ECONOMETRIC MODEL (Accuracy)

Mr Chng Hee Kok asked the Minister for Trade and Industry what is the degree of accuracy of his Ministry's econometric model for Singapore in view of the sharp variance between the actual and forecast 1988 economic performance.

The Minister of State for Trade and Industry (Mr Mah Bow Tan) (for the Minister for Trade and Industry): Mr Speaker, Sir, in December 1987, the MTI forecast growth for 1988 at 5-6%. Actual growth turned out to be 10.9%, much better than expected.

On a sectoral basis, MTI predicted growth of the construction and the financial and business services sectors reasonably well, but grossly under-estimated growth in the other three sectors, namely, manufacturing, commerce, and transport and communications. Commerce grew three times faster than expected, while manufacturing and transport and communications grew twice as fast.

The economy in 1988 benefitted from unforeseen changes in external conditions. In December 1987, when we made our first forecast, our index of leading indicators had been declining for three months. After the stock market crash of 19 October 1987, the outlook for the world economy was subdued. But as it turned out, there was no post-crash global recession in 1988. Instead of slowing down, the US economy grew by nearly 4%, or twice the original forecast.

There were sectoral factors as well. Strong worldwide demand for disk drives buoyed the manufacturing sector. The high Yen and the upturn in regional economies brought us record numbers of tourists. Transport and communications expanded on the heels of the manufacturing and commerce sectors.

Sir, economic forecasting is always uncertain, however good the econometric model. Many events are unforeseeable and cannot be anticipated by any model. We can only update our models with fresh data on a regular basis and revise our forecasts during the year to improve their accuracy.

MTI also has to bear in mind that an over-optimistic forecast may do more damage, in terms of mistaken Government policies, ill-founded business expectations and excessive wage settlements than a slightly conservative one. MTI therefore prefers, if need be, to err on the side of caution. When growth exceeds our expectations, as it did in 1988, we are glad to be proved wrong.

Mr Chng Hee Kok: Sir, the general belief is that to the Government's prediction we should add 1% more to the growth to get the actual. In view of what has happened, how many percentage points will the Minister advise that we should take into account now?

Mr Mah Bow Tan: Sir, MTI forecast growth for 1989 at 6-7% and the Member realizes that this was only released some three weeks ago. Barring unforeseen circumstances and in the light of the data that we have, I must advise that the estimate is the best that is currently available. But of course, as I said, if we are proven wrong and growth exceeds our forecast we will be glad.

Dr Wan Soon Bee (Pasir Panjang GRC): Sir, the Minister has said that the forecast is not accurate and I do agree that a variance of 10-15% is acceptable. But in this case, I think the variance has been very big, from 6-7% to about 10.9%. This will make the forecast quite valueless and meaningless.

Mr Mah Bow Tan: Mr Speaker, Sir, as I explained in my reply just now, the forecast was made at the end of 1987 and, as it turned out, actual growth, as the Member rightly pointed out, was much more. But there were extenuating circumstances as to why the forecast in the end turned out to be wrong. As I have indicated earlier, MTI will revise its forecast on a quarterly basis.