The real estate risk of hospitality firms: Examining stock-return sensitivity to property values

International Journal of Hospitality Management 31 (2012) 695–702

Contents lists available at SciVerse ScienceDirect

International Journal of Hospitality Management journal homepage: www.elsevier.com/locate/ijhosman

The real estate risk of hospitality firms: Examining stock-return sensitivity to property values Seul Ki Lee a,1 , SooCheong (Shawn) Jang b,∗ a b

School of Hospitality and Tourism Management, Purdue University, Marriott Hall, 700 State Street, West Lafayette, IN 47907, United States School of Hospitality and Tourism Management, Purdue University, Marriott Hall, 700 West State Street, West Lafayette, IN 47907, United States

a r t i c l e

i n f o

Keywords: Stock-return exposure Real estate risk Arbitrage pricing theory (APT) Two-factor model Risk premium

a b s t r a c t The value of a hospitality firm is often believed to be dependent on the market price of the properties they own. However, the core business of a hospitality firm is the production of products and services. Since the real estate assets are depreciated throughout their useful life, short-term covariance of firm value with real estate prices seems implausible. Using a two-factor model, the current study examined the real estate exposure of US hospitality firms through daily stock return data from 2005 to 2009. Results indicate that the majority (88%) of the hospitality firms were exposed to real estate risk at some point during the sample period, while the second-stage analysis of real estate betas suggests that exposure is conditional on the financial status of the hospitality firm. Implications and suggestions for future research are presented with the findings of the study. © 2011 Elsevier Ltd. All rights reserved.

1. Introduction Real estate is an essential asset for hospitality businesses. The extent to which a business uses its real estate is directly linked with the production capacity and demand accessibility of all hospitality firms. In order to expand sales, a hospitality firm must increase real estate inputs at a certain point. Further, to tap into remote demand hospitality firms must acquire or lease real estate at the geographic location of interest. Accordingly, many researchers have proposed that the value of hospitality firm is dependent on the value of their properties. For example, Gyourko and Keim (1993) posited that stock returns on vacation and restaurant businesses should be related to real estate returns since the companies own valuable properties. Parrino (1997) argued that Marriott’s unfavorable financial status in the 1990s was due to a decline in operating cash flows and a weak market in the properties it owned. Ling and Naranjo (1999) implicitly suggested that returns on hotels and motels are related to property appreciation returns. More recently, Newell and Seabrook (2006) asserted that hotels comprise both a business and a property risk. Formally, the supposition that a hospitality firm’s value is correlated with property prices can be interpreted as the firm’s exposure to real estate risk. The return series on real estate assets is

∗ Corresponding author. Tel.: +1 765 496 3610; fax: +1 765 494 0327. E-mail addresses: [email protected] (S.K. Lee), [email protected] (S. Jang). 1 Tel.: +1 765 337 6249. 0278-4319/$ – see front matter © 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.ijhm.2011.09.005

generally regarded as exhibiting random walk behavior (Kleiman et al., 2002). If the firm value is influenced by random changes in the price of assets it owns the firm would be perceived as susceptible to this specific uncertainty or risk, hence being exposed to the asset price of interest (Adler and Dumas, 1984). With real estate risk exposure, the stock returns of hospitality firms become a function of the real estate return factor, as well as other random return-generating factors. However, as intuitive as the above reasoning may seem a critical question remains. Hospitality firms’ real estate assets are primarily deployed to produce the products and services that constitute their core business and generate recurrent earnings. The acquisition and construction of properties are based on the premise that these book assets will be depreciated throughout their useful life in order to generate cash flow from operating activities (Dalbor and Upneja, 2004; Upneja and Dalbor, 1999) just as manufacturing firms utilize their property, plant, and equipment. Since hospitality firms have rather long depreciation schedules or lease agreements, temporary (i.e. daily) shifts in property prices do not affect a firm’s balance sheet (book value of assets) or income statement (user cost of capital). Thus, a theoretical gap is identified. If the core business of hospitality firms requires the use of real estate as factor inputs, temporary variations in the market price of properties should not affect firm value. Theoretically, a firm could sell its real estate assets if the realization of the sale was more desirable than the expected operating cash flows from utilizing the asset. However, this is not expected to significantly influence firm value, as valuation of the hotel is likely to be made based on the sum of future cash flows it

696

S.K. Lee, S. Jang / International Journal of Hospitality Management 31 (2012) 695–702

Table 1 Sample industry groupings. Industry Hotel firms Hotels Casinos and Casino hotels Restaurant firms Full service restaurants Limited-service restaurants

NAICS

No. of firms

Average real estate holdings ($ millions)

% Total assets

721110 721120/713210

8 12

1995 1613

63.15 65.17

722100 722211

26 11

1604 2268

65.64 63.45

provides through operations (Corgel and deRoos, 1993). Moreover, if hospitality firms do not carry real estate assets with the primary objective of selling them at a profit, firm value should not be systematically related to the market price of these assets. Thus, in an asset-pricing scheme, unexpected returns from property transactions, such as gains from salvage value, are best explained by the abnormal return (alpha) rather than exposure to systematic risk (beta). Meanwhile, the limited empirical evidence to date allows only an inconclusive perspective. Using national data from Singapore, Ong and Yong (2000) found that hotels and restaurants had the highest positive (increasing returns from property appreciation) real estate exposure among non-real estate industries. On the other hand, using U.S. data, Hsieh and Peterson (2000) reported that the lodging industry was not exposed to real estate risk, while the restaurant industry was negatively exposed (decreasing returns from property appreciation). Nevertheless, understanding this prominent risk is important. If the returns on hospitality firms are correlated with property prices, corporate financial managers should take the variations in property prices into consideration, as the firm values would be dependent on the random movements of the real estate market. Furthermore, since real estate risk is likely to be systematic (Tuzel, 2010), valuation by investors and portfolio managers of a firm’s capital assets would be dependent on real estate prices. Therefore, this study intends to fill in this theoretical and evidentiary gap by examining hospitality firms’ exposure to real estate risk and the potential determinants of exposure. Specifically, the objectives of this study were to (1) examine individual and time-variant exposure of hospitality firms to real estate risk at the firm-level and (2) test potential determinants of real estate exposure based on hypotheses developed from a review of the previous literature. Implications, limitations, and suggestions for future research are discussed along with the findings of the study. 2. Literature review and hypotheses 2.1. Real estate as a systematic risk Real estate accounts for a significant portion of corporate wealth. On average, real estate accounts for 25% of firms’ net worth and 19% of total corporate assets at historic cost (Laposa and Charlton, 2001). Further, 30% of all companies own their land and buildings rather than to lease them (Brounen and Eichholtz, 2005). Regardless of size, age, and sector, all firms use real estate in one way or another. Manufacturing firms deploy real estate to accommodate their inventory, production equipment, and finished goods, whereas hi-tech industries need real estate to house research facilities and staff. On balance sheets of any firm, real estate assets are listed under the long-term asset category. Similarly, a company’s income statement accounts for the real estate ‘costs’ in generating revenue through such items as depreciation, occupancy expenses, or rental expenses (Tuzel, 2010). Because real estate is a common asset category across all industries and businesses, many distinguish it as a source of systematic risk (He, 2002). If all firms own the same type of asset, it seems

convincing that exposure to this risk is non-diversifiable. Drawing on the efficient market hypothesis (Fama, 1970) and the arbitrage pricing theory (Ross, 1976; Roll and Ross, 1980) and assuming that real estate returns are random (Kleiman et al., 2002); this implies that a firm’s returns are generated by the real estate factor as well as a number of other non-diversifiable risk factors: ri,t = rf,t +



ˇi,k Fk,t + ˇRE RRE,t

(1)

where ri is the return on hospitality firm i stock, rf the risk-free rate of return, F1 ∼ Fk the k common return-generating factors, ˇ1,1 − ˇi,k the firm-specific exposures to the k risk factors, RRE the return from real estate, ˇRE the exposure to real estate risk, and t the time subscript. Consequently, a number of studies have tested the role of real estate risk in asset pricing, or equivalently, the significance of ˇRE . Testing with a single-factor model, Liu et al. (1990) first reported the existence of real estate risk premium in the market. Mei and Lee (1994) provided evidence for a significant real estate factor premium in addition to stock and bond factors. Hsieh and Peterson (2000) revealed the systematic relation between stock returns and REITs returns between 19 out of 53 industries in the U.S. He (2002) found that there is a sixth real estate factor, in addition to the three stock factors: market, size, book-to-market, and two bond factors: term structure and default risk, of Fama and French (1993). Kullmann (2001) and He (2002) reported that the explanatory power of multifactor asset-pricing models improves when the real estate factor is added. More recently, Tuzel (2010) found higher industry-adjusted returns on firms with more real estate holdings, after accounting for the Fama–French stock factors and the momentum factor (Jegadeesh and Titman, 1993; i.e. the inertia of stock performance proxied by a zero-investment portfolio constructed from a long position on stocks that have performed well in the past and from a short position on stocks that have performed poor in the past). Consequently, there seems to be a general consensus on the role of real estate risk as a return-generating factor. Further, given that the risk is common (Hsieh and Peterson, 2000), much of corporate wealth is concentrated in real estate (Kullmann, 2001), and essentially a macroeconomic risk is also theoretically appealing to categorize real estate risk as non-diversifiable (He, 2002). Yet there is still a missing link. Even though real estate risk is priced in the market, in terms of providing higher returns it is not risk but exposure that is relevant in the asset-pricing scheme. To date, studies have implicitly assumed that having more real estate will increase exposure to real estate risk (Tuzel, 2010; Hwa, 2006). This may seem intuitive, but it leaves room to consider additional determinants of real estate exposure. Evidence also supports this notion, as Hsieh and Peterson (2000) failed to find that the lodging industry was exposed to real estate risk. If real estate exposure was governed solely by the real estate holdings of a firm, hotel firms would be among the most exposed industries. However, at the industry level hotel firms did not show significant exposure, whereas the restaurant firms showed a surprising net negative exposure despite their sizable real estate holdings (see Table 1). In turn, we deduced three

S.K. Lee, S. Jang / International Journal of Hospitality Management 31 (2012) 695–702

possible explanations as follows: first, the studies outlined above use a set of other empirical ‘factors’ in their multifactor model approach as their primary concern was to find the significance of the real estate factor in excess of the well-established factors, which may be problematic. Because the real estate factor is seen as systematically related to size and book-to-market factors (Kullmann, 2001; Peterson and Hsieh, 1997), the estimated real estate betas may be residual betas. Second, aggregating the data at the industry level may ‘cancel out’ or blur the exposures, as the underlying logic of industry-level firm pooling is that firms in an industry behave similarly with respect to real estate risk (Bartov and Bodnar, 1994). This may lead to noise in the estimation or insignificant results. Third, estimating a single beta for respective firms during a period involves the implicit assumption that the firms will have, on average, a consistent exposure coefficient to real estate risk throughout the time span. However, risk exposure may be dependent on firm characteristics that are time-variant (He and Ng, 1998). In order to address these issues the current study used Jorion’s (1990) two-factor model, which is widely used to estimate exposure to macroeconomic risks, such as interest rate and exchange rates (Bartram, 2007; Williamson, 2001). The current study also estimates yearly real estate betas at the firm level in order to allow for variation in the coefficients among firms and across time. Meanwhile, if real estate risk is a unique, systematic risk, it should precede any other empirically based factor, such as size or book-to-market factors, in explaining stock returns. The real estate betas obtained from two-factor estimations were then used for a second-stage analysis by regressing them on a set of firm-specific time-variant variables identified through the literature. The procedure is explained in details in the following subsection. 2.2. Determinants of real estate exposure Using real estate through ownership is generally identified as a major source of real estate risk due to the adjustment costs associated with production shocks (Tuzel, 2010). For example, a lodging firm that owns hotels, in contrast to franchising or managing them, is considered riskier (Binkley, 2001) because the cost to adjust capacity is higher. It also seems intuitive that real estate risk is introduced when a firm lists the property on its balance sheet, since this property can be sold off at a price determined by the market. Consistent with Tuzel’s (2010) argument, we hypothesized that ownership of real estate will increase the exposure to real estate risk. Excluding ownership, there are two major options for using a real estate input, namely capitalized lease and operating lease. Among the two types of leases, capitalized leases closely resemble ownership by the lessee. SFAS No. 13, “Accounting for Leases,” requires the lessee to list both the leased asset and the pertinent debt amount on the asset and liability sections of the balance sheet when a lease is capitalized. Throughout its usage, the asset is depreciated while the liability is amortized. Capitalized leases often involve the transfer of the ownership, including bargain purchase options for the lessee at the end of the lease term, setting the lease term at least 75% or more of the residual value of the asset, or requiring disclosures on the balance sheet similar to acquisitions (Sharpe and Nguyen, 1995). In many ways the lessee has the benefits and risks associated with the capitalized lease, which are similar to owning real estate, while the lessor has a mere security interest. Upneja and Schmidgall (2001) also noted that ownership transfer and bargain purchase options are the major drivers of capitalizing leases for hotel financial managers, indicating a common motivation for ownership and capitalized leases. Therefore, Hypothesis 1 was formed as follows:

697

Hypothesis 1. The use of real estate through ownership or a capitalized lease increases the real estate exposure of a hospitality firm. On the other hand, operating leases are rather an off-financing tool for the lessee. On the lessee’s income statement, the real estate use is recorded as a rental expense, while the lessor depreciates the leased real estate. According to SFAS No. 13, the benefits and risks of owning real estate are not significantly transferred from lessor to the lessee through operating leases. Eisfeldt and Rampini (2009) argued that the operating lease is the only ‘true’ lease, as the lessor retains effective ownership of the asset. Also, an operating lease is more flexible and more easily reversed, as the lessee can either assume or reject the lease. Koh and Jang (2009) also implied that the market price of properties under an operating lease may not be important to the lessees, as lessors can repossess their assets when the lessee fails to make lease payments. Following this line of reasoning, we hypothesized that for the real estate assets used under operating lease contracts, the market price will be of importance to the value of the lessor and not the lessee. Hypothesis 2. The use of real estate through an operating lease does not affect the real estate exposure of a hospitality firm. Finally, we considered the motivation for hospitality firms’ real estate transactions. Since the real estate assets are intended to be a fixed input that has a rather long (i.e. 30 years) production life, it is intuitive to believe that the short-term changes in market prices should not matter. However, some potential drivers of real estate transactions, typically not considered during the acquisition stage, are noted throughout the literature. Parrino (1997) pointed out that hospitality firms’ properties can be sold off to cover debt and debt-related expenses. Financially constrained firms can also voluntarily engage in asset sales to improve the liquidity of the company and secure funds for reinvestment (Hovakimian and Titman, 2006). For example, a hospitality firm may have a significant amount of real estate assets but can have less or no exposure to real estate risk, if the cash flows from the operation are stable and sufficient liquidity is maintained to carry out debt-covering and investing activities; the need for disinvestment is eliminated. Toward the other extreme, when a hospitality firm is in severe financial constraint and becomes insolvent the firm’s liquidation value would be close to the market price of its assets, the majority of which is real estate. Simply put, the price of real estate can be important to firm value when the firm is selling it. In this case, the firm value may be aligned with movements in property prices. Investors are aware that the firm has an incentive to disinvest and the expected cash from the transaction is dependent on the market price. Yet, real estate assets can be used to support the firms’ core businesses when it has no immediate, urgent need for cash. Therefore, we hypothesized the following: Hypothesis 3. Financial constraint increases the real estate exposure of hospitality firms. 3. Data and methods 3.1. Sample and data Data from 2005 through 2009 was retrieved from the Compustat Database based on the North American Industry Classification System (NAICS) codes. After retrieving all firms under the NAICS codes for hotels (721110), casino hotels (721120), casinos (713210), full-service restaurants (722100), and limited-service restaurants (722211), two steps were taken to ensure the reliability of the results and the testability of the data. First, we excluded any firm identified as international as we used the Dow Jones REIT Composite Index, which is a proxy for returns in the US real estate market.

698

S.K. Lee, S. Jang / International Journal of Hospitality Management 31 (2012) 695–702

Table 2 Descriptive statistics of variables. Variable

Definition

Mean

St. dev.

Minimum

ˇRE

Real estate exposure estimated from model (2) Owned real estate gross buildings divided by total assets Capitalized real estate lease gross capitalized leases divided by total assets Operating lease expense rental expense divided by total assets Operating cash flows operating cash flows divided by total assets Long-term debt long-term debt divided by total assets Quick ratio Quick assets divided by current liabilities

0.296

0.248

−0.497

1.050

0.284

0.267

0.000

0.960

0.224

0.280

0.000

0.819

0.056

0.063

0.000

0.475

0.113

0.090

–0.213

0.350

0.326

0.421

0.000

3.675

0.938

1.507

0.047

21.858

OWN CLEASE OLEASE OCF LEV QR

The second step involved excluding firms whose daily stock price data did not appear on the CRSP tape. After these procedures, there were a total of 57 firms in the sample: 8 hotels, 7 casino hotels, 5 casinos, 26 full service restaurants, and 11 limited service restaurants. Accordingly, the firms’ daily stock return data and the value-weighted stock market index spanning throughout the accounting periods were obtained from the Center for Research on Stock Prices (CRSP) tape. As the firms have different fiscal years, the daily return data were matched to the firm’s fiscal year from the starting day to the ending day of the accounting periods. Real estate returns were obtained from the Dow Jones REIT Composite Index as the daily return on the index. The Dow Jones REIT Composite Index is constructed from the returns of companies that qualify as Equity REITs in the “Dow Jones stock universe.” This index covered 95% of the entire US REITs market until early 2005 (Dow Jones Indexes, 2011). Table 1 provides a description of the selected industries, the NAICS code, and the significance of the real estate held by firms in these industries. As observed, the real estate assets of hospitality firms and the percentage to total assets are fairly sizable and there is little variation across the industry groupings. This supports the exclusive and collective examination of the hospitality industries in terms of their real estate holdings and related risk. 3.2. Estimation of real estate betas The firms’ stock-returns exposure to real estate risk, or equivalently the real estate betas, were estimated using Jorion’s (1990) two-factor model, which can be written as a single-equation regression model of the following form: ri,t = ˛i,t + ˇMi,y RM,t + ˇREi,y RREt + εi,t

(2)

where ri,t is the ith firm’s stock return on day t, ˛ is the intercept, RM,t is the market return on day t, RREt is the daily return on the Dow Jones REITs Index, εt is random disturbance, and y the year subscript. Real estate betas were estimated yearly for each firm, requiring five estimations per firm. The model was estimated using ordinary least squares (OLS), consistent with recent approaches in exposure estimation (Sweeney and Warga, 1986; Bartram, 2007). By using the two-factor model, the estimated betas are exposure coefficients obtained after the market risk is ‘partialled out.’ Since the security market and real estate market are integrated to some extent (Liu et al., 1990) the twofactor model may underestimate the exposure to real estate risk. However, the current study preferred a conservative approach, as omitting the market factor could conversely overestimate the exposure.

Maximum

3.3. Variables and pooled regression for estimated betas Compustat divides real estate assets into four categories: buildings, capitalized leases, construction in progress, and land and improvements. Buildings are firm-owned properties and capitalized leases are lease contracts that are capitalized. Construction in progress is not yet utilizable or marketable by the firm, and land is excluded as it is not reproducible (Tuzel, 2010). Operating leases do not appear on the balance sheet, but are listed as an expense on income statements. The real estate betas obtained from model (2) were collected for a second-stage pooled regression on a set of firm-specific variables. To test our hypotheses we used the following functional form: ˇREi,y = 0 + 1 OWNi,y + 2 CLEASEi,y + 3 OLEASEi,y + 4 OCFi,y + 5 LEVi,y + 6 QRi,y + εi,y

(3)

where 0 is the intercept, OWN is ownership of property, and CLEASE is a capitalized lease on property. The variables OWN and CLEASE were scaled by the book value of total assets and jointly tested for Hypothesis 1. The ex ante expectation for the signs of both coefficients were positive (+), as they were expected to increase exposure to real estate risk. OLEASE is the operating lease expense scaled by the book value of total assets and is included to test for Hypothesis 2. The ex ante expectation for the parameter was 3 = 0, as an operating lease is a flexible, off-financing tool for the lessee that does not transfer the benefits (of real estate return) or risks (of real estate loss) from the lessor. OCF is operating cash flows scaled by total assets, LEV is longterm liability scaled by total assets, and QR is quick ratio, defined as quick assets divided by current liability. Significances of these three variables are used to jointly test Hypothesis 3. The ex ante expectations for the signs of the coefficients was negative (−) for OCF, as greater cash flows from operation would free the firm from the need to disinvest its assets and positive (+) for LEV, as debt-related expenses and risks, such as bankruptcy risk, increase with the size of debt (Baxter, 1967). The expectation on the sign of coefficient for QR was also negative (−), as short maturity exceeding quick assets indicates a liquidity risk and may lead to underinvestment problems (Johnson, 2003). It is also noted that the firms’ financial distress can be proxied by the corporate bond rating. However, a search of the database indicated that more than half of the sampled firms did not have data on bond ratings. Therefore, the three variables outlined above were used to collectively measure the financial distress. Subscripts i and y denote the firm and the year corresponding to observation. Detailed definitions and descriptive statistics of the variables are provided in Table 2.

S.K. Lee, S. Jang / International Journal of Hospitality Management 31 (2012) 695–702 Table 3 Result of real estate beta estimation.

699

Table 4 Result of second-stage pooled regression.

Risk factor

ˇM

ˇRE

˛

Variable

Coefficient

Standard error

t-Statistic

p-Value

VIF

Number of observations (57 × 5) Mean Standard deviation Maximum Minimum

285 0.854 0.528 2.440 −0.502

285 0.134 0.233 1.050 −0.571

285 −0.001 0.087 0.500 −0.633

OWN CLEASE OLEASE OCF LEV QR ␭0

0.189* 0.224** 0.646 −0.001*** 0.088** −0.046*** 0.342** *

0.099 0.109 0.609 0.000 0.042 0.011 0.070

1.92 2.05 1.06 −4.53 2.11 −4.33 4.91

0.058 0.043 0.291 0.000 0.037 0.000 0.000

1.655 2.331 1.995 1.457 1.087 1.260 –

F-statistic R2 Adj. R2

7.27*** 0.306 0.264

Coefficients statistically different from 0

ˇRE

ˇM

Number Statistically significant at 226 10% level 213 5% level 197 1% level

˛

Mean

Number

Mean

Number

Mean

1.025 1.056 0.951

106 88 64

0.296 0.332 0.390

19 8 2

−0.017 −0.129 −0.510

n = 106. * p < 0.1. ** p < 0.05. *** p < 0.01.

Firms with statistically significant RE Betas for at least 1 year Total Hotels Casinos and Casino hotels Full-service restaurants Limited-service restaurants

50/57 (88%) 6/ 8 (75%) 12/12 (100%) 22/26 (85%) 9/11 (82%)

Abbreviations: ˇM , market beta; ˇRE , real estate beta; ˛, alpha. Average R2 of two-factor regressions: 0.211.

4. Results and discussion 4.1. Real estate beta estimation The results of the beta estimation are summarized in Table 3 and the complete regression results appear in Appendices A–C. Among 285 estimations (5 betas for 57 firms), 106 (37%) showed significant beta coefficients at p < 0.1 level, supporting the sensitivity of firm value to real estate prices. On average, the market factor seemed to have more weight than the real estate factor on the stock returns of hospitality firms. However, 7 firms had greater exposure to real estate than to the market factor, illustrating the importance of real estate value for hospitality firms. Interestingly, 7 firms showed negative real estate exposure at some point: 4 casino/casino hotel firms and 3 restaurant firms, while 5 firms had a net negative real estate exposure on average (see Appendices A–C). Alpha was rarely observed, and when it was significant it was negative on average (abnormal loss). There seems to be no significant difference in likelihood of exposure across industry groupings. Another noteworthy aspect of these results is that although a significant percentage of firms (88%) had exposure to real estate risk at some point, there no firm was consistently exposed to real estate risk for all 5 years. This supports the idea that a firm’s real estate exposure may be conditional, and possibly a function of firm-specific characteristics that are time-variant. As a result, the second-stage analysis on the determinants of exposure was warranted. 4.2. Second-stage pooled regression Results of the pooled regression are shown in Table 4. In order to check the robustness of the results, heteroscedasticity and multicollinearity tests were performed. The Breusch–Pagan test did not reject the null of constant variance at even the p-level of 0.1 (2 : 1.18; p > 2 : 0.278) and no multicollinearity problem seems to exist, as shown by the variance inflation factor (no variable exceeding the VIF of 3) of the variables in Table 4. Consistent with our ex ante expectations, variables OWN and CLEASE were both significant, which indicates the real estate exposure of hospitality firms increases with the ownership of real estate

and capitalized leases. The results are consistent with the idea that using long-term assets through ownership or a capitalized lease will add to a firm’s real estate risk by incurring adjustment costs. Real estate assets will be exposed to transaction exposure when there is need for adjustment through demand and production shocks. As a result, Hypothesis 1 was supported. Operating leases, despite having a positive sign, were not significant even at the p < 0.1 level. The null of H2 : 3 = 0 was not rejected, which is consistent with the idea that real estate prices will affect a hospitality firm’s earnings when the assets are listed on their balance sheets and market expectations are formed based on the possible future sale of the assets. Consequently, Hypothesis 2 was supported. The variables OCF, LEV, and QR were highly significant. The direction of effects reveal an interpretive finding as the signs on these variables are negative, positive, and negative, respectively. A firm’s financial constraint will increase with leverage and decrease with cash flow and quick ratio. It should also be noted that the estimated betas are not the real estate risk itself, but rather the susceptibility of the firm to real estate risk. As such, property prices do not influence owner hospitality firms if they have good liquidity or a secure financial position. However, when hospitality firms face immediate liquidity problems, the value of the firm starts to align with the market price of real estate assets as these assets are expected to be sold to finance the firms’ debt-covering or investment activities. The results supported Hypothesis 3, and imply that the real estate exposure of hospitality firms is conditional on the liquidity, or equivalently financial distress, of the firm. 4.3. Additional robustness checks After the initial estimation several concerns arose, as it is likely that financially constrained firms are more likely to use leases than the less constrained counterparts (Koh and Jang, 2009). Thus, there may be unobserved heterogeneity between the firms that use some kind of lease and those that do not use any type of lease. Accordingly, two robustness checks were done to test the reliability of the model. First, a potential ‘selection’ effect was tested on the real estate exposure variable by using financial constraint proxies (OCF, LEV, and QR) as selection variables to predict leasing (1 = lease; 2 = do not lease; Lee and Jang, 2010). The Inverse-Mills ratio () of the two-step regression was insignificant at even p < 0.1 (p-value of 0.349), suggesting that there was no selection bias in the multiple regression models and no uncontrolled heterogeneity between lessees and non-lessees. Second, we also suspected that the ownership and financial constraint variables may interact as firms in financial distress choose to use various tools to convert real assets into cash, such

700

S.K. Lee, S. Jang / International Journal of Hospitality Management 31 (2012) 695–702

as sales–leaseback or sales–management back options (Whittaker, 2008). Therefore, we generated (3 × 2) = 6 interaction variables between the financial constraint variables and the real estate ownership variables to test for a potential interaction effect between the two. The results did not support interaction effects; the adjusted R2 of the interaction-augmented model only increased by a small amount, 0.01, while the joint test on all six interaction terms (null being at least one of the coefficients is not equal to zero) was not rejected at even p < 0.1 (p-value of 0.145). Third was the effect of recessionary period on the real estate exposure. Although the fiscal years do not necessarily coincide with the NBER-defined recessionary period, which is from December 2007 to June 2009, we tested the effect of the recessionary period by introducing a dummy variable RECDUMMY (unity if year 2008 and zero if not) in the model. Estimation showed that the result was qualitatively not different, while the F-statistic and adjusted R-squared slightly dropped and the RECDUMMY was statistically insignificant at p < 0.1 level. Therefore, we concluded in favor of the original model, which is more parsimonious and has a slightly better parameter-adjusted model fit. Lastly, we tested for the firm-specific and year-specific effects using the two-way fixed effects model with two-way cluster-robust standard errors (Gow et al., 2010; Peterson, 2009). Result of the two-way effects model in terms of R-squared and F-statistic suggested that incorporation of the year effects would lead to spurious estimation, as 17 firms (30% of the total sample) had fiscal years that do not conform to the calendar years. When the betweeneffects model (where only the firm-specific effects were accounted for) was employed (Himmelberg and Peterson, 1994), the signs on coefficients were consistent with the original model, but the model fit and efficiency of estimation was significantly impaired with the given limited sample size (Greene, 2008). Results of these procedures are appendicized at the end of manuscript. 5. Conclusion Real estate risk is a common, non-diversifiable, and macroeconomic risk that is priced in the capital market (Tuzel, 2010). According to Ross’s Arbitrage Pricing Theory (1976), such systematic risk should be a return-generating factor. Accordingly, ownership of the capital asset with a greater exposure requires a higher return. While the recent stream of studies has advocated the presence of real estate risk premium in the capital market, the actual evidence is limited and mixed. It offers little insight into the hospitality industry, despite being arguably the most real estateintensive industry among the various industries that constitute the market. In an attempt to address some of the limitations identified in prior studies and further investigate the real estate exposure of hospitality firms, the current study utilized daily, firm-level data and Jorion’s (1990) two-factor model to allow for firm-specific and time-varying real estate exposure. The real estate betas estimated from the two-factor models were collected for a second-stage analysis that matched the coefficients with potential determinants of real estate exposure. The results largely supported our ex ante expectations. Real estate exposure was pervasive among the sampled hospitality firms, highlighted by the large proportion of hospitality firms (88%) that were exposed to real estate risk at some point during the data period. However, no firm was consistently exposed to real estate risk throughout the 5-year horizon. This motivated a second-stage analysis on firm-specific, time-variant determinants of exposure. In the second stage analysis, the conditional nature of real estate exposure was revealed; property prices only affected the hospitality firms’ returns if they were subject to potential asset sales driven by financial or liquidity constraints.

The theoretical contribution of this study can be summarized two-fold. First, our results contradict the findings of the past study which reports that return of hospitality industries are not positively exposed to real estate risk. Second, we reveal the conditional nature of real estate risk exposure for hospitality firms. Utilizing daily data to estimate the real estate betas for respective hospitality firms, we found that the majority of stock returns were significantly and positively exposed to real estate returns. This result contradicts Hsieh and Peterson’s (2000) argument that hotels and restaurants may not be (or negatively) exposed to real estate risk. The current study also attempted to provide an answer to the conceptual question of whether a firm’s stock returns are correlated with real estate returns even though the real estate is a factor input with an extended useful life. In examining possible motivations for real estate transactions by hospitality firms, this study discovered that hospitality firms’ exposure to real estate risk is conditional, and largely related to financial and liquidity constraints. These results have some meaningful implications for the industry, especially for hospitality financial managers. Since hospitality firms that own more real estate risk are perceived as riskier, shareholders are likely to require higher returns. Thus, hospitality firms with a significant portion of their asset portfolio in real estate will be required to provide higher returns than their non-owning counterparts. Equivalently, the capital assets of these firms will be less valuable even if their cash flows are comparable. In this regard, utilization of the sales–leaseback option, off-financing tools, management contracts, or franchise agreements may help increase firm value (Slovin et al., 1990; Combs et al., 2004), as it would reduce the exposure to real estate risk that would inevitably follow property ownership. Also of importance is the conditional nature of real estate exposure. If a firm’s liquidity is maintained and there is no immediate need to liquefy the book assets, the firm’s exposure to real estate price decreases; the firm is viewed as less risky, and hence the investors may accept lower returns. In this regard, the joint role of corporate asset and financial managers are important in managing the real estate exposure of hospitality firms by maintaining liquidity and reserve cash. Furthermore, investors and portfolio managers should also manage real estate exposure when constructing their portfolios (Hsieh and Peterson, 2000), either through hedging or diversification. Hedging seems to be of limited value as the means to short-sell real estate assets seems obscure, even though some related approach may be feasible such as using the real estate investment trust securities (REITs) or the interest rate derivative securities to hedge against real estate risk (Hartzell et al., 1987). The firm-level beta results show that although scarce, there may be some gaming and restaurant firms exhibiting a negative covariance with the real estate factor, which allows for a complete hedging of portfolio real estate risk. As real estate risk is likely to be common across industries and firms, these capital assets (REITs, interest rate derivative securities, and all other securities of negative historical covariance with the real estate market) will be of great investment value in mitigating real estate the risk and constructing a mean–variance efficient portfolio. Despite the significance of this study, it is not without limitations. The sample size is limited, and the number of firms traded on a daily basis is even smaller. The time span (one 5-year period) is relatively short as the hospitality firms are frequently listed and delisted in the stock market. For example, the 5-year period starting from January 2000 lacks data on 17 of the firms sampled for current study. Firms reported in the Compustat cannot effectively represent an industry that consists of many small-sized, independent operations. Although numerous studies have found theoretical justification for and evidence of real estate risk, the efforts to empirically estimate real estate exposure cannot

S.K. Lee, S. Jang / International Journal of Hospitality Management 31 (2012) 695–702

completely avoid the complicated effects of greater macroeconomic movements. All capital and real estate markets will be integrated to some extent (Ling and Naranjo, 1999), and suspicion persists that there are unobserved communality and return-generating factors that continue to be difficult to isolate out. In the future, further investigation into real estate risk using different multifactor models, a broad examination of related and other real estate asset-intensive industries, and refined real estate sales price data will all contribute to improvement of understanding on this area. A joint study with exposure to other factor input prices, such as wage and raw materials, would yield meaningful implications for corporate managers, while identification of sectors

701

or firms that allow hedging or diversification of real estate risk will be of great value to investors and portfolio managers. Due to its pivotal role in economic theory and asset pricing, quantifying, diversifying, and hedging real estate risk are important tasks for managers, investors, and researchers alike. Since real estate risk is priced, ignoring the risk implies that the firm or the investor may not be pricing the risk correctly, and hence pursuing a suboptimal strategy for maximizing firm or portfolio value. Although real estate risk is common and pervasive throughout all sectors of the economy, the real-estate intensive nature of the hospitality industry demands further research efforts on this topic. Appendix A. Complete regression results

ID

Company name

Ticker

NAICS

sig. coefs

Mean ˇM

sig. coefs

Mean ˇRE

sig. coefs

Mean ˛

Mean R2

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57

ARK Restaurants Ameristar Casinos Flanigans Enterprises BJ’S Restaurants Benihana Bob Evans Farms Buffalo Wild Wings BOYD Gaming Corp. Cheesecake Factory Mexican Restaurants Cracker Barrel CEC Entertainment O’Charley’s Century Casinos COSI Inc. California Pizza Kitchen Famous Daves Domino’s Pizza Darden Restaurants Brinker Intl ELBIT Imaging Frisch’s Rest Granite City Gaylord Ent Good Times Rest Starwood Intl Game Tech ISLE of Capri Casinos J. Alexander’s Corp. Jack in the Box Krispy Kreme Landrys Lubys Las Vegas Sands Marriott Int Mcdonald’s Corp. Monarch Casino MGM Resorts INT Maui Land and Pineapple Mccormick and Schmicks Nathan’s Famous Empire Resorts Orient-Express P F Changs Pinnacle Ent Panera Bread Papa Johns Int Red Robin Starbucks Sonesta Intl Sonic Corp. Star Buffet Texas Roadhouse Nevada Gold and Casino Great Wolf Resorts Wynn Resorts Ltd. Yum Brands

ARKR ASCA BDL BJRI BNHNA BOBE BWLD BYD CAKE CASA CBRL CEC CHUX CNTY COSI CPKI DAVE DPZ DRI EAT EMITF FRS GCFB GET GTIM HOT IGT ISLE JAX JBX KKD LNY LUB LVS MAR MCD MCRI MGM MLP MSSR NATH NYNY OEH PFCB PNK PNRA PZZA RRGB SBUX SNSTA SONC STRZ TXRH UWN WOLF WYNN YUM

722100 713210 722100 722100 722100 722100 722100 713210 722100 722100 722100 722100 722100 713210 722100 722100 722100 722100 722100 722100 721110 722100 722100 721110 722211 721110 721120 721120 722100 722211 722211 722100 722211 713210 721110 722211 713210 713210 721110 722100 722211 721120 721110 722100 721120 722211 722211 722100 722211 721110 722211 722100 722100 721120 721110 713210 722211

0 5 1 5 2 5 5 5 4 2 5 5 5 3 3 3 3 5 5 4 4 3 1 5 0 5 5 5 2 5 5 5 5 5 5 5 5 5 5 4 1 3 5 5 5 5 5 5 5 0 5 1 4 3 5 5 5

– 1.28 0.59 1.00 0.99 0.94 0.87 1.40 0.91 0.80 0.89 0.77 1.33 0.83 1.49 1.11 0.81 0.73 0.82 0.93 1.34 0.30 0.69 1.23 – 1.06 1.06 1.28 0.59 1.05 1.18 1.20 1.06 1.56 0.85 0.73 1.36 1.10 0.81 1.10 0.55 1.63 1.00 0.80 1.45 0.82 0.71 0.96 0.99 – 0.70 0.49 0.82 1.03 1.37 1.36 0.80

1 2 3 2 1 2 1 1 4 0 2 3 1 1 1 3 1 3 3 2 0 0 0 3 0 5 3 1 1 2 1 2 2 1 4 0 3 2 0 3 1 1 3 3 3 2 4 2 2 1 3 0 3 2 1 2 3

0.17 0.39 −0.14 0.42 0.23 0.41 0.45 0.32 0.31 – 0.43 0.32 0.25 −0.50 0.81 0.48 0.27 0.30 0.26 0.45 – – – 0.44 – 0.35 0.07 0.35 −0.22 0.38 0.83 0.31 0.34 0.48 0.30 – 0.13 0.45 – 0.41 −0.30 1.05 0.36 0.31 0.30 0.24 0.26 0.33 0.24 0.29 0.30 – 0.47 −0.35 0.33 0.43 0.14

1 0 1 0 1 2 0 0 1 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 0 0 2 1 0 1 1 0 0 0 0 2 0 0 0 0 0 0 0 0 0 1 1 0 1 0 0 0 0 1 0 0 0

0.21 – 0.20 – 0.30 0.07 – – −0.23 – – – −0.63 – – – – – – – 0.50 – – – – – 0.00 −0.41 – 0.19 −0.54 – – – – 0.15 – – – – – – – – – 0.36 0.20 – −0.16 – – – – −0.62 – – –

0.01 0.31 0.02 0.27 0.10 0.34 0.16 0.33 0.27 0.03 0.31 0.26 0.21 0.04 0.09 0.27 0.07 0.25 0.27 0.22 0.18 0.02 0.01 0.47 0.01 0.50 0.33 0.25 0.01 0.34 0.15 0.27 0.22 0.24 0.47 0.31 0.25 0.23 0.16 0.17 0.01 0.07 0.28 0.25 0.32 0.20 0.29 0.25 0.35 0.02 0.27 0.01 0.22 0.04 0.28 0.33 0.36

702

S.K. Lee, S. Jang / International Journal of Hospitality Management 31 (2012) 695–702

Appendix B. Result of two-way fixed effects model with two-way cluster-robust errors Variable

Coefficient

Standard error

t-Statistic

p-Value

OWN CLEASE OLEASE OCF LEV QR 0

0.579 −0.386 −1.533 −3.308*** −0.210 −0.042*** 0.822***

0.390 0.784 2.326 1.050 0.147 0.015 0.216

1.480 −0.490 −0.660 −3.150 −1.430 −2.790 3.810

0.144 0.625 0.513 0.003 0.159 0.008 0.000

F-statistic R2

3.92** 0.034

n = 106. ** p < 0.05. *** p < 0.01.

Appendix C. Result of between-effects model Variable

Coefficient

Standard Error

t-statistic

p-value

OWN CLEASE OLEASE OCF LEV QR ␭0

0.075 0.202 0.448 −1.474*** 0.117* −0.076*** 0.416***

0.162 0.166 0.704 0.418 0.069 0.020 0.119

0.460 1.220 0.640 −3.530 1.710 −3.850 3.500

0.646 0.228 0.528 0.001 0.095 0.000 0.001

F-statistic R2

6.01*** 0.276

n = 106. * p < 0.1. *** p < 0.01.

References Adler, M., Dumas, B., 1984. Exposure to currency risk: definition and measurement. Financial Management 13, 41–50. Bartov, E., Bodnar, G.M., 1994. Firm valuation, earnings expectations, and the exchange-rate exposure effect. Journal of Finance 49, 1755–1785. Bartram, S.M., 2007. Corporate cash flow and stock price exposures to foreign exchange rate risk. Journal of Corporate Finance 13, 981–994. Baxter, N.D., 1967. Leverage, risk of ruin and the cost of capital. Journal of Finance 22, 395–403. Binkley, C., 2001. Checking out hotel stocks: it may not be check-in time. The Wall Street Journal 26 (March), C1. Brounen, D., Eichholtz, P.M., 2005. Corporate real estate ownership implications: international performance evidence. Journal of Real Estate Finance and Economics 30, 429–445. Combs, J.G., Ketchen, D.J., Hoover, V.L., 2004. A strategic groups approach to the franchising–performance relationship. Journal of Business Venturing 19, 877–897. Corgel, J.B., deRoos, J.A., 1993. The ADR rule-of-thumb as predictor of lodging property values. International Journal of Hospitality Management 12, 353–365. Dalbor, M.C., Upneja, A., 2004. The investment opportunity set and the long-term debt decision of U.S. lodging firms. Journal of Hospitality and Tourism Research 28, 346–355. Dow Jones Indexes, 2011. Dow Jones equity all REIT index fact sheet. Retrieved January 31, 2011 from http://www.djindexes.com/mdsidx/downloads/fact info/Dow Jones Equity All REIT Index Fact Sheet.pdf. Eisfeldt, A.L., Rampini, A.A., 2009. Leasing, ability to repossess, and debt capacity. Review of Financial Studies 22, 1621–1657. Fama, E.F., 1970. Efficient capital markets: a review of theory and empirical work. Journal of Finance 25, 383–417. Fama, E.F., French, K.R., 1993. Common risk factors in the returns on stocks and bonds. Journal of Finance 33, 3–56. Gow, I., Ormazabal, G., Taylor, D., 2010. Correcting for cross-sectional and time-series dependence in accounting research. Accounting Review 85 (2), 483–512.

Greene, W.H., 2008. Econometric Analysis, 6th ed. Prentice Hall. Gyourko, J., Keim, D.B., 1993. Risk and return in real estate: evidence from a real estate stock index. Financial Analysts Journal 49, 39–46. Hartzell, D., Hekman, J.S., Miles, M.E., 1987. Real estate returns and inflation. Real Estate Economics 15, 617–637. He, J., Ng, L.K., 1998. The foreign exchange exposure of Japanese multinational corporations. Journal of Finance 53, 733–753. He, L.T., 2002. Excess returns of industrial stocks and the real estate factor. Southern Economic Journal 68, 632–645. Himmelberg, C.P., Peterson, B.C., 1994. R&D and internal finance: a panel study of small firms in high-tech industries. Review of Economics and Statistics 76 (1), 38–51. Hovakimian, G., Titman, S., 2006. Corporate investment with financial constraints: sensitivity of investment to funds from voluntary asset sales. Journal of Money, Credit and Banking 38, 357–374. Hsieh, C., Peterson, J.D., 2000. Book assets, real estate, and returns on common stock. Journal of Real Estate Finance and Economics 21, 221–233. Hwa, T.K., 2006. Corporate real estate holdings and impact on firm returns. In: Paper presented at the 12th Pacific RIM Real Estate Society Annual Conference, Auckland, New Zealand. Jegadeesh, N., Titman, S., 1993. Returns to buying winners and selling losers: implications for stock market efficiency. Journal of Finance 48, 65–91. Johnson, S.A., 2003. Debt maturity and the effects of growth opportunities and liquidity risk on leverage. Review of Financial Studies 16, 209–236. Jorion, P., 1990. The exchange-rate exposure of US multinationals. Journal of Business 63, 331–345. Kleiman, R.T., Payne, J.E., Sahu, A.P., 2002. Random walks and market efficiency: evidence from international real estate markets. Journal of Real Estate Research 24, 279–298. Koh, J.H., Jang, S., 2009. Determinants of using operating lease in the hotel industry. International Journal of Hospitality Management 28, 638–640. Kullmann, C., 2001. Real estate and its role in household portfolio choice. Working paper, University of British Columbia. Laposa, S., Charlton, M., 2001. European versus US corporation: a comparison of property holdings. Journal of Corporate Real Estate 4, 34–47. Lee, S.K., Jang, S., 2010. Internationalization and exposure to foreign currency risk: an examination of lodging firms. International Journal of Hospitality Management 29, 701–710. Ling, D.C., Naranjo, A., 1999. The integration of commercial real estate markets and stock markets. Real Estate Economics 27, 483–515. Liu, C.H., Hartzell, D.J., Greig, W., Grissom, T.V., 1990. The integration of the real estate market and the stock market: some preliminary evidence. Journal of Real Estate Finance and Economics 3, 261–280. Mei, J., Lee, A., 1994. Is there a real estate factor premium? Journal of Real Estate Finance and Economics 9, 113–126. Newell, G., Seabrook, R., 2006. Factors influencing hotel investment decision making. Journal of Property Investment and Finance 24, 279–294. Sharpe, S.A., Nguyen, H.H., 1995. Capital market imperfections and the incentive to lease. Journal of Financial Economics 39, 271–294. Ong, S.-E., Yong, Y.Y., 2000. Real estate exposure and asset intensity. Journal of Real Estate Portfolio Management 6, 27–35. Parrino, R., 1997. Spinoffs and wealth transfers: the Marriott case. Journal of Financial Economics 43, 241–274. Peterson, J.D., Hsieh, C., 1997. Do common risk factors in the returns on stocks and bonds explain returns on REITs? Real Estate Economics 25, 321–345. Peterson, M.A., 2009. Estimating standard errors in finance panel data sets: comparing approaches. Review of Financial Studies 22 (1), 435–480. Roll, R., Ross, S.A., 1980. An empirical investigation of the arbitrage pricing theory. Journal of Finance 35, 1073–1103. Ross, S.A., 1976. The arbitrage theory of capital asset pricing. Journal of Economic Theory 13, 341–360. Slovin, M.B., Sushka, M.E., Poloncheck, J.A., 1990. Corporate sale–leasebacks and shareholder wealth. Journal of Finance 45, 289–299. Sweeney, R.J., Warga, A.D., 1986. The pricing of interest-rate risk: evidence from the stock market. Journal of Finance 41, 393–410. Tuzel, S., 2010. Corporate real estate holdings and the cross-section of stock returns. Review of Financial Studies 23, 2269–2302. Upneja, A., Dalbor, M.C., 1999. An examination of leasing policy, tax rates, and financial stability in the restaurant industry. Journal of Hospitality and Tourism Research 23, 85–99. Upneja, A., Schmidgall, R., 2001. Equipment leasing in the U.S. lodging industry. Cornell Hotel and Restaurant Administration Quarterly 42, 56–61. Whittaker, C., 2008. Hotel operator motives in UK sale and leaseback/managementback transactions. International Journal of Hospitality Management 27, 641–648. Williamson, R., 2001. Exchange rate exposure and competition: evidence from the automotive industry. Journal of Financial Economics 59, 441–475.