, studies to the effect of rating changes using daily data will be discussed. Hand an Holthausen (1992) investigate the daily excess bond and stock returns of companies between 1977 and 1982. They found significant excess bond and stock returns resulting from a downgrade and weaker positive returns following from an upgrade. The sample consisted of companies that had experienced a rating change either performed by Moody’s or Standard and Poor’s. Holthausen and Leftwich (1986) have also found empirical evidence on negative abnormal returns surrounding the time of a downgrade. These results are in line with the argument that CRAs provide new information to the market. Though, they found no significant evidence for abnormal returns at the time of a credit rating upgrade. The data consisted of all companies listed on either the New York or American Stock Exchange that had undergone a credit rating change during 1977 till 1982. The same results apply for a study performed by Nayar and Rozeff (1994). They used daily data from 1977 to 1985 to study the effect of commercial paper rating changes on stock prices. They also found significant negative abnormal returns resulting from a downgrade. These results imply that there is information asymmetry and that the CRAs convey new and private information to the market through credit rating changes.

Stickel (1986) studied not only the effect of a credit rating change on common stock prices, but also on preferred stock prices. His sample contained data of companies between 1972 and 1980 listed either on the New York or American Stock Exchange. Stickel found significant abnormal returns regarding preferred stock prices, but no impact on common stock prices. Following this research, CRAs provide the market with new information, but only to some extent, as the common stock prices are not affected.

(9)

8 2.3 Hypotheses

Based on the previous studies and their results, hypotheses for this study can be formed. In this section the hypotheses of this research will be discussed. By forming null hypotheses and by rejecting them or not, information about the effects tested will become clear and conclusions can be drawn. The main hypothesis will be split into six different sub-hypotheses.

Main Hypothesis: credit rating changes have an impact on stock prices.

As the previous researches show, the general results imply a negative response following from a rating downgrade. From this it can be concluded that credit rating changes have an impact on stock prices, but only to some extent. Therefore, the hypothesis is formed such that a reaction following from a rating change is expected. Even though most results display only a significant negative response to a credit rating downgrade, there have been positive effects following a rating upgrade, but only with lower significance. As both results have been present, despite the positive effect being weaker, is it expected that an upgrade has a positive effect and a downgrade has a negative effect on the stock price. Therefore, the hypotheses are as follows:

Sub-Hypothesis 1: a credit rating upgrade has a positive impact on the stock price. Statistical Sub-Hypothesis 1: H0: β1=0 vs H1: β>0

Sub-Hypothesis 2: a credit rating downgrade has a negative impact on the stock price. Statistical Sub-Hypothesis 2: H0: β2=0 vs H1: β2<0

Whether the main hypothesis will be rejected or not is an answer to the research question. If there is a significant effect, independent of the nature of the change, the main hypothesis will be true. Expected here is that the CRAs provide the public with new information through a rating change, which will lead to a reaction from investors in the form of buying or selling stocks. Also expected is that a credit rating upgrade will have a positive impact on the stock price, as the credit quality improves. A credit rating downgrade on the other hand is expected to have a negative impact on the stock price, as the credit quality deteriorates. The CRAs are assumed to provide the public with new information, hence the information asymmetry would decrease if this were to be true.

Besides the formal hypotheses, the statistical hypotheses are also included. The statistical null hypothesis always displays no effect, and the alternative hypothesis is aligned with the formal hypothesis.

Sub-Hypothesis 3: the impact of a credit rating upgrade on the stock price is larger for companies with a high beta.

(10)

9 Statistical sub-hypothesis 3: H0: β3=0 vs H1: β3>0

Sub-Hypothesis 4: the impact of a credit rating downgrade on the stock price is larger for companies with a high beta.

Statistical sub-hypothesis 4: H0: β4=0 vs β4>0

The reason to include these hypotheses is to uncover whether a credit rating change has a larger effect when the risk of the company is high. Beta is the measure of market risk for a company and is used in this paper to differentiate between high and low levels of risk. The beta is included as a dummy variable equaling one if the beta is higher or equal to one and zero if the beta is lower than one. It is expected that companies with a high beta are more sensitive to rating changes, as the risk of these companies is higher, leading to more responsiveness from investors after a rating change. Sub-Hypothesis 5: the impact of a credit rating upgrade on the stock price is larger for companies with leveraged securities.

Statistical sub-hypothesis 5: H0: β5=0 vs H1: β5>0

Sub-Hypothesis 6: the impact of a credit rating downgrade on the stock price is larger for companies with leveraged securities.

By including these hypotheses, it is tested whether the credit quality of the company has an

additional impact on the effect of a credit rating change on stock prices. It is expected that the effect of a rating change is higher for leveraged securities. This is because of the bigger consequences of inaccurate ratings for leveraged securities, as they have a bigger chance of default. As been stated before, bigger consequences coincide with bigger responsibilities, which leads to more accurate timing and more new information. New information results in abnormal returns and the hypothesis being true.

(11)

10

3. Data and Methodology 3.1 Data Description

The data used in this paper consists of US companies that have had a credit rating upgrade or downgrade, performed by Moody’s, between the years 2014 and 2015. Solely credit rating data applied by Moody’s is used. For the data to be included in the model, the stock prices within the selected period of the entity had to be available. Besides including the type of credit rating change and the stock prices, the beta and the credit quality of the stocks was added. After this selection process, the sample data consists of 62 companies of which 32 companies have had an upgrade and 30 companies have experienced a downgrade. The research will focus on the short-term effect of a credit rating change, therefore daily stock prices are used. The reason to focus on the short-term is to research the possible information asymmetry between the market and the CRAs. If rating changes convey private information, this will become visible in the first days after the change, by an

adjustment in the stock price. 3.2 Data Collection

Firstly, data of all the US companies that have had a credit rating change between 2014 and 2015 must be collected. Such a list is available on Moody’s official site. Moody’s provides a historical dataset of all publicly traded companies that have undergone a credit rating change, going back to 1996.

Secondly, daily stock price information of every company selected must be collected. This can be found by using ‘The Center for Research in Security Prices’ database, which is a specification in the ‘Wharton Research Data Services’. Companies of which the stock price information wasn’t available were excluded from the model.

Thirdly, the Beta’s of the companies selected must be included in the research. This can also be found using the CRSP database under the caption ‘Stock/Portfolio Assignments’. Once selected, every company’s Beta can be collected using the option ‘Beta Deciles’.

Lastly, the credit quality of the company’s securities was added, being either investment-grade or leveraged. This information can also be found on Moody’s official site.

3.3 Methodology

In this paper, an Event Study is used as a statistical analysis to the effect of a credit rating change on a company’s stock price. An event study, also known as abnormal performance index test, involves the analyses of security price behavior around the time of an information announcement or an event

(12)

11

(Bowman, 1983). The Event Study is performed following a panel regression, as different dates from various companies are used.

To research potential abnormal returns, the normal returns must be included in the analysis to check whether the returns in the event window differentiate from the normal returns. The normal returns are represented on the timeline by the period between t=-30and t=-10

and between

t=10

and

t=30.

The days before or after the actual event are represented as ‘t’.

For every company, sixty daily

returns are included in the sample of which forty are normal returns. Hence, the estimation window and the post-event window. In this research, the effects of a credit rating change will be estimated in three different periods within the event window. Firstly, the abnormal returns will be measured for the ten days before the rating change, so t=-10 to t=-1. This is done to find out whether investors are already anticipating the rating change before it occurred. If this were to be true, abnormal returns should be present during this period. Secondly, the effects of a rating change within three days after the event will be measured. This is the main purpose of the paper and the result will be an answer to the research question. The abnormal results here will be measured in the period between t=0 and t=3. As t=0 is the actual event day, t=3 means that the returns will be measured for the 3 days after. Lastly, the returns will be examined for the ten days after the event, so between t=0 and t=10. This is investigated as investors might have a delayed response to rating changes, which will become visible by examining the stock prices till ten days after the change.

The effect of a rating up- or downgrade will be assessed using the log of the difference in stock prices. This way the statistical significance of a change in the stock price caused by a rating change can be measured. The stock price in this study is represented by the closing ask price of the company. The potential abnormal returns(AR) can be measured against the benchmark of normal returns. In other words, the changing stock prices during the estimation window and the post-event window will be compared to the changing stock prices during the event window. Hence, the following will be measured: ARᵢ = Rᵢ - NRᵢ.The returns from the estimation window and post-event window are stated as NRᵢ. The returns from the event-window are stated as Rᵢ. The difference between these returns is the abnormal return, indicated by ARᵢ.

Because panel regression is applied in this research, multiple stock prices of various companies must be included. To measure the total abnormal performance of the complete sample, the cumulative

(13)

12

abnormal returns(CAR) must be calculated. This is simply the sum of all abnormal returns. To finally

see what implications a credit rating change has for a single stock price, the average cumulative

abnormal returns(ACAR) must be calculated, which is the CAR divided by the number of abnormal

returns. Basically, the statistical hypothesis can be stated as follows: H0: E(ACAR) = 0. If this null hypothesis is rejected, there is evidence for abnormal returns, hence there is an effect of rating changes on stock prices.

3.4 Econometrical Challenges

The regression in this study is performed using panel data, which is data in which behavior of different entities is observed across time. In this study, the fixed-effects method is used to estimate the parameters in the model. Stock and Watson (2011) explain that when using the fixed-effects method, the entities are expected to each have its own individual characteristics that may or may not affect the dependent variable in a unique way. In this study, every company could have individual characteristics that influence the stock price. Each entity is different and therefore the entity’s error term and constant term should not correlate with another. Using this method, time-invariant

characteristics are removed, so the net effect of the independent variable on the dependent variable can be captured.

In the analysis, heteroskedasticity-robust standard errors are used, meaning that the error term is expected to be heteroskedastic. If the error term is heteroskedastic, the variance of the error term is dependent on the independent variable. If the variance of the error term isn’t dependent of the independent variable, the error term is homoscedastic (Stock and Watson, 2011). In this study, it is expected that the variance of the error term differs among the various companies, so

heteroskedasticity-robust standard errors are used. 3.5 The model

The dependent variable in the model is log of the difference between the stock prices. The difference is equal to the daily return without dividends. By using the log, the percentage change of the stock prices following from a credit rating change can be assessed. Dividend payouts are not included in this model, as they are not relevant in analyzing the correlation between a rating change and a change in stock prices. It would either deteriorate or strengthen the effect being tested. The main explanatory variable is the credit rating change, with a distinction made between an upgrade and a downgrade. Hereby, an upgrade meaning an improvement of the creditworthiness of the company reflected by an improved credit rating and a downgrade meaning the exact opposite. The creditworthiness is a valuation of the likelihood of default on a borrower’s financial obligations

(14)

13

and the expected repayment at the event of default (Becker & Milbourn, 2011). The change of a credit rating is represented in the model by two dummy variables. Dummy variable ‘UP’ reflects a credit rating upgrade and equals one if the company has experienced an upgrade. Dummy variable ‘DOWN’ reflects a credit rating downgrade and equals one if the company has experienced a

downgrade. Both the up- and downgrade will be compared to the situation where there has been no credit rating change. This means that in both situations the dummy variable equals zero if there has neither been an up- or downgrade.

As a control variable, the beta is added to the model. It’s also included as a dummy variable in the form of an interaction term with both the variable ‘UP’ and the variable ‘DOWN’. The dummy variable ‘BETA’ equals one if the beta of the company is higher or equal to one and zero if the beta is lower than one. In the model this holds the following: β3(UPᵢ * BETAᵢ) + β4(DOWNᵢ * BETAᵢ). The effect being examined here is whether a beta of higher than one implies an extra effect on the stock price given an up- or downgrade. Because beta is the measure of market risk and a high beta implies higher overall risk, the stock price is expected to be more sensitive to credit rating changes when the company has a high beta.

Another control variable included is the credit quality of a company’s securities, being either investment-grade or leveraged. Investment-grade securities are characterized by their relatively higher credit quality and leveraged securities by their relatively lower credit quality. CRAs might respond differently to news of either investment-grade or leveraged securities, which influences the accuracy and the timing of the rating. As previously discussed, this could have an impact on the effect of a rating change. The variable will also be included as a dummy variable in the form of an

interaction term with both ‘UP’ and ‘DOWN’. The dummy variable, named ‘LEV’, will equal one if the security is leveraged and zero if the stock is an investment-grade security. This way the potential additional effect of an up-or downgrade caused by a stock being leveraged can be measured. In the model this implies the following: β5(UPᵢ * LEVᵢ) + β6(DOWNᵢ * LEVᵢ).

The complete model will contain the log of the difference in stock prices as the dependent variable and the dummy variable ‘UP’ and the dummy variable ‘DOWN’ as the main explanatory variables. Also included are the control variables ‘BETA’ and ‘LEV’, both in the form of two interaction terms with ‘UP’ and ‘DOWN’. Besides these factors, a constant termand the error term will be included. The model will be stated as the following:

LN(StockPricet+1 / StockPricet

)

= β0 + β1UPᵢ + β2DOWNᵢ + β3(UPᵢ * BETAᵢ) + β4(DOWNᵢ * BETAᵢ) +

(15)

14 4. Results

In this section the results of the analysis will be reviewed. Based on the hypotheses, the statistical significance of every variable in de model will be discussed and inferences can be made.

The statistical significance of every variable included in the research is tested at a 99%, 95% and a 90% confidence level. Besides the variation in confidence levels, inferences will also be made about the average cumulative abnormal returns(ACARs) in different time windows within the event window. As been said in the previous part, the event window is split up in three time windows. The first one tests the abnormal returns in the ten days before the credit rating change (t=-10to t=-1), the second one tests the abnormal returns between the day of the rating change and three days after (t=0to t=3), and the last one tests the abnormal returns between the day of the rating change and ten days after (t=0to t=10).

4.1 The effect of a credit rating change

In this part the results of a credit rating up-or downgrade on the stock price will be analyzed. Based on a t-test and the P-value, the hypotheses can either be rejected or not. Both the test for upgrades and downgrades are one-tailed. For the upgrades, it is tested whether the effect on stock prices will be positive, so larger than zero. For the downgrades, a negative effect on the stock prices is tested, so an effect of smaller than zero. The hypothesis regarding the upgrades is stated as follows: Sub-Hypothesis 1: a credit rating upgrade has a positive impact on the stock price.

Statistical Sub-Hypothesis 1: H0: β1=0 vs H1: β>0

Table 4.1.1

Window Upgrades ACAR Test-Statistic P-Value

t=-10 to t=-1 -.0033 -0.22 0.5875

t= 0 to t=3 -.0057 -0.68 0.7515

t=0 to t=10 -.0069 -0.82 0.7925

It is tested whether the average cumulative abnormal returns (ACARs) are different from zero in the three event windows. *** significant at a 99% confidence level

** significant at a 95% confidence level * significant at a 90% confidence level

From the table above it can be concluded that there is no significant effect on the stock price caused by a credit rating upgrade. The null-hypothesis cannot be rejected at any of the confidence levels. Although the predicted effect was positive following from an upgrade, the effects observed in the

(16)

15

table are negative. As they’re not significant at any confidence level, these values are not relevant in this analysis, but it might be interested to consider these negative effects in other studies.

Sub-Hypothesis 2: a credit rating downgrade has a negative impact on the stock price. Statistical Sub-Hypothesis 2: H0: β2=0 vs H1: β2<0

Table 4.1.2

Window Downgrades ACAR Test-Statistic P-Value

t=-10 to t=-1 .0280 0.96 0.8305

t= 0 to t=3 -.0430 ** -1.93 0.0295

t=0 to t=10 -.0442 ** -1.92 0.0295

As the table shows, there are significant effects observed following from a credit rating downgrade. These effects have been shown before in previous studies. In this study, the effects in the three days after and the ten days after the rating change are significant at a 90% and 95% confidence level. The null-hypothesis can be rejected at those confidence levels. In the first three days after companies have experienced a credit rating downgrade, stock prices are expected to decrease by 4,3%. In the ten days after the rating change, stock prices are expected to decrease even more, as this decrease is 4,42%.

4.2 The effect of a credit rating change with high risk levels

This section of the results will display the significance of the effects of credit rating changes on stock prices for companies of which the beta is high. The tests for these variables are both one-tailed. Different form the previous two sub-hypotheses, both effects are tested against an alternative hypothesis stating a positive effect, as a high beta is expected to enlarge the effect of a credit rating change. Therefore, the ACARs are tested for an effect of greater than zero.

Sub-Hypothesis 3: the impact of a credit rating upgrade on the stock price is larger for companies with a high beta.

(17)

16

Table 4.2.1

Window High Beta ACAR for Upgrades

Test-Statistic P-Value

t=-10 to t=-1 .0089 0.58 0.2830

t= 0 to t=3 -.0047 -0.43 0.6650

t=0 to t=10 .0001 0.01 0.4965

From the results above can be concluded there is no significant additional effect on the stock price following from an upgrade when the company has a high beta. It can be concluded that the level of market risk has no influence on the relationship between the stock price and a credit rating upgrade. The null-hypothesis cannot be rejected at any of the confidence levels.

Sub-Hypothesis 4: the impact of a credit rating downgrade on the stock price is larger for companies with a high beta.

Statistical sub-hypothesis 4: H0: β4=0 vs β4>0

Table 4.2.2

Window High Beta ACAR for Downgrades

Test-Statistic P-Value

t=-10 to t=-1 .0224 0.52 0.3020

t= 0 to t=3 -.0058 -0.18 0.5720

t=0 to t=10 -.0043 -0.13 0.5510

The results in this table also imply that a high level of market risk has no influence on the impact of a credit rating change. As shown in table 4.1.2, negative ACARs are observed following from a

downgrade. It can be concluded that a high beta has no additional impact on these effects. There is no evidence to reject the null-hypothesis. High levels of market risk also have no influence on the effect of downgrades on stock prices.

(18)

17

4.3 The effect of a credit rating change for leveraged securities

This part of the results focusses on the differences of the effects of a credit rating change on stock prices between leveraged stocks and investment-grade stocks. It is expected that leveraged stocks are more sensitive to rating changes than investment-grade stocks. Therefore, it is expected that either for an upgrade or downgrade, the effect on the stock price is bigger when the stock is

leveraged. This implies that the null-hypotheses are testes against one-tailed alternative hypotheses. The alternative hypothesis for both cases states a positive effect resulting from the security being leveraged, so the ACARs are tested for an effect of greater than zero.

Sub-Hypothesis 5: the impact of a credit rating upgrade on the stock price is larger for companies with leveraged securities.

Table 4.3.1

Window Leveraged Stocks ACAR for Upgrades Test-Statistic P-Value t=-10 to t=-1 .0060 0.42 0.3365 t= 0 to t=3 .0095 0.88 0.1915 t=0 to t=10 .0038 0.31 0.3780

The results in the table imply no difference between the effects of an upgrade for leveraged stocks and investment-grade stocks. It is expected that the magnitude of the effect of a rating upgrade increases for leveraged stocks, but as there is no effect observed for rating upgrades, there is also no additional effect for leveraged stocks. No evidence results from the test-statistics to reject the null-hypothesis.

Sub-Hypothesis 6: the impact of a credit rating downgrade on the stock price is larger for companies with leveraged securities.

(19)

18

Table 4.3.2

Window Leveraged Stocks ACAR for Downgrades Test-Statistic P-Value t=-10 to t=-1 -.0336 -0.68 0.7520 t= 0 to t=3 .0499 * 1.37 0.0880 t=0 to t=10 .0497 1.29 0.1000

From the results below can be concluded that a distinction between leveraged and investment-grade stocks is valuable, as the effect of a rating downgrade has a greater impact for leveraged stocks within the first three days (t=0 to t=3). At a 90% confidence level, there is evidence to conclude that the additional effect of a stock being leveraged is 4,99%. Table 4.1.2 showed the ACARs of 4.3% following from a rating downgrade. The results from the table above imply an additional effect of 4,99%, which means that a downgrade for a company with leveraged stocks is expected to experience a decrease of 9.29% in its stock prices within the first three days. At a 90% confidence level, there is evidence to reject the null-hypothesis.

(20)

19

5. Conclusion

In this section the results of this research will be analyzed and possible explanations for the findings will be given. Besides that, the limitations of this study and suggestions for further research will be discussed.

5.1 Review of results

In this study, the short-term effects of a credit rating change on a US company’s stock price were examined. The results revealed negative cumulative abnormal returns regarding a credit rating downgrade, but no response to rating upgrades. This implies that CRAs have access to private company information and that a credit rating announcement conveys new information to the market. Hence, there is information asymmetry between the market and the CRAs. These results are consistent with earlier studies by Griffin and Sanvicente (1982), Holthausen and Leftwich (1986) and Nayar and Rozeff (1994) who also found significant negative returns following from a downgrade, but no significant returns following from an upgrade.

One reason why only a negative response is observed after a rating downgrade and no response is observed after an upgrade might be the CRAs higher responsiveness to negative news, as discussed in the literature review. Because of the higher consequences of a too high rating compared to a too low rating, the CRAs are more responsive to negative news. Higher responsiveness leads to more accurate timing, which leads to more new information revealed by the CRAs. Another reason might be that companies are more willing to disclose positive news to the market, which reduces the information asymmetry. When the CRAs decide on a credit rating upgrade, the information is probably already revealed to the public by the company itself, as it increases the value of the company. Negative news could affect the company in a bad way, so companies are unwilling to communicate negative news to the market. This negative news will eventually be disclosed by the CRAs through a credit rating downgrade. Therefore, CRAs convey new information to the market concerning downgrades and not upgrades.

Besides the sole effect a credit rating downgrade has on stock prices, an additional effect of

downgrades is observed for companies with leverages stocks. This effect is only significant at a 90% confidence level, but might be worth explaining. A possible explanation might be the higher risk when grading a leveraged company. Again, this is consistent with the higher responsiveness of CRAs regarding downgrades. The consequences of a too high rating are higher than too low rating and these consequences worsen for companies with leveraged stocks, as the chances of bankruptcy are higher for these companies. The implications of inaccurate ratings are worse for leveraged

(21)

20

companies than companies with investment-grade stocks, which leads to more accurate timing and more information disclosed by the CRAs.

5.2 Suggestions for further research

In further research to this topic it could be useful to distinguish between different magnitudes of rating changes, as a rating change of ‘Baa’ to ‘B’ might have a bigger influence than a change of ‘Baa’ to ‘Ba’ for example. As this is a bigger change, it is possible that the stock price is more affected. These changes of bigger magnitude do not appear very often, so a bigger sample is needed if this relationship is going to be studied.

It also might be interesting to study the effect a credit rating has on companies in different industries. Some industries could be more sensitive to changes, as they are riskier or more dependent on a good rating.

Another improvement could be to examine the effect on stock prices when companies are placed on the CreditWatch list. When companies are placed on this list, it implies that they are being examined and possibly be given a different credit rating. There could be a reaction from investors, as they anticipate a rating change, which could lead to a change in stock prices.

(22)

21

6. References

Altman, E. L., & Rijken, H. A. (2005). The Impact of the Rating Agencies' Through‐the‐cycle Methodology on Rating Dynamics. Economic Notes, 34, 127–154.

Becker, B., & Milbourn, T. (2011). How did increased competition affect credit ratings? Journal of

Financial Economics, 101, 493-514.

Bolton, P., Freixas, X., & Shapiro, J. (2012). The Credit Ratings Game. The Journal of Finance, 67, 85-111.

Boone, A. L., & White, J. T. (2015). The effect of institutional ownership on firm transparency and information production. Journal of Financial Economics, 117, 508-533.

Bowman, R. (1983). Understanding and conducting event studies. Journal of Business Finance and

Accounting, 10, 561-584.

Crawford, C., & Wolfson, J. (2010). Lessons from the current financial crisis: should credit rating agencies be re-structured? Journal of Business Economics Research, 8, 85-91.

Dichev, I. D., & Piotroski, J. D. (2001). The long‐run stock returns following bond ratings changes. The

Journal of Finance, 56, 173-203.

Elayan, F. A., Maris, B. A., & Young, P. J. (1996). The Effect of Commercial Paper Rating Changes and Credit‐Watch Placement on Common Stock Prices. Financial Review, 31, 149-167.

Fama, E. F. (1970). Efficient capital markets: A review of theory and empirical work. The journal of

Finance, 25, 383-417.

Graham, J. R., & Harvey, C. R. (2001). The theory and practice of corporate finance: Evidence from the field. Journal of Financial Economics, 60, 187-243.

Griffin, P. A., & Sanvicente, A. Z. (1982). Common stock returns and rating changes: a methodological comparison. The Journal of Finance, 37, 103–119.

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, 373–398.

Hand, J. R. M., & Holthausen, R. W. (1992). The effect of bond rating agency announcements on bond and stock prices. The Journal of Finance, 47, 733–752.

Holthausen, R. W., & Leftwich, R. W. (1986). The effect of bond rating changes on common stock prices. Journal of Financial Economics, 17, 57–89.

Hsueh, L. P., & Liu, Y. A. (1992). Market anticipation and the effect of bond rating changes on common stock prices. Journal of Business Research, 24, 225-239.

Moody`s Investor Service. (2002). Retrieved from Rating Policy - Understanding Moody`s Corporate

Bond.

Moody's Investor Service. (2006). Moody's Rating System. Moody's Investor Service. Retrieved from Moody`s Ratings System.

(23)

22

Moody's Investor Service. (2016). Rating Symbols and Definitions. Moody's Investor Service. Nayar, N., & Rozeff, M. S. (1994). Ratings, commercial paper, and equity returns. The Journal of

finance, 49, 1431–1449.

Pinches, G. E., & Singleton, J.C. (1978). The adjustment of stock prices to bond rating changes. The

Journal of Finance, 33, 29–44.

Schwendiman, C. J., & Pinches, G. E. (1975). An analysis of alternative measures of investment risk.

The Journal of Finance, 30, 193–200.

Stickel, S. E. (1986). The effect of preferred stock rating changes on preferred and common stock prices. The Journal of Accounting and Economics, 8, 197.

Stock, J. H., & Watson, M. W. (2012). Introduction to Econometrics (3rd Edition). Pearson Education. White, L. J. (2010). Markets: The credit rating agencies. The Journal of Economic Perspectives, 24,

(24)

23

7. Appendix

Table 7.1. Complete list of companies included

Firm

Date

Action

New

Rating

Characteristic

Advance Auto Parts 05-08-2015 Upgrade Baa2 Investment-grade

Allegheny Technologies 16-12-2015 Downgrade B2 Leveraged

Alon USA

Avis Budget Group BE Aerospace

Becton Dickinson & Co Best Buy Company Bio-Rad Laboratories Boardwalk Pipeline Bon-Ton Stores CVR Refining Campbell Soup Celgene Corp

Clear Channel Outdoor Corning Deluxe Dillard’s eBay Emmis Communication Entravision Communication Express Scripts FMC Foot Locker

Goodyear Tire & Rubber HCA Harley Davidson Hormel Foods JetBlue Airways Kate Spade Kirby

Kratos Defense and Security LMI Lamar Advertising Lattice Semiconductor M.D.C. MSCI Manpower Meritor Milacron Minerals Technologies Mueller Water Products Murphy Oil

NCR NRG Energy Netflix Owens Illinois

Penn National Gaming Pentair Pilgrim’s Pride 18-11-2015 18-07-2014 10-11-2014 17-03-2015 24-08-2015 17-07-2014 05-11-2014 02-12-2015 31-03-2015 06-10-2014 03-08-2015 08-12-2015 28-10-2015 02-10-2014 27-05-2015 20-07-2015 22-04-2015 12-05-2014 16-09-2015 06-10-2014 23-07-2014 06-08-2015 21-09-2015 19-12-2014 27-01-2015 15-07-2015 01-07-2015 22-09-2015 16-11-2015 29-06-2015 14-09-2015 29-09-2015 05-11-2015 03-08-2015 23-07-2015 09-03-2015 16-07-2015 11-06-2015 15-10-2015 30-06-2015 12-11-2015 06-10-2015 22-01-2015 06-08-2015 05-11-2015 09-09-2015 04-03-2015 Downgrade Upgrade Downgrade Downgrade Upgrade Upgrade Downgrade Downgrade Upgrade Downgrade Downgrade Downgrade Downgrade Upgrade Upgrade Downgrade Downgrade Upgrade Upgrade Downgrade Upgrade Upgrade Upgrade Upgrade Upgrade Upgrade Upgrade Upgrade Downgrade Downgrade Upgrade Downgrade Downgrade Downgrade Upgrade Upgrade Upgrade Upgrade Upgrade Upgrade Downgrade Downgrade Downgrade Downgrade Downgrade Downgrade Upgrade B2 Ba3 Ba2 Baa2 Baa1 Baa3 Baa3 Caa1 Ba3 A3 Baa2 B3 Baa1 Ba1 Baa3 Caa1 B3 B1 Baa2 Baa2 Ba1 Ba2 Ba2 A3 A1 Ba3 Ba3 Baa2 Caa2 B3 Ba2 B2 Ba2 Ba2 Baa1 B1 B2 Ba2 Ba3 Ba1 Ba3 Ba2 B1 Ba3 Ba3 Baa3 Ba3 Leveraged Leveraged Leveraged Investment-grade Investment-grade Investment-grade Investment-grade Leveraged Leveraged Investment-grade Investment-grade Leveraged Investment-grade Leveraged Investment-grade Leveraged Leveraged Leveraged Investment-grade Investment-grade Leveraged Leveraged Leveraged Investment-grade Investment-grade Leveraged Leveraged Investment-grade Leveraged Leveraged Leveraged Leveraged Leveraged Leveraged Investment-grade Leveraged Leveraged Leveraged Leveraged Leveraged Leveraged Leveraged Leveraged Leveraged Leveraged Investment-grade Leveraged

(25)

24

Platform Specialty Products Ralph Lauren

Scientific Games

Scripps Networks Interactive Snap-on Starbucks Timken TriMas United Continental Valmont Wabash National Waste Management YRC Worldwide 28-10-2015 26-06-2015 25-11-2014 03-09-2015 17-01-2014 15-09-2015 31-08-2015 03-06-2015 15-06-2015 26-08-2015 12-03-2015 17-06-2014 14-02-2014 Downgrade Upgrade Downgrade Downgrade Upgrade Upgrade Downgrade Downgrade Upgrade Downgrade Upgrade Upgrade Upgrade B2 A2 B2 Baa3 A3 A2 Baa3 Ba3 Ba3 Baa3 Ba3 Baa2 B3 Leveraged Investment-grade Leveraged Investment-grade Investment-grade Investment-grade Investment-grade Leveraged Leveraged Investment-grade Leveraged Investment-grade Leveraged

Table 7.2. Variable description

Variable Description Units

StockPrice The closing ask price of a company’s stock.

Percentage

UP A dummy variable equaling 1 for upgrades and 0 for no change.

Dummy

DOWN A dummy variable equaling 1 for downgrades and 0 for no change.

Dummy

BETA A dummy variable equaling 1 for betas of higher or equal to 1 and equaling 0 for betas of lower than 1.

Dummy

LEV A dummy variable equaling 1 for leveraged stocks and 0 for investment-grade stocks.

(26)

25 Table 7.3. Global Long-Term Rating Scale

Rating Description

Aaa Obligations rated Aaa are judged to be of the highest quality, subject to the lowest level of credit risk.

Aa Obligations rated Aa are judged to be of high quality and are subject to very low credit risk. A Obligations rated A are judged to be upper-medium grade and are subject to low credit risk. Baa Obligations rated Baa are judged to be medium-grade and subject to moderate credit risk and

as such may possess certain speculative characteristics.

Ba Obligations rated Ba are judged to be speculative and are subject to substantial credit risk. B Obligations rated B are considered speculative and are subject to high credit risk.

Caa Obligations rated Caa are judged to be speculative of poor standing and are subject to very high credit risk.

Ca Obligations rated Ca are highly speculative and are likely in, or very near, default, with some prospect of recovery of principal and interest.

C Obligations rated C are the lowest rated and are typically in default, with little prospect for recovery of principal or interest.

Retrieved from Moody’s investor Service (2016).

Table 7.4. Global Short-Term Rating Scale Rating Description

P-1 Issuers (or supporting institutions) rated Prime-1 have a superior ability to repay short-term debt obligations.

P-2 Issuers (or supporting institutions) rated Prime-2 have a strong ability to repay short-term debt obligations.

P-3 Issuers (or supporting institutions) rated Prime-3 have an acceptable ability to repay short-term obligations.

NP Issuers (or supporting institutions) rated Not Prime do not fall within any of the Prime rating categories.

The effects of credit rating changes on US stock prices

Luk Jongsma

10772871