So let us concentrate on the tourism business. The second assumption is that the decline in total spending by visitors was roughly proportional to the decline in their numbers.

I hear an objection. That decline in total spending was probably much greater, you say. Hotels and restaurants cut their prices in an effort to keep business and tourists were therefore spending less per day than they normally would.

Well, yes and no. Hotel occupancy collapsed in May to 17 per cent from more than 80 per cent earlier in the year, but the latest figures suggest that hotel room rates were on average down by only about 20 per cent from their normal May level, not as much by far as the 50 per cent collapse that followed the Asian financial crisis in 1997-98. In money matters, the biggest trouble is always money trouble.

Meanwhile, despite great screams of pain from restaurateurs, the latest inflation figures say that the cost of meals away from home in the second quarter dropped by only 1.6 per cent from the previous year, this while overall consumer prices fell by 2.5 per cent. I shall give you a little of that factor of lower prices for visitors but not much.

Then we turn to the consideration that the impact of tourism on our economy is a net impact. It is what visitors spend here less what Hong Kong residents spend on trips abroad. We shall assume that the decline in spending by Hong Kong visitors abroad in the second quarter was also roughly proportional to the decline in the numbers of trips they made. I do not hear any great objection to the assumption this time.

We now take the resulting figure, net tourist spending, as a percentage of gross domestic product and we get the red line in the chart. It says that our economy was a net loser on tourism following the 1997-98 troubles but then slowly recovered lost ground and last year was in a plus position again.

The blue line shows you my calculation of what net tourist spending was as a percentage of GDP in the second quarter, drawn from those assumptions that net spending was in line with visitor arrivals and resident departures. It comes out as a negative 1.83 per cent

But this is not enough. The chart shows a pronounced quarter-to-quarter change in the figures, with the first quarter usually down because so many Hong Kong people spend the Lunar New Year holiday in the mainland. The second quarter usually shows a bounce upward again.

What might we have expected net tourist spending to be if the Sars epidemic had not been visited on us? Take it as a five-year average of the second quarter over the previous year and the figure you then get is plus 0.5 per cent, represented by the green line.

So there you have a rough estimate of it. The difference between minus 1.83 per cent and plus 0.5 per cent is 2.23 per cent and that is as good an estimate as anyone could make of how much GDP growth was likely to be down in the second quarter because of Sars.

Take it over the full year and you get only 0.56 per cent down, which is not much and the real impact is likely to be even less. Most of what visitors consume here is imported goods and imports are deducted from exports in calculating GDP growth. If visitors spend less we also import less.

I cannot give you a figure for this but it will certainly bring that minus 0.56 per cent to a lower figure. Note here that there is no offsetting upwards adjustment for Hong Kong residents buying goods that we exported to the countries they visit on trips abroad. That figure would be minuscule.

And people still say, do they, despite the growing evidence of a full recovery from Sars on every front and the growing evidence that its impact was concentrated on tourism, that you ought to lop 3 per cent off our GDP growth rate this year because of Sars?

I don't think so.