The complex model can be described as taking an optimistic perspective to forecasting international tourism demand for Greece since it generates rather high forecasts. Although diagnostic statistics show that the complex model provides a better fit to the historic data, ex post forecasts indicate that the forecasting accuracy of the simpler model is considerably better. Still, forecasts from the complex model are quite similar to some other predictions, suggesting that in the period 2001 to 2005 inbound tourism in Greece is likely to grow rapidly. The 2004 Olympic Games will be responsible for much of this growth. Satellite events that will begin as early as 2002 are expected to cause the preparations and the publicity effect to start rather early.

It is estimated that during the period that the games will be taking place, 235,000 international tourists will visit Greece. Overall, international tourists arrivals are expected to exceed 12 million by the year 2002 and to reach 15 million by 2005. During the Olympic year, this figure is expected to lie somewhere in the middle of the 12.9 millions and 15.3 millions region.

In its very basic form, forecasting involves making inferences about the future from reviewing the past. Tourism demand forecasting can help those involved in the tourism industry, a) to estimate the future levels of demand for tourist products/services, b) to identify the determinants of the demand for tourist products/services, c) to assess the impact of policy changes on tourism demand. The need for forecasts is particularly acute in the tourism industry because tourism products cannot be inventoried, production and consumption occur simultaneously, and most operations involve substantial amounts of fixed costs and investment. Indeed, forecasting is a core element of the planning process and a prerequisite to effective planning. Forecasts, whether explicit or implicit, are necessary in day to day operations as well as for taking strategic decisions.

Accurate forecasts of international tourism demand are particularly important to both public and private organizations. In the public sector, accurate forecasts on the number of arrivals can be used to assess the adequacy of the existing infrastructure and facilities, estimate future revenues and assist government budgeting. The private sector can use forecasts to avoid surpluses or shortages in the provision of goods and services, estimate the correct number of personnel required and assess the viability of expansion projects.

Quite often, tourism demand forecasting contributes information which is used to take crucial decisions. The value of any forecasting model thus lies to its ability to generate accurate forecasts. Nonetheless, the majority of previous forecasting models test theoretical specifications (Witt and Martin, 1987a, Uysal and Crompton, 1984, Dharmaratne, 1995). In fact, model evaluation has traditionally been confined to a comparison of different forecasting techniques in terms of a single accuracy criterion. These comparisons have often shown simpler forecasting methods to outperform the more complicated ones. However, it is believed that this is due to using some very basic econometric and time series methods for comparisons. Hence, it is possible that more advanced forecasting such as ARIMA are capable of generating more accurate forecasts than the commonly used simpler ones.

Tourism in Greece as well as in several other Mediterranean countries is one of the important sectors of the economy. Apart from its important contribution to the gross domestic product, tourism has repeatedly helped Greece to cover its trade deficit. However, the decline in international tourism arrivals during 1995 and 1996 reveals that tourism in Greece has undergone a serious crisis. Nonetheless, inbound tourism started to increase again from 1997, and this growth has been sustained until now. Since, the problems which brought about this recent crisis have not been resolved yet, both the private and public sector need to be conservative rather than optimistic about the growth of international tourism demand. Therefore, accurate forecasts of international tourist arrivals are believed to be necessary in planning for the future in tourism in Greece.

Nevertheless, some optimism about the growth of inbound tourism is justified, especially since Greece will host the 2004 Olympic Games. Normally this event attracts a large number of visitors, and not only during the period of the games. The extensive publicity can act as a magnet, attracting tourists before and after the event has taken place. Yet, attendance to the Olympic Games has been rather difficult to predict so far. Indeed, efforts to forecast the demand for past venues have largely been unsuccessful, with hope triumphing over experience. Thus given the critical state of tourism and the economy of Greece, authorities have to include in their plans for the future the impact of the Olympics on international tourism arrivals.

There is a recognized need for accurate forecasts of tourism demand. According a number of criteria, two models are suitable for forecasting international tourist arrivals to Greece. They belong to the general class of seasonal autoregressive and moving average (ARIMA) models but the first model is of higher order and thus slightly more complicated than the second one.