Forecasting annual water demands dominated by seasonal variations: the case of water demands in Mecca

Abstract

Water usage in Mecca is dominated by trends and cyclical variations in the number of foreign pilgrims visiting for the annual Hajj Pilgrimage. Time series regressions show that the mean temperature in Mecca at the time of the Hajj affects the number of pilgrims, inducing a long-term cyclical pattern for this variable and therefore water usage. The cointegrating relation between water usage, number of external pilgrims and temperature produces long-run forecasts of Mecca water demand.

Keywords:

• cyclical demand
• cointegration
• long-run forecasts
• water usage
• Hajj pilgrimage

JEL Classification:

• C22
• Q25

Notes

¹ The correspondence between Islamic and Gregorian years is only approximate.

² Numbers in parentheses are SEs of the estimated coefficients. R² is the square of the multiple correlation coefficient. BG-LM(2) is the Breusch–Godfrey Lagrange multiplier statistic for testing second order autocorrelation; under the null hypothesis of no autocorrelation, it is distributed as chi-squared with two degrees of freedom and the p-value of the test statistic is reported by p₂. White is the chi-squared test statistic for heteroscedasticity that is related to the explanatory variables in the regression plus their squares and cross-products; its associated p-value is given by p₄.

³ For example, long-run forecasts of GDP growth are often based on assumed rates of growth in primary inputs and technology, as in Foure et al. (2012) and Wilson and Purushothaman (2003).