Prepare slides for the article

Value Relevance of Financial
and Non-Financial Information:
Evidence from the Gaming
Industry
Lisa Goh
Kevin C.K. Lam
Hong Weng Lawrence Lei
Abstract
Using financial and non-financial data from casino gaming firms listed in the
United States from 1999–2017, we explore two research questions: (1) Is financial
information value relevant to financial markets in the casino gaming industry? (2) Does
non-financial information have incremental explanatory power over financial information?
In general, we find that accounting numbers can explain a firm’s market value and stock
returns in the casino gaming industry, except for accounting accruals, which may behave
differently compared to other industries. We also find that non-financial information, such
as the number of table games, number of slot machines, and their relative proportion, have
significant value relevance in explaining market valuation. Our findings contribute to a
better understanding of the value relevance of financial and non-financial information in
the casino gaming industry. We also provide analysis of firms characterized by these nonfinancial attributes.
Dr. Lisa Goh
Assistant Professor,
Department of
Accountancy, School of
Business, The Hang
Seng University of Hong
Kong,
lisagoh@hsu.edu.hk
Keywords: hospitality, casino, gaming, value relevance, table games, slot machines
JEL Code: L83, M19, M41
Professor Kevin C.K. Lam
Professor, Department of
Accountancy, School of
Business, The
Hang Seng University of
Hong Kong,
kevinlam@hsu.edu.hk
Dr. Hong Weng
(Lawrence) Lei
Assistant Professor,
Department of
Accountancy, School of
Business, The Hang Seng
University of Hong Kong,
lawrencelei@hsu.edu.hk
UNLV Gaming Research & Review Journal t Volume 23 Issue 1
41
Understanding how the stock markets use financial and non-financial information
is of core importance to companies, investors, regulators, and other stakeholders. “Value
relevance,” or the ability of a firm’s financial and non-financial information to explain and
predict firm value and stock prices, is a key topic in accounting and financial research.
Information is considered value relevant to investors if it informs their investment
decision-making, and the relationships between information, firm value, and stock prices
can also be informative for accounting standard-setters and financial regulators when
implementing policy. In this study, we examine the power of financial and non-financial
information to explain stock market valuation and stock returns in a subset of firms in
the hospitality industry, namely casino gaming firms. Prior research in accounting and
finance has examined the relationship between accounting information and stock market
prices and returns (Francis & Schipper, 1999; Sami & Zhou, 2004) and found that primary
financial information such as earnings, book value of assets, and cash flows contributes to
explaining market values (Ou & Penman, 1989; Lev, 1989; Collins et al., 1997; Francis &
Schipper, 1999; Lev & Zarowin, 1999; Landsman & Maydew, 2002). In addition, studies
have found that accounting accruals have incremental explanatory power for market
values and stock returns (Rayburn, 1986; Dechow, 1994; Barth et al., 1999).
Recent evidence indicates that financial information has become less value
relevant over time, as business models evolve. Researchers have therefore started to
examine the value relevance of non-financial information. Amir and Lev (1996) document
that financial and non-financial information are complementary, and several studies even
find that non-financial information is more relevant than financial information (Graham et
al., 2001; Ittner & Larcker, 1998; Hughes, 2000; Riley et al., 2003).
In this study, we use both financial and non-financial measures from casino
gaming firms that are publicly traded in the United States.1 We study the casino gaming
industry for several reasons. First, it contributes substantially to the national and regional
economy. Gaming revenues in the U.S. were around $115 billion in 2016, and are forecast
to reach $130 billion in 2019 (Lock, 2018). Second, there is a high level of idiosyncratic
risk in the casino gaming industry; the role of casino gaming firms as quasi-financial
firms raises the issue of whether their financial numbers can explain their stock values.
Lastly, publicly listed casino gaming firms have long been classified as “sin” stocks whose
activities are frowned upon and considered unethical or immoral; there is evidence of
a consequent discounting of their share prices (Cheung & Lam, 2015). Thus, how the
financial information of casino gaming firms is valued and trusted by financial markets
may be different from the case of other firms.
Using data from casino gaming operators publicly listed in the United States from
1999 to 2017, we evaluate the usefulness of financial and non-financial information for
market valuation and stock returns. Our financial variables include assets, liabilities, book
value, earnings, accrual and cash flow from operations. As casino gaming revenues are
generated from customer spending on table games and slot machines, we use the number
of table games and slot machines as our major non-financial variables.
We find that financial measures, except for accrual numbers, have strong
explanatory power for market valuation and market-adjusted returns in the casino gaming
industry, which reflect a firm’s stock market performance compared to that of a benchmark
index. Our findings are of interest to accounting researchers because, while confirming
the validity of most financial variables, unlike prior accounting research (Rayburn,
1986; Dechow, 1994; Barth et al., 1999; Barth et al., 2001), we find that accounting
accruals are not value relevant to market valuation and stock returns in the casino gaming
industry. One possible explanation is that a large proportion of casino gaming revenues
are generated directly from cash spending by customers, and accruals, such as receivables
from gamblers, are significantly lower than in other industries and may be priced
differently.
1 As revenues of the U.S. gaming industry come largely from casinos, we use these terms interchangeably, with
casino revenues representing substantially all gaming revenues, excluding the video game industry.
42
UNLV Gaming Research & Review Journal t Volume 23 Issue 1
Value Relevance of Financial and Non-Financial Information
In terms of non-financial variables, we find that the number of slot machines
and number of gaming tables, and the ratio between them, have significant power in
explaining stock market values, incremental to financial information. Even after financial
variables are incorporated, market valuations are higher for casino gaming firms with
more table games, fewer slot machines, and a higher ratio of gaming tables to slot
machines. Our results are reasonable based on our survey of the casino gaming literature;
Thalheimer and Ali (2008) find that the addition of table games to the gaming mix
can significantly increase revenue. Recent evidence from the media also suggests that
casinos are rebalancing away from slots towards table games (Heim, 2015). However,
slot machines have become increasingly important in generating gaming revenues. For
example, Boylan (2016) suggests that technological developments have led to increased
revenue and efficiency from slot machines, offsetting some “economic challenges” from
table games. However, our non-financial variables are significant primarily in explaining
market valuation, and are only weakly related to stock returns. Hence, these non-financial
variables are better able to explain variation in long-term patterns, rather than variations
in annual returns.
As the first study of the value relevance of financial and non-financial
information in the casino gaming industry, this study makes several contributions. First,
we contribute to literature in accounting by studying the importance of financial and nonfinancial measures to stock prices in the casino gaming industry. We find that financial
information is generally value relevant while accrual measures are not. We also find
that non-financial variables in the casino gaming industry are generally value relevant
in explaining market valuation. Both results are new. Second, we provide insights for
hospitality management based on our study of our non-financial variables and their
relationships with the underlying gaming operations.
Literature Review
The Casino Gaming Industry in the United States
Several regulatory changes have led to significant developments in the casino
gaming industry (Eadington, 1999). Nevada’s Corporate Gaming Act of 1969 allowed
listed companies to hold gaming licenses for the first time. Subsequently, major
hospitality firms such as Hilton, Holiday Inn, MGM, and Ramada entered the casino
market. The period after 1988 saw a rapid expansion of casino gaming activities after the
passing of the Indian Gaming Regulatory Act by Congress in 1988 (Eadington, 1999).
Growth in the number of casinos and gaming opportunities has continued in recent
years, including both online and land-based gaming (Huber, 2015). Commercial casinos
now exist in 24 states (American Gaming Association, 2015) and Native Americanoperated casino gaming facilities now exist in 28 states. Industry estimates suggest that
commercial casinos provide up to 350,000 jobs and have a wider economic impact of
$240 billion (American Gaming Association, 2016), illustrating the economic significance
of the casino gaming industry.
Economic Costs and Benefits of the Casino Gaming Industry
Research has examined the economic benefits and costs of casino gaming
development. First, in terms of benefits, casinos “yield positive economic benefits on net
to [their] host economy” (Rose and Associates, 1998), including job creation (Regional
Economics Applications Laboratory, 2003; Morse & Goss, 2007; Eadington, 1999), and
increased government revenues through gaming taxes, which can be used for public
services to benefit local citizens (Williams et al., 2011; Eadington, 1999). The economic
costs associated with an increased number of casino gaming facilities include regulatory
costs, infrastructural upgrades, and social costs (Williams et al., 2011). Regulatory costs
are the cost of government oversight over gaming operations. Infrastructure costs relate
UNLV Gaming Research & Review Journal t Volume 23 Issue 1
43
to infrastructure upgrades when new casino gaming facilities are introduced (e.g. public
transportation, police and emergency services, roads, water services, sewage treatment).
The most well-known social cost associated with casinos is the possibility of problematic
gambling (addiction), which can lead to mental health issues, bankruptcy, and divorce,
and related problems.
In the casino gaming industry, gaming revenues derive primarily from two
sources: table games and slot machines. Overall, table games are likely to contribute
more to total gaming revenues, as the value of bets placed at table games is much larger
than at slot machines. For example, Siu and Eadington (2009) find that table games
generate much higher revenue and profit per square foot utilization, compared to slot
machines. They also find that table games generate a substantially higher proportion of
revenue for casinos in Macau than in Europe and North America, providing evidence
of international variations in sources of revenue generation. However, Boylan (2016)
suggests that technological developments have led to significant increases in slot machine
revenue-generation over time. Evidence also suggests that, to some degree, revenues
from tables and slots are inversely related. Thalheimer and Ali (2008) find that revenue
from slot machines increases with the number of slot machines, but decreases with the
number of table games, based on a sample of 27 racinos and riverboat casinos in the
Midwestern United States.
Researchers have also examined non-gaming revenues generated by casino
gaming operators. Competition due to legalization of gambling and the increasing
number of casinos in the U.S. means that non-gaming facilities are important as a source
of revenue-generation. The prior literature on these non-gaming revenues suggests
that the wider customer experience plays an important role in determining the revenues
and profitability of hotel-casino resorts (Suh, 2011; Tanford & Suh, 2012). Anecdotal
evidence from industry executives also suggests that senior managers view investment in
these non-gaming facilities as important in attracting customers and increasing gaming
revenues (Brinkerhoff-Jacobs, 2015).
Value Relevance of Financial and Non-Financial Information
A well-established body of research in the accounting literature has investigated
the importance of both financial and non-financial information to stock market prices.
This body of research measures the informativeness of accounting information concerning
market value and returns, i.e. the explanatory power of financial statements (namely
earnings, cash flows, and book value) for market prices. This helps investors assess the
informativeness of financial statements in their investment decisions; the higher the value
relevance, the more investors can rely on financial statements when making investment
decisions (Francis & Schipper, 1999; Sami & Zhou, 2004). However, researchers have
found mixed results on the stability of financial statement value relevance over time. Lev
(1989) suggests that the quality of the information content matters to stock prices, and
that the informativeness of financial statements declines with lower quality information.
For example, accounting standards allow for some intangible assets, but the cost and
value of these intangible assets cannot be fully recognized and measured by traditional
reporting models (Lev & Zarowin, 1999).
Accounting standards in the U.S. impose a standardized set of rules on businesses,
regardless of business model. As a result, the ability of accounting information to fully
reflect a firm’s condition is constrained. Thus, researchers have investigated the impact
of non-financial measures on stock prices. Amir and Lev (1996) find that for firms in
the wireless industry, including non-financial measures improves the informativeness
of financial measures, suggesting that financial and non-financial information are
complementary. Ittner and Larcker (1998) find that a customer satisfaction measure
is useful in explaining financial performance and stock prices. Riley et al. (2003) find
that several non-financial measures are value relevant for the airline industry, such as
44
UNLV Gaming Research & Review Journal t Volume 23 Issue 1
Value Relevance of Financial and Non-Financial Information
load factor, available capacity, and market share. Hughes (2000) finds that air pollution
measures are value relevant in explaining the market value of equity for electricity
utility firms. Graham et al. (2001) find that non-financial measures of Internet usage,
like unique users for service companies, and page views for e-retailers and content or
community firms, are more value relevant than other non-financial variables. Singal
(2012) suggests that consumer sentiment, i.e. confidence in the economy, affects stock
returns in the hospitality industry.
Accounting and Financial Research in the Casino Gaming Industry
Accounting and financial research specific to the casino gaming industry is
scarce and has largely concentrated on investment risk. In the decade before 2007, a
significant growth trend in both revenue and stock price was observed in the casino
industry, but right after 2007 revenue and stock price in the casino industry started to
decline due to recession (Repetti & Kim, 2010). Using financial data on casino gaming
operators from 1992–1994, Gu and Kim (1998) find that over 90% of investment risk
in the casino gaming industry is unsystematic. Casino gaming stocks also have higher
abnormal returns and variability than stocks of non-casino firms (Cheung & Lam,
2015), with IPOs that tend to underperform (Borghesi et al., 2015). Gu and Kim (1998)
evaluate different financial ratios and find that only asset turnover ratio is significantly
negatively related to beta, while other ratios have no statistically significant correlation.
The findings of Repetti and Kim (2010) are somewhat consistent with those of Gu and
Kim (1998). They find that the asset turnover ratio is significantly related to beta during
times of recession. Repetti and Kim (2010) also find that liabilities as a percentage of
assets are significantly positive during recession, suggesting that gaming companies are
very sensitive to economic downturn. Thus, gaming firms need to be careful to manage
their debts during recession, due to the associated financial risk. Repetti and Kim (2010)
find that other financial ratios such as the quick ratio, return on assets, and EBIT growth
rate are insignificant determinants of beta. We are not aware of any prior research on the
informativeness of financial or non-financial information in the casino gaming industry,
even though its “non-traditional” business model suggests that non-financial measures
may be particularly informative to investors.
Research Design
Sample Selection
Our sample consists of all U.S.-listed firms in the Compustat database with
returns data from CRSP, that reported gaming-related revenues for the period 1999 to
2017, the most recently available year at the time of data collection. We identify 38
casino gaming operators with the required data in Compustat. We delete four firms that
are primarily manufacturers of slot machines, rather than gaming operators. A further
14 firms are missing returns data, resulting in a final sample of 20 firms for analyse
of returns.2 We delete firms with negative equity value, as they are not going concern
operations, and winsorize outliers at the 1st and 99th percentiles. We apply a panel data
approach and examine cross-sectional (subscript j) and time series data (subscript t), for
the period from 1999, the first year for which industry-specific data for casino gaming
firms are available, until 2017. Our final sample consists of 244 firm-years from 20
unique firms. Appendix A lists the 20 casino gaming firms and the number of years that
each firm appears in our sample.
Hypothesis Development
The first objective of our study is to examine how established models of
valuation apply to the casino gaming industry. In general, empirical evidence thus
far from the accounting and finance literature has established significant relationships
2 Non-financial measures for the casino gaming industry can be found on Compustat’s Capital IQ data set under
“Industry-specific annual” data (S&P, 2008).
UNLV Gaming Research & Review Journal t Volume 23 Issue 1
45
between financial information (positively with assets, earnings, book value of equity,
accruals, and cash flows, and negatively with liabilities) and market valuation (Collins
et al., 1997; Francis & Schipper, 1999; Landsman & Maydew, 2002). However, as
noted earlier, researchers have also suggested that this simple characterization is not
applicable to all industries, as specific industry characteristics may lead investors to rely
differently on financial information, and on non-financial information to complement or
contextualize financial information (Amir & Lev, 1996; Riley et al., 2003). Moreover,
casino gaming firms may behave somewhat differently from other firms, as they have
higher abnormal returns and higher idiosyncratic risk, which suggests that investors rely
on more firm-specific information (Cheung & Lam, 2015). Here, we adopt a baseline
analysis using financial variables from prior research (earnings, book value of equity,
assets, liabilities, accruals, and cash flows), and predict that the market’s use of financial
information will behave in a manner similar to other industries, and that all of these
variables are value relevant to equity prices and stock market returns. We predict positive
relationships with earnings, book value of equity, assets, accruals, and cash flows, and a
negative relationship with liabilities; our hypothesis is stated as follows:
H1: Accounting information is value relevant to equity market values and stock
market returns in the casino gaming industry.
We then examine whether non-financial information provides incremental
information content to markets in valuing casino gaming firms. Researchers including
Amir and Lev (1996), Ittner and Larcker (1998), and Riley et al. (2003) find that nonfinancial information has additional explanatory power in predicting market values
and returns, complementing financial information. Therefore, including non-financial
information in a valuation model may be superior to using financial information alone.
For casino gaming firms, our non-financial measures include the number of table games,
number of slot machines, table game to slot machine ratio, and relative importance of
each in the casino’s operations. We expect non-financial information to be incrementally
value relevant over financial information. However, we make no directional prediction
for these variables. We develop our second hypothesis as follows:
H2: Non-financial information is incrementally value relevant to market prices
and stock returns in the casino gaming industry.
Regression Models
To test H1, we adopt the model of Francis and Schipper (1999) and Sami and
Zhou (2004) to measure the power of financial information to explain market values and
stock returns. We estimate the following regression models to examine the explanatory
power of assets and liabilities (Model 1A) and book value of equity and earnings (Model
2A) for market values, as follows:
Power of assets and liabilities to explain market value
MV_PSj,t = α0,t + α1,t ASSET_PSj,t + α2,t LIAB_PSj,t
(1A)

Power of book value of equity and earnings to explain market value
MV_LBVj,t = β0,t + β1,t BV_LBVj,t + β2,t EARN_LBVj,t
(2A)
where MV_PSj,t is the market value per share for company j at the end of year t, ASSET_
PSj,t and LIAB_PSj,t are the book value of assets and liabilities per share, respectively,
MV_LBVj,t is the market value of equity, BV_LBVj,t is the book value of equity, and
EARN_LBVj,t is earnings before extraordinary items. MV_LBVj,t, BV_LBVj,t, and
EARN_LBVj,t are deflated by the book value of equity at the end of year t−1 (i.e. the
lagged book value of equity). We expect a positive coefficient for ASSET_PSj,t and a
negative coefficient for LIAB_PSj,t, and positive coefficients for both BV_LBVj,t and
EARN_LBVj,t. Appendix B presents definitions of all variables.
46
UNLV Gaming Research & Review Journal t Volume 23 Issue 1
Value Relevance of Financial and Non-Financial Information
To examine the explanatory power of earnings and changes in earnings (Model
3A) and accruals and cash flows (Model 4A) for stock returns,3 we estimate the following
regression models:
Power of earnings and change in earnings to explain returns
RETj,t = ε0,t + ε1,t ΔEARN_LMVj,t + ε2,t EARN_LMVj,t
(3A)
Power of accruals and cash flow to explain returns
RETj,t = π0,t + π1,t ACCR_LMVj,t + π2,t CFO_LMVj,t
(4A)
where RETj,t is the cumulative market-adjusted return for company j for the 12 months
ending three months after the end of fiscal year, 4 ΔEARN_LMVj,t is the change
in earnings before extraordinary items, and EARN_LMVj,t is the earnings before
extraordinary items. Our accruals measure, ACCR_LMVj,t, is defined as (ΔCA – ΔCash)
– (ΔCL – ΔSTD – ΔTP) – DEP, following Aboody et al. (2002), where ΔCA is the
change in current assets, ΔCash is the change in cash and cash equivalents, ΔCL is the
change in current liabilities, ΔSTD is the change in debt included in current liabilities,
ΔTP is the change in tax payable, and DEP is the depreciation and amortization expense.
CFO_LMVj,t is defined as the cash flow from operating activities. LMV denotes variables
deflated by the market value of equity at the end of year t−1 (lagged market value of
equity). We expect positive coefficients on ΔEARN_LMVj,t, EARN_LMVj,t, ACCR_
LMVj,t, and CFO_LMVj,t.
The preceding models include only financial measures. To test H2, which
focuses on the incremental relevance of non-financial measures, we estimate variations of
the models adding non-financial variables specific to the casino gaming industry. First, to
estimate market value per share, we add two non-financial measures: LogNSLOTS, the
logarithmic (base 10) transformation of the number of slot machines, and LogNTABLES,
the logarithmic (base 10) transformation of the number of table games. We specifically
examine these two measures because casinos generate revenues directly from table games
and slot machines. We extend Models 1A and 2A as follows:
Power of assets, liabilities, slot machines, and table games to explain market value
MV_PSj,t = α0,t + α1,t ASSET_PSj,t + α2,t LIAB_PSj,t + α3,tLogNSLOTSj,t +
α4,tLogNTABLES j,t
(1B)
Power of book value, earnings, slot machines, and table games to explain market value
MV_LBVj,t = β0,t + β1,tBV_LBVj,t + β2,tEARN_LBVj,t + β3,tLogNSLOTSj,t +
β4,tLogNTABLES j,t
(2B)
We expect the coefficient signs of the financial variables in Models 1B and 2B
to be consistent with those in Models 1A and 2A. As noted earlier, for the non-financial
variables LogNTABLES and LogNSLOTS, we make no directional predictions, but
expect the coefficients to be significant.
Similarly, when examining stock returns, we also modify the models to include
non-financial variables as follows, and expect the coefficient of these variables to be
3 Market-adjusted returns refer to the raw stock market return of an individual firm, less the return of a
benchmark index over the same period. In this paper, the cumulative abnormal return is calculated as the sum of
a firm’s daily market-adjusted stock return over the 12-month period, ending three months after the fiscal yearend.
⁴ We calculate our return ending three months after the fiscal year end following the method of Lam et al.
(2013). This allows for the lag between the fiscal year end date and the release and filing of audited year-end
financial statements, which must be filed within 60-90 days of the fiscal year-end (SEC, 2016), depending on the
size of the issuer.
UNLV Gaming Research & Review Journal t Volume 23 Issue 1
47
significant:
Power of earnings, change in earnings, slot machines, and table games to explain returns
RETj,t = ε0,t + ε1,tΔEARN_LMVj,t + ε2,tEARN_LMVj,t + ε3,tLogNSLOTSj,t +
ε4,tLogNTABLESj,t
(3B)
Power of accruals, cash flow, slot machines, and table games to explain returns
RETj,t = π0,t + π1,tACCR_LMVj,t + π2,tCFO_LMVj,t + π3,tLogNSLOTSj,t +
π4,t LogNTABLESj,t
(4B)
Furthermore, to test for non-linear effects, we include the interaction term
(LogNSLOTS × LogNTABLES) of the number of table games and number of slot
machines in Models 1–4, to form Models 1C–4C. In Models 1D–4D, we replace the
log of the number of table games and the log of the number of slot machines with the
ratio between them, TABLE_SLOT_RATIO, calculated as the number of gaming tables
divided by the number of slot machines. In Models 1E–4E we add TABLE_SLOT_
RATIO_SQ, which is the squared term of TABLE_SLOT_RATIO, to examine the
potential nonlinear effect.
Empirical Results
Table 1 provides descriptive statistics for the financial, non-financial, and
gaming variables used in our regression models. Our sample casino gaming operators
have a mean market price of approximately $21 per share, ranging from 75 cents to
$141.49. The mean value of assets and liabilities per share are $30.83 and $22.21,
respectively, for a net positive asset value of approximately $8 per share. Among the
non-financial variables, the mean number of table games is 398 and the mean number of
slot machines is 8,864. Firms in our sample have a minimum of 4 gaming tables and 191
slot machines. The mean value of the ratio of table games to slot machines (TABLE_
SLOT_RATIO) is 4.921. That is, for every 100 slot machines, there are approximately
4.9 table games. Table 2 provides descriptive statistics of the number of slot machines
and table games by year from 1999 to 2017.
48
UNLV Gaming Research & Review Journal t Volume 23 Issue 1
Value Relevance of Financial and Non-Financial Information
Table 1
Descriptive Statistics
Variable
N
Mean
Median
Std. Dev.
Min
Max
Dependent variables
MV_PS
MV_LBV
RET
244
244
244
21.800
4.449
0.140
12.360
2.161
0.105
25.426
10.879
0.564
0.750
0.234
−2.019
141.490
109.945
2.694
Independent variables
ASSET_PS
LIAB_PS
BV_LBV
EARN_LBV
EARN_LMV
∆EARN_LMV
ACCRUAL_LMV
CFO_LMV
244
244
244
244
244
244
244
244
30.829
22.207
1.292
0.328
0.168
0.004
0.298
0.204
25.548
16.548
1.072
0.232
0.110
−0.005
0.171
0.151
26.937
22.427
2.012
0.849
0.264
0.295
0.518
0.226
0.980
0.259
−1.223
−2.311
−0.589
−1.922
−0.984
−0.296
136.630
110.966
21.285
6.082
2.114
2.107
3.496
1.183
Non−financial variables
NSLOTS
NTABLES
LogNSLOTS
LogNTABLES
TABLE_SLOT_RATIO
TABLE_SLOT_RATIO_SQ
244
244
244
244
244
244
8864.4
398.1
3.655
2.122
4.921
62.541
3500.0
92.5
3.544
1.966
2.552
6.511
9762.8
569.8
0.547
0.714
6.203
153.431
191.0
4.0
2.281
0.602
0.324
0.105
41136.0
2735.0
4.614
3.437
29.507
870.655
Table 2
Descriptive Statistics of NSLOTS and NTABLES by Year
Year
NSLOTS
NTABLES
Mean
Min
Max
Mean
Min
Max
1999
4,960
608
17,136
147
8
539
2000
7,325
858
28,546
232
15
1,135
2001
7,577
1050
28,779
220
11
1,120
2002
8,484
1099
25,928
234
11
1,017
2003
8,309
1103
26,439
229
11
1,101
2004
8,529
1186
24,809
234
7
1,050
2005
9,229
1158
39,175
308
7
1,638
2006
8,955
195
41,136
330
4
1,761
2007
8,187
195
36,686
367
4
1,956
2008
8,762
195
35,735
377
4
1,928
2009
8,921
195
34,966
376
4
1,985
2010
8,596
191
30,465
437
4
2,070
2011
8,747
191
29,673
449
4
2,080
2012
9,569
718
33,115
475
15
1,830
2013
10,391
558
37,083
585
10
2,380
2014
10,398
577
37,083
627
15
2,530
2015
10,260
718
37,083
630
30
2,542
2016
9,189
1070
31,006
610
25
2,735
2017
9,333
1133
24,460
904
52
1,699
UNLV Gaming Research & Review Journal t Volume 23 Issue 1
49
A table matrix of correlation coefficients between key regression variables is
provided in the Online Appendix, which is available upon request to the corresponding
author. The top and bottom diagonal provide Spearman and Pearson correlation
coefficients, respectively. In the discussion that follows, we refer to Pearson
correlations for continuous variables and Spearman correlations for discrete variables.
First, the correlations among the market valuation variables (MV_PS, MV_LBV,
ASSET_PS, LIAB_PS, BV_LBV, EARN_LBV, EARN_LMV, ACCRUAL_LMV,
and CFO_LMV) are generally highly significant and positive. Second, among the
variables in the returns models, the correlations between annual returns (RET) and
financial variables are often not significant. This suggests that our financial variables
may be more capable of explaining market valuation than of explaining stock returns.
Third, for non-financial variables, both LogNTABLES and LogNSLOTS are positively
correlated, and are significantly and positively correlated with MV_PS, ASSET_PS,
LIAB_PS, EARN_LBV, and ΔEARN_LMV. However, while TABLE_SLOT_RATIO
is significantly and positively correlated with MV_PS, it is negatively correlated with
EARN_LMV, ΔEARN_LMV and ACCRUAL_LMV. We note that the non-financial
variables, LogNTABLES, LogNSLOTS, and TABLE_SLOT_RATIO, are highly
correlated with each other (correlations from .52 to .81). All of the correlations are
reasonable and consistent with those in the literature. In the design of our regression
analysis, we gauge the effect of multicollinearity, which may lead to the wrong sign or
wide swings in parameter estimates (Greene, 1993). For this reason, we estimate the
effect of TABLE_SLOT_RATIO separately from LogNTABLES and LogNSLOTS.
Table 3
Value Relevance of Financial Information (MV_PS)
Variable
MV_PS
(1A)
MV_PS
(1B)
MV_PS
(1C)
MV_PS
(1D)
MV_PS
(1E)
Intercept
−0.8641
(−0.20)
1.2337***
(7.62)
55.3657***
(3.51)
1.3919***
(8.56)
33.2854
(1.38)
1.3937***
(8.58)
−8.6675**
(−2.13)
1.3375***
(8.59)
−3.4976
(−0.78)
1.3370***
(8.750)
−0.7451***
(−4.03)
−0.8792***
(−4.81)
−0.8607***
(−4.70)
−0.8934***
(−5.03)
−0.8943***
(−5.10)
LogNSLOTS
−22.4915***
(−3.64)
−14.9489*
(−1.71)
LogNTABLES
11.0083**
(2.49)
21.4507**
(2.20)
1.5278***
(5.10)
−0.3327
(−0.45)
ASSET_PS
LIAB_PS
LogNSLOTS × LogNTABLES
−3.4904
(−1.21)
TABLE_SLOT_ RATIO
TABLE_SLOT_ RATIO_SQ
0.0668***
(2.77)
Number of Cross Sections
20
20
20
20
20
Time Series Length
19
19
19
19
19
N
244
244
244
244
244
Firm Fixed Effects
Yes
Yes
Yes
Yes
Yes
Year Fixed Effects
Yes
Yes
Yes
Yes
Yes
R2
.3738
.4061
.4008
.4321
.4487
Note: The dependent variable is MV_PS. *** p < .01, ** p < .05, * p < .10. 50 UNLV Gaming Research & Review Journal t Volume 23 Issue 1 Value Relevance of Financial and Non-Financial Information Regression Results: Financial and Non-Financial Variables Table 3 reports the coefficients and t-statistics of regressions estimating the relationship between financial information and share prices. Results from Model 1A show a positive and significant relationship between ASSET_PS and MV_PS (coefficient = 1.2337, t = 7.62), and a negative and significant relationship between LIAB_PS and MV_PS (coefficient = −.7451, t = −4.03), consistent with relationships established in prior accounting research, which has found that assets are valued positively and liabilities negatively. Our results here indicate that in terms of assets and liabilities, the pricing behavior of investors for casino gaming firms is consistent with behavior observed for firms in other industries. In Model 1B, where we add LogNTABLES and LogNSLOTS to examine the incremental explanatory power of non-financial variables for market values, the coefficient to LogNSLOTS is negative and significant (coefficient = −22.4915; t = −3.64), while the coefficient to LogNTABLES is positive and significant (coefficient = 11.0083; t = 2.49). These results are interesting and surprising. The stock market seems to significantly positively value table games, but significantly negatively value slot machines. In this section, we document our empirical results; in later sections of this paper, we conjecture about the possible reasons for these valuations. In Model 1C, the coefficient of LogNSLOTS × LogNTABLES is not significant, suggesting that their effects are separate from each other. In Models 1D and 1E, the coefficients of TABLE_SLOT_RATIO (coefficient = 1.5278; t = 5.10) and TABLE_SLOT_RATIO_SQ (coefficient = .0668; t = 2.77) are both positive and highly significant, with results of Model 1E showing that the effect is non-linear. In Table 4, Model 2A, both earnings and book value of equity are positively priced in market values, with BV_LBV (coefficient = 4.3693; t = 26.22) and EARN_LBV (coefficient = 1.3318; t = 3.01) both positive and significant, consistent with results from prior literature. In Models 2B and 2C, the coefficients of LogNSLOTS and LogNTABLES and their interaction term are insignificant. In Model 2D, the coefficient of TABLE_SLOT_RATIO (coefficient = .1734; t = 2.00) is positive and significant. This suggests that rather than the individual numbers of slots and tables, it is the relative proportion of the two that is relevant for market values. When adding the square term, however, neither is significant. UNLV Gaming Research & Review Journal t Volume 23 Issue 1 51 Table 4 Value Relevance of Financial Information (MV_LBV) Variable MV_LBV (2A) MV_LBV (2B) MV_LBV (2C) MV_LBV (2D) MV_LBV (2E) Intercept −1.3586* (−1.69) 1.0130 (0.22) −2.7036 (−0.33) −2.2704*** (−2.64) −1.5628 (−1.45) BV_LBV 4.3693*** (26.22) 4.3684*** (25.63) 4.3591*** (25.35) 4.3659*** (26.16) 4.3889*** (26.16) EARN_LBV 1.3318*** (3.01) 1.2831*** (2.89) 1.2742*** (2.86) 1.2321*** (2.81) 1.2357*** (2.80) LogNSLOTS −1.5300 (−0.84) −0.3525 (−0.13) LogNTABLES 1.4976 (1.12) 3.3331 (0.92) 0.1734** (2.00) 0.1159 (0.44) LogNSLOTS × LogNTABLES −0.5541 (−0.55) TABLE_SLOT_ RATIO TABLE_SLOT_ RATIO_SQ 0. 0112 (1.17) Number of Cross Sections 20 20 20 20 20 Time Series Length 19 19 19 19 19 N 244 244 244 244 244 Firm Fixed Effects Yes Yes Yes Yes Yes Year Fixed Effects Yes Yes Yes Yes Yes R .7685 .7684 .7686 .7700 .7719 2 Note: The dependent variable is MV_LBV. *** p < .01, ** p