This summary is focused on showing the forecasting techniques used to determine the likely demand in tourism and argues that given the importance of the tourism sector to the economy of any tourist country, accurate forecasts of tourist arrivals are of importance for planning by both the private and public sectors. First we should answer the question what tourism is itself. It is obvious that tourism industry is not one company. It combines thousands of products and services. A company sets goals and uses its production, marketing and managerial resources to achieve them through its management process.
And in tourism there are too many companies involved and too many goals are set, but almost everything in this industry depends upon the visitor numbers in other words demand. This is the main target of forecasting. It has been pointed out that forecasting is useful in shaping demand and anticipating it to avoid unsold inventories and unfulfilled demand. Moreover since consumer satisfaction depends on complementary services, forecasting can help to anticipate the demand for such services.
As well it helps optimizing the use of public funds, in other words save money.
It should be mentioned that a fall in demand can bring about decreases in living standards following the rise in unemployment, while increased demand can lead to higher employment, income, output and inflation as well may threaten environmental quality and sustainability. Moreover, tourism firms are confronted by changing revenue and profits and governments experience changing tax revenue and expenditure. Thus, tourism demand effect can be observed in all sectors of economy – households and individuals, public sector and private businesses.
For example, decisions on tourist expenditures, the tourism markets structure and decision-making nature between them, cross-country linkages between tourism firms, the contribution of environmental resources and their relevance to policies for sustainable tourism have not been fully investigated and need further economic analysis. Aim. The paper is aiming on showing the existing forecasting techniques, their positive and negative features for better understanding the importance of demand forecasting in tourism, and the necessity of using these or those methods for obtaining the most accurate and precise results.
It is obvious that one of the more complex aspects of tourism is the tourism demand. As a rule it is defined and measured in a variety of ways and at a range of scales. Generally, there are economic, psychological and social psychological methods used in forecasting. For example, decision to purchase holidays are often made with friends and family so that consumer demand theory based on individual decision-making must take account of individuals` and groups` social contexts.
As well as the analysis of travel patterns and modes has been dominated by geographical analytical frameworks, while the study of demand outside economics tends to be underpinned by psychological or social psychological methods. ‘The many studies of tourism demand in different countries and time periods are reviewed by Archer, Johnson and Ashworth, Sheldon and Sinclair while Witt and Martin examined alternative approaches to tourism demand forecasting. ’ (Sinclair, 1997). The significance of tourism demand provides a strong case for better understanding of the decision-making process nature among tourists.
In case of using an inappropriate theoretical framework in empirical studies of demand can result in incorrect specification to estimate tourism demand and biased measures of the responsiveness of demand to changes in its determinants. It should be mentioned that ‘empirical studies help to explain and understand the level and pattern of tourism demand and its sensitivity to changes in the variables it is dependant on. For example, it helps in observing income in origin areas, exchange rates between different destinations and origins as well as relative rates of inflation.
This type of information is of importance to public sector policy-making and the private sector. ’ (Sinclair, 1997). But only in case of appropriate theoretical specification of the studying model used the estimates can be accurate and precise. Hence, explicit consideration of the consumer decision-making supporting empirical models is of importance in presenting the provided estimates are neither misleading nor inaccurate in their policy implications. Thus there are two approaches used to model tourism demand.
First one is the single equation model and the second is the system of equation model. ‘The first one single equation model has been used in studies of tourism demand for numerous countries and time periods and states that demand is a function of a number of determining variables. ’ (Sinclair, 1997). This technique permits the calculation of the demand sensitivity to changes in these variables. Contrary to the first approach, the system of equations model requires the simultaneous estimation of a tourism demand equations range for the countries or types of tourism expenditure considered.
The system of equations methodology tries to explain the sensitivity of the budget shares of tourism demand across a range of origins and destinations (or tourism types) to changes in the underlying determinants. There exists one more forecasting technique which is more recent and can be compared with the single equation approach. It is the Almost Ideal Demand System (AIDS). (Maria De Mello,1999). This model is theoretically better than the mentioned above and offers a range of useful information concerning the sensitivity of tourism demand to changes in relative prices and in tourists` expenditure budget.
This approach was used in examining the UK demand for tourism in its geographical neighbor-countries as France, Spain and Portugal. The result of such investigation indicated that ‘the UK demand for tourism in Spain increased more than proportionately with respect to a rise in the UK expenditure budget for tourism in three countries, the demand for tourism in France increased less than proportionately and the demand for tourism in Portugal remained stable.
The sensitivity of the UK demand for tourism in Spain to changes in effective prices in Spain is increasing and exceeds the corresponding values of the sensitivities of the demand for tourism in France and Portugal to changes in French and Portuguese prices, respectively. (Maria De Mello,1999). ‘In contrast, the UK demand for tourism in Spain is insensitive with respect to changes in prices in its smaller Portuguese neighbour.
The UK demand for Portugal is sensitive to changes in prices in Spain, although the degree of sensitivity appears to be decreasing over time, and the demand for France (Portugal) is insensitive with respect to a change in prices in Portugal (France)’(Maria De Mello,1999). So it is obvious that stability of demand in the face of rising prices may be observed as signals of success, and contrary outcomes mean a possible case for rethinking policy toward tourism demand. Scientists have used a variety of other forecasting techniques during the past decades for tourist industry.
Among them are quantitative forecasting methods. They may be classified into two categories: causal methods (regression and structural models) and time series methods (basic, intermediate, and advanced explorative methods). For further explanation we should mention that causal methods represent methodologies for identifying relationships between independent and dependent variables and attempt to incorporate the interdependences of various variables in the real world. However, there is certain difficulty of applying the causal methods. It is identifying the independent variables that affect the forecast variables.
So the accurateness and reliability of final forecast outputs made under causal methods depend on the quality of other variables. The second group of methods, time series quantitative methods, offers many advantages. It is pointed out that ‘the use at time t of available observations from a time series to forecast its value at some future time t+1 can provide a basis for (1) economic and business planning, (2) production planning, (3) inventory and production control, and (4) control and optimization of industrial processes’(Chen, 2003).
Time series methods offer techniques and concepts facilitating specification, estimation and evaluation. They acquire more precise forecasting results than those yielded by causal quantitative techniques. It should be mentioned as an example that forecasting is complicated by the strong seasonality of most tourism series. It is pointed out that to see seasonality as a form of data contamination is one of typical approaches to the analysis of macroeconomic time series. This was the approach often used in many census and statistical departments.
In the case of tourism analysis seasonality is integral to the process and is of high importance for the timing of the issuance of policy measures in addition to studying the long run trend. ‘As significant features of quantitative tourism forecasting (involving the numerical analysis of historical data) we see that while it is particularly useful for existing tourism elements, it is limited in its application to new ones where no previous data exists’. (Smith, 1996). This technique was used in forecasting potential UK demand for space tourism. Appendix 1, 2). (Barrett, 1999).
As well univariate forecasting techniques may be used to forecast arrivals. This limited methodology relative to structural models allowing policy makers to determine how changes in particular variables can help to improve the industry. The weak point of the technique is that the models have no explanatory variables so it is difficult to interpret the individual components.
Therefore, the forecasting record of many univariate models have considerable forecasting accuracy. Lim and McAleer employed univariate techniques to forecast quarterly tourist arrivals to Australia and to determine their forecasting accuracy using a variety of seasonal filters. Kulendran and King also employed a variety of models to rank forecasting performance of various tourist arrival series using seasonal unit root testing’ (Alleyne, 2002). Conclusions and Recommendations. It should be mentioned that forecasting techniques and forecasting itself have some weak points. Firstly, current forecasting is mostly the domain of policy makers.
It is beneficial for three groups: public sector tourism organizations as it helps justify budget allocations; managers of public and private sector tourism projects as they may encourage investors, and the forecasters themselves. There are no actual benefits from forecasting for tourism operators and suppliers because the results are not actionable and unrelated to the real needs of the majority of tourism businesses. The problem with the results may be illustrated by such an example. (March, 1993). ‘The BTR’s “Australian Tourism Forecasts” report released in April 1990 forecasts 4. 85 million visitors by the year 2000.
The BTR’s latest forecast for 2000 is 4. 824 million visitors. And only last month The Australian newspaper (Oct 12 1993:p. 6) reported the results of “a respected private sector forecaster” who has forecast 5. 33 million by the end of the decade’(March, 1993). So you see numbers keep changing and this is the evidence that sometimes the forecasting results become not actionable. Summarizing all the mentioned above we may say that there is a wide range of techniques used for forecasting demand in tourism. In this paper we mentioned only some of them which to our mind deserve attention and may be used in forecasting the demand.
As you may see investigation of tourism demand involves specific problems because it has some special nature attributed to the complexity of the motivational structure concerning decision-making process as well as the limited availability of the necessary data for econometric modeling. Quantitative approach for tourism demand needs the framework of a formal mathematical model providing estimates of sensitivity to changes in the variables the demand depends on. Econometric modelling offers a good basis for accurate forecasting which is of great importance to the public sector making investments in the industry.
The single equation model often ignores the dynamic nature of tourism demand, disregarding the possibility that the sensitivity of tourism demand to its determinants may differ between periods of time. The alternative model is the Almost Ideal Demand System. It is originally developed by Deaton and Muellbauer. This model not only permits the estimation of the complete set of relevant elasticities, but also allows for formal tests of the validity of the assumptions about consumer behaviour within the sample set of observations.
The AIDS allows to test assumptions and estimate parameters in a way which is not possible with other alternative models. So for now, we may say that there are no completely bad or good techniques used for forecasting tourism demand. But there are preferable models for getting more accurate results. It is better using models based on old theoretical knowledge but with new trends able to cover all the necessary aspects in forecasting tourism demand.