International Journal of Hospitality Management 38 (2014) 3038

Contents lists available at ScienceDirect

International Journal of Hospitality Management

journal homepage: www.elsevier.com/locate/ijhosman

Compensation practices in the lodging industry: Does top management pay affect corporate performance?
Arun Upneja a , Ozgur Ozdemir b,
a b

School of Hospitality Administration, Boston University, 928 Commonwealth Avenue, Boston, MA 02215, USA School of Applied Sciences, Department of Hotel Management, Ozyegin University, Cekmekoy, Istanbul 34794, Turkey

a r t i c l e
Keywords: Firm performance Compensation CEO CFO Agency theory

i n f o

a b s t r a c t
The current study examines the relationship between executive compensation and rm performance in the U.S. lodging industry. It is not clear-cut whether performance leads to compensation or compensation drives rm performance. Our contention is that cash and lagged equity-based compensation drive the rm performance. Our ndings suggest that chief executive ofcers (CEO) contemporaneous cashcompensation and one-year lagged equity-compensation positively affect the accounting performance measures return on assets and Tobins Q; but neither compensation components affects the marketperformance measure, stock returns, in the lodging industry. Quantitatively similar ndings are found for the chief nancial ofcer (CFO). Further robustness test show that further lags of equity compensation of both named executives do not result in increased stock performance in the lodging industry. 2013 Published by Elsevier Ltd.

1. Introduction Executive compensation is paid by rms to ensure adequate economic return to all stockholders of the rm. For example, while shareholders might be interested in an increase in rm value (short or long term), debt holders are interested in return of their capital with interest. Therefore, rms expect that managers will be good stewards of the rm. However, the root of the problem is that the interests of the managers might not coincide with the interests of the owners of the rm. This is the classic agency theory problem. Managers act in their own self-interest, while appearing to act in the interest of the various stakeholders of the rm. Concerted efforts have been made to structure compensation contracts to align the interests of the owners with those of the managers. This is the basic crux of the pay-for-performance literature. However, there is yet no consensus in the literature about the efcacy of this relationship (Jensen and Murphy, 1990; Duru and Iyengar, 1999). We are not yet sure if the evidence points to a pay-forperformance or perform-and-get-paid situation. In other words, it is not yet clear if executives are paid in advance with exhortation to perform or are they paid after they have performed. In many situations, the compensation contracts clearly state the benchmarks to be achieved before earning bonuses. Frequently the contracts are modied ex-post if the ex-ante benchmarks are not met.

Corresponding author. Tel.: +90 216 564 9476. E-mail addresses: aupneja@gmail.com (A. Upneja), ozgur.ozdemir@ozyegin.edu.tr (O. Ozdemir). 0278-4319/$ see front matter 2013 Published by Elsevier Ltd. http://dx.doi.org/10.1016/j.ijhm.2013.12.007

Agency theory argues that aligning executives personal interests with those of shareholders is crucial in alleviating the conicting interests of these two parties and increasing the likelihood of higher rm performance (Crumley, 2008). Compensating executives with company stock is an effective tool to align their interests with owners interests. Because executives total compensation is linked to stock performance of the company, stock-based incentive compensation plans permit a fair risk sharing between executives and owners (Veliyath and Bishop, 1995). Previous research in the compensation literature provides evidences from both perspectives. Some studies investigate the effect of executive compensation on rm performance (Sigler and Haley, 1995; Loderer and Martin, 1997; Mehran, 1995), while others examine this relationship in the opposite direction with the presumption that nancial performance is a determinant of executive compensation (Core et al., 1999; Attaway, 2000; Crumley, 2008). In this study, we extend the existing pay-for-performance literature along many dimensions. Although most studies investigate this issue using all available publicly traded rms, there are plenty of recent studies that look at specic industries to uncover the payperformance relationship. We focus our analysis on the U.S. lodging industry. While there are few compensation studies conducted in the restaurant industry, to the best of our knowledge, no previous study has looked at the pay-performance relationship in the U.S. lodging industry. One of the reasons hospitality researchers avoid investigating this phenomenon in the lodging industry is the limited sample size of the publicly traded U.S. lodging companies. Yet, given the different characteristics between lodging and restaurant industries, it may be misleading to interpret the ndings of the previous restaurant studies in the more complex lodging

A. Upneja, O. Ozdemir / International Journal of Hospitality Management 38 (2014) 3038

31

industry. Many concerns make it difcult to extend the results of the restaurant industry-specic studies to lodging rms. First, lodging rms are more complex and larger than the restaurant rms, so they are structurally different from restaurant rms. Second, lodging chains operate in a more global business environment and therefore are exposed to a more diverse customer base and varying customer demands. Third, lodging chains have a long history of corporatization versus the restaurant chains, which entails the expectation that they have stronger corporate practices (including executive compensation contracting) than the more recently appearing restaurant chains. These distinct characteristics of the lodging rms separate them from the restaurant rms and necessitate different sets of tests to understand the pay-performance phenomenon within the dynamics of the lodging industry. In terms of the differences from the larger manufacturing rms, lodging rms are usually smaller in size (Paryani et al., 2010) and are concerned more with service qualities rather than with manufacturing qualities. Also, the high turnover rate in the lodging industry, in contrast to that of the manufacturing industry, generates a negative image in the labor market and is likely to be a deterrent factor for top executives to work in this unstable working environment. Additionally, lower compensation levels in the lodging industry lead to a potential disadvantage for lodging rms to acquire the most-talented executives in the labor market, which then is likely to result in less-than optimal rm-executive matches and likely poor rm performance. Due to these differences, lodging rms largely require different characteristics and skill sets for the executive positions. Hence, we assume that it is very likely that there exist different dynamics underlying compensation practices in the lodging industry. While earlier research in the compensation literature focused on the chief executive ofcers (CEO) compensation, most recent studies have started to include other senior executives compensation packages (Jiang et al., 2010). Chief nancial ofcer (CFO) is one of these top senior executives that has strong control and responsibility for the nancial stability of a rm. Especially after the passage of the Sarbanes-Oxley Act of 2002, CFOs have been burdened with unprecedented legal responsibilities and have shared the risk of any corporate failure with the CEOs. Thus, CFOs can be regarded as the second most powerful executive in the corporate structure of an organization and take on a key role in the nancial performance and success of any rm. Moreover, the nancial decisions that the company makes involve both the involvement and approval of CFOs. In light of this argument, we propose that the increased responsibility on CFOs must be compensated with favorable pay-packages for increased rm performance. As far as the lodging industry is concerned, there has been no study that specically investigates the CFO compensation and its relation to rm performance. Hence, the current study attempts to examine the CFO compensation with respect to rm performance in the U.S. lodging industry in addition to providing insights to the CEO pay-rm performance relationship. We expect to see a positive association between a rms performance and CFOs compensation. Finding similar results as the CEO would suggest robustness for the impact of executive compensation on rm performance while dissimilar results suggest hierarchical differences in the pay level and rm performance, which may induce further research to scrutinize the discrepancy between CEO and CFO (and other senior executives) as it relates to rm performance. Previous studies in the hospitality research used either only salary or salary plus bonus (cash compensation) compensation packages to investigate the so-called pay-performance relationship (Kim and Gu, 2005; Gu and Choi, 2004). The main reason for excluding stock-based compensation is primarily the lack of data. FAS 123R was issued in 2004 (became effective in 2006) and required rms to disclose individual components of

executive compensation and expense stock options as of grant date. From the effective date of the FAS 123R, publicly traded U.S. rms have begun to disclose all relevant cash compensation (salary and cash bonuses) and equity compensation including restricted stocks, stock options, stock awards, and long-term pension contributions. Making use of these recently available pay components, the current study examines both the relationship between cash-compensation rm performance and stock-based compensation and rm performance using both accounting and market performance measures. We posit that accounting performance of a lodging rm is related to the cash component of the named executives compensation, whereas market performance is related more to the equity component of the executives compensation.1 2. Literature review 2.1. Agency theory The focus of the studies on the pay-performance relationship has shifted to direct linkages between executive compensation and rm performance (Canarella and Nourayi, 2008). Therefore, many researchers have utilized an agency theory approach to examine this relationship. An agency problem in an organization occurs when the executives of the company, who run the company on behalf of the owners, pursue goals and objectives that are not consistent with those of the organization (Welbourne and Cyr, 1999). This in turn leads to conicts with the interests of the stockholders who own the organization (Attaway, 2000). Boyd (1994) claims that agency problems mostly occur when the agents of the organization have no interest in the nancial outcomes of the decisions made. In addition to this, Welbourne and Cyr (1999) add that agency problems arise because risk preferences of agents are different from those of principals, which leads to decisions that are less than optimal. They further continue that this fact leads to two assumptions. First, agents are risk averse and avoid taking risks because they follow personal goals; and second, owners are risk neutral and they pursue organizational goals. Therefore, it could be argued that though the agents and owners are partners collaborating for the same purposes, they act differently when it comes to risk taking (Huang et al., 2004). The varying risk taking propensities further extend the gap between executives and owners interests. In order to lessen this gap among the two parties interests, agents incentives must be aligned with those of shareholders through compensation arrangements that reward agents on the basis of rm performance (Canarella and Nourayi, 2008; Murphy, 1985; Glassman and Rhoades, 1980; Welbourne and Cyr, 1999). In order to mitigate harmful effects of agency theory, stock-based incentives have been offered as an alternative pay method to top-level executives. Some companies even forced their executives to own company shares (Kay, 1999). One study showed that large companies, for example, J.C Penney, require that their CEOs maintain a large amount of company stock, as large as seven times their base salaries (Kay, 1999). Likewise, Tosi and Gomez-Mejia (1994) and Crumley (2008) stated that one-way of incentive alignment is stock ownership, which serves to create a situation in which the goals of the executives are similar to those of owners. These practices are in line with the propositions of outcome-based contractual agreements as suggested by the proponents of positivism stream of agency theory (Eisenhardt, 1989). This proposition argues that outcome-based contracts limit the agent opportunism by co-aligning the interests of the agents and the principals on the same grounds because the rewards of both

1 Equity-compensation is used interchangeably with stock-compensation throughout the paper.

32

A. Upneja, O. Ozdemir / International Journal of Hospitality Management 38 (2014) 3038

parties depend on the similar action preferences of both parties (Eisenhardt, 1989). This proposition was previously evidenced in Jensen and Meckling (1976), where they demonstrated the role of outcome-based contracts, increasing managerial ownership, in reducing the managerial opportunism in the compensation contracting practices. Risk averseness of agents is one critical factor that moderates the relationship between agents and principals in the context of contracting practices. If the agents characterize a signicantly risk neutral position, it becomes increasingly attractive to pass on risk to the agents and tie their reward to outcomes (Eisenhardt, 1989). 2.2. Previous research on pay-performance relationship The relationship between executive compensation and rm performance is a controversial one (Jensen and Murphy, 1990; Duru and Iyengar, 1999) and the relationship could be examined either way. While some researchers investigated the effect of executive compensation on rm performance (Sigler and Haley, 1995; Loderer and Martin, 1997; Mehran, 1995), others studied the relationship in the reverse direction (Core et al., 1998; Attaway, 2000; Crumley, 2008). In what follows, we describe some of the previous studies investigating pay-performance relationship. Attaway (2000) tested the pay-performance relationship in the computer and electronics industry using a sample of 42 rms. Running a multivariate regression analysis, he demonstrated that ROE positively affects CEO cash compensation at the conventional statistical signicance levels. Crumley (2008) conducted a similar study in the U.S. commercial banking industry and reported a positive relationship between CEO compensation and ROE. Gu and Choi (2004) provided further support for the positive relationship between CEO pay and ROA by providing evidence from the U.S. casino industry. When the stock prices were used as the measure of performance, Barber et al. (2006) reported a positive relationship between stock price and total CEO compensation, running their analysis over 73 restaurant rms. Contrary to the line of research that report a positive association between CEO pay and accounting performance, there is another stream of research that indicate no signicant relationship between CEO pay and accounting performance. Among those, for instance, Madura et al. (1996) studied small businesses that have a market value of less than 300 million as their sample rms and documented no signicant relationship between ROE and CEO compensation. Likewise, Skalpe (2007) reported that one-year lagged ROA does not signicantly affect CEO compensation. In a similar context, Kim and Gu (2005) examined the U.S. restaurant industry and also reported no signicant relationship between return on investment and CEO cash compensation. Several studies, like Crumleys (2008), found a weak relationship between executive compensation and stock return. As the previous studies show, executive compensation and company performance are not always directly related and show uctuations with respect to industry specic characteristics. 3. Research hypotheses Jensen and Murphy (1990) stated that it is difcult to determine the empirical order of the relationship between compensation and performance. Most often, researchers investigated this alleged relationship from the view that compensation committees structure CEO compensation packages based on rm performance measures, and whether CEOs were able to achieve or exceed those performance benchmarks. This view suggests that performance benchmarks are core determinants to reward (punish) CEOs for their success (failure). However, a signicant number of studies suggest that accomplishing performance benchmarks does not

guarantee increased compensation. There is also some evidence that the proposed positive relationship between performance and compensation is indeed in the inverse direction, which suggests that compensation committees do not put a pay-premium on attaining performance benchmarks. As opposed to this line of research, our contemplation is that executives performance is dependent on the level of compensation they are paid. Compensation contracts contain several cash and equity based components, each of which poses a different effect on motivating executives to excel for company performance. Cash compensation is usually determined in advance at the beginning of each calendar year, whereas the equity compensation is contingent on the performance that is usually observed at the end of the scal year. Given this fact, we argue that there is a direct effect of known cash compensation on the company performance in the current year, while we expect equity compensation to have a future effect on the compensation. The delayed effect of equity compensation is due to the fact that equity compensation awarded in time t will be valued in the upcoming periods (t + 1, t + 2,. . .). Because equity compensation is comprised of securities (i.e. stocks, options, option rights, etc.), executives (also the owners of these securities) should strive to increase the values of these securities in order to amplify their own wealth. Hence, we would expect equity compensation awarded in year t to boost the performance in future years within the horizon of these securities. For simplicity and due to limited data size, we adhere to a short-time horizon for lagged effect and predict that equity compensation paid in year t 1 to executives should have a positive effect on rm performance in the current year (year t). Hence, we test following hypotheses to investigate whether contemporaneous cash compensation and one-year lagged equity compensation paid to CEOs and CFOs affect the accounting performance of lodging rms. We expect both compensation variables to return positive coefcients. H1a. There is a positive relationship between CEO cash compensation and accounting-based rm performance in the U.S. lodging industry. H1b. There is a positive relationship between CFO cash compensation and accounting-based rm performance in the U.S. lodging industry. As noted above, we expect equity compensation to have a future effect on the rm performance. Considering that, we also examine the stock returns because equity compensation components are all tied to stock performance in the upcoming years. For instance, if we take stock options, we can see that this compensation component ties executives future total wealth to the appreciation of stock price till the exercise day. By logic, executives who hold a signicant number of stock options as part of their overall compensation packages need to exert a greater effort to maximize the stock price within the holding period so that they can realize the difference between the strike price and the spot price on the exercise day. The same thinking holds for almost all equity compensation components whose values are tied to future stock prices. With that reasoning, we expect that the previous periods equity compensation should be positively related with the stock return in the current period, which is tested with below hypotheses for CEOs and CFOs. H2a. There is a positive relationship between CEOs equitycompensation and stock return for companies in the U.S. lodging industry. H2b. There is a positive relationship between CFOs equitycompensation and stock return for companies in the U.S. lodging industry. Fig. 1 depicts the proposed empirical relationship among the variables of interests as well as summarizing the hypotheses.

A. Upneja, O. Ozdemir / International Journal of Hospitality Management 38 (2014) 3038 Table 1 The sample rms included in our study and the data sources. Firm name Starwood Hotels & Resorts Worldwide Inc. Marriott International Inc. Red Lion Hotels Corporation Interstate Hotels and Resorts Great Wolf Resorts Inc Morgans Hotel Group Co. Lodgian Inc. Sonesta International Hotels Marcus Corporation Vail Resorts Inc. Choice Hotels Corporation Bluegreen Corporation Data source

33

Fig. 1. Proposed relationship between pay and performance.

4. Data and methodology We run three regression models via ordinary least square (OLS) method to test hypotheses of the study. We identify ROA and Tobins Q as the proxy of accounting performance, and test whether they are positively related to contemporaneous cash compensation hypothesis 1a and 1b. Although we do not specically formulate a direct association between equity-based compensation and accounting performance, we still include this variable along with other control variables rm size, leverage, CEO tenure and CEO change to alleviate the effect of a potential omitted variables bias problem.2 Hypothesis 2a and 2b test the relationship between the lagged equity-based compensation and stock returns for CEO and CFO respectively. As we discussed previously, we expect current stock performance to be positively related to lagged equity-based compensation because value of these equity-based pay components are tied to future stock price, specically to the stock price at the exercise date. 4.1. Sample and data collection We use data compiled from several data sources. We rst start with COMPUSTAT to obtain a list of publicly traded companies whose main line of business is lodging services. We searched the COMPUSTAT database for North American Industry Classication System (NAICS) code 721110 for hotels (except casino hotels) and motels. COMPUSTAT returned 27 rms, of which twelve were inactive and fteen were active during the period of 20022012, the years for which we collected data. Out of these 15 active companies, two were privatized, and one led bankruptcy. Therefore, those companies were excluded from the sample list. Additionally, three of the remaining twelve active companies were foreign companies, and one company did not provide executive compensation information for the period 20022012. After exclusion of these three foreign rms and one domestic rm that did not provide compensation data, we were left with eight rms that constituted our initial sample from COMPUSTAT. Not all publicly traded rms are listed on COMPUSTAT. Therefore, we searched for publicly-listed rms whose main line of business are hotels and lodging related services but which are not listed on COMPUSTAT. We were able to nd four additional rms that match our search criteria listed under the Travel and Leisure Consumer Services classication in New York Stock Exchange (NYSE). The sample rms are listed in Table 1. Most of the data was hand-collected from annual reports (10-K) and annual proxy reports (DEF 14) that are led with Securities Exchange Commission (SEC). While annual reports provide accounting data for calculation of performance measures and control variables, proxy statements provide detailed compensation data for each named executive. Particularly, proxy statements include individual pay components of salary, bonus, stock awards,

COMPUSTAT COMPUSTAT COMPUSTAT COMPUSTAT COMPUSTAT COMPUSTAT COMPUSTAT COMPUSTAT NYSE (not on COMPUSTAT) NYSE (not on COMPUSTAT) NYSE (not on COMPUSTAT) NYSE (not on COMPUSTAT)

and option awards. Closing stock prices for each year are obtained from www.nance.yahoo.com. 4.2. Variables and models Consistent with prior research we use both accounting-based and market-based performance measures. Mehran (1995) uses ROA and Tobins Q as the proxy of accounting-based performance measures in a similar study where he investigated the executive compensation structure and rm performance. We follow Mehran (1995) in this regard and use ROA and Tobins Q. Whether or not ROA is an appropriate proxy for performance is a long-debated issue in the literature. Some argue that accounting returns, including ROA, convey little information about economic rates of return (Fisher and McGowan, 1983; Benston, 1985). On the other hand, Paul (1992) discusses that accounting returns, including ROA, provide information to the board about the value CEO added to the rm. In line with Paul (1992) we argue that our choice of ROA as the proxy for accounting return in this study is reasonable to our research question. ROA is operationalized as the ratio of net income to average total assets. Tobins Q is computed as the ratio of the sum of the equity market value and liabilities book value to the sum of equity book value and liabilities book value. Our third dependent variable is stock return, which has been used in previous studies as the proxy of market-based performance measure. Consistent with the literature, we use the adjusted stock return as our proxy for market-based performance measure. Stock return is operationalized as closing price at scal year-end plus dividends divided by the closing price of the prior scal year-end (Canarella and Nourayi, 2008). Our main independent variables are CEO (CFO) contemporaneous cash compensation and one-year lagged equity-compensation. Cash compensation is simply operationalized as the natural logarithm of salary plus bonus paid to the executives (Sigler and Haley, 1995). Equity-compensation is operationalized as the natural logarithm of sum of stock awards and stock options as reported in the proxy statements. Prior to FASB 123R, which became effective in 2006, companies reported the dollar value of stock awards granted and number of stock options granted in the proxy statements along with the estimated potential realizable value at assumed annual rates of stock price appreciation for option term (at 5% and 10%). Thus, we operationalize the stock awards as the dollar amount represented in the proxy statements and the option awards as the potential realizable value of options granted at 5% annual rate of stock price appreciation. Beginning with 2006, companies started to comply with FASB 123R, and they have disclosed compensation expense pertaining to the option awards and stock awards for the named executives recognized in year t. We take the sum of compensation expenses recognized in year t for option awards and stock awards for named executive as the proxy of equity-compensation for the after-2006 period. Following Mehran (1995), we include

2 The data set in this study is limited to twelve publicly traded lodging companies that provided 95 (72) rm-year observations for CEO (CFO) analysis for the period of 20022012. Small sample size prevents us from adding more control variables.

34

A. Upneja, O. Ozdemir / International Journal of Hospitality Management 38 (2014) 3038 Table 2 Descriptive statistics. Mean ROA Tobins Q Stock Return CashComp CEO EquityComp CEO Size Leverage Tenure CEO New CEO CashComp CFO EquityComp CFO Tenure CFO New CFO 0.0383 1.6954 0.14134 $959,732 $1,937,223 20.6409 0.4316 9.56 0.18 $483,357 $908,103 4.02 0.17 Std. dev. 0.1028 1.9873 0.60697 $688,965 $3,176,559 1.3147 0.2433 12.27 0.39 $254,663 1,271,651 2.26 0.38 Minimum 0.1342 0.4140 0.84302 $216,000 $0 18.2066 0.0964 1.00 0 $122,019 $0 1.00 0 Maximum 0.3967 11.8973 3.15385 $3,802,365 $20,844,526 23.2453 1.6586 40.00 1 $1,199,536 $6,753,728 10.00 1

four control variables: rm size, leverage, CEO (CFO) tenure, and new-CEO (CFO) in our models. Size is operationalized as the natural logarithm of total assets. Because larger rms are expected to pay more to their executives, it is reasonable to assume that they expect higher company performance. Based on this assumption, it is reasonable to control for rm size. Leverage, on the other hand, follows the previous literature and is used as the ratio of long-term debt to total assets. CEO (CFO) tenure is included to control for the effect of tenure and expertise on the job. The last control variable new-CEO (CFO) is included in the models in case a change in these positions distorts the lagged relationship between equitycompensation and rm performance. Being a dummy indicator, this variable takes on a value of 1 if a new CEO (CFO) takes on the position in year t relative to year t 1, 0 otherwise. All regression models in the paper include year dummies to control for the macro-level factors that could have similar effects on all the rms with regards to compensation practices in the lodging industry.3 We estimate the following regression models via OLS: ROAj,t = 0 + 1 CashCompj,t + 2 EquityCompj,t 1 + 3 Sizej,t + 4 Leveragej,t + 5 Tenurej,t + 6 New-CEO(CFO)j,t + D.Year + Tobin s Qj,t = 0 + 1 CashCompj,t + 2 EquityCompj,t 1 + 3 Sizej,t + 4 Leveragej,t + 5 Tenurej,t + 6 New-CEO(CFO)j,t + D.Year + RTRj,t = 0 + 1 CashCompj,t + 2 EquityCompj,t 1 + 3 Sizej,t + 4 Leveragej,t + 5 Tenurej,t + 6 New-CEO(CFO)j,t + D.Year + where, ROA is rm js return on assets in year t and operationalized as the net income divided by average total assets. CashComp is the natural logarithm of sum of salary plus bonus. EquityComp is the natural logarithm of sum of stock awards and stock options. Size is a control variable estimated as the natural logarithm of the book value of assets. Leverage is a control variable and estimated as the ratio of long-term debt to total assets. Tenure is the number of years passed in the position since starting the position. New-CEO (CFO) is a dummy variable indicating whether a new executive is in charge in year t relative to year t 1. It takes on a value of 1 if change takes place, 0 otherwise. Tobins Q is ratio of equity market value plus book value of liabilities to equity book value plus book value of liabilities. RTR is rm js stock return in year t. In accordance with previous research, we use Pearson Productmoment Correlation to analyze the correlations among variables and OLS to test our hypotheses. We regress each dependent variable (ROA, Tobins Q and stock return) on our variables of interest, cash compensation, and equity-compensation, along with the control variables. 5. Results 5.1. Descriptive statistics Table 2 displays summary statistics for the main and control variables employed in the current study. On average, lodging rms in our sample yielded 4% return on assets spanning a range of 13% to 40%. Likewise, other accounting performance measure, Tobins Q, produced positive value for the lodging rms within the observations years. Average Tobins Q was about 1.7, implying an average value gain over cost of the lodging rms assets.

Stock returns exhibit large swings, and present a widely dispersed range of a minimum of 84% and a maximum of 315%. On average, lodging rms annual stock return was found to be 14%. Lodging rms CEOs average equity-compensation exceeds average cashcompensation by about 50%, which is consistent with the idea that equity-based compensation is becoming the more signicant component of CEOs total pay. The highest value of CEO equity-pay ($20,844,526) observed in our sample suggests that some large rms make substantial use of equity-based compensation as an incentive tool. In comparison, the highest value of CEO cash compensation remains at $3,802,365, which is far below the maximum equity-compensation. CFOs equity-compensation is also substantially higher than their cash-compensation. Lodging rms CFOs, on average, earned $908,103 equity-based compensation and $483,357 cash-based compensation. This is about a 50% difference and qualitatively similar to CEOs compensation mix. This particular nding suggests a consistent application of a compensation mix across senior positions; however, we avoid deriving further implications from this nding without examining other senior executives compensation mix. 5.2. Empirical ndings Table 3 displays regression coefcients of the models that capture the impact of lodging rms CEOs cash-based and equity-based compensation on three distinct corporate performance measures: ROA, Tobins Q and stock return. Our results demonstrate that CEOs contemporaneous cash compensation and one-year lagged equity compensation are positively related to their rms accounting-base performance measures, which provides evidence in the favor of hypothesis 1a. The evidence in Table 3 suggests that 1% increase in CEOs contemporaneous cash compensation tends to increase average lodging rms ROA by 0.00037 percentage points, and Tobins Q by 0.00578% points. These relationships are both statistically significant at the 10% level, with respective p-values of 0.053 for the ROA model and 0.058 for the Tobins Q model. Likewise, ROA and Tobins Q in the current year increase by 0.00004% points (p-value = 0.026) and 0.0009% points (p-value = 0.000) respectively when these rms CEOs are compensated with one percent more equity-based compensation in the prior year; and as the p-values show the effects are strongly signicant. To make interpretations about the economic impact of compensation variables on the performance measures, we divide the coefcients of both compensation variables reported in the Table 3 by 100 because these variables are log-transformed.4

3 For brevity, we do not report the year-dummy coefcients in the tables displaying the regression results.

4 In a regression model where the dependent variables is in its original metric, but the independent variable is log-transformed the economic interpretation follows as: one percent change in the independent variable leads to coefcient value/100 units change in the dependent variable.

A. Upneja, O. Ozdemir / International Journal of Hospitality Management 38 (2014) 3038

35

Table 3 Relationship between CEO compensation and rm performance. In the table below, we summarize the ndings of our main regression analysis. We investigate the relationship between three different performance measures, ROA, Tobins Q and Stock Return, and CEO compensation measured by cash compensation and equity-based compensation. We regress each performance measures individually on the compensation variables plus the control variables. ROA Coefcient CashCompt EquityCompt1 Sizet Leveraget Tenuret New (CEO)t Constant F-statistic Adj. R-Sq N 0.037 0.004 0.022 0.140 0.002 0.005 0.1351 2.91 0.233 95 p-Value 0.053 0.026 0.041 0.001 0.019 0.828 0.553 Tobins Q Coefcient 0.5785 0.0909 0.2930 4.8339 0.0321 0.6130 1.7625 8.58 0.647 79 p-Value 0.058 0.000 0.046 0.000 0.001 0.035 0.694 Return Coefcient 0.0731 0.0081 0.0161 0.0401 0.0010 0.2292 0.6223 6.87 0.492 92 p-Value 0.370 0.211 0.707 0.804 0.761 0.022 0.600

ROA is the ratio of net income to the average total asset. Tobins Q is the ratio of equity market value plus book value of liabilities to equity book value plus book value of liabilities. Return is the ratio of the difference between the closing stock price of the current year and the prior year to the closing stock price of the prior year. CashComp is the natural logarithm of total cash compensation (salary + bonus). EquityComp is the natural logarithm of sum of stock options and stock awards. Size is the natural logarithm of the book value of assets. Leverage is the ratio of long-term debt to total assets. Tenure represents the number of years passed in the CEO position in the same rm. New is an indicator variable whether a new CEO takes on the position in year t compared to year t 1. It takes a value of 1 if a new CEO is in charge in year t, 0 otherwise. To eliminate the extreme ROA values that distort the normality of the residuals in the ROA equation, ROA is winsorized at 5% level. All the three regressions include year dummies but the coefcients are not reported for clarity and brevity.

Table 4 Relationship between CFO compensation and rm performance. In the table below, we summarize the ndings of our main regression analysis. We investigate the relationship between three different performance measures, ROA, Tobins Q and Stock Return, and CFO compensation measured by cash compensation and equity-based compensation. We regress each performance measures individually on the compensation variables plus the control variables. ROA Coefcient CashCompt EquityCompt1 Sizet Leveraget Tenuret New (CFO)t Constant F-statistic Adj. R-Sq N 0.0456 0.0016 0.0031 0.1200 0.0041 0.0013 0.5717 3.33 0.336 72 p-Value 0.037 0.406 0.727 0.000 0.354 0.958 0.007 Tobins Q Coefcient 0.8762 0.0601 0.1282 5.0538 0.0161 0.2142 9.6003 4.39 0.459 66 p-Value 0.033 0.108 0.451 0.000 0.867 0.664 0.016 Return Coefcient 0.0898 0.0006 0.0222 0.1063 0.0020 0.0522 0.3346 8.16 0.599 73 p-Value 0.426 0.950 0.631 0.486 0.925 0.674 0.739

ROA is the ratio of net income to the average total asset. Tobins Q is the ratio of equity market value plus book value of liabilities to equity book value plus book value of liabilities. Return is the ratio of the difference between the closing stock price of the current year and the prior year to the closing stock price of the prior year. CashComp is the natural logarithm of total cash compensation (salary + bonus). EquityComp is the natural logarithm of sum of stock options and stock awards. Size is the natural logarithm of the book value of assets. Leverage is the ratio of long-term debt to total assets. Tenure represent the number of years passed in the CEO position in the same rm. New is an indicator variable whether a new CEO takes on the position in year t compared to year t 1. It takes a value of 1 if a new CEO is in charge in year t, 0 otherwise. To eliminate the extreme ROA values that distort the normality of the residuals in the ROA equation, ROA is winsorized at 5% level. All the three regressions include year dummies but the results are not reported for clarity and brevity.

In both models, leverage and CEO tenure signicantly and positively affect the performance variables, and rm size is inversely related to rm performance. New-CEO variable is not signicant in the ROA model and highly signicant in the Tobins Q model. Contrary to accounting-based performance measures, market-based performance measure, stock return yields non-signicant results. Neither of the compensation variables was found to be related to the lodging rms stock returns, which rejects hypothesis 2a. This is in direct contrast to our assumption that higher equity-based compensation would motivate CEOs to achieve higher stock returns, which, in turn, will raise their own compensation. In fact, this nding is also contrary to our previous observations outlined in the descriptive statistics. The average rm in our sample paid about 50% more equity-based compensation to their CEOs compared to cashcompensation, yet this incentive seems not to result in a higher stock return for lodging rms. Additionally, we observed no significant relationship between stock returns and control variables.

We repeat the same analysis for CFOs. The results are reported in Table 4. Findings of this analysis suggest that CFOs contemporaneous cash compensation positively affects ROA and Tobins Q. This nding provides support for hypothesis 1b. The positive coefcients of 0.0456 (p-value = 0.037) and 0.8762 (p-value = 0.033) imply 0.00046% point increase in ROA and 0.0088% point increase in Tobins Q as CFOs are awarded with 1% more cash compensation. One-year lagged equity-compensation, on the other hand, is not signicant at the conventional signicance levels in both ROA and Tobins Q models with p-values of 0.406 and 0.108 respectively. In both models, only signicant control variable is the leverage. Results further exhibit that stock returns are not statistically related to contemporaneous cash compensation and one-year lagged equity-based compensation. Particularly, insignicant relationship between lagged equitybased compensation and stock returns leads us to reject hypothesis 2b.

36 Table 5 Robustness tests.

A. Upneja, O. Ozdemir / International Journal of Hospitality Management 38 (2014) 3038

Panel A. Further lag periods of equity-compensation CEO analysis ROA CashCompt EquityCompt1 EquityCompt2 EquityCompt3 Adj. R-Sq N 0.028 (0.84) 0.001 (0.42) 0.003 (1.04) 0.001 (0.53) 0.205 71 Tobins Q 1.629 (1.77) 0.011 (0.19) 0.076 (1.38) 0.0968* (2.25) 0.572 55 Return 0.146 (0.85) 0.006 (0.36) 0.005 (0.36) 0.003 (0.26) 0.274 71 CFO analysis ROA 0.045 (0.69) 0.002 (0.34) 0.002 (0.49) 0.004 (0.99) 0.081 57 Tobins Q 0.206 (0.19) 0.039 (0.47) 0.049 (0.58) 0.128* (2.09) 0.471 46 Return 0.061 (0.36) 0.004 (0.27) 0.014 (1.11) 0.001 (0.13) 0.593 57

Panel B. Effect of future compensation on current rm performance CEO analysis ROA CashCompt+1 EquityComp EquityCompt+1 EquityCompt+2 Adj. R-Sq N 0.028 (1.26) 0.003 (1.40) 0.002 (0.77) 0.002 (0.79) 0.411 71 Tobins Q 0.550 (1.50) 0.057 (1.44) 0.066 (1.55) 0.065 (1.61) 0.53 75 Return 0.110 (0.64) 0.030 (1.68) 0.003 (0.15) 0.017 (0.94) 0.042 70 CFO analysis ROA 0.013 (0.31) 0.003 (0.82) 0.002 (0.43) 0.003 (0.56) 0.138 56.0 Tobins Q 0.376 (0.58) 0.042 (0.68) 0.088 (1.20) 0.055 (0.68) 0.375 60.0 Return 0.339 (1.62) 0.026 (1.29) 0.022 (0.98) 0.008 (0.29) 0.025 55.0

t statistics are presented in parentheses. * Signicant at p < 0.05.

5.3. Robustness tests 5.3.1. Further lag periods of equity-compensation Further lag periods of equity-compensation are introduced to the base model to examine whether equity-based compensation paid to named executives in farther prior years (year t 2, year t 3) subsequently lead to higher rm performance. A priori, the expectation in this specication is that equity-pay components such as stock options have vesting periods and only be cashed out in time. Therefore, it would be expected that performance measures, at least the stock returns, should improve during the holding period so that holders of these options can realize compensation gains on their stock options. Yet, the ndings reported in Panel A of Table 5 suggest no such implications. As further lags of equity-compensation (2-year and 3-year lagged equity-based compensation) are introduced to the base models, no statistical signicance is achieved. Inclusion of further lags distorts the ndings of the original models as well. These ndings are consistent across the CFO analyses. Neither of the further lags of the equity-based compensation is signicant in explaining current lodging rm performance in the CFO analysis as well.

ahead equity-based compensation of the CEO is signicant in the models, which provides further evidence to strengthen the pay-toperformance direction rather than performance-to-pay direction, at least for the lodging rms observed in this study. Similar results are produced when the CFO compensation is used as the interest of the analysis.

6. Concluding remarks This paper examines the empirical evidence on whether executives cash and equity-based compensation give them an incentive to increase nancial performance of their companies. We investigate this phenomenon in the U.S. lodging industry and nd evidence that lodging rms CEOs contemporaneous cash compensation and one-year lagged equity-based compensation positively impact the accounting performance measures of ROA and Tobins Q. These ndings are in line with previous studies ndings (Crumley, 2008; Attaway, 2000; Gu and Choi, 2004). However, our results for accounting returns are contrary to those of Core et al. (1999) and Madura et al. (1996). While Core et al. (1999) report an insignicant relationship between compensation and ROA; Madura et al. (1996) document an insignicant relationship between compensation and ROE. In contrast to accounting performance measures, market performance measure, stock return, is not related to any compensation variable we tested. Particularly, we are surprised with the nding that stock performance is not related to equity-based compensation in our sample of lodging rms. Yet, this particular nding is consistent with the ndings of previous studies such as Attaway (2000), Murthy and Salter, 1975, and Madura et al. (1996), which reported substantially weak or no correlation between stock return and compensation. A further intriguing nding reveals itself in the robustness tests when further lags of equity compensation are introduced to the models. Compensation components such as stock options are normally distributed to managers/executives with pre-determined vesting periods. Hence, for an executive to make excessive gains on his/her option holdings, stock value at the time of the exercise date should exceed the pre-determined contractual striking price, which intuitively makes one assume that the executive should outperform during the holding period to maximize his/her wealth

5.3.2. Pay-performance simultaneity In this specication, we examine whether the inverse of the proposed relationship as evidenced in the long-argued payperformance literature holds for the sample of lodging rms included in the current study. We regress sample lodging rms performance (ROA, Tobins Q and stock returns) measures in year t on the following years (year t + 1) cash compensation and current equity-based compensation. We include year t + 1 and year t + 2 equity compensation variables into the base model in the further model specications. Although this specication does not solve the simultaneity problem entirely, it gives some assurance for the direction of the relationship evidenced in the current studys research design, which simply proposes that compensation is given rst and then performance is expected. Results of these specications are reported in Panel B of Table 5. The results show that CEOs future cash compensation is not related to current years performance, which is consistent across all three performance indicators employed. Likewise, neither one-year ahead or two-year

A. Upneja, O. Ozdemir / International Journal of Hospitality Management 38 (2014) 3038

37

on this particular compensation. In this context, stock awards along with stock options distributed to CEO and other executives in prior years should motivate these senior executives to maximize the stock value in excess of strike price so that they can increase their own wealth. Yet, for our sample of lodging rms, we do not see this expectation realize. Particularly, we do not observe any signicant effect of lagged equity-compensation on stock returns. The negative coefcient on the lagged-equity compensation variables is very intriguing and deserves more future research. Based on our current ndings, we abstain from jumping to a quick conclusion to state that equity-compensation is not used as a valuable motivational tool to increase a lodging rms stock performance, but rather conclude that there exists an anomaly with regards to equity-based compensation and its effect on the stock performance, which is evident with the consistently negative coefcients on lagged equity-compensation variables added to the model. Overall, these ndings provide signicant insights to the board compensation committees of the lodging rms. A foremost implication is that compensating CEOs with higher equity-based compensation does not seem to produce higher stock return performance for the lodging rms. Indeed, in a competitive world, tying a managers overall wealth to improved rm value via higher stock price performance seems to be a logical and effective choice, yet this does not seem to occur in the lodging industry. Secondly, our analysis suggests that accounting returns still seem to be the benchmarks for the compensation contracting purposes in the lodging industry. When CEOs are awarded with signicant cash compensation and equity-compensation, they achieve substantial accounting returns. The particularly signicant relationship between accounting performance measures and cash and equitybased compensation highlights the fact that lodging CEOs take the accounting performance measures as their benchmark even though a large part of their compensation is strictly tied to stock performance. One explanation for this particular nding is that CEOs might be viewing the stock performance as an externally driven market measure with management having limited control over it. Results for the CFO analysis are qualitatively similar to those of CEO. With that said, we can say there seems to be a consistent pattern across two senior positions in term of the effect of compensation on inuencing a lodging rm accounting performance.

disclosures based on FASB 123R regulations to proxy for the stockbased compensation. The advantage of this approach is that these numbers are widely reported and are consistent across different rms. Any alternative algorithm we develop would have faced the problem of not having sufcient underlying data for us to be able to estimate stock-compensation values consistently for each rm. Moreover, given that the direction of the pay-performance relationship is not clear-cut, one can also examine the alleged pay-performance relationship in a simultaneity context with a model that simultaneously treats compensation and rm performance as endogenous variables. As another extension to the eld, researchers might be interested in combining other research areas to the pay-performance research. For instance, internationalization increases the rm risk and researchers can investigate how this increased risk via internationalization impacts executive pay and rm performance.

References
Attaway, M.C., 2000. A study of the relationship between company performance and CEO compensation. American Business Review 18 (1), 7786. Barber, N., Ghiselli, R., Deale, C., 2006. Assessing the relationship of CEO compensation and company nancial performance in the restaurant segment of the hospitality industry. Journal of Food Service Business Research 9 (4), 6582. Benston, G.J., 1985. The validity of prots-structure studies with particular reference to the FTCS line of business data. American Economic Review 75, 3767. Boyd, B.K., 1994. Board control and CEO compensation. Strategic Management Journal 15 (5), 335344. Canarella, G., Nourayi, M.M., 2008. Executive compensation and rm performance: adjustment dynamics, non-linearity, and asymmetry. Managerial and Decision Economics 29, 293315. Core, J., Holthausen, R.W., Larcker, D.F., 1999. Corporate governance, chief executive ofcer compensation, and rm performance. Journal of Financial Economics 51, 371406. Crumley, C.R., 2008. A study of relationship between rm performance and CEO compensation in the U.S. commercial banking industry. The Journal of Applied Management and Entrepreneurship 13, 2645. Duru, A.I., Iyengar, R.J., 1999. Linking CEO pay to rm performance: empirical evidence from the electric utility industry. Managerial Finance 25 (9), 2134. Eisenhardt, K.M., 1989. Agency theory: an assessment and review. Academy of Management Review 14 (1), 5774. Fisher, F.M., McGowan, J.J., 1983. On the misuse of accounting rates of return to infer monopoly prots. American Economic Review 73, 8297. Glassman, C.A., Rhoades, S.A., 1980. Owner vs. manager control effects on bank performance. Review of Economics and Statistics 27, 263270. Gu, Z., Choi, Y.H., 2004. CEO compensation determinants in the casino industry. Journal of Hospitality & Tourism Research 28 (2), 143155. Huang, I., Huang, P., Chen, Y., 2004. A study of Taiwans travel agent salary system: an agency theory perspective. Tourism Management 25 (4), 483490. Jensen, M.C., Meckling, W.H., 1976. Theory of the rm: managerial behavior, agency costs and ownership structure. Journal of Financial Economics 3, 305360. Jensen, M.C., Murphy, K.J., 1990. Performance pay and top management incentives. Journal of Political Economy 98 (2), 225264. Jiang, J., Petroni, K., Wang, I., 2010. CFOs and CEOs: who have the most inuence on earnings management? Journal of Financial Economics 96 (3), 513526. Kay, I.T., 1999. Growing shareholder value: why executive stock ownership works. Compensation and Benets Review 31 (1), 3237. Kim, H., Gu, Z., 2005. A preliminary examination of determinants of CEO cash compensation in the U.S. restaurant industry from an agency theory perspective. Journal of Hospitality & Tourism Research 29 (3), 341355. Loderer, C., Martin, K., 1997. Executive stock ownership and performance tracking faint traces. Journal of Financial Economics 45, 223355. Madura, J., Martin, A.D., Jessel, K.A., 1996. Determinants of CEO compensation in small publicly-traded businesses. American Business Review 14, 8088. Mehran, H., 1995. Executive compensation structure, ownership, and rm performance. Journal of Financial Economics 38, 163184. Murphy, K., 1985. Corporate performance and managerial remuneration. Journal of Accounting and Economics 7, 1142. Murthy, K.R.S., Salter, M.S., 1975. Should CEO pay be linked to results? Harvard Business Review 53 (3), 6673. Paryani, K., Masoudi, A., Cudney, E.A., 2010. QFD application in the hospitality industry. The Quality Management Journal 17 (1), 728. Paul, J.M., 1992. On the efciency of stock-based compensation. Review of Financial Studies 5, 471502.

7. Limitations and future research We acknowledge that there are certain limitations in our research design. A major limitation of this study is the relatively small sample size. In fact, it could be more appropriate to say that the data set we use is indeed the true population for the period 20022012 to the extent that we were able to identify all publicly traded lodging companies in this time frame. We tried to be as thorough in identifying all the publicly traded lodging companies as possible while considering the data availability for the period. However, we acknowledge that results from running 95-rm year observations could only be relevant to the rms in our data set for the study period. Yet, one also needs to recognize that we worked with all the data available for lodging rms, and these rms are leaders in the lodging industry, which are likely to inuence the compensation practices of smaller rms that may be going public in the near future. We further recognize that more data points would truly yield more reliable parameter estimates to back our results in this study. One way of increasing the data size might be going back farther in the time period to include the pre-2002 period. Another limitation is the accuracy of our proxy for stock-based compensation. We use the valuation

38

A. Upneja, O. Ozdemir / International Journal of Hospitality Management 38 (2014) 3038 Veliyath, R., Bishop, J.W., 1995. Relationship between CEO compensation and rm performance: empirical evidence of labor market norms. The International Journal of Organizational Analysis 3 (3), 268283. Welbourne, T.M., Cyr, L.A., 1999. Using ownership as an incentive does the too many chiefs rule apply in entrepreneurial rms? Group & Organization Management 24 (4), 438460.

Sigler, K.J., Haley, J.P., 1995. CEO pay and company performance. Managerial Finance 21 (2), 3141. Skalpe, O., 2007. The CEO gender pay gap in the tourism industry evidence from Norway. Tourism Management 28, 845853. Tosi, H.L., Gomez-Mejia, L.R., 1994. CEO compensation monitoring and rm performance. Academy of Management Journal 37 (4), 216.