September 2011

This section reports on our recent forecasting track record. The forecasts of construction activity that were published on this website in April 2011 are compared with outcomes for construction activity that have been published subsequently by the ABS. These outcomes reflect actual data published by the ABS covering the March quarter of 2011. Monitoring our track record helps identify strengths and weaknesses in the forecasting process that can be used to continuously improve that process.

Non-Residential Building

In April 2011, we forecast non-residential building to be down by 13 per cent in the March quarter of 2011 (compared to the March quarter of 2010). Actual data shows that non-residential building did in fact fall by 15 per cent over that time period. In level terms, our forecast was within 2 per cent of the actual value. As such, the actual contraction in non-residential building activity was more or less in line with our expectations.

The largest detractor was the educational building sector. Educational building fell by more than expected, down by 35 per cent compared to a forecast decline of 27 per cent. Predicted falls in other commercial building, and the miscellaneous category were also underestimated.

However, predicted falls for health and aged care building, and entertainment and recreation building, were overestimated. For example, health and aged care was forecast to decline by 28 per cent, but in fact only fell by 2 per cent. Likewise, entertainment and recreation building was forecast to decline by 16 per cent, compared with an actual decline of 3 per cent.

Notably, the direction of all non-residential building sectors was accurately forecast, except for accommodation building which recorded a decline in activity rather than the expected rise.

Residential Building

In April 2011, we forecast residential building to rise by 11 per cent in the March quarter of 2011 compared to the March quarter of 2009. Actual data shows that our forecast was relatively close with a 9 per cent rise in activity.

For most categories, our forecast was reasonably accurate. New other residential construction (apartments and townhouses) was forecast to expand by 22 per cent, which compares favorably with the actual growth rate of 26 per cent. Likewise, our forecast of 5 per cent growth for large alterations and additions was very close to the actual growth rate of 4 per cent. Our forecast for the largest category – small alterations and additions – was more or less exactly on the money, with a forecast and actual growth rate of 13 per cent.
The new houses category is the only exception; having been forecast to grow by 7 per cent, new house construction actually contracted by 1 per cent.

Overall, our forecasts for residential building activity were very close to actual data.

Engineering Construction

In April 2011, we forecast that total engineering construction for the March quarter of 2011 would grow by 3 per cent (compared with the March quarter of 2011). Actual data shows that growth in engineering construction over the year to March 2011 was much stronger than predicted, at 17 per cent.

With the exception of the telecommunications category (which was forecast to grow by 22 per cent but actually contract by 3 per cent), activity growth in all categories was underestimated. The largest discrepancies were in the heavy industry (mining) category, and the bridges, railways and harbours category, reflecting stronger than expected activity on resource related projects. Heavy industry construction was forecast to grow by 13 per cent, but actually grew by 24 per cent. The gap was even wider for bridges, railways and harbours construction, which was forecast to grow by 18 per cent but actually grew by 55 per cent.

While these variances are notable, the general direction and magnitude of the forecasts has been proven to be broadly correct.

Growth Forecasts for 2010-11