Sovereign downgrades and their spillover effects to the Dutch stock market during the recent financial crisis : an event study

(1)

Sovereign Downgrades and their Spillover

Effects to the Dutch Stock Market during

the Recent Financial Crisis: an Event Study

Olaf van der Wardt

5957648

Finance

26/06/2013

Abstract

For my thesis I want to investigate whether spillover effects are seen on the Dutch stock market due to credit rating downgrades of other sovereigns in the Euro zone. This thesis focuses on the period of September 2008 to the end of 2012, covering the recent financial crisis. By conducting an event study, abnormal returns are analysed for several event windows. This thesis concludes that negative abnormal returns exist for the overall Dutch market. For the financial sector of the Netherlands results are even less expected; several (high) positive abnormal are seen. Summarizing, the results give no significant evidence on spillover effects at neither the whole Dutch stock market, nor its financial sector.

(2)

- 2 -

Introduction

Credit rating agencies, like Standard & Poor’s, Moody’s and Fitch, have been subjected to much debate and criticism during the recent global financial crisis. These rating institutions have been criticised for instance for the lack of ability to foresee major flaws in the financial system preceding the recent financial crisis. These major flaws first came to surface at September 2008, with the bankruptcy of Lehmann Brothers, one of the major investment banks in the United States at the time. Just as Lehmann Brothers, several other banks went into bankruptcy, while many others needed governmental assistance. Besides many firms, also governments were having trouble to meet their financial obligations. These economic developments led eventually to a major credit crisis in the Eurozone, catalysed by a sovereign credit rating change of Greece in April 2010. At that time, Greece saw their sovereign rating change into a speculative grade (from BBB- to BB+), causing a snowball effect of sovereign downgrades among other (South-)European countries. During these events, credit rating agencies were blamed for pro cyclical behaviour, and thereby encouraging financial distress in the European zone.

For my thesis I want to explore this latter statement some more. Much literature on this topic focus on bond yields return differences and stock market return changes in countries that experienced sovereign downgrading by credit rating agencies. Less literature is available on the relationship between sovereign rating changes and possible spillovers with respect to bond and stock markets. Therefore, I want to explore spillover effects on the latter market some further. Specifically, I want to focus on spillover effects caused by sovereign credit rating downgrades of Eurozone members to the Dutch stock market. Next to my main research question, I want to examine whether these possible spillover effects are industry dependent. Furthermore I want to investigate whether these effects became more anticipated over time, i.e. less market reaction after a sovereign downgrade at a later stage compared to such rating changes at an earlier stage.

I contribute to existing literature by providing empirical evidence on the possible existence of spillover effects to the Dutch stock market, because there is minimal or even no information available regarding these effects on the stock market in the Netherlands during the recent financial crisis. Next thing that isn’t explored by prior literature yet are the earlier mentioned focus on the financial sector dependence of such possible spillovers. Also for this matters I will focus on the Dutch stock market. Above this all, I think it could be of importance to gain some further knowledge for investors, with respect to possible increasing Dutch stock risk because of sovereign downgrades of other Euro members. It is trivial that risk is an important component of portfolios and therefore inducing portfolio managers, in the case of increasing risk, to alter their portfolios.

Findings gained from the event study used in this thesis, show that negative abnormal returns are seen on the Dutch stock market after sovereign downgrades of other Euro members. However these findings are not significant. For the financial sector these findings are even less significant, if there are spillover effects seen at all.

(4)

- 4 -

In my thesis, I will first examine existing literature about the subject matter I want to cover in order to obtain good understanding on all relevant parts concerning my thesis. In this literature review I will briefly discuss credit rating agencies, European financial integration as an environmental condition for possible spillover effects, which are discussed later on in the section. Using these outcomes of prior literature, I will be able to develop my hypothesises. The next section will contain the research method implemented and information about the data collection to conduct the empirical analysis. In section 4, I will show and analyse the empirical results. Finally, I will draw conclusions from my results and discuss them in the last section.

2. Literature review

2.1 Credit ratings agencies

The three most well-known credit rating agencies (CRAs) are Standard & Poor’s, Moody’s and Fitch. The first two agencies are the most dominant ones. These credit rating agencies are independent, and rate corporations and countries on their ability and credibility to meet their liabilities, i.e. their credit worthiness. The latter implies that CRAs take into account the likelihood of default of the debtor, that is to be rated. When it comes to sovereign debt rating, aspects as politics, external liquidity, economic structure and prospects and fiscal and monetary flexibility are taken into account to conduct an appropriate rating for a particular country.

Table 1 provides the credit rates of both of the dominant credit rating agencies (i.e. Standard & Poor’s and Moody’s) are compared to each other. In the last column we can see whether a rating is seen as an investment grade or as a non-investment grade, also known as a speculative grade. These grades are not absolute, but rather are relative. For corporations, this implies that a AAA-rated firm is less likely to default on its financial obligations than an Aa1-rated firm is. This Aa1-rated company is, on its turn, less likely to default than a BBB-rated company is. This does not exclude the opportunity of default of an AAA-rated company. For instance, in the case of Lehmann Brothers and AIG in September 2008, all three credit rating agencies granted both companies an investment grade of at least an ‘A’ up until their collapse. However, most of the times an AAA-rated company is considered to be safe, i.e. unlikely to default, for at least the near future. For countries, credit ratings are of importance because a better rating induces countries to borrow at a lower rate, compared to countries that are given a lower rating.

(5)

- 5 -

Table 1. Standard & Poor’s and Moody’s rating comparison S&P’s Moody’s

AAA Aaa Investment grades

AA+ Aa1 A Aa2 AA- Aa3 A+ A1 A A2 A- A3 BBB+ Baa1 BBB Baa2 BBB- Baa3

BB+ Ba1 Speculative grades

BB Ba2 BB- Ba3 B+ B1 B B2 B- B3 CCC+ Ca1 CCC Ca2 CCC- Ca3

Next to credit ratings, CRAs provide outlooks and credit watches. An outlook represents the opinion of the CRA about how creditworthiness of a firm will develop over the next two years. An outlook is attached to all ratings and can be trivially negative, stable or positive. On the other hand, credit watches are more short-term focused. Credit watches reflect new firm/country specific developments which could affect its next rating in the near future. These outlooks and credit watches can help investors to forecast eventual rating changes. However, the existence of outlooks and credit watches do not necessarily infer that a rating change needs to be preceded by either one of these forecasts. Simultaneously, credit watches or outlook changes don’t need to be followed by a change in credit rating.

Several recent studies investigate whether financial markets react differently to credit rating changes preceded by a credit watch in comparison to credit rating chances which are not preceded by such a credit watch. In other words, these papers investigate whether credit watches are informative. Boot et al. (2006) argue that credit watches cannot be informative because of the fact that the information used by CRA’s lose their power once it is published. Also Purda (2007) didn’t find evidence of credit watches providing significant new information. She compares rating changes with preceding credit watches with “surprise” rating changes. She finds no difference in stock market returns between the two, at the time of the rating change announcement. These theories contradict statements by S&P that CRAs help to establish more market efficiency by solving (part of) the information asymmetry in the market.

(6)

- 6 -

However, other studies are in line with the view that CRAs contribute to a more efficient market by providing new information to investors. Kaminsky and Schmukler (2002) state that credit watchlists will cause a mostly anticipated rating change, if such a change occurs. In their opinion rating changes will not provide much information content. An older study by Holthausen and Leftwich (1986) supports this argument empirically. However, their obtained results are of limited extent due the use of a small sample size.

Section 2.3 will later show that spillover effects are seen between stock markets caused by sovereign downgrading (e.g. Pukthuanthong-Le, et al (2007) and Ferreira and Gama (2007)). We can extrapolate these findings to come to the conclusion that a significant amount of studies show that credit rating changes do have at least some informational power, which is reflected by abnormal market returns around the event date. Therefore in my thesis, I will only make use of actual sovereign rating downgrades. More specifically, I will only focus on downgrades because of the fact that these are main changes seen during the time frame I will focus on for my research, i.e. the recent debt crisis in Europe.

2.2 Economic integration

Since the introduction of the European Monetary System (EMS), and later the European Monetary Union (EMU), Europe has become more and more economically integrated. This resulted in a harmonized economic policy to pursue economic growth by converging the separate markets of the members. This did lead to better resource allocation, for instance more merger and acquisition activity between countries to improve intra-European efficiency (Weiler and Kocjan, 2005, p.5). Above that, Weiler and Kocjan argue that the introduction of a common currency led to a decreased level of exchange rate volatility in comparison to pre-euro currency volatility risk of the separate currencies. On top of that, also transaction costs are heavily reduced between the seventeen members that have adopted the currency.

Many studies have been published on the matter of economic integration in Europe. Kim, Moshirian and Wu (2005) for instance, find a clear shift in stock market co-movements and deeper stock market linkages since the introduction of the euro. In an analyses of the European stock market integration, Fratzscher (2002) offers empirical evidence that the unification process led to a significant level of integration, especially for countries that chose for adopting the Euro. Next to the reduction of currency risk for these countries, convergence of inflation rates and interest rates catalysed European integration among stock markets. Other recent studies conducted further quantitative evidence for this development of integration. For instance Adjaouté and Danthine (2004) show that the converging equity premiums and the decreasing cost of capital in the Euro zone is indicating strong financial integration. Another event confirming this development are the converging industry earnings yields among the Euro members. Specifically an earnings yield difference

(7)

- 7 -

between 60 and 330 basis points lower between countries is seen if both of them are members of the Eurozone (Bekaert, 2011).

2.3 Sovereign downgrades and spillover effects

In prior literature we found that credit rating downgrades lead to market reactions, reflected by abnormal returns at or around the announcement date of such a change in credit rating. Further we have seen that Europe has become a well-integrated economic market where co-movements in stock markets and other economic convergences are seen, the last decades. In this section, I want to explore the literature available on the relationship between the market reactions to credit downgrades and a well-integrated economic landscape, in the form of spillover effects. Despite of the numerous advantages of this monetary union, such as the elimination of exchange rate volatility, major downsides came to surface during the recent financial crisis. One of these downsides were the spillover effects between countries regarding sovereign downgrades. Spillover effects refer to the transmission of shocks between markets and countries. We will discuss these effects further in this section.

As mentioned in the introduction, most literature on the relationship between sovereign rating changes and possible spillovers focuses on bond markets. However, for my thesis I am specifically interested in spillovers to other countries’ stock markets caused by sovereign downgrades in a particular country. Several studies show that such effects are seen in different circumstances. For instance Ferreira and Gama (2007) focus in their paper on cross-country and cross-market effects of debt rating. Their major finding is that sovereign debt ratings news spill over to other stock markets. This spillover effect is asymmetric, indicating a negative impact of 51 basis points on average across the non-event countries in case of a sovereign downgrade in the event country, while there is no significant impact on a sovereign upgrade. Kaminsky and Schmukler (2002) agree upon these findings. In their paper on spillover effects in emerging economies they find a significant change in specific bonds and stocks due to a change in credit rating of that particular asset. Furthermore they conclude that rating changes of those assets in one country also contribute to spillover effects on similar assets in other emerging economies. These empirical findings show that contagion among quick developing economies occurs. Another study on the matter of spillover effects in emerging countries is that of Pukthuanthong-Le, et al (2007). Their paper is about the impact of changes in sovereign ratings and outlooks on international bond and stock markets and also the difference of such effects between country characteristics. Next to the several macro-economic components influencing the effect of spillovers, Pukthuanthong-Le, et al (2007) show that geographical distance is inversely related to the spillover impact. A more pronounced downgrade impact in emerging markets is also found just as a negative effect after a downgrade in cross-country industries. These results are gained while controlling for

(8)

- 8 -

country characteristics as the level of economic and financial development and as well for the cultural, regional en institutional environment. On top of that, the empirical results of Pukthuanthong-Le, et al (2007) are in line with the findings of Ferreira and Gama (2007) about the asymmetric effects of spillovers. Their conclusions come from empirical findings that resulted from an event study method and a multivariate regression analysis to control for macroeconomic variables. They found that stock returns tend to be more pronounced during a downgrade in the case of a high inflation rate and low budget deficits plus the fact that these effects are asymmetric, implying that sovereign downgrades have a bigger effect on stock returns than sovereign upgrades do. A couple of studies provide possible explanations for these asymmetric effects found in several papers. Gande and Parsley (2007) suggest that CRAs are cautious when it comes to sovereign downgrading because of the fear for retaliation by governments denying further access to essential information, causing the shock of a downgrade to be greater. Another possible explanation is provided by Holthausen and Leftwich (1986). They think of governments having an incentive to share positive news in economically prosperous times and on the other hand are not eager to share negative news to the public in bad times. Downgrades should therefore have greater information extent than upgrades which is reflected by the difference in market response between the two. Another way to think of these asymmetric effects is that investors can be seen as somewhat risk-averse on the average, and therefore react more severe to credit rating downgrades as they do to upgrades in credit rating by the same amount (Steiner and Heinke, 2001).

It is also relevant for my thesis that there are several studies on market linkages (e.g. contagion and spillovers)in times of crises (e.g. Kaminsky and Reinhart (2000); Dungey and Martin (2007)). They focus on crisis periods in Mexico, South-East Asia and Russia of more than a decade ago. From their empirical results they both find significant evidence for the existence of contagion in stock markets across borders.These findings are in line with those of the earlier described paper of Pukthuanthong-Le, et al (2007). In their paper they examined crises as well. From their empirical results they concluded significant indications of contagion and spillover effects in time of financial disturbances between countries in emerging economic region. Therefore, it is interesting to see whether the same findings are seen for Europe, during the recent crisis.

3. Empirical methodology

3.1 Hypotheses

Having explored the prior literature to achieve understanding of the subject matter to cover, I can now formulate my hypothesises to my research question(s).

My main research question was about whether spillover effects were seen on the Dutch stock market as a consequence of sovereign downgrades of other Euro members

(9)

- 9 -

during the recent crisis. From the literature we have seen that (sovereign) credit downgrades lead to abnormal returns on stock markets. Moreover, literature shows that in well integrated economic markets spillover effects caused by these downgrades are common. Since the Netherlands is part of the EMU and have adopted the Euro, I expect that the answer to my main question is positive:

H1: The Dutch stock market did experience spillovers from other Euro members caused by sovereign downgrades in those countries, during the recent financial crisis.

Because I expect the answer to my research question to be positive, I want to investigate further whether these possible spillovers are seen in the financial sector more severely, i.e. does the Dutch financial sector experienced higher negative abnormal returns than is seen in the overall Dutch index, due to sovereign downgrades of other EU-members. Because there is no prior literature on this particular topic, from my intuition, I expect the answer to this subquestion to be positive. That is, I expect that the financial sector is hurt more severely than the Dutch stock market as a whole did:

H2: These spillover effects are industry dependant.

.

3.2 Method

To investigate whether spillover effects are seen on the Dutch stock market, as a cause of sovereign downgrades of other Euro members during the recent crisis, I will make use of the event study method used by Pukthuanthong-Le, et al (2007). In their paper they focus on emerging economies and test whether market reactions, i.e. abnormal returns, are seen after sovereign downgrades of other countries. Using the same technique I will estimate the Dutch stock market reaction after such sovereign downgrades.

To examine whether such reactions are seen during the recent crisis, we have to look for abnormal returns at, or around the announcement date of the downgrade (event window). This is calculated by making use of the simple market model:

ARt = Rt - αi - βiSTOXX1800Globalt

Where Ri,t is the value weighted rate of return index for all Dutch listed companies at time t.

STOXX1800Globalt represents the value weighted return of the STOXX 1800 global world

market index at time t. Prior to the calculation of the abnormal return, αi and βi are

estimated by performing OLS on a regression of the world market index as the independent variable and the stock i’s return as the dependent variable, over a 120-day (t-210; t-91) period 90 days before the event (t=0). From the abnormal returns the cumulative abnormal returns (CARs) are obtained by adding up the results:

CARi,t = Σ (Ri,t - αi - βiSTOXX1800Globalt)

(10)

- 10 -

CAARi,t = 1/N Σ (Ri,t - αi - βiSTOXX1800Globalt)

These CAARs can be used to investigate whether spillover effects did occur. To show spillover effects, the calculated CAARs have to be significant. To test for this significance, I will make use of the Student t-test, assuming the CAARs to be independently and identically distributed with a mean of zero and variance σ. Then I will be able to perform a t-test whether these CAARs are significantly different from the world index returns. For this t-test I use the standard deviations of the abnormal returns found in the earlier mentioned estimation period. For my research I will make use of four different CAAR-horizons:

- CAAR1; abnormal return at t0

- CAAR3; abnormal returns from t-1 to t1

- CAAR5; abnormal returns from t-6 to t-2

- CAAR11; abnormal returns from t-5 to t5

The first two are used to focus on the abnormal returns at, or just around, the event date. The market reactions for these two time frames show whether the actual implementation of the sovereign downgrade rating changes have an immediate impact. The third time frame is used to control for anticipation effects, i.e. expected or leaked information effecting the market returns before the event. The last time horizon is used to check whether market returns are adjust over a longer time period, in contrast to immediate alteration.

To answer my subqueston I have to divide all Dutch listed companies into sector classifications. After that I follow the same procedure, described above, for the financial sector classification to see whether spillover effects are different between the overall Dutch market and the financial sector of this market, i.e. different abnormal returns (and CAARs) between each other.

3.3 Data

For the estimation of αi and βi,and consequently the calculation of abnormal returns, all

daily stock data returns available on the Dutch market and the global market index are needed. I retrieved that information from DataStream by conducting for both data types a time series list for a time frame of the 1st January 2008 to the 1st of January 2013, covering the interval of the recent crisis. This resulted in a sample of 84 listed companies, each with stock return data of 1306 days. This totals to 109,704 observations. Next to that, I was able to obtain the value weighted global index of MSCI directly from DataStream over the same days. However, no such list exists for all Dutch listed companies. Only the AEX, AMX, EMX indices are value weighted. I had to manually calculate the value weighted returns on a manually composed Dutch index of the three existing indices (therefore including all listed companies). The 15 companies with the highest market capitalization are listed in

(11)

- 11 -

appendix1.

Furthermore, I had to divide all listed companies into different sectors. I used the same sector classifications as DataStream provides; ten different main sectors according the Industry Classification Benchmark (ICB). Appendix 1, depicted in the appendix, provides information on both the classifications and allocation of all Dutch listed companies to these classifications.

The data for credit rating downgrades are obtained from the websites of both major credit rating agencies, i.e. S&P’s and Moody’s, for the same period as I used for obtaining daily stock returns. Appendix 2, included in the appendix, lists all downgrades of countries which adopted the euro currency during the recent financial crisis. All downgrades totalled to the amount of 61, 37 by S&P’s and 24 by Moody’s. Table 2 provides an overview of the raw returns of the different industries at the announcement date of fifteen selected downgrade events

Table 2. Raw returns (in %)

Sovereign downgrade Dutch market index Financials index

Belgium 1.54 2.31 France -0.60 0.41 Greece1 Greece2 -0.16 -0.76 -3.80 -1.17 Greece3 -0.38 -0.77 Greece4 -0.02 1.90 Italy1 -1.28 1.52 Italy2 -2.04 -1.53 Ireland1 -0.36 -2.27 Ireland2 -0.97 0.44 Portugal1 -2.72 -1.44 Portugal2 -0.38 -0.77 Spain1 -0.53 -1.39 Spain2 -1.81 -2.69 Spain3 1.77 2.17

(12)

- 12 -

5. Results

We will first look at the results of the abnormal returns for the entire value weighted Dutch stock market return, listed In table 5. In this table, the empirical results of the different CAARs are presented for different sovereign downgrade events. We can clearly see that for the announcement day (CAAR1) most abnormal returns are, as expected, negative. We note that six of them are statistically significant. We see that the abnormal returns of the entire Dutch stock market after the downgrades of Greece 1, Italy 1 & 2, and Portugal 1 are significant at a 1% level. Next to this observation, we see that the sovereign downgrades of Belgium and Spain2 are followed by unexpectedly small positive returns. Maybe other information at that day (or event window) did influence these outcomes.

In the column of CAAR3, we see that negative abnormal returns are still most common, i.e. only three positive returns. Furthermore, in this column we encounter two significant abnormal returns. The abnormal returns after the Greece3 downgrade is significant at 5% and for Ireland2 the negative abnormal returns are significant at 10%. Interesting to see is that most abnormal returns in the first two columns are negative, as was expected by prior literature.

Now, it is important to see what results the bigger event windows will show. In the case of event window t-6 – t-2, we see that only four negative abnormal returns appear. This

implies that at this event window eleven sovereign downgrades are followed by positive returns on the Dutch stock market. Important to note is that none of these results are significant. Also, for the last column positive abnormal returns are most common. In this last column two significant results did occur. One significant positive result after the Italy2 downgrade and one significant negative result after the Spain1 downgrade, both at a 5% significance level.

These results are in line with the literature that states that spillover effect do occur (e.g. Pukthuanthong-Le, et al (2007)). We clearly see negative abnormal returns for the first two columns, which indicates that the Dutch market did respond to sovereign downgrades of other Euro members. But we have to note that not all results in these columns are significant. Therefore, it does not provide clear evidence of Dutch market response after other Euro countries’ sovereign downgrades. The intuition of the hypothesis can, to some extent, be confirmed by the evidence from these empirical results. That is, the expectation of negative abnormal returns are seen in table 3 at t0 and t-1 – t1.

(13)

- 13 -

Table 3. CAARs per sovereign downgrade

Downgrade CAAR1 CAAR3 CAAR5 CAAR11

Belgium 0.11 0.46 -0.35 0.15 (0.18) (-1.24) (-0.79) (0.76) France -0.58 0.00 -0.19 -0.06 (-0.92) (0.01) (-0.67) (-0.30) Greece1 Greece2 Greece3 Greece4 Italy1 Italy2 Ireland1 Ireland2 Portugal1 Portugal2 Spain1 Spain2 Spain3 -2.47*** (-2.76) -0.86 (-1.04) -0.17 (-0.21) -1.97** (-2.52) -2.10*** (-3.30) -1.98*** (-3.09) -0.39 (-0.58) -1.47* (-1.88) -2.84*** (-3.16) -0.17 (-0.20) -1.29 (-1.59) 0.12 (0.13) -0.92 (-1.44) -1.14** (-2.20) -0.40 (-0.83) -0.15 (-0.31) -0.17 (-0.38) 0.12 (0.34) -0.22 (-0.60) -0.14 (-0.34) -0.88* (-1.85) -0.62 (-1.19) -0.15 (-0.30) -0.47 (-1.00) -0.13 (-0.25) 0.04 (0.10) -0.02 (-0.04) 0.00 (0.01) 0.11 (0.29) -0.20 (-0.56) 0.26 (0.90) 0.04 (0.14) 0.21 (0.69) 0.09 (0.25) 0.04 (0.10) 0.11 (0.27) 0.03 (0.07) 0.02 (0.05) 0.40 (1.40) -0.37 (-1.39) -0.28 (-1.10) 0.08 (0.33) -0.22 (-0.95) -0.13 (-0.66) 0.45** (2.34) 0.18 (0.87) 0.41 (1.62) 0.08 (0.31) 0.08 (0.31) -0.53** (-2.16) 0.17 (0.62) 0.28 (1.46)

Test-statistics between parentheses. *, **, and *** indicate significance levels at 10%, 5% and 1% respectively.

Next to these findings, I examined what results are seen in the financial sector. As stated in the literature review, the Euro zone is a heavily integrated financial region. Therefore, we will now look at abnormal returns seen in the financial sector at the different event windows. From table 6, we see that the results are a little different from the ones we have seen at the total Dutch stock market. In the financial sector we see somewhat more randomized results. This observation is also reflected by insignificant betas and alphas seen in the OLS to value the estimators in favor of being able to forecast the normal returns in the sector.

(14)

- 14 -

Table 4. CAARs from the financial sector per sovereign downgrade

Downgrade CAAR1 CAAR3 CAAR5 CAAR11

Belgium 0.44 0.28 0.48 -0.22 (0.22) (0.25) (0.66) (-0.36) France -0.77 -1.31 -0.18 0.11 (-0.39) (-1.16) (-0.24) (0.17) Greece1 Greece2 Greece3 Greece4 Italy1 Italy2 Ireland1 Ireland2 Portugal1 Portugal2 Spain1 Spain2 Spain3 -3.80 (-1.00) -1.38 (-0.50) -0.83 (-0.48) 1.84 (1.12) -1.68 (-0.90) 1.32 (0.71) -2.38 (-0.53) 0.27 (0.13) -1.44 (-0.64) -0.65 (-0.79) -1.53 (-0.52) -2.84 (-1.66) 3.41* (1.86) -1.66 (-0.76) 0.06 (0.03) -2.13 (-2.11)** 0.58 (0.62) -1.96 -(1.82)* 1.99 (1.83) -1.11 (-0.43) -0.82 (-0.69) 2.57 (1.95)* -2.37 (-1.97)** -0.47 (-0.28) 0.69 (0.69) 0.61 (0.57) -0.96 (-0.66) 0.74 (0.71) 0.20 (0.29) -0.84 (-1.35) -1.39 (-1.98)** -0.47 (-0.69) -0.07 (-0.05) -0.06 (-0.08) 2.00 (2.33)** -0.04 (-0.05) 0.10 (0.11) 0.33 (0.50) 0.68 (0.97) -0.53 (-0.46) 0.35 (0.40) 0.00 (0.00) -0.05 (-0.10) -1.02 (-1.82)* -0.15 (-0.27) 0.27 (0.20) -0.07 (-0.10) 0.44 (0.63) -0.03 (-0.07) -0.23 (-0.27) 0.32 (0.63) 0.03 (0.05)

Test-statistics between parentheses. *, **, and *** indicate significance levels at 10%, 5% and 1% respectively.

When we look somewhat closer to the table, we can identify mostly negative abnormal returns for the CAAR1, i.e. the date of the event. However, we have to note that in the same column relatively high positive returns are seen after the downgrades of Belgium, Greece4 and Spain3. In the second column, that of CAAR3, the same results are shown as in the first column. Most abnormal returns seen for the period of t-1 to t1 are negative ones. However,

also for this timeframe, (large) positive returns are seen. In contrast however to the first column, there are four significant abnormal returns. For Portugal1 this abnormal returns is unexpectedly positive at a 10% significance level. For both Greece3 and Portugal2, the abnormal returns with the most significance are seen. Both of them are negative, indicating a spillover effect of at least some extent to the financial sector. The next two columns are somewhat similar. Also, in these two columns most abnormal returns are still negative. In

(15)

- 15 -

both of the columns the results for Italy1 are significant at 5% and 10% for the third and fourth column, respectively. However, we also have to account for the number of positive abnormal returns. We even see a significant positive abnormal return for Portugal1, as was also the case for column two. For these positive results in the table for the financials it is possible that evident information, other than the particular downgrades, causes some of the abnormal return to be surprisingly positive.

As was the case for the results of the entire Dutch stock market, also for the financial sector we didn’t find clear evidence of spillover effects to the Dutch stock market as a result of sovereign downgrading in other countries. However, also in table 4 we find that most abnormal returns are negative ones, despite some (high) positive abnormal returns we stated earlier. These negative returns are in line with the preceding literature, despite the (mostly) insignificant values of these returns for the obtained results.

Next to these findings, we also have to compare the different CAARs for both tables. We can identify declining abnormal returns, but because in these abnormal returns are averaged over the period we have to look for the significance of the abnormal returns. We cannot find clear results with respect to existing differences between the CAARs. In some cases, significant abnormal returns are seen at the event date only, while other cases show significant results for other event windows. From these results, we can infer that market reactions do exist after sovereign downgrading in other countries. But we cannot identify exactly in which time interval these alterations will occur.

5. Conclusion

The role of credit rating agencies has been an on-going topic for a long while, especially during the last financial crisis. The Netherlands have not been downgraded during the recent crisis. For me, it was however interesting to see whether the Dutch market did still get hurt through the dynamics of potential spillovers of other Euro countries due to sovereign downgrading by these agencies. Several of those countries had been downgraded at least once during the recent crisis.

From the literature section I found that spillover effects and contagion effects do occur in some circumstances, including recent crises of the last decade. To answer my research question I examined the returns seen on the Dutch stock market for several sovereign downgrades in the Eurozone. I calculated the CAARs over different windows to control for anticipation effects before the event and ambivalently for possible late market reactions to the sovereign downgrade. I didn’t find significant evidence to claim that spillover effects are indeed seen on the Dutch stock market during the recent crisis. Although I did find results showing that for the event windows at, or around the event the abnormal returns are mostly negative. This implies that maybe little spillover effect is seen on the Dutch stock market.

(16)

- 16 -

did show different abnormal returns than those seen on the overall market, taking into account from the preceding literature, that the Eurozone is a well-integrated financial system. Therefore, I did expect the abnormal returns to be more negative than the ones seen on the overall market. However this was not the case. Surprisingly, more positive abnormal returns are seen in the financial sector compared the earlier investigated overall market.

To test my hypothesises, I used the event study method described earlier, the same that is commonly used in the literature. However we have to take into account that the estimation periods are in an economic environment that is distressed, inducing that the benchmark can be somewhat biased. On the other hand taking a benchmark of before the crisis is also biased to some extent, because such a benchmark is taken from an economical state that is different than that of the event. Therefore it is difficult to conduct totally unbiased estimators. The same thing is true for the estimation of the standard deviation of the abnormal returns. That is, we cannot assume the standard deviation to be constant over time, especially when one extrapolates standard deviations from an economically stable period to a time of crisis, in the case of the usage of a benchmark before the crisis.

Above that, the sample for both the overall equity market and the financial market are dominated by a small number of companies. This induces that large deviations in the returns of these companies have a large impact on the outcomes, because I make use of value weighted indices. However, to get rid of value weighted indices is not an option. Otherwise we would get a clearly biased outcome for the index, i.e. treating elephants and mice the same. Next to these discussions, we have to mention that the estimator of beta for the overall is significantly different from zero for all estimation periods of each downgrade. On the contrary, this was not the case for the estimation of beta for the financial sector benchmark. This induces a problem interpreting the results found for the analysis of the subquestion. This may explain the somewhat surprising results seen in the analysis. Furthermore, the benchmark included also European equities which can explain why the results (abnormal returns) for either the overall Dutch stock market and the financial sector of the Netherlands are mostly not significant. Even more, when we take into account that not only Europe was struggling for the time period of the analysis, also the US were having financial distress in the same interval.

It could also be the case that other evident information surfaced at the time of a particular event (i.e. downgrade), causing the abnormal return to be different than expected. Besides the proposed explanations, we also can think of a protecting Dutch government (maybe more than other governments). During the recent crisis we have seen that the Dutch government was not afraid of supporting several financial institutions (e.g. ING) and even nationalized the ABN AMRO Bank. Default of larger institutions in the Netherlands are therefore to be considered unlikely. Maybe the Dutch government did support its financial sector better than most other governments, all over the world, did. This could be the case because the Netherlands were never downgraded and therefore held their triple A status. All of this could induce positive abnormal returns for the financial

(17)

- 17 -

sector, as is seen in a number of results. Lastly, I have to bring up the possibility that downgrades are anticipated way before the downgrades are actually implemented by the credit rating agencies , as is stated in several papers we have seen in the literature review. The tools that catalyse such possible dynamics are the mentioned credit watchlists and outlooks by these papers. This possibility that the Efficient Market Hypothesis applies to my research area for at least a substantial extend is supported by the moderate negative abnormal returns seen in the overall Dutch stock market and the even more surprising results for the financial sector.

To examine spillover effects further, one could think of taking into account economic country characteristics as is done by some other research in the past. The recent crisis is very interesting for conducting such research because of the amount of downgrades that have taken place over the period. In more stable economic environments these downgrades do not occur that frequently. Above that, future research could think of better benchmarks than the ones that are now commonly used. Another interesting research can be of the different effects of sovereign downgrades for countries that are in the same close region as the researched country, compared to that of spillover effects from countries outside this close region.

(18)

- 18 -

6. Reference list

Adjaouté, K. and J. Danthine (2004). Equity returns and integration: Is Europe changing?

Oxford Review of Economic Policy, 20, (4).

Bekaert, G. et al. (2010). The European Union, the Euro, and equity market integration.

NBER Working paper series, 16583, (December)

Brooks, R. et al. (2004). The National Market Impact of Sovereign Rating Changes. Journal

of Banking and Finance, 28, (1), 233-250

Dungey, M. and Martin, V.L. (2007). Unraveling Financial Market Linkages During Crises.

Journal of Applied Econometrics, 22, 89-119.

Fratzscher, M. (2002). Financial market integration in Europe: On the effects of the EMU on Stock markets. International Journal of Finance and Economics, 7, 165-193.

Ferreira, M. & P. Gama (2007). Does sovereign debt ratings news spill over to international stock markets? Journal of Banking and Finance, 31, (10), 3162-3182

Gande, A. and D. Parsley (2005). News spillovers in the sovereign debt market. Journal of

Financial Economics, 75, (3), 691-734

Holthausen, R. and R. Leftwich (1986). The Effect of Bond Rating Changes on Common Stock Prices. Journal of Financial Economics, 17, (1), 57-89

Kaminsky G. and C. Reinhart (2000). On crises, contagion and confusion. Journal of

International Economics, 51, (1), 145-168

Kaminsky, G., and S.L. Schmukler. 2002. Emerging market instability: Do sovereign ratings affect country risk and stock returns? World Bank Economic Review, 16, 171−195 Kim, S., Moshirian, F. and E. Wu (2005). Dynamic stock market integration driven by the

European Monetary Union: An empirical analysis. Journal of Banking and Finance, 29, (10), 2475-2502

Moody’s’ website: moodys.com

Pukthuanthong-Le, K., F.A. Elayan, and L.C. Rose. (2007). Equity and debt market

responses to sovereign credit ratings announcement. Global Finance Journal, 18, 47-83.

(19)

- 19 -

Purda, D. (2007). Stock market reaction to anticipated versus surprised rating changes.

Journal of Financial Research, 30, (10), 301-320

Steiner, M., and V.G. Heinke. 2001. Event study concerning international bond price effects of credit rating actions. International Journal of Finance and Economics, 6,

139-157

Standard and Poor’s’ website: standardandpoors.com

Weiler, J.H.H. and M. Kocjan. 2005. The Law of the European Union. NYU School of Law, unit 1-3, 5-6.

(20)

- 20 -

Appendix 1. 15 companies with the highest market capitalization (01/01/2008 – 01/01/2013).

__________________________________________________________________________________ # Corporation Market Capitalization

(%)

Sector

1 Royal Dutch Shell 2 Unilever

23.6 9.7

Oil & Gas

Consumer Goods

3 ING 7.5 Financials

4 Philips 5.4 Industrials

5 Heineken 5.3 Consumer Goods

6 KPN 4.7 Telecommunication

7 AHOLD 3.1 Consumer Services

8 ASM Holding 3.0 Technology

9 Akzo Nobel 2.8 Basic Materials

10 AEGON 2.7 Financials

(21)

- 21 -

Appendix 2 . All Dutch stock allocated to the Industries Classifications Benchmark (ICB)

____________________________________________________________________________________________________ OIL & GAS

ROYAL DUTCH SHELL A VOPAK FUGRO SBM OFFSHORE BASIC MATERIALS AKZO NOBEL DSM KONINKLIJKE INDUSTRIALS PHILIPS ELTN.KONINKLIJKE RANDSTAD HOLDING BOSKALIS WESTMINSTER AALBERTS INDUSTRIES ARCADIS POSTNL ROYAL IMTEC USG PEOPLE BAM GROEP KON. BRUNEL INTL. HUNTER DOUGLAS TEN CATE TKH GROUP BALLAST NEDAM BATENBURG TECHNIEK CROWN VAN GELDER DOCDATA DPA GROUP GRONTMIJ HEIJMANS HES - BEHEER HOLLAND COLOURS HYDRATEC INDUSTRIES KENDRION MACINTOSH RETAIL NEDAP NEWAYS ELEC.INTL. ORANJEWOUD 'A' ROTO SMEETS STERN GROEP HEALTH CARE FORNIX BIOSCIENCES PHARMING GROUP TELECOMMUNICATIONS KPN KON UTILITY <NONE> CONSUMER GOODS UNILEVER CERTS. HEINEKEN HEINEKEN HLDG. NUTRECO CSM CERTS. UNILEVER ACCELL GROUP BETER BED HOLDING WESSANEN KON.CERTS. CONSUMER SERVICES AHOLD KON.

REED ELSEVIER (AMS) WOLTERS KLUWER SLIGRO FOOD GROUP TELEGRAAF MEDIA GROEP AFC AJAX BRILL (KON.) FINANCIAL ING GROEP AEGON HAL TRUST CORIO ASM INTERNATIONAL WERELDHAVE BINCKBANK VAN LANSCHOT VASTNED RETAIL BEVER HOLDING KARDAN N V KAS BANK

NEW SOURCES ENERGY NIEUWE STEEN INV. VALUE8

WITTE MOLEN TECHNOLOGY

(22)

22

ASML HOLDING TOM TOM

AND INTL.PUBLISHERS BE SEMICONDUCTOR CTAC NM - MARKET VALUE EXACT HOLDING ICT AUTOMATISERING NEDSENSE ENTERPRISES ORDINA QURIUS ROODMICROTEC SIMAC TECHNIEK TIE KINETIX

(23)

23

Appendix 3. All sovereign downgrades in the Euro zone during financial crisis (2008-2012).

__________________________________________________________________________________

Country S&P’s Moody’s Total

Austria Belgium Cyprus Estonia France Greece Ireland Italy Malta Portugal Slovakia Slovenia Spain Total 1 1 5 1 1 8 6 2 1 5 1 1 4 37 1 2 6 5 1 1 4 1 3 24 1 2 7 1 1 14 11 3 2 9 1 2 7 61

Appendix 4. 15 Selected downgrades

___________________________________________________________________________

Downgrade Rating agency Date (dd/mm/yyyy)

Belgium Moody’s 16/12/2011 France S&P’s 13/01/2012 Greece 1 S&P’s 14/01/2009 Greece 2 Moody’s 22/04/2010 Greece 3 S&P’s 29/03/2011 Greece 4 S&P’s 09/05/2011 Ireland 1 Moody’s 19/07/2010 Ireland 2 S&P’s 23/11/2012 Italy 1 S&P’s 19/09/2011 Italy 2 Moody’s 04/10/2011 Portugal 1 S&P’s 21/01/2009 Portugal 2 S&P’s 29/03/2011 Spain 1 S&P’s 28/04/2010 Spain 2 Moody’s 10/03/2011 Spain 3 Moody’s 18/10/2011

Sovereign downgrades and their spillover effects to the Dutch stock market during the recent financial crisis : an event study