Forecasting Tourism Indicators In Syria Using A Bootstrap ARIMA Model

Abstract

The study aims to forecast tourism indicators in Syria, represented by the number of tourists, tourist nights, and tourism revenues, using annual data for the period 2000–2023 obtained from publications of the Central Bureau of Statistics. The research adopts a quantitative analytical approach within a time-series modeling and forecasting framework. Phillips–Perron tests indicate that the series are integrated of order one, and accordingly an ARIMA(1,1,1) model is estimated for each indicator. Model adequacy is evaluated through diagnostic checks for autocorrelation, residual distribution, and heteroskedasticity, and the results are satisfactory. Forecasting performance is strengthened using residual bootstrap with 1,000 resamples to generate an empirical distribution of forecasts and construct 95% confidence intervals. Forecasts for 2024–2029 suggest a gradual recovery of the sector, with uncertainty explicitly quantified by the 95% confidence bands. Over the forecast horizon, the number of tourists ranges between 2,082,500 and 3,197,000, with an uncertainty width of about 735 thousand in 2024 increasing to 834 thousand in 2029. The results provide a quantitative tool to support decision-makers and investors in capacity planning, marketing strategies, and expected-return assessment.