The impact of the Quantitative Easing on the casino industry in the U.S.A.

(1)

The impact of the Quantitative Easing

on the casino industry in the U.S.A.

Abstract

This paper investigates whether the Quantitative Easing improved the share prices of the casino companies in the U.S. in 2008. In order to examine the impact of the QE on the casino industry, the event-study method and the cross-sectional regression were conducted. The result from the T-test indicates that the Quantitative Easing significantly improved the share prices of the casino industry. In addition, the result from the cross-sectional regression implies that the higher leverage ratio and the enterprise value positively affected the returns of the share prices in the casino companies.

Keywords: Quantitative Easing, Event-study, Casino Industry, Recession, Cross-sectional

Regression

JFL Classification: G12, G14, L83

Name: Jaehoon Jung Student Number: 11086408 Supervisor: Dr. Jan Lemmen

Study: Bachelor of Science in Economics and Business Track: Economics and Finance

Number of Credits Thesis: 12 ECTS

(2)

Statement of Originality

This document is written by Jaehoon Jung, who declares to take full responsibility for

the contents of this document. I declare that the text and the work presented in this

document are 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)

List of tables

Title Page

1. Financial Variables ……… 15

2. T-test for Abnormal Return ………... 15

3. T-test for Cumulative Average Abnormal Return ………. 16

4. The detail of 10 announcements of the Quantitative Easing ………. 16

5. Glejser test ………18

6. Regression Analysis for CAAR(0,+1) ………... 19

(5)

5 1 Introduction

The financial turmoil in 2008-2009 had detrimental impacts on the real economy of the United States of America. The crisis led to substantial declines in consumer spending on gambling. Horváth & Paap (2012) argue that when the economy is in slump, consumers anticipate that there will be higher unemployment, lower salaries and thereby reduced incomes. Consumers cut back their spending in response to the downturn. However, there have been researches that show the positive relationship between the economic recession and the revenue on the gambling. High unemployment increases consumers` spare time for leisure activities and thereby expenditures on the gambling increase (Horváth, 2012). Horváth (2012) also reports that during the recession, consumers tend to bet more money on a small chance of winning millions of dollars than during the period of the economic boom. In spite of the positive relationship, the casino sector was the most adversely affected industry during the financial crisis. The share prices of major gaming companies in the U.S. plummeted from 2007 to 2010. For example, the share price of MGM Mirage and Las Vegas Sands fell 98% and 99% respectively (Eadington, 2011). There are several reasons for the significant declines in the share prices of the casino companies in spite of the positive relationship between the recession and the revenue on gambling.

The first reason is that the casino companies had transformed themselves into luxury entertainments over the past two decades. The casino venues have been coupled with high-quality restaurants, bars, luxury hotels and retail shopping malls (Eadington, 2011). Non-gaming revenues continuously increased over the last two decades. Even though the recession leads to higher spending on the gambling activities, people are disinclined to spend money on luxury restaurants, bars, and hotels. The crisis led to significant reductions in the non-gaming revenue such as luxury hotels, bars, and high-class restaurants, and the gaming companies that make most profits from the non-gaming sectors showed significant declines in the revenue and had difficulty with sustaining their business.

The second reason is that the introduction of smoking bans in several states negatively affected the revenues in the casino industry. It is supported by the fact that the neighboring states that had not prohibited smoking in a public place showed significant increase in the earnings whereas the states that introduced the smoking ban experience the dramatic decline in the revenue.

The third reason is that the casino industry is approaching a saturation point. The older casino companies enjoyed the monopoly positions over the last several decades. According to the economic theory, suppliers that are in the monopoly position can affect market prices as they wish and yield abnormally high return. Dadayan (2016) demonstrates that the number of commercial casinos has been on the rise since 2001. As the revenues from the commercial casinos play an essential role in balancing states` budget, the states legalized the casino operations increasingly. In addition, as the

(6)

6 states targeted not only its local gamblers, but also customers from other states, the authorized casino facilities were operated near borders (Dadayan, 2016). It deteriorated the existing commercial casino`s revenues, which had enjoyed the monopoly positions.

The last reason is that many states legalized other types of gambling activities in an attempt to raise revenue. The most prominent examples are lottery and pari-mutuel betting. There has been a shift in demand from the commercial casinos to those forms of new gambling activities.

While the casino industry was struggling with those four challenges, the subprime-mortgage crisis in the summer of 2007 worsened the casino sector. The unprecedented magnitude of the economic slump changed consumer`s spending pattern and led to a substantial reduction in the gambling activities. The sub-prime mortgage crisis in 2008 had the most detrimental impacts on the Federal budget since the Great Depression. Therefore, many states legalized a variety of forms of gambling in order to compensate for the reduced tax revenues. The surge in the casino facilities and lottery that were authorized by the state governments in the wake of the financial crisis made it harder for the existing companies to sustain the business. The substantial loss in the revenue was not only from consumers` cutback in the discretionary spending, but also a displacement of their customers to other new established casinos or newly legalized lotteries. While most sectors struggled with the crisis to sustain their business in 2008, the Federal Reserve undertook the QE, aiming for an increase in the aggregate demand. The idea of the QE is to stimulate the economy by increasing the liquidity with large quantities of asset purchases and inducing the increase in the spending. A plenty of researches have demonstrated that the QE had positive impacts on the equity market and the real economy. The QE succeeded in bouncing the market index back to the level of the previous crisis. However, there have been rare studies concerning the impact of the Quantitative Easing on the sectors that were brutally affected by the financial crisis. Among many industries that were critically damaged by the economic turmoil in 2008, the casino industry will be examined in this paper. As the purpose of the QE was to stimulate the real economy, it would be interesting to discover how broadly the QE affected other business sectors. This paper will research whether or not the unconventional monetary policy helped the casino industry exit from the economic downturn even though they had struggled with other challenges not relevant to the financial crisis such as the smoking ban, the growing competitions from neighboring states.

2 Literature Review

2.1 Quantitative Easing

In general, Central banks can break through the recession by lowering the short-term interest rate without directly purchasing government bonds and corporate bonds or lending to private sector,

(7)

7 achieving the target of the inflation rate (Smaghi, 2009). However, when the unprecedented magnitude of the recession hit the world economy in 2008, the short-term interest rate was at zero bound and the Federal Reserve was facing the challenge with expanding the monetary policy further to break through the crisis. As an alternative, the Federal Reserve initiated the QE, aiming to boost the real economy through three transmission channels. They are confidence channel, asset price channel, and bank lending channel, respectively (Joyce, Tong & Woods, 2011). Under the confidence channel, people believe that the QE would turn the economy around and thereby increase their spending. In the asset price channel, the vast quantities of bond purchase by the Federal Reserve lead to an increase in the bond and share prices. The increase in the bond price leads to lower interest rates and lower borrowing cost for firms and households. It will lead to a rise in the investment and the spending. As for the banking lending channel, large amounts of government bonds, corporate bonds and mortgage-backed securities purchases inject more liquidity in the banking sector. The higher reserves make the banks lend more, which leads to an increase in the consumer spending and the investment.

2.2 Recession Proof

The casino industry was one of the least sensitive sectors to the economic conditions. This is because the casino companies were in the monopoly position in the past. The states had regulated the operations of commercial casinos strictly with the concern about higher rates of bankruptcy, crime, divorces, and addictions to the gambling in the local community (Dadayen, 2016). Even though the economy was hit by the recession, in the absence of competitions, the casino companies were able to steer the price such that they earned higher returns. Not only that, there have been plenty of researches that prove the positive relationship between the economic downturn and the revenue of games. For example, Horváth (2012) argues that the recession positively affects the casino industry. Horváth (2012) states that people more rely on gambling to earn their living during the economic slump than normal times. In addition, the increase in the unemployment which is caused by the recession contributes to the higher revenue. Horváth (2012) reports that the increased spare time that results from the higher unemployment rate bore people and thereby induce them to engage in the gambling activity. For example, during the recession of 2001, in contrast to the decline in the economic activities in other industries, the gaming revenues in the casino industry increased substantially (Legg & Tang, 2011). However, the economic downturn in 2008 proves that the casino industry is not recession proof anymore. The industry was the most detrimentally affected sector during the financial turmoil of 2008. The substantial decline in the revenue is attributed to four factors; The growing competitions, the introduction of the smoking ban, the substitution effect, and the transformation into luxury facilities with hotels, bars, and restaurants.

(8)

8 2.3 Factor 1: The growing competition

The casino companies that were established before 1970 enjoyed the monopoly position in the market. However, many states have legalized casino operations for the following reasons. Firstly, the states attempted to improve budget deficits with the new sources of revenues from the casino operations. While raising the tax rate brings about oppositions from voters, the authorization of casino operation generates new streams of tax revenue without enraging taxpayers. Especially, most states authorized the casino operations near the borders in order to attract customers from adjacent states and it triggered interstate competitions. Some states suffered the tax-revenue losses from the new legalized casinos from other states and thereby legalized the casino operations near the border to compensate for the tax-revenue loss. This interstate competition continuously reduced the revenues of the casino companies (Dadayan, 2016). His study supports the cannibalization by the interstate competition. For instance, There were substantial amounts of declines in the tax revenues collected from the casino sector between 2008 and 2015 (Dadayan, 2016). In his findings, while all the existing casinos experienced dramatic declines in the revenues, the new casino showed the increase in the revenue. This proves that the casino companies cannibalize each other.

2.4 Factor 2: The introduction of the smoking ban

There were 15 states that prohibited people from smoking in a public place by law in 2009 (Garrett, Pakko & Board, 2009). The negative impacts of the smoking ban on the casino industry arise from two factors. Firstly, smoking gamblers choose to gamble in other states that do not ban smoking. Secondly, since smokers have to pause games whenever they smoke, the gaming revenues are adversely affected. In addition, Zhang, Tam, & Zhou (2016) put forward that people might reconsider continuing gambling and end the game during the break time. Besides, smoking gamblers have a tendency of spending more money than non-gamblers. Smoking gamblers and non-smoking ones were researched with regards to how much time and money they spend on gambling activities (Petry & Oncken, 2002). The research shows that gamblers spend more money and time on gambling activities. For these reasons, the smoking ban led to decreases in smoking gamblers` visitations and thereby damaged the casino industry.

2.5 Factor 3: The substitution effect

In this paper, the substitution refers to the shift from the casino activities to other forms of gambling activities. When the recession hits the economy, not only do the states expand the casino operations in an effort to raise revenue, but also they authorize a host of forms of gambling such as lotteries, sports betting, video games, and iGaming (Dadayan, 2016). The lottery negatively influences casino

(9)

9 gambling revenues as it is an economical substitute for the casino gambling (Elliott & Navin, 2002). Elliot (2002) investigated the relationship between the casino and the pari-mutuel betting on lotteries using cross-section data and proved the cannibalization effects between two. These types of gambling activities provide the same service as the casino but lighten the burden of traveling long distance. Some people decide to play other forms of gambling instead of traveling to casinos. It led to reductions in visitors and the revenues of the casino companies.

2.6 Factor 4: Transformation into luxury casinos

Over the past several decades, the casino operators had transformed themselves into mega-casinos coupled with luxury hotels, bars, restaurants and plenty of leisure facilities (Eadington, 2011). The transformation into the mega-casinos led to an increase in visitors and the revenues. The luxury casinos led to more significant increases in the non-gaming revenues than in the gaming revenues (Eadington,2011). The higher ratio of the non-gaming revenues to the total revenues made the industry vulnerable to the recession. The economic slump tends to increase the gaming revenue as more people rely on the gambling as a source of income. The recession discourages people from spending their income on luxury hotels or bars and thereby decrease the non-gaming revenues. In addition, Eadington (2011) reports that the severity of the Great Recession led to poor returns on the invested capital. Most mega casino companies have higher leverage ratio as tremendous amounts of capitals are needed to build luxury hotels and entertainment venues. The reductions in visitations and revenues made the companies harder to redeem their debt and led to lower ROA.

With those four challenges that the casino industry was facing, the share prices in the casino sector plummeted at the greater magnitude than in any other sectors. The QE raised consumers` confidence and boosted the consumer spending. It would have led to an increase in visitors to the casino companies. Also, the asset price channel of the QE lowered the borrowing costs for the households. Consumers became less constrained to the liquidity and increased their discretionary spending, which led to increased revenues in the casino companies. The QE also signaled the market that the economy would recover from the downturn. The positive prospect of the QE was incorporated in the stock market. Whether those positive impacts of the QE outweighed the negative four factors, the growing competition, the substitution effects, the smoking ban, and the transformation into mega-casino, will be investigated.

3 Methodology 3.1 Event-study

(10)

10 The event-study is widely used to measure the impact of a certain event on firms` share prices. This methodology is based on the market efficiency hypothesis. Security prices in the efficient market incorporate all information available and any changes in the security prices reflect new information. The idea of the event-study for the impact of the Quantitative Easing is that the announcements of the QE, which involved the Large-Scale Asset Purchase (LSAPs) of government and agency debt and Mortgage-backed securities (MBS), would have made the market believe that the economy would exit from the recession and this belief would have been reflected on the security prices as soon as the scheme went public. The huge swing in the share prices implies that the event had influences on the share prices. However, since there are plenty of factors that affect the share prices, such as firm-specific risk, exchange rate, inflation and so on, several techniques are needed to isolate the pure effect of the announcements. Thornton (2017) suggests that monthly or weekly data might cause the endogeneity problem. This is because when data is collected on the basis of a month or week, alternations in the share prices can be attributable to not only the announcements of the QE, but also other factors (Thornton, 2017). He argues that the endogeneity problem can be prevented by selecting daily or higher frequencies data. For this reason, daily data during a certain period before and after the announcements over 19 casino companies will be collected. The abnormal return that was caused by the event is estimated as the difference between the stock`s actual return and a benchmark (De Jong, 2007). The benchmark is the average return of the certain period before the event takes place (De Jong, 2007). Therefore, the constant mean model is used in this study. The benchmark return is defined as:

BR

i

=

1

𝑇𝑇

∑

𝑇𝑇𝑖𝑖=1

𝑅𝑅

𝑖𝑖𝑖𝑖

(1)

The letter ‘T’ refers to the time period that is used to calculate the benchmark return. The letter ‘i’’ stands for each of the 19 casino companies that are investigated. In this paper, the estimation period for the benchmark is 30 trading days before the event day. Therefore, T that refers to the time period is 30. In this paper, the negative number stands for the days before the event date.

BR

i

=

₃₀1

∑

−31𝑖𝑖=−2

𝑅𝑅

𝑖𝑖𝑖𝑖

(2)

The abnormal return is defined as:

(11)

11 The idea of the abnormal return is that the return on the event date must be larger than the average return on normal times if the event had substantial influences on the share prices and thereby the abnormal return is larger than zero. According to the market efficiency hypothesis, the announcement effect should be incorporated on the share prices immediately. Nevertheless, in reality, it may take more than one day for the market to adjust to new events. Not only that, there might be possibilities that the scheme of the QE had been leaked before it went public. Considering those two possibilities, the abnormal return for the day before the announcement, 𝐴𝐴𝑅𝑅₋₁, the abnormal return for the day of the announcements, 𝐴𝐴𝑅𝑅₀, the abnormal return for the next day of the announcements, 𝐴𝐴𝑅𝑅₊₁, the cumulative abnormal return from the day of the announcement to the next day, 𝐶𝐶𝐴𝐴𝑅𝑅_(0,+1) and the cumulative abnormal return from the day before the announcement to the next day of the announcement, 𝐶𝐶𝐴𝐴𝑅𝑅_(−1,+1) will be examined. In this study, 0 indicates the event day and the sign “+” tells us how many days passed from the event day. Thus, the cumulative abnormal return is defined as:

𝐶𝐶𝐴𝐴𝑅𝑅

_{𝑖𝑖,(0,+1)}

=

𝐴𝐴𝑅𝑅

_𝑖𝑖,0

+

𝐴𝐴𝑅𝑅

_𝑖𝑖,+1

(4)

𝐶𝐶𝐴𝐴𝑅𝑅

_{𝑖𝑖,(−1,+1)}

=

𝐴𝐴𝑅𝑅

_{𝑖𝑖,−1}

+

𝐴𝐴𝑅𝑅

_𝑖𝑖,0

+ 𝐴𝐴𝑅𝑅

_𝑖𝑖,+1

(5)

This paper will investigate 10 announcements of the QE, which took place from November 2008 to January 2010. This period is referred to the 1st round of the QE. The 2nd and 3rd rounds of the QE began in November 2010 and September 2012 respectively. The 2nd and 3rd rounds of the QE were quite far from the point when the Great Recession hit the U.S. More factors must be taken into account to gauge the effectiveness of the 2nd and 3rd rounds of the QE. Therefore, they are excluded from this paper and passed to next researchers. To measure the effectiveness of the QE, we excluded the announcements that contain other news apart from the QE. When other news apart from the scheme of the QE were included in the announcements, it is hard to say that the abnormal return derived from the effectiveness of the QE (Thornton, 2017). Each of the ten announcements will be denoted with 𝑄𝑄𝑄𝑄_𝑖𝑖. Thus, the corresponding abnormal return and cumulative abnormal return will be defined as 𝑄𝑄𝑄𝑄_𝑖𝑖𝐴𝐴𝑅𝑅_𝑖𝑖 and 𝑄𝑄𝑄𝑄_𝑖𝑖𝐶𝐶𝐴𝐴𝑅𝑅_𝐼𝐼 respectively. The average of 19 companies` abnormal return is defined as:

𝑄𝑄𝑄𝑄

_𝑇𝑇

𝐴𝐴𝐴𝐴𝑅𝑅 =

19

∑

19𝑖𝑖=1

𝐶𝐶𝐴𝐴𝑅𝑅

𝑖𝑖,(−1,+1)

0

:

𝑄𝑄𝑄𝑄

𝑇𝑇

𝐶𝐶𝐴𝐴𝐴𝐴𝑅𝑅

(−1,+1)

=0

(11)

𝐻𝐻

1

:

𝑄𝑄𝑄𝑄

𝑇𝑇

𝐶𝐶𝐴𝐴𝐴𝐴𝑅𝑅

(0,+1)

>0

𝐻𝐻

1

:

𝑄𝑄𝑄𝑄

𝑇𝑇

𝐶𝐶𝐴𝐴𝐴𝐴𝑅𝑅

(−1,+1)

>0

(12)

Due to the small size of the sample, the t-test will be employed to test the null hypothesis. Before we proceed with the t-test, several assumptions need to be made. First of all, the abnormal return of each casino company is identically and independently distributed. The independence of abnormal returns indicates that the abnormal return from one period is not correlated to the one from other periods (De Jong, 2007). That is, Cov(𝐴𝐴𝑅𝑅_{𝐼𝐼,𝑎𝑎} , 𝐴𝐴𝑅𝑅_{𝐼𝐼,𝑏𝑏} ) = 0. The second assumption that must be made is that those abnormal returns are normally distributed with the mean of zero and the variance of 𝜎𝜎2. The third assumption is that the samples are drawn randomly. The non-randomly drawn samples lead to the biasedness of the t-statistic. Under those assumptions met, the t-statistic follows the t-student distribution with N-1 degrees of freedom. As the standard deviations of the abnormal return are not known, it will be estimated in the following manner.

(13)

13

S

t

=

�

1 𝑁𝑁−1

∑ (𝑄𝑄𝑄𝑄

𝑁𝑁𝑖𝑖=1 𝑖𝑖

𝐴𝐴𝑅𝑅

𝑖𝑖

− 𝑄𝑄𝑄𝑄

𝑖𝑖

𝐴𝐴𝐴𝐴𝑅𝑅)

2

(13)

S

t

=

�

1 𝑁𝑁−1

∑ (𝑄𝑄𝑄𝑄

𝑁𝑁𝑖𝑖=1 𝑖𝑖

𝐶𝐶𝐴𝐴𝑅𝑅

𝑖𝑖

− 𝑄𝑄𝑄𝑄

𝑖𝑖

𝐶𝐶𝐴𝐴𝐴𝐴𝑅𝑅)

2

(14)

With the five assumptions above made, the estimated standard deviation, and the average abnormal returns, the t-statistics is defined as:

T =

𝑄𝑄𝑄𝑄

𝑡𝑡

𝐴𝐴𝐴𝐴𝐴𝐴

−1

−0

St

⁄

√𝑁𝑁

~ t

N-1

(15)

T =

𝑄𝑄𝑄𝑄

𝑡𝑡

𝐴𝐴𝐴𝐴𝐴𝐴

0

−0

St

√𝑁𝑁

~ t

N-1

(19)

Since the purpose of this paper is to examine whether the QE had the positive impact on the share prices of the casino companies, the one-sided t-test will be conducted with the significance level of 5%. That is, if the t-statistics is larger than 1.684, there is significant evidence that the null hypothesis can be rejected. The rejected t-value indicates that the QE affected the share prices positively.

3.2 Cross-sectional Regression

1

) + 𝜖𝜖

) + 𝜖𝜖

18

(21)

4 Data

The share prices of 19 casino companies and the financial variables of 18 casino companies in the U.S. were collected from the Wharton Database. Since the financial variables of Melco Resorts & Entertainment were not available, the company is excluded from the cross-sectional regression. Thornton (2017) asserts that we can only conclude that the abnormal returns were caused by the announcements of the QE when the news exclusively contained the QE. The 10 announcements that contain the QE news solely are found in his article and we investigate those 10 announcements. Apart from those ten announcements, other announcements that included not only the QE, but also other economic policies are excluded. This is because the abnormal returns could have been from the other economic policies, not the QE. Further research for those excluded announcements is needed whether the abnormal returns arose from the QE or other economic policies. The following tables are the leverage ratio, the ROA, the enterprise values, and the natural logarithm of the enterprise value of 18 casino companies.

(15)

15

Table 1

Financial Variables

Leverage ROA Enterprise Value Ln(Enterprise Value) Century Casino 31.20% 3.07% $45,275.53 10.62 Pinnacle Entertainment 52.07% 4.89% $490,257.01 13.06

Nevada Gold Casino 31.88% -7.44% $10,499.99 9.22

Canterbury Park 31.88% -7.44% $10,499.99 9.22

Wynn Resort 68.49% 9.77% $5,623,184.29 15.47

Monarch Casino 23.76% 16.64% $137,798.99 11.8

MGM Resort 78.46% 7.03% $3,383,796.92 14.91

Full House 18.45% -0.88% $33,995.48 10.34

Las Vegas Sands 74.31% 5.31% $6,099,763.48 15.34

Churchill Down 13.05% 10.36% $469,390.99 13.05

Empire Resort 299.29% -3.73% $71,559.15 11.0314

Dover Gaming Ent 50.14% 20.90% $461,918.07 11

Penn National 63.71% 11.66% $1,939,252.37 14.46

Boyd Gaming 67.11% 7.85% $589,391.99 13.21

Ameristar 82.51% 13.67% $738,006.17 13.43

SHFL Entertainment 71.33% 13.90% $292,226.74 12.51

Lakes Entertainment 8.47% 1.26% $78,635.31 11.26

Isle of Capri Casino 87.59% 9.61% $204,995.16 12.08

Table 1 provides the leverage ratio, the ROA, the enterprise value and Ln (Enterprise value) of 18 companies. Those ratios are average values calculated from November 2008 to March 2010. The leverage ratio is calculated by dividing the value of debt by the value of equity. The ROA is calculated by dividing the net income by the total asset. The enterprise value is provided by multiplying the stock price by the number of shares outstanding.

5 Result and Discussion

5.1 The result of the t-test for the AR(-1), AR(0),AR(+1), CAAR(0,+1), and CAAR(-1,+1)

The following tables are the result of the T-test for 10 announcements. Table 2

(16)

16 T-value

(t= -1) Sig T-value (t= 0) Sig T-value (t= +1) Sig QE1 3.624 0.002*** 2.941 0.009* 3.037 0.007*** QE2 2.337 0.031** -6.217 0 4.215 0.001*** QE3 0.19 0.851 5.801 0.000*** 2.708 0.014*** QE4 -0.657 0.519 5.243 0.000*** -8.013 0.000 QE5 3.206 0.005*** 1.636 0.119 2.01 0.060* QE6 -1.026 0.318 3.047 0.007*** 1.96 0.066* QE7 -5.981 0.000 1.671 0.112 1.67 0.112 QE8 1.84 0.082 -2.925 0.009 -3.241 0.005 QE9 3.084 0.006*** 1.256 0.225 3.821 0.001*** QE10 -1.701 0.106 2.988 0.008*** 0.572 0.574

*** Statistically significant at the 1 percent level.

**Statistically significant at the 2.5 percent level .

*Statistically significant at the 5 percent level.

Null hypothesis: Abnormal return is not significantly bigger than zero.

Table 3

T-test for the Cumulative Average Abnormal Return T-value

(0,+1) Sig T-value (-1,+1) Sig QE1 4.864 0.000*** 5.024 0.000*** QE2 -2.397 0.028 -0.829 0.418 QE3 6.079 0.000*** 2.83 0.011*** QE4 -2.659 0.016 -2.427 0.026 QE5 2.342 0.031** 3.682 0.002*** QE6 2.812 0.012*** 2.044 0.056* QE7 2.273 0.036** -1.065 0.301 QE8 -3.66 0.002 -2.758 0.013 QE9 3.758 0.001*** 5.471 0.000*** QE10 3.376 0.003*** 2.691 0.015***

*Statistically significant at the 5 percent level

Null hypothesis: the Cumulative Average Abnormal Return is not significantly bigger than zero.

Table 4

(17)

17

Announcement Date Description

QE1 11/25/2008 Initial Large-Scale-Asset Purchase program announcement

QE2 12/01/2008 Chairman speech

QE3 12/16/2008 Federal Open Market Committee Statement QE4 01/28/2009 Federal Open Market Committee Statement QE5 03/18/2009 Federal Open Market Committee Statement QE6 04/29/2009 Federal Open Market Committee Statement QE7 08/12/2009 Federal Open Market Committee Statement QE8 09/23/2009 Federal Open Market Committee Statement QE9 04/11/2009 Federal Open Market Committee Statement QE10 03/17/2010 Federal Open Market Committee Statement

In the previous days of the 1st, 2nd, 5th, 8th, and 9th announcements, the abnormal returns were detected. This indicates that the information about the announcement of the QE might have been leaked before it went public. However, on the days of the announcements, 2nd, 5th, 8th, 9th announcements are not statistically significant. This might have resulted from the leakages of the information. The synergies of the announcements were already incorporated at the share prices on the previous days and the share prices did not fluctuate on the days of the announcements. This is consistent with the fact that on the days of 3rd, 4th, 6th, 10th announcements, the abnormal returns are statistically significant.

On the previous day of those four announcements, the abnormal returns were not statistically significant. This indicates that there were no leakage of the information and the market learned the information about the QE on the days of the announcements for the first time and the share prices substantially reacted to the news on the days of the announcements, but not on the previous day. Further research with regard to the relationship between the abnormal return on the days of the announcements and the previous day is needed.

As for the results of the t-test for the next day of the announcements, 1st, 2nd, 3rd, 5th, 6th, and 9th announcements are statistically significant. It might demonstrate that it takes the market more than one day to incorporate new information and the market efficiency hypothesis is not valid in practice. Further research regarding whether or not the abnormal returns on those five days arose from the slagging adjusting time to the new information is needed.

On the results of the CAAR(0, +1), the scenario that the scheme of the QE had been leaked before it went public was ruled out. On the days of 1st, 2nd, 5th, 6th, 7th, 9th, and 10th announcements, the CAARs are statistically significant.

For the outcomes for the CAAR(-1, +1) where the possibilities of the leakage of information are taken into account, the CAAR(-1,+1)s are statistically significant on the days of 1st, 3rd, 5th, 6th, 9th, and 10th

(18)

18 announcements. The significant CAARs indicates that the market believed that the economy would recover from the recession with the positive impacts of the QE. Also, it implies that the market forecasted that the three mechanisms of the QE, the confidence channel, the asset price channel, and the bank lending channel, would lead to increases in the spending and the investment on the casino industry. Only for the 8th announcement, the abnormal returns are negative in five measurements(AR-1, AR0,AR+measurements(AR-1,CAAR(0,+1),CAAR(-measurements(AR-1,+1)). This might be either because the magnitude of the QE which the Federal Reserve planned to conduct did not suffice or there were other firm-specific factors that negatively affected the casino industry. Further research is needed for the 8th announcement.

5.2 The result of the cross-sectional regression.

In order to find out the economic variables that affected the cumulative abnormal returns, we conducted the cross-sectional regression of the CAAR(0,+1) and the CAAR(-1,+1) of the 18 casino companies on the leverage ratio, the ROA, and ln (Enterprise Value). The regression of each company is defined as:

𝐶𝐶𝐴𝐴𝐴𝐴𝑅𝑅_𝑖𝑖(0, +1) = 𝛽𝛽₀+ 𝛽𝛽₁( 𝐷𝐷𝐷𝐷𝑏𝑏𝑖𝑖

𝑄𝑄𝐸𝐸𝐸𝐸𝑖𝑖𝑖𝑖𝐸𝐸𝑖𝑖) + 𝛽𝛽2(𝑅𝑅𝑅𝑅𝐴𝐴𝑖𝑖) + 𝛽𝛽3ln(𝑄𝑄𝐸𝐸𝑖𝑖) + 𝜖𝜖𝑖𝑖 (22)

𝐶𝐶𝐴𝐴𝐴𝐴𝑅𝑅_𝑖𝑖(−1, +1) = 𝛽𝛽₀+ 𝛽𝛽₁(𝐷𝐷𝐷𝐷𝑏𝑏𝑖𝑖

𝑄𝑄𝐸𝐸𝐸𝐸𝑖𝑖𝑖𝑖𝐸𝐸𝑖𝑖) + 𝛽𝛽2(𝑅𝑅𝑅𝑅𝐴𝐴𝑖𝑖) + 𝛽𝛽3ln(𝑄𝑄𝐸𝐸𝑖𝑖) + 𝜖𝜖𝑖𝑖 (23)

The following tables are the result of the Glejser test for the CAAR(0,+1) and the CAAR (-1,+1). This test tells us whether the residual of the dependent variables suffer from heteroscedasticity or not.

Table 5 Glejser test

Residual of CAAR(0,+1) Residual of CAAR(-1,+1)

T-value Sig T-value Sig

(Constant) 0.12 0.906 0.626 0.541

Leverage -1.363 0.194 -2.043 0.06

ROA -0.577 0.573 0.902 0.383

Ln(Enterprise Value) 0.927 0.37 0.292 0.775

The residuals of the CAAR(0,+1) and the CAAR(-1,+1) are regressed on the leverage ratio, the ROA, and ln (Enterprise Value).

If the t-value of each explanatory variable were statistically significant, it implies that the residual has linear relationships with explanatory variables and thereby suffers from heteroscedasticity.

(19)

19

The p-values of the three explanatory variables are larger than the significance level of 5%. Thereby, it can be concluded that the residuals of the CAAR are not linearly related to the three explanatory variables. Thus, as the error term is constant, the robust standard error will be applied in the cross-sectional regression.

The following tables are the results of the cross-sectional regression of the Cumulative Average Abnormal Return for the 18 companies from the event day to the next day and the previous day of the event to the next day. Since financial variables of Melco Resorts & Entertainment in 2008 are not available, it is excluded from the regression.

Table 6

Regression Analysis for CAAR(0,+1)

Beta T-value Sig R Square Adjusted R Square

(Constant) -0.037 -1.286 0.219 0.629 _0.549

Leverage 0.006 2.774 0.015***

ROA -0.071 -1.109 0.286

Ln(Enterprise Value) 0.006 2.293 0.038**

The Cumulative Average Abnormal Return of 18 companies are regressed on the leverage ratio, the ROA, and ln(Enterprise Value) from the day of the QE announcements to the next day.

**Statistically significant at the 2.5 percent level.

Table 7

Regression Analysis for CAAR(-1,+1)

Beta T-value Sig R Square Adjusted R Square

(Constant) -0.041 -1.119 0.282 0.605 _0.521

Leverage 0.005 1.979 0.068*

ROA -0.251 -3.064 0.008***

Ln(Enterprise Value) 0.009 2.715 0.017***

The Cumulative Average Abnormal Return of 18 companies are regressed on the leverage ratio, the ROA, and ln(Enterprise Value) from the previous day of the QE announcement to the next day of the QE announcement.

(20)

20

**Statistically significant at the 2.5 percent level.

The first explanatory variable, the leverage ratio, is statistically significant in both regressions. It displays the positive relationship between the stock returns and the leverage ratio. Most states impose higher tax rate on the casino revenue. Not only that, states tend to change tax structures subject to the casino industry during the recession so that they can collect more tax revenues from the casino industry. Due to the higher tax rate, the casino companies benefit from the high leverage ratio. The higher interest payment from the higher leverage ratio reduces the taxable income. This is called the tax-shield. Corporations generate high cash flow and increase the value of companies through the tax –shield. The amount of tax-shield generated from the interest expense is defined as :

Tax Shield = 𝜏𝜏*(Interest Expense) (24)

This indicates that companies with high leverage ratio generate large amounts of cash flow when the tax rate is high. The casino sector which higher tax-rate is levied on has more incentives than other sectors to carry more debt. For this reason, the leverage ratio contributes to the increase in the stock returns of the casino companies. The second explanatory variable, the Return On Assets, is only statistically significant on the regression of the CAAR( -1, +1). The ROA measures how much net income companies generate with their assets. In general, the ROA positively affects the stock returns. This is because the ROA refers to how profitable companies are with their invested capital and companies with higher ROA are more likely to perform well. Nevertheless, the coefficient of the ROA is negative. Further research is needed regarding why the ROA negatively affected the CAAR on those periods. The last explanatory variable, the enterprise value, are statistically significant in both regressions. The significant coefficients of the ln( 𝑄𝑄𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸 𝐸𝐸𝑉𝑉𝑉𝑉𝑉𝑉𝐸𝐸) indicate that the size of companies positively affected the stock returns. This is because large-sized firms are in the leading position and have various sources to finance themselves in comparison with small and middle-sized firms. With those various tools to finance themselves, large-sized firms can readily respond to the market changes.

6 Conclusion

This paper investigated how the unconventional monetary policy called the Quantitative Easing affected the share prices in the casino industry. With the event-study method, 19 casino companies` share prices were examined on the days of the 10 announcements. As a consequence, there were significant positive abnormal returns on the previous day of five announcements. This proves that the schemes of the QE were already leaked to the market. On top of that, the abnormal returns on the day

(21)

21 of the announcements are not statistically significant for those five announcements of which abnormal returns are significant on the previous day. This implies that the information about the QE was already incorporated on the stock prices due to the leakages on the previous day and the share prices on the day of the announcements were not affected. Among the ten announcements for the next day of the announcement, six of them were statistically significant. This implies that the market efficiency hypothesis is not applicable in practice. Further research regarding the abnormal returns for the next few days after the event is needed as there are possibilities that it might take the market a couple of days to adjust to new information.

The CAAR(0, +1) and the CAAR(-1, +1) that incorporate the abnormal returns on the non-announcement days indicate that the 1st,2nd, 3rd, 5th, 6th, 7th, 9th, and 10th QE contributed to the increase in the share price of the casino companies.

The cross-sectional regression of the Cumulative Average Abnormal Returns was conducted in order to find out what caused the higher returns on the event days. While the leverage ratio and the ln (𝑄𝑄𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸 𝐸𝐸𝑉𝑉𝑉𝑉𝑉𝑉𝐸𝐸) are positively related to the CAARs, the ROA displays the negative relationship with the CAAR (-1, +1) or does not show linear relationships with the CAAR(0,+1).

However, this paper has several limitations. To begin with, since most casino companies are not publicly listed and thereby their financial data is not available, only 19 companies` share prices were investigated. It is hard to apply those results to the entire casino companies in the U.S. Further research for the remainder casino companies is needed. Secondly, the fluctuations of the share prices are the estimated impacts of the Quantitative Easing which stock investors measured. This does not necessarily translate into an improvement in profits in the casino industry. Further research is needed to find out whether the QE improved the companies` revenues over a long period of time. Lastly, this paper only investigated the impact of the first round of the QE. There were 2nd and 3rd rounds of the QEs. This paper excluded the 2nd and 3rd rounds of the QE for the reason that they occurred more than 2 years later when the financial crisis hit the U.S and thereby there are more factors to be considered in order to investigate the impact of the 2nd and 3rd rounds of the QE. Further research for the remainder QEs is needed.

(22)

22

References

Dadayan, L. (2016). State revenues from gambling. Rockefeller Institute, April.

De Jong, F. (2007). Event studies methodology. Lecture Notes.

Eadington, W. R. (2011). After the great recession: The future of casino gaming in America and Europe. Economic Affairs, 31(1), 27-33.

Elliott, D. S., & Navin, J. C. (2002). Has riverboat gambling reduced state lottery revenue?. Public

Finance Review, 30(3), 235-247.

Gagnon, J. (2010). Large-Scale Asset Purchases by the Federal Reserve: Do They Work?. DIANE Publishing.

Garrett, T. A., Pakko, M. R., & Board, G. (2009). Casino revenue and the Illinois smoking ban (No. 2009-027).

Horváth, C., & Paap, R. (2012). The effect of recessions on gambling expenditures. Journal of

Gambling Studies, 28(4), 703-717.

Joyce, M., Tong, M., & Woods, R. (2011). The United Kingdom’s quantitative easing policy: design, operation and impact.

Legg, M. P., & Tang, H. (2011). Why Casinos are not recession proof: An business cycle econometric case study of the Las Vegas region.

MacKinlay, A. C. (1997). Event studies in economics and finance. Journal of economic

literature, 35(1), 13-39.

Petry, N. M., & Oncken, C. (2002). Cigarette smoking is associated with increased severity of gambling problems in treatment‐seeking gamblers. Addiction, 97(6), 745-753.

Smaghi, L. B. (2009). Conventional and unconventional monetary policy. Speech at the Center for

Monetary and Banking Studies, Geneva, 28.

Thalheimer, R., & Ali, M. M. (2008). The demand for casino gaming with special reference to a smoking ban. Economic Inquiry, 46(2), 273-282.

Thornton, D. L. (2017). Effectiveness of QE: An assessment of event-study evidence. Journal of

Macroeconomics, 52, 56-74.

Zhang, J. H., Tam, K. P., & Zhou, N. (2016). Do smoking bans always hurt the gaming industry? Differentiated impacts on the market value of casino firms in Macao. Economics, 10(28), 1.

(23)

23

Appendix

Appendix 1. Financial variables of the casino companies

Century Casino

N Minimum Maximum Mean Std. Deviation

Statistic Statistic Statistic Statistic Std. Error Statistic

Leverage 10 .13 .36 .3120 .02105 .06655

ROA 10 .02 .05 .0307 .00302 .00956

Enterprise Value 10 20540.24 71000.82 45275.5280 6275.17916 19843.85886

Ln(EV) 10 9.93 11.17 10.6247 .14980 .47370

Valid N (listwise) 10

Pinnacle Entertainment INC

Leverage 10 .45 .68 .5207 .02301 .07276 ROA 10 .03 .07 .0489 .00458 .01447 Enterprise Value 10 268647.68 736052.76 490257.0100 46463.10046 146929.22461 Ln(EV) 10 12.50 13.51 13.0575 .10358 .32753 Valid N (listwise) 10 Nevada Gold

Leverage 10 .15 .55 .3188 .05207 .16466 ROA 10 -.08 -.06 -.0744 .00228 .00720 Enterprise Value 10 5563.77 15397.41 10499.9985 993.66737 3142.25213 Ln(EV) 10 8.62 9.64 9.2168 .09894 .31287 Valid N (listwise) 10 Canterbury

Leverage 10 .15 .55 .3188 .05207 .16466 ROA 10 -.08 -.06 -.0744 .00228 .00720 Enterprise Value 10 5563.77 15397.41 10499.9985 993.66737 3142.25213 Ln(EV) 10 8.62 9.64 9.2168 .09894 .31287 Valid N (listwise) 10 Wynn Resort

(24)

24

Leverage 10 .53 .74 .6849 .02074 .06557 ROA 10 .08 .12 .0977 .00574 .01814 Enterprise Value 10 2371336.38 9078284.48 5623184.2990 715506.06364 2262628.84076 Ln(EV) 10 14.68 16.02 15.4651 .13404 .42386 Valid N (listwise) 10 Monarch Casino

Leverage 10 .07 .35 .2376 .04018 .12706 ROA 10 .13 .21 .1664 .01118 .03535 Enterprise Value 10 78514.14 181533.72 137798.9940 10395.38512 32873.09414 Ln(EV) 10 11.27 12.11 11.8046 .08288 .26208 Valid N (listwise) 10 MGM Resort

Leverage 10 .62 .98 .7846 .04742 .14995 ROA 10 .05 .08 .0703 .00398 .01260 Enterprise Value 10 793563.61 5748249.65 3383796.9200 481840.74361 1523714.21929 Ln(EV) 10 13.58 15.56 14.9098 .18468 .58401 Valid N (listwise) 10 Full House

Leverage 10 .03 .30 .1845 .03038 .09606

ROA 10 -.03 .14 -.0088 .01666 .05269

Ln(EV) 10 9.87 10.96 10.3447 .14121 .44656

Valid N (listwise) 10 Las Vegas Sands

Leverage 10 .61 .81 .7431 .02139 .06763

ROA 10 .04 .07 .0531 .00263 .00832

Ln(EV) 10 14.22 16.40 15.3411 .26200 .82851

(25)

25

Churchill Down

Leverage 10 .10 .22 .1305 .01160 .03669 ROA 10 .09 .11 .1036 .00206 .00652 Enterprise Value 10 391094.73 530112.60 469390.9980 17170.26407 54297.14249 Ln(EV) 10 12.88 13.18 13.0530 .03737 .11818 Valid N (listwise) 10 Empire Resort

Leverage 10 1.69 4.18 2.9929 .29604 .93615

ROA 10 -.07 .00 -.0373 .00653 .02064

Ln(EV) 10 9.84 11.76 11.0314 .18849 .59605

Dover Down Gaming Entertainment

Leverage 10 .47 .52 .5014 .00450 .01425 ROA 10 .15 .25 .2090 .01036 .03277 Enterprise Value 10 44109.00 94491.15 61918.0670 5523.89336 17468.08457 Ln(EV) 10 10.69 11.46 11.0008 .08347 .26395 Valid N (listwise) 10 Penn National

Leverage 10 .56 .74 .6371 .02704 .08552 ROA 10 .11 .13 .1166 .00328 .01037 Enterprise Value 10 1476543.60 2582678.07 1939252.3690 130567.04007 412889.23398 Ln(EV) 10 14.21 14.76 14.4577 .06659 .21059 Valid N (listwise) 10 Boyd Gaming

Leverage 10 .63 .70 .6711 .01037 .03280

ROA 10 .06 .09 .0785 .00358 .01131

(26)

26

Ln(EV) 10 12.71 13.82 13.2073 .13335 .42169

Valid N (listwise) 10 Ameristar

Leverage 10 .79 .90 .8251 .01065 .03368 ROA 10 .13 .15 .1367 .00198 .00627 Enterprise Value 10 339539.94 1048744.45 738006.1670 89300.28430 282392.29410 Ln(EV) 10 12.74 13.86 13.4326 .13872 .43866 Valid N (listwise) 10 SHFL Entertainment

Leverage 10 .38 .88 .7133 .05289 .16724 ROA 10 .12 .18 .1390 .00537 .01698 Enterprise Value 10 155019.60 477218.00 292226.7410 38895.47031 122998.27685 Ln(EV) 10 11.95 13.08 12.5058 .13251 .41904 Valid N (listwise) 10 Lakes Entertainment

Leverage 10 .00 .17 .0847 .02262 .07152

ROA 10 -.06 .07 .0126 .01827 .05779

Ln(EV) 10 10.91 11.53 11.2590 .05566 .17602

Valid N (listwise) 10 Isle of Capri Casino

Leverage 10 .86 .90 .8759 .00445 .01407

ROA 10 .09 .10 .0961 .00152 .00482

Ln(EV) 10 11.37 12.83 12.0846 .18511 .58536

(27)

27

Test Value = 0

t df Sig. (2-tailed) Mean Difference

95% Confidence Interval of the Difference Lower Upper QE1 3.624 18 .002*** .09324 .0392 .1473 QE2 2.337 18 .031** .02578 .0026 .0490 QE3 .190 18 .851 .00657 -.0661 .0792 QE4 -.657 18 .519 -.00455 -.0191 .0100 QE5 3.206 18 .005*** .06960 .0240 .1152 QE6 -1.026 18 .318 -.01199 -.0365 .0126 QE7 -5.981 18 .000 -.03441 -.0465 -.0223 QE8 1.840 18 .082 .01252 -.0018 .0268 QE9 3.084 18 .006*** .03732 .0119 .0628 QE10 -1.701 18 .106 -.00950 -.0212 .0022

Appendix 3. Statistic result of the t-test on the announcement days

Test Value = 0

95% Confidence Interval of the Difference Lower Upper QE1 2.941 18 .009* .06291 .0180 .1079 QE2 -6.217 18 .000 -.11512 -.1540 -.0762 QE3 5.801 18 .000*** .07437 .0474 .1013 QE4 5.243 18 .000*** .03561 .0213 .0499 QE5 1.636 18 .119 .01864 -.0053 .0426 QE6 3.047 18 .007*** .04251 .0132 .0718 QE7 1.671 18 .112 .01161 -.0030 .0262 QE8 -2.925 18 .009 -.01729 -.0297 -.0049 QE9 1.256 18 .225 .01232 -.0083 .0329 QE10 2.988 18 .008*** .01914 .0057 .0326

(28)

28

Test Value = 0

95% Confidence Interval of the Difference Lower Upper QE1 3.037 18 .007*** .07467 .0230 .1263 QE2 4.215 18 .001*** .07252 .0364 .1087 QE3 2.708 18 .014*** .03773 .0085 .0670 QE4 -8.013 18 .000 -.05621 -.0709 -.0415 QE5 2.010 18 .060* .02568 -.0012 .0525 QE6 1.960 18 .066* .04517 -.0033 .0936 QE7 1.670 18 .112 .01252 -.0032 .0283 QE8 -3.241 18 .005 -.03186 -.0525 -.0112 QE9 3.821 18 .001*** .04239 .0191 .0657 QE10 .572 18 .574 .00278 -.0074 .0130

Appendix 5. Statistic result of the t-test for the CAAR

(0, +1)

Test Value = 0

95% Confidence Interval of the Difference Lower Upper QE1 4.864 18 .000*** .13758 .0782 .1970 QE2 -2.397 18 .028 -.04260 -.0799 -.0053 QE3 6.079 18 .000*** .11210 .0734 .1508 QE4 -2.659 18 .016 -.02059 -.0369 -.0043 QE5 2.342 18 .031** .04432 .0046 .0841 QE6 2.812 18 .012*** .08768 .0222 .1532 QE7 2.273 18 .036** .02413 .0018 .0464 QE8 -3.660 18 .002 -.04915 -.0774 -.0209 QE9 3.758 18 .001*** .05471 .0241 .0853 QE10 3.376 18 .003*** .02192 .0083 .0356

Appendix 6. Statistic result of the t-test for the CAAR

(-1 , +1)

Test Value = 0

95% Confidence Interval of the Difference

(29)

29 QE1 5.024 18 .000*** .23082 .1343 .3273 QE2 -.829 18 .418 -.01682 -.0594 .0258 QE3 2.830 18 .011*** .11866 .0306 .2067 QE4 -2.427 18 .026 -.02514 -.0469 -.0034 QE5 3.682 18 .002*** .11392 .0489 .1789 QE6 2.044 18 .056* .07568 -.0021 .1535 QE7 -1.065 18 .301 -.01028 -.0306 .0100 QE8 -2.758 18 .013 -.03664 -.0645 -.0087 QE9 5.471 18 .000*** .09203 .0567 .1274 QE10 2.691 18 .015*** .01242 .0027 .0221

Appendix 7. Results of the cross-sectional regression

Model Summary Model R R Square Adjusted R Square Std. Error of the Estimate 1 .793a .629 .549 .01658 a. Predictors: (Constant), Ln (Enterprise Value), ROA, Leverage

ANOVAa

Model Sum of Squares df Mean Square F Sig.

1 Regression .007 3 .002 7.900 .003***

Residual .004 14 .000

Total .010 17

a. Dependent Variable: CAAR(0,+1)

b. Predictors: (Constant), Ln (Enterprise Value), ROA, Leverage

Coefficientsa Model Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta 1 (Constant) -.037 .029 -1.286 .219 Leverage .006 .002 .543 2.774 .015*** ROA -.071 .064 -.199 -1.109 .286 Ln (Enterprise Value) .006 .002 .441 2.293 .038**

a. Dependent Variable: CAAR(0,+1)

Model Summary Model R R Square Adjusted R Square Std. Error of the Estimate

(30)

30

1 .778a .605 .521 .02138 a. Predictors: (Constant), Ln (Enterprise Value), ROA, Leverage

ANOVAa

1 Regression .010 3 .003 7.162 .004***

Residual .006 14 .000

Total .016 17

a. Dependent Variable: CAAR(-1,+1)

Coefficientsa Model Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta 1 (Constant) -.041 .037 -1.119 .282 Leverage .005 .003 .399 1.979 .068* ROA -.251 .082 -.567 -3.064 .008*** Ln(Enterprise Value) .009 .003 .538 2.715 .017***

Model Summary Model R R Square Adjusted R Square Std. Error of the Estimate 1 .778a .605 .521 .02138 a. Predictors: (Constant), Ln (Enterprise Value), ROA, Leverage

ANOVAa

1 Regression .010 3 .003 7.162 .004***

Residual .006 14 .000

Total .016 17

Coefficientsa Model Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta

(31)

31

1 (Constant) -.041 .037 -1.119 .282

Leverage .005 .003 .399 1.979 .068*

ROA -.251 .082 -.567 -3.064 .008***

Ln(Enterprise Value) .009 .003 .538 2.715 .017***

Appendix 8. Results of the Glejser test

Coefficientsa Model Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta 1 (Constant) .002 .015 .120 .906 Leverage -.001 .001 -.400 -1.363 .194 ROA -.019 .032 -.155 -.577 .573 LN (Enterprise Value) .001 .001 .267 .927 .370

a. Dependent Variable: Residual of CAAR(0,+1)

Coefficientsa Model Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta 1 (Constant) .012 .019 .626 .541 Leverage -.003 .001 -.568 -2.043 .060 ROA .037 .042 .230 .902 .383 Ln (Enterprise Value) .000 .002 .080 .292 .775

The impact of the Quantitative Easing on the casino industry in the U.S.A.