Article, Annotated Bibliographies

Contents lists available at ScienceDirect

Tourism Management

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

Do environmentally sustainable practices make hotels more efficient? A study of major hotels in Sri Lanka Thamarasi Kularatnea, Clevo Wilsona,∗, Jonas Månssonb, Vincent Hoanga, Boon Leea a School of Economics and Finance, Queensland University of Technology, Brisbane, QLD, 4000, Australia b School of Business and Economics, Linnaeus University, Sweden & Thammasat Centre for Efficiency and Productivity Analysis, Thammasat University, Thailand

A R T I C L E I N F O

Keywords: Hotel efficiency Productivity Green practices Data envelopment analysis (DEA) Malmquist Bootstrap estimation

A B S T R A C T

As environmentally sustainable practices are becoming popular in the hotel industry, their impact on efficiency is an important issue. To investigate the performance of hotels in this context, we use data envelopment analysis (DEA) double bootstrap approach to firstly assess the technical efficiency and its determinants of a sample of medium and large scale hotels in Sri Lanka for the period 2010–2014. Secondly, we evaluate the impact of a number of explanatory variables in determining hotel efficiency. The results reveal that the average technical efficiency is 61% with the maximum being 71.5% and the minimum 46.8%. The results conclude that being environmentally responsible enhances the efficiency of hotels, specifically in terms of improving energy effi- ciency and waste management. Water consumption is shown to have a contradictory result in relation of im- proving efficiency. Finally, we estimate the bootstrapped Malmquist productivity index to examine the level of productivity in the Sri Lankan hotel industry. The results of this study provide hotel operators and government with insights into the nature of competitive advantage which can assist them with strategic decision making to improve the technical and environmental management of hotels.

1. Introduction

Tourism industry in Sri Lanka is one of the fastest growing sectors in the post-civil war era since 2009. Tourism is the fourth largest foreign exchange earner and directly contributed 4.6% to Sri Lanka's gross domestic product (GDP) in 2015 and 4.2% of direct employment (World Travel and Tourism Council, 2016). Since 2009 the tourism industry has staged a rapid recovery with tourist arrivals more than quadrupling from 447,890 in 2009 to 2,050,832 in 2016. This contrasts with the 30 years period up to 2009, during which the average annual occupancy rate in graded accommodation averaged below 60% - a product of the war and civil unrest. However due to the post-war tourism boom, this figure had risen to over 74% by 2015 (Sri Lanka Tourism Development Authority, 2015). In accommodating this sub- stantial increase, hotels - especially medium and large scale – have made an important contribution to the economy. Of the hotels, 14.2% of total room capacity is accounted for by those with a 5 star rating, 10.8% by 4 star hotels and 7.6% by 3 star hotels. To cater for the ra- pidly rising tourism demand, hotels must now raise their efficiency levels.

Despite the importance of the tourism industry to the Sri Lankan economy, its future growth is now being challenged by a new force - the

changing nature of tourists who now demand eco-friendly practices. Many hotels acknowledge that adopting eco-friendly practices is not only an ethical practice but also beneficial in reducing costs, image enhancement, creating market differentiation and corporate social re- sponsibility (Mair & Jago, 2010; Radwan, Jones, & Minoli, 2012). This has led to attempts by both the Sri Lankan government and hoteliers to develop programs to promote the tourism industry's environmental sustainability. The “Greening Sri Lankan Hotels” campaign has been focused on small and medium scale hotels and the hospitality industry in general, providing a supporting role in terms of enhancing knowhow on improving energy and water utilisation efficiencies. According to the International Finance Corporation (IFC) report, these programmes play an important role in developing knowledge about energy efficiency and better water utilisation (IFC, 2013).

The hospitality industry is currently subject to persistent demands from customers for environmentally friendly practices. This is particu- larly so for hotels located in destinations which primarily attract tour- ists for their environmental diversity. In these cases, guests typically have strong preferences for green consumption alternatives (see, for example, Han & Hyun, 2018). Meeting this demand therefore becomes a factor contributing to demand attraction. This preference for ‘green’ hotels which address environmental concerns is borne out by a number

https://doi.org/10.1016/j.tourman.2018.09.009 Received 25 February 2017; Received in revised form 17 September 2018; Accepted 19 September 2018

∗ Corresponding author. E-mail address: [email protected] (C. Wilson).

Tourism Management 71 (2019) 213–225

Available online 18 October 2018 0261-5177/ © 2018 Published by Elsevier Ltd.

T

of studies (Bohdanowicz, 2006; Chan & Wong, 2006; Han & Yoon, 2015; Hathroubi, Peypoch, & Robinot, 2014). Moreover, research also suggests that in such cases, clients are willing to pay a premium and are more likely to pay a repeat visit (Laroche, Bergeron, & Barbaro-Forleo, 2001; Szuchnicki, 2009). In addition, a green hotel image has been found to positively affect customers' perceptions and increases their revisit intention for a future stay (Lee, Hsu, Han, & Kim, 2010). It is therefore important for hoteliers to know whether a decision to adopt eco-friendly practices in their hotel operations will be effective.

From a supply side perspective, the tourism industry regularly face issues related to environmental degradation (see e.g. Halkos & Matsiori, 2018; Ndebele & Forgie, 2017) - a fundamental concern for govern- ments and stakeholders (Hathroubi et al., 2014). In recognition of the negative environmental impacts of production, governments along with the green movement within the hotel and tourism industry have be- come aware of the need to implement effective measures to protect the environment. In this context, hotel managers have an incentive to adopt strategies that incorporate environmental sustainability.

The importance of addressing environmental issues is well-under- stood as a form of, variously, fulfilling customer needs and social re- sponsibility or business ethics. However, hotel managers need to be convinced that eco-friendly or green practices will be a cost effective, performance-improving exercise in the long run. For example, is reuse of towels or use of energy efficient lighting simply measures which demonstrate a hotel's participation in sustainable initiatives or merely a cost savings mechanism for the hotel? And ultimately does it contribute to improving overall efficiency? It is this latter question which this study is designed to answer.

The reasons for improving efficiency in terms of boosting profit- ability of a firm's operations is well documented in the literature (see e.g. Shang, Wang, & Hung, 2009). Given the high level of competition in the hotel industry there is considerable pressure to upgrade their efficiency relative to competitors. This creates a particular need for benchmarking analyses that can identify best practices. This is parti- cularly so at the micro level, where the hotel industry is becoming in- creasingly sensitive to the changing tastes and preferences of tourists seeking accommodation (Assaf & Agbola, 2011). Efficiency evaluation is also beneficial as it provides guidelines for policy makers to correct inefficient management directions and promote positive practices which are a product of the competitive environment.

While the importance of eco-friendly practices in the hotel industry is increasing, the operational practices involve a number of issues. There is the likely tension between pro-environmental operations and costs. That is, installation of eco-friendly practices could increase costs. However, they could equally reduce costs if alternative technologies are used. For example, in a given developing country's socio-economic and policy environment, adopting eco-friendly practices may involve ex- pensive inputs. As a result certain hotels tend to use environmentally detrimental inputs. For example, considerably greater initial costs are involved in the installation of efficient HVAC (heating, ventilation, air conditioning) systems, energy efficient lighting (Light-Emitting Diode-

LEDs and Compact Fluorescent Lamp-CFLs) and use of renewable en- ergy sources. During the introduction stage of the concept of a green hotel, hoteliers invest more to make their operations eco-friendly, thereby impacting cost efficiency (Shieh, 2012).

Nevertheless, studies conducted by the International Hotels Environment Initiative (IHEI) reveal that 90% of hotel guests prefer to stay in a hotel that cares for the environment. Therefore, in the long run, investment in environmentally green practices by hotels can be cost effective, and improve their competitiveness by attracting en- vironmentally concerned consumers. Government efforts to promote environmentally friendly practices in firms should therefore be seen as having the potential to be cost effective, client appreciative and a means to both improve the competitiveness of firms and overcome negative environmental impacts.

In this context, it is important to understand the nature of energy and water consumption patterns of tourist hotels. The hotel industry in Sri Lanka consumes 2% of Sri Lanka's total electricity usage and 1% of water of total consumption. On average 50% of hotels' electricity use is accounted for by air-conditioning, 20% for lighting and the rest for kitchen, laundry and other purposes. 38% of typical water usage in tourist hotels is by guests in their rooms and 21% by kitchens (IFC, 2013). Hence it is clear that efforts to conserve water should be directed at these respective departments. In terms of waste generation, the IFC reports that 46% of the waste generated by hotels is food and non- recyclables (solid waste), 25% paper and the rest is in the form of cardboard, plastics, glass and metal. Energy, water and waste are therefore the major areas being improved by hotels in their efforts to become ecofriendly in their operations.

The above information reflects a positive business environment for ecofriendly initiatives and the appropriateness of the Sri Lankan gov- ernment's encouragement of hotels to address environmental issues. However, in adopting eco-friendly practices, hotel managers need to benchmark such strategies so as to measure the extent to which they actually improve their efficiency. The objective of this paper is there- fore twofold. First, it is to measure the technical efficiency and pro- ductivity of a sample of medium and large scale Sri Lankan hotels and second, to evaluate the impact of eco-friendly practices on efficiency. This study incorporates three different eco-friendly measures (i.e. en- ergy consumption, water consumption and waste management) that Sri Lankan hotels typically incorporate in their operations as given in Table 1.

This research paper proceeds as follows. In section 2 we review related prior studies in the areas of efficiency in the hotel industry. Section 3 sets out the research method. Section 4 provides a description of the data and section 5 details empirical results and interpretations. Section 6 provides concluding remarks.

2. Literature review

Studies on the efficiency of hotel industry operations have provided insights into issues relating to the industry in general and to hotel

Table 1 Eco friendly practices.

Type of green practice Criteria

Energy The hotel's use of renewable energy, as a % of total energy used (e.g. solar, biomass, wind) The hotel's implementation of energy saving measures through energy efficient lighting and HVAC (heating, ventilation, and air conditioning) improvements

Water The hotel harvests and utilises rainwater for purposes such as gardening and washing The hotel controls the quality and quantity of effluent discharge with proper treatment prior to discharge into the environment and sewerage network

Waste The hotel attempts to reduce generation of waste in its daily operations (e.g. by reducing the usage of paper, plastic and glass containers, food waste and used oil, cardboard and paper etc.) The hotel reuses where possible The hotel recycles waste (in the hotel and/or through external parties) The hotel treats hazardous waste before sending to landfill

T. Kularatne et al. Tourism Management 71 (2019) 213–225

214

managers in particular (Ben Aissa & Goaied, 2016; Hathroubi et al., 2014). There are a number of recent studies addressing efficiency of hotels which adopt two main methods to estimate efficiency. They are data envelopment analysis (DEA) and the stochastic frontier approach (SFA).1 We provide in Table 2 a summary of studies which uses DEA to measure efficiency of hotels. DEA was first applied in the hotel industry studies by Morey & Dittman, 1995 to measure performance of 54 US hotels. Since then, hotel efficiency literature has extensively employed CCR and BCC models2 (Banker, Charnes, & Cooper, 1984; Charnes, Cooper, & Rhodes, 1978). Hwang and Chang (2003) use a DEA CCR model to measure the managerial performance of 45 hotels in Taiwan. Anderson, Fok, and Scott (2000) utilise both BCC and CCR models to investigate the performance of 48 hotels in the US in 1994 and find that the overall efficiency value was an exceptionally low as result of poor technical efficiency and scale efficiency. Other studies which use BCC and CCR models include Tsaur (2001), Brown and Ragsdale (2002), Reynolds (2003), Chiang, Tsai, and Wang (2004) and Barros (2005a). However, none of these studies seek to measure the efficiency of the hotel industry in Sri Lanka.

The literature has shown there are many variables which are not under the control of a firm but can influence firm efficiency (Coelli, Prasada Rao, O'Donnell, & Battese, 1998). Therefore properly ac- counting for these variables is important for the results to be reliable. However, most existing efficiency studies of the hotel industry do not account for the impact of explanatory variables (determinants ex- plaining variations in efficiency levels across decision making units (DMUs) on efficiency. The existing literature only accounts for such variables as ownership or managerial factors. For example, Barros and Dieke (2008) test the impact of ownership (on both chains and in- dependent hotels) on efficiency and find that hotels' membership in a group increases efficiency. Assaf, Barros, and Josiassen (2012) find that ownership, classification and size of a particular hotel have a significant impact on its efficiency. Shang et al. (2009) measured the impact of hotel management factors (i.e. age, location and management style) on efficiency. It is also important to note that some studies divide their samples according to size, type and ownership of the hotel to in- vestigate efficiency based on specific groups (Assaf et al., 2012), instead of using criteria such as explanatory variables as used in the DEA second stage analysis.

Performance measurement studies in the hotel industry using the Malmquist productivity index are rare. Among them is the study by Assaf and Barros (2011) which uses the Malmquist index with bias correction to analyse the performance of hotel chains from the UAE, Saudi Arabia and Oman. Barros (2005a) estimated the Malmquist index for 42 hotels in Portugal by breaking it down to technical efficiency and technological change. A bootstrapped Malmquist index to measure the productivity of the Sri Lankan hotel industry would be a novel addition to this literature.

There is a rich source of literature which studies the effect of en- vironmental concerns on firms' operations in the manufacturing sector (see, for example, Egilmez & Park, 2014; Kumar Mandal & Madheswaran, 2010). Service sector firms which have been described as the “silent destroyers of the environment” have only received at- tention of researchers in recent years (Shieh, 2012). The study by Bohdanowicz (2006) of Swedish and Polish hotel industries reveals that the magnitude of the impacts of the hotel industry is often under- estimated and that it consumes a large amount of local and imported non-durable goods, water and energy, and emits large amounts of carbon dioxide. Consequently, while individual hotels may not have a

sizeable negative impact on the environment, collectively they can be highly and wastefully resource intensive. It has been estimated that 75% of hotels' environmental impacts can be directly attributed to ex- cessive consumption (Bohdanowicz, 2006). This is wasteful not only in terms of resources but also in terms of unnecessarily high operational costs. As a result, there is a strong incentive for hotels to increasingly adopt green practices to their production processes.

Accounting for the impact of eco-friendly or green practices in the study of hotel efficiency is therefore much needed. The Green Hotels Association (2014) articulates the concept of a green hotel as “en- vironmentally-friendly properties whose managers are eager to institute programs that save water, save energy and reduce solid waste—while saving money—to help protect our one and only earth!”. The 1992 United Nations Conference on Environment and Development pointed out that hotels can minimise their environmental impacts by installing visible eco-friendly technology (such as solar panels, low flow sho- werheads, recycling bins, etc.), and in this way gain the attention of customers (Kang, Stein, Heo, & Lee, 2012).

Initial moves to use green practices in international hotel manage- ment were centred on cost saving initiatives by reducing waste, energy usage and government regulation (Shieh, 2012). However, due to the escalating demand for green hotels, adopting green practices has be- come not only a cost saving method, but equally associated with cus- tomer expectations (Chan, 2013), corporate image (Penny, 2007) and the willingness to pay a premium for green hotels (Kang et al., 2012; Laroche et al., 2001). Chen, Chen, Zhang, and Xu (2018) identified that green management in a hotel helps not only in terms of profitability and customer retention but also demonstrates improved social responsi- bility and good reputations. Given the complexities of implementing strategies related to green management of a hotel, hotel managers need to have a sophisticated appreciation of what actually drives a hotel's efficiency and hence, profitability. Although managerial implications of using green practices in hotels have been well examined, the efficiency of hotels adopting green practices is yet to be investigated. In parti- cular, empirical testing of the relationship between green practices and technical efficiency has become an important and timely concern in the Sri Lankan context where environmentalism is increasingly being pro- moted by the government.

In highlighting the importance of green processes in hotel produc- tion, it is also useful to look at the literature examining consumer preferences toward green practices. Studies by Kasim (2004) and Lee et al. (2010) found positive perceptions of consumers regarding a ho- tel's green initiatives. In particular, tourists were willing to accept ho- tels' water conservation, recycling and energy conservation actions that contributed to positive environmental impacts. However other studies indicated that the perception of some hotels guests is that hotels may just be using green practices as a marketing tool or to gain financial benefits (Yi, Li, & Jai, 2016).

3. Methodology

3.1. DEA efficiency analysis

DEA is a linear programming procedure for a frontier analysis of inputs and outputs. It was first introduced by Charnes et al. (1978) building on the frontier efficiency concept first elucidated by Farrell (1957). This latter concept was based on a mathematical programming approach to the construction of production frontiers and designed to measure of efficiency in relation to the estimated frontiers. This model, which assumes constant returns to scale (CRS), i.e. long run efficiency. This model is also known as the CCR model. Banker et al. (1984) first introduced the assumption of variable returns to scale (VRS) which has become known in the literature as the BCC model. The BCC model is also known as short run efficiency. In this study we will assume variable returns to scale for the repeated cross-sectional analysis, i.e. short run efficiency, and constant returns to scale for the analysis of productivity.

1 This paper focuses only on the DEA literature on hotel efficiency. 2 The CCR model named after Charnes et al. (1978) is based on the as-

sumption of constant returns to scale (CRS). The BCC model, developed under the assumption of variable reruns to scale (VRS) was named after Banker et al. (1984).

T. Kularatne et al. Tourism Management 71 (2019) 213–225

215

Ta bl e 2

Li te ra tu re su rv ey

of D EA

m od el s on

to ur is m .

St ud y

M et ho d

U ni ts

In pu ts

O ut pu ts

Se co nd

st ag e ex pl an at or y va ri ab le s

(M or ey

& D itt m an ,

19 95 )

D EA

54 U S ho te ls

(1 ) ro om

ex pe nd itu re s; (2 ) en er gy

co st s, (3 ) sa la ry ;

(4 ) ad ve rt is in g ex pe nd itu re s; (5 ) no n- sa la ry

ex pe ns es ;( 6) fix ed

ex pe nd itu re s.

(1 ) to ta lr ev en ue ;( 2) le ve lo fs er vi ce

de liv er ed ;( 3)

ra te of gr ow

th .

–

(A nd er so n et al .,

20 00 )

D EA

(t ec hn ic al an d

al lo ca tiv e)

48 U S ho te ls

(1 ) fu llt im e eq ui va le nt em

pl oy ee s; (2 ) nu m be r of

ro om

s; (3 ) to ta lg am

in g re la te d ex pe ns es ;( 4) to ta l

fo od

an d be ve ra ge ;e xp en se s; (5 ) ot he r ex pe ns es .

(1 ) to ta lr ev en ue s; (2 ) ot he r re ve nu es .

–

(T sa ur ,2 00 1)

D EA

53 Ta iw an es e ho te ls

(1 ) to ta lo pe ra tin g ex pe ns es ;( 2) nu m be r of ro om

s oc cu pi ed ;( 3)

to ta lfl oo r sp ac e; (4 ) nu m be r of

em pl oy ee si n th e ca te ri ng

di vi si on ;( 5) ca te ri ng

co st s.

(1 )t ot al op er at in g re ve nu es ;( 2) nu m be ro fr oo m s; (3 )

av er ag e da ily

ra te ;( 3) to ta lo pe ra tin g re ve nu e of th e

ca te ri ng

di vi si on

–

(H w an g & Ch an g,

20 03 )

D EA

CC R m od el ;

M al m qu is t

45 Ta iw an es e ho te ls

(1 ) nu m be r of fu llt im e em

pl oy ee s; (2 ) nu m be r of

gu es tr oo m s; (3 )t ot al di m en si on

of m ea ld ep ar tm en t;

(4 ) op er at in g ex pe ns es .

(1 )r oo m re ve nu e; (2 )f oo d an d be ve ra ge

re ve nu e; (3 )

ot he r re ve nu e.

–

(B ar ro s, 20 05 a)

D EA

M al m qu is t w ith

se co nd -s ta ge

To bi t

re gr es si on

42 Po rt ug ue se ho te ls

(1 99 9– 20 01 )

(1 ) fu llt im e em

pl oy ee s; (2 ) co st of la bo ur ;( 3) bo ok

va lu e of pr op er ty ;( 4) O pe ra tin g co st s.

(1 ) sa le s; (2 ) nu m be r of gu es ts ;( 3) nu m be r of ni gh ts

oc cu pi ed .

(1 ) lo ca tio na lc om

pe tit iv e po si tio n; (2 ) ci ty or

re m ot e; (3 )d is ta nc e to th e ai rp or t; (4 )n o. of ro om

s; (5 ) va lu e of ye ar ly in ve st m en t; (6 ) ra tio

of th e

w or ke rs of th e ho te lv er su s th e to ta ln um

be r of

En at ur 's w or ke rs

(B ar ro s, 20 05 b)

D EA -C CR

an d D EA -B CC

m od el

42 En at ur

ho te ls in

Po rt ug al

(1 )f ul lti m e em

pl oy ee s; (2 )c os to fl ab ou r; (3 )r oo m s;

(4 ) su rf ac e ar ea

of th e ho te l; (5 ) bo ok

va lu e of

pr op er ty ;( 6) op er at io na lc os ts ;( 7) ex te rn al co st s.

(1 ) sa le s; (2 ) nu m be r of gu es ts ;( 3) ni gh ts sp en t.

–

(B ar ro s & Sa nt os ,

20 06 )

D EA

al lo ca tiv e m od el

15 Po rt ug ue se ho te ls

(1 99 8– 20 02 )

(1 ) fu ll- tim

e em

pl oy ee s; (2 ) bo ok

va lu e of as se ts

(1 ) sa le s; (2 ) ad de d va lu e; (3 ) ea rn in gs .

–

(B ar ro s & D ie ke ,

20 08 )

D EA

1s t st ag e M al m qu is t

w ith

2n d st ag e

bo ot st ra pp ed

To bi t m od el

12 Lu an da -A fr ic an

ho te ls

(1 ) to ta lc os t; (2 ) in ve st m en te xp en di tu re

(1 ) re ve nu e pe r av ai la bl e ro om

(1 ) tr en d; (2 ) sq ua re tr en d; (3 ) m ar ke t sh ar e; (4 )

du m m y fo r gr ou p (5 ) du m m y fo r in te rn at io na l

ex pa ns io n st ra te gy

(C he n, 20 09 )

D EA

w ith

sl ac ks

7 ho te ls in Ta iw an

(1 )n um

be ro fe m pl oy ee s; (2 )s ur fa ce d ar ea ;( 3) gu es t

ro om

s; (4 ) op er at in g ex pe ns es ;( 5) de pr ec ia tio n

ex pe ns es

(1 ) oc cu pa nc y ra te ;( 2) ra te of gu es t sa tis fa ct io n; (3 )

nu m be r of gu es ts ;( 4) ro om

re ve nu e; (5 ) ot he r

re ve nu e

–

(A ss af et al ., 20 12 )

D EA

m et af ro nt ie r

78 Ta iw an es e ho te ls

1) nu m be r of ro om

s; (2 ) nu m be r of fu ll tim

e eq ui va le nt em

pl oy ee s in th e ro om

di vi si on ;( 3)

nu m be ro ff ul lt im e eq ui va le nt em

pl oy ee si n th e fo od

an d be ve ra ge

di vi si on ;( 4) nu m be r of fu ll tim

e eq ui va le nt em

pl oy ee s in ot he r de pa rt m en ts .

(1 ) to ta lr oo m re ve nu es ;( 2)

to ta lf oo d an d be ve ra ge

re ve nu es ;( 3) to ta lo fo th er re ve nu es

–

(H si eh

& Li n, 20 10 )

Re la tio na ln et w or k D EA

57 in te rn at io na l

to ur is t ho te ls in

Ta iw an

(( 1) ac co m m od at io n co st s; (2 ) em

pl oy ee s of th e

ac co m m od at io n de pa rt m en t; (3 ) ca te ri ng

co st ;( 3)

em pl oy ee s of th e ca te ri ng

de pa rt m en t

(1 ) ro om

s; (2 ) ca te ri ng

flo or s as in te rm ed ia te ou tp ut s

an d; (1 ) re ve nu e of th e ac co m m od at io ns

(2 ) re ve nu e

of th e ca te ri ng

de pa rt m en ts as ou tp ut s

–

(H ua ng ,H

o, & Ch iu ,

20 14 )

D EA

tw o st ag e m od el

(e xt en si on

to CZ

tw o- st ag e

m od el )

58 in te rn at io na l

to ur is t ho te ls in

Ta iw an

(1 )o pe ra tin g ex pe ns es ;( 2) ro om

s; (3 )c at er in g sp ac e;

(4 ) em

pl oy ee s; (5 ) m ar ke tin g as in pu ts an d (1 )

m ar ke tin g ex pe ns e as in te rm ed ia te in pu t

(1 ) oc cu pa nc y re ve nu e; (2 ) nu m be r of lo dg in g gu es ts

as fin al ou tp ut s of oc cu pa nc y di vi si on ;( 1) oc cu pa nc y

se rv ic e ca pa ci ty ;( 2) ca te ri ng

se rv ic e ca pa ci ty as

in te rm ed ia te ou tp ut ;a nd

(1 )c at er in g re ve nu e as fin al

ou tp ut of ca te ri ng

di vi si on .

–

(O uk il, Ch an no uf ,&

A l-Z ai di ,2 01 6)

D EA

CC R m od el 2n d st ag e

bo ot st ra pp ed

To bi t m od el

58 ho te ls in O m an

(1 ) N um

be r of be ds ;( 2) sa la ry of em

pl oy ee s.

(1 ) A nn ua lr ev en ue ;( 2) N um

be r of gu es ts ;( 3)

N um

be r of ni gh ts ;( 4) oc cu pa nc y ra te .

(1 )o w ne rs hi p; (2 )s iz e; (3 )s ta rr at in g; (3 )n at ur e; (3 )

cu ltu re ;( 5) ac tiv iti es

T. Kularatne et al. Tourism Management 71 (2019) 213–225

216

DEA is applied to unit assessment of homogeneous units, such as hotels, which are referred to as DMUs. The efficiency of a DMU is measured as the ratio of weighted outputs to weighted inputs. The ef- ficiency of each DMU can be calculated, once the frontier is con- structed, by comparing distances from the points on the frontier with the points that are below the frontier. However, DEA is sensitive to outliers which might exaggerate the actual frontier. The researcher must specify three characteristics of the model: the input-output or- ientation system, the returns to scale and the weights of the evaluation system in order to solve the linear programming problem.

Before stating the problem to be solved a few notations is needed. Let there be K (k=1,2, ...,K) hotels. These hotels producesM (m=1,2, ...,M) different outputs with the use of N (n=1,2, …,N) different in- puts. For the hotel under examination (‘O’) the input-oriented DEA ef- ficiency estimator o can be derived by solving the following linear programming;

=TE s t

min . .

o (1)

= =

z y y m M, 1,2, .., k

K

k km o 1 (2)

= =

z x x n N, 1,2, ..., k

K

k kn o o 1 (3)

=z k K CRS0, 1,2, ..., ( )k (4)

= =

z VRS1 ( ) k

K

k 1 (5)

The value of o obtained is the technical efficiency score for the hotel under examination. A measure of = 1o indicates that the hotel is technically efficient, and inefficient if < 1o . This linear programming problem must be solved K times, once for each hotel in the sample. If only rows [1] – [4] are used the DEA model is solved under the as- sumption of constant returns to scale (CRS). By including restriction [5] we impose a variable return to scale (VRS) assumption.

In order to test the hypotheses that efficiency of hotels is determined by different contextual variables, this study uses the two-step approach as suggested by Coelli et al. (1998). It is recognized in DEA literature that the efficiency scores obtained in the first stage can be correlated with the explanatory variables used in the second stage regression. This can lead to inconsistency and biased estimates of the second stage re- gression. To overcome this problem, a bootstrap procedure is needed as pointed out by Simar and Wilson (2007). The bootstrap is a resampling technique used as a means of approximating the properties of the sampling distribution of an estimator and, hence, allowing the conduct of hypothesis testing and construction of confidence intervals. We es- timate the following specification:

= +k k k (6)

where k is a vector of environmental variables which is expected to explain the efficiency variations and where refers to a vector of parameters to be estimated and k is an error term.

In hotel efficiency literature, the bootstrapping method was first used by Barros and Dieke (2008), who estimated the technical effi- ciency of 12 hotels in Africa over the years 2000–2006. In the first stage they used a DEA model to rank hotels and in the second stage, the Simar and Wilson (2007) procedure was used to double bootstrap DEA scores with a truncated regression. Assaf and Agbola (2011) employed a double bootstrap approach to assess the technical efficiency of Aus- tralian hotels for the period 2004–2007. These studies indicate that the DEA bootstrap approach corrects for the bias inherent in traditional DEA models.

The model explained in equation (6) can be further expanded to explain each environmental variable as presented in equation (7). We

examine the determinants of technical efficiency of Sri Lankan hotels as follows:

= + + + + + + +

+ + + + +

age star size type ecol eco eco

D D D D

2 3

2011 2012 2013 2014 k

k

0 1 2 3 4 5 6 7

8 9 10 11 (7)

where k represents the technical efficient score. Age is the number of years the hotel has been in operation, size is a dummy variable re- presenting 1 for large hotels (more than 100 rooms) and 0 otherwise. star is the star rating of the hotel, type is a dummy variable which is 1 for resorts and 0 otherwise. Thus these variables aim to capture the impact of the type of hotel on efficiency. eco1, eco2 and eco3 are scores given to each hotel based on their involvement in eco-friendly practices in day to day operations representing energy-saving, water-saving and waste management practices respectively. We incorporated criteria defined in Table 1 in constructing each eco variable. D2011, D2012, D2013 and D2014 are dummy variables constructed for each year (2010 is taken as the reference).

The contextual variables of Age (Assaf & Agbola, 2011;Shang et al., 2009, Size (Barros & Dieke, 2008; Oukil et al., 2016), Star (Oliveira, Pedro, & Marques, 2013; Oukil et al., 2016) and Type (Wang, Hung, & Shang, 2006; Barros, 2005a; Tsaur, Chiang, & Chang, 1999) are com- monly discussed in relation to efficiency in the literature. These are examined in detail in section 4.2. The three eco variables represent three different criteria explaining green practices of a hotel in Sri Lanka. The first criteria (eco1) relate to a hotel's energy conservation mea- sures. This was evaluated based on two factors; firstly the use of re- newable energy as a percentage of total energy used and secondly, the use of energy efficient lighting and HVAC systems. Studies have shown that the hotel industry is facing substantial difficulties in controlling the quality of water it uses and in reducing over-consumption of water which could lower the cost of water and support water conservation (Mensah, 2006). Hence the next criteria was based on water efficiency by looking at hotels' reuse of rainwater and the extent to which effluent discharges have been properly treated prior to discharge into the en- vironment (eco2). Waste management (eco3) was taken as the third criteria given solid waste generation and disposal is considered as one of the main negative impacts of hotels on the environment (Radwan et al., 2012). The questions used indicated hotels' attempts to reduce, reuse and recycle energy in their daily operations (Table 1).

3.2. Productivity measurement using Malmquist productivity index

This study measured productivity improvements in the Sri Lankan hotel industry using the nonparametric Malmquist index (for pio- neering works see Färe, Grosskopf, Lindgren, & Roos, 1992; 1994). We establish the sources of productivity change by decomposing the Malmquist productivity index into efficiency change and technical change in the production frontier. Following Färe, Grosskopf, Norris, and Zhang (1994), the input-oriented Malmquist productivity change index is expressed as:

= =

× + + + + +

+ + + +

M y x y x D y x D y x

D y x D y x

( , , , ) ([ ( , )/ ( , )]

[ ( , )/ ( , )]) t I

t t t t t I

t t t I

t t

t I

t t t I

t t

1 1 1 1 1

1 1 1 1 1/2 (8)

where the superscript I indicated input-orientated technology, M is the productivity of the most recent production point + +x y( , )t t1 1 using period t+1, relative to the earlier production point x y( , )t t using period t+1 technology and D represents input distance functions. Values greater than 1.0 indicate productivity growth between the two periods.

Equation (3) can be re-written as:

=

× × + + + + + +

+ + + + + +

M y x y x D y x D y x

D y x D y x D y x D y x

( , , , ) [ ( , )/ ( , )]

[( ( , )/ ( , )) ( , )/ ( , )] t I

t t t t t I

t t t I

t t

t I

t t t I

t t t I

t t t I

t t

1 1 1 1 1 1

1 1 1 1 1 1 1/2 (9)

where

T. Kularatne et al. Tourism Management 71 (2019) 213–225

217

= + + +E D y x D y x[( ( , )/ ( , )]t I

t t t I

t t1 1 1 (10)

= ×+ + + + + +T D y x D y x D y x D y x[( ( , )/ ( , )) ( ( , )/ ( , ))]t I

t t t I

t t t I

t t t I

t t1 1 1 1 1 1 1/2 (11)

Malmquist (M) productivity is the product of a measure of technical change (T) measured by shifts in frontier between period t+1 and t. The efficiency change (E) over this period measures how much closer the DMU or hotel is to the frontier.

In this study, however, we used the bootstrap approach (Simar & Wilson, 1999) to calculate confidence intervals for hotel-specific Malmquist productivity indices for the period 2010–2014. The inputs and outputs used are described in section 4.1. The bootstrap Malmquist indices allow calculation of the bias in the results. In this study, 2000 bootstrap iterations were performed and 95% confidence intervals were constructed. The R software and FEAR package were used for the above analysis.

4. Data and results

4.1. Data

We used panel data for 24 hotels over the 5 annual periods 2010–2014. The data is obtained from a field survey undertaken in Sri Lanka between November 2014 and May 2015. The target population for this study was medium and large scale hotel establishments located in Sri Lanka. The number of DMUs which falls into this category is relatively low in a small developing economy such as Sri Lanka. Thus, although the current boom in the tourism industry has given rise to heavy investments in the hotel sector, medium and large scale hotels which have been in operation for 5 years-are limited in number. These factors explain the size of the population under investigation: currently there are 46 three to five star hotels (Sri Lanka Tourism Development Authority, 2015). Hence a sample of 24 represents more than 50% of the total population of interest.

DEA methodology can be applied to a small DMU population (Evanoff & Israilevich, 1991; Perrigot, Cliquet, & Piot-Lepetit, 2009). For example, Chen (2009) uses DEA to investigate the efficiency of seven hotel chains in Thailand while Perrigot et al. (2009) assessed the efficiency of fifteen hotel chains in France. However, the number of inputs and outputs entered in the DEA model must be reasonably de- fined. The DEA literature suggests that number of DMUs should be at least twice the product of the number of inputs and outputs in the model (Dyson et al., 2001). Moreover, the bootstrap technique becomes particularly useful - especially for small samples - in achieving reliable results (Atkinson & Wilson, 1995).

There are two outputs and three inputs in our model as shown in Table 3. The two outputs are room revenue and other revenue which includes food and beverage revenue and revenue from other depart- ments. Inputs include the number of rooms, the number of employees and the book value of assets. The chosen inputs and outputs are based on two criteria; firstly the availability of data and secondly a literature review of previous studies.

The explanatory variables used in this study include the age of the hotel, star rating, size3 (based on the number of rooms), type (whether is it a resort or a city hotel), and use of eco-friendly practices. These variables are explained in detail in section 5.2.

5. Results and discussion

Given the extensive competition within the hotel industry during the post-war development era, it is reasonable to consider the input orientation case where hotels are assumed to minimise inputs given the levels of outputs are fixed. The efficiency results presented in this sec- tion are relative. That is, a hotel assessed to be more or less efficient in relation to the other 23 hotels in the sample. Any improvements sug- gested in increasing efficiency are relative to best-practice hotels in the analysis.4

5.1. Short run technical efficiency

Table 4 presents the bias corrected technical efficiency scores for each year in the study period.5 We observe an improvement in mean efficiency scores from 2010 to 2012 (55%–68%) followed by a decline in efficiency in 2013. In 2014 a slight increase up to 65% is observed. The bias corrected mean efficiency for the study period is 61.1% with the maximum being 71.5% and the minimum 51.2%. Except for the first year, five and ten hotels are considered to be efficient. In this study this is defined as having an efficiency score that is not significant from one. The detailed results are provided in Appendix A.

5.2. Determinants of efficiency

Table 5 reports the estimated truncated second stage regression results as specified in equation (7). The truncated regression model with a bootstrap fits the data well, with statistical significance for most parameters. The estimations generally confirm our prior expectations.

The positive and statistically significant coefficient of the age vari- able is consistent with past research. The literature suggests that as hotels mature with age, they tend to earn a certain level of reputation and brand status which induces them to maintain a high level of effi- ciency in their operations (see, for example, Assaf & Agbola, 2011; Shang et al., 2009). Likewise hotels which are mature in age tend to rely on reputation and brand to improve sales (Shang, Hung, Lo, & Wang, 2008). This reliance on brand and reputation may affect how hotels utilise the available resources, hence contributing to efficiency.

The star ratings of hotels are an indication of the level of luxury of the services provided, the quality of food and beverages, entertainment facilities, views and the variety of rooms of different sizes aligned with international standards (Oliveira et al., 2013). Higher star ratings are

Table 3 Characteristics of the inputs and outputs.a.

Variables Units Range Mean SD

Outputs Room revenue LKR 20–552.42 194.22 123.16 Other revenues LKR 18–589 160.43 113.58 Inputs Number of employees Number 42–416 212.48 95.28 Number of rooms Number 31–200 101.53 41.93 Book value of assets LKR 32.18–5766 987.27 1067.78 explanatory variables Age Number 1–122 23.63 23.06 Star Number 0–5 3.3 1.62 Size (1 if more than 100 rooms; 0

otherwise) Dummy 0–1 – –

Type (1 for resorts; 0 for city hotels) Dummy 0–1 – – Eco 1 (Energy) Number 5–80 45.00 26.46 Eco 2 (Water) Number 5–100 66.04 19.74 Eco 3 (Waste) Number 17.5–97.5 71.04 22.31

Note: LKR: Sri Lankan rupees; a the monetary values are in millions LKR.

3 It may be assumed that star rating of the hotel can be related to the size of the hotel and hence potential multicollinearity between the two variables. We tested for multicollinearity between size and star rating using the Spearman's rank correlation test and found that these two variables are independent in our study sample. 4 The DEA technical efficiency scores were estimated using RDEA package

applied in R-software. 5 A potential problem with DEA is that hotels get an efficiency score equal to

one if they are unique. Uniqueness is defined as where the efficiency score is one, but the hotel is not used in the reference to any other hotel. The number of unique hotels is: 2010 2, 2011 1, 2012 1, 2013 1 and 2014 4.

T. Kularatne et al. Tourism Management 71 (2019) 213–225

218

generally associated with higher efficiency (Assaf & Agbola, 2011; Assaf et al., 2012; Barros & Dieke, 2008; Pine & Phillips, 2005). It is noted that Oliveira et al. (2013) found that star rating is not a significant determinant of efficiency. In a more recent study by Fernández and Becerra (2015), they discovered that the effect of star-rating on effi- ciency can be diverse and is based on other factors. Also, Chen (2007) indicated that there are particularly large differences in efficiency be- tween 5 star and 4 star hotels and that the latter are in fact generally more efficient that the former. The same conclusion was reached by Tarim, Dener, and Tarim (2000). In the present study, the double bootstrap regression results indicate that the star rating is positively associated with efficiency (but not significant).

The coefficient of the size variable is also found to impact negatively on the technical efficiency of Sri Lankan hotels. The relationship

between size and efficiency is an area of contradictory findings in the hotel efficiency literature. For example, Chen (2007) and Hwang and Chang (2003) indicate that no difference in efficiency exists between large and small scale hotels. However, Barros and Dieke (2008) con- clude that large hotels are more efficient than small hotels in their African study (Luanda) and the same is found to be true for some stu- dies of Portuguese hotels (Barros, 2005b, 2006). The negative re- lationship between hotel size and efficiency in our context could be attributed to the fact that medium sized hotels can improve their effi- ciency with a greater size whereas for large hotels the growth process can sometimes cause diseconomies of scale (examples from other in- dustries include Romano & Guerrini, 2011, Torres & Morrison, 2006).

The type variable is found to have a positive relationship with effi- ciency, although it is not significant. This implies resorts are more ef- ficient than city hotels. On this issue, the literature provides mixed assessments; for example Tsaur et al. (1999) find that city hotels are more efficient than resorts whereas Wang et al. (2006) conclude otherwise. It is also true that resorts, given their usually attractive lo- cations, generally have higher occupancies at weekends (Wang et al., 2006).

Adopting eco-friendly or green practices is shown to contribute positively to efficiency in relation to energy efficiency (Eco1) and waste management (Eco3). In these cases the coefficients are positive and highly significant partly confirming our prior hypothesis. Although initial adaptation of green practices may incur substantive costs, in the long run, a hotel benefits in terms of costs by reducing the amount of energy used and managing waste. Green practices also helps to estab- lish a brand image to attract more environmentally conscious custo- mers. Pertinently, Hathroubi et al. (2014) find that hotels using clean and renewable energy are more efficient whereas Kim, Li, Han, and Kim (2016) show hotel green practices have a significant influence on cus- tomers' overall ratings and hotel performance.

Clearly, hotels could improve efficiency of their operations by in- vestigating their energy use and by discovering possibilities of moving towards the use of renewable energy sources such as solar which is abundantly available in all parts of Sri Lanka. Installing energy efficient HVAC systems (especially air conditioning) could be useful for warm

Table 4 Bias corrected technical efficiency scores for Sri Lankan hotels 2000–2014, Variable returns to scale.

Hotel 2010 2011 2012 2013 2014 Arithmetic mean

1 0.624 0.657 0.811 0.637** 0.719** 0.690 2 0.725 0.676** 0.753 0.637 0.673 0.693 3 0.537** 0.591 0.699 0.627 0.666 0.624 4 0.527 0.363** 0.598** 0.660** 0.793 0.588 5 0.601** 0.687** 0.731 0.650 0.708 0.675 6 0.534** 0.59 0.697 0.627 0.662 0.622 7 0.500 0.559** 0.563** 0.394** 0.623** 0.528 8 0.418 0.514** 0.590** 0.492** 0.545** 0.512 9 0.547 0.597** 0.823 0.737** 0.624** 0.666 10 0.469 0.462** 0.536** 0.428** 0.505** 0.480 11 0.414 0.540** 0.560** 0.466** 0.375** 0.471 12 0.594** 0.591 0.744 0.665 0.703 0.659 13 0.543** 0.584 0.753 0.769** 0.752** 0.680 14 0.676** 0.796** 0.697 0.657 0.747** 0.715 15 0.630 0.676** 0.736 0.668 0.699** 0.682 16 0.518 0.727** 0.767** 0.462** 0.807** 0.656 17 0.552 0.429** 0.702** 0.612** 0.718** 0.603 18 0.609 0.592** 0.744** 0.608** 0.557** 0.622 19 0.647 0.651** 0.857** 0.680** 0.733** 0.714 20 0.651 0.618** 0.667** 0.544** 0.548** 0.606 21 0.451 0.607** 0.531** 0.433** 0.667 0.538 22 0.245 0.323** 0.368** 0.634** 0.769** 0.468 23 0.571 0.695** 0.617** 0.636** 0.486** 0.601 24 0.560 0.501** 0.743** 0.573** 0.476** 0.571 Number of efficient hotels (nonsignificant) 18 5 10 7 7 Arithmetic mean 0.548 0.584 0.679 0.596 0.648 0.611

**Indicate significance at 5 per cent level.

Table 5 Determinants of efficiency.

Simar and Wilson (2007) double bootstrap regression results

Variable Bootstrap efficiency VRS efficiency

Coefficient Bootstrap standard error

Coefficient Bootstrap standard error

Age 0.0005 0.000 0.0050** 0.003 Star 0.0025 0.007 −0.0077 0.017 Size −0.0417** 0.022 −0.0259 0.038 Type 0.0293 0.062 0.0972 0.113 Eco1 0.0016*** 0.000 0.0021*** 0.001 Eco2 −0.0024*** 0.001 −0.0035 0.002 Eco3 0.0018*** 0.001 0.0041*** 0.001 D2010 0.0363 0.027 0.0591 0.042 D2011 0.1302*** 0.028 0.0562 0.043 D2012 0.0465** 0.027 0.0162 0.043 D2013 0.0985*** 0.027 0.0096 0.048 Constant 0.5088*** 0.053 0.4902*** 0.153 Sigma 0.0944 0.006 0.1152 0.011

**and *** indicate statistical significance at 5% and 1% levels respectively. 1Number of bootstrap replications= 1000.

T. Kularatne et al. Tourism Management 71 (2019) 213–225

219

tropical climate conditions in reducing costs. According to Ali, Mustafa, Al-Mashaqbah, Mashal, and Mohsen (2008), electricity used for oper- ating air conditioning systems in a hotel corresponds approximately to 30% or more of total expenditures on energy. It should be noted, however, that given the high initial installation costs, energy efficient systems are relatively easier to acquire for large scale hotels. A study conducted for the New Zealand accommodation sector indicted that there is a substantial gap between the positive perceptions about energy efficiency systems held by operators and the actual levels of im- plementation (Becken, 2013).

Effective waste management contributes positively to a hotel's technical efficiency. Thus, managing food and other solid waste through reducing, reusing and recycling can be financially beneficial. For example, recycling helps to reduce landfill and the cost of waste disposal (Radwan et al., 2012). Also, green purchasing reduces waste at source and increases the potential for reusing and recycling (Min & Galle, 1997).

Reducing water consumption is found to have a negative relation- ship with a hotel's technical efficiency. This is an unexpected result. The literature identifies positive initiatives in water saving in hotel opera- tions (see for example, Warren & Becken, 2017), although no study looks at a direct relationship with water use and a hotel's technical efficiency. Water is a critical resource in the hotel industry associated with guest comfort and therefore controlling water consumption may impact guest perceptions and therefore the demand. It is useful to note that water is a widely available resource in Sri Lanka and the cost is relatively low (than electricity). Moreover, water saving measures are not systematically in place in the hotel industry yet and therefore fur- ther discussion and research in this area is needed to explain the given result.

5.3. Malmquist productivity index

One source to variation over time can relate to the fact that hotels become more efficient or that there is some type of technological in- novation that makes it possible to, overall, use less inputs to produce a given amount of output. Information about efficiency change and technological change over time can be studied by using a Malmquist productivity index. To calculate the level of productivity in the Sri Lankan hotel industry, we use the Malmquist productivity index. The average productivity changes from 2010/11 to 2013/14 for each hotel are presented in Table 7. Table 8 presents the Malmquist productivity, efficiency change and technical change for each hotel averaged over the study period. The detailed results and the estimated confidence inter- vals are given in Appendix B.

A Malmquist index value of greater than 1 indicates a productivity increase whereas a value less than 1 indicates a productivity decline. A value of 1 indicates neither an increase nor a decrease in productivity. The results presented in Table 6 show an average negative productivity growth of 19% over the study period. On average, productivity de- creased by 23% in 2010/11, 17% by 2011/12, 13% by 2012/13 and 22% by 2013/14.

Table 7 represents the decomposition of productivity change with the use of the Malmquist index, into technological change and effi- ciency change. Efficiency change implies that a hotel has moved closer to the existing frontier whereas technical change refers to the changes of the production possibility set. The overall productivity decline of 19% includes a 1% of efficiency change and a 17% of technological change.

Malmquist productivity indices, efficiency change and technical change are presented in Table 8 for each year. We observe a −4.2% efficiency change in 2010/11 followed by a 2.1% in 2011/12, 4.2% in 2012/13 and 1.6% in 2013/14. Technical change is observed to be −20% for 2010/11, −19.2% for 2011/12, 16.7% for 2012/13 and 20% for 2013/14.

6. Concluding remarks

This paper measures the technical efficiency of a sample of medium and large scale hotels in Sri Lanka. It is noted that Sri Lanka is one of the most important tourist destinations in Asia and is experiencing a post war tourism boom. Thus, a performance evaluation of the hotel in- dustry is found to be an important factor for the development of the tourism industry. By applying the non-parametric DEA approach we

Table 6 Productivity development using the Malmquist productivity index.

Hotel 2010/2011. 2011/2012 2012/2013 2013/2014

1 0.762** 0.893** 0.969 0.807** 2 0.635** 0.753** 0.792** 0.700** 3 0.745** 1.008 0.843** 0.723** 4 0.933** 0.623** 0.562** 0.383** 5 0.800** 0.802** 0.894** 0.844** 6 0.794** 0.916** 0.859** 0.923** 7 0.743** 1.066 0.990 0.535** 8 0.624** 0.777** 1.031 0.734** 9 0.728** 0.813** 0.851** 1.053 10 0.732** 0.793** 0.974 0.638** 11 0.621** 0.740** 0.985 0.993 12 0.767** 1.393** 0.878** 0.812** 13 0.787** 0.940** 0.939** 0.871** 14 0.827** 0.686** 0.942** 0.893** 15 0.789** 0.810** 0.832** 0.858** 16 0.516** 0.922** 1.095 0.672** 17 1.028 0.537** 0.910** 0.824** 18 0.767** 0.740** 0.824** 0.791** 19 0.757** 0.568** 0.779** 0.432** 20 1.193** 0.734** 0.863** 0.789** 21 0.660** 0.855** 0.912** 0.674** 22 0.680** 0.738** 0.444** 0.422** 23 0.655** 0.881** 0.74** 1.066** 24 0.874** 0.755** 0.953 1.153** Arithmetic mean 0.767 0.823 0.869 0.775

**indicate significance at 5 per cent level.

Table 7 Malmquist productivity, efficiency change & technical change averaged over the study period.a.

Hotel Malmquist productivity

Efficiency change

Technical change

hotel 1 0.866 1.000 0.849 hotel 2 0.720 0.959 0.751 hotel 3 0.828 1.000 0.867 hotel 4 0.625 0.773 0.811 hotel 5 0.835 1.000 0.901 hotel6 0.873 1.000 1.000 hotel 7 0.820 0.916 0.865 hotel 8 0.784 1.007 0.801 hotel 9 0.848 1.064 0.850 hotel10 0.791 0.955 0.790 hotel11 0.840 1.106 0.761 hotel12 0.963 1.131 0.942 hotel13 0.884 1.000 1.000 hotel14 0.837 1.000 0.767 hotel15 0.822 1.000 0.900 hotel16 0.778 0.987 0.752 hotel17 0.818 1.005 0.761 hotel18 0.781 1.000 0.741 hotel19 0.634 0.846 0.761 hotel20 0.895 1.000 0.812 hotel21 0.775 0.995 0.808 hotel22 0.571 0.748 0.829 hotel23 0.836 1.142 0.806 hotel24 0.946 1.096 0.842 Arithmetic mean 0.807 0.989 0.832

a If the value is insignificant, replaced by a value of 1.000.

T. Kularatne et al. Tourism Management 71 (2019) 213–225

220

estimate the output oriented technical efficiency for the period 2010–2014 of a sample of 24 hotels using the CRS and VRS models. This study also uses the DEA double bootstrap method in order to correct for bias in the estimation of technical efficiency. Overall, the average technical efficiency score (bias-corrected) of Sri Lankan hotels is found to be 61% for the study period.

The results reveal that medium and large scale hotels in Sri Lanka operate at a level of 39% inefficiency or in other words, hotels on average have 39% resources wastage. In the face of the tourism boom, there is a clear need for hotels to improve their efficiency by which production outputs are created in order to reach an optimum scale of production. Given there are no previous studies examining the level of efficiency in Sri Lankan hotels, future research is warranted to confirm these findings.

In the second stage of our research, we investigate the determinants of efficiency. We found that the hotel's age has a positive relationship with efficiency. Size demonstrates a negative impact on efficiency. The results suggest that, expansion of operations by large hotels can lead to diseconomies of scale. More importantly, we investigate the notion that ‘going green’ largely has a positive impact on improving technical ef- ficiency of hotels. The criteria used to measure green practices include energy and water consumption, waste management and other opera- tional practices. Results reveal that, while energy efficiency and waste management measures improve hotels' technical efficiency, reducing water consumption cause inefficiencies. On the other hand, a hotel's environmental initiatives in relation to energy and water management helps to improve their competitiveness, save money and attract en- vironmentally concerned customers. We identify that water is a critical component of guest comfort and therefore may impact guest percep- tions and hence, the demand.

The success of a hotel as a hospitality services provider is largely dependent on its energy and water consumption (that is, they are re- quired to provide facilities with a high degree of comfort such as air- conditioning). Justifying a hotel's energy and water consumption levels can be highly contextual depending on climatic conditions and con- sumer demand. Given Sri Lanka has a warm tropical climate with average yearly temperatures ranging from 28 °C to nearly 31 °C, a

hotel's energy and water consumption can be relatively high. Therefore, analysing the impact of eco variables on efficiency needs to be based on a complete understanding of hotel operations of Sri Lanka. Moreover, exploring new types of technology that facilitates eco-friendly con- sumption initiatives that also ensure guest satisfaction may be neces- sary for the future of green hotel operations.

From the perspective of individual hotels, managerial implications for increasing efficiency indicate the importance of maximising total revenue of hotels. At a strategic level, hotels need to learn that maturity in terms of age and size is likely to have an impact on levels of effi- ciency (diseconomies of scale). From a national perspective, the results are important in confirming that the government initiatives of adopting environmentally sustainable practices for hotels are essentially bene- ficial.

As an extension to this analysis, we estimated the bootstrapped Malmquist productivity index for the study period. The results found that a large number of medium and large scale hotels in Sri Lanka did not achieve significant progress in terms of productivity. The industry was found to experience a negative productivity growth of approxi- mately 19% per year. A year-by-year comparison shows little im- provement in productivity from 2010 to 2014. One problem is that the room capacity of Sri Lankan hotels after the post war tourism devel- opment stage is still inadequate to cater to anticipated tourism demand. Additionally, room rates have increased considerably during the post war tourism boom and are now commonly known to be over-priced compared to other Asian destinations. An emphasis on providing greater room capacity and competitive prices would therefore help improve overall productivity of the industry in the future.

Although this research provides interesting insights for the devel- opment of the hotel industry, it is important to identify its limitations. Firstly, due to data restrictions, the number of DMUs used in this study was small. As a result, we could only use a limited range of input/ output factors in the performance evaluation. Secondly, it would be interesting to include variables such as tourist satisfaction (Hathroubi et al., 2014) or service quality so as to undertake a more comprehensive analysis. Thirdly, the sample used for the analysis is taken from Sri Lanka thus the generalisability of the results remains to be tested. Fu- ture research could therefore usefully be carried out to test these con- clusions on a wider international basis.

Author contributions

The contributions of the authors to the paper is as follows:

Dr Thamarasi Kularatne=45 Professor Clevo Wilson= 20 Dr Jonas Månsson= 15 Dr Boon Lee= 10 Dr Vincent Hoang=10

Appendix A. Bias corrected technical efficiency scores (input-oriented) for Sri Lankan hotel (2010–2014)

Year 2010 2011 2012

Hotel Efficiency Upper bound Lower bound Hotel Efficiency Upper bound Lower bound Hotel Efficiency Upper bound

1 0.624 1.042 0.464 1 0.657 1.006 0.504 1 0.811 1.145 2 0.725 1.148 0.597 2 0.676** 0.938 0.448 2 0.753 1.073 3 0.537** 0.632 0.167 3 0.591 1.262 0.279 3 0.699 1.653 4 0.527 1.524 0.429 4 0.363** 0.452 0.302 4 0.598** 0.704 5 0.601** 0.925 0.297 5 0.687** 0.948 0.467 5 0.731 1.214 6 0.534** 0.649 0.161 6 0.59 1.297 0.274 6 0.697 1.683 7 0.5 1.617 0.4 7 0.559** 0.709 0.452 7 0.563** 0.69 8 0.418 1.726 0.319 8 0.514** 0.601 0.458 8 0.59** 0.719 9 0.547 1.401 0.439 9 0.597** 0.789 0.478 9 0.823 1.016

Table 8 Estimated Malmquist productivity, efficiency change and technical change by year.a.

Year Malmquist productivity Efficiency change Technical change

2010/11 0.766 0.958 0.800 2011/12 0.820 1.021 0.808 2012/13 0.867 1.042 0.833 2013/14 0.773 1.016 0.797

a If the value is insignificant, replaced by a value of 1.000.

T. Kularatne et al. Tourism Management 71 (2019) 213–225

221

10 0.469 1.765 0.384 10 0.462** 0.622 0.364 10 0.536** 0.649 11 0.414 2.007 0.343 11 0.54** 0.612 0.488 11 0.56** 0.665 12 0.594** 0.931 0.28 12 0.591 1.356 0.277 12 0.744 1.151 13 0.543** 0.626 0.178 13 0.584 1.072 0.259 13 0.753 1.096 14 0.676** 0.911 0.444 14 0.796** 0.97 0.694 14 0.697 1.662 15 0.63 0.96 0.355 15 0.676** 0.986 0.449 15 0.736 1.187 16 0.518 1.633 0.446 16 0.727** 0.893 0.61 16 0.767** 0.917 17 0.552 1.413 0.452 17 0.429** 0.522 0.363 17 0.702** 0.903 18 0.609 1.384 0.525 18 0.592** 0.719 0.507 18 0.744** 0.883 19 0.647 1.289 0.545 19 0.651** 0.787 0.557 19 0.857** 0.995 20 0.651 1.027 0.396 20 0.618** 0.727 0.548 20 0.667** 0.782 21 0.451 1.786 0.373 21 0.607** 0.707 0.547 21 0.531** 0.618 22 0.245 2.857 0.182 22 0.323** 0.46 0.25 22 0.368** 0.437 23 0.571 1.423 0.468 23 0.695** 0.807 0.626 23 0.617** 0.762 24 0.56 1.461 0.455 24 0.501** 0.631 0.418 24 0.743** 0.847

Year 2012 2013 2014

Hotel Lower bound Hotel Efficiency Upper bound Lower bound Hotel Efficiency Upper bound Lower bound

1 0.684 1 0.637** 0.872 0.511 1 0.719** 0.989 0.584 2 0.533 2 0.637 1.093 0.363 2 0.673 1.14 0.424 3 0.428 3 0.627 1.207 0.344 3 0.666 1.238 0.408 4 0.526 4 0.66** 0.813 0.565 4 0.793 1.01 0.657 5 0.49 5 0.65 1.092 0.388 5 0.708 1.065 0.488 6 0.422 6 0.627 1.204 0.345 6 0.662 1.267 0.401 7 0.49 7 0.394** 0.515 0.324 7 0.623** 0.744 0.542 8 0.518 8 0.492** 0.627 0.416 8 0.545** 0.651 0.479 9 0.718 9 0.737** 0.916 0.613 9 0.624** 0.713 0.568 10 0.471 10 0.428** 0.552 0.354 10 0.505** 0.614 0.428 11 0.501 11 0.466** 0.551 0.408 11 0.375** 0.454 0.322 12 0.516 12 0.665 1.047 0.417 12 0.703 1.058 0.481 13 0.532 13 0.769** 0.963 0.63 13 0.752** 0.942 0.625 14 0.423 14 0.657 1.081 0.402 14 0.747** 0.975 0.569 15 0.5 15 0.668 1.12 0.423 15 0.699** 0.994 0.567 16 0.681 16 0.462** 0.541 0.409 16 0.807** 0.972 0.688 17 0.585 17 0.612** 0.816 0.487 17 0.718** 0.892 0.612 18 0.662 18 0.608** 0.779 0.498 18 0.557** 0.691 0.474 19 0.744 19 0.68** 0.935 0.452 19 0.733** 0.963 0.541 20 0.589 20 0.544** 0.675 0.451 20 0.548** 0.687 0.459 21 0.467 21 0.433** 0.525 0.364 21 0.667 1.238 0.409 22 0.327 22 0.634** 0.823 0.514 22 0.769** 0.986 0.612 23 0.535 23 0.636** 0.788 0.533 23 0.486** 0.606 0.403 24 0.681 24 0.573** 0.686 0.48 24 0.476** 0.548 0.426

Appendix B. Bootstrapped Malmquist productivity, efficiency change and technical change for Sri Lankan hotels

Years Hotel Malmquist Productivity

Lower bound

Upper bound

Efficiency change

Lower bound

Upper bound

Technical Change

Lower bound

Upper bound

2010_11 1 0.762** 0.740 0.811 0.961 0.748 1.112 0.793 0.664 0.939 2010_11 2 0.635** 0.558 0.702 0.837 0.630 0.939 0.758 0.622 0.893 2010_11 3 0.745** 0.693 0.832 1.000 0.665 1.209 0.745 0.545 0.918 2010_11 4 0.933** 0.933 0.933 1.124 0.735 1.181 0.831 0.786 1.044 2010_11 5 0.800** 0.719 0.825 1.000 0.690 1.166 0.800 0.642 0.947 2010_11 6 0.794** 0.770 0.806 1.000 0.600 1.208 0.794 0.590 1.011 2010_11 7 0.743** 0.721 0.752 0.899 0.735 1.034 0.826 0.689 0.958 2010_11 8 0.624** 0.621 0.634 0.759 0.480 0.821 0.822 0.752 1.045 2010_11 9 0.728** 0.725 0.801 0.898 0.669 1.038 0.810 0.682 0.998 2010_11 10 0.732** 0.661 0.765 0.966 0.796 1.147 0.758 0.602 0.850 2010_11 11 0.621** 0.600 0.628 0.753 0.544 0.789 0.824 0.777 0.995 2010_11 12 0.767** 0.614 0.852 1.000 0.711 1.161 0.767 0.624 0.927 2010_11 13 0.787** 0.748 0.841 1.000 0.593 1.222 0.787 0.599 1.008 2010_11 14 0.827** 0.779 0.903 1.036 0.816 1.128 0.798 0.739 0.957 2010_11 15 0.789** 0.713 0.858 1.000 0.687 1.163 0.789 0.655 0.964

T. Kularatne et al. Tourism Management 71 (2019) 213–225

222

2010_11 16 0.516** 0.463 0.568 0.675 0.532 0.751 0.765 0.663 0.909 2010_11 17 1.028 0.965 1.092 1.309 1.017 1.423 0.786 0.713 0.932 2010_11 18 0.767** 0.686 0.857 0.987 0.777 1.083 0.777 0.683 0.939 2010_11 19 0.757** 0.756 0.840 0.934 0.744 1.020 0.811 0.764 0.978 2010_11 20 1.193** 1.156 1.267 1.341 0.786 1.460 0.890 0.839 1.147 2010_11 21 0.660** 0.658 0.660 0.779 0.568 0.830 0.847 0.788 1.028 2010_11 22 0.680** 0.647 0.740 0.843 0.528 0.949 0.807 0.713 1.033 2010_11 23 0.655** 0.643 0.670 0.794 0.499 0.862 0.826 0.753 1.051 2010_11 24 0.874** 0.824 0.895 1.106 0.822 1.254 0.791 0.661 0.951 2011_12 1 0.893** 0.870 0.916 0.928 0.702 1.034 0.963 0.834 1.148 2011_12 2 0.753** 0.697 0.794 1.000 0.815 1.261 0.753 0.494 0.848 2011_12 3 1.008 0.981 1.048 1.000 0.609 1.178 1.008 0.800 1.291 2011_12 4 0.623** 0.598 0.631 0.837 0.719 0.969 0.744 0.610 0.825 2011_12 5 0.802** 0.798 0.807 1.000 0.741 1.213 0.802 0.590 0.965 2011_12 6 0.916** 0.836 0.921 1.000 0.681 1.217 0.916 0.618 1.104 2011_12 7 1.066 0.915 1.211 1.292 0.992 1.605 0.825 0.642 0.957 2011_12 8 0.777** 0.637 0.795 1.093 0.952 1.262 0.711 0.534 0.762 2011_12 9 0.813** 0.806 0.922 0.837 0.658 0.973 0.972 0.835 1.168 2011_12 10 0.793** 0.625 0.839 1.025 0.766 1.177 0.773 0.601 0.863 2011_12 11 0.740** 0.689 0.742 1.023 0.900 1.194 0.723 0.564 0.787 2011_12 12 1.393** 1.079 1.671 1.523 1.082 1.818 0.914 0.774 1.105 2011_12 13 0.940** 0.916 0.963 1.000 0.618 1.176 0.940 0.751 1.194 2011_12 14 0.686** 0.648 0.712 0.965 0.856 1.145 0.710 0.552 0.769 2011_12 15 0.810** 0.713 0.923 1.000 0.773 1.264 0.810 0.593 0.929 2011_12 16 0.922** 0.849 0.940 1.274 1.103 1.563 0.724 0.490 0.811 2011_12 17 0.537** 0.493 0.602 0.712 0.672 0.842 0.754 0.586 0.822 2011_12 18 0.740** 0.708 0.755 1.052 0.932 1.317 0.704 0.456 0.779 2011_12 19 0.568** 0.474 0.578 0.796 0.662 0.905 0.714 0.577 0.771 2011_12 20 0.734** 0.721 0.764 1.010 0.909 1.177 0.727 0.582 0.799 2011_12 21 0.855** 0.792 0.855 1.200 1.076 1.380 0.712 0.582 0.769 2011_12 22 0.738** 0.634 0.751 0.919 0.697 1.042 0.803 0.650 0.918 2011_12 23 0.881** 0.809 0.920 1.230 1.084 1.500 0.716 0.508 0.783 2011_12 24 0.755** 0.708 0.755 0.780 0.586 0.859 0.967 0.849 1.148 2012_13 1 0.969 0.938 1.009 1.153 0.948 1.331 0.840 0.696 0.971 2012_13 2 0.792** 0.774 0.813 1.000 0.664 1.313 0.792 0.455 0.986 2012_13 3 0.843** 0.771 0.866 1.000 0.693 1.233 0.843 0.565 1.017 2012_13 4 0.562** 0.544 0.594 0.734 0.592 0.897 0.765 0.576 0.891 2012_13 5 0.894** 0.850 0.920 1.000 0.675 1.202 0.894 0.667 1.090 2012_13 6 0.859** 0.839 0.882 1.000 0.640 1.288 0.859 0.532 1.083 2012_13 7 0.990 0.774 1.035 1.095 0.623 1.238 0.903 0.739 1.079 2012_13 8 1.031 0.999 1.046 1.267 1.098 1.477 0.813 0.651 0.909 2012_13 9 0.851** 0.727 0.858 1.006 0.781 1.156 0.846 0.662 0.977 2012_13 10 0.974 0.933 1.030 1.148 0.922 1.356 0.848 0.676 0.986 2012_13 11 0.985 0.948 1.001 1.208 1.042 1.412 0.815 0.642 0.910 2012_13 12 0.878** 0.839 0.941 0.962 0.719 1.098 0.913 0.763 1.112 2012_13 13 0.939** 0.809 0.944 1.022 0.617 1.191 0.918 0.705 1.126 2012_13 14 0.942** 0.912 0.952 1.145 0.901 1.327 0.822 0.664 0.961 2012_13 15 0.832** 0.790 0.862 1.000 0.624 1.284 0.832 0.523 1.036 2012_13 16 1.095 0.765 1.141 1.341 0.914 1.519 0.817 0.656 0.931 2012_13 17 0.910** 0.880 0.945 1.141 0.944 1.366 0.798 0.621 0.906 2012_13 18 0.824** 0.687 0.834 1.022 0.754 1.214 0.806 0.565 0.954 2012_13 19 0.779** 0.765 0.896 0.998 0.886 1.255 0.780 0.621 0.873 2012_13 20 0.863** 0.811 0.871 1.086 0.899 1.267 0.795 0.619 0.902 2012_13 21 0.912** 0.896 0.950 1.105 0.959 1.291 0.826 0.677 0.922 2012_13 22 0.444** 0.394 0.472 0.553 0.437 0.636 0.803 0.646 0.917 2012_13 23 0.740** 0.658 0.744 0.907 0.722 1.081 0.817 0.588 0.941 2012_13 24 0.953 0.870 1.089 1.112 0.932 1.315 0.857 0.719 0.976 2013–14 1 0.807** 0.776 0.821 1.057 0.835 1.191 0.764 0.652 0.892 2013–14 2 0.700** 0.670 0.725 1.000 0.454 1.299 0.700 0.424 0.938 2013–14 3 0.723** 0.720 0.731 1.000 0.615 1.255 0.723 0.482 0.923 2013–14 4 0.383** 0.381 0.385 0.521 0.379 0.600 0.736 0.603 0.893 2013–14 5 0.844** 0.802 0.862 1.000 0.461 1.183 0.844 0.661 1.114 2013–14 6 0.923** 0.867 0.958 1.000 0.350 1.262 0.923 0.612 1.264 2013–14 7 0.535** 0.526 0.558 0.664 0.478 0.737 0.807 0.713 0.995 2013–14 8 0.734** 0.697 0.737 1.077 0.865 1.259 0.681 0.531 0.788 2013–14 9 1.053 0.955 1.099 1.420 1.127 1.646 0.742 0.598 0.843

T. Kularatne et al. Tourism Management 71 (2019) 213–225

223

2013–14 10 0.638** 0.577 0.655 0.819 0.563 0.915 0.779 0.666 0.952 2013–14 11 0.993 0.988 1.048 1.461 1.203 1.771 0.680 0.517 0.790 2013–14 12 0.812** 0.799 0.861 0.972 0.668 1.135 0.836 0.665 1.050 2013–14 13 0.871** 0.738 0.918 1.070 0.590 1.260 0.813 0.655 1.016 2013–14 14 0.893** 0.867 0.896 1.211 0.907 1.412 0.737 0.586 0.882 2013–14 15 0.858** 0.761 0.925 1.123 0.502 1.385 0.764 0.565 1.026 2013–14 16 0.672** 0.645 0.702 0.957 0.778 1.140 0.702 0.538 0.813 2013–14 17 0.824** 0.780 0.830 1.166 0.884 1.355 0.707 0.558 0.834 2013–14 18 0.791** 0.716 0.794 1.170 0.828 1.404 0.676 0.492 0.820 2013–14 19 0.432** 0.418 0.450 0.587 0.447 0.688 0.737 0.595 0.887 2013–14 20 0.789** 0.762 0.801 1.085 0.819 1.261 0.727 0.584 0.858 2013–14 21 0.674** 0.672 0.736 0.971 0.837 1.200 0.694 0.512 0.790 2013–14 22 0.422** 0.414 0.434 0.594 0.476 0.710 0.710 0.543 0.828 2013–14 23 1.066** 1.035 1.163 1.545 1.198 1.903 0.690 0.499 0.831 2013–14 24 1.153** 1.105 1.172 1.602 1.297 1.893 0.720 0.555 0.825

References

Ali, Y., Mustafa, M., Al-Mashaqbah, S., Mashal, K., & Mohsen, M. (2008). Potential of energy savings in the hotel sector in Jordan. Energy Conversion and Management, 49(11), 3391–3397.

Anderson, R. I., Fok, R., & Scott, J. (2000). Hotel industry efficiency: An advanced linear programming examination. American Business Review, 18(1), 40–48.

Assaf, A. G., & Agbola, F. W. (2011). Modelling the performance of australian hotels: A DEA double bootstrap approach. Tourism Economics, 17(1), 73–89.

Assaf, A., & Barros, C. P. (2011). Performance analysis of the gulf hotel industry: A malmquist index with bias correction. International Journal of Hospitality Management, 30(4), 819–826.

Assaf, A., Barros, C. P., & Josiassen, A. (2012). Hotel efficiency: A bootstrapped meta- frontier approach. International Journal of Hospitality Management, 31(2), 621–629.

Atkinson, S., & Wilson, P. (1995). Comparing mean efficiency and productivity scores from small samples: A bootstrap methodology. Journal of Productivity Analysis, 6(2), 137–152.

Banker, R. D., Charnes, A., & Cooper, W. W. (1984). Some models for estimating technical and scale inefficiencies in data envelopment analysis. Management Science, 30(9), 1078.

Barros, C. P. (2005a). Evaluating the efficiency of a small hotel chain with a Malmquist productivity index. International Journal of Tourism Research, 7(3), 173–184.

Barros, C. P. (2005b). Measuring efficiency in the hotel sector. Annals of Tourism Research, 32(2), 456–477.

Barros, C. P. (2006). Analysing the rate of technical change in the Portuguese hotel in- dustry. Tourism Economics, 12, 325–346.

Barros, C. P., & Dieke, P. U. C. (2008). Technical efficiency of African hotels. International Journal of Hospitality Management, 27(3), 438–447.

Barros, C. A. P., & Santos, C. A. (2006). The measurement of efficiency in Portuguese hotels using data envelopment analysis. Journal of Hospitality & Tourism Research, 30(3), 378–400.

Becken, S. (2013). Operators' perceptions of energy use and actual saving opportunities for tourism accommodation. Asia Pacific Journal of Tourism Research, 18(1–2), 72–91.

Ben Aissa, S., & Goaied, M. (2016). Determinants of Tunisian hotel profitability: The role of managerial efficiency. Tourism Management, 52, 478–487.

Bohdanowicz, P. (2006). Environmental awareness and initiatives in the Swedish and Polish hotel industries—survey results. International Journal of Hospitality Management, 25(4), 662–682.

Brown, J. R., & Ragsdale, C. T. (2002). The competitive market efficiency of hotel brands: An applications of data envelopment analysis. Journal of Hospitality & Tourism Research, 26(4), 332–360.

Chan, E. S. W. (2013). Gap analysis of green hotel marketing. International Journal of Contemporary Hospitality Management, 25(7), 1017–1048.

Chan, E. S. W., & Wong, S. C. K. (2006). Motivations for ISO 14001 in the hotel industry. Tourism Management, 27(3), 481–492.

Charnes, A., Cooper, W. W., & Rhodes, E. (1978). Measuring the efficiency of decision making units. European Journal of Operational Research, 2(6), 429–444.

Chen, C.-F. (2007). Applying the stochastic frontier approach to measure hotel manage- rial efficiency in Taiwan. Tourism Management, 28(3), 696–702.

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(3), 415–422.

Chen, S., Chen, H. H., Zhang, K. Q., & Xu, X.-l. (2018). A comprehensive theoretical framework for examining learning effects in green and conventionally managed ho- tels. Journal of Cleaner Production, 174, 1392–1399.

Chiang, W.-E., Tsai, M.-H., & Wang, L. S.-M. (2004). A DEA evaluation of Taipei hotels. Annals of Tourism Research, 31(3), 712–715.

Coelli, T., Prasada Rao, D. S., O'Donnell, C. J., & Battese, G. E. (1998). An introduction to efficiency and productivity analysis. Boston: Kluwer Academic Publishers.

Dyson, R. G., Allen, R., Camanho, A. S., Podinovski, V. V., Sarrico, C. S., & Shale, E. A. (2001). Pitfalls and protocols in DEA. European Journal of Operational Research, 132(2), 245–259.

Egilmez, G., & Park, Y. S. (2014). Transportation related carbon, energy and water footprint analysis of U.S. manufacturing: An eco-efficiency assessment. Transportation Research Part D: Transport and Environment, 32, 143–159.

Evanoff, D. D., & Israilevich, P. R. (1991). Productive efficiency in banking. Economic Perspectives, 15(4), 11–32.

Färe, R., Grosskopf, S., Lindgren, B., & Roos, P. (1992). Productivity changes in Swedish pharamacies 1980–1989: A non-parametric malmquist approach. Journal of Productivity Analysis, 3(1), 85–101.

Färe, R., Grosskopf, S., Norris, M., & Zhang, Z. (1994). Productivity growth, technical progress, and efficiency change in industrialized countries. The American Economic Review, 84(1), 66–83.

Farrell, M. J. (1957). The measurement of productive efficiency. Journal of the Royal Statistical Society: Series A, 120(3), 253–290.

Fernández, M. A., & Becerra, R. (2015). An analysis of Spanish Hotel efficiency. Cornell Hospitality Quarterly, 56(3), 248–257.

Green Hotels Association. (2014). Retrieved 18-02-2014, from http://www.greenhotels. com/.

Halkos, G., & Matsiori, S. (2018). Environmental attitudes and preferences for coastal zone improvements. Economic Analysis and Policy, 58, 153–166.

Han, H., & Hyun, S. S. (2018). What influences water conservation and towel reuse practices of hotel guests? Tourism Management, 64, 87–97.

Han, H., & Yoon, H. J. (2015). Hotel customers' environmentally responsible behavioral intention: Impact of key constructs on decision in green consumerism. International Journal of Hospitality Management, 45, 22–33.

Hathroubi, S., Peypoch, N., & Robinot, E. (2014). Technical efficiency and environmental management: The Tunisian case. Journal of Hospitality and Tourism Management, 21, 27–33.

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(1), 14–24.

Huang, C.-W., Ho, F. N., & Chiu, Y.-h. (2014). Measurement of tourist hotels׳ productive efficiency, occupancy, and catering service effectiveness using a modified two-stage DEA model in Taiwan. Omega, 48, 49–59.

Hwang, S.-N., & Chang, T.-Y. (2003). Using data envelopment analysis to measure hotel managerial efficiency change in Taiwan. Tourism Management, 24(4), 357–369.

IFC (2013). Ensuring sustainability in Sri Lanka's growing hotel industry. Retrieved from http://www.ifc.org/wps/wcm/connect/publications_ext_content/ifc_external_ publication_site/publications/all+publications.

Kang, K. H., Stein, L., Heo, C. Y., & Lee, S. (2012). Consumers' willingness to pay for green initiatives of the hotel industry. International Journal of Hospitality Management, 31(2), 564–572.

Kasim, A. (2004). Socio-environmentally responsible hotel business: Do tourists to Penang island, Malaysia care? Journal of Hospitality & Leisure Marketing, 11(4), 5–28.

Kim, W. G., Li, J. J., Han, J. S., & Kim, Y. (2016). The influence of recent hotel amenities and green practices on guests' price premium and revisit intention. Tourism Economics, 1–17. http://journals.sagepub.com.ezp01.library.qut.edu.au/doi/pdf/10. 5367/te.2015.0531.

Kumar Mandal, S., & Madheswaran, S. (2010). Environmental efficiency of the Indian cement industry: An interstate analysis. Energy Policy, 38(2), 1108–1118.

Laroche, M., Bergeron, J., & Barbaro-Forleo, G. (2001). Targeting consumers who are willing to pay more for environmentally friendly products. Journal of Consumer Marketing, 18(6), 503–520.

Lee, J.-S., Hsu, L.-T., Han, H., & Kim, Y. (2010). Understanding how consumers view green hotels: How a hotel's green image can influence behavioural intentions. Journal of Sustainable Tourism, 18(7), 901–914.

Mair, J., & Jago, L. (2010). The development of a conceptual model of greening in the business events tourism sector. Journal of Sustainable Tourism, 18(1), 77–94.

Mensah, I. (2006). Environmental management practices among hotels in the greater Accra region. International Journal of Hospitality Management, 25(3), 414–431.

Min, H., & Galle, W. P. (1997). Green purchasing strategies: Trends and implications. International Journal of Purchasing and Materials Management, 33(3), 10–17.

Morey, R. C., & Dittman, D. A. (1995). Evaluating a hotel GM's performance. Cornell Hotel and Restaurant Administration Quarterly, 36(5), 30–35.

T. Kularatne et al. Tourism Management 71 (2019) 213–225

224

Ndebele, T., & Forgie, V. (2017). Estimating the conomic benefits of a wetland restoration programme in New Zealand: A contingent valuation approach. Economic Analysis and Policy, 55, 75–89.

Oliveira, R., Pedro, M. I., & Marques, R. C. (2013). Efficiency and its determinants in Portuguese hotels in the Algarve. Tourism Management, 36, 641–649.

Oukil, A., Channouf, N., & Al-Zaidi, A. (2016). Performance evaluation of the hotel in- dustry in an emerging tourism destination: The case of Oman. Journal of Hospitality and Tourism Management, 29, 60–68.

Penny, W. Y. K. (2007). The use of environmental management as a facilities management tool in the Macao hotel sector. Facilities, 25(7/8), 286–295.

Perrigot, R., Cliquet, G., & Piot-Lepetit, I. (2009). Plural form chain and efficiency: Insights from the French hotel chains and the DEA methodology. European Management Journal, 27(4), 268–280.

Pine, R., & Phillips, P. (2005). Performance comparisons of hotels in China. International Journal of Hospitality Management, 24(1), 57–73.

Radwan, H. R. I., Jones, E., & Minoli, D. (2012). Solid waste management in small hotels: A comparison of green and non-green small hotels in Wales. Journal of Sustainable Tourism, 20(4), 533–550.

Reynolds, D. (2003). Hospitality-productivity assessment using data-envelopment ana- lysis. Cornell Hotel and Restaurant Administration Quarterly, 44(2), 130–137.

Romano, G., & Guerrini, A. (2011). Measuring and comparing the efficiency of water utility companies: A data envelopment analysis approach. Utilities Policy, 19(3), 202–209.

Shang, J.-K., Hung, W.-T., Lo, C.-F., & Wang, F.-C. (2008). Ecommerce and hotel per- formance: Three-stage DEA analysis. Service Industries Journal, 28(4), 529–540.

Shang, J.-K., Wang, F.-C., & Hung, W.-T. (2009). A stochastic DEA study of hotel effi- ciency. Applied Economics, 42(19), 2505–2518.

Shieh, H.-S. (2012). The greener, the more cost efficient? An empirical study of inter- national tourist hotels in Taiwan. The International Journal of Sustainable Development and World Ecology, 19(6), 536–545.

Simar, L., & Wilson, P. W. (1999). Estimating and bootstrapping Malmquist indices. European Journal of Operational Research, 115(3), 459–471.

Simar, L., & Wilson, P. W. (2007). Estimation and inference in two-stage, semi-parametric models of production processes. Journal of Econometrics, 136(1), 31–64.

Sri Lanka Tourism Development Authority. (2015). Annual statistical reportColombo: SLTDA.

Szuchnicki, A. L. (2009). Examining the influence of restaurant green practices on customer return intentionUNLV Theses, Dissertations, Professional Papers, and Capstones. Paper 155. Retrieved fromhttp://digitalscholarship.unlv.edu/thesesdissertations/155.

Tarim, Ş., Dener, H. I., & Tarim, Ş. A. (2000). Efficiency measurement in the hotel in- dustry: Output factor constrained DEA application. Anatolia, 11(2), 111–123.

Torres, M., & Morrison, P. C. (2006). Driving forces for consolidation or fragmentation of the US water utility industry: A cost function approach with endogenous output. Journal of Urban Economics, 59 104e120.

Tsaur, S. H. (2001). The operating efficiency of international tourist hotels in Taiwan. Asia Pacific Journal of Tourism Research, 6(1), 73–81.

Tsaur, S. H., Chiang, C. I., & Chang, T. Y. (1999). Evaluating the operating efficiency of international tourist hotels using the modified DEA model. Asia Pacific Journal of Tourism Research, 4(2), 73–78.

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.

Warren, C., & Becken, S. (2017). Saving energy and water in tourist accommodation: A systematic literature review (1987–2015). International Journal of Tourism Research, 19(3), 289–303.

World Travel and Tourism Council (2016). Travel and tourism economic impact 2014. Retrieved from https://www.wttc.org/.

Yi, S., Li, X., & Jai, T.-M. (2016). Hotel guests' perception of best green practices: A content analysis of online reviews. Tourism and Hospitality Research, 18(2), 191–202.

Professor Clevo Wilson specializes in tourism, environ- mental, ecological, agricultural, transport, energy, and de- velopment economics with a special interest in using en- vironmental valuation techniques, both revealed and stated. His research also focus on supply chain value ana- lysis, efficiency and productivity analysis, structural equa- tion modelling, and cost benefit analyses. He has under- taken research and published papers in diverse topics including tourism (with a special focus on nature-based tourism and economic value analysis), aquaculture, energy and water conservation, agriculture, transport, natural disasters (floods and wildfires), impact of major projects and pollution on property values, environmental sustain- ability and conservation of wildlife. He has been involved in

25 + major surveys and interviews with extensive data collection and analysis, including

Sri Lanka and Australia since 1996. Of these, 10 + studies have been in relation to wildlife tourism and management with an emphasis on economic value analysis of these resources. In 2012 he co-authored a book entitled Nature-based Tourism and Conservation: New Economic Insights and Case Studies published by Edward Elgar, Cheltenham, UK.

Dr Boon Lee is a senior lecturer in the School of economics and finance at Queensland University of Technology. His research interests focuses on efficiency and productivity analysis and international comparisons of output and pro- ductivity. He has undertaken research and published ex- tensively in the areas of efficiency and productivity ana- lysis, International comparisons of output and productivity and decomposition of income inequality. In 2009, he was invited to participate in a public hearing of the House of Representatives Standing Committee on Economics on in- quiry into raising the level of productivity growth in the Australian economy.

Dr Vincent Hoang is a senior lecturer in the School of economics and finance at Queensland University of Technology. His research interests include compassion and altruism, efficiency and productivity analysis, agricultural, natural resource, environmental economics and ecological economics, energy economics and sustainable business and development. He also acts as chief investigator in two grants funded by Vietnam's Ministry of Agriculture and Rural Development and Japan's Nomura Foundation. Before moving into academia, he had spent five years working in economic, business and public relations con- sulting industry in Vietnam at the positions of Account Supervisors, Account Director and Business Development Director. At these roles, he had worked with many mutli-

national corporations including Unilever, Microsoft, Hewlett-Packard, P&G, ADB, Intel, Coca Cola, Mercedes Benze, etc. In 2006, Vincent set up a public relations firm in Ho Chi Minh City.

Dr Thamarasi Kularatne is a Research Assistant in the School of Economics and Finance at the Queensland University of Technology, Brisbane, Australia. Her primary interest of research includes tourism and environmental economics, non-market valuation techniques and efficiency and productivity analysis.

Dr Jonas Månsson is Associate Professor at Linnaeus University, Växjö, Sweden and affiliated with Thammasat Centre for Efficiency and Productivity Analysis, Bangkok, Thailand. He is also appointed special advisor in Efficiency and productivity analysis at the Swedish National Audit Office. His main areas of research are efficiency, pro- ductivity and policy impacts.

T. Kularatne et al. Tourism Management 71 (2019) 213–225

225