Untapped Demand For Heritage A Contingen PDF

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 13

ARTICLE IN PRESS

Tourism Management 28 (2007) 83–95


www.elsevier.com/locate/tourman

Research Article

Untapped demand for heritage: A contingent valuation study


of Prinsep Ghat, Calcutta
Mousumi Duttaa,, Sarmila Banerjeeb, Zakir Husainc
a
Department of Economics, Bethune College, India
b
Department of Economics, Calcutta University, India
c
Department of Economics, Rabindra Bharati University, India
Received 9 March 2005; accepted 26 July 2005

Abstract

In view of the scarcity of financial resources, urban planners in developing countries are attempting to conserve heritage sites on a self-
financing basis through their development as visitor attraction sites. The economic justification of such programmes can be sought in
terms of the potential revenue from heritage tourism.
This approach, however, neglects the significant non-use values derived from heritage resources. Based on the contingent valuation
method, the Total Economic Value for Prinsep Ghat in Calcutta, India, is estimated using limited dependent regression models. While a
TOBIT model is used to determine the determinants of the willingness to pay, a truncated model is used to identify the characteristics of
the potential clientele. The results show the presence of a substantial untapped demand for the site that can be utilized by suitable
marketing of the site without affecting its historic appeal.
r 2005 Elsevier Ltd. All rights reserved.

Keywords: Asia; India; Built heritage; Contingent valuation; Limited dependent analysis; Urban regeneration

1. Heritage as cultural capital to the gradual decay of built-heritage with an associated


loss of cultural capital.
Many cities in developing countries are cradles of history In recent years, however, urban planners in developing
and culture, abounding in heritage buildings and sites. In countries have recognized the merit of preserving the past.
many cases, they are also areas in which economic, social This realization was greatly influenced by the recent
and environmental problems are concentrated. Urban attempts of developed countries to conserve heritage in a
planners, however, have tended to neglect the issue of number of innovative ways. The emergence of the concept
conserving heritage sites in the urban planning and of ‘adaptive reuse’ has been of crucial importance in this
management process. This is partly due to their pre- context. It calls for a movement away from preservation to
occupation with proper provisioning of essential amenities. conservation.2 This implies a shift of emphasis from
Moreover, legal and institutional barriers1 have also accurate and reverential restoration to a freer and more
contributed to this problem by raising the transaction creative attitude to the changes that an old building may
costs of preserving and maintaining heritage. This has led undergo. Instead of preserving the old heritage building as
an object of art it may now be looked upon as a product
integrated to the whole socio-economic system. Thus
Corresponding author. Residence: 19 Bangur Avenue, Block B, planners should concentrate not only on the architectural,
Calcutta 700 055, India. Tel.: +91 33 2574 5395, 98306 27937(M). cultural and historic merit of a site but view the entire stock
E-mail address: dmousumi@hotmail.com (M. Dutta).
1
For instance, in Calcutta, a city in India, the Tenancy Act has freezed
2
rent to low levels. This reduces the economic incentive to maintain sites. Preservation refers to keeping the heritage in its original form while
Secondly, the act—along with the legal system—make removal of tenants conservation means rebuilding or reusing the heritage site keeping it intact
or even illegal squatters a very costly and time-consuming process. as far as possible (Strike, 1994).

0261-5177/$ - see front matter r 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tourman.2005.07.021
ARTICLE IN PRESS
84 M. Dutta et al. / Tourism Management 28 (2007) 83–95

of heritage as potentially useful for sound economic, social used to estimate Total Economic Value (TEV) of the site.
and ecological purposes by creating opportunities for The last section draws policy conclusions.
urban regeneration based on their possible use and intrinsic
existence value (Cantacuzzino, 1989). 2. Estimating demand for heritage—a review of literature
While restored buildings and sites can be reused in many
ways, an important category of such uses has been to The feasibility of using heritage as a capital can be
utilize them as visitor attraction sites to tap the rapidly studied using two approaches. One category of studies has
burgeoning demand for tourism in general, and heritage attempted to forecast tourism demand using demand
tourism in particular (Cossons, 1989; Davies & Prentice, elasticities (Gunadhi & Boey, 1986) and various versions
1995; Hovinen, 1995; Light & Prentice, 1994; Pearce, 1989). of the ARIMA model (Akal, 2004; Chu, 2004; Goh & Law,
The principal resources for heritage tourism are the 2002). However, these methods are based on the existing
material remains of the past—in the form of historic demand for a site or region that has already been revealed
buildings, archaeological sites, monuments and cultural in the tourism market by visitors to the site.
artefacts (Konrad, 1982; McNulty, 1993). With the The second approach takes a broader perspective. It
expansion of tourism, however, the term heritage is highlights all possible types of benefits accruing from built
extended to include those things inherited from the past heritage. Based on Mitchell and Carson (1989), these
that can be selected for promotion as tourism products benefits can be classified as follows:
(Prentice, 1992) through development of new infrastructure (1) Use Value: (a) Commercial value: Tourism to such
(e.g., the expansion of the hotel and transport sector) and heritage sites can generate income and employment in
new functions, like leisure, amusement, entertainment, ancillary industries like hotel, transport, etc. This flow of
shopping, etc. (Peattie & Moutinho, 2000). income and real estate value comprise of commercial value
However, the problem facing cash strapped policy- of the site. (b) Recreational value is the satisfaction derived
makers in LDCs is whether the huge expenditure involved from leisure-based activities.
in restoring and maintaining heritage sites is economically (2) Non-use value: (a) Existence value: The mere existence
justified. This involves estimating the benefit from conser- of the site affords satisfaction to a person in the form of
ving the site. The total value of conservation consists not deriving pride in his heritage. (b) Vicarious value: This is
only of the revenue generated from promoting heritage the satisfaction derived from leaving open the option to
tourism, but the benefits that can be enjoyed even without another member of his generation to derive direct use value
accessing the site. While the former can be estimated on the from the built heritage. (c) Bequest value: The fact that an
basis of observed market behaviour, the total value can be individual allows the future generation the option to enjoy
estimated using only Stated Preference methods.3 the site also has some value.
This paper focuses on a heritage site of a specific South- (3) In the presence of uncertainty, two types of benefits
Asian city, Calcutta—an old but active metropolitan city in emerge: (a) Option Value: Preservation of the heritage site
India. The site is Prinsep Ghat, a river front structure allows its adaptive reuse in the future. Option value is the
having an important historical and aesthetic value. The site risk premium arising from risk aversion of individuals with
has easy accessibility from different corners of the city and respect to uncertainty about future availability. (b) Quasi-
also from the outskirts. The riverfront is a popular option value: This is the premium paid to prevent an
attraction for Calcutta’s residents. Consequently, it can activity that would have foreclosed a better-informed
be commercially developed utilizing its heritage appeal. decision at a latter time. The underlying assumption is
The paper reports the results of a survey employing that a learning process may enable the agent to obtain a
conjoint choice questions to elicit people’s preferences for new use value from a heritage site.
an urban regeneration project that changes the aesthetic To estimate all these values the TEV (comprising of both
and use character of the heritage site. use and non-use values) must be taken into account.
The paper is organized as follows. In the next section the Recent definitions of conservation have increasingly fo-
existing literature is briefly reviewed. The history of the site cussed on such non-marketed and non-use benefits gener-
is presented in the third section. This is followed by a ated by cultural resources (Mossetto, 1994, pp. 81–84). It is
description of the survey methodology. In the next section, difficult to estimate the total value for this type of urban
motivations underling decision (not) to contribute is regeneration and restoration project because many of the
examined; this is followed by a descriptive and econometric services that they provide to the public—including aesthetic
analysis of the determinants of WTP. These results are then quality, comfort, sense of neighbourhood identity, city
character, preservation of cultural and historical heritage,
access to outdoor space—are non-market benefits (Alberini,
3
The problem, therefore, is being approached from a social perspective Riganti, & Longo, 2003).
and not from that of manager’s. The latter perspective would involve a Environmental economists have shown that, in the
comparison of the costs with the revenue that can be generated from
heritage tourism; from a social perspective, it is the total value of the presence of a strong link between a marketed good and
conservation programme, and not the potential revenue from tourism, the resource being valued, Willingness to Pay (WTP) for an
that is relevant. increase in non-marketed goods/services can be measured
ARTICLE IN PRESS
M. Dutta et al. / Tourism Management 28 (2007) 83–95 85

using revealed preference methods like Travel Cost method Peru and Tanzania. However, many developing Asian
or Hedonic Pricing method. However, when there is weak countries also abound in heritage sites. Given low levels of
or strong seperability between the resource being valued per capita income in these countries, it is necessary to
and a marketed good, the problem of missing markets for evaluate the demand for such sites using the CVM in order
non-use values reduces the ability of revealed preference to judge their potential as a possible tourist attraction site.
methods to capture the TEV of goods with significant non- The present study differs with respect to its methodology
use values; stated preference methods involving direct from similar CVM studies. Firstly, unlike other studies
questioning of respondents become relevant. Of stated based mostly on OLS regression method (Hooi & Ying,
preference methods, the Contingent Valuation Method 2004) or dichotomous choice models (Laarman & Greger-
(CVM), first used by Davies (1963) for the marginal sen, 1996; Lee, 1997; Lee & Han, 2002), it uses a TOBIT
valuation of Maine Woods in USA, is currently ‘‘by far the regression model to identify the determinants of WTP for
most popular technique for measuring total value for improving the site. Secondly, the study employs the
goods having significant passive use value’’ (Carson, regression results to quantify the TEV of the site. Thirdly,
Flores, & Mitchell, 2001, pp. 109). the target group to whom heritage can be marketed has
Over the years, a consensus has emerged that CVM can also been identified using a truncated regression model.
estimate the use value accurately—stated WTP for use
values have been found to be similar to actual WTP in field 3. A brief history of Prinsep Ghat5
and laboratory experiments (Bishop, Heberlein, & Kealy,
1983; Dickie, Fisher, & Gerking, 1987; Kealy, Dovidio, & The Prinsep Ghat was the largest, the grandest, and the
Rockell, 1988). Whether this is also true for non-use values, most beautiful of the different Ghats of Imperial Calcutta.
however has remained a major source of contention It is slightly to the south of a popular promenade by the
(Diamond & Hausman, 1994; Hausman, 1993). Critics of River Hooghly. The Ghat was rectangular in shape,
CVM also point to the different types of strategic biases designed like the Greek Parthenum. There were two marble
that may also creep in as a result of the hypothetical nature lions—one reclining, the other standing—near the landing
of the survey. Free riding (Brubaker, 1982; Marwell & place.6 The Ghat itself was named after Sir James Prinsep
Ames, 1981) or warm glow effects (Seip & Strand, 1992) (1799–1830), former Assay Master in the Calcutta Mint.
has been observed in CVM studies. These biases, however, The exact date of construction of Prinsep Ghat is
are not specific to CVM and do not render the method unknown, though it is documented to be in existence at
invalid (Arrow et al., 1993). least since 1843. Lord Ellenborough was the first Viceroy to
In spite of persistent criticism, the cautious acceptance of depart from this Ghat. Their Royal Highnesses, Edward
CVM by the Blue Ribbon Committee4 set up by National VII and George V also landed here during their visit to
Oceanographic and Atmospheric Administration (NOAA) Calcutta. Ships used to land here directly from England
to evaluate the validity of CVM has set the final seal of carrying the Viceroy, troops and civilians from different
approval. The report also spells out a set of rigorous sections of the population. Of the latter, the most colourful
guidelines according to which such surveys should be and eagerly awaited were the ‘fishing fleet’—the group of
undertaken (Arrow et al., 1993). This has led to the unmarried English girls sent to India by their desperate
increasing adoption of this method in various fields, parents to capture a husband during the Winter Balls.
including cultural economics. Epstein (2003) argues that, The construction of the Howrah Station opened up an
despite all its limitations, CV remains the best method to alternative land route between Bombay and Calcutta. This
aggregate preferences for cultural resources. led to the decay in importance of the Prinsep Ghat. Lord
Despite the plethora of CV studies (Carson et al., 2001) Curzon was the last Viceroy to depart from this Ghat, in
only a small fraction is devoted to valuing cultural 1905. Later on the river receded westwards, away from the
resources—and even less for heritage resources (Noonan, Ghat. It slowly fell into disuse and its landing place became
2003). Examples of such studies valuing heritage are ruined. It is only in 2000, the Public Works Department
Alberini et al. (2003), Bille Hansen (1997), Morey and (PWD) of the West Bengal Government started a program
Rossman (2003), Navraud and Ready (2002), Santagata to renovate the Ghat and beautify its surroundings.
and Signorello (2000), and Sanz, Herrero, and Bedate
(2003). CVM has also been used for checking feasibility of
4. Methodology
promoting tourism for natural resources in a number of
cases (Laarman & Gregersen, 1996; Lee, 1997; Lee & Han,
4.1. Selection of the sample
2002; Hooi & Ying, 2004). Noonan (2003) reports that
most of these applications are in the context of developed
A CVM study was undertaken to explore the possibility
countries. Out of the 65 CV surveys of cultural resources
of urban regeneration via development of heritage tourism
covered by him only five studies were for developing
countries—one each for Argentina, Mexico, Morocco, 5
The history of Prinsep Ghat is based on Doig (ns), Mitra (1980), Mitra
and Mitra (1988).
4 6
Chaired by Kenneth Arrow and Robert Solow. In 1977 the lions were removed to Fort William.
ARTICLE IN PRESS
86 M. Dutta et al. / Tourism Management 28 (2007) 83–95

Table 1
Description of the data

Income groups % of sample Occupational % of sample Educational groups % of Religion % of


population groups population sample sample
population population

0–5000 21.54 Formal 45.30 Primary education & less 9.39 Hindu 75.69
services
5001–10,000 19.34 Business 12.15 High school education 16.02 Islam 12.71
10,001–15,000 15.47 Professional 13.81 Graduate 30.95 Christian 7.18
15,001–20,000 17.68 Dependents 14.93 Post-graduate and above 22.65 Others 4.42
20,001 & above 25.97 Manual and 13.81 Professional qualification 20.99
informal
services

Note: Dependents include housewives, unemployed and retired persons.

in a commercial basis around this site. The survey period As pointed out before, the CVM has been criticized for
began in August 2002 and continued up to January 2003. its hypothetical nature—as Scott claimed ‘‘ask a hypothe-
The survey elicits preferences of the residents of Calcutta tical question and you will get a hypothetical response’’
because it was presumed that any change in the site would (Scott, 1965, p. 37). This may generate invalid responses or
primarily influence the welfare of the residents of the city. outliers. In defence, CVM practitioners argue that the
Others may hold both use and non-use values for Prinsep share of invalid or inconsistent responses is low and focus
Ghat, but this has been ignored at this stage of the on identifying such observations.9 Unfortunately there is
research. no established theoretical guideline to identify outliers or
The sample size was initially 203 individuals. Despite the invalid responses. Further, as pointed out by Hanley and
desirability of using the random sampling method, the Spash (1993), very large bids need not necessarily represent
sample was drawn in a purposive manner to achieve a outliers. They argue that economic value is based on
balance between different income, occupation education effective demand; thus if such bids represent a genuine
and religious group (Table 1). While trying to obtain a response and if the income level of respondents is
sample with a socio-economic pattern similar to that of commensurate with their bid, then there is no justification
Calcutta, the primary focus was on ensuring a sufficient for excluding them. Therefore, non-zero bids have been
number of observations in each category. While this limits checked for their consistency by comparing WTP with
the extent to which the results can be generalized respect to variables like attitudes towards heritage, knowl-
(population mean, for instance, can no longer be estimated edge about the site, socio-economic background, etc.
from the sample means), this enables us to avoid possible Scatter plots of WTP against income were prepared and
problems that may have been encountered in estimating the observations with WTP significantly higher/lower than the
regression models used to estimate TEV in Section 7. average WTP of his/her income group were identified. The
CV surveys can be administered in various ways— attitude, educational level and specific traits of the
through mail, telephonic and face-to-face interviews. Mail respondent were considered using his/her filled-in ques-
surveys can be easily administered, but as literacy levels are tionnaire to check whether the response was justified;
low in Calcutta, it was expected that the response rate otherwise the observation was dropped. This process was
would be very low following this method. Further, it is repeated for variables like literacy level, knowledge about
impossible to check whether the respondent understands Prinsep Ghat, attitude towards development and heritage,
the questions. While this problem can be avoided in the and occupation groups. It was found that 22 responses
case of telephonic surveys, visual cues cannot be used. In were inconsistent; these were discarded. The results are,
both these methods, low-income respondents get dropped therefore, for 181 observations only.
from the survey.7 Therefore, the face-to-face interview
method was employed, following the recommendations of
the NOAA panel (Arrow et al., 1993). This method allows (footnote continued)
manner and avoid giving cues was emphasized during training of
greater control over the interview and the use of visual aids interviewers. The possibility of any interviewer bias manifested through
arousing interest and awareness.8 variations between the WTPs obtained by the interviewers was subse-
quently tested using non-parametric methods (the Kruskall-Wallis H-test
7
In Calcutta, many respondents will not be able to read the mail order and Mann-Whitney U-test) and found to be absent.
9
questionnaire. At least 25% of the total population do not have a postal It has also been pointed out that even market data reported by seller
address; even more lack telephones. and buyers during marketing surveys is misstated in an attempt to
8
The danger with this method, however, is that the response may influence market conditions (Boyle & Bergstorm, 2001, p. 199). Thus the
become dependent on the manner in which the survey is administered by problem of invalid responses is not unique to CVM and does not
the interviewers. To avoid this problem the need to speak in a neutral invalidate this method.
ARTICLE IN PRESS
M. Dutta et al. / Tourism Management 28 (2007) 83–95 87

4.2. Contingent scenario and respondents’ preferences believed that the spot could be developed as a tourist
attraction with theme restaurants and outlets selling
Since, the respondents have to indicate their choice in an artefacts with historical associations. Even peddlers re-
(constructed and therefore) unfamiliar market, they must sponded positively to such ideas as they realize that this
be provided with complete information about the resource would generate employment and income. However the
being valued (Fischoff & Furby, 1988; Sagoff, 1988). While commercial use value could be increased further if the site
it has been acknowledged that consumers in actual markets is developed properly. Only a few argued that commercial
do not base their choices on the entire information set, the use of the site may affect the historical and aesthetic value
subset on which the choice is made is not known and varies of the site—that is, commercial development was not
across consumers. This calls for provisioning of complete complementary to conservation.
information in the contingent scenario (Munro & Hanley,
2001). However, too much of information can lead to
‘‘cognitive overload’’ (Boyle & Bergstorm, 2001); further, 4.3. Elicitation of bid and willingness to pay
the information provided in actual markets is rarely perfect
(Munro & Hanley, 2001), so that the issue becomes one There are various methods of eliciting the WTP—open-
of providing the optimal information about the commodity ended, referendum, iterative bidding game and payments
characteristics and services provided by it, and the card. It has been pointed out that, given imperfect
availability of substitutes.10 It is also essential for the information and the time constraint facing the respondent,
scenario to be believable and neutral (Mitchell & Carson, the open-ended format typically leads to understatement
1985). (Bateman, Langford, & Rasbash, 2001).
After briefing respondents about the history of the site In contrast, dichotomous formats resemble market
and on the recent Government step undertaken by the transactions (Desvousges, Smith, & McGivney, 1983)
Government to restore Prinsep Ghat, a hypothetical and, under certain conditions, result in truth telling
scheme was presented before the respondents. This becoming the optimal strategy (Hoehn & Randall, 1987).
scenario described how the Government was planning to Within the dichotomous format, the referendum method
promote the tourism potential of the site by beautifying the has the defect that it fails to elicit actual WTP. So this
riverside, introducing boating facilities and river based survey employed the iterative bidding method.12
recreational activities, setting up theme restaurants and In the absence of any accepted method to determine the
retail outlets selling artefacts would generate recreational starting bids, the suggestion made by Cooper (1993) and
value, income and employment opportunities. But, in view Kanninen (1995)—that they be determined on the basis of
of scarcity of funds, the Government was proposing to information gained from pilot surveys—has been followed.
raise funds directly from the public. The respondents were Three starting bids (SBID) were used—yearly amounts of
asked to state their maximum WTP for this programme. Rs. 575 (the highest start bid—referred to as HSBID), Rs.
To avoid any spurious emotions affecting responses, 375 (referred to as MSBID) and Rs. 125 (termed LSBID).13
respondents were reminded that their income is limited One problem with starting bids is that respondents faced
and has alternative uses. Again to avoid any embededness with an unfamiliar situation may interpret them as
bias, respondents were requested to think only about indicative of the true value of the resource and use it as a
Prinsep Ghat and not about the other heritage sites of reference point.14 This may create an anchoring bias
Calcutta.11 (Boyle, Bishop, & Welsh, 1986; Rowe, d’Arge, & Brook-
CVM studies have showed that the payment vehicle and shire, 1980). However, the relationship between the initial
its institutional context can influence response (Mitchell & bid and the response is not inconsistent with utility
Carson, 1985; Rowe & Chestnut, 1983). The pilot survey maximizing behaviour (Farmer & Randall, 1995; Hane-
for Prinsep Ghat had revealed that, in view of the disparity mann, 1994). Further, anchoring has been observed even
in socio-economic status of the respondents, it was difficult for market goods (Cooper & Loomis, 1992 and Shapiro,
to design a payment vehicle that would be commonly 1968).
accepted. To avoid a bias against the bid vehicle, CVM surveys ask respondents whether they would
alternative means of payment were offered to the contribute for a specific period or not.15 During the pilot
respondent—cash or cheque, monthly deductions through survey, however, information was sought on both the
electricity bills, or any other medium of payment. 12
The payments card is not popularly used as it is difficult to construct,
The survey reveals that both recreational and commer- particularly for heterogeneous communities in LDCs.
cial use values are significant. Many of the respondents 13
The starting bid has been stated in Rupees (Rs.)/month. Restating the
chosen WTP in Rs./year and allowing the respondent the option to revise
the choice made earlier has rechecked the response. Rupee is the Indian
10
Again, the problem is the absence of a set of accepted operational currency; it is approximately equal to about two US cents.
14
guidelines. For a citation of literature, refer to Bateman et al. (1995).
11 15
Given the unique nature of the conservation programme in Calcutta In CVM studies the bid vehicle can bias stated WTP. To avoid bias
and nature of the site, the chances of embeddedness occurring is possibly respondents have been offered a choice of alternative instruments of
low. payment: cash/cheque, via electricity bills, user fees and any other forms.
ARTICLE IN PRESS
88 M. Dutta et al. / Tourism Management 28 (2007) 83–95

Table 2 for instance. However, the non-use values and potential


Estimates of willingness to pay commercial use value are important factors explaining
Group Observations Median Mean Standard
willingness to contribute.16
deviation
6. Analysis of preferences
All 181 375.00 613.06 732.75
observations
Positive 140 550.00 792.60 742.87 The next step of the analysis attempts to find out the
LWTP causal factors influencing the responses.

6.1. Selection of the attributes


contribution amount and preferred period of payment. It
was found that the period of payment varied across The causal factors expected to be relevant in explaining
individuals significantly. For instance, most of the higher WTP are as follows:
income respondents prefer one-shot payment, while low-
Family Income (FINCOME), Expenditure on travelling
income respondents prefer to pay a low sum over a longer
for leisure to places outside Calcutta (EXP), Education
period. In the study, therefore, Lifetime Willingness to Pay
(GRAD), Attitude towards heritage (revealed through
(LWTP) was used instead of WTP. This is defined as WTP
the extent of his agreement with five statements—AT1,
per annum multiplied by number of years for which the
AT2, y AT5), Knowledge about Calcutta’s heritage
individual is willing to pay. In the cases where an individual
sites (KNOW), Number of museum visits in last three
was willing to pay for the rest of his or her life, the
years (MUSEUM), and Attraction of built heritage
expectation of life at present age was calculated from Life
during travels (BUILDING) are expected to influence
Tables to find the approximate years of payment. Taking
WTP favourably (expected sign of the regression
LWTP, rather than WTP, can be justified on the grounds
coefficient is positive). Whether s/he had heard the
that preservation of heritage sites is not a once-and-for-all
name of the site (HEARD), or had visited it (VISIT) and
task, but a continuous responsibility.
the respondent’s knowledge about the present condition
The average and median LWTP for Prinsep Ghat is
of the site (AWARE) are also important, and were
given in Table 2. While Mean is the commonly used
expected to affect WTP positively. In addition, the value
measure of Central Tendency, Median WTP is also
of the Start Bid (SBID) and the gender of the interviewer
reported in studies as it remains unaffected in the presence
(GINT) were also hypothesized to influence WTP.
of extreme values. It can be seen that the median is lower
Finally, the responses were expected to be sensitive with
than mean. This implies that majority of the respondents
respect to occupation of the respondent, his/her age,
are willing to pay less than the mean WTP.
gender and religion.

5. Motivations underlying WTP decision A list of variables on which information has been
collected is given in Appendix A.
Subsequently, the motivations underlying WTP decision
are analysed on the basis of the reasons cited by 6.2. Correlation analysis
respondents for contributing or not after obtaining the
WTP for the site. This once again tests the consistency of The correlation coefficient between each of these
preferences. variables and LWTP has been estimated to check their
It is found that 77.35% of respondents are willing to pay degree of association. In many cases the correlation
whereas 22.65% are not willing to pay. The major causes coefficients are found to be significantly high (Appendix B).
for zero response are summarized in Table 3. An important As expected, the simple correlation between LWTP and
trend emerges from a study of the diverse motives. Very educational qualification (whether the respondent is at
few people (only 4% of respondents not paying) feel that least a graduate), family income and travelling expenditure
conservation of Prinsep Ghat is unimportant. Analysing are significantly positive. On the other hand, the correla-
the income profile of the non-contributors and their stated tion coefficient of LWTP with occupation is significantly
reasons for not paying indicates that lack of income is the negative. The codes for occupations have been arranged in
main reason for not contributing. This leads them to argue such a manner that higher ranks imply both lower income
that the funds should be raised from visitors or provided by and awareness levels. Hence negative correlation between
the Government. LWTP and occupational structure is quite expected. The
An analysis of the reasons for positive willingness to relation between LWTP and start bid is also significantly
contribute (Table 4) reveals that the presence of recrea- positive.
tional use value is not an important factor. This is because
of the availability of substitutes in other parts of the city, 16
Epstein (2003), however, warns that even CVM may miss some
Calcutta—the lawns of the nearby Victoria Memorial Hall components of the non-use value of cultural resources.
ARTICLE IN PRESS
M. Dutta et al. / Tourism Management 28 (2007) 83–95 89

Table 3
Distribution of persons by reasons for not paying

Reasons for not paying Number of respondents Percentage

I feel that conservation of Prinsep Ghat is unimportant 4 6.15


There are more serious problems requiring our attention 11 16.92
The Fund should be raised from visitors to Prinsep Ghat 9 13.85
Only high income groups should contribute 8 12.31
I want to contribute, but I have no spare income 10 15.38
It is the responsibility of the Government, etc. 17 26.15
Others 6 9.23
Total 65 100.00

Note: Respondents had been allowed to opt for more than one option. This explains why the column total of ‘Number of Respondents’ (65) does not tally
with number of persons refusing to pay (41).

Table 4
Distribution of persons by reasons for paying

Reasons for not paying Number of respondents Percentage

Regular/occasional visitor to Prinsep Ghat 7 2.28


Prinsep Ghat should be protected to keep option of visiting open for future 103 33.55
It is our duty to protect heritage sites 97 31.60
Prinsep Ghat should be protected to generate income from tourism 97 31.60
Payment is mandatory 3 0.98
Total 307 100.00

Note: Again, there is a discrepancy between the total of 307 and number of respondents with positive WTP (140); this is owing to the fact that respondents
were allowed to cite three reasons for their positive WTP.

Contrary to maintained belief, no significant association In this study, all persons who are not willing (or able) to
is observed between LWTP and attitudes towards Calcutta’s pay for the restoration and development of the site are
heritage, whether a person heard about the site or visited the reported to have a zero WTP. However, this group is not
site, whether s/he is aware of the history and current homogenous and can be further divided into two sub-
position of the site, age of the respondent, his/her attraction groups: While some people have zero WTP, others may
towards built heritage while travelling for leisure, and the have value the site negatively. The latter group is against
number of museum visits in last three years. Correlation conservation of the site for various reasons: they may feel
analysis, by itself, is incomplete and has to be supplemented that there are other important problems requiring atten-
by further statistical analysis examining the existence of tion, or they may think that any attempt to reuse the site
causal relationship between a group of variables and a will degrade its historical and aesthetic value, etc. For all
dependent variable. practical purposes it is not possible to obtain any estimate
of the negative WTP. So the negative value of the
dependent variable is conceptually possible, but not
6.3. Choice of appropriate econometric model actually observable. Thus, the dependent variable is
practically censored and the problem of ‘limited depen-
The method of estimation to be applied in case of a dence’ is encountered.
CVM study depends on the structure of bidding. If the Thus,
bidding is of close-ended type, the relevant estimation
method would be LOGIT/PROBIT.17 These dichotomous LWTP ¼ LWTP if LWTP 40;
(1)
choice models are subjected to severe criticism in terms of LWTP ¼ 0 otherwise:
their limited ability to account for the restrictions imposed The resulting sample is said to be a censored sample. For
by underlying economic theory (Hanemann & Kanninen, all our observations with LWTP ¼ 0, we only know that
2001). In fact the censored regressions based on open- LWTP*p0; that is,
ended referendum are expected to produce more realistic
estimates of the WTP in a hypothetical market. PðLWTP ¼ 0Þ ¼ PðLWTP  p0Þ. (2)

17
In other words, a probability mass (a ‘spike’) at LWTP ¼ 0
A LOGIT model has been estimated. However, as the results are is being inserted. Since cases where LWTP* ¼ 0 (the first
disappointing (with only occupation and as significant variables) the
results have not reported. One of the causes of poor result with LOGIT sub-group) cannot be distinguished from cases where
model is it works with binary relations, thereby loosing valuable LWTP*o0 (the second sub-group), the use of the TOBIT
information. model to identify the factors affecting LWTP for entire
ARTICLE IN PRESS
90 M. Dutta et al. / Tourism Management 28 (2007) 83–95

sample seems appropriate. Such a model can be used to In Models I and II, AT3, EXP and GRAD are
predict the total economic value of the site. significantly positive. The truncated regression further
Some additional analyses are undertaken based exclu- shows that occupation of the respondents is also important
sively on the positive responses. Thus the regression is in explaining LWTP. In Model II the coefficients of
truncated at LWTP ¼ 0. This truncated analysis would be DAGE1 and DAGE3 are negative. The negative sign of
especially useful in studying the characteristics of the DAGE1 results from the younger people having less
potential clients of the renovated site. This is useful in nostalgic feelings about Calcutta’s past compared to
designing the nature of renovation and the marketing middle-aged persons. With regard to DAGE3, the older
strategy. For instance, if the major tourists are educated people may have a greater feeling for the past, but it is
middle-class family groups then the provision of entertain- unlikely that this feeling will be reflected in the form of
ment services should be designed accordingly to match monetary payments as these people generally have less
their tastes and preferences. command over buying decisions and face greater uncer-
In this paper two regression models are estimated using tainty due to their dependence. The results of the truncated
the LIMDEP package: models also indicate that respondents generally state a
higher WTP to impress female interviewers. This aspect,
(a) Model I: TOBIT Model; however, is not relevant in identifying the target group—
(b) Model II: Truncated Model.18 which was the objective for estimating the truncated
model.
6.4. Regression results The TOBIT model considers all individuals—both who
are willing and not willing to pay. It is used to project the
Initially, LWTP has been regressed on all variables demand for developing the site as tourist attraction site.
expected to have significant association with it. Since the The truncated model, on the other hand, provides a picture
t-values of some variables were very low, the models were of the effect of different motivations underlying the
re-estimated dropping the variable with the lowest t-value response of individuals. It indicates that educated persons,
in each step. The results of the final iterations are presented high-income households, middle-aged persons, working in
below: service sector or engaged in business or professional
TOBIT Model: activities are potential user of the resource. The latter
dimension becomes important in developing marketing
LWTP ¼ 1632:45 þ 583:09 AT3 þ 0:003 EXP
ð2:23Þ ð1:86Þ ð2:07Þ strategies for the post-restoration phase.
þ 374:57 GRAD þ 2:17 SBID . ð3Þ
ð2:73Þ ð7:54Þ 6.5. Omitted variables
Log Likelihood Function: 1149.801.
Truncated Model: Some variables—gender, religion and family income of
the respondent, his/her attitudes (except AT3) towards
LWTP ¼ 12627:37 þ 2808:80 AT3 þ 0:02 EXP heritage, his/her awareness about the site, his/her museum
ð2:18Þ ð1:85Þ ð3:64Þ
visits in the last 3 years, knowledge about Calcutta’s
þ 2212:54 GRAD þ 1443:64 SERVICE
ð2:88Þ ð2:57Þ heritage sites, etc.—expected to be related to WTP had to
þ 1339:61 PROF þ 1565:59 BUS be dropped from the final regression because of their low
ð2:16Þ ð2:33Þ t-values. One explanation is that the effects of these
 927:59 DAGE1  1242:13 DAGE3 þ 957:14 GINT . variables are captured through other variables, which are
ð2:59Þ ð2:18Þ ð2:68Þ
present in the final iteration of the models. To check the
ð4Þ validity of this possibility the correlation matrix between
Log Likelihood Function: 1149.801. the variables appearing in the final results and some of the
[The t-ratios are given in parentheses.] important omitted variables has been estimated using the
Pearson Correlation Coefficient19 (Tables 5 and 6).
18
In Models I and II the exact values for SBID have been used and From these correlation matrices, it is apparent that the
treated as a quantitative regressor. In another variant of the truncated included regressors and several excluded variables are
model (say, Model III), SBID has been used as a qualitative variable. To highly correlated. So explicit inclusion of all the expected
capture the differential effect of SBID variations on WTP two intercept influences are barred by the potential problem of multi-
dummies were constructed. In construction of these dummies the relevant
question becomes: which SBID to take as the reference category? collinearity. The list of factors, finally identified as
Theoretically, the estimation results should be the same irrespective of determinants of LWTP are reported in Table 7.
which SBID is used as the reference category. The estimation results are
the same whether LSBID or HSBID is taken as the reference category.
Surprisingly, slightly different results are obtained on taking MSBID as
19
the reference category—the coefficients of this model differ from the Although we have presented the regression results for Attitudes taken
earlier variants even after the final iteration. Since, Model II captures the separately, we had also experimented with an index based on [1] all 5
effects of a greater number of variables, only this Model has been attitudes [2] the last 3 attitudes. The best regression result was obtained for
reported. AT3.
ARTICLE IN PRESS
M. Dutta et al. / Tourism Management 28 (2007) 83–95 91

Table 5 is run to obtain estimated values of a, b1 and b2.20 The


Alternative attitudes: a correlation analysis alternative values of the Starting Bid (0, 125, 375 and 575)
Incorporated variable AT1 AT2 AT5
and the average expenditure on travelling (EXP) estimated
for each sub-group are substituted in the above equation to
AT3 0.337** 0.166* 0.149* get the estimated LWTP (corresponding to each value of
the SBID) for that particular sub-group.
**Correlation is significant at 0.01 level (2-tailed).
*Correlation is significant at 0.05 level (2-tailed). Though there should have been four equations, the first
sub-group has only one observation; so the stated value
of the respondent’s LWTP is used for the first category
Table 6 (Table 8).
Socio-economic factors: a correlation analysis From the estimated LWTP of five equations median
LWTP is calculated for non-graduates. Similarly, median
Incorporated variable HEARD VISIT AWARE FINCOME MUSEUM
LWTP is estimated for the group GRAD ¼ 1 from eight
DAGE 0.213** 0.190* 0.177* equations of relevant sub-groups. The median values are
GRAD 0.217** 0.264** 0.460** 0.210** then multiplied by the total number of non-graduates and
EXP 0.428** 0.215** graduates respectively. Using Census data and the Socio-
**Correlation is significant at 0.01 level (2-tailed). economic Profile for Calcutta prepared by CMDA
*Correlation is significant at 0.05 level (2-tailed). (Chatterjee et al., 1999), the total number of non-graduates
and graduates above 25 years in the city have been
estimated to be 2,684,199 and 499,314, respectively.
Multiplying the class specific median LWTP by this figure,
Table 7 and adding the two figures, TEV of Prinsep Ghat is
Relation between regressors appearing in final iteration and excluded estimated to be Rs. 686,186,540 only. Although the figure
regressors obtained for TEV appears very large, it is actually based on
a very modest payment of less than Rs. 300 (about USD
Included variables Include effects of Excluded variables
6.67) spread over a period. However, despite precautions,
AT3 AT1, AT2, AT5 the estimates of TEV may suffer from embededness. So we
DAGE HEARD, VISIT, have scaled down the TEV by assuming that 50% of the
AWARE
promised value will actually be paid. This corresponds to
GRAD HEARD, AWARE,
FINCOME, the recommendations of the NOAA Panel (Arrow et al.,
MUSEUM 1993). The resultant TEV is Rs. 34,309,327.
EXP FINCOME, The total costs of renovating Prinsep Ghat was Rs.
MUSEUM 1,900,000 only. Assuming that 10% of this amount is spent
AT4, BUILDING,
as annual maintenance, the annual expenditure on Prinsep
KNOW, RELIGION
Ghat turns out to be Rs. 190,000 only, which is much less
than the value of the site. The surplus indicates the
presence of an untapped demand for urban regeneration
7. Total economic value of Prinsep Ghat centred around the Prinsep Ghat. It is further estimated
that the principal amount necessary to generate the annual
Finally the total economic value (TEV) of Prinsep Ghat maintenance cost of the site is Rs. 2,714,285.71. This
to residents of Calcutta Metropolitan Area is estimated amount can be obtained if respondents pay only 0.4 % of
using the TOBIT regression model. Here the relevant their stated WTP!
causal determinants of WTP are AT3, EXP, GRAD and Before passing on to the concluding section, one
SBID. Information on AT3, EXP and SBID are not observation is in order. The TEV reported has been
available for the relevant population as a whole. But estimated for a given scenario. It is possible to increase the
information on GRAD is obtained for the residents of TEV by modifying the scenario enhancing the value of
Calcutta Metropolitan Area from the Calcutta Metropo- Prinsep Ghat even more. This has important implications
litan Development Authority Socio-economic Household for heritage preservation and optimal land use in densely
Survey of 1996–97 (Chatterjee, Bhattacharya, & Halder, populated areas.
1999) and Directorate of Census Operation, West Bengal
(1991). 8. Conclusion
Based on the educational levels and attitudes towards
heritage the sample is divided into four sub-groups (see In recent years, an important justification of conserva-
Fig. 1). For each of these sub-groups the regression tion in developing countries was that they could be used as
equation
20
The intercept dummy includes the effect of other variables (AT3 and
LWTP ¼ a þ b1 EXP þ b2 SBID (5) GRAD).
ARTICLE IN PRESS
92 M. Dutta et al. / Tourism Management 28 (2007) 83–95

GRAD=0 GRAD=1

AT3 = 0 AT3 = 1 AT3=0 AT3 = 1

SBID=0 SBID=125 SBID=375 SBID=575SBID=0 SBID=125 SBID=375 SBID=575 SBID=0 SBID=125 SBID=375 SBID=575SBID=0 SBID=125 SBID=375 SBID=575

Median LWTP1 Median LWTP2

Fig. 1. Categories used for estimating TEV.

Table 8
Estimation of class specific median LWTP

GRAD Attitude Sample SBID Intercept Coeff of SBID Coeff of EXP AverageExp Projected LWTP Median

GRAD ¼ 0 AT3 ¼ 0 1 No regression could be fitted 275 205.1026


AT3 ¼ 1 46 0 123.686 1.1703 0.006 6035.56 87.4724
125 58.81506
375 351.3901
575 585.4501
GRAD ¼ 1 AT3 ¼ 0 6 0 169.494 0.9705 0.0037 21500 89.944 273.9935
125 31.3685
375 273.9935
575 468.0935
AT3 ¼ 1 129 0 147.551 2.4354 0.0027 21804.65 88.6787
125 215.7463
375 824.5963
575 1311.676

cultural capital to maintain the site on a self-funding basis starting proposition that the use of heritage as a cultural
by developing it as a visitor attraction site. This approach capital can be economically justified. However, it is not
has been justified by estimating tourism demand (based enough to test for viability. Given the high opportunity
only on use values) for the heritage sites. On the basis of costs of funds in LDCs policy makers also have to initiate
this study of Prinsep Ghat, this paper argues that this steps to mobilize funds and ensure that conservation
approach underestimates the total benefits from such programmes are financially viable from a management
programmes as it ignores the considerable non-use and perspective. This calls for steps like converting our
non-market values stemming from such programmes. hypothetical scenario to reality and mobilizing funds from
The survey reported in this paper estimates median potential clientele. The results of the truncated model
LWTP of Prinsep Ghat to be Rs. 375. Using a TOBIT indicates the socio-economic characteristics of the group
model, the bid function was estimated identifying the that can be targeted during commercial development of
determinants of LWTP. It was found that attitude of the such sites: educated persons, persons with high income
respondent, educational level, and expenditure on tourism levels, middle aged persons, those working in the service
determined WTP. Further, these variables also captured sector, or engaged in business, or professionals, and their
the effects of other variables like family income, religion, families would form the prospective customer group.
knowledge about the site, frequency of museum visits, etc. Although this paper has focussed on the demand side,
on WTP. Using this regression model the TEV was re- supply side considerations may also play important role in
estimated. It was found that even if respondents paid only this context. The extent to which this special type of
0.4% of their stated WTP, this would justify the conserva- tourism development is possible depends on the presence
tion of Prinsep Ghat. of suitable structures in terms of accessibility, space
Although the purposive sampling method used in this availability (Shackley, 1998; Swarbrooke, 1995), on the
survey limits the confidence with which these results can be financial capacity of public and private partners, on their
generalized for policy making, this study does validate the planning and marketing ability, and on the political will
ARTICLE IN PRESS
M. Dutta et al. / Tourism Management 28 (2007) 83–95 93

(Jansen-Verbeke & Lievois, 1999). It may also involve BUILDING ¼ if built heritage attracts the respondent
selection between alternative sites. Tools like multinomial while travelling for leisure
LOGIT with random utility modelling (RUM) may be KNOW ¼ respondent’s knowledge about heritage sites
appropriate for such purposes. This type of exercise would in Calcutta
be essential for the civic authority to carry out in SBID ¼ start bid
prioritising the sites to be selected for such development. HSBID ¼ high start bid ( ¼ Rs. 575.00)
MSBID ¼ medium start bid ( ¼ Rs. 375.00)
Acknowledgements LSBID ¼ low start bid ( ¼ Rs. 125.00)
GINT ¼ 1 if the interviewer is female; 0, otherwise.
The study was funded by a Minor Research Project
grant of the University Grants Commission, Eastern Zone,
India. Insightful comments and suggestions from Robin Appendix B. Correlation matrix between LWTP and socio-
Mukherjee, Dipankar Coondoo and Sugata Sen Roy are economic variables
gratefully acknowledged. The usual disclaimers apply.

Appendix A. List of variables Variables Para- Non- Variables Para- Non-


metric parametric metric parametric
AT1 ¼ 1 for agreement that Calcutta’s heritage is AT1 0.060 0.052 AWARE 0.058 0.005
something to be proud of AT2 0.008 0.014 DAGE 0.012 0.032
AT2 ¼ 1 for agreement that value of the city would not AT3 0.104 0.084 RELIGION 0.012 0.002
be the same without its heritage buildings and sites AT4 0.015 0.056 GRAD 0.283** 0.266**
AT3 ¼ 1 for agreement that heritage sites need to be AT5 0.008 0.074 FINCOME 0.209** 0.235**
conserved to promote socio-cultural environment of the HEARD 0.061 0.062 OCCUP 0.263** 0.218**
city VISIT 0.030 0.063 EXP 0.209** 0.236**
AT4 ¼ 1 for agreement that all heritage buildings and MUSEUM 0.121 0.135 BUILDING 0.084 0.071
KNOW 0.052 0.064 SBID 0.414** 0.626**
sites should be conserved
AT5 ¼ 1 for agreement that developmental projects **Correlation is significant at 0.01 level (2-tailed).
harming heritage should not be implemented
HEARD ¼ 1 if the respondent heard of the Prinsep
Ghat References
AWARE ¼ 1 if the respondent is aware of the past
history and present condition of the Ghat Akal, M. (2004). Forecasting Turkey’s tourism revenues by ARMAX
VISIT ¼ 1 if the respondent has visited the site model. Tourism Management, 25(5), 565–580.
Alberini, A., Riganti, P., & Longo, A. (2003). Can people value the
FINCOME ¼ family income of the respondent
aesthetic and use services of urban sites? Evidence from a survey of
EXP ¼ respondent’s travelling expenditure in the last 3 Belfast residents. Journal of Cultural Economics, 27(3-4), 193–213.
years Arrow, K., Solow, R., Portney, P., Leamer, E., Radner, R., & Schuman,
OCCUP ¼ occupation of the respondent ( ¼ 1 for H. (1993). Report of the NOAA panel on contingent valuation. US
service, ¼ 2 for business, ¼ 3 for professionals, ¼ 4 Federal Register, 58, 4601–4614.
for dependents (housewives and retired persons), ¼ 5 Bateman, I. J., Langford, I. H., & Rasbash, J. (2001). In I. J. Bateman,
K. G. Willis (Eds.), Elicitation effects in contingent valuation studies
for manual labourers) (pp. 511–539).
SERVICE ¼ if the respondent is a service holder Bille Hansen, T. (1997). The willingness-to-pay for the Royal Theatre in
BUS ¼ if the respondent is a business man Copenhagen as a public good. Journal of Cultural Economics, 21(1),
PROF ¼ if the respondent is a self-employed person 193–213.
Bishop, R. C., Heberlein, T., & Kealy, M. J. (1983). Contingent valuation
DEP ¼ if the respondent is a housewife or an
of environmental assets: Comparison with a simulated market. Natural
unemployed or a retired person Resources Journal, 23, 619–634.
MLAB ¼ if the respondent is a manual labourer or Boyle, K., Bishop, R., & Welsh, M. (1986). Starting point bias in
works in an informal sector contingent valuation survey. Land Economics, 65, 57–63.
GRAD ¼ 1 if the respondent has graduation, post- Boyle, K. G., & Bergstorm, J. C. (2001). In I. J. Bateman, K. G. Willis
graduation or above, or professional degrees; 0, other- (Ed.), Doubt, doubts and doubters: The genesis of a new research agenda
(pp. 183–206).
wise Brubaker, E. (1982). Sixty-eight percent free revelation and thirty-two
MUSEUM ¼ number of visits to a museum in the last percent free ride? Demand disclosures under varying conditions of
3 years exclusion. In V. L. Smith (Ed.), Research in experimental economics,
DAGE1 ¼ age groups (as a dummy variable) Vol. II (pp. 151–166). Greenwich, CT: JAI Press.
GENDER ¼ 1 if the respondent is a female; 0, Cantacuzzino, S. (1989). Rearchitecture, old buildings/new uses. London:
Thames and Hudson.
otherwise Carson, R. T., Flores, N. E., & Mitchell, R. C. (2001). In I. J. Bateman,
RELIGION ¼ religion of the respondent (as dummy K. G. Willis (Eds.), The theory and measurement of passive use value
variable) (pp. 97–130).
ARTICLE IN PRESS
94 M. Dutta et al. / Tourism Management 28 (2007) 83–95

Chatterjee, N., Bhattacharya, N., & Halder, A. (1999). Socio-economic Kanninen, B. J. (1995). Bias in discrete response contingent valuation.
profile of households in Calcutta metropolitan area: 1996–97. Calcutta: Journal of Environmental Economics and Management, 28(1), 114–125.
Calcutta Metropolitan Development Authority. Kealy, M. J., Dovidio, J. F., & Rockell, M. L. (1988). Accuracy in
Chu, F. L. (2004). Forecasting tourism demand: A cubic polynomial valuation is a matter of degree. Land Economics, 64, 158–171.
approach. Tourism Management, 25(2), 209–218. Konrad, V. A. (1982). Historical artefacts as recreational resources. In G.
Cooper, J. C. (1993). Optimal bid selection for dichotomous choice Wall, & G. S. Marsh (Eds.), Recreational land use: perspectives on its
contingent valuation surveys. Journal of Environmental Economics and evolution in Canada (pp. 392–416). Ottawa: Carleton University Press.
Management, 24, 25–40. Laarman, J. G., & Gregersen, H. M. (1996). Pricing policies in nature-
Cooper, J. C., & Loomis, J. (1992). Sensitivity of willingness-to-pay based tourism. Tourism Management, 17(4), 247–254.
estimates to bid design in dichotomous choice contingent valuation Lee, C. K. (1997). Valuation of nature-based tourism resources using
models. Land Economics, 68, 211–224. dichotomous choice contingent valuation method. Tourism Manage-
Cossons, N. (1989). Heritage tourism—Trends and tribulations. Tourism ment, 18(8), 587–591.
Management, 10(3), 192–194. Lee, C. K., & Han, S. Y. (2002). Estimating the use and preservation
Davies, A., & Prentice, R. (1995). Conceptualising the latent visitor to values of national park’s tourism resources using a contingent
heritage attractions. Tourism Management, 16(7), 491–500. valuation method. Tourism Management, 23(5), 531–540.
Davies, R. K. (1963). Recreation planning as an economic problem. Light, D., & Prentice, R. (1994). Who consumes the heritage product?
Natural Resources Journal, 3, 239–249. Implications for European heritage tourism. In G. J. Ashworth, & P. J.
Desvousges, W. H., Smith, V. K., & McGivney, M. P. (1983). A Larkham (Eds.), Building a new heritage: Tourism, culture and identity
comparison of alternative approaches for estimating recreations and in the new Europe (pp. 90–116). London: Routledge.
related benefits of water quality improvements. EPA Report, 230-05-83- Marwell, G., & Ames, R. E. (1981). Economists free ride: Does anyone
001. Washington: US Environmental Protection Agency, Office of else? Journal of Public Economics, 15, 295–310.
Policy Analysis. McNulty, R. (1993). Cultural tourism and sustainable development.
Diamond, P. A., & Hausman, J. A. (1994). Contingent valuation: Is some World Travel and Tourism Review, 156–162.
number better than no number? Journal of Economic Perspectives, 8(4), Mitchell, R. C., & Carson, R. T. (1985). Comment on option value:
45–64. Empirical evidence from case study of recreation and water quality.
Dickie, M., Fisher, A., & Gerking, S. (1987). Market transactions and Quarterly Journal of Economics, 100, 292–294.
hypothetical demand data: A comparative study. Journal of the Mitchell, R. C., & Carson, R. T. (1989). Using surveys to value public
American Statistical Association, 81, 69–75. goods: The contingent valuation method. Washington, DC: Resources
Directorate of Census Operation, West Bengal. (1991). Town directory and for the Future.
primary census abstract-1991: West Bengal district census handbook: Mitra, R. (1980). Kalikata darpan. Calcutta: Subarnarekha.
Calcutta, Vol. 26: XIIA and XIIB, New Delhi. Mitra, R., & Mitra, R. (1988). Ghater katha. Calcutta: Pratikhan Press
Doig, D. (ns). Calcutta—An artist’s impression. Calcutta: The Statesman. Private Limited.
Epstein, R. A. (2003). The regrettable necessity of contingent valuation. Morey, E., & Rossman, K. G. (2003). Using stated preference questions to
Journal of Cultural Economics, 27(3–4), 259–274. investigate variations in willingness to pay for preserving marble
Farmer, M. C., & Randall, A. (1995). Understanding starting price effects monuments: Classic heterogeneity, random parameters, and mixture
in contingent valuation data sets. Paper presented at sixth annual models. Journal of Cultural Economics, 27(3-4), 215–229.
meeting of the European association of environmental and resource Mossetto, G. (1994). The economic dilemma of heritage preservation. In
economists, University of Umea, Sweden, 17–20 June. A. Peacock, & P. Rizzo (Eds.), Cultural economics and cultural policies
Fischoff, B., & Furby, L. (1988). Measuring values: A conceptual (pp. 81–96). The Netherlands: Kluwer Academic Publishers.
framework for interpreting transactions with special reference to Munro, A., & Hanley, N. D. (2001). In I. J. Bateman, K. G. Willis (Eds.),
contingent valuation of visibility. Journal of Risk and Uncertainty, 1, Information, uncertainty and contingent valuation (pp. 258–279).
147–184. Navraud, S., & Ready, R. C. (2002). In Valuing cultural heritage: Applying
Goh, C., & Law, R. (2002). Modelling and forecasting tourism demand environmental valuation techniques to historic buildings, monuments and
for arrivals with stochastic non-stationary seasonality and interven- artifacts. Northampton, MA: Edward Elgar.
tion. Tourism Management., 23(5), 499–510. Noonan, D. S. (2003). Contingent valuation and cultural resources: A
Gunadhi, H., & Boey, C. K. (1986). Demand elasticities of tourism in meta-analytic review of the literature. Journal of Cultural Economics,
Singapore. Tourism Management., 7(4), 239–253. 27(3-4), 151–176.
Hanemann, W. M. (1994). Valuing the environment through contingent Pearce, D. (1989). Conservation today. London: Routledge.
valuation. Journal of Economic Perspectives, 8(4), 19–43. Peattie, K., & Moutinho, L. (2000). The marketing environment for travel
Hanemann, W. M., & Kanninen, B. (2001). The statistical analysis of and tourism. In L. Moutinho (Ed.), Strategic management in tourism
discrete response CV data. In I. J. Bateman, Willis (Eds.) (pp. 302–441). (pp. 17–37). Wallingford: CABI Publishing.
Hanley, N., & Spash, J. (1993). Cost benefit analysis and the environment. Prentice, R. (1992). Tourism and heritage places. London: Routledge.
London: Aldershot. Rowe, R. D., & Chestnut, L. G. (1983). Valuing environmental
Hausman, J. A. (Ed.). (1993). Contingent valuation: A critical assessment. commodities revisited. Land Economics, 59, 404–410.
Amsterdam: Elsevier Science Publishers BV. Rowe, R., d’Arge, R., & Brookshire, D. (1980). An experiment on the
Hoehn, J. P., & Randall, A. (1987). A satisfactory benefit cost indicator economic value of visibility. Journal of Environmental Economics and
from contingent valuation. Journal of Environmental Economics and Management, 7, 1–19.
Management, 14, 226–247. Sagoff, M. (1988). Some problems with environmental economics.
Hooi, T. D., & Ying, S. P. (2004). A study on the determinants of tourists’ Environmental Ethics, 10, 55–74.
willingness to pay for beach vacation: The case of Cherating Beach. Santagata, W., & Signorello, G. (2000). Contingent valuation of a cultural
Journal of Travel and Tourism Research, 4(2), 100–130. public good and policy design: A case study of ‘Napoli Musei Aperti’.
Hovinen, G. R. (1995). Heritage issues in urban tourism: An assessment of Journal of Cultural Economics, 24(3), 181–203.
new trends in Lancaster country. Tourism Management, 16(5), Sanz, J. A., Herrero, L. C., & Bedate, A. M. (2003). Contingent valuation and
381–388. semiparametric methods: A case study of the national museum of sculpture
Jansen-Verbeke, M., & Lievois, E. (1999). Analysing heritage resources for in Valladolid, Spain. Journal of Cultural Economics, 27(3-4), 241–257.
urban tourism in European cities. In D. Pearce, & R. Butler (Eds.), Scott, A. (1965). The valuation of game resources: Some theoretical aspects.
Contemporary issues in tourism development (pp. 81–107). London: Canadian Fisheries Report, iv. Ottawa, Ontario: Department of
Routledge. Fisheries of Canada.
ARTICLE IN PRESS
M. Dutta et al. / Tourism Management 28 (2007) 83–95 95

Seip, K., & Strand, J. (1992). Willingness to pay for environmental Shapiro, B. P. (1968). The psychology of pricing. Harvard Business
goods in Norway: A contingent valuation study with real pay- Review, 46, 14–25.
ments in Norway. Environmental and Resource Economics, 2, Strike, J. (1994). Architecture in conservation: Managing development at
91–106. historic sites. London: Routledge.
Shackley, M. (Ed.). (1998). Visitor management: Case studies from world Swarbrooke, J. (1995). The development and management of visitor
heritage sites. London: Butterworth and Heinemann. attractions. London: Butterworth and Heinemann.

You might also like