The basic approach taken here is to start by estimating a projected number of (international) tourism visits "without" the fires and haze. This projection is used to estimate the number of visits that could have been expected during September to mid-November 1997 without the fires.
The average expenditure per tourist is then calculated from data on total tourism expenditure and the number of visits. This is applied to September, October, and November to estimate the expected expenditure during that period.