A PRELIMINARY ECONOMIC IMPACT FORECASTING MODEL FOR SHORT-TERM TOURISM EVENTS: A PRELIMINARY FRAMEWORK

The tourism industry is the fastest growing industry and the third largest employer in the State of Rhode Island. In terms of industrial output, tourism is currently estimated to generate about five percent of the State's gross domestic product. Rhode Island attracts an estimated 5 to 8 million visitors annually. Among these are visitors to Rhode Island's short-term events. The sales impact that is generated in the business sector, the personal income and employment that is generated for individuals and households and the government revenue provided through taxes qualify it as a major economic force for the future. The measurement of such impacts is a challenge to researchers because of the diverse nature of the spending groups present and the fact that the services they demand are also demanded by the residents of the State. Traditionally, economic impact studies have focused on estimating sales impacts of short-term tourism events. Predicting these impacts has not yet been attempted. This study attempts to make some headway in this direction by developing a framework for forecasting the impacts of short-term tourism events. The economic impact of a short-term event can be described as the product of an average participant's expenditures and total participation in visitor-days. The forecasting framework developed in this study is composed of these two parts: a participation forecast and an expenditure forecast. The expenditure forecast is based on a model which relates expenditure per person-day to numbers of persons per group and days per visitor at the event. This model is used to predict total expenditures. An allocation model is used to estimate expenditures on particular categories of event goods and services. The participation forecast is a synthesis of empirical and expert judgmental prediction estimates. The empirical forecast is based on a model relating numbers of visitors per event to prices, weather and consumer perception. The empirical and judgment estimates of participation are weighted according to the proportions of total variance and apded together. 1his synthesis is related to the Bayesian procedure of revising belief in the light of new information. Some of the questions which remain unanswered by this research include: How to devise methods of measuring the reliability of subjective estimates by experts; How to construct a detailed typology of events for impact assessment; and How to incorporate event and site reputation into a dynamic forecasting model. Specific answers to these questions could have a strong influence on future data collection and updating procedures. General answers will enable transference of the forecasting procedure to other states. The preliminary model derived in this thesis should serve as a useful guide to researchers and potential sponsors of future Rhode Island events. It is hoped that the State of Rhode Island Department of Economic Development might use the results for logistic and

2) The Boaters, i.e owners and competitors in race events

3)
The exhibitors and dealers participating in show events

4)
The Trade Patrons attending the event

5)
The spectators or general patrons.

ai = ai(A).
The consumer's utility function U(X(ai(A)),Y), is assumed to be quasi-concave and twice differentiable. The possibility of "product improvement 11 .and its effects on demand enter through the perception variable.
The consumer's problem can be stated as a constrained maximization problem and is expressed as Maximize where c represents the personal cost of information to consumer.
In this formulation, the information or advertising variable could be exogenous or endogenous 58 depending on whether participants actively seek information on the quality of the events.
Given the above maximization problem, and assuming information is exogenous (i.e., C=O), the optimal policy for consumers is determined by the first order conditions of the Lagrangian: where subscripts denote first partials.
The first order conditions suggest that the demand for the event is a function of pricesof X and Y and A which is an exogenous variable. Implicitly the function x 1 = X 1 (Px,Py,A) represents the demand for x 1 for any combination of prices and advertising.    As before, consider an individual who maximizes utility subject to a budget constraint but now includes congestion as a negative characteristic of a tourism event.
If all individuals react the same way to congestion, the average consumer will reflect market behavior.
Panel A in Figure   In a generalized form, a dynamic model of reputation could be experessed as  Substituting the estimated parameters of c 0 and fA into 4.3.5 will yield p(a1 + 262A + 63R) = a 1 (a 1 + 2a 2 A + a 3 R) + Aa 1 +Ao( al + 2 a2A + 63R) + d.
Solving for A: (pal -Ao)262 Therefore A= f(p,d~R).    Thus it has been reported that an industry in which the region specializes will have an LQ greater than one while an industry in which the region does not specialize will have an LQ less than one.
To determine the proportion of the total sales of the regional industry that is accounted for, the sum of the national direct-requirements coefficient is multiplied by the regional location quotient if LQ(i) is less than one or one if LQ(i) is greater than one. The implication, therefore, is that the supplying industry will certainly not supply more than the demanding industry requires, even if the supplying industry is substantial. A regional matrix for all the industries within the region could thus be dev$loped while omitting the household sector. v.2.2 The REMI Models: The Regional Economic Models Incorporated Most recent work in time series modeling has followed the Box-Jenkins (1970) method whereby seasonality in the data is accounted for by using a stochastic difference operator. Wu (1977) reported that the use of this operator may be ineffective loss.  Green and Srinivasan (1978).
The discrete choice judgment models employ the use of probit and logit analysis to study discrete choice judgments.
With this procedure, participants are asked to make such judgments as "will attend" or "will not attend" for each event.

That is
Weight of a Specific Event                            The sign of the coefficient on travel cost is consistent with a priori theory and its t statistic is significant.