DID THE ASSET-LIGHT STRATEGY HELP HOTEL STOCKS STAY RESILIENT DURING COVID-19? A Thesis Presented to the Faculty of the Graduate School of Cornell University In Partial Fulfillment of the Requirements for the Degree of Master of Science by Yanjun Wang August 2025 © 2025 Yanjun Wang ABSTRACT This study investigates the role of the asset-light strategy in shaping the stock performance of hotel firms across the pre-COVID-19 and post-COVID-19 periods. By comparing publicly listed asset-light hotel C-corporations and asset-heavy hotel REITs from 2017 to 2019 and from 2020 to 2022, the research assesses whether structural differences in asset ownership influence hotel stocks’ risk exposure and return characteristics. The analysis first constructs Markowitz mean-variance frontiers to evaluate whether asset-light firms exhibit more favorable risk-return profiles than their asset-heavy counterparts. It then conducts panel regressions to examine whether such differences in performance are statistically associated with the degree of asset-light structure, as measured by a composite asset-light score. Results show that stocks of asset-light hotel firms consistently outperformed asset-heavy peers in both periods. Regression results further indicate that firms with higher asset-light scores experienced smaller increases in market beta during the pandemic, suggesting stronger downside protection. iii BIOGRAPHICAL SKETCH Yanjun Wang is an M.S. student in the School of Hotel Administration at Cornell University. iv ACKNOWLEDGMENTS To Cornell University, for providing me with the opportunity to learn, grow, and explore beyond boundaries. v TABLE OF CONTENTS BIOGRAPHICAL SKETCH ......................................................................................... iii ACKNOWLEDGMENTS ............................................................................................. iv TABLE OF CONTENTS ................................................................................................ v LIST OF FIGURES ....................................................................................................... vi LIST OF TABLES ........................................................................................................ vii 1 INTRODUCTION ....................................................................................................... 1 2 LITERATURE REVIEW ............................................................................................ 2 3 DATA AND METHODOLOGY ................................................................................. 3 3.1 Data Sources and Sample Selection ..................................................................... 3 3.2 Analytical Framework and Model Implementation .............................................. 5 3.2.1 Portfolio-Based Risk-Return Analysis........................................................... 6 3.2.2 Regression Model Specification .................................................................. 10 4 EMPIRICAL RESULTS ............................................................................................ 16 4.1 Stock Performance Patterns Across Two Periods ............................................... 16 4.1.1 Cumulative Return Trends ........................................................................... 16 4.1.2 Return, Risk, and Sharpe Ratio Patterns...................................................... 18 4.1.3 Mean-Variance Frontier Dynamics .............................................................. 22 4.2 Regression Results on Market Exposure ............................................................ 27 5 CONCLUSION ......................................................................................................... 32 REFERENCES ............................................................................................................. 33 APPENDIX .................................................................................................................. 35 vi LIST OF FIGURES Figure 1 Cumulative Stock Returns of Hotel C-Corporations ..................................... 17 Figure 2 Cumulative Stock Returns of Hotel REITs .................................................... 17 Figure 3 Average Daily Return Trends by Year .......................................................... 20 Figure 4 Average Daily Standard Deviation Trends by Year ...................................... 21 Figure 5 Sharpe Ratio Comparison: Pre- vs. Post-Pandemic ....................................... 21 Figure 6 Mean-Variance Frontier of Hotel C-Corporations (Short Allowed) .............. 24 Figure 7 Mean-Variance Frontier of Hotel REITs (Short Allowed) ............................ 25 Figure 8 Efficient Frontier Comparison of Hotel C-Corporations and REITs ............. 25 Figure 9 Heatmap of Correlations ................................................................................ 26 Figure 10 Beta Comparison: Pre- vs. Post-Pandemic .................................................. 28 Figure 11 Relationships Between Beta and Individual Asset-Light Variables ............ 31 vii LIST OF TABLES Table 1 Sample Firms by Asset Strategy Classification ................................................ 5 Table 2 Financial Ratios Used in Regression Analysis ................................................ 12 Table 3 Components and Z-Scores for Asset Light Score 1 (Partial Sample) ............. 13 Table 4 Components and Z-Scores for Asset Light Score 2 (Partial Sample) ............. 13 Table 5 Return, Risk, and Sharpe Ratios ...................................................................... 19 Table 6 Summary of Portfolio-Level Risk-Return Metrics .......................................... 23 Table 7 Estimated Betas of Hotel C-Corporations and REITs ..................................... 27 Table 8 Panel Regression Results, Model 1 ................................................................. 29 Table 9 Panel Regression Results, Model 2 ................................................................. 29 1 1 INTRODUCTION Hotel firms differ in how they structure capital and generate revenue. Firms such as Marriott and Hilton, which adopt the asset-light strategy, generate income primarily through franchise and management fees. These firms operate with minimal ownership of physical assets, allowing them to maintain lower fixed costs and more flexible capital structures (Li and Singal, 2019). In contrast, asset-heavy hotel firms, including many hotel REITs, retain ownership of real estate assets and derive income directly from hotel operations and property-level cash flows. Such structural variation may lead to asymmetric performance before and during the COVID-19 pandemic. This pattern is theoretically supported by Sohn, Tang, and Jang (2014), who find that the effectiveness of the asset-light strategy varies across the business cycle. Empirical evidence during the COVID-19 crisis further confirms this tendency, as shown by García Gómez et al. (2021). Asset-light hotel firms are expected to demonstrate greater resilience during periods of crisis. Their fee-based revenue streams are less sensitive to short-term fluctuations in occupancy and room rates, and their leaner balance sheets reduce the burden of debt servicing and fixed obligations when demand weakens. By minimizing capital exposure and decoupling earnings from asset ownership, asset-light firms are better positioned to preserve margins and avoid sharp declines in financial performance during the pandemic. In contrast, asset-heavy hotel firms may perform better during relatively favorable periods, such as the pre-COVID-19 years. Because they own the underlying real estate, these firms are positioned to capture a larger share of upside gains, including RevPAR growth, operating leverage, and property appreciation. Their capital- intensive structures can amplify earnings in such conditions, enabling them to benefit 2 more from increases in travel demand and real asset values observed during that time (Kim and Jang, 2012; Kim, Noh, and Lee 2018). To evaluate how the asset-light strategy interacts with macroeconomic shocks, this study compares the stock performance of publicly listed hotel C-corporations and hotel REITs before and during the COVID-19 pandemic. Specifically, it first constructs Markowitz mean-variance frontiers to compare changes in the risk-return profiles of these two groups across pre- and post-pandemic periods. Regression analysis is then employed to examine whether the asset-light strategy had a significant effect on stock performance, building on prior work by Sohn, Tang, and Jang (2013) and Li and Singal (2019), who used similar methods to assess the impact of the asset strategy on market exposure. 2 LITERATURE REVIEW The asset-light strategy has fundamentally transformed how hotel firms operate. Rather than owning hotel properties directly, many firms now focus on generating income through franchise and management contracts, resulting in reduced capital intensity and enhanced financial flexibility. This strategy has been associated with lower leverage, stronger interest coverage, and improved return on invested capital, as shown by Li and Singal (2019) and Sohn, Tang, and Jang (2013). These financial characteristics allow hotel firms to scale more easily and generate more stable fee- based income with lower exposure to operational volatility. It also plays a critical role in shaping stock performance, especially during economic downturns. Kim, Gu, and Mattila (2002) and Lee (2012) find that firms retaining real estate ownership, such as hotel REITs, typically face higher market betas and greater sensitivity to real estate price fluctuations. While REITs benefit from 3 asset ownership during favorable market conditions, Kim and Jang (2012) find that they tend to underperform C-Corporations on a risk-adjusted basis, suggesting that their higher return volatility offsets potential upside gains. Supporting this view, Kim, Noh, and Lee (2018) document stronger co-movement between REITs and property indices, further highlighting their exposure to real estate market risk. During periods of crisis, these differences become more pronounced. Pal (2021) notes that volatility in the hospitality sector reacts asymmetrically to external shocks, with negative events such as health or financial crises inducing stronger volatility spikes than positive developments, and García Gómez et al. (2021) show that firms with higher fee-based income shares experienced smaller valuation losses during COVID-19. In addition to employing firm-level regression analysis, this study incorporates a portfolio-level perspective to compare the performance of hotel REITs and C- corporations. By constructing mean-variance frontiers and examining risk, return and Sharpe ratio of minimum-variance portfolios before and during the COVID-19 pandemic, the analysis offers a more objective view of how these two asset structures differ in overall risk-return characteristics. This method reduces the influence of firm- specific outliers and provides a clearer picture of group behavior under varying market conditions. Prior work, such as Low, Das, and Piffaretti (2015), applies a similar framework to assess hotel assets in broader mixed-asset portfolios. 3 DATA AND METHODOLOGY 3.1 Data Sources and Sample Selection This study relies on three categories of data: daily stock prices, asset pricing factors, and firm-level financial ratios. Daily stock price data for publicly listed hotel C- corporations and hotel REITs were obtained from Yahoo Finance, covering the period 4 from January 2017 through December 2022. These data were used to calculate cumulative returns, average daily returns, average daily standard deviations, and beta coefficients. The Fama French five-factor data required for beta estimation were accessed from Kenneth R. French’s data library hosted by Dartmouth College. Firm- level financial ratios used as independent variables in the regression analysis were retrieved from S&P Capital IQ under the Financials section, specifically from the Financial Ratios category. The selected variables include Owned Hotel Share, Net PP&E to Total Assets Ratio, Depreciation to Total Assets Ratio, Fixed Asset Turnover (FAT), CapEx to Revenue Ratio, Debt to Capital Ratio, and Return on Invested Capital (ROIC). For a small number of firms with missing values for certain variables, supplementary information was manually collected from their 10-K filings to ensure completeness. To ensure consistency in reporting and comparability across firms, the sample was selected based on three criteria. First, all firms must be publicly traded and have complete daily stock price data available from 2017 through 2022. Second, the sample includes only firms operating under a dedicated hotel business model, thereby excluding firms primarily involved in online travel booking, platform-based services, or diversified operations, such as Booking Holdings and Airbnb. Third, each firm must be clearly identifiable as either asset-light or asset-heavy based on business model disclosures, financial filings, and industry classifications. Asset-light firms, such as Marriott International, Hilton Worldwide, and InterContinental Hotels Group, primarily generate revenue through brand licensing and management contracts, while asset-heavy hotel REITs, including Host Hotels & Resorts, Park Hotels & Resorts, and Ryman Hospitality Properties, own and operate portfolios of hotel real estate. The 5 final sample consists of six hotel C-corporations1 classified as asset-light firms and thirteen hotel REITs classified as asset-heavy firms. A complete list of sample firms, along with their classification and ticker symbols, is presented in Table 1. Table 1 Sample Firms by Asset Strategy Classification Asset-Light Hotel C-Corporations Asset-Heavy Hotel REITs Firm Name Ticker Firm Name Ticker Accor S.A. AC Apple Hospitality REIT, Inc. APLE Choice Hotels International, Inc. CHH Ashford Hospitality Trust, Inc. AHT Hilton Worldwide Holdings Inc. HLT Braemar Hotels & Resorts Inc. BHR Hyatt Hotels Corporation H DiamondRock Hospitality Company DRH InterContinental Hotels Group PLC IHG Host Hotels & Resorts, Inc. HST Marriott International, Inc. MAR Park Hotels & Resorts Inc. PK Pebblebrook Hotel Trust PEB RLJ Lodging Trust RLJ Ryman Hospitality Properties, Inc. RHP Service Properties Trust SVC Summit Hotel Properties, Inc. INN Sunstone Hotel Investors, Inc. SHO Xenia Hotels & Resorts, Inc. XHR 3.2 Analytical Framework and Model Implementation The empirical analysis consists of two components. First, Markowitz mean-variance frontiers are constructed to compare changes in the risk-return profiles of two groups of hotel firms: asset light and asset heavy, before and during the COVID-19 pandemic. This portfolio-level analysis complements firm-level observations by providing a clearer view of group-level risk-return trade-offs, especially given that individual firm 1 Wyndham Hotels & Resorts was excluded from the sample due to the absence of complete data for the 2017–2019 period. The company became publicly traded on June 1, 2018, following its spin-off from Wyndham Worldwide Corporation. 6 metrics can vary considerably. Second, panel regression models are used to test whether firms that adopt asset-light strategies exhibit significantly different levels of market exposure, measured by beta, and whether these differences vary across two periods. 3.2.1 Portfolio-Based Risk-Return Analysis To analyze the aggregate performance of asset-light and asset-heavy hotel firms, the mean-variance frontier is constructed in three steps. The first step computes cumulative returns based on daily adjusted closing prices. The second step calculates average daily returns and average daily standard deviations for each firm by year, which serve as the key inputs for the third step. Building on these statistics, the third step constructs the mean-variance frontier through portfolio optimization, evaluating the trade-off between return and risk across asset-light and asset-heavy hotel groups. The first step calculates the cumulative return to measure the total investment gain before and during the COVID-19 pandemic. Let Pt represent the adjusted closing price of a stock on trading day 𝑡, and let Pt−1 represent the adjusted closing price on the previous trading day 𝑡 − 1. To calculate the daily return rt, the price on day t – 1 is first subtracted from the price on day 𝑡, then the result is divided by the price on day 𝑡 − 1. This gives the percentage change in price from one day to the next, as shown in Equation 3.2.1. After obtaining the series of daily returns, the cumulative return over 𝑇 trading days is calculated by first adding 1 to each daily return rt, then multiplying all resulting terms from 𝑡 = 1 to t = T. Finally, 1 is subtracted from the product to get the total compounded return over the full period, as shown in Equation 3.2.2. Equation 3.2.1 𝑟𝑡 = 𝑃𝑡 − 𝑃𝑡−1 𝑃𝑡−1 7 Equation 3.2.2 𝑅cumulative = ∏(1 + 𝑟𝑡) 𝑇 𝑡=1 − 1 The second step calculates the average daily return and average daily standard deviation for each stock i within a specific year 𝑦. Let 𝑟𝑖,𝑡 denote the daily return of stock 𝑖 on trading day 𝑡. Let 𝑇𝑦 represent the set of all trading days in year 𝑦, and |𝑇𝑦| denote the total number of trading days in that year. To compute the average daily return 𝑟𝑖,𝑦, the individual daily returns ri,t are summed across all trading days 𝑡 in year 𝑦, and the sum is then divided by |𝑇𝑦|, the number of trading days. This gives the arithmetic mean return per day for that stock during the year, as shown in Equation 3.2.3. To calculate the average daily standard deviation Si,y, the squared difference between each daily return ri,t and the average return 𝑟𝑖,𝑦 is first computed. These squared deviations are then summed across all trading days. The sum is divided by |𝑇𝑦| − 1 , which adjusts for sample size when estimating the population variance. Finally, the square root of the result is taken to obtain the standard deviation, which reflects the average size of daily return fluctuations over the year, as shown in Equation 3.2.4. Equation 3.2.3 𝑟𝑖,𝑦 = 1 |𝑇𝑦| ∑ 𝑟𝑖,𝑡 𝑡∈𝑇𝑦 Equation 3.2.4 𝑆𝑖,𝑦 = √ 1 |𝑇𝑦| − 1 ∑(𝑟𝑖,𝑡 − 𝑟𝑖,𝑦) 2 𝑡∈𝑇𝑦 8 The third step constructs the mean-variance frontier by solving for the minimum- variance portfolio and the tangency portfolio. These two optimized portfolios define the boundary of the mean-variance frontier. Let 𝒘 represent the portfolio weight vector, 𝚺 represent the covariance matrix of asset returns, μ represent the expected return vector, and 𝑟𝑓 represent the risk-free rate. To derive the minimum-variance portfolio, the objective is to minimize the total portfolio variance. This is calculated by multiplying the transposed weight vector 𝒘⊤ with the covariance matrix 𝚺 , then multiplying the result by the weight vector 𝒘 again. This yields the scalar portfolio variance. The constraint imposed is that the sum of portfolio weights ∑ 𝑤𝑖 𝑛 𝑖=1 must equal 1, ensuring that all capital is fully invested. This optimization is presented in Equation 3.2.5. To construct the tangency portfolio, the Sharpe ratio is maximized. The numerator is the expected excess return, computed by subtracting the risk-free rate 𝑟𝑓 times a vector of ones 𝟏 from the expected return vector 𝛍, then multiplying the result by the transposed weight vector 𝒘⊤. The denominator is the standard deviation of the portfolio, calculated by taking the square root of 𝒘⊤𝚺𝒘 . The same full investment constraint 𝒘⊤𝟏 = 1 is applied. This formulation is shown in Equation 3.2.6. Equation 3.2.5 The minimum-variance portfolio minimizes the portfolio variance, defined as: σ𝑝 2   =  𝒘⊤ 𝚺 𝒘 The optimization problem is: min 𝑤 𝒘⊤𝚺𝒘 subject to ∑ 𝑤𝑖 𝑛 𝑖=1 = 1 The solution for the weights is: 𝑤min-var = 𝚺−𝟏𝟏 𝟏⊤𝚺−𝟏𝟏 9 Equation 3.2.6 The tangency portfolio maximizes the Sharpe ratio, defined as: S = μ𝑝 − 𝑟𝑓 σ𝑝 where: • 𝜇𝑝: expected portfolio return, 𝜇𝑝 = 𝒘⊤𝝁. • 𝑟𝑓: risk-free rate, using the 1-month Treasury Bill rate. • 𝜎𝑝: portfolio standard deviation, 𝜎𝑝 = √𝒘⊤Σ𝒘. The optimization problem is: max 𝑤 𝒘⊤(𝛍 − 𝑟𝑓𝟏) √𝒘⊤𝚺𝒘 subject to 𝒘⊤𝟏 = 1 The solution for the weights is: 𝑤tan = 𝚺−𝟏(𝛍 − 𝑟𝑓𝟏) 𝟏⊤𝚺−𝟏(𝛍 − 𝑟𝑓𝟏) To evaluate portfolio performance, three metrics are computed: the expected return, the standard deviation, and the Sharpe ratio, as shown in Equations 3.2.7. These measures allow for a clear comparison of how the risk-return profiles of the two groups evolved from the pre-pandemic to the post-pandemic period. Equations 3.2.7 Expected Return (𝜇𝑝) 𝜇𝑝 = 𝒘⊤𝛍 Standard Deviation (σ𝑝) σ𝑝 = √𝒘⊤𝚺𝒘 Sharpe Ratio (S) S = μ𝑝−𝑟𝑓 σ𝑝 10 3.2.2 Regression Model Specification To examine whether the asset-light strategy is associated with differences in systematic risk exposure before and during the COVID-19 pandemic, the regression analysis proceeds in three steps. First, firm-level market beta is estimated using the Fama-French five-factor model. Second, a composite Asset-Light Score is constructed to measure the extent to which each firm adopts an asset-light strategy. Third, two panel regression models are run to examine the relationship between the asset-light strategy and market exposure across two periods. The first step estimates firm-level market beta to serve as the dependent variable in the regression models. Market beta measures the sensitivity of a firm’s stock return to movements in the overall market and captures its exposure to systematic risk. To obtain annual beta estimates, the Fama-French five-factor model is applied to daily excess returns for each firm from 2017 to 2019 and 2020 to 2022, as shown in Equation 3.2.8. Excess return is defined as the stock return minus the risk-free rate. The regression includes five explanatory factors: the market premium, size (SMB), value (HML), profitability (RMW), and investment (CMA), where the coefficient on the market premium represents the firm’s market beta for that year. Equation 3.2.8 𝑅𝑖𝑡 − 𝑅𝑓𝑡 = α𝑖 + β𝑚(𝑅𝑚𝑡 − 𝑅𝑓𝑡) + β𝑠SMBt + βℎHMLt + β𝑟RMWt + β𝑐CMAt + ϵ𝑖𝑡 • 𝑅𝑖𝑡: return on firm i at time t • 𝑅𝑓𝑡: risk-free rate at time t • 𝑅𝑚𝑡: market return at time t • α𝑖: firm-specific intercept • β𝑚: loading on the market risk premium 11 • β𝑠: loading on SMB (Small Minus Big, size factor) • βℎ: loading on HML (High Minus Low, value factor) • β𝑟: loading on RMW (Robust Minus Weak, profitability factor) • β𝑐: loading on CMA (Conservative Minus Aggressive, investment factor) • ϵ𝑖𝑡: error term at time t The second step constructs the Asset-Light Score to serve as the key explanatory variable in the regression models. This composite score measures the extent to which each firm adopts an asset-light strategy. It draws on prior research that identifies key indicators of asset-light models, such as tangible asset intensity and fee-income ratio (Li and Singal, 2019; Sohn, Tang, and Jang, 2013). Two versions of the score are constructed. The first version is based on three asset-related ratios: Non-Owned Hotel Share, Light Depreciation Ratio, and Fixed Asset Turnover, as shown in Table 2. These ratios respectively reflect the share of hotels that a firm franchises or manages rather than owns, the extent to which it maintains a lower depreciation-to-asset ratio, and how efficiently it utilizes fixed assets to generate revenue. It primarily captures differences in hotel ownership structure. To construct the composite score, each of the three component ratios is first standardized by year. Let 𝑋𝑖𝑡 represent the value of a given ratio for firm 𝑖 in year 𝑡, 𝑋𝑡 denote the cross-sectional mean of that ratio in year t, and σ𝑡 denote the corresponding cross-sectional standard deviation. The standardized value, or z-score, is computed by subtracting the mean 𝑋𝑡 from the firm- level value 𝑋𝑖𝑡 , then dividing the result by 𝜎𝑡 , as shown in Equation 3.2.9. The resulting z-score expresses how far a given firm’s value deviates from the mean in units of standard deviation. Standardization is necessary because the original ratios differ in units. Some are expressed as percentages and others as multiples. Without standardization, variables with larger numeric ranges would disproportionately 12 influence the composite score. After standardizing all three ratios, the composite Asset-Light Score is obtained by taking the simple average of the z-scores for each firm in each year, as shown in Equation 3.2.10 and Table 3. The second version of the Asset-Light Score is constructed using Light PPE Ratio, Light Depreciation Ratio, and Fixed Asset Turnover, as shown in Equation 3.2.11 and Table 4. While it shares two components with the first version, it replaces Non-Owned Hotel Share with Light PPE Ratio, which captures the relative size of property, plant, and equipment holdings. This version emphasizes the firm’s balance sheet asset composition and overall physical asset intensity rather than the extent of direct hotel ownership. Table 2 Financial Ratios Used in Regression Analysis Variable Name Formula 1 Non-Owned Hotel Share 1 − 𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑂𝑤𝑛𝑒𝑑 𝐻𝑜𝑡𝑒𝑙𝑠 𝑇𝑜𝑡𝑎𝑙 𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝐻𝑜𝑡𝑒𝑙𝑠 2 Light PPE Ratio 1 − 𝑁𝑒𝑡 𝑃𝑃&𝐸 𝑇𝑜𝑡𝑎𝑙 𝐴𝑠𝑠𝑒𝑡𝑠 3 Light Depreciation Ratio 1 − 𝐷𝑒𝑝𝑟𝑒𝑐𝑖𝑎𝑡𝑖𝑜𝑛 𝐸𝑥𝑝𝑒𝑛𝑠𝑒 𝑇𝑜𝑡𝑎𝑙 𝐴𝑠𝑠𝑒𝑡𝑠 4 Fixed Asset Turnover 𝑇𝑜𝑡𝑎𝑙 𝑅𝑒𝑣𝑒𝑛𝑢𝑒 (𝐵𝑒𝑔𝑖𝑛𝑛𝑖𝑛𝑔 𝑁𝑒𝑡 𝑃𝑃&𝐸 + 𝐸𝑛𝑑𝑖𝑛𝑔 𝑁𝑒𝑡 𝑃𝑃&𝐸)/2 5 CapEx Ratio 𝐶𝑎𝑝𝑖𝑡𝑎𝑙 𝐸𝑥𝑝𝑒𝑛𝑑𝑖𝑡𝑢𝑟𝑒 𝑇𝑜𝑡𝑎𝑙 𝑅𝑒𝑣𝑒𝑛𝑢𝑒 6 Leverage Ratio 𝑇𝑜𝑡𝑎𝑙 𝐷𝑒𝑏𝑡 𝑇𝑜𝑡𝑎𝑙 𝐶𝑎𝑝𝑖𝑡𝑎𝑙 7 Return on Invested Capital 𝑁𝑒𝑡 𝑂𝑝𝑒𝑟𝑎𝑡𝑖𝑛𝑔 𝑃𝑟𝑜𝑓𝑖𝑡 𝐴𝑓𝑡𝑒𝑟 𝑇𝑎𝑥 𝐼𝑛𝑣𝑒𝑠𝑡𝑒𝑑 𝐶𝑎𝑝𝑖𝑡𝑎𝑙 Equation 3.2.9 z(𝑋𝑖𝑡) = 𝑋𝑖𝑡 − 𝑋𝑡 σ𝑡 13 Equation 3.2.10 Asset Light Score𝑖𝑡 1 = 1 3 [𝑧(Non Owned Hotel Share𝑖𝑡) + 𝑧(Light Depreciation Ratio𝑖𝑡) + 𝑧(Fixed Asset Turnover𝑖𝑡)] Equation 3.2.11 Asset Light Score𝑖𝑡 2 = 1 3 [𝑧(Light PPE Ratio𝑖𝑡) + 𝑧(Light Depreciation Ratio𝑖𝑡) + 𝑧(Fixed Asset Turnover𝑖𝑡)] Table 3 Components and Z-Scores for Asset Light Score 1 (Partial Sample) Asset Year Non-Owned Hotel Share Light Depreciation Ratio Fixed Asset Turnover Asset Light Score 1 Original Value Z-Score Original Value Z-Score Original Value Z-Score AC 2017 0.748 0.945 0.990 1.715 4.533 1.373 1.345 AC 2018 0.948 1.392 0.991 1.796 3.558 0.952 1.380 AC 2019 0.951 1.397 0.976 0.821 3.452 0.906 1.042 AC 2020 0.969 1.438 0.974 0.656 1.819 0.202 0.765 AC 2021 0.978 1.458 0.977 0.850 3.777 1.047 1.118 AC 2022 0.979 1.461 0.981 1.099 6.585 2.259 1.606 CHH 2017 1.000 1.507 0.993 1.969 11.244 4.270 2.582 CHH 2018 1.000 1.507 0.987 1.568 9.874 3.679 2.251 CHH 2019 0.999 1.506 0.986 1.501 4.432 1.329 1.445 CHH 2020 0.999 1.506 0.984 1.317 1.020 -0.143 0.893 CHH 2021 0.999 1.505 0.987 1.552 1.415 0.027 1.028 CHH 2022 0.999 1.505 0.986 1.440 1.490 0.060 1.001 H 2017 0.944 1.382 0.954 -0.709 0.650 -0.303 0.123 H 2018 0.954 1.405 0.957 -0.492 0.654 -0.301 0.204 H 2019 0.960 1.417 0.961 -0.240 0.677 -0.291 0.295 H 2020 0.962 1.422 0.966 0.110 0.207 -0.494 0.346 H 2021 0.968 1.435 0.975 0.749 0.419 -0.403 0.594 H 2022 0.974 1.448 0.965 0.066 1.079 -0.118 0.466 Table 4 Components and Z-Scores for Asset Light Score 2 (Partial Sample) Asset Year Light PPE Ratio Light Depreciation Ratio Fixed Asset Turnover Asset Light Score 2 Original Value Z-Score Original Value Z-Score Original Value Z-Score AC 2017 0.945 1.711 0.990 1.715 4.533 1.373 1.600 AC 2018 0.909 1.607 0.991 1.796 3.558 0.952 1.452 AC 2019 0.917 1.629 0.976 0.821 3.452 0.906 1.119 AC 2020 0.941 1.700 0.974 0.656 1.819 0.202 0.852 AC 2021 0.949 1.722 0.977 0.850 3.777 1.047 1.206 AC 2022 0.937 1.688 0.981 1.099 6.585 2.259 1.682 CHH 2017 0.916 1.628 0.993 1.969 11.244 4.270 2.622 CHH 2018 0.888 1.547 0.987 1.568 9.874 3.679 2.265 CHH 2019 0.729 1.091 0.986 1.501 4.432 1.329 1.307 CHH 2020 0.778 1.231 0.984 1.317 1.020 -0.143 0.802 CHH 2021 0.787 1.257 0.987 1.552 1.415 0.027 0.945 14 Asset Year Light PPE Ratio Light Depreciation Ratio Fixed Asset Turnover Asset Light Score 2 Original Value Z-Score Original Value Z-Score Original Value Z-Score CHH 2022 0.764 1.191 0.986 1.440 1.490 0.060 0.897 H 2017 0.467 0.340 0.954 -0.709 0.650 -0.303 -0.224 H 2018 0.528 0.514 0.957 -0.492 0.654 -0.301 -0.093 H 2019 0.531 0.522 0.961 -0.240 0.677 -0.291 -0.003 H 2020 0.606 0.737 0.966 0.110 0.207 -0.494 0.118 H 2021 0.739 1.118 0.975 0.749 0.419 -0.403 0.488 H 2022 0.775 1.223 0.965 0.066 1.079 -0.118 0.391 The third step employs panel regression models to examine whether asset-light strategies are associated with systematic differences in market exposure, as shown in Equation 3.2.12. In the first model, the dependent variable is the firm-level market beta β𝑖𝑡, as defined in the beta estimation procedure. The key explanatory variable is the asset-light score based on Non-Owned Hotel Share, Light Depreciation Ratio, and Fixed Asset Turnover, as introduced in the asset-light score construction process. A dummy variable for the COVID-19 period, which equals 1 during the COVID-19 period and 0 otherwise, is included to capture shifts in market exposure associated with the crisis. The interaction term between the asset-light score and the COVID-19 dummy tests whether the relationship between asset-light strategy and beta changes during the pandemic. Control variables 𝑋𝑖𝑡, including CapEx Ratio, Leverage Ratio, and Return on Invested Capital, account for differences in firm investment policy, capital structure, and operating performance. Firm fixed effects μ𝑖 are included to control for unobserved, time-invariant firm characteristics that may influence market beta. The error term ε𝑖𝑡 captures variation not explained by the model. The second model adopts the same structure but uses an alternative version of the asset-light score that emphasizes physical asset intensity, as shown in Equation 3.2.13. Specifically, Light PP&E Ratio replaces Non-Owned Hotel Share, shifting the focus from hotel ownership structure to balance sheet asset composition. All other elements of the regression specification remain unchanged. 15 Equation 3.2.12 Model 1 Specification 𝛽𝑖𝑡 = 𝛼𝑖 + 𝛾1 ⋅ 𝐴𝑠𝑠𝑒𝑡𝐿𝑖𝑔ℎ𝑡𝑆𝑐𝑜𝑟𝑒𝑖𝑡 1 + 𝛾2 ⋅ COVID − 19t + 𝛾3 ⋅ (𝐴𝑠𝑠𝑒𝑡𝐿𝑖𝑔ℎ𝑡𝑆𝑐𝑜𝑟𝑒𝑖𝑡 1 ⋅ COVID − 19t) + 𝛿′𝑋𝑖𝑡 + 𝜇𝑖 + 𝜀𝑖𝑡 • βit: firm-level beta for firm i at time t • AssetLightScoreit 1 : composite score based on Non-Owned Hotel Share, Light Depreciation Ratio, and Fixed Asset Turnover • COVID − 19t: dummy variable indicating the COVID-19 period • Xit: control variables, including CapEx Ratio, Leverage Ratio, and ROIC • μi: firm fixed effects • εit: error term Equation 3.2.13 Model 2 Specification 𝛽𝑖𝑡 = 𝛼𝑖 + 𝛾1 ⋅ 𝐴𝑠𝑠𝑒𝑡𝐿𝑖𝑔ℎ𝑡𝑆𝑐𝑜𝑟𝑒𝑖𝑡 2 + 𝛾2 ⋅ COVID − 19t + 𝛾3 ⋅ (𝐴𝑠𝑠𝑒𝑡𝐿𝑖𝑔ℎ𝑡𝑆𝑐𝑜𝑟𝑒𝑖𝑡 2 ⋅ COVID − 19t) + 𝛿′𝑋𝑖𝑡 + 𝜇𝑖 + 𝜀𝑖𝑡 • βit: firm-level beta for firm i at time t • AssetLightScoreit 2 : composite score based on Light PP&E Ratio, Light Depreciation Ratio, and Fixed Asset Turnover • COVID − 19t: dummy variable indicating the COVID-19 period • Xit: control variables, including CapEx Ratio, Leverage Ratio and ROIC • μi: firm fixed effects • εit: error term 16 4 EMPIRICAL RESULTS 4.1 Stock Performance Patterns Across Two Periods Both hotel C-Corporations and REITs experienced declines during the COVID-19 pandemic, but their return patterns and recovery trajectories differed. This section provides a descriptive comparison of the risk-return profiles of hotel C-Corporations and hotel REITs across the two periods, offering preliminary insight into how their performance responded to the market disruptions brought by the COVID-19 pandemic. 4.1.1 Cumulative Return Trends The cumulative stock return patterns of hotel C-Corporations and REITs over the two periods reveal some differences in both performance levels and trajectory stability. During the 2017-2019, hotel C-Corporations demonstrated a consistent upward trend, with cumulative returns for most firms reaching between 0.5 and 1.0 by the end of the period, as seen in Figure 1. This indicates relatively strong performance across the group, with limited volatility. The shape and slope of the return curves suggest that investor expectations remained optimistic and stable throughout the expansionary phase, supported by favorable macroeconomic conditions and steady revenue growth. In contrast, hotel REITs recorded more modest and volatile returns in the same period. As seen in Figure 2, the majority of REITs fluctuated within a range of -0.2 to 0.2, with several firms experiencing sustained negative performance. This pattern reflects greater sensitivity to capital market conditions and potentially weaker growth prospects during the pre-pandemic. The wider dispersion in returns indicates less consistency in performance within this group, compared to the relatively cohesive trend observed among C-Corporations. 17 The divergence became more pronounced during the 2020-2022. Following the outbreak of COVID-19, both groups experienced sharp initial declines, yet their recovery paths differed. C-Corporations exhibited a moderate rebound, with cumulative returns stabilizing between -0.1 and 0.1 by the end of the period. Although these values remained below pre-pandemic levels, the recovery trajectories were more consistent and less dispersed. In contrast, hotel REITs remained in a more distressed range of -0.5 to 0, with limited upward momentum and persistent underperformance across the group. This sustained gap in cumulative returns highlights the contrasting return dynamics of the two groups under adverse conditions and provides a preliminary indication of their differential exposure to systemic shocks. Figure 1 Cumulative Stock Returns of Hotel C-Corporations 2017-2019 2020-2022 Figure 2 Cumulative Stock Returns of Hotel REITs 2017-2019 18 2020-2022 4.1.2 Return, Risk, and Sharpe Ratio Patterns Hotel C-Corporations and REITs were both exposed to same macroeconomic shocks during the observation period. As a result, they exhibit some shared performance characteristics, including a sharp decline in returns and a surge in risk around 2020. However, as seen in Figure 3, Figure 4, and Figure 5, differences emerge when examining return behavior, risk response, and risk-adjusted performance across the two groups. In terms of return trends, measured by the average daily return within each year, both C-Corporations and REITs experienced declines in 2020, followed by partial rebounds. C-Corporations displayed relatively more consistent recovery in 2021 and 2022. While return paths were not uniformly upward, most firms hovered around zero, with returns generally staying near the breakeven point and varying within a narrower band. In contrast, REITs exhibited more irregular and volatile return movements, with several firms continuing to report negative or near-zero returns throughout the post- pandemic period. These patterns suggest that although both sectors were affected by 19 the same external shock, the recovery in C-Corporations was generally more stable and less fragmented. With respect to risk, measured by the average daily standard deviation within each year, both groups experienced a pronounced increase in 2020, which aligns with heightened market-wide uncertainty. C-Corporations exhibited a relatively faster return to pre-pandemic volatility levels by 2022. In contrast, REITs showed persistently elevated volatility in 2021 and beyond, with less consistent convergence across firms. This sustained volatility implies greater uncertainty surrounding REIT earnings and valuations. Regarding risk-adjusted performance, measured by the Sharpe ratio, both groups experienced a decline during the pandemic. C-Corporations generally remained in positive territory, indicating that returns continued to offer some compensation for risk. REITs, however, experienced a more pronounced deterioration, with Sharpe ratios falling into negative territory for most firms. This outcome indicates that increased risk was not offset by sufficient returns. These differences underscore the greater vulnerability of REITs under adverse conditions and suggest that their structural characteristics may limit their ability to sustain performance when market conditions deteriorate. Table 5 Return, Risk, and Sharpe Ratios Asset Average Daily Return Average Daily Standard Deviation Sharpe Ratio Pre Post Pre Post Pre Post Hotel C-Corp AC 0.00043 0.00073 0.01827 0.03595 0.02015 0.01950 CHH 0.00094 0.00043 0.01218 0.02461 0.07188 0.01627 H 0.00076 0.00048 0.01283 0.03102 0.05397 0.01458 HLT 0.00104 0.00050 0.01337 0.02543 0.07296 0.01857 IHG 0.00064 0.00021 0.01143 0.02809 0.05067 0.00646 MAR 0.00096 0.00043 0.01397 0.03004 0.06400 0.01344 Hotel REITs 20 Asset Average Daily Return Average Daily Standard Deviation Sharpe Ratio Pre Post Pre Post Pre Post AHT -0.00085 -0.00018 0.02491 0.12730 -0.03673 -0.00159 APLE 0.00003 0.00064 0.00918 0.03433 -0.00332 0.01798 BHR -0.00008 0.00098 0.02067 0.06261 -0.00688 0.01526 DRH 0.00022 0.00071 0.01315 0.04785 0.01169 0.01423 HST 0.00026 0.00034 0.01327 0.03138 0.01451 0.00988 INN -0.00004 0.00022 0.01394 0.04153 -0.00740 0.00462 PEB 0.00016 -0.00018 0.01483 0.03833 0.00683 -0.00544 PK 0.00028 0.00001 0.01351 0.04420 0.01620 -0.00030 RHP 0.00070 0.00097 0.01358 0.04497 0.04686 0.02106 RLJ -0.00005 0.00021 0.01425 0.04254 -0.00813 0.00434 SHO 0.00016 -0.00001 0.01264 0.03017 0.00742 -0.00121 SVC 0.00001 -0.00015 0.01134 0.05295 -0.00461 -0.00339 XHR 0.00046 0.00025 0.01399 0.04138 0.02837 0.00535 Figure 3 Average Daily Return Trends by Year 21 Figure 4 Average Daily Standard Deviation Trends by Year Figure 5 Sharpe Ratio Comparison: Pre- vs. Post-Pandemic 22 4.1.3 Mean-Variance Frontier Dynamics The mean-variance frontiers for hotel C-Corporations and REITs are constructed based on two standard portfolio strategies: the minimum-variance portfolio, which minimizes total risk, and the tangency portfolio, which maximizes the Sharpe ratio, as seen in Table 6. From the pre-pandemic to the pandemic period, both frontiers shift downward and rightward, indicating that portfolio returns declined while risk levels increased. In addition, the frontier exhibits reduced curvature and a compressed structure, indicating increased asset co-movements and reduced diversification potential during the pandemic. As shown in Figure 9, higher asset correlations contributed to a general decline in risk-adjusted portfolio efficiency. For the hotel C-Corp portfolio, the frontier shifts in a consistent and structured manner, as seen in Figure 6. The tangency portfolio’s average daily return declines from 0.0010 to 0.0009, while its standard deviation rises from 0.0111 to 0.0293. The minimum-variance portfolio shows a similar trend, with the average daily return falling from 0.0007 to 0.0005 and standard deviation increasing from 0.0093 to 0.0217. The frontier also exhibits reduced curvature and a more compact distribution, reflecting a diminished ability to enhance returns through risk diversification. The REIT frontier exhibits a less uniform pattern, as seen in Figure 7 2 . The minimum-variance portfolio’s average daily return increases slightly from 0.0001 to 0.0003, while its standard deviation also rises from 0.0088 to 0.0238. The tangency portfolio shows extreme values in both periods, with an average daily return of 0.0231 and a standard deviation of 0.2217 in the pre-pandemic period, and 0.0102 and 0.1390, respectively, during the pandemic. These results are driven by highly concentrated 2 The REIT plots use different axis ranges from the C-Corp plots to better accommodate the distribution of most assets and enhance visual clarity. AHT (2020-2022) is excluded due to its exceptionally high risk level. 23 weights in a small number of volatile REITs, which disproportionately influenced the optimization due to outlier performance. As these tangency portfolios do not reflect the broader REIT group, they are excluded from direct performance comparison across firm types. While both groups exhibit reduced portfolio efficiency during the pandemic, C- Corporations maintain stronger performance in minimum-variance portfolios relative to REITs in both the pre-pandemic and pandemic periods, as shown in Figure 8. In the pre-pandemic period, the C-Corporation portfolio achieves a higher average daily return (0.0007 vs. 0.0001) and slightly higher standard deviation (0.0093 vs. 0.0088), resulting in a higher Sharpe ratio (0.0680 vs. 0.0041). During the pandemic period, it again shows a higher average daily return (0.0005 vs. 0.0003), lower standard deviation (0.0217 vs. 0.0238), and a higher Sharpe ratio (0.0216 vs. 0.0125). This pattern challenges the initial expectation that asset-heavy firms would exhibit stronger performance before the pandemic, and highlights the need to further examine the mechanisms underlying the consistent advantage of C-Corporations. Table 6 Summary of Portfolio-Level Risk-Return Metrics Asset Average Daily Return Average Daily Standard Deviation Sharpe Ratio Pre Post Pre Post Pre Post Hotel C-Corp Min-Var Portfolio 0.0007 0.0005 0.0093 0.0217 0.0680 0.0216 Tangency Portfolio 0.0010 0.0009 0.0111 0.0293 0.0810 0.0292 Hotel REITs Min-Var Portfolio 0.0001 0.0003 0.0088 0.0238 0.0041 0.0125 Tangency Portfolio 0.0231 0.0102 0.2217 0.1390 0.1037 0.0732 24 Figure 6 Mean-Variance Frontier of Hotel C-Corporations (Short Allowed) 2017-2019 2020-2022 25 Figure 7 Mean-Variance Frontier of Hotel REITs (Short Allowed) 2017-2019 2020-2022 Figure 8 Efficient Frontier Comparison of Hotel C-Corporations and REITs 26 Figure 9 Heatmap of Correlations 2017-2019 2020-2022 27 4.2 Regression Results on Market Exposure To evaluate whether the asset-light strategy influences market exposure, firm-level betas were estimated using the Fama-French five-factor model, as shown in Table 7. The estimated betas are plotted for the pre-pandemic and pandemic periods to allow visual inspection of changes across time. As shown in Figure 10, most hotel firms experienced an increase in beta during the pandemic, indicating heightened exposure to systematic risk. Among them, hotel REITs showed more pronounced increases, while changes among C-Corporation firms were generally milder. Two C-Corporation firms, Hilton and Marriott, even recorded slight declines in beta during this period. Table 7 Estimated Betas of Hotel C-Corporations and REITs 28 Figure 10 Beta Comparison: Pre- vs. Post-Pandemic To test whether asset-light strategies are associated with smaller increases in market beta during the pandemic, panel regression models were estimated using firm- level betas as the dependent variable. Regression results are presented in Table 8 for Model 1, which defines the asset-light score based on Non-Owned Hotel Share, Light Depreciation Ratio, and Fixed Asset Turnover. Table 9 reports results for Model 2, where the score is constructed using Light PPE Ratio, Light Depreciation Ratio, and Fixed Asset Turnover. 29 Table 8 Panel Regression Results, Model 1 Table 9 Panel Regression Results, Model 2 The coefficient on the Asset-Light Score reflects its association with firm-level beta during the pre-pandemic period, as the pandemic dummy equals zero in this 30 specification. In the first regression, a one-unit increase in the asset-light score is associated with an increase in beta of approximately 0.160. The positive sign suggests that asset-light firms exhibit slightly higher exposure to systematic risk before the pandemic, potentially allowing them to capture a greater share of market gains under stable conditions. However, this result is not statistically significant. The COVID-19 period dummy captures the average change in beta associated with the pandemic period, holding other factors constant. The estimated coefficient is positive and statistically significant at the 1% level across both models, with values around 0.412 and 0.414. This indicates that, relative to the pre-pandemic period, firms experienced a substantial increase in exposure to systematic risk during the pandemic. The interaction term between the Asset-Light Score and the COVID-19 dummy captures how the relationship between the asset-light strategy and market beta changes during the pandemic. The coefficient is negative and statistically significant in both regressions, with values of -0.171 and -0.165. These results indicate that firms with stronger asset-light characteristics experienced smaller increases in beta during the pandemic. Specifically, for each one-unit increase in the asset-light score, the rise in beta during the pandemic is reduced by approximately 0.171 and 0.165, relative to firms following asset-heavy strategies. To complement the regression results, Figure 11 illustrates the relationships between firm-level beta and four asset-light indicators: the non-owned hotels share, the light PPE ratio, the light depreciation ratio, and fixed asset turnover. For each indicator, beta values are plotted against firm-level measures, with separate linear trend lines fitted for the pre-pandemic and post-pandemic periods. In the post- pandemic period, all four plots display a negative slope, indicating that firms with more asset-light characteristics such as higher non-ownership share, lighter fixed assets, lower depreciation intensity, or higher asset turnover tend to exhibit lower 31 market betas. In contrast, the pre-pandemic fitted lines mostly exhibit a positive slope, although the corresponding regression results from this earlier period are not statistically significant. These plots provide additional context and suggest that the asset-light strategy may have been associated with lower market risk, particularly during periods of heightened uncertainty. Figure 11 Relationships Between Beta and Individual Asset-Light Variables 32 5 CONCLUSION Our findings show that asset-light hotel firms consistently outperform asset-heavy peers in terms of stock returns and exhibit lower return volatility across both the pre- pandemic and pandemic periods. In the pre-pandemic period, the relationship between the asset-light strategy and market beta is positive but not statistically significant, indicating no clear association with systematic risk under stable conditions. During the pandemic, both groups experienced an increase in beta, but the rise was significantly smaller for asset-light firms, suggesting a smaller increase in market sensitivity relative to asset-heavy firms. This contrast in beta behavior indicates that the asset- light strategy provides greater downside protection during the pandemic. Considering both the stronger risk-return profile and reduced sensitivity to market shocks, the results support the conclusion that the asset-light strategy helped hotel stocks stay resilient during COVID-19. 33 REFERENCES García-Gómez, C.-D., Demir, E., Esteban, J., & Bilan, Y. (2021). The impact of COVID-19 outbreak on hotels’ value compared to previous diseases: The role of ALFO strategy. Heliyon, 7, e07836. https://doi.org/10.1016/j.heliyon.2021.e07836 Kim, J., & Jang, S. (2012). Comparative analyses of hotel REITs: Examining risk- return and performance characteristics. International Journal of Contemporary Hospitality Management, 24, 594–613. https://doi.org/10.1108/09596111211226842 Kim, H., Gu, Z., & Mattila, A. S. (2002). Hotel real estate investment trusts’ risk features and beta determinants. Journal of Hospitality & Tourism Research, 26(2), 138–154. Kim, S., Noh, S., & Lee, S. K. (2018). Asset-light strategy and real estate risk of lodging C-corps and REITs. International Journal of Hospitality Management, 78. https://doi.org/10.1016/j.ijhm.2018.09.004 Lee, S. K. (2012). The real estate risk of hospitality firms: Examining stock-return sensitivity to property values. International Journal of Hospitality Management, 31. https://doi.org/10.1016/j.ijhm.2011.09.005 Li, Y., & Singal, M. (2019). Capital structure in the hospitality industry: The role of the asset-light and fee-oriented strategy. Tourism Management, 70, 124–133. https://doi.org/10.1016/j.tourman.2018.08.004 34 Low, W., Das, P., & Piffaretti, C. (2015). The role of hotels in mixed asset portfolios: Revisiting the asset-heavy versus asset-light debate. International Journal of the Built Environment and Asset Management, 1, 273. https://doi.org/10.1504/IJBEAM.2015.075061 Pal, D. (2021). Does hospitality industry stock volatility react asymmetrically to health and economic crises? Economic Modelling, 108, 105739. https://doi.org/10.1016/j.econmod.2021.105739 Sohn, J., Tang, C.-H., & Jang, S. (2013). Does the asset-light and fee-oriented strategy create value? International Journal of Hospitality Management, 32, 270–277. https://doi.org/10.1016/j.ijhm.2012.07.004 Sohn, J., Tang, C.-H., & Jang, S. (2014). Asymmetric impacts of the asset-light and fee-oriented strategy: The business cycle matters! International Journal of Hospitality Management, 40, 100–108. https://doi.org/10.1016/j.ijhm.2014.03.007 35 APPENDIX Appendix A. Average Daily Return and Standard Deviation by Firm and Year Hotel C-Corps Year Average Daily Return Average Daily Standard Deviation Hotel REITs Year Average Daily Return Average Daily Standard Deviation AC 2017 0.00030 0.01521 AHT 2017 -0.00003 0.02161 AC 2018 0.00037 0.02090 AHT 2018 -0.00149 0.02209 AC 2019 0.00062 0.01833 AHT 2019 -0.00060 0.03071 AC 2020 0.00094 0.05376 AHT 2020 0.00277 0.20242 AC 2021 0.00107 0.02246 AHT 2021 -0.00149 0.07001 AC 2022 0.00016 0.02191 AHT 2022 -0.00183 0.05148 CHH 2017 0.00142 0.01201 APLE 2017 0.00019 0.00811 CHH 2018 -0.00019 0.01334 APLE 2018 -0.00092 0.00993 CHH 2019 0.00158 0.01105 APLE 2019 0.00083 0.00936 CHH 2020 0.00070 0.03358 APLE 2020 0.00057 0.05245 CHH 2021 0.00163 0.01510 APLE 2021 0.00108 0.01925 CHH 2022 -0.00106 0.02141 APLE 2022 0.00028 0.02026 H 2017 0.00120 0.01127 BHR 2017 -0.00087 0.02086 H 2018 -0.00020 0.01453 BHR 2018 0.00010 0.01953 H 2019 0.00126 0.01249 BHR 2019 0.00053 0.02163 H 2020 0.00013 0.04188 BHR 2020 0.00216 0.09718 H 2021 0.00125 0.02162 BHR 2021 0.00098 0.03360 H 2022 0.00005 0.02582 BHR 2022 -0.00021 0.03422 HLT 2017 0.00152 0.00998 DRH 2017 0.00018 0.01305 HLT 2018 -0.00027 0.01556 DRH 2018 -0.00064 0.01395 HLT 2019 0.00187 0.01391 DRH 2019 0.00111 0.01240 HLT 2020 0.00054 0.03235 DRH 2020 0.00140 0.07348 HLT 2021 0.00152 0.01877 DRH 2021 0.00097 0.02691 HLT 2022 -0.00057 0.02325 DRH 2022 -0.00025 0.02713 IHG 2017 0.00138 0.00925 HST 2017 0.00047 0.01281 IHG 2018 -0.00044 0.01334 HST 2018 -0.00041 0.01455 IHG 2019 0.00099 0.01129 HST 2019 0.00071 0.01235 IHG 2020 0.00057 0.03952 HST 2020 0.00005 0.04355 IHG 2021 0.00015 0.01734 HST 2021 0.00090 0.02080 IHG 2022 -0.00011 0.02249 HST 2022 0.00006 0.02501 MAR 2017 0.00208 0.01061 INN 2017 0.00006 0.01364 MAR 2018 -0.00071 0.01633 INN 2018 -0.00146 0.01440 MAR 2019 0.00150 0.01427 INN 2019 0.00128 0.01369 MAR 2020 0.00034 0.04202 INN 2020 0.00088 0.06271 MAR 2021 0.00109 0.01981 INN 2021 0.00059 0.02353 MAR 2022 -0.00014 0.02345 INN 2022 -0.00083 0.02613 PEB 2017 0.00116 0.01360 PEB 2018 -0.00079 0.01584 PEB 2019 0.00012 0.01497 PEB 2020 0.00017 0.05582 PEB 2021 0.00097 0.02338 PEB 2022 -0.00170 0.02722 PK 2017 0.00041 0.01296 PK 2018 0.00008 0.01378 PK 2019 0.00036 0.01383 PK 2020 0.00069 0.06538 PK 2021 0.00074 0.02683 PK 2022 -0.00140 0.02937 RHP 2017 0.00062 0.01089 RHP 2018 0.00015 0.01513 RHP 2019 0.00132 0.01437 RHP 2020 0.00152 0.06913 RHP 2021 0.00153 0.02515 RHP 2022 -0.00014 0.02546 RLJ 2017 -0.00008 0.01465 RLJ 2018 -0.00078 0.01528 RLJ 2019 0.00070 0.01273 RLJ 2020 0.00112 0.06418 RLJ 2021 0.00025 0.02488 RLJ 2022 -0.00075 0.02616 SHO 2017 0.00057 0.01185 SHO 2018 -0.00066 0.01382 SHO 2019 0.00056 0.01217 SHO 2020 0.00004 0.04101 SHO 2021 0.00035 0.02073 SHO 2022 -0.00043 0.02491 SVC 2017 0.00008 0.01040 SVC 2018 -0.00050 0.01249 SVC 2019 0.00045 0.01105 SVC 2020 -0.00001 0.07702 SVC 2021 -0.00060 0.03028 SVC 2022 0.00015 0.03953 XHR 2017 0.00071 0.01160 XHR 2018 -0.00058 0.01491 XHR 2019 0.00125 0.01516 XHR 2020 0.00062 0.06158 XHR 2021 0.00097 0.02363 XHR 2022 -0.00086 0.02796 36 Appendix B. Hotel Stock Risk-Return Metrics: Pre- vs. Post-Pandemic Comparison Asset Average Daily Return Average Daily Standard Deviation Sharpe Ratio Pre-Pandemic Post-Pandemic Pre-Pandemic Post-Pandemic Pre-Pandemic Post-Pandemic Hotel C-Corps AC 0.00043 0.00073 0.01827 0.03595 0.02015 0.01950 CHH 0.00094 0.00043 0.01218 0.02461 0.07188 0.01627 H 0.00076 0.00048 0.01283 0.03102 0.05397 0.01458 HLT 0.00104 0.00050 0.01337 0.02543 0.07296 0.01857 IHG 0.00064 0.00021 0.01143 0.02809 0.05067 0.00646 MAR 0.00096 0.00043 0.01397 0.03004 0.06400 0.01344 Hotel REITs AHT -0.00071 -0.00018 0.02513 0.12730 -0.03058 -0.00159 APLE 0.00003 0.00064 0.00918 0.03433 -0.00332 0.01798 BHR -0.00008 0.00098 0.02067 0.06261 -0.00688 0.01526 DRH 0.00022 0.00071 0.01315 0.04785 0.01169 0.01423 HST 0.00026 0.00034 0.01327 0.03138 0.01451 0.00988 INN -0.00004 0.00022 0.01394 0.04153 -0.00740 0.00462 PEB 0.00016 -0.00018 0.01483 0.03833 0.00683 -0.00544 PK 0.00028 0.00001 0.01351 0.04420 0.01620 -0.00030 RHP 0.00070 0.00097 0.01358 0.04497 0.04686 0.02106 RLJ -0.00005 0.00021 0.01425 0.04254 -0.00813 0.00434 SHO 0.00016 -0.00001 0.01264 0.03017 0.00742 -0.00121 SVC 0.00001 -0.00015 0.01134 0.05295 -0.00461 -0.00339 XHR 0.00046 0.00025 0.01399 0.04138 0.02837 0.00535 Appendix C. Asset-Light Score Construction Components Asset Year Non-Owned Hotel Share (Score 1 Component) Light PPE Ratio (Score 2 Component) Light Depreciation Ratio (Score 1 & 2 Component) Fixed Asset Turnover (Score 1 & 2 Component) Asset Light Score 1 Asset Light Score 2 Original Value Z-Score Original Value Z-Score Original Value Z-Score Original Value Z-Score Hotel C-Corps AC 2017 0.748 0.945 0.945 1.711 0.990 1.715 4.533 1.373 1.345 1.600 AC 2018 0.948 1.392 0.909 1.607 0.991 1.796 3.558 0.952 1.380 1.452 AC 2019 0.951 1.397 0.917 1.629 0.976 0.821 3.452 0.906 1.042 1.119 AC 2020 0.969 1.438 0.941 1.700 0.974 0.656 1.819 0.202 0.765 0.852 AC 2021 0.978 1.458 0.949 1.722 0.977 0.850 3.777 1.047 1.118 1.206 AC 2022 0.979 1.461 0.937 1.688 0.981 1.099 6.585 2.259 1.606 1.682 CHH 2017 1.000 1.507 0.916 1.628 0.993 1.969 11.244 4.270 2.582 2.622 CHH 2018 1.000 1.507 0.888 1.547 0.987 1.568 9.874 3.679 2.251 2.265 CHH 2019 0.999 1.506 0.729 1.091 0.986 1.501 4.432 1.329 1.445 1.307 CHH 2020 0.999 1.506 0.778 1.231 0.984 1.317 1.020 -0.143 0.893 0.802 CHH 2021 0.999 1.505 0.787 1.257 0.987 1.552 1.415 0.027 1.028 0.945 CHH 2022 0.999 1.505 0.764 1.191 0.986 1.440 1.490 0.060 1.001 0.897 H 2017 0.944 1.382 0.467 0.340 0.954 -0.709 0.650 -0.303 0.123 -0.224 H 2018 0.954 1.405 0.528 0.514 0.957 -0.492 0.654 -0.301 0.204 -0.093 H 2019 0.960 1.417 0.531 0.522 0.961 -0.240 0.677 -0.291 0.295 -0.003 H 2020 0.962 1.422 0.606 0.737 0.966 0.110 0.207 -0.494 0.346 0.118 H 2021 0.968 1.435 0.739 1.118 0.975 0.749 0.419 -0.403 0.594 0.488 H 2022 0.974 1.448 0.775 1.223 0.965 0.066 1.079 -0.118 0.466 0.391 HLT 2017 0.986 1.476 0.975 1.797 0.976 0.816 9.810 3.651 1.981 2.088 HLT 2018 0.988 1.479 0.974 1.793 0.977 0.843 10.189 3.814 2.046 2.150 HLT 2019 0.989 1.484 0.917 1.629 0.977 0.849 4.667 1.431 1.254 1.303 HLT 2020 0.991 1.486 0.933 1.677 0.980 1.079 1.353 0.000 0.855 0.919 HLT 2021 0.992 1.490 0.935 1.683 0.988 1.596 2.309 0.413 1.166 1.231 HLT 2022 0.993 1.491 0.939 1.694 0.990 1.715 3.850 1.078 1.428 1.496 IHG 2017 0.999 1.504 0.778 1.233 0.966 0.127 5.146 1.638 1.090 0.999 IHG 2018 0.996 1.498 0.817 1.342 0.973 0.596 4.160 1.212 1.102 1.050 IHG 2019 0.996 1.497 0.809 1.322 0.972 0.540 4.361 1.299 1.112 1.053 IHG 2020 0.996 1.499 0.900 1.581 0.978 0.938 2.697 0.580 1.006 1.033 IHG 2021 0.997 1.500 0.913 1.618 0.979 1.009 5.067 1.603 1.371 1.410 IHG 2022 0.997 1.501 0.896 1.571 0.984 1.327 9.179 3.379 2.069 2.092 MAR 2017 0.990 1.484 0.925 1.653 0.990 1.772 2.421 0.461 1.239 1.295 MAR 2018 0.991 1.487 0.917 1.631 0.990 1.776 2.782 0.617 1.293 1.342 MAR 2019 0.991 1.486 0.889 1.549 0.990 1.768 2.263 0.393 1.216 1.237 MAR 2020 0.991 1.488 0.908 1.605 0.991 1.792 0.838 -0.222 1.019 1.058 MAR 2021 0.992 1.489 0.900 1.580 0.991 1.840 1.414 0.027 1.119 1.149 MAR 2022 0.992 1.490 0.896 1.571 0.992 1.896 2.085 0.316 1.234 1.261 Hotel REITs AHT 2017 0.000 -0.719 0.136 -0.611 0.947 -1.178 0.350 -0.433 -0.776 -0.741 AHT 2018 0.000 -0.719 0.124 -0.645 0.945 -1.336 0.351 -0.432 -0.829 -0.805 AHT 2019 0.000 -0.719 0.114 -0.675 0.943 -1.485 0.363 -0.427 -0.877 -0.862 AHT 2020 0.000 -0.719 0.070 -0.799 0.932 -2.191 0.133 -0.526 -1.145 -1.172 AHT 2021 0.000 -0.719 0.201 -0.425 0.947 -1.216 0.239 -0.481 -0.805 -0.707 AHT 2022 0.000 -0.719 0.193 -0.448 0.948 -1.088 0.385 -0.417 -0.741 -0.651 APLE 2017 0.000 -0.719 0.022 -0.937 0.964 -0.029 0.258 -0.472 -0.407 -0.480 APLE 2018 0.000 -0.719 0.023 -0.936 0.963 -0.113 0.264 -0.469 -0.434 -0.506 APLE 2019 0.000 -0.719 0.018 -0.950 0.961 -0.240 0.262 -0.471 -0.477 -0.554 APLE 2020 0.000 -0.719 0.014 -0.960 0.959 -0.395 0.125 -0.530 -0.548 -0.628 APLE 2021 0.000 -0.719 0.018 -0.949 0.961 -0.200 0.197 -0.498 -0.472 -0.549 APLE 2022 0.000 -0.719 0.028 -0.920 0.962 -0.170 0.265 -0.469 -0.453 -0.520 BHR 2017 0.000 -0.719 0.195 -0.441 0.963 -0.077 0.380 -0.420 -0.405 -0.313 37 Asset Year Non-Owned Hotel Share (Score 1 Component) Light PPE Ratio (Score 2 Component) Light Depreciation Ratio (Score 1 & 2 Component) Fixed Asset Turnover (Score 1 & 2 Component) Asset Light Score 1 Asset Light Score 2 Original Value Z-Score Original Value Z-Score Original Value Z-Score Original Value Z-Score BHR 2018 0.000 -0.719 0.206 -0.411 0.965 0.035 0.337 -0.438 -0.374 -0.271 BHR 2019 0.000 -0.719 0.111 -0.683 0.960 -0.292 0.327 -0.443 -0.485 -0.473 BHR 2020 0.000 -0.719 0.100 -0.713 0.956 -0.563 0.145 -0.521 -0.601 -0.599 BHR 2021 0.000 -0.719 0.188 -0.461 0.961 -0.251 0.282 -0.462 -0.477 -0.391 BHR 2022 0.000 -0.719 0.181 -0.482 0.967 0.204 0.384 -0.418 -0.311 -0.232 DRH 2017 0.000 -0.719 0.132 -0.623 0.968 0.247 0.326 -0.443 -0.305 -0.273 DRH 2018 0.000 -0.719 0.079 -0.774 0.967 0.197 0.306 -0.451 -0.324 -0.343 DRH 2019 0.000 -0.719 0.088 -0.749 0.966 0.075 0.309 -0.450 -0.365 -0.375 DRH 2020 0.000 -0.719 0.074 -0.789 0.964 -0.060 0.099 -0.541 -0.440 -0.463 DRH 2021 0.000 -0.719 0.072 -0.794 0.965 0.058 0.200 -0.497 -0.386 -0.411 DRH 2022 0.000 -0.719 0.112 -0.679 0.966 0.112 0.358 -0.429 -0.345 -0.332 HST 2017 0.183 -0.312 0.171 -0.510 0.939 -1.704 0.546 -0.348 -0.788 -0.854 HST 2018 0.097 -0.503 0.193 -0.448 0.943 -1.433 0.571 -0.337 -0.758 -0.739 HST 2019 0.111 -0.471 0.166 -0.526 0.946 -1.244 0.548 -0.347 -0.687 -0.705 HST 2020 0.112 -0.469 0.223 -0.361 0.948 -1.093 0.157 -0.516 -0.692 -0.656 HST 2021 0.111 -0.471 0.146 -0.581 0.946 -1.274 0.284 -0.461 -0.735 -0.772 HST 2022 0.228 -0.212 0.160 -0.541 0.946 -1.265 0.471 -0.380 -0.619 -0.729 INN 2017 0.000 -0.719 0.067 -0.808 0.961 -0.226 0.286 -0.460 -0.468 -0.498 INN 2018 0.000 -0.719 0.080 -0.772 0.955 -0.674 0.276 -0.464 -0.619 -0.637 INN 2019 0.000 -0.719 0.065 -0.815 0.958 -0.453 0.259 -0.472 -0.548 -0.580 INN 2020 0.000 -0.719 0.049 -0.860 0.951 -0.922 0.108 -0.537 -0.726 -0.773 INN 2021 0.000 -0.719 0.069 -0.802 0.953 -0.765 0.171 -0.510 -0.664 -0.692 INN 2022 0.000 -0.719 0.060 -0.830 0.950 -0.963 0.273 -0.466 -0.716 -0.753 PEB 2017 0.000 -0.719 0.052 -0.852 0.961 -0.267 0.300 -0.454 -0.480 -0.524 PEB 2018 0.000 -0.719 0.064 -0.818 0.984 1.367 0.184 -0.504 0.048 0.015 PEB 2019 0.000 -0.719 0.026 -0.928 0.964 -0.039 0.251 -0.475 -0.411 -0.481 PEB 2020 0.000 -0.719 0.032 -0.909 0.963 -0.094 0.073 -0.552 -0.455 -0.519 PEB 2021 0.000 -0.719 0.029 -0.918 0.964 -0.017 0.123 -0.531 -0.422 -0.488 PEB 2022 0.000 -0.719 0.042 -0.880 0.961 -0.238 0.233 -0.483 -0.480 -0.534 PK 2017 0.134 -0.420 0.144 -0.587 0.970 0.404 0.336 -0.439 -0.151 -0.207 PK 2018 0.151 -0.383 0.148 -0.576 0.970 0.409 0.335 -0.439 -0.138 -0.202 PK 2019 0.113 -0.467 0.128 -0.633 0.977 0.832 0.318 -0.446 -0.027 -0.083 PK 2020 0.117 -0.459 0.110 -0.685 0.972 0.507 0.086 -0.546 -0.166 -0.242 PK 2021 0.111 -0.471 0.105 -0.700 0.971 0.459 0.149 -0.519 -0.177 -0.253 PK 2022 0.087 -0.525 0.125 -0.642 0.972 0.541 0.292 -0.458 -0.147 -0.186 RHP 2017 0.000 -0.719 0.182 -0.480 0.956 -0.599 0.581 -0.333 -0.550 -0.471 RHP 2018 0.000 -0.719 0.183 -0.476 0.969 0.287 0.487 -0.374 -0.268 -0.188 RHP 2019 0.000 -0.719 0.234 -0.329 0.948 -1.142 0.511 -0.363 -0.741 -0.611 RHP 2020 0.000 -0.719 0.124 -0.647 0.940 -1.699 0.166 -0.512 -0.977 -0.953 RHP 2021 0.000 -0.719 0.150 -0.569 0.938 -1.772 0.302 -0.453 -0.981 -0.931 RHP 2022 0.000 -0.719 0.210 -0.397 0.948 -1.095 0.576 -0.335 -0.716 -0.609 RLJ 2017 0.006 -0.705 0.148 -0.578 0.972 0.549 0.296 -0.456 -0.204 -0.161 RLJ 2018 0.007 -0.704 0.104 -0.703 0.960 -0.320 0.316 -0.447 -0.490 -0.490 RLJ 2019 0.010 -0.697 0.187 -0.465 0.964 -0.040 0.309 -0.450 -0.396 -0.319 RLJ 2020 0.010 -0.697 0.176 -0.496 0.965 0.069 0.099 -0.541 -0.390 -0.323 RLJ 2021 0.020 -0.673 0.152 -0.564 0.964 -0.061 0.175 -0.508 -0.414 -0.378 RLJ 2022 0.010 -0.696 0.133 -0.620 0.963 -0.107 0.275 -0.465 -0.422 -0.397 SHO 2017 0.000 -0.719 0.204 -0.417 0.959 -0.378 0.385 -0.417 -0.505 -0.404 SHO 2018 0.000 -0.719 0.248 -0.290 0.963 -0.088 0.381 -0.419 -0.409 -0.266 SHO 2019 0.000 -0.719 0.246 -0.296 0.962 -0.145 0.375 -0.422 -0.429 -0.288 SHO 2020 0.000 -0.719 0.156 -0.554 0.954 -0.705 0.098 -0.541 -0.655 -0.600 SHO 2021 0.000 -0.719 0.096 -0.726 0.958 -0.460 0.193 -0.500 -0.560 -0.562 SHO 2022 0.000 -0.719 0.074 -0.790 0.959 -0.370 0.325 -0.443 -0.511 -0.534 SVC 2017 0.000 -0.719 0.071 -0.797 0.946 -1.262 0.338 -0.438 -0.806 -0.833 SVC 2018 0.000 -0.719 0.087 -0.750 0.944 -1.405 0.348 -0.433 -0.852 -0.863 SVC 2019 0.000 -0.719 0.035 -0.900 0.953 -0.809 0.303 -0.453 -0.660 -0.721 SVC 2020 0.000 -0.719 0.083 -0.763 0.943 -1.491 0.152 -0.518 -0.909 -0.924 SVC 2021 0.000 -0.719 0.222 -0.364 0.947 -1.195 0.198 -0.498 -0.804 -0.686 SVC 2022 0.000 -0.719 0.110 -0.686 0.946 -1.228 0.270 -0.467 -0.804 -0.794 XHR 2017 0.000 -0.719 0.136 -0.610 0.951 -0.923 0.368 -0.425 -0.689 -0.653 XHR 2018 0.000 -0.719 0.093 -0.734 0.950 -0.970 0.380 -0.419 -0.703 -0.708 XHR 2019 0.000 -0.719 0.103 -0.705 0.952 -0.817 0.396 -0.413 -0.649 -0.645 XHR 2020 0.000 -0.719 0.166 -0.525 0.952 -0.819 0.135 -0.526 -0.688 -0.623 XHR 2021 0.000 -0.719 0.223 -0.362 0.958 -0.432 0.248 -0.477 -0.543 -0.424 XHR 2022 0.000 -0.719 0.155 -0.555 0.957 -0.511 0.399 -0.411 -0.547 -0.493