Instituto Tecnologico y de Estudios Superiores de Monterrey Campus MonterreyDeterminants of Tourism DemandFor MexicoEmilio Noe Hernandez Kelly*Fernando Mendoza Lopez** Econometria IIDr. Hector RodriguezMonterrey, Nuevo Leon, November 29th, 2004 I. INTRODUCTIONTourism has long been considered a viable option for growth in many less Developed nations. It has been widely accepted that tourism is a low investment, high return industry making its profitability extremely high. In Mexico, the tourism industry directly provides around 8% of the GDP, and leaves even more through spillover effects.
Mexico owns many tourist attractions of different types. For example, Mexico has world renowned beaches and resorts, such as Cancun or Los Cabos. It also has many colonial cities, which mix a classic Spanish-European type architecture with an indigenous touch. Ancient ruins of lost civilizations still persist in the central and southern areas of the country. With all this and more Mexico has much to offer for international tourists. International tourist flows have increased amazingly since the 60’s and 70’s creating a market that is tremendously important on a global level.
It is an industry that can be common to both developed and less developed nations. Mexico, being one of the top tourist destinations in the world has seen this growth first hand, and seemingly has also done very well in this respect.
It is interesting to remark that despite Mexico’s apparent diversity and cultural heritage, its development on foreign tourism, measured in currency flows, is represented on a 54% by five destinies that are promoted by the FONATUR (National Fund of Tourism Foment) which are: Los Cabos, Cancun, Loreto, Ixtapa and Huatulco. These destinations represent a tax income of over 300 million dollars per year. Obviously, there are many people who believe that this industry also brings many disadvantages to having a strong tourism sector .There are many negative factors to consider: first of all, there are some negative social and environmental impacts that are brought upon by foreign tourists. Also there are some distributive issues that should be assessed although none of these matters will be touched in the rest of this paper.
We focus on measuring performance so it is necessary to know what variables have some kind of explanatory properties. A logical question arises: What factors determine growth of tourism industry? In the literature found, that speak of this topic; we see estimations, basically defining the demand side of the equation. We may infer that is because of the difficulties that are represented by the development of a production function for this service oriented industry (although it is possible to estimate).
The objective of this investigation is to try to explain the behavior of the industry. We try to find the determinants of income growth in the tourism sector for Mexico.
We will explain the performance of the Mexican industry based on a time series econometric model estimated with ordinary least squares (OLS) for the 1980-2002 period.
II. A TOURISM INDUSTRY MODELThe model used in the elaboration of this paper is a demand model, and for this reasons takes in account prices and income. The treatment of this model assumes that the measuring of foreign tourism revenues can be put into a model consistent with a model for export demand set forth in other work. The model we are proposing is based on a typical model described by Muscatelli, Stevenson and Montagna, (1995), who establish an equation for Volume of Exports as a function of a price vector (that includes domestic and foreign prices) and a variable that captures the world demand conditions. Following the next form:X = level of exports.
Pw = world prices.
PX = domestic prices.
Yw = world share imports of domestic products.
In this paper we define our model for foreign tourism revenues as follows:TURISMEX = an index of foreign tourism revenues.
ERI = real exchange rate index (MXP/USD).
USGDP = United States of America GDPIt is plausible to select these variables, due to the implication of prices in the calculation of the real exchange rate (Muscatelli, Stevenson and Montagna, 1995). In the case of the second variable, the GDP for the US, we use this as a proxy for Income due to the United States’ predominant position in tourist flows to Mexico. The latter justified on the fact that around the 82% of foreign tourists that visit Mexico are Americans. The variable is measured in dollars.
Indexing the revenues of tourist activity with a 1994 = 100 basis, is on the seek of dismissing problems of seasonalitySo the model looks to explain tourism flows to Mexico. This variable is the amount spent by tourists (not accounting, border visits nor cruise revenues), measured in dollars. This variable was seasonally adjusted and indexed in two different ways. The data for the United States was already adjusted. The data for Mexico, the dependent variable and the real exchange rate (pesos per dollar) were obtained from the Mexican Central Bank (Banco de Mexico). The data for the US GDP was obtained from the Bureau of Economic Analysis for the United States Government.
An indexed real exchange rate depicts the trends of this financial indicator in order to use it as a determinant of our goal model. We expect that if the exchange rate increases, the revenues of tourism will be impacted positively. Based on the PPP analysis, this means that, the larger the exchange rate is, the cheaper it is for Americans to visit and spend vacations in Mexico.
We consider, also that the inclusion of the income part of the demand for tourism, if economic activity is encouraged in the U.S., would result in the American people assigning more of their income to vacation spending. Hence, if this resource assignment increases, it should be expected that tourism revenues for Mexico would also increase.
Now that we have established and defined the variables and expectations, we move to the results and interpretation of this model.
III. Results The model we have just established used data from 1980 to 2003, with trimestral periodicity. The data found includes 96 observationsThe dependent variable used here was converted into to different indexes that bring us to different results. The main model, that from here on will be referred to as model 1, used all data in logarithmic fashion. The first Index used (Model 1), was obtained by standardizing each observation, all this while considering that each trimester has a slightly different behavior. We obtained the average for each trimester and their respective standard deviations. After which, we proceeded to normalize the variables by trimester following Ba, Sulin and Paul A. Pavlou (2002). These values were later converted to an index that runs from 0 to 100, 0 being the lowest observation and 100 being the highest. With this we expect the variable to take a relative value form, compared to usual behavior for those trimesters.
We must also note that the independent variable, US GDP was taken as the lag for 4 periods. We base this on papers that argument that a vacation depends more on the income of the preceding year, and not on the actual one in progress. Models with Current trimester US GDP were ran, but returned ambiguous results, and seriously affected the outcome of the model in many cases. Model 1’s results returned as shown in the following table:VARIABLEBP-VALUE(Constante)0.7371541.18E-07LNERI0.2764988E-15LNGDP0.0261420.042706Table 3.1. Coefficients and P-values for model 1.
As we can see the signs are in accordance with what was expected and the variables are statistically significant to a five percent level. The R2 for this model was of .483, a modest yet significant explanatory power for this model.
The results show that for a 1 percent change in Real exchange rate, tourism relatively grows .28 %, and the US GDP helps out in .03 % percent. We must also mention that these variables were all corrected for autocorrelation by the Theil and Nagar method.
Dummy variables for the period beyond 1987 were taken in account, and did not show much significance. This was done because of the notable expansion of the sector in this year, mainly due to the devaluation of the peso, and low gasoline prices.
This model has the same independent variables only that the dependant was modified. We took only the observation for tourism and divided it between the averages of the respective trimesters that the observations were in. The R2 for the second model was of .726, there was a much better fit for this model. The model was also corrected by the Theil and Nagar method.
The Results for model to are as follow:VARIABLEBP-VALUE(Constante)-20.64791.71E-05LNERI1.7043934.83E-06LNGDP0.0748370.095059Both variables returned with the expected signs and were statistically significant (GDP was only significant at a 10% level).
This model shows that a 1.704% change in the level of tourism, as measured by our index, would come from a 1% change in the exchange rate. It also shows that the elasticity of tourism, with respect to US GDP lagged four periods, is of .07 percent.
Both these models show us that these two variables explain up to some extent the evolution of Mexican revenue from international tourists.
IV. Evaluation of MLS Assumptions1. LinearityGraph 4.1 In the first graph we can se the positive lineal relationship held by the US GDP and the revenues from foreign tourism, axis x and y respectively.
Graph 4.2In Graph 4.2 we can see the relationship between the revenues from tourism on the Y axis, and the Real Exchange rate on the other. Linearity holds well enough for these variables.
2. No Multicolinearity The measurement used here is the VIF indicator shows the following, for model 1.
VARIABLEVIF(Constante)LNERI1.053806LNGDP1.053806Table 4.1VARIABLEVIF(Constante)LNERI5.811601LNGDP5.811601These results show that they are clearly within acceptable parameters. For these reasons none of the variables were removed from the process.
3. HeteroscedasticityGraph 4.3With the graphic method we can see that gdp has no apparent relation with the error for the first model.
Graph 4.4In the second model the graph shows the same relationship.
Graph 4.5In Graph 4.5 we can see no clear relationship for the errors and the variable. The Graph for the second model appears almost exactly the same. (Not displayed)We checked for Heteroscedasticity with the Breusch-Pagan Method. The method returned good results for us. It showed no relation whatsoever between the independent variables and the errors. With this we consider the homoscedasticity supposition covered.
4. Autocorrelation For this assumption we used the VIF indicator, also having good results, for both models.
The results are the following: VARIABLEVIFVIF(Constant)Model 1Model 2LNERI1.0538065.811601LNGDP1.0538065.811601With this result it is obvious are model does not have the problem mentioned, so its validity cannot be questioned in this point.
V. Conclusions and Possible ExtensionsIt has been established and empirically demonstrated through the model proposed that there is a direct relationship between the foreign tourism revenues and the real exchange rate, and in a similar way, a direct relationship between the same variable and the US GDP.
Mexico, with its large offer in tourism attractions, can sustain a larger demand for tourism. We can deduct by now that random shocks (impacting the exchange rates) could alter the performance of the industry. Similarly in this sense, is recommendable to look for inner markets to develop in order to lose dependency of unmanageable situations, as the growth of US.
Another option for Mexico could be, diversifying the target market. It is easily sustained that spatial distance (Williams, 1970) plays a major role in determining tourism flow, but this does not mean, efforts should not be made towards seeking for vacationers from other parts of the world.
We can say that this is influenced in a great proportion by the American tourists’ decisions on vacationing and the planning of their authorities towards economic development. Then, as a very important conclusion, we should consider that the two variables of our model are exogenous for Mexico. We suggest estimating the regression adding variables which denote endogenous aspects as investment on tourism infrastructure or advertising of Mexican destinies in the US. One of the stronger barriers for achieving this analysis is the lack of information, because, many of the variables are recently measured and there is not enough data.
VI. BibliographyAscanio, Alfredo. PhD, (2004) Un modelo de correlacion multiple para el turismo internacional que llegaba a Venezuela desde 1971 hasta 1991. Caracas-Venezuela. Universidad Simon Bolivar. Departamento de Ciencias EconomicasMuscatelli, Vito Antonio, Stevenson, Andrew A, Montagna, Catia. Modeling Aggregate Manufactured Exports for Some Asian Newly Industrialized Economies, The Review of Economics an Statistics, Vol. 77, No. 1 (Feb, 1995), 147-155. MIT. PressLewis, H. Alexander. (1953) The impact of tourism on the economy of Cape Cod, Massachussets. Economic Geography, Vol. 29. No. 4, 320-326.
Williams, Anthony V, Zelinsky, Wilbur (1970). On some patterns in international tourist flows. Economic Geography, Vol. 46. No. 4, 549-567.
Zagal Arreguin, Hector. La acapulquenizacion de la economia. Articulo del portal mexicoweb.com. Descargado el: 30 de octubre de 2004.
VII. DataBank of Mexico.
http://www.banxico.org.mx/. Consulted on November 2004.
Bureau of Economics, http://www.bea.gov/. Consulted: November 2004National Fund of Tourism Foment (FONATUR)http://www.fonatur.gob.mx/Consulted on November 2004.
National Institute of Statistics, Geography and Informatics, (INEGI)http://www.inegi.gob.mx/Consulted on November 2004.
Tourism Promotion Council (PROMOTUR)http://www.promotur.com.mx/Consulted on November 2004.
Tourism Secretary (SECTUR)http://www.sectur.gob.mx/Consulted on November 2004.
World Tourism Organization.
Consulted on November 2004.
Cite this Determinants of Tourism Demand: For Mexico
Determinants of Tourism Demand: For Mexico. (2018, Nov 05). Retrieved from https://graduateway.com/determinants-of-tourism-demand-for-mexico/