Alkiviadis Panagopoulos

Department of Tourism Management, TEI of Western Greece, Patras, Greece

Sonia Malefaki

Department of Mechanical Engineering & Aeronautic, University of Patras, Rio, Greece, e-mail: smalefaki@upatras.gr

Ioannis A. Nikas

Department of Tourism Management, TEI of Western Greece, Patras, Greece

 

 

Abstract

The present work studies the well-timed issue of forecasting the tourism demand for the three prefectures of the Region of Western Greece: Achaia, Etoloakarnania and Ilia. The proposed approach consists in finding a proper model, for each prefecture, based mostly on the effectiveness and complexity of the proposed models. Utilizing the official statistical data for the tourist occupancy of all tourist accommodations (except camping sites) from the above three prefectures, for all months between 2005 and 2012, a trend and seasonality analysis has been realized in order to construct suitable models using the well-known ARIMA (Box-Jenkins) methodology. Then a series of statistical test have been employed to select the best fitted model to given data. The forecasting effectiveness of the chosen model is measured using the last twelve observations as a training set. Finally, a 12-period prediction for the three prefectures is proposed.

Keywords: Prefectures of Achaia, Ilia & Etoloakarnania, ARIMA, forecasting model, Western Greece tourism.

 

Introduction

It is a common truth that tourism is a perishable product which strongly depends on various factors as the current financial status, natural disasters, cultural events, social behaviors, marketing policies etc. In this work the issue of forecasting the tourism demand will be studied in the prefectures constitutes the Region of Western Greece.

Tourism industry in Greece is composed from very crucial economic activities and constitutes a valuable source of earnings; tourism employment, contribution in gross domestic product and multiplier effect investments. A key-factor of Greek tourism industry is its ability to host visitors in various places. Therefore, a desirable task is to forecast monthly and/or annually percentages of occupancy in Greece and in the various regions of this country. Potentially, this task leads to more effective use (allocation) of the available sources for the Greece visitors.

The current global and local financial crisis comes, also, to demonstrate the decisive role of tourism in Greek economy. In fact, Greek tourism managed to maintain its strength and prove to the state - but also the society - that with an effective support, the sector can become a driving force for the creation of more income in the country and the improvement of the economy competitiveness.

The perspective of local communities to gain strong currency, with high travel receipts fast enough, charms the national government, local authorities and groups of owners of capital, who pursue and encourage tourism. Tourism is an important source of income for many Greek regions, and especially for those with less developed modern service/industrial based economies, such as the Region of Western Greece.

The possibility of forecasting tourism demand at regional and local level, and particularly at different scales of the Region of Western Greece, will give firstly a clear picture of the development of tourism in the study region and secondly the possibility of continuous updating of this picture, and also prediction of the type and intensity of tourism demand (and especially hotel demand), at various spatial scales. So that any attempt for policy making of tourism development, in public and private sector, be based on a reliable depiction of the trends and patterns (such as seasonality) of hotel demand at different spatio-temporal scales. The continuous quantitative and qualitative expansion of such data would allow the formation of a framework for monitoring the evolutionary progress of hotel demand and consumption, and thus will play a key role in creating an integrated system of planning tourism development at regional and local level.

Forecasting is about predicting the behavior of future events (Makridakis & Hibon, 1979; Frees, 1996; Franses, 2004) and plays a significant role in tourism planning. Tourism investments should be based on professional business planning and on achievement vision of the industry future. The tourism industry needs to reduce the risks of poor decisions. One prompt way to reduce this risk is by discerning future events or environments more clearly (Smith, 1995; Burger et al, 2001). Benefits derived from forecasting are imaginable. In the case of forecasts of demands turning out too high, accommodation firms will suffer; there might be, for instance empty rooms in hotels, unoccupied apartments, and so on. If, on the other hand, the case turned out to be that forecasts of demand are too low, then firms will lose opportunities; for example, there may be inadequate hotel accommodation etc. (Chu, 2004). In practice, time series forecasts are extrapolations in future times of the available time series values. A good projection should provide a forecaster with a sense of the reliability of the forecast. A convenient way to capture this sense is the prediction interval, which provides a measure of the reliability of the forecast (Psillakis, Panagopoulos, & Kanellopoulos, 2009). An exhaustive review on forecasting time series can be found in (Song, & Li, 2008).

In this paper, motivated mostly from absence of systematic research of the Region of Western Greece (Panagopoulos, & Panagopoulos, 2005), we used the well-known Box-Jenkins method (Box, & Jenkins, 1976; Box, Jenkins, & Reinsel, 1994) to model the hotel tourism occupancy (except camping sites) for the three prefectures of the Region of Western Greece, for the period 2005-2012. The choice of an ARIMA model consists mostly in its flexibility and generality as it can handle different types of data. Furthermore, despite of the hard programming implementation of Box-Jenkins method it can be easily found and applied in many computational and statistical packages (e.g. Minitab) and it can produce reliable predictions when the appropriate model is chosen.

The structure of the paper is as follows: In the next section we present a short profile of Western Greece and in section 3 we develop the forecasting models for each prefecture, as well as, we demonstrate the prediction results. Finally, in the last section we discuss some conclusions and remarks.

 

Short Profile of Western Greece

The Region of Western Greece occupies the northwest Peloponnese and the western tip of mainland Greece. It includes the prefectures of Etoloakarnania, Achaia and Ilia. For the most part the land is mountainous (45.3%) and half-mountainous (25.6%), while only 29.1% are lowlands. It has extensive coastline at all three prefectures, which are bounded by the Ionian Sea and the gulfs of Amvrakikos, Patras and Corinth. The main lines in the investment profile of Western Greece are described by the followings (Invest in Greece Agency, 2012; Research Institute for Tourism, 2014):

Main economic activities include agriculture and tourism services.

As in many other Regions of Greece, production of wine and olive oil is significant. Dairy products are also important to the local economy as well as fish farming, unique to the area and a traditional source of income.

Western Greece is quickly becoming one of the top tourism destinations in Greece. The emergence new hotel units and new investments in the area have strengthened the local economy and are currently changing the overall profile of economic activity.

The geomorphology of Western Greece has great diversity. It includes mountains with a very high altitude, large natural lakes and rivers.

The Region of Western Greece is privileged in terms of accommodating many, various and significantly sensitive ecosystems.

Significant tourism infrastructures (Western Greece is served by 2 airports and 6 ports)

In 2013 there were in Western Greece 292 Hotels with 11368 rooms and 21417 beds.

The tourism development of the Region’s prefectures is dynamic but spatially restricted, since specific areas (as enclaves) are developing a complex superstructure, where tourist accommodation and services, for organized package tourist plays a prominent role, but as important is the development of holiday home enclaves, and camping units. It is mainly a Region for domestic tourism. Greek tourists present a stable consumption (in terms of hotel nights spent) all over the year, particularly in the prefectures of Achaia and Etoloakarnania. In contrast, the consumption of foreign tourists presents a relatively high seasonality, especially in the prefecture of Ilia (May to September).

 

The proposed models

For each of the studied prefecture a forecasting is proposed based on ARIMA forecasting models. For the evaluation of the proposed models, we used the monthly occupancy of all tourist accommodations (except from camping sites) from the Prefectures of Etoloakarnania, Achaia and Hlia for January of 2005 till December 2012. The used data were obtained from the official records of the Hellenic Statistical Authority. It is underlined that Hellenic Statistical Authority has not released any similar data for the period 2013 until now. The development of proper models for each prefecture was made using the Minitab package.

For all three prefectures, the plotted data (Figure 1) reveal a strong seasonality with considerable variation in the rates of occupancy between summer and winter months. Maximum occupancy is observed in August in all three prefectures and for the three summer months a significantly increased occupancy is noticed, contrary to the winter months where the occupancy rates are very low. The greater volatility in occupancy occurs in Ilia which during the summer months the occupancy exceeds 70%, while during the winter months fall below 10%. On the other hand, the smallest fluctuation in occupancy between summer and winter months is observed in the prefecture of Etoloakarnania (19% - 57%).

It is of particular interest the underlying trend in the plotted datasets. The entire region of Western Greece (Figure 1a) discloses a clear decreasing trend over the last 3 years. In prefecture of Achaia (Figure 1c) is much more intense and occurs throughout the study period. On the other hand this downward trend is not as pronounced in Ilia (Figure 1b), and a fall is observed in Etoloakarnania (Figure 1d), especially during the summer months, where the maximum values of this time series occur.

Figure 1. Time series plot for three prefectures and the Region of Western Greece for period 2005-2012