THE THIRD-QUARTER economic growth report is due within days and obviously many people are waiting to see whether it confirms recovery from the Sars outbreak and a return to better days. These figures are likely to have an impact on confidence in Hong Kong. The trouble, however, is that GDP growth measurements are not very exact. Although statisticians like to claim accuracy for them, the reality is that they are more like a car with a four-speed manual transmission. Your rough measures are 0 to 3 per cent growth, 3 to 6 per cent, 6 to 9 per cent and probable nonsense. As with the car, there is only one gear for reverse and do not bother setting numbers on it. But I like to play with GDP numbers on spreadsheets to see what underlying indicators they may show and this exercise has produced a very good coincident indicator for economic growth - that of electricity consumption growth. This makes absolute sense. We use electricity for almost everything and the wealthier we grow the more electricity we burn. The first chart shows you the correlation between electricity consumption and economic growth in Asia excluding Japan and China. As you can see, the fit is an excellent one, up and down at the same time, with electricity growth showing a fairly consistent margin of about 2.5 per cent over GDP growth, which suggests either that Asian governments underestimate GDP growth (fat chance) or that Asia is becoming more profligate in electricity use (more likely). I excluded Japan from the sample because the size of the Japanese economy swamps all others in Asia and you would only really see Japan in the chart if I included it. The parallel between electricity and GDP in Japan is a very close one, however. I excluded the mainland because its figures are all over the shop. Electricity consumption growth was down to 2 per cent in 1998 and is now over 14 per cent. Yes indeed, I think this tells you the real story of mainland economic growth in recent years. Take GDP growth as being 2 per cent lower than electricity growth and you get a stagnant economy in 1998 with a booming one now, but that is not what the official figures say. I would like to show you that chart for the mainland, and the one for Japan, too, but I need the second of my permitted two charts a day for Hong Kong. If I sent in four of them I would have the boss telling me again that he will hire a photographer when he wants my column to consist of pictures alone. Good idea, actually, just charts with little inscriptions that say things like 'Wow!' and 'Look at this!' It certainly works in the investment trade. But back to Hong Kong, for which you can see the correlation between GDP and electricity in the second chart. Once again, the correlation has been reasonably close. That is, until 1998, when power consumption growth soared while GDP growth plummeted, with the reverse happening two years later. The anomaly probably has something to do with unseasonable weather, a cold summer in 1997 (remember the rain dropping off Prince Charles's cap) and a scorcher in 1998, although I cannot be certain. At a rough guess, about half of our electricity consumption in summer conditions air. But let us play with that final figure for electricity, 0.64 per cent year-over-year growth in September on a 12-month average basis and definitely swinging back up. If this is any indication, then third-quarter GDP growth will be in first gear of my four-gear model - 0 to 3 per cent. Actually, I think it will be more like an immediate jump to second gear, 3 per cent plus, but that guess comes from holding a damp finger to the wind. It is not a bad technique in GDP forecasting, however.