Measuring the Effectiveness of Mutual Learning for Taiwan’s Tourist Hotels with the DEA Approach

Abstract

Taiwan’s international business and leisure hotels have created specific, divergent service operating systems to gain competitive advantage based on their distinctive target markets. For instance, the leisure properties focus on a hospitable, welcoming staff, while the business hotels aim for speedy service by employing relatively more workers. Given that hotel operators frequently benchmark direct competitors to improve on performance, researchers have suggested that hotel operators use a mutual learning strategy by benchmarking strategic techniques from hotels in disparate market segments. This study evaluates the effectiveness of the suggested mutual learning strategy between Taiwan’s business hotels and its leisure hotels, using the Different Systems model of data envelopment analysis (DEA) to examine potential improvements in efficiency. Empirical results show that few of Taiwan’s business hotels can gain efficiency through mutual learning from the leisure hotels—and some business hotels would actually lose efficiency if they adopted leisure properties’ operating practices. On the other hand, more than half of the leisure hotels in this sample would be able to achieve best practices from the mutual learning approach. On balance, both types of hotel can gain efficiency from the more common approach of benchmarking their direct competitors.

Keywords

data envelopment analysis efficiency Taiwan hotels mutual learning tourist hotels

The number of international tourist hotels in Taiwan has grown rapidly in recent years, in accord with government development policies. The total number of international hotels increased from 44 to 60 between 1985 and 2007, with an estimated 92 international tourist hotels in operation in 2011. Confronted with this competitive market, hotel operators have recognized the need to increase their performance, often by benchmarking competitors. The question this raises is which competitors to benchmark, along with a related question of whether learning from other hotels will increase efficiency.

To address these issues, we apply data envelopment analysis (DEA), a measure of operational efficiency, to hotels in Taiwan. DEA has become increasingly interesting and important in academic research studies (Tsaur 2000; Hwang and Chang 2003; Chiang, Tsai, and Wang 2004; Wang, Hung, and Shang 2006; Sun and Lu 2005; Yang and Lu 2006; Chen 2007; Hu et al. 2009; Hu et al. 2010). As we explain next, we particularly explore the extent to which Taiwan’s hotels can use a mechanism known as mutual learning to improve operational efficiency.

Taiwan’s hotel industry comprises two major types of hotel: leisure hotels, which are chiefly located in resort areas and target the leisure market, and business properties, which are generally located in metropolitan areas and offer accommodations to a target market of business travelers (Steadmon and Kasavana 1988; Hwang and Chang 2003). Compared to business hotels, leisure hotels provide greater general amenities, such as child care programs and in-room kitchenettes, and recreation facilities to guests. Thus, leisure hotels incur more expenses than business hotels do in terms of building maintenance and repair. Leisure hotels enjoy strong occupancy rates during weekends, unlike business properties, but leisure hotels face considerable seasonality and must readily deal with both off-peak and busy seasons. During the off-peak season, leisure operators need to come up with more advertising and promotional activities to attract guests, such as preferential off-peak season prices or collaborative marketing with theme parks located within the resort area. In contrast to leisure properties, business hotels are characterized by speedy service, business-specific facilities, and business-related information. To enable their speedy service, business hotels usually employ more attendants than do leisure hotels. They also perform customer relationship management to raise brand loyalty and to stabilize inflow of customers.

Given their different market environments and targets, Taiwan’s business and leisure hotels have developed differential competitive advantages. Instead of the normal approach of benchmarking similar hotels, researchers have suggested that the two types of hotel could learn from each other.

This concept of mutual learning (or cross-learning), which indicates the process of knowledge transfer among individuals, was proposed by Argyris and Schon (1978). In addition to individuals, cross-learning is also considered as one firm’s acquisition of another firm’s implicit knowledge through “observation, imitation, and practice” (Nonaka 1994). The act of groups of specialists transferring their specific knowledge, which encompasses different concepts, modes, and views among one another, is also identified as cross-learning (Schmick and Kieser 2008). Demeester and Qi (2005) proved that the cross-learning strategy can expedite the rate of product upgrades. Under an increasingly competitive environment, extending to competitor markets is essential to survival and sustainable development. To that end, cross-learning the operational technology of competitors is useful in improving the performance of leisure and business hotels (Hwang and Chang 2003; Sun and Lu 2005).

The main aim of this study is to examine the effectiveness of mutual learning and efficiency evaluation in Taiwan’s leisure and business hotels. We evaluated the operational efficiency of Taiwan’s international hotels, targeting the leisure and business market, using the Different Systems model of data envelopment analysis, which was proposed by Tone (1993). We have seen no research investigating learning effectiveness in the hotel industry. For empirical evaluation, the effectiveness of learning in inefficient hotels, both leisure and business types, is calculated based on the efficiency score as compared to the most efficient hotels (which is called the DEA frontier).

This study is organized as follows. First, previous studies on hotel efficiency are illustrated. Then, the Different Systems model and the calculations for effectiveness are introduced. Empirical results and analysis are described next. The conclusion and suggestions are offered in the final section.

Literature Review

A considerable number of studies apply efficiency measures to evaluate the performance of tourist hotels by simultaneously considering both multiple inputs consumed and multiple outputs generated. Some research studies (Anderson et al. 1999; Barros 2004; Chen 2007; Hu et al. 2010) have applied parametric efficiency measures, but DEA, a nonparametric efficiency measure, is frequently applied in investigations on tourist-hotel performance. Several tourist-hotel studies assume return-to-scales as constant when measuring operational efficiency (Morey and Dittman 1995; Johns, Howcroft, and Drake 1997; Tsaur 2000; Sanjeev 2007). In contrast, to explore causes of inefficiency, operational efficiency is decomposed into structural components of technical, pure, scale efficiency, and cost efficiency, as cited in some studies (Anderson, Fok, and Scott 2000; Chiang, Tsai, and Wang 2004; Barros 2005b; Hu et al. 2009). The nonradial DEA models based on slacks (or inefficiencies) of the input–output evaluation have been used by Sun and Lu (2005) and Yang and Lu (2006) for tourist hotel efficiency measures when simultaneously considering possible input decreases as well as output increases (Chen 2011). Hwang and Chang (2003), Barros and Alves (2004), and Barros (2005a) discussed the time-series performance of tourist hotels, in which efficiency change and technology change for different periods were evaluated using the Malmquist model. Meanwhile, Sigala et al. (2005) and Wang, Hung, and Shang (2006) used multiple DEA models to screen out external effects in efficiency and to purely compute managerial efficiency of tourist hotels. Other DEA studies have applied to Thailand (Lin 2011) and Malaysia (Kim 2011). Yu and Lee (2009) and Hsieh and Lin (2010) used a two-stage framework to measure the productive efficiency and service effectiveness of tourist hotels. Differences in efficiency between business and leisure hotels have been described in many tourist-hotel efficiency researches, such as Hwang and Chang (2003), Sun and Lu (2005), Wang, Hung, and Shang (2006), Yang and Lu (2006), and Chen (2007).

To provide for the learning targets for inefficient hotels, researchers have indicated references (i.e., the learning targets). Six business hotels and seven leisure hotels were suggested as learning targets in the evaluation by Sun and Lu (2005). Meanwhile, among the eleven learning targets identified by Hwang and Chang (2003), only two hotels were categorized as leisure type. However, the effectiveness of learning has not been specifically revealed in these studies.

Empirical Model and Data Resource

Different Systems Model

Farrell (1957) proposed the concept of the “frontier” to measure efficiency and productivity, where the frontier is the boundary created by graphing the most efficient operations, and the remaining operations are “enveloped” by that frontier. The first DEA model, by Charnes, Cooper, and Rhodes (CCR model 1978) is also based on the concept of the frontier and has been used to measure the relative efficiency of similar decision-making units (DMUs). Banker, Charnes, and Cooper (1984) advanced this CCR approach by assuming constant return-to-scales (the BCC model). Subsequent models have been based on these two models, such as super-efficiency (Andersen and Petersen 1993), two-stage (Seiford and Zhu 1999; Kao and Hwang 2008), undesirable factors (Seiford and Zhu 2002), and nonradial measure models (Ali, Lerme, and Seiford 1995; Thrall 1996).

Until Tone (1993) proposed the Different Systems model, the assumption in the models was that the technology set is essentially uniform. Cooper, Seiford, and Tone (2000) argued that one cannot reasonably assume a common production possibility set between DMUs that employ different kinds of technology. Considering the presence of technology indivisibility, Tone and Sahoo (2003) specified that the assumption of heterogeneous DMUs having an identical technology set holds when the output produced is very large and that the small production processes can be replicated. However, most real-life production processes fail to satisfy these stringent criteria. O’Donnell et al. (2008) offered the existence of gaps between the frontier technologies composed of heterogeneous and homogeneous firms.

Tone’s (1993) Different Systems model of DEA considers that the technology set and the frontier are composed of different technology types. The mathematical program of the different systems DEA model is formulated as follows

\begin{array}{l} \underset{θ, λ}{Min} : θ \\ s . t θ_{j} x_{e j} \geq \sum_{i \in A} x_{i j} λ_{i} + \sum_{i \in B} x_{i j} λ_{i}, j = 1, \dots, J \\ y_{e r} \leq \sum_{i \in A} y_{i r} λ_{i} + \sum_{i \in B} y_{i r} λ_{i}, r = 1, \dots, R \\ \sum_{i \in A} λ_{i} = p_{A}, \sum_{i \in B} λ_{i} = p_{B} \\ p_{A} + p_{B} = 1, λ_{i} \geq 0, i = 1, \dots, n \\ p_{A}, p_{B} = 0 or 1 . \end{array}

(1)

In formula (1), θ denotes the objective score; x_ij, j = 1…j, and y_ir, r = 1,…,R, denote the jth input and rth output, respectively, of DMU i; i = 1…,n; x_ej and y_er denote the jth input and rth output, respectively, of the DMU under evaluation; and λ_i represents the composed weight of benchmarks. (We demonstrate a solution to the Different Systems model in the appendix.)

Suppose that all DMUs are distinguished into two different technology types and the DMU under evaluation (labeled as DMU_e) is categorized into the A group. If the model assumes p_A = 1 and p_B = 0, then the score θ^Acan be computed by referring to the frontier of the A group. If the model assumes p_A = 0 and p_B = 1 then the score, θ^Bcan be computed by referring to the frontier of the B group. The outer boundary will then be chosen as the different systems frontier for DMU_e, and the efficiency score, θ*, can then obtained by using formula (2).

θ^{*} = MIN {θ^{A}, θ^{B}} .

(2)

Effectiveness of Learning from Different Systems

This study further computes the effectiveness of learning from different groups based on the scores calculated from the Different Systems model. The effectiveness indicators are illustrated as follows:

η^{A} = {\begin{cases} {θ^{B} + (1 - θ^{A})}_{{if}_{θ * = θ^{A}}^{θ * = θ^{B}}} \\ 1 \end{cases}

(3)

η^{B} = {\begin{cases} {θ^{A} + (1 - θ^{B})}_{{if}_{θ * = θ^{B}}^{θ * = θ^{A}}} \\ 1 \end{cases}

(4)

In formula (3), η^A represents the effectiveness indicator of learning from group A of the inefficient DMU_e, which presumably belongs to the A group. If the different systems frontier belongs to the other group (i.e., θ* = θ^B), the effectiveness indicator obtained is θ^A+(1-θ^B), with a value less than one. If the different systems frontier belongs to the “itself group” (i.e., θ* = θ^A), the effectiveness indicator will be equal to 1. In formula (4), η^A represents the effectiveness of learning from group B of the inefficient DMU_e, which belongs to the A group. If the different systems frontier belongs to the itself group (i.e., θ* = θ^A), the effectiveness indicator obtained is θ^B+(1−θ^A), with a value less than 1. If the different systems frontier belongs to the other group (i.e., θ* = θ^B), the effectiveness indicator is equal to one.

The interpretation of the effectiveness of learning in relation to different systems can be observed in the following graphic example. Exhibit 1 illustrates DMUs in two different technology groups, A and B, with two outputs. Technology A depicted as $T^{A} {(x, y) : \sum_{j \in A} x_{l j} λ_{j}^{A} \leq x_{l}, l = 1, 2, \sum_{j \in A} y_{j} λ_{j}^{A} \geq y, λ_{j}^{A} \geq 0}$ and technology B is $T^{B} {(x, y) : \sum_{j \in B} x_{l j} λ_{j}^{B} \leq x_{l}, l = 1, 2, \sum_{j \in B} y_{j} λ_{j}^{B} \geq y, λ_{j}^{B} \geq 0}$ . The efficiency frontiers of groups A and B are formed with the surface of technology sets A and B, which are the lines linking A₁-A₂-A₃-A₄ and B₁-B₂-B₃-B₄. According to the definition of Tone (1993), the different systems frontier is formed with the union of the frontiers of groups A and B, represented by the line linking B₁-B₂-T-A₃-A₄. (This is similar to the meta-frontier concept proposed by O’Donnell et al. 2008.) We can then assume that there are three inefficient units, DMU A₅, A₆, and A₇, in group A.

Exhibit 1:

The Frontier of Two Systems

The efficiency of DMU A₅ is θ* = θ^B. If DMU A₅ acquires learning from group B (point Q), then it will learn best practices and may become efficient, as depicted by a new position at the Different Systems efficiency frontier. This implies that the inefficiency, as represented by the gap between A₅ and Q, can be completely improved. If DMU A₅ acquires learning from the target of group A (as shown by point P), it only achieves the frontier of technology A. This implies that only the gap between A₅ and P (i.e., 1−θ^A) can be improved; the efficiency score is increased to θ^B +(1−θ^A). Therefore, we define η^A = θ^B +(1−θ^A) as the effectiveness of learning from group A, and η^B = 1 as the effectiveness of learning from group B.

Similarly, the efficiency score of DMU A₆ is obtained as θ* = θ^A. If DMU A₆ acquires learning from the target of group A (point S), it will learn best practices and may become efficient. If DMU A₆ obtains learning from group B (point R), only the gap between A₆ and R (i.e.,1−θ^B ) can be improved; the best efficiency after learning is θ^A+(1−θ^B) . The effectiveness of learning from group A is η^A=1 and from group B: is η^B = θ^A+(1−θ^B).

When the relatively inefficient unit is placed outside the frontier of the other group (i.e., it is already more efficient than the dissimilar hotels), the unit will be unable to learn from the opposite group. For example, the DMU A₇ is located outside the frontier of group B. If DMU A₇ acquires learning from the target of group A (point S), it will learn best practices and be efficient. If DMU A₇ obtains learning from group B (point R), on the other hand, its efficiency score will be reduced. In other words, the different systems frontier (the line linking B₁-B₂-T-A₃-A₄) represents the best practice for overall firms and can be measured by the surface of union sets T^A∪T^B. If DMU A₅ attempts to achieve best practice, it needs to learn from technology B and refer to the composition of the input and output of TB. Therefore, through mutual learning, DMU A₅ will understand what an optimum composition of factors is, which will help it to achieve best practice and become efficient.

Data Source

Empirical data that we analyze are drawn from annual statistics from 58 Taiwan international tourist hotels, obtained from the Operating Report of International Tourist Hotels in Taiwan 2008 published by the Taiwan Tourist Bureau in 2009. For the choice of inputs, we rely on previous studies evaluating Taiwan’s international tourist hotels that used operating expenses, rooms, catering space, and employees as inputs (e.g., Hsieh and Lin 2010; Yang and Lu 2006; Sun and Lu 2005). The occupancy rate and total revenues are identified as outputs, as used by Chen (2009) and Yang and Lu (2006). The descriptive statistics of the four inputs (labeled as x₁,x₂,x₃, andx₄) and two outputs (labeled as y₁ and y₂) are listed in Exhibit 2.

Exhibit 2:

Descriptive Statistics

	x ₁	x ₂	x ₃	x ₄	y ₁	y ₂
Mean	311.7	1,166.8	298.2	216.5	605.9	66.5
SD	207.0	924.0	149.4	227.0	600.0	13.7
Maximum	912.0	4,777.0	873.0	1,112.7	2,927.1	92.3
Minimum	32.0	92.0	50.0	6.3	39.3	27.2
Business type
Mean	364.1	1,307.2	339.5	269.9	773.0	73.1
SD	234.1	859.3	168.8	270.8	707.7	8.6
Leisure type
Mean	242.6	981.5	243.7	146.1	385.3	57.7
SD	141.3	990.1	97.9	124.3	313.1	14.5

The inputs are as follows:

Number of employees (x₁): number of individual employees involved in the operation of hotels, including guest rooms, catering, and management staff;

Area of catering space (x₂): refers to the square measurement of floor space utilized for providing food and beverage to room guests, banquets, and weddings (measured in square meters):

Number of rooms (x₃): number of rooms that are provided for rent to guests (without regard to size, room rate, or quality); and

Operating expenses (x₄): refers to the expenditures of hotel operations, including water and electricity fuel expenses, food and beverage costs, insurance premium, and maintenance and repair expenses (measured in million New Taiwan Dollars, NT$).

The outputs are:

Total revenues (y₁): refers to the operational revenues of hotels, including guest room revenues, food and beverage revenues, clothes washing revenue, store rent incomes, night club revenues, and service charges (measured in million NT$); and

Occupancy rate (y₂): refers to the ratio of the actual rooms rented and those available to be rented.

The descriptive statistics indicate that most of the business hotels are larger than leisure hotels, as shown by the fact that the average inputs and outputs of the business type are higher than the leisure type.

Empirical Evaluation

First we wanted to ensure that the data distinguished the leisure hotels from the business hotels, that is, whether different systems exist under the frontiers of business and leisure types. For this purpose, this study used the Mann-Whitney U test proposed by Brockett and Golany (1996). The null hypothesis for comparison of different groups is “the samples of different types come from identical populations.” Results show that the statistic of the empirical data set is significant at the 1.6-percent level (refer to Exhibit 3), indicating that business and leisure hotels present obvious differences in terms of individual frontiers, and using the Different Systems model to evaluate the empirical data set is justified.

Exhibit 3:

Mann-Whitney U Test Results for Different Systems

	Mann-Whitney U	Z	P
Value	602	−2.408	0.016

Note: When z ≤ − z_1−α/2 or z ≥ −z_1−α/2 at the significant α % level, the null hypothesis is rejected.

Results of business hotel efficiency scores and effectiveness of learning are reported in Exhibit 4. The first column shows the number code for each hotel. Scores were evaluated by referring to the itself group (business hotels) and the other group (leisure hotels), θ^itself and, θ^otheras shown in the second and third columns. For the different systems, the efficiency scores, θ*, obtained from θ* = Min {θ^itself, θ^other}, are shown in the fourth column. Twelve business hotels have performed efficiently according to these data, while 21 business hotels are operating less efficiently than their competitors. For five of the inefficient hotels (namely, B13, B15, B16, B17, and B20), the score referring to the other group (θ^other) could not be feasibly calculated, a result that indicates that these hotels are unable to find any learning target among any of the leisure hotels. This is shown in the “Effectiveness” column, which shows the effectiveness of learning from the other group. The effectiveness indicator of these five hotels is equal to the original efficiency score; the reason for this is that no reference could be found in the opposite group. In other words, the five business hotels cannot gain any effectiveness by learning from leisure hotels.

Exhibit 4:

Efficiency Analysis for the Business Type

				Effectiveness
No.	θ^itself	θ^other	θ*	Learning from the Other	Learning from Itself
B01	1	4.127	1
B02	1	Infeasible	1
B03	1	Infeasible	1
B04	1	Infeasible	1
B05	1	Infeasible	1
B06	1	Infeasible	1
B07	1	1.082	1
B08	1	Infeasible	1
B09	1	Infeasible	1
B10	1	1.516	1
B11	1	2.094	1
B12	1	Infeasible	1
B13	0.904	Infeasible	0.904	0.904	1
B14	0.875	0.960	0.875	0.915	1
B15	0.849	Infeasible	0.849	0.849	1
B16	0.821	Infeasible	0.821	0.821	1
B17	0.765	Infeasible	0.765	0.765	1
B18	0.753	1.015	0.753	0.738^a	1
B19	0.862	0.745	0.745	1	0.882
B20	0.744	Infeasible	0.744	0.744	1
B21	0.740	1.071	0.740	0.669^a	1
B22	0.731	1.052	0.731	0.679^a	1
B23	0.726	0.776	0.726	0.951	1
B24	0.715	0.963	0.715	0.752	1
B25	0.713	0.960	0.713	0.753	1
B26	0.712	1.185	0.712	0.527^a	1
B27	0.689	1.137	0.689	0.552^a	1
B28	0.621	0.953	0.621	0.669	1
B29	0.589	0.816	0.589	0.772	1
B30	0.801	0.578	0.578	1	0.777
B31	0.575	0.767	0.575	0.809	1
B32	0.550	0.789	0.550	0.760	1
B33	0.498	0.624	0.498	0.874	1

Represents the hotel having a regressive learning effectiveness.

In contrast, five other business hotels, B18, B21, B22, B26, and B27, would actually lose effectiveness if they benchmarked the leisure hotels. This is shown by an effectiveness-of-learning score that is lower than the original efficiency score. These business hotels, which are located outside the leisure frontier and surpass all leisure hotels in efficiency, would likely regress in performance if they “learn” from the leisure hotels.

The nine business hotels with effectiveness indicators larger than their original efficiency score but lower than 1 can gain some benefit from benchmarking the leisure hotels, but they will not gain full efficiency by learning from their leisure-type counterparts. The remaining two business hotels, B19 and B30, have an effectiveness indicator equal to 1; they can learn the best practices and may become efficient when they learn from the leisure hotels.

Thus, we conclude that aside from the two hotels that could successfully benchmark the leisure hotels (the two with an effectiveness indicator equal to 1), the other 19 inefficient business hotels would do best to benchmark their direct competitors and learn the best practices from the other efficient business hotels.

On balance, the results on the leisure hotels’ efficiency scores and effectiveness of learning (Exhibit 5) show considerable room for mutual learning. There are four efficient leisure hotels, and just two would regress if they attempted to benchmark business properties. The remaining 21 leisure properties are inefficient. No infeasible measure of the score, θ^other, existed among leisure hotels, which implies that each hotel can find at least a reference within the business type.

Exhibit 5:

Efficiency Analysis for the Leisure Type

				Effectiveness
No.	θ^itself	θ^other	θ*	Learning from the Other	Learning from Itself
L01	1	Infeasible	1
L02	1	1.258	1
L03	1	2.128	1
L04	1	2.016	1
L05	1	0.952	0.952	1	0.952
L06	1	0.910	0.910	1	0.910
L07	0.872	1.029	0.872	0.842^a	1
L08	1	0.867	0.867	1	0.867
L09	1	0.791	0.791	1	0.791
L10	0.775	1.097	0.775	0.679^a	1
L11	0.764	0.841	0.764	0.923	1
L12	0.802	0.756	0.756	1	0.954
L13	0.749	0.854	0.749	0.895	1
L14	0.734	0.727	0.727	1	0.993
L15	0.761	0.654	0.654	1	0.893
L16	0.616	0.881	0.616	0.736	1
L17	0.650	0.600	0.600	1	0.950
L18	0.596	0.625	0.596	0.971	1
L19	0.592	0.596	0.592	0.996	1
L20	0.583	0.641	0.583	0.943	1
L21	0.725	0.549	0.549	1	0.824
L22	0.585	0.548	0.548	1	0.963
L23	0.795	0.512	0.512	1	0.716
L24	0.508	0.604	0.508	0.904	1
L25	0.716	0.499	0.499	1	0.783

Represents the hotel having a regressive learning effectiveness.

The two hotels that would regress from “learning” from business hotels are L07 and L10, for which the effectiveness of learning from the business type is lower than the original efficiency score. There are seven leisure hotels with an effectiveness indicator larger than the original efficiency and lower than 1, implying that these operations would progress in efficiency but cannot gain full efficiency from business type learning. Nine leisure hotels may become efficient when they benchmark their direct competitors and learn from other leisure hotels (the itself group). Meanwhile, the 12 leisure hotels, which have an effectiveness indicator of learning from the other type equal to 1, may become efficient by learning from the business type.

Summarizing the empirical results above, only two inefficient business hotels can learn best practices from the leisure type and possibly become more efficient. The bulk of the business hotels would not really learn much from benchmarking the leisure-type properties. In particular, five of them would not be affected by the learning, and the other five could experience efficiency regression. On the other hand, by learning from the business type, nine inefficient leisure hotels may not learn the best practices, but the other 12 inefficient units may become efficient.

To identify the significance of the differences in efficiency scores and effectiveness indicators between business and leisure types, this study further compared the means using the Mann-Whitney U test, as shown in Exhibit 6. The average of (original) efficiency for business types, at 0.815, is not significantly greater than that of the leisure type, at 0.737. When learning from the itself group, the average of inefficient leisure hotels could increase to 0.984 and that of business hotels could rise to 0.933. When learning from the other group, however, the average of inefficient leisure hotels could increase to 0.947, but the average of business types will be significantly reduced to 0.786.

Exhibit 6:

Difference Verification of the Efficiency and Learning Effectiveness

	Hotel type	Mean	SD	Z statistic
Original efficiency	Business	0.815	0.166	−1.547
	Leisure	0.737	0.174	(0.122)
Effectiveness of learning from the itself group	Business	0.984	0.054	−3.833^a
	Leisure	0.933	0.088	(0.000)
Effectiveness of learning from the other group	Business	0.786	0.129	−2.968^a
	Leisure	0.947	0.092	(0.003)

Note: Values in parentheses represent the p value.

Denotes significance at the 0.01 level.

Results show that learning from different hotel types to improve efficiency is beneficial for inefficient leisure hotels, but not for relatively inefficient business hotels. The inefficient business and leisure hotels are both able to effectively increase their efficiency via learning from targets within their same hotel types.

Conclusion

Although Taiwan’s international business and leisure hotels target different consumer markets, they could potentially improve their operational efficiency by benchmarking operations at hotels that are not in their market segment (as well as developing their own competitive advantages based on their market environment). Several studies have suggested this as a strategy to improve efficiency. To measure the effectiveness of this concept of mutual learning, this study used the Different Systems DEA model to calculate the effectiveness indicators of mutual learning between the business and leisure hotels.

The empirical results indicate that most business hotels are unable to learn (or apply) the best practices from mutual learning. Indeed, five of the business hotels had mastered their technology to the extent that their efficiency regressed if they attempted to apply leisure hotels’ systems. On the other hand, more than half of leisure hotels could improve their operations by learning best practices from the most efficient business hotels and thus may become efficient through mutual learning. The study demonstrated the value of benchmarking direct competitors, as both business and leisure hotels experienced positive effects by learning from their own type of hotel. On balance, then, the efficiency of leisure-type hotels can increase from mutual learning, but the efficiency of business types will be reduced. Therefore, we posit that mutual learning is effective for leisure hotels but not for business hotels.

Our study did not indicate the specific reasons for this outcome, but we can infer the reasons from other research findings. Quality personal interaction, friendliness, and helpful reception are the critical components of a leisure hotel’s service (Parasuraman et al. 1988), while speedy service is the most important factor influencing hotel selection by all business travelers. To avoid service failures, such as slow check-in or check-out, slow service, and long wait for tables at breakfast, business hotels must employ a sufficient number of service employees (Lewis and McCann 2004). Business hotels that rely on a large staff for speed in service are unable to achieve positive results by referring to leisure hotels’ model of fewer, but more engaging, service employees. Demand-based variable pricing practices are frequently used in hotels’ yield management systems (Choi and Mattila 2004; Rao and Smith 2006), but the two types of hotel apply different yield management strategies. Price-sensitive leisure travelers are more concerned with room rates in their hotel selection than are business travelers (Lewis 1985). Leisure hotels also conduct promotions for price reductions to increase revenue with higher occupancy (Hwang and Chang 2003). Instead, business travelers are focused on hotels that provide rapid service and adequate business facilities, such as meeting rooms, secretarial services, and international direct dial service.

Of course, the two types of hotel guest do have certain features in common, and events have caused leisure hotels to look at business travelers as a potential market segment. Convenient communication and location are the important factors for both business and leisure travelers (Knutson 1988). The completion of Taiwan’s high-speed railway and the increase in expressway density have gradually removed geographical restrictions for suburban leisure hotels. Furthermore, the Taiwanese government has actively expanded the exhibition and convention industry. The development of the conference market has become a future target for leisure hotels to attract customers in the off-peak season (Hwang and Chang 2003). However, the services required by the conference market are similar to those of business hotels (e.g., business meetings, conference attendance, sales meetings, executive meetings, and training sessions) (Yavas and Babakus 2005). In addition, conference hotels must ensure easy access to information technology services, such as computers, wi-fi, and video phones. This is vastly different from merely offering comfortable overnight accommodations with gracious amenities to leisure travelers. These are some reasons that we believe leisure hotels will benefit from benchmarking business hotels.

A final note on potential future research. In addition to market conditions, the differences between chain and independent-operated hotels have been frequently discussed in previous studies. Evaluating the effect of mutual learning on various operating types is a potential topic for future research.

Footnotes

Appendix

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Ali

A. I.

Lerme

C. S.

Seiford

L. M.

1995. Components of efficiency evaluation in data envelopment analysis. European Journal of Operational Research 80:462-73.

Anderson

R. I.

Fok

Scott

2000. Hotel industry efficiency: An advanced linear programming examination. American Business Review 18 (1): 40-48.

Andersen

Petersen

N. C.

1993. A procedure for ranking efficient units in data envelopment analysis. Management Science 39:1261-64.

Anderson

R. I.

Xia

Fish

Mixhello

1999. Measuring efficiency in the hotel industry: A stochastic approach. International Journal of Hospitality Management 18 (1): 45-47.

Argyris

Schon

D. A.

1978. Organizational Learning: A theory of Action Perspective. Addison-Wesley, Reading, MA.

Banker

R. D.

Charnes

A. C.

Cooper

W. W.

1984. Some models for estimating technical and scale inefficiencies in data envelopment analysis. Management Science 30:1078-92.

Barros

C. P.

2004. A stochastic cost frontier in the Portuguese hotel industry. Tourism Economics 10:177-92.

Barros

C. P.

2005a. Evaluating the efficiency of a small hotel chain with a Malmquist productivity index. International Journal of Tourism Research 7:173-84.

Barros

C. P.

2005b. Measuring efficiency in the hotel sector. Annals of Tourism Research 32 (2): 456-77.

10.

Barros

C. P.

Alves

F. P.

2004. Productivity in the tourist industry. International Advances in Economic Research 10 (3): 215-25.

11.

Brockett

P. L.

Golany

1996. Using rank statistic for determining programmatic efficiency differences in data envelopment analysis. Management Science 42:467-72.

12.

Charnes

A. C.

Cooper

W. W.

Rhodes

E. L.

1978. Measuring the efficiency of decision making units. European Journal of Operational Research 2:429-44.

13.

Chen

C. F.

2007. Applying the stochastic frontier approach to measure hotel managerial efficiency in Taiwan. Tourism Management 28:696-702.

14.

Chen

T-H.

2009. Performance measurement of an enterprise and business units with an application to a Taiwanese hotel chain. International Journal of Hospitality Management 28: 415-422.

15.

Chen

T-H.

2011. Performance measurement in a small Taiwanese hotel chain. Cornell Hospitality Quarterly 52 (3): 354-62.

16.

Chiang

W. E.

Tsai

M. H.

Wang

L. S.

2004. A DEA evaluation of Taipei hotels. Annals of Tourism Research 31 (3): 712-15.

17.

Choi

Mattila

2004. Hotel revenue management and its impact on customer fairness perceptions. Journal of Revenue and Pricing Management 2 (4): 303-314.

18.

Cooper

W. W.

Seiford

L. M.

Tone

2000. Data envelopment analysis: A comprehensive text with models, applications, references and DEA-solver software. Boston: Kluwer Academic Publishers.

19.

Demeester

L. L.

2005. Managing learning resources for consecutive product generations. International Journal of Production Economics 95:265-83.

20.

Farrell

M. J.

1957. The measurement of productive efficiency. Journal of Royal Statistical Society 120 (3): 253-90.

21.

Hsieh

L. F.

Lin

L. H.

2010. A performance evaluation model for international tourist hotels in Taiwan: An application of the relational network DEA. International Journal of Hospitality Management 29: 14-24.

22.

J. L.

Chiu

C. N.

Shieh

H. S.

Huang

C. H.

2010. A stochastic cost efficiency analysis of international tourist hotels in Taiwan. International Journal of Hospitality Management 29 (1): 99-107.

23.

J. L.

Shieh

H. S.

Huang

C. H.

Chiu

C. N.

(2009). Cost efficiency of international tourist hotels in Taiwan: a DEA application. Asia-Pacific Journal of Tourism Research 14 (4): 371-84.

24.

Hwang

S. N.

Chang

T. Y.

2003. Using data envelopment analysis to measure hotel managerial efficiency change in Taiwan. Tourism Management 23:357-69.

25.

Johns

Howcroft

Drake

1997. The use of data envelopment analysis to monitor hotel productivity. Progress in Hospitality and Tourism 3 (2):119-27.

26.

Kao

Hwang

S. N.

2008. Efficiency decomposition in two-stage data envelopment analysis: An application to non-life insurance companies in Taiwan. European Journal of Operational Research 185:418-29.

27.

Kim

2011. Factor determinants of total factor productivity growth in the Malaysian hotel industry: A stochastic frontier approach. Cornell Hospitality Quarterly 52 (1): 35-47.

28.

Knutson

B. J.

1988. Frequent travelers: Making them happy and bringing them back. Cornell Hotel and Restaurant Administration Quarterly 29 (1): 83-87.

29.

Lewis

R. C.

1985. Predicting hotel choice: The factors underlying perception. Cornell Hotel and Restaurant Administration Quarterly 26 (3): 82-96.

30.

Lewis

B. R.

McCann

2004. Service failure and recovery: Evidence from the hotel industry. International Journal of Contemporary Hospitality Management 16 (1): 6-17.

31.

Lin

Y-S.

2011. Estimating cost efficiency and the technology gap ratio using the metafrontier approach for Taiwanese international tourist hotels. Cornell Hospitality Quarterly 52 (3): 341-53.

32.

Morey

Dittman

1995. Evaluating a hotel GM’s performance: A case study in benchmarking. Cornell Hotel Restaurant and Administration Quarterly 36 (5): 30-35.

33.

Nonaka

1994. A dynamic theory of organizational knowledge creation. Organization Science 5:15-37.

34.

O’Donnell

C. J.

Rao

D. S. P.

Battese

G. E.

2008. Metafrontier frameworks for the study of firm-level efficiencies and technology ratios. Empirical Economics 34: 231–255.

35.

Parasuraman

Zeithaml

V. A.

Berry

L. L.

1988. SERVQUAL: A multiple-item scale for measuring consumer perceptions of service quality. Journal of Retailing 64 (1):12-40.

36.

Rao

B. V.

Smith

B. C.

2006. Decision support in online travel retailing. Journal of Revenue and Pricing Management 5 (1): 72-80.

37.

Sanjeev

G. M.

2007. Measuring efficiency of the hotel and restaurant sector: The case of India. International Journal of Contemporary Hospitality Management 19 (5):378-87.

38.

Schmick

Kieser

2008. How much do specialists have to learn from each other when they jointly develop radical product innovations? Research Policy 37: 1148-63.

39.

Seiford

L. M.

Zhu

1999. Profitability and marketability of the top 55 U.S. commercial banks. Management Science 45 (9): 1270-88.

40.

Seiford

L. M.

Zhu

2002. Modeling undesirable factors in efficiency evaluation. European Journal of Operational Research 142:16-20.

41.

Sigala

Jones

Lockwood

Airey

2005. Productivity in hotels: A stepwise data envelopment analysis of hotels’ rooms division processes. Service Industries Journal 25 (1): 61-81.

42.

Steadmon

C. E.

Kasavana

M. L.

1988. Managing front office operations. 2nd ed. Orlando, FL: Educational Institute of the American Hotel and Motel Association.

43.

Sun

W. M.

2005. Evaluating the performance of the Taiwanese hotel industry using a weight slacks-based measure. Asia-Pacific Journal of Operational Research 22 (4): 487-512.

44.

Thrall

R. M.

1996. Duality, classification and slacks in DEA. Annals of Operations Research 66: 109-38.

45.

Tone

1993. Data envelopment analysis (in Japanese). Tokyo: JUSE Press.

46.

Tone

Sahoo

B. K.

2003. Scale, indivisibilities, and production function in data envelopment analysis. International Journal of Production Economics 84:165-92.

47.

Tsaur

2000. The operating of international tourist hotels in Taiwan. Asia Pacific Journal of Tourist Research 6 (1): 29-37.

48.

Wang

F. C.

Hung

W. T.

Shang

J. K.

2006. Measuring pure managerial efficiency of international tourist hotels in Taiwan. Service Industries Journal 26 (1): 59-71.

49.

Yang

W. M.

2006. Performance benchmarking for Taiwan’s international tourist hotels. INFOR 44 (3): 229-45.

50.

Yavas

Babakus

2003. What do guests look for in a hotel: A multi-attribute approach. Services Marketing Quarterly 25 (2): 1-9.

51.

M. M.

Lee

B. C. Y.

2009. Efficiency and effectiveness of service business: Evidence from international tourist hotels in Taiwan. Tourism Management 30: 571-580.