When will the trend bend? - Imperial College London economist examines the value of forecasting

For immediate use
19 May 2003

The benefits and pitfalls of forecasting will be examined in an Inaugural lecture given by Nigel Meade, Professor of Quantative Finance at the Business School, Imperial College London. 'When will the trend bend? - the value of forecasting' will be delivered on Tuesday 20 May at Imperial's South Kensington campus.

"Forecasting is not an activity at which human intuition excels" says Professor Meade, who highlights his lecture with instances from the world of technology to show how wide of the mark forecasts can be. He quotes the Building News, which on 29 June 1894 expressed great pessimism concerning the future of Tower Bridge in London:

"The Tower Bridge is, on the whole, a fine piece of work... foredoomed to become obsolete. Repeated opening and lowering will be found tedious. By and by the bridge will become a fixed bridge as it should have been from the first and a million in money would have been saved."

Professor Meade explains that human forecasting is so unreliable because we tend to be influenced by what we wish to happen, and by under-estimation of uncertainty within the model. In many situations our predictions are also overly influenced by one-off occasions that show extreme results.

One major reason for incorrect forecasts is that they do not anticipate the growth in activity that new technologies facilitate. So IBM Chairman TJ Watson's 1943 prediction: "I think there is a world market for about five computers" was wrong because he was thinking of current activities that would benefit from computers rather than what computers would enable organisations like banks and airlines to do.

Similarly, when K Olsen, President of Digital Equipment Corporation, said in 1977: "There is no reason for any individual to have a computer in their home", he was thinking of computers purely as specialised equipment for professionals and not taking into account their potential for other uses.

Accuracy in forecasting can be vital to the successful running of businesses. To illustrate this, Professor Meade examines three examples of the type of decision a telecoms company may face. On a simple level it will want to forecast demand for a spare-part in its transmission system. It will also need to run a call centre, requiring a forecast of call frequency to construct staffing schedules. On the strategic side of the business is a forecast for the demand for a new generation of mobile phones.

In forecasting the demand for spare parts, the data available is the past history of demand for the parts. This problem requires an automatic forecasting system that will need a minimum of manual intervention. The penalties for an under-estimate will be transmission systems offering a diminished service. If demand is over-estimated capital will be tied up in unused inventory.

When looking at the efficiency of the call centre, Professor Meade estimates that around 65 per cent of costs are due to staffing, making accurate forecasting of call frequency extremely important in controlling costs. An over-estimate would result in idle operators and wasted wages while an under-estimate would mean delayed response, lost calls and lost business. A forecast for this problem should consist of the expected number of calls along with some measure of its uncertainty.

The forecast for the demand of new mobile phone technology will be the basis of the decision on whether to offer the service at all, and if so how much capacity should be installed and how much should be paid for an operating licence. Professor Meade estimates that this forecast will be labour intensive the result of a careful estimation process involving comparisons with the way markets have behaved in other countries and for similar technologies.

Professor Meade looks at different forecasting models to arrive at the most effective solution for each of these problems. In doing this, he identifies two important factors that underpin the success of any forecast close examination of all the data available and quantifying the uncertainty in the forecast. He concludes: Recognising that the trend will bend is more important than knowing when the trend will bend.

Professor Meade's lecture will take place on Tuesday 20 May at 17.30. In the chair will be Professor David Begg, Principal of the Business School.

For further information contact:

Abigail Smith
Imperial College London Press Office
Tel: 020 7594 6701
Email: abigail.smith@imperial.ac.uk

Notes to editors:

1. With a Teaching Quality Assessment of 'excellent' and a 5 rating in the 2001 Research Assessment Exercise, the main objective of the Business School is to contribute to advanced teaching and research at the interfaces of management, technology and innovation. The School currently offers six postgraduate programmes of education, a full-time and part-time (executive) MBA, a Singapore MBA, an MSc in Finance, an MSc in Health Management and a new Distance Learning MBA. In addition, the School offers a Joint Honours programme and various business and management modules to undergraduate students at Imperial. Website: www.imperial.ac.uk/business

2. Consistently rated in the top three UK university institutions, Imperial College London is a world leading science-based university whose reputation for excellence in teaching and research attracts students (10,000) and staff (5,000) of the highest international quality.
Innovative research at the College explores the interface between science, medicine, engineering and management and delivers practical solutions, which enhance the quality of life and the environment - underpinned by a dynamic enterprise culture.
www.imperial.ac.uk