Featured Article / Invited Peer-reviewed Article
Liquidity Traps in Hotel Revenue Management during the Pandemic and Actions Taken by Hoteliers
2024 Volume 44 Issue 2 Pages 129-137
Details
2024 Volume 44 Issue 2 Pages 129-137
During the COVID-19 pandemic period from March 2020 to September 2022 in Japan, the hotel revenue management mechanism through dynamic pricing for room revenue maximization did not work because of a lack of traditional room demand regardless of price level. This situation resembles a liquidity trap in monetary policy in the Keynesian economics world, where an interest rate drop does not stimulate investment volume. In this situation, hotel managers need to generate revenues with other methods and/or business models for survival rather than mere dynamic pricing. This research demonstrates that there was a liquidity trap in hotel revenue management where a drop in the room rate did not cultivate room demand during the pandemic in the case of Tokyo’s select-service hotel sector. Under such a liquidity trap, there were several ways for hotel managers to generate revenue by capturing nontraditional demands that emerged through the realignment of the management resources at hand.
2020年3月から2022年9月までの日本における新型コロナウィルスの流行期には，価格水準に関係なく従来の宿泊需要が不足したため，ダイナミック・プライシングによる客室売上最大化のためのホテルレベニューマネジメントのメカニズムが機能しなかった。このような状況は，ケインズ経済学の世界における金融政策上の流動性の罠に似ており，その状況下では金利がゼロに近づいて下がっても投資が喚起されない。かかる状況下，ホテル経営者は生き残るために，単なるダイナミック・プライシングではなく，他の方法やビジネスモデルで収益を上げる必要があった。本研究は，東京のビジネスホテルセクターのケースにおいて，客室料金引き下げがパンデミック時の客室需要を開拓しないという流動性の罠がホテルレベニューマネジメントにおいて存在したことを証明するものである。また，このような流動性の罠の下で，ホテルマネジャーが手持ちの経営資源を再配置することで，勃興する非伝統的な需要を取り込み，売上を生み出す方法がいくつかあったことも明らかになった。
Revenue management (“RM” in short) is one of the management methods used by hotel managers to maximize revenue per available room (“RevPAR” in short), mainly through dynamic pricing on the basis of the expected room demand volume. While RM might mean maximizing a whole hotel’s revenue (not only from the Rooms Department but also from the Food & Beverage/Banquet/Other Departments), this research focuses on room revenue.
Its basic assumption is that demand volume will increase as prices drop. Therefore, it is believed that hoteliers can increase occupancy rates (“OCCs” in short) by lowering room rates when OCCs are estimated to be low on their booking curve at hand. In such a world, OCC is low when the average daily room rate (“ADR” in short) is low. This positive correlation is expected to be observed in the market in a normal hotel market environment.
However, it seems that RM did not work during the COVID-19 pandemic phase (“Pandemic Phase” in short; the exact definition of the pandemic phase is described later). Even when hoteliers dropped ADR, room demands were not well stimulated. Consequently, there was no positive correlation between the ADR and OCC. This situation resembles a liquidity trap in monetary policy in the Keynesian economics world, where investment volume no longer increases when an interest rate decreases toward zero.
This research is designed to determine whether such a liquidity trap in hotel revenue management existed during the pandemic, while the pre- and postpandemic periods highlight the effectiveness of RM in the Tokyo select-service hotel market.
On a related note, how hoteliers survive under such a liquidity trap in hotel RM is another question to answer. Some hoteliers were creative in the roll out of new business models to capture emerging demands during the pandemic, if not the satisfactory level of revenue they wanted. This research also addresses these survival plans from a resource management perspective.
There is little research available regarding the impact of the pandemic on RM. In particular, among those in the Japanese hotel market, only one, authored by Aoki and Uetake (2021), was found in this research effort. However, it is based on interviews with hotel managers regarding changes in RM policy during the pandemic and does not address the actual relationship between ADR and OCC via statistical analysis. The lack of statistical analysis in this field is not surprising, as obtaining such data is extremely challenging for third-party researchers because of the confidentiality of managerial information.
However, the effectiveness of RM has been verified by mega-chain operators such as Marriott and Hyatt as well as large hotel consulting firms in the past. For instance, according to Montan et al. (2008), an exercise of RM can increase the profit margin by 2 to 7%. As there are a few cost increases associated with RM implementation other than the system subscription fee and the labor cost of a revenue manager to be in place, the revenue increase tends to flow to the bottom line well.
The introduction of an automatic RM system (which suggests ideal room rates and that are automatically reflected to the booking engine and online travel agents (“OTAs” in short)) requires the hotel to integrate room department systems such as a booking engine and a channel manager as well as a property management system that controls room inventory. Owing to the complexity of the interfaces among such systems and the typical cost-conscious attitude of hoteliers, most independent hotels and ryokans in Japan do not yet have an automatic RM system. Therefore, an RM implementation tends to be more hand-crafted rather than automatic. The status of the RM system introduction in Japan is discussed in Section III.2.
Given this situation, there is a fear of not finding a valid correlation between OCC and ADRs even during the pre- and postpandemic periods. Therefore, this research first needed to verify that the effectiveness of RM existed pre- and postpandemic.
To conduct a statistical analysis of the effectiveness of hotel RM, an analyzable database of OCC and ADR in a homogeneous hotel property pool covering the pre- and postpandemic periods must be obtained. The next section addresses the database selection. After a necessary data cleaning process, the data from each phase are analyzed with a regression model to determine the correlation between the OCC and ADR.
Another question to answer in this research is what hoteliers did to generate hotel revenue during the pandemic period when a liquidity trap of RM existed. This research lists several cases that are categorized by a combination of the resources that each hotelier has.
1. Database selectionThe most difficult part of this research is to obtain the data for statistical analysis. Most hotel operators do not share daily or monthly OCCs and ADRs with a third party, even for research purposes. In addition, even a general manager is often not authorized to disclose information to a third party. Furthermore, unlike in the U.S., where the American Hotel and Lodging Association represents the majority of hotel operators in the U.S., there are multiple Japanese hotel/ryokan operator organizations in Japan, and none of them represent the majority of the players. They do not collect monthly operational information for research publication purposes.
Another consideration point of database selection is the hotel type; a full-service hotel category has several issues for RM analysis, whereas the select-service hotel category has fewer concerns. Therefore, the select-service hotel category is desirable for this research. One of the concerns is that there might be inconsistency in service charge treatment for the ADR calculation. Most Japanese hotel companies do not adopt the uniform system of accounts for the lodging industry, which defines OCC and ADR. Thus, the definition of ADR at Hotel A may differ from that at Hotel B. In this sense, a select-service hotel category is better because it does not have a service charge.
There are several other reasons why it is better to have select-service hotels for this research:
(1) Compared with purely looking at the room rate, full-service hotels tend to have many businesses groups that have a longer lead time and unique reasons for selecting a hotel to stay in.
(2) In the case of resort hotels and ryokans, they often sell dinner packages. They may want to lower the room rate to sell more rooms for more dinner revenue.
(3) Similarly, MICE hotels (which focus on meeting, incentive, convention, and exhibition-related businesses) may want to lower the room rate to obtain more meeting rooms and food and beverage (“F&B” in short) businesses.
(4) Immediately after the pandemic, some full-service hotels were forced to limit the number of rooms sold due to a shortage of labor, especially in the housekeeping sector. This situation may distort the relationship between OCC and ADR in terms of RM outcome.
On the basis of the above discussion, it is common for researchers to use STR data in Japan, which provides customized monthly performance data by identifying duration and hotel grade categories such as luxury and midscale. STR is the largest hotel operational information research firm in the world, and it serves in Japan as well. However, their categorization is based on grade. To select the “select-service” category, it is likely that “mid-scale” and “economy” in their definitions would be suitable. However, some mid-scale hotels could be full-service, and some upper mid-scale hotels could be select-service. This problem can be managed by reviewing all the participant names in the STR data pool for manual screening.
Another concern is that they provide data after having weighted-average calculations by themselves for the confidentiality of individual hotel performance. The data pool may contain some outliers because of misinputs and/or no report submission, which should be eliminated from the statistical analysis. In addition, Japanese operators’ participation rate is generally lower than that of international chains, which cannot rely on information exchanged among local hotels because they are concerned about potential breaches of anti-trust laws in their home country. Moreover, the majority of select-service hotels are operated by Japanese operators. Thus, the participant mix of the data pool available from STR may not be an ideal representation of the market.
While it is challenging to find a clear solution to these issues, one promising solution is to use the database of the Tokyo Hotel Association (Tokyo Hotel Kai in Japanese). It does not have a legal entity but rather a closed member-only association where some 250 member hotels exchange monthly ADR and OCC as a form of raw data. As its monthly report shows the actual ADR/OCC/RevPAR of individual member hotels, a hotel manager can analyze where his or her hotel stands in the market.
The Association’s members are mostly select-service hotels, and they do not have service charges. Additionally, they are mixed in terms of chains and independent hotels, and most of the members are domestic hotel operators. This characteristic seems to resemble that of Tokyo’s select-service hotel market itself. Therefore, it seems practically ideal to observe the relationship between ADR and OCC on the basis of these data.
Moreover, one of the drawbacks of using this database is the large amount of missing data and typos because of the nature of the Association, where each hotel manager voluntarily reports the data to the Association’s HQ, which cannot force the members to submit and/or verify the accuracy of their input. Thus, it is necessary to go through each datum to eliminate invalid and/or vacant input from the master pool.
On the basis of the above discussion, the database of the Tokyo Hotel Association is adopted for this research. Thanks to Mr. Hikoji Takabe, a representative of the Association, the data pool, which includes 261 member hotels’ monthly ADR and OCC from 1 January 2018 to 31 December 2023, is obtained for this research.
2. Preparation for analysisThere was monthly ADR/OCC information for 254 hotels with 47,979 rooms in the master file after the deletion of seven hotels without their room count information, which is required to calculate the key-count weighted average ADR and OCC. Another consideration point was that there were four full-service hotels with 6,017 rooms included in the data, which were also deleted from the data pool to have a pure select-service hotel pool.
Consequently, 250 hotels with 41,962 rooms remained before data cleaning.
The next step is data cleaning, which eliminates (i) data of hotels that were not opened as of January 2018 and (ii) data of hotels that did not properly report monthly ADRs/OCCs to the Association either on the basis of no reports or obvious typos throughout the examined period (from January 2018 to December 2023).
Notably, some hotels should have achieved 100% occupancy during the pandemic phase, as they were leased to municipality governments to accommodate mild case patients who were not able to be hospitalized owing to the capacity limitations of hospitals. However, such hotels were not found in the data pool. Lease-out to municipality governments is not regarded as a hotel operation in this research.
After the abovementioned data-cleaning exercise, 53 hotels with 8,588 keys remained in the database, with an average key count of 162 per property. While the volume of the analyzable data is only one-fifth that of the original database, it still maintains a good volume for statistical analysis. Among the 53 hotels, 45 belong to a hotel chain, which represents 84.9% of the data pool. The monthly ADR and OCC are calculated with a key count weight for the 53 hotels.
To understand the characteristics of Tokyo Hotel Association members, a quick survey was conducted on the status of the RM system introduction. In January 2024, the Association’s HQ asked the members to respond to the survey by e-mail, and 48 responded. Notably, these 48 hotels may not be the same as the 53 hotels used for the statistical analysis. The survey asked each manager—which category his or her hotel is in:
① A sophisticated RM system package has already been in place;
② RM is manually conducted mainly on the basis of demand forecasts;
③ RM is manually conducted mainly on the basis of competitors’ pricing; or
④ No real-time dynamic pricing is conducted.
Figure 1 shows the results of the survey.
RM System Adoption
Interestingly, most of the responders somewhat adopt RM, whereas the majority take a manual approach. On the basis of these survey results, database analysis is anticipated to reveal a strong correlation between ADR and OCC in a normal market environment (i.e., during the pre- and postpandemic periods).
3. Split of phasesThe next step is to split the 6-year-long data period into several phases. This research splits the period into the following four phases:
(1) Pre-Pandemic phase: From January 2018 to February 2020 (when the first infection case was reported in Japan)
(2) Pandemic phase: From March 2020 to September 2021 (when the last state of emergency was declared by the Japanese central government)
(3) Recovery phase: From October 2021 to September 2022 (the government restarted issuing visas to inbound leisure travelers in October 2022)
(4) Post-Pandemic Phase: From October 2022 to December 2023
Figure 2 shows the ADR and OCC trends of the Tokyo Hotel Association in the four phases discussed above.
Tokyo Hotel Association ADR/OCC Trend
Figure 3 shows scatter plots of the monthly OCC/ADR during the prepandemic phase, when the Tokyo select-service hotel market experienced healthy growth in inbound leisure volume and, consequently, RevPAR. With 0.6610 R square, 0.0000 significance F, and 0.0000 P value, the ADR can predict the OCC. The RM is effective.
Pre-Pandemic Phase
Figure 4 shows scatter plots of the monthly OCC/ADR during the pandemic phase when the market OCC fell below 55% in most months. Notably, some hotels shut down the operation, and some were used as lodging facilities for mild-case patients. According to a presentation document prepared by A Card Hotel System (2022), there were some 90,000 rooms used for lodging facilities for mild-case patients at that time. This represents approximately 5% of the 1.76 million room stock in Japan estimated by the Ministry of Health, Labor and Welfare (2023). Therefore, such out-of-service room stocks should have pushed the market OCC up.
Pandemic phase
In any event, there is no correlation between the OCC and ADR. Under these circumstances, hoteliers were not able to gain OCC by dropping ADR. A liquidity trap in hotel RM is observed.
3. Recovery phaseFigure 5 shows scatter plots of the monthly OCC/ADR during the recovery phase, when the market lacked inbound guests but the market OCC was maintained above 60%. With 0.7214 R square, 0.0005 significance F, and 0.0005 P value, the ADR can predict the OCC. The effectiveness of the RM returns.
Recovery phase
Figure 6 shows scatter plots of the monthly OCC/ADR during the postpandemic phase when the market welcomes inbound guests and the market OCC is maintained above 80%. With an R squared value of 0.3619, a significant F value of 0.0177, and a P value of 0.0177, the ADR can predict the OCC.
Postpandemic phase
Notably, compared with the prepandemic and recovery phases, the coefficient of determination (R square) is lower. In addition, the intercept is higher, and the coefficients of ADR are lower. This could mean that the room rate is no longer a predictor of OCC as the market OCC approaches 90%.
While individual hotels can constitute more than 90% of OCC, a market OCC has difficulty doing so. Thus, hoteliers are gradually increasing their room rates with few losses on OCC during the postpandemic phase. This bullish attitude of hoteliers has been backed by the steady increase in inbound visitor arrivals to Japan reported by the Japan National Tourism Organization (2023, 2024). It is anticipated that the majority of such foreign travelers stay in Tokyo, so the ADR of the Association is now driven by the number of visitor arrivals in Japan. According to Tourism Statistics issued by the Japan Tourism Agency on 29 February 2024, the total number of accommodation guests in Japan in 2023 was 59,275,000, and the share of foreign visitors was 19.3%. While Japanese demand for stay increased by 10.2% compared with that in 2022, foreign visitors’ demand increased by 592.8%.
Figure 7 shows the relationship between Tokyo ADR and Japan Visitor Arrivals. With 0.8859 R Square, 0.0000 Significance F, and 0.000 P value, Japan Visitor Arrivals drive Tokyo ADR, and consequently, RevPAR, as the OCC does not change much at 88%. As the ongoing construction cost hike tends to prevent developers from developing new hotels, the pace of new supply in Japan is slowing. Consequently, the OCC will remain high, and the situation in which Japan-visitor arrivals drive Tokyo ADRs should continue as long as the inbound volume is also high.
Tokyo ADR vs. Japan Visitor Arrivals
A question remains unanswered: what did hoteliers do to earn revenues from their hotels during the pandemic phase in the Japanese hotel market? The room rate drop did not work because of the liquidity trap. As the travel demands for both business and leisure activities heavily decreased due to the pandemic, hotels needed to find other uses for their guestrooms and ancillary facilities than accommodation use for travelers.
However, there were steady room demands by essential workers for medical, transportation, and infrastructure maintenance-related work. They formed a base business for OCC building even during the pandemic phase, if not enough.
Before we list the examples of “other usages”, let us look into the breakdown of the resources a hotel has. A typical hotel has guestrooms, restaurants and their kitchens, ballrooms, human resources, guest information, etc., depending on the services provided. Table 1 shows the breakdown of resources a hotel uses for each demand.
Resource realignment in hotel management during the pandemic
Typically, travelers use guestrooms, and local residents and workers use restaurants and ballrooms. However, this was not the case during the pandemic. Therefore, hoteliers needed to think about who else uses them for what. Some creative hoteliers found solutions. The examples are as follows.
1. Room-focused products(1) Serviced Apartment Package (for long stay)—This idea was not new. Normally, long-stay demands come with certain discounts on room rates, so hoteliers had been somewhat hesitant to take such discounts under the normal business environment. The key point of a service apartment package is that a staying guest will not be a traveler but rather a “resident”. Furthermore, some guests became curious about this package, given the low price tag. To some extent, this is a small win of dynamic pricing to stimulate demand during the pandemic phase. For example, Imperial Hotel Tokyo started selling the service apartment package of a 30-sqm studio room on 1 February 2021 at the price of JPY360,000 for a month, i.e., JPY12,000 per day (Imperial Hotel, 2021). This is less than half of the ADR the hotel previously achieved. Ninety-nine rooms for this package were sold out in a day (Mynavi News, 2021).
(2) Lodging Facilities for mild-case patients—An immediate need for municipality governments was to secure lodging facilities that can house mild-case patients for week-long cure stays so that hospitals can free up rooms to care for more severe-case patients. The Tokyo Metropolitan Government, for example, set up five facilities with 2,865 keys on 1 May 2020 (Mainichi Shimbun, 2020), three months after COVID-19 landed in Japan. The hotels who lost traveler accommodation demands seemed suitable candidates for this purpose. However, at the initial stage of the pandemic, many hoteliers were hesitant to surrender their buildings to local governments because they believed that (i) the pandemic may not last long, (ii) while the occupancy rate would be 100%, the room rate the governments offered may not be as high as the room rate they could earn from normal operations, and (iii) guests may have perceived that the virus still existed in the building even after sanitization and resumption of normal hotel operation. Notably, wages for hotel staff who cannot work at a mild-case patient-stay facility were not an issue, as the central government provided subsidies to such hotels and restaurants. The quick movers on this emerging demand during the pandemic were APA Hotels and Toyoko Inn, who were the top 2 largest select-service hotel chains in Japan (Travel News at, 2021). Both were private-held companies, so quick decisions were made by the founder’s family/management. Later, many hotels signed this type of contract with local governments, as it was proven that it was the best way to earn money during the pandemic. The “b” Ikebukuro was one of them. According to its owner, Japan Hotel REIT (2022, 2023), it was leased to the Tokyo Metropolitan Government from August 2020 to March 2023 at a rate of JPY 9,100 (JPY5,300 until March 2021).
2. Kitchen-focused products(1) Take-out Lunch Box—This is one of the easiest options every hotel could have taken. However, a lunch box sells at less than 1,000 yen, and taking into account the subsidy for layoff staff, many hoteliers opted not to offer such a service.
(2) Food Truck—Because of the abovementioned reasons, hotels did not hire a food truck to cater lunch boxes outside of their buildings. However, the Imperial Hotel Tokyo was different. It dispatched a food truck to the Tokyo International Forum, where it served contracted food services, which ceased operation during the pandemic (Imperial Hotel, 2022). This is to maintain brand awareness in the local community and keep chefs working to maintain their cooking skills, not necessarily for profit. The Imperial Hotel also took back the kitchen operation at the staff canteen from the third-party contracted food service provider during the pandemic so that its chefs could keep cooking.
3. HR-focused products(1) Temporary transfer to other companies in other industries—Although the central government offered subsidies to laid-off employees, many service companies affected by the pandemic transferred some of their staff to other less-affected industries for a certain period to keep their employees working and learning. Hoshino Resort, one of the largest ryokan/hotel operators in Japan, for example, transferred some staff to a farm that supplies vegetables to its ryokan (Japanese Inn.) (Kohsaka, 2021). While the labor cost at Hoshino might be higher than the wage that transferred staff can earn at the farm, they learned how to grow and harvest vegetables. After returning to their ryokans, they can vividly describe their experiences to the staying guests, which adds value to the guest experience. Therefore, this business model is viewed not for pure money earnings but for training and future marketing.
4. Cross-selling products(1) Super In-room Dining—This business model was, most likely, invented by Hotel New Otani Tokyo. During the pandemic, the government periodically declared a state of emergency. Once declared, restaurants were prohibited from offering alcoholic drinks. Consequently, all the high-end restaurants ceased operation. New Otani reported that there was room to serve in-room dining at a guestroom and that its sommeliers, who did not work at their restaurants, could come with the dishes to the guestroom for wine selection and serving (Hotel New Otani Tokyo, 2021). With this package, local residents, rather than travelers, can enjoy staying at a guestroom at New Otani for a fine dining experience with a personal-serviced sommelier. This package was called Super Room Service (room service means in-room dining in Japan). This package seemed to sell well, and New Otani repeatedly offered this package when a state of emergency was declared later. This business model uses not only guest rooms but also kitchen and restaurant staff (i.e., sommeliers) as well as guest databases for direct marketing, as shown in Table 1.
(2) Sport Live Viewing—As shown in Table 1, most alternative uses of hotel resources during the pandemic did not use ballrooms. This is because ballrooms are considered places in which the virus can cluster in a confined space, which may lead to the spread-out of the virus. However, according to the Japan Tourism Facilities Association (2020), Japan’s regulation of the ventilation capability of ballrooms at a qualified hotel as a specified building (Tokutei Kenchikubutsu), air replacement occurs frequently enough to prevent virus spread. Therefore, as long as a low density of people is maintained, a ballroom could be utilized even during the pandemic. Shinagawa Prince Hotel tried to make this “mission impossible” possible. It sold tickets for live viewing of the Seibu Lions game, a professional baseball team owned by Seibu Holdings, which also owns Prince Hotel Group (Prince Hotel, 2020). The seats were set at a low density in a large ballroom at Shinagawa Prince Hotel.
As discussed above, hoteliers were innovative in creating new business models through the realignment of their management resources during the pandemic. However, such businesses may not have been as lucrative as their normal hotel operations because of their low price and/or limited volume. Therefore, as the pandemic calmed down, such business models disappeared. They worked as, in a sense, a survival kit under the emergency caused by the pandemic.
Under the pandemic, the volume of demand will not increase even if the price decreases. This is because the demand curve itself shifts down so that a low price cannot help a hotel find more guests just because of the low price. Under this situation, hoteliers tried to earn revenue by realigning the resources they had.
In contrast, the demand volume of select-service hotels in single-bed rooms in Japan is known to decrease significantly once the room rate exceeds a threshold. This is because most Japanese corporate business travelers, who constitute the core guest category of such rooms, have a certain budget ceiling per company policy. In the case of Tokyo, it is at approximately JPY12,000, and in the case of Osaka, it is at approximately JPY10,000, for example. This is another thesis topic to explore.
To observe this “rate ceiling”, we need to segregate the room rate of single-bed rooms from the whole-room inventory. Such information is not available from the database of the Tokyo Hotel Association. If we can invite a major revenue management system vendor to contribute to the research, this rate ceiling research could be conducted. If such a rate ceiling exists, they need to consider nonprice competition even in a commodity market where differentiation is difficult.
The dataset generated and analyzed in this article is publicly unavailable because confidentiality is maintained at the data source.
Tomohiko Sawayanagi
Bachelor of Economics, Hitotsubashi University (1987). Master of Management in Hospitality, Cornell University (1998). After working at a bank and a security firm, joined Jones Lang LaSalle Hotels & Hospitality Group as the Department Head for its Japan entity (2000–2020). Currently, Representative Director, Brain Picks Incorporated, and Specially Appointed Professor, College of Tourism, Rikkyo University.
