Introduction

1UNESCO’s Convention concerning the Protection of the World Cultural and Natural Heritage, adopted in 1972, has a number of goals designed principally to encourage the protection of important sites. Increasing tourism is not one its goals and excessive visits indeed threaten some World Heritage sites. UNESCO has responded with a World Heritage Tourism Program, designed to encourage benefit for local populations with sustainable management practices.

2The Historic town of Guanajuato and adjacent mines were named to the World Heritage list in 1988. The 450-year-old city was by the 18^th century the world’s largest silver producing zone. The enormous wealth generated by these mines is mirrored in the distinctive and elaborate architecture of its extensive array of historic buildings. In the 20^th century, the economic influence of the mines faded and Guanajuato’s economy has increasingly been driven by three influences : 1) the government of the State of Guanajuato, headquartered in the city of Guanajuato, and 2) the University of Guanajuato, which has 20,000 students, and 3) tourism.

3Guanajuato City had a population of 141,000 in 2000. The state’s population in 2000 was 4.7 million. Other prominent cities in Guanajuato state are Leon (1.1 million), Irapuato (440,134), Celaya (382,958), and Salamanca (226,654). (State of Guanajuato, 2006 : 120).

4There are several other principal sites in Guanajuato state with heritage tourism appeal, such as San Miguel de Allende, the birthplace of Mexican Independence, Dolores Hidalgo, and the semi-abandoned town of Pozos Town of Pozos, which some promoters are calling the next San Miguel de Allende. As for San Miguel, there is a push for its inscription on the World Heritage list. The town and the Sanctuary of "Jesus Nazareno" (Jesus of Nazareth) of Atotonilco have been on UNESCO’s Tentative List for future World Heritage inscriptions since 2006.

5Guanajuato was inscribed on the list for meeting for meeting four of UNESCO’s criteria for inclusion. Sites meeting at least one criterion merit acceptance onto the list.

6There is sufficient anecdotal evidence that World Heritage listing encourages tourism and that a new breed of educated and affluent tourists are checking off "life lists” as bird watchers have been known to do. One expert notes : "Tour companies and tourist accommodation with access to World Heritage areas commonly advertise that fact in their marketing material ; and tourism developers and entrepreneurs preferentially pursue opportunities in and around World Heritage areas." (Buckley 2002 : 2).

7Few studies have examined the link between World Heritage listing and tourism. A study entitled "The Effects of World Heritage Listing on Tourism to Australian National Parks, " attempted to demonstrate the marginal contribution of World Heritage listing to tourism by examining time series data for listed sites and comparable unlisted sites during simultaneous periods. Buckley observed several analytical problems. For example, he notes that World Heritage actually implies a "bundle of attributes which includes heritage value, branding, marketing and often increased infrastructure funding." (Buckley, 2007 : 70). In his analysis the branding effect of listing was not identified separately. Buckley’s analysis suffered from admitted data insufficiencies, which evidently is quite common for time-series studies of tourism.

8Guanajuato state tourism officials recognize the importance of the World Heritage inscription, but emphasize it only selectively in its marketing efforts. Their current marketing thrust appeals to younger people with higher disposable income. (Ayala, 2007). New marketing materials emphasize emotional aspects and downplay the region’s history. Recent brochures show attractive and fit young people in romantic situations surrounded by sumptuous food and gorgeous flowers. In these publications one occasionally might find historic structures as background.

9The Guanajuato officials confirm the notion that heritage tourism is not particularly popular with people from the United States, although a new campaign will target third-generation "paisanos" in the United States who may wish to rediscover their heritage. (Ayala, 2007).

10Still, heritage tourism continues to have a big role for selective markets, such as Europe and Asia. Reportedly, UNESCO complaints about garbage in Guanajuato and a subsequent threaten to withdraw the city’s World Heritage designation, prompted a major wave in investment in the past four years, which resulted in a change in design for streets and street fixtures in the heart of the tourist district. Critics point out that many of the new street paving materials and European-inspired designs for their layout on the streets and alleyways are inappropriate to the location and history of Guanajuato. Likewise, new street lamps have been criticized as inauthentic and inappropriate.

11Critics also highlight the sociological drawbacks of world heritage tourism in places like Guanajuato, where the "sustainability of cultural heritage is threatened" and a veritable "class ghetto" is created. (Diaz, 2007). A profound example of this in Guanajuato is the Mina de Guadalupe Golf Course created from the massive ruins of an 18^th century mine in nearby Valenciana.

12A golf course, 100-room hotel, and 21 adjacent private houses may provide a financial mechanism to sustain the ruins of an imposing historic structure (also known as the "Elephant Mine" for its flying buttresses). (Club de Golf Mina Guadalupe, 2007). However ; the World Heritage inscription for Guanajuato includes this adjacent mine. Limiting access to this source of local heritage to those with the ability to pay $40 U.S. for nine holes of golf (plus $20 for a cart) obviously creates barriers for many to appreciate the architectural and historical contributions of a World Heritage Site. With a target market of well-off Americans, the price of preservation in this case has been loss of access for the majority of Mexicans to part of their heritage.

13Other critics note that another project adjacent to the Mina de Guadalupe golf project threatens ecological resources in El Orito ecological preserve, including medicinal plants. (Patlán, 2006). A group wants to reconfigure Guanajuato’s main public market, Mercado Hidalgo, into an upscale mall, including a Sanborn’s facility. (Diaz, 2007). Numerous econometric analyses have been performed over the years to develop forecasting tools to better understand the variables that drive tourism and develop better-informed projections for the economic benefits of tourism. While they don’t specifically address the independent effect of World Heritage, discussion of specification issues they highlight may be helpful for someone approaching the general subject.

14This preliminary analysis seeks to develop a regression model which would indicate the relative importance of the 1988 World Heritage listing vis-à-vis other variables which might be found in Buckley’s above-mentioned bundle of attributes.

15The instant analysis confronted many of the same problems confronted by Buckley. It is difficult to find consistent time-series data for both tourism and economics going back to the years before 1988. Often data are unavailable. At other times, data may show unexplainable bumps or dips in some years. Conversations with people closer to the issue suggest consistent reporting is often a problem in an economy such as Mexico’s where commerce is relatively unregulated and data or taxes are collected less stringently than in other countries.

The Problem

16Since 1972 UNESCO and the World Heritage List has sought “to encourage the identification, protection, and preservation of cultural and natural heritage around the world considered to be of outstanding value to humanity." (UNESCO, 2000). There is anecdotal evidence that a subculture of well-educated and well-healed tourists seeks to visit World Heritage Sites. The extent of this tourism apparently has not been successfully analyzed.

17This analysis seeks to explore the statistical relationship between membership in the World Heritage List and tourism for one particular site – the Mexican City of Guanajuato and Adjacent Mines.

18If a statistical relationship is proven a model could be used by planners and economists to better accommodate the effects of a World Heritage Listing. For example, if the listing can be demonstrated to have produced an increase in tourism in Guanajuato that is independent of other possible influences, then statistical inferences can be drawn for the model for other prospective sites. These inferences can be used, for example, to better plan needs for additional infrastructure, the overall impact on the economics of a site and affected adjacent territory, and the need for additional facilities to accommodate tourism.

19Discussions with local officials and academics in Guanajuato intended to identify a wide array of variables and solid data which will contribute to a useful analytical tool. Downstream, further analysis and additional input from sources in Mexico and elsewhere may be refined to better identify and refine the independent statistical relationship between the World Heritage Listing and tourism.

Review of Previous Work

20Economic models have frequently been used to understand the effect of tourism on an economy. However ; many are prospective models used to forecast the effects of tourism.

21A study developed to assess the economic impacts of designating the Cornish mining area as a World Heritage Site concluded : "The evidence from other WHS sites and tourism and economic development professionals that there will be positive effects of up to 10 % increase in visitor numbers to Cornwall and West Devon." (Atlantic Consultants, 2003 : 79). The study also observed that success in tourism development from inscription in the World Heritage List is highly dependent upon marketing efforts.

22A study by Ralf Buckley of Griffith University in Queensland, Australia attempted to demonstrate the marginal contribution of World Heritage listing to tourism by examining time series data for listed sites and comparable unlisted sites during simultaneous periods. Buckley identified numerous analytical problems which would be helpful for the instant analysis. For example, he notes that World Heritage actually implies a "bundle of attributes which includes heritage value, branding, marketing and often increased infrastructure funding." (2004 : 72). In his analysis it the branding effect of listing was not identified separately

23Buckley concluded that "Most of the World Heritage Areas (WHAs) considered here received several times more visitors than the control sites, but it is not clear whether the difference is because WHAs are larger or more accessible, because they are better-known, because they are listed as World Heritage, or because they contain features of natural or cultural heritage which the others do not." (2004 : 82).

24Literature on the economics of tourism has been useful for better identifying model specification issues, such as influential variables, or the need to lag time-series data. With respect to the latter issue, in the case of Guanajuato, the effect of World Heritage Listing might not necessarily be immediate insofar as there may be a lag between the original designation and tourist awareness of that designation. Likewise, the local tourism industry cannot immediately respond to the World Heritage inscription with expanded tourism facilities. Without new hotels, restaurants, and transportation improvements, increased tourism may be impossible within the short term1.

25A work entitled The Economics of Tourism – in particular, Chapter 3, "Empirical Studies of Tourism Demand, " provides some useful guidance for model specification. (Sinclair and Stabler, 1997 : 35-47). For international tourism, exchange rates might be relevant. It has been argued that tourists are usually unaware of inflation in visited countries and that effective exchange rates (adjusted for inflation) may not be meaningful indicators. There is some suggestion that lagged variables should be considered. In the context of this analysis, it might be useful to consider that consumer awareness of World Heritage and World Heritage in Guanajuato might have lagged the 1988 designation.

26The book indicates “It is likely that the demand for tourism in a particular destination in a given period depends upon demand in the previous period. This is because the demand for tourism in an alternative location may be deterred by consumers’ lack of experience and sometimes knowledge of it. Thus, it is commonly assumed that the more information consumers have about a destination, the greater the demand for it." The authors also observe that it has been determined from various studies that transport price variables have often been insignificant, citing J.S. Little, 1980 and Strong and Redman in 1982.

27The authors suggest most studies of demand have included independent variables on an ad hoc basis. One solution, they suggest is to include all possible relevant independent variables and then exclude those that are insignificant.

Data

28Unfortunately, it is impossible to follow that last suggestion if the data are not freely available. In an ideal world, one would observe the relationships of a range of possible independent variables to tourism in Guanajuato before and after the World Heritage inscription in 1988.

29These variables might include the following :

• Income growth for visitors – a measure which would probably be more useful if the analysis is stratified by visitors’ country of origin.
• Infrastructure improvement expenditures to Guanajuato’s historic center and the mine area.
• Incidence of strikes, economic and political upheaval, or natural events (such as floods or storms) which may interfere with tourism.
• Expenditures for tourism marketing.

30Many of these data for the relevant time series are either unavailable or not recoverable during the time available for completing this analysis. The most important data identified are 1986-2006 figures for foreign and domestic tourism which were provided by El Consejo de Promoción Turística de Mexico, the Mexican Tourist Promotion Board. A look at these data and some of the trends they describe would be instructive before attempting any model specification. (The figures are reproduced in Appendix A.) Figure 1 below demonstrates a few trends. With respect to both Guanajuato tourism and Mexican tourism in general, an upheaval of sorts occurred after 2001 – evidently a result of the events of September 11, 2001 and the post-technology bubble economic downturn

31in the early years of the decade, which had a negative effect on economic activity worldwide.

32Note, however ; a puzzling difference in the trend in visits to Guanajuato and to Mexico in general during that period. Between 2002 and 2003, total visits to Guanajuato dropped 22 %, while there was a 5 % increase for all of Mexico. The decline in visits to Guanajuato was shared by both Mexican tourists and those from other countries, with Mexican tourism declining 22 %, while foreign visitors declined 29 %.

33Figure 2 below further illustrates and perhaps amplifies this trend. While there was a precipitous decline in visits from both Mexican and non-Mexicans between 2002 and 2003, foreign visitors to Guanajuato declined by over 40 % between 2001 and 2002, while Mexican visitors increased by nearly 5 %.

34If World Heritage inscription is an important draw for tourists, especially culturally curious, affluent types, then one would likely expect the share of foreign tourist visits (versus domestic Mexican visits) to increase and perhaps reach a plateau.

35These data appear to indicate a general trend of increased non-Mexican share until 2000, when over 13 % of Guanajuato’s visitors were non-Mexicans. The sharp decline in non-Mexican tourism after 2000 might be explained by the combined effects of 9/11/2001 and economic slowdown. As non-Mexicans are more likely to arrive by airplane than Mexicans, the public’s reluctance to fly post-9/11 probably had more impact on the foreign segment of Guanajuato’s tourist market.

36Interestingly, Figures 1 and 2 suggest the economic turmoil in Mexico which was experienced in 1994 as a result of peso devaluation (CIA World Factbook, 2007) had no obvious effect on tourism to either the city of Guanajuato or Mexico at large. The relationship between the relative value of the Mexican peso and other securities is highlighted in Figure 3 below, which shows annual foreign visitors to Guanajuato and pesos per Special Drawing Right (SDR). SDRs are the unit of account of the International Monetary Fund and other international agencies and represent a basket of securities. Presently it includes the euro, Japanese yen, pound sterling and U.S. Dollar. (International Monetary Fund, 2007).

37The observed data (and the graph) show a fairly strong positive correlation (r=.831) between the number of Mexican pesos per SDR and foreign visitors to Guanajuato prior to 2001. This means that as the peso weakens (and it requires more of them to buy a basket of world currencies) foreign tourist visits to Guanajuato simultaneously increase. Curiously, after 2001 the data show a slightly weaker negative correlation (r=-.704) between a weak peso and foreign visits to Guanajuato.

38One should consider both the accuracy and relevance of the tourist data for Mexicans. Presumably, data for foreign tourists would come off the "Migratory Form for Foreign Tourist" or tourist card, collected at entry points. The responsibility of this information is the National Institute of Immigration (Sistema Integral de Operación Migratoria del Instituto Nacional de Migración). This system is relatively new. Before 2004, the number of foreign tourists was calculated by the Mexican central bank : Banco de México through an official survey called the "Cuenta Satélite de Turismo". Reportedly, these earlier data were not completely accurate as they were calculated rather than determined from a census. It is unclear how information for Mexican visits are collected or how accurate they would be.

39Furthermore, Mexican visitors to Guanajuato have a wider range of purposes than one would generally encounter with foreign tourists. While local tourism officials may wish to target affluent third-generation Mexican-Americans, many are far more likely to come to Guanajuato to visit family than non-Mexicans. Consequently, even accurate data about Mexican arrivals would be more likely to have more "noise" masking trends in World Heritage tourism.

40Another variable of likely importance identified in time-series data from the Mexican Tourist Promotion Board is available rooms. Figure 4 below illustrates the strong correlation (.929) or large linear dependence between values for hotel room availability and visitors – at least for the years between 1986 and 2000. Intuitively, one would think that room availability would affect tourism. Clearly there is a relationship, although, of course, correlation does not mean causation.

Statistical Analysis

41Absent – or, should we say, pending – much-needed data on infrastructure investment and tourist marketing expense in Guanajuato, it nevertheless remains possible to examine the statistical relationship between the World Heritage inscription and tourism. As discussed, we have reasonable data for visitors and hotel rooms in the city between 1986 and 2006. We know there is a strong relationship between available rooms and foreign visitors. We have seen evidence of a strong correlation between a weak Mexican peso and foreigner visits to Guanajuato – a relationship which vanished in 2001. When we add another variable to express the presence or lack of the UNESCO World Heritage Listing, do we see that there is explanatory value from that additional variable ?

42In statistics, one cans use a "dummy variable" to represent an attribute or categorical variable in a multiple regression – such as presence or lack of presence of designation as a World Heritage site.

43One might expect the World Heritage Listing to have a delayed effect on tourism. Guanajuato should not be inscribed on the list in December, 1988, and expect an immediate increase in foreign visitors beginning several days later. Indeed, the graph seen on Figure 4 suggests that if there is a relationship between the designation and foreign visitors to Guanajuato, there was a lag of several years.

44For this analysis, we initially assign a value of zero to the dummy variable during the period from 1986 to 1991 and one for ensuing years. In an iterative process, the year of interest is advanced until statistically significant coefficients can be estimated. Figure 5 above is a visual depiction of the relationships. Four variables are graphed : 1) foreign tourists (in tens of thousands), 2) hotel rooms (in thousands), 3) the number of Mexican pesos per SDR, and 4) a dummy variable with a value of zero for the period up to and including 1990 (the property during the twelfth session of the World Heritage Committee on 5-8 December 1988) and a value of one for years beginning with 1991.

45The trend lines for visitors, rooms, and pesos per SDR show a general linearity over the period, with more substantial peaks for incoming foreign visitors in 1991 and 1995-2000 and a sharp elevation in the number of pesos per SDR with the currency revaluation in 1994. As noted earlier, the relationship between the value of the Mexican peso and foreign visits to Guanajuato changed radically in 2001. Further, the trend for available rooms generally shows a slow and steady increase, curiously with some declines in several years – most notably 2003 (down 163 rooms) and 2006 (down 163 rooms).

46In light of different behavior of visitor and currency data after 2001, a second dummy variable was added to capture the independent effect of events in that year. This variable was assigned a value of zero for the years up to and including 2000 and a value of one for the 2001-2006.

47The expected model would isolate the independent effects of these variables, enabling those interested to best understand the effect of the World Heritage listing on the number of foreign visitors to Guanajuato. Data were analyzed using fixed-effect models including the two dummy variables where the covariance structure is autoregressive of lag 1. Seven models were analyzed using the Proc Mixed procedure of SAS.

48Using the data shown in the chart above, the values for hotel rooms (variable "x₁" pesos per SDR (variable "x₂), and the dummy variables for World Heritage list inscription (variable "d₁") and 9/11 (variable "d₂" were regressed against foreign visitors to Guanajuato ("y"). This procedure was repeated with variable d₁assuming a value of one in years subsequent to 1991 until a p-value of less than .05 was found for this variable. Above this level, one could not conclude the probability coefficient of that variable is significant at that confidence interval. Because the data corresponding to each year were exhibited some dependency on data from the previous year, the analysis employed an autoregressive covariance structure of lag 1. These results appear below in Figure 6.

49As Figure 6 indicates, the p-value for d₁ (dummy variable for World Heritage designation) only achieved a value below .05 when 1997 was assume to be the year at which the World Heritage designation had an effect on the number of foreign tourists to Guanajuato. The results suggest there was an eight-year lag before there was an unquestionable positive effect of the World Heritage designation on foreign visitors to Guanajuato, Mexico. In terms of goodness of fit, when d₁ is assigned a value beginning in 1997, the Second-Order Akaike Information Criterion (AICC) has a value of 335.3. This is lower than those calculated for all other years in the sample except for the model with the designation taking effect in 1991, with a value of 335.1, and 1992, with a value of 333.9. The AICC weight for the selected model is .220, versus .243 and .443, respectively, for the 1991 and 1992 models. This suggests the 1992 model is twice as likely to have a better fit than the model where the designation has effect in 1997. However ; as shown in Figure 6, the coefficients for that model are not significant at a .95 confidence interval and were rejected. Further, the coefficient for d₁ in the 1992 model has a negative sign, which defies logic because it is impossible to imagine the World Heritage designation would depress tourism.

50On this basis, we select the model which is specified with the designation having an effect in 1997 and beyond.

51As Exhibit 6 shows, all coefficients, except for x₂ – pesos per SDR – are significant at a .95 confidence interval. (Pesos per SDR has a p-value of .189, indicating it is significant at .811.) (SAS output for the model is reproduced in Appendix B.)

52On the basis of the given data, the estimated coefficient for d₁ suggests Guanajuato’s inscription on the World Heritage list was responsible for an increase of 13,845 annual visitors to Guanajuato which occurred by 1997 – all else remaining equal. In 1997 there were 44, 475 non-Mexican visitors to Guanajuato. Consequently, the model implies 31.1 % of these resulted directly from the Guanajuato’s 1989 inscription in the World Heritage list.

53Likewise, based on the given data, the coefficient for d₂ – the 9/11 variable – implies a decline of over 28,000 foreign visitors to Guanajuato beginning in 2001 – all other variables remaining constant. The difference between the reported number of foreign visitors in 2000 (62,536) and the average number of visitors between 2001 and 2006 (22,810) is 27,399 – 2.28 % less than the change forecast by the model.

Implications and Further Research

54The above analysis suggests a bump of over 13,000 annual non-Mexican tourists to Guanajuato over the study period (1986-2006) as a consequence of the World Heritage Listing. This increase is independent of any increases in hotel rooms or decreases in the value of the Mexican peso, which would also encourage visits by foreigners. It also is independent of the negative effect on tourism which resulted from the 9/11 attacks and changes in their aftermath. These numbers should be construed as order-of-magnitude estimates of the independent effect of inscribing Guanajuato on the World Heritage List in 1988.

55One may also conclude that – at least from a statistical perspective – Guanajuato’s World Heritage inscription in December, 1988, had no statistically significant impact on the number of non-Mexican visitors until 1997.

56Down the road, it would be useful to validate these findings with similar data for other Mexican towns with World Heritage inscriptions, such as Morelia, Oaxaca, Zacatecas, Campeche, Morelia, or Querétaro. (See Figure 7 below.)

57Figure 7 (above) compares the trend in foreign visitors to five Mexican cities inscribed on the World Heritage list and highlights the year of inscription. The image suggests similar relationships may be found in cities other than Guanajuato. Likewise, it would be interesting to use a similar technique to compare the Guanajuato results with those demonstrated by World Heritage cities in other parts of the world.